z

.SLAC-PUB-5199 April 1990 (El

-.

PRODUCTION OF 400 MIRRORS WITH HIGH REFLECTIVITY FOR USE IN THE SLD CERENKOV RING IMAGING DETECTOR*

VUV

K. Abe, P. Antilogus, D. Astoq2 A. Bean, T. Bienz, F. Bird,2(a) D. Caldwell, P. Co~le,~ D. Co~ne,~ P. Datte,3(b) J. Dubosq3 W. Dunwoodie,2 P. Gagnon: D. Hale,3 G. Hallewell,2 K. Hasegawa, M. Hilton,3(c) J. Hubeq3 P. Jacques5 R. A. Johnsoq6 H. Kawahara,2 Y, Kwoq2 D. W. G. S. Leith,2 A. Lq3 J. Martinez, L. Mathys,3 S. McHugh,3 R. Morrisoq3 D. Muller,2 T. Nagarnine,2 M. Nussbarq6 T. PaveL2 R. P1an0,~ B. Ratcliff, P. Rensing,2 A. K. S. Santha, D. Schultq2 S. Shapiro,2 A. Shoup, C. Sirnopoulos,2 E. Solodov,2(d) P. Stamer,5 I. Stockdahq6 F. Suekane, N. Togq2 J. Va vrq2 J. S. Whitaker,7 D. A. Willianq4 S. H. Williams,2 M. Witherell,3 R. J. Wilson,7 S. Yellin, and H. Yuta

THE SLD

CRID

COLLABORATION

of Physics, Tohoku University, Aramaki, Sendai 980, Japan 2Stanjord Linear Accelerator Center, Stanford University, Stanford, CA 94309, USA Department of Physics, University of California, Santa Barbara, CA 93106, USA 4Santa Cruz Institute for Particle Physics, University of California, Santa Crux, CA 95064, USA cs Laboratory, Rutgers University, 5Serin Phy sa PO Box 849, Piscataway, NJ 08855, USA 6Department of Physics, University of Cincinnati, Cincinnati, OH 45221, USA 7Department of Physics, Boston University, Boston, MA 02215, USA

- Department

Submitted
*

to Nuclear Instruments

and Methods B.

Work supported by Department of Energy contract DE-ACO3-76SFO0515; and National Science Foundation grants PHY88-13669 and PHY88-13018. (a) Present address: EP Division, CERN, CH1211, Geneva 23, Switzerland. (b) Present address: Texas Accelerator Center, 2319 Timberloch Drive, The Woodlands, TX 77380, USA. (c) Permanent address: Applied Magnetics Company, Goleta, CA 93117. (d)

Permanent address:

Institute

of Nuclear

Physics,

Novosibirsk,

630090,

USSR.

.-

ABSTRACT
Large Detector for experimental particle physics detection (CRID). at the

_c

The Stanford

SLAC Linear Collider contains a Cerenkov Ring Imaging Detector -. CRID mirrors have been successfully produced and installed.

The barrel mirror

The industrial

production

process, the quality (VUV)

control reflectivity

of the mirrors

produced,

and the results of

the vacuum ultraviolet An average reflectivity 180-230 nm wavelength of a typical measurements .._ incident

and mirror-shape

accuracy are described.

of at least 80% for light at 160 nm and 85% for light in the range has been achieved in the production The surface roughness loss due to scattering of over 400 mirrors distortion of the

size of 30 by 30 cm. imply that the light

and optical

is a few percent

light and the angular error due to shape distortion

is less than 1 mrad.

.-

1. INTRODUCTION
under construction particle physics

_*
-. A large Cerenkov Ring Imaging Detector (GRID) y2 is currently to be included detection in the Stanford Linear Large Detector Collider (SLC). for experimental This CRID

at the SLAC

consists of a cylindrical

barrel and two endcaps, as shown in fig. l(a).

This device is capable of determining which,

the actual radius of the Cerenkov ring associated with each charged particle along with a momentum In a gas ring imaging charged particle traveling measurement, provides particle identification.

detector 3y4 the Cerenkov through the radiator

light emitted is reflected

by a relativistic by a mirror to

medium

focus a ring image onto the photosensitive - detector consists of Time Projection tetrakis[dimethylamine]ethylene the photocathode mixed with

detector plane. The single-photon

position additive

Chambers (TPCs) with the photosensitive with an ionization gas. The TPCs threshold (drift

(TMAE, the drift

of 5.4 eV) as

boxes) use quartz

windows that are transparent -

to VUV light.

This apparatus permits the reconstruction

in three dimensions of the position where the photon ionized a TMAE molecule. The angle of Cerenkov radiation emission, from which the velocity of the relativistic particle can be determined, is found by reconstructing the ring image as registered

by the photodetector

for all the photons associated with a given charged particle. includes both liquid and gas radiators, gaseous TMAE spherical focusing as the photocathode, Each

The SLD barrel CRID mirrors,

long drift boxes of the TPC type containing chambers with charge division into 10 azimuthal

and proportional

readout, as shown in fig. l(b).

barrel half is divided associated liquid of 5 mirrors

sectors which contain 2 drift Each drift

boxes and their

radiators

and 20 mirrors.

box detects light from 2 rows The SLD barrel CRID

and all of the 10 mirrors

have a unique shape.

contains a total of 400 spherical mirrors.

3?he gas radiator c -. emitted

perfluoro-n-pentane5 for light

(C5F12)

in which

the Cerenkov

light

is

is only transparent (TMAE)

of wavelengths quantum

longer than efficiency of photons

165 nm, whereas

the photocathode below 220 nm. the constraint mirrors

h as a significant

only at wavelengths as possible within of the

In order to detect as large a number of a reasonable

cost, it was necessary to require the reflectivity

to be not less than 85% for light in the range of 180-220 nm. with which the CRID measures the Cerenkov angle of a particle

The accuracy

from the gas radiator of error: geometric, chromatic

is determined

by the contributions

from five separate sources multiple scattering, mrad

aberration

in the gas radiator,

measurement,

and momentum

smearing. 6 The angular reflection of the mirrors

error of less than fl

for the two focal lengths geometric contribution

of 479 mm and 508 mm ensured

a small

from optical uncertainty

distortions

to the allowed overall photoelectron and surface as

spatial reconstruction quality of mirrors

of about l-2 mm 3 . The focal properties for ring imaging but cost. 430 were coated, (UCSB)

which are suitable

work are not as demanding to maintain

those for an astronomical quantities A total completed CRID. with thin mirrors of 472 mirror by the University

telescope -mirror, and reasonable substrates,

are difficult

in large

of which

were successfully for the SLD barrel into of

of California,

Santa Barbara

The 400 mirrors

that best met our acceptance criteria production of VUV

have been installed control

the CRID the mirrors

vessel. We describe the mirror produced,

process, the quality reflectivity.

and the results obtained

2. MIRROR 2.1 Mirror Substrate

substrate

SUBSTRATE

Manufacturing
was required to have a surface finish light. This requirement which would limit

The mirror scattering

to a few percent of the incident

is due to the small

number _c

of photons

produced,

typically

10 or so. In addition, error of <l mrad.

the mirror

substrate were

was required

to give an angular reflection manufacturing

These requirements

achieved in a two-step

process.7 (In the slumping process,

(i) The 6 mm x 1 m x 1 m glass blanks were slumped. soda-lime

float glass was heated and pulled into spherical

molds under vacuum

while soft .)
-

(ii)

The spherical blank was ground and polished to produce a substrate roughness After passing mirror 53 nm (rms). visual inspection, the mirror substrate

with surface

was cut into one of the

10 different ._ _.substrate within

sizes, and the edges were chamfered

0.16 cm or less. The mirror were kept to

was placed into a size gauge to easily verify that the dimensions The 10 mirror

60.076 cm of the specifications.

sizes varied from approximately

26 x 23 cm to 35 x 28 cm. 2.2 Mounting

After at UCSB,

Fixtures

and Epoxy
passed the visual mounting epoxy inspection and dimension check

the mirror an annealed

substrate aluminum

spool was epoxied was used with

onto the substrate. jig designed to

Approximately

0.15 mm of DP190

a gluing

ensure careful spool placement. mirror substrates.

Figure 2 shows the mounting

spool shape used on all

Before settling spools, rigorous both

on the use of 3M DP190 epoxy to hold the mirror was carried out. The epoxy was required that absorb

to the mounting to adhere well to UV light in the

testing

glass and aluminum,

not to outgas by-products

range of 160-220 nm, and to have an acceptably

small thermal

expansion

coefficient.

The choice of DP190 was made from a small list of possible adhesives that had been previously tested for outgassing and approved 5 by the SLD CRID group. The epoxy

.-

was further -. f -I

tested at UCSB

by gluing a spool to a mirror

substrate,

and hanging

weights from the spool over a long period of time. In the course of production, undergone it was noticed that mirror substrates that had

a normal gluing procedure and had been subjected voids in the epoxy-glass substrate. interface.

to high temperatures, from only process was

._

developed one mirror -

The epoxy released totally used in this normal gluing

The epoxy thickness

not controlled

so a new gluing procedure

was developed in which the glue thickness

was increased to a minimum strength . to the spool. the epoxy-glass

of 0.15 mm using a spacer in the gluing jig. The epoxy temperature (40C) with

was then tested in an oven at the CRID operating

weights ranging from 0.91 kg (approximate Small voids (about interface 1 mm*)

mirror substrate weight) to 4.5 kg attached evolved into large voids (about 1 cm*) in

after a week, but all epoxy bonds held.

The area of the epoxy was also increased to give further confidence in the mounting spool-mirror substrate connection. To increase this area without system, an epoxy injection a complete redesign method was made

and reassembly of the mirror to introduce

support

1 cc of epoxy to fill the region between a previously after injection

glued spool s feet. distortion wa,s

A coated mirror remeasured.

showed no change when the optical

Such spools did not develop voids in the injection oven tests.

regions while voids A dramatic test of mirror 24 hours

did develop on the foot regions under low-weight this increased adhesion was provided substrate and subjecting

by taking an injected

and an uninjected

both of them to a 22.7 kg weight at 40C. Within

the uninjected attached

mirror substrate released the spool, while the injected one still remains One cc of epoxy was injected beneath the spool for all mirrors coefficients adhesive

1 year later.

as a result.

The most probable

cause of release is that the mismatched

of expansion of the glass, aluminum, thickness

and epoxy conspired with the minimal voids.

to stress the interfaces into developing

#&-an additional in the detector, f -. The secondary of the mirror mirror

protection

against a mirror

detaching

from its spool and falling on all mirrors.

a secondary restraint restraint

system was designed and installed

system consisted of four wires connecting spool. This restraint doing further

clips on the corners

and the mirror

system was designed to keep any damage, but not to hold it in a

whose glue joint rigid position

failed from

sufficiently
-

to be useful for focusing.

2.3 Surface

Roughness
of better than 3 nm (rms) is necessary in order to keep the 5%. The surface profile of about 3-D Micro-Surface Measurement

Surface smoothness

light loss due to scattering 70%

to below approximately using a WYCOg lens. The method

of the mirrors

was tested8

.System with technique structure.

a 5x magnification computer

makes use of an interferometer of the surface stylus profilometer system was

and an online

to provide

real time 3-D displays

This type of measurement At the beginning

is vastly superior to a traditional of each testing

in many respects. calibrated mirror

session, the WYCO

using a reference with a known surface roughness of 0.24 nm. Each tested at two or three different locations, and the peak to valley and of the

was measured

surface statistics surface roughness

were calculated

at each location. mirror

Figure 3 shows an example

profile of a typical criterion of nowhere substrates

substrate. of less than 3 nm, and a

The acceptance surface roughness mirrors.

of an rms surface roughness greater

than 5 nm, was met by 91% of all the tested criterion were repolished

The mirror

that failed the acceptance

or replaced

at no additional

cost.

2.4 Optical

Distortion
distortion of each mirror was measured by placing

The optical grid of accurately

a mask with

positioned

holes of a diameter

of 1 cm (made on a numerically

controlled f grid pattern

milling

machine)

in front

of the mirror (approximately

and illuminating 1 m away).

it with

a point

light source at the center of curvature was projected

The image of the of

onto a frosted

glass screen, which was set a distance a demagnification with

10 cm behind ._

the center of curvature.

This provided

of 10 : 1 of

the grid pattern.

This image was then photographed

a 1 : 1 macro lens (see on the grid points corresponding to

fig. 4). An ideal mirror scribed on the frosted angular

would produce light spots which fell exactly glass. Departures from the ideal curvature,

errors of approximately patterns

1 mrad at the ground glass screen, were easy to see. to the ideal pattern to give Two

The photographed the reflection photographs error

were measured and compared over between to different

averaged

20 and 36 spots for each mirror. of the mirror

corresponding

masked positions reflection

were taken. that

The requirement

that the average angular

error be 51 mrad implied

the average deviation the negative. 0.2 mrad level.

-

of a light spot from its corresponding the film was developed

grid point be 51 mm on at the

After

it was easy to measure deviations

For an ideal mirror the center of curvature create an undistorted imaged pattern centerline computer

illuminated of the mirrors,

by a point

light

source placed

very close to

the square grid pattern

of the mask would of the

square grid image pattern.

To make the measurement

possible, the point light source was situated and the frosted

off-axis below the mirror A

by a few centimeters, simulation with

glass was above the centerline. the actual positioning

ray tracing positioning

was run to duplicate

of elements. the expected vertical

The off-axis

of the light source and the frosted rectangular,

glass cause

square grid to become slightly The ideal image pattern

with the longer side in the rectangle,

direction.

was taken to be this computed in the mirror.

and any other effects found were due to imperfections

The acceptance f negative

criterion

of a maximum

average reflection substrates.

error of 1 mm on each The mirror substrates cost.

was satisfied

by 96.5% of all the mirror criterion

that failed the acceptance

were replaced again at no additional

3. COATING 3.1 Coating

-

AND

REFLECTIVITY

Before

coating,

the mirror

substrates

were cleaned

by conventional

methods.

After extensive testing of coatings by many vendors, Acton Research Corporation slo standard coating ARC1600 was chosen. purity The mirror aluminum substrates were coated with

approximately (MgF2) -by

80 nm of ultra-high

and 40 nm of magnesium 10s7 T). The mirrors

fluoride were held

overcoat under high vacuum

(approximately

the spool pieces so no shaded areas were left on the mirrors. in a few seconds and followed surface. Figure immediately

The aluminum

was deposited oxidation to insure (280%

by the MgF2 to minimize rotated

of the reflective an even coating.

The depositions

took place as the mirrors reflectance coating, humidity

5 shows the minimum After

specifications were

at X = 160 nm and 285% for X 2 180 nm).

the mirrors at 20C.

stored in a dry-air

clean room with less than 15% relative

3.2 Reflectivity
Three spherical

Testing
mirror substrates and one planar witness coupon Research Corporation. using the Acton ranging (5.1 x 5.1 cm) of the

were coated in each batch by Acton witness Vacuum coupon was measured

The reflectivity Model

by the vendor

VRTMS-502

Measurement

System at 10 wavelengths coupon

between

160 and 230 nm. specifications, as

If the reflectivity

of the witness

was above the minimum spherical mirrors

shown in fig. 5, the reflectivity acceptable and the mirrors

of the full-size

was assumed to be

were shipped.

Witness

coupons were necessary because

the A&on mirrors. f -. After measured standard designed

-

VUV

spectrometer

was not large enough to contain

the full-size

spherical

delivery

to UCSB,

the reflectivity wavelengths

of each full-size from

spherical

mirror

was to a was mirror and to be

at eight different whose absolute and constructed It consisted rotators

ranging

160 to 230 nm relative A custom instrument

reflectivity

was well measured.

to measure the relative of a large vacuum which permitted small with mirror

reflectivity steering

of each full-size mirrors,

(see fig. 6). stepper compared vacuum motor

box with

stands, mirror

the reflectivity mounted

of the full-size

to the standard UV monochromator

in the box.

A McPhearson a narrow

218

a deuterium

lamp l1 directed

beam of tube that

light onto the test mirror, was coated with p-terphenyl

which reflected wave shifter.l* and rotated All

the light into a photomultiplier The full-size test mirror about its center point, the motors

could be pivoted allowing any point

about its center of curvature, on the mirror controlled surface

to be tested.

and settings

were computer mirror

via CAMAC.

At each wavelength,

three positions

of each full-size

were tested. curved mirror.

The light from the full-size (5.1 cm diameter) reflectivity

test mirror

was compared

to that from a of the test

standard

to determine

the relative

reflectivity

The absolute

of the curved standard

was tested at least monthly of the

using a second custom instrument full-size After spherical delivery mirrors

described later; thus the absolute reflectance rather accurately.

could be calculated

to UCSB, the reflectivity

of the witness coupons were measured at in the range 160 to 230 nm. This was to measure down absolute reflectivity vacuum

the same 10 wavelengths done using a custom of small mirror

as chosen by Acton which instrument

instrument, This

was built

samples.

was pumped

to a high

(approximately occur.

10 -6 T) so no VUV

absorption

from oxygen or water with

vapor could lamp

A second McPhearson

218 VUV
10

monochromator

a deuterium

dire&d ..f -. the light

a narrow

beam

of light mirror,

onto the test mirror. which reflected

The test mirror

reflected

onto a standard

the light

into a photomultiplier. when the the

The light from this measurement standard deuterium standard reflectance mirror was pivoted

was then compared

to the light obtained the narrow

into a position

to reflect

beam from

lamp directly mirror

into the same photomultiplier.

The same position

on the

was illuminated

by the narrow beam in both cases; thus the absolute directly.

of the mirror

sample was determined

3.3 Reflectivity
Figure coated .steadily

Results
measured at UCSB, of all 430 full-size The reflectivity of all the

5 shows the average reflectivities, at 8 wavelengths

mirrors

in the range of 160 to 230 nm.

increases from 84% at 160 nm to 89% at 220 nm. are presented at four wavelengths

The reflectivities showing

430 mirrors number

in fig. 7 as histograms

of mirrors

N having reflectivity

R in percent. as measured of the witness from 180 to

The average reflectivities

of the witness coupons for all 430 mirrors

at UCSB are shown in fig. 8. The agreement between the measurements

-

coupons by UCSB and Acton 230 nm. However, Acton

Corp. is good for the wavelengths

ranging

the UCSB averages are as much as 3.6% lower than those made by in the range of 160 to 175 nm (as shown in fig. 8). The the measuring measurements errors over all wavelengths. of the witness coupons at

Corp. for wavelengths

average results are in agreement well within The agreement between the reflectivity

UCSB and those of the associated full-size mirrors (see fig. 5). The average reflectance differ by a maximum of the full-size

at UCSB is good at all wavelengths mirrors and the witness coupons

of 1.5% (which occurs at 190 nm).

11

.-

4. ALIGNMENT .f
-.

4.1 Mounting
The mirror

Spool
alignment process involved the rotating and bending of the mounting and then fully

spools (see fig. 2). The spools were machined from 6061T-6 aluminum annealed.

The purpose of the spool design was to allow the bending of the stem to in any direction with high precision. It had to satisfy the double bumps) but easily position. the amount

orient the mirror constraint

of not bending too easily (to prevent droop or accidental without springing

enough to make small adjustments The effect of a variable of deflection . original position

back to the original

torque on the spools was tested, measuring

when the torque was applied and the ability when the torque was removed. the elastic limit

of the spool to return to its

It was found using a stem diameter Since the average mirror

of 7.62 mm that

was reached at 3.1 kg.

weighed less than 1 kg, this was deemed adequate to prevent droop yet also permit alignment changes of 1 or 2 mrad.

4.2 Mirror
-

Alignment
table was designed to simulate the position of the mirror array as if

An alignment

it were in the SLD CRID vessel. In the CRID each mirror a different position along a drift box. The alignment

type focuses light towards

table had two sides, the mirror

array side and the point light source side, as shown in fig. 9. An array consisted of two ladders attached 5 mirrors mounted to a permanently via mirror mounted mirror main channel. Each ladder had

arms onto a side mount channel. The arms were made orient the mirror. The source

of stainless steel tubing side of the alignment accurately

cut and welded to nominally

table had a point light source-screen assembly, that could be positioned by means of precision alignment holes in the in turn.

and reproducibly

table to bring the source to the center of curvature Each mirror was then illuminated

of each of the 10 mirrors

by the point light source, and the image projected

12

onto-a

frosted

glass screen.

The radius of curvature

was measured

by moving

the

screen position
t

along the axis of the mirror and bending

until the image reached its smallest point. the spool stem (see sec. 4.1 and fig. 2). The alignment was done to

Aligning -. Typical a typical restraint further

was done by rotating adjustments accuracy

needed were of the order of 10 mrad. of f2 mm (f2

mrad) from the center of the screen. A secondary at the time of alignment to provide

system was installed protection

on each mirror

against a mirror were aligned,

releasing from its spool (see sec. 2.2). the ladder with its side channel was detached box to

After

all 5 mirrors

from the mirror be sent to SLAC. with

main channel of the alignment Special shipping pertaining ability.

table and mounted

in a shipping

boxes were designed and tested in order to comply to transport-environment that driving resistance, cleanliness,

the requirements

--and shock-absorbing affect the alignment. packets filled with The desiccant nitrogen

Tests indicated

over rough roads did not several small

The steam-cleaned

aluminum

boxes contained

a desiccating

agent; the boxes were then filled with dry nitrogen. box was purged with dry

packets

were replaced- and the shipping

after each use.

4.3 Mirror After

Ladder
their arrival

Installation
at SLAC, the mirror Previously, ladders were placed in clean, dry (15% 40 main mirror channels identical to the

RH at 20C)

storage shelves.

one on the alignment argon calorimeter the main channels. on a large turret. instrument

table had been mounted Installation

onto the inside surface of the liquid consisted of placing the ladders onto mounted The

vacuum jacket.

This was done by means of a sophisticated The turret allowed access to all positions along three directions to accurately
13

instrument

around

the barrel.

itself permitted

translation

and rotations

about two

axes. This freedom of motion

was required

mate the ladder to the main

channel without _f

bending.

The ladder was held onto the instrument

by a compressed

air system which could be remotely step was to wedge a bar between

released after the ladder was in place. The final the ladder and the channel by means of a screw

which drove the bar and locked the assembly in place.

4.4 In-Place
Alignment f2 mrad.

Mirror

Measurement
to an accuracy barrel of approximately to be considerably of the main or side

was done on a bench as described the actual deviation

We expected contributing

in the CRID

larger-the

factors may include minor misalignments

channels which would be magnified

-

by a factor of 100 or so due to the long lever arm. where each mirror pointed onto the particular in mechanical

In addition, drift

we wanted

to determine

box to which it was focused. of the barrel structure

This would remove any deviations itself. supported

tolerances -

We have constructed as the particular simple in concept. drift

a device which is rigidly box associated with

at the same three points The device is very which is free to pivot The steering bounces off the position in which the at three wellof fl or

a ladder of mirrors. at a small mirror

A laser beam is directed

about two axes so as to direct the beam onto the mirror mirror is positioned mirror

being measured.

so that the reflected beam from the measured mirror and back into the laser opening. By determining the direction

the steering

and angles of the steering mirror, measured mirror is pointing.

we can easily compute each mirror

In practice,

was measured

spaced points 2 mrad.

to determine

the focal length

of the mirror

to an accuracy

The actual measurements aimed-for spot was f20

indicated

that the average deviation

from the original for us since we

mrad, approximately.
14

This posed no problem

now-have _T
-.

the actual direction

of pointing

to an order-of-mag;litude measurement.

better

accuracy,

but it stressed the need for making

an in-place

5. CONCLUDING
_ A mirror and a coating to produce quality slumping, grinding, and polishing

REMARKS
process at Lancaster Glass Company implemented counter. The

process at Acton

Research Corp. have been successfully mirrors for the SLD Barrel CRID

more than 400 spherical

control measurements and confirm

performed

at Acton Research Corp.

and at UCSB show

good consistency _ surface roughness . limited

that the average reflectivity

is at least 80% for light at range for these mirrors. imply that the scattering The is

160 nm and 85% for light in the 180-230 nm wavelength and optical distortion measurements

to a few percent

of the incident

light and the angular

reflection

error is less

than 1 mrad.

ACKNOWLEDGMENTS
-

We would like to thank 3-D Micro-Surface to produce invaluable

Applied

Magnetics

Co. for the free use of their WYCO who worked

Measurement

System, all the UCSB undergraduates and Tim Montagne

and test the mirrors, assistance.

and his crew at SLAC for their

15

.-

REFERENCES
(1984); UC-340 (1987); IEEE (May 1984).

-. f
-. 1. SLD Design Report, SLAC-REPORT-273 2. G. Hallewell et al., SLAC-PUB-4405

Trans. Nucl. Sci. NS-35,

398 (1988).
3. R. J. Apsimon, 4. P. Baillon et al., Nucl. Instrum.

Methods A241,

339 (1985).

et al., Nzlcl. Instrum.

Methods A277,
Bristol, U.K.

338 (1989).

5. I.S.C. Chemicals

Ltd., Avoumouth,

6. F. Bird, SLAC-REPORT-332 7. Mirror 8. Applied 9. WYCO 10. Acton substrates Magnetics manufactured

(1988). by Lancaster Glass Co., Lancaster, OH 43130.

Co., Goleta, CA 93117.

Corp., 1955 East 6th St., Tucson, AZ, 85719. Research Corp., Acton, MA 01720. Originally both the deuterium lamp

11. Hamamatsu
-

L879-01, with MgF2 window. were brought

and the monochromator lamp lifetime resulted.

under vacuum but a short deuterium with a

The lamp was then closed in a container

continuous

nitrogen flow; thus the lifetime of the deuterium

lamp was extended.

12. There was a gradual loss of the p-terphenyl window frequently since it was within

wave shifter coating on the MgF2 coating was

the vacuum chamber. The p-terphenyl light output.

replaced to ensure maximum

16

.-

FIGURE

CAPTIONS
the e* beams of one quadrant of the

1. (a) Vertical

section in the plane including

SLD detector.

(b) S ch ematic of CRID barrel longitudinal See (a) for scale.

section with Cerenkov

photons pictured. 2. Aluminum

mounting

spool epoxied onto the back of the mirror

substrate.

3. Surface roughness profile of a typical mirror substrate,

showing a 2 x 2 mm area of 1.020 m.

with a surface roughness of 1.91 nm (rms) and a curvature 4. Bench setup used to measure the optical 5. The average VUV reflectance results distortion

from each mirror. for the witness

as measured by UCSB

coupons and full-size mirrors, 6. The custom instrument mirrors: 7. %sults

and the minimum

acceptable reflectance specifications. reflectivity of device. of all full-size

used to measure the relative and (b) photograph

(a) light path schematic, of the reflectivity

measurement

of the full-size mirrors for all 430 mirrors (a) 160 nm, (b) 180 nm, (c) 200 nm, and

at UCSB for the light at wavelength

-

(d) 220 nm. 8. Results of the reflectivity at Acton measurement of the witness coupons for all 430 mirrors 160 thru 230 nm.

Corp. and UCSB for the light at wavelengths

9. Table used to align mirrors

on a ladder, as discussed in text.

17

..f -.

_-

Warm

Iron Calorimeter

(mm)

Fig. la

_-

Gas Radiator

e+

e-

11-89 6507Al

Fig. lb

_-

SPOOL PIECE Top View

5.08 cm Dia. c- 3.18 cm Dia.I

n \ --w
0
-

I n t\

0 0

> 0
:I>:
-

Side View
Attaches to Arm 0.48 cm 7
\ /

1 t

Stem -+ 0.76 cm 2-90 6576A4

1 ---I.-

L Attaches to I Mirror

Fig. 2

.f -.

_-

13Q455316 2I p; : RR: 3 - !J: 1 . 9 2 r-1m 1 . 5 1 f-i m 1 9 . 7 f-1rn

IS:32

E33/~16r89

-Pef

5.

m,<

SURFRCE

: 650. lnm N V L E 111 .I : 1 . El 2 5 m R c r 60m R 2:~ 1 : -76.

44.40 Orrentot1or-0

cl
0 IFron: 9.8 > 4.9 -

-0.1

-4

-5-u

SIJ 3 a , I nr -9.9nm

c
0

1
496 D1rtanc2 992 Lrllcronr 14aa 1

p1912 1984

-9 u: L:

-.

WYKO

7.6

-7

.o

-14.4 -196 Ul stance 992 i4aa rrllcronr~ Isa.4

U: lS.I3rtm L:-14.4nm

WYKO
5-90 6576A5

Fig. 3

_-

FROSTED GLASS SCREEN ILLUMINATED / GRID \ CAMERA

MASK

1 1

I
5OURCE

I
-

A
6576~6

RADIUS OF CURVATURE

1110 R

5-90

Fig. 4

.-

- Min. Acceptable Reflectance l W itness Coupon Reflectances o Full-Size Reflectances -

0.8
I I I I

160
2-90

180 200 Wavelength (nm)

220
6576A3

Fig. 5

.-

.-----.-a

\I

MgF2 WINDOW
II

! I/

APERATURE LENS

-a
l--e--------

i \I
iv

&!

FLAT VUV REFLECTIVE MIRROR TM= FULL-SIZE TEST MIRROR or SM=SMALL STANDARD MIRROR

xxxx P-TERPHENYL COATING

---

1 NITROGEN FLOW REGION

5-90

I :________ i TURBO PUMP REGION (n 10M6torr) ---I ,I -1 LARGE MECHANICAL PUMP REGION (- 10e3 torr) 6576A7

Fig. 6a

I
.-

-1

.-CT, LL

-.

.-

Reflectance at 160 nm
I I I I 1 I I 1

Reflectance at 180 nm 80 I 60

60 2 40 2 .k 2 20
n

(a) -

(W
I

0.75

0.85 0.95 Reflectance Reflectance at 200 nm I I I I

01. 0.75

0.85 0.95 Reflectance Reflectance at 220 nm

I I

I 60 i 40 .2 20

(cl
I 2

80 60

00 0.75
2-90

0.85 0.95 Reflectance

0 0.75
Fig. 7

0.85 0.95 Reflectance

6576Al

.-

-1

0 UCSB Measurements l Acton Measurements

-

5:
-

I
0.8 160 180 200

PP
220

2-90

Wavelength
Fig. 8

(nm)

6576A2

.-

f -.