Fracture modes and acoustic emission characteristics of hydrogen-assisted cracking in high-strength low-alloy steel weldment

Fracture Modes and Acoustic Emission Characteristics of Hydrogen-Assisted Cracking in High-Strength Low-Alloy Steel Weldment

Roseana da Exaltação Trevisan

Hipolito Carvajal Fals

Departamento de Engenharia de Fabricação

Faculdade de Engenharia Mecânica

Universidade Estadual de Campinas

13083-970 Campinas, SP Brazil

[e-mail]

Abstract

The aim of the present paper is to study the relationship between the fracture modes in hydrogen-assisted cracking (HAC) in microalloied steel and the emission of acoustic signals during the fracturing process. For this reason, a flux-cored arc weld (FCAW) was used in a high-strength low-alloy steel. The consumable used were the commercially available AWS E120T5-K4 and had a diameter of 1.6 mm. Two different shielding gases were used (CO₂ and CO₂+5% H₂) to obtain complete phenomenon characterization. The implant test was applied with three levels of restriction stresses. An acoustic emission measurement system (AEMS) was coupled to the implant test apparatus. The output signal from the acoustic emission sensor was passed through an electronic amplifier and processed by a root mean square (RMS) voltage converter. Fracture surfaces were examined by scanning electron microscopy (SEM) and image analysis. Fracture modes were related with the intensity, the energy and the number of the peaks of the acoustic emission signal. The shielding gas CO₂+5% H₂ proved to be very useful in the experiments. Basically, three different fracture modes were identified in terms of fracture appearance: microvoid coalescence (MVC), intergranular (IG) and quasi-cleavage (QC). The results show that each mode of fracture presents a characteristic acoustic signal.

Keywords: Microalloyed steel, welding, hydrogen-assisted cracking, fracture modes, acoustic emission

Introduction

Hydrogen-assisted cracking in a low carbon microalloyed steel occurs mainly in weld metal, influenced basically by the accumulation of hydrogen, the microstructure and the particular properties of this steel.

The flux-cored arc-welding (FCAW) process combines versatility and high productivity for industrial use. Recent investigations have shown that flux-cored filler metals can produce hydrogen content in the weld metal of 0.9 to 6.6 ppm deposited metal (Nippes and Xiong, 1988). The low hydrogen potential of the FCAW process makes it a superior choice for the welding of mild steel. However, the amount of hydrogen necessary to cause cold cracking in steels is inversely related to yield strength. Therefore, high-strength steels may be susceptible to hydrogen-assisted cracking when welded by the FCAW process.

In a high-strength steel, crack initiation occurs predominantly in the heat-affected zone. By contrast, in a low-carbon microalloyed steel, crack begins to form and grows in the weld metal (Vuick,1992).

According to the existing model for hydrogen-assisted cracking (Beachem, 1972), intergranular (IG), quasi-cleavage (QC), or microvoid coalescence (MVC) fracture modes operate depending upon the microstructure, the crack-tip stress intensity and the concentration of hydrogen.

Beachem (1972) also suggested the interrelationship, shown in Fig. 1, between the stress intensity factor, the dissolved hydrogen content, and the hydrogen-assisted cracking fracture mode. The IG mode predominates when the stress intensity factor is low. At higher stress intensities, the energetically favorable IG process is replaced by the QC and MVC proceses; MVC and QC are relatively faster fracture modes compared with the IG mode.

Fig. 1 Beacham diagram (Yurioka and Suzuki, 1990)

Using the implant test, Vasudevan, Stout and Pense (1981) observed that the crack morphology in the microalloyed steel was predominantly intergranular at a low stress intensity factor, progressively giving way, to quasi-cleavage and microvoid coalescence with an increasing stress intensity factor. Those results were also reported by Yurioka and Suzuki (1990) and were well interpreted by the Beachem diagram shown in Fig. 1. Point X in Fig. 1 denotes the combination of stress intensity and hydrogen concentration just after welding. As time passes, hydrogen accumulates at the notch tip. When the hydrogen concentration reaches the critical level at point A, a crack begins to form. In a self-restraint weld, residual stress is relaxed and stress intensity decreases with crack growth, and then the crack eventually stops. In an external weld cracking test (implant test), however, stress intensity continues to increase and the crack growth rate is accelerated as the crack grows.

Gedeon and Eagar (1990), observed the interrelationship between the stress intensity factor and hydrogen-assisted cracking fracture mode; as advanced by the Beachem theory. They also found that crack propagation at a higher hydrogen level is predominantly IG.

It is now generally accepted that the kinetics and morphology of growth of hydrogen-assisted cracking is sensitively influenced by microstructural features and other factors including hydrogen content, strain rate, temperature and the stress intensity factor.

Acoustic emission is a technique now used extensively to monitor the propagation of cracks and other flaws in stressed structures. This technology has been a wide variety of applications, such as proof testing and failure mechanism discrimination of composite structures, monitoring of manufacturing processes, etc.

Acoustic signals are emitted during crack formation and propagation. These signals can be detected by an acoustic emission measurement system (AEMS) attached to the test piece. Many authors have worked with such a technique to detect the crack formation and growth.

Racko (1987) used acoustic emission signals to confirm theoretical models of determining cold cracking. The author related the acoustic emission spectrum to crack length.

Hippsley, Buttle and Scruby (1988) discussed the phenomenon of stress-relief cracking and the low-ductility intergranular fracture mode for the temperature of 500 ^oC, using an acoustic emission system. The authors showed that the crack propagation time of a low-ductility intergranular fracture is approximately 10 milliseconds. The authors also showed that fractures with small dimples are more difficult to detect with acoustic emission systems. This is because the signal amplitude is ten times smaller than in the intergranular fracture mode.

Using a AEMS attached to the modified implant test, Ferraresi and Trevisan (1997) showed the possibility of using acoustic emission signals to detect the initiation and growth of reheat cracking in high-strength low-alloy steel.

The aim of the present research is to identify the different hydrogen-assisted cracking fracture modes using the implant test and the acoustic emission measurement system.

Experimental Procedure

All tests were performed on a commercially graded, quenched and tempered high-strength low-alloy steel. Chemical composition and mechanical properties of the material used are given in Table 1.

Table 1 Composition (Wt.%) and Mechanical Properties of HSLA Steel

Input was maintained at 0.659 kJ/mm. A T-120 (E120-T5-K4) wire was used with CO₂ and CO₂+5%H₂ shielding.

All tests were carried out using the flux-cored arc-welding (FCAW) process, and the heat Diffusible Hydrogen Analysis

The diffusible hydrogen contents of FCA welds were determined by gas chromatography using an Oerlinkon chromatograph model GS-1006H in conformity with ISO 3690 and ANSI/AWS A4.3-86. Table 2 presents the diffusible hydrogen content for welds made with the two different gases.

Table 2 Average Levels of Diffusible Hydrogen for Each Gas

Implant Test

Among the various testing methods for assessing hydrogen embrittlement, the implant test has become one of the most popular for scientific investigation of the cracking phenomenon in welds. This is due to the fact that stress, hydrogen level and microstructure can be independently varied and controlled.

Test procedures were taken from the French standards for hydrogen cracking tests (NF A 89-100) in order to guarantee the reproducibility of the test, since helical grooves are safer than circular grooves, in terms of the position in the HAZ. A simplified scheme of the equipment is shown in Fig. 2.

Fig. 2 Experimental Setup for Monitoring AE and Force Signals in Implant Test Rig

Tests were carried out using three levels of initial load, namely 80%, 76% and 66% of the material yield point having CO₂ shielding and 66%, 57% and 52% of the material yield point having CO₂+5%H₂ shielding. During the tests, the loading history was monitored by a data acquisition system, as shown in Fig. 2.

A scanning electron microscope (SEM) was used to examine the fracture surfaces of the implant specimens. Quantitative fractography was performed to map the various fracture mode zones across the surface of the implant specimens, using Global image analyzer.

Acoustic Emission System

A block diagram of a computerized AE system directly coupled to an implant test rig is illustrated in Fig. 2.

The AE system used in this work to monitor the instant of formation and propagation of cracks consists of an AE sensor, an amplifier, a root mean square (RMS) voltmeter and a data acquisition system. This configuration is considered to be robust, cheap, highly sensitive and flexible, and easy to assemble.

The output signal from the AE sensor (piezoelectric transducer) was passed through an electronic amplifier (X1000 gain) and processed by a RMS voltage converter. The RMS signal pulse was recorded in a microcomputer by an analog/digital conversion card.

The AE sensor is classified as a wideband sensor, model WD, with a typical operation range of 100-1000 kHz. The amplifier functions at 40/60 dB. The RMS converter was built in the Manufacturing Department of FEM/UNICAMP. This equipment determines the AE energy that is proportional to the integral of the square of the AE sensor output voltage.

Results and Discussion

The results of metallographic observation show that all the cracks began to form in the coarse-grained region of the HAZ and grew towards the fusion zone (FZ). This was primarily related to the hardness and concentration of hydrogen in the weld deposit. The cracking behavior of this steel was confirmed by Vuick (1992). The cracking zone in all implant fractures was in the weld fusion zone.

Figure 3 shows that the rupture time drops remarkably when the applied stress increases at every hydrogen level examined, confirming Gedeon and Eagar (1990) studies.

Fig. 3 Implant Test Results. Time to Failure versus Stress

As can be seen in the same figure, time to failure decreases when the hydrogen concentration increases. In all cases, the load was applied to the implant specimen and maintained until either the rupture occurred. The lower critical stress (LCS) was determined for each shielding gas used. The values of LSC were 459 MPa for low hydrogen and 318 MPa for high hydrogen shielding gas.

Using SEM, it is shown that hydrogen crack fracture surfaces can exhibit a wide variety of modes. Crack fracture surfaces observed covered the range of microvoid coalescence (MVC), quasi-cleavage (QC) and intergranular (IG) modes.

Figure 4(a) shows a typical overall view of the fracture surface of an implant specimen. Figure 4 also shows the location of each of the magnified photos taken of this surface. Figure 4(b) shows the

Fig. 4

The samples were also mapped to show regions of the different fracture modes. A typical quantitative fracture map is shown in Fig. 5.

Fig. 5 Quantitative Fracture Map Showing the Regions of Fracture Modes

The quantitative relationship between stress and the areas of fracture modes for each level of hydrogen used is shown in Table 3. It can be seen that the proportions of the QC, MVC and IG areas depend on the amount of hydrogen dissolved in the weld metal and also on stress intensity. The results also show (Table 3) that IG morphology increased with increasing stress when CO₂+5% H₂ was used, as suggested by Gedeon and Eagar (1990).

Table 3

Correlation between Stress and Average Area of the Fracture Mode for each gas (YS-Yield Strength)

The resulting plot of applied stress versus QC area is shown in Fig. 6. It can be observed that the area of QC morphology increased with increasing stress when CO₂+5% H₂ was used. The opposite tendency is observed for pure CO₂. At the same time, it can be observed that the QC morphology area decreases when MVC morphology increases, which agrees with the Beachem microplasticity model (Beachem, C.D., 1972).

Fig. 6 Correlation Between Stress and QC Average Area

SEM analysis and quantitative fracture maps were correlated with the acoustic emission measurements. It was initially observed that in all tests, events of acoustic emission with a higher amplitude were preceded by several events of a lower amplitude, which is related to coexistence with the IG, MVC and QC modes on the fracture surface.

The experimental data from AE signals which depend on the applied stress levels are shown in Table 4. These data are also related to different gases.

Table 4 Listing of Acoustic Emission Data.

Table 4 reveals an increase in all acoustic emission data for an increasing stress, when CO₂ was used, opposite to the results observed when a higher level of hydrogen was used. This result is due to the relation between the fracture modes that act in the cracking of the samples, under the conditions used, as will be analyzed below.

A sample of the results is illustrated in the plots in Fig. 7(a), which shows the behavior of average events of acoustic emission signals, as a function of QC area. The lower values of acoustical emission data were presumably caused because this fracture mode is characterized by little plastic deformation. On the other hand, it was proved that the higher values of acoustic emission data were caused by an increase in the MVC area, as can be observed in Fig. 7(b).

Fig. 7

In fact, the increase in acoustic emission data was produced by the occurrence of more plastic deformation in the MVC mode than in the QC mode. The Beachem model adequately explains the presence of plastic deformation preceding hydrogen cracking, caused by the nucleation of a void around the main crack, presumably from decohesion of inclusions, when large volumes of material are plastically deformed at the crack tip in the MVC mode.

In the analysis of Fig 7(b), it is verified that in the tests using CO₂+5% H₂, in spite of the existence of a smaller fracture area for MVC, the average number of AE signal events was approximately the same, when compared with the experiments made with a lower level of hydrogen. This result is caused by the increase in the IG fracture area because of the increment in the level of hydrogen. It should be explained that the IG fracture generates an increase in the number of AE events; but these possess a low amplitude and a low level of energy as can be seen checked in the analysis in Table 4, also suggested by Carpenter and Smith (1990).

Also, it was observed that when a fracture surface display a larger transition region compound morphology in the MVC and QC modes in the samples welded with higher level of hydrogen, it causes an increase in plastic deformation and consequently higher AE signals.

Conclusions

The implant test with the introduction of CO₂ + 5% H₂ shielding was shown to be an adequate tool to induce HAC with lower times and stresses.

Using the proposed procedure, the study was able to confirm that all fractures occurred in the fusion weld zone and the HAC by three different fracture modes: Intergranular (IG), Microvoid Coalescence (MVC), Quasi-cleavage (QC).

An excellent correlation between the acoustic emission data and the quantitative fracture map was obtained.

The AE signals were helpful to detect the HAC. Larger areas of the QC mode were associated with lower AE signals the larger areas of MVC fracture modes were associated with higher AE signals, regardless of hydrogen level.

Acknowledgment

The authors would like to express their gratitude to the Sao Paulo State Council, FAPESP, for their financial support of this project.

Manuscript received: Frebruary 1999. Technical Editor: Allison Rocha Machado

• Beachem, C.D., 1972, "A New Model for Hydrogen-Assisted Cracking (Hydrogen Embrittlement)"; Metallurgical Transactions; vol. 3; p. 437.
• Beattie, A.G., 1983, "Acoustic Emission, Principles and Instrumentation", Journal of Acoustic Emission, vol.2. pp. 95-128.
• Carpenter, S.H. and Smith, D.R., 1990,"The Effects of Cathodic Charging on the Acoustic Emission Generated by Intergranular Cracking in Sensitized 304 Stainless Steel", Metallurgical Transactions A, vol. 21A, p. 1933.
• Fang, C.K.; Kannatey-Asibu, E. and Barber, J.R., 1995, "Acoustic Emission Investigation of Cold Cracking in Gas Metal Arc Welding of AISI 4340 Steel"; Welding Journal, June, p. 178s.
• Ferraresi, V.A and Trevisan R.E., 1997, "Utilização da Emissão Acústica no Estudo de Reaquecimento", Proceedings, XIV Brazilian Congress of Mechanical Engineering, Bauru, SP, Brazil.
• Hippsley, C.A; Buttle, D.J. and Scruby, C.B., 1988, A Study of the Dinamics of High Temperature Brittle Intergranular Fracture by Acoustic Emission. Acta Metallurgica; vol. 36; n. 2; p. 441.
• Gedeon, S.A. and Eagar; T.W., 1990, "Assessing Hydrogen-Assisted Cracking Fracture Modes in High-Strength Steel Weldments"; Welding Journal, June, p. 213s.
• Mashino, S., et. al., 1996, "Analysis of Microfracture Mechanism of Titanium Alloy by Acoustic Emission Technique", Materials Science and Engineering, A 213, pp. 66-70.
• Moorthy, V., Jayakumar, T. and Baldev, R., 1996, "Acoustic Emission for Identification of Stage II-a to II-b Transition During Stage II Fatigue Crack Growth in Parent and Weld Metal in AISI 316 austenitic Stainless Steel", Materials Science and Technology, vol. 12, p. 56.
• Nippes, E.F.and Xiong, D.J., 1988, "Investigation of Hydrogen-Assisted Cracking in FCA Welds on HY-80 Steel", Welding Journal, June, p. 132s.
• NF A 89-100, 1982, "Cold Cracking Test Methods Using Implants", April, 10p.
• Racko, D., 1987, "Acoustic Emission from Welds as Indicator of Cracking", Materials Science and Technology, December, vol. 3, p. 32.
• Savage, W. F., Nippes, E.F. and Homma, H., 1976, "Hydrogen Induced Cracking in HY-80 Steel Weldments", Welding Journal, vol. 61(8), p. 368s.
• Vasudevan, R.; Stout, R.D. and Pense, A.W., 1981, "Hydrogen-Assisted Cracking in HSLA Pipeline Steels", Welding Journal, 60 (9), p. 155s.
• Vuick, J., 1992, "An Update of the State-of-the-Art of Weld Metal Hydrogen Cracking", Welding in the Word (Le Soudage Dans Le Monde), vol. 31, n. 5, pp. 308-321.
• Yurioka. N. and Suzuki. H., 1990, "Hydrogen Assisted Cracking in C-Mn and Low Alloy Steel Weldment", International Materials Reviews; vol. 35; n. 4; p. 56.

Publication Dates

• Publication in this collection
  11 Oct 2001
• Date of issue
  Dec 1999

History

• Received
  Feb 1999