Effect of iron dilution on corrosion resistance of

Ni-Cr-Mo alloy cladding

M.F.GITTOS Corrosion tests were carried out on Ni-Cr-Mo overlays deposited by welding on to C-Mn
T. G. GOOCH steel substrates. Manual metal arc, pulsed metal inert gas, and hot wire tungsten inert gas
processes were used, with conditions varied to obtain a range of dilutions of the overlay by
the parent steel. Critical pitting temperatures were determined in FeCIJ solution and in an
oxidising H2S04-HCl environment. Overlay pitting resistance was reduced by increased
dilution from the substrate steel. The results support a general recommendation that iron
content should be below 5% for optimum performance, although corrosion behaviour was
determined not by the iron level per se but by the bulk composition and the minimum alloy
content developed in the solidification structure. Manual metal arc deposits tended to be
inferior to gas shielded overlays. Only general correlation was found between pitting resist-
ance in the two test environments so caution is necessary when using FeCl3 testing for weld
procedure qualification purposes.
© 1996 The Institute of Materials. Manuscript received 10 July 1996; in final form
15 October 1996. The authors are at TWI, Abington Hall, Abington, Cambridge CBl 6AL.

INTRODUCTION EXPERIMENTAL
Weld cladding with nickel alloys is widely used to enhance Materials
the corrosion performance of steel components. In partic- The base plates for the claddings were C-Mn steels.
ular, extensive use has been made of Ni-Cr-Mo welding Manual metal arc (MMA) electrodes of diameter 4 mm,
consumables for sea water1 and flue gas desulphurisation according to American Welding Society (AWS) specification
(FGD)2 environments. It has long been recognised that the A5.11-90 for ENiCrMo-3, and 1·2mm wire for gas shielded
corrosion resistance of high alloy weld metals can be welding, to AWS specification A5.14-89 for ERNiCrMo-3,
affected by elemental segregation during solidification3 and, were obtained. For comparative purposes, N06625 parent
in the case of dissimilar welds, by dilution from the parent material was used in the form of 25 mm diameter bar.
materials.1•2 The effects of deposition procedure on the Material compositions are given in Table 1, where MIG
level of substrate dilution of Ni-Cr-Mo weld overlays1,4,5 refers to metal inert gas and TI G to tungsten inert gas.
have been studied, but little quantitative information has
been published relating dilution directly to corrosion Welding
performance. Perhaps as a consequence, very restrictive
Each deposit covered an area of approximately 200 x
compositional limits are sometimes imposed for the clad-
75 mm. Manual metal arc welding was used to make one,
ding surface6 which can be difficult and costly to achieve.
two, and three layer claddings and the hot wire TIG and
The present work was therefore undertaken to investigate
the effect of dilution by steel substrates on the corrosion synergic pulsed MIG process were employed to produce
single layer deposits. Two MIG welding conditions were
resistance of Ni-Cr-Mo claddings. Studies were based on
test media appropriate to chloride and oxidising acid used to obtain nominally high and low dilutions. Argon
environments, as experienced in marine service and FGD shielding was used for the MIG overlays and a 75He-25Ar
plant. Claddings on the C-Mn steel were prepared using mixture for the TIG specimen. Welding conditions are
coated electrodes and inert gas shielded welding tech- summarised in Table 2.
niques were used to deposit UNS N06625 type alloys. A
range of dilutions, essentially by iron, was obtained. Material examination
Coupons cut from the overlays were used to determine Transverse metallographic sections were prepared from the
critical pitting temperatures in lO%FeCl3 solution and in overlays for examination by light and scanning electron
H2S04-RCI. microscopy (SEM). Etching was carried out electrolytically
in 20%H2S04 with 0·1 g L -1 of NH4CNS.

Table 1 Compositions of parent materials and consumables, wt-%

Specimen C Si Mn P S Cr Mo Ni Al N Nb Ti Fe

Steel A used for

MMA deposits 0·15 0·35 1·35 0·029 0·021 0·09 0·01 0·07 0·053 ... 0·028 <0·002 "Bal.
Steel B used for
MIG and TIG
deposits 0·17 0·43 1.46 0·017 <0·002 0·05 <0·005 0·01 0·031 0·020 0·004 Bal.
N06625 0·010 0·12 0·07 <0·005 0·002 22·1 9·33 Bal. 0·16 0·017 3·40 0·19 2·04
MMA electodes* <0·10 <0·75 <1·0 <0·03 <0·02 20'0-23'0 8'0-10'0 >55·0 ···t 3·15-4'15 ···t <7·0
MIG and TIG
wiret <0·10 <0·50 <0·50 <0·02 <0'015 20·0-23'0 8'0-10·0 > 58·0 <0·40· 3·15-4·15 <0·40 <5·0

* According to AWS specification A5.11-90 for ENiCrMo-3.

t Not specified according to A5.11-90.
t According to AWS specification A5.14-89 for ERNiCrMo-3.

British Corrosion Journal 1996 Vol. 31 " NO.4 309

310 Gittos and Gooch Corrosion resistance of Ni-Cr-Mo cladding

a
2 Scanning electron micrograph of interdendritic N b rich phase in
MMA single layer deposit: nominal magnification x 4110

Two aqueous test solutions were employed, namely

100/oFeCI3.6H20 as in ASTM G48-76 Method A and a
simulated FGD environment of 7H2S04-3HCI-ICuCI2-
1FeCIJ (commonly termed 'green death'). In both media,
testing was carried out to define a critical pitting temper-
ature (CPT) above which pitting occurred.
The prepared coupons were degreased, weighed, and
immersed in 225 mL of each test solution in individual
conical flasks, housed in a constant temperature bath. The
initial exposure temperature was 5°C for both the FeCI]
and 'green death' solutions. After exposure for 23 h at the
required temperature, the coupons were removed, weighed.
and examined for pitting using a x 10 magnifier. If there
b was no evidence of appreciable attack or weight loss this
a high dilution MIG; h hot wire TIG procedure was repeated with a temperature increment of
Typical microstructures of given clad layers electrolytically 2·5 or 5 K. The onset of pitting with increasing temperature
etched in dilute H zSO 4 was thus determined both visually and by weight loss in
each test environment.

Chemical analyses were obtained from the top surface of RESULTS

each cladding, by X-ray fluorescence for major elements Cladding compositions
and by Leco combustion for carbon and sulphur. Nitrogen Bulk chemical analyses of the cladding layers are given in
and oxygen determinations were made for the top beads of Table 3. Appreciable variation in dilution is apparent, with
all clad dings by Leco inert gas fusion. Segregation within iron content in the range 3'4-270/0 and molybdenum
the deposits was assessed using energy dispersive X-ray content 6'5-8·8%, and deposits were produced covering a
(EDX) microanalysis in a scanning electron microscope. similar compositional range for both flux coated and gas
shielded consumables. The extremes of dilution and iron
Corrosion testing content were provided by the three layer MMA (3'4%
Coupons measuring 40 x 20 x 2 mm were cut from the iron) and one layer hot wire TIG (27'60/0 iron) specimens.
claddings to study the top layer in the case of the MMA
two and three layer deposits and from the weld metal only ~icrostructures
for the single layers. Coupons of diameter 25 mm and The microstructures of all deposits were similar, consisting
thickness 2·5 mm were cut from the parent N06625 bar. All of a cored dendritic structure with minor phases tending
specimens were wet ground to a 120 grit finish using to be concentrated in the interdendritic spaces (Fig. 1). The
SiC paper. most abundant particles were niobium and were probably
a Laves phase.4 They tended towards a dendritic appearance
themselves (Fig. 2) and shared the interdendritic areas with
Table 2 Welding conditions smaller particles enriched in chromium and iron, which
were probably carbides. In the MMA deposits there were
Low High Hot also numerous spherical inclusions, up to a few micrometres
dilution dilution wire in diameter, which contained aluminium and titanium
Parameter MMA MIG MIG TIG (Fig. 3) and were presumably deoxidation products.
The dendrite cores of MMA claddings had a dark
Voltage, V 24 28 28 15
appearance and were surrounded by a distinct pale grey
Current, A 117 250-50* 305-60* 300
Traverse speed,
region in secondary electron SEM images (Fig. 4). The
mm min-1 135 120 120 135 interdendritic areas, which principally contained the white
Average arc energy, Laves phase, exhibited similar contrast to that of the cores.
kJ mm-1 1·3 1·8 1·8 2·0 Results of EDX microanalysis showed the expected segre-
gation pattern for the dendrites, with partitioning of nickel
* Peak current minus background current. and iron to the core and molybdenum and niobium

British Corrosion Journal 1996 Vol. 31 NO.4

 Gittos and Gooch Corrosion resistance of Ni-Cr-Mo cladding 311

a a

Ti

Cr

Fe Ni

b
Ni

a high contrast secondary electron image in as polished condition,

0.1 keV 10.4 nominal magnification x 1000; h secondary electron image after
electrolytic etching in dilute H2S04, nominal magnification x 1020
a back scattered scanning electron micrograph with inclusions shown as 4 Given scanning electron micrographs (~fsingle layer MM A deposit
black particles, nominal magnification x 497; h EDX analysis from
inclusion
3 Single layer MMA deposit above about 5 mg was taken to indicate the occurrence
of pitting. Overall, the duplicate specimens gave similar
results; the initial M MA single layer test in FeCl3 was
interdendritically (Table 4). The pale grey region proved to
stopped prematurely at sec because of an apparent pit
have an intermediate composition, but the SEM results
indication, although the weight loss was below 0·5 mg.
showed no such intermediate contrast for the gas shielded
With this possible exception, the tests also indicate fairly
deposits. The level of segregation in the higher dilution
good agreement between CPT values derived visually and
MIG overlay appeared similar to that in the MMA
those from weight loss.
cladding of comparable iron content, although segregation
The parent N06625 bar material was not attacked over
seemed less marked in the hot wire TIG overlay (Table 4).
the temperature range employed, but the claddings gave a
range of CPT values. The single layer MMA deposit
Pitting tests experienced pitting at the lowest temperatures, in contrast
The results of the duplicate tests in FeCI) are given in to the behaviour of the MIG cladding of similar dilution,
Table 5, which shows critical pitting temperature (CPT) there being a difference of over 40 K in CPT. However, the
values derived from both the weight changes and the visual two and three layer MMA deposits achieved high pitting
indications of pitting. In the former case, a weight loss temperatures similar to those of the compositionally

Table 3 Analyses of top surfaces of overlays

Content, wt-%

Specimen C Si Mn P S Cr Mo Ni Al N Nb Ti 0 Fe PRE

One layer
MMA 0·062 0·31 0·26 0·011 0·005 16·7 6·44 49·7 0'07 0·028 2·66 0·06 0·050 23·0 40·2
Two layer
MMA 0·046 0·33 0·09 0·007 0·004 19·8 8·32 58·8 0'08 0·016 3'14 0·07 0·074 9·11 47·5
Three layer
MMA 0·041 0·32 0·03 <0·005 0·003 20·9 8·83 62·8 0'06 0·021 3·23 0·06 0·073 3-43 50-4
Low dilution
MIG 0·030 0·06 0·09 <0·005 <0·001 20·3 8·53 62'1 0'29 0·006 3·32 0·22 0·002 4·81 48·6
High dilution
MIG 0·051 0·11 0·34 0·008 <0·001 16·4 6·94 50·3 0'24 0·008 2'70 O·U~ 0·002 22·5 39'4
Hot wire
TIG 0·056 0·13 0·42 0·010 <0·001 15·4 6·49 46·9 0·23 0·010 2·52 0·17 0·011 27·6 37·0

British Corrosion Journal 1996 Vol. 31 NO.4

312 Gittos and Gooch Corrosion resistance of Ni-Cr-Mo cladding

Table 4 Results of EDX microanalyses from segregated

regions

Normalised content, wt- %

Cladding Region Ni Cr Fe Mo Nb

One layer MMA Dendrite core 48·8 16·3 27·7 5·5 1·1
Intermediate 47·7 16·7 26·1 6·9 2·0
Interdendritic 46·3 16·8 24·0 8·5 3·8
High dilution MIG Dendritic core 47·6 15·7 29·2 5·5 1·1
Interdendritic 44·8 15·6 25·8 8·6 4·5
Hot wire TIG Dendritic core 48·0 15·7 29·2 5·9 1·0
Interdendritic 46·9 16·6 26·4 6·3 3·1

comparable MIG overlay. Despite it having the highest

dilution, the CPT of the hot wire TIG deposit was relatively
high at about 45°C.
The 'green death' test results are also given in Table 5. 5 Cross-section through pitting in single layer unetched MMA
Again, the weight losses recorded correlated well with the deposit ex posed to 'green death ' solution x 50
observed onset of pitting. At 80°C, no attack was found on
the parent material, nor on the three lay er MMA cladding;
all other specimens showed pitting at lower temperatures,
to the base metal and control of dilution to below 5% is
including the low dilution MIG overlay.
unlikely to be of practical advantage.
Sectioning of the tested coupons showed that classical
At the same time , as pointed out by Ahluwalia et al.,l1
pitting had occurred, as shown in Fig. 5. Pit morphology
concern attaches not so much to the iron level per se
was similar in the two media and it was apparent that,
having an adverse effect but to the total content of alloying
within the pits , the dendrite cores were co rroded preferen-
elements which promote corrosion resistance, primarily
tially to the interdendritic regions (Fig. 6).
chromium and molybdenum. There may , therefore, be little
practical advantage in pursuing extremely low dilution
DISCUSSION levels and iron pickup; a change in dilution of, say , 5-10%
The results in both test media clearly demonstrate the would reduce a nominal 9% molybdenum content by onl y
adverse effect of dilution from the base steel on the pitting about 0·5%. The effect of composition on pitting resistance
resistance of the nominal 9% molybdenum nickel based has been well studied for stainless steels and the concept
consumables employed; CPT values for th e higher dilution has reached general acceptance of a pitting resistance
claddings were appreciably lower than th ose for the N06625 equivalent (PRE) such as 3 . 12 , 13
base metal specimens. The most commonly practised
PRE = %Cr + 3·3 %Mo + 16%N (1)
control on dilution is to restrict the iron content''v" and
the CPT test results are plotted against measured iron A directly analogous relationship has not been obtained
levels in Fig . 7. A maximum of 5% iron in the cladding for nickel based materials, but, using equation (1) , the
layer next to the service environment ha s been considered FeCl 3 and 'green death' results are plotted against PRE in
desirable to match base N06625 material, although the Fig. 8. In both media, a PRE of about 50 or above is
range of permissible iron contents in this alloy was required for the claddings to approach the resistance to
originally set for solid solution strengthening rather than attack of N06625 parent material.
for corrosion resistance." Welding consumables of very low Comparison of CPT data from different investigations is
iron content ( < 2 0/0) have been propounded as advantageous difficult because the results obtained are influenced by the
in obtaining deposits with less than 5% iron ' ? and to some experimental procedure employed. In Fig. 9, comparison is
extent this level is supported by the data in Fig. 7. It might made with work by Suutala and Kurkela on stainless
be argued that, to ensure performance almost matching steels.' These investigators employed a factor of 13 for
that of wrought N06625 material, a low er limit may be nitrogen in the PRE relationship, but the difference from
necessary for critical applications; however, even with no equation (1) is insignificant at the low nitrogen levels of
iron dilution, the inevitable segregation of alloying elements the weld deposited claddings and base metal. From Fig. 9
in a weld deposit will reduce corrosion resistance relative it can be seen that the pre sent results compare very

Table 5 Summary of pitting test results

Pitting temperature, DC

FeCl 3 'G reen death'

Specimen Visual Weight loss Visu al Weight loss PRE

One lay er MMA 5, 32·5 > 5, 32·5 35,40 40,40 40·2

Two layer MMA 65, 70 65,70 52' 5, 72·5 52'5 , 72·5 47·5
Three la yer MMA 62,5, > 75 62'5 , > 75 >80, >80 > 80, > 80 50-4
Lo w dilution MIG > 75, > 75 > 75, > 75 55, 57·5 55, 57·5 48·6
High dilution MIG 65, > 75 70, > 75 42'5 , 50 42'5 , 50 39-4
Hot wir e TIG 45, 47·5 42'5 , 47·5 40,40 40,40 37·0
N06625 > 75, > 75 > 75, > 75 > 80 >80 53·2

British Corrosion Journal 1996 Vol. 31 N o.4

 Gittos and Gooch Corrosion resistance of Ni-Cr- M o cladding 313

90
(a)
80
±~1 1
70
•• ...
60 •
50

~ 40

~
~ 30 •
~
<ll
~ I
20 ' - - - - -----'-- I
- -'---- - - - - -I ' - - - - ' - -I- - - - - - - 'I- - - - - - - '
~ 0 5 10 15 20 25 30
Ol
'S
'0.
90
(b) :l •
] 80

a
;S 70 •
60

• ...
50

40
...•
30

20 '---- - - 'I - - -- - I
'---- - I
----'-- - - ' -I - -- I - ' - - -----'
----
o 5 10 15 20 25 30

Fe content, wt %

... TlG/MI G • M MA X PM

a in FeCl 3 solution; b in 'green death ' solution

7 Effect oj Fe content 011 CPT

directly to indicate behaviour in service in other environ-

b
mental conditions, unless an appropriate correlation has
a x 500; h x 1000 previously been esta blished. Caution is therefore necessary
6 Detail of FeCl 3 pitting in single layer MMA deposit in setting FeCl 3 solution test pass/fai l criteria for welding
procedure qualification. In this respect it would not
necessarily be mandatory in all applications that weld
favourably with those of Suutala and Kurkela, although
the caveat must be expressed that use of a single test
90
specimen with multiple exposure at increasing temperatures, (a)
as in the present work, is known to lead to relatively high 80
CPT values being obtained.!" This is pro bably because of 1 ...
a pickling effect resulting from exposure at lower temper- 70 •
atures before passive film breakdown and pitting initiation
60
• •
take place during sub sequent higher temperature exposure.
Figures 7 and 8 show a consistent trend that, at higher
dilution levels, CPT values in the two medi a employed are
50 ......
40
lower for deposits produced using coated electrodes than ;>
for the gas shielded overlays. The reason for this observation
is not clear. Table 4 does not show any marked variation
~
.:: 30 •
~
<ll
I I I I
between different processes in terms of segregation pattern Q.
E
20
and the resultant minimum alloy content and PRE ~ 30 35 40 45 50 55
Ol
developed in dendrite cores. It may be that the effect stems 'S
,~
90
(b)
from different inclusion populations in the overlays, and
the relatively high oxygen levels in the MMA deposits will
~
Q.

.g 80 ... :l
';:
be noted, but any effect of inclusions in constituting pit {,j
70 •
initiation sites would be expected to hav e been reduced
•
by the incremental temperature test exposure method
employed.
Pit initiation and growth in the two media were very
60

50 .......
similar and were clearly controlled by th e alloy element
segregation during solidification with th e interdendritic
40
....
30
second phase particles having less effect. The results from
FeCl 3 and 'green death' solutions are compared in Fig. 10. 20
While there is a general correlation in the pitting resistance 30 35 40 45 50 55
of an overla y in the two media, the agreement is not exact PRE
despite the similar pit initiation sites. It is becoming
... TlG/MIG • MMA X PM
increasingly common to specify pitting testing in FeCl 3
solution during weld procedure qualification. Clearly, a in FeCl 3 solution; b in 'green death' solution
however, performance in an FeC l3 test can not be used 8 Relationship between PRE and CPT

British Corrosion Journal 1996 Vol. 31 No .4

314 Gittos and Gooch Corrosion resistance of Ni-Cr-Mo cladding

80 90
.1 .1 t
70 ..•.. 8& •
0
80

~
e-
60
0
• ~~ 70

~ 000 x x
e 50 ~ 60
0
ell
Q. ~
'S
E
.! 40 x ~c 50
01

~~
Ql
c:: '" Weld
:g 30
xX metals
Q. 40

~ 20
8 30

10
X X 20
X 20 30 40 50 60 70 80 90
0
25 30 35 40 45 50 55
Ferric chloride CPT,°C

PRE
..•.. TlG/MIG • MMA X PM
..•.. TIG/MIG • MMA :x PM

10 Comparison of CPT values in FeCl3 and 'green death' solutions

9 Comparison of FeCl3 data from present work with data
from Ref. 3 .
ACKNOWLEDGEMENTS
overlays of the alloy type concerned should be restricted The authors would like to thank colleagues at TWI for
to a maximum of 5% iron. It can be inferred from Fig. 8 advice and assistance during the present work, which was
that, for many purposes, the PRE obtained with higher jointly funded by Research Member Companies of TWI
dilution levels may well be perfectly adequate for satisfac- and the Minerals and Metals Division of the UK DTI.
tory service to be achieved. In fact, dilution of up to 10%
has been regarded as acceptable for sea water service of REFERENCES
weld deposits from ERNiCrMo-3 wire consumables.1•4 1. D. F. HASSON, C. ZANIS, L. APRIGLIANO, and c. FRASER: J. Mater.
Energy Syst., 1985, 7, (3), 356-264.
2. E. L. HIBNER, L. E. SHOEMAKER, R. W. ROSS, Jr, and s. L. MORSE:
CONCLUSIONS Proc. Conf. Corrosion '85, Boston, MA, USA, March 1985,
1. Corrosion resistance of Ni-Cr-Mo claddings Paper 53.
deposited on to C-Mn steel has been studied. The MMA, 3. N. SUUTALA and M. KURKELA: Proc. Conf. Stainless Steels '84,
Gothenburg, Sweden, September 1984, The Metals Society,
pulsed MIG, and hot wire TIG processes were used to
240-247.
obtain a range of dilution levels and CPT values were 4. c. S. DIMBYLOW and K. M. CffiPPERFIELD: Weld. Met. Fabr.,
obtained both in FeCl3 solution and in an oxidising 1983, 51, (5), 229-235.
H2S04-HCI environment. 5. c. s. DIMBYLOW: Met. Constr., 1985, 17, (1), 35-39.
2. Dilution of the overlay by steel reduced pitting 6. P. KOSHY: Proc. 17th Annual Offshore Technology Conf.,
resistance in both media relative to N06625 parent metal. Houston, TX, USA, May 1985, Paper OTC 4973.
3. The results support a general recommendation that 7. M. J. CIESLAK: Weld. J., 1991, 70, (2), 49s-56s.
overlay iron content should be below 5% for optimum 8. J. B. CHUBB and J. BILLINGHAM: 'Effect of iron on the

pitting resistance, although, in less demanding environ- performance of IN 625 weld deposits', Report BR91456
(N85-17376/3), Cranfield Institute of Technology, Cranfield,
ments, single layer overlays with higher iron dilution may 1984.
provide adequate performance. 9. S. FLOREEN, G. E. FUCHS, and w. J. YANG: 'Superalloys 718, 625,
4. Pit initiation took place preferentially in alloy depleted 706 and various derivatives', (ed. E. A. Loria), 13-37; 1994,
regions in the solidification structure of the overlays. Hence, Warrendale, PA, Minerals, Metals and Materials Society.
pitting resistance is determined not by iron per se but by 10. M. N. MALIGAS, J. C. VICIC, D. C. AGARWAL, and M. KOHLER:
the total alloy content in terms of the PRE and the Proc. Conf. Corrosion '94, Baltimore, MD, USA,
minimum alloy level reached in dendrite centres. February-March 1994, NACE, Paper 63.
5. Pitting resistance of MMA overlays was inferior to 11. H. S. AHLUWAHLIA, L. H. FLASCHE, M. N. MALIGAS, and J. VICIC:

that of gas shielded ·overlays. Proc. Conf. Corrosion '93, New Orleans, LA, USA, March
1993, NACE, Paper 138.
6. General correlation was found between pitting resist- 12. S. BERNHARDSSON: Proc. Conf. Duplex Stainless Steels '91,
ance of given overlays in the two media, but this was not Beaune, France, October 1991, SF2M, 185-210.
exact. Thus, caution is necessary in setting pass/fail criteria, 13. G. RABENSTEINER: Weld. World, 1989, 27, (1/2), 2-13.
if FeCl3 solution testing is employed for weld procedure 14. R. A. WALKER and T. G. GOOCH: Br. Corros. J., 1991, 26,
qualification purposes. (1), 51-59.

British Corrosion Journal 1996 Vol. 31 NO.4