• No results found

Corrosion Handbook on Stainless Steel

N/A
N/A
Protected

Academic year: 2021

Share "Corrosion Handbook on Stainless Steel"

Copied!
142
0
0

Loading.... (view fulltext now)

Full text

(1)

Sandvik Steel

Corrosion Handbook

Stainless Steels

(2)

Corrosion Handbook

Stainless Steels

(3)
(4)

Sandvik Steel

Corrosion Handbook

Stainless Steels

(5)

S-811 81 Sandviken, Sweden

Phone +46 26 26 30 00

Fax +46 26 25 17 10

www.steel.sandvik.com

Corrosion Tables

© 1999

Page 1 – 45 AB Sandvik Steel

Page 1 – 88 Avesta Sheffield AB and AB Sandvik Steel

ISBN: 91-630-2124-2

(6)

Corrosion Handbook

for stainless steels

P R E F A C E

When first introduced in 1994 this “Corrosion handbook for stainless steels”

replaced an earlier edition published by Jernkontoret, Stockholm, Sweden,

which was jointly produced by the Scandinavian manufacturers of stainless

steel in 1979. Being a unique source of information for material specialists and

designers, it has been highly appreciated.

Continued materials research has resulted in new grades and improved

properties of the existing grades. New corrosion tests are continuously being

carried out, often reflecting the more aggressive environments to which the

materials are being exposed. A combination of these factors has motivated a

revision of the corrosion tables. The present revised and extended edition is

the result of a cooperation between Avesta Sheffield AB and AB Sandvik

Steel in Sweden.

As an introduction, a series of papers are presented on the corrosion theories

in connection with stainless steels. Corrosion testing, different steel types and

grades, general aspects on different applications, as well as fabrication aspects

are also discussed. These papers are then followed by corrosion tables and

graphs describing the resistance of various materials to different environments

(in alphabetical order), concentrations and temperatures.

We are pleased to see that this work, which is based on more than 70 years'

experience in solving corrosion problems with stainless steel, is made

avail-able to industry. It is my belief that this corrosion handbook will be a valuavail-able

tool for all material specifiers when designing the process plant and equipment

of today and tomorrow.

Sandviken, March 1999

Per Ericson

(7)
(8)

Introduction

I:8

Corrosion of metals

I:9

Different corrosion types and test methods

I:10

General corrosion

I:10

Galvanic or two-metal corrosion

I:11

Intergranular corrosion

I:12

Pitting corrosion

I:14

Crevice corrosion

I:16

Stress corrosion cracking

I:17

High temperature corrosion

I:21

Introductrion

I:21

Oxidation

I:22

Catastrophic oxidation

I:22

Sulphidation

I:22

Carburisation and nitration

I:23

Molten metal corrosion

I:23

Halogen corrosion

I:23

Erosion-corrosion

I:23

Applications

I:23

Composite tube applications

I:25

Stainless steels

I:26

Introduction

I:26

Austenitic stainless steels and

Duplex stainless steels

I:27

Manufacturing programme

I:28

Special stainless steel grades

I:28

Nickel base alloys

I:29

Titanium

I:30

Zirconium

I:30

Applications for stainless steels

I:31

Chemical industry

I:31

Urea production

I:35

Oil and Gas industry

I:37

Corrosion in petroleum refining and

petrochemical applications

I:40

The pulp and paper industry

I:43

Fabrication

I:46

Constructional design

I:46

Bending

I:48

Expanding into tube sheets

I:48

Surface properties

I:49

Steel grades – Manufacturing programme

I:50

Corrosion tables

II:1

Isocorrosion diagrams:

Acetic acid

II:2

Chromic acid

II:16

Citric acid

II:16

Fluosilicic acid

II:20

Formic acid

II:22

Hydrochloric acid

II:24

Hydrochloric acid with chlorine

II:24

Hydrofluoric acid

II:26

Lactic acid

II:30

Nitric acid

II:35

Oxalic acid

II:39

Phosphoric acid

II:41

– with chloride additions II:42

– with fluoride additions II:42

Sodium hydroxide

II:54

Sulphuric acid

– deareated

II:59

– naturally areated

II:59

– with chloride additions II:60

– with iron sulphate

II:60

– with chromic acid

II:61

– with copper sulphate

II:61

Tartaric acid

II:70

Physical tables

II:74

Density, modulus of elasticity and coefficient

of linear expansion of stainless steels

II:74

Thermal conductivity of stainless steels

II:74

Physical properties of certain chemical elements II:75

Temperature conversion table

II:76

Chemical elements

II:78

Degrees Baumé

II:79

Vapour pressure of water

II:79

pH values:

alkaline solutions

II:80

acid solutions

II:80

foods

II:80

substances in human body

II:80

hydrochloric acid, nitric acid and sulphuric

acid solutions

II:81

Relationship between weight-% and density,

molarity, volume-%, kg/litre and degrees Baumé:

acetic acid

II:81

ammonium hydroxide

II:82

formic acid

II:82

hydrochloric acid

II:83

nitric acid

II:83

phosphoric acid

II:84

potassium hydroxide

II:84

sodium hydroxide

II:85

sulphuric acid

II:85

Glossary

II:86

Disclaimer

II:88

(9)
(10)

1.

Introduction

The first part of this handbook constitutes an introduction to corrosion together with a brief description of stainless steel grades and special metals. Different corrosion types are dis-cussed comprehensively in connection with some relevant test methods, some of which have been used to gather the data in the latter part of the handbook. Fabrication of stainless steel products is described and some advice in this area is given. Finally, a number of applications are reviewed for which the corrosion problems are discussed in more detail and materials selection for especially demanding environ-ments is suggested.

The corrosion tables comprising the last half of this corrosion handbook have been produced in cooperation between AB Sandvik Steel and Avesta Sheffield AB. They are intended to constitute a guide to the corrosion resistance of the included stainless steel grades. Different levels of corrosion rates are shown together with indications of local attack, such as pit-ting. A large number of different environments are included and in many cases also a wide range of concentrations and temperatures. You are hereby provided with a useful key in

the choice of material for a certain application. The data can also be used when the corrosion resistance of used or recom-mended grades is discussed with regard to changes in con-centration or temperature. To further illustrate the corrosion resistance of stainless steel grades several diagrams have been included, showing isocorrosion curves etc.

The subject of stainless steel corrosion is a vast area and can-not be completely covered in this handbook. For detailed information about corrosion types in different environments several references may be given [cf ref 1-4]. Standard tests of corrosion resistance are thoroughly described in reference 5. Corrosion research is an ever ongoing process with innumer-able articles being published every year. The state of the art can be found in the journals focussing on corrosion, of which reference 6 to 9 are recommended here. Lectures are also continuously being published within AB Sandvik Steel con-cerning the properties and experiences of Sandvik steel grades and special metals. A catalogue of titles can be re-quested from your nearest Sandvik Steel sales office.

1. M.G. Fontana, Corrosion Engineering, 3rd Edition, McGraw-Hill, 1987.

2. Corrosion Mechanisms in Theory and Practice, P. Marcus and J. Oudar, Eds., Marcel Dekker Inc, 1995. 3. H.H. Uhlig and R. Winston Revie, Corrosion and Corrosion Control, 3rd Edition, John Wiley and sons, 1985. 4. G. Wranglén, An Introduction to Corrosion and Protection of Metals, Institut för metallskydd, 1972. 5. Corrosion tests and standards: Application and Interpretation, John Baboian, Ed., American Society for

Testing and Materials, 1995 6. Corrosion Science 7. British Corrosion Journal 8. Corrosion, NACE 9. Werkstoffe und Korrosion

(11)

C o r r o s i o n i s g e n e r a l l y defined as a dissolution of

a material due to a reaction with the surrounding environ-ment. Most metals are in a thermodynamically unstable form, and corrosion often means that there exists a thermodynamic driving force for recombination of the unstable elemental form to the chemically stable oxidized form found in nature. Many metals can maintain their unstable form, in spite of the thermodynamic driving force, thanks to their ability to pas-sivate. Passivation means that the surface of the metal is cove-red by a, usually very thin, layer of corrosion product. This so called passive layer, which in most cases consists of an oxide film, separates the metal from the surrounding environment and hence the corrosion resistance is considerably increased. The ability to passivate for pure iron is limited, which means that iron will show a relatively low corrosion resistance in most media. The ability to passivate is however increased by alloying with chromium. At a chromium content of about 13% the alloy shows a considerably better resistance to

cor-rosion and has become what is usually called stainless. The reason is that the chromium forms an oxide layer on the sur-face, and this layer sufficiently protects the metal from the surrounding environment when the chromium content is about 13% or more. Other common elements that are used to improve the ability to passivate are molybdenum and nitro-gen.

When passivity cannot be maintained, due to a too aggressive environment, the metal will be exposed to the surrounding environment and a dissolution of the metal may take place. The attack initiates at weak spots on the surface, such as scratches or contaminated areas, and is then continued as local or uniform corrosion. The rate at which the dissolution takes place is usually measured in terms of e.g. mm/year. In the case of local attack, such as pitting or stress corrosion cracking, the corrosion rate is not a relevant parameter. Instead, any signs of local attack should be seen as a warning not to use the material under those specific circumstances.

2.

Corrosion of metals

Heat-exchanger tube of Sanicro 29 heavily attacked by erosion corrosion on the inside of the bend. The cause was sand in the cooling water.

(12)

W h e n c o r r o s i o n o c c u r s the attack is characterised

by the way in which the metal is dissolved. Many forms of corrosion exist, and here the most common types occurring for stainless steels are described.

General corrosion

General corrosion is characterised by a uniform attack over the surface of the material when exposed to a corrosive me-dium. It is therefore possible to define a corrosion rate (r), often stated as a mass loss per unit surface area (g/m2h), or as

a mean metal loss per unit time (mm/year). The latter unit is used for tabled data as well as in the diagrams presented in this book. It is, however, sometimes desired that the corrosion rate should be expressed in mils/year (milli-inch per year or mpy). Corrosion rates in mm/year are easily translated to rates in mpy from the relationship:

r (mpy) = 39.4 x r (mm/year)

Isocorrosion diagrams illustrate the resistance of metallic materials to general corrosion. Each isocorrosion line repres-ents a fixed corrosion rate and the dependence of concentra-tion and temperature on the corrosive medium can be shown.

Figure 1. Curves representing a corrosion rate of 0.3 mm/year for Sanicro 28 and three other alloys in wet-process phosphoric acid at 100°C. The combined effect of chloride and fluoride is shown.

The most common environments where general corrosion occurs on stainless steels are strongly acidic or alkaline solu-tions. The specific composition of the environment is crucial for the corrosiveness, and may be drastically changed if ox-idising or reducing compounds are added. The performance of stainless steel grades can vary considerably in the same envi-ronment and to different additives. It is therefore of great importance that the environment where a product is to be used is thoroughly characterised. When this is done a suitable material can usually be selected. The economic advantages of choosing a grade with high corrosion resistance, sometimes acquired at a higher price per kilo, can be illustrated by esti-mations of the life cycle cost.

T E S T I N G O F G E N E R A L C O R RO S I O N

Testing of general corrosion is usually performed by exposing samples to the corrosive environments over specific time intervals and calculating the corrosion rates from weight los-ses. To avoid irregularities from the initiation period the sam-ple is weighed after each period and the first value is dis-regarded. At least two samples of each material should be used and the corrosion rate is determined as the calculated mean value for two or more periods.

Figure 2. Sandvik 5R60 (middle) and 5R10 (right) tested in 60% H2SO4 at 100°C for 24 hours. Unexposed test coupon to the left.

3.

Different corrosion types

and test methods

F -, % 1.5 1.0 0.5 0 200 400 600 800 Cl-, ppm 2302b H3PO4 H2PO4 Fe3+ = 70 % = 4 % = 0.45 % Sanicro 28 Alloy 20Cb3 Alloy 904 Alloy 825

(13)

Another way to calculate corrosion rates is to use an electro-chemical technique. The corrosion current, I, is proportional to the amount of oxidised metal if no competing oxidation reaction is taking place. By this method changes in corrosion rate can be registered whilst adding corrosives or inhibitors, raising the temperature etc.

Galvanic or two-metal corrosion

When two different metals are used in the same environment they often obtain different potentials. If they are in contact or otherwise electrically connected a sufficient potential differ-ence might produce a flow of electrons between them. The more noble material becomes cathodic and the less corrosion resistant anodic. This often results in increased corrosion of the anodic material and a decreased attack of the cathodic. This phenomenon is used for cathodic protection, where a so called sacrificial anode is connected to the material to be pro-tected. When coupling different stainless steel grades poten-tial differences are generally too small to cause galvanic cor-rosion problems.

The ranking of materials with regard to potential can be found in galvanic or EMF-series (electromotive force). Low EMF values are found for magnesium and zinc, among others, whe-reas copper, platinum and gold have high values. Standard EMF values have been calculated for standard conditions but the order between metals often differs depending on the envi-ronment. For metals such as titanium or aluminium, which form very protective oxide layers even at room temperature in air, the low standard values of potentials can seldom be repro-duced. For a more realistic ranking of metals galvanic series have been determined empirically for special environments, such as sea water (see table 1). When evaluating the risk of galvanic corrosion it is of great importance that the correct potential values are used, determined for the right solution and temperature. In the same way, when sacrificial anodes are used, and thus galvanic corrosion is desired, the material which functions as an anode can become cathodic to the material to be protected in certain environments. This is the case when zinc is used for the protection of steel, if the con-centrations of carbonates become too high. Galvanic corro-sion of stainless steels in the passivated state is unusual. When in connection to other metals, such as copper or carbon steel, stainless steels will generally be cathodic.

The relative area of the anode compared to the cathode greatly influences the corrosion rate. The larger the cathode area is compared to the anode area the faster the corrosion will be.

Table 1. Approximate EMF-values in sea water

Material Free corrosion potential / volts SCE*

Magnesium -1.6

Zinc -1.0

Aluminium alloys -1.0 to -0.75

Mild steel, cast iron -0.65

Copper -0.34

Admiralty brass -0.30

Ferritic stainless steel, 430 -0.24

Nickel 200 -0.2 to -0.1

304-type stainless steel** -0.08 316-type stainless steel** -0.10 to 0

Titanium -0.05 to +0.05

Platinum +0.2

*Saturated Calomel Electrode

** For passivated material values well above +0.1 can be obtained

As an example of this steel bolts in a more noble copper sheet corrode more quickly than steel sheet with copper bolts, in the same environment. When the metals are in contact with one another the rate of anodic corrosion often decreases with increasing distance from the cathode, i.e. the material loss is greatest close to the cathode. This is especially pronounced when the surrounding electrolyte has low conductivity. There also exists a related corrosion type sometimes called indirect galvanic corrosion. In this case the metals need not be in direct contact. Instead the more noble metal corrodes by uniform corrosion and its metal ions are transported, e.g. in solution, to the surface of the anodic metal. They are there reduced and at the same time enhanced oxidation of the anodic metal occurs. The corrosion of the anodic metal can accelerate even further if the noble metal atoms on its surface can act as sites for more effici-ent reduction of other species, e.g. from the electrolyte.

C O R RO S I O N T E S T S F O R G A LVA N I C C O R RO S I O N

Tests for galvanic corrosion are rarely performed as this is more for the designer's concern. It might, however, be very useful to check the metal combinations in the environment to be used. The corrosion rates for the metals in contact can then be compared to the rates of uniform corrosion of the isolated metals in the same environment. Another possibility is to measure the potentials of each metal against a reference elec-trode, or the potential difference between them, in the environ-ment in question. It is important, however, to remember that increased corrosion does not automatically follow from dis-similar potentials. Generally a potential difference of several tens of volts is required and in addition to this a decrease in galvanic corrosion rate is often observed if passivation occurs.

(14)

Intergranular corrosion

Intergranular corrosion is characterised by preferential dissolu-tion at the grain boundaries. Usually intergranular corrosion occurs in stainless steels that have been exposed prolonged time to certain temperatures (typically 600-900°C), so that a formation of chromium-rich carbides in the grain boundaries has taken place. Immediately adjacent to these carbides, in the outer parts of the grains, there will be a chromium depleted zone. These regions, which consequently have a lower corro-sion resistance than the rest of the alloy, will suffer from pref-erential dissolution.

Figure 3. Microstructure of a material surface with intergranular attacks.

To avoid intergranular corrosion, resulting from chromium carbides, one can choose a grade with lower carbon content, one alloyed with stabilising elements like titanium or niobium (typically in grades like 321, 347 etc), or avoid heating in the sensitising temperature range. The latter can be difficult to achieve in practice when, for example welding is carried out. Nitrides may also precipitate in grain boundaries during heat treatment, and in molybdenum-bearing stainless steels also the intermetallic phases, such as χ and σ. In a TTT-diagram (time-temperature-transformation) the time required for the forma-tion of precipitates during heat treatment is illustrated.

T E S T I N G F O R R E S I S TA N C E TO I N T E R G R A N U L A R C O R RO S I O N

Several methods exist for the testing of intergranular corrosion. Generally an oxidising, acidic solution is used, but pH, poten-tial and temperature depend on the method used. Because of their differences one must choose a method which is suitable for the steel grade and grain boundary composition to be tested. The applicability for some ASTM tests to austenitic stainless steels are summarised in table 2.

Intergranular corrosion in stainless steels may result from pre-cipitation of carbides, nitrides or intermetallic phases. Only in the most highly oxidizing solutions can intergranular attack be caused by intermetallic phases. When a test is to be restricted to carbides, in a material containing nitrides or intermetallic phases, a less oxidizing solution should therefore be chosen. Corrosion potentials of wrought stainless steel in different test solutions and the detectable phases are summarised in table 3. Table 2. Applicability of some ASTM standard practices in A 262 for testing of intergranular corrosion in austenitic stainless steels

PRACTICE TEST TEMPERATURE TIME APPLICABILITY EVALUATION

A Oxalic Acid Etch ambient 1.5 min Chromium carbide Microscopic: classification

Screening Test sensitization only of etch structure

B Ferric Sulphate boiling 120 h Chromium carbide Weight loss/

50% Sulphuric Acid corrosion rate

C 65% Nitric Acid boiling 240 h Chromium carbide Weight loss/

and sigma phase corrosion rate

D 10% Nitric Acid 70°C 4 h Chromium carbide Corrosion ratio

3% Hydrofluoric in 316, 316L, 317 compared to solution

Acid and 317L annealed specimen

E 6% Copper Sulphate boiling 24 h Chromium carbide Examination for

16% Sulphuric Acid fissures after bending

Metallic Copper

F Copper Sulphate boiling 120 h Chromium carbide Weight loss/

50% Sulphuric Acid in cast 316 and 316L corrosion rate

(15)

H U E Y T E S T

The Huey test (ASTM A262, practice C) means that the samples are boiled for 5 periods of 48 hours each in 65% nitric acid. The corrosion rate is calculated for each period from weight losses. For further information the maximum depth of attack may be measured, but this is not included in the stand-ard evaluation. The environment is strongly oxidising and should only be used as a check on whether the material has been correctly heat treated. The Huey test can therefore not be used to compare the corrosion resistance of different steels to other, less oxidising, environments. This test is suitable for the detection of chromium depleted regions as well as intermetal-lic precipitations, like sigma phase, in the material.

Figure 4. Intergranular corrosion testing by the Huey method.

S T R A U S S T E S T

A common way of investigating the resistance to intergranu-lar corrosion is to heat treat the sample in the sensitising tem-perature range and carry out a Strauss test (SIS 117105, DIN 50914, ASTM A262 practice E etc.). The samples are boiled in a solution of copper sulphate and sulphuric acid with cop-per turnings. The test time (15, 20 or 24 hours) depends on the standard used and the evaluation consists of a visual exam-ination for cracks originating from intergranular corrosion attacks. The samples are usually bent before examination. If cracks are suspected to arise from poor ductility a similar but unexposed sample should be used for reference.

Figure 5. Time-temperature-sensitisation curve for AISI 304 and Sandvik 3R12 (AISI 304L). Results from testing in boiling Strauss solution (12% H2SO4, 6% CuSO4). There is a risk of intergranular corrosion to the right of the curves.

100 Annealing time, h Temperature, °C (°F) 500 (930) 600 (1110) 700 (1290) 800 (1470) 0.1 0.5 1.0 5.0 10 50 Sandvik 3R12 AISI 304 1302b 0.05

Table 3. Corrosion potentials and detectable phases for wrought stainless steels in some acid solutions

SOLUTION CORROSION AUSTENITIC STEELS FERRITIC STEELS

POTENTIAL Cr-carbide Sigma Carbides and Nitrides Intermetallics

(VSCE) Fe-Cr Fe-Cr-Mo

65% HNO3 0.75-1.0 yes 316, 316L, yes yes yes

317, 317L, 321

Fe2(SO4)3 0.6 yes no yes yes yes

H2SO4 (321 possible (not σor χin

exception) unstabilised Fe-Cr-Mo alloys)

CuSO4 0.35 yes no yes yes no

H2SO4

As above but 0.1 yes no yes yes no

with metallic Cu

10% HNO3 -0.1-0.3 yes no yes yes no

3% HF

5% H2SO4 -0.6 yes no no yes no

(not σor χin unstabilised grades)

(16)

Whenever the origin of cracks is questionable a detailed metallographic examination should be performed to deter-mine the absence or presence of intergranular attack. This test method can detect chromium depleted regions in the material but cannot detect other possibly detrimental inhomogeneities, like precipitations of sigma phase. Figure 5 shows a so called time-temperature-sensitisation curve, where the minimum time for heat treatment before intergranular attacks appear is shown as a function of temperature.

S T R E I C H E R T E S T

The Streicher test (ASTM A262 practice B, ASTM G28) requires the samples to be immersed in a boiling solution of ferric sulphate and sulphuric acid for a period of up to five days. The test can detect chromium depleted regions in stain-less steels but cannot be used to detect susceptibility to inter-granular attacks associated with sigma phase in wrought stainless steels. The evalution of samples is done by calcu-lating the corrosion rates and may also be completed with the measured depths of attack.

Pitting corrosion

When pitting corrosion occurs the attack is localised to small areas on the surface where a break through of the passive layer takes place. This will create pits and possibly eventually holes in the metal. This form of corrosion is often more detri-mental than general corrosion, due to the local dissolution which can cause rapid penetration of the metal thickness. The nucleation time for pits depends on factors such as the ox-idising character of the environment, the concentration of aggressive ions such as chlorides, pH and the alloy composi-tion of the metal. The properties of the surface, for instance the presence of initiation sites at defects and inclusions, also effect the nucleation time. During the attack, however, sev-eral mechanisms act together to result in an autocatalytic pro-gress of pit growth. The environment within the pit becomes increasingly aggressive, due to anion selective diffusion into

the pit. Often a lid of corrosion products is formed leaving a very small hole at the surface which prevents dilution of the pit contents. Finally, the metal surface surrounding the pit mouth becomes cathodically protected through electron migration from the pit.

In many cases pitting corrosion is not detected until it has caused severe damage, such as a complete penetration in sheet or tube material. This is due to the very small pit holes formed on the surface and to the fact that the metal surfaces in many applications become covered with precipitates from process fluids or with thick layers of more or less protective corrosion products. The corrosion products from a pit attack are often found to create a lid on top of the pit, with only a very small opening. When examining the metal surface for pits it should therefore be thoroughly cleaned in order to reveal the pitted areas.

Figure 7. Test coupons after pitting corrosion test according to ASTM G48. From the left to the right: Sandvik 3R60, Sandvik 2RK65, SAF 2304, SAF 2205 and SAF 2507.

P R E - VA L U E S

The chromium content of stainless steel grades is important and alloying with molybdenum and nitrogen has proved very beneficial for the pitting resistance. From experimental data, relations between elemental composition and pitting resist-ance have been developed. These values are generally called PRE, pitting resistance equivalents, and can be used for an approximative ranking of stainless steel grades. Several forms are known of which one often used expression is showed below.

PRE = %Cr + 3.3% Mo + 16%N

In the duplex austenitic-ferritic stainless steels the nitrogen content is high, which promotes the pitting resistance. The high nitrogen content in SAF 2205 (nominal value: 0.18%) compared to some other 2205 type grades with a minimum of 0.08% N (for UNS S31803), is therefore beneficial. In the Corrosion products 7326 film Cl_ Cl_ Cl_ Fe2+ Fe2+ H+ H+ OH_ OH_ e _ e _ Metal Passive H2O+O2 H2O+O2

Figure 6. A corrosion pit with some possible chemical reactions.

(17)

duplex grades the PRE might, however, differ between the two phases. To avoid this duplex steel grades must be de-veloped with a balanced composition so that the elements of importance are partitioned to equal benefit in the two phases. This is the case for SAF 2507, for which PRE in both phases is greater than 40.

Table 4. Minimum PRE-values for some Sandvik stainless steel grades

Steel grade PRE (as defined above)

Sandvik 3R12 18.7 Sandvik 3R60 25.6 Sandvik 2RK65 33.0 SAF 2304 24.1 SAF 2205 35 SAF 2507 41 Sanicro 28 37.7 T H E E F F E C T S O F A L L OY I N G E L E M E N T S O N P I T T I N G R E S I S TA N C E

The effects on pitting corrosion resistance of alloying with for instance molybdenum or nitrogen have been investigated, but the picture is not yet completely clear. In the case of alloying with molybdenum improved metal passivation has been found. When pitting attack occurs the molybdenum assists in repairing the passive layer so that pit nucleation is stopped. According to one theory molybdate ions are formed from dis-solved molybdenum. The molybdate ions then remain at the outer surface of the diffusion layer so that it becomes cation selective. The aggressive anions, such as chlorides, are there-by prevented from reaching the surface. At the interface between the oxide and the diffusion layer anion selectivity prevails so that oxide growth can continue. After the initiation of attack increased amounts of molybdate ions have also been detected in pitted areas.

The effects of molybdenum seems to be enhanced by nitrogen which influences the molybdate concentrations at the surface. This has been explained by the production of ammonium ions which increases the pH which, in turn, makes the formation of molybdate ions more likely. Surface analysis has also proved that iron dissolution is increased with increased nitro-gen amounts, whereas the dissolution of chromium and molybdenum decreases. In alloys with increased nitrogen amounts the passive films have been found to contain higher ratios of chromium in the outer layer. Below this exists a thin layer enriched in nitrogen, nickel and molybdenum.

T E S T I N G F O R P I T T I N G R E S I S TA N C E

Ordinary methods using weight losses are not recommended for pitting tests because the overall material loss may be very small even though severe pits have developed. Evaluating the number of pits, their depths or localisation can give useful information about the mechanisms of initiation and growth, but is not adequate for evaluation of pitting resistance. It is therefore preferable to simulate conditions which result in pit initiation to provide a relevant means of assessing resistance to this type of corrosion attack. This is done by exposing the test material to an aggressive environment for a certain time. If corrosion pits are observed the material has failed.

T H E A S T M G 4 8 - T E S T

One immersion test method for pitting corrosion is the ASTM G48A-test, where samples are immersed in a 6% FeCl3 solu-tion. This solution is very corrosive due to the simultaneous presence of chloride ions and oxidising ferric ions. The temper-ature is held constant (e.g. at 22±2 or 50±2°C) and the recom-mended time for exposure is 72 hours. After cleaning the samples are weighed and the surfaces are examined for pits.

C P T D E T E R M I N AT I O N S

The temperature and the presence of oxidising agents are important parameters for pitting corrosion. A number of test methods have therefore been developed to determine the cri-tical values for these. The cricri-tical pitting temperature, CPT, is a value often requested. It should be noted that CPT values are specific for the environment and the test method used and these must therefore always be reported. CPT can be deter-mined by a modified version of the ASTM G48 test described above. A sequence of test periods are used at increasing tem-peratures. Between immersion periods the samples are removed, cleaned and examined for pits. The minimum tem-perature at which the material experiences pitting corrosion is defined as the CPT.

CPT can also be determined electrochemically. The most common ways are either to use a potentiodynamic method and measure at which temperature the break-through potenti-al drops, or to use a constant potentipotenti-al (potentiostatic method) and measure the temperature at which the current increases drastically. A discrepancy exists between results from diffe-rent methods but they are both frequently used to study differ-ences in pitting resistance of stainless steels.

Figure 8 shows the critical pitting temperature for different alloys measured potentiostatically at +300 mV SCE and with varying chloride concentrations. The potential +300 mV SCE is often found in natural sea water. Figure 9 shows the same

(18)

figure for other alloys and with a higher applied potential (+600 mV SCE), which corresponds to a potential often found in chlorinated sea water.

Figure 8. CPT-values for SAF 2205, AISI 304 and AISI 316 at varying concentrations of chloride. A potentio-static determination at +300 mV SCE and pH=6.0.

Figure 9. CPT-values at varying concentrations of sodi-um chloride, from 3 to 25%. A potentiostatic determi-nation at +600 mV SCE. CPT, °C (°F), 300 mV SCE 0 (32) 20 (68) 40 (105) 60 (140) 80 (175) 100 (210) Cl–, weight-% 4159b 0.01 0.02 0.05 0.10 0.20 0.50 1.0 2.0 AISI 304 AISI 316 SAF 2205 SAF 2507 6185b 5 10 15 20 25 Cl–,% 3 6 9 12 15 SAF 2205 25 Cr – Duplex CPT,°C (°F), 600 mV SCE 40 (105) 50 (120) 60 (140) 70 (160) 80 (175) 90 (195) 100 (210) NaCl, weight-%

Crevice corrosion

This form of corrosion is in principle the same as pitting cor-rosion, but occurs in crevices, e.g. between flange joints, under deposits on the metal surface or in welds with incom-plete penetration. A concentration cell is created with the anode in the crevice and the cathode on the outer surface. This corrosion form may be hard to spot, in the same way as for pitting, as it occurs in concealed places. When examining

the material on location crevices should be opened and the surface be thoroughly cleaned. In practice it is very difficult to entirely eliminate crevices in constructions. Crevice corro-sion often occurs at lower temperatures and at lower chloride concentrations than for pitting corrosion. Up to a certain limit, the risk for attack increases the more narrow the cre-vice is.

Figure 10. Schematic illustration of crevice corrosion under a washer.

(19)

T E S T F O R C R E V I C E C O R R O S I O N

Evaluation of crevice corrosion resistance may be done accor-ding to the ASTM G48B test. As in the pitting test (practice A) samples are immersed in a 6% FeCl3 solution. Before immersion crevice formers with specified properties are mounted on the samples. The critical temperature for crevice corrosion, CCT, may be determined in the same way as for CPT. One method for this is described in the MTI-2 standard, where the temperature is increased by 2.5°C. For reasons of investigating specific crevice formers, e.g. when the extrac-tion of aggressive ions may be suspected, modified tests are sometimes performed.

Figure 11. Comparison of CPT- and CCT-values for some stainless steels (obtained by the modified ASTM G48 method).

Stress corrosion cracking

Stress corrosion cracking (SCC) is an environmentally as-sisted cracking process, where a specific environment combined with tensile stress induces cracks on the metal surface. Stress corrosion cracking often occurs at increased temperatures, i.e. above 60°C, but cases where SCC has occurred at lower temperatures exist. The most common media where stress corrosion cracking occurs are chloride containing solutions, but in other environments, such as caus-tics and polythionic acid, problems with SCC may also appe-ar. Some enviroments that may cause stress corrosion crack-ing of stainless steels are listed below.

Table 5. Some environments where stainless steels are prone to stress corrosion cracking.

Acid chloride solutions Seawater

Condensing steam from chloride waters H2S + chlorides

Polythionic acid (sensitised material) NaCl-H2O2

NaOH-H2S

The mechanism of stress corrosion cracking is not well understood. This is mainly due to the specific features of SCC being the result of a complex interplay of metal, interface and environment properties. As a result of this different combina-tions of solution and stress are seldom comparable and the most reliable information is obtained from empirical experi-ments. During SCC the material does not undergo general corrosion and the phenomenon is sometimes considered to be one of activation/passivation interaction. It has been found that cracks often initiate in trenches or pits on the surface, which can act as stress raisers. The isolated times for pit ini-tiation, pit growth, crack initiation and fracture may, however, differ considerably between different materials.

Figure 12. Stress corrosion cracking of a tube.

In some cases crack initiation has been associated with the formation of a brittle film at the surface. The film developed at grain boundaries might, for instance, have lower ductility due to a different metal composition than the bulk material. At a certain film thickness and under stress this brittle film will crack and expose the underlying metal. New film growth will proceed with subsequent continued crack growth and so forth. The developed crack tip has a small radius and will develop a very high stress concentration. Even so, the stress condition alone is not sufficient for crack growth, but corrosion still plays a very large part. It has been shown experimentally that stress corrosion cracking can be stopped when applying cat-hodic protection, i.e. when corrosion is stopped but the stress conditions remain unchanged.

904L SAF 2205 25 Cr Duplex* 6Mo+N austenitic SAF 2507 CPT (°C) CCT (°C) *25Cr - 3Mo - .2N 90 80 70 60 50 40 30 20 10 6247b Temperature (°C)

(20)

Figure 13. Transgranular stress corrosion crack in Sandvik grade 2RE69 after autoclave testing in 1000 ppm chloride at 250°C.

Cracking may be either transgranular (TGSCC) or intergranu-lar (IGSCC) or, perhaps most usual, a combination of both. The material microstructure and alloying components are of major importance for crack paths as well as for SCC resist-ance. Alloying with Ni can make materials less prone to SCC and the duplex microstructure of the austenitic-ferritic grades is also beneficial. Standard austenitic stainless steels, like AISI 304 and AISI 316, are generally prone to SCC in chlo-ide containing environments at temperatures above 60°C, except at very low chloride contents, and therefore higher alloyed austenitics or duplex stainless steels should be used.

H Y D RO G E N E M B R I T T L E M E N T

Hydrogen embrittlement (HE) is sometimes stated to be a kind of SCC. This might, however, lead to serious misunder-standings as many discrepancies exist. Perhaps most import-ant is that HE cannot be reduced by cathodic protection, but might instead increase under such circumstances. The reason for this is that HE is caused by the penetration of atomic hydrogen into the metal structure. This, in turn, might occur when reduction of H+ is taking place on the metal surface, e.g. during cathodic protection in acidic environments. Several deposition techniques, such as electroplating, also involve reduction processes at the metal surface with the fol-lowing risk of hydrogen penetration and embrittlement. To avoid this, treated articles are often baked before use to remo-ve the hydrogen. The risk for HE is increased for harder metals, but the tendency to hydrogen cracking decreases with increasing temperature. Some differences between HE and SCC are illustrated in figure 14.

Figure 14. Possible mechanisms of cracking due to SCC and HE respectively.

S U L P H I D E S T R E S S C R A C K I N G

Sulphide stress cracking (SSC) might be defined as a variant of HE, but is sometimes treated as a special corrosion type. Sulphides are hydrogen-evolution poisons and as such pre-vent the hydrogen atoms formed on the metal surface from pairing up and dissolving as H2 into the surrounding solution. SSC has been found to cause severe problems especially in the oil and gas industry. A standard for material requirements in so-called sour environments has therefore been developed: NACE MR0175. Among the acceptable steel grades are SAF 2205, SAF 2507 and Sanicro 28. New grades can be accepted in NACE MR0175 after successful testing according to one of four methods described in NACE TM 0177. (See also chapter 7, oil & gas industry.)

T E S T M E T H O D S

The stress corrosion cracking resistance can be tested in a laboratory by different methods. Usually a constant load, a constant elongation or a slow strain rate is applied to the sample in a chloride containing environment. Several stan-dards have been developed regarding the stress application for SCC testing. U-bends, C-rings and bent-beam specimen are some examples of this.

7325 σ σ Hydrogen embrittlement H++ H H H H H Crack growth direction

σ σ

Men+ Men+

Anodic stress

corrosion cracking Time to cracking

Anodic current M Men+ + n e-

Immunity

Cathodic current 2e- + 2H+ 2H Crack growth direction

(21)

Constant load can be applied to the sample simply by the use of weights as shown in figure 16. Another way is to strain the sample with the aid of a spring. Regardless of method certain requirements must always be met by the testing equipment. Crevices and galvanic corrosion must be avoided and all exposed parts must be resistant to any kind of corrosion in the environment.

Figure 17. Results of SCC test with constant load in 40% CaCl2 at 100°C (210°F) with aerated test solution.

Figure 17 shows examples of results from constant load tests according to a modified ASTM G36 method, where 40% CaCl2 at 100°C is used as the corrosive medium. The time to failure versus load is recorded and a threshold stress is deter-mined, below which SCC does not occur. In figure 18 results from constant load tests (stress equals yield strength) are presented for which a spring was used to apply the stress. The mounted samples were tested in an autoclave at different tem-peratures and chloride concentrations. This method has been verified to correspond to practical cases, as illustrated in the figure.

Figure 18. A compilation of practical experience and laboratory SCC test data of 3RE60.

The drop evaporation test is another common method to mea-sure SCC-resistance. With this method a specimen is electri-cally heated and subjected to a constant load and at the same time a dilute sodium chloride solution is dripped onto the spe-cimen. At the heated surface the chloride solution evaporates, leaving a highly concentrated chloride environment. The test is continued until the specimen cracks or up to a specified time, usually 500 hours. In the same manner as in the modifi-ed ASTM G36-method the time to failure versus load is recor-ded and a threshold stress is determined below which SCC does not occur.

4229b 1. 0. 0. 0. 0. Stress/Tensile strength 100°C (210°F) AISI 316L 10 20 30 40 50 60 Time to fracture, h SAF 2205 AISI 304L No SCC SCC CaCl2 1373b 300 (572) 250 (482) 200 (392) 150 (302) 100 (212) 50 (122) 0.0001 0.001 0.01 0.1 1 10 % Cl_ Temperature, °C (°F)

Laboratory test, aerated solution

SCC No SCC

Service experience, Cl_ Service experience, Cl_ +S/H2S

Figure 15. U-bent sample for stress corrosion testing with constant strain.

Figure 16. Constant load SCC testing in 40% CaCl2 at 100°C, using weights.

(22)

In slow strain rate testing (SSRT) the stress over the sample is continously increased. The strain rate must be chosen cor-rectly so that SCC is induced. Too low values result in repassivation of the attacks, whereas too high values give purely mechanical fractures. This method of loading resulting in continuous breaking of the passive layer and monotonically increasing load is very aggressive and does not reflect the conditions in practical service of the material. The results from an SSRT test are conveniently presented in a stress-strain curve, in the same way as mechanical tests (figure 19). In this way several parameters can be compared for tests in corrosive and inert environments. The ratio between such parameters is often used to rank the susceptibility to SCC for different materials.

For the testing of sulphide induced cracking in sour service the NACE standard TM 0177 may be used. A choice between four methods is given, in which stress is applied in different ways. The test solution is selected to give hydrogen absorp-tion condiabsorp-tions equal to that expected of the most severe well environment. It should have a partial pressure of hydrogen sulphide of 1 bar, 5% chlorides and pH=3. The test tempera-ture is set to 20-25°C and the samples must resist a period of up to 30 days without cracking.

fracture corrosive environment inert environment (ultimate) tensile strenght 0 0 Nom. stress , F/A °

Nominal strain (I - I0)/I0

7590

Figure 19. Stress-strain curve for corrosive and inert environment respectively.

(23)

4.

High temperature corrosion

Introduction

The temperature region for high temperature steels, ranges from about 400˚C up to temperatures where the mechanical properties of the alloy will be severely reduced. High temper-ature applications are found in various processes, as for example wire annealing, combustion/incineration, hydro-carbon cracking, air heating (recuperators) and shieldings. In these applications the selection of an alloy becomes difficult, as the material is subjected not only to corrosive chemicals, but also to a degeneration due to the temperature. That is, the alloy may become brittle due to structural changes, e.g. formation of sigma-phase or 475°C embrittlement. This is particularly so in the case of ferritic steel with a Cr-content higher than 15%, and in the temperature ranges 400-800˚C. Furthermore, for materials that are designed for carrying a load, own weight or pressure, the creep properties will beco-me a dominating factor. In most cases a compromise between the desired properties must be made, as no single alloy is lik-ely to have excellent values in all these fields (e.g. high Cr is good for corrosion, but bad regarding structural stability). Most common material temperatures in high temperature applications are in the range 400-900˚C. At these temperat-ures the alloy will react with the surrounding gas phase and degenerate chemically, forming what is referred to as corro-sion products. The formation rates of these products are strongly temperature dependent. In aqueous solutions a 10 times higher corrosion rate is not uncommon for a temper-ature change of 30˚C. The same 10 fold change (or worse) may occur with a 20˚C change under high temperature condi-tions. Note that if the temperature is too low it may be diffi-cult for a protective oxide scale to form. This can become a problem particularly in applications where the material is internally cooled, e.g. furnace tubes.

Most of the corrosive elements can be found in the upper right corner of the periodic table. The elements in this area have similarities in chemistry and will, thus, have similar reactions paths. The corrosion product formed may be "beneficial" or "detrimental" to the alloy, depending upon the specific prod-uct formed. In practice the life time for an alloy relies on it's ability to form a dense, adherent and continuous oxide layer with a low growth rate. If other products are formed, i.e. sulphides, carbides, nitrides, or halides, the scale will be less protective or even more corrosive than no scale at all.

However, a good oxide scale will not last for ever, as the scale will continue to grow until it reaches a certain thickness. At this point stresses due to the differences in volume between the scale and the alloy will become so large that the scale will crack. The cracks will cause the oxide to spall off and then the protection of the scale is gone. The spalling can be measured and is often used to compare and rank the alloys. The spalling temperature is measured by heating the alloy in air at increas-ing temperature steps for a certain time and the temperature where the weight change of the alloy exceeds a pre-defined value is set as the scaling temperature (usually this value is 1.5 g/m2h). For comparison, the scaling temperature for car-bon steel is about 550˚C and for 353MA (25% Cr, 35% Ni) the scaling temperature is about 1225˚C. The maximum ser-vice temperature of an alloy is usually set 50˚C below the scaling temperature.

Figure 1. Rapid oxide growth (dark) in 18-8 material caused by too high temperatures. The oxide growth rate has been fast enough to form pieces of metal “islands”. This oxide scale is not protective!

As there are large differences, not only in the volume between the oxide and the alloy but also in the thermal expansion coef-ficient scaling may become a severe problem when the pro-cess includes a temperature cycle.

Scaling is not the only mechanism that can disrupt a protect-ive oxide. In many applications a melt deposit may also be formed, especially in combustion processes. This melt can dissolve the oxide layer and open up the alloy for further and rapid corrosion. This particular reaction is often referred to as

(24)

“hot corrosion”. Hot corrosion is a well known phenomena in the gas turbine industry. There is an on-going argument on the mechanism of this type of corrosion and so far the only agree-ment is that there are two types of hot corrosion, which pear at different temperatures. In many cases studied it ap-peared that the onset of rapid corrosion was dependent on some degree of thermal cycling and that there was an inde-fined initial period before the corrosion rate increased. (See also catastrophic oxidation below.)

In the subsequent text some corrosion reactions will be dis-cussed, but first it should be mentioned that there is a great difference between high temperature corrosion and “wet” corrosion. This means that material developed for high temperature applications may not have as good corrosion resistance at low temperatures and in “wet” applications as it has in the high temperature range.

Oxidation

The main corrosion reaction is oxidation. Oxidation of an alloy may occur at any temperature (the oxidation rate will increase with the temperature) if the amount of oxygen is high enough. The advantage with oxidation is that the oxide layer formed may serve as a protection from further corrosion, that is, if the oxide layer formed is dense, continuous, and ad-herent. Some alloying elements like Al, Si, and Cr may form a dense layer, and of these three elements Al and Si will form the most effective oxide scale. Unfortunately, the amount of Al or Si needed in the alloy for forming this oxide scale will make the alloy rather brittle and it's fabrication difficult. In carbon steel and in low alloyed steel, iron is the main oxide former. Iron may form a rather dense oxide, Fe2O3, which is protect-ive. However, if the temperature is higher than 550˚C, wüstite (FeO) is formed. This phase is porous, it is not protective and rapid oxide growth will occur causing the low scaling temper-ature of carbon steel. To enhance the protective effects of the oxide scale small amounts of RE (Reactive Elements, = Sc, Y, La, Ce, Pr, Nd, Pm, Sm, (sometimes also called REM, Rare Earth Metals) may be added to the alloy. It is not fully clear why this should occur, but several investigations have shown that these small (< 0.1%) additions result in reduced growth rate and improved adherence of the Cr- and Al- oxides.

Catastrophic oxidation

Even if the formation of oxides in general is beneficial for the alloy, there are some oxides that should be avoided. Oxides, such as for example MoO3, have rather low melting points, or may form compounds with other oxides that have low melt-ing points, especially in static (non-flowmelt-ing) systems. The melts formed may dissolve the remaining oxide layer and

then the material will be destroyed within a very short period of time. This behaviour is often referred to as “catastrophic oxidation”. Other alloying elements that are sensitive to catastrophic oxidation are V, Pb, W, Ta, and Nb.

Sulphidation

In many high temperature processes sulphur is present, e.g. in most processes where coal or oil is combusted. The chemistry of sulphur is rather similar to that of oxygen, and sulphur will thus react in competition with oxygen. However, sulphur will not form such a dense layer as in the case of oxides and more-over, the melting point of the sulphides formed may some-times be several hundred degrees lower than that of the oxide, e.g. Ni3S4, melting point ~650˚C. If Ni3S4 is formed in the grain boundaries this will also decrease the strength of the alloy. Hence Ni-containing alloys should be avoided in sul-phurous environments if they are unable to form a protective oxide scale.

In a sulphur-containing atmosphere where sufficient oxygen is present to allow an oxide scale to form the corrosion resist-ance is determined by the properties of that layer, but if cracks (see scaling above) start to form in the oxide layer, then sul-phur will attack at these points. This means that the corrosion resistance in sulphur containing atmospheres will depend on the scaling temperature of the alloy, and the maximum service temperature will be dependant upon both the scaling temper-ature and the amount of sulphur in the gas. On the other hand, if the alloy cannot form a protective oxide the corrosion resistance will be greatly reduced and more dependant upon the alloy composition. Under these conditions it is favourable to use an alloy that is high in Cr and has either no Ni (e.g. 4C54, provided that embrittlement is not a problem) or a lower level of Ni (e.g. 253MA).

Figure 2. Catastrophic oxidation of a Mo-rich duplex alloy, after 24h in 1060˚C

(25)

Carburisation and nitration

In some applications the atmosphere is reducing with a high content of carbon (C) or nitrogen (N), e.g. in wire annealing, ethylene furnaces, carbon black production, or steam-methane reforming. In these processes the material temper-ature may be as high as 900-1150˚C. If carbon is used to generate a reducing environment the carbon species in the atmosphere may react with the alloy, forming carbides. Sometimes the carburisation reaction will occur rapidly and this phenomena, “metal dusting” is mainly observed in waste heat boilers in steam reforming processes in the temperature range 500-800˚C. Even if metal dusting does not occur, the formation of carbides is detrimental as this reaction often catalyses coking, and the coke may then block the tube. In some wire annealing processes cracked ammonia is used to create a non-oxidising environment. In this case the nitrogen activity is high, which will lead to the formation of nitrides in the alloy.

The chemistry of the alloy may change due to carburisa-tion/nitridation, as it is often one alloying element that reacts with the carbon/nitrogen, e.g. Cr. This element will then be tied up in the precipitates and will not be able to form a pro-tective oxide scale. Furthermore, the bulk material below the surface will be depleted in Cr and this will reuce the pos-sibility of oxide formation.

A heavy carbon or nitrogen pickup may also embrittle the alloy as the precipitation of the carbides/nitrides are concen-trated at the grain boundaries. Also, the creep properties and the ductility will be affected by the carburisation/nitridation reaction. The resistance to carbon and nitrogen pickup is improved by increased Ni-content. In oxidising environments strong oxide forms such as Cr, Al, and Si are beneficial. Al is of special interest, particularly in environments where carbu-risation is the main problem.

Molten metal corrosion

The main corrosion mechanism here is either massive or selective dissolution. Some general observations are that austenitic Ni-Cr-Fe alloys dissolve more rapidly with increas-ing Ni-content, and that ferritic Fe-Cr stainless steels tend to be more resistant. Note that the embrittlement caused by the liquid metal is more hazardous than corrosion. This embrittle-ment is more pronounced for melts containing silver (Ag), copper (Cu), and zinc (Zn). When molten, these metals will penetrate austenitic alloys intergranularly and may cause rup-ture within a few seconds. Cases where this has been reported are for example, welding of stainless steel to galvanised car-bon steel, weld cracking due to copper contamination, and cracking caused by copper containing anti-seize compound.

Halogen corrosion

Halogen (i.e. F, Cl, Br and I) containing gases are often very aggressive against all metallic materials. The ones of greatest interest are F and Cl. The former is regarded to be the more corrosive one while the latter is more frequent. If the envir-onment is reducing, Ni and Cr generally improve the corro-sion resistance. In oxidising environments, Cr, and especially Mo and W, are detrimental due to their tendency to form vol-atile oxychlorides. The general mechanism of corrosion is the same as for oxidation and sulphidation. High chloride vapour pressure can result in penetration and disruption of the oxide scale. Voids may be formed under the scale as the chlorides evaporate. There has been limited study of halogen reaction with metals and general material performance criteria are still under development.

Erosion-corrosion

Erosion can enhance or retard corrosion attacks; increasing them by removing the protective layer or decreasing them by the removal of corrosive deposits. similarly, corrosion may increase or decrease erosion rates; increasing them by at-tacking the eroded surface or decreasing them by forming an oxide layer that is more erosion resistant then the parent metal. This means that in addition to having a good resistance against the corrosive atmosphere, materials exposed to ero-sion must be able to develop an adherent, ductile and self-healing oxide layer.

The addition of Rare Earth Metals has a beneficial influence on all these three oxide layer properties. Application areas where erosion may occur are for example fluidized bed com-bustion (FBC) and cement production.

There are still a lot of work to be done to determine the effects of erosion-corrosion on different materials, hence, no specific data is available on erosion-corrosion of alloys.

Applications

The high temperature applications that Sandvik targets are presently recuperators, muffle tubes, thermocouple protection tubes, boiler tubes and ethylene furnace tubes. The selection of the material in each application is very much dependent on the environment. For example, the atmosphere in a muffle tube may be oxidising (air), reducing (cracked ammonia), or carburizing e.g. bundy tube production). Each environment demands specific properties of the alloy. In table 1-3 some Sandvik recommendations on material selection are pre-sented, varying with application and environment.

(26)

P O W E R B O I L E R S

Power utility boilers, especially those burning biomass- or fossil fuels which often contain high levels of sulphur and chlorine, represent one of the most demanding applications for stainless steel tubes. The outer tube surface is attacked at high temperature by the combustion products with differing corrosion and/or erosion mechanisms.

The inside surface is often subjected to steam oxidation cor-rosion: and the material is expected to support high loads resulting from high internal steam pressure – often under con-ditions where creep is a significant factor.

Currently, temperatures and pressures are generally being increased, with the intention of improving thermal efficiency and to reduce pollution. Conventional high strength alloys can

be susceptible to the corrosion attack, whilst corrosion resis-tant alloys often have insufficient strength at the temperatures involved. Tube failures result in having to shut down the power station.

A proven solution is to use composite tube. There are two main areas of application: evaporator tubes and the super heater/reheater tubes.

In the evaporator section, tubes typically have a highly corro-sion resistant outer layer, 25Cr20Ni variant, and a load bearing carbon- or low alloy inner component. In the super-heater/reheater elements a typical solution utilises a high alloy outer component, e.g. 310Nb, Sanicro 28, alloy 825, or alloy 625 bonded to a creep resistant steel inner component.

353 MA 253 MA 310, Sanicro 31HT, 4C54 304H, 316H, 321H In air Temp °C 500 600 700 800 900 1000 1100 1175

In oxidising sulphurous gases

In reducing sulphurous gases 353 MA 4C54 253 MA 310 Sanicro 31HT 304H, 316H, 321H 4C54 253 MA 310 353 MA Sanicro 31HT 304H, 316H, 321H

(High humidity may lower temperature 50-150 °C)

The maximum temperature is depending on the level of flue gas impurities (S, Na, V)

Grade Temp °C 400 500 600 700 800 900 1000 Sanicro 31HT* 253 MA 353 MA 309S 310S 4C54 γ σ phase σ phase σ phase σ phase σ phase 475°C-embrittlement

Table 1. Maximum working temperature in different gases.

(27)

Composite tube applications

In many cases high temperature materials from Sandvik Steel are delivered for applications where both corrosion resistance and pressure vessel approval must be fulfilled. Such applica-tions are black liquor recovery boilers (BLRB): municipal waste incinerators, and power utility boilers.

Table 3. High temperature corrosion properties A comparision between Sandvik High Temperature Steels and TP 310.

Oxidation Carburisation Nitriding*

TP 310 0 0 0

353 MA +++ +++ ++

253 MA + + 0

Sanicro 31HT = ++ ++

4C54 0 – ** –

* in cracked ammonia atmosphere. ** 4C54 has very good resistance to

metal dusting corrosion.

0 = reference value + = superior to – = inferior to.

A BLRB is a boiler about 30-70 m high where the floor and the walls are made of panel welded tubes. The boiler is used to burn the residues from the wood cooking and to recycle the cooking chemicals (mainly sulphides). The combustion of the residue is controlled in such way that the sulphides in the fuel (black liquor) will form smelt that accumulates at the bottom of the boiler. This smelt (now called green liquor) is then taken out through smelt spouts at the bottom of the boiler and processed further to be reused in the cooking. The tubes are water cooled, and the water/steam pressure is generally be-tween 60-100 bar. This will correspond to a water/steam saturation temperature of 250-300˚C in the walls and steam temperatures of around 400-480˚C in the superheater. The

actual material temperature may be some 20-50˚C higher than the water, or the steam temperature.

This boiler is the heart of the pulp industry, i.e. if the boiler is shut down, the whole plant must stop.

There are high safety demands in this process because if a tube bursts, and water contacts the melt, there is a high risk for a boiler explosion. The environment in a municipal waste incinerator boiler (MWB) is much more aggressive to metals that that in the BLRB, and thus the corrosion rates here are high. If a tube bursts or fails here, the plant must be shut down for repair which is costly.

For these two applications Sandvik has developed composite tube solutions. Composite tubes consist of two alloys that are co-extruded to form a tube with outer and inner components. The bonding between these components is a chemical metal-to-metal bond (sometimes referred to as a metallurgical bond). One of the components is there to serve as the load carrier and the other as a corrosion protection layer. The load carrier is often a carbon steel type 4L7 (SA210-A1) for tem-peratures below 450˚C, a low alloyed carbon steel like HT8 (T22) for temperatures up to 550˚C, or higher alloyed steel like T91 for temperatures up to 600˚C.

In the BLRB's the combinations Sanicro 38/4L7 (floor), 3R12/4L7 (wall) and, 3RE28/HT8 (superheater) are recom-mended.

For waste incinerators the recommendations are 3R12/4L7, Sanicro 28/4L7 or Sanicro 63/4L7, depending on temperat-ures and type of pollutants in the fuel. There is on-going deve-lopment of composite tube combinations for various applica-tions.

(28)

Introduction

Stainless steels is a designation for a group of iron-base alloys with such a composition that they are able to passivate, i.e. form a passive layer which protects the metal from the sur-rounding environment, and thus hinder metal dissolution (corrosion). The chief alloying element in stainless steels is chromium (Cr), which in concentrations above 12-13% forms a passive layer on the metal. Increasing the chromium content leads to a stronger passivity and thus a higher corrosion resist-ance. Chromium is a so called ferrite stabiliser, which means that chromium does not alter the structure of iron, which has a ferritic structure. The physical properties of an alloy with only chromium added therefore does not differ much from pure iron. These types of alloys are called ferritic stainless steels.

Although chromium makes the steel stainless it cannot resist more aggressive environments, and the formability of ferritic stainless steels is limited. Other elements are therefore added to modify the structure, the mechanical properties and the corrosion resistance.

Nickel (Ni) is added to alter the structure of the steel but may also improve the corrosion resistance if sufficient amounts are added. Nickel is an austenite stabiliser, which means that an addition of nickel will alter the structure from ferritic to auste-nitic. In an alloy with 18% chromium about 10% nickel is required to alter the structure to almost purely austenitic.

These alloys, the austenitic stainless steels, have an improved formability, greater toughness and high temperature prop-erties, as well as improved weldability compared to ferritic stainless steels. The physical properties will also change, e.g. fully austenitic stainless steels are non-magnetic.

Molybdenum (Mo) has an effect similar to chromium with regard to structure and corrosion resistance. Molybdenum alloyed steels are what is usually called ”acid proof”, which refers to the beneficial effect of molybdenum on the corrosion resistance in sulphite digester liquor. In some media, like strongly oxidising acids, molybdenum may impair the corro-sion resistance, and should therefore be avoided in such appli-cations.

Nitrogen (N) increases the strength, the corrosion resistance and also improves the structural stability of stainless steels. In many cases, especially for duplex stainless steels, it also improves the weldability.

Copper (Cu) is beneficial for the corrosion resistance in cer-tain acids. Titanium (Ti) and niobium (Nb) are used as carb-ide formers, which means that carbon is preferentally bound to these elements, thus reducing the risk for precipitation of deleterious chromium carbides in the grain boundaries, and hence decreasing the risk of intergranular corrosion.

Four main types of structures exist in stainless steels; ferritic, martensitic, austenitic and duplex (austenitic-ferritic).

(29)

Austenitic stainless steels

and Duplex stainless steels

The austenitic family of stainless steels covers a broad inter-val of alloying elements, from standard 18-9 to super austeni-tics with up to 7 % molybdenum and matching contents of chromium and nickel. The super austenitics are often also alloyed with nitrogen. Their corrosion resistance is therefore adapted to a great variety of corrosive environments. The super austenitics are most often designed to resist pitting and crevice corrosion in chloride containing environments, e.g. sea water. They also have good resistance to stress corrosion cracking, in solutions containing hydrogen sulphide and in alkaline solutions.

The austenitic steels have good ductility, at both low and high temperatures. Their weldability is good. The austenitic steels are readily welded with all normally used techniques. High alloy steels can be welded with over-alloyed filler metals, thus matching the corrosion resistance of the base metal. Duplex stainless steels are stainless steels with a microstruc-ture comprising typically 40-50% ferrite and the rest

aus-tenite. These steels combine important properties from both ferritic and austenitic stainless steels. They show good stress corrosion resistance and also good ductility and weldability. All modern duplex stainless steels have a low carbon content. Duplex grades with PRE-number (PRE = Pitting Resistance Equivalent = % Cr + 3.3% Mo + 16% N) greater than 40 are called super duplex. These steels possess very good corrosion properties, especially in chloride containing environments. The duplex structure gives high mechanical strength, approx-imately twice that of austenitics, combined with low thermal expansion, close to that of carbon steel. The low Ni-content is cost saving and high mechanical strength means lighter con-struction, which gives cost advantages.

The upper service temperature of duplex stainless steels is around 300°C due to the risk of embrittlement and formation of precipitates. The weldability of duplex stainless steels is good. Welding of duplex stainless steels with proper welding parameters and matching filler metals gives good corrosion and mechanical properties.

Figure 1. Microstructure of austenitic stainless steel. Figure 2. Microstructure of duplex stainless steel. The dark areas are ferrite and the light areas are austenite.

References

Related documents

Expected Utility with q=0.5 (Benchmark case).. Unlike the case q = 0 , with sudden stops the optimal value of η from the debtor’s point of view is always less than the maximum

THE HOME CARE ALLIANCE OF MASSACHUSETTS is pleased to present this opportunity for companies to exhibit and sponsor at our 2010 Spring Conference &amp; Trade Show.. The Home

After processing of the miRNA from double stranded precursors, by a variety of different proteins, including the RNase III enzyme Dicer and its double-stranded RNA

So far, only path prediction models for single pedestrians in static environments ignoring interactions with other pedestri- ans have been reviewed. This section will consider models

Stream surface w ater sam ples have large seasonal and inter-annual geochem ical variation, however, source w aters show distinct chemical signatures allow ing the

As discussed in section 4.1, to build the next-level coarser graph, each step of coarsening of the multilevel graph partitioning consists of vertex matching followed by