Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
WELDING RESEARCH
WELDING RESEARCH
Experimental Procedures
Materials
Table 2 — Hot Wire GTAW Parameters for Alloy 625 Overlay on F22 Steel Forging
Step Voltage (V) Peak Current (A) Background Hot Wire Wire Feed Preheat
Increment (in.) Current (A) Voltage (V) Speed (in./min) Temperature (°F)
0.150 11.0 240.0 120.0 3.3 66.9–82.2 350
Interpass Travel Hot Wire Hot Wire Heat Deposition Shielding Gas
Temperature (°F) Speed (in./min) Peak Current (A) Background Input (kJ/in.) Rate (lb/h) Flow Rate (ft3/h)
Current (A)
550 7.9 179.0 55.0 10.0 2.0 30.0
Table 3 — Chemical Composition (wt%) of F22 Steel and NiBased Filler Metal Alloy 625
C Ni Cr Mn Si Mo S P Al Cu Ti Nb Fe
F22 0.15 0.11 2.28 0.60 0.30 0.98 0.009 0.009 0.022 0.12 0.001 0.002 bal.
625 <0.01 64.0 22.7 <0.01 0.04 9.0 0.001 <0.01 0.12 <0.011 0.23 3.59 0.3
WELDING RESEARCH
A B
C D
Fig. 2 — F22/625 sample in aswelded condition: A — HV0.1 hardness map; B — hardness indents matrix; C — overlap of hardness map on
the indent matrix; D — hardness distribution across the fusion boundary.
to tempering heat treatments, not practical under actual fabrication pered in argon for oxidation protec-
F22/625 samples were subjected to conditions. These values were calcu- tion using a horizontal Lindberg
the 11 tempering heat treatments as lated using Equation 1 with tempera- 59544 type furnace. After tempering,
shown in Table 4. The first six heat ture (T) in Kelvin and time (t) in samples were quenched in room-tem-
treatments were selected to represent hours (Ref. 18). During procedure de- perature water.
the HJPs, which span a PWHT proce- velopment, thermocouple monitoring
dure commonly used in industrial was used to verify that samples
practice. The other five heat treat- reached furnace temperature in ap- HJP = T (20 + log(t)) (1)
ments were selected with larger HJPs proximately 10–20 min. This heating
that would accentuate the tempering time was not considered in the calcu-
and carbon diffusion effects, but are lation of the HJP. Samples were tem- Microstructure Analysis and
Hardness Measurements
Table 4 — Tempering Treatments on F22/625 DMW Samples Samples were mounted in a conduc-
tive Bakelite ground using successive
Heat Treatment Temperature Time HJP steps and polished through 1 m dia-
mond. Each sample was first chemical-
HT1 620°C 4h 18398 ly etched using 5% nital (5 vol-% nitric
HT2 640°C 2h 18535 acid 95 vol-% ethyl alcohol) to reveal
HT3 640°C 6h 18970 the steel microstructure and then elec-
HT4 650°C 5h 19105
HT5 650°C 10 h 19383
trolytically etched in 10% chromic acid
HT6 660°C 10 h 19593 to reveal the weld metal. Electrolytic
HT7 650°C 20 h 19661 etching conditions were 5.0 V for 5 s.
HT8 670°C 10 h 19803 Microhardness maps were developed
HT9 660°C 50 h 20154 using a Vickers hardness mapping de-
HT10 660°C 100 h 20526 vice (LECO LM100AT) with a 100-g
HT11 660°C 500 h 21178 load (VHN0.1). Typical hardness maps
WELDING RESEARCH
A B
C D
Fig. 3 — Representative microstructure in the F22/625 overlay in the aswelded condition: A — Base metal; B — FGHAZ; C — CGHAZ; D —
Alloy 625 weld metal.
spanned the steel HAZ, fusion bound- could still be obtained. This is in con- Results
ary region, and weld metal. Each hard- trast to the conventional view that
ness map was on a 40 78 grid with sufficient distance between indents is
100 m spacing between indents. The needed to attain reliable hardness val-
AsWelded Hardness
hardness grids were located such that ues in microhardness testing. Samples Figure 2 shows the hardness map of
40 indents spanned the fusion bound- for nanoindentation were polished the as-welded F22/625 interface and
ary and the 78 indents included two ad- through 0.5 m using colloidal silica. demonstrates the methodology for
jacent overlay weld beads. Using this ap- All testing was conducted with sam- collecting and displaying hardness
proach, reheating effects of adjacent ples in the as-polished (unetched) data. The transition from the high
weld beads could be determined. condition. hardness HAZ to the low hardness
Nanohardness determined in terms weld metal corresponded to the fusion
Nanoindentation of force, gigapascal (GPa), was then boundary region of the sample. The
converted to Vickers hardness. Since hardness in the HAZ was not homoge-
Nanoindentation was conducted there is not a standardized method to neous due to the tempering effect of
using an MTS Nano Indenter® XP sys- correlate nanohardness and Vickers the overlapping, multibead overlay
tem. Indents were made at a constant hardness, a direct method to correlate process. Because of this, it was diffi-
penetration depth of 500 nm (0.5 the two was used in this study. After cult to assign an average HAZ hard-
m). The spacing between indents was making 20 nanoindents, the sample ness based on the map. For example,
6 m. Kim et al. (Ref. 19) reported the was indented using a Vickers hardness some regions of the CGHAZ exceeded
distance between neighboring indents indenter (HV0.1) at exactly the same lo- 420 HV0.1 while adjacent regions were
has little effect on the measured hard- cations of the nanoindents. The in the range of 320 to 350 HV0.1.
ness. Even when the distance is as nanohardness values were plotted vs. By plotting all the data as a function
small as an indent diameter, reliable their corresponding HV0.1 values, and a of distance perpendicular to the fusion
hardness and elastic modulus values correlation relationship was obtained. boundary, the average and maximum
WELDING RESEARCH
WELDING RESEARCH
A B C
D E F
WELDING RESEARCH
A B
C D
Fig. 6 — Chemical composition gradient across the fusion boundary: A — Fusion boundary with narrow PGZ of aswelded sample; B —
fusion boundary with wide PGZ of aswelded sample; C — fusion boundary with narrow PGZ of sample after 660˚C/50 h; D — fusion
boundary with wide PGZ of sample after 660˚C/50 h.
eleven PWHT conditions and their cor- an average value of 250 HV0.1 while the The maximum and average hard-
responding hardness distribution weld metal hardened to more than 300 ness of CGHAZ, weld metal, and fu-
across the fusion boundary region. At HV0.1. Under these same tempering con- sion boundary PGZ from Fig. 7 are
620˚C/4 h, the metal and HAZ have ditions, there was no hardness peak at summarized in Fig. 8 as a function of
similar hardness but there is a narrow the fusion boundary. the HJP. The dashed line at 250 VHN
hardness decrease along the fusion At 660°C/50 h (Fig. 7F), the HAZ represents the hardness limit desig-
boundary on the base metal side (Fig. hardness decreased to approximately nated by the NACE Standard
7A), resulting from the carbon deple- 200 HV0.1 and a hardness peak began to MR0175/ISO15156 (Ref. 2) to ensure
tion in this area. appear at the fusion boundary. At resistance to SSC. Note that at HJP
Carbon diffusion and pileup in the 660°C/100 h (Fig. 7G) and 660°C/ values below approximately 19,500,
fusion boundary PGZ is not significant 500 h (Fig. 7H), the weld metal and the PWHT does not soften the CGHAZ
under PWHT conditions 620˚C/4 h, HAZ hardness remained relatively con- sufficiently to meet the NACE require-
640˚C/6 h, and 650˚C/5 h, and no hard- stant, but a distinct hardness peak was ment. PWHTs above this value de-
ness peak was observed along the fu- established at the fusion boundary. This creased the CGHAZ below 250 VHN,
sion boundary. As the tempering tem- is the result of carbon diffusion from but the hardness of the fusion bound-
perature increased, the weld metal con- the steel into the weld metal and the ary PGZ increased. Above ~20,000
tinued to harden while the CGHAZ soft- subsequent carbon “pileup” at the HJP, the PGZ began to harden signifi-
ened. At 670°C/10 h (Fig. 7E), the CG- boundary as a result of limited carbon cantly, reaching values above 350
HAZ had an almost uniform hardness at diffusivity in the Alloy 625 weld metal. HV0.1 due to carbon pileup. The hard-
WELDING RESEARCH
Fig. 7 — Hardness maps and hardness distributions of postweld heat treated samples: A — 620˚C/4 h; B — 640˚C/6 h; C — 650˚C/5 h.
WELDING RESEARCH
Fig. 7 — Hardness maps and hardness distributions of postweld heat treated samples: D — 650˚C/10 h; E — 670˚C/10 h; F — 660˚C/50 h.
WELDING RESEARCH
Fig. 7 — Hardness maps and hardness distributions of postweld heat treated samples: G — 660˚C/100 h; H — 660˚C/500 h.
WELDING RESEARCH
A B
C D
WELDING RESEARCH
y 59.552x 29.251, R2 0.9114. ing. Hodgson et al. (Ref. 1) discussed (Fig. 5C) should be Mo2C. On the
the tempering behavior of different trend curve of average hardness of
Using these linear equations, the heats of F22 steels starting from an the CGHAZ (the black square dot
peak values of nanohardness of samples as-welded condition, but they did not line) in Fig. 8, we can also see that the
650˚C/10 h, 670˚C/10 h, and investigate the microstructure evolu- CGHAZ hardness of sample 650˚C/10
660˚C/100 h can be converted to Vick- tion of the actual F22 CGHAZ as a h (HJP 19383) increased slightly
ers hardness values. The converted peak function of the HJP. In this investiga- compared to samples 650˚C/5 h and
VHN values are compared with actual tion, a single representative heat of 640˚C/6 h, most likely because of
measured HV0.1 values (from Fig. 8) in F22 steel was overlaid with Alloy 625 more Mo2C precipitation. This is the
Table 5. As expected, the converted filler metal using production condi- so-called “secondary hardening” ef-
peak nanohardness values exceed the tions. This resulted in a CGHAZ that fect caused by molybdenum carbides
peak HV0.1 values in all cases. This dif- is truly representative of actual serv- precipitation in low-alloy carbon
ference is largest under conditions ice components where the CGHAZ steels (Ref. 21). The secondary hard-
where carbon diffusion is the greatest. generated by the overlay pass may be ening occurs roughly in the range of
partially tempered by the adjacent HJP 19,100~19,400 (Fig. 8). Hodg-
pass or the one from the layer above son et al. (Ref. 1) also found the sec-
Discussion it. Thus, this investigation describes ondary hardening effect during
the effect of PWHT conditions on the PWHT of F22 steels in the range of
CGHAZ Microstructure and HAZ and fusion boundary mi- HJP 19000~19100. Considering the
Secondary Hardening crostructure that is closer to produc- slight difference in composition be-
tion conditions. tween what is reported here and
Postweld heat treatment of As a first approximation, it is possi- Hodgson’s work, as well as the differ-
F22/625 overlays was applied to re- ble to predict carbide formation in the ence of as-welded conditions, the sec-
duce the HAZ hardness to meet the CGHAZ of the postweld heat treated ondary hardening observed in this
NACE standard MR0175/ISO15156 F22/625 DMWs by plotting each PWHT study appears to confirm Hodgson’s
(Ref. 2) and to reduce susceptibility condition in the carbide precipitation result (Ref. 1).
to SSC. In the process of tempering diagram proposed by Baker and Nutting Upon further tempering
the martensitic HAZ of the steel, car- and shown in Fig. 11 (Ref. 20). For the (660˚C/10 h, HJP 19593), the sec-
bon diffusion from the CGHAZ to the PWHT condition 620˚C/4 h, the cemen- ondary hardening effect disappears
PGZ of the Alloy 625 weld metal can tite (Fe3C) preferentially formed at or (Fig. 8) because the Mo carbides dis-
potentially embrittle the interface. In near the prior austenite grain bound- solved in favor of Cr7C3 carbides. In
this investigation, a number of PWHT aries. When martensite was tempered the CGHAZ microstructures of sam-
conditions were used to identify a po- to ferrite, super-saturated carbon in the ples at 670˚C/10 h (Fig. 5D) and
tential “middle ground” between the martensite diffused to the prior austen- 660˚C/50 h (Fig. 5E), fine carbides
HAZ tempering effect and interface ite grain boundaries. Therefore, higher can be seen, which are probably Cr7C3.
embrittlement phenomenon to en- carbon concentration at or near the Further tempering to 660˚C/500 h re-
sure resistance to HAC and SSC. The grain boundaries resulted in the forma- sulted in the precipitation of M23C6,
as-welded CGHAZ contained mainly tion of cementite. according to Fig. 11. However, to con-
martensite. After PWHT, the marten- According to the diagram, the larg- firm the carbides type with certainty,
site was decomposed into carbides or er fraction of carbides in the CGHAZ transmission electron microscopy
ferrite based on the degree of temper- microstructure of sample 650˚C/10 h work needs to be done.
WELDING RESEARCH
Chemical Composition Gradient example, after tempering for response. It showed that a narrow
660˚C/50 h (HJP 20,154), the peak HJP range exists where the HAZ hard-
The composition gradient of substi- hardness at the interface was essen- ness drops below 250 VHN while
tutional elements iron, nickel, and tially equivalent to that of the hardening at the fusion boundary
chromium was not altered from the 670˚C/10 h (HJP 19,803) PWHT, as does not exceed 350 VHN. This sug-
as-welded condition even after tem- shown in Fig. 7E and F. However, an gested that PWHT conditions may ex-
pering at 660˚C/50 h. This result is additional 50 r at 660°C (HJP ist that meet the NACE hardness re-
consistent with Dodge’s work (Ref. 7). 20,526) resulted in a large increase in quirement while preventing embrit-
Carbon diffusion is strongly affected interface hardness, with values of tlement of the weld interface due to
by tempering conditions and repre- over 500 HV0.1. This suggests that carbon diffusion and pileup. Specific
sents the fundamental issue affecting temperature has more influence on conclusions from this investigation
the mechanical and corrosion behavior the tempering of the CGHAZ because are listed below.
of the weld interface region. Alexan- only short range diffusion is required, 1) The fusion boundary of the
drov et al. and Dodge studied the car- while longer times at the tempering F22/625 overlay exhibited a distinct
bon diffusion during PWHT using an temperature allow long range diffu- planar growth region, often referred to
electron probe microanalyzer and a sion that leads to the carbon deple- as a featureless zone. The width of this
DictraTM simulation (Refs. 5, 7). Based tion in the HAZ and carbon pileup at zone varied from approximately 10 to
on their studies, this investigation fo- the interface. Note that there is little 50 microns depending on location
cused on the hardness variation in the evidence of carbon depletion in the along the interface and was the widest
area very near to the fusion boundary HAZ, except at the most extreme when a base metal swirl was present.
using both standard microhardness PWHT conditions. 2) In the CGHAZ, martensite de-
and nanoindentation. Work is ongoing As shown in Table 5, the composed into ferrite and carbides
to better model carbon diffusion ki- nanohardness values of the PGZ are during PWHT, resulting in a reduction
netics to support both tempering and uniformly higher than the Vickers in hardness. Secondary hardening can
carbon pileup effects at the interface. hardness measurements. There are occur at HJP values above 19,000 due
two basic reasons for this difference. to the formation of alloy carbides pre-
Vickers Hardness and First, there is a size effect of the in- sumed to be Mo2C.
Nanohardness dent for hardness testing. This is 3) For HJP values in the range from
commonly observed in Vickers hard- 18,500 to 19,500, the HAZ hardness re-
The hardness data in Fig. 8 sug- ness testing, where hardness values mained relatively stable in the range of
gests that a PWHT “sweet spot” may increased with smaller loads (indent 250 to 300 HV0.1, and there was little
exist in the HJP range from 19,500 to size). Nanoindentation uses a much hardening at the fusion boundary.
20,000 where the CGHAZ hardness is smaller indent size than with the 4) In the HJP range from 19,500 to
below 250 HV0.1 and hardening at the 100-g Vickers load used here. Second, 20,000, the HAZ hardness decreased
fusion boundary is limited to below nanoindentation can differentiate a to below 250 HV0.1 with little increase
350 HV0.1. In this HJP range, the re- very small microstructural region in the interface hardness. For example,
sistance to HAC and/or SSC should be that has very high hardness and thus a PWHT at 660°C for 10 h resulted in
optimum for the F22/625 DMW. The avoids the averaging effect that oc- an average HAZ hardness of 250 HV0.1
PGZ hardness at the fusion boundary curs with the Vickers indentations and an interface hardness of approxi-
can potentially tolerate values in ex- (Ref. 25). The converted VHN values mately 300 HV0.1.
cess of 250 HV0.1. This is because the in Table 5 are within the correspon- 5) At HJP values above 20,000, sig-
area of the high hardness in the PGZ ding error-bar ranges of measured nificant hardening of the interface oc-
after PWHT is much narrower than Vickers hardness shown in Fig. 8. cured due to carbon diffusion and
the CGHAZ, and the diffusivity of hy- Thus, the linear conversion relation- “pileup” in the PGZ. Peak interface
drogen in the austenitic PGZ is much ship from nanohardness to Vickers hardness values exceeded 450 HV0.1
lower than in the CGHAZ (Refs. 5, hardness appears reasonable. Such under these PWHT conditions.
22–24). Thus, there is a higher proba- linear conversion relationships were 6) Nanohardness measurements
bility for hydrogen to diffuse into the also reported by other researchers, in- at the interface indicated that the
CGHAZ, and initiate cracking. At low- cluding Mencin et al. (Ref. 13), Rice et hardness levels may be higher than
er values of HJP, the CGHAZ does not al. (Ref. 14), Sawa (Ref. 15), and determined using standard Vickers
meet the NACE hardness requirement, Zhang et al. (Ref. 16). measurements.
while at higher values local failure at 7) Nanohardness and Vickers hard-
the fusion boundary would be expect- Summary and Conclusions ness have a linear proportional corre-
ed. Ongoing work using the delayed lation relationship with relatively high
hydrogen cracking test (DHCT) is de- This investigation has shown the reliability (R2 80%).
signed to verify this hypothesis. influence of PWHT on the tempering 8) The PWHT response of the
The HJP is most effective at pre- response of the F22 HAZ and con- F22/625 dissimilar weld overlay can
dicting the tempering response of the comitant hardening of the weld fusion be used to select PWHT conditions to
CGHAZ, but provides only limited in- boundary and Alloy 625 weld metal. determine optimum resistance to hy-
sight into the formation of the high The Hollomon-Jaffe parameter (HJP) drogen-assisted cracking and sulfide
hardness zone at the interface. For was used to quantify the tempering stress cracking.
WELDING RESEARCH
Acknowledgments
ment of Ni–base alloy overlays applied to 15. Sawa, T. 2010. Correlation between
carbon steel. Welding in the World 57: nanoindentation test result and vickers
39–53. hardness. IMEKO 2010 TC3, TC5, and
6. Fenske, J. A. 2010. Microstructure TC22 Conferences, Pattaya, Chonburi,
This work was supported by and hydrogen induced failure mechanisms Thailand.
Cameron International (now Shlum- in iron-nickel weldments. Dissertation, 16. Zhang, L., Ohmura, T., and Tsuzaki,
berger) through the NSF I/UCRC, University of Illinois at Urbana-Cham- K. 2012. Application of nanoindentation
Manufacturing and Materials Joining paign, Urbana, Illinois. technique in martensitic structures.
Innovation Center (MA2JIC) at The 7. Dodge, M. F. 2014. The effect of heat Nanoindentation in Materials Science, In-
Ohio State University. Special thanks treatment on the embrittlement of dissim- Tech: 109–130.
to Acute Technological Services for ilar welded joints. Doctoral thesis, Univer- 17. Bourgeois, D. 2015. Hydrogen as-
producing the Alloy 625 overlays on sity of Leicester, Leicester, UK. sisted crack in dissimilar metal welds for
F22 forgings. Also, the technical sup- 8. Gedeon, S. A., and Eagar, T. W. 1990. subsea service under cathodic protection.
Thermomechanical analysis of hydrogen Dissertation, Columbus: The Ohio State
port and insight provided by Dean
absortion in welding. Welding Journal University.
Hannam and Nash Ubale from Shlum- 69(7): 264-s to 271-s. 18. Hollomon, J., and Jaffe, L. 1945.
berger was greatly appreciated. 9. Ouden, G., and Griebling, O. 1990. Time-temperature relations in tempering
Recent Trends in Welding Science and Tech- steel. Metal Technology 12: 223–249.
nology. Materials Park, Ohio: ASM 19. Kim, H. J., and Kim, D. E. 2013. Ef-
References International. fects of proximity on hardness and elastic
10. Beaugrand, V. C., Smith, L. S., and modulus measurements of SiO2 and Cu by
Gittos, M. F. 2009. Subsea dissimilar joints: nanoindentation. Tribol Lett 49: 85–94.
1. Hodgson, D. K., Dai, T., and Lippold, Failure mechanisms and opportunities for 20. Baker, R., and Nutting, J. 1959. The
J. C. 2015. Transformation and tempering mitigation. Corrosion Paper #09305. tempering of 2.25Cr-1Mo steel after
behavior of the heat-affected zone of 11. Dodge, M. F., Dong, H. B., Militit- quenching and normalizing. The Journal of
2.25Cr-1Mo steel. Welding Journal 94(8): sky, M., Barnett, R. P., Marques, V. F., and the Iron and Steel Institute 192: 257–268.
250-s to 256-s. Gittos, M. F. 2012. Environment-induced 21. Totten, G. E. 2007. Steel Heat Treat-
2. ISO 15156-2:2009, Petroleum and cracking in welding joints in subsea oil and ment Handbook, 2nd edition, Boca Raton:
natural gas industries — Materials for use in gas systems — Part I. The Welding Insiti- Taylor & Francis Group.
H2S-containing environments in oil and gas tute, Cambridge, UK. 22. Easterling, K. 1983. Introduction to
production — Part 2: Cracking-resistant car- 12. Fenske, J. A., Robertson, I. M., Ayer, the Physical Metallurgy of Welding, 1st edi-
bon and low alloy steels and the use of cast R., Husle, M., Lillig, D., and Newbury, B. tion. Butterworths Monographs in
irons. 2012. Microstructure and hydrogen- Materials.
3. Troiano, A. R., and Hehemann, R. F. induced failure mechanisms in Fe and Ni 23. Bailey, N., Coe, F. R., Gooch, P. H.,
1982. Hydrogen embrittlement and stress alloy weldments. Metallurgical and Materi- Hart, M., Jenkins, N., and Pargeter, R. J.
corrosion cracking in sour environment. als Transactions A 43A: 3011–3022. 1993. Welding Steels without Hydrogen
Current solutions to hydrogen problems in 13. Mencin, P., Van Tyne, C. J., and Cracking, 2nd edition. Nashville: Abington
steel. Proc. of the First Int’l., C. G. I. a. G. Levy, B. S. 2009. A method for measuring Publishing. Vol. 12A.
M. Pressouyre, Ed., Ohio: ASM Metals the hardness and elastic modulus of the 24. Sykes, C., Burton, H. H., and Gregg,
Park. 299–302. surface layer on hot forging dies using a C. C. Hydrogen in steel manufacture. J.
4. Berkowitz, B. J., and Heubaum, H. nano indentation. Journal of Materials En- Iron Steel Inst. 156: 155–180.
1984. The role of hydrogen in sulfide gineering and Performance 18(8): 25. Oliver, W. C., and Pharr, G. M.
stress cracking of low alloy steels. Corrosion 1067–1072. 2004. Measurement of hardness and elas-
40(S): 240–245. 14. Rice, P. M., and Stoller, R. E. 2000. tic modulus by instrumented indentation:
5. Alexandrov, B. T., Lippold, J. C., Correlation of Nanoindentation and Conven- Advances in understanding and refine-
Sowards, J. W., Hope, A. T., and Saltzmann, tional Mechanical Property Measurements. ments to methodology. Journal of Materials
D. R. 2013. Fusion boundary microstruc- MRS Online Proceeding Library Archive, p. Research (JMR) 19: 3–20.
ture evolution associated with embrittle- 649.
TAO DAI (dai.234@osu.edu) and JOHN C. LIPPOLD (lippold.1@osu.edu) are in the Welding Engineering Program, The Ohio State University,
Columbus, Ohio.