Page 32 of 34

8.0 OTHER REQUIREMENTS, CONVENTIONS AND GUIDELINES (cont’d)

8.5 Electronic File Names and Filing (cont’d)

Use font size 8 for the file name. It should not be noticeable since it is not part of the text, but it
should still be legible on the printed document.

For maximum effectiveness, reliability and security, documents and other files
e.g. spreadsheets, must be prepared, developed, filed and stored in the common electronic work
folders for your group, which are on the LAN. There are two important reasons for this
requirement. Firstly, the current files must always be accessible to the whole team. Secondly,
the LAN is automatically backed up, so in theory, it ought to be more reliable than other
devices. Nevertheless, all files must also be backed up.

Documents and other files must be backed up onto your hard drive or onto CD or memory stick
(jump drive). This is especially important when the document is being developed, to ensure
that work is not lost.

As a minimum, a hard copy must be filed in the Outgoing Correspondence Folders for your
group, and in the work file for the job.
ABSTRACT

Precipitation of second phase particles is a key factor dominating the mechanical properties
of high temperature alloys. In order to control and optimize the precipitation effect it is of
great importance to study the role of alloying elements in the formation and stability of
precipitates. As a favored family of corrosion and creep resistant austenitic stainless steels the
20Cr-25Ni alloy was modified by addition of copper, molybdenum, nitrogen, niobium and
vanadium. A set of alloys with similar matrix but varying contents of niobium, vanadium and
nitrogen were prepared. Sample preparation process included melting, hot forging, solution
annealing and finally aging for 500 h at 700, 800 and 850 ºC.

Light optical and scanning electron microscopy revealed micron-scale precipitates on grain
and twin boundaries as well as sub-micron intragranular precipitates in all samples.
Characterization of precipitates was carried out by means of energy dispersive X-ray
spectroscopy (EDS). Among micron-scale precipitates M23C6 carbide was the dominant
phase at 700 ºC aging temperature; whereas silicon-rich eta phase (M5SiC) was the main
precipitate in samples aged at 800 and 850 ºC. A few sigma phase particles were found in one
of the niobium containing samples aged at 700 and 800 ºC. Sub-micron intragranular
precipitates were analyzed using low voltage EDS. The spatial resolution of EDS
microanalysis at 5 kV accelerating voltage was estimated as almost 100 nm which was at
least eight times better than that using the ordinary 20 kV voltage. Also, low voltage EDS
revealed the presence of light elements (carbon, nitrogen and boron) in the composition of
sub-micron particles thanks to the less matrix effect in absorption of low energy X-rays of
light elements. In samples aged at 700 ºC niobium-rich and vanadium-rich carbonitrides were
found as the dominant precipitates; whereas they contained much less carbon in samples aged
at 800 ºC and mostly became carbon-free nitrides with well-defined cuboidal shapes at 850
ºC aging temperature. This showed that niobium/vanadium-rich nitride phases are stable
precipitates at aging temperatures above 700 ºC.

The drawbacks of low voltage EDS were indicated as high detection limit (no detection of
low- content elements), poor accuracy of quantitative analysis and high sensitivity to surface
contamination. Some possible ways to improve the accuracy of low voltage EDS, e.g. longer
acquisition time were examined and some other suggestions are proposed for future works.

Keywords: austenitic stainles steel, precipitation, EDS, low voltage EDS, micron-scale
precipitates, sub-micron intragranular precipitates

iii
 ROYAL INSTITUTE
OF TECHNOLOGY

Precipitation Study in a High Temperature Austenitic Stainless Steel

using Low Voltage Energy Dispersive X-ray Spectroscopy

By:

Ali Gharehbaghi

Supervisor:

Fredrik Olsson

Examiner:

Malin Selleby

Master Degree Project

School of Industrial Engineering and Management

Department of Materials Science and Engineering
Royal Institute of Technology (KTH)
SE-100 44 Stockholm, Sweden

March 2012
Contents

INTRODUCTION ..................................................................................................................... 1

1.1 Demand for advanced materials in energy production, an environmental concern .... 1

1.2 Development of austenitic stainless steel for high temperature applications.............. 2

BASICS AND BACKGROUND ............................................................................................... 5

2.1 Stainless steel .............................................................................................................. 5

2.2 Austenitic stainless steels ............................................................................................ 6

2.3 Effect of alloying elements in austenitic stainless steels............................................. 8

2.4 Precipitation in austenitic stainless steels.................................................................. 10

2.5 Energy dispersive X-ray spectroscopy (EDS) microanalysis ................................... 16

2.5.1 Beam electron-specimen interaction .................................................................. 16

2.5.2 Origin of characteristic X-rays........................................................................... 18

2.5.3 Continuum X-ray ............................................................................................... 20

2.5.4 EDS X-ray spectrum .......................................................................................... 21

2.5.5 Accuracy of EDS microanalysis ........................................................................ 23

2.6 Low voltage EDS microanalysis ............................................................................... 24

2.6.1 Drawbacks of low voltage EDS ......................................................................... 26

EXPERIMENTAL WORK ...................................................................................................... 29

3.1 Alloy selection........................................................................................................... 29

3.2 Base melt preparation ................................................................................................ 29

3.3 Making mini-melts .................................................................................................... 30

3.4 Forging and solution annealing ................................................................................. 32

3.5 Aging ......................................................................................................................... 33

3.6 Metallography and microstructural examination ...................................................... 33

3.7 Characterization of precipitates using EDS microanalysis ....................................... 33

RESULTS AND DISCUSSION .............................................................................................. 35

v
 4.1 Sample preparation process ....................................................................................... 35

4.1.1 High oxygen content .......................................................................................... 35

4.1.2 Solution annealing ............................................................................................. 36

4.1.3 Catastrophic oxidation ....................................................................................... 38

4.2 Optical microscopy ................................................................................................... 39

4.3 Scanning electron microscopy .................................................................................. 40

4.4 Identification of precipitates using EDS microanalysis ............................................ 44

4.4.1 M23C6 ................................................................................................................. 44

4.4.2 Silicon-rich eta (η) phase ................................................................................... 46

4.4.3 G-phase .............................................................................................................. 48

4.4.4 High-carbon carbide........................................................................................... 50

4.4.5 Sigma phase ....................................................................................................... 51

4.4.6 Intragranular sub-micron precipitates ................................................................ 51

4.5 Low voltage EDS for identification of sub-micron precipitates ............................... 52

4.5.1 Residual phase sub-micron particles .................................................................. 54

4.5.2 Niobium-rich carbonitride and nitride ............................................................... 55

4.5.3 Vanadium-rich carbonitride and nitride ............................................................. 57

4.5.4 Copper-rich precipitates ..................................................................................... 60

4.6 Limitations of low voltage EDS microanalysis......................................................... 61

4.7 Accuracy of low voltage EDS ................................................................................... 64

4.8 Determining the spatial resolution of low voltage EDS ............................................ 66

SUMMARY AND CONCLUSION ........................................................................................ 69

FUTURE STUDIES................................................................................................................. 72

REFERENCES ........................................................................................................................ 74

vi
 Chapter 1

1 INTRODUCTION

1.1 Demand for advanced materials in energy production, an environmental concern

Climate researchers consider global warming as the major threat to the world’s environment.
The increased concentration of greenhouse gases in the earth’s atmosphere is the main factor
accelerating this phenomenon. For instance, the concentration of carbon dioxide (CO2), as the
major greenhouse gas in the atmosphere, has increased by nearly 35% since mid-1800 with
[1]
the highest growth during the last decades . Among different sources of carbon dioxide,
electricity and heat generation sector has the largest proportion with over 40% of the global
[2]
CO2 emission . Coal fired power plants with the largest share in this sector are responsible
[3]
for 28% of global CO2 emission . The worldwide dominance of coal as the major fuel in
electricity generation is expected to continue in long term future thanks to its relatively low
price, availability and stable market around the world [2, 4].

The worldwide environmental concerns since 1990’s besides the economic reasons have
provided an extra incentive to increase the efficiency of coal fired power plants. According to
studies each percentage point of efficiency increase in a coal fired power plant can result in
[3, 5]
2.5% reduction in CO2 emission . Therefore, an international effort including several
R&D projects have been conducted to improve the efficiency of coal-fired power plants.

Coal is burnt to boil water; then the generated steam is heated up and pressurized before
passing to turn the steam turbine. The efficiency of a coal fired power plant depends strongly
[1, 4-6]
on the steam temperature and pressure . The conventional power plants operated with
the steam temperature and pressure up to 455 ºC and 190 bar, respectively (subcritical
[5]
condition) which resulted in efficiencies under 40% . This operating condition has already
been raised up to 45% by using the steam in supercritical condition at 540 ºC and 250 bar [5].
Improvements are in progress to reach an efficiency of 46-49% under the ultra-supercritical
steam condition at 600 ºC and 300 bar [5]. Eventually, the next generation of coal-fired power
plants are planned to operate with steam at 700 ºC and 350 bar or higher to get the net
efficiencies above 50% and emit almost 35% less carbon dioxide than conventional coal-fired
power plants [5].

1
In order to build up such high-efficiency power plant, parallel to optimizing the
thermodynamic cycles, it is required to develop new materials for the critical sections dealing
with high temperature, high pressure and corrosive atmosphere. In this context, the boiler is a
key section where the hot pressurized steam is generated. As it is defined in AD700, the
European project for the next generation of coal-fired power plants, service condition for the
most critical stage of steam-cycle in super-heater/re-heater section of boiler (SH/RH tubes)
includes working temperature (metal) of 700 ºC and the steam pressure of 350 bar (35 MPa)
[5]
in addition to the fireside corrosion in sulfuric atmosphere as well as the steam side
[7]
oxidation and spallation . A suitable material to withstand this condition is supposed to
have minimum creep rupture strength of 100 MPa in 100,000 h at 700 ºC and less than 2 mm
[5]
thickness loss in 200,000 h exposure time . Nickel base alloys like Inconel 617 and 740
fulfill these requirements but they are comparatively expensive. As a more economic
candidate, the high temperature grades of austenitic stainless steel are under development.

1.2 Development of austenitic stainless steel for high temperature applications

Stainless steel is a major alloy in energy generation industry where a combination of

corrosion resistance and creep strength is required in a wide range of elevated temperatures.
Most boiler components are made of the high temperature ferritic and austenitic stainless
steels. In thick sections, like headers, which are subjected to the thermal fatigue the ferritic
[6, 7]
grades are preferable due to their lower thermal expansion coefficient . But where high
creep strength and corrosion resistance at temperatures (metal) above 600 ºC is required (e.g.
in the finishing stages of SH/RH tubing) austenitic grades are dominant [6, 7].

In the recent evolution of the steam condition, various grades of austenitic stainless steels
have been developed. Table 1.1 gives the chemical composition of recently developed grades
as boiler materials in the temperature range 650–700 ºC [6, 7].

Table 1.1. Chemical composition (wt%) of the recently developed austenitic stainless steels as boiler
materials in the temperature range 650–700 ºC

Material C Cr Ni Mn Si Mo Nb Cu N Others
Super 304H 0.1 18.0 9.0 0.8 0.2 ... 0.4 3.0 0.1 ...
NF709 0.15 20.0 25.0 1.0 0.5 1.5 0.20 ... ... Ti 0.10
Save 25 0.1 23.0 18.5 1.0 0.1 ... 0.45 3.0 0.2 W 1.5
HR3C 0.06 25.0 20.0 1.2 0.4 ... 0.45 … 0.2 …
Alloy 800H 0.08 21.0 32.0 1.2 0.5 … … … … Ti 0.5, Al 0.4
Sanicro25 0.08 22.0 25.0 1.0 0.1 ... 0.50 3.0 0.2 W 3.0, Co 1.5

2
Chromium content is the main factor determining the corrosion resistance and usually for
extremely harsh corrosive conditions beyond 650 ºC (e.g. SH/RH tubes) a chromium content
[6]
in excess of 20% is required . Nickel, as the main austenite stabilizer element, has also a
[6, 9]
beneficial effect on high temperature corrosion resistance . Precipitation of the second
phase particles dominates creep properties. The effect of alloying elements has been
[8]
extensively studied as the major factor in formation and stability of precipitates . Some
elements such as niobium and titanium have been well understood to stabilize the steel
(corrosion point of view) and to form effective precipitates (creep point of view) [6]. Nitrogen
is frequently used thanks to its beneficial effects like solution hardening, nitride and
carbonitride precipitation and corrosion resistance improvement [6]. Copper has been recently
recognized as a former of ultra-fine precipitates during long-term aging which is highly
[6, 8]
beneficial to the creep resistance . Some elements form intermetallic precipitates (e.g.
titanium and aluminum) and some others have solid solution strengthening effect (e.g.
tungsten and cobalt) [6, 8]. In spite of the above understood properties extensive researches are
still ongoing to optimize these effects as well as developing new alloying systems.

The present work is a precipitation study on a 20Cr-25Ni austenitic stainless steel which has
been a favored family of the creep and corrosion resistant alloys. It was carried out as a part
of an ongoing research in the high temperature group of Outokumpu Avesta Research Center.
The main goal was to study the role of vanadium, niobium, nitrogen and copper on the
precipitation in low-carbon nitrogen-containing 20Cr-25Ni alloy at elevated temperatures.
For this purpose a base melt with 20Cr-25Ni matrix and from that a number of sample melts
(mini-melts) with varying amounts of vanadium, niobium and nitrogen but almost similar in
the contents of copper, molybdenum, manganese and silicon were prepared. Samples were
hot-forged, solution annealed and then aged for 500 h at three temperatures 700, 800 and 850
ºC. The microstructures of aged samples were observed using light optical and scanning
electron microscopy (SEM). Energy dispersive X-ray spectroscopy (EDS) was used for
compositional analysis to characterize the precipitates. For sub-micron precipitates which
were not possible to be identified by ordinary EDS analysis the use of low voltage EDS in the
available field emission gun SEM was proposed and examined. By that, the spatial resolution
of analysis improved from around 1 μm to almost 100 nm. The advantages and limitations of
low voltage EDS were also studied and to some extent practically examined.

3
4
 Chapter 2

2 BASICS AND BACKGROUND

2.1 Stainless steel

Stainless steel is a generic term for alloy steels containing at least 11 wt% of chromium [9]. At
this level chromium can form a coherent, adherent and regenerating layer of chromium oxide
(Cr2O3) on the steel surface when exposed to oxygen. The Cr2O3 film is impervious to air and
[10, 11]
water, protecting the steel from corrosion . Improved corrosion resistance and also a
wide range of microstructures and properties may be obtained by further alloying addition
[15]
. Stainless steels can be divided into five groups where each group involves various grades
with defined standard composition ranges. The first four groups are ferritic, austenitic, duplex
(ferritic-austenitic) and martensitic stainless steels characterizing with their crystallographic
microstructure. The fifth group, precipitation hardening (PH), are characterized with their
capability to be heat treated [12]. Table 2.1 gives a summary of the nominal composition range
of five groups with their characteristics [9].

Table 2.1. Composition range and characteristic of five different stainless steel groups

Steel Nominal composition range (wt%)

Characteristics
category C Cr Ni Mo Other

Martensitic 0.03-2.2 10.5-18 0-7 0-4 Mn, Si, V, N, Cu high strength and hardness

Ferritic <0.25 10.5-30 0-4.5 0-4.5 Mn, Si, N, Ti/ lowest-cost corrosion
Nb, Al resistance alloy

Duplex <0.05 18-30 1-8 1-5 Mn, Si, N, Cu, W economic high corrosion
resistance, good strength

Austenitic <0.08 16-30 8-35 0-7 N, Nb, Ti, Cu high corrosion resistance,
formable, weldable
Precipitation <0.2 10-17 4.3-26 0-5 Mn, Si, Al, Ti, high strength and
Hardening Nb, V, Cu, W toughness

The most common crystallographic forms of stainless steels are body-centred cubic (BCC,
ferrite) and face-centered cubic (FCC, austenite) which determines the steel properties
(essentially mechanical properties). The effect of alloying elements is usually quantified as
whether they promote formation of BCC structure (ferrite stabilizing) or FCC (austenite
stabilizing). Cr and Ni, as the most important alloying elements in stainless steel, are
powerful ferrite and austenite stabilizers respectively. So they are considered as reference to
categorize the effect of other alloying elements; those promoting ferrite are classified under

5
Cr as ferrite stabilizers and the austenite stabilizer elements are classified under Ni. Different
formulas such as equations 2.1 and 2.2 has been suggested to quantify the potency of alloying
elements compared to those of Cr and Ni to form ferrite and austenite, respectively [11, 13, 14]:

Cr equivalent = (Cr)+2(Si)+1.5(Mo)+5(V)+5.5(Al)+1.75(Nb)+1.5(Ti)+0.75(W) (Eq.2.1)

Ni equivalent = (Ni)+(Co)+0.5(Mn)+0.3(Cu)+25(N)+30(C) (Eq.2.2)

The element symbols in the equation denote their content in wt%. However, the use of the
above equations is not straightforward as they are suggested for austenitic alloys [13]. With the
aid of the Schaeffler-Delong diagram (Figure 2.1) it is possible to predict the present phases
in as-solidified condition of stainless steels. Chromium equivalent (Creq) and nickel
equivalent (Nieq), which are formulated more generally than in equations 2.1 and 2.2,
represent the combined effect of the important ferrite and austenite stabilizer elements,
respectively. The main use of this diagram is to predict the weld metal when designing the
welding process for stainless steels; however, it may also be used for cast products [9].
Nieq = Ni + 30 × C + 0.5 × Mn

Creq = Cr + Mo + 1.5 × Si + 0.5 × Nb

Figure 2.1. Schaeffler-Delong diagram to predict the stainless steel weld metal [9].

2.2 Austenitic stainless steels

Austenitic stainless steels contain 16 to 28 wt% chromium and sufficient amount of austenite
stabilizer elements like nickel, manganese and carbon to have austenitic microstructure at
room temperature. Nickel is the basic substitutional element used to stabilize austenite, so the

6
austenitic stainless steels are usually considered as Fe-Cr-Ni ternary alloys. Figure 2.2 shows
an isothermal section of the ternary Fe-Cr-Ni equilibrium diagram at 750 ºC [8]. The positions
of two well-known austenitic grades, 18Cr-12Ni and 20Cr-25Ni are marked in the diagram.
Other alloying elements either interstitial like C and N, or substitutional such as Mn, Mo, Si,
Ti, Nb, W, V, Cu, Al, etc. may also be added to give the desired properties. The effect of
important alloying elements is discussed later.

Figure 2.2. Calculated isothermal section of Fe-Cr-Ni diagram at 750 ºC: typical 18-12 and 20-25
stainless alloys lie in the austenite region [8]

Austenitic stainless steels are the most common used stainless alloys due to their superior
corrosion resistance, excellent formability and weldability. The FCC austenite provides
characteristic properties like high ductility and toughness, no impact transition at cryogenic
temperatures and low stacking fault energy. The latter leads to high work-hardening rates; so
that the strengthening method for austenitic stainless steels is work-hardening rather than heat
treatment. This can result in a broad range of strength from about 200 MPa in annealed state
to even 2000 MPa in work-hardened state. Furthermore, austenitic stainless steels owing to
the FCC structure are in many cases more proper to be used at elevated temperatures
(especially above 550 ºC) since they do not lose their strength as quickly as other iron-based
structures. Therefore, high temperature austenitic grades are ideal candidates where advanced
creep properties are required at temperatures above 550 ºC [9].

7
2.3 Effect of alloying elements in austenitic stainless steels

Chromium

Chromium is the most important alloying element in stainless steels because the stainless
character essentially occurs when chromium is over a minimum level (~11 wt%). The higher
the chromium content is, the denser and stronger is the oxide passive film. However, too
much chromium (e.g. over 25%) may promote formation of intermetallic sigma phase and
cause embrittlement [8]. Chromium is a strong ferrite stabilizer, so a combination of austenite
stabilizer elements (Ni, Mn, N, C, etc.) is always required to maintain the austenitic
microstructure. This competitive effect is expressed by some relations, like equation 2.3,
which shows the condition to have an austenitic microstructure when the effect of Cr, Ni, N
and C are taken into account [15]:

30(%C) + 26(%N) + (%Ni) + 11.1 ≥ 1.3 (%Cr) (Eq.2.3)

In creep resistant alloys chromium takes part in the precipitation of different second phases as
both main constituent element (e.g. in M23C6 and sigma phase) and substituent one (e.g. in G
phase). It has high affinity for carbon to form chromium rich M23C6 carbide which is
undesirable in high amounts on grain boundaries; because the local concentration of Cr
around carbides decreases below the necessary level for corrosion resistance (sensitization).

Nickel

Nickel is an austenite stabilizer and increases the toughness, ductility, weldability and general
[15]
corrosion resistance of stainless steel, especially against pitting corrosion . High nickel
content in the matrix can make the thermal expansion coefficient of the alloy closer to that of
the oxide layer reducing the risk of oxide rupture and giving an increase in corrosion
[15]
resistance at high temperatures . Nickel usually dissolves in carbides as partially
substituted for chromium; whereas it is a major constituent in some other types of
precipitates, like G phase.

Carbon

Carbon atoms occupy the interstitial sites within the FCC structure of austenitic stainless steel
providing a solid solution strengthening effect. On a percentage basis carbon is the strongest
austenite stabilizer element (factor of 30 in eq. 2.2) but its use is limited due to low solubility
in austenite and also sensitization. As mentioned earlier M23C6 chromium rich carbides form

8
on grain boundaries leading to chromium depletion and consequently sensitivity to the
intergranular corrosion and/or stress corrosion cracking [9, 12].

Nitrogen

Nitrogen is an other interstitial solute and a strong austenite stabilizer. The solid solution
strengthening effect of N is more pronounced than C and it also enhances the strain hardening
[9, 15]
effect . Nitrogen has a limited solubility that can be increased by addition of manganese
[8]
to obtain superior pitting corrosion resistance . Its addition is also beneficial for creep
strength by promoting the precipitation of nitrides and carbonitrides.

Manganese

Manganese is an austenite former and sometimes used as a replacement for nickel; however,
the alloy exhibits less corrosion resistance [13]. It increases the nitrogen solubility in austenite
without forming nitrides and also contributes with nitrogen strengthening effect. Manganese
addition has some disadvantages like its weakening effect on the passive film and also
increasing the risk of pitting corrosion at manganese sulphide inclusion sites [16].

Silicon

Silicon, which is a ferrite stabilizer, improves the oxidation resistance of the stainless steel by
forming a dense silicon oxide sub-layer on the surface. It is an effective element where
[9]
sulfuric and nitric acids are present in the environment . Also, silicon prevents carburizing
of austenitic grades at high temperatures [17].

Stabilizing elements

Where it is not desirable and/or possible to extremely lower the carbon content (e.g. <0.02
wt%) strong carbide former elements like titanium and niobium are added to the stainless
steel to prevent chromium depletion. They have higher affinity for carbon than chromium,
diminishing M23C6 precipitation along grain boundaries and prevent localized chromium
depletion, so-called stabilizing process [9]. These elements are ferrite stabilizers.

Molybdenum

Molybdenum is a strong ferrite former and necessitates addition of more austenite stabilizers
to keep the austenitic structure. It enhances the general corrosion resistance of stainless steel,
especially against pitting corrosion. Molybdenum improves creep strength by solid solution
[8]
hardening and facilitating carbide precipitation . However, it promotes sigma phase and
Laves phase formation during the long-term aging. Also, molybdenum oxide, MoO3, has low

9
melting point (795 ºC) so that its molten oxide and/or its eutectic with some other low
melting point oxides (like Na2O) can cause catastrophic oxidation at high temperatures. Thus
in many cases the operating temperature of molybdenum containing alloys is limited to 700-
750 ºC [18].

Vanadium

Vanadium is used as a precipitate former in creep resistant alloys, usually nitrogen-containing

low carbon grades with high resistance to intergranular corrosion. Vanadium forms nitrides
[19]
such as VN and CrVN, which have been understood as effective and stable precipitates .
Like molybdenum, vanadium has also low melting point oxide, V2O5 (690 ºC). To avoid
catastrophic oxidation in vanadium containing alloys (so-called vanadium attack), in most
cases the vanadium content must be kept lower than 0.5% [19, 20].

Copper

Copper enhances the corrosion resistance of stainless steel in certain acids and promotes
austenitic structure. In high temperature grades, copper has been recently known to form
[20-
nano-scale precipitates in long term aging at high temperatures improving creep strength
22]
but it needs more investigation. Some results show that copper-rich precipitates form
[13]
independently of the other phases ; also the optimum effect of copper in different creep
[23, 24]
resistant alloys of 18-10 family is attained at 2.5-3 wt% copper content with aging
temperatures below 850 ºC [23].

Rare earth elements (REM)

Rare earth metals (REM), like cerium (Ce), are added in small amounts to certain heat
resistant austenitic stainless steels in order to increase the resistance to oxidation and high
temperature corrosion. Cerium promotes formation of the protective chromium oxide layer
and increases its adhesion to the underlying substrate [20].

2.4 Precipitation in austenitic stainless steels

Precipitation is the most dominant factor on mechanical properties at high temperatures.

Formation, dissolution and coarsening of precipitates have to be understood to achieve good
[13]
creep properties .The extensive research is still ongoing for better control over the
precipitation process to get more reliable performance of the alloys over the long service
periods. The precipitation process in austenitic stainless steel, similar to the other alloying
systems, is based on the decreasing in solubility limit of alloying elements in austenite when

10
it is cooled down. Before using the alloy, solution annealing is done at proper temperature
and time to dissolve the alloying elements in austenite; usually following by rapid cooling to
form a supersaturated solid solution. When the alloy is aged in-service, supersaturated
elements diffuse and react to form precipitates. Type, morphology, size, distribution and
stability of precipitates are factors dominating the creep behavior of the alloy. In this section
the main precipitates forming in creep resistance austenitic stainless steels are reviewed.

Carbides, nitrides and carbonitrides

Carbon and nitrogen are the interstitial elements and present in almost all grades of austenitic
stainless steels. They have high affinity for carbide/nitride former elements such as Ti, Nb, V
and Cr. The reaction product may be either beneficial or detrimental to the mechanical
properties; however, most of them improve creep properties under controlled formation.

M23C6

The term M23C6 is a general notation for Cr23C6 chromium carbide in which other metallic
[9, 13]
elements like Fe, Mo, V, W and Ni often substitute partially for Cr . It is known as the
main precipitate in many grades of austenitic stainless steel. Carbon has low supersaturation
level in austenite and high affinity for chromium [9]. Even very little carbon content (e.g. 0.02
wt%) is supersaturated at low temperatures (Figure 2.3); but due to very low diffusion rate
practically no carbide forms below a certain temperature.
Temperature (ºC)
Temperature (ºF)

C (wt%)

Figure 2.3. Equilibrium diagram for carbon in a typical 18-10 austenitic stainless steel [9].

The precipitation of M23C6 is accelerated above 550 ºC and usually starts on grain boundaries
and incoherent twin boundaries where diffusion is much more rapid than in the bulk material
[9, 25]
. As aging advances, M23C6 can precipitate on other sites like coherent twin boundaries,

11
interior grains on dislocations and around other particles, like residual carbides/nitrides [13, 25].
Some alloying elements can hugely affect the precipitation of M23C6 carbide. Elements like
molybdenum and nickel enhance it by decreasing the solubility of carbon in austenite; unlike
nitrogen, which increases the solubility of carbon and by that diminishes M23C6 precipitation
[9]
. With respect to both corrosion and creep properties it is desired for M23C6 carbide to
precipitate as fine intragranular particles instead of chain-shape ones on grain boundaries.

MX

MX refers to precipitates forming by reaction of a strong carbide/nitride former such as Ti,

[8, 13]
Nb and V (as M) and either of carbon or nitrogen or both (X) . As mentioned earlier
these elements are usually used to stabilize the steel against intergranular corrosion sensitivity
by trapping the excess carbon in the form of MC precipitate instead of the detrimental
chromium rich M23C6. Stabilizing heat treatment has to be done after solution annealing of
the alloy prior to use it by heating in the temperature range of 840–900 ºC for several hours
[13]
. Addition of M group to austenitic stainless steel has also creep strengthening effect when
fine MX precipitates form, mostly intragranular, during in-service aging. For instance NbX
particles are known to have more strengthening effect than intragranular M23C6 particles;
however, NbX particles may grow faster under creep condition since stress promotes
[26]
diffusion of Nb atoms but not C atoms . MX has a low energy NaCl-like atomic structure
forming a cuboidal shape after sufficient aging. The stoichiometric composition of MX gives
it high stability with no compositional change in long-term aging; however, the
stoichiometric availability of M and X species is not possible in practice and thus, available
[13]
elements like chromium may substitute for M . Many factors like M/X ratio in the alloy,
solution annealing process and change in local equilibrium by nucleation of MX affect the
precipitation and growth of MX particles. For instance some studies show an optimum creep
[13]
life for 20-25 steel with 1:1 atomic ratio of niobium and carbon ; also furnace cooling
instead of common water quenching after solution annealing has been found to markedly
increase the number of fine niobium rich MX intragranular precipitates [27].

Z phase

Z phase is defined as a niobium rich phase forming in nitrogen containing niobium stabilized
austenitic stainless steels. It has a tetragonal crystal with suggested formula per unit cell as
Cr2Nb2N2 in which carbon may be partially dissolved and replaced with nitrogen to form
[13]
CrNb(CN) composition . It is still unclear whether Z phase forms directly or from MX

12
precipitates (such as Nb(C,N) in 20-25 alloy) but as a general agreement it is known as a
stable fine dispersed phase (more stable than MX precipitates) at high temperatures, usually
[13]
above 700 ºC . For instance the networks of fine CrNbN particles in a 22-16 alloy were
found to effectively block the dislocation movement after 18 kh aging at 750 ºC [24].

M 6C

M6C, called eta carbide in the literature, is another type of FCC carbide which usually forms
after long-term aging. M can be varying combination of elements like chromium, iron,
[13]
molybdenum, nickel, niobium and silicon . The niobium rich M6C has been reported to
form in 20-25 alloy after long term aging (e.g. 50 kh) above 600 ºC [13]. This was confirmed
[28]
by Sourmail et al when the silicon containing eta phase formulated as Cr3Ni2SiX was
identified in NF709 steel of 20-25 family after 2500 hr aging at 750 and 800 ºC, where X
refers to C or/and N. This phase forms on the similar sites as M23C6 with also similar
morphologies so that, a precise compositional and diffraction analysis is required to
distinguish Cr3Ni2SiX from M23C6 [28]. Also, eta carbide has been often confused in the early
studies with G phase due to their similar structure and presence of silicon; especially when
considering the varying content of silicon in eta phase (2-10 wt%) which makes uncertainty
[29]
in its identification . For nitrogen-containing 20-25 steel with molybdenum in excess of
[29, 30]
4.5 wt%, Jargelius-Pettesson has shown both thermodynamically and experimentally
that nitrogen can stabilize the silicon-rich eta phase, but with nitride composition (M5SiN).

Intermetallic precipitates

G phase

G phase is a silicon-containing intermetallic precipitate with FCC structure forming in

austenitic stainless steels stabilized by niobium or titanium. The general formula A16D6Si7
has been suggested for G phase where A is mostly nickel and D is usually niobium (or
titanium). Also other elements like iron and chromium can substitute for A and D to limited
extent [13]. In niobium stabilized 20-25 alloy with sufficient silicon content (e.g. 0.5 wt%), G
[31]
phase can form in temperature range 500–850 ºC even prior to M23C6 . The preferential
site for its first nucleation is the residual NbC particles on grain boundaries as a source of
[31]
niobium which has low diffusion rate . The volume fraction of G phase depends strongly
on the silicon content of the alloy. As mentioned earlier, G phase is always confused with
silicon-rich eta phase; but generally it is agreed that increasing of nickel content promote G
phase as more stable precipitate than silicon-rich eta phase especially at low carbon level

13
 [13] [29, 30]
(e.g. 0.02 wt%) . As a possible way to distinguish these two, Jargelius-Pettesson
suggests that eta phase exhibits silicon content in the range 4.5-21 at%, whereas for G phase,
with respect to the general formula A16D6Si7, this amount is 24 at%. Also the absence of
interstitial elements such as nitrogen and carbon is considered as evidence for the
intermetallic G phase [13].

Sigma phase

Sigma phase is a well-known intermetallic phase with tetragonal atomic structure forming in
Fe-Cr alloying systems. In high alloyed austenitic stainless steels other elements such as Mo,
Ni, Mn and Si may dissolve in sigma phase, replacing for iron and chromium atoms. The
precipitation of sigma phase causes embrittlement in the alloy especially when it forms
continuously on grain boundaries; however, its dispersed intragranular form improves
slightly the creep properties. The precipitation of sigma phase is usually a very slow process
occurring in long term aging. Some studies conclude that M23C6 and δ-ferrite can be the
[13]
nucleation sites for sigma phase . Also, triple grain boundary junction is known as a
preferential nucleation site for sigma phase. Alloying elements like Mo, Nb, Ti, Cr and
especially Si are known to promote the precipitation of sigma phase [13].

Laves phase

Laves phase is an intermetallic component of iron and other metallic elements like Mo, Nb
and Ti. It has a hexagonal unit cell which necessitates close atomic size of its constituent
elements. Laves phase is generally known as more stable precipitates in shape and size than
[32]
carbides/nitrides in long term aging at temperatures above 700 ºC . In niobium stabilized
steels Fe2Nb Laves phase is frequently reported. Also, in 20-25 alloy containing molybdenum
[13]
the Laves phase may form with Fe, Mo, Cr, Ni and Si atoms . Some alloy additions can
modify the precipitation of Laves phase to get better creep strengthening effect; for instance
addition of zirconium has been reported to promote the precipitation of fine dispersions of
intermetallic Laves phases like Fe2Nb and δ-Ni3Nb in a 20-30-2Nb austenitic alloy [32].

A number of other phases can also precipitate in different grades of austenitic stainless steel
under varying condition. For instance Cr2N in a high nitrogen alloy above 900 ºC, π nitride in
a 20-25 high nitrogen grade at 850 ºC and copper-rich nano-scale particles. The latter, as
mentioned earlier, are effective creep strengthening precipitates during long term aging.

There are preferential nucleation sites for the second phase precipitates, which are indeed
different crystallographic defects within the matrix structure. The stored energy at the locus

14
of each defect reduces the net interfacial free energy required for nucleation and growth of
the new phase particles [33]. Grain boundary is considered as the most favored nucleation site
because of its high interfacial energy, its ability to accommodate the structural mismatch
between precipitate and matrix and also high density of vacancies which facilitates the
diffusion of solute atoms [34]. Twin boundaries are the next preferential nucleation sites where
a strain field is associated with stacking fault, especially in incoherent twin boundaries which
[35]
have higher density of crystal imperfections . Interior the grains, the slip bands which are
associated with a high density of tangled dislocations are preferential sites for nucleation of
[25, 36]
precipitates . Slip bands are parallel due to the preferred parallel planes for dislocation
slip within each grain and thus, precipitates forming on slip bands appear like parallel arrays
in the microstructure. Other crystallographic defects such as vacancies and dislocations can
be the locus of precipitate nucleation to form single particles. Also, the inclusion particles as
well as initial precipitates forming in the early stages of aging are the other common
nucleation sites for further precipitation.

The kinetics of precipitation processes in niobium stabilized austenitic stainless steel with 20-
25 matrix has been summarized in a time-temperature-precipitation (TTP) diagram by Powell
[31]
et al (Figure 2.4). Each curve in this diagram represents the temperature and time
condition at which the corresponding precipitate has been firstly observed. The exception is
grain boundary G phase which was supposed to form from the residual Nb(CN) particles on
grain boundaries; so that the time-temperature condition at which these Nb(CN) particles
were disappeared has been considered as the finish curve for grain boundary G phase [31].

850

800

750
Temperature (oC)

700

650

600

550

500

1 10 102 103 104 Time (h)

Figure 2.4. TTP diagram for niobium stabilized 20-25 austenitic stainless steel by Powell et al [31].

15
2.5 Energy dispersive X-ray spectroscopy (EDS) microanalysis

When a solid sample is bombarded with a focused beam of high-energy electrons, as an

interaction product of the electron beam and sample atoms, X-ray emits from the sample with
an energy spectrum proportional to the chemical analysis of the bombarded area. Energy
dispersive X-ray spectroscopy (EDS) makes use of this X-ray spectrum to determine the
localized element composition. All elements from atomic number 4 (Be) can be detected
based on the characteristic energy of X-ray photons emitting from each element [37].

For the first time Moseley in 1913 found that the frequency of characteristic X-ray emitting is
a function of the element atomic number. This led to the early spectrochemical analysis using
X-ray spectrum from the excited specimen to determine its element composition; however the
analyzed area was quite large (>1 mm2). The first instrument using a focused electron beam
was built by R. Castaingin in early 1950’s. Further development led to the electron probe
microanalyzer (EPMA) with a gas-proportional counting detector and wavelength dispersive
spectrometer (WDS) working based on X-ray diffraction on a crystal at a certain angle
proportional to the X-ray wavelength. Finally lithium-drifted silicon solid-state detectors,
[38]
Si(Li), were developed (Fitzgerald et al., 1968 ) to introduce the energy dispersive X-ray
spectroscopy (EDS) as a fast and accurate analyzing method. Since scanning electron
microscope (SEM) works with the same focused electron beam as EPMA, nowadays the X-
ray microanalyzing is usually done by SEMs equipped with EDS and/or WDS detectors to
have a more effective function.

2.5.1 Beam electron-specimen interaction

High energy beam electrons undergo variety of interactions with specimen atoms resulting in
different products which carry information such as specimen’s composition, crystallography
and topography. Beam electrons are scattered within the specimen either elastic or inelastic.
In elastic scattering a beam electron changes its trajectory without losing kinetic energy. This
interaction is responsible for electron backscattering (high-angle elastic scattering) which is
an important signal for imaging. Inelastic scattering occurs when all or a part of electron
kinetic energy is transferred to the specimen atoms generating different products such as
secondary and Auger electrons as well as characteristic X-rays [38]. Trajectories of elastic and
inelastic scattering determine the volume in which beam electrons interact with specimen
atoms, so-called interaction volume. Figure 2.5 shows a scheme of typical interaction volume
with different interaction zones and a Monte Carlo electron trajectory simulation [38, 39].

16
 Electron
beam 0.5 μm

a b

Figure 2.5. (a): Schematic beam electrons-specimen interaction volume [39]; (b): Monte Carlo
simulation of electron trajectory in iron for 20 keV beam energy [38].

The pear shape of the interaction volume is due to the characteristics of elastic and inelastic
scattering. As electrons penetrate into the specimen they lose energy and thus probability of
elastic scattering increases which results in more trajectory deviations (broadening of the
interaction volume). This also provides several interaction zones relative to the beam entering
point differing in product. Auger electrons coming from very close-to-surface are used for
surface analyzing whereas secondary and backscattered electrons are mostly used for SEM
imaging. The characteristic X-rays are the main signals in EDS microanalysis to determine
the chemical composition of the specimen. However other X-ray radiations like continuum
X-ray and the secondary fluorescence may also be used as the sources of other information.
Characteristic of the interaction volume is a function of beam energy, specimen composition
and beam incident angle. There are some empirical expressions to determine the size of the
interaction volume. For instance equations 2.4 and 2.5 have been derived to estimate depth
and width of the interaction volume respectively for a perpendicular electron beam [39]:

) . Eq. 2.4
0.1(
( )=

( ) .
0.0 Eq. 2.5
( )=

Where E0 is the beam energy (keV) and ρ is the specimen density (gr/cm3). The beam energy
is the kinetic energy of beam electrons which is equal to the SEM accelerating voltage
multiplied by the charge of electron. Therefore the accelerating voltage is a major factor to
control the size of the interaction volume.

17
There are also theoretical methods to determine the interaction volume. Monte Carlo electron
trajectory simulation is a common method to visualize the interaction volume. Figure 2.6
shows the effect of SEM accelerating voltage on the size of the interaction volume in iron.

0.5 μm 0.5 μm 0.5 μm

a b c

Figure 2.6. Monte Carlo simulation of the interaction volume in iron as a function of accelerating
voltage, (a): 10 kV, (b): 20 kV, (c): 30 kV [38].

In EDS microanalysis the spatial resolution is determined by the size of the interaction
volume. The smaller the interaction 1volume is, the better the spatial resolution of
microanalysis will be. Usually the larger dimension of interaction volume is considered as the
spatial resolution since the characteristic X-rays come from very deep region of the
interaction volume.

2.5.2 Origin of characteristic X-rays

When a beam electron hits a bounded inner-shell electron of the specimen it can knock the
atomic electron and make an electron vacancy in that shell. The minimum necessary beam
electron energy for this interaction is called critical excitation/ionization energy (Ec). To
resume the ground state a limited set of allowed electron transitions from the outer shells may
occur to fill the vacancy. Since the energy of shell electrons are sharply defined as the
characteristic values of atomic species, the energy difference of the ejected electron and the
substituent one is also sharply defined and can be measured as a characteristic value. This
energy excess can emit from the atom as either the kinetic energy of an Auger electron or the
[38]
energy of an X-ray photon . Figure 2.7 shows the excitation process and two possible
transitions to generate X-ray; a single transition from an outer shell to the vacancy producing
a single X-ray photon and a multiple transition which results in the emission of two X-ray
[39]
photons . The characteristic energy of X-ray can be understood by the energy level
diagram of the atom. Electrons orbit around the nucleus on the specific orbital states with
specific energies; each is identified by a unique set of quantum numbers (n, l, j, mj). The
principle quantum number, n, denotes the shell in which all electrons have close energies

18
while l, j and mj characterize the subshell, magnetic momentum and spin of the electron
respectively.

Figure 2.7. Excitation of atom by incident electron generating single or multiple X-ray photons [39]

The number of energy states or electron population in each shell is determined based on the
defined possible values for quantum numbers l, j and mj per each principle number, n. In the
X-ray notation the closest shell to the nucleus (n=1) is known as K with 2 electrons; the next
is L (n=2) containing 8 electrons, then M (n=3) with 18 electrons, etc. Except the inner shell
K other shells are divided into subshells to be able to have high number of energy states.
Figure 2.8 is a schematic diagram of the inner atomic electron shells.

Figure 2.8. Schematic diagram of the inner atomic electron shells

The characteristic X-rays are named based on the shell/subshell in which the initial vacancy
[40]
occurs and the one from which an electron drops to fill that vacancy . The general
expression for a characteristic X-ray is Spi where “S” is the shell with inner vacancy (K, L or
M), “p” is a Greek letter indicating the inter-shell transition probability in decreasing order

19
(α, β, etc.) and “i” is a number denoting the spectral resolvability of the X-rays within a given
shell in descending order (1, 2, etc.). For instance the most probable characteristic X-ray is
Kα1 while the less probable, Kα2, has very close value which is usually not resolved, and thus
[38]
they are considered as a single characteristic X-ray denoted as Kα . Lots of other X-rays
are also irresolvable from the adjacent ones or have very low probability especially when the
atomic number increases. So a simplified expression of the most observed characteristic X-
rays is usually used (Figure 2.9).

Figure 2.9. Energy level diagram of atom and the most probable transitions producing resolvable
characteristic X-rays [38].

2.5.3 Continuum X-ray

In addition to the characteristic X-ray photons continuum X-ray is also produced from the
beam-specimen interaction. Beam electrons are deflected by strong electromagnetic fields of
the specimen atoms nuclei, decelerated and lose a part of their kinetic energy. This energy
loss is emitted in the form of X-ray photons called braking radiation (bremsstrahlung in
German). Because of the random nature of interaction, all degrees of electron deceleration are
possible, so the radiation has a continuous energy spectrum from zero to the entire beam
electron energy and, thus, is also known as continuum X-ray. The intensity of continuum has
been formulated as equation 2.6 [38]:

̅( )
. .

20
Where ̅ is the average atomic number of the specimen, is the beam energy, i is the beam
current and is the continuum radiation energy. According to the above equation in low
photon energies ( ) the continuum intensity will be higher. Also when heavy elements (high
Z) are being analyzed their stronger nucleus fields cause higher density of electron deflection
[38]
and thus the intensity of continuum radiation increases . The importance and effect of
continuum X-ray on EDS analysis is discussed in the following section.

2.5.4 EDS X-ray spectrum

X-rays are detected by a solid state detector which is usually a lithium-drifted silicon Si(Li)
semiconductor crystal. The captured X-rays create electron-hole pairs forming charge pulse
under an applied bias. The pulses are then converted to voltage signals by a charge-to-voltage
converter, so that each voltage signal is proportional to the energy of the incoming X-ray
photon. These signals are amplified by a linear amplifier and finally passed to a computer
analyzer where the data is digitized and displayed as a histogram of characteristic lines with
varying intensities over voltage (energy) [37-40].

Figure 2.10 shows a typical EDS spectrum of an austenitic stainless steel (one of the alloys of
the present work) consisting of the characteristic lines (peaks) and continuum X-ray.
Fe Kα

Cr Kα
Ni Kα

Mo Lα ,2
Number of counts

Lα lines Cu Kα
Cr Kβ Fe Kβ

Ni Kβ
Si Kα
Continuum X-ray Cu Kβ

keV
FWHM
Energy (keV)

Figure 2.10. EDS spectrum of an austenitic alloy of the present work and position of characteristic
lines, taken by SEM with EDS detector at 20 kV accelerating voltage. Continuum X-ray is also
illustrated. The energy resolution of nickel Kα peak is pointed by a red dash and FWHM letters.

Although the characteristic lines have sharply defined energies, in EDS spectrum they are
broadened to Gaussian profiles (peaks) because of the convolution of detector-analyzer

21
 [38]
response effect . The position of each peak corresponds to the X-ray characteristic energy
of a certain element and thus the present elements within the analyzed volume can be
identified, so-called qualitative analysis. Also the peak height (intensity) shows the detection
frequency of its corresponding characteristic X-ray which is proportional to the concentration
of the excited element in the analyzed volume. Therefore the local composition of the
specimen can also be obtained, so-called quantitative analysis. EDS spectrum has some
important features determining the quality and accuracy of the analysis. Here some major
ones are discussed.

Continuum X-ray

An EDS detector collects both characteristic and continuum X-ray photons. Since the
continuum X-rays have a continuous energy spectrum they produce spectral background
intensity under the characteristic peaks (Figure 2.10). This background is usually considered
as a nuisance to the EDS analysis because the low intensity peaks coming from low
[38]
concentrated elements may be hidden under the background ; so the lowest detectable
concentration of alloying elements (detection limit) is strongly dependent on the intensity of
continuum X-rays. As it was discussed under the equation 2.6 it can be added that phases
containing heavy elements have higher detection limits because of producing more intense
continuum [37].

Energy resolution

The energy resolution (FWHM) is defined as the width of the peak at its half maximum.
Usually manganese Kα peak at 5.89 keV is measured as a standard reference. The energy
resolution of the common EDS systems using Si(Li) detector is in the range of 130–150 eV,
[37]
however, by Ge detectors this amount has been lowered to 115 eV . As an example the
energy resolution for the above EDS spectrum can be measured as ~160 eV for nickel Kα
peak (red dash in Figure 2.10). FWHM depends on the variety of factors like intrinsic and
thermal electronic noises in detector and the so-called dead time. A background level of noise
is always present which can be lowered by choosing longer process time, i.e. giving more
time for averaging the incoming voltage signals from detector [40] resulting in narrower peaks.
Dead time is the time given to the detection system after the arrival of an X-ray photon to be
[38]
processed and added to the recorded spectrum before entering further photons ; so
increasing the dead time can intensify the peaks. However, there is always a trade-off
between the process time and the dead time for the least peak broadening (best FWHM)

22
because the longer the process time, the longer pause to measure each X-ray, and the fewer
events that can be detected [40].

Peak-to-background ratio

Peak-to-background ratio (P/B) is an important factor determining the ability of the analyzer
to detect and measure the characteristic peaks from the underlying background. Higher peak-
to-background ratio provides more accurate analysis and lowers the detection limit. In EDS
spectrum Kα lines have much larger P/B ratio than L and M lines (Figure 2.10) because at
sufficient beam energy Kα characteristic X-rays have the highest probability to be generated
(section 2.5.2); so analysis using Kα lines provides high accuracy. Although increasing the
accelerating voltage provides more intense peaks P/B ratio may not necessarily improve since
continuum X-ray also intensifies at higher voltages (equation 2.6). The best way to improve
P/B ratio is prolonging the acquisition time with optimum selection of other parameters such
as process time and dead time.

Spectral artifacts

EDS spectrum usually contains different artifacts where most of them originate from the
detector electronics. The two most common artifacts are pulse pileup peaks and silicon
escape peaks. Pileup peaks occur when two consecutive X-ray photons are recorded as a
single one because of their nearly same arrival time to the detector. It can be reduced by
lowering the beam current and/or process time [40].

2.5.5 Accuracy of EDS microanalysis

The accuracy of EDS, both qualitative and quantitative analyses is a direct function of the
above mentioned characteristics of EDS spectrum. Apart from the instrument capability, the
quality of EDS spectrum can be improved to some extent by adequate operational control.
The effect of some controllable parameters such as acquisition time, process time, dead time
and beam current were shortly discussed in the last section.

Regarding the qualitative analysis, the accuracy depends on the ability of analyzer to
distinguish the characteristic lines (peaks) from each other as well as from the background,
measure their energy and compare them to those of the predefined values to identify the
present elements. For this purpose [38]:

1. Spectrometer needs to be properly calibrated with a reference element having the main X-
ray line (Kα line) close to that of the interest.

23
2. Operating condition must be properly set up to give sufficient counts of the present
characteristic X-rays providing adequate P/B ratio. The smallest detectable peak is usually
considered as three times the standard deviation of the background [37, 40]
3. Serious peak overlaps must be avoided and in the case of inevitable overlaps like MoLα
(2.293 keV) and SKα (2.307 keV) the proper deconvolution must be done.

In quantitative analysis “the intensity of present peaks in the EDS spectrum corresponding to
preselected elements for both sample and standard are measured under the same operating
[38]
condition” . Then the intensity ratios (peak/standard) are converted to the chemical
[38]
concentration of preselected elements . In addition to the above listed considerations in
operating the EDS analyzer a number of corrections such as correction of matrix atomic
number-absorption-fluorescence effect (ZAF correction) is also carried out by the analyzer
itself to improve the accuracy of quantitative analysis (discussed in detail in reference 31).
[37, 38]
The common accuracy of EDS analyzers is ~1% for major elements which means 1%
error of the measured content from the actual/accepted content. This amount can be higher
(poorer) for too light or too heavy elements because of different factors such as matrix effect
(for light elements) and bounding effect [41] (for heavy elements). Also the common detection
limit of EDS analysis is 0.1% when K lines are used giving adequate P/B ratio.

2.6 Low voltage EDS microanalysis

As mentioned earlier the energy of electron beam, as a function of accelerating voltage, must
overcome the critical excitation energy of the X-ray line of interest. However, higher energies
with a factor of 1.5 to 3, termed overvoltage, are always needed to have adequate excitation
efficiency. A common used SEM accelerating voltage for EDS microanalysis of iron-base
alloys is 20 kV. This voltage provides sufficient over-voltage for effectively exciting the
majority of present elements to give Kα lines, by which an accurate analysis can be made.
The exception is a group of heavy elements, commonly being alloyed to steel such as
niobium and molybdenum with high energy Kα lines (>16 keV) but instead having resolvable
lower energy L lines. According to the empirical equations 2.4 and 2.5 the estimated
dimension of the interaction volume at 20 kV is ~1.1 μm in depth and ~0.9 μm in width
which means a spatial resolution of almost 1 μm for EDS microanalysis. Therefore
characterization of sub-micron particles becomes so difficult or even impossible in case of
nano-scale particles due to interference of the interaction volume with matrix. The main
advantage of scanning transition electron microscopy (STEM) is its thin specimen and extra

24
high voltage which limit the interaction volume and also X-ray absorption [42] resulting in the
superior spatial resolution of a few nanometers [43]. However, an upper limit for the effective
spatial resolution of STEM is usually considered as 0.1 μm with respect to the common
[44]
thickness of its thin foil samples . The gap between the chemical resolution capability of
SEM-EDS (>1 μm) and that of the STEM-EDS (<0.1) can be filled by using low accelerating
[44]
voltage in high resolution SEM . For this application field, emission gun SEM (FE-SEM)
is preferable due to its highly concentrated beam to minimize the effect of chromatic
aberration and get adequate resolution [45].

When reducing the accelerating voltage from 20 to 5 kV the estimated interaction volume can
be reduced for the same specimen to ~140110 nm which means almost 8 times better spatial
resolution. Also, the matrix absorption effect is reduced thanks to the shorter X-ray escape
path in matrix when the beam penetration depth is smaller at low voltage. Therefore, Low
voltage EDS is suitable for analyzing fine particles, light elements and thin coatings [43].

Figure 2.11 shows the characterization of a sub-micron precipitate in an austenitic stainless

steel studied in the present work using 5 kV accelerating voltage in FE-SEM with attached
EDS analyzer. The appearance of niobium, boron, carbon and nitrogen peaks in the EDS
spectrum shows that these elements have segregated to the analyzed particle.

a b

200 nm
keV

Figure 2.11. (a): SEM image of a sub-micron precipitate in an austenitic stainless steel of the present
work; (b): The corresponding EDS spectrum of the precipitate at 5 kV accelerating voltage showing
enrichment of niobium, carbon, nitrogen and boron in the particle

In the case of a thin coating layer Figure 2.12 shows how decreasing the accelerating voltage
can correct the chemical analysis of a 0.18 μm aluminum film on a silicon wafer. The
interference of the interaction volume with substrate decreases with reducing the voltage and
is fully eliminated in 4 kV (c) [44].

25
 a b c

1.487 1.740 keV

Figure 2.12. EDS spectrum of an aluminum coating film (0.18 μm thick) on the silicon wafer at
different accelerating voltages, (c): correct analysis can be obtained at 4 kV [44]

In the recent improvements of high-resolution FE-SEM with optimized EDS analyzers the
spatial resolution of 20 nm has been obtained [43]. This has markedly improved the capability
of SEM-EDS microanalyzers to cover a wide range of spatial resolutions above few tens of
nanometer. Within this range the low voltage SEM-EDS can even compete with STEM
thanks to its advantages such as easier instrument operation, no need of thin foil sample
preparation, possibility of analyzing bulk samples and better counting statistics due to the
larger available analyzing area [42-44].

2.6.1 Drawbacks of low voltage EDS

There are some weak points in using low voltage EDS microanalysis which are discussed in
detail in section 4.6; here some major ones are reviewed.

The main problem is the limited number of characteristic lines in the X-ray spectrum at low
voltage, especially K lines as the most probable and resolvable lines [43, 45]. As it can be seen
in Table 2.2, giving the main characteristic energies (Kα and Lα) of common metallic
elements, except magnesium and aluminum with Kα at 1.254 and 1.487 keV respectively
other common metallic elements cannot be effectively excited at voltages below 10 kV to
give Kα lines. Therefore, only L lines are available in low voltage EDS having much lower
P/B ratios than Kα peaks. Also, in the low energy region of the X-ray spectrum, e.g. below 3
keV, there are lots of compact low-intensity L and M lines and thus, peak overlap can be a
serious problem reducing the accuracy of both qualitative and quantitative analyses [43].

Table 2.2. X-ray characteristic energies (keV) of common metallic elements [47]

Element Mg Al Ti V Cr Mn Fe Ni Cu Zn Nb Mo

Kα 1.254 1.487 4.511 4.952 5.406 5.899 6.404 7.478 8.048 8.639 16.615 17.479

Lα — — 0.452 0.511 0.573 0.637 0.705 0.852 0.930 1.012 2.166 2.293

26
The other problem of low voltage EDS is related to the characteristic of continuum X-ray at
low beam energy. According to the equation 2.6 the intensity of continuum X-ray for a given
beam energy (E0) is higher in the low energy region of X-ray spectrum (low Ev) and also
increases with increasing the average atomic number of the specimen. This can reduce the
P/B ratio, especially for L lines (high-Z) and deteriorate the detection limit. Moreover, at low
accelerating voltage the analysis is generally more sensitive to surface condition due to the
smaller interaction depth. This can increase the negative effect of surface contamination in
producing nuisance peaks in EDS spectrum.

27
28
 Chapter 3

3 EXPERIMENTAL WORK

3.1 Alloy selection

As discussed earlier the 20Cr-25Ni austenitic stainless steel family has been extensively
studied and was mainly developed to be used as SH/RH tube material although its application
is by no means limited to this sector; but there is still potential for further development. The
aim of the present work was to study the effect of simultaneous presence of vanadium,
niobium, nitrogen and copper on the microstructure of a 20Cr-25Ni alloy at elevated
temperatures. The designed composition for the alloy was according to Table 3.1.

Table 3.1. Designed chemical composition ranges for studied alloys

Element Cr Ni Mn Si Mo C Cu N V Nb
Wt % 20 25 1 ≤0.6 1.5 ≤0.015 2.8 0.1-0.25 0-0.4 0-0.4

Manganese and silicon were decided to be at the similar contents as previous studies on
[28, 31]
20Cr-25Ni alloys . Molybdenum was supposed to improve the corrosion properties at
this level with no detrimental effect on precipitation. Carbon was kept at the lowest possible
level to minimize the risk of chromium depletion along grain boundaries. Copper, as
discussed earlier, with contents in excess of 2.5% can effectively form fine precipitates [23, 24];
however, no more than 3 wt% copper has been used in creep resistant alloys (Table 1.1).
Therefore, a copper content of 2.8% was selected for this study as a constant alloying element
in the matrix. Nitrogen varied from 0.1 to 0.25 to see how it affect the precipitation of nitrides
and carbonitrides in the presence of vanadium and/or niobium. Finally, the contents of
vanadium and niobium varied in such a way to be able to investigate their competitive effects
on the precipitation of different phases such as carbides, nitrides and carbonitrides as well as
their probable effect on the precipitation of nano-scale copper-rich phase.

3.2 Base melt preparation

A medium-frequency high vacuum melting furnace was used to make the base melt. When
vacuum was applied the chamber was filled by pure nitrogen at 300 torr pressure to set an
equilibrium state with the dissolved nitrogen in order to avoid nitrogen escape from the melt.
Table 3.2 gives mass and chemical composition of charged materials. The melting process

29
took almost 5.5 hours. At the final stage, before casting the melt into the crucible, aluminum
was added for deoxidization.

Table 3.2. Chemical composition and mass of charged materials for the base melt

Charged material Chemical composition (Wt %) Mass (gr)

Stainless steel 725LN 25.1Cr, 22.15Ni, 2Mo, 1.7Mn. 0.2Si, 0.13N, 0.012C 17717
Pure iron ingot 99.86Fe, 0.009C, 0.064Mn, 0,004Si, 0.02Ni, 0.015Cr 4785
Fe-Cr 30.4Fe, 69Cr, 0.052C, 0.545Si 595
Fe-Si 25Fe, 75Si 160
Fe-Mo 31Fe, 68.4Mo, 0.5Cu, 0.02C 103
Fe-Cr-N 24Fe, 65.4Cr, 10.55N, .04C 125
Cu 100Cu 755
Cr 100Cr 700
Ni 100Ni 3100
Mn 100Mn 10
Al 100Al 20

The resulted chemical composition of the ingot is given in Table 3.3.

Table 3.3. Chemical composition of the base melt ingot

Element Cr Ni Mn Mo Si C N Cu P S Nb V
Wt % 19.88 25.23 1.1 1.54 0.57 0.014 0.111 2.81 0.011 0.003 0.004 0.043

The oxygen content of the base melt was analyzed as 119 ppm. This relatively high amount
was due to a little leakage in the furnace and failure in aluminum addition which caused an
incomplete deoxidization.

3.3 Making mini-melts

In order to have desired variation in the chemical composition, smaller melts (mini-melts)
had to be made from the base melt with re-melting it and addition of Nb, V and N. For this
purpose a laboratory scale high frequency melting furnace (Figure 3.1) was used. A number
of similar slices were cut from the base melt ingot with dimension not bigger than 81.51
cm. Then the slices were precisely weighed and charged into small crucibles with addition of
certain amounts of alloying elements to get the desired chemical analysis for each mini-melt.
Figure 3.1 shows different parts of the laboratory scale high frequency melting furnace used
for making mini-melts. The main challenges of using this method were difficulty in repetition
of similar melting process, controlling the low level of carbon and obtaining the desired

30
nitrogen content. Every minor change in charging and operating of the furnace, e.g. size and
contact area of the charged pieces and the relative position of the crucible and the coil
affected hugely the melting time and the final analysis of the melt.

Ventilation hood

Refractory cap

Quartz chamber

Charged
Crucible Induction coil

Protection gas
Electric lift inlet
Table

Figure 3.1. The laboratory scale high frequency melting furnace used to make mini-melts.

Nitrogen was used as the protection gas inside the quartz chamber. The average melting time
was 9 minutes. When the heat was completely molten it was held for 30 to 60 seconds before
switching off the furnace to give the melt enough time to be stirred by its internal eddy flow.
In order to check the homogeneity of the melt, one sample was analyzed from two sections
with 10 millimeters distance. The result showed a similar analysis confirming that the given
time to the melt is sufficient for the alloying elements to be mixed homogeneously. Finally
thirteen mini-melts weighing around 700 gr were made by this method. Figure 3.2 shows one
of the typical mini-melt samples and the analyzed section.

10 mm

Figure 3.2. One of the mini-melt samples and the analyzed section

31
Of the thirteen mini-melts four (no. 5, 7, 10 and 11) were selected for further experiments
with the chemical analyses as listed in Table 3.4. A more easily understood nomenclature is
used to indicate the main alloying concept of each sample.

Table 3.4. Chemical analysis of four selected samples

Alloying elements (wt%)

Mini-melt
Cr Ni Mn Mo Si C N Cu P S Nb V O (ppm)
5-V 20.15 25.18 1.0 1.53 0.65 0.015 0.119 2.83 0.01 0.003 0.004 0.412 270
7-Nb 20.12 25.19 1.08 1.49 0.59 0.017 0.121 2.79 0.012 0.003 0.36 0.039 188
10-VNb 20.15 25.29 1.08 1.51 0.59 0.016 0.12 2.82 0.012 0.003 0.44 0.41 221
11-VNbN 20.19 25.31 1.09 1.51 0.63 0.017 0.213 2.83 0.013 0.002 0.46 0.415 106

The oxygen content of mini-melts varied from 106 to 270 ppm. This variation was due to the
technical difficulties for deoxidization in the laboratory scale furnace and also changes in the
melting condition of different mini-melts. As it is discussed later in chapter 4 this oxygen has
no significant effect on the study.

3.4 Forging and solution annealing

To break up the solidification microstructure of mini-melts, make them more homogenous

and also form the desired dimension for final samples, a hot forging process was used. For
this purpose a proper handle was welded to each mini-melt. Then they were heated up to
1200 ºC and pressed in multiple steps using a hydraulic open-die press. Figure 3.3 shows a
specimen under forging and its final shape.

a b c

Figure 3.3. Forging at 1200 ºC in multiple steps in thickness (a) and width (b) to form the final cross
section of 1150 mm (c)

Eight steps of pressing were done to reduce the thickness and four steps for the width.
Between the steps specimens were kept in the furnace at 1200 ºC temperature. The forging
process was set up to get the final cross section of 1150 mm which lasted almost 2 h for

32
each sample. After the final pressing step samples were quenched in water. Then solution
annealing was carried out on the forged pieces at 1150 ºC for 30 minutes followed by
quenching in water to get supersaturated solution of alloying elements in austenite.

3.5 Aging

Solution annealed samples were cut out into smaller pieces in order to make three groups of
samples each containing all 4 alloys. Then they were arranged in 3 furnaces to be aged at
three temperatures of 700, 800 and 850 ºC for 500 h. Sample number 10 and 11 suffered
catastrophic oxidation at 800 and 850 ºC after different aging times (Figure 3.4).

10 mm

Figure 3.4. Catastrophic oxidation in sample 10-VNb after 260 h aging in 850 ºC

3.6 Metallography and microstructural examination

Polishing of the samples ended with 3 µm diamond paste. Then electrolytic etching was used
to reveal grain/twin boundaries and precipitates. The etchant was Aqua regia or V2A (10 HCl
: 10 H2O : 1 HNO3) at 50 ºC. The selected etching method was according to the company’s
internal instruction [48] and the proper etching time was found as 35 seconds by checking the
microstructure at different times. The instruments used for the microstructural observation
were light optical microscope and ZEISS Ultra 55 field emission scanning electron
microscope (FE-SEM). The FE-SEM provided high resolution imaging thanks to its
concentrated beam and a set of sensitive high efficiency secondary and backscattered electron
detectors.

3.7 Characterization of precipitates using EDS microanalysis

To identify the precipitates energy dispersive X-ray spectroscopy (EDS) was carried out. The
instrument was FE-SEM equipped with an Oxford EDS microanalyzer (INCA-Penta Fet X3).
The microanalysis was done at high and low accelerating voltages (20 and 5 kV) to identify

33
micron-sized and sub-micron-sized precipitates, respectively. Pure cobalt (Kα=6.93 and
Lα=0.776 keV) as a common element for quantitative optimization of the EDS system was
used for 20kV accelerating voltage. For optimization at 5kV voltage an element was required
with Kα line below 2 keV to get adequate accuracy in the EDS analysis. Of the few available
pure elements, silicon (Kα=1.74) was selected for this purpose.

To study the effect of operational controls on the EDS analysis different manipulations on the
acquisition setup was carried out and also 2.5 kV accelerating voltage was tested in order to
see whether it can improve the accuracy of the analysis.

34
 Chapter 4

4 RESULTS AND DISCUSSION

4.1 Sample preparation process

The sample preparation process has always effects on the entire study. Here some important
aspects of the sample preparation are discussed.

4.1.1 High oxygen content

As mentioned earlier, due to some technical problems in the melt preparation the
deoxidization was inadequate and resulted in high oxygen content varying between 106 to
270 ppm in different samples. Since the solubility of oxygen in the solid steel is low (~10
[49]
ppm) the excess amount of oxygen forms nonmetallic phases such as MnO and SiO2 .
Figure 4.1-(a) shows SEM image of sample 5-V (containing 270 ppm oxygen) after solution
annealing in which the dark phase is a large inclusion. The EDS spectrum of the inclusion (4-
b) shows peaks of oxygen, silicon, manganese and chromium.

a b

15 μm keV

Figure 4.1. (a): SEM image of sample 5-V in solution annealed state with an inclusion particle, (b):
EDS spectrum of the inclusion at 20 keV beam energy showing the presence of O, Si, Mn and Cr.

The same EDS analysis was carried out for the inclusions in all samples. The resulted average
chemical composition of inclusions (three analyses per sample) was according to Table 4.1.
The oxygen content of each sample is also given in the Table 4.1. As it can be seen,
inclusions are oxides where O, Cr and Mn have the highest proportion. In sample 5-V
comprising the highest oxygen content (270 ppm) silicon is also a major oxide former.

35
 Table 4.1. EDS quantitative analysis of the inclusion particles (average of 3 analyses per sample)

Oxygen Chemical analysis of inclusion (wt%)

Sample content
(ppm) O Fe Cr Ni Mn V Cu Si Al Other
5-V 270 36.2 2.8 34.0 0.8 20.5 2.6 — 6.1 1.6 —
7-Nb 194 30.4 14.4 31.2 6.5 14.3 — 0.9 0.2 1.7 0.5 Ti
10-VNb 221 28.8 1.8 38.2 0.9 21.2 3.7 — 0.3 0.8 0.7 Ti
11-VNbN 106 28.9 7.1 33.8 0.6 25.0 2.4 — — 6.3 0.9 Ti

The average volume fraction of inclusions in sample 5-V was estimated as 0.13% using a
manual point counting method according to ASTM E 562. From that the amount of extracted
elements from the matrix to form oxides can be obtained by multiplying their concentration
in the oxide (Table 4.1) by the volume fraction of inclusions. For instance the amount of Mn
existing in the matrix (1.0 wt%, Table 3.4) will reduce by (0.20510.13) or ~0.027 when it
segregates into the oxide inclusions. The same calculation for other elements taking part in
the oxide phase was carried out with results as given in Table 4.2.

Table 4.2. Reduction of the matrix elements in wt% to form oxide inclusions in sample 5-V

O Fe Cr Ni Mn V Cu Si
0.047 0.004 0.044 0.001 0.027 0.003 — 0.008

This range of reduction in the alloying elements is negligible. The calculated amount for
oxygen (470 ppm) is higher than its initial analyzed amount (270 ppm) indicating that almost
all oxygen in the matrix has formed oxide inclusions rather than remaining in solution.

4.1.2 Solution annealing

Solution annealing treatment was 30 min heating at 1150 ºC followed by water quenching
which was selected based on the previous works in Outokumpu and also those that have been
[28, 31, 50]
suggested in literature for similar alloys . The SEM microstructure of solution
annealed samples can be seen in Figure 4.2. In sample 5-V almost no residual phase was
observed which verified the selected annealing temperature and time for this alloy; but in
niobium containing samples (7-Nb, 10-VNb and 11-VNbN) some residual phases remained
undissolved in the microstructure (bright second phase particles in micrographs (b), (c) and
(d) of Figure 4.2). They might have been formed in any of the previous steps of sample
preparation (e.g. solidification of mini-melts, hot forging, etc.), especially when considering
the variation in their morphology and formation site. In this step only micron-scale particles
are analyzed and sub-micron ones are discussed in section 4.5.

36
 a b 1

2
10 μm 5 μm

c d
1

2 1

3 μm 400 nm

Figure 4.2. Solution annealed SEM images; (a): sample 5-V with no residual phase; (b): sample 7-Nb,
two locations of residual particles are pointed as 1 (grain boundary) and 2 (interior grain); (c): sample
10-VNb, residual phase particles on twin boundary; (d): sample 11-VNbN, sub-micron residual phase
particles located on grain boundary (1) and interior grain (2).

EDS microanalysis was carried out to identify the residual phase with the resulted average
chemical analysis (at least five particles per sample) according to Table 4.3. The average
volume fraction of residual phase was also estimated using a manual point counting method
according to ASTM E 562 and the result is given in Table 4.3.

Table 4.3. Average chemical analysis of residual phase in samples 7-Nb, 10-VNb and 11-VNbN

wt% of detected elements volume

Sample fraction
B C N V Cr Fe Ni Nb (%)
7-Nb 13.7 6.6 8.1 0.1 19.3 7.1 2.2 43.0 0.39
10-VNb 11.5 6.9 7.8 2.5 19.5 7.1 2.4 42.4 0.28
11-VNbN 10.7 6.4 7.5 3.7 20.0 10.1 4.2 35.4 0.32

It can be seen that the composition of residual phase is very close in three samples with Nb-
C-N-B base and some other dissolved elements. The high concentration of niobium and

37
nitrogen in the residual phase is undesirable due to the reduction of their content in the matrix
and consequently their effect in precipitation under aging; especially when considering the
relatively high volume fraction of residual phase in all samples. For instance in sample 7-Nb
the amount of niobium extraction from the matrix to the residual phase can be calculated as
(0.39×43.0)/100 or ~0.17 wt% which means that almost half of the initial niobium content
has been absorbed by the residual phase and thus, its precipitation effect during aging is
almost halved. Also, high boron content can be seen in the analysis of the residual phase,
though it had a very low content in mini-melts (<0.001 wt%). Boron is recognized as a highly
active element in substitution for nitrogen and carbon even at very low concentration [13], so it
can be suggested that almost the whole boron content of alloys, though too low, has dissolved
in the niobium carbonitride phase.

The average grain size of the solution annealed samples was estimated in the range 96–129
µm (Table 4.4) using ASTM E112 standard. This range of grain size is at least twice of the
[50]
common standard grain size for commercial 20-25 alloy products such as SH/RH tubes .
Some reasons for this grain enlargement can be suggested as slow solidification of the mini-
melts in refractory crucibles, a long hot forging process at 1200 ºC which might have caused
grain growth and no cold working during the process.

Table 4.4. Estimated average grain size of the solution annealed samples using ASTM E112

Sample No. Average grain size (μm)

5-V 129
7-Nb 101
10-VNb 108
11-VNbN 96

With respect to the fact that the niobium-rich residual phase has always been observed in
niobium stabilized 20-25 alloys [12, 28, and 31], in order to avoid further grain growth the solution
annealing was not redone at higher temperature for complete dissolving of the residual phase.

4.1.3 Catastrophic oxidation

Samples 10 and 11 suffered complete catastrophic oxidation at aging temperatures 800 and
850 ºC. As discussed earlier, molybdenum and vanadium have low melting point oxides
which can cause catastrophic oxidation under unfavorable conditions. In the case of
molybdenum one of the conditions for catastrophic oxidation of stainless steel has been found
as stagnant aging atmosphere [51]. To try to eliminate this condition a sample was aged inside

38
a tube furnace at 800 ºC with some internal air flow. After 200 h aging this sample was also
found to be catastrophically oxidized. Since the other available ways to seal the samples and
safely age them at 800 and 850 ºC were time-consuming and also the time and material was
limited, it was decided to proceed with remaining samples. Therefore, at aging temperatures
above 700 ºC only samples 5-V and 7-Nb were available to be examined.

4.2 Optical microscopy

The used etching method (35 sec in V2A at 50 ºC) revealed grain and twin boundaries,
precipitates on these boundaries and also fine intragranular precipitates. Figure 4.3 illustrates
some microstructures of etched samples.

a b
3

1
20 μm 20 μm

c d

20 μm 20 μm

Figure 4.3. Microstructure of samples after 500 h aging, etched in V2A at 50 ºC for 35 sec. (a):
sample 5-V with precipitates on grain boundaries (1), twin boundaries (2, 3) and interior grains (dark
dots in the matrix). (b): sample 5-V aged at 850 ºC, coarsened precipitates are arrowed. (c): sample 7-
Nb aged at 800 ºC and (d): sample 10-VNb aged at 700 ºC where the residual phases are arrowed.

In microstructure (a) of Figure 4.3 for sample 5-V aged at 700 ºC, precipitates on grain
boundaries, coherent and incoherent twin boundaries and within grains are visible. In (b),
(Sample 5-V aged at 850 ºC) had coarsened grain boundary precipitates which are pointed by

39
red arrows. In sample 7-Nb aged 500 h at 850 ºC (microstructure c) grain/twin boundary
precipitates are coarsened and their surrounding regions are precipitate free. Also residual
phases pointed with red arrows are present. In microstructure (d) for sample 10-VNb aged at
700 ºC, residual phases are pointed with arrows and the typical grain/twin boundary
precipitates as well as intragranular particles can be seen. Since a high resolution FE-SEM
was also available for microstructural study, the use of optical microscopy was mostly limited
to the sample preparation process where a quick verification of the microstructure was
necessary; e.g. after solution annealing and etching.

4.3 Scanning electron microscopy

The FE-SEM had several detectors to be used for imaging. Figure 4.4 shows the SEM image
of sample 5-V, aged 500 h at 700 ºC, taken by different detectors.

a b

c
d

Figure 4.4. SEM image of sample 5-V aged for 500 h at 700 ºC using different detectors to image the
same site; (a): SE (secondary electron) detector, (b): high efficiency In-lens SE detector, (c): EsB
(energy selective back-scattered electron) detector and (d): AsB (angle selective backscattered
electron) detector. In-lens SE detector (b) gives the highest contrast with the maximum number of
resolvable second phase particles.

40
It can be seen in Figure 4.4 that the In-lens SE detector (micrograph b) gives the highest
contrast with the maximum number of resolvable second phase particles (a bit better than
micrograph (c) by EsB detector). The EsB detector provides high resolution contrast
especially at low voltage; while AsB detector is used for compositional and crystal
orientation imaging [46, 47]. Therefore, SEM imaging was done using the In-lens SE detector.

Figure 4.5 shows SEM image of sample 5-V aged 500 h at 700 ºC. In comparison to the
solution annealed state (Figure 4.2-a) the second phase particles have precipitated on
different sites with varying morphologies. The precipitates could be generally divided into
three groups in terms of their formation site and in the order of decreasing in size: precipitates
on grain boundaries, those on twin boundaries and precipitates located in the interior of
grains. As it can be seen in the micrograph, grain boundary precipitates (pointed as 1) have
sizes mostly in the range 0.5–1 µm and are either separate particles or connected strings.

2 4
1
3

3 μm

Figure 4.5. SEM microstructure of sample 5-V, aged at 700 ºC for 500 h containing three groups of
precipitates; (1): grain boundary precipitates, (2), (3) and (4): twin boundary precipitates on
incoherent, vicinity of incoherent and coherent boundaries respectively and (5): intragranular aligned
precipitates locating on the slip bands.

The next group are twin boundary precipitates which have appeared in three forms:
precipitates on incoherent twin boundaries (pointed as 2), plate-shaped precipitates starting
from the incoherent boundaries and elongated parallel to the coherent boundaries (pointed as
3) and also those formed along coherent boundaries (pointed as 4). This precipitation pattern
on and in the vicinity of twin boundaries is typical in the microstructure of high temperature
[25, 35]
austenitic grades and can be frequently found in the literature . The precipitation on
incoherent twin boundaries and in their vicinity starts earlier with faster growth rate;

41
however, precipitates on coherent and incoherent boundaries have been often found to be of
[25]
the same type . Due to this fact and also the larger size of the incoherent twin boundary
precipitates, they are mostly analyzed and commonly termed as “twin boundary precipitates”
in this work.

The third group, intragranular precipitates sizing below few hundreds of nanometer can be
seen in this magnification as bright dots in the dark matrix. They are mostly located on the
straight lines parallel to a certain direction within each grain (pointed as 5). These lines are
recognized as the slip bands containing a group of slip planes where dislocations are aligned
[25, 28, 36]
, providing high-energy sites for precipitates nucleation. This pattern of precipitation
was common in all samples with aging temperature of 700 ºC. Figure 4.6 illustrates SEM
image of sample 10-VNb aged at 700 ºC for 500 h where the letters represent particles of the
above mentioned three groups: G, grain boundary precipitates; T, twin boundary precipitates
and I, intragranular sub-micron precipitates. Also, the residual phase particles are denoted by
R.

R T G

R
G

8 μm

Figure 4.6. SEM image of sample 10-VNb, aged 500 h at 700 ºC; G: grain boundary precipitates, T:
twin boundary precipitates, I: intragranular aligned particles and R: residual phase particles

The reason why the intragranular precipitates do not seem aligned in all grains refers to the
effect of sectioning direction with respect to the slip bands direction which differs from one
grain to another. Also there are always dispersed precipitates with no specific orientation
relationship which nucleate randomly within the matrix in the locus of different defects.

42
At higher aging temperatures (800 and 850 ºC) a similar precipitation pattern was observed;
however, grain and twin boundary precipitates had larger average size and more spherical
shape compare to those at 700 ºC. Moreover, the surrounding regions of boundaries were free
of the intragranular precipitates apparently due to the local depletion of alloying elements
which have segregated into the coarse grain/twin boundary precipitates (Figure 4.7).

2 μm

Figure 4.7. SEM image of sample 5-V, aged at 800 ºC for 500 h; grain and twin boundary precipitates
(G and T respectively) are coarser and more spherical than those at 700 ºC. Precipitation free zones
around grain/twin boundaries are obvious. Intragranular precipitates (I) are still oriented.

In higher magnifications, the intragranular precipitates can be seen to have more well-defined
shapes by increasing the aging temperature. This morphological change was more obvious in
sample 5-V where cuboidal precipitates appeared by increasing the aging temperature from
700 ºC to 800 and 850 ºC (Figure 4.8).

a b

250 nm 250 nm

Figure 4.8. (a): SEM image of sample 5-V aged 500 h at 700 ºC, few sub-micron precipitates have
well-defined shapes (pointed by arrows); (b): SEM image of the same sample with 800 ºC aging
temperature, sub-micron precipitates have well-defined shapes (mostly cuboidal).

43
4.4 Identification of precipitates using EDS microanalysis

This section deals with the identification of micron-scale precipitates using EDS at the
ordinary voltage of 20 kV. As mentioned earlier, EDS gives the chemical composition which
is not necessarily enough for phase identification. Supplementary methods such as X-ray
diffraction are usually needed for crystallographic study and by that, more reliable phase
identification. Also, there is uncertainty in the obtained chemical composition from EDS
because of possible interference between the analyzed precipitate and matrix within the
interaction volume and also the effect of matrix and surface contamination on detection of
light elements. Therefore, in this study the EDS quantitative analyses are categorized into
different composition ranges in a way that each range is similar to a composition which is
specified for a certain phase in the literature. In other words, the nearest phase composition
from literature is suggested for each obtained composition range as the most probable phase.
However, it should be added that this method has always some degree of uncertainty since
only the compositional similarity is taken into account. Chemical compositions are given in
the atomic percentage so that, the atomic ratio of the present elements can be easily obtained
to facilitate the estimation of the possible chemical formula.

4.4.1 M23C6

As the most common precipitate in austenitic stainless steels, M23C6 carbide was expected to
be found in the microstructure of aged samples. Figure 4.9 shows the SEM microstructure of
sample 5-V, aged for 500 h at 700 ºC and the corresponding EDS spectra of the matrix and a
grain boundary precipitate. The precipitates pointed as 1 (grain boundary), 2 (incoherent twin
boundary) and 3 (around the dark phase inclusion) were suggested to be M23C6 carbides. The
EDS spectrum of the grain boundary precipitate (b) can be compared with that of the matrix
(c) where the chromium Kα peak has intensified in spectrum (b) accompanied by the
appearance of carbon peak and strengthening of molybdenum Lα peak relative to the matrix
spectrum (c). The resulted quantitative analysis of the above particles is given in Table 4.5.

Table 4.5. Quantitative analysis of three analyzed precipitates pointed as 1, 2 and 3 in Figure 4.9.

Atomic% of detected elements

Analyzed site
C Si V Cr Mn Fe Ni Cu Mo
Grain boundary 18.6 1.3 0.5 39.9 1.8 24.0 10.3 1.0 2.8
Twin boundary 15.0 1.8 0.9 29.6 — 30.4 18.2 2.1 2.0
Around inclusion 17.9 1.5 0.6 33.6 2.5 28.0 13.3 1.0 1.4

44
 a b

1
3 keV

5 μm
keV

Figure 4.9. (a): SEM microstructure of sample 5-V, aged at 700 ºC for 500 h (a): M23C6 precipitates
on grain boundary (1), twin boundary (2) and around the inclusion particle (3); EDS spectrum of
grain boundary carbide (b) and that of the matrix (c) at 20kV accelerating voltage.

Although the composition of M23C6 carbide is not exactly specified in the literature, its
[13, 30]
chromium content is often reported above 50 at% and thus, the above measured
chromium contents are lower than its common level in M23C6. Also, the atomic percentage of
carbon is lower than its level in M23C6 formula which is ( × 100) or ~20.7%. With respect
2

to the size of the precipitates in Figure 4.9, varying in the approximated range of 0.5–1.5 μm
(larger dimension), some interference between the interaction volume (~1 μm spatial
resolution at 20 kV, section 2.6) and the matrix can be expected when the precipitates were
being analyzed. Therefore, the resulted analysis for precipitates could have slightly shifted to
that of the matrix in different degrees depending on the size of the particle. Moreover, matrix
effect which is considerable for low energy X-rays of carbon, may result in lower analyzed
carbon content than its actual concentration. So these precipitates can be most probably
M23C6 carbides.

The same analysis on other samples aged at 700 ºC suggested that M23C6 carbide was the
dominant precipitate, mostly forming on grain/twin boundaries. M23C6 was also found in
higher aging temperatures (samples 5-V and 7-Nb) but in less number. Table 4.6 gives the
obtained compositional ranges for precipitates suggested being M23C6 carbides in all samples
with the number of observed precipitates over the total analyzed particles. All precipitates

45
were located on grain and twin boundaries and only the important constituent elements of
M23C6 carbide (C, Si, Cr, Fe, Ni and Mo) are given in the table.

Table 4.6. Compositional ranges of the precipitates suggested being M23C6 carbides with the number
of observed precipitates over total number of analyzed particles

Atomic % range Number/

Sample total
C Si Cr Fe Ni Mo analyzed

5-V at 700 ºC 14.6-32.0 0.9-2.5 28.4-40.5 24.0-42.9 10.3-19.6 1.4-2.8 31/55

700
7-NB ºC ºC
at 700 13.9-28.4 0.8-2.3 26.6-39.2 28.3-41.0 12.2-20.4 1.0-3.1 26/36
700atºC700 ºC
10-VNb 15.5-25.4 0.6-2.0 29.7-41.2 27.5-40.9 12.3-18.1 1.3-2.1 18/42
11-VNbN at 700 ºC 16.8-24.0 1.1-1.9 28.9-42.5 29.3-38.7 11.6-17.5 1.3-2.7 23/44
5-V at 800 ºC 19.4-30.9 1.5-3.0 32.0-46.6 19.7-24.3 9.4-12.2 2.8-4.5 5/33
7-Nb at 800 ºC 16.7-32.4 0.9-2.7 30.8-44.2 20.1-30.2 10.3-13.5 2.2-3.6 15/39
5-V at 850 ºC 18.9-26.8 1.7-3.8 31.7-48.6 14.0-23.7 8.3-10.9 1.9-3.9 5/31
7-Nb at 850 ºC 16.2-28.0 0.8-3.6 34.9-49.3 21.1-32.0 13.3-16.1 2.1-3.7 6/23

It can be seen that the fraction of M23C6 carbides in sample 5-V reduces significantly from
56% at 700 ºC to 15% at 800 ºC but remains almost constant at higher aging temperature
(16% at 850 ºC); meanwhile, in sample 7-Nb it exhibits a continuous reduction from 72% at
700 ºC to 38% at 800 ºC and 26% at 850 ºC. Precipitates with composition of M23C6 also
have high fraction in samples 10-VNb and 11-VNbN (43% and 52% respectively) suggesting
that M23C6 is the dominant precipitate at 700 ºC. At higher aging temperatures, though with
lower fraction, carbides have been enriched with chromium, silicon and molybdenum which
is in agreement with the literature [13]. When comparing samples 5-V and 7-Nb, the relatively
higher fraction of M23C6 carbides in sample 7-Nb can be addressed to the different effects of
stabilizing elements (vanadium and niobium). In sample 7-Nb incomplete dissolving of
niobium due to the existence of residual phases prior to the aging (discussed in section 4.1.2),
might have led to the insufficient stabilizing and consequently higher fraction of carbide.
Whereas, in sample 5-V with no residual phase (complete dissolving of vanadium prior to the
aging) higher degree of stabilizing has occurred resulting in lower level of free carbon atoms
and consequently lower fraction of M23C6 carbide.

4.4.2 Silicon-rich eta (η) phase

A group of carbides containing over 4 at% silicon and less carbon content than its level in
M23C6 were observed with relatively high fraction in all samples on the same sites as M23C6.

46
Figure 4.10 shows three of the analyzed precipitates of this type in sample 5-V, aged 500 h at
800 ºC and the corresponding EDS spectrum of one analyzed particle.

a b

M23C6

M23C6

4 μm keV

Figure 4.10. (a): SEM image of sample 5-V aged 500 h at 800 ºC, precipitates identified as eta (η)
phase in association with M23C6 carbides. (b): EDS spectrum of one eta phase having strong peaks of
Si, Mo and Cr.

In the above EDS spectrum, presence of carbon peak in association with relatively intense
silicon, chromium and molybdenum peaks suggests that the analyzed particle can be silicon-
containing carbide or eta phase (η). Similar precipitates were found in all samples with
compositional ranges given in Table 4.7.

Table 4.7. Compositional ranges of precipitates suggested being silicon-rich eta phase with the
number of observed precipitates over the total number of analyzed particles

Atomic % range Number/

Sample total
C Si Cr Fe Ni Mo analyzed

5-V at 700 ºC 4.4-16.3 3.8-15.7 26.1-41.7 13.5-20.5 9.4-19.2 3.0-7.6 14/55

700
7-NB ºC ºC
at 700 3.9-11.6 4.61-8.0 25.5-38.1 33.4-38.0 20.4-23.1 2.2-3.7 4/36
700atºC700 ºC
10-VNb 5.3-14.1 3.5-8.3 26.7-41.0 26.5-35.0 15.7-21.3 1.8-5.7 16/42
11-VNbN at 700 ºC 4.4-18.3 3.8-7.2 23.4-31.9 20.0-39.8 16.7-22.7 1.7-3.3 15/44
5-V at 800 ºC 5.9-21.7 4.7-20.4 26.1-46.7 13.3-30.7 9.7-20.1 3.8-7.9 15/33
7-Nb at 800 ºC 4.8-23.4 6.7-12.1 27.4-44.7 11.6-28.3 9.3-24.6 4.2-6.5 12/39
5-V at 850 ºC 5.0-28.0 4.2-19.7 30.3-40.6 10.7-24.9 9.6-22.2 3.7-8.6 17/31
7-Nb at 850 ºC 6.4 4.6 28.5 37.6 19.4 2.8 1/23

As it can be seen in Table 4.7, the composition of precipitates has large variation which is
reported for eta phase in the literature [13, 28-30]. In the general formula of M6C, carbon content
is supposed to be 1/7 or 14.3 at%, whereas in the above compositions it has large variation

47
from 4 at% to even 28 at% (in sample 5-V at 850 ºC). But as mentioned earlier, EDS has low
accuracy in analysis of light elements giving either lower content than its actual amount due
to the matrix effect or higher because of surface contamination. With respect to the silicon
content in excess of 4 at% and presence of carbon, regardless of its exact content, it can be
suggested that the current precipitates are silicon-rich eta phase. The relatively high
molybdenum content is also an evidence for eta phase due to the frequently observation of
[13, 28-30]
molybdenum containing eta phase . Also, in samples 5-V and 7-Nb the average
content of silicon, molybdenum and chromium as well as the fraction of eta phase increases
with aging temperature which is in agreement with the literature [13, 28]; except in sample 7-Nb
at 850 ºC where, only one eta phase precipitate was found (in section 4.4.4 another type of
carbide as the dominant precipitate in this sample is discussed). So, silicon-rich eta phase
dominates precipitation at high aging temperatures by a fraction of 45% and 55% for sample
5-V aged at 800 ºC and 850 ºC respectively and also 31% for sample 7-Nb at 800 ºC.

An unclear feature of the above observed eta phase is the absence of nitrogen. In the literature
[28-30]
, nitrogen is reported to be eta phase stabilizer with general formula M5SiN, but in the
current analyses there was no sign of nitrogen; even though its content was relatively high in
the alloys. This can be due to the low accuracy of high voltage EDS for detecting nitrogen.
Also, accurate methods for quantitative measurement of carbon and nitrogen such as electron
energy loss spectroscopy are required to determine the share of these two elements in
composition of the observed eta phase.

4.4.3 G-phase

Another silicon-rich phase was observed having no signal of carbon in its EDS spectrum. The
silicon content of this phase was fairly high, but in many cases lower than the specified level
for G phase in the literature (24 at%). Since this phase was free of carbon with high silicon
content, it may be possible to assume that it is G phase in its early stages of formation or a
transient phase from earlier formed precipitates to G phase. Figure 4.11 shows SEM images
of samples 5-V and 7-Nb aged at 850 and 800 ºC respectively for 500 h where precipitates
pointed as “G” are suggested to be G phase. In the given EDS spectra, strong silicon peaks in
the absence of carbon signal accompanying with intense molybdenum and chromium peaks
show a silicon-rich composition which is not carbide. In spectrum (d) for sample 7-Nb, there
[31]
is also intense peak of niobium Lα. This composition can be found in the literature as G
phase which forms from the initial NbC precipitate on grain boundaries, as a preferential site
for G phase in 20-25 alloys.

48
 a b
η

M23C6

3 μm
keV

c d

4 μm
M23C6
keV

Figure 4.11. SEM images of (a): sample 5-V aged 500 h at 850 ºC and (c): sample 7-Nb aged 500 h at
800 ºC. Precipitates pointed as “G” are suggested to be G phase in association with M23C6 and eta
carbides. (b) and (d): corresponding EDS spectra of “G” particles with no carbon signal, but strong
peaks of silicon, molybdenum and chromium. In (d) niobium Lα is also present.

Table 4.8 gives the compositional ranges of precipitates suggested to be G phase and the
number of observed particle of this type over the total analyzed precipitates. In sample 7-Nb
only at 800 ºC aging temperature this type of precipitate was found.

Table 4.8. Compositional ranges of precipitates suggested being G phase with the number of observed
precipitates over the total number of analyzed particles

Atomic % range Number/

Sample total
Si Cr Fe Ni Mo Other analyzed

5-V at 700 ºC 14.7 30.9 25.6 21.0 4.1 2.2 V 1/55

700atºC700 ºC
10-VNb 11.3-13.1 26.7-34.6 26.5-31.0 20.7-26.3 2.8-3.9 ─ 3/42
11-VNbN at 700 ºC 10.0-13.8 25.3-29.6 28.0-33.8 21.4-26.5 3.0-3.3 1.7-2.4 V 2/44
5-V at 800 ºC 13.0-18.1 27.2-40.1 12.1-20.8 19.1-27.7 4.1-7.0 2.8-6.0 V 6/33
7-Nb at 800 ºC 11.8-17.7 24.0-39.0 13.0-29.6 20.8-27.9 5.6-9.6 0.0-7.7 Nb 4/39
5-V at 850 ºC 16.0-23.2 28.4-40.3 10.8-26.2 19.9-28.1 5.5-9.3 5.0-8.8 V 9/31

49
The volume fraction of G phase has increased in sample 5-V with increasing the aging
temperature accompanying with higher contents of silicon. This can support the above
assumption that this phase can be a transient phase to from G phase; because higher aging
temperature has facilitated the diffusion of atoms (especially silicon) and thus, a closer
composition to the reported G phase in the literature (24 at% silicon) has been obtained.

4.4.4 High-carbon carbide

A carbide with extremely high carbon content and relatively high chromium was found in
sample 7-Nb aged at 850 ºC. Figure 4.12 shows a number of analyzed precipitates of this type
and a sample of corresponding EDS spectrum.

a b
C

M23C6
4 μm keV

Figure 4.12. (a): SEM image of sample 7-Nb aged for 500 h at 850 ºC, analyzed precipitates pointed
as C contained high chromium and carbon contents. One M23C6 precipitate was also identified; (b):
EDS spectrum of a precipitate of type C with intense chromium and carbon peaks.

The EDS quantitative analysis of these carbides gave the compositional range as in Table 4.9.
The carbon content is higher than that in common carbides (e.g. M23C6 and M6C); while
chromium content has large variation with relatively high upper limit.

Table 4.9. Composition range of high-carbon carbides in sample 7-Nb aged 500 h at 850 ºC

Atomic % range Number/

total
C Si Cr Fe Ni Mo Nb analyzed

26.3-40.1 0.0-0.8 39.5-61.0 9.4-27.9 2.9-10.4 2.1-4.0 0.0-2.3 14/23

High-carbon carbides such as M7C3 and M5C2 are not probable to form in the current alloy
since they usually need very high concentration of carbon, e.g. in carburized steel. One
suggestion is that the observed carbides are M23C6 where extremely high carbon content is
due to the effect of contamination on the EDS analysis. The other probable case is that the
niobium-containing carbides are M23C6 which has formed around the initial NbC particles on

50
grain boundaries; so that higher carbon content in association with some niobium have been
detected by EDS.

4.4.5 Sigma phase

Very few particles situated on grain boundaries of sample 7-Nb aged at 700 and 800 ºC were
found to have a sigma phase-like chemical analyses (Table 4.10); even though the aging time
was limited to 500 h. The main evidences supporting the suggestion of sigma phase were the
absence of carbon, low silicon content and some manganese segregation into the particles.

Table 4.10. Chemical composition range of precipitates suggested to be sigma phase; observed on
grain boundaries of sample 7-Nb aged at 700 and 800 ºC with number of particles of this type over the
total analyzed precipitates

Aging Atomic % ranges Number/ total

temperature analyzed
Si Cr Mn Fe Ni Mo
700 ºC 2.0 31.8 1.6 44.5 15.3 3.1 1/36
800 ºC 1.7-2.1 36.0-41.5 1.5-2.0 37.6-42.5 11.9-15.1 3.8-5.9 3/39

With respect to the fact that sigma phase forms in long-term aging and also very few particles
of this type were observed, it can be argued that formation of this phase has been limited to
few sites with local equilibrium condition for precipitation of sigma phase.

4.4.6 Intragranular sub-micron precipitates

The intragranular sub-micron precipitates were firstly analyzed by EDS at ordinary 20 kV

voltage. One analyzed site and resulted EDS spectra are shown in Figure 4.13 with the
corresponding quantitative analysis as given in Table 4.11. It can be seen that some
segregation of niobium and chromium has occurred from the matrix to the precipitates (also
very little for vanadium and silicon); but this compositional information is not enough to be
able to identify the precipitates.

Table 4.11. Quantitative analysis (atomic%) of two sub-micron particles in sample 11-VNbN aged
500 h at 700 ºC pointed as 1 and 2 in the SEM image of Figure 4.13 and that of the matrix

Site Si V Cr Fe Ni Cu Nb Mo Other
1 0.9 1.1 25.5 43.9 21.3 2.9 4.2 — —
2 1.1 1.2 23.5 44.4 22.8 2.8 2.2 1.5 0.5 Mn
Matrix 0.7 0.8 21.7 47.7 24.3 2.9 — 1.3 0.7 Mn

The minor compositional difference between particles and the matrix is due to the
interference of the relatively large interaction volume at 20 kV voltage (~1 μm) with matrix

51
when fine particles are being analyzed (~160 nm and ~100 nm in this case). Also, no light
element (e.g. carbon and nitrogen) were detected although they are expected to be present in
the sub-micron precipitates. This is due to the matrix effect in absorption of low energy X-
rays of light elements having long escape path. Because of these uncertainties a lower voltage
(5 kV) was used to identify sub-micron precipitates.

a b

200 nm keV

c d

keV keV

Figure 4.13. (a): sub-micron precipitates in sample 11-VNbN aged 500 h at 700 ºC; EDS spectra of
(b): matrix, (c): precipitate 1 contains niobium Lα peak but neither carbon nor nitrogen and (d):
precipitate 2 with less intense niobium Lα peak but again no carbon or nitrogen.

4.5 Low voltage EDS for identification of sub-micron precipitates

In order to characterize the sub-micron precipitates, low voltage EDS was carried out to have
adequate spatial resolution relative to the size of particles. The same particles which were
found impossible to be identified at 20 kV (Figure 4.13, section 4.4.6) were re-analyzed, this
time using 5 kV accelerating voltage. The resulted EDS spectra, as can be seen in Figure
4.14, showed a substantial increase in the information taken from the particles compare to
that at 20 kV. The shape of EDS spectra changes significantly from the matrix to the
precipitates: the intensity of iron and nickel Lα peaks decreases by several orders of

52
magnitude, strong niobium Lα peaks appear, as well as vanadium Lα peak (broadening of
chromium peak especially in spectrum c) and also Kα peaks of boron, carbon and nitrogen
are quite resolvable.

a b c

keV keV keV

Figure 4.14. EDS spectra of sample 11-VNbN aged 500 h at 700 ºC using 5 kV accelerating voltage;
(a): matrix, (b) and (c): sub-micron precipitates pointed as 1 and 2 respectively in Figure 4.13 where
intense niobium Lα peak as well as K lines of light elements (carbon, boron and nitrogen) are present.

The corresponding quantitative analysis in Table 4.12 shows that the segregation of alloying
elements (e.g. niobium and vanadium) from matrix to the precipitates is several times higher
than that in Table 4.11 for the high voltage analysis. Also, enrichment by carbon and
presence of nitrogen and boron are new information from the particles that reveals their full
composition showing that both are carbonitrides with different contents of niobium,
vanadium, chromium, iron, nickel and boron. However, in low voltage EDS some elements
with lower contents such as copper and molybdenum have not been detected and also a
carbon peak in matrix spectrum is always present due to the contamination effect.

Table 4.12. EDS quantitative analysis (wt%) of sub-micron precipitates in samples 11-VNbN aged
500 h at 700 ºC marked as 1 and 2 in Figure 4.13 and that of the matrix

Site C N Si V Cr Fe Ni Cu Nb Mo Other
1 4.9 5.1 — 5.7 31.1 15.7 5.6 — 29.0 — 2.8 B
2 5.4 1.8 — 8.8 25.8 32.2 17.1 — 8.9 — —
Matrix 1.6 — — — 21.9 49.8 26.8 — — — —

From the above comparison between high and low voltage EDS analyses of the same site,
low voltage EDS has lots of advantages over the high voltage in qualitative analysis by
improving the spatial resolution and also revealing of the light elements. However, the
quantitative analysis of low voltage EDS is controversial due to some problems such as a
constantly present carbon peak in all spectra and no detection of other alloying elements such
as copper, molybdenum and silicon.

53
The drawbacks of low voltage EDS are discussed later in detail (section 4.7.2), but here it
should be mentioned that low voltage EDS had a poor accuracy in quantitative analysis and
thus, for identification of sub-micron precipitates the qualitative analysis has only been used.
Indeed, by comparing the EDS spectra of precipitates and that of the matrix the segregation
of alloying elements from the matrix to the precipitates has been observed and used as the
main indication for phase identification. In section 4.7 some aspects of low voltage EDS
quantitative analysis are also discussed. Following types of sub-micron precipitates were
observed in the aged samples by EDS microanalysis at 5 kV accelerating voltage.

4.5.1 Residual phase sub-micron particles

As mentioned earlier (section 4.1.2), residual phase sub-micron particles were observed in the
microstructures of all niobium-containing samples after solution annealing, having no
particular formation site. It was necessary to identify the residual particles to avoid confusing
them with precipitates formed under aging. For this purpose a number of sub-micron residual
particles were analyzed in the unaged state of niobium-containing samples. The result
showed that the majority of residual particles were niobium-rich nitrides. Figure 4.15 shows
two analyzed residual phase sub-micron particles and their corresponding EDS spectra.

2 1

a 200 nm

b c d

keV keV keV

Figure 4.15. (a): SEM image of two sub-micron residual particles in solution annealed state of sample
11-VNbN; EDS spectra of (b): matrix, (c) and (d): particles 1 and 2, respectively with intense niobium
Lα peak as well as K lines of boron and nitrogen. EDS was at 5 kV accelerating voltage.

54
It can be seen in the SEM image that the residual sub-micron particles have well-defined
shapes (close to cuboidal). The spectra show that almost no carbon enrichment has occurred
in the particles; unlike niobium, nitrogen, boron and vanadium which have highly segregated
from the matrix to the particles. The quantitative analysis (Table 4.13) verifies this niobium-
rich nitro-boride composition.

Table 4.13. EDS quantitative analysis (wt%) of sub-micron residual particles in sample 11-VNbN
denoted as 1 and 2 in Figure 4.15 and that of the matrix

Site B C N Si V Cr Fe Ni Cu Nb Mo
1 3.2 1.5 9.0 — 7.1 32.6 7.9 2.5 — 36.2 —
2 6.2 1.9 8.0 0.3 10.2 31.5 6.2 1.0 — 34.7 —
Matrix — 1.3 — — 1.5 23.4 45.9 24.2 1.9 — 1.8

The composition of matrix is not accurate (e.g. the presence of carbon and relatively high
vanadium and chromium contents; discussed later in section 4.6), but its difference from the
composition of particles suggests that no silicon, copper or molybdenum and very low
(almost negligible) amounts of carbon and nickel are present in the sub-micron residual
particles. Almost all analyzed particles of this type in other niobium-containing samples had
similar morphology, EDS spectrum and chemical analysis as given above. This can be a
guideline when analyzing the sub-micron precipitates in aged samples to indicate whether a
particle is residual or has formed during aging. For instance, particle 1 in Table 4.12 has a
close composition to that of the above residual particles suggesting that it could be a residual
particle which has dissolved some carbon and also some substituted iron and nickel during
aging.

4.5.2 Niobium-rich carbonitride and nitride

A high number of niobium-rich carbonitride precipitates were present in all aged niobium-
containing samples. Figure 4.16 shows three analyzed sub-micron precipitates in sample 7-
Nb, aged 500 h at 700 ºC using 5 kV voltage and their corresponding EDS spectra.
Segregation of niobium, carbon and nitrogen from matrix to the particles can be seen when
comparing their spectra. It can be inferred from the peak intensities that particle 3 has the
highest contents of niobium, nitrogen and carbon. The carbonitride composition in
association with well-defined cuboidal shape is similar to the characteristics of Nb(CN)
[13]
precipitates ; however, this is just a morphological and compositional similarity. For a
certain judgment whether the analyzed particles are Nb(CN), it is also required to determine
their exact chemical formula and also crystallographic structure. The latter was not in the

55
scope of this work and the exact formula was impossible due to poor accuracy of low voltage
EDS quantitative analysis.

a b
1

200 nm
keV

c d e

keV keV keV

Figure 4.16. (a): sub-micron precipitates in sample 7-Nb aged 500 h at 700 ºC. (b): EDS spectrum of
the matrix giving Fe, Cr, Ni and C peaks; (c), (d) and (e): EDS spectra of precipitates 1, 2 and 3,
respectively all giving niobium Lα peak and also carbon and nitrogen K peaks. Boron peak intensifies
gradually from spectrum (c) to (e).

The same analysis was carried out on sample 7-Nb aged at higher temperatures. At 800 ºC
much less number of shapeless particles were observed; instead, high number of cuboidal
particles was present with nearly similar compositions to those at 700 ºC (containing both
carbon and nitrogen). In contrast, sample aged at 850 ºC had almost no shapeless particle
where the majority of observed precipitates were cuboidal. For the latter sample, as it can be
seen in Figure 4.17, in the EDS spectra of precipitates 4 and 5 (typical for all cuboidal
particles) the intensity of carbon peak remains constant from matrix to the precipitates
suggesting that the particles didn’t contain carbon anymore; whereas nitrogen has resolvable
intense peak. Also substituted chromium, iron and nickel and dissolved boron are still
present. These changes in the morphology and composition of niobium-rich precipitates from
700 ºC to 800 and 850 ºC aging temperatures show that they mostly have carbonitride

56
composition below around 800 ºC, but at higher aging temperatures, e.g. 850 ºC, niobium-
rich nitride is the preferential phase forming with sharply defined cuboidal shape.

4 a b

200 nm keV

c d

keV keV

Figure 4.17. (a): Cuboidal sub-micron precipitates in the SEM image of sample 7-Nb aged 500 h at
850 ºC; EDS spectra of (b): matrix, (c) and (d): precipitates 4 and 5, respectively. The intensity of
carbon peak in the spectra of precipitates is almost equal to that of the matrix.

Furthermore, with respect to the similar EDS spectra and morphology of these niobium-rich
precipitates and that of the sub-micron residual particles (section 4.5.1) it can be suggested
that they could have formed under similar condition. Also, due to very few number of
observed sub-micron residual particles in the solution annealed state, it can be added that the
upper temperature limit for the stability of cuboidal niobium-rich precipitates is close to the
selected temperature for solution annealing, i.e. 1150 ºC.

4.5.3 Vanadium-rich carbonitride and nitride

In sample 5-V, with the same chemical composition as 7-Nb but containing vanadium instead
of niobium, various vanadium-rich precipitates were found mostly with carbonitride and

57
nitride compositions. Figure 4.18 shows an analyzed site in sample 5-V aged at 700 ºC with
two analyzed sub-micron precipitates and the corresponding EDS spectra.

a b
1

300 nm keV

c d

keV keV

Figure 4.18. (a): SEM image of sample 5-V aged 500 h at 700 ºC, sub-micron precipitates with no well-
defined shapes; EDS spectra of (b): matrix; (c), (d): precipitates 1 and 2 respectively, with vanadium Lα
peak. In (c), compared to (b), the nitrogen peak is a bit stronger than the carbon peak and vice versa.

At higher aging temperatures cuboidal precipitates were observed in higher number, mostly
had a vanadium nitride composition with very little substitution of other elements. Figures
4.19 shows two sites in sample 5-V aged at 800 ºC for 500 h analyzed with different EDS
setups and the corresponding spectra of analyzed particles.

The aligned precipitates in micrograph (a) of Figure 4.19 are clearly cuboidal. In spectrum
(c), which was typical for all precipitates, neither chromium nor carbon was detected by the
EDS analyzer whereas vanadium and nitrogen are the most intense peaks. This suggests that
precipitates are vanadium nitride with little substituted iron and nickel. Almost the same
result observed for precipitates 2 and 3 in micrograph (d); but with little dissolved carbon
(spectra f and g). However, very few particles had a carbonitride composition like spectrum
(h) where carbon and chromium peaks and a small, but resolvable nitrogen peak are present.
These particles were mostly unorientated, shapeless and larger than the cuboidal ones.

58
 a b c

200 nm keV keV

d e f

2
keV keV

g
h

300 nm

keV keV

Figure 4.19. (a) and (d): SEM images of sample 5-V aged 500 h at 800 ºC; (b) and (e): EDS spectrum
of the matrix with carbon background; (c), (f) and (g): spectra of precipitates 1, 2 and 3, respectively
giving intense vanadium and nitrogen peaks where the carbon peak is almost equivalent to the carbon
background level in the matrix; (h): spectrum of particle 4 showing the carbonitride composition.

All above observations can propose that the dominant precipitate at 800 ºC aging temperature
of sample 5-V was vanadium nitride with a characteristic cuboidal shape containing almost
no dissolved carbon (if any, very little) and much lower amounts of substituted elements like
iron, nickel and chromium compared to the vanadium carbonitrides at 700 ºC. This type of
precipitate was also frequently found in sample aged at 850 ºC (Figure 4.20).
As it is shown in Figure 4.20, EDS spectra of precipitates 1 and 2 shows no carbon
enrichment and thus they are vanadium nitrides with less amount of substituted elements than
those in 800 ºC. Also, particle 1 with cuboidal shape seems to be purer than 2 which show the
characteristic morphology of vanadium nitride at high temperature. Since the observed
precipitates of this type were mostly aligned towards dislocation lying directions, then

59
increasing the density of dislocations (e.g. by cold work prior to aging) provides more
preferential sites for nucleation of stable precipitates and thus higher strengthening effect.

a
b

2
200 nm keV

c d e

keV keV keV

Figure 4.20. (a): SEM image of sample 5-V aged 500 h at 850 ºC, from aligned precipitates 1 is
cuboidal; (b): matrix spectrum with carbon background; (c), (d) EDS spectra of precipitates 1, 2
respectively with no intensified carbon peak showing vanadium nitride composition, (e): spectrum of
precipitate 3 shows some carbon enrichment and thus a carbonitride composition.

4.5.4 Copper-rich precipitates

One main goal of the present work was finding the probable copper-rich nano-scale
precipitates. From over 1000 spectra taken by low voltage EDS very few analyses (≤10)
showed the presence of copper in sub-micron precipitates. Figure 4.21 shows one of the
detected copper-rich precipitates in sample 5-V aged for 500 h at 700 ºC, where the particle is
analyzed from two points resulting in the EDS spectra (d) and (e) which show copper
enrichment. The copper-rich particle sizing around 100 nm is located among the aligned
precipitates all with spectrum similar to (c) which was discussed earlier as vanadium-rich
carbonitride. Also, no morphological difference can be seen between the copper-rich particle
and the neighboring carbonitrides; which was a common feature for all detected copper-rich
particles in other samples. By contrast, the copper-rich nano-scale precipitates in other studies

60
of 20-25 alloy have been observed with characteristic morphologies; for instance, round-
shape dispersed particles [21] or cluster-shape nucleating on dislocations [28], both sizing below
50 nm even after 30 kh aging time [21].

a b c

keV keV
3
1 d e

200 nm

keV keV

Figure 4.21. (a): Aligned precipitates in SEM image of sample 5-V aged at 700 ºC for 500 h. (b):
matrix EDS spectrum; (c): spectrum of precipitate 1 showing a typical vanadium-rich carbonitride;
(d) and (e): EDS spectra of copper-rich precipitate taken from two points (2 and 3) having intense
copper Lα peak at ~0.93 keV just right to the nickel Lα

This dissimilarity in combination with statistically very low number of observed copper-rich
precipitates (≤1%) suggests that the current precipitates have formed under a restricted
localized equilibrium condition rather than a regular homogenous one. Therefore, it can be
argued that sub-micron copper-rich precipitates either have not formed under the applied
aging condition or have formed but with no detectable signals in their EDS analysis. The
latter can be more probable when considering the reported size for precipitates of this type in
[21, 28]
the literature to be below 50 nm which is at least two times smaller than the spatial
resolution of the analysis at 5 kV voltage (~100 nm) and thus, particles become so difficult to
be distinguished from the matrix.

4.6 Limitations of low voltage EDS microanalysis

Although using 5 kV voltage for EDS microanalysis provided better spatial resolution to
identify sub-micron precipitates, there were limitations which affect the quality and extent of
this application. Here the main problems are discussed.

61
Peak overlap

Austenitic stainless steel has a compact region of L lines of its main matrix elements (iron,
chromium and nickel) below 1 keV in the EDS spectrum. For alloys of the present study this
region became more compact by L lines of alloying elements such as manganese, vanadium
and copper, all in the range 0.5–1 keV. Since in low voltage EDS (e.g. 5 kV) all mentioned
elements were detected on the reliance of their L lines, this high number of elements with
close L lines decreased the accuracy of analysis due to inevitable peak overlaps. Table 4.14
gives the characteristic L line energies of the present metallic elements in samples 10-VNb
and 11-VNbN.

Table 4.14. Photon energies of X-ray characteristic L lines for vanadium, chromium, manganese, iron,
nickel, copper, niobium and molybdenum [47]

Element V Cr Mn Fe Ni Cu Nb Mo

Lα1 energy (keV) 0,511 0,573 0,637 0,705 0,851 0,930 2,166 2,293

Lα2 energy (keV) 0,511 0,573 0,637 0,705 0,851 0,930 2,163 2,290

Lβ1 energy (keV) 0,519 0,583 0,649 0,719 0,869 0,950 2,257 2,395

Lβ2 energy (keV) — — — — — — 2,367 2,518

It was mentioned before that the energy resolution (FWHM) of common EDS analyzers is
around 100 eV for molybdenum Lα peak in the energy range below 3 keV. As it can be seen
in Table 4.14 the energy differences between L lines of the adjacent elements are all less than
100 eV except for Fe-Ni and Nb-Mo. Therefore, peak overlap is expected for the pairs V-Cr,
Cr- Mn, Mn- Fe and Ni-Cu in the EDS spectrum. This problem is more serious when the
content of elements in the analyzed region differs a lot; so that the element with low content
may be missed in the analysis since its low intensity characteristic peak can be entirely
hidden under the broadened peaks of the adjacent elements with much higher compositional
contents.

Low peak intensities and peak-to-background ratio

X-ray generation is reduced by decreasing the beam energy; so for a given acquisition time
lower accelerating voltage gives lower intensity peaks in the EDS spectrum. This
consequently results in lower peak-to-background ratio (P/B) which increases the detection
limit. Therefore, low-concentrated elements may be missed since their characteristic peaks
may have too low intensity from being resolved from the background; even out of the above
mentioned compact region, e.g. silicon (Kα=1.74 keV) and molybdenum (Lα=2.29 keV). To

62
get high enough intensities to detect such elements several times longer acquisition time must
be used which contradicts the main advantage of EDS as a quick analysis.

High sensitivity to contaminations

Low voltage EDS is more sensitive to the specimen surface due to the smaller penetration
depth of the beam electrons. Therefore, surface contamination can strongly affect the analysis
by producing disturbing X-rays. Organic matters as the most common contamination are
polymerized on the surface by beam bombardment. The main product is usually carbon
which causes an increase in carbon Kα count [38].

The above problems are demonstrated in Figure 4.22 showing an analyzed precipitate in
sample 10-VNb aged 500 h at 700 ºC and the resulted EDS spectra at 5 kV voltage.

150 nm

b c

VLα

NKα

R1
Si Mo R2
Mn Cu

keV keV

Figure 4.22. (a): SEM image of sample 10-VNb aged 500 h at 700 ºC, the analyzed precipitate is
pointed; (b): EDS spectrum of the matrix, R1 shows the energy resolution of chromium Lα, position
of non-detected alloying elements are pointed on the energy axis; (c): EDS spectrum of the
precipitate, R2 is the chromium Lα peak resolution, locus of vanadium and nitrogen peaks are marked.

It can be seen that the width of chromium Lα peak at its half maximum increases from the
matrix spectrum (R1) to that of the precipitate (R2) due to the presence of vanadium Lα (511

63
eV) just behind the chromium Lα peak (573 eV); but EDS analyzer has not detected
vanadium due to the severe peak overlap and thus, proper deconvolution must be done to
avoid missing of vanadium. The other problem is no detection of alloying elements such as
manganese, copper, silicon and molybdenum. Manganese and copper had no chance to be
detected since their characteristic L lines were always buried under the peaks of their adjacent
elements (Cr, Fe and Ni) with much higher contents. Also, silicon Kα and molybdenum Lα
had not enough intensity to be distinguished from the background. The same problem can be
observed for nitrogen peak in the precipitate spectrum due to its low P/B ratio; however, a
resolvable intensity rise at the locus of nitrogen Kα (391.4 eV) from ~35 counts in the matrix
spectrum to ~75 counts in that of the precipitate shows nitrogen segregation. Regarding
carbon Kα, it has been detected in both spectra with an intensity increase from ~70 counts in
matrixes to ~115 counts in particle; however, carbon peak in the matrix spectrum produced
by contamination is a major problem in quantitative analysis.

4.7 Accuracy of low voltage EDS

The above discussion shows a poor accuracy for low voltage EDS. The qualitative analysis
has missed low concentrated elements in the matrix, e.g. molybdenum at 1.5 wt% and copper
at 2.8 wt%. Also, nitrogen peak has been ignored in the precipitate spectrum due to its low
P/B ratio. Subsequently, the quantitative analysis had low accuracy even after deconvolution
of overlapped peaks. Table 4.15 gives the corresponding quantitative analysis of the above
discussed spectra with and without deconvolution.

Table 4.15. Quantitative analysis (wt%) of the spectra in Figure 4.22 before and after deconvolution
of overlapped peaks; the term “corrected” means after deconvolution

Site C N Si V Cr Mn Fe Ni Cu Nb Mo
Matrix 1.0 — — — 24.4 — 50.3 24.3 — — —
Precipitate 2.7 — — — 27.0 — 39.6 20.6 — 10.1 —
Corrected matrix 0.9 — 0.3 2.0 22.8 0.5 46.4 23.7 2.0 — 1.3
Corrected precipitate 2.8 1.9 — 7.4 25.6 0.3 32.5 16.6 1.3 8.8 2.8
Matrix at 20 kV — — 0.6 0.5 21.1 0.7 47.9 24.6 3.0 — 1.5

The matrix analysis taken from high voltage EDS (20 kV) is also given as a relatively
accurate one as reference. When comparing the initial matrix composition at low voltage with
that at 20 kV (as reference) the high degree of deviation can be seen. By deconvolution of
overlapped peaks to add the elements which are known to be present in the matrix, although
the composition is slightly corrected towards the reference (e.g. the contents of copper,

64
molybdenum and silicon) but still has low accuracy. For instance, the given vanadium
content (2.0 wt%) is obviously wrong which can be due to the overlap of vanadium and
chromium Lα peaks. For precipitate, the initial analysis shows niobium, chromium and
carbon enrichment; whereas deconvolution reveals also segregation of vanadium, nitrogen
and molybdenum to the particle. As discussed earlier, the P/B ratio and subsequently the
accuracy of EDS analysis may be improved at a given voltage by increasing the acquisition
time. This was investigated by increasing the acquisition time from 25 s in the last analysis to
100 s for analysis of other carbonitride precipitates in sample 10-VNb. Figure 4.23 shows the
resulted EDS spectra in which more characteristic peaks than before have been automatically
detected (e.g. N, V and Si) which means higher P/B ratio.

a b c

keV keV keV

Figure 4.23. EDS spectra of sample 10-VNb aged 500 h at 700 ºC analyzed at 100 sec acquisition
time; (a) and (b): carbonitride sub-micron precipitates; (c): matrix; more characteristic lines are
automatically detected (e.g. nitrogen, silicon and vanadium).

Although the present peaks in the above spectra are not necessarily more resolvable than
those in shorter acquisition time (Figure 4.22-c), for instance vanadium Lα from chromium
peak, because of higher P/B ratio the analyzer has detected more number of characteristic
lines without need of deconvolution; which means a more accurate analysis can be obtained.
However, in longer acquisition time the effect of contamination in producing the background
carbon peak was more serious. As it can be seen in the matrix spectrum (c) the relative
intensity of carbon peak is much higher than that in the shorter acquisition time (Figure 4.22-
b) which can be explained by more polymerization of contaminations in longer beam
bombardment (more carbon deposition on the surface). However, a part of the intense carbon
peak in spectrum (c) is due to the sequence of analysis, where it was taken after spectra (a)
and (b) and thus, some deposited carbon from the previous analyses had been present on the
surface prior to the analysis of matrix.

65
In order to estimate the accuracy of low voltage EDS, the resulted quantitative analysis was
compared with a full chemical analysis carried out by the Outokumpu steel- work laboratory.
For this purpose, the solution annealed state of sample 5-V was used since it contained almost
no residual phase and thus, had the most homogeneous composition. Table 4.16 shows the
resulted average quantitative analysis of sample 5-V by low voltage EDS and that given by
the laboratory. Light elements (carbon and nitrogen) were skipped due to the contamination
effect on the low-energy region of EDS spectrum. The error (in %) of measured content for
each element was calculated as a measure of accuracy using equation 4.1:

C
er en rror ( ur )= × 100 . .1

Where CEDS is the content of the element given by low voltage EDS quantitative analysis and
CLab is that given by the laboratory.

Table 4.16. Average quantitative analysis (wt%) of sample 10-VNb in solution annealed state, the
given analysis by the laboratory and accuracy of each measured content

Element Si V Cr Mn Fe Ni Cu Mo
Low voltage EDS 0.5 1.3 20.6 1.9 47.2 24.7 2.6 1.3
Laboratory 0.7 0.4 20.2 1.0 48.1 25.2 2.8 1.5
Accuracy (%) -29 220 2 88 -2 -2 -10 -17

In comparison with the common accuracy of ±1% for EDS quantitative analysis at ordinary
[37, 38]
voltages it can be seen that the accuracy of EDS at 5 kV can be reasonable only for
iron, chromium and nickel (main matrix elements) which is in the range ±2%. But the
accuracy is quite poor for the measured contents of other elements; either higher than the
accepted value due to peak overlap with high-content elements (e.g. for vanadium and
manganese) or lower than the accepted value because of the low P/B ratio (e.g. for silicon and
molybdenum) or lower than the accepted value, although there is peak overlap (for copper).

4.8 Determining the spatial resolution of low voltage EDS

The minimum particle size which is possible to be analyzed is an important feature of each
analyzing system; particularly for SEM-EDS it is the main parameter determining the spatial
resolution. Regardless of the theoretical way to estimate the minimum analysable particle size
(e.g. by determining the size of the interaction volume) a practical method was used to
determine this size and from that, evaluate the spatial resolution of the instrument used in the
present work. For this purpose, a set of analyses was done on the precipitates sizing from a

66
few hundreds of nanometers to below one hundred nanometer at 5 kV accelerating voltage.
Figure 4.24 shows the SEM image of the analyzed particles in sample 7-Nb aged 500 h at 700
ºC and their corresponding EDS spectra. The precipitates are denoted with numbers 1 to 5 by
the order of their size, starting from ~200 nm for particle 1 down to ~60 nm for number 5.
The auto-identification mode of the analyzer was used to have a common criterion for all
spectra.

5
3
4

200 nm a

b c d

keV keV keV

e f g

keV keV keV

Figure 4.24. (a): SEM image of sample 7-Nb aged 500 h at 700 ºC with analyzed sub-micron
precipitates sizing from ~200 nm (no. 1) to ~60 nm (no. 5); EDS spectra of (b):1, (c):2, (d):3, (e):4,
(f):5 and (g):matrix. The last auto-identification of niobium Lα line occurred in (e).

67
It can be seen that niobium Lα peak was automatically identified until precipitate 4 (spectrum
e) sizing around 100 nm in the SEM image; however, in spectrum (f) at the locus of niobium
Lα, peak is still distinguishable from the background (pointed by red arrow) but has not
enough P/B ratio to be detected. Therefore, the minimum analyzable particle size can be
suggested to be around 100 nm for the current instrument at 5 kV accelerating voltage. This
amount has good agreement with the estimated size of the interaction volume as 11090 nm
(section 2.6). The same experiments on samples 10-VNb and 11-VNbN gave almost the same
result. Niobium containing precipitates were analyzed since niobium Lα line (2.167 keV) is
clearly resolvable compared to that of the vanadium.
[43, 54]
In the literature , niobium detection in the EDS spectra of niobium-rich precipitates has
been frequently used as a criterion for determining the spatial resolution of low voltage FE-
SEM, where the minimum precipitate size in which the niobium Lα peak is detectable is
recognized as the special resolution. Therefore, the above resulted minimum analyzable size
can be suggested as the estimated spatial resolution of the FE-SEM used in the present work
at 5 kV accelerating voltage. However, in this definition only the qualitative analysis is
considered; i.e. distinguishing of the particle from the matrix has been defined as the criterion
for indicating the spatial resolution of the instrument.

68
 Chapter 5

5 SUMMARY AND CONCLUSION

The present work covered the whole process of sample preparation from melting to
metallography plus an extensive EDS microanalysis. Although the original goal was
precipitation study, low voltage EDS came to the focus from the middle of the work.

The precipitate observation in different samples aged for 500 h at 700, 800 and 850 ºC can be
summarized as Table 5.1.

Table 5.1. Summary of precipitate observation in different samples aged for 500 h at 700, 800 and
850 ºC with morphology, size and location of precipitates

solution Aging temperature (ºC)

Sample Morphology, size, location
annealed state 700 800 850
M23C6 (+) M23C6 M23C6 globular/string/plate shape,
η phase η phase (+) η phase (+) 0.5-1.5 µm,
5-V ─ G phase (-) G phase G phase grain/ twin boundaries
V(CN) (+) V(CN) V(CN) (-) cuboidal above 800 ºC, 100-200
─ V(N) V(N) (+) nm, intragranular on slip bands

Nb(BCN) Nb(BCN) Nb(BCN) Nb(BCN) residual phase, globular, >1 µm,

everywhere in the matrix
M23C6 (+) M23C6 M23C6 (+)
η phase η phase η phase (-) globular/chain shape,
7-Nb 0.5-1.5 µm,
─ G phase ─ grain/ twin boundaries
σ phase (-) σ phase (-) ─
Nb(BCN) Nb(BCN) Nb(BN) cuboidal, 100-200 nm,
Nb(CN) Nb(CN) Nb(N) intragranular on slip bands

Nb(BCN) Nb(BCN) residual phase, globular, >1 µm,

everywhere in the matrix

10-VNb M23C6 (+)

globular/chain shape, 0.5-1.5 µm,
and η phase
grain/ twin boundaries
11-VNbN G phase (-)
Nb(BCN) shapeless/cuboidal, 100-200 nm,
Nb(CN) intragranular on slip bands

(+): Phase observed in high fraction (≥ 40% of the total analysed particles of the same morphology)
(-): Phase observed in low fraction (≤ 10% of the total analysed particles of the same morphology)

69
Phase identification was done by comparing the obtained composition ranges from EDS and
that in the literature to find the closest phase to each composition range. Therefore, there was
always some degree of uncertainty in the phase identification. The micron-scale particles
were analyzed by EDS at 20 kV; whereas for intragranular sub-micron precipitates low
voltage EDS (at 5 kV) was used. In Table 5.1 precipitates with fraction of 40% and more are
considered as the dominant phase and indicated by (+). Samples 10-VNb and 11-VNbN were
available only for 700 ºC aging temperature and contained similar precipitates. Sigma phase
was observed only in sample 7-Nb with very low fraction (< 8%).

The intragranular sub-micron precipitates were niobium and vanadium-rich carbonitrides

with high contents of substituted matrix elements; locating on slip bands as the preferential
nucleation sites. With increasing the aging temperature from 700 ºC to 800 and 850 ºC
carbon content of these precipitates decreased and simultaneously their shapes became more
well-defined; so that in samples aged at 850 ºC most of the particles (especially vanadium-
rich phase) were cuboidal carbon- free nitrides. Very few copper-rich precipitates were found
in different samples (less than 1% of total analyzed particles) with no characteristic
morphology and formation site and thus, were considered as localized single precipitates and
were not put in the above summary table.

Besides the above precipitate observations, the following conclusions can be drawn from the
rest of the study:

1. The selected method for making mini-melts in the laboratory scale induction furnace had
low reproducibility with poor control over the melting, alloying and deoxidization
process. This increased the number of variables differing between samples such as alloy
homogeneity, solidification structure and inclusion content, all affecting the precipitation
process.

2. EDS microanalysis using 5 kV accelerating voltage decreased the spatial resolution of

analysis from ~1 µm in the ordinary 20 kV voltage to ~100 nm. Low voltage EDS also
revealed the light elements such as carbon, nitrogen and boron in the composition of
precipitates thanks to the less matrix effect in absorption of low energy X-rays of light
elements.

3. The drawbacks of low voltage EDS were indicated as peak overlap below 1 keV in the
EDS spectrum, low P/B ratio, high sensitivity to contamination resulting in high

70
 detection limit (no detection of low-content elements), poor accuracy of quantitative
analysis and the continuous presence of a carbon background peak.

4. Low voltage EDS was ideal for identification of niobium rich sub-micron precipitates
since niobium Lα peak was well resolvable in the EDS spectrum; unlike vanadium Lα
peak which had overlap with chromium Lα. Therefore, the low voltage EDS was not
suitable for identification of vanadium-rich precipitates when vanadium content was not
high enough to provide adequate P/B ratio to be resolved from chromium.

5. Increasing the acquisition time provided better P/B ratio and more automatically detected
characteristic lines in the EDS spectrum (improving the qualitative analysis). However, it
increased the carbon background peak due to the increase in the contamination effect
(deteriorating the quantitative analysis).

6. EDS at 5 kV was estimated to be almost 2.5 times less accurate than the common EDS
analysis at high voltages for the main matrix elements (iron, nickel and chromium). Also,
for other elements with lower content such as copper, molybdenum and silicon the
accuracy was quite poor in the range ±30%.

7. The spatial resolution of the instrument used for low voltage EDS was estimated as 100
nm at 5 kV accelerating voltage which is at least eight times better than that using 20 kV.
However, only qualitative analysis was taken into the account for this estimation.

71
6 FUTURE STUDIES

When the effect of alloying elements on precipitation is being studied it is necessary to make
the alloys with exactly similar process, including melting condition as well as mechanical and
thermal process prior to the aging. This provides the same thermodynamic condition for
precipitation process in all alloys and limits the variables to that of interest (here, alloying
content). For the future works it is highly recommended that different alloys are made by the
same melting process which is reproducible to avoid increasing the affecting variables on the
precipitation such as degree of homogeneity, solidification structure and inclusion content.

The alloys of this study can be aged for longer times in order to investigate the stability of
observed precipitates in 500 h aging. Also, phases forming in long-term aging such as sigma
phase and probably copper-rich nano-scale precipitates may be found in higher fraction when
the aging time increases. Supplementary characterization methods can be used for
crystallographic examination to verify the findings of this work.

Regarding the low voltage EDS lots of possible settings in the SEM-EDS can be examined in
order to find the optimum setup for the most accurate analysis. Factors such as probe current,
size of aperture and accelerating voltage in SEM; and the acquisition parameters such as live
time, process time and the resulted dead time in EDS analyzer should be optimized to get
higher rate of X-ray generation, minimize the peak overlap and improve the P/B ratio.

To minimize the contamination effect and subsequently weaken the carbon background peak,
two possible ways are proposed. First, preheating the samples prior to the analysis in a proper
temperature range to evaporate common organic contaminations (e.g. oils, breath products,
Figure print, etc.) without detrimental effect on the sample; e.g. in the range of 100−150 ºC.
The second way is to equip the SEM chamber with so-called cold finger which is specifically
designed to absorb the polymerization products of contaminations and significantly decrease
the amount of carbon deposition on the sample surface.

The other suggestion is to use adequate standard samples for optimization of the EDS
analyzer. In this work the optimization for all low voltage EDS analyses was done by using
pure silicon; whereas for each element of interest, especially light elements, a standard
sample containing that specific element should be used to improve the accuracy of analysis
[38]
. For instance, when carbon is concerned a standard sample of Fe3C can be used for
optimization; whereas for nitrogen analysis a nitrogen containing standard sample shall be
used, e.g. chromium nitride.

72
ACKKNOWLEDGEMENT
The author is grateful to Rachel Pettersson who granted the chance to perform this master
thesis in Outokumpu Avesta Research Center (ARC) and gave illuminating comments on the
report. Thanks are also expressed to Fredrik Olsson for supervision of the work as well as
reviewing the report, former high-temperature group of ARC for valuable advices, Staffan
Hertzman and Malin Selleby for reviewing and commenting the report and all ARC staff for
their aids, advices and training to use laboratory facilitates.

73
7 REFERENCES

1. Atmosphere Changes, U.S. environmental protection agency, EPA,

http://www.epa.gov/climatechange/science/recentac.html, Dec. 2011.
2. CO2 emission from fuel combustion, 2010 edition, International Energy Agency
(EIA), http://www.iea.org/co2highlights/co2highlights.pdf, Aug. 2011.
3. Power generation from coal, Measuring and reporting efficiency performance and
CO2 emissions, Coal industry advisory board, International energy agency (EIA),
http://www.iea.org/ciab/papers/power_generation_from_coal.pdf, Dec. 2011.
4. R. Blum, S. Kjær and J. Bugge, Development of a PF fired high efficiency power
plant (AD700), New development on metallurgy and applications of high strength
steels international conference, Buenos Aires, 2008.
5. Advanced fossil fuel power generation, SETIS, European commission,
http://setis.ec.europa.eu/newsroom-items-folder/advanced-fossil-fuel-powergeneration, 2011.
6. Viswanathan and W. Bakker, Materials for Ultra supercritical Coal Power Plants-
Boiler Materials, JMEPEG (2001) 10:81-95.
7. R. Viswanathan, J.F. Henry, J. Tanzosh, G. Stanko, J. Shingledecker, B. Vitalis and
R. Purgert , U.S. Program on Materials Technology for Ultra-Supercritical Coal
Power Plants, JMEPEG (2005) 14:281-292.
8. T. Sourmail, simultaneous precipitation reactions in creep resistant austenitic stainless
steels, Doctoral thesis, University of Cambridge, 2001.
9. M.F. McGuire, Stainless steels for design engineers, ASM International, chap. 1 and
6, 2008.
10. International Stainless Steel Forum, http://www.worldstainless.org/NR/RDONLYRES/B2617D50-
73AE-4FAB-BDCD- 88ABD7891B97/4933/THESTAINLESSSTEELFAMILY.PDF, Sep. 2011.

11. T. Sourmail and H.K.D.H. Bhadeshia, Stainless steels, University of Cambridge,

http://www.msm.cam.ac.uk/phase-trans/2005/Stainless_steels/stainless.html, 2011.
12. R. Naraghi, Martensitic transformation in austenitic stainless steels, Master of Science
thesis, KTH, 2009.
13. T. Sourmail, Precipitation in creep resistant austenitic stainless steels, Materials
science and technology, Jan.2001, Vol.17.
14. Th. Nebel and D. Eifler, Cyclic deformation behavior of austenitic steels at ambient
and elevated temperatures, Sadhana, Vol. 28, Parts 1&2, Feb/Apr 2003, pp. 187–208.

74
15. D. Peckner, I.M. Bernstein, Handbook of stainless steels, McGraw-Hill, 1977, Chap.
4, 11 and 12.
16. J. Degerbeck and E. Wold, Some aspects of the influence of manganese in austenitic
stainless steels, Materials and Corrosion, Vol. 25, Issue 3, pages 172–174, 1974.
17. P.J. Cunat, Alloying elements in stainless steel and other chromium-containing alloys,
Euro Inox 2004, http://www.euro-inox.org/pdf/map/AlloyingElements_EN.pdf, Dec. 2011.
18. S. L. Chawla, R. K. Gupta, Materials selection for corrosion control, ASM
International, 1993, pp.15.
19. T. Ishitsuka, H. Mimura, H. Morimoto, M. Matsumoto, K. Nagashima, M.
Mizumoto, J. Okamoto, Development of new austenitic stainless steel boiler tube with
high strength at elevated temperatures and intergranular corrosion resistance, Nippon
steel technical report, 2004.
20. Béla Leffler, Stainless steels and their properties, http://www.outokumpu.com/files/
Group/HR/Documents/STAINLESS20.pdf, Oct.2011.
21. Chai, J.O. Nilsson, M. Boström, J. Högberg and U. Forsberg, Advanced heat resistant
austenitic stainless steel, ADVANCED STEELS, 2011, Part 5, 385-397.
22. A. Tohyama, M. Miyauchi and Y. Minami, Development of 18Cr-10Ni-3Cu-Ti-Nb
stainless steel for ultra-super critical boilers false, Materials for Advanced Power
Engineering 1994, Liege, Belgium, 3-6 Oct. 1994, pp. 495-504.
23. J. P. Shingledecker, P. J. Maziasz, N. D. Evans and M. J. Pollard, Alloy additions for
improved creep-rupture properties of a cast austenitic alloy, http://ornl.org/sci/de_
materials/documents/ShingledeckerMST2005Paper-Final.pdf, Nov. 2011.
24. S. Caminada, G. Cumino, L. Cipolla, D.Venditti, A. Di Gianfrancesco, Y. Minami, T.
Ono, Creep properties and microstructural evolution of austenitic TEMPALOY steels,
International journal of pressure vessels and piping, 87 (2010), 336-344.
25. B. Sasmal, Mechanism of the formation of lamellar M23C6 at and near twin
boundaries in austenitic stainless steels, Metallurgical and materials transactions A,
Vol. 30A, Nov. 1999-2791.
26. B. Peng, H. Zhang, J. Hong, J. Gao, H. Zhang, J. Li, Q. Wang, The evolution of
precipitates of 22Cr–25Ni–Mo–Nb–N heat-resistant austenitic steel in long-term
creep, Materials science and engineering A, 527 (2010), 4424–4430.
27. Y.U. Kimb, S.I. Kwunb, J.H. Shima, D.B. Parkb, Y.W. Choa, Y.H. Chunga, W.S.
Jung, The effect of cooling conditions after solution treatment on the creep rupture

75
 strength of sanicro25, The 3rd symposium on heat resistant steels and alloys for high
efficiency USC power plants, Japan, 2009.
28. T. Sourmail and H.K.D.H. Bhadeshia, Microstructure evolution in two variants of
NF709 at 1023 and 1073 K, Metallurgical and Materials Transactions A, Vol. 36A,
Jan. 2005-23.
29. R. F. A. Jargelius-Pettersson, Precipitation in a nitrogen-alloyed stainless steel at 850
ºC, Scripta Metallurgica et Materialia, Vol. 28, pp. 1399-1403, 1993.
30. R. F. A. Jargelius-Pettersson, Precipitation trends in highly alloyed austenitic stainless
steels, Z. Metallkd., 1998, 79, 177-183.
31. D.J. Powell, R. Pilkington and A. Miller, The precipitation characteristics of 20% Cr-
25% Ni-Nb stabilized stainless steel, Acta metall. Vol. 36, No. 3, pp. 713-724, 1998.
32. Y. Yamamoto, M. Takeyama, Z.P. Lu, C.T. Liu, N.D. Evans, P.J. Maziasz, M.P.
Brady, Alloying effects on creep and oxidation resistance of austenitic stainless steel
alloys employing intermetallic precipitates, Intermetallics 16 (2008) 453-462.
33. K. C. Russel and H. I. Aaronson, Sequences of precipitate nucleation, Journal of
materials science 10 (1975) 1991-1999.
34. G. W. Lorimer, Precipitation at grain boundaries, EDP Sciences,
http://dx.doi.org/10.1051/jphyscol:1975423, Feb. 2012.
35. N. Terao and B. Sasmal, Precipitation of M23C6 type carbide on twin boundaries in
austenitic stainless steels, Metallography 13: 117-133 (1980).
36. T. Furuhara, T. Maki, Variant selection in heterogeneous nucleation on defects in
diffusional phase transformation and precipitation, Materials science and engineering
A312 (2001) 145–154.
37. Introduction to Energy Dispersive X-ray Spectrometry (EDS), micron.ucr.edu/public/
manuals/EDS-intro.pdf, Nov.2011.
38. J. I. Goldstein, A. D. Romig, D. E. Newbury, C. E. Lyman, P. Echlin, C. Fiori, D. C.
Joy, E. Lifshin, Scanning electron microscopy and X-ray microanalysis, Second
edition, Plenum Press, 1994, ISBN 0-306-44175-6.
39. Types of signals in SEM, http://www4.nau.edu/microanalysis/Microprobe-
SEM/Signals.html, Nov. 2011.
40. B. Hafner, Energy Dispersive Spectroscopy on the SEM: A Primer, Characterization
Facility, University of Minnesota, http://www.charfac.umn.edu/instruments/eds_on_sem_primer.pdf,

Nov. 2011.

76
41. P. Poelt, Low voltage EDXS and elements of the first transition series, Mikrochim.
Acta 132, 129-135, 2000.
42. A.M. Elwazri, R. Varano, F. Siciliano, D. Bai and S. Yue, Effect of cool deformation
on mechanical properties of a high-strength pipeline steel, METALLURGICAL AND
MATERIALS TRANSACTIONS A, Vol. 36A, Nov. 2005—2929.
43. R. Gauvin, K. Robertson, P. Horny, A. M. Elwazri, and S. Yue, Materials
characterization using high-resolution scanning-electron microscopy and X-ray
microanalysis, JOM, March 2006.
44. E.D. Boyes, On low voltage scanning electron microscopy and chemical
microanalysis, Microsc. Microanal. 6, 307–316, 2000.
45. D. C. Joy and C. S. Joy, Low voltage scanning electron microscopy, Micron, Vol. 27,
No. 3-4, pp. 247-263, 1996.
46. E.D. Boyes, Analytical potential of EDS at low voltages, Mikrochim. Acta 138, 225-
234, 2002.
47. X-Ray Data Booklet, http://www.physics.uwo.ca/~lgonchar/courses/p9826/xdb.pdf,
Nov. 2011.
48. J. O. Andersson, Etching of stainless steel, internal instruction for section of
metallography, Outokumpu Avesta research center.
49. M. Shiga et al, Heat resisting steel and steam turbine rotor shaft, European Patent
EP0806490, 2001, http://www.freepatentsonline.com/EP0806490.html, Nov. 2011.
50. ASTM A 213/A 213M-09a, Standard specification for seamless ferritic and austenitic
alloy-steel boiler, superheater and heat-exchanger tubes, ASTM international, 2009.
51. E.C. Green, Catastrophic oxidation to stainless steel superheater steam lines, UK
National Corrosion Conference 1982, pp. 103-104. 1982.
52. Carl Zeiss SMT AG - ULTRA 55 FESEM, http://www.microscopy-
analysis.com/news /carl-zeiss-smt-ag-ultra-55-fesem?c=, Nov.2011.
53. Instrument Database, Carl Zeiss AG - ULTRA 55, http://www.speciation.net/
Database/Instruments/Carl-Zeiss-AG/ULTRA-55-;i675, Nov. 2011.
54. R. Varano, A.M. Elwazri, F. Siciliano, D.Q. Bai, R. Gauvin, S. Yue, FE-SEM study
of fine Nb precipitates in carbon extraction replicas, Materials Science Forum, Vols
500-501 (2005), pp. 663-668.

77
 TECHNICAL WRITING

STYLE GUIDE

Prepared By

Gregory M. Lilburne
Hafedh Al Qassab
Savas Soydaner

Technical Services Department

September 2007

©Technical Services Department Bapco 2007

Published by Crambeth Allen Publishing Ltd 2007. Individual copies of this Style Guide may be freely
downloaded from www.eptq.com but further distribution or any form of reproduction, in whole or part, is
prohibited without prior written consent of the publisher.
 Page i

CONTENTS

Page

Contents ……………………………………………………………………..… i

1.0 Introduction ………………………………………………………………….. 1

2.0 What is Style and Why Is It Important? …………………………………… 3

3.0 Overall Appearance and Physical Layout ……………………………….…. 4

3.1 Page Layout ………………………………………………………………… 5
3.2 Font Type and Size …………………..…..………………….……………… 6
3.3 Text Alignment and Word Spacing ………………………………………… 8
3.4 Section Indentation and Numbering ……………………………………….. 9
3.5 Headings, Bold and Underline …………………………...………………… 10

4.0 Tables in the Text ……………………………………………………………. 11

5.0 Numbering, Bullets and Lists ……………………………………………….. 12

6.0 Writing Style …………………………………………………………………. 14

6.1 Language and Expression …………………………………………………... 14
a. Active or Passive Voice ………………………………………………… 14
b. Word Selection …………………………………………………………. 15
c. Spelling ………………………………………………………………… 17
d. Punctuation …………………………………………………………….. 17
e. Grammar ……………………………………………………………….. 18
f. Pedantry ………………………………………………………………... 18
6.2 Sentence Length ……………………………………………………………. 18
6.3 Numbers in the Text ………………………………………………………... 19
6.4 Dates ………………………………………………………………………... 19
6.5 Abbreviations ……………………………………………………………….. 20
6.6 References in the Text and Lists ……………………………………………. 20
6.7 Brackets …………………………………………………………………….. 21
6.8 Dashes ………………………………………………………………………. 22
6.9 Page Numbering ……………………………………………………………. 22
6.10 Other Guidelines ……………………………………………………………. 23

7.0 Emails at Work (‘Engineering Emails’) ……………………………………... 25

7.1 Introduction …………………………………………………………………. 25
7.2 Main Governing Points ……………………………………………………... 25
7.3 Email Etiquette and Conventions …………………………………………... 27
 Page ii

CONTENTS (cont’d)

Page

8.0 Other Requirements, Conventions and Guidelines …….………………….. 30

8.1 The First Page …………………..........…….……………………………….. 30
8.2 The Signature Page …………………………………………………………. 30
8.3 Attachments ………………………………………………………………… 30
8.4 Distribution Lists …………………………………………………………… 30
8.5 Electronic File Names and Filing …………..................……………………. 31

9.0 References ……………………………………………………………………. 33

 Page 1 of 34

1.0 INTRODUCTION
“Engineers are rarely criticised for lack of knowledge of thermodynamics, mass transfer or
mathematics. They are, however, often criticised for their inability to write clear, concise and
correct English. It is tragic to find engineers who are technically competent but unable to
demonstrate this competence to others. The advancement of their careers and the achievement of
personal satisfaction from their jobs, depend on their ability to communicate ideas and findings
successfully to colleagues and to the public.” 13

In our profession, the importance of good communications skills, especially in written

communication, cannot be over-emphasised. Good technical writing is a must, to ensure that
your message, ideas and recommendations are unequivocally understood by all readers.

You work hard all day, sweating under the hot sun to make field measurements and collect
samples, analysing copious quantities of data and test results, evaluating unit performance,
running simulations, reviewing historical records, solving the problems and creating new
concepts. But all this effort is wasted if the resulting written communication fails to deliver the
right message in the right manner. At best a poor writing style distracts - perhaps even amuses -
the reader; at worst it irritates and confuses the very people whom you wish to impress and
convince. Busy decision-makers routinely have many documents of all types pass over their
desks and have better things to do with their time than try to puzzle out unclear communications.

It’s a hard life. There may have been many reasons for your choosing a career in engineering.
The opportunity to excel in written English is unlikely to have been one of them. But the fact
remains that the ability to communicate in a clear and unambiguous manner lies at the heart of
your job. It’s not a luxury but an operating necessity.7

Throughout your working life clear written communication will enable your technical expertise
and hard work to shine through, helping you to achieve your ambitions. Within Bapco it will, at
the very least, assist you in getting your point across. In your dealings with the outside world it
will, quite simply, mark out the professionalism of the department and the company. In short,
without the ability to communicate clearly in writing you are handicapping yourself at every
level.

You may accept the truth of all this, but still find the thought of technical writing daunting and
onerous, particularly if English is not your first language. Take comfort.

Firstly, if you have what it takes to be a competent process engineer, you have the ability to
produce good technical writing. Such writing may not be easier, but it’s certainly not harder,
than sound engineering work. Like engineering, good writing requires knowledge of the field,
good planning, critical thinking and attention to detail. It’s unlikely to be perfect first time
around and will require revision.11 This is entirely normal.

Secondly, you are likely to have people around you who can help. If their writing skills and
knowledge of the subject are greater than yours, that’s excellent. Show them your work and
listen and learn from their observations. Supervisors, and those proficient in technical writing,
have a duty to help their colleagues improve their writing skills; and you, in turn, should accept
their constructive criticism as an opportunity to improve your skills. Moreover, just as you would
have a colleague check your engineering calculations, so should you ask them to review your
written work.

Thirdly, at the risk of sounding immodest, you now have this Technical Writing Style Guide. As
process engineers, working in Technical Services, our only product is our written work and the
same is also true for so many engineers of other disciplines. Therefore it’s surprising that style
guides aimed specifically at engineers are something of a rarity. We hope we can at least start to
rectify this state of affairs with this document. This is your guide to style, tailor-made to our
 Page 2 of 34

1.0 INTRODUCTION (cont’d)

environment and objectives, tested over time, the result of years of experience here at Bapco and
elsewhere. It won’t tell you how to produce great written work, but we hope it will show you
how style, format and choice of language might assist your daily tasks – every one of them
involving written communication – of generating ideas, developing concepts, ordering catalysts,
solving problems or optimising the operation of a unit.

At this point, a note of caution is timely. Adopting the correct style for your written work is only
part of the story and does not, by itself, ensure that the document meets the standards that we set
for ourselves. Organisation of one’s thinking, the structure of the text and the document, and the
presentation of the logic and the technical arguments are also vital to success, and it is equally
imperative that the content of the final document is accurate, consistent and unambiguous. All
these aspects taken together make for the best document.

Finally, take the same pride in your technical writing as you would in developing a calculation,
piping and instrument diagram, process flow diagram, drawing, sketch or a painting. With
guidance and practice, technical writing should become second nature and in the long run it is
efficient, powerful and satisfying.
 Page 3 of 34

2.0 WHAT IS STYLE and WHY IS IT IMPORTANT?

For the purpose of this document, the term style means the set of rules, conventions and
guidelines which covers most aspects of the written work produced in Technical Services
Department. It encompasses a diverse range of topics from overall document appearance and
layout, to the choice of words and the way in which they are used and presented in the form of
sentences, paragraphs, sections and subsections.

This style guide mainly concentrates on the way the document is presented and how it looks.
This area is easy to get right for maximum effectiveness. The choice of words and sentence
construction i.e. language and expression are also addressed, but in lesser detail. This is a more
esoteric aspect of writing, but can be improved by reading, asking questions, taking advice from
someone who has better writing skills and more experience and, most of all, practice, at work
and outside.

No matter who receives your correspondence, the way it looks, the physical layout, its neatness,
the structure and clear organization and presentation of ideas and sections, length of the
sentences and the paragraphs, correctness of spelling, the choice of words and the style of
writing, all show a professional attitude and respect for your reader. In addition to having a
positive effect on the reader, you want your written work to reflect well on you the writer, your
group and the department.4,5

Overall, style is about passing on your message in a pleasant, clear and effective way.
 Page 4 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT

Obviously the information in any document must be presented in a clear, logical and structured
way, but how it looks is also vitally important. This is about making the reader’s life easy or
difficult. The reader should get the ideas with the greatest possible ease and not waste his energy
fighting through your presentation.

The overall appearance and physical layout of your document has a major impact on how well
it’s received, read, understood and ultimately accepted. If the document doesn’t look good, then
there’s a high risk that it will discourage, frustrate, amuse or even irritate the readers to such an
extent that people will put off reading it for as long as possible.4 In addition, your work, your
message, the conclusions and the recommendations will be poorly received and possibly unfairly
judged or ignored, or even rejected outright.

There have been many articles and books written on the subject of improving technical writing.
One of the important common themes is that attention must be paid to the presentation of the
work. It’s important to understand that the same material, better presented, can have a much
higher impact and improve your success rate. Presentation can be enhanced in many simple
ways, which are often ignored, to the detriment of the document and the writer. The more
important steps which can be taken to improve appearance and readability are summarised
below.

Features like double line spacing, generous margins, headings, subheadings and indentations, all
help make a report look reader friendly. Numbering of sections, with a logical, consistent,
sequence of major and minor numbers, also improves clarity.4 It is not the reader’s job to
decipher your document – it should not be a puzzle.

To make your report appealing to the eye and enhance readability, break your writing up into
short sections. Pages that are crammed with text from margin to margin, and long, unbroken
blocks of text are uninviting and intimidate and bore the reader. Moreover, fast reading is done
with the eyes, and eyes cannot follow big blocks of text – they just slow down the reader.
Breaking your writing up into short sections and short paragraphs, and using headings and
subheadings, makes it easier to read.1c,5,6 Therefore be generous with white space.

Something as simple as a blank line between sections serves to open up the piece and make it
less intimidating. Ten pages of solid text often seems like more of a chore than twelve pages
with some extra white space thrown in for good measure.6

Include sufficient white space in your document by using wide margins and generous spacing
between paragraphs and titles, headings and your main text. This will make your document more
inviting and easier to follow. To check whether a document has sufficient white space, hold a
page of text at arm’s length and squint at it, so that it becomes blurred. If you see nothing but big
blocks of text, more white space is required between the elements of text.3 Remember, that a
document, when viewed from a distance, should be pleasant and balanced, just like a good
painting is.

Modern word processing packages have a tremendous number of features which simply weren’t
available in the days of the typewriter and the first dedicated word processors. Although these
features can be used to great advantage, many people fall into the trap of using a myriad of
inappropriate features just because they’re readily available. This can have disastrous effects on
the document. The final document must still be well presented and readable. To use every
gimmick and font available might be fun for the writer, but in most cases it will only irritate or
confuse the reader. Given an appropriate font, plenty of paragraphs and enough white space,
your document will convey its message as clearly as the words allow.1c
 Page 5 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT (cont’d)

The main factors which affect the appearance of the document are:

♦ Page Layout
¾ Margins
¾ Header – with page numbering, and possibly with file name and reference
♦ Font Selection
¾ Font Type
¾ Font Size
♦ Text Alignment and Word Spacing
♦ Indentation
♦ Paragraph spacing
♦ Use of bullets
♦ Use of the hard space i.e. CTRL-shift-space or hyphen

For maximum effectiveness and readability of your memoranda, reports, and email attachments,
follow the rules given below. These are well established, tried and proven. Everyone has a
different view but these work.

3.1 Page Layout

The rules for page layout are:

♦ Use A-4 paper size for all text documents – ensure that A-4 is selected in the ‘Page Setup’
and in the ‘Printer Properties’. Other paper sizes, especially ‘Legal’ will lead to margins
which look ridiculous on A-4 paper.
♦ Margins should be as follows:
¾ Left: 1 inch (2.54 cm)
¾ Right: 1 inch (2.54 cm)
¾ Top: 0.7 to 1 inch (1.78 to 2.54 cm), depending on what, if anything, is in the header –
the aim is to ensure a ‘reasonable’ space between the top of the page or the last line in
the header, and the first line of the main body of the text.
¾ Bottom: 0.7 inch (1.78 cm)
♦ Headers and footers should be 0.5 inches (1.27 cm) inside the edges.
♦ If there is more than one page, then automatically number each page as ‘Page x of y’, in
font size 9, at the top right hand corner of the header box. Use a right tab at the right
margin to locate the text. Ensure that at least one blank line in the header separates the
header text from the main body of the text.
♦ Try to avoid text in the footer box unless it has a special purpose e.g. copyright or
confidentiality.
♦ Do not put page numbers in the footer.

In finalising the document, use page breaks to ensure that the layout is frozen and that
unsightly gaps or empty pages do not appear.
 Page 6 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT (cont’d)

3.2 Font Type and Size

The type and size of font selected for the text are extremely important to the overall appearance
and readability of the document. Although there are a large number of fonts available in the
word processing package, very few are suitable for the documents which TSD engineers
produce.

The rules and guidelines for font type and size are presented below. However, in order to
understand the rationale for these guidelines, a few definitions and explanations are required.

Fonts
In printing, publishing and word processing, a font is defined as a complete set of characters
i.e. letters, figures, punctuation marks, and other symbols, which are available in a particular
type face. A font family is a group of fonts with the same name e.g. Times.1d

Roman is the upright form of type used for all general purposes. It contrasts with the sloping
italic type.2c Roman font is much easier to read and looks more pleasant. Avoid long runs of
bold or italic font.1c

Serif Fonts
A serif font is one that has short strokes called serifs, which cross the ends of the stems, arms
and tails of the characters, as shown in the following example:1c

The characters in a serif font look better than characters which do not have a serif i.e. sanserif
characters, and they are thought to have a psychological benefit in promoting readability by
drawing the reader’s eye along a line of text.3 Most people agree that serif type is easier to
read than sanserif fonts.2c

Sanserif Fonts
Fonts which have characters without serifs are called sanserif. Sanserif faces such as Arial may
appear to be neater or cleaner than serif faces such as Times New Roman, but they are more
difficult to read, especially in long stretches in large slabs of text.1c,3 The problem is
aggravated by the use of justified text. “Being exceptionally clear, Sanserif fonts are ideal for
signposts”.1c

Although they are sometimes considered for headings and captions, sanserif fonts should not
be used for TSD reports and memoranda. However, given their simple definition and clarity,
they can successfully be used for faxes and email messages, but take care with the font size to
avoid an inconsistent and unsightly appearance e.g. The font used in this part of the
sentence is Arial, with a font size of 11, whereas the rest of the sentence is in Times New
Roman with a font size of 11. As is evident, the Times New Roman font is easier on the eye.
 Page 7 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT (cont’d)

3.2 Font Type and Size (cont’d)

Italics
Italic font is sometimes called cursive or oblique type because it is not vertically upright but
sloped. It was developed in the sixteenth century from handwriting.1f

Italics can be difficult to read and are best suited to serif rather than sanserif fonts. They have
limited application in a technical document.

Italics should be used sparingly to emphasise a word or group of words, or to draw attention to
an unusual word or one being used in an unusual way, or to highlight technical terms, if
necessary. Note however, that technical terms should only be italicised for their first
appearance.2a

Summary of Conventions for Fonts

For maximum effectiveness use the following rules and guidelines:

♦ Use ‘Times New Roman’ with a font size of 11 or 12 points.

♦ Do not use smaller font sizes for the body of the text – it makes the document more
difficult to read, and distracts the reader. Small passages can be presented in a smaller font
to achieve some special effect or emphasis.
♦ Do not use much larger font sizes for the body of the text – it can make the document look
somewhat childish and primitive, it distracts the reader and casts doubts on the credibility
and attitude of the writer.
♦ Do not use a mixture of fonts in the text – this is highly disconcerting for the reader. At the
very least it is distracting, but it can also frustrate and irritate the reader, and cast doubt on
the author’s attitude and motivation.
♦ Use upper and lower case according to the basic rules of grammar e.g. use upper case for
proper names, to start a new sentence after a full stop and for the names of the months and
the days of the week. However, in general, there is no single correct rule in this regard that
covers all situations, and some judgement may also be necessary.12 More advice relating to
refinery stream names and process equipment is given in Sections 6.5 and 6.10.
♦ Do not write in all capitals – this is the word processing equivalent of using a crayon. It
distracts and irritates the reader, it may appear threatening, it reduces the credibility of the
writer and detracts from the message.
♦ As an example, the font type used in this sentence is Times New Roman, with a font size
of 11. The characters have a serif i.e. a short stroke on the feet and tails, which is important
for readability.2c,3
♦ Do not use sanserif fonts like Arial in the main business documents e.g. memoranda, file
notes, reports, email attachments and letters.
♦ Sanserif fonts may be used for faxes and emails with a font size of 10.
 Page 8 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT (cont’d)

3.3 Text Alignment and Word Spacing

Alignment and justification describe how the text is aligned with respect to the left and right
margins of the document. In our word processing package there are four choices:

♦ Align Left – in this format the text is flush with the left margin, the length of each line
varies and the right side of the text is ‘ragged’.
♦ Centre – means that the text is aligned about an imaginary vertical line down the centre of
the page, and therefore the left and right side of the text is ragged.
♦ Align Right – the text is flush with the right margin and the left side is ragged.
♦ Justified – in this format, the text is set flush against both the left and the right margins, as
is done in most books and journals.

For maximum readability, the spacing between the words in each line of text should be
consistent and even.1g This can only be achieved by using ‘left aligned’ text, with a ragged
right margin. This also gives the reader’s eye a rest at the end of each line, which makes the
document easier to read.3 The other advantages of left aligned text are that hyphenation can be
significantly reduced, which improves readability, and the layout is simple, more natural and
more pleasant for the reader.1e

Do not use justified text i.e. the text flush against both margins. This has far too many
disadvantages which combine to degrade the overall presentation and devalue the content of
the document. Justified format creates large blocks of text which always appear dense and are
difficult to read.3 It can also be more difficult to edit justified text e.g. when words are deleted
on one line, other words are continually moved up from the following line to ensure that the
line length remains constant, and this can be disconcerting and lead to errors. Moreover,
justified text also introduces unintended large gaps between the words. Even and consistent
spacing can rarely, if ever, be achieved with justified text.

Too much word space can also “create distracting rivers of white space that run vertically
through the text”, as shown in the following example:1g

Hence the rule is – only use left aligned text, with a ragged right margin, for maximum
benefit.
 Page 9 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT (cont’d)

3.4 Section Indentation and Numbering

Give careful consideration to numbering of sections and subsections. As discussed above, this
will enhance the layout, avoid confusion and any apparent inconsistency. It will also force the
writer to organise his or her thoughts on the structure of the document and how best to present
the ideas in a logical manner. Therefore:

♦ Number the sections and subsections logically. In general, manual numbering is preferred
since it is tried and tested, will avoid formatting problems and allow the section numbering
to be continued at the top of each page, if appropriate. Automatic numbering is also a
powerful tool since it can also be linked to the index or table of contents. However,
historically this system has been complicated, user-unfriendly, and allows little or no
flexibility. The writer must really be skilled to use this system effectively and avoid
introducing other problems.

♦ Indent the subsections and, if necessary, the sub-subsections.

♦ For greatest effectiveness, keep the structure simple. Therefore, in general, the strong
preference is to use only two levels of subsections i.e. sections and subsections. It should
be fairly easy to stay within this limit in medium size memoranda and reports. However
there may be occasions when you are forced to use another level, i.e. sub-subsections,
especially for longer, more comprehensive documents. However, even in these types of
documents, do not use more than three levels of subsections. Use of more numbered
subdivisions i.e. lower order sections, severely upsets the appearance and organisation of
the document and is unworkable.

♦ The following tab settings, in inches, are recommended: 0.15, 0.25, 0.35, 0.50 and 0.70,
with:

¾ The section number aligned at the left-hand margin of the page and the first word of
the section heading at 0.35”.
¾ The subsection number at 0.15” from the left-hand margin and the first word of the
subsection heading at 0.50”. This means that the first number of the subsection will be
approximately in line with the ‘.0’ in the section number.
¾ The sub-subsection number at 0.25” from the left-hand margin and the first word of the
sub-subsection title at 0.70”. This means that the first number in the sub-subsection
will be about in line with the ‘.X’ at the end of the subsection number.
¾ It is not necessary or possible to have all the titles of each section and subsection
aligned. The aim is to end up with titles which are indented for emphasis and clarity.

Note: The corresponding tab settings in centimetres, are: 0.38, 0.63, 0.89, 1.27, 1.78.

♦ If more lower level ‘subsections’ are required, use an alphabetical character or a simple
number, either with or without an underline. It is also acceptable to use just a heading
either underlined, or simply in bold.

♦ Do not use: 1-, 1.-, 1), 1)-, or a-, a.-, a), a.) or a)-, or any other combination of these types
of characters because they are ungainly.

An example of the use of effective indentation, alignment and designation of sections,

subsections and sub-subsections is given on the following page.
 Page 10 of 34

3.0 OVERALL APPEARANCE AND PHYSICAL LAYOUT (cont’d)

3.4 Section Indentation and Numbering (cont’d)

Example

Use the following example as a guide for indentation and alignment:

6.0 MODEL II UNIT DESIGN FEATURES (cont’d)

6.1 The Model II Unit (cont’d) Note: In this example the
‘6.0’ is aligned with the left
6.1.1 Layout and Flow Scheme margin.

6.1.2 Regenerated Catalyst Standpipes (RCSP’s)

a. Standpipe Aeration
b. Catalyst Circulation Rates and Standpipe Stability

3.5 Headings, Bold and Underline

The main heading or title at the beginning of the document must be located in the upper centre
of the first page. It may be written using upper case for the first letter in each main word and
lower case for the other letters, which looks better and is more inviting than all capitals. A less
preferred option is to use capitals for all words. In either case, the words should be in bold.
Underlining of the document title/heading on the first page is therefore not required, although it
may be used. If it is used, then consider underlining only the last line of the heading/title, rather
than all the lines, which can be rather forbidding and messy, depending on the number of lines.
If possible, keep the heading to one line.

As indicated in the above example:

♦ Section headings should be capitals and bold, without an underline.

♦ Subsections and sub-subsections should use upper case for the first letter in each main
word and lower case for the other letters, and be in bold. No underline is used.

♦ All other lower order sections should use upper case for the first letter in each main word
and lower case for the other letters. These headings should not be in bold.
♦ To emphasise selected words in the text use either an underline or show them in bold, but
do not use an underline and bold together.
 Page 11 of 34

4.0 TABLES IN THE TEXT

Tables really do add clarity, as well as information, but they must be simple, clear, informative,
reasonably small, and appealing to the eye. Tables in the text are preferred because they enhance
understanding and efficiency and the reader does not have to go back and forth to an attachment.
However, any table that is longer than half a page is a good candidate for being an attachment.

Therefore:
♦ The layout, column width, row height, and line densities all have to be carefully selected.
For maximum effectiveness:
¾ Do not show the left and right borders of the table. This opens up the table.
¾ Use higher density lines to highlight the title row and the bottom of the table. The
density should typically be 1 to 1½ points.
¾ Use higher density lines to delineate the internal column borders (density typically 1 to
1½ points).
¾ Use low density lines with a density of ½ to ¾ point, or no lines, for the rows within the
table.
♦ Do not use large, bulky, complicated and unsightly tables in the text – they defeat the
purpose of having a table
♦ Make the font size of the entries in the table one or two sizes smaller than that of the main
body of the text – this will allow more information to be fitted into each cell and will
minimise the size of the table
♦ Format the cells so that each entry is centred vertically
♦ Adjust the row height so that there is some space above and below each entry
♦ Centre and bold the column headings
♦ Centre align or left justify the entries, depending on their nature – be careful with numbers,
which should still be lined up according to the decimal points
♦ With numbers, use a comma to separate the numbers before the decimal point, into groups
of three to signify thousands, millions and so on. It is not acceptable to use a space instead
of the comma. Refer also to Section 6.3.
♦ If percentage compositions are used, which ought to add to 100.0, the total must be shown
and the numbers must indeed add to 100.0 percent. So check!

The tables given below are examples of the style which must be followed:

Source Unit Sulphur (tonnes/day)

Revamped 2HDU, diesel service 160.0
FCCU with HVGO feed 45.5
Others 17.5
Total 223.0

Source Unit Sulphur (tonnes/day)

Revamped 2HDU, diesel service 160.0
FCCU with HVGO feed 45.5
Others 17.5
Total 223.0
 Page 12 of 34

5.0 NUMBERING, BULLETS AND LISTS

Numbers and bullets are very effective when presenting lists of items which cannot be
appropriately or reasonably presented within the main body of the text. They can also be a useful
way of summarising findings, conclusions or action items. However, if they are not used or
presented carefully, they can cause confusion and frustration.

In general, numbered and bulleted items must be succinct, specific and relevant to the text. They
should not be complicated sentences or paragraphs which embody several ideas or issues. Since
it is difficult to link bulleted items, if a relationship does need to exist between some items then
they should be part of the text.

A discussion or introduction should precede a list of numbered and bulleted items, which puts
the list into context and indicates why it is being presented. The purpose of this text may be to
link ideas, issues, concerns or requirements, or summarise the outcome of the ideas that follow
and which support the text. Often a concluding paragraph is required at the end of the list.

The use of numbers or bullets is not a substitute for logical, well presented discussion in the
form of sentences and paragraphs. Bullets and numbers are generally used to present sets of
information but not the text of detailed discussions, logic solutions or arguments.

If there are a few items in a list which are not referenced later or elsewhere in the document, use
bullets. Otherwise use numbers or letters for these items.

The following guidelines and rules should be used for presenting numbered and bulleted lists:

♦ For long lists of items, the numbers or alphabetical characters can be entered manually or
automatically. The automatic numbering system is useful for consistency but it’s not without
its own problems, especially when the sequence has to be broken for an un-numbered
comment. In any case:
¾ For numbered items in a list use the following format:
1.
2.
3.

¾ For alphabetically designated items in a list use the following format:

a.
b.
c.

Do not use: 1-, 1.-, 1), 1)-, or a-, a.-, a), a.) or a)-, or any other combination of characters. To
do so, may confuse, frustrate and possibly irritate the reader.
♦ Select the bullet symbol carefully and conservatively e.g. solid circles (•), diamonds (♦) or
squares (▪), are acceptable.
♦ Bullet symbols should not distract from or overshadow the text. Therefore the symbol font
should be 3 to 4 points smaller than the text e.g. use font size 8 for a bullet with font size 11
or 12 text, as is done in this document.
♦ This is an example of a font size 11 bullet with font size 11 text. Clearly this bullet symbol
is too big and unsightly, and detracts from the text.
 Page 13 of 34

5.0 NUMBERING, BULLETS AND LISTS (cont’d)

♦ Minimise the use of sub-bullets but if there is a need to use them, then indent them and make
them less striking than the main bullet. For example:
♦ Example
o Aaa, or
¾ Aaaaa

♦ In general, the number or bullet symbol should be aligned left in line with the preceding
body of text, and the text of the bullet should be indented by 0.25 inches (0.63 cm).
♦ Do not use Tabs with bullets. This is not required and can lead to confusion and errors in
creating and revising your document.
♦ The same rules for presentation of text apply to bullets. Therefore avoid extensive long lists
without any white space. In general, numbered and bulleted items should be separated by a
spacing equivalent to one line space.
 Page 14 of 34

6.0 WRITING STYLE

6.1 Language and Expression

Much has been written on the use of language and expression in written work and several
relevant references are listed in this style guide. You are urged to read and consider the advice
given in these references as well as to find and consult other sources.

It is beyond the scope of this document to present a complete model for technical writing
which contains a rigid set of rules. The final style of writing which you adopt will vary to some
extent with the subject which is being addressed, the type of document which is being prepared
and the audience. However there are some typical guidelines and conventions which you
should use to improve the style of your writing, make it more interesting and increase your
standing and likelihood of success in achieving your objectives.

The guidelines and conventions that are summarised in this section are based on our
experience, and recommendations taken from the literature.

The most important point to understand and remember is that good communicators are made,
not born.7 Your writing ability and your writing style can be improved only by reading, asking
questions, taking advice from someone who has better writing skills and more experience and,
most of all, by practice, at work and outside.

a. Active or Passive Voice 5,6

There is a natural tendency for technical writing to be serious, dry and written in the passive
(impersonal) voice i.e. using third person pronouns and passive verbs. For example: “it has
been determined” or “various catalysts have been examined in the test reaction”.

This tendency is due to the training and advice we received in science and engineering subjects
from an early age, when the use of the active voice and pronouns like “we”, was just not
allowed. Unfortunately this approach makes for terrible reading and creates a heavy document.

Nowadays however, it is recognised that the active voice can be successfully used in many, but
not all documents, to make the text more lively and interesting for the reader. For example:
“we determined” or “we examined various catalysts in the test reaction”.

When using the active voice, your writing will be more direct and vigorous and your sentences
will be more concise. This promotes interest, readability, and understanding. As a general rule,
write using the active voice and explain the subject as you would verbally describe it to a
colleague i.e. be natural in your expression. However, don’t get too conversational or use
colloquialisms, because this can create the impression that the document is superficial, lacks
professionalism or is carelessly prepared. The following examples illustrate the benefits:5

Passive Voice Active Voice

Control of the bearing-oil supply is provided
by the shutoff valves.
Leaking of the seals is prevented by O-rings.
Fuel-cost savings were realised through the
installation of thermal insulation.
The shutoff valves control the bearing-oil
supply.
O-rings keep the seals from leaking.
The installation of thermal insulation cut
fuel costs.
 Page 15 of 34

6.0 WRITING STYLE (cont’d)

6.1 Language and Expression (cont’d)

a. Active or Passive Voice 5,6 (cont’d)

The exception to the use of this style i.e. the active voice, is the presentation or statement of
recommendations, when the passive voice is usually required e.g. “It is therefore recommended
that an award…..”.

In addition, it is generally better to use the word “we”, which is consistent with a team and
departmental approach, rather than the word “I”, which could suggest arrogance, selfishness
and egotism.

b. Word Selection 5,8

The choice of words used to express your ideas, observations and analysis is very important.
There are no absolute rules, but the key to readability is simplicity, in the choice of words and
how they are used. Try to stick to words which most people are familiar with. This does not
mean that you should “dumb down” your writing, but that obscure words should generally be
avoided, even if “technically” they fit the context. For example, use words like “enthusiasm”
and “promptness” which are well known to most people, rather than “alacrity”, which has the
same meanings, but which very few people would be familiar with. Never use obscure words
to impress the reader.

The number of words needed and used by people in daily life depends on what they do and
their level of education. It typically varies from 600 to 5000.8 As process engineers we need
additional specialist terms which must be used precisely in context.8

In addition, “we tend to use a wider range of words in writing than in talking. There are two
reasons: firstly, the day-to-day shorthand or speech is no longer appropriate for writing, and
secondly, you must introduce a greater variety of expression to avoid boring the reader”.8

Do not repeat words and phrases too often when they are close to each other in the text. You
must expand and develop your vocabulary so that you have a choice of different words
e.g. show, present, demonstrate, establish, prove, identify, highlight, determine. Read as much
as you can, and have a dictionary handy to check on the meaning and spelling of words you
haven’t come across before. Refer also to Section 9.0 for a discussion on dictionaries and
thesauruses.

Technical terms must be used in your written work because that is the nature of the job. They
are therefore used to enhance your writing. They should be used to communicate your meaning
precisely, and you may need to explain or define the term, depending on your readers’ level of
understanding and experience.5 When it comes to technical descriptions it is important to be as
specific as possible e.g. say ‘a 40 ft spray dryer’ rather than ‘a tall spray dryer’. Moreover, “do
not be content to say something is good, bad, fast or slow when you can say how good, how
bad, how fast, or how slow.”5 However, never use technical terms for the sake of using them or
to impress your audience.

You must learn and use standard words, phrases and terms that are relevant and appropriate to
the work you are involved in. For example, in describing the flow path of a stream, do not
write that “the naphtha goes to the separator” but use clear, acceptable process engineering
language. Therefore in place of the word “goes”, use “flows”, “is pumped”, “is pressured”,
“flows under gravity”, or “is routed”.

Furthermore, avoid meaningless jargon, which one writer has described as “technicalese”
e.g. never write “mobile dentition” when you really mean to say “loose teeth”.5
 Page 16 of 34

6.0 WRITING STYLE (cont’d)

6.1 Language and Expression (cont’d)

b. Word Selection 5,8 (cont’d)

Finally, avoid using words just for padding out your document. This practice bores, irritates
and possibly confuses the reader, and your message will certainly be diluted, if not lost
completely. There are two areas which are easily addressed. Firstly, avoid using many words
when one or two will suffice. For example:5

Wordy, Waffly Phrases Concise Substitute

During the course of During
In the form of As
Several, some, a few.
A number of This waffly phrase begs the
obvious question ‘how many?’

Avoid using tautologies or redundancies i.e. saying the same thing twice in the one phrase.1l,5
Some examples are given below:

Tautologies (Redundancies) Concise Substitute

Join together Join
Revert back Revert
Advance plan Plan
A new innovation An innovation
A paid professional A professional
Rise up Rise
Actual experience Experience

You can easily identify and rectify the “padding” and redundancies during your reviews of the
different drafts of your document. These are ideal opportunities to streamline, simplify and
enhance your writing.

Keep your writing simple and to the point, and as far as is practical, keep sentences reasonably
short and sharp. “Simplicity conveys a certain elegance – you don’t have to write long,
complex sentences to appear scientific.”6 However, as stated above, this does not mean dumb
down your writing.

Finally, remember, you can increase your word power only by reading and listening, and
maybe watching some television, and most of all, by practising your writing. “Be adventurous.
If you come across a new word, look up the meaning and mentally file it away for future use.
Over time your vocabulary will broaden and ripen, allowing you to express yourself in new and
richer ways.”8
 Page 17 of 34

6.0 WRITING STYLE (cont’d)

6.1 Language and Expression (cont’d)
c. Spelling
Despite what some “progressive” teachers may think and advocate, spelling is important.
If your knowledge of words and spelling is weak, then you must use a dictionary for assistance.
You can even use the spelling checker in the word processing package, but be aware that it has
its dangers. It won’t understand the context, so “deign” is acceptable and not a mistake, but the
real word you want to use is “design”. So using this tool does not relieve you of the
responsibility of proper, thorough checking.
It’s better to use British English spelling in most cases e.g. metre and not meter (which actually
means to measure), colour not color, modelling and not modeling, sulphur not sulfur. There are
always some exceptions. In some cases convention may override this guideline and in other
cases it may be what is more aesthetically appealing, which requires some judgement
e.g. ‘hydrodesulphurisation’ looks better than ‘hydrodesulphurization’. British English spelling
is also a corporate requirement.
Words that end in ‘ize’ and ‘ise’ can cause a problem. The English tend to use the latter form
e.g. “realise”, which is softer, but the word “realize”, which is the American form, is
historically correct and, in the majority of cases, is directly linked to the Greek origins of the
word. More important however, is to avoid using the ‘ize’ ending for all similar words
e.g. ‘advertise’ simply does not have an ‘ize’ ending under any circumstances and there are
many words which fall into this category. Practically speaking, it is safer to use ‘ise’
everywhere, but if you’re unsure, then use a dictionary. Whatever convention you use, you
must be consistent within the document to avoid irritating the reader.2g,12
Obvious mistakes, which could be classed as typographical errors, are easily found through
proper review and proof reading of your document. If you don’t find and correct such errors
and leave them scattered throughout your document, then the reader will be left with a bad
impression of your abilities and attitude. Your reader will value your document based on how
much you, yourself, value your own document.

d. Punctuation
Punctuation is always important. Full stops, commas, colons, semi-colons, brackets,
apostrophes, quotation marks (inverted commas), have all been developed to allow ideas to be
expressed in the best possible way. They complement and enhance the words and their usage.
Some examples follow:
♦ Commas give the reader time to pause and catch their breath, perhaps take in relevant
phrases, and to link different parts of the sentence. More importantly, they affect the
meaning of a sentence. For example, the following sentence is a lucid statement, but
without commas in the right locations it is meaningless, “That that is is that that is not is
not”. Where would you put the commas? (The answer is given at the end of Section 9.) A
second example shows how the meaning can be distorted by putting the comma in the
wrong place. It refers to the giant panda which “eats, shoots and leaves”. Does it actually
have a gun?10
♦ An apostrophe is used to denote ownership e.g. the refinery’s normal throughput, or to
indicate the omission of letters e.g. don’t (do not) and it’s (it is). It’s easy to forget or
misuse an apostrophe e.g. ‘last year’s performance’ should be used and not ‘last years
performance’. In addition, to indicate ownership where a plural is involved, the apostrophe
should be after the ‘s’ e.g. the companies’ combined profit. In general, contractions
e.g. don’t and it’s, are not preferred for technical writing. Recognise when an apostrophe is
needed and take care when using it. The correct use of apostrophes is something of a
defining notion in modern English in separating the literate from the illiterate.
 Page 18 of 34

6.0 WRITING STYLE (cont’d)

6.1 Language and Expression (cont’d)
e. Grammar

The terms nouns, verbs, adjectives, phrases and clauses may now seem a distant memory from
early schooling. However, there are certain rules which have to be followed. Without these
rules the words simply will not fit together and your ideas will be lost in clumsy and
fundamentally wrong sentences. In some cases the words will not even form proper sentences.

So pay attention to grammar and follow the basic rules.

You cannot rely on the word processing package to help you with grammar, except perhaps in
the most basic areas like the use of singular or plural nouns and the nature of the verb ending
e.g. the stream flows or the streams flow, but not a combination of both. Unfortunately, the
word processing package often gets this fairly straightforward rule wrong.

f. Pedantry

Do not be pedantic i.e. don’t be nit-picking, or split hairs, or be excessively formal or fussy.
You should not blindly follow rigorous or traditional rules if they make your sentence sound
clumsy or just wrong. An example is the use of the split infinitive, in which a word is placed
between the ‘to’ and the ‘verb’ e.g. to boldly go. In this case, ‘to go boldly’ sounds all right,
but ‘to monitor more accurately’ sounds clumsy compared with ‘to more accurately monitor’,
which sounds much better. A second example is an archaic convention that sentences should
not end with a preposition e.g. in, on, with, under… As Winston Churchill once remarked,
“This is the sort of nonsense up with which I will not put”. In other words, your sentences have
to sound good, reasonable, logical, as well as contain the right ideas. If it doesn’t sound right
when read out loud, then it probably isn’t.

6.2 Sentence Length

In general, for maximum effectiveness in your writing, avoid long sentences. Very long
sentences, especially those without proper punctuation, can become convoluted, and can very
easily confuse, bore, tire, and possibly irritate the reader. At the very least, they distract the
reader from the point you are trying to make.

Short sentences are easier to grasp than long ones, and the occasional short sentence provides
relief from the longer ones. However, too many short sentences in quick succession can be
disturbing and distract the reader as well.2d,5

Ideally, the text should contain a mixture of short and long sentences.

Much has been written about sentence length. For example:

♦ Research has shown that our working memories can only process about 25 words at a time,
so try to limit sentence length to 15 to 20 words.3
♦ A survey by Harvard professor D.H. Menzel indicates that in technical papers, sentences
become difficult to understand when they exceed 34 words in length.5
♦ The target for average sentence length in Plain English documents is 20 words. However
the average in fiction is around 15 words.2d
 Page 19 of 34

6.0 WRITING STYLE (cont’d)

6.2 Sentence Length (cont’d)
Based on our experience, the optimum average length of sentences in technical documents is
equivalent to about 25 to 30 words, and this should be your target. Sentences can obviously be
shorter or longer, depending on the subject, context and content. However, it is important that
you go over your text and break the long sentences into two or more separate sentences. A
good guide for keeping sentence length under control is to write sentences that can be spoken
aloud without losing your breath. Do not take a deep breath before doing this test!5

6.3 Numbers in the Text

In the main body of the text, numbers can be written in words or as digits.
♦ For smaller numbers e.g. one to ten, words may be preferable, depending on the context.
♦ For larger numbers use digits.
♦ At Bapco, and in many other countries, it is standard practice to use a ‘dot’ or ‘full stop’ to
designate the decimal point separating the integers from the decimal fraction part of the
number. However, in some countries a ‘comma’ is used instead of a ‘dot’ and this can
obviously lead to confusion. Therefore, be careful when reviewing the data in
correspondence from overseas companies.
♦ With numbers in the text, use a comma to separate the numbers before the decimal point
into groups of three i.e. into thousands and millions and so on.
♦ With numbers in the text, it is also possible to use a space to separate the numbers before
the decimal point, into groups of three. However, this approach is problematic, and may
cause confusion and upset the look of the sentence because the groups of 3 numbers could
appear as words. Although ISO 1000 17 infers this approach, and it is routinely used in
some countries, possibly because a comma is used in place of a decimal point, at Bapco it
is not recommended for text – use commas as indicated above.
♦ Use digits to express the full number if the order of magnitude is less than a million
e.g. 24,000.
♦ If the order of magnitude is millions then use the word “million” e.g. 3.8 million.
♦ For currency, express dollars as either US$ or ‘dollars’ depending on the context. The same
applies to euros (€). Do not leave a space between the symbols and the numbers e.g. use
US$52,500 and not US $ 52,500 or US$ 52,500.
♦ For currency, use the word ‘million’ e.g. US$3.8 million, or, less preferably, use the
abbreviation ‘MM’ e.g. US$3.8 MM.
♦ For a currency range in the millions, use numbers and words for maximum clarity and
appearance e.g. US$125 to 150 million per year. An acceptable alternative is also to use
abbreviations and dashes in the text e.g. US$125-150 MM/yr.

6.4 Dates1b,2b
The format for dates should be day, month, year, which is a logical sequence and requires no
punctuation. For the example of the date “8 October 2005”:
♦ As stated above, do not use punctuation e.g. 8 October, 2005
♦ Use the cardinal number i.e. 8, rather than the ordinal i.e. 8th. This has been a growing
worldwide trend which started many years ago, and should be followed. Use of the ordinal
adds nothing and distracts from the main text.
♦ Avoid expressing the date purely as numbers i.e. 8/10/2005, because this looks ugly in the
main body of the document and it can lead to confusion, because in the US, the normal
convention is to put the month first, followed by the day e.g. 10/8/2005.
 Page 20 of 34

6.0 WRITING STYLE (cont’d)

6.5 Abbreviations
Use appropriate, common abbreviations to enhance your writing and improve the presentation
and readability of the text, but use them properly.

Technical terms can be abbreviated but when they are first used in the document they should be
presented in their full form, followed by the abbreviation in brackets e.g. Heavy Vacuum Gas
Oil (HVGO). Thereafter the abbreviation should be used without further definition.

The same approach is taken with equipment and equipment numbers. When equipment is first
referenced, it is very important to use the full name of the item followed by the equipment
number. This is to ensure that the equipment service and function are properly defined
e.g. Product Stripper C8406. Do not introduce and use equipment numbers without the
equipment name and service because this will create confusion, doubt and irritation, which will
certainly detract from your message.

Latin abbreviations can be used very effectively to improve readability. Common abbreviations
and conventions are as follows:

♦ e.g. – means “for example” and comes from the Latin exempli gratia. A comma may be
used before, but not immediately after, the abbreviation. Full stops are required. This
abbreviation is best used with a short phrase or list. It should not be used if it precedes and
is connected with another long sentence or long list. In these cases use a separate sentence
starting with the complete words “For example”.
♦ i.e. – means “that is”, or “that is to say”. It comes from the Latin id est. A comma may be
used before, but not after, the abbreviation. Full stops are required.
♦ viz. – means “namely” and comes from the Latin videlicet. A comma may be used before,
but not after, the abbreviation. The full stop is required.

An exception to the above is the abbreviation “etc.” which stands for et cetera, and means “and
the rest” or “so on” or “and others”. This is a much over-used and largely meaningless
abbreviation. It is not appropriate in formal writing.1j It is often discouraged because of its
inelegance, its discourtesy to the reader, and the implication that the writer is being lazy or is
short of information, which is more often than not the situation.2e Therefore do not use “etc.”.

The Latin phrase status quo, which means “the existing situation”, can sometimes be used.
However, be careful not to use phrases like “the present status quo” or “the status quo
situation”, each of which contains a repetition of the same ideas.1i

6.6 References in the Text and Lists

Text
In most documents it will be necessary at some stage to reference other work within the main
body of the text e.g. other documents, articles or meeting notes. The best way of referencing
this external information is to use a superscript number or letter after the full stop at the end of
the relevant sentence. This is the convention which is followed in this style guide.

Do not use the full reference in the text because it is a gross distraction for the reader, the train
of thought embodied in the sentence is severely disrupted, and the whole point of the sentence
or paragraph can be lost.
 Page 21 of 34

6.0 WRITING STYLE (cont’d)

6.6 References in the Text and Lists (cont’d)

Lists
A list of “References” should be included as the last section in your document. In a long report
it will be a separate numbered section of the document. In shorter communications like a
memorandum, it will be below an underlined heading.

In either case, the references should be presented as a numbered list. Particular references can
be included as a subset of a main reference and these should be designated using letters. Refer
to the References section of this style guide for an example of the required format.

Use the following guidelines for arranging and presenting lists of references:

♦ The font size for the list of references should be one to two sizes smaller than the main
body of the text. The same requirement applies to the list of “Attachments”.
♦ It is not necessary to indent the main references in a long list. The reference numbers can
be aligned with the section number or the first letter ‘R’ in the word ‘References’ in the
section title.
♦ References to internal documents should be presented with the title in quotation marks,
followed by the type of document, the author/authority/sender, the recipient, the date and
the TSD General Office file number.
♦ Literature references should begin with the author’s surname, followed by the author’s
initials, the title in quotation marks, the name of the source journal, the year, volume and
number of the publication and the page numbers.

6.7 Brackets1a,2f

Brackets, which are also called “parentheses”, are sometimes used, but are often detrimental to
the document, and are best avoided. Unfortunately brackets tend to break up the flow of the
sentence and are visually disturbing. In most cases, you will find that commas, or less
preferred, dashes, are a satisfactory alternative to brackets.

If there is an overwhelming need to use brackets, then follow these rules:

♦ Round brackets are the most common type of parentheses i.e. ( ). They are used to enclose
abbreviations and material that provides additional information or comment. The sentence
must be grammatically complete with or without the bracketed information.
♦ Avoid the use of long phrases or even sentences inside the brackets – these are not
appropriate, severely disrupt the flow of the sentence and distract the reader from your
idea. If such information really is needed, then present it as discrete sentences, even if it
means slightly restructuring your page or document.
♦ A comma may be used after the closing bracket. However, never use a comma before the
opening bracket.
♦ If you do have to use brackets, then remember that the standard for punctuation and
grammar within the brackets remains the same as that for a normal sentence.
 Page 22 of 34

6.0 WRITING STYLE (cont’d)

6.7 Brackets1a,2f (cont’d)
The following example illustrates the use of brackets, the conventions, and the improvement
which can be made by deleting the brackets.

“I set up the easel (which I had recently purchased), laid out my paints and selected a brush.”

Note the comma after the closing bracket. More importantly, this sentence, without the
brackets, looks better and conveys the same meaning. For example:

“I set up the easel, which I had recently purchased, laid out my paints and selected a brush.”

6.8 Dashes1k
In general, dashes should be avoided in your technical writing.

Dashes can easily disrupt the flow of the sentence and readability. It is better to use commas or
separate sentences. However, there may be some occasions where dashes need to be used.

The words contained between a pair of dashes will usually emerge more prominently or
become more emphatic. Unfortunately the effect will be lost if the “dashed” section is too long.
In this case, the reader may be drawn to a different train of thought, which can be convoluted
or confusing, and then find it difficult to put this information into the context of the main
sentence. Unfortunately there are many people who use dashes to add extraneous or irrelevant
material, when really, separate sentences are needed in their own right. So, great care is needed
when using dashes. In addition:

♦ A single dash towards the end of a sentence is more acceptable than a pair of dashes in the
middle of a sentence.
♦ Dashes are most effective when used sparingly.
♦ Do not use dashes to take short cuts on grammar

6.9 Page Numbering

Page numbering and the position of the page number are important, especially in longer
documents. It is also important to let the reader know how many pages there are in the
document, a convention which will also highlight any missing pages.

The traditional position for the page number is the top outer corner of the page, either as part of
the running header or alone.1m Therefore this is where most people expect to find the page
number. This also means that in our technical documents, which are normally printed on one
side of the paper only, the page number should be at the top right hand corner, usually as part
of the header. It is particularly important for page numbers to be easy to find and to read, and a
reader will tend to look at the top of the page first.1m Page numbering at the top outer edge also
facilitates leafing through the pages.

Some books and journals have the page numbers at the bottom centre of the page, especially
when other important material needs to be at the top, but this is less preferred and should not be
used for our technical documents.
 Page 23 of 34

6.0 WRITING STYLE (cont’d)

6.9 Page Numbering (cont’d)
When numbering the pages of your documents:
♦ Use the automatic pagination in the header, i.e. ‘Page x of y’.
♦ Use Times New Roman with font size 9.
♦ In shorter documents, like a two or three page memorandum or letter, it is still important to
let the reader know how many pages there are. If it’s not appropriate to have a page number
on the first page, then use the normal page numbering convention on the following page(s).
In some exceptional cases, e.g. a formal company letter from the president, it is acceptable
to put a page continuation i.e. ‘…/2’, at the bottom right hand corner of the first page, and
then use the normal page numbering convention on the following page(s). However such
continuations should not be used for our technical documents.
♦ For longer documents, start the page numbering from the first page of the text or
introductory page. In such documents, use roman lower case numerals i.e. i, ii, iii and so
on, for preliminary pages e.g. a preface or separate summary, or a table of contents.

6.10 Other Guidelines

There are many other aspects to your writing that could be considered trivial and that can easily
be taken for granted, but which nevertheless can be used to greatly improve your document.
For example:
♦ Use subscripts and superscripts to enhance the presentation and readability e.g. H2S is
better than H2S and m3 always looks better than m3. The subscript and superscript forms
are both easier to read, are logical, and consistent with technical conventions.
♦ In general, leave a space between a numerical value and the associated unit(s), especially
when the units have more than two characters e.g. use ‘125 psig’ rather than ‘125psig’, and
‘75,000 BTU/hr-ft2’ instead of ‘75,000BTU/hr-ft2’. However, there are some exceptions
which should be followed e.g. use ‘120°F’ and not ‘120 °F’. Use some judgement to
ensure that the text looks right, reads well and that there is no confusion or ambiguity.
♦ Use the degree symbol (°) for a temperature value e.g. 120°F, and not the word ‘degree’ or
‘degrees’ or ‘deg’. Do not use a superscripted zero i.e. 1200F, as a substitute for the degree
symbol.
♦ A useful way of expressing and emphasising a temperature difference is to use ‘degF’ to
differentiate it from a temperature reading ‘°F’.
♦ Use a hard space or a hard hyphen at the end of a line of text to avoid stranding an odd
letter, number, word, or part of a phrase. Use CTRL-shift-space or hyphen.
♦ Avoid contractions or abbreviations of standard terms e.g. use ‘laboratory’ and not ‘lab’,
use ‘2Poly’ and not ‘poly plant’.
♦ Do not use phrases like ‘basis the above’ or ‘basis this’. This terminology is an old Caltex
convention which was originally used for telexes. It is outdated and grammatically
incorrect. Instead use the full expression e.g. ‘on the basis of the….’.
♦ The first letter of a product stream name should be a capital e.g. Light Isomate or Heavy
Catalytic Naphtha (HCN). However do not use capitals for general streams e.g. hydrogen,
naphtha (unless it is a rundown product), kerosene, diesel.
♦ In general, use the term ‘refinery’ rather than ‘Refinery’.
 Page 24 of 34

6.0 WRITING STYLE (cont’d)

6.10 Other Guidelines (cont’d)

♦ The first letter of equipment names should be a capital e.g. HP Separator, Main
Fractionator, Product Stripper. Do not use capitals with generic references to equipment
e.g. pumps, heat exchangers, coolers.
♦ When using concentrations expressed as ‘ppm’ and in most cases, ‘percentage’, remember
to show the basis i.e. weight, volume or molar. Use ‘wppm’ not ‘ppmw’, ‘vppm’ not
‘ppmv’ and ‘wt%’ not ‘wt %’ or ‘%wt’.
♦ For pressures, remember to state the basis i.e. gauge (g) or absolute (a). For example, use
100 psig or 100 psia. However, for pressure drop use only ‘psi’, which is technically
correct.
♦ When using a slash i.e. ‘/’ do not leave a space between the word on each side of the slash
and the slash itself. For example, use ‘and/or’ and not ‘and / or’.
♦ Do not use ‘MT’ as an abbreviation for ‘metric ton’. The latter term actually means
1000 kilograms (kg). It is written as ‘tonne’ and has the symbol ‘t’. Therefore define
‘tonne’ as 1000 kg in the document or table, and thereafter use either the word or the
symbol, as necessary.
♦ Use the symbol ‘M’ as the multiplier equivalent to 1000. This is the standard at Bapco. Do
not use the symbol ‘k’ or ‘K’. The latter actually means kelvins, which is a measure of
absolute temperature.
 Page 25 of 34

7.0 EMAILS AT WORK (‘Engineering Emails’)

7.1 Introduction
In business, the ever-present email has to some extent, replaced the formal memorandum as the
main medium for transferring information. Unfortunately, the same care, attention to detail and
application of rules and conventions for writing and presentation, which are used for formal
documents, are often missing in email messages. The email has become so important in our
normal business and there is so much potential for getting it wrong and creating potentially
serious problems, that it requires a separate, critical review.

We live in the electronic age where essentially instantaneous communications e.g. email and
text messaging, are a way of life. Consequently, there is a tendency at work to use the same
rather informal, colloquial or even abbreviated style of writing that is routinely used in
non-work related communications. More often than not, this means that grammatical
correctness, spelling, punctuation, and structure are not considered important, and that rules
and conventions may be ignored.

While this could be considered acceptable outside work, this is not the case in business.
Therefore, it’s very important to remember that emails are still technical documents and that
you, the engineer and author, must be steadfast and vigilant in guarding against bringing these
sorts of private communications habits into the workplace. Regard any email message as a
formal memorandum. As with more traditional formal documents, emails are still “a written
mode of communication that highlights a person’s ability - or inability - to write”.9

Generally, emails should be used for short messages, either to convey specific information,
short answers or advice on particular issues, or as the modern equivalent of a covering
memorandum which relates to attached documents.

Emails are not suitable for presenting a large body of information or tables. It is highly likely
that any formatting will be lost in the sending and receipt of the message. This means that the
message can come out garbled. Therefore, for reasonably sized documents or where formatting
is important e.g. tables are required, use attachments prepared in the word processor, which
allows normal rules and conventions to be followed.

7.2 Main Governing Points

There are some important general points governing the use of emails which must be observed:

1. Email is not a substitute for meeting and talking with the recipient(s), or for telephone
conversations. So in the majority of cases, talk to the recipient before sending your
message. The email is then used to summarise the discussion. The main advantages are that
the recipients will not be surprised, the likelihood of misunderstanding is minimised or
avoided, you might learn something which will influence or moderate your approach, or
you may change the emphasis.

Therefore, in general, if you can’t start the message with “As discussed”, then don’t bother
sending it.

2. Emails must have a definite purpose.8 If they don’t have an objective, then the more you
say, and the more often you say it, the less likely it is that anyone will pay attention and the
more likely it is that you’ll create problems for yourself and the department.
 Page 26 of 34

7.0 EMAILS AT WORK (‘Engineering Emails’) (cont’d)

7.2 Main Governing Points (cont’d)

3. Emails are forever. They can be rapidly and widely distributed, easily stored and can come
back to haunt the author. They are their own records and can be held against you.
Microsoft, Enron, and others learnt this lesson the hard way through the courts in the US.
Therefore take great care in creating your email message, choose the words carefully and
make sure that it’s properly reviewed and approved. Include only content that is
appropriate to your professional position at work.8 The normal steps taken in preparing and
issuing a memorandum must be followed when generating an email.

4. Thoroughly read and check your message for correctness and relevance. Before issuing
emails have them checked by a colleague or supervisor, as appropriate. Too frequently
people send emails following first drafts. This typically leads to errors. Precisely because
emails can be prepared quickly and easily, and they may seem trivial, many people ignore
the proof-reading and checking step. This has led to many embarrassing and unfortunate
instances e.g. one message came from a very high corporate level, addressed to all lower
levels and areas, which described arrangements for various work shifts. Unfortunately the
author didn’t pick up the error in which the letter ‘f’ was missing from the phrase ‘all shift
employees’.

5. Do not forward emails to others for checking. This will only lead to confusion. The
company is not adopting a paperless office concept. So print out a hard copy of the email
and hand deliver it to the checker. Provide background information to the email to assist
the checker in effectively reviewing it. This is efficient as well as professionally courteous.
It is also important to realise that proof reading on the screen is never very effective and
most people will print a hard copy if they don’t have one. This has the added advantage of
highlighting any miscellaneous editing comments which have been mistakenly left at the
bottom of the document.

6. Keep messages short and to the point. Most people don’t read below the bottom of the
screen. Moreover, as stated above, use attachments for longer documents including tables
of data. This approach also allows the attached document to be properly formatted and
specified with a formal title, date, revision number and with ‘prepared by’, ‘checked by’
and ‘approved by’ blocks, to be used as appropriate.

7. Never use a table of data in the body of an email message because of doubts on formatting
and the risk of it being garbled at the recipient’s end. Tables must be given as an
attachment to the email and properly titled and referenced.

8. Clearly state the action you want to be taken, or your recommendation. Make sure that it is
not inferred or buried in many lines of text.

9. Do not issue emails to colleagues within the same department, unless it is to document a
discussion or agreed course of action. Emails must never be a substitute for walking across
or down the corridor to talk.

10. Do not use emails to circumvent correct communication channels and schedules of
authorities. There have been examples of people at a lower corporate level writing to
general managers or higher levels, with complaints or asking for information or follow-up,
and employees bypassing several layers of the management structure and requesting
meetings with the manager. The company’s chain of command must be respected.

11. Decide carefully on the addressee(s) and the distribution list. Ensure that they are the
appropriate and relevant people for your message. It is equally important that you do not
include someone who should not or would not want to receive your message.
 Page 27 of 34

7.0 EMAILS AT WORK (‘Engineering Emails’) (cont’d)

7.2 Main Governing Points (cont’d)

12. When replying to emails, do not just hit the “reply to all” button. Ensure that the recipients
i.e. the addressee(s) and those on the copy circulation list are relevant to the subject and the
message. Too many times an email is initiated and copied to multiple parties, many of
whom do not need to be copied, and probably do not want to be copied. Hence do not add
to the problem by thoughtlessly replying to everyone.8

13. Sometimes you will receive an antagonistic or abusive email, which is very upsetting and
can easily move you to anger. This sort of unprofessional behaviour goes on in any
organisation. In such a situation do not reply immediately, in the heat of the moment.
Allow yourself some breathing space to calm down, and discuss the matter with your
supervisor and/or colleagues. If you decide to reply, then prepare a formal, cold, technical
and factual reply. Your reply must be kept on a technical level highlighting the
deficiencies, whatever they might be, in the original message. It should not be a personal
attack on the originator of the email message. Typically send your reply the next day.

14. Never use emails to call for a meeting. Always talk to people in person or via telephone to
establish the mutually convenient date and time, and then follow up with a final email of
confirmation. If you don’t do this then you can end up with many, many emails, all with
different views, suggestions, and changes, all of which clog up the inboxes with useless
information.

15. To avoid the tone of your message and your attitude being misunderstood, never use
emoticons, and be careful with your use of conventional punctuation e.g. even an
exclamation mark can appear as very strong and threatening.

16. As everyone is well aware, we are inundated with emails at work. Supervisors can easily
get more than 50 in a day. Most emails are not needed, but unfortunately all have to be
read at some stage in case there is a real request or action item. Therefore, contribute to
improving the system by ensuring that your messages really are required and by copying
only those who really are concerned with the subject. Do not use general copy lists.

7.3 Email Etiquette and Conventions

Although the presentation of emails is less formal than that of business letters, there is still a
certain etiquette and style to be followed.
♦ Use Arial size 10 font.
♦ Use black rather than different colours for the text, which do not print or copy well.
♦ If you have edited a message for forwarding, then it is likely that the changes are shown in
a different colour. Therefore, before sending the message, ensure that all the text is in
black.
♦ Do not use coloured, ‘flowery’ or ‘artistic’ backgrounds in your email messages. They are
distracting, frustrating and annoying and have no place in a professional document.
♦ Start the message with the name(s) of the recipients only, and ensure that the names are
spelled correctly to avoid inadvertent insults.
♦ It is also acceptable not to have any names but just start with the message.
♦ Do not open with “Dear” – this is reserved for formal letters only.
♦ Do not open with “Gents” – this is reserved for the Men’s Toilets. If you wish, use
“Gentlemen”, but ensure that you do not use this opening if the audience includes ladies.
Hence, as mentioned above, it is better not to use any opening at all.
 Page 28 of 34

7.0 EMAILS AT WORK (‘Engineering Emails’) (cont’d)

7.3 Email Etiquette and Conventions (cont’d)

♦ End the message with a simple complimentary closing of “Regards” or “Thanks”, followed
by your name and possibly your telephone extension. It is also acceptable to leave out the
complimentary closing. In any case, do not make the complimentary closing into a
grandiose or subservient gesture – it’s not needed and is not appropriate for our business
environment.
♦ In general, and especially within the company, do not add any other details after your name
e.g. position, title, telephone numbers (work, home, mobile), fax, and email address. This is
a waste of space. Besides, in most cases the recipient already has this information,
especially at Bapco, or if they need it, they’ll ask. This practice may also be regarded as
showy and egotistical. There are some exceptions e.g. when first contacting an outside
organisation or company.
♦ Choose the subject title carefully. It must be relevant, succinct, and in some cases may
even effectively summarise the purpose and content of the message e.g. a meeting time and
date.
♦ Do not leave the subject line blank, or use an irrelevant title – this leads to confusion and
creates some doubt about the subject.
♦ Do not reply to an email, retaining the subject, and then write about a different topic. This
is lazy and confuses the reader, and it’s unprofessional.
♦ If you need to revise an attachment to an email, save the attachment as a separate file with
a new name, revise and save the document, and then attach it to your outgoing email
message. Do not update the existing attachment and keep the same file name. This is
procedurally wrong, it is confusing at the very least and it could have serious implications
if the information is misinterpreted or wrongly used.
♦ Do not adopt the practice of repeatedly replying to emails on one subject, and allowing
previous messages to accumulate at the bottom of the current message. This is
unnecessary, certainly a waste of space and probably paper, and it doesn’t have any
benefit. No one seriously wades through pages and pages of earlier messages. Therefore
reply and keep only the last one or two messages, if necessary. File earlier messages.
Avoid the ridiculous situation of having ten or twenty pages of messages as part of your
current message. Discourage the practice in others. Use electronic filing and retrieval to
store and research the background to the subject.
♦ Do not request a ‘reply receipt’ on your email. This practice can appear presumptuous and
threatening and annoys the recipients. It suggests that the originator equates ‘receipt’ with
‘acceptance’ and ‘action’ on behalf of the recipient, which is just not true.
♦ Since the ‘send’ button and the other editing functions are on the top toolbar, it’s easy to
send a message before your have finished writing it, and before it has been reviewed by
others. The best way to avoid prematurely sending your message is to ensure that you
don’t put the addressee or cc list into your message until it is final i.e. until after it has been
reviewed and approved by your supervisor and others who are concerned.
♦ Remember to actually attach the attachment, and check that the attachment is what you
want it to be. Open the attachment and check it before sending any message.
♦ Use complete, grammatical sentences. Emails are not the modern equivalent of telexes in
which it was a point of honour to minimise the number of words, as well as the cost. There
are examples of people who still use this minimalist approach and their messages are
sometimes incomprehensible. In particular make sure that you use the indefinite article ‘a’,
and the definite article ‘the’, as appropriate.
 Page 29 of 34

7.0 EMAILS AT WORK (‘Engineering Emails’) (cont’d)

7.3 Email Etiquette and Conventions (cont’d)

♦ Emails must be grammatically correct. Do not use shortcuts and abbreviations that are
accepted for private communication and ‘sms-ing’ e.g. do not use “u r l8” but “you are
late”.
♦ In general be polite e.g. the genuine use of the words ‘please’ and ‘thanks’ will certainly
enhance your message. However there are also occasions when your message needs to be
very formal and firm, and to achieve this tone, the words, structure and even the
distribution list need to be carefully selected.
♦ Even though emails are generally short documents compared to memoranda and reports,
they must be structured in the same way. Use subtitles and bullets, as appropriate, to
present information clearly.
 Page 31 of 34

8.0 OTHER REQUIREMENTS, CONVENTIONS AND GUIDELINES (cont’d)

8.4 Distribution Lists (cont’d)

In general, avoid huge lists and make sure that you send the document only to people who
really need to know.8 This will include the people who may or will be making the decisions,
and those directly involved in the work or who will need to become involved, perhaps as part
of a team. Always include your immediate colleagues, other relevant engineers in the
department, and those who contributed to the work, especially those who provided information
and are quoted.

Not everyone needs to see all the attachments. Therefore you can have two lists: those to be
copied ‘without attach’, and those to be copied ‘with attach’.

For departmental recipients use initials e.g. SS, GML, HAQ, to indicate that each person gets a
separate copy. If the intention is to circulate one copy through several people, then use a slash
to separate the initials, e.g. SS/GML/HAQ. Do not put a space between the slash and the
initials. For recipients outside the department use abbreviated position titles, sometimes with
the initials or the name in brackets e.g. Supt FCCU-OPD or Supt P&U (EHM).

Include relevant TSD General Office file numbers at the end of the distribution list.

As the author you will probably need two or three copies e.g. one for the work file, one for the
group’s Outgoing Correspondence File and, if necessary, one personal copy for future
reference. This requirement should be signified by putting the number of copies in brackets
after your initials e.g. HAQ(3). Note that there is no space between the last initial and the
opening bracket.

Use font size 9 or 10 so that the distribution list doesn’t overshadow the conclusions,
recommendations and signature(s), which are normally on the final page. The font size should
preferably be less than that used for the attachments, but can also be the same size.

8.5 Electronic File Names and Filing

In the bottom left hand corner of the last written page of the main body of the text of the
document, state the document electronic file name and path so that it can be found and revised,
or used as a starting point for another document in future. Make it clear that it is the document
file name - use “file:” followed by the name. Do not put the file name in the footer.

Keep the file name appropriate and succinct. Use appropriate abbreviations. Be careful if using
the whole document title as the file name – this could mean ending up with a long sentence,
which would be a waste of space and could create a suggestion of laziness or lack of
imagination.

Consider starting the file name with a letter which indicates the type of document e.g. ‘l’ for
letters, ‘f’ for faxes, ‘m’ for memoranda, ‘t’ for transmittals, ‘n’ for notes and so on.

Since documents will be modified and you may want to keep earlier revisions, or there may be
several in a series e.g. routine reports, include in the file name an abbreviated date and revision
number e.g. ‘-01’ or ‘revA’. Examples of file names are:

♦ file: c:\mydocuments\work\tsd style guide-revB.doc

♦ file: c:\work\process descriptions\descr-rxn-regen-01.doc
♦ file: p:\catalytic\lbo\edp\l-min-aug04.doc
 Page 30 of 34

8.0 OTHER REQUIREMENTS, CONVENTIONS AND GUIDELINES

8.1 The First Page

In a memorandum or file memorandum:

♦ Put the document type, and if needed, the addressee, in bold, at the top left hand corner e.g.

Memorandum
To: Manager - OPD

♦ The subject heading for most documents must be centred at the top of the document. Do not
cram it into the top right hand corner as was done many years ago before word processing
was available.
♦ Do not put a date on the document. This will be done by the General Office when the
document is issued.

8.2 The Signature Page

Usually, the last page of the main body of the text will contain the signatures e.g. the author of
the file memorandum, the Manager Technical Services, or a list of authorities who will sign the
document indicating agreement, endorsement and approval.

The signature page must always include the “punch line” of the document i.e. the over-riding
conclusion(s) and recommendation(s), the requested action, which must be clearly stated, and
the costs which are involved. Apart from being professionally courteous and business-like, it is
vital to ensure that the signatories know what they are signing for, and why.

It is preferable to have the list of attachments, the references and the distribution list on the
same page but this is not always possible. It is therefore acceptable to put this information on
the following page, if necessary.

8.3 Attachments

The ‘Attachments’ or ‘List of Attachments’ should be presented as a proper numbered list

below the signatures. Use the full title of each attachment and show the number of pages in
brackets at the end of each attachment. For example:

1. “……………………………….” (3 pages)
2. “……………….” (2 pages)

Use a font size which is two sizes smaller than the text i.e. font size 9 or 10, to minimise space
requirements, and to ensure that the list does not overshadow the main text.

8.4 Distribution Lists

The distribution list i.e. the ‘cc: list’ always comes as the last section on the last page of text. It
is not numbered. Although distribution lists may seem trivial, they must be given very careful
consideration.
 Page 33 of 34

9.0 REFERENCES
The references used in the development of this Technical Writing Style Guide are given below.

As indicated in Section 6.1, a good, comprehensive dictionary is an invaluable aid when it

comes to written work. It can be used to check or confirm the correct spelling of words, their
pronunciation, meanings and origins. Depending on the depth of the book it may also contain a
wealth of other information e.g. Latin and Greek phrases, abbreviations, chemical elements and
conversion factors.

A thesaurus can also be a useful source of new words and can be used to enhance one’s
vocabulary. It is a good tool for experienced writers but it can also be a trap for inexperienced
engineers, and especially for those whose first language is not English. This is because although
some words may have a similar meaning, it may not be appropriate or possible to use them in
every context. For example, several years ago, one of our enterprising engineers, when
attempting to describe the unit performance as good, stated that the performance during the test
run was “blue riband”. This was the result of the use of the thesaurus and it was clearly the
wrong choice of words, but it was an interesting learning opportunity. Therefore be careful when
using a thesaurus.

Probably the best single-volume dictionary of British English is the Chambers 21ST Century
Dictionary.14 It’s possible to search both the Dictionary and its companion thesaurus, on the Web
at www.chambersharrap.co.uk but there is really no substitute for having a hard copy to hand.
Moreover, the very action of searching for a word and having to look at other words and subjects
is instructive in itself and a great learning activity.

The Chambers Thesaurus 15 is an excellent reference, but probably the most traditional, well
known and most comprehensive thesaurus is still Roget’s Thesaurus.16

The Chambers 21ST Century Dictionary also includes information on abbreviations and
acronyms, but a more extensive and up-to-date listing may be found on the web at
www.acronymfinder.com.

The references used in this style guide are:

1. Hughs, B. (General Editor), Drury, J., and Barrett, M., “The Penguin Working Words”, Penguin
Books Australia Ltd., 1993, ISBN 0140513515
a. pp 78 – conventions for the use of brackets
b. pp 159 – conventions for dates
c. pp 171 to 175 – general discussion of desktop publishing including the importance of readability,
white space, headings, avoidance of mixtures of different font types and large slabs of sanserif
text
d. pp 229 – definition and discussions of fonts
e. pp 298 – description and application of justified and unjustified text
f. pp 531 to 534 – technical definitions and review of type faces, fonts and different formats
g. pp 564 to 565 – the importance of word spacing
h. pp 293 – the use of italics
i. pp 504 – the meaning and use of “status quo”
j. pp 211 – the use and value of the abbreviation “etc.”
k. pp 158 – conventions for, and comments on, the use of dashes
l. pp 514 – tautology
m. pp 386 – page numbering conventions
 Page 34 of 34

9.0 REFERENCES (cont’d)

2. Peters, P., “Australian English Style Guide”, Cambridge University Press, 1995, ISBN 0521576342
a. pp 410 – the use of italics
b. pp 184 - convention for dates
c. pp 664, 685 - roman type face and serif fonts
d. pp 683 – sentences and style
e. pp 254 – the meaning and use of “et cetera”
f. pp 563 – the use of brackets
g. pp 405 – the etymology, conventions, and practicalities of ‘ise/ize’ endings

3. Eszes, R., “Communicating Your Message Effectively”, Engineering Dimensions, July/August 1994,
pp 25-26
4. Boland, A., “Got Report-O-Phobia?”, Chemical Engineering, March 1996, pp 131-132
5. Bly, R.W., “Avoid These Technical Writing Mistakes”, Chemical Engineering Progress, June 1998,
pp 107-112
6. DiGregorio, K., “Pick Up the Pace of Your Writing”, Chemical Engineering, December 1991,
pp 117-122
7. “Communication and Working Effectively”, issued 20 October 1999 by TSD Catalytic Processes
Section, prepared by G.M. Lilburne, S. Soydaner and S. Hamza Alawi
8. Venables, J. (editor), “Communication Skills for Engineers and Scientists”, Third edition, 2002,
published by the Institution of Chemical Engineers
9. Hunt, E., “The New Rules of Effective Communication”, Engineering Dimensions,
November/December 1997, pp 30-31
10. Truss, L., “Eats, Shoots & Leaves – The Zero Tolerance Approach to Punctuation”, 2003, published
by Gotham Books (Penguin Publications), ISBN 1-592-40087-6
11. Nadziejka, David.E., “Make Your Writing Stand Out”, Chemical Engineering, August 1989,
pp 147-152
12. “Fowler’s Modern English Usage”, second edition (1983), revised by Sir Ernest Gowers, pp 72, on
the use of capital letters, and pp 314, covering the –ize and –ise endings in verbs, origins, preferences
and practicalities
13. Mackay, D. “Writing for Engineers: A Guide to the Preparation of Technical Reports and Theses”,
Department of Chemical Engineering and Applied Chemistry, University of Toronto, 1992,
(http://chem-eng.utoronto.ca/~writing/index1.html)
14. “Chambers 21ST Century Dictionary”, second edition, 1999, ISBN 055014210X,
ISBN-13: 978-0550142108
15. “Chambers Thesaurus”, new edition 2004, ISBN 0550100679, ISBN-13: 978-0550100672
16. “Roget's Thesaurus of English Words and Phrases: 150th Anniversary Edition,”, Edited by George
Davidson, 2006, ISBN 0141004428, ISBN-13: 978-0141004426
17. ISO 1000:1992 – “SI units and recommendations for the use of their multiples and of certain other
units” and AS ISO 1000 – 1998, “The international system of units (SI) and its application”, see
examples in Section 5.2, and Annex B, which infer that multiples of 1000 should be separated by a
space rather than a comma, but note that the Standards use a comma rather than a full stop for the
decimal place

Note
The sentence in §6.1d should be punctuated as follows: “That, that is, is, that, that is not, is not.”