Hydrogen effects in high-strength steel weldments have been well-studied by many industries due to their role in structural failures. When diffusion and aggregation of hydrogen occurs in steels, a phenomenon called hydrogen induced cracking (HIC) can occur. HIC can reduce fatigue performance and ductility, and the fabrication controls needed to avoid it can be very costly. The four necessary factors that govern HIC susceptibility are: (1) the amount of diffusible hydrogen in the weld metal or heat affected zone (HAZ), (2) the amount of tensile residual stress, (3) the presence of a susceptible weld metal and/or HAZ microstructure, and (4) temperatures between 200°C and -100°C. Welding inherently produces high tensile residual stress (particularly in thick sections), and HIC-susceptible microstructures and service temperatures are common for high-strength steel structures.
Over the last decade, the U.S. Navy has been investigating a high-strength steel welding consumable containing 10 wt.% Ni, dubbed 10Ni steel, that has demonstrated a desirable combination of high strength and good toughness. The material is primarily martensitic but contains 3-5 wt.% retained austenite. It is hypothesized that the retained austenite in the 10Ni steel weld metal may provide a sink (or trap) for hydrogen, potentially reducing the necessary preheat temperature required to mitigate HIC. The objective of this work was to conduct a series of experiments to quantify the evolution of hydrogen from 10Ni steel weld deposits and evaluate its HIC susceptibility using small-scale weldability tests.
The hot carrier gas extraction method was used to quantify diffusible hydrogen in 10Ni steel, MIL-100S-1, and two MIL-120S-1 weld deposits. All wires used were 1.2mm diameter. 10Ni steel wire was bare, while MIL-100S-1 and MIL-120S-1 wire were copper coated. Deposits fabricated using 98% Ar-2% O2 and 95% Ar-5% CO2 shielding gases were evaluated at test temperatures between 400°C and 625°C. The tests performed at 400°C were in accordance with AWS A4.3, Addendum1, and higher temperature tests performed to release hydrogen from potential trapping sites.
The second segment of testing was performing GBOP tests, in accordance with AWS B4.0, using 10Ni steel, MIL-100S-1, and MIL-120S-1 weld deposits. The filler material, metal transfer mode, preheat, and heat input were the variables of interest with test conditions repeated in triplicate. The welds were clamped for a minimum of 97 hours before being heat tinted and fractured to reveal hydrogen cracking. Heat tinted areas indicative of hydrogen damage were measured using ImageJ image processing and analysis software to determine the percent area of cracking. Weld metal chemistry and microhardness were measured on one weld per parameter set to add additional context for differences in percent cracking.
Empirical models from the literature for prediction of critical preheat for 0% cracking in the GBOP test were evaluated against the experimental results collected in this study and typical high strength steel preheat requirements.
At an extraction temperature of 400°C, 10Ni steel demonstrated diffusible hydrogen values of 1.10 mL/100g and 1.15 mL/100g when deposited using the spray metal transfer gas metal arc welding (GMAW) process using 98% Ar-2% O2 and 95% Ar-5% CO2 shielding gas, respectively. The low diffusible hydrogen contents were attributed to the fact that the electrode was both lubricant and feed aid free. During the higher temperature diffusible hydrogen tests, the average measured values increased to 1.70 mL/100g at 500°C and 1.72 mL/100g at 625°C.
During GBOP testing of 10Ni steel, 0% cracking was achieved with spray transfer GMAW by preheating the test piece to 65.6°C for a heat input of 0.984 kJ/mm and at room temperature for a heat input of 2.322 kJ/mm. The weldments testing the effect of metal transfer mode showed that pulsed GMAW had a lower average percent cracking than spray transfer GMAW under the same heat input and preheat conditions. As a comparison to 10Ni steel, weld deposits from two different heats of MIL-120S-1 and one heat of MIL-100S-1 were tested at 65.6°C preheat and a heat input of 0.984 J/mm. The first heat of MIL-120S-1, which contained 7.78 mL/100g of diffusible hydrogen, demonstrated 72% cracking. The second heat of MIL-120S-1 contained 4.60 mL/100g of diffusible hydrogen and demonstrated 79% cracking. The large differences in cracking between 10Ni steel and MIL-120S require additional testing to determine if they were primarily caused by lower HIC susceptibility or the lower diffusible hydrogen content. The MIL-100S-1 wire contained 3.2 mL/100g of diffusible hydrogen and demonstrated 40% cracking.
The empirical models for prediction of critical preheat for 0% cracking had poor agreement with the weld deposits evaluated in this study. The models are based on various carbon equivalence (CE) equations derived for the select chemistry ranges used during their development. To better predict the necessary preheat for current high strength steels and the 10Ni steel chemistries being tested, different CE equations were applied to the literature models. However, none of the CE equations’ chemistry ranges fully encompassed the 10Ni steel’s chemistry. The changes in the empirical models’ predicted preheats were driven by differences in CE equations used.
10Ni steel weld deposits were investigated to better understand their hydrogen content and susceptibility to HIC. Diffusible hydrogen testing demonstrated low diffusible hydrogen contents of bare wire 10Ni steel weld deposits. The low amount of diffusible hydrogen was attributed to the lack of hygroscopic feed aids or lubricants on the 10Ni consumable. Future work using 10Ni steel wires with feed aid will be needed as a comparison.
GBOP testing showed that 0% cracking for 10Ni steel deposits can be achieved by increases in preheat temperature and heat input. Additionally, the 10Ni deposits appeared to require a lower preheat for 0% cracking than the MIL-120S-1 and MIL-100S-1 deposits. The cracking seen during testing using MIL-120S-1 was likely convoluted by the much higher diffusible hydrogen content of those deposits. However, the MIL-100S-1 results offer more conclusive evidence of the promising behavior of 10Ni steel.
The five equations investigated to predict necessary preheat and evaluate the severity of the GBOP testing showed varying results and were largely depended on the CE equation used.