WELDING RESEARCH SUPPLEMENT TO THE WELDING JOURNAL. MAY, 1984 Sponsored by the American Welding Society and the Welding Research Council A Critical Evaluation of the Glycerin Test Constituents other than hydrogen are present in gas collected during testing, and variations in glassware shape and size dramatically affect gas volumes actually collected BY M. A. QUINTANA ABSTRACT. In recent years there has of hydrogen gas and hydrogen-contain­ content of the resulting deposit (Refs. 3, developed a rising interest in the mea­ ing compounds during welding which 4). However, there are limitations. Since surement of the diffusible hydrogen con­ produces atomic hydrogen in the arc the relationship between coating mois­ tent of weldments. As a result, the glyc­ atmosphere. Some common sources of ture and hydrogen content can vary erin test has been adopted as a standard hydrogen are moisture contained in SMA significantly with electrode type (Ref. 3), a test method for diffusible hydrogen mea­ (shielded metal-arc) electrode coatings, given coating moisture level can result in surement throughout the welding indus­ lubricants on solid and cored wires, and a wide range of weld metal hydrogen try. However, recent work has revealed moisture and other contaminants present contents depending upon the coating that test results are of questionable valid­ in shielding gases. system. It, therefore, becomes necessary ity. Analyses of the gas collected utilizing Because of the detrimental effects of to evaluate the actual hydrogen content this technique indicate that constituents hydrogen uptake during welding, consid­ of the deposit. When considering a pro­ other than hydrogen are present in signif­ erable effort has been devoted to the cess such as FCAW (flux cored arc weld­ icant proportions and that those propor­ control of the hydrogen potential of ing) or GMAW (gas metal arc welding), tions exhibit significant variation. In addi­ welding consumables and processes. This indirect tests have limited meaning (Ref. tion, variations in hardware shape and has been achieved, in part, through the 5), and the hydrogen content of the size can dramatically affect the gas vol­ careful selection of lubricants for drawing deposit must be measured directly. ume actually collected. operations and an increased emphasis on This paper discusses these problems in cleanliness and control of contaminants. Diffusible Hydrogen Measurement Tests detail and identifies the probable causes. Perhaps the greatest effort has been Among the items discussed are the solu­ expended for SMA electrodes where the Several tests have been developed for bility properties and purity of the glycerin control of weld metal hydrogen levels the direct measurement of weld metal bath, the effect of environmental factors has been achieved through minimization diffusible hydrogen content. The diffus­ on the bath, the effect of sample prepa­ of coating moisture levels (Refs. 1-3). ible hydrogen content is that portion of ration and handling on test results, and This approach is reasonable for cov­ the total hydrogen content which is the size and shape of glassware. ered electrodes, since the coating mois­ evolved rapidly from a sample at or near ture does contribute to the hydrogen room temperature. Two widely accepted methods for measurement are very simi­ Introduction lar in that they involve the use of a For years it has been recognized that Paper presented at the 64th Annual AWS collecting fluid (Refs. 6, 7). Welded spec­ hydrogen present as an interstitial impuri­ Convention in Philadelphia, Pennsylvania, dur­ imens are allowed to degas in some form ty has an adverse effect on the mechani­ ing April 25-29, 1983, under sponsorship of the of eudiometer tube filled with fluid. cal properties of steels. Frequently, High Strength Steel Subcommittee of the Evolved gases are collected in the upper hydrogen uptake, resulting in a degrada­ Welding Research Council. portion of the tube over the fluid and the volume is measured. Results are reported tion of properties, occurs during process­ M. A. QUINTANA is a Welding Metallurgist, as the volume of gas collected per unit ing. For welded structures, hydrogen Electric Boat Division, General Dynamics, Grot- uptake occurs as a result of dissociation on, Connecticut. mass of weld deposit. WELDING RESEARCH SUPPLEMENT 1141-s Individual test procedures differ in the similar results (Refs. 14, 15). 18, 19), which can be appreciable after size and preparation of welded samples Procedural variations are likely respon­ extended periods of time. and in the collecting fluid used; however, sible for some of the observed problems the principle of operation is the same. In with the glycerin test. Two procedural Investigation Objectives one case, mercury is used as the collect­ variables of concern are the heating of As a result of the problems described ing fluid (Ref. 6), and the specimen is base material prior to welding and the above, the results obtained using the allowed to degas at room temperature. In time elapsed between completion of glycerin test are not reproducible the other case, glycerin is used as the welding and quenching. Prior heating of between laboratories (Ref. 20) and are of collecting fluid (Refs. 7, 8). This test the base material is thought to liberate questionable validity. The fact that such involves placing a welded specimen in a any hydrogen present in solution which results are used as acceptance criteria for glycerin filled eudiometer tube immersed might contribute to the weld metal diffus­ welding consumables is of tremendous ible hydrogen content (Ref. 9). The in a glycerin bath maintained at 45 ± 3°C concern. Because of this, a thorough expected result is a reduction in the (113 ± 5.4°F). As the specimen degases, evaluation of the glycerin test is in order. amount of hydrogen evolved by the bubbles pass up through the glycerin and It is the intent of the investigation specimen. collect in the upper, calibrated portion of described in this paper to characterize the eudiometer. The volume of gas col­ The time elapsed between completion some of the problems which are inherent lected is measured and normalized by the of welding and quenching is believed to in the glycerin test and to identify their mass of the weld deposit. be important from the standpoint of probable causes. diffusion (Refs. 9, 17). Since diffusion Glycerin Test Problems rates are higher at elevated tempera­ Experimental Due to the health hazards associated tures, appreciable amounts of hydrogen with the use of mercury, the glycerin test can be lost to the atmosphere if sufficient Specimens fabricated under identical has been widely used in the United time is available prior to quenching. conditions should, within reason, result in States. As a result, the glycerin test has Therefore, one would expect a greater the same amount of diffusible hydrogen become an accepted standard by which amount of hydrogen evolved from the per unit mass of weld deposit. With the welding consumables are evaluated (Ref. welded specimen while immersed in the hydrogen content maintained relatively 8). However, rather serious problems glycerin if the time elapsed to quenching constant for a series of tests, other with the glycerin test have been reported is kept to a minimum. conditions can be selectively varied in (Refs. 9-16). It has been demonstrated Other variables of concern are related order to determine what impact they that diffusible hydrogen levels deter­ to the condition of the glycerin bath itself. have on the volume and composition of mined by the glycerin method are consis­ Glycerin, upon exposure to the atmo­ the collected gas. To demonstrate this, tently lower (^50%) than comparable sphere, readily absorbs a variety of gases several tests were conducted. Initially, mercury test values (Refs. 9, 10, 13, 16). including water vapor, oxygen, and nitro­ using the flux cored arc welding (FCAW) The obvious implication is that the glycer­ gen. Even though a glycerin test assembly process, the effect of glassware size and in test does not recover all of the evolved is initially assembled with relatively pure shape was demonstrated. Subsequent hydrogen. In addition, glycerin test results glycerin, it does not necessarily remain in testing using the spray gas metal arc have been shown to exhibit far more that condition indefinitely. Further, it is welding (GMAW) process demonstrated scatter than comparable mercury test possible that contaminants absorbed by the effects of base material pretreatment, results (Ref. 9). the glycerin bath are responsible for con­ elapsed time to quench, and water con­ Recent investigations have also point­ tamination of the hydrogen collected tent of the glycerin bath. In addition, a ed out other problems associated with over glycerin during testing (Refs. 14, 15). series of tests was conducted simulta­ the glycerin test. Analysis of the gas To further complicate the situation, the neously with the GMAW tests to illustrate collected over glycerin conducted during solubility of hydrogen in glycerin (Refs. other factors which affect the composi­ this investigation indicates that gases oth­ 18, 19) can result in the loss of hydrogen tion of gas samples. er than hydrogen are present in the gas evolved from the test specimen to the The chemical compositions of base sample in significant proportions, making bath through dissolution. The extent to materials and filler materials used during the validity of the glycerin test question­ which dissolution occurs is affected by this investigation are presented in Table able. Other researchers have reported the water content of the glycerin (Refs. 1, The glycerin used was 99% pure, Table 1—Summary of Chemical Analyses of Base Materials and Filler Metals, Wt-% C S P Si Mn Cr Ni Cu Mo v Zr Specification Base material for FCAW .22 .013 .012 .12 .79 .04 .04 .03 .02 002 -C001 QQ-S-741 SIM. to and first of GMAW ASTM A36 tests Base material GMAW .21 .017 .004 .11 .80 .04 .06 .09 .01 01 <.001 QQ-S-741 SIM.