The Pennsylvania State College The Graduate School Department of Mineral Technology Division of Metallurgy Density and Hydrogen Occlusion of Some Ferrous Metals A Thesis by James Howard Keeler Submitted in partial fulfillm ent of the requirements for the degree of Doctor of Philosophy June 1951 Approved* 2.. 1 / Division of Metallurgy CMtef, Division of Metallurgy TABLE OF CONTENTS Page INTRODUCTION ' 1 EXPERIMENT 16 Materials 16 Density 18 Hydrogen Occlusion 25 RESULTS AND DISCUSSION 36 Density 36 Occlusive Capacity 59 SUMMARY AND CONCLUSIONS 68 REFERENCES 70 I ABSTRACT Whether a metal exhibits a decrease in density with cold re­ duction was found to depend upon impurities in the metal. A high- purity iron did not exhibit any decrease in density after cold re­ duction. In cold-reduced ingot-iron, only a 43 per cent recovery of density was found during annealing up to 1200 C, whereas SAE 1020 exhibited almost complete density recovery during recrystalli­ zation at 600 C. High-purity iron and ingot iron in the hot-rolled, the cold-reduced, or the annealed condition exhibited no appreciable differences in hydrogen occlusion between 250 and 550 C. Hot-rolled and annealed SAE 1020 displayed an occlusive capacity similar to that of the high-purity iron and ingot iron, whereas the cold-reduced SAE 1020 exhibited a remarkably large hydrogen ooolusion, especially at 250 C and 300 C. A theory based upon two varieties of voids in the metal formed during cold-reduction is offered to relate the various observations. ACKNOWLEDGEMENT It is a pleasure to acknowledge the aid and encouragement of my many friends in this research. To Dr. H. M. Davis, for his unlimited patience, and for his guidance as a friend and counselor, the author •wishes to express his deepest gratitude. The helpful suggestions, critical advice, and sympathetic assistance of Dr. R. W. Lindsay and Dr. J. W. Fredrickson cannot be overlooked. The author is indebted to the United States Steel Company for the commercial sheet materials of this investigation and for all analyses. Stimulating discussions with research personnel of the Company and with Dr. L. S. Darken of the Research Laboratoiy of the United States Steel Corporation lent impetus to much of this inves­ tigation. Special thanks are due to Mr. Maynard L. Hill for his many acts of assistance, both technical and personal, and to Professor W. J. Reagan and Professor L. F. Haller who helped obtain materials used in the thesis research. There are, of course, the unrecoided courtesies and suggestions of other members of the Division of Metallurgy, and of fellow students who have aided in the accomplishment of this work. 4, rapid until 65 per cent reduction was reached, after which the rate of decrease diminished, and the total density decrease approached a lim it of about 0*27 per cent* The m aterial, containing more of carbon and other elements, showed a rapid in itial decrease in density with cold working* Iftien 50 per cent cold reduction was reached, the lim iting change of density, about 0*26 per cent, had been obtained* It was thus observed that, for small reductions, D]_ showed the greater density loss, but both Ai and D]_ had approximately the same maximum ohange, about 0*26-0*27 per cent* The effects of cold reduction and subsequent annealing upon the recovery of density were studied for two cold-drawn specimens, (60.7 per cent reduction) and (62*5 per cent reduction)* "They (the specimens) were heated in vacuo to successive temperatures, held for 1/2 h r., and subsequently cooled*" (It is assumed by this statement that the same specimen was used in annealing at various temperatures, each succeeding temperature being higher than the pre­ vious temperature* It was also noted that Di had been normalized prior to cold-drawing whereas A^ had been annealed prior to cold re­ duction* ) Specimen Di, reduced 60*7 per cent, did not show any recovery of density below 400 C* However, between 400 C and 700 C, the den­ sity returned almost to that of the undrawn specimen* The 62*5 per cent cold-reduoed A^ also showed no recovery when annealed below 400 C. Unlike Di, there was only a slow recovery of density by Ai 4 5. as the temperature -was raised above 500 C. Even after 1/2 hour at 1000 C, the absolute recovery was only 0,09 per cent, which was about 40 per cent of the original change. Although the lack of a complete recovery of density during annealing in vacuo at 1000 C, well above the recrystallization temp­ erature of the Ax m aterial, was perhaps surprising, similar results have been reported by Ishigaki^^O for cold-hammered Armco iron. Ishigaki showed graphically that the Armco iron decreased 0.09 per cent in density, and recovered only about 70 per cent of the density change with annealing at 1000 C (time of annealing and whether in atmosphere or in vacuo were not given). Andrew made several experiments with another ingot iron (A3: 0.07 C, 0.01 Si, 0.04 Mn, 0.009 P, and 0.054 S) in order to determine if the decrease in density with cold work was, in some way, dependent upon the inclusions present in the iron. As stated by Andrew, "If cracks or flaws, submicroscopic or visible, were produoed in the inclusions on cold drawing, they might contribute to a lower­ ing of the bulk density of the specimen. On this assumption the den­ sity of a cold-worked specimen cannot be fully recovered until the cracks or flaws in the inclusions are completely healed or eliminated." On the basis of a lower original (annealed) density, and, apparently, a metallographic examination, the second ingot iron was considered to have more inclusions than did Ax* It was found that this iron had a greater density decrease with cdd reduction and showed a recovery 4 6 of density with subsequent annealing comparable to that obtained with the first ingot iron (Ai), On the basis of these tests Andrew thought that, with ingot-iron specimens, the decrease in density accompanying cold reduction varies with the composition of the iron and probably with the inclusion content* These deductions by Andrew, although in agreement with the concept that inclusions may vary the density loss on cold working, are perhaps controversial* The original (annealed) density may not have been the maximum density for the materials involved* Although the prior history of these materials was not given, it should be pointed out that the density of a commercially hot-rolled m aterial, particularly a low-carbon, low-alloy iron, is variable and may not be the maximum for the particular composition. Thus Andrew's observed mflTri mum change in density may have been less than the possible change of density by cold work. A second qualifying circumstance which might modify Andrew's deduction was the difference between the in­ clusion contents of the test materials* Of the ingot-iron specimens A^ and A3, the latter was reported as having the greater inclusion content* The chemical analyses might well support this observation* However, in such low-carbon irons, the oxygen content is usually con­ siderable, and the lack of oxygen analyses leaves some doubt about the inclusion contents* Metallographic estimations of inclusion con­ tent are aooeptable for wide variations in im purities, but are subject to recognized lim itations of particle size and distribution, both on i a macroscopic scale and on a microscopic scale. It would also seem questionable to attribute great significance to the difference in in­ clusion content between the two ingot irons whereas the same differ­ ence in density loss is obtained between A3 and the medium-carbon steel Di* Greater details of composition for the Di steel would have been welcome. It is not the contention of the writer that the deductions made by Andrew are incorrect. Rather, i t is to be recog­ nized that seme reservations are in order. In agreement with the work of Andrew on the influence of in­ clusion content in ingot iron is the study of density of extremely high-purity copper by Smart, Smith, and Phillips^3*0, Here it was found that the high-purity copper disclosed only a very slight de­ crease in density with cold drawing. However, the same high-purity copper with 0,037 per cent oxygen added showed considerable density decrease with cold drawing. This change was greater than the effect due to formation of oxide, and the authors commented, possibly by giving rise to the formation of minute internal voids," Zener^3)# on the other hand, has shown quantitative calculations indicating that grain distortion by itse lf can produce expansion of the order of mag­ nitude of the experimental observations, Maier^*^ showed that both the density of iron and that of copper passed through a minimum with increasing cold reduction. After this minimum (88 per cent reduction for iron) has been reached, the density begins to increase with fur­ ther cold reduction. No other investigator has mentioned this be- 8. havior. Finally, the different decreases in density reported by Andrew^*), Ishigaki^^^, and Maier^^^ should be noted* The maxi­ mum loss in density, 1 per cent, was obtained in torsion by Uaier* / QJ \ (J* W* Landon'' showed a 2 per cent decrease in density of wrought iron with torsion*) Andrew's specimens were drawn and show a density decrease of as much as 0*27 per cent, while Ishigaki measured the density of cold-hammered iron and obtained a density decrease of 0*09 per cent* It thus appears that the mode of cold reduction, which determines the characteristic type of plastic flow occurring, can govern the magnitude of the density loss* A large amount of infoxmation is available on the subject of hydrogen in metals* D. P* SndLth^'^ lists a bibliography of 1467 references which takes into account the literature to about the end of 1946* Zapffe and Sim s^^ published a bibliography on "Hydrogen in Steel" of over 500 references* It has been known, as a result of the classical work of Sieverts(^), that atomic and not molecular gases are dissolved in metals* Like other metals that take up hydrogen endothermically (e.g., Ni, Co, Cu, Pt), iron dissolves but a small amount of hydrogen, and the alloys of iron and hydrogen display but one phase* i 9. One of the first considerations in a discussion of hydrogen in iron and steel is that of the quantity of hydrogen generally found* Usually it is less than 0*001 per cent by weight and is usually considered almost insignificant* However, it should be re ­ membered that this figure is on a weight basis, whereas on an atomic basis there is fifty-six times as much, or in comparison with the 0*001 weight per cent given above, 0*056 atomic per cent* This fig­ ure again assumes greater significance when it is realized that this amount in the steel, 0.001 per cent by weight, depending upon its distribution, may produce pressures at ordinary temperature to exceed the strength of steel* (Examples of this are the flakes and shatter- cracks in steel ingots*) In ,tendothermic,, iron there is considerable change in the re­ ported solubility of hydrogen with temperature, the change being exaggerated by phase changes and by the melting of iron* For example, if relative volumes (the volume that the hydrogen extracted from a given piece of metal would occupy at standard conditions compared to the volume of the metal from which it was extracted) are used for com­ parative purposes, the amount of hydrogen in iron at various temper­ atures can easily be shown^^^. F irst, in liquid iron there is gen­ erally found about 2*2 relative volumes of hydrogen* With solidi­ fication there is a loss of 1*2 relative volumes leaving 1*0 rela­ tive volume of hydrogen s till in the iron* This amount decreases to 0*4 just above the austenite-to-ferrite transformation temperature i