Professional Documents
Culture Documents
John M. Kolyer, Donald E. Watson (Auth.) - ESD From A To Z - Electrostatic Discharge Control For Electronics-Springer US (1996)
John M. Kolyer, Donald E. Watson (Auth.) - ESD From A To Z - Electrostatic Discharge Control For Electronics-Springer US (1996)
FROMATOZ
ELECTROSTATIC DISCHARGE
CONTROL FOR ELECTRONICS
SECOND EDITION
ESD
FROMATOZ
ELECTROSTATIC DISCHARGE
CONTROL FOR ELECTRONICS
SECOND EDmON
John M. Kolyer
and
Donald E. Watson
....
~,
KLUWER ACADEMIC PUBLISHERS
BOSTONIDORDRECHTILONDON
Distributors for North, Central and South America:
Kluwer Academic Publishers
101 Philip Drive
Assinippi Park
Norwell, Massachusetts 02061 USA
Telephone (781) 871-6600
Fax (781) 871-6528
E-Mail <kluwer@wkap.com>
Appendix 167
Paper No.1. Selection of Packaging Materials for Electrostatic
Discharge-Sensitive (ESDS) Items 169
Paper No.2. Permanence of the Antistatic Property of Commercial
Antistatic Bags and Tote Boxes 175
v
vi CONTENTS
Paper No.3. Hazards of Static Charges and Fields at the Workstation 181
Paper No.4. Perforated Foil Bags: Partial Transparency and
Excellent ESD Protection 201
Paper No.5. Cost-Effective Methods ofTestingIMonitoring
Wrist Straps 207
Paper No.6. Methodology for Evaluation of Static-Limiting
Floor Finishes 218
Paper No.7. Tote Box Material: How Good Is It? 226
Paper No.8. Electrostatic Discharge (ESD) Control in an
Automated Process 233
Paper No.9. Corrosion and Contamination by Antistatic Additives
in Plastic Films 238
Paper No. 10. Controlling Voltage on Personnel 244
Paper No. 11. ESD-Control Myths, Old and New 251
Paper No. 12. ESD Testing of Silicon Wafers 258
Paper No. 13. CDM and Work Surface Selection 263
Paper No. 14. Is Your Work Surface CDM Safe? 269
Paper No. 15. Realistic Testing of ESD Materials 272
Paper No. 16. Testing Surfaces for ESD Safety 281
Paper No. 17. Humidity and Temperature Effects on Surface Resistivity 286
Paper No. 18. Packaging for High-Voltage Discharge Protection 294
Paper No. 19. Hidden Charges on ESD-Protective Packaging 299
Paper No. 20. Toward an Ideal ESD-Protective Package 305
Index 321
Preface to the Revised Edition
In the past five years, the field of electrostatic discharge (ESD) control has under-
gone some notable changes. Industry standards have multiplied, though not all of
these, in our view, are realistic and meaningful. Increasing importance has been
ascribed to the Charged Device Model (CDM) versus the Human Body Model
(HBM) as a cause of device damage and, presumably, premature (latent) failure.
Packaging materials have significantly evolved. Air ionization techniques have
improved, and usage has grown. Finally, and importantly, the government has ceased
imposing MIL-STD-1686 on all new contracts, leaving companies on their own to
formulate an ESD-control policy and write implementing documents. All these
changes are dealt with in five new chapters and ten new reprinted papers added to
this revised edition of ESD from A to Z.
Also, the original chapters have been augmented with new material such as more
troubleshooting examples in Chapter 8 and a 20-question multiple-choice test for
certifying operators in Chapter 9.
More than ever, the book seeks to provide advice, guidance, and practical ex-
amples, not just a jumble of facts and generalizations. For instance, the added
tailored versions of the model specifications for ESD-safe handling and packaging
are actually in use at medium-sized corporations and could serve as patterns for
many readers.
The new material in this edition is intended to aid your decision-making process
as directly as possible, given that you must make your own judgments under such
constraints as cost, personnel skill level, and available facilities. As in the first
edition, the essence of this book is the integration of conceptual "tools" (Chapter
4) into an overall approach to ESD control as codified in model specifications
(Chapter 9) and, now, in policy and program documents (Chapter 11).
The frontispiece, new for this revised edition, represents the essence of ESD
damage: a generalized electrical field or air discharge strikes an edge contact of
a board, runs along a circuit line (in white), and figuratively explodes the heart
of a device. To prevent this microdisaster is the object of all our procedures and
paperwork.
We hope our book will prove helpful, and we wish you success in the ever-
evolving, and always challenging, field of ESD control.
vII
Introduction
This book does not pretend to be an encyclopedic test, which would have to be a
compilation of contributions from experts in various facets of ESO control. It is
closer to being a handbook, but we feel that the most accurate descriptive term is
an approach: a coherent, logical, and cost-effective system. As the building blocks
of this system, about fifty real and conceptual tools from A to Z appear in bold
type throughout the text of the book in order to emphasize their usefulness and to
provide an organizing principle. Insofar as "A to Z" hints of thoroughness, the
thoroughness is not in the length of the A-to-Z list but in the applicability of our
approach to all ESO-control situations.
The first edition of ESD from A to Z differed from the self-published edition
of 1989 in that various improvements were made in the test; some of the papers in
the Appendix were condensed to remove extraneous data, and an extensive index
was added to make the book more useful. This revised edition adds much new
material.
The book is arranged as follows. First, Chapter 1 tells how to use the book, and
Chapters 2 and 3 give basic principles and fundamentals. Then the real and con-
ceptual tools are listed and explained in Chapter 4, and three of them are of such
importance that they are given their own chapters: Chapter 5 for the charged de-
vice model (CDM), Chapter 6 for the static-safe package (SSP), and Chapter 7 for
the static-safe workstation (SSW). Since the best way to teach is by examples, the
tools are put to use in Chapter 8 on ESO troubleshooting; this chapter is a collec-
tion of actual case histories illustrating the flexibility of our approach. This brings
us to the bottom line: codification of our ESO control methods into specifications
for packaging and for in-plant handling (Chapter 9) and into a program of maxi-
mum simplicity and minimum expense (Chapters 10, 11, and 12). The important
topic of disposition of mishandled hardware is discussed in Chapter 13. Finally,
Chapter 14 wraps up the book with checklists of what to buy and do and a conclu-
sion summarizing the elements of our approach, and Chapter 15 gives specula-
tions of the future of ESO control.
References are listed at the end of the main text, before the Appendix. As an ex-
ample of the numbering system, Ref. 4-3 is the third reference cited in Chapter 4.
The Appendix contains twenty of our published papers or articles which give
experimental justification for our rules and techniques as well as details of test
methods and discussions of special topics such as tote boxes and the permanence
Ix
x INTRODUCTION
First, take an hour to skim through the Appendix and read the abstracts and con-
clusions carefully. Note the locations of detailed infonnation, e.g., description of
test methods, for future reference.
Next, study Chapters 2, 3, and 4. Chapter 4 gives the Basic Rule and the key
concepts, including damage mechanisms, from which all else follows. When you
come to the Charged Device Model (CDM), static-safe package (SSP), and static-
safe workstation (SSW), stop and read Chapters 5, 6, and 7. Remember that every-
thing we're doing is aimed at detecting, avoiding, or preventing the two hazards of
fields and discharges.
Study Chapter 8 to see how unique problem situations can always be handled by
using the gamut of ideas and methods from Chapter 4. As problems arise, consult
this chapter again to refresh your memory.
Use the model specifications in Chapter 9 as the basis for your own, adding
more detail as necessary.
Use Chapters 10, II, and 12 as a guide when you're setting up a new program or
improving an existing program.
Note the approaches in Chapter 13 for the disposition of mishandled hardware;
you'll face this problem sooner or later.
Use the check lists in Chapter 14 when you're buying equipment, solving prob-
lems, initiating a program, etc.
Read Chapter 15 on the future for general infonnation.
In summary, learn the location of infonnation in the book so that it can be an
ever-useful guide for troubleshooting, material/equipment buying, specification-
writing, and program management.
When the book has been thoroughly understood, the reason for each provision
of the model specifications in Chapter 9 should be evident. The reader should see
how every rule, though based on limited data and more or less empirical, is in-
tended to implement the Basic Rule. Remember that control measures may be
primary (essential) or secondary (redundant or backup); for example, in Model
Specification I, requirement paragraph 3.6.11 for the CD Rule is a primary pre-
caution, but 3.6.12.2 for unnecessary touching ofleads and 3.6.12.3 for unnecessary
1
2 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Basic Physics
3
4 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
especially prominent in highly conductive materials, e.g., metals, over which and
through which electrons roam freely. In contrast, electrons cannot travel through
or over nonconductors such as common plastics (unless voltages are very high, as
will be discussed later), so "pools" of electrons--or, conversely, areas of deple-
tion-lie stagnant on nonconductive surfaces after triboelectrification (from Greek
tribein, meaning "to rub").
At relatively low voltage, under approximately 2000 V, or 2 kV, these stagnant
electron pools, or "dried-out" areas yearning to be moistened, cannot flow, and
they make their presence known only by an electrical or electrostatic field (E field),
which is either positive or negative. This field is "a voltage gradient between two
surfaces" (Ref. 2-6) and is a force field that pushes or pulls electrical charges. If
one surface is a charged plastic tote box, for example, and the other surface com-
prises the floor, ceiling, and walls of the room, the field will be intense only near
the box. Then, a particle in that field will be polarized, with the electrons gathered
on the side of the particle toward the box if the field is positive or away from the
box if the field is negative. Polarization results from electrons forcefully repelling
one another across space, or else from an electron deficit just as forcefully tugging
at distant electrons.
Indeed, individual electrons are analogous to misanthropes. They can be forced
into a crowd but are anxious to disperse. Thus, if a charge (quantity of electrons) is
deposited on a conductor, the electrons almost instantly scatter over the surface to
give a uniform distribution, with each particle as far as possible from its fellows.
These antisocial particles not only wish to avoid each other but yearn to hide away
in "holes" (mobile electron vacancies) in semiconductors or in the safe harbor of
the outer shell of any electron-hungry atom.
Electrons
Fig. 2-1.
Fig. 2-2.
6 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Fig. 2-3.
Fig. 2-4.
What happens? The water level on one side rises to break the dam (Fig. 2-3); the oxide
is punctured and the MOSFET has failed.
So much for the HHM. Now, let's consider the FIM, also described in Chapter
4. As shown in Fig. 2-4, stagnant pools of electrons on an approaching noncon-
ductor, e.g., a plastic circuit board, are repelling the electron pools in the reservoirs.
BASIC PHYSICS 7
The nearer reservoir pool surges higher, because the E field is more intense nearer
the pools on the charged plastic, and again the dam breaks as in Fig. 2-3-another
MOSFET failure. This creation of a voltage differential by an E field is called
"induction. "
The last but not least damage mechanism to be illustrated by the water analogy
is the CDM, described in Chapters 4 and 5. As seen in Fig. 2-5, stagnant pools of
COM
Field
Fig. 2-5.
COM
Field
,
/
.' /;
/
Fig. 2-6.
8 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
electrons lie near the rim of one reservoir, representing charges on a nonconduc-
tive part of a device, e.g., the plastic case of a dual-in-line package (DIP). The E
field, being motionless, causes no surge, and the slight difference in water heights
seen in Fig. 2-5 cannot crack the dam. However, when a grounded finger touches
a DIP lead and draws off electrons, which are always anxious to rush to ground
(Fig. 2-6), a large water height differential results, and again the dam crumbles as in
Fig. 2-3.
Note that the analogy works backward for opposite charges. A positive finger in
Fig. 2-3 sucks water out, the water on the right instead of the left side surges up in
Fig. 2-4, and the grounded finger dumps in electrons in Fig. 2-6.
The morals of this water analogy are: (I) a charged finger (HBM) or tool (MM,
see Chapter 4) should not be allowed to transfer electrons to or from ESD-sensi-
tive devices, (2) an intense E field should not be brought near an ESD-sensitive
device, and (3) leads of charged ESD-sensitive devices should not be grounded by
touching them with a finger or other conductor.
The water analogy is useful for training operators. Just remember to stress that
electricity is a sort of magical blue fluid that, unlike water, forcefully repels itself
while, conversely, desiccated regions attract the fluid. Of course, this analogy, like
any other, cannot be stretched too far and, in this case, "too far" means high voltage.
High Voltage
Above 3 kY, as an arbitrary dividing line, electricity is no longer "really rather simple."
Charges on nonconductors are not merely pools radiating a force field but are able
to discharge through the air by ionizing it. A large-scale example of a high-voltage
discharge is lightning, and a small-scale example is the spark thrown from a finger
to a doorknob. Furthermore, electrons at high voltages can move over nonconductive
surfaces to some extent and even punch through layers of material when the dielec-
tric strength is exceeded. The world of high voltage is a strange one.
To appreciate the nature of high static voltages, e.g., 40 kY, triboelectrically
charge a slab of rigid polystyrene packaging foam by stroking it on fabric or car-
pet. Bring the foam near the back of your hand, and you'll feel the hairs lift be-
cause a charge opposite to the charge on the foam is induced on the hairs, and then
opposite charges attract each other. If the foam is negatively charged, electrons are
driven off the hairs to leave them positively charged, and vice versa. Induction, as de-
monstrated here with hairs on the hand, is the CUlprit in FIM damage as we have seen.
Now we come to a more dramatic effect, the air discharge. When an NE-2 neon
bulb with one lead grounded is brought within 6 inches of the highly charged
foam, you will hear a "snap," and the light will flash; an invisible "lightning bolt"
has traversed half a foot of space. If the bulb, which flashes at 80 Y, had been a
MOSFET, it would have been destroyed.
Note that induction precedes such a discharge. Electrons are driven off into
ground when a negative charge approaches a grounded conductor-or electrons
BASIC PHYSICS 9
are drawn from ground when the charge is positive-and when the gap becomes
small enough, ionized air carries current as a spark.
You'll find that you can bring a grounded metallic plate with surface parallel to
the foam surface to 0.5 inch from the foam with no discharge, whereas a crackling
discharge will occur as far as 6 inches from a corner of the plate. The sharp corner
is acting as a lightning rod by concentrating the lines of force of the E field and
ionizing the air (corona discharge). Saint Elmo's fire is the corona glow on tips of
a sailing ship's rigging.
With a static meter, you can see the charge induced on an electrically isolated
plate, e.g., a 1 ft2 sheet of lin-inch aluminum mounted on a base of polystyrene
foam, when charged foam is brought near the plate. If negatively charged foam is
held in front of the plate, a meter behind the plate will show a negative charge
because electrons were driven to the back of the plate, and this charge will subside
when the foam is withdrawn. However, if the foam discharges to the isolated plate
by being brought near a corner, the plate will hold a negative charge that fades
slowly as positive air ions are attracted to neutralize the charge. This charged plate
is a capacitor, with the dielectric being air and the other plate, or plates, being a
tabletop, the walls of the room, and so forth. Such a charged conductor could
discharge to an ESD-sensitive item to cause damage by the MM mechanism (Chap-
ter 4). If a charged person were the capacitor, the mechanism would be the HBM
which, like the MM, is a special case of DI (Chapter 4).
By the way, all the triboelectricity (transferred electrons) on a person charged by
walking resides on his shoe soles. If the soles "steal electrons" from a carpet, the
negative E field on the surface of the shoe soles drives electrons away from the
skin of the sole of the foot, so that the sole of the foot becomes positive while the
rest of the skin becomes negative (Fig. 2-7). This is, of course, another example of
+ + + + + +
""
. ""
""""
"""""""""""""
""""
"" ' ",
'"
Carpet
Fig. 2-7.
10 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
induction. If a spark is thrown to ground by the person's finger, the minus signs
above the foot in Fig. 2-7 vanish (electrons have run off to ground), leaving the
positive charges still trapped on the sole of the foot. If the person stepped out of his
shoes and onto an insulating surface, his whole skin would be positive. This pro-
cess is called compound induction, and isolated conductors charged by compound
conduction are an ESD hazard (see Paper 19 in the Appendix). Meanwhile, the
negative charge on the nonconductive shoe sole persists because it is, you will
remember, static (immobile).
Now, ground your formerly isolated metal plate, and a static meter held behind
it will show no charge at all when a charged piece of foam approaches the front of
the plate. You'll find that a grounded metallic screen will work, too. The grounded
plate or screen provides shielding, which is one of the tools of ESD protection
discussed in Chapter 4.
From the above experiments, you can imagine how much the ESD hazard is
aggravated at low relative humidity, e.g., 10-20% , when voltages run high and
synthetic fabrics and other materials crackle with air discharges. Unprotected ESD-
sensitive items are at great risk, like bystanders caught in the crossfire of a riot.
Low-voltage H8M, FIM, and CDM are enough to worry about without DI from
all directions! However, remember that humidification will only lessen the ESD
problem, not eliminate it, as pointed out in Chapter 4.
Now, let's turn to the sources of static charges, of which common triboelec-
trification, e.g., an operator's clothing accidentally stroking a plastic tote tray is
only one. All possible static generators, some subtle and unexpected, must be iden-
tified in your workplace.
Static Generators
Contact Charging and Triboelectric Charging. As mentioned above, rub-
bing may merely increase contact area, so these two charging phenomena are hard
to disentangle. For a brief but in-depth discussion of contact charging, including
the concepts of Fermi level and valence bands versus conduction bands of elec-
trons, see Ref. 2-8. Among many practical examples of triboelectric charging is
the peeling of insulative tape from circuit boards, which can generate several kV.
Also, a nonconductive liquid can charge a solid, e.g., high-purity water impinging
on a silicon wafer (one remedy in this case is to add carbon dioxide to the water to
make it conductive).
Induction Charging. We have seen this phenomenon in the FIM (water anal-
ogy, above) and shoe-sole charging (Fig. 7).
Freezing. Large voltages have been seen between ice and water phases during
freezing. A theory is that "splinters" ejected from a freezing drop are charged op-
positely to ice particles remaining in the drop. This process electrifies clouds in
thunderstorms (Ref. 2-8).
BASIC PHYSICS 11
Radioactive Decay. This is the source of alpha particles used in nuclear air
ionization, as mentioned above.
Thermionic Emission. Hot filaments can emit positive ions (Ref. 2-8 and
Example 19 in Chapter 8). Beware of this effect with heat guns that blow hot air.
Field Emission. An E field polarizes particles and, if the field is intense enough,
electrons may be lost from the negative end of the particle while positive ions may
be extracted from the positive end. If more electrons than positive ions are lost, the
net charge on the particle is positive, and it will be drawn in the direction of the
field (Ref. 2-8). The "direction of the field" is by definition the direction in which
a positive charge is forced to move.
Corona. This is the basis of electrical ionizers in which air molecules become
charged (ionized) in the vicinity of sharp, high-voltage emitter points.
Hot, Blowing Air. Our tests showed no charging above 100 V in Example 20,
Chapter 8. In careful experiments, a blast of shop air gave very low voltages on
metal surfaces, presumably because of triboelectric charging by submicroscopic
particles, while filtered air caused no detectable charging (Ref. 2-10). Kinetic en-
ergy calculations indicate that pure gases cannot charge surfaces.
ESD Is a Surge
Electrostatic discharge has been called a "spark," sometimes with the admission
that the only "spark" is a high-temperature breakdown of gate oxide, metallization
melting, or other internal device failure. We are concerned, however, with the ex-
ternal ESD events that lead to internal device failure. If "spark" is defined as an air
discharge, the FIM obviously is sparkless, as is the HBM at low voltage, e.g., 1
kV, while the CDM mayor may not involve a spark (see Chapter 5).
To prove that the HBM can be a sparkless surge, don't ground your wrist strap
but attach it to the lid of a Static Event Detector™ (3M Co.), sensitive to 80-100 V,
with the body of the detector grounded. Sit in a chair, shuffle your feet on a carpet,
and quickly lift your feet. You have charged yourself (by induction from the tri-
boelectric charge on your shoe soles) and, as your feet rise, your capacitance
falls, your voltage surges, and the detector is tripped. A MOSFET, represented by
the detector, would have failed. Obviously, there was no gap for a spark to jump
between you and the detector lid. A surge was to blame.
Clearly, ESD, in the sense we use it, is a voltage surge that mayor may not
include a spark (air discharge). Indeed, in the failure of voltage-sensitive devices
like MOSFETs, the minimum required current (electron flow) to cause damage is
extremely feeble and far below sparking level.
Conclusion
To summarize, static electricity is a "self-repelling fluid" that creates a force field
in space. This field induces charges on nearby conductors by polarizing them but
is blocked (shielded) by a grounded conductive sheet or screen. The static electrical
BASIC PHYSICS 13
This chapter is reprinted, by pennission, from the Technical Record of the Expo
'92 International Conference on Electromagnetic Compatibility, sponsored by EMC
Technology Magazine, Reston, VA, May 18-22, 1992, pages 154-161.
Introduction
First, let's define "ESD." Electrostatic discharge (ESD) means a discharge (flow
of electrons) to or from a charge (deficit or surplus of electrons) that fonnerly had
been static (immobile).
Electrons are immobile under two conditions: (1) They reside on a nonconductive
surface, e.g., common plastic, over which they can't flow, or (2) they are trapped
on a conductive but electrically isolated object, e.g., a screwdriver blade with an
insulating plastic handle.
Now, if that screwdriver blade, bearing a sufficient charge, is brought near the
lead of an integrated circuit (IC), a spark jumps (if the voltage is high enough) and
we have an ESD. Or if an ungrounded electrically isolated person becomes stati-
cally charged by walking on a carpet and then throws a spark to a doorknob, we
have another ESD.
Even if the ESD is imperceptible and sparkless, it may still damage a highly
sensitive Ie. In any case, an ESD is fonnerly static (resting) electricity in motion,
and this motion, like the swing of a wrecking ball, can do damage and must be
controlled.
Threats
How do we control ESD? First, let's review the threatening fonns it may take.
These fonns are called ESD damage models, and Fig. 3-1 is a diagrammatic sum-
mary of the major ones. For a full discussion, see books such as Ref. 3-1.
For simplicity, Fig. 3-1 uses a silicon oxide layer, as on metal oxide semicon-
ductor field-effect transistor (MOSFET) gates, to represent all ESD-sensitive device
14
FUNDAMENTALS OF ESD CONTROL 15
elements, but these also include PN junctions, thin resistive films, and piezoelectric
crystals (Ref. 3-2). The destructive "zap" involves almost no current for voltage-
sensitive devices but more considerable current for current-sensitive devices.
With reference to Fig. 3-1 for definitions of abbreviations, the person-to-door-
knob ESO previously mentioned is an example of the HBM and also, if the knob
is not grounded, the FM. In the FM, the capacitance, or electron-sink ability, of
the knob (or of circuitry on a board) is a substitute for ground and accepts or
provides electron.
DI
PLATE
CONDUCTIVE
- FFB
SURFACE
~
CDM J}- ~
~
OR mIF
SYMBOLS ABBREVIATIONS
ril [?
AND
CONTACTS TO
OXIDE LAYER
~ RAPID
DISCHARGE
MM • MACHINE MODEL
FM s FLOATING MODEL
FIM ~ FIELD INDUCED MODEL
-L PUNCTURED COM • CHARGED DEVICE MODEL
- GROUND ~ OXIDE FFB = FIELD FROM BOARD MODEL
.
~ (OBVIOUS E = ELECTRICAL (E) FIELD
...L. FAILURE) LF • LATENT FAILURE
CAPACITANCE DI DIRECT INJECTION
~
Defenses
Basic defenses are diagrammed in Fig. 3-2. Design protection comes first, and
grounding the input through a diode shows the principle. In this oversimplified
representation, the diode is gated and opens to drain off excessive inputs. Typi-
cally, the damage threshold is raised from 100 V to 2000 V, so the item is still
ESD-sensitive and, in many cases, e.g., dies being assembled, input protection is
inapplicable. Therefore, design protection is no panacea, and the need for ESD
control by other means remains.
The first of these other means is grounding of conductors. All metallic or car-
bon-loaded conductors, from large structures such as ovens to the smallest tools,
must be grounded in a static-safe workstation (SSW) to prevent the MM (discharges
FUNDAMENTALS OF ESD CONTROL 17
~~ ±
~DESIGN
~~N~··
PROTECTION
vs SURFACE
I •
,-,'
LP ~
SLOW DRAIN
;::;;~
/'AI1 SPECIAL
//11\\",,-
-
V-I \ \ ~'lzED
-\ + AIR
HUMID FlOOR +
AIR
+++
U FINISH
~
"/
NEUTRALIZATION
FLOOR OF CHARGES
LlMITINGTC
SYMBOLS ABBREVIATIONS
~ CURRENT-LIMITING RESISTOR CM .. CONTINUOUS MONITOR
(FOR WRIST STRAP)
t ZENER DIODE
VS • VOLTAGE SUPPRESSION
LP • LEAKAGE PATH
TC .. TRIBOELECTRIC CHARGING
SLOW DISCHARGE E • aECTRICAL (E) FIELD
I!N .. DELTA VOLTAGE
OTHER SYMBOLS: SEEAGURE1
from charged conductors) and FIM (fields from charged conductors). To prevent
the H8M, operators' skin also must be grounded. Skin-grounding preferably
includes a continuous monitor because ESD damage happens in a tiny fraction of
a second and no occasional checks, however frequent, can put failed wrist straps
out of service fast enough (Ref. 3-4). The resistors in Fig. 3-2 limit the current to
ground to a safe maximum of 1.0 rnA in case the operator touches a hot lead-and
18 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
500 V-sensitive device might not "zap" it if static-limiting floor finish has mod-
erated his triboelectric charge from 1000 V to 300 V.
We have seen that a special finish reduces TC on floors, and the general defense
is to treat nonconductive surfaces of all kinds, e.g., upholstery, walls, or plastic
equipment housings, with a topical antistat that attracts atmospheric moisture to
form a "sweat layer" that drains away incipient charges.
An even more general defense that reaches all surfaces and "nips TC in the bud"
is humidification. However, high relative humidity has drawbacks. It reduces TC
but does not eliminate it, so no operator disciplines should be relaxed. However,
these disciplines may be violated, contrary to the rules, if humidification lulls
personnel into a false sense of security. Furthermore, humidification is expensive
and may cause personal discomfort (because of mugginess), corrosion, delamina-
tion of circuit boards, and solderability problems.
Another general defense is air ionization. As shown in Fig. 3-2, the principle is
to bathe charged surfaces in a stream of both positive and negative air ions. In the
diagram, the negative charge (minus symbols) on the device, which could cause a
CDM discharge, will soon attract positive air ions and be neutralized.
A basic limitation of ionization is that it does not prevent TC and can only
control charges that already exist and emanate an E field to attract the neutralizing
air ions. Thus, the method is not a panacea, but it does playa significant role in
ESD control as will be discussed later in this book.
Last but not least among defenses is operator skill, achieved by training and
evidenced by certification. This brings us to the subject of documenting and orga-
nizing your program.
Documentation/Organization
We've touched on the major defenses with simple illustrations but, in fact, ESD
countermeasures can become quite complicated. This is particularly true for spe-
cial processing such as cleanroom operations or automated assembly. Then, how
should all the various ESD-control procedures by documented? And how does a
manufacturer gear up organizationally?
Fig. 3-3 shows some key elements. First come the military documents if you
have a government contract. Or, even if you don't, the combination of MIL-STD-
1686 and MIL-HDBK-263 makes a good "bible." (Note: But see Chapter 10.)
Next is your own ESD-control program plan, which implements the govern-
ment documents. It carries out their intent by translating general "commandments"
to "situation ethics" fitting your circumstances, e.g., budget available and facili-
ties already in place. These circumstances also include your established business
system, whose policies will necessarily impact the plan.
From the plan flow nuts-and-bolts documents such as drawings and specifica-
tions, which are continually revised and improved by inputs from experimenta-
tion, failure analysis and, in accordance with Total Quality Management (TQM)
20 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
REQUIREMENTS SUGGESTIONS
(MIL-STD-1686. (MIl-HDBK-263.
•
DIDa. ETC.) MIl-HDBK-n3. ETC.)
ESD-CONTROl
PROGRAM PLAN
• BUSINESS
SYSTEM
APPROVED
!
MATERiAlS AND DRAWINGS EXPERIMENTATION
EQUIPMENT r-- SPECIFICATIONS
(MIL-STD-1000. ETC.) (ENGINEERING)
7
(LIVING LIST)
FAILURE
ANAlYSIS
FAClllTIZATION
(BENCHES.
OPERATOR
WORK
~ OPERATOR
EXPERIENCE
GROUNDING. ETC.) INSTRUCTIONS (TOM)
..
-
L
I
HANDLING. TRAINING
ASSEMBLY. AND
STORAGE CERTIFICATION
~ J
FINAl
TESTING OA
l
\
'"
MARKING.
PACKAGING. ~
SHIPPING
! COMPANY
HIGH YIELD,
HIGH RELIABILITY.
"'---- STATIC-
AWARENESS
PROGRAM
NO LATENT FAILURE
Conclusion
This thumbnail sketch has given textbook examples, but actual ESO-control situ-
ations are often unique, with no ready-made answer. You must arm yourself with
"real and conceptual tools" and run your own tests to get the facts for a cost-
effective solution (Ref. 3-9). Without facts, the only alternative is to try to buy
your way out of the problem with ignorant overkill.
At one extreme of ESO expertise is a naive customer sitting at a costly ESO-safe
assembly bench heaped with superfluous equipment touted by glib salesmen, and
at the other extreme is an ESO "guru" working at a wooden table without even a
wrist strap because he knows exactly what he's doing. Your ESO-safe workers
should be somewhere in the middle: knowledgeable but backed up by "safety nets"
to allow for human error. In "Defenses," above, static-limiting floor finish was
mentioned as an example of such a safety net.
To put ESO control succinctly: "Work smart," and "buy the basics, forget the
frills."
Most important of all, make every operator a "little guru" and "captain of his/her
own workstation," because operator skill is the key element in a successful pro-
gram. Remember that an inept sea captain can wreck the most modem ship equipped
with every sophisticated radar and sonar safety device.
Chapter 4
First, about fifty principles, rules, definitions, test methods, etc., are surveyed "from
A to Z" under the heading "Real and Conceptual Tools: Definitions." This is not a
glossary of ESD terms but a limited list of those that are especially important to
our approach; definitions of other terms, such as abbreviations for types of de-
vices, can be found in textbooks on electronics.
Second, the terms are elucidated under the heading "Real and Conceptual Tools:
Discussion," and three of the most important ones, the COM, SSP, and SSW, are
given chapters of their own, Chapters 5, 6, and 7. These tools serve to troubleshoot
in novel situations (Chapter 8), to develop specifications (Chapter 9), and to build
a logical and thorough ESD-control program (Chapter 12). As mentioned in the
Introduction, these key terms appear in bold type throughout the text.
ESD control has its complexities but is basically a simple matter of avoiding
discharges and dangerously intense E fields--not all E fields, which would be
impractical. (Note: In our usages, "E field" and "field" are synonymous.) If all
objects in the SSZ are groundable and grounded, and if the CD Rule is followed,
most of the ESD problems will have been solved. But the reader shouldn't be
lulled into false confidence by this statement, because achieving these basic re-
quirements of grounding and following the CD Rule is more easily said than
done, human nature being what it is, and a residue of problems, exemplified by
those in Chapter 8, will remain. These may be solved by diligent use of our con-
ceptual tool kit.
Table 4-1 categorizes the tools as primarily applicable to either fields or dis-
charges except when no distinction can be made. There is often much overlap;
humidification, for example, reduces not only fields but also discharges (which
may be visible sparks when the air is dry enough), but it is most often thought of as
lessening the static charge on nonconductors, especially common plastics, to
reduce the danger of the FIM. The point of the table is that every concept, mate-
rial, or object in this chapter plays a role in preventing OJ or the "internal OJ"
caused by induction ("pushing" or "pulling" of electrons by a field to create
polarization within a device).
22
REAL AND CONCEPTUAL TOOLS FROM A TO Z 23
holes, such as made by staples, are allowed. Wire screen, e.g., FED-SPEC-RR-W-365 ,
Type VII, 18 X 16 regular, is an excellent Faraday cage, equivalent to at least
6-mil foil. (Note: This is our own definition.)
Faraday Cup: A shielded cup with attached electrometer for measuring the
charge on any object dropped into the cup.
FFB: Field-From-Board damage mechanism.
Field: An E field in the context of this book.
Field Meter: An instrument for measuring E fields.
FIM: Field-Induced Model, a damage mechanism.
Grounding: Metallic connection with the earth to establish zero potential with
respect to earth. Grounds for ESD purposes, e.g., for an SSW, include water pipes
and building structural steel. Grounding also means the electrical connection of
conductive, static-dissipative, or antistatic materials with earth in any manner.
HBM: Human Body Model, a damage mechanism (special case of DI).
Humidification: Controlling relative humidity, e.g., to 30% or 40% minimum
at 70°F, to reduce the triboelectric charging of materials in the SSW.
Insulator: A nonconductor.
Ionization: The ionization of air for the purpose of neutralizing charged surfaces.
Latent Failure: Failure of an ESDS device because of previous degradation
and weakening by one or more ESD events.
MM: Machine Model, a damage mechanism.
MOSFET: Metal-oxide semiconductor field-effect transistor. Used in the dis-
charge test and in special tests.
Nonconductive: Having a surface resistivity of at least 1012 ohms/square (our
definition, consistent with that of MIL-HDBK-263 and the Electronic Industries
Association).
Nonconductor: A nonconductive material.
Operator Discipline: A rule of conduct for operators that tends to minimize
the chance of ESD damage to ESDS items being handled.
Personnel Voltage Tester: Trade name for a voltmeter with high input im-
pedance and low output capacitance, for measuring the instantaneous voltage on
people.
Shielding: Attenuation of an E field by an object between the charged surface
and a point in space at which the field strength is measured. The more conductive
the object blocking the field, the better the shielding.
Shunting: "Shorting out" or connecting leads together as a means of lowering
the ESD sensitivity of an ESDS item.
Special Test: A test designed to resolve a specific ESD-problem situation.
Typically, the test conditions are worst-case, yet realistic, so that confidence is
high if the test is passed.
SSP: Static-safe package. This is an SSZ within a Faraday cage or within
material providing sufficient shielding to pass a discharge test on the particular
configuration of shipped/stored ESDS item and packaging.
26 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Basic Rule: For ESDS items sensitive to 100 V by the OBM, this rule may be
rewritten: "An ESDS item shall never, even for a nanosecond, be exposed to an E
field in violation of the CD Rule or have its sensitive leads or terminals touched
to, or receive a discharge from, any surface at more than 50 V."
"Even for a nanosecond" refers to the rapidity at which ESD damage can occur.
Only 25 V applied for 100 nanoseconds, at the chip, can destroy memories or
microprocessors (Ref. 4-2). In the waveform required by MIL-STD-883, Test
Method 3015, referenced in MIL-STD-1686, the rise time for the OBM is a maxi-
mum of 10 nanoseconds (Ref. 4-3). A typical ESD event involves rapid discharge
with rise times of l00{}-3000 volts/nanosecond (Ref. 4-4).
Although a charge of 300 V on nonconductors is harmless for devices sensitive
to 100 V by the OBM (Figure 3 of Paper No.3 in the Appendix), it is usually not
practical to distinguish nonconductors from other materials, and nonconductors
shouldn't be in an SSW. Therefore, 50 V is stated as the limit for any surface.
A basic rule suggested by a supplier of conductive ESD-control materials is:
ESDS items "may be handled or approached only with highly conductive grounded
objects," and "storage or transportation can only be done in Faraday cages (maxi-
mum of 1()4 ohm/square material)." This is the "conductive approach." Note that
our definition of Faraday cage is more rigorous; see the discussion.
In contrast to the conductive approach, a supplier of antistatic materials has
long insisted that all surfaces in the SSW must be "nonsparking." This is the "an-
tistatic approach," and its merit versus the conductive approach is a notable con-
troversy in the very controversial field of ESD control. Which of these two
approaches is right? The conductive approach is correct in theory but lacks the
"forgivingness" or latitude for error needed in imperfect, real-life situations. Spe-
cifically, the CDM and DJ are not taken into account. ESDS items, materials, or
even careless operators might be charged, and conductive surfaces can increase
ESD damage by rapid discharges as discussed under "CDM" below. Also, con-
ductive items, e.g., bags, can carry lethal currents if a 1lO-volt hot lead falls onto
them (Paper No.7 in the Appendix); this danger is improbable. but safety is so
important that even unlikely possibilities must be considered.
There are also other arguments against conductive materials, such as sloughing
of metal flakes from metallized bags or particles from carbon-loaded plastic bags
or tote boxes; these conductive flakes or particles could cause shorts by falling
into open devices or onto printed circuit boards to bridge circuit lines. On the other
hand, antistatic materials of the conventional type (depending on a fugitive anti-
stat) have a permanence problem (Paper No.2 in the Appendix), and their rubbed-
off or volatilized antistats can cause contamination (Paper No.9 in the Appendix).
Nothing is perfect.
On balance, we use and recommend a mix of antistatic, static-dissipative, and
conductive materials, as stated just before the "Conclusions" section of Paper No.
3 in the Appendix. We prefer to avoid conductive materials (note that "conduc-
tive" is defined by the conductivity of the surface, so an externally antistatic foil-
laminate bag, for example, is not conductive), but we use them when no alternatives
REAL AND CONCEPTUAL TOOLS FROM A TO Z 29
are conveniently available. After all, conductive items such a screwdrivers and
other hand tools are ubiquitous in SSWs. We merely try to minimize the addition
of more conductors.
In conclusion, the Basic Rule guards against fields, by means of the CO Rule
and control of triboelectric charging, and also against discharges, by means of
grounding of materials and operators, by operator disciplines, and by using cou-
pons to make automated processes ESD-safe.
Capacitance: Capacitance C is related to the charge Q and the voltage V of an
object by the equation: Q = CV. This means that for a given Q, or quantity of
electrons present or missing, a decrease in C, which is the capacity to hold the
charge, results in a rise in V, which is the "looseness" of electrons (or "anxious-
ness" of their lack) that manifests itself as an E field. Thus, when the capacity to
hold the charge decreases, the charge is less tightly gripped or restrained and radi-
ates as a field; when the capacity increases, this field collapses.
In practice, a charged object is one plate of a condenser (capacitor) separated by
air as the dielectric layer from a ground plane such as a workbench surface or a
floor, which is the other plate of the condenser. When the distance between the
plates doubles, Q remains the same, C is halved, and V doubles. This relationship
is discussed in connection with the ARTG in Paper No. 10 in the Appendix. The
most familiar, and important, example of V rising as C falls is a charged person
lifting his foot.
COM: For a discussion of CDM, see Chapter 5. The nature of the COM has
been questioned, with the suggestion that damage occurs in the charging process
rather than during a discharge, so that the COM resembles the FIM. However,
our observations on the FFB, which is a kind of COM, indicate that a discharge
isn't necessary for ESD damage but increases its likelihood (Paper No.8 in the
Appendix).
Insofar as 01 is a part of the COM, a conductive surface tends to be more
damaging than an antistatic surface (Chapter 5). This was demonstrated in tests
which support the intuition that an antistatic surface slows the rate of a discharge
and is safer than a conductive surface. Hence we discourage the use of conduc-
tive surfaces but don't forbid it; see "Basic Rule" above.
CO Rule: This empirical rule is based on MOSFET tests as described in Pa-
per No.3 in the Appendix. Since the MOSFET board used in the tests is an exag-
geratedly sensitive design (large "antennas"), the CO Rule is worst-case, and this
worst-case quality relieves uncertainty in field-meter readings as explained under
"Apparent Charge" above. The equation d = V°.5/1.8, where d is the minimum
safe distance in inches and V is the apparent charge in volts, can be expressed as
a table as included in Model Specification 1 in Chapter 9.
The CO Rule is incorporated in a company in-plant-handling specification, as it
is in Model Specification 1, and several years of experience have proved this rule
to be extremely useful. It has helped solve many problems, a few of which are
described in Chapter 8. The rule is especially valuable in preventing needless ex-
pense; fears of low-intensity fields are discredited so that purchase of equipment
30 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
to control these unreal hazards is avoided. For example, conductive chairs were
shown to be unnecessary in Example 5 in Chapter 8. In other words, the rule fore-
stalls "overkill" and maximizes cost-effectiveness.
Based on the monetary savings provided by the CD Rule, field meters should
be generously distributed to operators. One field meter for each SSW is ideal. Not
only do the meters allow cost-effective solutions to ESD problems via the CD
Rule, but they are a prime tool of eternal vigilance to exclude common plastics
and other sources of fields from the SSW.
In conclusion, the CD Rule will be one of your best friends in ESD control. Use
it often.
Conductive, Conductor: Conductivity is defined by surface resistivity be-
cause surfaces tend to be the locus of currents in ESD-control work; examples are
vapor-deposited metal layers or antistat sweat layers. In the test (ASTM D257 or
equivalent), current flows through the bulk of a conductive material such as car-
bon-loaded polyolefin, not just on the surface, but what matters from a practical
standpoint is not where the current flows but how large it is, because the property
of real interest is the rate of bleed-off of static charges.
Incidentally, this facility to carry current may be too large for safety of person-
nel; see "Basic Rule" above.
For this reason of safety to people and because of slightly more danger to ESDS
devices (see "CDM" above), the use of conductive surfaces is discouraged but
not prohibited in our approach.
Do not use a field meter to measure the true voltage on a conductor. The read-
ing probably will be deceptively low, so that DI may be unrecognized as a threat.
See the precautions under "Field Meter."
Continuous Wrist-Strap Monitor: This subject is discussed in detail in Pa-
pers No.5 and lOin the Appendix. The Basic Rule requires continuous monitor-
ing "even for a nanosecond," and if a wrist-strap system is found to have failed in
a daily check, for example, the operator's skin voltage may have been sufficient to
cause damage by DI for not just a nanosecond but for a whole shift!
More and more continuous wrist-strap monitors, of the resistive type as we
require, are being used in assembly areas. There is no logical alternative, espe-
cially for high-reliability products in which latent failures, however few, could be
disastrous. However, try monitors before buying to be sure the audible alarm is not
annoying, etc.
Coupon: Paper No.8 in the Appendix describes the coupon approach and its
successful application to an automated assembly process. The complete version of
this paper also mentions damage mechanisms and ESD hazards.
01: Electrons flow directly in or out of a device. This is the most straightfor-
ward and common cause of ESD damage, with people being the major threat.
Therefore, controlling voltage on personnel (Paper No. lOin the Appendix) is one
of the most important parts of an ESD-control program.
The HBM is a model for 01 caused by people, but the conditions are not realis-
tically worst-case; see "HBM" below.
REAL AND CONCEPTUAL TOOLS FROM A TO Z 31
Discharge: Discharges and fields are the two ESO threats controlled by the
Basic Rule. A discharge may be very rapid, e.g., with a rise time of 1000-3000
volts/nanosecond, or slow, e.g., 1000 volts/minute, in which case it is called "bleed-
off' or "drain." Conductive surfaces favor rapid discharge, and there is evidence
that these surfaces are therefore more hazardous to ESDS items than are anti-
static surfaces (see "COM" above).
A discharge is a flow of electrons, with the quantity of electrons being mea-
sured in coulombs and the rate of flow (current) in amperes = coulombs/second. A
discharge creates a field which may itself cause ESO damage.
Discharge Test: In variations of this test, we use a resistance of 1500 ohms as
in the HBM (Chapter 9, Model Specification 2, para. 4.3.3, option I), 150 ohms
(same reference, option 2), or 0 ohms (Test No.1 in Paper No.4 in the Appendix).
For tote boxes, electrodes are used instead of a capacitance probe (see under "Test
Methods" in Paper No. 7 in the Appendix). The test becomes much more rigorous
as the resistance drops.
The shielding/discharge test with a capacitive sensor and oscilloscope (see Fig.
2 of Paper No.3 in the Appendix) has become a standard for bag materials (in EIA
541, "EIA" being the Electronic Industries Association), but this test is misleading
as explained in Paper No.3 in the Appendix. The flaws in this test are (1) it gives
falsely high (good) readings when metallization is exposed rather than buried, and
(2) the flat electrodes don't provide a spark discharge. A "real-life" spark from a
person's finger evaporates the metal from some see-through metallized bags so
that they afford poor protection; see Table 3 of Paper No.4 in the Appendix. An
analogy of an unrealistic test that is still on the books but has been largely sup-
planted by more realistic tests is ASTM 0635 for flame resistance of plastics.
Nylon passes this test, but glass-reinforced nylon, which if anything should be less
flammable, fails. Why? Because the low melt viscosity of nylon allows the molten
region to drip off, carrying the flame with it, unless glass fibers support the melt so
that the flame can propagate! The shielding/discharge or pulse test of EIA 541 is
just as unrealistic. Another test corrected ASTM 0635 by adding cotton below the
burning specimen with a provision that burning, falling drops must not ignite the
cotton for the test to be passed. Similarly, a correction to the pulse test would be to
add a spark discharge as we have done in our modified version. See Chapter 7 for
further discussion of unrealistic versus realistic tests.
It is interesting that researchers at a major corporation producing see-through
metallized bag material noted evaporation of the metallization by a spark and even
studied the area evaporated versus voltage of the discharge, but this information
was suppressed because it might have harmed sales. Our position is that spark
discharges are a significant threat and bags should resist them. See Paper No. 20
in the Appendix.
E Field: Fields are, of course, the threat, along with discharges, that the Ba-
sic Rule controls. For practical purposes (the CD Rule), field strength is mea-
sured in terms of apparent charge and the distance from the charged surface. See
the discussion under "Apparent Charge" above.
32 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
ESDS, ESDS Item: ESDS devices fail in two common ways. Voltage-induced
failure dominates for discrete MOS (metal-oxide semiconductor) devices. A volt-
age of 80-100 V exceeds the dielectric-breakdown strength of the common thick-
ness of 1000 angstroms of oxide. After breakdown, the part usually shows
permanently degraded leakage characteristics and is more susceptible to future
damage; it has a reduced operating margin and is likely to fail.
The other common failure mode is current-induced failure. This appears in
Schottky or PN (bipolar) junction ICs, which are subject to destructive reverse
breakdown in which the instantaneous power of the discharge (more than 5 kW
for less than I J.lsec) generates enough heat to melt silicon and metal. Thin-film
resistors are damaged by surface breakdown that bypasses part of the resistor and
leaves a shunt path, or else metal is vaporized or melted from the thin film. In
either case, the resistance value usually shifts, which is serious in a precision (less
than 0.1 % variation) part (Ref. 4-5).
Typical failure distributions for bipolar ICs are 90% junction burnout and 10%
metallization burnout. For MOS integrated circuits, this changes to 63% metalli-
zation burnout and 27% oxide punchthrough (Ref. 4-2). For a full discussion of
failure mechanisms, see MIL-HDBK-263.
MIL-STD-1686, which superseded DoD-STD-1686, defines three classes of
ESD-sensitivity: Class 1 (0-1,999 V), Class 2 (2,000-3,999 V), and Class 3
(4,000-15,999 V). Class 1, which is the concern of this book, includes microwave
devices (such as Schottky barrier diodes), M OSFETs, surface acoustic wave (SAW)
devices,junction field-effect transistors (JFETs), charged coupled devices (CCDs),
precision voltage regulator diodes, operational amplifiers (OPAMPs), thin-film
resistors, integrated circuits (ICs), hybrids using Class 1 parts, very high speed ICs
(VHSIC), and some silicon-controlled rectifiers (SCRs).
Note that components mounted on an assembly are relatively ESD-safe only if
the assembly has protection circuits at all sensitive nodes. The risk of ESD damage
can actually increase for mounted devices because each printed conductor con-
nects to several devices and a discharge to that conductor stresses several devices,
not just one. (Note: The word "conductor" in the last sentence is not in bold type
because it is used in the special sense of "circuit line.") Another factor is that
circuit lines can act as antennas to intensify the effect of fields and promote dam-
age by the FIM. Furthermore, CMOS (complementary metal-oxide semiconduc-
tor) circuits subjected to ESD while powered are subject to an additional risk called
"latchup," which is an avalanche effect causing overheating and catastrophic fail-
ure (Ref. 4-5). We repeat: ESDS devices mounted on boards can be at even more
risk than when unmounted (Ref. 4-6).
Faraday Cage: As defined in technical dictionaries, a "Faraday cage" is a
room enclosed with an earth-grounded metal screen, and the only commercial ESD-
protective bag in accordance with this definition (except for grounding) is con-
structed of aluminum wire screen (as used for windows) sandwiched between layers
of antistatic polyethylene (see Paper No.4 in the Appendix). However, for most
purposes, 0.25 mil of aluminum foil, as in commercial foil laminate bags (see
REAL AND CONCEPTUAL TOOLS FROM A TO Z 33
Table 10 of Paper No.3 and Table 3 of Paper No.4 in the Appendix) is sufficient.
We recommend these foil laminate bags for all in-plant handling as well as ship-
ping (but nonfoil package constructions may meet the discharge test of Model
Specification 6). However, consider the screen bag when visibility of the contents
is needed and extremely high ESD sensitivity, high cost, or need for high reliabil-
ity of the enclosed item justifies a relatively expensive package.
Note that being a Faraday cage or equivalent is not enough for an SSP, because
the inside of an SSP must be an SSZ in which triboelectric charging cannot violate
the Basic Rule by creating hazardous fields. The old MIL-B-81705, Revision B,
Type I, was a perfectly good Faraday cage, but its nonconductive liner was unsuit-
able for an SSZ; in a test, miniature circuit boards with MOSFETs were shaken in a
bag and found to be damaged due to triboelectric charging (Ref. 4-7). Our own
tests verified that this liner was nonconductive. The new MIL-B-81705, Revision
C, Type I is satisfactory. It contains a layer of opaque metallization, not foil.
In conclusion, use foil laminate bags or opaque, metallized bags (MIL-B-81705C,
Type I) with antistatic surfaces both inside and outside, or, in special cases, use
screen bags with antistatic plastic sandwiching the screen (see Paper No.4 in the
Appendix).
Faraday Cup: For examples of the use of this test equipment in studying the
propensity for triboelectric charging, see Papers No.6 and 7 in the Appendix. A
field meter measures charges indirectly and imperfectly, whereas the Faraday
cup gives accurate reading on any charged object dropped into it. However, when
the intensity of the E field is the property of real interest, as it is using the CD
Rule, the relevant parameter is the apparent charge read by a field meter and
expressed in volts.
FFB: This damage mechanism is a special case of the FIM. Triboelectric
charging of circuit boards, e.g., by sliding on stainless-steel conveyor belts or
oven shelves, creates a field which damages ESDS devices by induction as ex-
plained under "FIM" below.
Field: See "E Field."
Field Meter: This has been and still is the single most useful tool in ESD
control.
Field meters commonly are calibrated at the factory with a charged I-ft2 plate,
but in actual measurements the target areas vary, with a large effect on the reading
(Ref. 4-8). Thus, the actual voltage, or charge, on the surface can't be accurately
determined, but what matters for our purposes is the apparent charge as dis-
cussed under "Apparent Charge" above.
Though the real charge is irrelevant, we do want a properly determined appar-
ent charge as a measure of field strength, and there are a few requirements for
using the meter correctly.
First, the meter must be calibrated. This is suitably done annually, with a cali-
bration sticker being placed on the meter.
Second, if the meter has a grounding outlet, it must be grounded with a lead or,
if the meter has no such outlet, the operator holding the meter must be grounded.
34 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Third, the meter must be zeroed on a grounded conductive object, suitably the
grounded operator's hand. Note that meter readings are relative. If neither op-
erator nor meter is grounded and the operator is charged to -1000 V and has
zeroed the meter on his hand, he will read a grounded I-ft2 surface as +1000 V,
or he will read a surface at + 1000 V as +2000 V. Thus, to purposely read the
approximate charge on himself, the ungrounded operator can take a reading on a
grounded I-ftl surface and reverse the polarity. Or a larger surface can be used
for a rough result.
Fourth, the apparent charge must be read according to the meter manufacturer's
instructions; for example, the meter might read true at 6 inches from a surface, or
the distance might be increased to 12 inches, to keep the needle on scale, and the
reading doubled.
We recommend that every SSW have a field meter to be used in complying
with the invaluable CO Rule. However, the meter must be used only for apparent
charges. It is not to be used for reading voltages on conductors because the true
voltage is not shown by the meter unless the conductor happens to be a I-ft2 plate
(assuming the meter was calibrated with this). For example, because of its small
area, a wire at 500 V might read only 50 V on the meter. According to the CO
Rule, this reading of an apparent charge of 50 V prescribes a safe distance of 4
inches, which is fine, but a requirement of the Basic Rule for ESOS items sensi-
tive to 100 V is that a sensitive lead may not touch a surface at more than 50 V in
real voltage. In this example, if the bare wire were touched, the item would be
damaged by OJ. In other words, field meters, as their name indicates, are for
judging fields, via apparent charge and distance, not for measuring true volt-
ages! To measure voltages on conductors, use a contact or noncontact voltmeter.
A noncontact voltmeter is a type of field meter, but it is specially calibrated and
applicable only to certain sample geometries; it is not an ordinary, general-purpose
field meter.
Operators should be given detailed instructions on using the field meter
because it is such an important tool. For familiarization, they should charge a
sheet of non conductive plastic, e.g., polyethylene, by rubbing with polyester
or woolen fabric, and note the unevenness of the voltage readings over the
surface because electrons or electron-deficient areas are clustered irregularly;
in such cases, the highest reading, which is the most conservative for protec-
tion purposes, shall be used as the apparent charge in implementing the CO
Rule. Also, operators should rub the plastic against grounded stainless steel
and note that grounded conductors, though they never show a charge them-
selves because it immediately drains off, are very effective in charging non-
conductors including circuit-board laminates. Intuition may suggest that
grounded metals are harmless in regard to ESD, but this idea is dangerously
wrong! Not only can they cause triboelectric charging, but they can be tar-
gets for rapid discharges in the COM: see "COM" above. The skin is another
conductor that charges nonconductors by rubbing, as the operator can
REAL AND CONCEPTUAL TOOLS FROM A TO Z 35
STATIC VOLTAGES
Ionization for no valid, well-defined reason is useless and may actually be part of the
ESD problem by creating charges on ungrounded conductors; that such charging can
occur is indicated by GIDEP Alert H7-A-85-02, issued in 1985, on an electrically unbal-
anced AC ionizer. Hence, ionizer manufacturers have shown much concern about bal-
ancing, which in some models is now automatic.
REAL AND CONCEPTUAL TOOLS FROM A TO Z 37
better than dry nitrogen, so a nitrogen nozzle can be configured to allow air to be
aspirated into the gas stream ahead of the ionizing element; this feature appears in
electrical designs but not in nuclear ones, so the electrical type should be used on
nitrogen lines for best results. See Example 38 in Chapter 8 on air ionization.
Note that there are several possible causes of static charge formation in a blow-
off operation. These include (1) triboelectric charging by flying dust or other
particles or by sliding of displaced objects, (2) evaporation of nonaqueous sol-
vents, (3) deformation of solids (piezoelectric effect), and (4) cryogenic charging,
e.g., by freezing of water. In the last case, small splinters supposedly are ejected
from the freezing drop, and they either carry off or leave an excess of electrons so
that the remaining drop freezes to a charged ice particle. (See Chapter 2.)
In conclusion, ionization has its pros and cons. Pros are: (1) background static
charge levels are reduced, (2) there can be an air-cleaning effect, and (3) ionization
is the most practical method of neutralizing charged nonconductors; indeed, ioniz-
ers are sometimes essential, as in solving the problems in Examples 25 and 34 of
Chapter 8. The cons are: (1) dangerous space charges which charge ungrounded
(electrically isolated) conductors are possible, (2) ozone may be generated above
the OSHA limit, (3) the neutralization process (rate of charge decay by air ions
impinging on a surface) is relatively slow so that ESD damage can occur in the
course of triboelectric charging (the field created by pulling tape from a roll will
light a neon bulb in the midst of a stream of ionized air), (4) work areas are incom-
pletely covered by blowing ionized air because of shadowing by obstructions, and
(5) ionization is ineffective when a static charge is hidden by voltage suppression.
As an example of the latter, a charged sheet of plastic lying on a workbench surface
is immune to neutralization by ionized air because the field is collapsed and cannot
attract air ions of the opposite polarity. However, the field reappears and may cause
damage by the FIM when the plastic is lifted from the surface. To put it another way,
the plastic sheet lying on the surface makes an effective capacitor so that C is rela-
tively high and V is very low in the equation Q = CV (see "Capacitance" above).
In view of these advantages and disadvantages, we recommend using ioniza-
tion only "when risks are carefully controlled and are outweighed by benefits," as
we said above. Select ionizers from a major manufacturer with an established repu-
tation, and insist on monitoring and autobalancing features to minimize the space
charging hazard (see Example 42, Chapter 8).
In the years from 1989 to 1995, room ionization as a general precaution has
become increasingly popular, especially in electronics factories in the Far East
with large installations, e.g., 50,000 fe. Also, ionization to reduce airborne par-
ticle counts in c1eanrooms has gained more acceptance. Improved autobalancing,
erosion-resistant emitter points, and other evolutionary developments have en-
couraged the acceptance of ionization techniques.
Latent Failure: This is one of many controversial subjects in the field of ESD
control. Latent failure is accepted as being a real phenomenon (for example, Refs.
4-2,4-16, and 4-17). But the question remains: How frequent are latent failures?
The consensus of opinion by leading ESD experts is that these failures are indeed
real but are very rare. Nevertheless, one latent failure in a high-reliability aerospace
REAL AND CONCEPTUAL TOOLS FROM A TO Z 39
system, for example, might be catastrophic, so full ESD precautions must be taken
even if the chance of latent failure is remote. The above comment on rarity per-
tains to total failures, but some device experts claim that malfunctions such as
intermittency are not rarely but usually caused by ESD events in the manufactur-
ing process. In critical applications such as life-support systems or computerized
weaponry, such malfunctions, though short of complete failure, might themselves
be disastrous. Also, replacing ESD-damaged chips in a defective end product gets
"very expensive terribly fast" (Ref. 18). In conclusion, the possibility of latent
failure, or at least latent malfunction, is a good reason for rigorous ESD control.
MM: In the Machine Model, popular in Japanese ESD literature, a metallic
tool participates in DI. The resistance is nearly 0 ohm.
MOSFET: This has been the "white rat" in much of our research. It is appro-
priate for worst-case testing when the ESDS devices being protected are less sen-
sitive than the MOSFET. Other methods of recording voltage transients may be
used, but MOSFET damage is graphic and convincing because an actual compo-
nent is degraded. Endless special tests with MOSFETs can be devised; see Ex-
amples 10, 12, 18, 19,21,26,27,30,31,33, and 35 in Chapter 8.
Nonconductive, Nonconductor: MIL-HDBK-263 describes nonconductors
as having a surface resistivity of 10 12 ohms/square or more, because a surface
resistivity below this level is needed for rapid bleed-off of static charges as dis-
cussed under "Antistatic" above. On a nonconductor, a static charge, true to its
name, is static or motionless more than long enough (for only nanoseconds are
needed) for its associated field to threaten ESDS items by the FIM.
Operator Disciplines: This subject is discussed in Chapter 7 on the SSW, and
disciplines are incorporated in Model Specification 1 in Chapter 9. No matter how
excellent the ESD-control methods and equipment, the operator is all-important,
as the pilot of a plane is all-important, because the best equipment in the world can
be misused; ESD damage can be done, or the plane can crash. See the discussion
of the importance of operator skiIIs in Reference Document 6 in Chapter 9.
Personnel Voltage Tester: In using this instrument, a charged person merely
pushes a button to record his instantaneous voltage. For examples of test data, see
Paper No. lOin the Appendix. This is voltage on the skin, of course, not on garments.
Shielding: Even materials which are less than conductive provide partial
shielding from fields; for example, an antistatic polyethylene bag gave 75% shield-
ing in a shielding/discharge test (Table 9 of Paper No.3 in the Appendix). How-
ever, conductors are required for a high level of shielding. We prefer a "buried"
layer of at least 0.25 mil of aluminum foil (or the electrical equivalent) for bags or
other containers for in-plant handling; see "Faraday Cage" above. For trouble-
shooting or solving unique ESD problem situations, shielding can be provided by
various conductive barriers, e.g., a chain link fence protecting an SSW in Example
29 of Chapter 8 or a connector shell protecting ESDS pins in Example 30.
Like grounding, ionization (used only when necessary), or the use of topical
antistat, shielding is a basic ESD-control technique.
Shunting: Shunting of ESDS leads or terminals does not assure complete
ESD safety. For the best measure of safety, the consensus of opinion at present is
40 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
and making the second one negative. Only if the surface resistivity falls below
about 10 12 ohms/square will the electrons cease to be static (stationary) and slide
from the surface to give the rapid bleed-off that an antistatic surface needs to be
useful in ESD control.
A static charge commonly results from triboelectric charging, but there are
other sources, including freezing of water, evaporation of solvents, ion and elec-
tron beams, atomizing of liquids (spray charging), photoelectric charging, corona
charging, and deformation of solids (piezoelectric effect of crystals). See Chapter 2.
Static-Dissipative: We use the old definition from DoD-HDBK-263.
Static Field: See "E Field" above.
Static-Limiting Floor Finish: This subject is thoroughly discussed in Paper
No.6 in the Appendix.
A static-limiting floor finish is a backup or safety-net precaution like
humidification. Another is room ionization, which we don't recommend except
for cleanrooms when advantages, e.g., cleaning of the air, are demonstrable and
risks minimal. Unlike these other methods, the special floor finish is inexpensive
and quite unobtrusive. Tile floors must be "waxed" somehow, so why not use a
static-limiting floor finish? Choose one that is durable and easily maintained by
standard, simple procedures. Triboelectric charging of shoe soles is the property
of interest, but surface resistivity is related to this for a given finish and is conve-
niently monitored; 10" ohms/square by a Voyager SRM-110 meter is a typical
allowable maximum before the floor coating must be restored. (Note: Among dif-
ferent finishes, the correlation between triboelectric charging and surface resis-
tivity is imperfect; see Paper No.6 in the Appendix.)
Remember that basic floor-finish properties, as well as ESD properties, are vi-
tal. For example, slip resistance must be retained throughout the wear cycle for
employee safety.
Be careful to follow the floor-finish supplier's instructions for maintenance of
the finish. For example, don't use diluted finish as a cleaner, because the resulting
sticky deposit will collect dirt; use a special floor-finish cleaner, containing an
antistat, which will restore the antistatic property, without stickiness, to the surface.
Surface Resistivity: Values obtained for this property depend on electrode
configuration, pressure on the electrodes, roughness of the surface being mea-
sured, applied voltage, and other factors, so that methods of measurement are a
deep subject (Ref. 4-19). However, for our purposes, an approximate, easily made
determination by a Voyager SRM-II 0 meter or equivalent is sufficient. The criti-
cal property is bleed-off time for static charges and, since we don't have an exact
criterion for bleed-off time, which is arbitrary, we don't need precise values for the
property, surface resistivity, that controls bleed-off time. As we said in a discus-
sion of test methods (Ref. 4-1), "We need only to know if a feeble current will run
across [a plastic surface). So trying to split hairs on the exact resistance is not very
productive thinking." In other words, an acceptable surface resistivity value, e.g.,
below 10 12 ohms/square, is just an indication that enough current will flow across
a surface so static charges can't accumulate.
42 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
One surface resistivity meter, such as the Voyager SRM-ll 0 or SRMJRTG Meter,
which measures resistance to ground also, is desirable for every assembly area
where ESDS items are being handled. The meter will see much use in checking
floor finish, questionable plastics, performance of antistatic materials when the
relative humidity drops, etc.
Topical Antistat: Topical antistat treatment is almost a panacea for ridding
the SSW of nonconductors. We say "almost" because topical antistats have their
deficiencies: invisibility, fugitivity, and possible contamination.
Because of the invisibility of the antistat, treated surfaces must be "so labeled," as
required in paragraph 3.6 10 of Model Specification 1 in Chapter 9. Fugitivity means
that the antistat will disappear by vaporizing or wearing off at some unpredictable
rate, so periodic checks must be made with a field meter; fortunately, this is easily
done. For possible contamination effects, see Paper No.9 in the Appendix. If corro-
sion by chloride ion is feared, e.g., when silicon wafers are being handled, use one of
the "chloride-free" (actually, very low in chloride) topical antistats on the market.
Despite the above limitations, topical antistats are a good friend in ESD con-
trol, but don't overdo your reliance on them. Eliminate as many nonconductors as
possible before antistat-treating the necessary ones remaining in the SSW.
Triboelectric Charging: The triboelectric series is only approximately valid;
for example, an anomaly is noted in the complete version of Paper No. 6 con-
densed in the Appendix. However, for reference, here's a brief version including
modem materials and deleting the irrelevant fur and sealing wax:
Positive End
Air
Hands
Glass
Hair
Quartz
Nylon
Wool
Silk
Aluminum
Paper
Cotton
Steel
Wood
Rubber
Gold
Polyester
Polyurethane
Polyethylene
Polypropylene
Vinyl
REAL AND CONCEPTUAL TOOLS FROM A TO Z 43
Aclar (polychlorotrifluoroethylene)
Silicon
Teflon
Negative End
In theory, when two substances from the list are rubbed together, the one higher
on the list becomes positively charged. Thus Aclar, which is a useful material
when triboelectric charging (or "triboelectrification") is wanted, almost always
takes a negative charge. Note that the same material, e.g., polyethylene, can charge
itself so that one area of a sheet may be positive and another negative. Also, "circles"
occur in which A is charged positively by B, B is charged positively by C, but C is
charged positively by A instead of being charged negatively as it "should" be.
Experimentally, triboelectric charging is notoriously erratic.
Triboelectric charging tests include the bag shaker method using quartz and Teflon
disks to bracket most of the triboelectric series (Appendix G of EIA 541) and the
roller test or drag test illustrated in Figs. 5 and 6 of Paper No.6 in the Appendix.
Voltage Suppression: This is both a concept and a technique for ESD control.
A good example of voltage suppression is the minimal apparent charge that can
be observed on painted, grounded metals (allowed in the SSW as necessary non-
conductors in paragraph 3.6.10 of Model Specification 1 in Chapter 9) or plastic-
coated metals (also allowed in the same paragraph of Model Specification 1; note
the vinyl-coated steel tote box in Paper No.7 in the Appendix). Another example
is the relatively low apparent charge on hair hanging close to the skin (Example
2 in Chapter 8).
Use voltage suppression whenever possible to "kill" fields and achieve com-
pliance with the CD Rule.
Walk Test: This test is illustrated in Fig. 7 of Paper No.6 in the Appendix. It
may be augmented or replaced by measurements with a PVT-300 Personnel Volt-
age Tester as utilized in Paper No. lOin the Appendix. However, when using the
PVT-300, be sure to take at least 10 readings in order to catch, or partially catch,
the voltage spikes occurring when the shoe leaves the floor and capacitance drops.
The walk test is a good example of a bottom-line practical and realistic test that
directly measures the property of concern-voltage on people in this case. Labo-
ratory tests tend to be more reproducible but less relevant to real-life use of materials.
Zapflash: This inexpensive and convenient continuity-checker can also be
used to demonstrate various ESD phenomena such as voltage surges on a person
caused by triboelectric charging when he shuffles his feet on a carpet. We recom-
mend that every SSW have a Zapflash for verifying grounding as well as a field
meter for use in complying with the CD Rule.
"Zapflash" is a trade name of Anderson Effects, Redlands, CA. Besides show-
ing continuity up to about 7 megohms, the Zapflash can be used for checking
batteries, lamp filaments, and fuses, for detennining diode or DC voltage polarity,
for revealing AC power leakage to grounded equipment cases or machinery (which
could be a DI hazard), and for identifying the hot lead of an AC outlet.
Chapter 5
This model was born in 1974 with the proposal by Speakman that a part such as an
integrated circuit might be damaged or destroyed by rapid discharge of static elec-
tricity accumulated on the part's own body (Ref. 5-1). In 1980, the CDM was
judged the predominant failure mode at AT&T, as typified by dual-in-line pack-
ages (DIPs) sliding within their plastic packaging tubes to become triboelectrically
charged with most of the charge residing on the lead frame (Ref. 5-2). In 1992, it
was said, "It has become clear in the past five years that the CDM, not the 8BM,
is responsible for the vast majority of ESD damage, certainly to discrete devices
and probably to entire circuit cards and assemblies" (Ref. 5-3). Reflecting industry
concern about the CDM, certain packaging materials are now being advertised as
"CDM-safe."
Because of concern about the CDM, we attempted to define how conductive a
"conductive" surface must be to cause the "rapid discharge" required by this
damage mechanism. Our results appear in detail in the appended Papers No. 12
and 13 and in summary form in Papers No. 14, 15, and 16. The importance of
CDM safety in packaging is mentioned in Paper No. 20.
In essence, we found that thin-film resistors (current-sensitive) were relatively
CDM-resistant, but MOSFETs, represented by a Static Event Detector™ (3M
Co.), were very vulnerable. In fact, even antistatic polyethylene could be a threat
to MOSFETs at 50% relative humidity though not at 35% (Paper No. 16). Resis-
tance to ground measured through a device lead was 10 10 ohms for no CDM fail-
ures, 109 ohms for some failures, and 105-108 ohms for consistent failures (Paper
No. 16). Skin resistance through the device lead was }Q6-107 ohms (Papers No. 13
and 16), identifying fingers as a prime CDM danger. Hence we say, "A grounded
operator is the lesser of two evils versus an ungrounded operator" among the mot-
toes in Chapter 14.
We've mentioned the CDM danger to DIPs, and a remedy is to let the parts fall
onto an antistatic surface, preferably 1010-1011 ohms/square, when they are emp-
tied from a tube or rail. But the Field-From-Board (FFB) mechanism (Chapter 4)
is another problem. In this damage mechanism, ESDS parts such as integrated
circuits are mounted on a circuit board which becomes triboelectrically charged;
44
THE CHARGED DEVICE MODEL (CDM) 45
Finger
of Grounded
Operator
Negatively
Charged
Board with
Devices
then, touching a lead or terminal with a tool or a finger could cause a damaging
discharge.
Figure 5-1 diagrams the FFB. The negative charge on the board has polarized an
approaching fingertip by driving electrons off onto the operator, and a discharge,
which may involve a spark depending on voltage and capacitance, is imminent.
What must be done to assure that handling of ESDS items is CDM-safe and
FFB-safe? As mentioned in Chapter 4 in the discussion of the Basic Rule, we
favor a practical, cost-effective mix of antistatic, static-dissipative, and conduc-
tive materials with minimization of the conductive ones. Our suggestion is that
ESDS items with a sensitivity of over 100 V by the HBM can be handled safely
with operators' bare fingers if circuit boards are frequently checked for E fields
with a static meter and "washed" with ionized air from a blower as required. The
CDM and FFB require two elements-a static charge and a conductive surface-
and in this case we are removing the charge.
However, ifthe ESDS items are supersensitive, e.g., 12 V, extreme precautions
will be needed (Chapter 15), and it will be prudent to attempt to eliminate both
static charges and conductive surfaces. This means that operators must wear fin-
ger cots, which should be antistatic (at least 109 but not more than 10 12 ohms/
square), not static-dissipative, as demonstrated in Paper No. 16. Now the operator
is no longer the lesser of two evils but has been eliminated as a CDM/FFB hazard.
We've confined the wearing of cots to handling supersensitive items because
the cots are inconvenient to use and interfere with the tactile sensation, or "feel,"
46 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
needed for delicate assembly work. Perhaps antistatic cots of thin, tight-fitting
latex (or elastomer resembling latex) like the fingertips of surgical gloves would
solve the "feel" problem.
In conclusion, we suggest that you read the detailed information in the Appen-
dix and be on guard for the CDM danger in your operations. Strictly enforce the
CD Rule, with special attention to circuit boards with mounted devices. "No charge,
no discharge." Also, insist on ESD-protective packaging with a CDM-safe (anti-
static) surface.
Chapter 6
By the Basic Rule, an ESDS item must always be in an SSZ. This SSZ will be in
(1) an SSP, (2) an SSW, or (3) elsewhere under continuous operator supervision
(paragraph 3.5.2 of Model Specification 1 in Chapter 9). An example of "else-
where" is an open tote box being carried between SSWs by an operator whose
wrist-strap cord is connected to an overhead grounded trolley.
In this chapter we're concerned with the SSP, which is defined in Chapter 4. The
SSP must have an outer Faraday cage or at least pass a discharge test in accor-
dance with paragraph 4.3.3 of Model Specification 2 in Chapter 9.
But how is the SSZ maintained inside the SSP? Remember that we're guarding
from fields and discharges in accordance with the Basic Rule. Since the Fara-
day-cage wall of the SSP excludes external hazards, and there's no source of in-
ternal discharges except internal fields, the only internal hazard is the creation of
fields by triboelectric charging. But this is a big "only," because triboelectric
charging is a mysterious or at least erratic process affected by the least amount of
surface contamination. For example, shoe soles can become charged the opposite
of their original polarity when they become invisibly soiled, and sometimes a sheet
of plastic, oddly enough, can be folded over and rubbed to charge itself, becoming
positive in one spot and negative in another, perhaps because no two surfaces are
quite identical (Ref. 4-14).
Thus, the problem of designing an SSP is attaining high confidence that tri-
boelectric charging inside the package will be minimal. Approaches are: (1) run-
ning laboratory triboelectric charging tests, e.g., the test in Appendix G of EIA
541 or the roller test illustrated in Fig. 5 of Paper No.6 in the Appendix, to screen
candidate materials for an intimate wrap, (2) running a more realistic special test
in which coupons with MOSFETs are shaken inside the package as in Ref. 4-7,
(3) mounting the packaged ESDS item so that it cannot slide within the package
during shipment, and (4) protecting the item (but not completely, as mentioned
under "Shunting" in Chapter 4) by shunting the leads or terminals. In approaches
I and 2, the surface(s) of the item to be packaged should be tested in the triboelec-
tric charging test; the EIA 541 bag-shaking test uses quartz and Teflon to bracket
the triboelectric series (see "Triboelectric Charging" in Chapter 4) because they're
47
48 ESD FROM A TO Z: ELECTROSTAnC DISCHARGE CONTROL FOR ELECTRONICS
at opposite ends of it, but, since the series is fallible, this method of scoping all
materials may fail in practice.
At the present state of ESD-control knowledge and material development, we
recommend using antistatic materials as the intimate wrap, keeping in mind the
question of their permanence (Paper No.2 in the Appendix) and taking no chances
of contamination (Ref. 6-1) or corrosion (Paper No.9 in the Appendix). If at all
possible, secure the item so that it can't slide when the package is vibrated in
shipment, and use the technique of shunting. Probably the worst hazard is static-
generating foam such as polystyrene foam "peanuts." If such a material must be
used, make very sure it's antistatic by a practical shaking test followed by a check
with a field meter. Remember that the package may find itself in relatively dirty
air. Will the cushioning material remain antistatic at, say, 10% relative humidity
at 70°F? Testing is necessary. In our experience, a good brand of antistatic bubble-
wrap is a safer cushioning material than "peanuts." The bubble-wrap can't break
open to expose fresh, nonconductive surfaces, and its resilience tends to hold
packaged items in place. Use it whenever possible. Cardboard boxes with a buried
layer of aluminum foil and with antistatic bubble-wrap inside are excellent pack-
ages.
Though we allow a variety of shielding materials to accommodate suppliers, as
long as the discharge test in Model Specification 2 can be passed, we much prefer
a Faraday cage (by our definition). In a hierarchy of tests (Paper No. 4 in the
Appendix), a foil laminate bag was arbitrarily ranked as "good," while a bag con-
struction with heavy foil (opaque, of course) or metal screen (partially transpar-
ent) was "excellent." Bags with see-through metallization were at best "fair." (Note:
In later tests, a "metal-in" see-through bag was "good"; see Paper No. 15.) Signifi-
cantly, two groups of investigators at major corporations arrived independently at
the same conclusion that foil provides very effective shielding whereas see-through
metallization, which is on the order of only 100 angstroms thick, is dubious. At the
1984 EOS/ESD Symposium, the Rockwell group said, "Only foil-containing lami-
nate bags are recommended for Faraday-cage protection against worst-case fields
and discharges" (conclusion 12 of Paper No.3 in the Appendix), and the British
Telecom group said, "Highly static-sensitive components should always be pro-
tected by metal-foil bags" (Ref. 6-2).
See-through metallized bags have become popular because of their partial trans-
parency, and a capacitance-probe (shielding/discharge) test (in EIA 541) has
been contrived by the bag manufacturers and accepted by naive users. The first
problem with this test is that the voltage is too low (1000 V); the British Telecom
tests (Ref. 6-2) used higher voltages, e.g., 1600 V, which can easily be encountered
in practice. Furthermore, Ref. 6-2 shows that as the area of the electrodes falls, the
shielding effectiveness of see-through metallized bags decreases. This is in the
direction of a spark discharge, and the second problem with the test is that spark
discharges, which are not accounted for by the large flat electrodes used, burn off
metallization and can damage ESDS items in the bag; a person can walk on a
carpet at 2~30% relative humidity at 70°F and touch his finger to a see-through
THE STATIC-8AFE PACKAGE (SSP) 49
metallized bag with the metal protected only by a thin polymeric coating to "blow"
a MOSFET inside; see Fig. 2 and Table 3 of Paper No.4 in the Appendix. Use
only "metal-in" see-through bags, with the metallization protected by a O.S-mil
Mylar (polyester) film.
Admittedly, the MOSFET was configured in the worst possible way (leads con-
nected to upper and lower "antennas"), so that the chance of ESD damage was
maximized, but why take chances? We agree with Ref. 6-3 that "packaging cost
differences of as much as one or two dollars for product protection in these areas
(precision medical, communications, personal hazard, or defense application) are
a minor consideration ... EMIIRFI testing may also be appropriate in such cases."
Yes! And foil or screen gives EMI/RFI protection along with protection from spark
discharges. Furthermore, foil laminate bags cost little or no more than see-through
metallized bags! Screen bags (see Paper No.4 in the Appendix) do cost consider-
ably more, but an extra "one or two dollars" is trivial when the product in the bag
is worth $100,000 or its latent failure might cause a $10,000,000 catastrophe.
Being penny-wise and pound-foolish in this case is not only risky but may be
unethical where health or defense is involved. Note the suppression of data men-
tioned under "Discharge Test" in Chapter 4. Also, MIL-HDBK-773 (Ref. 6-4)
insists on foil bags. However, the opaque metallization in the new MIL-B-8170SC,
Type I, is not foil but is very effective (Paper No. 15).
But we're tired of being a voice crying in the wilderness. Let the world go its
way and make believe that a wispy 100 angstroms of metal, which is not an EMI/RFI
shield and, if unprotected, is easily evaporated by a spark from a fmger, is a "Fara-
day cage"!
For recommended Faraday-cage tote box constructions, see Paper No. 7 in the
Appendix; the shielding layer in these is foil as a minimum.
Here's a final word on corrosion and contamination (the subject of Paper No. 10
in the Appendix). We advise running an accelerated special test with your ESDS
item inside of, or against the surface of, the proposed intimate wrap, as a test was
done with a connector and the shunt bar proposed for it in Example 14 of Chapter
8. Generally suitable high-humidity conditions are MIL-STD-202, Method 106,
with or without the vibration or freeze cycles as appropriate, for one month. This
rigorous accelerated test probably will disclose long-term problems, but we sug-
gest an accompanying real-time test at slightly elevated temperature, e.g., lOO°F,
with samples being checked every few months. The results of the real-time test
will give you a good idea of what is happening to your product in storage. Where
corrosion is concerned, "never assume anything."
For suppliers of ESD-control packaging materials, see buyers' guides such as
the one from Evaluation Engineering magazine, Nokomis, FL.
Chapter 7
50
THE STATIC-SAFE WORKSTATION (SSW) 51
that this cost-effective approach is feasible. The operator need only remember that
ESDS items must be "kept at arm's length" and not be unnecessarily brought near
hair, clothing, or chair upholstery.
Similarly, walls and floors with fields harmless by the CD Rule may be inside
the SSW. In fact, the SSW should be considerably larger that the SSZ so that
sources of fields near the SSZ are controlled. Consider Example 29 in Chapter 9.
In this case, the boundary of the SSZ was 18 inches from a chain link fence which
provided shielding from fields on passers-by. Should the boundary of the SSW be
the same as that of the SSZ, or should the SSW extend all the way to the fence? It
should extend to the fence so that the 18-inch "buffer zone" along the SSZ is
subject to SSW requirements and new field sources won't be added. If the SSW
ended 18 inches from the fence, common polyethylene packaging film carrying a
high apparent charge, e.g., -10,000 V, might be "legally" heaped between the
fence and the SSW. In theory, the operator would note any violation of the CD
Rule by using his or her field meter and would remove the polyethylene, but if the
SSW had extended all the way to the fence, the plastic wouldn't have been lying
there (assuming compliance with SSW rules) in the first place.
Though the language was unclear, the old "one-meter rule" of DoD-STD-1686,
now superseded by MIL-STD-1686, seemed intended to provide such a buffer
zone. The trouble with this rule was that it was arbitrary, and one meter might not
be a safe distance, by the CD Rule, from a high apparent charge. Since we have
the quantitative CD Rule, the extent of the buffer zone between the working sur-
face and the boundary of the SSW is optional; it may be nil where a workbench
touches a wall but may be a few feet from the other end of the bench as marked by
a yellow line on the floor. In this case, common nonconductive plastic couldn't
"legally" be left lying on the floor near the bench. Without wasting expensive
space, of course, such buffer zones should be maximized.
Certification of SSWs
See Reference Document 2 to model Specification 1 in Chapter 9.
Humidification, Ionization,
Static-Limiting Floor Finish
These subjects are discussed in Chapter 4. In brief, use humidification if your
ambient air is often dry, if the expense of humidification equipment is bearable,
THE STATIC-SAFE WORKSTATION (SSW) 53
and if there are no hannful side effects such as corrosion; also, beware of false
confidence inspired in personnel. Use ionization only when there is no alterna-
tive, as sometimes happens; Examples 24 and 25 in Chapter 8 are cases where
triboelectric charging was unavoidable and only ionization could mitigate it.
Make the most of static-limiting floor finish, which is a very economical "safety
net" as seen in the cost calculation in Table 1 of Paper No.6 in the Appendix.
Materials
Allow "necessary nonconductors" as listed in paragraph 3.6.10 of Model Speci-
fication 1 in Chapter 9. Nonconductors are either ignored because their apparent
charge is low and they are safely far from the ESDS item, in accordance with the
CD Rule, or they are controlled by treating with topical antistat or by having
their fields attenuated by shielding, as by the chain link fence in Example 29 of
Chapter 8, or by voltage suppression, as by keeping hair close to the skin in Ex-
ample 2 of Chapter 8.
Allow conductive materials only when no antistatic or static-dissipative alter-
native is convenient. (Note: As of this writing, the more convenient EIA and MIL-
STD-1686 designation of "dissipative" is coming into use to replace these two
classifications with only one. This is a sensible change in nomenclature, because
the difference between antistatic and static-dissipative is arbitrary and is seldom
a useful distinction.) For categories of materials and criteria for their selection, see
Reference Documents 4 and 5 appended to Model Specification 1 in Chapter 9.
When selecting materials, remember that standard tests may not be sufficient!
Laboratory tests often give misleading results because real-life conditions are not
simulated. Hence, special tests with realistic configurations are often needed. An
example of an unrealistic test is the shielding/discharge or pulse test in EIA 541,
which ignores the realistic possibility of a spark discharge as discussed in
Chapter 6.
Another, far more serious example of unrealistic testing involves a type of aero-
space wire. The insulation on this wire has three different properties from conven-
tional wire insulations: (1) possible susceptibility to hydrolysis because it's a
condensation instead of an addition polymer, (2) unusual stiffness because the
insulation is high-modulus and must be wrapped on the conductor as tape, and (3)
high aromaticity (high carbon content in the molecule).
If realistic testing had been done before the wire was installed in airplanes and
space vehicles, the following would have happened: (1) exposure to wann water
and high-pH cleaning solutions, simulating wheel-well conditions in Navy planes,
would have caused hydrolytic breakdown, so the wire would have been restricted
from wheel-well applications, (2) vibration tests with simulated airframe assem-
blies would have shown chafing and shorting when the stiff wires rubbed together
(as opposed to no chafing with conventional wire whose soft insulation makes it
limp), so installation would have been done with this problem in mind, and (3)
54 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
flashover and arc propagation tests would have shown that an arc propagated even
in vacuum because of carbonization of the highly aromatic polymer, and care would
have been taken to prevent shorting that might start an arc, or the wire would have
been excluded from use in applications where the current was sufficient to sustain
an arc.
What really happened was that applications went forward based on standard,
unrealistic lab tests. For example, laboratory abrasion tests gave excellent re-
sults because they did not include the chafing situation where stiff wires rubbed
together, and standard tests for electrical properties failed to predict the arcing
problem. Thus untold millions of feet of the wire have been installed, and all
three problems described above have reared their ugly heads in the course of a
few decades (disasters can be slow in coming). The wire is now "in big trouble"
because realistic tests weren't run long ago. But enough. A word to the wise is
sufficient.
Equipment
Model Specification 1 in Chapter 9 requires grounding of all conductors, which
is not always easily done. For example, a stainless-steel oven shelf is grounded
directly, but the blade of a screwdriver is grounded to the operator's hand only
through the "sweat layer" of topical antistat applied to the nonconductive handle.
Microscopes present a special problem requiring judgment in bending the ground-
ing rules; metal eyepiece housings, for example, have small capacitance and are
not expected to touch an ESDS item, so they can be ungrounded. However, we
recommend grounding the major sections of the microscope even if wires must be
added to link them. Use the Zapflash to check continuity.
Soldering iron tips must be grounded so that there is a potential of no greater
than 2 mV between the hot tip and the grounded ESDS item being soldered, if
MIL-STD-2000 is being followed. Also, that standard requires less than 5.0 ohms
resistance between tip and workstation ground.
Electrical equipment must be checked for E fields, of course. Remember that
space charging can be a problem with ionizers, as discussed under "Ionization" in
Chapter 4, and don't let ionizers become part of the ESD problem they're sup-
posed to solve.
Treat plastic parts of equipment, e.g., housings, with topical antistat unless the
plastic is a coating that is relatively thin (less than 20 mils) and the metal is thick
enough (over 60 mils) to provide effective voltage suppression (paragraph 3.6.10(2)
in Model Specification 1 in Chapter 9). Beware of stresscracking polycarbonate
(Paper No.9 in the Appendix).
In special cases, beware of H fields (magnetic fields). MIL-STD-2000 requires
that H fields of soldering irons, soldering machines, and associated processing
equipment be less than 2 gauss measured at any surface of an item being
processed.
THE STA1lC-8AFE WORKSTA1l0N (SSW) 55
Operator Disciplines,
Training, and Certification
In general, operators must maintain the SSZ and always know its boundaries within
the SSW. ESDS items should never be brought unnecessarily near or touched to
any surface, nor should an operator fidget or shuffle his or her feet unnecessarily.
In other words, though the best materials and equipment have been selected and
the ARTG is being met by using a continuous wrist-strap monitor, don't look
for trouble. See the "Conclusions" section of Paper No.3 in the Appendix. Opera-
tor disciplines are included as requirements in Model Specification 1, paragraph
3.6.12, in Chapter 9, and training and certification of operators are covered in
Reference Document 6 appended to that specification.
Chapter 8
ESD Troubleshooting:
Illustrative Examples
1. Ann Hair
Problem: Reference 8-1 suggested that arm hair might be an ESD hazard, with
charges up to 900 V having been measured. We suspected that such voltages are
atypical. What new operator disciplines, if any, should be introduced to control
fields from arm hair?
Solution: The approach was a special test. Polyester garment sleeves or sheets
of common nonconductive polyethylene were stroked across the upper surface of
operators' forearms, and the apparent charge was read. Because of voltage sup-
pression by the skin, in no case did the reading exceed 300 V, which represents a
harmless field for ESDS items sensitive to 100 V (Fig. 3 of Paper No.3 in the
Appendix). Therefore, no new operator disciplines were introduced. However,
operators were warned to keep the potential hazard in mind and run checks with
their field meters on especially dry days. Frequent field-meter checks of the whole
SSW, to enforce the CD Rule, are of course a vital part of our program.
56
ESD TROUBLESHOOTING: ILLUSTRATIVE EXAMPLES 57
2. Head Hair
Problem: Apparent charges of up to + 10,000 V were observed on the hair of
female operators; by the CD Rule, the safe distance from 10,000 V is 56 inches
(paragraph 3.6.11 of Model Specification 1 in Chapter 9). What new operator
disciplines should be introduced?
Solution: The approach was a special test. Apparent charges of up to + 10,000
V were found with hair hanging 6 to 8 inches from the skin, but when the hair was
2 inches from the skin, the maximum apparent charge was only +300 V, which is
a harmless level as noted in Example 1 above, because of voltage suppression.
Therefore, a new operator discipline was introduced: hair must be kept within 2
inches of the skin by tying back as required; this requirement appears in paragraph
3.6.12.6 of Model Specification 1 in Chapter 9.
3. Operator Clothing
Problem: What precautions should be taken regarding the static charges on cloth-
ing? Should special antistatic or other type of ESD-control smocks be required?
Solution: The approach was a special test. Apparent charges were measured
on shirts, blouses, and sleeves being worn by operators, and a few months' worth
of data showed that fields were sufficiently weak in the vicinity of the operator's
hands, where the ESDS item was, to meet the CD Rule as long as sleeves were
short or rolled to above the elbow. Therefore, street clothes are allowed if the
sleeves are short or rolled; with long, unrolled sleeves, ESD-protective smocks are
required (paragraph 3.6.12.10 of Model Specification 1 in Chapter 9). Smocks,
though optional, may raise ESD awareness by being "uniforms" and give factories
a more businesslike appearance. However, these marginal advantages must be
weighed against the initial cost of the smocks as well as the cleaning costs, which
may be considerable. A compromise is to use rolled-up sleeves in engineering and
test labs and smocks in assembly areas.
4. Identification Badges
Problem: Apparent charges of up to -2000 V were observed on plastic identi-
fication badges (Table 4 of Paper No.3 in the Appendix). What precautions should
betaken?
Solution: The apparent charges of up to -2000 V were observed when the
badge was held away from a surface; in fact, by triboelectric charging of badges
with polycholorotrifluoroethylene (Aclar, by Allied-Signal), the apparent charge
could be pushed as high as + 10,000 V. However, in special tests, when the charged
badge was held close to the skin, as a badge is worn in practice, the apparent
charge dropped to 500 V or less because of voltage suppression. The CD Rule
allows a distance of 12 inches from 500 V (paragraph 3.6.11 of Model Specifica-
tion 1 in Chapter 9), and ESDS items are seldom brought much closer than this to
58 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
a badge. Furthermore, badges are not highly charged in practice because they're
not rubbed as in our worst-case tests. The conclusion is that badges are not a prac-
tical hazard, but if there is any doubt, they can be easily treated with a topical
antistat.
5. Vinyl Chair
Problem: An apparent charge of +500 V was noted on a vinyl-upholstered chair
(Table of Paper No.3 in the Appendix), and higher voltages could be produced in
contrived triboelectric charging tests. Therefore, should ordinary chairs be al-
lowed in the SSW, or are conductive chairs necessary?
Solution: The approach was analysis. First, the operator's sitting in the chair
attenuates the chair's field by voltage suppression. Second, the bench top pro-
vides partial shielding. Third, field-meter checks in assembly areas have always
confirmed compliance with the CD Rule in regard to chairs. The conclusion is that
ordinary chairs are acceptable (paragraph 3.6.10 of Model Specification 1 in Chapter
9), but an operator discipline requires that ESDS items shall never be held be-
yond the front edge of the work surface (paragraph 3.6.12.2 of the same specifica-
tion).
7. Charged Walls
Problem: Do charged walls present a field hazard when they bound an SSW,
e.g., when a workbench is against a wall?
Solution: The approach was a special test. Triboelectric charging of walls
with garment fabric or Aclar film gave less that 1000 V apparent charge because
fields tend to collapse on large surfaces (voltage suppression). At 1000 V, the safe
distance in inches by the CD Rule =(1000)0.5/1.8 = 18 (or use the convenient table
in paragraph 3.6.11 of Model Specification 1 in Chapter 9), and in practice ESDS
items are never nearer than this to walls in our operations. Therefore, walls are
judged not a practical hazard. If they were, they would be treated with topical
ESD TROUBLESHOOTING: ILLUSTRATIVE EXAMPLES 59
antistat in accordance with paragraph 3.6.10 of the same specification; note that
the topical antistat must be renewed at least every three months.
8. Charged Windows
Problem: Under dry conditions (relative humidity approximately 20% at 70°F),
an apparent charge of 5000 V was noted on the glass of a window. Triboelectric
charging is not caused by pure air or any other gas (Ref. 4-14), so the charging
was attributed either to electrically unbalanced air or to dust carried by the warm
desert wind which also caused the low humidity. How should the hazard of charged
windows be controlled?
Solution: The approach was analysis. By the CD Rule, the safe distance from
the window in inches =(5000)°.5/1.8 =39, so SSWs were arranged so that their
boundaries were at least 39 inches from windows. Otherwise, a nonsmearing topi.
cal antistat could have been applied to the glass, with field·meter checks to make
sure the charge was sufficiently lowered and remained in control.
touched the dust cover; then another five survived with the cover removed. When
the dust cover was replaced by aluminum foil wrapping the connector face but the
cable was unshielded, a MOSFET was damaged by the charged finger being
touched to the foil.
These tests show that ESD protection of the connector face was irrelevant. The
problem was that the field from the charged person's hand was being "picked up"
by the unshielded wire to the MOSFET so that this device was damaged by the
FIM; the dust cover could be charged to 10,000 V by triboelectric charging at
low relative humidity (10-15% at 70°F), so it itself might have caused damage by
the FIM with the cable unshielded.
The conclusion was that, when the cable was shielded, the dust cover could be
any material, even ordinary nonconductive plastic. However, good ESD-control
practice demands that an antistatic or conductive plastic cover, either of which is
commercially available, be used. (Note: A static-dissipative cover also would be
fine; this is an example of the uselessness of distinguishing between antistatic and
static-dissipative, which are suitably lumped together as "dissipative" in the cur-
rent revision of MIL-HDBK-263.) The most important result of the test was the
demonstration of the need for shielding the wire bundle of the cable.
Note that these conclusions pertain only to the particular connector-cable con-
figuration tested. For example, if a sensitive pin was less deeply recessed and was
closer to the dust cover, a charged finger might "zap" the pin through antistatic or
even conductive plastic; then a true Faraday-cage cover (buried foil or better)
would be needed.
11. Brushes
Problem: The suitability of four types of brushes for use in scrubbing ESDS
modules in SSWs was to be determined.
Solution: The approach was analysis, special test, and standard test (for sur-
face resistivity). The first brush had a varnished wooden handle, a phenolic band
(seemingly wood-flour filled) holding the bristles, and natural (horsehair) bristles.
At 65% relative humidity and 71 OF, the surface resistivity of the handle was 2 x
1012 ohms/square at 500 V and of the phenolic band was 4 x 10 10 ohms/square. The
bristles, being natural rather than synthetic, were approved by analysis in accor-
dance with MIL-HDBK-263. The conclusion was that the only part of the brush
that was a problem was the handle, which had to be treated with topical antistat
and so labeled (as an insulated tool handle, paragraph 3.6.10(9) of Model Specifi-
cation 1 in Chapter 9).
The second brush had an unfinished wooden handle and a metal band holding
the horsehair bristles, while the third brush was similar but with hog bristles. The
wood was found to be antistatic, and the bristles were approved by analysis as for
the first brush.
The fourth brush was a novel design created to solve a problem. The steel handle
of an acid brush was scraping module surfaces, so a one-inch section of the
ESD TROUBLESHOOTING: ILLUSTRATIVE EXAMPLES 61
nonconductive polyolefin heat-shrink tubing was shrunk over the handle to cover
the first eighth-inch of bristles (natural) and prevent the steel from touching the
surface being scrubbed. But would the polyolefin hold an unacceptable field? Tri-
boelectric charging of the detached polyolefin gave -390 V apparent charge,
which would require a safe distance in inches, by the CD Rule, of (390)0.5/1.8 =11
between the polyolefin and an ESDS item. However, when the polyolefin was
shrunk in place onto the steel handle, which was held by a grounded person, the
apparent charge was reduced by voltage suppression to -110 V, which is accept-
able on nonconductors (Fig. 3 of Paper No. 3 in the Appendix). The conclusion
was that the brush was acceptable for use without antistat-treating the polyolefin,
which in essence falls under the plastic-coated metals provision of paragraph
3.6.10(2) of Model Specification 1 in Chapter 9.
13. Smocks
Problem: Were three different designs of antistatic or static-dissipative smocks
suitable for use inside SSWs?
Solution: The approach was a special test. The first smock, containing 1%
stainless-steel fibers, was tested by clamping the cloth between two sets of vise-
like electrodes; at 1~1000 V, the surface resistivity gauged by a megohmmeter
was 2-4 X 1010 ohms/square. Alternatively, the resistance was measured between
dammed-in pools of mercury to give 2 X 1010 ohms/square at 10 V.
The second smock had black conductive threads in a crisscross grid and mea-
sured 6 X lOS ohms/square (static-dissipative) by the clamp method. A third smock
had black conductive threads 4 mm apart and running all the same way so that the
result of the clamp test was 8 X 107 ohms/square with the black threads perpen-
dicular to the electrodes but greater than 1013 ohms/square with the black threads
parallel to the electrodes.
The conclusion was that the first two smocks were suitable in terms of surface
resistivity, but the third smock was doubtful. Note that the use of smocks in general
62 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
was judged to be optional in Example 3 above. Before the first design could be
used, the possibility of its shedding stainless-steel fibers would have to be checked,
because these might cause shorts in ESDS items, e.g., between adjacent circuit
lines on printed wiring boards. GIDEP Alert D5-A-84-0l (April 25, 1984) de-
scribes just such an occurrence.
possibly be transferred, and (4) they should not be needlessly rubbed on any surface
because the less they are triboelectrically charged, the better, even though the voltage
is suppressed. Furthermore, it must be realized that charges on items being handled
cannot drain off onto the fingers when these nonconductive gloves are being worn.
18. AirGuns
Problem: Were two brands of nuclear ionized air guns equivalent in performance?
What was a safe distance, in terms of the field, from these guns and from two
brands of guns of the electrical type?
64 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Solution: The approach was to use special tests. A 6 x 8 inch piece of FR-4
epoxy-glass circuit-board laminate was given an apparent charge of -5000 V by
triboelectric charging and blown with the competitive nuclear air guns at 3 feet.
After 1 second, the residual apparent charge was -4000 V in both cases. After 2
seconds, one was -1500 V, and the other was -1000 V. After 3 seconds, both were
-500 V. The conclusion was that the brands were equivalent. Note, however, that
certain polonium-210 ionizers were recalled in February 1988 because of an in-
stance of escape of microspheres in which the radioactive isotope was encapsu-
lated (Ref. 4-12). A brand of nuclear air gun using a different method of
encapsulating the polonium-21O was still allowed at that time.
The field from nuclear ionizers is very low (Ref. 4-15 indicates charging of
isolated conductors to less than 50 V), and a safe distance of 1 inch from the
nozzle was established in special tests using MOSFETs. However, electrical ion-
ized air guns caused MOSFET damage at 1 inch from the nozzle; the safe distance
was established as 2 inches for one model and 3 inches for the other (Table 5 of
Paper No.3 in the Appendix).
1. The subject heat gun was and ESD hazard for MOSFETs sensitive to
100V.
2. Positive ions in the airstream were involved in charging the surface being
heated.
3. Worst-case laboratory tests should be run with the specific ESDS devices
installed in hardware on which the heat gun might be used. For example, the gun
might prove safe for 1000-V-sensitive devices. However, a surer solution to the
problem would be to replace the gun by a different heating tool, such as a Glo-
Ring, which depends on radiant heating rather than blowing hot air and is said to
be "ideal for many applications including heat-shrinking."
66 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
3 feet was too great a distance to be practical, so the pulsed DC ionizer was put out
of service; all units were collected and discarded.
GIDEPAlert E9-A-86-011 stated that pulsed DC ionizers did not protect ESDS
devices, but later testing is said to have shown no ion imbalance at 6 inches with
the ionizer operating at 5 Hz (cycles/second) and an imbalance of only 60 V at 2
Hz; no MOSFETs (Texas Instruments 3N163) were damaged (Ref. 8-4). How-
ever, MOSFETs were damaged at 0.4 Hz (22 cycles/minute) as stated in the GIDEP
Alert and found in our tests, above. It would seem that the frequency must be
about 5 Hz, minimum. We advise the potential buyer of pulsed DC equipment to
satisfy himself, preferably with his own tests, that isolated conductors won't be
charged to cause damage by DI.
similar result was found for FFB damage (see the complete version of Paper No.8
condensed in the Appendix); here, again, no grounding was needed for MOSFET
damage, but the damaging charge came from a field rather than from DI.
10,000/500 = 20 times. If the voltage was 10,oooVon passers-by, the safe distance
from the fence by the CD Rule in inches would be (500)0.5/1.8 = 12. If the voltage
was 20,000 V on passers-by, which is considered a worst-case maximum, the safe
distance in inches would be (1000)°·5/1.8 = 18. In conclusion, SSZs were main-
tained at 18 inches, minimum, from the fence.
Next, the MOSFET board in the plain polyethylene bag was placed in a paper
envelope and mailed 5 times across a corporate facility, involving sorting in a
mailroom. None of the MOSFETs was damaged (°/5 damage).
The conclusion is that ESD damage is not inevitable even with very sensitive
devices in a bag with minimal shielding capability, but even the remote possibility
of damage of high-reliability parts is reason enough to use an SSP.
Problem: How could the ESD safety of a complex automated process, with vari-
ous hazards involving several ESD damage mechanisms (DI, FIM, CDM, FFB),
be assured with high confidence?
Solution: The approach was a combination of analysis and special tests. Pa-
per No. 8 in the Appendix outlines how the problem was solved by the coupon
approach, which is uniquely applicable to automation and gives very high confi-
dence in ESD safety because of the absence of unpredictable operator error. Mis-
handling cannot occur unless the process goes awry, as would be detected in periodic
coupon checks.
Note that the coupon procedure has been made part of Model Specification 1
(paragraph 3.9) in Chapter 9. This is another example of the way in which specifi-
cation grow by absorbing the solutions found by troubleshooting. In terms of this
book, an expanding Chapter 8 continues to feed into Chapter 9.
Solution: Two conductive-fiber carpets were evaluated by the walk test (Paper
No. 15). Carpet 2 in Table 1 of this paper utilized the same fibers as the cordless
wrist strap in the above example. Conductive fibers may function by drawing the
charge from each tiny nonconductive shoe sole area contacted by a fiber at the
moment of separation of sole from carpet. Though not as effective as static-limit-
ing floor finish, conductive carpets may make the difference between, say, 8 kV
on a walking person who touches a conductor and throws a spark and 2 k V so that
he throws no spark, thus preventing the upset (Ref. 8-5) of equipment such as
computers in carpeted rooms.
Gas o 2 6 12
Air Yes Yes Yes No
Nitrogen Yes Yes (barely) No No
In further testing with nitrogen, a safety nozzle with a hand lever gave too long
a path from the cartridge for effective neutralization. The conclusion was that a
half-inch length of rubber tubing, with tip serrated to brush gently against the
work, should be attached directly to the cartridge for nitrogen use.
normal relative humidities. In contrast. the conductive tweezers had very low sur-
face resistivUy (5-6 ohms/square) at 0-93% relative humidity and were definitely
a CDM hazard. 1be antistatic tweezers were chosen for handling ESDS dice.
megohmmeter while the high (positive) lead was connected to an NFPA 99 (or
ASTM F 150) electrode on the tile layout (l ft 2 tiles) shown in Fig. 8-1. The tiles
were electrically connected underneath. Results, averaged for 10 subjects, follow.
The new footwear was more effective than heel-grounders on a soiled surface,
presumably because of the greater contact area (the entire sole) of the new foot-
wear. In conclusion, the new footwear was judged to provide superior grounding.
It also was easy to put on and take off and reasonably comfortable to wear.
or extractable) antistat. A second polymeric phase dispersed through the vinyl ma-
trix provides weak conductivity through the bulk of the material as well as on the
surface and cannot be extracted by common solvents. Also, in a special test, this
material did not make a nonconductive surface undesirably conductive by con-
tact (Paper No. 20) so that electrical leakage could occur; this is important for
packaged items such as resistors.
The same technology with polymeric antistatic additives has been applied to
other plastics, e.g., nylon (Example 43, above), ABS, and polypropylene, and is a
breakthrough especially applicable to plastic packaging (Paper No. 20).
Chapter 9
Model Specifications
Introduction
These model specifications are not claimed to be "perfect" or all-inclusive. Their
purpose is to illustrate and codify the elements of our approach, e.g., the SSZ
concept and the CD Rule, in specification fonnat. Here the threads of our preced-
ing A-to-Z discussions are pulled together in a system of required procedures.
Note how the results of special studies (troubleshooting) have been incorporated
as pointed out in the last chapter.
Specifications may be written as a list of general requirements without "how-to"
instructions or detailed examples, but the problem with this method is that the
reader is puzzled by sterile, though correct, generalizations. Alternatively, specific
materials, procedures, etc., can be given, but the problem here is lack of flexibility
for meeting general requirements in new, creative ways. Our specifications are
compromises; we give a general requirement and then illustrate it with specifics.
Thus the intent is stated along with methods-but not the only methods-to achieve
it.
Model Specification 1 concerns the SSW and gives rules for maintaining an
SSZ within it using the various "tools" of Chapter 4. Similarly, Model Specifica-
tion 2 utilizes these tools for the SSP. You should be able to adapt these specifica-
tions to your own use by appropriately modifying and expanding them. Use them
freely; see the copyright page.
1. Scope
This specification established requirements for in-plant handling aspects of an ESD-
control program as required by MIL-STD-1686 and contains detailed requirements
n
78 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
which are called for but not given in that document. The purpose of the ESD-control
program is to protect electrical and electronic items (including components, as-
semblies, and units as defined in 5.1) that are susceptible to damage or degradation
by ESD. The ESD-control program does not apply to explosives or other hazard-
ous materials.
2. Applicable Documents
The following documents, of the latest issue in effect except as otherwise indi-
cated, form a part of this specification to the extent specified herein. In the event of
conflict between documents referenced herein and the contents of this specifica-
tion, the contents of this specification shall govern.
SPECIFICATIONS
Military
MIL-M-38510 Microcircuits, General Specification
for
STANDARDS
Military
MIL-STD-120 Marking for Shipment and Storage
MIL-STD-1686 Electrostatic Discharge Control
Program for Protection of
Electrical and Electronic Parts,
Assemblies, and Equipment
(Excluding Electrically Initiated
Explosive Devices)
Department of Defense Soldering Technology, High Quality/
DoD-STD-2000-lB* High Reliability
OTHER PUBLICATIONS
MIL-HDBK-263 Electrostatic Discharge Control
Handbook for Protection of
Electrical and Electronic Parts,
Assemblies, and Equipment
(Excluding Electrically Initiated
Explosive Devices)
NAVSEA SE 003-AA-TRN-OlO Electrostatic Discharge
Training Manual
SPECIFICATIONS
Model Specification 2 Packaging of ESDS Items for
Shjpment, Storage, and
In-Plant Transfer
OTHER PUBLICATIONS·
Reference Document 1 General Design, Construction,
and Maintenance Guidelines
for ESO-Protected Areas
Reference Document 2 Certification of Static-Safe
Workstations (SSWs)
Reference Oocument 3 Procurement Requirements for
Continuous Wrist-Strap
Monitor Unit
Reference Document 4 Approved ESO-Control Materials
and Equipment
Reference Document 5 Criteria for Selection of Approved
ESO-Control Materials and
Equipment
Reference Document 6 Certification of Personnel for
ESOControl
3. Requirements
3.3. Design protection. The need for design protection shall be determined in
accordance with MIL-HOBK-263 and NAVSEA SE 003-AA-TRN-OIO.
"'Reference Documents 1-6 are made independent of the specification so that they can be revised
separately. This is especially important for Reference Document 4, which is a "living document" that is
always growing and changing.
80 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
3.5. Static-safe zone (SSZ). While being protected (3.4), ESDS items shall
never leave an SSZ, which is defined as a volume in space at every point of which
the Basic Rule (3.5.1) is followed.
3.5.1. Basic Rule. An ESDS item shall never, even for a nanosecond, be ex-
posed to an E field in violation of the CD Rule (3.6.11) or have its sensitive
leads or terminals touched to, or receive a discharge from, any surface at
more than 50 V.
3.5.2. Location ofSSZ. The SSZ may be (1) inside a static-safe package (SSP)
as defined by Model Specification 2, (2) within a static-safe workstation (SSW) as
defined in 3.6.1, or (3) around the ESDS item while this item is being relocated
from one SSW to another under the continuous supervision of ESD-certified per-
sonnel (3.6.13). An ESDS item shall be removed from an SSP only within the
SSZ of an SSW.
Distribution wire
Other electrical
equipment
don't buy these monitors, or any other major ESD-control equipment, without a
trial in your own assembly area.)
3.6.5.3.4. Ground/ault circuit interruptors (GFCls). GFCIs shall be installed
when voltages above 225 V are being used, so that personnel will be protected
even if electrically overstressed resistors should be carbonized so that they allow a
current of more than I mA.
3.6.5.3.5. Conductive floor mats. Conductive floor mats or other methods of
grounding personnel other than monitored wrist straps (3.6.5.3.3) shall be used
only when necessary. An example is the grounding of mobile personnel when
wrist straps even with long cords or connected to a sliding trolley overhead are
impractical. Floor mats shall be connected to the static-ground distribution wire
(3.6.5.2.3) through a I-megohm resistor, or they may be connected directly to the
grounding lug of the work surface. Conductive shoes or shoe straps with a mini-
mum resistance of I megohm shall be used in conjunction with conductive floor
mats.
3.6.6. Materials and equipment.
3.6.6.1. Materials. ESD-protective materials used in accordance with this speci-
fication shall have a surface resistivity of less than 1012 ohms/square when mea-
sured at 70°F and a maximum of 50% relative humidity with a Voyager SRM-II 0
MODEL SPECIFICATIONS 83
As necessary, these shall be treated with topical antistat to make them anti-
static so that their field will not violate the CD Rule (3.6.11). If corrosion is feared,
a chloride-free topical antistat may be selected. In the above list, Nos. 1 and 2
will not need antistat treatment because of voltage suppression, No. 3 will be
used without antistat treatment but subject to the conditions of 3.6.12.11, Nos. 4-8
probably will not need antistat treatment (for badges, see Examples 4 in Chapter
8), and Nos. 9 and 10 shall be treated with topical antistat and so labeled. The
treatment shall be renewed every three months (Note: or more often, as indicated
by experience), and its effectiveness shall be monitored by field-meter checks.
3.6.11. CD Rule. The CD Rule shall never be violated, the requirement being
that ESDS items must be kept at least a safe distance d from a surface with appar-
ent charge V in volts according to the equation: d =yo.s/l.8. The apparent charge
is the charge in volts of a surface or object as read by a hand-held field meter
according to the meter manufacturer's instructions; this is an uncorrected reading
and represents the field strength at the point in space where the meter is held.
(Note: See the discussion of the CD Rule in Chapter 4.) For convenience, some
solutions of the equation are:
300 to 6,000 43
500 12 8,000 50
1,000 18 to,OOO 56
2,000 25 15,000 68
3,000 30 20,000 79
4,000 35 30,000 96
SSZ will end at an operator's wrists, at least 12 inches from the points of an ion-
ized air blower or 3 inches from a nozzle (3.6.8), and at distance d from any charged
surface according to the CD Rule (3.6.11).
3.6.12.2. Parts handling. The operator shall not unnecessarily touch ESDS
leads, terminals, or contacts even though he or she is grounded. Parts shall be
handled by their cases as size permits. An ESDS item shall never be held beyond
the front edge of the work surface because of possible fields from the chair or
operator's lower garments.
3.6.12.3. Touching 0/ surfaces. The operator shall not allow ESDS items to
touch surfaces or objects other than the approved work surface, approved materi-
als and equipment, or the assembly to which the item is being attached. For ex-
ample, clothing shall not be touched.
3.6.12.4. Triboelectric charging by body movement. The operator shall not
fidget, squirm, or shuffle his or her feet unnecessarily. The object of this discipline
is to minimize triboelectric charging of the operator's body or apparel.
3.6.12.5. Voltage surges by capacitance drop. The operator shall not touch
ESDS leads, terminals, or contacts while arising from a chair or lifting his or her
feet. This prohibition protects ESDS items from surges in V when C drops in the
equation Q = CV.
3.6.12.6. Hair. Hair shall be tied back as necessary so that it is within 2 inches
from the skin in order to limit the apparent charge by voltage suppression. (Note:
See Example 2 in Chapter 8. This requirement on hair is an example of the incor-
poration of the results of troubleshooting studies into the specification.)
3.6.12.7. Shorting clips and protective containers. The operator shall not re-
move shorting clips or protective containers until ready to use an ESDS device.
3.6.12.8. Exclusion o/nonconductors. Static-generating nonconductors, e.g.,
ordinary, untreated tote boxes, plastic envelopes or notebooks (for work instruc-
tions, etc.), glazed paper, or ordinary plastic packaging material, shall be excluded
from the SSW. Only necessary nonconductors (3.6.10) shall be allowed.
3.6.12.9. Cleanliness. A clean and orderly SSW shall be maintained. Lunches,
coffee cups, candy wrappers, etc., shall be excluded; many of these materials are
unnecessary nonconductors.
3.6.12.10. Wearing apparel. Operators shall wear (1) close-fitting blouses or
shirts with short sleeves (above the elbow), (2) close-fitting blouses or shirts with
sleeves rolled up above the elbow, or (3) ESD-protective smocks. (Note: In assem-
bly areas, these smocks are an attractive "uniform" that raises ESD-awareness and
may impress visitors, but periodic cleaning and replacement of the smocks may
prove relatively expensive.)
3.6.12.11. Gloves and finger cots. Cotton or other antistatic gloves and finger
cots shall be worn, with the exception that nonconductive gloves or cots, e.g.,
latex gloves for cleanroom work where antistatic additives might rub off to cause
contamination, may be used when they are tightly fitting and are judged to be
necessary nonconductors (3.6.10(3». Of course, nonconductive gloves shall not
be put on or removed at the SSW in such a way that their field, uncontrolled by
86 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
voltage suppression as it is when they are being worn, violates the CD Rule
(3.6.11). In general, nonconductive latex gloves shall be put on or removed at
least 5 feet from the boundary of the SSW. (Note: For precautions on using these
gloves, see Example 15 in Chapter 8.)
3.6.12.12. Working near chain link/ences. The SSZ (3.5) shall be at least 18
inches from a chain link fence bounding the SSW when people in ordinary clothes
may pass near the other side of the fence. (Note: This paragraph would not be
included in most specifications. It's an example of a special provision based on an
unusual situation, namely, the presence of fences in assembly areas; see Example
29 in Chapter 8 and the real need for a fence provision in Chapter 11.)
3.6.12.13. Transient personnel. Operators shall ensure that transient person-
nel, e.g., supervisors or visitors, do not enter SSWs unless such personnel are
grounded and closely supervised by the operator.
3.6.13. Certification and training 0/personnel. Operators and other personnel
who specify, acquire, design, assemble, inspect, test, package, repair, rework, in-
stall, or maintain ESDS items shall be certified in accordance with Reference
Document 6 by completing a training course in ESD control (at least 8 hours in-
struction and one test), then recertified annually (at least 2 hours brush-up instruc-
tion and one test). Records of certification and recertification shall be maintained.
(Note: Operator training should correspond to the specification; for example, if
this model specification were used, our book with appropriate explanations and
enlargements would serve as the text or basis of course material. The need for
consistency of specification and training material seems obvious, but in practice
there may be contradictions because the ESD-control program was developed piece-
meal and the training course and the evolving specification went in different direc-
tions. Another point we wish to emphasize is that operators should not be kept
ignorant by a management jealously guarding facts because "knowledge is power."
Rather, in regard to ESD control, the operator should be a "senior employee" able
to act independently and intelligently like an airplane pilot or police officer.)
3.6.14. Testing.
3.6.14.1 Dielectric strength or insulation resistance tests. Equipment contain-
ing ESDS items shall not be subjected to dielectric strength or insulation resis-
tance tests.
3.6.14.2. Applied voltage. ESDS items shall not be subjected to applied volt-
age, e.g., from an ohmmeter, at an energy level that could create a damaging
current.
3.6.14.3. Removal from test sockets. ESDS items shall not be removed from
test sockets with power applied.
3.6.14.4. Power to connectors. All power shall be turned off to the connectors
in a system before printed circuit boards or other assemblies containing ESDS
items are inserted into or removed from the connectors.
3.6.14.5. Proper polarity o/voltage. Test equipment setups shall be arranged
for proper polarity of voltage before conducting parametric or functional testing.
3.6.14.6. Supply voltages and signal voltages. Direct-current (DC) supply
MODEL SPECIFICATIONS 87
voltages shall always be applied before signal voltages except for memory subas-
semblies. Signal voltages shall always be removed before DC supply voltages
unless the function removes signal voltages simultaneously with DC supply
voltages.
3.6.15. Installation site. The following minimum installation procedures shall
be enforced:
a. Protective coverings shall be maintained intact when ESDS units are to be stored
prior to installation.
b. Protective caps, shunts, etc., on unit receptacles shall not be removed during
unit installation until the completion of required interconnecting cable and
connector assemblies.
c. Connector and receptacle surfaces shall be momentarily contacted to ground in
order to discharge any electrostatic potentials on the cable or installer's body
just prior to engaging a cable connector with a mating receptacle and prior to
removing protective covering.
CAUTION
OBSERVE PRECAUTIONS
FOR HANDLING
ELECTROSTATIC
SENSITIVE
DEVICES
3.7. Audits and reviews. Audits and reviews of the ESD-control program shall
be conducted in accordance with contractual requirements. (Note: A thorough au-
dit can be a disturbing revelation. The ARTG may be found to be exceeded in a
majority of cases if continuous wrist-strap monitors, as required in this specifi-
cation, are not used, and many SSWs may be found to have faulty grounding. An
88 ESD FROM A TO Z: ELECTROSTAnC DISCHARGE CONTROL FOR ELECTRONICS
adverse audit could result in much assembled hardware being considered suspect
and perhaps rejected.)
5. Notes
will bloom to the surface and regenerate the sweat layer a few times until the
internal reservoir of antistat is exhausted. The material must be cleaned only with
an antis tat solution, without drying; this procedure leaves an antistatic film on the
surface. Other products such as permanently antistatic or static-dissipative film
with a radiation-cured coating may also be solvent-sensitive and may lose their
effectiveness after a single brief exposure.
(Note: Paragraphs 5.2 and 5.3 illustrate the kind of tutorial information that may
be included in Section 5. Even though personnel have been taught these facts,
caution notes are desirable to heighten the awareness of hazards that may damage
hardware and, even more importantly, may endanger health or cause personal
injury.)
3. Housekeeping. The SSW shall be kept clean, orderly, and free of unneces-
sary nonconductors and unapproved materials and equipment.
3. Certitication procedures.
3.1. Certification decal. Check that each SSW has a current certification decal
to the requirements herein.
3.2. Electrical resistance from work surface to ground.
3.2.1. Benches with static-dissipative or antistatic work sUrfaces with a buried
conductive layer. Measure the resistance from work surface to ground with a
megohmmeter and an NFPA 99 electrode. Connect the negative ("low") lead of
the megohmmeter to the electrode and the positive ("high") lead to the grounding
lug. Place the electrode on the bench surface 18 plus or minus 2 inches from the
lug. Measure the resistance from the electrode to the lug at SOO V. The resistance
shall be between 1 X lOS and I x 1011 ohms. Alternatively, a Voyager RTG-210
Resistance to Ground meter or equivalent can be used. (When the resistance is
under about 7 megohms using the palm of the hand as the electrode, a Zapflash
will light. This is a useful quick check when the expected resistance to ground is
toward the low end of the required range. If the Zapflash doesn't light, the resis-
tance might still be acceptable, of course, by being in the middle or high end of the
required range.)
3.2.2. Benches with conductive work surfaces. Using a multimeter at 0.5-S0
V, probe between the work surface and the distribution wire (static ground). The
resistance shall be 1 megohm, provided by a current-limiting resistor as done for
static-dissipative or antistatic work surfaces. Laminar-flow booths, exhaust booths,
and similar facilities where the conductive surface is electrically common with the
structure shall have a resistance of O-S ohms, verified by probing between the sur-
face and commercial power ground. Note that conductive bench tops (work sur-
faces) are used in many cleanroom and special-process applications where the nature
of the task requires highly conductive materials such as stainless steel or aluminum.
3.2.3. Processing, storage, and test equipment. Using a multimeter at O.S-SO
V, probe between the commercial power ground connection on the chassis and
bare metal on the front (or on the inside for ovens and desiccators) of the equip-
ment. The resistance shall be O-S ohms. Verify the grounding of baffles in ovens
and environmental chambers by probing in a similar way; again, the requirement
is O-S ohms.
3.3. Potential difference between static ground and commercial power ground.
Static ground shall be used for assembly benches (3.2.1 and 3.2.2). Other SSWs
(3.2.3) may use commercial power ground. These two grounds shall not differ by
more than S V dc or rms. Continuous wrist-strap monitors use both these grounds
and will detect a gross failure in either.
94 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
3.4. Wrist-strap and ground wire connection. Check that each SSW has wrist-
strap and ground wire connections in good condition. Alligator clips and twisted
wire connections shall not be certified. All wrist straps must have continuous
wrist-strap monitors.
3.5. Resistance between the grounding lug ofthe work surface and static ground
distribution wire. Check that this resistance is 1 megohm, provided by a series
resistor in the lead wire which connects the grounding lug and the distribution
wire. Make sure that the lead wire is a minimum of 18 AWG and the distribution
wire is a minimum of 12 AWG. ("Minimum" refers to size; the higher the AWG,
the smaller the conductor.)
3.6. Resistance between static ground distribution wire and commercial power
ground. Check that this resistance is less than 25 ohms.
3. Requirements.
3.1. General requirements.
3.1.1. Overall design. The unit shall consist of a monitor (alarm device) and a
special wrist strap. The wrist strap shall consist of a two-snap wrist band and a two-
conductor coil cord for connection to the monitor. The unit shall provide two points
of skin contact and two ground connections. An alert shall be indicated in the
event of an "open" (high resistance) or "short" (low resistance) in either circuit.
3.1.2. Workmanship. Workmanship shall equal or exceed industry standard,
and fabrication shall be from good-quality components.
3.1.3. Provision for calibration. The monitor shall have provision for calibra-
tion and calibration seals.
3.1.4. Documentation. Service instructions, calibration procedures, and draw-
ings, including a schematic of the circuitry, shall be provided by the supplier.
MODEL SPECIFICATIONS 95
4. Approved Source List. Units for purchase shall be selected from among
those listed in Reference Document 4, paragraph 3.3.
96 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
(Note: This document lists materials and equipment selected according to the cri-
teria of Reference Document 5 as well as other criteria such as cost. The list gives
the name of the item, supplier's part number, supplier's location (city and state),
and the basis for approval of the item. Approval may be by standard test, special
test, or analysis, and the basis-for-approval column cites internal notebook page
references, letters, and test reports, as well as suppliers' data. See Reference Docu-
ment 5 for categories of materials and equipment listed.
This document is a "living list" which is constantly being added to and updated.
Listing approved items directly in the specification has proved impractical in our
experience because of the need for frequent revisions of the specification to keep
the list current, each revision requiring a time-consuming coordination with the
Engineering, Manufacturing, and Quality Assurance organizations.
Since ESD control is still a "young" field, novel products are always appearing.
For example, a digital desoldering apparatus meeting all the requirements of 000-
STD-2000-1B was advertised in late 1988 as featuring "the only static-dissipa-
tive desoldering handpiece on the market." Since the need for treating a plastic
handle with topical antistat would be avoided, this equipment is a good example
of a candidate worth evaluating. (Of course, we can't list commercial products in
this book, so now we'll proceed to Reference Document 5, which gives the criteria
for choosing them.)
Table o/Contents
1. Introduction
2. Publications cited
3. Materials/equipment and criteria
3.1. Bag materials and plastic films/sheets
3.2. Foam and bubble-wrap materials for cushioning
3.3. Apparel and operator grounding equipment
3.4. Tote boxes, bins trays, boxes, containers
3.5. Workbench tops and mats
3.6. Chairs, floor mats for chairs, seat covers
3.7. Floor materials (mats, tiles, carpets, static-limiting floor finish)
3.8. Static meters
MODEL SPECIFICATIONS 97
2. Publications cited.
2.1. Government documents.
SPECIFICATIONS (Military):
STANDARDS
Federal
FED-STD-l 0 1 Preservation, Packaging, and
Packing Materials: Test Procedures
98 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Military
STANDARDS
American Society for Testing and Materials
OTHER PUBLICATIONS
(Note: Many other documents could be referenced. The ones given are merely
illustrative.)
as well as an antistatic
polyethylene liner with sur-
face resistivity less than 1012
ohms/square. (MIL-B-81705B
had an unsatisfactory liner.)
*Optional items.
MODEL SPECIFICATIONS 103
*Optional items.
104 ESD FROM A TO Z: ELECTROSTAnC DISCHARGE CONTROL FOR ELECTRONICS
during triboelectric charging. Concepts like the ARTG are a little more difficult
to explain; in this case, a momentary surge of voltage must be envisioned, requir-
ing limited resistance (enough conductivity) to drain off the electrons so quickly
that they can't accumulate to the danger level.
Visual aids such as videotapes are valuable if they are tailor-made to fit the
specification. A demonstration with a curve-tracer, showing a MOSFET being
damaged by an ungrounded person's touch, is a vivid demonstration of proven
value in convincing the audience that ESD damage really happens. We remember
one videotape which showed an airplane crashing in a ball of flame because of
latent failure. This may be a bit exaggerated, but horror stories of ESD destruc-
tion of expensive assemblies (see Example 25 in Chapter 8) are documented and
cogent. No employee would want to be a guilty party in such an infraction, and no
one needs to be if he follows the rule: "If there's doubt about ESD protection in
this process, ask Engineering for advice." As in any field, the operator or engineer
on the shop floor must know the limits of his knowledge so that he won't overstep
them and risk either overt damage or the undetected "wounding" of devices that
can cause those hard-to-prove but always possible latent failures.
A suitable test is 20 questions with multiple-choice answers and a passing grade
of, say, 70%. The test should reveal whether the student understands how precau-
tions and rules flow from and implement the Basic Rule. People tend to follow
instructions when good reasons are given but to ignore "senseless" orders. See the
sample test at the end of this chapter.
We have found that decals applied to identification badges are useful in screen-
ing unqualified people from ESD-protected areas. Also, the decals raise the ESD-
awareness of their wearers and, like miniature diplomas, are a source of pride.
Thorough records of certification and recertification are required by MIL-STD-
1686 and should be kept in any event. If hardware should be mishandled, the tes-
timony of certified personnel will be more credible when their training is
documented. For example, an operator might say, "When my wrist strap pulled
loose, I was so absorbed in my work, which involved touching ESDS leads sensi-
tive to 300 V, that I ignored the alarm of the continuous monitor for a minute. But
as soon as I realized what had happened, I measured my voltage, by pointing a
field meter at the grounded work surface, and found it to be less than 100 V even
when I shuffled my feet (on static-limiting floor finish), which I hadn't been
doing during the infraction." Based on this report by an astute and certified person,
if the hardware tested O.K., it could be assumed to be safe from latent failure.
We emphasize once again that an operator is analogous to an airplane pilot who
can either "fly" or "crash" the most costly, well-equipped SSW. Since the operator's
skills are vital, he or she should have considerable autonomy and, as Dan Ander-
son says, should be "vice-president" of the SSW. An authoritarian management
system which withholds information because "knowledge is power" will not only
keep operators ignorant but make them indifferent and even hostile. In ESD con-
trol, personnel must be trusted as well as trained.
MODEL SPECIFICATIONS 107
I. Scope
This specification establishes requirements for packaging constructions which meet
the definition (3.2, below) for a static-safe package (SSP). An SSP protects the
packaged ESDS item from both fields and discharges.
2. Applicable Documents
The following documents, of the latest issue in effect except as otherwise indi-
cated, form a part of this specification to the extent specified herein. In the event of
conflict between documents referenced herein and the contents of this specifica-
tion, the contents of this specification shall govern.
Federal
RR-W-365 Wire Fabric (Insect Screening)
PPP-C-795 Cushioning Material, Packaging
(Flexible Cellular, Plastic
Film) for Packaging Applications
PPP-C-1752 Cushioning Material, Unicellular
Polyethylene Foam, Flexible
(for Packaging Purposes)
PPC-C-1797 Cushioning Material, Resilient,
Low Density, Unicellular,
Polypropylene Foam
PPP-C-1842 Cushioning Material, Plastic; Open
Cell for Packaging Applications
Military
MIL-B-131 Barrier Material, Waterproof,
Greaseproof, Flexible, Heat-Sealable
MIL-S-19491 Semiconductor Devices, Packaging of
MIL-M-38510 Microcircuits, General
Specification for
MIL-R-39032 Resistors, Packaging of
MIL-B-81705 Barrier Materials, Flexible, Electrostatic-
Free, Heat-Sealable
108 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
STANDARDS
American Society for Testing and Materials
3. Requirements
3.3. Materials. The following materials shall be used subject to the require-
ments of this specification.
3.3.1. Specified materials.
3.3.2. Other materials. The materials shall conform to the following require-
ments.
3.3.2.1. Conductive material. Noncorrosive conductive ESD-protective ma-
terial shall have a surface resistivity of less than lOS ohms/square.
110 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
3.5. Corrosivity. The interior of the package shall be noncorrosive when tested
in accordance with 4.3.6.
MODEL SPECIFICATIONS 111
3.9. Cardboard. Glazed cardboard shall not be used unless the glaze is a spe-
cial antistatic composition.
3.11. Cleanliness. Prior to packaging, ESDS items shall be clean and free
from foreign material in accordance with the applicable component or assembly
specification.
3.12. Uniformity. Containers shall be uniform in size and shape for shipment
but not necessarily for storage or in-plant transfer.
Note on symbols:
The ESD Association (Rome, NY) states that items that are
susceptible to ESD, such as ICs and assembled boards, should
use the triangle, hand, and slash symbol (Fig. a). ESD-protective
(a) products, such as static-shielding bags, should use the triangle,
hand, and arc symbol (Fig. b).
Some manufacturers use a lightningbolt symbol to indicate
ESD susceptibility, but that symbol should not be used because
it can be confused with one for electrical safety. Likewise, a
circle with three arrows entering it (Fig. c), which is often used
(b) for ESD protection, should not be used because it is generic
for electromagnetic, magnetic radioactive, or electrostatic fields.
3.1S. Other packaging requirements. (Note: Many special requirements are pos-
sible, especially for physical protection, but the purpose of this model specification
is to emphasize ESD-protection requirements. On military contracts, documents such
as MIL-STD-794, MIL-S-19491, MIL-M-3851O, and MIL-R-39032 may be cited.)
5. Notes
5.1. EMIIRFI shielding. EMI/RFI shielding is outside the scope of this model
specification. However, protection is afforded by foil or screen, e.g., 3.3.1.a, or by
3.3.1.i, or by opaque, vapor-deposited metal (MIL-B-81705C, Type I).
5.2. Magnetic shielding. Shielding from H fields is outside the scope of this
specification. Iron foil may be used.
Multiple-Choice Test
for Personnel Certification
The following twenty questions are designed to test comprehension of Model Speci-
fications 1 and 2 and a course in ESD principles. A suggested passing grade is 70%
(no more than six wrong answers). Correct answers are given at the end of the
questions.
1. An ESD event that may cause device damage is: (a) always a surge (electron
flow), which may include a spark, (b) always a spark (air discharge), (c) always a
discharge from a person's finger or a tool, (d) always the result of high voltage
(over 3 kV).
2. Damage from the Human Body Model (8BM) is: (a) prevented by ground-
ing the skin, (b) the sole cause of ESD damage when people handle devices, (c)
caused by charges residing on the devices themselves, (d) prevented by operators
wearing conductive footwear on nonconductive floors.
3. The Charged Device Model (CDM) is prevented by: (a) grounding the skin
or tools that may touch a charged device, (b) grounding plastic or ceramic cases
of devices, (c) neutralizing charges on ESDS items with ionized air, when pos-
sible, and touching the leads only with antistatic materials, (d) touching leads
only with bare fingers, never a metallic tool.
116 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
4. An antistatic material is: (a) always a flexible plastic film, (b) electrically
nonconductive so that charges can't flow on or off it, (c) a material with low
surface conductivity but high volume conductivity, (d) a CDM-safe surface but a
poor barrier to external discharges when used for packaging.
5. For our purposes, a Faraday cage is: (a) an antistatic plastic bag or box, (b)
a carbon-loaded plastic bag or box, (c) a laminate containing aluminum foil, (d) an
electrically isolated open cup used for measuring static charges.
6. A static field or E field is: (a) a force field in space, (b) measured by a field
meter pointed at a charged surface, (c) the cause of ESD damage by the FIM, (d)
all of the above.
7. Relative humidity: (a) prevents all ESD events if high enough, (b) is not a
factor in ESD events, (c) affects the performance of most antistatic materials, (d)
must be controlled in all assembly areas and warehouses.
8. Air ionization: (a) is effective when used for specific reasons, (b) is a cure-
all making skin-grounding unnecessary, (c) is effective only at high relative hu-
midity, (d) neutralizes charges on conductors but not on nonconductors.
9. Allowable resistance to ground (ARTG) pertains to: (a) tote boxes lying on
workstations, (b) operators wearing wrist straps, (c) static-dissipative bench tops,
(d) conductive surfaces such as stainless-steel oven shelves.
to. The CD Rule is: (a) applied using measurements taken with a field meter,
(b) a technique of static-safe packaging, (c) calculated on the basis of relative
humidity, (d) never violated when continuous wrist-strap monitors are worn.
12. A static-safe package (SSP): (a) uses antistatic film as a shield against
external discharges, (b) is opened only at a static-safe workstation (SSW), (c)
must be a rigid, metallic enclosure, (d) must be an opaque, foil-containing
laminate.
13. When your continuous wrist-strap monitor alarms, you must: (a) cease
work and determine the reason for the alarm, such as the need for applying anti-
static lotion to your wrist, (b) disconnect the strap to silence the alarm and finish
working, (c) ignore the alarm while you return the ESDS item to its SSP, (d)
continue work until your supervisor hears the alarm and comes to advise you.
MODEL SPECIFICATIONS 117
15. You may allow a visitor to enter your SSW if he: (a) is ungrounded but
refrains from touching ESDS items, (b) wears an antistatic smock, (c) is grounded
and closely supervised by you, (d) keeps his feet outside the SSW boundary while
he leans in to observe.
16. If your hair is long, you must: (a) tie it back so it is within 2 inches of your
skin, (b) let it hang loose but don't let it brush ESDS items, (c) brush it to the left
side of your head and work to the right, (d) let it hang loose but keep ESDS items
at least 15 inches from it.
17. An ESDS item must always be in an SSP or SSW unless: (a) personnel
remain at least 5 feet from the item, (b) the relative humidity exceeds 50%, (c) the
item is under the continuous supervision of certified personnel while being trans-
ferred between SSWs, (d) the item is lying on antistatic plastic.
18. When a plastic tool handle is found to have an apparent charge of7 kY: (a)
the tool must be excluded from the SSW, (b) the handle should be given a one-
time treatment with topical antistat, (c) the handle should be treated every three
months with topical antistat unless field-meter checks show need for more fre-
quent treatment, (d) the handle is left untreated but not allowed to touch the leads
of ESDS items.
19. Pink antistatic bags: (a) should be washed with water when soiled, (b) should
be cleaned only with an antistat solution, (c) should be excluded from the SSW at
relative humidities below 30%, (d) are examples of Faraday-cage packaging.
20. A static-dissipative mat must be grounded by: (a) an alligator clip on the
snap with a lead to a metallic water pipe or building structural steel, (b) an ap-
proved snap connection with a lead to commercial power ground, (c) an approved
snap connection with a lead wire, containing a resistor, connected to the static
ground distribution wire, (d) an approved snap connection with a lead to a plastic
water pipe.
Industry Standards
118
INDUSTRY STANDARDS 119
In recent years, the ESO Association, formerly called the EOS/ESO Associa-
tion, in Rome, New York, has created several more standards (Ref. 10-1 and 10-2).
In the list below, "s" means Standard, and "OS" means Draft Standard. The Oraft
Standards have been released to industry for comments and suggestions. "AOV"
means Advisory Program. Note that some standards are called "EOS/ESO" and
others "ESO."The "EOS/ESO" designation is to be changed to "ESO" in the course
of time when the standards are revised. In the right-hand column are Model Speci-
fication I paragraphs to which the standard might be applied. The standard marked
"N/A" is inappropriate in our view because it does not involve a spark discharge
(see "Discharge Test" in Chapter 4).
Further ESD Association standards are planned. As of this writing, the Associa-
tion is considering preliminary drafts on component handling equipment, solder-
ing equipment, volume resistance of materials, and measurement systems for
waveforms, radiation effects, and equipment effects. Also, a committee will focus
on ESD issues in cleanrooms.
ISO 9000 standards are becoming important in ESD control. They promote pro-
tection measures such as continuous wrist-strap monitoring (Ref. 10-3), which
is a mainstay of our recommended approach.
In conclusion, you should (1) review the government and industry documents
available on CD-ROM to make sure you have the latest versions and (2) contact
the ESD Association for a list of their available standards, handbooks, and advi-
sory programs. If the test methods seem practical and meaningful in terms of your
program, work them into your handling and packaging specifications.
Chapter 11
1. Scope
This document establishes X Corporations's policy and ESD-control program for
ensuring product quality by preventing damage or degradation by ESD during
handling, assembly, testing, shipment, storage, and in-plant transfer.
2. Applicability
This document pertains to all ESD-sensitive assemblies manufactured by all of X
Corporation's divisions and subsidiaries. Procedures required by this policy re-
main applicable to all ESD-sensitive components and subassemblies until the prod-
uct has been completely assembled and becomes a Faraday cage (shielded
enclosure) immune to ESD damage. All items shall be handled with the same
precautions as the most sensitive item.
122
LIVING WITHOUT MIL·STD-1686 123
4. Implementation
The ESD-control policy is implemented by the following documents:
4.1 Procedures for Handling, Assembly, and Testing of Electrostatic Discharge
(ESD)-Sensitive Items
4.2 Procedures for Packaging of Electrostatic Discharge (ESD)-Sensitive Items
for Shipment, Storage, and In-Plant Transfer
4.3 Program Bulletins for Government Contracts
Note that this policy document serves as the ESD-control program plan. Also
note that there is no Quality Assurance (QA) organization. The workers monitor
their own quality, including ESD control, as part of a Total Quality Management
(TQM) approach. Another approach, to be seen below, is to have department man-
agers play the QA role.
As examples of 4.1 and 4.2 from the above policy document, the following Case
1 and Case 2 are tailored versions of the model specifications in Chapter 9. The
tailoring was done by the authors acting as consultants to two medium-sized com-
panies (approximately 500 employees each), and so these are real-life examples.
In both cases, all notes at the ends of paragraphs in the models were deleted.
124 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Case 1
PARAGRAPH CHANGES
PARAGRAPH CHANGES
Case 2
Paragraph Changes
First sentence changed to: "This document establishes in-plant
procedures for implementing the X Corporation ESD-Control
Program Plan."
2 "Specification" changed to "document."
2.1.2.2 Deleted.
3.1 Changed to: "In the event of a conflict between the requirements of
this document and drawings calling out this document. the
requirements of the drawings shall take precedence."
3.2 Changed to: "ESDS items shall all be considered sensitive to lOOV
by the Human Body Model."
3.3 Changed to: "The need for design protection shall be determined in
collaboration with component suppliers."
3.3.\ Changed to: "The least sensitive component that will meet
performance requirements shall be used. When necessary. protective
circuitry shall be incorporated at the lowest practical level of
assembly to limit the sensitivity of that assembly."
3.3.2 Changed to: "Unit terminals. including test points. located on the
surface of the unit enclosure shall contain protective circuitry."
3.6.2 Deleted.
3.6.4 "Reference Document 2" changed to "Note 5.3."
3.6.5.2 Added: "Also. voltages surges from equipment may occur on
commercial power ground."
Figure I "Carbon paper layer" changed to "Static-dissipative mat." "Lug"
changed to "Snap."
3.6.5.3.5 Deleted.
3.6.6.1 "Surface resistivity meter" made generic. Reference to Ref. Doc. 4
and 5 deleted.
3.6.6.2. Deleted.
3.6.7
3.6.8 Changed to: "Air ionizers shall be used only when absolutely
necessary. e.g .• in peeling nonconductlve tape from rolls. Ionized
air blowers shall be electrical. but nozzles (guns) may be either
electrical or nuclear. The blowers shall have an autobalancing
feature. All Ionization ... of the ionizers."
3.6.9 Changed to: "An approved static-limiting floor finish shall be
applied to all tile floors and maintained so that appearance is
attractive. slip resistance is adequate. and the surface resistivity never
exceeds 10" ohms as measured between two NFPA 99 electrodes
1 inch apart."
3.6.10.4 Changed to: "TIle floors (carpets are unacceptable)."
3.6.10.11 New item added: "Plastic panels on equipment."
3.6.12.6 Changed to: "Hair shall not hang in front of the operator so that it
touches or closely approaches ESDS items."
3.6.12.12 Added: ..... other side of the fence or nonconductlve plastic is stored
there."
LIVING WITHOUT MIL-STD-1686 127
3.6.13 "Ref. Doc. 6" changed to "Note 5.4." "Annually" changed to "every
two years."
3.6.15 Deleted.
3.6.16.1 MIL-STD-129 deleted.
3.6.16.2 Sealing of tags in envelopes deleted.
3.7 Changed to: "Audits and reviews of the ESD-control program shall
be conducted as required."
3.9 Changed to: "Automated processes shall be periodically reviewed
for ESD safety."
3.10 Changed to: "As soon as possible after the infraction, the following
ten data items shall be recorded. Disposition of the hardware then
shall be decided by a Material Review Board." (The ten items are
listed from Chapter 13.)
4.1 Changed to: "All personnel handling ESDS items shall be
responsible for compliance with the requirements specified."
4.2 Changed to: "SSWs, ionization equipment, and personnel shall be
certified. "
4.5,4.6, Deleted.
4.7,4.8
4.10 Changed to: "Verification of automated processes. Verification of
automated processes shall be done monthly."
5.3,5.3.1, Deleted.
5.3.2, Ref.
Doc. 1
Ref. Doc. 2 Relabeled "Para. 5.3."
New 5.3, "Document" changed to "note."
subpara. 1
New 5.3, Changed to: "AN SSW is ... "
subpara.2
Ref. Docs. 3, Deleted.
4,5, and 6
New 5.4 Added: "Operator training."
New 5.4.1 Added: "Certification. Operators shall receive eight hours of
instruction, including a 30-minute videotape reviewing the basic
reasons for ESD control and illustrating the key procedures required
by this document, and operators shall be given a test with a
minimum of 20 multiple-choice questions (passing score 70%)."
New 5.4.2 Added: "Recertification. Operators shall receive two hours of
instruction, including a 15-30 minute video illustrating key
procedures, as well as a test as in 5.4.1."
New 5.4.3 Added: "Badge decal. Certified operators shall have a distinctive
decal applied to their company badges."
128 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
PARAGRAPH CHANGES
Program Organization
and Implementation
This chapter includes several topics: reasons for an ESD-control program; history
of ESD control and purpose of this chapter; objectives and overview of program;
what? when? and how?; cost-effectiveness; plan of action; the enforcement prob-
lem; advice to small companies; the role of the program coordinator; and safety.
130
PROGRAM ORGANIZATION AND IMPLEMENTATION 131
the chief executive to the new hire in the floor-maintenance crew. All too often the
ESD-control program is viewed as a set of rules inflicted on assemblers and tech-
nicians who process hardware. This distorted viewpoint destines a program to fail,
in the sense that the program costs more that it saves. Remember that operators
stand in the front line of the ESD battle, but without the logistics of the rear ech-
elon providing materials, equipment, and direction, the battle will be lost. There-
fore, a total management system must be developed to support the front lines.
A realistic and systematic approach is needed. This requires information on the
flow of product through the plant, from receipt and inspection of raw materials to
quality control and shipment of finished hardware. The factory is a complex, deli-
cately balanced structure whose operation is almost impossible to diagram be-
cause many support functions are not clearly defined. However, overlooking one
seemingly obscure function can jeopardize an ESD-control program. Engineering
organizations may be equally bewildering, but none of their elements can be slighted.
Only a total, integrated effort can succeed.
The beginning of the development of an ESD-control program is an informa-
tion-gathering process in which the following questions must be answered:
I. What is the lowest V-zap level (see "HBM" in Chapter 4) in each process?
2. What control methods are best suited to each step in the process? For example,
are wrist straps preferable to conductive floors combined with conductive foot-
wear? (We recommend wrist straps, with long cords and overhead trolleys if nec-
essary, whenever possible. Conductive floors can be hazardous, as pointed out in
Reference Document 1 in Chapter 9.)
3. What level of personnel training will be required?
4. Are existing policies supportive of an ESD-control program?
5. What new company policies will be required to enforce ESD-control
disciplines?
6. What special skills will be required in each area, e.g., assembly versus testing?
profile. "How?" is the tough question, so let's look at "when?" first. It is impracti-
cal to implement total ESO control all at once. The first control measures should
be introduced concurrently with personnel training in the areas of most vulnerabil-
ity to ESO. These initial measures are easiest to sell because their cost-effective-
ness is readily demonstrated. After the first disciplines are in place, refinements
and expansion of ESO controls to the balance of the factory will commence. With
refinements, diligent attention to detail will be needed to prevent reverting to the
old way. Workers will have to break old habits and establish new ones, and
management's failure to recognize and support the habit-changing aspects of the
program will cause it to be ineffective and fail to produce the desired cost savings.
"How" involves the most controversy. Each salesperson selling ESO-protective
materials and equipment claims to have the best "how" and will quickly tell you
why competitors' "hows" will not work. The only practical solution to this di-
lemma is to buy only products that clearly implement the Basic Rule. Remember
that three main damage mechanisms are at work. First is DI, of which the HBM is
the imperfect simulation (too high resistance, for one thing) of the most common
source of current flow: touching by people. Second is the CDM, in which the
device is charged and subsequently is damaged by DI when an ESDS lead touches
a surface with some conductivity, especially a truly conductive one. Third is the
FIM, in which a field causes a difference in potential to be induced in the device
so that the device "zaps" itself from within. A special case of this is the FFB, in
which a field develops on the board holding the devices; when the damage occurs
during subsequent discharge, the FFB resembles the CDM. The FIM is the most
insidious of the damage models because of the complex nature of internal circuitry
and the unknown locus of electrical energy imparted by a field. Repeatability in
FIM testing is difficult, and data can't predict the outcome of a subsequent test
where one of the parameters such as field strength is changed. There is no stan-
dard test established by government or industry for establishing reliable tolerance
thresholds. However, a conservative safe-distance range has been empirically de-
veloped and over the years has proven safe and reliable; this is our own CD Rule.
With the above three damage models in mind, materials and equipment should be
selected to combat specific threats. For example, antistatic materials, as opposed
to conductive, will be chosen to discourage the CDM by lessening the rapidity of
discharges.
Cost-Effectiveness
Special ESO-control program design objectives should be developed around facili-
ties and equipment as configured in the shops and laboratories engaged in producing
ESDS hardware. Therefore, no one program will fit all occasions. Cost-effective-
ness, which is the driving force with top management, is attained by accurately de-
fining control requirements and fitting them to existing conditions. Factors in selecting
these requirements are: (1) lowest level of ESO-sensitivity, (2) mission criticality of
134 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
the product, (3) cost of ESDS components, (4) cost and feasibility of rework, (5)
schedule impact of ESD failures, and (6) willingness and ability of workers to accept
and implement more or less complex disciplines of ESD control.
In this book we have made the following assumptions about the program illus-
trated by our model specifications in Chapter 9:
1. "A wrist strap will render the human skin harmless." The truth is that ESD is
first and foremost a people problem, and a grounded operator is only the lesser of
two evils versus an ungrounded operator because of the CDM. Furthermore, the
mere wearing of a wrist strap doesn't guarantee that the ARTG is being met; con-
tinuous monitoring is necessary.
2. "Higher humidity will solve all ESD handling problems." The truth is that
increasing the humidity is far from a panacea and will only help. It would be more
correct to say that low humidity will aggravate problems when ESD-control
measures are marginal.
PROGRAM ORGANIZATION AND IMPLEMENTATION 135
3. "Components are safe once they are mounted on a printed circuit board."
The surprising truth is that they can be at greater risk, because circuit lines in-
crease the target area for DI and CDM events and act as antennas for FlM events.
At best the board can provide protective networks which will reduce the ESD
hazard but not eliminate it.
4. "Faraday cages are constructed of carbon-loaded plastic." The truth is that
carbon-loaded conductive plastic provides some field attenuation but insufficient
protection from DI. See "Faraday Cage" in Chapter 4.
5. "Ionization will allow the use of plain, nonconductive plastics in an ESD-
controlled area." The truth is that ionization will neutralize charges on ungrounded
nonconductors but will not prevent triboelectric charging in the first place. This
charging can occur in an ionized atmosphere and zap a part before ionization can
neutralize the charge! Therefore, ionization cannot eliminate requirements for
passive control measures.
6. "Ionization is a substitute for the use of wrist straps or other grounding of
personnel." This is an egregious myth that has been proposed in the past by at least
one equipment supplier. The truth is that grounding is necessary whether the at-
mosphere is ionized or not, because ionization takes time to neutralize a charge
and ESDS items can be damaged before the ions have done their work. Remember
that the Basic Rule requires protection "even for a nanosecond."
7. "Devices with protective networks (designed and built in) are not ESDS."
The unfortunate truth is that protective networks make devices less sensitive but
can't provide total protection.
Plan of Action
A plan of action can now be fonnulated to assign the responsibilities and actions
required to put the ESD-control program into place. This plan should include, but
not be limited to, the following elements:
2. Contracting. The negotiators for both the contractor and the customer must
have a meeting of the minds on ESD control before the contract is signed. An
imprudent contractor can give away the store with overkill that imposes unneces-
sary, excessive costs on himself or he can underkill and leave himself open to
quality and reliability problems. Contract boilerplate drafted in the vacuum-tube
era is no longer appropriate.
ionization, and the like, is a must for all F&IE personnel. Facilitization of an
ESD-protected area can be one of the more costly elements of the program and
therefore should be specified and designed by people well-trained in electrostatics.
8. Engineering. Engineering will determine and specify the V-zap levels (see
"HBM" in Chapter 4) of the product in every configuration from component
through module, subsystem, and system levels. Testing procedures, failure analy-
sis, and designing to meet customer ESD requirements are only the tip of the ice-
berg. The engineer who takes the easy way out when specifying ESD characteristics
can send production costs out of sight. Few engineers have had formal training in
electrostatics, and even fewer have experienced the unpredictable behavior of elec-
trostatic phenomena, e.g., tribolelectric charging, at assembly stations. There-
fore, educating engineers in ESD is an absolute must. Even then, standards,
handbooks, and specifications will not be understood in the same way by all engi-
neers, and the burden of specification interpretation will fall on the shoulders of
Engineering management.
10. Quality Assurance (QA). QA, with its many subset activities, will inter-
face with every part of the ESD-control program from procurement to final test
and customer acceptance. The ESD training of inspectors and first-line manage-
ment should be more extensive than that of manufacturing operators. As in manu-
facturing, the integrity of personnel is primary to success of ESD control.
need for minor changes will be obvious only to persons close to the program, and
they must ensure that it grows as required to maintain high yields of product and
prevent latent failures.
In conclusion, an ESD-control program is unlike anything else experienced in
the electronics industry because of the latent failure aspect: the product cannot be
routinely inspected to determine if the program is effective. Thus the program
depends on good practice, not testing, and requires integrity by everyone from top
to bottom. It is expensive and not easy to justify. If the program is realistic and
clearly presented to the people who can make or break it, it will be successful.
Also, it must be cost-effective, which is the key feature of the approach we are
outlining in this book. Unnecessary overprotection or overkill tends to result in an
unprofitable project-but remember that inexpensive overkill, as with foillami-
nate bags as Faraday cages, is all to the good. No matter how effective in theory,
the program will receive multilateral support only if it is comfortable to imple-
ment. That is, it must be rational, understandable, and practical, and this is the kind
of program we are attempting to sketch for you.
List A
NEEDED FOR EACH OPERATOR COST, $ (ILLUSTRATIVE ONLY)
List B
NEEDED FOR EACH OPERATOR COST, $ (ILLUSTRATIVE ONLY)
What does list A buy in ESD safety over list B? List A includes a conductive
mat and footwear as a backup to the unmonitored wrist strap, but dirt on the mat
and/or shoe soles can cause resistance to ground to exceed the ARTG, and the
operator might lift both feet and be completely ungrounded. Use of a continuous
wrist-strap monitor is the only positive method of enforcing the Basic Rule,
which does not allow the ARTG to be exceeded "even for a nanosecond."
The personal field meter in List B puts the operator in command of the FIM
danger and makes an antistatic smock and a conductive chair unnecessary. Sleeves
will be rolled up, and if necessary, as shown by apparent charge, the ordinary
chair upholstery could be treated with topical antistat to meet the CD Rule. List
B gives the means to meet the requirements of Model Specification 1 in Chapter 9.
Furthermore, list B includes static-limiting floor finish as a low-cost, unobtru-
sive backup or "safety net." ("Do it right, but keep the safety net tight.") For ex-
ample, if a supervisor with ordinary footwear entered the SSW and touched an
ESDS lead, damage would be far less likely with the special floor finish than with
a conductive mat, which tends to produce high charges on rubber-soled shoes by
triboelectric charging.
The ionizer in list A has no clearly defined purpose and is needless in this case;
see "Ionization" in Chapter 4.
Finally, list B includes a versatile Zapflash for checking grounding and, among
other things, proving that black plastic items are conductive (highly loaded with
graphitic carbon) and not merely pigmented black (with about 1% of nonconduc-
tive carbon).
In conclusion, list B is far superior to A but costs 27% less! This is a good
example of getting more for your money with basics instead of frills.
An even better example is seen in Lists C and D, based on the real-life case of
Corporation X, Case 2 (see specification tailoring in Chapter 11). Costs are in
1994 dollars. List C was being contemplated, and our consultation reduced the
expense by 54% by trimming superfluous or excessive equipment in favor of con-
tinuous wrist-strap monitors and some important but economical items in List D.
The principle of cost-effectiveness illustrated here is built into this entire book,
which we have written with the needs of small companies in mind. But remember
PROGRAM ORGANIZATION AND IMPLEMENTATION 141
that an optimum choice of materials and equipment is the minor part of the battle
against ESD. The major part is the training and encouraging of operators. For ex-
ample, that hypothetical $300 for a field meter is well spent only if the meter is used
properly! See "Field Meter" in Chapter 4. Thus, more important than "Buy the
basics, forget the frills" is another motto: "Work smart." A small company may have
an advantage here, because there are fewer people to be educated and perhaps some-
thing of a patriarchal system that encourages personnel to identify themselves with
the company and its policies and to be craftsmen rather than bored, alienated hirelings.
Incidentally, the work surface is most economically a static.dissipative table
mat or, if solvent- and solder-resistance is required, a hard laminate retrofitted to a
standard (but grounded) workbench. Much money can be saved by converting
LlstC
APPROXIMATE NUMBER APPROXIMATE
EQUIPMENT COST, S/UNIT OF UNITS TOTAL COST, $
List 0
APPROXIMATE NUMBER APPROXIMATE
EQUIPMENT COST, $/UNIT OF UNITS TOTAL COST, $
existing benches this way rather than buying new ESD-control benches as a big
company would tend to do.
A final example of cost-saving on equipment is the use of unglazed cardboard
or hard vulcanized fiber tote boxes. Assuming that tote boxes are allowed to go in
and out of SSWs, the boxes must be SSPs (Faraday cages) if unprotected ESDS
items are placed in them but need be only antistatic if the items are safely inside
Faraday-cage packages such as foil laminate bags. In the latter case, unglazed
cardboard (paragraph 3.6.6.1 of Model Specification I, Chapter 9) or hard vulca-
nized fiber (Paper No.7 in the Appendix) is satisfactory. As of early 1989, ap-
proximate prices for tote boxes 11 x 17 inches by 6 inches deep in a quantity of
100 were:
We mention safety last but, of course, it is not least. Be sure to rigidly enforce the
safety provisions of Model Specification 1, including use of ground fault circuit
interruptors (GFCls) if necessary, so that personnel cannot be electrocuted by con-
nection to ground as a result of ESD precautions; accidents can happen, of course,
whenever a hot lead and a hard ground can be touched simultaneously. Remember
to provide springs on the continuous wrist-strap monitors for plugging in the
wrist-strap cords, as required in Reference Document 3 in Chapter 9, so that the
operator can pull free from any direction. Discourage the use of conductive mate-
rials in favor of antistatic or static-dissipative ones that cannot carry a lethal
current (above 1.0 rnA). If conductive floors are used-which we don't recom-
mend unless absolutely necessary-beware of low resistance to ground that might
allow a person with wet, leather-soled shoes to be electrocuted; we found an ex-
ample of this condition (Reference Document 1 in Chapter 9).
In conclusion, when you design a safe system, consider every possible path to
ground and every worst-case scenario, e.g., wet leather-soled shoes on a conductive
144 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
floor as mentioned above. If I-megohm resistors are installed in every path to ground
and GFCIs are used when high voltage is being handled, electrocution as a result
ofESD control is virtually impossible-but this is a big "if' that requires vigilance
and enforcement. Therefore, QA and the company Safety function must be deeply
involved in the ESD-control program.
Finally, we must mention that many marginal safety questions will arise. Ion-
ization, for example, is considered somewhat dangerous when polonium-210 is
the ion source; see "Ionization" in Chapter 4. Many would call this concern far-
fetched, and an even more unlikely danger is the low voltage (0.1-5.0 V, Refer-
ence Document 3 in Chapter 9) imposed on the wearer of a wrist strap with a
continuous wrist-strap monitor; in this case, the voltage is not only low but it
is imposed across two skin-contact points on the wrist so that current will flow
only between these and cannot pass through vital organs. Concern over such
matters may seem ridiculous, but some nonengineers are easily alarmed by any
new "hazard" in the workplace, and Safety personnel must take their concerns
seriously and provide reassurance. Genuine hazards, such as those noted in
Government-Industry Data Exchange Program (GIDEP) Alerts, must be promptly
removed.
Chapter 13
This is a vital subject when expensive products, e.g., modules worth $100,000
each, might have to be scrapped because of ESD damage that is measured or even
only suspected.
The general definition of "mishandling" includes overstress by dropping, bump-
ing, temperature excursions (heating or cooling), vibration, or other mechanical or
environmental events, but we're concerned here only with ESD.
It is assumed that the attending employee, or another employee who observes
the ESD overstress, will take personal responsibility and promptly report the inci-
dent to a Quality Assurance (QA) representative, who will fill out the appropriate
paperwork ("squawk sheet"). Then a Reliability Engineer will review the data and
make a judgment. These data are:
These are the basic data to be recorded in the squawk. They mayor may not be
sufficient for Reliability, with advice from ESD specialists in other departments,
to make a decision on disposition of the item.
145
146 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
humidification, and especially ionization, must be used with caution; see the
discussion of these subjects in Chapter 4. For example, the reenactment of an in-
fraction in which the operator touching ESDS leads was accidentally ungrounded
might show 100 V on his skin with static-limiting floor finish versus 800 V with
standard floor finish. Then, if the mishandled item was sensitive to 500 V, damage
by DI from the operator's fingers could be discounted for the static-limiting floor
finish but not the standard one; the safety net proved useful.
Obviously, judgments often will have to be based on incomplete data, just as
business decisions are made, but safety nets and worst-case special tests, e.g.,
with coupons as described above, will help greatly. The more critical and high-
reliability the ESDS item, the less leniently should infractions be treated. Remem-
ber that latent failures are real though they may be rare (see "Latent Failure" in
Chapter 4). Ethics forbids the use of sophistry to excuse any and every infraction
not causing measurable damage, and some mishandled hardware will have to be
scrapped even though it passes all its functional tests.
Chapter 14
Checklists, General
148
CHECKLISTS, WHAT TO BUY AND DO. CONCLUSION 149
MIL-HDBK-773
ESD Controls Study, Final Report, prepared for NASA by Reliability Analysis
Center, Rome Air Development Center, September 1981
NAVSEA SE 003-AA-TRN-OtO: Electrostatic Discharge Training Manual
Mil-STD-129: Marking for Shipment and Storage
EIA 541 : Packaging Material Standards for ESD-Sensitive Items (was Interim Standard
5-A or 5-A-l) (order from Electronic Industries Association, Washington, DC)
NFPA 99: Standard for Health Care Facilities (order from National Fire Protection
Association, Quincy, MA) (formerly NFPA 56A; defines electrode for measur-
ing resistance of surfaces to ground)
ASTM F 150 (defines same electrode as NFPA 99)
DoD-SID-2000-lB (soldering) (replaced by MIL-STD-2000 in Jamuary 1989)
MIL-M-385to (microcircuits)
MIL-B-81705 (packaging)
FED-SID-tO 1 (test procedures)
ESD Association Standards (Chapter to)
Test Equipment
Field meter (preferably one for every operator)
Zapflash (one for every operator)
Surface resistivity meter (Voyager SRM-l1 0 or SRMlRTG, which also measures
resistance to ground, or equivalent) (one for every ESD-protected area)
CHECKLISTS, WHAT TO BUY AND DO. CONCLUSION 151
Multimeter
Megohmmeter and NFPA 99 (or ASTM F 150) electrode for measuring resistance
to ground of surfaces (or use Voyager SRMlRTG Meter)
Faraday cup
Personnel Voltage Tester (Voyager PVT-300 or equivalent)
Discharge test apparatus, including high-voltage power supply
MOSFETs and curve-tracer
Static Event Detector™
Charged plate monitor (for testing ionizers)
Topaka slip test equipment (for static-limiting floor finish)
Gloss meter (for the floor finish)
Samples and ramp for roller test for triboelectric charging (see Paper No.6 in the
Appendix)
Miscellaneous equipment for special tests
Checklists, What to Do
Operator Disciplines
Maintain SSZ and know its boundaries at all times.
Don't touch ESDS leads unnecessarily.
Handle parts by cases when size permits
Don't touch ESDS items to unapproved surfaces such as clothing.
Don't fidget or shuffle feet unnecessarily.
Don't touch ESDS leads while rising from chair or lifting foot.
Wear wrist strap with continuous wrist-strap monitor.
Keep hair less than 2 inches from skin and tie back as required.
Don't remove shorting clips until ready to use item.
Exclude nonconductors from SSW.
152 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Mottoes
Buy the basics, forget the frills.
Do it right, but keep the safety net tight.
Humidify if you can justify; ionize only when it's wise.
ESD-wax the floor; it doesn't cost much more.
People are the prime ESD problem-and the prime solution.
Work smart.
Don't look for trouble. (Don't touch leads unnecessarily, etc.)
A grounded operator is the lesser of two evils versus an ungrounded operator.
(Because of the COM.)
No charge, no discharge. (Preventing COM.)
When in doubt, try it out. (Run special tests.)
Setting Up a Program
Reasons.
Required by government contract.
Saves money by lowering reject rates.
Prevents latent failures.
CHECKLISTS, WHAT TO BUY AND DO. CONCLUSION 153
Overview.
Use realistic, systematic approach.
Gather information:
V-zap levels (by (HBM).
Best control methods.
Needed levels of personnel training.
Existing policies supportive of program?
New policies needed.
Special personnel skills needed.
Arrange information in meaningful format.
Factors in Cost-Effectiveness.
Lowest level of ESO sensitivity.
Mission criticality.
Cost of ESDS components.
Cost and feasibility of rework.
Schedule impact of ESO failures.
Willingness and ability of workers.
Dangerous Myths.
"Wrist straps render skin harmless."
"High humidity solves all ESO handling problems."
"Components are safe when mounted on boards."
154 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Enforcement Problem.
Bend lesser rules while still meeting Basic Rule.
Justify rule-bending by analysis or special test.
Teach personnel reasons for rules.
Make rules "comfortable."
Get workers' suggestions.
Use democracy but not too much.
Safety.
Use GFCIs when necessary (high voltage).
Install springs on continuous wrist.strap monitors.
Discourage conductive materials in the SSW.
Beware of conductive floors.
Consider all possible paths to ground and worst-case scenarios for low-resistance
grounding of people.
Be sure the Safety function is deeply involved in the ESD-control program.
Deal with employees' concerns about remote safety hazards of new ESD-control
equipment.
Conclusion
What are we guarding against? Fields and discharges. The SSP fully protects
against both. When out of the SSP, and ESDS item is either in transit under con-
tinuous operator supervision or else is in an SSW.
In the SSW, where operator disciplines are vital, fields are controlled by ground·
ing conductors, by excluding unnecessary nonconductors, and by controlling nec-
essary nonconductors with the CD Rule and topical antistat. The antistat makes
surface resistivity low enough for incipient static charges to bleed off; also, lubric-
ity of the antistat layer on the surface limits the triboelectric charging propensity.
To control static-generating processes such as grit-blasting, the techniques are ground-
ing, voltage suppression, and, if special tests demonstrate the need for it, ionization.
Discharges from people are avoided by using continuous wrist-strap moni·
tors to meet the ARTG and keep skin near ground potential along with other
grounded surfaces. Also, isolated conductors could participate in discharges if
charged by fields, so fields from electrical machinery, CRTs, ionizers, etc., must
be controlled. Of course, there shouldn't be any isolated conductors, but human
fallibility requires redundant (overlapping) precautions.
Backups to the above primary precautions or defenses include static-limiting
floor finish, certain operator disciplines such as prohibitions against touching,
and possibly humidification. We do not recommend ionization as a general backup;
it should be used only for specific purposes when other methods fail. (But, as of
1995, ionization is becoming more popular as a backup; see Chapter 4.)
Overall, the strategy is to be cost-effective by keeping fields weak rather than
eliminating them and by substituting astute operator disciplines for expensive
156 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Given these variables, we can't be more explicit than we've been. Besides, no
matter how much detail we gave, we could only tell you what to do technically, not
contractually or "politically." You must absorb our advice and then cast it in your
own mold, arriving at specifications extrapolated from our models.
This book may serve as a "consultant's bible" because an ingenious ESD-control
specialist could review a customer's specific needs and adapt our approach to them.
In other words, we've provided the cloth to be tailored to fit any program.
In conclusion, our approach is cost-effective yet conservative. By means of strict
operator disciplines and concepts such as the SSZ and the CD Rule, we forgo
purchases of frills such as conductive chairs which might otherwise be bought
"just in case." But we're conservative in the sense of insisting on protection with
high confidence by such means as Faraday-cage foil packages (or at least pack-
ages passing a discharge test), continuous wrist-strap monitors, and a relatively
low ARTG. We are also wary of dangers such as space charging by ionization,
corrosion by humidification, or contamination by antistats or particles or fibers
from conductive bags, boxes, wrist bands, etc. "Buy the basics, forget the frills."
The money saved on frills such as an unjustified room ionization system can
procure not only essentials, e.g., continuous wrist-strap monitors, but highly
recommended options such as a one-to-one ratio of field meters to operators.
Besides a cost-effective program, we also stress troubleshooting to deal with
any situation using the A-to-Z real and conceptual tools in Chapter 4. In summary,
what we're propounding is a logical, commonsense, economical approach based
on data from meaningful, realistic tests-as opposed to unsupported intuition and
data from misleading tests contrived by tendentious suppliers.
CHECKUSTS, WHAT TO BUY AND DO, CONCLUSION 157
We repeat that we are defending against fields and discharges, and every re-
quirement in the model specifications of Chapter 9 has a purpose in this "war," By
way of a summary, Table 14-1, which is a variation on Table 4-1 in Chapter 4,
relates some defenses to hazards and damage mechanisms; Table 14-2 distinguishes
between primary and secondary defenses. (Note: These tables are simplified for
clarity; not all defenses are listed.) Operators as well as engineers should fully
understand these two tables. "Do it right, but keep the safety net tight."
A book on industrial accidents makes the important point that the immediate
cause, e.g., a spark which ignites a flammable vapor mixed with air, is just the last
of a series of causes such as a broken valve letting gas leak, the failure of a detector
to note this leaking gas, etc. By the time the final cause occurred, the accident was
inevitable. Therefore, the question is not "What was the cause?" but "What can be
changed to prevent another similar accident?" Regarding ESD, the immediate cause
of damage by DI might be a spark from an operator's finger, but the underlying
causes could include failure to require continuous wrist-strap monitors, inad-
equate training of personnel, and the use of ordinary floor wax instead of static-
limiting floor finish. Thus, the chain of causes of ESD damage can lead back
through the failure of primary defenses as well as lack of secondary defenses (safety
nets) which make the ESD climate more forgiving. When an aerialist falls to his
death, the ftrst cause might be a frayed rope on the flying trapeze while the second
cause might be a tom or missing safety net.
A capsule summary of this book is:
This book is not a conventional text but a cost-effective approach to ESO control based
on extensive data in the authors' twenty published papers reprinted in the Appendix.
Equipment expense is minimized by superior handling techniques which let you "buy
the basics, forget the frills." Chapters include real and conceptual tools from A to Z,
damage mechanisms, troubleshooting (special tests), model specifications, industry stan-
dards, program management, and lists of what to buy and do.
Chapter 15
The Future
Materials
As we said earlier in this book, the EIA's single classification "dissipative" (at
least 105 but less than 10 12 ohms/square) is sensible, but we continue to use the
antistatic and static-dissipative categories in this chapter because they appear
in all the old literature on ESD control. Also, the distinction is meaningful in
terms of the CDM; antistatic surfaces are safer for charged devices than are
static-dissipative surfaces.
Antistatic packaging materials which depend on antistats that create a surface
sweat layer have problems with the fugitivity of the antistat, causing imperma-
nence of the antistatic property (Paper No. 2 in the Appendix) and sometimes
corrosion and contamination (Paper No.9 in the Appendix). Another deficiency is
the humidity dependence of the sweat layer, which involves moisture attracted
from the air. Therefore, hoped-for improvements in antistatic or static-dissipative
materials include permanence and effectiveness at very low relative humidity.
In fact, permanently antistatic or static-dissipative plastics have been created.
For example, "static dissipation originates within the polymer molecule rather than
being added" for a new static-dissipative ABS alloy (Ref. 15-1). Various plastics
are made permanently antistatic with polymeric additives (Paper No. 20 in the
Appendix). Also, a bag made of cellophane has the unusual properties of true
permanence and no antistat rub-off (because it contains no antistat) as well as
biodegradability. This material is listed in paragraph 3.3 .1.k of Model Specification
2 in Chapter 9.
159
160 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
Conductive plastic materials are filled not only with graphitic carbon but
stainless-steel or other conductive fibers, and there are proprietary compositions
whose conductive additives are unrevealed. Many new products can be expected
in this area, a prime application being tote boxes. (But note that we discourage
conductive surfaces in the SSW; see Paper No.7 in the Appendix.)
Coatings are another active field. Early antistatic or static-dissipative ones we
have tested either blistered or shifted to nonconductive after a I-day water soak,
representing tote boxes being washed or left outdoors in the rain (which has hap-
pened!). Static-dissipative packaging film with a radiation-crosslinked coating
(Ref. 15-2) is noncontaminating and essentially humidity-independent but has the
drawback of becoming nonconductive after brief exposure to water or various
organic solvents; we found this in our work, and the same result was described in
a NASA Materials Science Laboratory report issued in April 1988. Attempts have
been made by the manufacturer to increase solvent resistance, and current ver-
sions must be evaluated. Polyvinyl chloride film with a relatively heavy radiation-
cured coating has been found to resist solvent extraction of the antistat.
In general, materials of the future will be noncontaminating, permanent by an
oven test, e.g., at 160°F, solvent-resistant, and humidity-independent. The user
will be able to pick any surface resistivity he wants. However, triboelectric charg-
ing may be a slight problem. The fugitivity of old-fashioned antistats has been
beneficial in controlling charging by the lubricity effect, in which a microscopic
amount of liquid antistat rubs off onto the other surface and liquid separates from
liquid to cause minimal tearing loose of electrons. Without this purposeful con-
tamination, charging of packaged items will tend to be higher. But whether the
loss of lubricity for new, permanently antistatic packaging materials will make
much practical difference remains to be seen. A compromise between contamina-
tion and triboelectric charging is to minimize the amount of liquid antistat on the
surface, as has been done in the new generation of MIL-B-81705, Type II, films
mentioned in Paper No.9 in the Appendix.
For the future of packaging materials, see Paper No. 20 in the Appendix.
Equipment
Equipment will evolve; for example, ionization will become safer in regard to
space charging (causing charging of ungrounded conductors) and more fully au-
tomatic in operation. Various test equipment will appear as called for in new stan-
dards and specifications. We hope that less expensive versions of popular items
such as field meters and surface resistivity meters will appear on the market so
that these essential tools can be more generously distributed to personnel and
preferably given to every operator.
Among new items of equipment are rigid ESD-protective boxes and cases that
are reusable and recyclable (Ref. 15-3) and a low-cost spring-like shunting device
to protect device leads or the edge contacts of a circuit-board module (Ref. 15-4).
THE FUTURE 161
Standards
New standards are evolving. Unfortunately, some of them may be unrealistic, as
discussed for the shielding/discharge test for bags in Chapter 6. The astute con-
sumer will take salesmen's claims cum grano salis and devise his own special
tests. Consider the case of static-limiting floor finish. If the objective is to limit
charges on personnel wearing their customary shoes, then the "bottom-line" test is
the walk test (Paper No.6 in the Appendix) or measurements with a Personnel
Voltage Tester (Paper No. to in the Appendix), whereas surface resistivity is
useful only insofar as it correlates with the voltage on walking people-and this
correlation holds only for a given floor finish, not for different finishes being com-
pared. The slip test (Paper No.6 in the Appendix) is also a critical bottom-line test,
though unconnected with ESD, because a slippery floor can bring lawsuits. Let's
forget irrelevant tests and concentrate on those which reflect reality.
Among the newer standards are the new verion of Test Method 3015.6 in Notice
7 to MIL-STD-883C (February 1988), EIA 541, MIL-STD-1686 (replacing DoD-
STD-1686), and MIL-HDBK-773. The last will "change the ways things are done
in-plant" (Ref. 15-5); for example, an interesting requirement in paragraph 5 of
this document discourages the use of conductive materials because they can ac-
cept spark discharges.
Also, ESD Association Standards are appearing rapidly (Chapter to).
ESDS Items
Several years ago, there was speculation that protective circuitry might greatly
reduce the burden of ESD control by leaving few ESDS devices sensitive to less
than 2000 V (by the HBM) when installed. However, at the same time, device
geometries were shrinking to less than 0.5 microns, putting some devices into the
12-volt category. On balance, ESD remains a major threat, only partly alleviated
by protective circuitry, and ESD control is difficult indeed for new components
sensitive to well under the old MOSFET level of 100 V.
For one thing, a wrist strap grounded through 5 megohms will not protect a
12-volt device from an operator stroking a garment sleeve (represented by a plastic
film in our test) on a table in such a way that he would generate 3800 V ungrounded;
see the curve for Aclar in Fig. 1 of Paper No. lOin the Appendix. To protect such
a device, operator movement would have to be restricted and a low ARTG, per-
haps 1 megohm, required. Also, antistatic garments, conductive chairs, and
humidification (if justified) might be needed.
What about fields? Making the assumptions of Paper No.3 in the Appendix, the
equation for the CD Rule for a device of any HBM sensitivity level becomes:
1,000 41 4.2
5,000 113 9.4
10,000 160 13
20,000 226 19
163
164 ESD FROM A TO Z: ELECTROSTATIC DISCHARGE CONTROL FOR ELECTRONICS
8-2. Scott's Standard Methods of Chemical Analysis, 6th Edition, Volume I, page 334.
8-3. NAV SEA SE 003-AA-TRN-0IO, "Electrostatic Discharge Training Manual," pages 35 and 36.
8-4. R. D. Anderson, "Alert Error Corrected: How Sweet It Is!" EOSIESD Technology Magazine,
OctoberlNovember 1987, page 8.
8-5. "Designing for Compliance: Immunity to ESD," Application Note 106, Special Supplement,
Compliance Engineering Magazine, 1991.
9-1. N. B. Fuqua and R. C. Walker, "ESD Controls Study, Final Report," prepared for NASA by the
Reliability Analysis Center, Rome Air Development Center, September 1981, page 30.
10-1. S. A. Halperin, "Anything But Static: The EOSIESD Association Takes Charge," Compliance
Engineering Magazine, Winter 1991, pages 13-28.
10-2. S. Weitz, "New Trends in ESDTest Methods," EMC Test and Design Magazine, February 1993,
pages 22-26.
10-3. 1. C. Hoigaard, "ISO 9000 Promotes Automated Continuous Monitoring," Evaluation Engi-
neering Magazine, February 1995, pages 105-108.
12-1. K.1essen and 1. Barto, "Static Control Team Concept-Implementation of ESD Protection in
Manufacturing," Evaluation Engineering Magazine, November/December 1983, page 94.
12-2. 1. R. Giuliano, "SD Program Nets Large Financial Gains," Evaluation Engineering Magazine,
December 1986, page 26.
15-1. "New Plastics Harvest Adds Variety," Machine Design Magazine, 1uly 21,1988, page 12.
15-2. A. H. Keough, "Antistatic Resin Composition," U.S. Patent 4,623,594 (November 18, 1986).
15-3. P. O'Shea, "Totes/Bins/packaging Go Environmental," Evaluation Engineering Magazine, March
1995, pages 92-95.
15-4. D. Cronin, "Cro-Bar: A New Technique for ESD Protection," EMC Test and Design Magazine,
1anuary 1993.
15-5. "MIL-HDBK-773: Sighs of Relief," EOSIESD Technology Magazine, OctoberlNovember 1988,
page 7.
Appendix
Some of the following papers have been condensed to highlight conclusions and
save space, but the majority have been reproduced in their entirety.
Paper No. 1 was reprinted with permission of lIT Research Institute/Reliability
Analysis Center. Paper No.5 was reprinted with permission of Evaluation Engineering
Magazine. Paper No.7 was reprinted with permission of EOSIESD Technology
Magazine. Papers No.2, 3, 4, 6, 8, 9, and 10 were reprinted with permission of the
EOS/ESD Association. Papers No. 11-20 were reprinted with permission as noted
in each case.
For a quick overview, leaf through and read the abstracts.
167
PaperNo. 1
Note: The following is a condensed version of this paper. See the Symposium
Proceedings for the complete text.
Abstract
Seven materials were evaluated. Properties such as transparency, puncture resistance, and
heat seal strength were measured. Also, ESDS devices in bags were subjected to the exter-
nal static field of a model "human finger" charged to 25,000 volts as well as to the extreme
condition of a high-voltage continuous discharge. The assumption was that if the packaged
devices survived ESD overstressing with no measurable damage they would not be subtly
damaged under actual handling/transit conditions. Such damage could reduce lifetime and
reliability. For the highest confidence in protection from ESD, at an acceptable cost and
with military specification coverage, a double-bagging system was adopted: antistatic poly-
ethylene, which is MIL-B-81705, Type II, as the intimate wrap with a foil bag, MIL-B-
81705, Type I, as the exterior "Faraday cage." Another foil bag, proposed as MIL-B-81705,
Type III, is effective without a Type II inner bag. A brief review is given of the context of
static control plant procedures in which the protective bags are used.
Introduction
The selection of packaging materials for ESDS items is difficult because many products are
on the market and the claims made by rival manufacturers are often questionable and even
contradictory.
The early products included antistatic polyethylene or "pink poly," which is MIL-B-
81705, Type II; conductive (carbon-loaded) polyethylene; and a laminate incorporating alu-
minum foil, which is MIL-B-81705, Type I. Later, variations on "pink poly" and foil bags
169
170 APPENDIX
were developed; these are "blue poly" and a foil laminate bag proposed as MIL-B-81705,
Type III, and so abbreviated in quotes as "Type III." Also, a new concept in "static-protec-
tive" bags appeared: polyester film with an exterior coating of vacuum-deposited nickel,
protected by a thin lacquer coat, and an antistatic polyethylene lining; this construction has
the advantage of partial transparency. The proponents of the latter bag made the industry
aware of the need for "Faraday cage" protection from external static fields/discharges and
the insufficiency of "pink poly" alone for worst-case handling conditions. An aluminum-
coated, partially-transparent bag is now competitive with the original, nickel-coated ver-
sion. In 1980, Department of Defense documents were issued, DOD-STD-1686 and
DOD-HDBK-263, which emphasized that the interior of a bag must be incapable of
triboelectric charging, as true of "pink poly," while the exterior should be a conductive
"Faraday cage" when external static fields might be encountered.
Our evaluation included all seven materials mentioned above. The objective was high
confidence in protection from external fields/discharges, at an acceptable cost, even for
items of the greatest ESD sensitivity. The bags were for general use: they were not for clean
room applications, handling of open devices, or the packaging of items sensitive to trace
contamination. Functional properties such as heat sealability were checked, and packaged
MOSFETs were exposed to external static overstresses to provide assurance of freedom
from damage by realistic stresses. The chosen material(s), of course, had to be cost-effective.
Experimental Methods
Figures I through 4 show some of the experimental methods. Figures 5 and 6 are a typical
example of ESD damage.
Materials Tested
The seven materials are described in Table I. All were received as samples in the form of
bags. "Pink poly" is MIL-B-81705, Type II; by "MIL-B-81705" we are referring, through-
out this paper, to the current Revision B. Two leading brands of carbon-loaded polyethylene
were tested; the volume resistivities, which were calculated from surface resistivity mea-
surements with bar-clampled samples at low voltage, were 235 and 875 ohm-em. Using
bar-clamped samples and a megohmeter at 10 volts, surface resistivities were in the ex-
pected range for the antistatic lining of the nickel-coated material, 2 x 1010 ohms/square,
and the aluminum-coated material, 7 x 1011 ohms/square. "Pink poly" and "blue poly" gave,
respectively, 8 x 1010 and I x 1011 ohms/square. The black lining of the Type I bag gave an
apparent 8 x 107 ohms/square, but edge effects were possible. "Type III" has a "pink poly"
lining for which 2 x 1011 ohms/square was found. Surface resistivity measurements are
neither closely reproducible nor a critical test for antistatic properties; static bleed-off time
(Federal Standard 101, Method 4046) is better for single materials but not for the Type I
laminate, whose foil layer drains the charge in terms of perceived voltage.
CAPACITOR
NONCONDUCTING IURFACE
-11.aoo OR -.aoo V
co. 1. 5. OR • SECONDS)
.......
NONCOND_ ....UCTING SURFACE
. . . .ITIIIIMCI
• • KW PEAK lOWER
• ,. GIGAHERTZ
.,0SECONOS
If an ESD shielding rating of "good" and a degree of transparency are required, the
metal-coated bags seem adequate. The aluminum version was more transparent than the
nickel version and performed as well in our other tests. However, potential users should
conduct their own evaluations.
In our case, "excellent" ESD shielding was desired. No sacrifice in confidence in ESD
protection was made in order to gain partial transparency. This position left only two candi-
dates: 'JYpe I and "'JYpe m." These foil bags, incidentally, not only give maximum ESD
protection but have superior sturdiness (abrasion and puncture resistance) and are excellent
moisture barriers. Since 'JYpe I has a questionably antistatic lining, as discussed above, only
PAPER NO.1 173
Fig. 5. Punch-through of gate oxide induced by "finger" (25,000 volts, 167 pF) with
MOSFET in black bag (1200><).
''Type III" with its "pink poly" lining remained as a single-bagging candidate. ''Type III" is
not yet covered by a military specification, but approval is said to be pending.
Our MX missile module assembly facility, for which a bagging system was being se-
lected, required cushioning material around assembled modules. The choice was "pink poly"
bubble-pack with skins of Richmond RCAS-1200 (the only material thoroughly tested by
us for antistatic properties). Since this material is not a "Faraday cage," double-bagging
was necessary; a Type I bag was chosen as the ESD shield. Type I was preferred to ''Type
III" because the former has military specification coverage and the latter's "safe" lining
was not required. Also, Type I is less expensive (Table 1).
When cushioning is unnecessary, a MIL-B-81705 system comprises a "pink poly" (Type
II) inner bag and a Type I outer bag; both materials are bought from suppliers on QPL-
81705-6 (issued December 30, 1980). The inner "pink poly" bag provides high at-work-
station transparency; an operator can withdraw the pink bag and easily see its contents. If
174 APPENDIX
Code: E = Excellent
G= Good
F=Fair
P= Poor
x= None
the latter feature is not considered important, the "Type DI" bag is effective but does not yet
have military specification coverage.
The above benefits of double-bagging with Type II inside Type I are bought at a cost
penalty of 13-21% over single-bagging with meta1-coated or '7ype III" bags. Using the
prices in Table I, a Type I bag at 21¢ and a Type II bag at 13¢ together cost 34¢, which is 4,
5, or 6¢ more than the nickel-coated. ''Type III," or aluminum-coated bags, respectively.
The above selections are based on present knowledge, but there are unanswered ques-
tions concerning bag materials. For example, how permanent are their ESD shielding or
antistatic properties? Accelerated aging tests combined with real-time exposures should
give reassuring, or disturbing, answers. Also, new materials may be expected to appear on
the market. The "bag of the future" may have most or all of the following properties: low
cost, >75% transparency, high ESD shielding effectiveness (like Type I or "Type III", per-
manent antistatic and ESD shielding properties, no dependence on a minimum relative hu-
midity, no propensity to contaminate packaged items. minimum sloughing (with no possible
shedding of conductive particles), excellent moisture barrier properties, good heat-sealability,
and, last but not least, cost-effectiveness.
PaperNo. 2
Note: The following is a condensed version of this paper. See the Symposium
Proceedings for the complete text.
Abstract
The shelf life of antistatic materials has been questioned. Therefore, accelerated life tests
were conducted with several commercial bag materials. One criterion for retention of the
antistatic property was ability of the bag lining to triboelectrically charge NEMA FR-4
glass-reinforced epoxy circuit board material. It was found that light rubbing, as occurs on
withdrawing boards from bags in in-plant handling, gave the same result-a low and pre-
sumably tolerable charge-whether or not antistat was present. Contamination caused higher
charging. Thus, shelf life was judged to be unlimited, by this test, if bags are stored closed
and clean. Another criterion was the charge on the bag itself. The static field is suppressed
by foil or metallization, but surface conductivity is required to control the field for antistatic
polyethylene alone (MIL-B-81705, Type II). For a leading brand of the latter, incorporating
a low-volatility antistat, the shelf life was estimated as at least 8 years if bags are used at
above 20 percent relative humidity. Heavy or repeated rubbing, as may be caused by vibra-
tion in shipment, is another matter. Transfer of antistat from bag lining to surface of the
repeatedly rubbed item seems necessary to prevent high charging. Again, a shelf life of over
8 years was assigned to the tested brand of MIL-B-81705, Type II, but data were insuffi-
cient for a prediction for foil or metallized bags. Antistatic polyethylene tote boxes were
also studied. Used, worn boxes can accept a triboelectric charge. Such boxes might be
periodically treated with topical antistat solution. Conductive boxes are an alternative but
have potential problems such as sloughing of conductive particles or triboelectric charging
of ESD-sensitive items being added or removed.
175
176 APPENDIX
Introduction
The pennanence of antistatic plastic packaging materials is in doubt because large increases
in surface resitivity during a I-year shelf life of films or foam sheets have been reported.'
Increased susceptibility to triboelectric charging by polystyrene foam also was found for
aged antistatic materials. The indication is that the antistat on the surface of the plastic
volatizes into the air or is removed by contact with absorbent materials such as cardboard.
Therefore, minimization of exposure to moving air or paper products was recommended.'
In view of the above, our purpose was to establish the pennanence of certain antistatic
materials, in the fonn of bags, which we use or might use. Of particular interest was the
propensity of bag materials to triboelectrically charge a printed circuit board module on
insertion or withdrawal. Therefore, a static charging test was devised in which a bare glass-
reinforced epoxy board was pulled in a reproducible manner from aged bags and the charges
on the board measured. Also, surface resistivities of aged bags were noted. Accelerated
aging was achieved by exposing bags, both closed and propped open, to 160°F in a forced-
draft oven. Bags were also exposed to blotting paper.
Tote boxes were also of concern because used, worn, antistatic polyethylene boxes found
in the plant were observed to take charges of up to a few kilovolts when stroked with a
nylon cloth. Whether a given charge was really a practical danger was judged by dropping
highly ESD-sensitive devices [metal oxide semiconductor field effect transistors (MOSFETs)]
into a nonantistatic tote box charged to various levels and electrically testing the devices for
degradation.
Experimental Methods
Figures I and 2 show the principal experimental methods.
Conclusions
Table 1.
Type I \0 MIL-B-SI705, Type I (vapor-deposited aluminum +
(brands A and B) spun-bonded polyethylene + aluminum foil +
carbon-loaded polyethylene)
Pink Poly 6 MIL-B-S1705, Type n (antistatic polyethylene)
Foil + Antistatic Poly S Antistatic spun-bonded polyethylene + aluminum foil
+ antistatic polyethylene
Nickel-Coated 2.5 100 Angstroms nickel + 1 mil polyester + 1.5 mils
antistatic polyethylene
Stainless-Steel-Coated 3 Polyester + stainless steel + antistatic polyethylene
Black 4 Carbon-loaded polyethylene (conductive)
Grid 4 Antistatic nylon + barrier film + conductive ink grid
+ barrier film + antistatic copolymer
Antistatic Foam 250 Layered, antistat-treated polypropylene foam
3/4-IN. THICK
WOOD(o.IILI'~ ~
1/2·IN. THICK •• ~
ANTISTATIC FOAM (0.031 LI'~ .. '. " , ...
ANTISTATIC lAG. Ix 10 IN. ~
FR ... EPOXY CIRCUIT IOARD (IARE.
NO COI'I'ERI. e x 12 x O.OIM IN.
"
(1.5 IN. EXTENDED INTO IAGI
TA8LESWITH
MElAMINE.
FORMALDEHYDE
lAMINATE SURFACE
CARRIAGE (CARRIAGE PLUS CIRCUIT
BOARD· 1.1 LII
LEADS TO MEGOHMMITE"
Y1:=. ~-
lUIt'ACE OF lAG
NOTE: A U·LI. WEIGHT WAI PLACED
tl1t4N. ALUMINUM
IIlOUNOED EDGI'
has a long record of successful use. Also our tote box tests suggest that limited static fields
may be tolerable. Therefore, it is presumed that the relatively low triboelectric charges
detected on the circuit board would not endanger attached devices. In contrast, the high
charges observed when the board was stroked with carbon-loaded polyethylene might well
be damaging. However, voltage suppression would limit the field on a multilayer board
with internal planes of copper.
2. In tenns of surface resistivity, shelf life at usual relative humidities (over 20 percent)
is conservatively estimated as 8 years for 6-mil "pink poly" bags when kept closed. This
relatively good pennanence is probably due to the low vapor pressure of the antistat, which
is extruded into the polymer and gradually bleeds to the surface to fonn a "sweat layer"
with atmospheric moisture. See Fig. 3 and Table 2.
3. Bags should be stored closed to exclude contamination such as dust, kept out of
contact with paper or other absorbent materials, and, of course, never washed with water or
organic solvents.
4. Two extreme situations for using bags are: (I) light, unrepeated rubbing by the pack-
aged item against the bag lining (in-plant handling) and (2) heavy or repeated rubbing due
to vibration (shipment). Antistat depletion is irrelevant for light rubbing, in tenns of our
board-charging test, but may be critical for heavy or repeated rubbing. In the latter case, the
key factor may be neither resistivity nor lubricity but antistat transfer from bag to packaged
item. The role of relative humidity is not clear. Note that the lining of MIL-B-81705, type I,
which is without antistat, has been reported to develop damaging charges on MOSFETs in
a vibration situation.3 Also note that a surface layer of peanut oil, instead of antistat, can
prevent triboelectric charging of DIPs by polyethylene under heavy-rubbing conditions. 4
5. Antistatic polyethylene bags without foil or metallization to control the static field
can develop significant charges on themselves at some combination of antistat depletion
and low moisture content of the "sweat layer." Here, relative humidity could be a critical
factor. Thus our prediction of a long shelf life for "pink poly" (MIL-B-81705, Type II) is
limited to use at >20 percent relative humidity.
PAPER NO.2 179
1I&0Il10" ATUVRI
tOl.L...----,!'---~2---,l-3- - -..!----:!.f----:!.
TIME. MONTHS
Fig. 3. Surface resistivity versus time for aged 6-mil ''pink poly" bags.
EXTREMBUSE MIL-S-817OS
CONDmON TYPE n FOIL OR METALLIZED
3. Conductive tote boxes have no permanence problem but may triboelectrically charge
devices being removed from them, slough conductive particles, or cause ESD damage by
rapid discharges as could happen with any conductor.
4. Each used must select a tote box material based on his unique parts-handling situation.
References
I. G. O. Head, "Drastic Losses of Conductivity in Antistatic Plastics," Reliability Analysis Center
EOSIESD Symposium Proceedings, pp. 120-123, Orlando, FL, 1982.
2. J. M. Kolyer and W. E. Anderson, "Selection of Packaging Materials for Electrostatic Dis-
charge-Sensitive Items," Reliability Analysis Center EOSIESD Symposium Proceedings, pp.
75-84, Las Vegas, NY, 1981.
3. D. M. Yenni, Jr., and J. R. Huntsman, "The Deficiencies in Military Specification MIL-B-81705:
Considerations and a Simple Model for Static Protection," presented at the Reliability Analysis
Center EOSIESD Symposium, Denver, CO, 1979.
4. J. R. Huntsman and D. M. Yenni, Jr., "Test Methods for Static Control Products," Reliability
Analysis Center EOSIESD Symposium Proceedings, pp. 94-109, Orlando, FL, 1982.
PaperNo. 3
Abstract
Tests were conducted to determine under what practical conditions electrostatic discharge-
sensitive (ESDS) items can be damaged at the work station by static charges and fields.
Various materials and equipment were evaluated for possible hazards (e.g., static fields
created by electrical air ionizers) which must be avoided by careful operating techniques.
The data led to important conclusions and guidelines for an ESD control program in accor-
dance with DoD-STD-1686. Examples are given of equipment, materials, and techniques
which complement one another, and basic rules for electronic assemblers are suggested.
Introduction
A charged surface always creates a static field (E-field). However, in discussing "hazards of
charges" we ignore the field and are concerned with discharges between conductors, for ex-
ample from a person's finger to the copper line on a circuit-board module. In discussing
"hazards of fields" we are concemed with ESD damage caused by induction. The effects of
RF (radio-frequency radiation) or of magnetic fields (H-fields) are not considered in this paper.
Following a description of the test methods used, we evaluate charge and field hazards,
including those from electrical equipment, and discuss some methods of coping with them,
Finally we briefly describe a synthesis of materials/equipment and operator techniques for
handling ESDS items in accordance with DoD-STD-1686.
Test Methods
Surface Resistivity
The electrodes have been illustrated. I A 5-lb weight rested on each electrode, and strips of soft
carbon-filled polyolefin were placed under the blades to improve contact with static-dissipa-
tive or antistatic surfaces. With this arrangement, a steel surface read lOS ohms/square, so lOS
was subtracted as a correction factor. Readings were made with a Beckman Model L-1O
megohmmeter at 500 volts unless otherwise noted. Multimeters were used for lower voltages.
COPPER
CIRCUIT
LINE
6 x 9 x 0.06 INCH
PR-4 GLASS-EPOXY
LAMINATE
ESD damage to the MOSFETs by static fields. 2 Similar printed circuit boards with MOSFETs
and circuit lines for antennas have been used in other investigations3•4 but for different
purposes. Our board, held by the operator as shown in Fig. I, was intended as a worst-case
simulation of ESDS modules being handled at the work station.
MOSFETs were often shorted, as noted in the data, but they were also considered damaged
if the current-voltage curve changed and/or VOS(1lI) shifted by more than 0.1 volt (VOS(1ll) read-
ings for undamaged MOSFETs were reproducible to ±O.02 volt). Since latent ESD failures
are a reality,S a VOS(1ll) shift of only 0.1 volt is considered significant. Damage is reported as a
fraction; for example, "2/5" means that two MOSFETs were damaged out of five tested.
Shielding/Discharge Test
This test for evaluating packaging materials was conducted with apparatus in accordance
with published reports. 6•7 In brief, a capacitive sensor, comprising two 0.75-inch-diameter
aluminum disks (0.06 inch thick) separated by 0.5 inch of acrylic plastic, was placed inside
a 3 x 3 inch pouch of packaging material clamped between flat aluminum electrodes. See
Fig. 2. In accordance with a published procedure,8 a pulse of 1200 volts was applied from a
200 pF capacitor through a 400 kilohm resistor to the upper electrode, the lower electrode
being grounded. The pulse picked up by the capacitive sensor was delivered to a storage
oscilloscope. Results were conveniently expressed as "percent attenuation;" for example, if
the pulse recorded by the sensor was 300 volts, the percent attenuation was (1200 - 3(0)
(100)/1200 = 75. One edge of the pouch under test was always continous so that the mate-
rial had electrical continuity from side to side.
SHIELDIMG/DISCHAIGI TEST
Bench Tops
In discussing bench tops we are concerned not with the top holding a charge but with its
ability to drain charges from objects placed on it.
Static-dissipative tops (surface resistivity 1()5-109ohms/square) are favorably discussed
in DoD-HDBK-263 and have been suggested as ideal. 9Antistatic tops may drain charges
too slowly, as will be seen below, while conductive tops are an electrical safety hazard and
might damage charged ESDS items by discharging them too rapidly.9.10 (See also DoD-
HDBK-263 and the discussion below.)
A top may have a conductive sublayer without being satisfactory, because without suffi-
cient surface conductivity the charge will be merely voltage-suppressed and not drained."
In our tests, conductive or antistatic tote boxes charged to ±1O,OOO volts drained incom-
pletely (e.g., 8000 volts remained on an antistatic box) after resting 1 minute on an antistatic
top (lOll ohms/square) but drained completely «50 volts remained) on a static-dissipative
top (109 ohms/square). One successful construction for a static-dissipative top is, in our
experience, a high-pressure laminate with linen fabric in the surface layer and conductive
carbon paper (grounded through 1 megohm) as the sublayer.12 The sol vent resistance of this
laminate is reported to be very good,I2 and in our tests 50 daily rinses with methyl ethyl
ketone had no effect on surface resistivity or appearance.
Static-dissipative tops are safer than conductive tops for operators, but can static-dissipative
tops also be safer for devices as suggested above? To answer this question, tests were con-
ducted with the MOSFET board (Fig. 1). An operator was charged to various levels by
PAPER NO.3 185
contacting the dc power supply, and a second after releasing the supply he touched the lead
of the MOSFET board to either a static-dissipative bench top or a conductive work surface
represented by an aluminum sheet. Both surfaces were grounded through 1 megohm.
Results are given in Table 1. Both surfaces allowed damage to occur, but the static dissi-
pative surface was less destructive. That a more resistive surface can be less damaging will
be seen again in Table 2.
In conclusion, an antistatic top will not itself hold a charge, but it drains charges too
slowly from objects placed on it. The best choice seems a static-dissipative top with a bur-
ied conductive layer and a securely mounted grounding lug.
Tote Boxes
Tote box selection is a controversial subject. Antistatic tote boxes can lose their "sweat
layer" and should be periodically treated with an antistat solution. \ Carbon-loaded plastic
boxes, on the other hand, are permanently conductive but can slough conductive carbon
particles. If sloughing is not considered a problem, one question remains: Are carbon-loaded
boxes too conductive? The fear is that a charge on a conductive box might "zap" an ESDS
item, whereas under the same conditions the charge would bleed off harmlessly from an
antistatic box. Or, conversely, a charged ESDS item might discharge too rapidly to the
"sparking surface" of a conductive box. 10,\3
We attempted to settle this question using the MOSFET board (Fig. 1). Either the opera-
tor or the tote box was charged, and the lead of the MOSFET board was touched to the box.
Two boxes were tested: an antistatic box (surface resistivity 2 x 109 ohms/square at 48
percent relative humidity and 72°F) and a black conductive tote box (2.2 x 10" ohms/square
at 32 volts). Both boxes were injection-molded by the same manufacturer and had the same
dimensions (approx. 7 x 10 x 3 inches deep, with a 0.08-inch wall). Results are given in
Tables 2 and 3, which also include an aluminum sheet (12 x 12 x 0.06 inch) as a reference.
Table 2 also includes a static-shielding bag, discussed below. When the operator was charged,
the box or sheet lay on a static-dissipative bench top grounded through 1 megohm; when
the box or sheet was charged, it stood on insulating stand-offs (plastic breakers) and the
STATIC-DISSIPATIVE
SURFACE ALUMINUM
VOLTAGE (8 x 108 OHMS/SQUARE) SHEET
PAJmALLY-
TRANSPARENT BAG
(II x 14 INCH)
ANTISTATIC CONDUCTIVE ALUMINUM WITH EXTERNAL
VOLTAGE TOTE BOX TOTE BOX SHEET METALLIZATION
operator was grounded through 1 megohm with a wrist strap. The tests were conducted at
68 percent relative humidity and n°F.
The result was that the antistatic tote box was clearly less damaging than the conductive
box when the operator was charged but only slightly less damaging when the box was
charged. An explanation is that when the box was charged (Table 3) induction became the
dominant factor; that is, MOSFETs were damaged by capacitive coupling when the lead of
the MOSFET board penetrated the field of the charged box. On the other hand, when the
operator was charged (Table 2) the controlling factor was the resistance of the surface being
touched by the lead.
The conclusion is that antistatic boxes are less of a discharge hazard than conductive
boxes if handling techniques are poor. With good operator discipline, conductive boxes are
safe and have the advantage of not relying on a fugitive antistatic additive; however, the
user must assure himself that sloughing of conductive particles is not a problem.
In our assembly operations, modules are placed in Faraday-cage (foil-containing lami-
nate) bags, and then the bags are placed in tote boxes. When a tote box is received at a work
station, the operator removes the bag and sets the tote box aside before removing the mod-
ule. Thus, the tote box is not a threat no matter what its electrical properties. Even ordinary
plastic tote boxes could be used in this manner, but good practice calls for selection of
conductive or antistatic materials in an ESD control program. The blanket rule is that ordi-
nary plastics must be excluded from work stations.
Nonconductive Plastics
When a polyethylene film (0.004 inch) at -13,000 volts (apparent charge, as are all charges
reported in this paper) lay flat on grounded aluminum foil, the apparent charge was reduced
to -300 volts because of voltage suppression, and MOSFET damage by a grounded opera-
tor touching the MOSFET board lead to the plastic surface was 0/3. Similarly, for a vinyl
sheet (0.058 inch) at -8000 volts the apparent charge fell to -300 volts and MOSFET dam-
age was 0/5. The nonconductive plastics were unable to deliver their high charges, and the
suppressed fields were too low to cause damage.
The conclusion is that fields, rather than charges per se, are the hazard in this case. Thus
the MOSFET damage reported in Table 4 when charged nonconductors were touched would
presumably have been caused by a close approach without contact.
Operators
An operator's skin should always be grounded. However, a grounded operator can cause
ESD damage, as seen in Tables 3 and 4, and is only the lesser of two evils versus an un-
grounded operator. Good handling techniques are needed to protect ESDS items from
grounded operators when charges on the items are unavoidable. For example, a module
might be triboelectrically charged by spraying with a conformal coating. Then, if a grounded
operator were to touch a contact a damaging discharge could occur as explained in the
literature. 10.14 A solution to this problem in terms of handling techniques is for the operator
to be conscious of the hazard and avoid touching leads or contacts of ESDS items.
A possibility is that an operator's fingers might be reduced from conductive to static-
dissipative or antistatic surfaces by his wearing appropriate gloves, possibly cotton. This
idea is especially applicable to static-producing assembly or rework operations such as grit
blasting.
whose slope represents field strength. The same result is given by plotting rv
ted against d 2 , where V is voltage and d is distance from the surface, will give a straight line
versus d,
which was done for convenience. Of course, the source of the field is an area rather than a
point, and the data are not precise enough or numerous enough to establish an exact thresh-
0 l
V
7000
1
sooo
--• i-.
JIAZAIDOUS SAFE
~
S 3300 .oil.
... ...
.
Q
/
;0.
r.:I
...../ v
~ 2000 -..., ...
~
i
1700
:;
1300
~ 1000 .A
500
~ .. ~ ...
300
100 / ...
~NONCONDUcroRS
50 "/
o 10 20 30 40 so 60
DISTANCE, INCHES
This estimate of minimum hazardous field strength is, of course, worst-case because the
long antennas of the MOSFET board served to "gather" static fields. At the other extreme
was a "MOSFET module" in which the antennas (insulated wires) were sandwiched be-
tween aluminum ground planes while the projecting lead shown in Fig. 1 was eliminated.
At 1 foot from a surface at -7000 volts, the "MOSFET module" gave MOSFET damage 0/5
versus 4/5 (two shorted) for the MOSFET board. However, the "MOSFET module" did
allow damage (2/3, 0 shorted) at 2 inches from a surface at -7000 volts. Therefore, the
constant C for the "MOSFET module" is more than 7 but less than 42. Real-life module
assembly situations will lie between the two extremes represented by the MOSFET board
and module.
Air Ionizers
A few ionized-air blowers and nozzles (guns) were evaluated for field hazards by waving
the MOSFET board (Fig. 1) at various distances from their tips while they were running.
The operator was grounded through 1 megohm as usual. Table 5 gives the data.
A safe working distance from the blowers or the pulsed dc ionizer was 10 inches. Inci-
dentally, the pulsed de ionizer had the advantage of not blowing air. An airstream is often
190 APPENDIX
objectionable; for example, it causes undesirable cooling during soldering. However, this
fanless ionizer was as effective as the blowers in rapidity of neutralizing positive or nega-
tive charges on plastics in tests at a range of 3 feet.
The nozzles were operated at 30 Ib/in. 2 gage air pressure. Model C was less damaging
than the others; in this model, the corona-discharge point was mounted "piggyback" out of
the airstream. The nuclear type caused no damage because it has no electrical field. Another
advantage of the nuclear type is its compact size. However, its cost may exceed that of
electrical nozzles over a period of years because of an annual leasing fee, and the poisonous
radioactive material (polonium-210) presents a remote safely hazard. For very close work,
on the basis of field hazard as well as bulk, the nuclear-type nozzle is preferred.
The conclusion is that electrical air ionizers do not present a field hazard if the work is
kept 10 inches or more away from them. However, when nozzles must be held very close to
ESDS items, as in cleaning a module by blowing off particles, the nuclear type is required.
The equipment which we tested is meant to be operated with the positive and negative
emitters at similar voltages so that there is little net field strength or ion imbalance at the
workbench level 5 feet or more beneath the emitters. However, if the system were acciden-
tally unbalanced would an electrically isolated conductor on the bench become sufficiently
charged to damage ESOS items touched to it?
To answer this question, tests were conducted in the booth described above under "Test
Methods" using only one of the two emitters. The bench surface was static-dissipative as
described under "Bench Tops," above. An aluminum sheet (12 x 12 x ~ inch) was isolated
from the bench surface by two different thicknesses of nonconductive plastic (bubble-wrap)
as listed in Table 6. Thus, a capacitor was produced by the aluminum sheet and the carbon-
paper sublayer of the bench; when the two were separated by ~ inch. for example, the
calculated capacitance of the aluminum sheet was 129 pF. The voltage on the aluminum
sheet was measured with an Electrostatic Field Meter Model 970 from Static Control Ser-
vices; in this case the field meter usually employed was not sensitive enough. Table 6 gives
the data.
The system is normally operated at about 8200 volts; a slight increase, to perhaps 9000
volts, may be required at low humidity. Therefore. Table 6 indicates that even if all the
emitters of one polarity should fail completely there will not be dangerous charging of
isolated conductors lying on the bench. In conclusion, worst-case testing showed no hazard.
Having gained confidence that the room ionization system would not be part of the ESO
problem, we tested its efficacy at neutralizing charges on plastics. Table 7 shows that neu-
tralization of charges at the bench level was reasonably rapid for a variety of materials; the
relative humidity during this test was 56 percent at 74°F. Even with the system somewhat
unbalanced (15,000 positive ions/ml, 22,000 negative ions/ml), results were similar. To
check the effect of humidity. the booth was sealed and the air inside it was partially dried
with a desiccant. At 10-20 percent relative humidity, the decay rates of the materials in
Table 7 remained rapid.
Another question was the effect of position of charged objects under the emitters; it was
feared, for example, that a positively charged material under a positive emitter might re-
ceive relatively few negative ions and so exhibit a slow decay rate. Therefore a test was run
with a positively charged plastic and a negatively charged plastic at 51 percent relative
humidity and 73°F. Table 8 shows that charge decay rates remained far more rapid at any
position with the system on than when it was off. Tests at commercial installations of this
and a competitive system gave better uniformity, presumably because a grid of emitters
covered the ceiling instead of only two emitters being confined to a booth. Of course, per-
fect uniformity of decay rate across the bench is unnecessary.
It was desired to compare the effectiveness of the room ionization system with that of a
typical bench-model ionized-air blower (as in Table 5). Polyethylene bubble-wrap
triboelectrically charged to -10,000 volts was found to lose half its charge after a I-minute
exposure (at 60 percent relative humidity and 74°F) either to the emitters at ±8200 volts at
bench level or to the blower at a distance of 5 feet.
In conclusion, the room ionization system tested did not present a field or charge hazard
at the bench level and was indeed effective in removing standing charges on nonconductors
such as plastic packaging materials. Fig. 4 (for 49 percent relative humidity and 73°F), is a
typical example of the utility of room ionization in removing a stubborn charge on plastic;
similar curves appear in a published report. 16 Note that the electrical field of a charged item
attracts ions to cause neutralization. Therefore, a thin sheet of plastic which lies flat on a
workbench and has a largely collapsed field will be neutralized only slowly and incom-
pletely. The bubble-wrap in our test was i inch thick and was folded over on itself so that
the measured surface was elevated about I inch.
Assembly Operations
Many static-producing manufacturing operations are listed in DoD-HDBK-263 and a Navy
training manual,I7 but only two will be discussed here: peeling masking tape from a roll and
grit-blasting a module to remove the coating from components for rework.
A conductive masking tape would have virtually no static field because of charge drain-
age and voltage suppression. Our tests showed that a commercial aluminum foil/fiber tape
had sufficient conformability and tear resistance for masking, but its acrylic adhesive re-
leased incompletely after oven-baking of the masked parts at 150°F. Therefore, the only
deficiency of this tape might be overcome by substituting a silicone adhesive, with good
release, for the unsatisfactory acrylic.
A Micro Blaster (Comco, Inc.) was found to create a charge of +700 volts on FR-4
circuit board material when sodium bicarbonate "grit" was sprayed against the surface using
10 Ib/in.2 gage air pressure. By adding a "piggyback" corona-discharge ionizer (Simco PIN
-18
II)
-----
~160 WITHOUT ROOM IONIZATION
I~~
="'-140
--.......
~-12 0 I~
,
~
a- 100
\
1 ---.
i =:
0
0
~ -4 0
\
-2 0
\. ~WITH 100M IONIZATION
I'
10 20 30 40 so 60
TIME. MINUTES
4100034) operating at 7000 volts, with the point 0.6 inch from the spray nozzle, the charge
developed was reduced to only +60 volts. This charge is well below the danger level for
nonconductors (Fig. 2).
Operators
A properly grounded operator has no charge on his skin. However, arm hair has been re-
ported to develop up to 900 volts, even at 50 percent relative humidity, when skirt sleeves
are rolled up; at least one major corporation considers this a serious static problem. 18 Note
that in our worst-case tests as little as 500 volts (on vinyl plastic, Table 4) was sufficient to
damage a MOSFET. Obviously, head hair which dangles and might touch ESDS items is
also a hazard.
A synthetic-fiber smock can carry damaging fields, as seen in Table 4, and DoD-HDBK-
263 makes the point that clothing should never touch ESDS items. An "antistatic" smock is
at least a partial solution, but it should not contain stainless-steel fibers which might fall
onto circuit boards and cause shorts. 19 The operator's chair, if nonconductive, is also a
hazard (Table 4).
The general approach to the problem of unavoidable static fields on an operator's hair or
clothing-or on anything in the work station- is for him or her not to bring ESDS items
near charged surfaces unnecessarily. In other words, do not look for trouble. In addition,
head hair can be tied back, smock sleeves rolled up (with the proviso that ESDS items are
kept away from hair on the forearms), etc.
Wrist Straps
Various designs were tested. Bead chains gave intermittent contact but succeeded in pro-
tecting MOSFETs when the operator charged himself by shuffling his feet on a carpet and
touched the lead of the MOSFET board to ground. The main objection to bead chains is
their tendency to ride up over sleeves and lose contact with the skin. Any design which
presses grounded metal snugly against the skin seems adequate; this includes expanding
stainless-steel watchband types or designs in which a metal "wristwatch" element is held
against the skin by an elastic band. However, wrist straps which depend on the conductivity
of carbon-loaded plastic are sUSpect,20 and a conductive fabric band shed steel fibers up to
0.09 inch long in our tests.
Faraday-Cage Bags
It is well recognized that ESD-protective bags must guard against ( I) internal triboelectric
charging and (2) external static fields/discharges. 9
The bags in Tables 9 and 10 all provide some shielding against external fields. In terms
of the pulse attenuation test (Table 9), 93 percent attenuation must be sufficient because the
foil-containing laminate bag is an excellent Faraday cage. 21
PAPER NO.3 195
Antistatic Polyethylene 75
Carbon-Loaded Polyolefin (Conductive) 97
With Conductive Ink Grid, A 78
With Conductive Ink Grid, B 50
Partially Transparent with External Metallization 99,8
Same, but Handled (Metallization Cracked on Crease) 59
Partially Transparent with "Buried" Metallization, A 39
Partially Transparent with "Buried" Metallization, B 87
Laminate Containing Aluminum Foil, with Antistatic 93
Polyethylene Lining
Antistatic Polyethylene Bags Inside of and Outside of Carbon-Loaded 92
Polyolefin Bag
Antistatic Polyethylene Bag Outside of Partially Transparent Bag with 93
External Metallization
The high values given by a partially transparent, externally metallized bag (99.7 percent)
and a volume-conductive bag (97 percent) are artifacts, in a way, because conductive mate-
rial "shorts out" the upper and lower plates in the test; wrapping these bags in antistatic
polyethylene lowered their pulse attenuation to the 92-93 percent range. By this test it would
seem that carbon-loaded polyolefin is as good a Faraday cage as aluminum foil, but such is
not the case; discharges from a person's finger can damage ESDS items inside conductive
plastic bags. 3,21
A more significant test is the device-in-bag procedure of Table 10. The conditions were
certainly worst-case because a very low resistance (150 ohms) was used; a resistance of
*The EDS-200 probe was replaced by the fmger of it person triboelectrically charged by walking on a
carpet at 24% relative humidity and 73°F.
**Surface resistivity between sides = 256 ohms/square at 0.2 volt.
***Surface resistivity between sides = 2 x 10" ohms/square at I 00 volts.
196 APPENDIX
1500 ohms is specified in the human ESD model of DoD-HDBK-263, and a realistic esti-
mate for human resistance is said to be 350 kilohms.6 However, the foil-containing laminate
bag survived even this test.
MOSFETs in a partially transparent, externally metallized bag were damaged by a dis-
charge from the finger of a statically charged person (Table 10), and the spark burned off a
small area of metal. Published data6 seem to disagree, but conditions differed somewhat
from ours; for one thing, the capacitive probe had no leads as ours did. Also, the bag may
not have been pulled tightly over the upper plate of the probe as it was on our test. As
expected, MOSFET damage was worse (Table 10) when the externally metallized bag had
a low-conductivity crease caused by cracking of the metallization by handling. Also note
that externally metallized bags, being good conductors, can participate in discharges at the
work station (Table 2).
The conclusion, which we had reached earlier,21 is that only foil-containing laminate
bags provide absolute protection against worst-case fields and discharges. A suitable com-
mercial product is a three layer laminate with aluminum foil (0.00035 inch thick) sand-
wiched between antistat-treated spun-bonded polyethylene on the outside and antistatic
polyethylene on the inside. Note that Mll..-B-81705, Type I, is a foil-containing laminate
but has an unsatisfactory liner which does not prevent internal triboelectric charging. 4•9
Equipment/Material Selection
and Handling Techniques
One school of thought recommends exclusion of conductive materials 13 while another, the
"conductive approach," visualizes most objects in the work station as conductive, grounded,
and therefore at zero potential. 22 Either approach, or a mixture of the two approaches, will
be effective with proper maintenance of materials (e.g., periodic treatment of antistatic
surfaces with topical antistat) and handling techniques which accommodate the limitations
of the equipment/materials. Table 11 illustrates the principle of complementary materials
and techniques.
The fact is that ESD-protective materials must be used correctly to be effective. 23 In
other words, operators are more important than equipment. For example, a grounded con-
ductive chair would be more dangerous to ESDS items then an ordinary chair if a careless
operator without a wrist strap caused discharges by brushing device leads against the uphol-
stery. Conversely, a very skilled operator can protect ESDS items with only the bare essen-
tials of equipment, including a wrist strap, because he knows where the dangers lie. Therefore,
operator training is essential for ESD control programs as prescribed in DoD-STD-1686. A
wide range of equipment/materials is described in DoD-HDBK-263, and it is up to the user
to integrate these with operator disciplines in a unified approach.
In our program, handling techniques accommodate a mixture of conductive,
static-dissipative, and antistatic materials. Some basic rules are: (1) Never touch ESDS
device leads with the fingers unnecessarily, and keep ESDS items away from all surfaces
except the bench top and the assembly to which the item is being attached. (2) Keep ESDS
items in Faraday cages, such as foil-containing laminate bags, when away from the static-
safe work station. (3) Clear the immediate work area of all but essential objects. A tote box
might fail to drain completely and bear a charge-but if kept to one side it will be harmless.
(4) Be sure to ground conductors, such as soldering iron tips, which touch leads or contacts
and can discharge directly into ESDS devices. (5) Recognize the limitations of supplementary
PAPER NO.3 197
work station. As a last resort, ordinary plastics which definitely cannot be excluded must be
treated with topical antis tat.
In general, equipment and materials must be selected carefully from the gamut of prod-
ucts available, and then handling techniques must be adapted to the products chosen.
Conclusions
1. Tests confirmed that static-dissipative bench tops are preferred to conductive sur-
faces not only for personnel safety reasons but because a slower discharge rate can reduce
damage to ESDS items.
2. A charged conductive tote box is only slightly more hazardous than a charged anti-
static box when an ESDS item enters the static field of the box. But when boxes rest on a
grounded static-dissipative work surface, a conductive box participates much more readily
in harmful discharges than does an antistatic box. Both types of boxes have their pros and
cons, and either is suitable if used properly.
3. Statically charged nonconductive plastics can present a field hazard when the appar-
ent charge is as little as 500 volts. However, a charge of as much as 13,000 volts is not
deliverable to an ESDS item by contact (discharge) when the field is suppressed to 300
volts apparent charge by proximity to a ground plane.
4. Under worst-case conditions, using a MOSFET with long "antennas," a rough equa-
tion for the minimum safe distance d (in inches) from a surface charged to voltage V is: d =
.[V/I.8. When the antennas are shielded, as they often are in practice on circuit-board
modules, the constant in this equation rises from 1.8 to at least 7.
5. Based on the above data, basic rules are: (1) Ground all conductors so that none can
have a charge of over 50 volts. (2) Check with a field meter and allow no more than 300
volts apparent charge on nonconductors such as ordinary plastics. (3) If static fields are
unavoidable, keep ESDS items d inches away from charged surfaces according to the worst-
case equation d = .[V/I.8.
6. Electrical air-ionizing blowers or nozzles were found not to present a field hazard if
ESDS items are kept at least 10 inches away from the corona-discharge points. For close
work, nuclear-type air nozzles are indicated.
7. A pulsed dc benchtop ionizer, with no fan, was found to be as effective as conven-
tional ac blowers. The lack of an airstream is advantageous. (Note: Slow-pulsed ionizers
later proved hazardous; see "Ionization" in Chapter 2.)
8. A room ionization system with dc emitters proved to be effective-about equal to a
typical bench-model blower at 5 feet- and relatively uniform regardless of the position of
charged objects on the bench surface under the emitters. At worst-case conditions (emitter
of only one polarity operating), an isolated conductor at bench level (capacitance 129 pF)
was not charged sufficiently to damage MOSFETs.
9. The possibility of a conductive masking tape of foillfiber construction is suggested.
The aluminum foil prevents a static field by charge drainage and voltage suppression.
10. A corona-discharge point mounted above a small-scale grit blaster reduced the charge
developed on circuit-board material from 700 volts to the harmless level of 60 volts.
II. A bead-chain wrist strap is effective but tends to ride up over sleeves. Snug-fitting
straps in which metal contacts the skin are recommended.
12. In out tests, MOSFETs were damaged in partially transparent metallized bags, under
worst-case conditions, when the bags were touched by an operator charged to an estimated
PAPER NO.3 199
(Note: Later research modified some of the above conclusions, as seen in the text of this
book. For example, we became more cautious about ionization.)
References
I. I. M. Kolyer and W. E. Anderson, "Pennanence of the Antistatic Property of Commercial Anti-
static Bags and Tote Boxes," Reliability Analysis Center EOSIESD Symposium Proceedings,
pp. 87-94, Las Vegas, NY, 1983.
2. B. A. Unger, "Electrostatic Discharge Failures of Semiconductor Devices," Reliability Physics
Symposium, April 1981.
3. I. R. Huntsman, D. M. Yenni, Jr., and O. E. Mueller, "Fundamental Requirements for Static
Protective Containers," presented at 1980 NEPCON/West Conference, Anaheim, CA.
4. D. M. Yenni, Ir., and I. R. Huntsman, "The Deficiencies in Military Specification MIL-B-81705:
Considerations and a Simple Model for Static Protection," presented at the Reliability Analysis
Center EOS/ESD Symposium, Denver, CO, 1979.
5. "Latent ESD Failures: A Reality," Evaluation Engineering Magazine, p. 80, April 1982.
6. I. R. Huntsman and D. M. Yenni, Ir., "Test Methods for Static Control Products," Reliability
Analysis Center EOSIESD Symposium Proceedings, pp. 94-109, Orlando, FL, 1982.
7. Electronics Industry Association Interim Standard No.5, January 1983.
8. "Electromagnetic Shielding Effectiveness," technical infonnation sheet from the Bemis Co.,
Inc., ESD Protective Materials Dept., undated, received 1983.
9. N. B. Fuqua and R. C. Walker, "ESD Controls Study, Final Report," Reliability Analysis Center
No. 01115-30-2, September 1981.
10. B. Unger, R. Chemelli, P. Bossard, and M. Hudock, "Evaluation of Integrated Circuit Shipping
Tubes," Reliability Analysis Center EOSIESD Symposium Proceedings, pp. 57-64, Las Vegas,
NV, 1981.
II. J. R. Huntsman and D. M. Yenni, Jr., "Charge Drainage vs. Voltage Suppression by Static Con-
trol Table Tops," Evaluation Engineering Magazine, March 1982.
12. I. R. Mileham and N. I. Safeer, "Selection of Static Eliminating Decorative Table Top Mats or
Laminates," presented at 1984 NEPCON/West Conference, Anaheim, CA. The material we
tested is No.6 in the tables of this reference.
13. D. C. Anderson, "ESD Control: To Prevent the Spark that Kills," Evaluation Engineering Maga-
zine, pp. 120-131, July 1984.
200 APPENDIX
14. R. G. Chemelli, B. A. Unger, and P. R. Bossard, "ESD by Static Induction," Reliability Analysis
Center EOSIESD Symposium Proceedings, pp. 29-35, Las Vegas, NY, 1983.
15. J. E. Berry, "Static Control in Bare Board Testing," Electronic Packaging and Production, pp.
161-162, August 1981.
16. C. F. Mykkanen and D. R. Blinde, "The Room Ionization System: An Alternative to 40 Percent
RH," Evaluation Engineering Magazine, pp. 76-88, September 1983.
17. "Electrostatic Discharge Training Manual," NAVSEA SE 003-AA-TRN-0IO, Published by Di-
rection of Commander, Naval Sea Systems Command.
18. "Arm Hair Pin-Pointed as New ESD Hazard," Evaluation Engineering Magazine, p. 70, May
1984.
19. GIDEP (Government-Industry Exchange Program) Alert No. D5-A-84-OI, May 21,1984.
20. GIDEP (Government-Industry Exchange Program) Alert No. MX-A-82-02, March 21, 1983.
21. J. M. Kolyer and W. E. Anderson, " Selection of Packaging Materials for Electrostatic Dis-
charge-Sensitive Items," Reliability Analysis Center EOSIESD Symposium Proceedings, pp.
75-84, Las Vegas, NY, 1981.
22. R. Euker, "ESD in Integrated Circuit Assembly," Static Digest, published by Static Control
Systems Division, 3M Co., July 1983.
23. G. E. Hansel, "The Production Operator: Weak Link or Warrior in the ESD Battle?" Reliability
Analysis Center EOSIESD Symposium Proceedings, pp. 12-16, Las Vegas, NY, 1983.
PaperNo. 4
Note: The following is a condensed version of this paper. See the Symnposium
Proceedings for the complete text.
Abstract
Commercial foil-containing laminate bags provide "Faraday-cage" protection for electro-
static discharge (ESD)-sensitive items but are opaque. However, foil may be perforated to
give see-through capability if the metal is thick enough and the stamped holes have out-
wardly directed burrs which may act as "lightning rods" to intercept discharges. A wire
screen serves the same purpose, with its high points presumably receiving the discharges.
This principle is applicable to various other containers, e.g., tote boxes. Thus, partial trans-
parency need not mean a sacrifice in ESD protection.
Introduction
Commercial "Faraday-cage" bags are intended to protect against electrostatic fields and
discharges. A secondary goal is protection against EMI (electromagnetic interference), spe-
cifically at 1-10 GHz per MIL-B-81705, Type I. These bags fall into four classes: (1)
foil-containing laminates which give good ESD and EMI protection but are completely
opaque, (2) metallized bags in which the metallized layer is thin enough to allow partial
transparency with a considerable sacrifice of ESD and EMI protection (also, external met-
allization can crack on handling, thus lowering shielding effectiveness), (3) bags with a grid
of conductive ink that are fairly transparent but give poor ESD protection, and (4) metal-
fiber-containing bags that are also fairly transparent but give poor ESD protection.
Therefore, our objective was a new bag laminate which combines the "best of both worlds"
to a considerable extent: see-through capability along with good ESD protection and EMI
protection in the 1-10 GHz (radar) range.
201
202 APPENDIX
Materials Tested
Nine commercial ESD-protective bag materials were tested (Table 1): antistatic poly,
carbon-loaded, ink grid, buried metallization A, external nickel, copper fibers, buried met-
allization B, buried metallization C, and foil laminate. Several of these products have been
described and characterized by us. 1.2,3
The other six materials in Table 1 were fabricated in the laboratory from antistatic
polyethylene (6 mils) and either aluminum foil, perforated aluminum foil, or metallic
screens.
Experimental Methods
The test conditions are listed in Table 2 and diagrammed in Figs. 1,2, and 3.
2 3 4 5 6 7 8
~
a
~
....
fA
204 APPENDIX
ROHCOlIDUCTlVE SURFACE
us ISTOI.
PR>BI
(TISTS 3.5.7)
HONCONDUCTlVE SURFACE
Bag Material
Antlatatlc Poly F
Carbon·Loaded F
Copper Fiber. P
Burled Metallization B P F f
2·mll Perforated P (F) f F
foil
,
Burled Metallization C P P (F) F F F F
5-mll Perforated p p p p F
foil
Rating
Results
The data are summarized in Table 3. Note the excellent results given by screens. After
publication of this paper a screen bag was commercialized; the aluminum screen was sand-
wiched between layers of antistatic polyethylene. Perforated foil was not as effective as
screen and just as expensive. so it was never commercialized.
Conclusions
1. Metallized. carbon-loaded. or ink grid bags have been reported not to withstand a
direct static charge above 25 kV. Our own tests indicate a much lower threshold for these
materials. Note that personnel walking on a carpet at 1~20 percent relative humidity gen-
erate typically 35 kV.
2. Two independent groups of investigators have recommended (EOS/ESD Sympo-
sium. 1984) that highly ESD-sensitive items should always be protected by metal-foil bags.
206 APPENDIX
3. We have now found that metal foil can be perforated and still shield against ESD and
radar. Wire screen is even better. Outwardly directed burrs on the foil, or prominences on
the screen, may act as "lightning rods" to intercept discharges.
4. The effectiveness of perforated foil or screen laminates as a moisture vapor barrier
would depend on the permeability of the plastic film used.
5. Preliminary designs with performed foil or wire screen would cost twice as much as
a commercial foil laminate. However, added expense might be justified in protecting valu-
able, high-reliability electronics.
References
1. J. M. Kolyer and W. E. Anderson, "Selection of Packaging Materials for Electrostatic Discharge-
Sensitive Items," Reliability Analysis Center EOSIESD Symposium Proceedings, pp. 75-84,
Las Vegas, NV, 1981.
2. J. M. Kolyer and W. E. Anderson, "Permanence of the Antistatic Property of Commercial Anti-
static Bags and Toite Boxes," Reliability Analysis Center EOSIESD Symposium Proceedings,
pp. 87-94, Las Vegas, NV, 1983.
3. J. M. Kolyer. W. E. Anderson. and D. E. Watson. "Hazards of Static Charges and Fields at the
Work Station." Reliability Analysis Center EOSIESD Symposium Proceedings. pp. 7-19. Phila-
delphia. PA, 1984.
PaperNo. 5
COST-EFFECTIVE METIIODS OF
TESTING/MONITORING WRIST STRAPS
Abstract
An ungrounded operator is extremely dangerous to ESD (electrostatic discharge)-sensitive
devices as illustrated by a test at high relative humidity (54 percent) in which the operator
damaged transistors by rolling his chair. An ESD-protective smock did not help to protect
devices. A maximum resistance from operator to ground of 10 megohms is recommended,
and this value was exceeded in 3 of 36 tests under manufacturing conditions. Because of
flexing of the cords and other abuse, wrist straps have a limited lifetime, but they are too
expensive to scrap in a replacement cycle and instead must be monitored. Periodic monitor-
ing is unsatisfactory because ESD-sensitive items will be exposed to an ungrounded opera-
tor for some period of time, even if only an hour, when a strap fails. Then material review
action might be required on all hardware handled since the last successful test of that strap.
Therefore, continuous monitors may be "the wave of the future." Commercial continuous
monitors work on the capacitance principle, but another method is a continuity check across
the operator's skin, up and down a two-conductor cord, and in and out two separate ground
connections. Prototype monitors of this two-conductor design are being tested in a produc-
tion environment. High-cost items, e.g. $1000 modules, especially deserve continuous
monitoring.
Introduction
This paper is presented in the form of slides shown at the Electrical Overstress Exposition
(EOE) in Anaheim, CA, January 21,1986. Accompanying each slide is additional discus-
sion and clarification as required.
Wrist straps have a limited lifetime I because of frequent flexing and other abuse. Since a
time-based replacement cycle would scrap many good straps and be wasteful, monitoring is
necessary. This monitoring is preferably continuous, because the failure of a wrist strap in a
207
208 APPENDIX
periodic check raises the questions: "How long ago did this strap fail?" and "Have products
been damaged in the meantime?"
As shown in the following outline, the danger of a failed strap was demonstrated, simple
methods for testing resistance were devised, straps in an assembly plant were checked, and
means of monitoring were investigated.
OUTLINE
• DAMAGE CAUSED BY LACK OF WRIST STRAP
• RESISTANCE VS CHARGE ON OPERATOR
• ALLOWABLE RESISTANCE, OPERATOR TO GROUND
• RESISTANCE TEST METHODS
• THROUGH STRAP
• OPERATOR TO GROUND
• EQUIPMENT
• MEGOHMMETER
• ZAPFLASH
• RESISTANCE TESTING DATA
• FACTORS IN RESISTANCE, OPERATOR TO GROUND
• MONITORING
• PERIODIC VS CONTINUOUS
• CONTINUOUS DESIGNS
• CAPACITANCE TYPE
• TWO-CONDUCTOR TYPE
• ADVANTAGES OF TWQ..CONDUCTOR TYPE
• CONCLUSIONS
The charge on the operator was read by means of a Trek Model 512 field meter aimed at
an electrically isolated I-ft2 aluminum plate to which the operator was attached by a wrist
strap. The temperature was nOE
The transistor was a Motorola 2N4351 MOSFET. The testing procedure and damage
criteria have been described. 2
PAPER NO.5 209
DAMAGE CAUSED BY
LACK OF WRIST STRAP
• GROUNDED OPERATOR IS LESSER OF TWO EVILS VS
UNGROUNDED OPERATOR
(KOLYER ET AL, EOSIESD SYMPOSIUM 1984)
A charged operator is a very great danger because he can damage ESD-sensitive devices
by direct discharge or "injection" rather than indirectly by induction from a static field.
The published data3 were obtained with an operator holding a probe attached to an elec-
trometer and an oscilloscope. The Trek meter and l_ft2 plate were used as described under
the preceding slide. Since 10 volts was our limit of detection. the result at I megohm agrees
with that obtained by the more sophisticated method: a I-megohm resistor in the wrist strap
allows negligible voltage on the operator.
ALLOWABLE RESISTANCE,
OPERATOR TO GROUND
These thresholds are based on the data of the preceding slide. A charge of 11 volts on
the operator is hannless to most devices at the present time; in the future this allowable
resistance might have to be lowered. A I-megohm resistor is presently standard for wrist
straps.
As an added precaution, a static-limiting floor finish (UL288, Hanson-Loran Chemical
Company) was used. The chair-rolling or foot-shuffling operator in the slide before last
would have generated little charge with UL288. In the test, a conventional acrylic finish
was on the tile floor.
A Beckman L-lO megohmeter at 10 volts was used. For quick tests, a Zapflash (see two
slides below) is suitable.
The same megohmeter at 10 volts was used. The metal can (tin-coated steel) is a conve-
nient, hand-filling probe.
The Zapflash was tested with various resistance levels. At 1.0 megohm the light (indicat-
ing continuity) was very bright. At 3.0 megohms the light was relatively bright, but at 5.2
megohms it was dim and at 6.2 megohms it was barely visible.
Various combinations of wrist straps and operators gave a total of 36 tests. These mea-
surements were made in a manufacturing situation. One operator had spliced the cord of his
bead chain, thus increasing its resistance from 1.0 to 1.8 megohms; this is not a serious
increase, but it might have been. Another operator was pleased with her loose-fitting expan-
sion band; resistance from her skin to ground was 30 megohms for this reason.
212 APPENDIX
These examples illustrate the need for operator discipline, especially in the absence of con-
tinuous monitoring of the straps.
The three out-of-spec cases occurred in a program of periodic (weekly) checks and illus-
trate the desirability of continuous monitoring. In a periodic check, the operator may slide
the band up his arm to tighten it or make some other adjustment to pass the test. An analogy
is annual inspection of automobile exhaust emissions; once the driver has passed inspec-
tion, he might readjust his engine to exceed the allowable level of smog-producing gases.
FACTORS IN RESISTANCE,
OPERATOR TO GROUND
• OVERSIZED EXPANSION BAND
• DRYSKIN
• PERIODIC
• CONnNUOUS
CONTINUOUS MONITOR,
CAPACITANCE TYPE
• CHECKS CAPACITANCE, EG, 150 pF
Commercial products include the Simco Wrist Strap Monitor Model M50A, the Ground
Gard from Static Prevention, Incorporated, and the GM-ICT from Westek. Charleswater
Products has a developmental product to be introduced early in 1986.
214 APPENDIX
Note that in the absence of continuous monitoring, all hardware handled since the last
successful test of a failed strap might be subject to a material review action.
CONTINUOUS MONITOR,
CAPACITANCE TYPE: DIAGRAM
1·MEGOHM
RESISTOR
This is a very simplified representation of the Simco Wrist Sgrap Monitor Model M50A.
CONTINUOUS MONITOR,
TWO-CONDUCTOR TYPE
• CHECKS LOOP ACROSS SKIN, INTO FIRST GROUND
CONNECTION, AND OUT SECOND GROUND CONNECTION
CONTINUOUS MONITOR,
TWO-CONDUCTOR TYPE: DIAGRAM
",
.......... ------------_/T
ADVANTAGES OF TWO-CONDUCTOR
TYPE CONTINUOUS MONITOR
At least one maker of the capacitance-type monitor expects to add a feature to check
ground connections. Also, the problem of the strap being deceived by a high-capacitance
object may possibly be overcome. With these improvements, the capacitance-type monitor
would more nearly equal the two-conductor type, so that a choice between the two might
largely depend on cost.
Cost effectiveness is not easy to calculate at this point, since the costs of periodic inspec-
tion and of hardware damaged due to lack of continuous monitoring must enter the equa-
tion. In general, periodic checks may be adequate for low-cost products, but continuous
monitoring is indicated for expensive assemblies. After all, most of us consider the wrist
strap our first line of defense, and if it fails we have dropped our guard, however briefly, in
the absence of continuous monitoring.
CONCLUSIONS
References
1. A. P. Hahl, "A Wrist Strap Life Test Program," Reliability Analysis Center EOSIESD Sympo-
sium Proceedings, pp. 94-96, Philadelphia, PA, 1984.
2. J. M. Kolyer, W. E. Anderson, and D. E. Watson, "Hazards of Static Charges and Fields at the
Work Station," Reliability Analysis Center EOSIESD Symposium Proceedings, pp. 7-19, Phila-
delphia, PA, 1984.
3. J. R. Huntsman and D. M. Yenni, Jr., ''Test Methods for Static Control Products," Reliability
Analysis Center EOSIESD Symposium Proceedings, pp. 94-109, Orlando, FL, 1982.
PaperNo. 6
Note: The following is a condensed version of this paper. See the Symnposium
Proceedings for the complete test.
Summary
Methodology is described for evaluating floor finishes which limit static buildup on per-
sonnel wearing ordinary shoes and drain charges from personnel wearing conductive foot-
wear or heel grounders. Procedures include triboelectric charging of various synthetic shoe
sole materials and an accelerated scrubbing test to predict durability. The objective is a low-
charging. easily maintained. and cost-effective product.
Introduction
Several commercial ESD-control floor finishes were evaluated in mid-1985. Some were
unsatisfactory in limiting static generation. while others had questionable floor finish
proterties such as scrub resistance; in general. non-ESD properties seemed compromised
by antistat addition. All the products were judged too expensive to be cost-effective. These
deficiencies prompted our study.
Our objective was a low-charging. easily maintained. and cost-effective acrylic coating.
and the search involved development of methodology for evaluating commercial finishes
and perfecting new and better ones. The emphasis was on simple. relatively inexpensive
equipment and realistic performance tests.
218
PAPER NO.6 219
We prefer the tenn "static-limiting" because this property, not conductivity per se, is
what is desired if ordinary shoes are worn. "Static dissipative" and "antistatic" refer to
surface resistivity ranges, and surface resistivity correlates imperfectly with triboelectric
charging of shoe soles, as will be seen. "Static-preventing" would be a misnomer or false
claim because all antistatic materials penn it some charging, especially of certain "problem"
synthetics.
When conductive footwear or heel grounders are prescribed, low surface resistivity is
the key ESD property, and tests for triboelectric charging are superfluous. However, our
methodology remains pertinent because it predicts durability in tenns of surface resistivity
as well as triboelectric charging propensity.
Experimental Methods
The various experimental methods are shown in Fig. I through 7. Fig. 8 shows that the
actual voltage in the walk test can deviate considerably from the meter reading because of
inertia of the needle. To be conservative, we multiply meter readings by a factor of 2.
Cost Effectiveness
Cost effectiveness is a vital "property" and must be considered in selecting a floor
finish.
In conjunction with ordinary ungrounded shoes, the use of a static-limiting floor finish,
like humidification and room ionization, is a supplementary procedure, as opposed to pri-
mary ESD control methods such as the use of wrist straps. Supplementary procedures are
intended to create a relatively benign background or "safety net" for imperfect primary
methods. A perfect acrobat doesn't need a safety net, and a perfect operator who always
wears his wrist strap doesn't need a static-limiting floor finish, but in reality the reduction
of background hazards is worth some expense. The question is: how much expense?
220 APPENDIX
ADJUSTABLE
SCALE
RUIIER
RING
STRING
(TO WINDLASS)
CROSS SECTION
CABLE
(TO RECIPROCATING
MECHANISM)
Table I compares estimated costs of using a static-limiting floor finish derived from
experimental product No.4 versus a standard nonantistatic fmish of similar chemical nature
(acrylic and polyethylenic polymers). It is interesting that the chemicals account for only a
small fraction of the annual cost of floor maintenance: 4.2 percent for the conventional
finish and 6.5 percent for the static-limiting finish. For 10,000 ft2 of floor space, the pre-
mium for using the static-limiting finish is $235/year, which seems a low cost for any sig-
nificant additional EDS protection.
Note that the special finish could help justify saving mishandled items in material review
actions; for example, if an operator touched an ESD-sensitive item without wearing his
wrist strap, the argument might be made that the probable charge on his skin was below the
sensitivity level of the item based on walk-test monitoring data.
222 APPENDIX
f ....
, ...-c..': \
':.:.~
.... ' !'.' FARADAY
.........
.."':' . . "
.~..
":
. .
CUP
AUJillrul
SHEET (1 "2)
STAND-OFF INSULATOR
(POLYETHYLENE BEAKER WITH
SLIT FOR ALlJUNUM SHEET)
0
1000 0
.... 0
~ 0
i
-i= 0 00
f
t
500 cP
0
...
vi
-'
i
0
VOLTS, ACTUAL
Premium o 235
An advantage of the use of static-limiting floor finish with ordinary shoes as a supple-
mentary procedure is that no disturbing or even visible change is made in the work environ-
ment. In contrast, humidification can cause discomfort due to "mugginess," and room
ionization has been known to create anxiety about its imagined electrical effects on people.
However, thorough testing must assure that voltages are in fact sufficiently controlled with
the great variety of shoe soles which personnel may wear. For example, certain vinyl com-
pounds may be "problem" materials, as indicated by high charges in both the walk test and
laboratory tests.
Static-limiting floor finishes may be applied to conductive tile without destroying its
grounding capability. For example, two coats of the product derived from experimental
product No.4 increased the surface resistivity of a popular brand of conductive tile from
1 x 107 to 3 X 107 ohms/square and increased the resistance to ground (through a 5-lb,
2.5-inch-diameter NFPA 56A electrode) from 7 x 106 to 3 X 107 ohms at 10 volts. These
threefold or fourfold increases do not prevent rapid electrical drainage from conductive
footwear, and the finish provides wear-resistance and an attractive gloss.
Conclusions
1. Nine products commercially available in mid-1985 were evaluated to illustrate our
methodology and found to be generally unsatisfactory in performance and too high in cost.
Resistance to scrubbing and foot traffic seemed compromised by antistat addition. How-
ever, formulations are easily changed, and today's versions may be improved.
2. The products tested were not literally "zero-charge." They limited triboelectric charg-
ing of shoe soles but did not stop it. As of May 1986, some suppliers felt that ordinary,
ungrounded shoes could be worn without charging enough to damage ESD-sensitive de-
vices, but others felt that conductive footwear or heel grounders were needed to drain charges
including these created by movement of clothing.
3. The "acid test" of a static-limiting floor finish is the measurement of charges on
people who walk on it. When ordinary shoes are worn without conductive straps, surface
resistivity correlates imperfectly with charging and can be misleading.
4. The ever-changing shoe sole market presently includes a variety of synthetics such
as polyurethane, vinyl, ethylene-vinyl acetate, and styrene-butadiene rubber (SBR). These
materials vary in susceptibility to triboelectric charging; vinyls are high and polyurethanes
tend to be low (hence they are often used, with additives, in ESD-control footwear). Also,
different synthetics may respond in varied and unexpected ways to different floor finishes.
Therefore, an averaging effect is desired, and screening of candidate finishes should in-
clude several synthetics, not just Neolite (an SBR compound) as in AATCC Test Method
134-1979.
5. Compared to synthetics, leather soles are a minor problem. They hold only 15 per-
cent of the market, are usually used only on expensive shoes, and in general triboelectrically
charge much less than synthetics.
6. In laboratory tests, cylinders of various synthetics were rolled down a ramp of coated
tile, or pieces cut from actual shoe soles were dragged across the surface; in both cases the
specimens fell into a Faraday cup to measure the charge. The results roughly agreed with
voltages observed on people with random shoe soles when a variety of synthetics and sev-
eral people were measured to get average effects.
PAPER NO.6 225
Note: This paper appeared as an article in EOSIESD Technology Magazine, April 1987,
page 20:
~
. ... By John M. Kolyer, Rockwell International and
.-....... Dale M. Cullop, Hanson loran Chemical Co.
,1~::=; These tests are good candidates for a standard method of
~'!:...... ___ evaluating the static·limiting properties of floor finishes.
PaperNo. 7
Until recently. all tote boxes for ESD control were made of polyolefin and either topically
applied antistat. extruded-in antistat. or extruded-in graphitic (conductive) carbon. Each of
these materials has its limitations.
Topically applied antistat is only an expedient because the antistat wears off after some
undetermined period of use. Also. a wall of plain polyolefin. even with an antistatic surface.
provides little Faraday-cage shielding protection from external static fields or discharges
(Table I).
Extruded-in antistat provides a longer life than topically applied antistat because it
continues to bleed to the surface for some time. creating a weakly conductive sweat layer
from atmospheric moisture. However. handling. heat. contact with paper products. or
exposure to solvents will eventually deplete all the antistat (Ref 1). Again. shielding is
poor.
Extruded-in conductive carbon offers the advantage of permanence. but it has several
problems. A carbon-loaded polyolefin tote-box wall is conductive enough to endanger people;
a current of over 100 rnA, which is usually fatal (DOD-HDBK-263). can be carried at 110
V (Table 2). If the conductive box itself is charged it is more dangerous to devices than
antistatic or nonconductive boxes (Table 3 and Ref 2).
A high-conductivity surface is also dangerous to devices (Ref 3). especially if the opera-
tor should be charged and touches a sensitive lead to the box (Table 4).
However. the wall of a carbon-loaded polyolefin box is not conductive enough to be a
good Faraday cage; note that the volume resistivity of carbon-loaded polyolefin is on the
order of 102 Q-cm versus 1~ Q-cm for aluminum foil. Refs 3 and 4 agree that highly ESD-
sensitive components should always be protected by metal-foil bags. not carbon-loaded
bags or bags with thin. see-through metallization. This conclusion is applicable to other
containers such as boxes. Furthermore. carbon-loaded polyolefin imparts high triboelectric
charges to nonconductors stroked on it (Table 5 and Ref 1) and sloughs conductive particles
that could fall into open microelectronic devices and cause shorts.
226
PAPER NO.7 227
*For example, 2/5 (IS) would mean that five MOSFETs were tested, two were damaged and one of those damaged
was shorted.
**I-in. gap between each electrode and inner surface of box.
and vinyl-metal sheet. Multiple layers are necessary because a homogeneous wall is not
capable of providing both a safe antistatic or nonconductive surface and Faraday-cage
protection.
Commercially available fiberboard-foil construction ("Corshield" by Conductive Con-
tainers Inc.) consists of aluminum foil sandwiched between layers of fiberboard. The fiber-
board has a naturally antistatic surface but is covered with an antistatic coating to seal in
Antistatic 0/5
Carbon-loaded polyolefin 3/5 (15)
Corshield 0/5
Vinyl-aluminum sheet 0/5
Bare aluminum sheet 4/5 (35)
sulfur-containing impurities that might tarnish silver-plated leads. The foil provides su-
perior Faraday-cage protection. For example, a discharge from a key held by a person
charged to 8000 V caused no damage to field effect transistors (MOSFETs), whereas an
impractically heavy wall (0.140 in.) of carbon-loaded polyolefin allowed transistor dam-
age (Table 1).
Even the Tesla coil test was passed when the Corshield box was folded to give two layers
of foil (Table 1). In previous work (Ref 5), only constructions with heavy foil or metal
screen passed this test. Note that the Tesla-coil test is the worst case in electrical stress and
positioning of the device in the box. However, this test is not the worst case in statistical
significance because only five parts are tested for a "pass" rating. Furthermore, our Tesla-
coil test does not use worst-case acceptance criteria because subtle damage may not be seen
by the curve tracer, and the MOSFETs used are sensitive to 100 to 200 V whereas some new
devices may be affected by only 20 volts. If a cost-effective material can pass this test, we
would use that material.
PAPER NO.7 229
Fiberboard-foil costs less than other materials. Also, it can be stored easily as flat sheets
and then folded into boxes when needed. Its only major defect is its limited durability, but
heavier fiberboard will probably prove sturdy enough for most applications.
The vinyl-metal sheet design should satisfy the market niche requiring extreme durabil-
ity. The metal, either steel or aluminum, provides high structural strength and is coated on
both sides with tough vinyl, e.g., 0.010 in. thick, by either lamination or powder-coating.
The resulting nonconductive surface is safe for people. Also, in a contrived charge-device
model test (Table 6), the nonconductive surface was even safer for devices than an antistatic
surface.
The relatively heavy metal wall, 0.0375-in. aluminum (20 guage), is a virtually impreg-
nable Faraday cage and suppresses the voltage of a static charge, no matter how much the
surface may be stroked, so that the box never has an appreciable E field when the metal is
grounded via bare metal feet on the bottom. In our test, the effectiveness of draining off
charges onto an antistatic bench-top was better for a vinyl-metal sheet box than for a
carbon-loaded polyolefin box (Table 7).
However, good contact with the bench surface, aided by flatness of the bottom of the
box, can be critical. The slightly flexible Corshield box benefited from being conformable
and lying flat, whereas the rigid antistatic or carbon-loaded polyolefin boxes were slightly
"dished" (concave) so that only edges or corners made contact.
The nonconductive vinyl surface's only defect is that it can triboelectrically charge con-
ductors, whereas an antistatic or conductive surface cannot (Table 5). However, charging of
nonconductive surfaces, e.g., conform ally coated circuit-board modules, seems a more im-
portant issue, and nonconductive vinyl was less of an offender than carbon-loaded polyolefin
(Table 5).
Test Methods
Below are the various test methods that were used to derive the results shown in each of the
tables accompanying this article.
Shielding/Discharge Test
(Results in Table 1)
An electrode 1.5 in. square was taped against the inside surface of one wall of the tote box and
connected to the substrate-case lead of a Motorola 2N4351 MOSFET, and a similar elec-
trode was taped against the inner surface of the bottom of the box and connected to the gate
lead. Then a discharge was made to the outer wall of the box over the electrode. This dis-
charge was from a charged person holding a key, from a capacitor connected to a resistor
and a steel probe, or from a Tesla coil operated for 30 sec. The box sat on a grounded plate
during the test.
Summary
Table 8 summarizes our evaluations. Together, the two new multilayer boxes should satisfy
all in-plant handling needs for an ESD-control program. These boxes are able to afford
secure Faraday-cage protection for even the most sensitive items when electrically continu-
ous lids are being used.
TableS.
TOlE BOX TYPE
Cost advantage depends on many factors, most notably the number of units produced
and the fabrication method. In general, a vinyl-aluminum sheet box would be competitive
with an injection-molded carbon-loaded polyolefin box. A vinyl-steel sheet box, though
cheaper than aluminum, would be almost three times heavier in the same gauge. In contrast
with the above choices, the fiberboard-foil box is inherently inexpensive (Table 8).
Last of all, fiberboard boxes (Corshield) are commercially available at this time: how-
ever, vinyl-metal sheet designs are still in the prototype stage.
Another interesting multilayer design is one which utilizes an aluminum screen or foil
sandwiched between layers of hard vulcanized fiber (Table 1). The vulcanized fiber mate-
rial is naturally antistatic, even at low humidity (5 x 1011 O/sq after seven weeks' storage at
nop and 12% RH). As in the case of the vinyl-metal sheet box, the commercial success of
this design would depend upon the development of a practicable fabrication method, but
both appear to be excellent alternatives for future ESD control.
References
1. J. M. Kolyer and W. E. Anderson, "Pennanence of the Antistatic Property of Commercial Anti-
static Bags and Tote Boxes," Reliability Analysis Center EOSI.ESD Symposium Proceedings,
EOS-5, Las Vegas, NY (1983): 87-94.
2. J. M. Kolyer, W. E. Anderson and O. E. Watson, "Hazards of Static Charges and Fields at the
Work Station," Reliability Analysis Center EOSIESD Symposium Proceedings, EOS-6,
Philadlephia, PA (1984): 7-19.
3. R. O. Enoch and R. N. Shaw, "An Experimental Validation of the Field-Induced ESO Model,"
Reliability Analysis Center EOSIESD Symposium Proceedings, EOS-8, Las Vegas, NV (1986):
224-231.
4. G. C. Holmes, P. J. Huff and R. L. Johnson, "An Experimental Study of the ESO Screening
Effectiveness of Antistatic Bags," Reliability Analysis Center EOSIESD Symposium Proceed-
ings, EOS-6, Philadelphia, PA (1984): 78-84.
5. J. M. Kolyer and W. E. Anderson, "Perforated Foil Bags: Partial Transparency and Excellent
ESO Protection," Reliability Analysis Center EOSIESD Symposium Proceedings, EOS-7, Min-
neapolis, MN (1985): 111-117.
PaperNo. 8
Note: The following is a condensed version of this paper. See the Symposium
Proceedings for the complete text.
Summary
In the robot-based, computer-controlled installation of devices such as OP-AMPs and CMOS
on printed circuit boards, an exaggeratedly sensitive model assembly served to locate ESD
hazards and later monitor the process for their absence. This "coupon" approach applies
uniquely to automation with its freedom from unpredictable operator error.
Introduction
The objectives were (l) to reveal ESD hazards in an automated, robot-based manufacturing
process for mounting components on printed circuit boards (PCBs), (2) to remove or con-
trol these hazards, and (3) to monitor the process for continued ESD safety.
The process and the ESD-sensitive components of concern will be described. Then ESD
damage mechanisms, hazards, and remedies will be reviewed, and possible hazards will be
related to process steps. Next, the "coupon principle" of using an exaggeratedly sensitive
model assembly to locate and evaluate ESD hazards will be explained and illustrated with
specific coupon designs. Finally, test data obtained with coupons will be discussed. Haz-
ards were eliminated to give an ESD-safe process which could be checked periodically by
means of fresh coupons.
233
234 APPENDIX
Coupon Principle
An automated process, being free of unpredictable operator error, uniquely lends itself to
a "coupon" approach in which an exaggeratedly ESO-sensitive model assembly is passed
through the process as a coupon to reveal ESO hazards. Various coupon designs used in
our tests all contained a 2N4351 MOSFET. The assumption was that if this MOSFET
showed no measurable damage, in terms of gate-source threshold voltage and shape of
the curve on a curve-tracer, there would be no subtle damage (not detected in quality
control tests) to devices on a real module. Undetected damage could lead to latent failure
and hence shorter lifetime and reduced reliability. Latent failures are known to occur with
CMOS devices.
A coupon is shown in Fig. 2, and a close-up of the MOSFETs on this coupon is shown in
Fig. 3.
Experimental Methods
The unedited paper describes how the coupons were tested before and after passing through
various process steps. ESO hazards revealed by the coupons were removed, and fmally
coupons were passed without damage through the entire process and the AMC was declared
ESO-safe.
@ ® CD
VAPOR
o o
O6 PHASE
SOLDER ING
AMC
\
7"
.:0-
--
ROBonc
O4 ASSEMB LY
WORK 7'
STAnON
SOLDER PASTE
APPLlC AnON
WORK STATION
~
;II
:u
~
011
Fig. 2. Coupon.
Etll,,,,'s Note: AI,b," IIUIO"""- bly 11M to IISSU" 111111 qUllllty de-
etlllSsembly ,limllUl'" "","y of ,II, ,p/It 0 mucII "'uc,tI IIu""," _k
p,06/,ms CIIUS,tI by ,II, mowmelll fDlU, Rockwtll 1"'HlUltlllllGl IIltd
11M mlS/obs of IIumllll bellllS. til, _tllods d"cribetl below. TII,y
,II,n nmllllll ,II, probltm of tI"ecl- "" btu,d 011 ,II, coupOll prlllcipl"
illl 11M cOII,ro/lill, ESD I" ,,., lib- w"",11I 0 ,ubtu_bly _ Itllll-
se/lCt of IIumllll supI/'V/II011. III II tlw '0 ESD tllllll ,1101' '0 be _
limltttl WilY, ESD COIItroi becomlS fllCtll"d Itllnds III f", til, IlIltH till,..
l1li t1u,cilt III nmolt Stllli/ll. III, qUIIlijiClltlOllIlM "9UIIlijiclltlOll
III ",tI" '0 spo, IIn11S of ESD vul- of til, f«lIity.
",,,,bill,y tlurilll lIu'omlltttlllssem-
PaperNo. 9
Abstract
One commercial brand of MIL-B-81705, Type n film contained organic acid and caused
corrosion of solder-coated device leads on circuitry. However, solderability was unaffected
in accelerated tests. Even acid-free antistats can stresscrack polycarbonate, fog instrument
mirrors, weaken adhesive bonds, and discolor epoxy paint. These problems are reduced by
a new generation of Type n films.
Introduction
Contamination of surfaces by antistats from antistatic films can cause (l) corrosion of sol-
der (when the antistat contains organic acid), (2) stresscracking of polycarbonate plastic,
(3) fogging of instrument mirrors, (4) loss of strength of adhesive bonds, and (5) discolora-
tion of epoxy-polyamide paint Other adverse effects, e.g., contamination of gyroscope balls,
have been discussed in the literature. I
Corrosion of Solder
In early 1987 a Government-Industry Data Exchange Program (GIDEP) Alert2 described
corrosion of solder-coated leads by traces of n-octanoic acid in one commercial brand of
"pink poly" antistatic film (MIL-B-81705, Type II). Therefore, we ran tests to determine
the amount of acid present and the nature of corrosion caused by it
In our tests, extraction of a suspect (acid-contaminated) bag with refluxing ethanol gave
acidity corresponding to 514 ppm by weight n-octanoic (caprylic) acid, in agreement with
400-650 ppm found by methylene chloride extraction by other investigators.3 The acidity
238
PAPER NO.9 239
of a sample of neat (undiluted) antistat from the manufacturer of the suspect film corre-
sponded to 8.9 wt. % n-octanoic acid, and other investigators3 found up to 15 wt. %. From
= =
our acidity data, the wt. % of antistat in the film (514) (100)/89,000 0.6, which is
roughly the level the manufacturer said had been added to the polyethylene by blending and
extrusion.
Our corrosion testing is summarized in Table I. Degradation of solderability has been
suggested2.3 but was not found in our tests. When painted with neat acid-contaminated anti-
stat, without aging, leads remained wettable by solder, and strength of solder joints was
normal. In long-term tests, circuit-board specimens and resistors with freshly tinned copper
leads were stored in suspect bags at elevated temperature and humidity or immersed in
moist antistat or moist n-octanoic acid at 120°F. The conclusion was that corrosion of pack-
aged items does occur but is only superficial (dull, gray tarnish) because so little acid reaches
the solder surface. There was no effect on solderability except in the drastic test with neat n-
octanoic acid where enough corrosion products built up to cause dewetting of solder. When
the acid content was very low, e.g., 0.05 wt. % as n-octanoic acid in the neat antistat, antistats
of the diethanolamide type ("proposed antistat" in Table I) were noncorrosive, and the
cosmetic problem of tarnishing disappeared. Most items already stored in acid-contaminated
packaging were left as is, but new items were packed in acid-free containers. Panicky
repackaging that could deform leads or cause electrostatic discharge (ESD) damage was
avoided.
Stresscracking of Polycarbonate
Polycarbonate, which is notably subject to stresscracking, was attacked by all antistats tested,
though to different degrees. Stresscrack agents for polycarbonate include commonplace
liquids such as ethylene glycol, heptane, and com oil in the literature4 and, in our tests, pink
liquid hand soap, No. 30 motor oil, and RMA (rosin, mildly activated) solder flux. There-
fore, it was not surprising that liquid antistats, which have considerable solvent power, were
also stresscrack agents. At 2.4% tensile strain on the surface of ~-inch-thick bent strips of
polycarbonate, ethoxylated tertiary amine or diethanolamide-type antistats caused cracks to
propagate completely through the plastic within 24 hours. At 0.3% strain, the antistats still
caused catastrophic cracking, but the time was much longer: 2 to 7 weeks.
When conventional MIL-B-81705, Type II films or new-generation "dry" films (dis-
cussed below) were rubbed on the surface of bent polycarbonate strips at 2.4% strain, the
new films caused significantly less cracking than did the "old." The explanation is that the
new films have minimal liquid antistat on their surface. Thus they may be safe when the
stress level on the polycarbonate surface is very low, but the cautious recommendation is to
avoid contacting polycarbonate items with any plastic that contains a liquid antistat.
the bag (as detennined by rinsing a fresh bag) had migrated as vapor to contaminate the
plate. The same level of transfer is estimated to occur in 6 months at room temperature,
based on a Cox chart which relates vapor pressure of organic compounds to temperature.
The solution to the fogging problem obviously is to use ESD-protective packaging mate-
rial free of fugitive, surface-seeking additives. Another example of the undesirability of
transferable antistats is the packaging of items such as resistors whose surface must main-
tain high electrical resistivity. In this instance the weakly conductive "sweat layer" that is
desired on the package surface becomes a liability on the surface of the stored item.
5. Assuming that their permanence6 proves adequate, the new-generation bags and
bubble-wrap will be a great impovement over the old "wet" or "greasy" products but are not
a panacea because the difference is quantitative, not qualitative. Packaging materials free of
liquid antistat still must be used in critical applications such as shipment or storage of poly-
carbonate items. Potential users should run tests to find the most cost-effective packaging
system.
Acknowledgment
The assistance of R. Rushworth, D. R. Violette, A. E. Carmellini, and T. J. Hester in per-
forming chemical analyses and corrosion tests is greatly appreciated.
References
I. M. K. Bernett, H. Ravner, and D. C. Weber, "Electroactive Polymers as Alternate ESD Protec-
tive Materials," EODIESD Symposium Proceedings, pages 115-119, orlando, Fl., 1982.
2. GIDEP Alert No. E9-A-86-02, on "Materials, Plastic, Antistatic," issued January 23, 1987.
3. J. Anderson, R. Denton, and M. Smith, "Antistatic Polyethylene Package Corrosion," EOSIESD
Symposium Proceedings, pages 36-40, Orlando, Fl., 1987.
4. Modern Plastics Encyclopedia, 1978-79, pages 528-529.
5. "Standard Test Method for Strength Properties of Adhesives in Shear by Tension Loading (Metal-
to-Metal)," ASTM D 1002-72 (reapproved 1983).
6. J. M. Kolyer and W. E. Anderson, "Permanence of the Antistatic Property of Commercial Anti-
static Bags and Tote Boxes," EOSIESD Symposium Proceedings, pages 87-94, Las Vegas, NY,
1983.
PaperNo. 10
Note: The following is a condensed version of this paper. See the Symposium
Proceedings for the complete text.
Abstract
Controlling voltage on personnel is a vital and demanding requirement in ESD protection.
New data show that the allowable resistance to ground must be lower than was thought.
Furthermore, wrist-strap systems often fail, so monitoring must be continuous; resistive-
type monitors are best. A cost-effective program is described.
Introduction
Controlling voltage on personnel has long been recognized as of primary importance in
preventing electrostatic discharge (ESD) damage to electronics. The reason for concern is
that operators' fingers often touch leads or contacts of ESD-sensitive (ESDS) devices, and
a flow of current from person to device can destroy or "wound" it by the direct injection
(01) mechanism. A "wound," passing undetected in routine tests, may result in latent fail-
ure. Thus, improperly grounded people are the principal ESD hazard or at least prime sources
of ESD for destroying parts.
A resistance of about I megohm is generally agreed upon for current-limiting resistors to
protect operators wearing wrist straps. However, questions remain: (I) What should be the
244
PAPER NO. 10 245
maximum allowable resistance to ground (ARTG) for protection of ESDS items as opposed
to protection of operators? (2) How can the triboelectric charging of operators be mini-
mized to make the resistance to ground less critical? (3) Should this resistance be checked
at intervals or continuously? (4) If a continuous wrist-strap monitor is used, which design is
best? (5) How can skin-voltage control and monitoring be made part of a cost-effective
program? We have attempted to answer these questions based on experience as well as
experimental data.
Experimental Methods
An operator sat at a workbench and charged himself either by stroking a sheet of Aclar film
on the bench surface or by shuffling his feet. The resulting data are shown in Fig. I. Fig. I
represents worst-case charging conditions, but surveys of operator grounding in manufac-
turing areas showed many bad cords and instances of high resistance to ground because of
operators' dry skin. Therefore, we recommend continuous monitoring with units like the
one illustrated in Fig. 2 and 3.
A Cost-Effective Program
Including Skin-Voltage Control
We emphasize basic, proven methods of ESD control and deemphasize methods which are
of questionable value and may even be part of the problem. The major elements in our
suggested program for ESDS items in Class I of MIL-STD-1686A are:
1. Operator Disciplines
A sufficiently skilled operator could work with almost no special equipment, and, con-
versely, the most expensive appliances won't prevent damage by an ignorant or careless
operator. Remember that a grounded operator is only the lesser of two evils versus an un-
grounded operator because a charged device could be damaged by rapid discharge to the
grounded operator's finger (the Charged Device Model or CDM). To maintain their skills,
operators must be retrained periodically.
CONTINUOUS MONITOR,
TWO-CONDUCTOR TYPE: DIAGRAM
.ALARM
\' "',,
\
,
4. Exclusion of Nonconductors
from Workstations
This precaution must be rigidly enforced, with frequent checks. A field meter and surface
resistivity meter are essential. Nonconductive surfaces which can't be eliminated may be
treated with topical antistat, necessarily chloride-free in some applications, and so labeled.
Bench tops should have static-dissipative surfaces and a buried ground plane to suppress
voltage.
7. Minimization of Triboelectric
Charging of Operators
Static-limiting floor finish is very cost-effective as seen in Paper No.6. Choose a brand
based on its ability to limit tribroelectric charging of operators wearing shoes with various
types of soles. We recommend a "walk test" with a dozen of your personnel: the voltage on
them is measured with a field meter or with a Voyager PVT-300 Personnel Voltage Tester.
Surface resistivity of the finish should be at most 10" ohms/square, but note that surface
resistivity does not correlate with triboelectric charging for materials in general and corre-
lates imperfectly for floor fmishes. No matter what finish is used, operators should not
fidget or shuffle their feet unnecessarily; there is no sense in "looking for trouble."
PAPER NO. 10 249
8. Use of Ionizers
Be cautious with room ionization as mentioned above. Cleanrooms, by the way, are a
special case; an ionization system, correctly engineered, may reduce the number of airborne
particles. Use local ionizers only for a clearly defined purpose, such as to control charge
buildup in processes like grit-blasting. Never use ionization as a vaguely conceived "safety
net."
9. Humidification
In very dry areas, humidification is desirable because it makes antistatic materials with
"sweat layers" function better and it reduces triboelectric charging. But don't let it build
false conficdence, and beware of corrosion or other problems as mentioned above. As with
ionization, the use of humidity control must be judicious.
By "myth" we mean a more or less widely held belief that sounds credible but proves false
when examined critically. Many such beliefs have clouded ESD-control standards. First,
we'll discuss some old myths, then some new ones.
Among old ESD-control myths, now put to rest in most people's minds by long repeti-
tion of the truth, are the following:
Wrist straps render human skin harmless. The truth is that a grounded operator
versus an ungrounded operator is only the lesser of two evils because of the Charged
Device Model (CDM). A charged device may discharge to the grounded operator's fin-
gers and be damaged. If the operator is ungrounded and chances to be at the same poten-
tial as the device, no damage will occur. So grounding, in this unlikely case, would be an
"evil."
Components are safe once they're mounted on circuit boards. Circuit lines may act
as antennas to increase the likelihood of ESD damage and discharges caused by the Field
Induced Model (FIM). Also, more than one device may be damaged by Direct Injection (01)
and electrical overstress (EOS) when a single discharge occurs to a branched circuit line.
Carbon-loaded plastic provides a Faraday cage. Carbon-loaded plastic, which was
the earliest ESD-protective material (in the sense of not holding triboelectric charges) is not
conductive enough to shunt a direct discharge safely away from an ESD-sensitive item
inside a bag. Instead, the discharge acts through the wall of the bag and can cause damage if
the device leads are in a vulnerable position.
High humidity solves all ESD handling problems. Triboelectric charging is merely
reduced by high humidity, not eliminated. Operator disciplines cannot be relaxed when
humidification is provided. Humidification is expensive and may introduce problems such
as corrosion or operator discomfort.
Ionization allows the use of plain, nonconductive plastics. This statement is untrue
because nonconductive plastics can become triboelectrically charged and cause ESD dam-
age by the FIM before the ions have time to neutralize the charge.
Ionization is a substitute for grounding of personnel (by wrist straps, for
example). This egregious idea was actually promoted by an equipment supplier several
years ago. The fact is that operators can become triboelectrically charged (by shuffling their
feet, for example) and damage ESD-sensitive items with 01 by touching them despite a
shower of ions that neutralize their skin relatively slowly.
251
252 APPENDIX
Devices with protective networks are not ESD-sensitive. Protective networks raise
the threshold for ESO damage to a limited extent, such as from 100 V to 2,000 V by the
Human Body Model (HBM), not infinitely. In other words, the sensitivity class in accor-
dance with MIL-STD-1686A may be raised (from Class 1 to Class 2 in the previous ex-
ample), but the device will remain ESO-sensitive. In our system of ESO control, the
precautions wouldn't change because we treat all ESO-sensitive classes the same as a mat-
ter of convenience; maintaining "islands" of different disciplines within a plant is impractical.
So much for old myths. Unfortunately, a second generation of myths has arisen, created
partly by vendors' biases. We'll debunk several of these new, less obvious, and more insidi-
ous untruths. The following discussion includes our own biases, of course, but these are
based on data and not influenced by sales concerns. Readers are welcome to contact us c/o
ESDlEOS Technology on points of disagreement.
Background Ionization Is needed when highly ESD-sensitive Items are being
handled. This statement suggests a cure-all or panacean view of ionization, but ioniza-
tion is not a cure-all. In fact, ionization for no valid, well-defined reason can be useless and
actually become part of an ESO problem by creating unbalanced charges on ungrounded
conductors. The possible occurrence of such charging has been noted for an electrically
unbalanced AC ionizer (Ref. 1). Hence, ionizer manufacturers have shown much concern
about balancing which in many models is now automatic.
In contrast to humidification, which reduces triboelectric charges by inhibiting them in
the first place, ionization can neutralize a charge only after it has formed. Thus, ionization is
a process of correction rather than prevention. However, it does have its place in ESO
control when used judiciously. Local ionization helps control triboelectric-charging pro-
cesses, such as peeling tape from a roll, and room ionization does control standing charges
on common plastics and garments. Also, room ionization may have an incidental air-clean-
ing effect in cleanrooms, but potential users should run tests or at least review published
data to prove that vendors' claims of reduced particles are true.
While malfunctioning electrical ionizers could charge isolated conductors as we men-
tioned earlier, nuclear ionizers have a possible problem of polonium 210 contamination.
One brand of nuclear ionizers was recalled by the Nuclear Regulatory Commission in Feb-
ruary 1988. Actually, this danger is remote. In terms of cost, nuclear and electrical ionizers
may be similar; you must make your own calculations.
In sum, ionization is not a harmless vitamin pill but strong medicine that has possible
dangerous side effects and should be used only when risks are carefully controlled and
outweighed by benefits.
The resistance-to-ground for personnel may be 100 mn or more because charges
drain otT even at very high resistance as shown by steady-state tests. The fallacy of
this statement is that a steady-state test is unrealistic. The voltage spikes on seated operators
being triboelectrically charged under real conditions, such as shuffling feet or brushing a
garment sleeve over the work surface, must be measured. We did these measurements and
found that MOSFETs were damaged at only about 8 mn resistance-to-ground when an
operator stroked a plastic film, representing a sleeve, on a workbench under conditions
where the individual generated -3,800 V when ungrounded (Ref. 2). The implications of
such data are serious, and we recommend an allowable resistance-to-ground of 10 mn
maximum (with the usual 1 mn minimum for safety). The maximum resistance should be
controlled with a continuous wrist-strap monitor (see the next myth).
Occasional checks of wrist straps are sufficient. Logically, this statement is inde-
fensible because almost everyone agrees that 01 from operators' fingers is a major, or
PAPER NO. 11 253
even the major, ESD-damage mechanism, and it is well known that this damage occurs in
a matter of nanoseconds. Therefore, we can't drop our guard for even a fraction of a
second-monitoring of skin resistance-to-ground must be continuous. Even if wrist straps
were checked every hour (the real interval is usually daily or even weekly), when a failed
strap is discovered, all the ESD-sensitive items handled by the wearer during the last hour
are suspect. The only reasons continuous monitoring is illogically avoided are its per-
ceived cost and trouble. But, in reality, this monitoring is practical. More and more con-
tinuous monitors are being used in industry, especially on government programs, and we
have found certain models to be quite cost-effective. We prefer the two-conductor resis-
tive rather than the capacitance type.
Note that dryness of operators' skin may raise the resistance-ta-ground far above 10 mO,
even with a perfect wrist strap and good skin contact. The monitor also reminds operators to
apply antistatic lotion, usually once or twice during a shift.
All ESD damage is caused by a spark. As a bare, unqualified statement, this is mean-
ingless. The first problem is semantic: What is a "spark"? Is it a visible, audible discharge?
If so, an ungrounded operator's finger at 1000 V will damage a MOSFET, for example, by
touching a lead even if there is no spark visible or audible.
Even if a spark is defined as "any discharge causing damage by 01," electrical fields can
act through space to destroy MOSFET's by the FIM with no spark at all. In the latter case,
the spark occurs inside the device, but this definition of spark as "an internal discharge"
results in a useless truism that all ESD damage is caused by discharges inside devices,
which gives us no guidance in preventing these internal discharges. We want to know what
happens outside a device to cause the "spark" inside. Therefore, a more productive state-
ment is: "All ESD damage is caused by fields (FIM) or discharges (01, CDM)."
The only useful idea in the "spark theory" is that a rapid discharge may cause damage
under conditions where a slower discharge won't (see the next myth).
Conductive surfaces are dangerous; antistatic surfaces are safe. We object to this
generalization, though it contains more than a grain of truth. MIL-HDK-773 strongly
favors antistatic surfaces as being safer for ESD-sensitive items, and our own data sup-
port this. For human safety, conductive surfaces are obviously undesirable because they
can carry a lethal current (over I milliampere). However, even static-dissipative surfaces
can cause ESD damage to sensitive devices such as MOSFET's by the CDM, and avoid-
ing conductive surfaces completely is impracticable because tweezers, soldering iron tips,
and operators' fingers are ubiquitous in workstations. Our approach, which we consider a
prudent compromise, is to use a mix of antistatic and conductive surfaces, with the latter
being minimized.
If handling procedures are correctly developed and followed (admittedly this is a big
"if'), conductors will be grounded and ESD-sensitive items will be uncharged so the CDM
can't happen with either a conductive or an antistatic surface. In other words, we're avoid-
ing the trouble of covering fingertips with antistatic cots and so on by relying on operator
skills to prevent 01 and CDM, which can still happen with antistatic surfaces, though to a
much lesser extent. A relatively (not completely) "forgiving" work environment in terms of
discharges, with all antistatic or static-dissipative surfaces, seems unfeasible to us, and damage
by fields (by the FIM) could still happen.
Antistatic surfaces (of DIP tubes, for example) do not triboelectrically charge pack-
aged ESD-sensitive items by rubbing against them. In 1983, we were surprised to find
that antistatic polyethylene (MIL-B-81705, Type II) charged epoxy-glass circuit-board
material, and researchers at another major corporation were skeptical at first but verified
254 APPENDIX
our results (Ref. 3). The truth is that the contamination effect (antistat transferring to the
rubbed surface so that liquid separates from liquid with minimum charge generation) greatly
reduces triboelectrification but does not prevent it. An antistatic material whose surface
layer of antistat has been removed by volatilization or wear will not benefit from the con-
tamination effect nor will permanently antistatic materials with dry surfaces. Furthermore,
packaging materials that depend on a fugitive antistat will hold a charge when the antistat is
sufficiently depleted.
Triboelectric charging is quite material-specific, so tests should be run with the particu-
lar packaged item and packaging material (both fresh and aged) in question.
Highly conductive surfaces (of metals, for example) do not participate in triboelec-
tric charging and are safe for ESD·sensltive items. Nothing could be further from the
truth! Conductors only seem not to be triboelectrically charged because when they are
grounded, the charge runs off almost instantaneously. An isolated sheet of aluminum or
steel takes a charge of equal magnitude and opposite polarity to the charge taken by a
plastic sheet stroked against it. In fact, grounded metals and carbon-loaded, conductive
plastics are effective in charging nonconductors and causing damage by the COM, which
can occur during the charging process with no discharge when, for example, an intense,
uneven field is produced on a circuit board with mounted devices (we have called this the
field-from-board or FFB damage mechanism). For example, epoxy-glass circuit-board lami-
nate is readily charged by sliding on stainless-steel shelves in an oven, and people are charged
by walking with insulated, rubber-soled shoes on a conductive floor. Ironically, such a floor,
which was designed for use with conductive footwear, actually can make matters worse
with ordinary shoes.
Conductive flooring is always beneflcial for ESD control. We've just mentioned
that a conductive floor may be the opposite of beneficial when ordinary, nonconductive
shoes are worn. We say "may be" because shoe sole materials vary greatly in charging
propensity. Rubber and certain vinyl compositions tend to give especially high charges.
Another problem with conductive floors is safety. In one of our tests, operators with wet,
leather-soled shoes (as they might have been ifthey'd come in out of a rainstorm) showed a
resistance-to-ground low enough for a dangerous current of over I milliampere to pass
through their bodies if they touched a 110 V hot lead. Ground-fault interrupters may be
needed when such situations could occur. The point is that special footwear must be used
with conductive floors, and the safety issue must be considered. Only then are conductive
floors beneficial.
Conductive chairs and special smocks are essential In a rigorous ESD·control
program and allow some operator disciplines to be relaxed. Conductive chairs are
possibly dangerous to both people and devices, insofar as a lethal current might be carried
or the COM might be encouraged by the conductive surface, but we do not reject them on
that basis alone. Our main reason not to use conductive chairs is that they are unnecessary;
ordinary vinyl-upholstered chairs are not a hazard according to our measurements of
electrical-field strength in the work area on the bench.
Though vinyl does exhibit static charges, the fields from them are sufficiently attenuated
by distance and the shielding effect of the operator's body to be harmless at the bench
surface where ESO-sensitive items are being handled. This setup assumes that operators
will not bring the items near the upholstery, as one of our operator disciplines specifically
forbids. Thus, we are depending on operator skill to avoid the cost of expensive, foolproof
equipment, but operators are always the critical factor no matter how much money is thrown
at the problem of ESO control.
PAPER NO. 11 255
As for smocks, the shielding effect of fabric with widely spaced conductive threads is
dubious and, in any event, isn't needed with tight-fitting shirts or blouses with short or
rolled-up sleeves. Proximity to the skin suppresses fields on the cloth, and distance to the
operator's hands, in the normal working position, further weakens fields to the point of
being harmless. You can make measurements with a hand-held field meter to support or
refute this conclusion in your own assembly area. Another consideration is cost. Smocks
are good uniforms for raising ESD awareness and impressing visitors, but they are rather
expensive to collect, clean, and reissue.
No equipment justifies relaxing operator disciplines. Operators are the "captains of the
ship," and ineptitude on their part can sink the most lavishly capitalized ESD-control pro-
gram. Forget expensive frills and invest in operator training.
The hair hazard is eliminated by area ionization. Here's another case of expecting
too much from ionization. As stated earlier, area (room) or any other kind of ionization
should be used only for clearly defined purposes, and controlling static charges on hair is
not one of them. Long hair will constantly recharge as it swings around, and the resulting
field could cause damage by the FIM before air ions neutralize the charge. The best course
is to require hair to be tied back so static voltages on it are suppressed by nearness to the
skin.
Use a hand-held meter to verify that the field in the vicinity of ESD-sensitive items is
harmlessly weak, even with the operator's head bowed for close examination. We use the
"Charge-Distance Rule," an empirical relation between the apparent voltage on a surface
and the safe distance from that surface (Ref. 4). The equation is d = ~~:, where d is the
distance in inches and V is the apparent charge in volts measured vith a hand-held meter.
For example when V is 1,000 V, d is 18 in.
Ordinary floors in ESD-controlled areas must always be resurfaced or specially
treated. This assertion seems reasonable at first, but concrete floors, especially new ones,
often contain enough moisture to be weakly conductive. The presence of salts in the con-
crete also may help. Sealers may be soapy and afford low triboelectric charging of ordinary
shoe soles, as does the thin layer of antistat on floors coated with a static-limiting floor
finish. Therefore, concrete floors should be checked for surface resistivity if conductive
footwear is used, or triboelectric charging of people (which correlates imperfectly with
surface resistivity) should be checked if ordinary shoes are worn.
Soiled. uncoated vinyl-asbestos tile, though unattractive in appearance, may also impart
relatively low charges to people walking on it. In general, evaluate the floor before you
consider altering it. If nothing else, baseline data will show the improvement obtained and
help, in retrospect, to justify the cost of resurfacing (with conductive tile, for example) or
treating (such as with static-limiting floor finish).
All cellulosic materials change from antistatic to nonconductive at moderately low
relative humidity (15 to 20%). We believed this myth until our testing disclosed a coated,
plasticized cellophane (regenerated cellulose) that remains antistatic down to about 4%
relative humidity at room temperature. (For further information, see EOS/ESD Technology,
OctoberlNovember 1989, p. 9.)
Antistatic masking tapes are ESD-safe. As with many myths, the fallacv of this one
lies in oversimplification. Yes, antistatic masking tapes are safe in the sense that they will
not show a field because incipient charges drain to the worker's hand, while any field from
the nonconductive adhesive layer is suppressed by the weakly conductive tape layer. How-
ever, when the tape is pulled from a nonconductive surface (for example, epoxy-glass
circuit-board laminate), that surface may be charged enough to damage mounted devices by
256 APPENDIX
the FFB mechanism. The point is that charging of the substrate, as well as the tape itself,
must be considered before generalizing that the tape is ESD-safe.
Short drain times of charges from objects, such as tote boxes, deposited on work
surfaces is of major importance and justifies expensive laminates and grounding
schemes. Our experience and judgment indicate that situations where short drain time is
critical are rare. A commonly cited scenario in support of short drain time is a grounded
operator reaching into a charged tote box at rest on a bench and touching a sensitive device
lead. Damage occurs by the CDM because the body of the device is charged by induction
along with the tote box. However, such an incident is unlikely because the grounded opera-
tor usually will touch the box and discharge it before reaching in. In fact, reaching in and
touching a device lead without touching the box wall might be a difficult trick. We're satis-
fied with a reasonably short drain time without taking extreme measures, such as using a
conductive surface, which also has dangers, to minimize the time. Drain time has always
been a secondary consideration, not a controlling factor, in our selection of work surfaces
and grounding techniques.
Overkill protection Is necessarily expensive. This myth may stem from the idea that
you get what you pay for. However, overkill can come cheap. An example is a stone house
built near a quarry; the stone will last for thousands of years but is cost-competitive with
wood, which decays in a lifetime. Similarly, aluminum foil is a far more effective ESD
shield than see-through, lOO-angstrom-thick metallization but is similar in cost.
A laminate bag consisting of wire screen sandwiched between layers of antistatic poly-
ethylene is indeed overkill, resisting even 35,000 V from a Tesla coil, but this packaging at
a few dollars for a 10 x 12 in. bag is not really expensive when a $100.000 module for a
vital aerospace or defense application must be protected. Continuous wrist-strap monitor-
ing. discussed earlier, also may seem like overkill-though logically it is a necessity-but
costly failures or material review actions are avoided. In fact, you can't afford not to have it.
Latent ESD failures are commonplace. Perhaps we should call this one a "sus-
pected myth" because no one is sure. A consensus among leading device experts, whom
we questioned two years ago, was that latent failures are real but rare. In 1988, British
Telecom researchers stressed components to 90% of their HBM threshold and then sub-
jected them to accelerated testing to look for premature (latent) failures; however, they
found none (Ref. 5). Of course, this negative result doesn't rule out latent failures under all
conditions; any negative generalization is difficult if not impossible to prove.
Intuition supports the latent failure concept when we picture stressed devices as "walking
wounded" because wounded soldiers indeed may die early, but devices are not human
nor wounded in the sense that flesh can be. Intuition often uses false premises and is
wrong-though it could be correct by accident in this instance. The point is that skepticism
is in order until strong evidence for latent failures accumulates. But even the possibility of
a few latent failures in critical equipment, such as missile guidance systems or life-supporting
medical appliances, is reason enough for a rigorous ESD-control program.
Myths, old and new, are oversimplifications or bad generalizations that tend to lull be-
lievers into a false sense of security while they rely on a product or approach that has fatal
shortcomings or is needlessly expensive. For example, someone concerned only with "sparks"
would ignore the FIM and FFB, and someone relying on ionization as a cure-all could have
ESD-sensitive items damaged by poorly trained operators. In stating the truth as we see it,
our object has been to dispel irrational fears, roused by some vendors' scare tactics, as well
as foolish complacency. We have suggested when to spend and when to save because throwing
money at the problem in the wrong places (for expensive, unnecessary equipment) and
PAPER NO. 11 257
misguided thrift (refusal to buy continuous wrist-strap monitors) are both counterproduc-
tive. Buy the basics, forget the frills.
Our opinions are often controversial but supported by realistic tests as well as experience
and common sense. They form the basis of a flexible, cost-effective system of ESD control
that has proven itself over years of practice (Ref. 6).
References
1. GIDEP Alert H7-A-85-02, 1985.
2. Kolyer,1. M., D. E. Watson, and W. E. Anderson. "Controlling Voltage on Personnel," EOSIESD
Symposium Proceedings, EOS-ll, 1989, p. 23.
3. Private communication with I. R. Huntsman and D. M. Yenni, 3M, 1983.
4. Kolyer, I. M., W. E. Anderson, and D. E. Watson. "Hazards of Static Charges and Fields at the
Workstation," EOSIESD Symposium Proceedings, EOS-6, 1984.
5. Woodhouse, I., and K. D. Lowe. "ESD Latency: A Failure Analysis Investigation," EOSIESD
Symposium Proceedings, EOS-IO, 1988. p. 47.
6. Kolyer, I. M., and D. E. Watson. ESD From A to Z: Electrostatic Discharge Controlfor Elec-
tronics. Van Nostrand Reinhold, New York, NY, 1990.
Paper No. 12
As a typical ESD-sensitive device, the Motorola 2N435 I Metal Oxide Semiconductor Field-
Effect Transistor (MOSFET) has made an excellent "white rat" for our research. 1 However.
this destructive testing has drawbacks: the cost of the MOSFETs as well as annoying delays
when more pieces must be ordered. Now we have found that hundreds or even thousands of
tests can be run on a single silicon wafer.
The wafer is P-type silicon coated with 1,000 A of oxide. With either MOSFETs or
wafers, a voltage differential, which may be momentary, punctures the oxide in a micro-
scopic dielectric strength test.
The wafer is 4" in diameter and 0.020" thick and has a resistivity of 5 to 60 n'cm,
tvpicallv 15 to 25 n·cm. The oxide thickness is 800 to 1,200 A, nominally 1,000 A. This
type of wafer serves to monitor sodium contamination in furnaces where the oxide is grown,
so the upper surface holds a hundred sputtered aluminum dots for making electrical mea-
surements to detect sodium.
To make electrical contact with the underlying semiconductive silicon, one side of the
wafer was broken off and solder drops (labeled "s" in Fig. I) were applied 2" apart on the
exposed edge and pressed by spring-loaded contacts with terminals A and B. Resistance
between A and B was I x lOs n at 10 V.
The probe, electrically isolated with Teflon™, was held by a pivoted wooden arm which
could be adjusted for reach or radius as well as rotated. The procedure was to start at maxi-
mum radius and sweep the wafer at successively smaller radii. At first, an attempt was made
to utilize the aluminum dots, but the best procedure, which provides an almost unlimited
number of test points, was to avoid the dots and probe directly on the oxide. The steel probe,
0.112" in diameter with the tip polished to a radius of O.OOS", was springloaded to press the
oxide surface with a force of 0.6 lb (Fig. 2).
Resistance across the oxide layer was determined with a Beckman L-12 Megohmmeter
with shielded leads (RG-58IU coaxial cables) at 10 V. The low or negative megohmmeter
lead was connected to terminal T (which was connected to probe terminal P) while the high
or positive megohmmeter lead was connected to A. Undamaged oxide read approximately
5 x 1012 n, with the reading influenced by capacitance of the wafer apparatus. Damaged
oxide almost always read less than 5 x 10 10 n, usually in the lOS to 109 range but sometimes
lower. Slight degradation, e.g., 5 x 1011 n, rarely occurred. Intermittents occasionally oc-
curred, as with damaged MOSFETs.
258
PAPER NO. 12 259
CURRENT FLOW ~
PROBE
ESD I
EVENT
r
.. : ....\.,:.........................:.... .
.....-
~-- P - TYPE SILICON
s s
(solder) (solder)
For comparison tests, MOSFETs were similarly measured with the low megohmmeter
lead to the substrate-case lead of the MOSFET and the high megohmmeter lead to the gate
lead of the MOSFET. Fresh MOSFETs read over 1013 0, and degraded MOSFETs usually
read less than 1010 0, but subtle damage, e.g., 2 x 1012 0, occasionally occurred. The arbi-
trary criterion for failure was a resistance of less than 1 x 1012 O. This is more rigorous than
our former definition of failure as a shift of more than 0.1 V in gate-source threshold volt-
age, which allowed considerable damage to the gate oxide. 2
The difference between the wafer and a MOSFET gate was leakage current. For ex-
ample, when the voltage across the gate of a MOSFET was slowly raised, the resistance fell
to 2 x 1011 0 at 102 V (0.5 nA current) and 1 x 1010 0 at 110 V (11 nA) before failure
occurred and the resistance fell several orders of magnitude. In a similar test, the wafer
oxide began to leak (2 x 1012 0,0.05 nA) just before failure at 107 V.
In MOSFETs, perhaps the major leakage path is over the surface of the chip to the nearby
source and drain metallization. This leakage is significant in Charged Device Model (COM)
tests involving charge drainage from a pin to a surface. However, in testing where the rise
time is fast, gate leakage does not protect the MOSFET and it is successfully modeled by
the wafer.
The voltage sensitivity of the wafer apparatus was checked with a modification of the
Human Body Model (HBM). 3 Terminal A was grounded, and a 1,500-0 resistor connected
to a charged l00-pF capacitor was touched to terminal T. MOSFETs were tested by simi-
larly touching the gate lead with the substrate-case lead grounded. Both wafer and MOSFETs
began to fail at about 80 or 100 V (Table 1) as expected for 800 to 1,000 A of silicon oxide.
The following four examples illustrate that the wafer method correlates with the MOSFET
device tests. The examples deal with:
Example I. To demonstrate the Field-Induced Model (AM) for ESO damage, a I ftl
aluminum plate on a nonconductive standoff was connected to terminal P, and terminal A
was grounded. AcIar™ film (9" x 12") triboelectrically charged to -15,000 V was slowly
moved toward the plate to induce a charge. In repeated tests, when the film was approxi-
mately 40" from the plate, the oxide on the wafer failed as noted by a burst of RF heard from
an AM radio tuned between stations. This result is consistent with our Charge-Distance
(CD) Rule, which specifies a conservative safe working distance of 68" from surfaces charged
to 15,000 V.2
Example 2. In tests for allowable resistance to ground (ARTG) for personnel, an un-
grounded person with a capacitance of approximately 200 pF stroked the Aclar™ film on a
carpet. The film was quickly lifted, which charged the person to -4,000 V as read by a field
meter on an electrically isolated aluminum plate connected to the individual's skin. The
charging process was repeated with the person grounded by a wrist strap with a I-MQ
resistor and a finger pressed against T (A was grounded). The failure rate (number
failed/number tested) was 0/10. However, when the resistance in the strap was increased to
15 MO, the failure rate was 6/10. This result is consistent with previous data using MOSFETs.4
The explanation is that the voltage surge or spike exceeded 100 V and damaged the oxide
when the resistance to ground was too high; the charge on the person could not be drained
fast enough. In retrospect, the test could have used a wafer instead of MOSFETs to demon-
strate this principle and set the maximum ARTG at 10 MO.
Example 3. The setup in Fig. 1 modeled a circuit-board module with the lead from T
to P representing circuitry leading to the oxide and the wafer representing circuitry beyond
the oxide. The capacitance on the probe side was determined by charging T to -1,000 V
with the probe resting on Teflon and reading the charge with a Keithley Model 617 Pro-
grammable Electrometer. The charge was 12 nCo Therefore, C = QIV = 12 pF. By this
method, a ceramic capacitor labeled 100 pF held 103 nCo On the wafer side, C was 34 pF. A
discharge to T represented the so-called floating model in which an ungrounded ESO-sen-
sitive item is touched by a charged surface.5 The question was: Are low-capacitance con-
ductive objects, such as small screwdrivers, at comparatively low voltages really a hazard?
To obtain an answer, the objects in Table 2 were touched to a power supply lead, and then
to T. Alternatively, the probe was rested on Teflon, and the gate lead of a MOSFET was
connected to P by means of a jumper while the substrate-case lead was connected via a
jumper to B. The jumpers raised the capacitance from 12 to 14 pF on the probe side and
from 34 to 36 pF on the wafer side. The voltages in Table 2 are at the moment of contact.
For example, the ceramic capacitor was touched to the power supply lead at -1,000 V but
lost 50 V, by leakage or by attracting air ions, during the one-second transfer to T.
The data in Table 2 show the usual statistical scatter, but the wafer and MOSFET tests
agree that small conductive objects are indeed hazardous at about 500 V/5 pF or 3,000 V/l
pF, even with considerable capacitance (12 pF) in circuitry leading to the oxide layer. The
inverse situation is the COM in which charged DIPs with a capacitance of only 1 to 3 pF are
damaged by discharge from a lead to a workbench or other surface. s
Summary
In conclusion, the wafer test duplicates results with actual devices when the rise time is fast.
Data are somewhat scattered, just as with MOSFETs. The practical object is to obtain a
failure rate of 0/5 in a worst-case test to establish safe working conditions and to select
ESD-protective packaging.6 A MOSFET test has been recommended for packaging specifi-
cations, but the method of Example 4 would do as well.·
References
1. J. M. Kolyer and O. E. Watson, ESD from A to Z: Electrostatic Discharge Control for Electron-
ics, Van Nostrand Reinhold, 1990, pp. 21 and 96.
2. J. M. Kolyer, W. E. Anderson, and O. E. Watson, "Hazards of Static Charges and Fields at the
Work Station," EOSIESD Symposium Proceedings, 1984, p. 7.
3. OoO-HDBK-263, May 2,1980, p. 24.
4. J. M. Kolyer, O. E. Watson, and W. E. Anderson, "Controlling Voltage on Personnel," EOSIESD
Symposium Proceedings, 1989, p. 23.
5. O. J. McAteer, Electrostatic Discharge Control, McGraw-Hili, 1989, pp. 176 and 189.
6. J. M. Kolyer and W. E. Anderson, "Perforated Foil Bags: Partial Transparency and Excellent
ESO Protection," EOSIESD Symposium Proceedings, 1985, p. 111.
PaperNo. 13
Work surfaces for ESD control must be conductive enough to allow charges to drain from
conductive objects, yet resistive enough to restrict the current flow to a safe level in terms
of the Charged Device Model (CDM).I An unsafe level of current flow has been equated
with sparking. 2•3 A study was undertaken to determine the best compromise between drain
time and CDM protection based on device damage data, not mere intuition.
Using a Beckman L-12 Megohmmeter set at 10 V with the high or positive lead con-
nected to the grounding lug of l-ft2 panels or mats, or clamped to the edge of bag samples,
the resistance to ground was measured in two ways:
• With the low or negative lead of the megohmmeter connected to a dual in-line package
(DIP) square-tipped lead 0.015" wide and 0.009" thick (terminal D in Fig. I) pressing
the surface (S) with a force of 0.008 lb.
• With the low lead of the megohmmeter connected to a National Fire Protection Asso-
ciation (NFPA) 99 electrode (5 lb., 2.5" diameter, aluminum foil over rubber) resting
on the surface.
Eleven surfaces were tested with the following results (resistance in ohms through D is
given in parenthesis):
• Aluminum alloy 6061-T4 sheet (approximately 10 0, but varied from 3 to 1,100 0).
• Carbon-loaded conductive polyolefin bag (l x l()!l 0).
• Hard laminate with conductive fibers (5 x 1()!l 0).
• Bag with external nickel (2 x 101 0).
• Rubber mat (I x 109 0).
• Special vinyl mat (6 x 109 0).
• Three hard laminates (2 x 1010 0, 5 X 10 10 0, 7 X 1010 0).
• Antistatic polyethylene (MlL-B-81705, Type II) (1 x 1011 0).
• Painted steel as used for office furniture (5 x 1012 0).
Except for the aluminum and the nickel-coated bag, the NFPA 99 values are plotted
against the D values as shown in Fig. 2.
263
264 APPENDIX
""--.11',
If.
\
\
\
\
IJ2
I
The CDM was modeled using the setup of Fig. 1 as explained in Ref. 4. To represent a
charged device, the switch was closed, and the powered (negative) lead from a calibrated
power supply (Power Designs. Inc., Model 3KlOB) was connected through an 8-MO resis-
tor to terminal A before turning on the power. With the power supply turned on before
connecting it to terminal A, the 1,000 A oxide layer on the wafer and a MOSFET with its
leads shunted were damaged. This illustrates the fact that shunting does not necessarily
provide ESD protection. The MOSFET is tested by moving the probe to a Teflon™ surface
and using the jumpers as shown in Fig. 1.
With the assembly charged, the power supply lead was disconnected from terminal A,
the switch opened, and block B (0.25" thick) removed to let D fall to the grounded surface
S under test. As shown schematically in Fig. 3, capacitor C2 (representing the capacitance of
the wiring from terminal T to P and the probe itself when contacting the wafer) then drained
into the test surface through resistance R 2, while C. (representing the capacitance from the
oxide layer through terminal A, the switch, the wire w and to terminal D) was insulated by
the wafer oxide or by the gate oxide of the MOSFET. The l4-pFcapacitanceofC. is worst-
case for a large DIP; ordinary DIPs are 1-3 pF at the work surface. s The RC time constant
for C2 was as expected, given measurement uncertainties, e.g., approximately 3 s for anti-
static polyethylene vs 1.7 s calculated, with the fall in voltage at terminal A being moni-
tored with a Trek Model 512 Field Meter.
After 5 s, the switch was closed and terminal T was touched with a grounded I50-MO
resistor to discharge the setup before measuring the resistance of the wafer oxide or MOSFET
PAPER NO. 13 265
gate with the megohmmeter. Failure was judged by a significant resistance drop as ex-
plained in Ref. 4.
The wafer oxide failed when the voltage at C2 decreased by more than 100 V, indicating
that the dielectric strength of the oxide was exceeded. Even with the resistance through D
(Rz in Fig. 3) at 4 X 1012 n, the wafer oxide finally failed at 20 s when drainage of CI was
allowed to continue. At the arbitrary drainage time of 5 s, only painted steel was safe among
the surfaces tested.
In contrast to the wafer oxide, MOSFETs were partially protected by leakage current, as
explained in Ref. 4. The failure rate (number failed/number tested) decreased with increas-
ing resistance R2• The rate did not approach zero until approximately 1010 n at 500 V/14 pF,
1011 nat 500 V/214 pF, or 1012 nat 1,000 V/14 pF as shown in Fig. 4. The failure rate
reflected no sudden "sparking point" as the resistance decreased; rather, there was a smooth
rise in the probability of failure. Without leakage current, MOSFETs would be damaged
even by antistatic polyethylene at 500 V/14 pF, as is the wafer oxide. But, in fact, as the
resistance through D increases, the leakage rate exceeds the drainage rate; therefore, the
MOSFETs are not damaged.
By shunting the wafer with a lOS n resistor between terminals P and A, wafer failure
began abruptly below lOS n as expected. Thus, the shunted wafer can serve for testing
purposes as an idealized 50%-failure-rate MOSFET as shown in Fig. 4.
The conclusion drawn from Fig. 4 is that the optimum resistance through D is in the 1010
to 1011 n range, corresponding to an NFPA 99 measurement of 107 to lOS n for the data
points in Fig. 2. The rubber mat with a resistance through D of 1 x 109 n (l x 10«' n by
NFPA 99) was judged too conductive (approximately 40% MOSFET failure), while the
~ 10
f
M
81
~
u.
z
1)
i
~
C!) 6
g
§ 5
III
Jj
I
a: 4
'0
J
3
4 5 6 7 8 9 10 11 12 13
Log of Resistance to Ground through 0, Ohms
Fig. 3. Schematic for CDM apparatus (Fig. 1) with switch open and D resting on S.
vinyl mat at6 x I()'I n was borderline, The hard laminate at 5 x 0 1 ' was clearly unsuitable,
but fortunately most hard laminates for ESD control give 1010 to 1011 n through D as desired.
Incidentally, the CDM danger of conductive bags lying on the work surface is apparent
from Fig. 4.
Though the preferred hard laminates are safe for MOSFETs at 500 V, damage does occur
at 1,000 V, especially at higher capacitance levels, as shown in Fig. 4 and reported previ.
ously.6.7 Therefore, eve n these surfaces must be considered possibly hazardous todevices in
Class I ofMIL·STD·1686A. If drain time is not an issue, a nonconductive surface with an
underlying conductive layer for voltage suppression is best.'
Skin resistance varies greatly but was on the order of 107 n through 0, corresponding to
a MOSFET failure rate of 60%. 'Therefore, a grounded operator is the lesser of two evils vs
an ungrounded operator.' This COM hazard can be alleviated by wearing antistatic gloves
or finger cots or avoided altogether by wearing nonconductive latex gloves or cots.' How·
ever, the wearer of nonconductive gloves or cots must remember that charged objects can·
not be discharged by touching as they nonnally are, and even small charged objects at
relatively low voltage, e.g., a screwdriver at 5 pF and 500 V, can cause ESO damage by
direct injection or the floating model.'
PAPER NO. 13 267
Tests also were run with 1% tolerance, 1/20-W thin-film resistors (Class 1 of MIL-STD-
1686A) shunted between terminals P and A in Fig. l. At 3,000 V and C1 = 214 pF, a hard
laminate (5 X 1010 0 through D) caused less than 0.01 % resistance shifts for 10-0, 100-0,
lO-kO, or 3OO-ill resistors. However, the carbon-loaded conductive bag shifted the resis-
tance of the 3OO-k 0 resistor by -0.05%, and the aluminum surface caused shifts in all the
resistors to a maximum of -0.38%. The conclusion is that the CDM is not a major hazard
for thin-film resistors with I % or even 0.1 % tolerance when the work surface is a hard
laminate with 107 to lOS 0 resistance to ground by NFPA 99.
ICs or discrete devices usually will not exceed 500 V when near the work surface be-
cause of voltage suppression. In fact, the inputs of most ICs now are protected by zener
diodes to at least 1,000 V. Therefore. the major concern may be for Class I devices mounted
on circuit boards that have relatively high capacitance and also high voltage because a large
board tilted with respect to the work surface escapes effective voltage suppression. Accord-
ingly modules should be checked with a field meter and, if necessary, neutralized with
5/5 ////////////////////////////////////~"""""""'~
•
~ \.;
/ ~
4/5
~ Unshunted Wafer. ~
~ 500 or l000V. ~
~ 12pF ~
500V ~ \.;
/ ~
~ \.;
/ ~
3/5 () ~ () ~
~\ \.;
____ ~%..:ai~re_ _ _ _ _ _ _ /'~ _ _(OOOV ~
,~ ~
•
~ ~
215
•
MOSFETpata
500V.14 pF
.\\~ () ~
~
,
\
() 1000V. 14 pF \ ~
D 500V. 214 pF
'. :-.;
1/5 ~~ . \r500V.
\~ 214pF
··
2 3 4 5 6 7 8 9 10 11 12 13
Log of Resistance to Ground through D, Ohms
ionized air before a sensitive tenninal or edge contact touches the bench surface or the
operator's fingers.
In conclusion, while supporting the popularity of work-surface laminates with an NFPA
99 resistance to ground of 101 to lOS 0, the data make the important point that these surfaces
are only relatively CDM-safe. They are nonsparking, but discharges can cause ESD dam-
age without an evident spark. For example, in many cases, D fell to surface S and damaged
MOSFETs with no RF heard from an AM radio tuned between stations. Therefore, operator
disciplines must assure that the ability of the surface to drain charged objects does not
become a liability by damaging voltage-sensitive devices.
References
1. Carlton, D., "Selecting an ESD Workstation," Evaluation Engineering, January 1991, p. 94.
2. "New Test Proposed for Checking ESD Safety of Materials," Compliance Engineering, Fall
1990, Vol. VII, Issue 5, p. 77.
3. "Electrostatic Discharge Protective Packaging," MIL-HDBK-773, April I, 1988.
4. Kolyer, J. M. and Watson, D. E., "ESD Testing With Silicon Wafers," Evaluation Engineering,
September 1991, p. liS.
5. McAteer, O. J., Electrostatic Discharge Control, McGraw-Hill, 1989, p.176.
6. Kolyer, J. M., Anderson, W. E., and Watson, D. E., "Hazards of Static Charges and Fields at the
Work Station," EOSIESD Symposium Proceedings, 1984, p. 7.
7. Kolyer, J. M., Anderson, W. E., and Watson, D. E., "Tote Box Material: How Good Is It?"
EOSIESD Technology, OctoberlNovember 1987, p. 13.
8. Kolyer, J. M. and Watson, D. E., ESD from A to Z: Electrostatic Discharge Control for Electronics,
Van Nostrand Reinhold, 1990, p. 121.
9. Kolyer, J. M. and Watson, D. E., ESD from A to Z: Electrostatic Discharge Control for Electronics,
Van Nostrand Reinhold, 1990, p. 46.
PaperNo. 14
John M. Kolyer
The charged device model (CDM), in which charge on the circuitry of an ESD-sensitive
device flows from a contact, lead, or terminal to a benchtop or other surface, has become
the focus of increasing concern. The flow has been found to create sufficient voltage differ-
ential or current within the device to puncture gate oxide, degrade thin resistor films, or
cause other overstress damage. I Because of this, MIL-HDBK-773 warns against conduc-
tive surfaces.
Obviously, a highly conductive surface, such as metal, is unsafe by the defmition of the
CDM as a discharge-caused failure model. Conversely, a nonconductive surface is assumed
to be safe because discharge is avoided. However, some surface conductivity is desired to
drain static charges from objects such as tote boxes. A tradeoff is needed; there must be a
surface-to-ground resistance high enough to suppress the CDM while not overextending
the drain time.
One study found that the resistance to ground for work surfaces was correlated with the
CDM failure rates of MOSFETs sensitive to lOOV. (These MOSFETs are more easily dam-
aged than most, but not all, devices).2 The conclusion was that the resistance through a lead
that contacts the surface should be above 1010 ohms, because the MOSFETs were protected
by a 1010 ohm leakage path. When the resistance through the lead was 109 ohms, for ex-
ample, leakage around the gate oxide of the MOSFET was not fast enough to prevent a
damaging voltage differential of lOOV.
As seen in Fig. I, a conductive rubber mat and a laminate with conductive fibers were
unsafe, while painted steel was safe but nonconductive. Between these extremes are the
best surfaces: hard laminates resistive enough to prevent CDM damage yet sufficiently
conductive to drain charged objects.
How can you evaluate your own work surface? Here are three methods:
269
270 APPENDIX
4 5 6 7 8
80
Brlghl
(I mAl
Z"
'# c:
oj ~
iii :>
60 CC 2-
I!!
~
~
I-
S
.r::
.~ Dim
W CO
LL (I l1li)
lJ) ,Q
0 :>
::E CO
40 c:
g
Z Very Dim
(0.3l1li)
Dark
(O.lI'A)
20
•• •• •
2 3 4 5 6 7 8 9 10 11 12
Fig. 1. MOSFET failure rate by CDM versus resistance to ground and neon bulb brightness.
the surface gave a resistance through a device lead of 5 x 10' ohms with the lead resting on
a fiber vs. 5 x 1011 ohms between fibers.
2. The method discussed in Ref. 2 measures resistance at 10 V through a dual in-line
package (DIP) lead (0.015 x 0.009 in. in cross section) pressing the surface with a force of
0.008 lb. The lead is shifted from place to place to identify the lowest resistance, which
should be above 1010 ohms. (See Fig. 1.)
3. A neon bulb test was evaluated using the DIP lead with the specified force, because
pressure on soft surfaces (e.g, a rubber mat) greaty affects resistanceY An N-2 neon bulb
with maximum striking voltage of 90 V and design current of 0.6 rnA was in series between
the DIP lead and the negative lead of a megohmmeter at 500 V, while the positive
megohmmeter lead was connected to the grounding snap or conductive layer of the material
under test. The bulb became very dim at a little above 109 ohms through the DIP lead,
corresponding to 0.3 microampere.
PAPER NO. 14 211
The test did not detect COM-hazardous surfaces at the low end of the MOSFET failure
rate curve. However, the bulb did provide a visual demonstration of the worst COM-offenders
by glowing brightly and flashing on contact. With unhomogeneous, ungrounded surfaces
not amenable to a resistance measurement, a flash of the bulb would suggest a COM hazard.
In conclusion, COM-safe laminates or other COM-safe materials, such as soft mats, can
be identified by a resistance measurement. Measurements can be taken using the standard
NFPA 99 electrode or a small-area probe that simulates a device lead and detects conduc-
tive inhomogeneities.
References
1. McAteer, O. I., Electrostatic Discharge Control, McGraw-Hill, 1989, p. 173.
2. Kolyer, 1. M. and Watson, D. E., "The Charged Device Model and Work Surface Selection,"
Evaluation Engineering, October 1991, pp. 110-\17.
3. "New Test Proposed for Checking ESD Safety of Materials," Compliance Engineering, Fall
1990, p. 77.
4. Anderson, D. C., "A Simple Approach to ESD Damage Prevention," EMC Technology,
March/April 1991, p. 38.
PaperNo. 15
John M. Kolyer
Data-sheet material properties such as surface resistivity. shielding effectiveness. and tri-
boelectric charging propensity are measured under arbitrary. fixed conditions and cannot
predict performance under the varied. dynamic conditions of use. Thsts that realistically
simulate electrostatic discharge (ESD) damage mechanisms are needed for selecting ESD-
protective materials with confidence. This article briefly describes each of the common
ESD damage mechanisms~2,3 and gives examples of relevant tests.
272
PAPER NO. 15 273
NE-2
SED Bulb
use of the SED in ESD process control has been published. 4 An advantage of the SED is
ease of use, because a liquid crystal display (LCD) changes color when the monitor is
tripped, and no measurements or readings are needed.
The SED used in this test had a capacitance of 150 pF and was tripped by voltage surges
of 92 V or more, so it represented a MOSFET with a damage threshold of approximately
100 V. In repeated trials, the seated strap-wearer tripped the SED, demonstrating that
MOSFETs would have been damaged by being handled under analogous conditions (gate
lead touched, substrate-case lead grounded). The conclusion was that the HBM hazard was
not significantly reduced by the cordless wrist strap. A cord to ground, including a current-
limiting resistor to protect the wearer from shock, remains a necessity.
To reduce the chance of HBM damage, the skin voltage of people wearing ordinary
shoes can be lowered by special floor materials. This partial control of skin voltage is a
"safety net" to protect ESD-sensitive items if they are accidentally touched by ungrounded
personnel. Also, there is less chance of fingers throwing sparks on dry days and upsetting
equipment in computer rooms. A practical test for candidate flooring is measuring the skin
voltage on twelve randomly selected people wearing their ordinary shoes and walking nor-
mally. The walker holds a probe connected to an electrically isolated plate monitored by a
static meter. 2
A static-limiting floor finish with a microscopically thin surface layer of antistat reduces
the skin voltage of a person on a tile floor to approximately 10 percent of that found for
conventional acrylic finishes or "waxes.''2 An alternative to this special finish, which must
be carefully applied and maintained, is conductive carpet. Two types of such carpet were
evaluated with the results shown in Table 1. Most of the shoe soles, and all the heels, hap-
pened to be rubber rather than leather. The relative humidity was 40-42 percent at 75°F.
As seen in Table 1, charging varies greatly among people in this test, so data must be
averaged. Also, the brand of acrylic finish, as well as the condition of the tile, including soil-
ing, affects the results. Carpet 1, which contained a proportion of conductive fibers, reduced
people's skin voltage to 42 percent of the level for the tile with conventional floor finish, while
Carpet 2, in which all the fibers were conductive, reduced the voltage to 25 percent. These
carpets did not perform as well as static-limiting floor fmish, but they may be suitable for
replacing high-charging conventional carpet in computer rooms or office areas.
whereas HBM tests involve a resistance of 150-1500 n to represent that of a person. Since
the MM is the more severe condition, it is appropriate for worst-case tests of packaging
materials.
A convenient method for testing bag materials is to place a protected SED (Fig. I) inside
the bag, which rests on a grounded aluminum plate, and then to touch a charged probe to the
bag above the cap of the SED. A probe with a diameter of 0.025 inch represents a sharp
hand-held tooLS More severe tests6 use a rounded probe with a diameter of 0.25 inch to
represent a blunt tool. With the 0.2S-inch probe at 10 kV, the equivalence of the SED to an
oxidized wafer7 failing at an average of 117 V was demonstrated as shown in Table 2, and
the wafer test (Fig. 2) in turn correlated with MOSFET testsY
As seen in Table 2, only packaging films with an opaque metallic layer passed the severe
lO-kV, blunt-tool test with no failures in five trials. Such materials are rated "excellent."
Materials failing the blunt-tool test at 10 kV but passing at 6 kV are rated "good"; these
include MIL-B-S170SC Type III (see-through film laminate with vapor-deposited alumi-
num protected by O.S-mil polyester film). Materials passing only the sharp-tool testS at 9 or
10 kV are rated "fair," and materials failing all three tests are "poor."
"Poor" materials include see-through film laminates with the metal "out" (protected only
by a thin organic coating rather than polyester film). The spark thrown by a finger on a dry
day will create a hole in the metallization and trip the SED inside metal-out bags, even
though these bags pass the EIA 541 (Electronic Industries Association) shielding test. The
problem with the EIAtest is that it uses large, flat electrodes at relatively low voltage (I kV)
and does not include real-life spark discharges. This misleading test gives a good rating to
volume-conductive, carbon-loaded bags which perform poorly in our tests as well as in
tests recently conducted in Germany using the EIA 541 network with a krypton switch.s A
realistic approach demands that if spark discharges are a threat, they must be accounted for
in the test method.
Another example of the MM is discharge between a charged, isolated conductor and
an ESD-sensitive device. Isolated conductors may attain considerable voltages by
PAPER NO. 15 275
space-charging in the vicinity of air ionizers, so tests were run with a 20-pF plate placed at
various distances from ceiling-mounted. pulsed DC emitters.
A potential relative to ground of 250 V was required for the 20-pF plate to trip the SED
(Fig. 1) with case grounded. This result is consistent with experience7 for low-capacitance
charged objects. When the plate touched the cap of the SED. the capacitance rose so that the
original 250 V fell to approximatelv the 92-V threshold voltage for tripping the sensor.
Packaging
Material
Discharge
Oxide
Conductive
P-Silicon
Adhesive
For 250 V on the plate, cycle time was plotted against emitter-to-plate distance as shown
in Fig. 3. The normal operating range for emitter voltage and working distance lay in the
safe area of the graph and the conclusion was that space charging of 20-pF isolated conduc-
tors would not threaten devices, such as MOSFETs, mimicked by the SED.
I
6
o
5
Danger Zone
(250 V or
More on Plate)
at 22 kV on
Emitters
.I....
...
~
GI
3
2
(No Danger
Zone at 10-
12 kV on
Emitters)
/0 fIj Nonnal Operating
1 , / R~
1 2 3 4 5
Probe
The RC time constant for draining C2 at 8 pF is 8 x 10-7 second (rapid discharge) for
= =
probe-to-surface R 1()5 ohms and 0.8 second (slow bleed-off) for R 1011 ohms. In prin-
ciple, this constant could be calculated. However. in practice, R and C usually cannot be
measured with confidence, so an empirical test is necessary.
In general, materials with a surface resistivity of at least 109 ohms/square are CDM-safe
= =
at I kV and C, C2 5-10 pF. But keep in mind that surface resistivity readings vary with
the manner of measurement, including electrode configuration.
Curves of surface resistivity versus relative humidity (RH) can be derived by confining
samples with saturated chemical solutions that control the relative humidity in the air above
them. II As seen in Fig. 5, the resistivity of many materials varies greatly with RH, so the
same material may be CDM-safe at low humidity but unsafe at high humidity. At approxi-
mately 50 percent RH, vulcanized fiber was well above 109 ohms/square and passed the
above-mentioned CDM test with the SED; this material was selected for tweezers used to
handle ESD-sensitive dice. Similarly, a recently developed antistatic nylon for probe tools
for assembling the same dice passed the CDM test at 50 percent RH. A permanently anti-
static vinyl plastic, suitable for tote boxes, also passed. In contrast, a carbon-impregnated
cardboard packaging material had a surface resistivity less than lOS ohms/square at 50 per-
cent RH and failed the CDM test.
v • Antistatic Vinyl
F • Vulcanized Fiber
N • Antistatic Nylon
C • Carbon-Illpregnated Cardboard
Experience shows that standard triboelectric charging tests with quartz and Teflon™ are
insufficient, and specific materials of concern must be evaluated. For example, a perma-
nently antistatic vinyl plastic, suitable for tote boxes. was tested (Table 3) using circuit-
board laminates (FR-4, GF, and poly imide) with the FlM in mind. Triboelectric charging is
notoriously erratic and unreproducible, so these tests were run in triplicate. Because of the
scatter of the data, the best that can be done in such testing is to make the rough classifica-
tions shown. Black conductive plastic is well known to produce high charges, whereas
antistatic polyethylene (MIL-B-SI705C, Type II) imparts very low charge because antistat
invisibly rubs off onto the surface of the other material. The antistatic vinyl, which had a
dry surface with no antistat layer, was a medium charger -like cardboard or conventional
vinyl.
Summary
Novel tests are appropriate when standard tests are insufficiently specific and realistic. The
object is to simulate worst-case ESD-hazard scenarios so that materials for tools, contain-
ers, work surfaces. and flooring can be selected with maximum confidence.
References
1. O. I. McAteer, Electrostatic Discharge Control, McGraw-Hili, 1989.
2. I. M. Kolyer and D. E. Watson. ESD from A to Z: Electrostatic Discharge Control for Electronics,
Van Nostrand Reinhold, 1990.
3. I. M. Kolyer. "Fundamentals of ESD Control," Technical Record of Expo '91, International
Conference on Electromagnetic Compatibility, EMC Technology Magazine, Reston, VA May
18-20,1992, pages 154-161.
4. I. D. Campbell, "ESD Process Control and Measurements," EMC Test and Design,
September/October 1992, page 63.
5. I. M. Kolyer and D. E. Watson, "Packaging for High-Voltage Discharge Protection," Evaluation
Engineering, March 1992, pages 96-100.
280 APPENDIX
John M. Kolyer
Whether we want to admit it or not, we are our own worst enemies when it comes to elec-
tronic equipment. Human beings are the major source of direct injection (01), the most
straightforward and common cause of ESD damage.
Keeping ESD in check requires a good control program and even better testing proce-
dures. Today, several practical test methods simulate 01, the ESD damage mechanism that
includes the human body model and machine model, the charged device model (CDM), and
the field-induced model (FlM).'·2
The culprit in all these models is a conductive surface that participates in rapid dis-
charges to or from an ESD-sensitive (ESDS) device or circuitry leading to the device. The
conductive approach to ESD control, in which ESDS items are handled or approached only
with highly conductive grounded objects, encourages rapid discharges and must be less
safe than the rival static dissipative-approach. 2 The desirability of slow discharge is re-
flected in MIL-HDBK-773, which cautions against damage by spark discharges. 3
Where do you draw the line on conductivity? Resistance measurements vary with factors
such as electrode design, applied voltage, and relative humidity (RH), and results are diffi-
cult to relate to discharge rates and ESD damage. The carbon-loaded film in Table 1 is a
well-known sparker and, presumably, is too conductive. The laminate has a resistance to
ground in the 107 to 108 n range, previously judged to be CDM-safe for hard laminates. 4
But what about materials such as skin or a rubber mat?
Tests simulating damage models are necessary. A DI or CDM test has been proposed in
which a neon bulb connected to a 500-V power supply (PS) flashes when contacting con-
ductive surfaces.' This method is subjective and does not necessarily correlate with device
damage. 6
A CDM test uses actual devices (MOSFETs) for credibility, and this test now has been
expanded to represent DI and the FlM.6
The expense of inventorying MOSFETs has been avoided by replacing them with a simu-
lated voltage-sensitive device, the Static Event Detector™ (SED) from 3M.
For our testing purposes, a MOSFET (2N4351) stressed between the gate lead (G) and
the substrate-case lead (SC) is essentially a silicon-oxide layer, with a dielectric strength of
approximately 100 V, bypassed by a high-resistance leakage path. This path protects
MOSFETs during CDM discharges to high-resistance surfaces. 4
281
282 APPENDIX
Table 1. Materlala.
RESISTANCE (OHMS)
TOOROUND BBTWEEN
SURFACE 11IROUOH NFPA99
RESISnvrrY, 11IROUOHD NFPA99 EU!C'I1tODES
OHMS/SQUARB, (FlO. I) EU!C'I1tODB I IN. APART
NO. DESCRIPI10N AT 100 V ATIOV AT IOV AT 100 V
To cause damage, the rise time of voltage need not be rapid, because the oxide must
rupture whenever the voltage differential across it exceeds approximately 100 V. MOSFETs
subjected to the very slow rise time of 10 V/s failed at an average of 117 V-only slightly
above the 80 V to 100 V found at very fast rise times. A MOSFET with SC grounded will
fail when an antenna attached to G is moved slowly into an electrical field. 2
In the same situation, the SED is tripped only by a very intense field. creating a differen-
tial voltage far above 100 V, because normal sensor response requires voltage surges. When
rise times are fast, as they are in most ESD events, the MOSFET and the SED give equiva-
lent results.
An SED is visibly tripped at the moment of the ESD event. On the other hand, MOSFETs
can give false indications of failure by being damaged at other points in the test process.
PAPER NO. 16 283
The SED used for these tests had a capacitance of 150 pF and was tripped at 92 V. SED
and MOSFET failure rates were identical in a CDM test, and similar agreement has been
found in packaging tests with spark discharges (Table 2).
As shown in Fig. 1, the sensor (SED, MOSFET, thin-film resistor, or neon bulb) was held
between alligator clips AI and A2• The SED cap was connected by a fine wire (26-gauge) for
flexibility. Terminal Ts was connected to one of two probes: a dual in-line package lead (the
D probe) or an edge contact (EC) comprising a 0.25-in.-wide copper circuit line on glass-
filled epoxy board material mounted on a Teflon™ block (the EC probe).4
Note: Failure rates for the SED are expressed as fraction of failed/tested. Resistance shifts (highest of three tests) are
in percent.
284 APPENDtX
MOSFn
I: SED ~"
-= =-_-
.... -
~"'"- _~-0
-~
•
Fig. 1. Test apparatus.
Tenninals T T1, T, and T. were claw clips. Tl and Tl were grounded. With AI and Al
" C measured between Tl and a metal plate beneath the apparatus was 4
empty, capacilaJ'lCe I
pF. Similarly, capacitance Cl measured between T. and the plate was 8 pF with !he 0 probe
installed and to pF with the EC probe installed.
Capacitances CI and C2 represent those of circuit lines on a board. A similar measure-
ment between the plate and Ees on aerospace computer boards of various designs gave
from 3 pF to 31 pF. To raiseC a capacilorsomelimes was held byT, and Tl , if desired, C1
"
could also be l'3ised by a capacitor held by Tl and T•.
Testing was done at 1000 V. The SED sometimes was tripped by the surge when the PS
was turned on with switch Sclosed. The MOSFETs were always damaged, so the voltage
was increased slowly; for example. at 100 VIs.
After 10 s at 1000 V, S was opened and block B was pulled aside to let Dfall to the
surface. The SED was tripped by a I'3pid discharge through D. For the carbon-loaded film,
with 10' n resistance through D (Table t), the 8-pF capacitorC 2 dl'3ined with an RC time
constant of 8 x 10..1 s, the time constant for the laminate (.5 x 1010 n through D) was 0.4 s.
These times represent a I'3pid discharge vs a relatively slow bleed-off.
In theory, RC might be estimated in all cases. But, in pl'3ctice, R and C may be difficult
or almost impossible to measure, so an empirical test is more pl'3ctical and convincing than
a calculation.
Table 2 gives examples of simulated real-life handling conditions for ESDS items The
RH was approximately 50% except for tests with antistatic polyethylene at 35% RH. With
PAPER NO. 16 285
the caveat that generalization beyond the conditions of the tests is unjustified, conclusions
drawn from Table 2 are:
o The rule of thumb that a resistance to ground of 101 to lOS Q is desired for COM safety
of hard benchtop laminates is supported by the data. 4 This rule does not extend to other
materials, such as antistatic polyethylene.
o The most meaningful resistance measurement was through O. In the first test in Table
2, 1010 Q to ground gave no failures; 109 Q gave several failures, especially at high C I ;
and lOS to lOS Q gave all failures.
o Elevated RH reduces the general ESO problem but makes certain materials become
discharge hazards. Antistatic polyethylene was safe at 35% RH but not at 50% RH.
o ESO-protective finger cots should be in the antistatic range (109 to 1011 Q/sq) rather
fell onto an EC. This result emphasizes that discharge safety depends on the mode of
contact.
o Isolated (ungrounded) conductors such as the carbon-loaded film are COM threats
even at low capacitance. The converse situation (conductor at 1000 V, device uncharged)
gave similar results in past work. A I pF conductor was harmless, a 3 pF conductor was
damaging.1
o Even with a very conductive surface and an unrealistically high C I , the resistance of
References
1. McAteer, 0.1., Electrostatic Discharge Control, McGrawHilI, 1989, pp. 173-188
2. Kolyer,l. M., and Watson, O. E. ESD from A to Z: Electrostatic Discharge Controlfor Electronics,
Van Nostrand Reinhold. 1990, pp. 5, 6, 9,157,160,169, and 205.
3. MIL-HDBK-773, April 1988.4.3.1.
4. Kolyer,1. M., and Watson, O. E., "COM and Work Surface Selection," Evaluation Engineering,
October 1991, p. 110.
5. "New Test Proposed for Checking ESO Safety of Materials," Compliance Engineering, Fall
1990, p. 77.
6. Kolyer,l. M., "Is Your Work Surface COM Safe?" EOSIESD Technologv, February/March 1992,
p.27.
7. Kolyer,l. M., and Watson. O. E., "ESO Testing of Silicon Wafers," Evaluation Engineering,
September 1991, p. \15.
8. OoO-HOBK-263, May 1980,7.3.1.3.
PaperNo. 17
This investigation evaluated antistatic materials used in ESD control for electronics. The
objectives were to define curves for surface resistivity (a measure of the antistatic property)
vs relative humidity, detennine the temperature effect, and set an effective lower limit for
controlled relative humidity in assembly areas.
Twelve materials were studied: four commercial antistatic films, including a polyethyl-
ene film with a radiation-cured coating, three cellulosic (cellophane) films, two kinds of
paper (present in static-controlled areas in the fonn of documents such as work instruc-
tions), a detergent (representing topical antistats), a static-limiting floor finish, and leather
(representing leather shoe soles).
With reference to Fig. A, antistatic polyethylene I met MIL-B-81705, Type II, and con-
tained an amide antistat; antistatic polyethylene 2 contained an ethoxylated tertiary amine
antistat; the antistatic nylon also contained an ethoxylated tertiary amine antistat; and the
paper was 20-lb bond made by the bleached kraft process. Also, the filter paper was Whatman
No.1 made from cotton linters; the detergent was Joy (Proctor & Gamble); the floor finish
was a high-solids acrylic; and the leather was split, vegetable-tanned, and unfinished. To
facilitate resistance measurements, both sides of a nonconductive film (cellulose acetate)
were dipped in the detergent and floor finish.
Strips of film or sheet were held between clamps to make a square, as in ASTM D 257,
and sealed inside jars with saturated reagent-grade salt solutions (MgCI 2, NaN03, KCI,
KN0 3, K2SO.) to control the relative humidity (see Table 1).1-3. A drying agent (anhydrous,
calcium sulfate with indicating blue color) afforded 0% RH.
The jars were stabilized at three fixed temperatures, 41 to 46°F, 75 to 80°F, and 108°F,
until the samples reached equilibrium moisture content and the resistance readings became
constant. Although no acidic or basic vapors that might affect resistance were found, confir-
matory detenninations were made at 75 to 80°F with an alternate set of salt solutions (KF,
K2C0 3, NaHSO., NaN0 3, NH4HlO.).
The apparatus is shown in Figs. 1-6. Figure 1 is an individual jar. Figure 2 shows a
leather sample in the clamps. The jars were placed in a refrigerator (Fig. 3) or a low-heat
oven (Fig. 4). Figure 5 shows the interior of the oven with a light bulb (bottom of picture) as
the thennostatted heater, then the fan, then the samples. In Fig. 6, a jar is tested under
ambient conditions.
286
PAPER NO. 17 287
Polyethylene with
RId.- Curld COetIng
Fig. A. continued.
PAPER NO. 17 289
Detergent
MgCI 2·6Hp 43 35
75 33
108 32
NaN0J 43 78
75 75
108 70
KCI 43 88
75 85
108 82
KNOJ 43 95
75 94
108 89
K2S04 43 98
75 97
108 96
KF·2Hp 75-80 27
K 2COJ·2H2O 75-80 43
NaHSO.. HP 75-80 52
NH4~P04 75-80 93
As seen in Fig. A, agreement of results with the standard vs alternate solutions generally
was good. Incidentally, sulfuric acid solutions, sometimes used to control relative humidity,
did evolve acidic vapor and were rejected for that reason.
The resistance across the films was measured at 100 V with a Beckman L-12 megohmmeter
with the cables shielded to prevent RF interference. Time of electrification was 5 s. Runs
were made in triplicate with the averages plotted in Fig. A.
Standard deviations generally were 5 to 60% of the mean. No data were rejected. The
clamps contacted both sides of the specimens, so R in Fig. A is the same as surface resistiv-
ity (in M Q/sq) only for the volume-conductive materials (paper, leather, and cellophane
films). For the other materials, the surface resistivity is twice the value of R in Fig. A.
The data provided several conclusions:
• Surface resistivity is controlled by two factors. First is the moisture content which, in
tum, is controlled by the relative humidity; second is the temperature. A similar tempera-
ture effect has been reported for charge decay time,4 which would be expected to corre-
late with surface resistivity for materials such as antistatic polyethylene. At constant
relative humidity, the resistance of the antistatic materials decreased exponentially with
temperature, as occurs with dielectrics' and semiconductor.6 Metals show the opposite
effect.
• The reality of the temperature effect is proved by the dry-air data. At high relative
humidity, the temperature effect becomes small or nil as the materials move from the
semiconductive toward the conductive state.
• Because of the significant thermal effect, the temperature should be given, along with
the relative humidity, in reporting surface resistivity data.
PAPER NO. 17 291
3 4
5 6
The plasticized, coated cellulosic film Ecostat,8 contains moisture and a hygroscopic
softener or plasticizer. No antis tat is concentrated at the surface where it can be removed by
rubbing.
After vigorous rubbing with this film, nonconductive surfaces remained nonconductive.
However, they were made antistatic by being rubbed with, or merely pressed against, the
antistatic polyethylenes or the polyethylene with the radiation-cured coating.
In contrast to other materials, Ecostat remained antistatic after brief rinsing with water.
Soaking at 81°F for 24 h extracted the plasticizer so the film reverted to unplasticized cello-
phane (Fig. 7) and the room-temperature cutoff relative humidity rose from 4 to 18%.
In preliminary static decay tests (FED-STD-101, Method 4046) at 8 to 10% RH and
75°F, Ecostat showed a decay time of 5 to 11 s. It was noncorrosive to copper or Sn62 solder
when exposed in contact with these metals at 100% RH and 75 to 80°F for one year or
120°F for one month.
This film appears to be suitable for a noncontaminating intimate wrap; e.g., inside MIL-
B-81705, Type I, to meet the intent ofMIL-HDBK-773. It also is biodegradable.
PAPER NO. 17 293
CI Cellophane, pla.t•
1~pp~-r~--. .~~--p-~-r. .
10'
•
~
10'
References
1. International Critical Tables, Vol. I, p. 67.
2. Stokes and Robinson, Industrial and Engineering Chemistry, Vol. 41,1949, p. 2013.
3. Handbook o/Chemistry, edited by N. A. Lange, McGraw-Hill, Revised 10th Edition, 1967, p.
1432.
4. Evaluation Engineering, April 1982, p. 74.
5. ASTM D 257-78, Appendix XI.
6. Encyclopaedia Britannica, 15th Edition, Vol. 18, 1989, p. 241.
7. GlDEP Alert C6-A-87-09, November 12, 1987.
8. Evaluation Engineering, March 1990, p. 96.
PaperNo. 18
The EIA-541 electrostatic shielding test for packaging materials uses the relatively low
voltage of 1.000 V on a large flat electrode. This may simulate fields or low-voltage dis-
charges. but possible damage from ESD includes high-voltage discharges--often as visible
sparks-from people or conductive objects. This article describes a test method for simulat-
ing such discharges and outlines packaging strategies to resist them.
The test is conducted at -9 kV. the minus sign meaning that electrons flow from a charged
capacitor toward the package. This is not a worst-case human voltage; the machine-model
resistance of nearly zero is worst-case. and the ESD sensitivity of 100 V is nearly worst-
case. The test conditions constitute a reasonable requirement. passed by Types I and III of
MIL-B-SI705.
Figure I represents an idealized. nearly zero-resistance person with 200-pF capacitance
discharging from a sharp-tipped. hand-held tool after being triboelectrified to -9 kV (for
example. by walking on a carpet at 20% RH). The material under test was pulled tight over
the capacitive probe. Then the tip of the O.025"-diameter capacitor lead was touched to the
material for 1 s. The vertical shield prevented the E field of the capacitor lead from inducing
false voltages in the circuitry.
The sensor was an oxide-coated silicon wafer modeling a MOSFET gate. I A pulse of 100
V or greater across the disks of the capacitive probe punctured the oxide under a steel probe
pressed against the surface. The resistance of the oxide measured through the steel probe
n n
was over 10 12 initially and usually less than 109 when the oxide had been damaged. so
the failure criterion was the upper limit of detection of 2 x 109 n bv a digital multimeter.
The tests were run at approximately 50% RH.
Wafer-oxide failure rates for the 14 commercial materials in Table 1 are consistent with
MOSFET data in past work. 2 Results of the tests indicate that a conductive layer was re-
quired for passing the test (failure rate = 0/5). Foil bags (Materials 4 and 5) are prescribed in
MIL-HDBK-773 and have been recommended for highly ESD-sensitive items and worst-
case discharges. 2•3
It was also concluded that semitransparent vapor-deposited metallization is effective only
when buried under a layer of plastic (Material 6). The screen bag (Material 8) is very
discharge-protective2 and relatively transparent. but expensive.
The S-mil skin of Material 14 essentially is plain paper on the outside but carbon-
impregnated on the inside. This resembles the successful shielding combination of Material 2
294
PAPER NO. 18 295
Fig. 1. Apparatus.
1 0/5
12 1/5
2 12 5/5
lor2 3,9 or 10 0/5
1 11 2/5
2 11 0/5
3 or 11 5/5
9 3/5
10 4/5
12 1/5
12 2,3,9, 10, 11 or 12 0/5
3,9,10 or 11 12 0/5
3 or 10 2 5/5
9 or 11 2 4/5
lor 13 15 (0.5") 0/5
outside and Material 3 inside in Table 2. The 0.14" gap (largely air) inside Material 14 adds
to the protective effect.
The conductive layer should be buried (Materials 4, 6, 7, 8 and 14) rather than exposed
(Materials 3, 5, 9, 10 and 11). This not only improves shielding but also discourages ESD
damage by the charged device model.4
Sparks did not correlate with ESD failure. Material 3 failed with a visible arc, but Mate-
rial5 passed; Materials 1,2,12 and 13 failed without an arc. The arc to Materials 9,10 and
11 may have contributed to failure by creating bare spots approximately 0.01" to 0.02"
diameter in the metallization. (Fig.2).
Choices are limited for single materials pressed against the ESD-sensitive item, but op-
tions multiply for combinations including double bagging and air gaps (Table 2). For ex-
ample, a pink poly bag (Material 1) or an externally, metallized bag (Materials 9 or 10) is
ineffective alone, but when combined correctly with pink poly on the outside-not the
inside-the metallization is buried and protected from spark damage.
Cellophane (Material 2) also is an effective cover layer. Materials 3, 9, 10 or 11 succeed
when combined with foam (Material 12).
Table 3, derived from Table 2, provides examples of six construction types with the
principles of their success: Faraday-cage shielding, a dielectric barrier of relatively heavy
plastic, and protective spacing. Even inexpensive, unmetallized materials are adequate with
air gaps, which need be only OS'.
Commercially recyclable and biodegradable containers use construction types 5 and 6
Other commercial packaging utilizes construction type 4, in which a transparent bag is
inflated with the least expensive material-air.
Common packaging materials in the right configurations can prevent the damage of high-
voltage discharge. The techniques of buried metallization and protective spacing offer many
possibilities for cost-effective package design.
PAPER NO. 18 297
Fig. 2. Bare spot 0.017 inch in diameter created on surface of Material 9 by spark discharge.
References
I. Kolyer, J. M., and Watson, D. E., "ESD Testing of Silicon Wafers," Evaluation Engineering,
September 1991, pp. 115-119.
2. Kolyer, J. M., and Watson, D. E., ESD from A to Z: Electrostatic Discharge Control for
Electronics, Van Nostrand Reinhold, 1990, pp. 173, 175-181.
3. Holmes, O. C., Huff, P. J., and Johnson, R. L., "An Experimental Study of the Screening
Effectiveness of Antistatic Bags," Reliabiliy Analysis Center EOSIESD Symposium Proceed-
ings, Philadelphia, PA, 1984, pp. 78-84.
4. Kolyer, J. M., and Watson, D. E., "The Charged Device Model and Work Surface Selection,"
Evaluation Engineering, October 1991, pp. 110-117.
Paper No. 19
A static charge is a static (immobile) excess or deficit of electrons on a surface. This charge
emanates an electrical field (E field) usually easily detected by a static meter.
However, the E field is weak and difficult to detect when suppressed by proximity to a
nearby conductor as in packaging laminates with nonconductive outer layers sandwiching a
conductive or static-dissipative interJayer. The hard-to-find charges on the laminates are
called cryptocharges (meaning hidden or secret).
A cryptocharge, being immobile, cannot flow into a device lead or tenninal to cause
ESD damage, and the associated E field generally is too weak to cause damage by the field-
induced model (FIM). Buried-metal ESD-shielding laminates can be antistat-depleted and
nonconductive without being an FIM threat. Cryptocharges, however, can be brought out of
hiding by induction.
A conductor brought within an E field is polarized by induction. If the outer end or side
of the conductor is momentarily grounded before the conductor is removed from the field,
the conductor will have an opposite charge to that of the field. This is compound induction. I
Although cryptocharge fields are weak, when a metal plate is pressed to the charged
surface, momentarily grounded, and then lifted from the surface, potentials of up to several
kilovolts appear on the plate. This effect is contact compound induction (CCI).
To study CCI, several laminated and unlaminated films or sheets were tested (Table I).
Material I was a clamshell pack with the film stretched over a 3.8" x 3.8" x 0.3" cardboard
frame with 2.3" x 2.3" windows. Materials 2, 3 and 5 were bags stretched over cardboard in
the same configuration as Material I. Material 4 was a 4" x 6" x 1.1" box.
Samples were electrically cryptocharged for I min with the window area, or the side of
the box, resting on a charged 2" x 2" x 0.03" aluminum plate. A 0.2-lb weight on the sample
facilitated contact. See Fig. I.
Alternatively, samples were tribocharged by stroking them on a nylon carpet at 70-75°F
and 55-65% relative humidity (RH).
CCI was accomplished by pressing a 2" x 2" x 0.03" aluminum induction plate with a
Teflon™ handle against the cryptocharged surface with a force of 0.5 lb for 5 s. Then the
plate was grounded momentarily by being touched with a grounded person's finger before
being lifted from the surface (Fig. 2). The voltage on the plate was read with a Trek Model
512 Field Meter calibrated with a known -7 kV on the same plate.
299
300 APPENDIX
Table 1. Materials.
SURFACE RESlmVITY,
OHMS!SQUARE@ 100 V
AND 65% RH (75°F)
MAn:RIAL TOTAL THICK- SURFACE
NO. DESCRIYI10N NESS, MIL LAYERS INn:RLAYER 0un:R INNER
+OR-
7kV
POWER
SUPPLY
Table 2 shows that antistatic surfaces were unfavorable for cryptocharging. Material 2
lost its charge relatively rapidly while Materials 3 and 4 held no charge. Thus, cryptocharges
would seem to reside on the outer surface, not inside the interlayer.
Cryptocharges also were localized in the window touching the electrified plate (Fig. 1),
whereas a charge within the interlayer should have spread throughout the sample. This led
to the conclusion that the role of the interlayer is suppressing voltage, not holding charges.
Like other static charges, cryptocharges decayed exponentially because the rate of decay
(-liVldt) was directly proportional to the diminishing potential or voltage (V) that drives
drainage of charge from the surface and, by means of the E field, attracts oppositely charged
air ions. Then
BV=-dVidt
InV = InVo-Bt
where V is voltage in kV at time t in days, In is the natural logarithm, and Vo is the initial
voltage in kV.
The values of B in Table 2 were obtained by exponential regression analysis of several
data points, r being the correlation coefficient. The decay time in days from 5 kV to 1 kV is
calculated by dividing In5 by B. The half-life of any cryptocharge is In2 divided by B.
The effect of RH on the decay rate is uncertain. Decreasing humidity lowers the rate of
charge drainage over surfaces but may slightly increase the mobility and neutralizing effec-
tiveness of air ions. 2 In these tests, the temperature was 75-80°F and the RH was 55-65% as
measured by psychrometry. The concentration of air ions (conductivity of the air) was un-
known, so the decay rates in Table 2 are only comparative.
Unlike static charges that are not voltage-suppressed, cryptocharges resisted neutraliza-
tion by ionized air from a blower because the E fields were very low (100 V to 300 V on the
surface as sensed by the field meter). Thus, cryptocharges are not only hard to detect but
also hard to remove.
CCI must be distinguished from tribocharging. Uncharged Material 6 strongly tribocharged
the induction plate with a positive voltage by separation, but uncharged Materials 1,2,3,4,
and 5 gave negligible voltages by separation. Also, CCI for Material I was not inhibited by
cementing Material 1 to the face of the induction plate, whereas tribocharging would have
been minimized.
In all cases, momentary grounding was necessary for CCI but would not have been for
tribocharging. With Materials I, 3W, and 5, the polarity of the plate would have been posi-
tive for tribocharging, as it was for a negative cryptocharge; but the plate was negative for
a positive cryptocharge (Table 2). Incidentally, weak CCI charges on the plate occurred
without momentary grounding and might be explained by bleed-off into the air.
The CCI process apparently involved negligible electron flow between the sample and the
induction plate because covering the plate with volume-nonconductive film (Material 1)
did not inhibit CCI. Electron exchange, as opposed to induction, also would have made the
plate the same polarity as the cryptocharge instead of the opposite polarity.
Discharges from the induction plate, with a capacitance of approximately 1 pF when
held by the handle, damaged wafer oxide. 3 The oxidized wafer was the kind from which IC
chips are cut and represents the gate of a MOSFET. Failure occurred when the discharge
created a voltage differential across the oxide in excess of its dielectric strength of approxi-
mately 100 V. See Figs. 3 and 4.
The inverse situation is the charged device model (COM) in which a charged DIP with
1-3 pF capacitance discharges to ground with resulting ESD damage. I
In terms of the l-pF plate, the last column of Table 2 gives the dangerous period for an
initial CCI voltage of 5 kV. This was approximately 3 days for Material 2W, a week for
Material I, and 2.5 weeks for Material 3W.
The persistence of the cryptocharge on Material 3W might be due to effective voltage
suppression by the relatively thick metallic interlayer or to very low conductivity of the
PAPER NO. 19 303
!!IS
3/S
2IS
115
VOLTAGE,kV
References
1. McAteer, O. 1., Electrostatic Discharge Control, McGraw-Hili, 1989, pp. 58,176 and 179.
2. Jonassen, N., ''The Physics of Electrostatics," distributed at the 1984 EOS/ESD Symposium.
3. Kolyer, J. M., and Watson, D. E., "ESD Testing of Silicon Wafers," Evaluation Engineering,
September 1991, pp. 115-119.
4. Kolyer, J. M., and Watson, D. E., ESD from A to Z: Electrostatic Discharge Control for Elec-
tronics, Van Nostrand Reinhold, 1990, pp. 160, 161, 172 and 210-215.
PaperNo. 20
This paper is reprinted, by pennission, from the Proceedings of the 1994 EMCIESD
1nternational Conference, held in Anaheim, CA, April 12-19, 1994.
John M. Kolyer
Abstract
An ideal electrostatic discharge (ESD )-protective package would shield against worst-case
external discharges and be permanently static-dissipative and volume-conductive. Also, it
would have non-ESD properties such as recyclabi/ity, biodegradability, and flame-resis-
tance. This paper discusses state-of-the-art packaging and proposes superior designs ofthe
future that may be achieved, without major breakthroughs, by consolidation and extension
ofpresent techniques. Basic to this development will be realistic ESD tests, several of which
are described with examples of data.
Introduction
The title of this paper says "toward" because an ideal product is, by definition, impossible. It
would have to include conflicting properties as in "biodegradable, heat-sealable, low-cost,
transparent, stainless steel." Furthennore, it would have to be universally suitable for count-
less applications with specialized requirements. However, laminations of paper, plastics, and
metals can combine diverse properties and approach an ideal package for specific purposes.
This paper discusses key electrostatic discharge (ESD) properties and realistic tests for
them. Non-ESD properties are also reviewed. Then promising packaging designs that are
now available or on the horizon are discussed, with pros and cons of the two major catego-
ries, plastic and paper.
305
306 APPENDIX
lack of undesirable antis tat transfer to other surfaces), abrasion resistance (no loss of static-
dissipativeness and no sloughing of conductive particles), minimal triboelectric charging
propensity for nonconductive surfaces touched by the packaging material, and volume con-
ductivity to prevent stored energy. These properties will be further explained in the follow-
ing discussion of tests.
CDM Safety. In the CDM damage mechanism, a charged ESD-sensitive device discharges
to a conductive surface. Tests have been devised to simulate this process,6-9 and Fig. 4
shows a method using an SED.· An indirect method for judging CDM safety is to measure
the resistance to ground of materials with a buried conductive layer (Fig. 5) or resistance
between electrodes for other materials (Fig. 6). The resistance reading can be related to tests
with metal oxide semiconductor field-effect transistors (MOSFETs) in simulated CDM
events.6 Table 3 gives examples.
Voltage,tV 10 6 10
Capacitance, pF 200 350 350
Probe diameter, inch 0.025 0.25 0.25
F Poor
p F F Fair
P p F Good
P P P Excellent
Package
Surface
'--
Discharge
Probe
Tel
material (Fig. 6). Resistance for a paperboard packaging material versus voltage and RH is
shown in Fig. 7. Table 4 gives current calculated from resistance for several materials.
Obviously, low current-carrying capability is desired for safety for people as well as de-
vices. An alternating current of 4-21 rnA causes reflex action (which could result in injury),
21-40 rnA causes muscular inhibition, and 40-100 rnA causes respiratory block.1O there-
fore, only the fIrst material in Table 4 is a current-carrying hazard for personnel at Ito V.
Surface Resistivity versus RH. Surface resistivity values are only approximate because
the measurement depends on factors such as the pressure of the electrodes against the sample.
Nevertheless, this property is useful as a guideline. At approximately 1012 ohms/square,
Megohmmeter
~
LL2
Sample
Buried Conductive Layer
ohmmeter
~
LD
materials become nonconductive and hold static charges for several seconds or more. Re-
sistivity rises with falling RH, and the RH at which 1012 ohms/square is reached is the cutoff
for retention of the antistatic property. Surface resistivity at various RH levels is measured
by sealing samples above humidity-regulating saturated chemical solutions (or anhydrous
calcium sulfate for 0% RH) as diagrammed in Fig. 8 and explained in detail in Ref. 11.
Examples of curves of surface resistivity versus RH are shown in Fig. 9. A low-RH cutoff is
desired for packaging materials because dry air may be encountered in shipping and han-
dling. During unpacking, for example, triboelectric charges not drained to ground (through
the work surface or operator's fingers) could cause damage by the FIM.
10 8
en
E
.c
0
cC 10 7
• 10V
106
• l00V
0 _ _ _ 500V
o 10 20 30 40 50 60 70 80 90 100
R!i%
Fig. 7. Resistance to ground for cardboard with buried conductive layer.
before and after exposure of samples to two common solvents. The antistatic coating failed
because the antistat in it was easily leached.
Antistat Transfer. Transfer of antistat to the surface of packaged items can cause unde-
sirable effects such as lowering of surface impedance of resistors, stresscracking of poly-
carbonate plastic, and discoloration of epoxy paint. 2 In a simple test (Fig. 10), samples are
pressed against nonconductive plastic (cellulose acetate) sheets, which become antistatic if
antistat is transferred. In Table 6, antis tat is seen to have been transferred from an antistatic
coating. Superior antistatic materials are inherently antistatic, without depending on an ad-
ditive, or they contain a polymeric or chemically bound additive that cannot migrate.
Abrasion ResIstance. This property is measured with a Thber Abraser in which an abra-
sive wheel presses the surface as shown in principle in Fig. 11. The difference in cycles
required to wear holes through the liner (skin) of the four conugated paperboard samples in
Thble 7 is caused by varying liner thickness. Superficial conductive coatings were quickly
removed, whereas a buried conductive layer was not reached for hundreds of cycles. Abra-
sion is of concern because it might cause loss of BSD shielding and, more importantly,
result in sloughing of conductive particles that could bridge circuit lines in open devices.
~
~
.
.......
~
>
...~
cI
I
1=PIp8I'boenI
2 • ConducIIvely coated papeltloerd with
heavy antIatatlc vamlah
3. hpeltMMlrd with bUried conductive
layers
4. Paperboercllrnpragnated with carbon
polyethylene, may owe their success to the undesirable transfer of antistat, causing liquid to
separate from liquid on a microscopic level. For rigid containers, which do not rub the
packaged item on loading or removal as bags do, a medium-charging rating should suffice.
Crypotocharges. These hidden charges are an example of stored energy. They are diffi-
cult to detect with a static meter and occur on packaging material with a buried conductive
layer that suppresses the voltage on the charged (electron-rich or electron-depleted) non-
conductive surface. 12 Cryptocharges are brought out of hiding by contacting the surface
with a conductor, which becomes polarized, and then momentarily grounding the conduc-
tor, e.g., by touching it as shown in Fig. 12, so that electrons flow on or off the conductor to
PAPER NO. 20 313
ISOPROPYL
WATER RINSE WATER SOAK ALCOHOL SOAK
MATERIAL ORIGINAL (30 SECONDS) (30 MINUTES) (30 MINUTES)
give it a net charge. This charge then manifests itself as an E field (electrical field) when the
conductor is lifted from the surface. This special case of compound induction. called con-
tact compound induction (CCO. can impart several kV to the conductor, which could dam-
age ESD-sensitive items by a direct discharge according to the MM or, if the conductor is
part of an ESD-sensitive item, by the CDM. Scenarios for damage to devices by cryptocharges
are unlikely but must be considered, especially for critical hardware such as defense or Iife-
support systems.
Transparency. This property is easily obtained with plastics but not with paper products.
Cellophane, like paper. is cellulosic, and the correct grade, e.g .• plasticized and nitrocellu-
lose-coated. is permanently antistatic, noncorrosive, and heat-sealable but has low tear
strength and is best used as a window, liner, or intimate wrap. '3
Samples
Low Corrosivity. A stringent test13 is to expose copper or Sn63 solder in contact with a
packaging material at 100% RH for three months at 7S-80"F or one month at 1lO°F. There
are many standard tests. e.g.• FED-STD-IOI. Method 3005. Paper products should have a
low reducible sulfur content by the Technical Association of the Pulp and Industry (TAPPI)
test T406 or FED-STD-IOl, Method 3005.
Strength and DurabiUty. The American Society for Testing and Materials (ASTM) has
many standard tests for properties such as tensile strength (ASTM D 882), tear resistance
(ASTM D 1004, D 1922). puncture resistance (ASTM D 2582), impact resistance (ASTM
D 3420, D 4272), and resistance to being dropped (ASTM D 959, D 4(69). Less general
properties, e.g., score resistance, are important in specific applications.
tr,\ Rotating
-
8 Sample
Recyclability. Plastics are recyclable but require sorting for best results; a single type of
plastic can be added as regrind to virgin plastic for molding, but mixed plastics can be used
only for noncritical items such as low-pressure sewer pipe. Another approach is to depoly-
merize mixed plastic waste to basic oil feedstock. 14 Germany is especially strict about recy-
cling plastics, with a goal of 80% of plastic packaging to be recycled by July 1995. Of
course, reuse of plastic items is preferred to recycling; some companies reuse plastic trays,
for example, to lower costs and protect the environment. Paper packaging has an advantage
in recycling because it need not be sorted (unless, for example, it has a high carbon content)
and can be mixed with general paper waste, such as newspapers, for repulping. The solid-
waste debate is discussed in a recent book" that challenges many conventional positions.
For example, conversion by U.S. fast-food chains from disposable to reusable packaging
would require consumption of huge arnounts of water, detergent, and energy for dishwashing.
Flame Resistance. There are many tests for plastics, e.g., Underwriters Laboratories test
UL94 for heavy sections or ASTM D 568 or D 1433 for sheet and film, and for paper
products, e.g., ASTM D 4433 for fire-resistant, treated paperboard. A new fire-retardant,
antistatic, recyclable, and biodegradable treatment for liner board for packaging was an-
nounced late in 1993. Some plastic materials are inherently fire-resistant, for example, one-mil
316 APPENDIX
RATING
OVERAU (PROPENSITY ro
VOLTAGE ON COUPON (tV). 11fREE
AVERAGE (kV). CHARGE CIRCUIT
READINGS AND AVERAGE
NEGLECTING BOARD
MATERIAL EPOXY I EPOXY 2 POLYIMIDE SIGNS LAMINATES)
Plain paperboard +1 +1 -7
(corrugated cardboard) +1 +2 -1 3 Medium
±l ±ll ::2
+1 +S -3
Carbon-loaded polyethylene +4 +18 +3
+1 +11 +1 8 High
ill ...n ±2
+S +16 +2
Antistatic polyethylene 0 0 0
(Mll..-B-8170SC, Type II) 0 0 0 0 Low
Q Q Q
0 0 0
Paperboard with buried +2 +2 -4
conductive layers +1 +4 -2 3 Medium
1:3. +7 =l
+2 +4 -2
Paperboard impregnated +1 +10 -1
with carbon +O.S +12 -2 4 Medium
-.Jl ill ::l.
+1 +11 -1
Paperboard with heavy +14 +14 +1
antistatic coating +11 +11 +1 8 High
+11 ±ll .J!
+12 +12 +1
polyvinyl chloride (PVC) packaging film that is made permanently antistatic by a propri-
etary polymeric additive. Aluminum trihydrate and magnesium hydroxide are environmen-
tally friendly flame retardants because they function by endothermically decomposing to
release harmless water rather than to toxic gases. 16 Other inorganic, nonhalogenated flame
retardants include intumescents like ammonium polyphosphate.1 7 All these additives can be
used in (or on) both plastic and paper packaging.
Other Properties. The many other properties of importance in packaging include perme-
ability, markability, heat-sealability, and gluability with both resin and hot-melt adhesives.
It is important to realize that just one "Achilles heel" in some special property can dis-
qualify an otherwise ideal package. In fact, it is questionable how many applications a
single "ideal" package design could satisfy. The "ideal package" may in fact be a family of
designs with intrinsic advantages, e.g., permanent antistaticity and volume conductivity, in
common.
PAPER NO. 20 317
Paper. Paperboard has the advantages of being inherently volume-conductive and biode-
gradable, as well as reusable and recyclable. Its weaknesses are permeability, opacity, flam-
mability, and limited durability. Several commercial corrugated paperboard designs, which
are stiff, strong, and lightweight, were tested with the following results. Paperboard with
carbon impregnated throughout the skin was too conductive (Table 3, Fig. 1), but conductively
coated paperboard with a heavy antistatic varnish was CDM-safe (Table 3), safe to person-
nel (Table 4), and permanent (Tables 5 and 6). Though the varnished paperboard lacked
high abrasion resistance (Table 7), a highly conductive surface was never exposed.
Conductively coated paperboard with a thin varnish was undesirably conductive (Tables 3
and 4). A particularly effective design has a conductive carbon layer on the inside of each
318 APPENDIX
liner (skin) in contact with the medium (fluted paper). Thus, the conductive layer is safely
buried and cannot be easily abraded or be a CDM or personnel hazard. This design has
generally excellent ESD properties (Tables 3-7 and 9, Figs. 7 and 8) but gives only fair
shielding against external discharges. Excellent shielding is achieved in this design if
metallized paper is substituted for the carbon layer, while all the other desirable proper-
ties are retained (Table 9); aluminum is, in this context, nontoxic, and the amount used
(approximately 1000 A) is too small to interfere with repulping or biodegradation. Two
layers of metallized paper, one inside each liner, are needed for the excellent shielding
PAPER NO. 20 319
rating; a spark discharge creates a hole in the metallization on the first layer but leaves the
second layer intact and is intercepted by it. Metallized cellophane is a barrier to gases but is
more expensive than metallized paper and, not being fibrous, can cause fish-eyes in re-
cycled paper. Aluminum foil, incidentally, gives excellent high-voltage discharge protec-
tion but presumably interferes with repulping.
Plastic. The antistat permanence and volume conductivity achieved for plastics in recent
years has been a major breakthrough. This corrects the deficiencies in these properties found
in MlL-B-81705C, Types I and III, which are used as a reference point in Table 9, and
leaves only lack of biodegradability and lack of flame resistance. These may not be major
deficiencies, but efforts can be expected to correct them, at least for special packaging
applications. In fact, plastic packaging already available meets most of the "future" ratings
in Table 9--{)r can meet them if the right materials are combined. For example, aluminum
metallization buried between layers of the PVC film mentioned above might give a perma-
nently antistatic, flame-resistant version of MIL-B-81705C, Type I. Biodegradability is a
more "blue sky" property because, for one thing, the plastic might degrade prematurely on
the shelf or in use, but new products are appearing on the market. IS Also, transparency and
excellent (rather than good) high-voltage discharge resistance are incompatible properties
with present methods of metallization, and an extremely good vapor barrier requires metal
foil, an opaque layer of vapor-deposited metal or, possibly, a layer of expensive
polychlorotrifluoroethylene (PCTFE) transparent film.
Paper. Paper has the inherent limitations of being opaque and highly permeable. Coated
cellophane is transparent and a good vapor barrier and is volume-conductive and biode-
gradable, 13 but it is not recyclable for repulping (only for producing more cellophane). Flame
retardants can be added to paperboard, but corrosivity must be avoided. A flame-retardant
treatment for liner board recently announced claims to be colorless, odorless, nontoxic,
biodegradable, and recyclable. Durability of paperboard packages can be improved by us-
ing heavy sections, but there will be penalties in weight and cost. The paperboard with
buried metallized paper proposed in Table 9 remains to be commercialized and would be a
good future product on which to improve with flame retardants, etc.
Plastic versus Paper. In general, plastic is stronger and more durable than paper, but
paper is more "natural"; it is made from a renewable resource, is biodegradable, and is
easily recycled. Also, in regard to ESD, paper has the advantage of being permanently
antistatic and volume-conductive, but state-of-the-art plastic compounds also have these
qualities. Therefore, paper and plastic are moving toward the ideal package from differ-
ent positions. In the foreseeable future, neither material will preempt the other, and paper
and plastic constructions will coexist along with mixed constructions, e.g., paperboard
boxes lined with antistatic plastic foam. Of course, combinations like the latter could be
recycled only by separating the materials. A foam-lined paperboard box would be reused
as many times as possible and then peeled apart for separate disposal of the foam and
paperboard.
320 APPENDIX
Conclusion
Much progress toward an ideal ESD-protective package has been made in recent years.
Advanced materials are available, realistic tests have been devised. and ingenious, cost-
effective packages can be created by metallization, lamination, incorporation of cushion-
ing such as foams, rigid mounting of items to prevent triboelectric charging, protective
spacing with air gaps, and other design features. Biogradability for plastic packages will
be difficult to achieve but may be unimportant, while flame retardance for either plastic
or paper can be attained to a considerable degree if demanded. Certainly, overly conduc-
tive surfaces, less than good high-voltage discharge resistance, fugitive antistats, and
stored energy will be - or should be - problems of the past.
Without major breakthroughs, consolidation and extension of present techniques should give
products approaching the "ideal" packages in Table 9. These may be more expensive than
current packages, but greater durability as well as emphasis on reuse will lower the effective
cost. Meanwhile, less toxicity and more recycling and reuse will protect the environment.
References
I. J. M. Kolyer, "Fundamentals of ESD Control," EMC Technology Magazine 1992 Expo Techni-
cal Record, May 1992, pages 154-161.
2. 1. M. Kolyer and D. E. Watson, ESD from A to Z: Electrostatic Discharge Controlfor Electron-
ics, Van Nostrand Reinhold, 1990.
3. 0.1. McAteer, Electrostatic Dischange Control, McGraw-Hili, 1989.
4. 1. M. Kolyer, "Realistic Testing of ESD Materials," EMC Test and Design, September 1993,
pages 28-31.
5. 1. M. Kolyer and D. E. Watson, "Packaging for High-Voltage Discharge Protection," Evalua-
tion Engineering, March 1992, pages 96-100.
6. 1. M. Kolyer and D. E. Watson, "COM and Work Surface Selection," Evaluation Engineering,
October 1991, pages II 0-117.
7. D. C. Anderson, "A Simple Approach to ESD Damage Prevention," EMC Technology, Marchi
April 1991, page 38.
8. "New Test Proposed for Checking ESD Safety of Materials," Compliance Engineering, Fall
1990, page 77.
9. J. M. Kolyer, "Is Your Work Surface COM-Safe?," EOSIESD Technology, February/March 1992,
pages 27-28.
10. DoD-HDBK-263, May 2,1980, page 46.
II. 1. M. Kolyer and R. Rushworth, "Humidity and Temperature Effects on Surface Resistivity,"
Evaluation Engineering, October 1990, pages 106-110.
12. J. M. Kolyer and D. E. Watson, "Hidden Charges on ESD-Protective Packaging," Evaluation
Engineering, September 1992, pages 94-100.
13. News notes in EOSIESD Technology, OctoberlNovember 1989, page 9, and Evaluation Engi-
neering, March 1990, page 96.
14. P. L. Layman, "Advances in Feedstock Recycling Offer Help with Plastic Waste," Chemical
and Engineering News, October 4, 1993, pages 11-14; News note: "Group Formed for Recy-
cling Plastics into Feedstocks," ibid, November 29,1993, page 41.
15. 1. H. Alexander,ln Defense of Garbage, Praeger Publishers, 1993.
16. S. Ainsworth, "Magnesium Oxide Finds New Applications," Chemical and Engineering News,
October 25, 1993, pages 15 and 16.
17. "Flame Retardants: Processors Learn to Work With Halogen-Free Systems," Modern Plastics,
September 1993, pages 55-60.
18. "New Players Emerging in Biodegradable Polymers," Modern Plastics, October 1994, pages
33-37.
Index
321
322 INDEX
275; illus., 295; illus., 307; for boxes for SSP, 48, 311; table,
tables, 195; table, 203; table, 275; table, 295; table, 300
205; table, 227; table, 262; in packaging specifications, 111
table, 275; table, 295; table, tote box, cost of, table, 142
307 Carpet
Bags, Faraday-cage. See Faraday antistat-treated, 18, 83, 91
cage charge from walking on, 14, 18, 73;
Basic Rule illus., 9; mus., 17; table, 36
categorization of, table, 23 charge generation on plastic by, 299
central in operator training, 105, with conductive fibers, criteria for
106 selection, 73, 102
definition of, 24, 80, 109 Carriers, requirements for, 111
discussion of, 28 Cathode ray rubes (CRrs)
importance of, 1-2 charges on, 59
in model handling and assembly screens for, 59, 105, 141
specification, 80 Caution tag, 87
in model packaging specification, CCI (Contact Compound Induction),
109 299,313; mus., 301; illus.,
Benchtops. See Workbenches 317
Bins, criteria for, 101 CDM
Biodegradability, 315 categorization of, table, 23
Bleed-off time (decay time), 41, 264, definition of, 24
277 discussion of, 7-8, 16, 29,44-46,
Blow-off nozzles, 37-38, 73; table, 115, 162, 253, 254, 260,
190 263-271, 281, 285, 305, 306;
Boxes, criteria for selection, 101 illus., 7; mus., 15; illus., 277;
Boxes, tote. See Tote boxes allus., 308; table, 309; table,
Brushes 318
Criteria for selection, 105 involving operator, 45, 162, 186,
evaluation of four types, 60-61 285
Bubble-wrap, antistatic. See Anti- test on tote boxes, 231; table, 229
static bubble-wrap CD Rule
Buyers' guides for packaging mate- applied to
rials, 49 brushes, 61
clothing, 57
Capacitance gloves, 85-86
categorization of, table, 23 screwdriver handle, 61
definition of, 24 walls, 58, 91
discussion of, 29 windows, 59
of capacitor (150 pF) used in as primary defense, table, 158
shielding/discharge test, 183, categorization of, table, 23
294; table, 307 definition of, 24, 198
probe used in shielding/discharge discussion of, 29,45, 116
test, 48, 262; illus., 295 experimental derivation of,
Cardboard 188-189; illus., 189
324 INDEX
tape for holding shields on module, required in SSW, 55, 94, 157,250
68 safety of, 95, 143
threads in smocks, 61 springs for safety, 143
trays, 101 two-conductor (resistive) type, 214,
Conductive tote boxes. See Tote 250; illus., 215; illus., 247
boxes, conductive versus periodic monitoring, 213,
Conductor 250
categorization of, table, 23 Contracting, 136
checking of grounding with Corona discharge, 9, 11
ZapOasb,247 Corrosion
definition of, 24 by workbench laminates, 63
discussion of, 30 of silicon wafers, 42, 52, 63
isolated, charging by ionizers, of solder by antistat containing
66-67 n-octanoic acid, 238-239,
when to allow in SSW, 53 242; table, 240
Connector dust cover, special test and special test for, in SSp, 49
standard test for, 59-60 Corrosivity in packaging, 110, 114,
Connectors 314
dust cover for, 59-60 Cost-effectiveness of ESD-control
electronic box, packaging of, 70 program, 134-135, 153
power turned off before inserting Cots. See Finger cots
ESDS item, 86 Coupon
shunt for, plastic, 39-40, 62, 104 categorization of, table, 23
Contact charging, 10 definition of, 24
Containers for storage or in-plant discussion of, 30, 234
transfer, 112 for automated process, 72,
Contamination 233-237; illus., 236
by antistatic polyethylene, for testing SSP, 47
238-243 CRrs. See Cathode ray rubes
of packaging, 111, 114 Cryptocharges, 299-304, 312-313
Continuous wrist-strap monitor Current-carrying capability, 143, 306,
alert signals, 95 308; illus., 309; table, 227
as primary defense, table, 158 Curve-tracer, 146, 151, 182
band,95 Cut-off relative humidity, 292, 309
calibration of, 94
capacitance type, 213; illus., 214 Damage mechanisms
categorization of, table, 23 check list of, 148
cord,95 in automated process, 236-237
cost-effectiveness of, 216 relation to defenses and hazards,
criteria for selection, 94-95, 100 table, 157
definition of, 24 Decals, on certified operators' badges,
discussion of, 30,157,207-217, 106
250 Defenses
procurement requirements for, primary and secondary, table, 158
94-95 relation to hazards and damage
326 INDEX
158 FFB
elimination of unnecessary, 249 categorization of, table, 23
in future, 160 definition of, 25
in SSW, 54-55 discussion of, 33, 45, 256; ilIus.,
ESDS 15; iIlus., 45
categorization of, table, 23 in automated process, 236
definition of, 24 Field
discussion of, 32 categorization of, table, 23
ESDS item controlled within SSW, 50
categorization of, table, 23 definition of, 24
definition of, 24 discussion of, 4-9, 11, 31, 116, 162,
discussion of, 32 299, 313; ilIus., 6; ilIus., 7;
in future, 161-162 ilIus., 17
selection of packaging materials emission, 11
for, 169-174, 294-320 from lights and light fixtures, 92
Evaporation, charge formation by, 12 from transformers or electric
motors, 91
Facilities and Industrial Engineering, from various objects, MOSFE1S
136-137 damaged by, 187
Failure analysis, iIlus., 20 guarding against, 155
Faraday cage hazard at the work station,
categorization of, table, 23 188-194,247
definition of, 24-25, 109 penetration of walls by, 91
discussion of, 32-33, 48-49, 116, read by rIeld meter to determine
170, 249, 251 charge on operator, 208
foil laminate relation to damage mechanisms and
bags, table, 170; tables, 195; defenses, table, 157
table, 202; table, 205; table, strategy of keeping weak, 155
262; table, 275; table, 295; Field meter
table, 318 as noncontact voltmeter, 34, 191;
criteria for selection, 99-100 table, 191
in model packaging specifica- categorization of, table, 23
tion, 109, 129 checks with, to enforce CD Rule,
MIlrB-81705, 1YPe I as, 170, 56; table, 197
201; table, 318 cost, 140
part of conservative approach, criteria for, 102
156,249 definition of, 25
for SSP, 48-49 discussion of, 9-10, 12, 33-35,
tote boxes, 49, 228, 331 140,141
Faraday cup for monitoring of necessary non-
categorization of, table, 23 conductors, 84
definition of, 25 in future, 160
discussion of, 33, 114 measuring apparent charge with,
in triboelectric charging tests, 27,181-182
230; iIlus., 221 use of, check list, 152
328 INDEX
Ionizers (continued) 92
cost, 139, 141 Manufacturing function, commitment
electrical vs. nuclear, 37, 189-190; to ESD-control program
table, 190 required, 137
evaluated with MOSFET board, Marking, caution
189 in handling, 87
fanless, 67 of packaging, 111, 115
nozzles, criteria for, 103 Masking tape, conductive, 193, 198
ozone generation by, 38, 89 Materials, ESD-control
"piggyback" for grit-blasting, 68, and equipment,
193-194 approved, 95-105
precautions, 89 complementary, table, 197
pulsed DC, criteria for selection, 96-105,
criteria for, 103 196-198
safe distance from, 66-67; table, what to buy, check list, 150
190 approved, as primary defense, table,
shadowing effect, 89 158
space-charging by, 75, 275-276; in future, 159-160
illus., 276 in model packaging specification,
Items, definition of, 89 109-110
in SSW, 53-55
Labeling. See Marking, caution Materials, packaging. See Packaging
Labels, paper, allowed in SSW, 58 Mats, floor. See Floor mats
Laminate, foil. See Faraday cage, foil Mats, table. See Table mats
laminate Megohmmeter, 93,104,151; illus.,
Laminates, workbench. See Work- 311
benches, laminates Metallized (see-through) bags, limita-
Latent fallure tions of, 31, 48-49; illus.,
categorization of, table, 23 297; table, 205; table, 262;
definition of, 25 table 275; table, 295
discussed in video training tape, Metals, criteria for selection, 104
106 Microscopes
discussion of, 38-39, 49, 138, 147, in SSW, 54, 247
234,244,256; illus., 15 ungrounded, with rubber feet, 37
possible, of mishandled hardware, MIlrSTD-1686, replacement of,
146 122-129
Lights and light fixtures, ftelds from, Mishandled hardware, disposition of,
92 88, 145-147
Logistics, 136 MM
definition of, 25
Magnetic fields (H fields) discussion of, 8, 9, 16, 39,
controlled by soldering standard, 273-276; illus., 15
54 MOSFET
shielding from, by iron foil, 115 categorization of, table, 23
Maintenance of ESD-protected areas, damage tests, procedure, 182-183
INDEX 331
Static-limiting (continued) in SSW, Tesla coil test, 228; illus., 171; illus.,
52-53 204; table, 203; table, 205
test methods for, mus., 220; mus., Test equipment
221; mus., 222; illus., 223 check list of, 150-151
Statistical process control, 123 criteria for selection, 104
Steel, painted, 84, 263; mus., 270 Testing of ESDS items
Steel work surfaces, 91, 93 precautions during, 86-87
Stresscracking of polycarbonate by procedures, inspection of, 88
antistats, 238, 239, 242 Test, multiple-choice, for certifying
Summary of book, 158 personnel, 115-117
Surface resistivity Thermionic emission, 11
categorization of, table, 23 1bpical antistat
definition of, 26 applied to plastic parts of equip-
discussion of, 41-42 ment, 54, 96
effect of relative humidity and tem- applied to screwdriver handle, 54,
perature on, 277, 286-293, 61
310, 317-319; illus., 311; categorization of, table, 23
illus., 312 chloride-free, 42, 104
measurement of, for packaging, 113 cost, 110
meter, for static-limiting Door fIn- criteria for selection, 104
ish, 83,150 definition of, 26
meters in future, 160 discussion of, 42; table, 197
method of measuring, 182; illus., for treating
178 antistatic tote boxes, 179
of brush handles, 60 identification badges, 58, 84
of smocks, 61 necessary nonconductors in
requirements for packaging, no SSW, 84, 91, 198
versus time for aged antistatic walls or windows, 58-59
bags, illus., 179 possible contamination effects of,
Symbols, 112 238-243
Tote boxes
Taber abraser, 311; illus., 314 antistatic, need for antistat treat-
Table mats, 101, 117, 263; illus., 270; ment of, 179
table, 282 cardboard, 142,226-229, 285;
Tape table, 231
antistatic, 74, 255-256 conductive,
charging by stretching, 12 cost of, 142; table, 231
conductive, for holding shields on disadvantages of, 180, 185-186,
module, 68 226
conductive, masking, 193, 198 Corshield, 227-228; table, 227;
Techniques vs. materials and equip- table, 231
ment, table, 197 cost of, 142; table, 231
Teflon criteria for, 101, 185-187
charged by heat gun, 65 current carried by, 227
in triboelectric series, 43, 279 discussion of, 4, 226-232
INDEX 337