Measurement of hydrogen entry into the heat affected zone during weld cladding

TWI Bulletin, June 1985

by David Noble

David Noble, BSc (Eng), ARSM, is a Research Metallurgist in the Stainless Steels and Corrosion section of the Materials Department.

A method for directly monitoring the hydrogen flux into a substrate HAZ during cladding with austenitic stainless steel consumables is described. The technique can in principle be used for a range of joint geometries and processes to examine more fully the roles of welding variables in controlling hydrogen ingress into the HAZ and therefore assist in the prevention of hydrogen cracking.

The embrittling effect hydrogen has on steel has been well documented, ^[1] and can cause serious problems during the fabrication of steel structures as well as in service. In many ways arc welding is an 'ideal' technique for charging a localised region of metal with hydrogen because of the heat associated with the arc, the development of a molten pool, and the inherent problems in keeping the weld area free from contamination and consequently hydrogen.

During welding, the hydrogen which enters the weld pool can originate from several sources, e.g. moisture or other hydrogen containing compounds in the flux or shielding gas, or hydrogen may enter the weld metal via the wires themselves if they are dirty or greasy.

With conventional G-Mn structural and pipeline steels, the risk of hydrogen cracking on welding can generally be minimised by adjusting the arc energy or preheat temperature so that the cooling rate is reduced; the aim is to promote non-hardened microstructures in the weld area since these have reduced susceptibility to hydrogen embrittlement.

The method is described fully in ref. ^[1] . For more highly alloyed grades, however, such microstructural control may not always be feasible, and in this case reliance must be placed on control of the joint temperature before and especially after welding so that the material is held at a sufficiently high temperature for hydrogen to diffuse away from the weld HAZ.

A Group Sponsored investigation was carried out at The Welding Institute into factors influencing the risk of substrate hydrogen cracking when weld cladding alloy steels as may be used for heavy section pressure vessels. In these situations, control of the microstructure is virtually impossible, and as part of this work, it was necessary to examine hydrogen entry to the HAZ during one and two layer cladding cycles. Mathematical analysis of HAZ hydrogen pick-up has been carried out, ^[2] but little information exists from direct measurements.

A technique was therefore developed to monitor hydrogen flux, and is described below.

Experimental method

Analytical technique

The work was based on locating a hole in the substrate HAZ close to the fusion boundary, and thereby directly measuring hydrogen flow from the weld metal. For this purpose, a sampling hole was drilled to within about 5mm of the surface to be clad since preliminary trials had shown that, with the hole this distance from the surface of the plate, it would be positioned within the coarse grained HAZ after welding and experience the maximum hydrogen flux.

The gas analysis technique had to be discriminating to ensure that only hydrogen gas was detected. The system also had to be sufficiently flexible to respond rapidly to changes in hydrogen flux, without giving spurious information as a consequence of sampling hole temperature fluctuations. For these reasons mass spectrometry was used, and to ensure high sensitivity, no carrier gas was employed in the system but the tubes were evacuated before sampling. The recording system employed enabled a hard copy of the hydrogen pressure versus time to be taken. A diagram showing the positioning of the sampling holes and the analytical equipment is shown in Fig.1.

Welding details

A two layer strip submerged arc welding process was used to simulate industrial cladding operations. Type 309L strip, 30mm wide by 0.5mm thick was the consumable used for the first layer, with type 308L strip of the same dimensions for the second layer. In both cases, a chromium compensating flux was used in the as-received condition. The hydrogen sampling holes were located beneath the centre of the top layer bead and below two first layer bead overlaps ( Fig.1) since this area has been considered by some to be the most prone to cracking. ^[3]

Re-austenitisation of the HAZ by second layer bead deposition has also been regarded as a significant factor in promoting cracking and hence the sampling holes, being under the centre of the second layer beads, were appropriately placed to record any surges in hydrogen flux when re-austenitisation occurred. Five weld beads were laid in total, three first layer beads, which overlapped by approximately 20%, and two second layer beads, each centred over the first layer overlap. The sequences are shown in Fig.2. Welding parameters employed were 650A, 110 mm/min travel speed and 30V.

Thermal cycle measurements were performed in weld HAZs, and a base plate of sufficient thickness to represent an effectively infinite heat sink was chosen: an 80mm thick block of ferritic steel 300mm square. Steel conforming to ASTM A533B specification was chosen because of availability and as representative of alloy steels used in heavy section pressure vessel construction.

In addition to a hole for hydrogen sampling, another hole was drilled to the same distance from the steel surface as the sampling hole, in line with the anticipated centre of the top layer weld bead and approximately 200mm 'behind' the hydrogen hole. A Pt/Pt 13% Rh thermocouple was placed in the second hole for the purpose of recording HAZ temperatures before, during and after welding.

Weldment temperature control

The test described here was supposed to simulate industrial cladding where the time elapsed between adjacent weld bead deposition would be significant because of the overall size of the component. Consequently there was a delay of about 1 hr between deposition of adjacent passes in the first and second layers. If the test was preheated, then the preheat remained on during the hour. There was a further delay of approximately 24 hr between welding of the first and second layers of cladding. Details of the thermal treatments for each test are given in Table 1.

Table 1 Heat treatments and sequence of events for test welds

Weld deposit hydrogen measurements

Prior to direct HAZ hydrogen measurements being made in the sponsored programme, effort was put into the generation of a bank of hydrogen data which evaluated potential and deposit hydrogen contents for a number of consumables in different conditions ( e.g. as received and dried) as well as for different welding processes e.g. MMA and strip submerged arc. Using the same welding conditions as would be employed in the test described above to simulate industrial cladding, a bead of 309L strip was deposited using the submerged arc process on to a ferritic steel base plate which was in turn clamped in a water cooled copper jig, and after welding the sample was quenched in liquid nitrogen. The purpose of this technique was to permit as little hydrogen as possible to escape from the test bead before vacuum hot extraction of the remaining hydrogen in the weld metal. The technique is commonly used and details can be found elsewhere. ^[4] A deposit hydrogen content of about 9ml/100g fused metal weight was recorded; this was obtained using the welding flux in the as-received condition. Such hydrogen levels can be considered typical of industrial practice.

Sampling sequence and procedure

For each weld base block, two hydrogen sampling holes were drilled, but only the flux through one hole could be analysed at a time. The first bead edge would fall over the sampling hole such that the adjacent bead would fully fuse this area as a consequence of overlap during deposition. Figure 2 shows the analytical sequence used. Sampling began as soon as welding started, but there was no apparent hydrogen flux until the welding head was almost directly above the sampling hole. Sampling continued for at least two minutes after the welding head had passed the sampling hole.

Results

Figures 3-6 show the hydrogen flux measurements for welds 1 and 2 in terms of hydrogen pressure versus time, with temperature against time plots superimposed. In some cases, thermal cycles were not recorded as a result of thermocouple failure. The relevant measurement hole and run concerned are indicated in each case.

Table 2 shows the hydrogen results in terms of peak hydrogen partial pressure reached during each measurement and the corresponding HAZ temperature. A figure proportional to hydrogen volume was also estimated in each case from the area under the partial pressure-time curve, and is given in Table 2.

Table 2 Peak hydrogen partial pressures and estimated hydrogen volume for successive weld runs. (NM = not measured)

The positions of the holes with respect to the first and second layer fusion boundaries and the overlap regions were generally close to that required. A section through the hydrogen analysis hole from the first test weld was taken ( Fig.7) and the hole can be seen to fall below the first layer bead overlap and the centre of the second layer bead. Such sections were also used for metallographic examination, which showed refinement in the HAZ compared with that seen for a first layer deposit alone. It appears that the second layer had re-austenitised the first layer coarse grained HAZ, an observation confirmed by the measurement of temperatures up to 800°C in the hole region during second layer deposition. The A _l temperature for this steel can be taken as about 800°C. ^[5]

The results show that the general features of the hydrogen evolution curves are similar for both welds 1 and 2. During first layer deposition, hydrogen partial pressure rises rapidly as the bead passes over the measurement hole, then declines steadily. Run 2 from test weld l gave the greatest hydrogen levels and also the highest HAZ temperatures recorded (~1000°C). This observation can be explained by the postulation that the higher the temperature in the hole region, the greater the diffusion rate (above A ₃ ) and the more hydrogen released into the sampling hole for a given HAZ hydrogen concentration.

Deposition of second layer runs ( Fig.3-6) produced further hydrogen evolution into the measurement holes. Results from welds 1 and 2 are similar indicating that an interlayer hold and a preheat of 150°C have little influence on HAZ hydrogen flux.

Discussion

The technique adopted was clearly successful in enabling hydrogen movement into the HAZ during welding to be monitored. The method relies on a pressure change to indicate the hydrogen flux, the gas being exhausted to the atmosphere, hence, no quantitative measurement of the amount of hydrogen driven through the HAZ was made.

This makes the system, as described here, unable to compare the quantity of hydrogen driven through the HAZ for different welding processes or different consumable drying procedures. However, the principle is established, and it should certainly be feasible to obtain such information with little modification to the mass spectrometry equipment.

In the context of the cladding programme undertaken, two aspects of behaviour were clarified: a) that the application of preheat for the second layer had little effect on HAZ hydrogen flux, and b) more significantly, that appreciable hydrogen flow took place during deposition of the second layer. Noting that the method was not fully quantitative, the latter finding indicates that HAZ hydrogen cracking could occur from the second layer deposit: hence, it must be advised that preheat should be employed in practice for the second layer and maintained after welding to encourage hydrogen diffusion away from the HAZ before it cools to room temperature.

Summary

A technique used for measuring HAZ hydrogen flux during welding has been described and the effectiveness of this technique has been illustrated by presenting results obtained during a recent Welding Institute investigation into underclad hydrogen cracking.

Hydrogen fluxes into the HAZ have been studied by drilling holes into the expected location of the high temperature HAZ and monitoring hydrogen evolution during welding with a mass spectrometer. Studies were made for two layer strip submerged arc cladding, with the measurement holes located at run overlaps for the first layer. Second layers were deposited with and without preheating. Using this technique, the following conclusions can be drawn:

1. Hydrogen passes into the HAZ during deposition of the second layer of cladding, as well as during the first layer.
2. The results are consistent with the deposition of a second layer releasing quantities of hydrogen into the HAZ of a magnitude approaching those from the first layer.
3. Preheating for the second layer deposition did not influence the HAZ hydrogen flux.

Acknowledgements

The author is grateful to the group of companies who sponsored the programme for their kind permission in allowing the publication of the results. The companies who participated in this work were: Atomic Energy Corporation 0f South Africa Ltd; Electricite de France; Equipos Nucleares SA; Framatome; National Nuclear Corporation Ltd; Nuclear Installations Inspectorate (HSE); Nucleare Italiana Reattori Avanzati (NIRA); United Kingdom Atomic Energy Authority.

Thanks are due to Dr T G Davey, C R Dye and Dr J L Robinson who were involved with the project at an early stage and who carried out the initial tests.

References