TS 320 .R3 Copy 2 DEPARTMENT OF COMMERCE Scientific Papers OF THE Bureau of Standards S. W. STRATTON, Director No. 397 A STUDY OF THE RELATION BETWEEN THE BRINELL HARDNESS AND THE GRAIN SIZE OF ANNEALED CARBON STEELS BY HENRY S. RAWDON, Physicist Bureau of Standards EMILIO JIMENO-GIL, Professor of Physical Chemistry University of Oviedo, Oviedo, Spain SEPTEMBER 20, 1920 I). J. McAdam, jr . True A. S. T. M . 17. Ft. II. p. 58; 1917. Also, 18, Pt. II, p. 6S; 1918. 7 A. Pomp. Emtluss der Wiirmbehandlung auf die Kerbzahigkeit. Korngrosse und Harte von kohlen stofTarmem Flusseisen. Ferrum, 13, p. 49; 1915-16. Rawdon ill Hardness and Grain Size of Steels TABLE 1. — Results of Chemical Analyses of Steels Usea a 559 Specimen Carbon Manganese Phosphorus Sulphur Silicon A Per cent 0.07 .19 .46 .70 1. 12 Per cent 0.27 .41 .36 .22 .23 Per cent 0.006 .004 .019 .023 .019 Per cent 0.054 .050 .047 .011 .013 Per cent B C D E o Much of the material was furnished through the courtesy of the Carnegie Steel Co. In order to develop grains varying widely in size, two different methods were adopted. In the first the specimens (two-inch lengths of bars, one by one-half inch cross section) were heated for six hours at various temperatures, as shown in Table 2. Two speci- mens of each composition were heated at each of the temperatures indicated ; one was cooled slowly in the furnace at the close of the six-hour period, the other allowed to cool more rapidly in the air. For each particular temperature the specimens were all heated together (likewise cooled together) in order to eliminate possible effects from variations in heating conditions. For heating, an electric-resistance furnace of the muffle type was used for the four lower temperatures, and the higher temperatures were obtained in a small gas-fired muffle furnace. The temperature measurements were made by means of a ehromel-alumel thermocouple and a small portable potentiometer. In order to avoid decarburization during the continued heating at the high temperatures (974, 1024, and 1112 C, Table 2), the specimens were packed in a mixture of amorphous silica and 5 per cent of powdered charcoal. This was found, by trial, to be very effective in retarding decarburization; only a relatively thin layer was affected. When the specimens were prepared for the microscopic examination and the determina- tion of hardness, they were ground deeply enough to remove entirely the decarburized layer. The second method for producing variations in grain size con- sisted in annealing bars after they had been given a preliminary cold working by stressing in tension. This is discussed in detail in Section VI. 560 Scientific Papers of the Bureau of Standards [va.16 TABLE 2. — Annealing Temperatures and Methods of Cooling Specimens Designation of specimens Temperature for 6 hours Metood Remarks of cooling Degrees centigrade A 1, B-1, C-l, D-l, E-l 674(672-676) In air... Below Ai transformation; original grain size preserved 674 (672-676) Do nace A 2, B-2, C-2, D-2, E-2 762(752-772) Just above Ai transformation; origi- nal grain size replaced by a finer one 762 (752-772) Do nace 849 (840 858) A 33, B-33, C 33, D-33, E-33... 849 (840-858) nace A 4, B-4, C-4, D-4, E-4 953 (946-960) In air.. . nace 2-A 4, 2-B-4, 2-C-4, 2-D-4, 974 (952-995) In air 2-E-4 2-A 44, 2-B-44, 2-C-44, 2-D-44, 974 (952-995) Above A3 transformation 2-E-44 nace 1 A s, 1 B 5, l-C-5, l-D-5, 1024 (1010-1038) l-E-5 1H24 (1010-1038) l-E-55 nace A-5, B-5, C-5, D-5, E-5 1112 1 1098-11261 A-55, B-55, C-55, D-55, E-55... 1112 (1098-1126) nace A-66, B-66, C-66, D-66, E-66. . . 754° for 35 minutes; 674 Quenched Grain refinement by "double an- ib72-67b for 6 hours in water, reheated and cooled in fur- nealing" nace III. MICROSCOPIC EXAMINATION The specimens were examined microscopically and a grain size determination was attempted. The method used was the plani- metric one, modified and described by Jeffries 8 and adopted by the American Society for Testing Materials in its tentative speci- fications for the preparation of micrographs. 9 In Table 3 are summarized the results of the grain count. Only for the steels of the low-carbon content (particularly those cooled in air) and heated at the highest temperatures could a satisfactory grain-size determination be made with certaintv. 'Z.Jeffries, Am. Insl Mm, Engrs., 54, p. 594; 1916, Also, Met, and Chem. Eng., 18, p, 185; 1918. 1 Tentative Spec. E z-19 T, I'm, A S T -M . 19, I'l I. p. 770; 1919. jtmm^Giil Hardness and Grain Size of Steels TABLE 3. —Results of Grain-Size Determination 56l Average Specimen, air cooled ""£"*" heating A-l... A-2... A-3... A-4... 2-A-4 l-A-5. A-5... B-l... B-2 ... B-3... B-4... 2-B-4. l-B-5 B-5... C-l... C-2 ... C-3... C-4... 2-C-4. l-C-5. C-5... D-l.. D-2.. D-3.. D-4. 2-D-4 l-D-5 D-5.. E-l.. E-2.. E-3.. E-4.. 2-E-4 l-E-5 E-5.. °C 674 762 849 953 974 1024 1112 674 762 849 953 974 1024 1112 674 762 849 953 974 1024 1112 674 762 819 953 974 1024 1112 674 762 849 953 974 1024 1112 Number of grains per mm 2 1690 2330 2010 1330 580 380 CO 2620 en 1520 I'M CM (") CO CO CO ("> 410 22 20 12. Area of grain in n 2 590 430 500 750 1930 2610 CO 380 CO 660 CO CO CO CO (-0 CO CO 2410 45 250 49 880 77 820 CO (») CO (") CO CM CO CM ('0 CO 21 47 620 12.5 80 000 CO CO CO CO 1'.) CO CO CM CM CO CM I'M CO CO Specimen, furnace cooled A-ll... A-22... A-33... A^t4... 2-A-44 l-A-55 A-55... B-ll... B-22... B-33... B-44... 2-B-44. l-B-55. B-55... C-ll... C-22... C-33... C-44... 2-C-44. l-C-55 C-55... D-ll.. D-22-. D-33.. D-44.. 2-D-44 l-D-55 D-55.. E-ll.. E-22-. E-33.. E-44.. 2-E-44 l-E-55 E-55.. Average tempera- ture of heating Number of grains per mm 2 °C 674 1800 762 CO 849 930 953 (■>) 974 92 1024 66 1112 32 674 2857 762 CO 849 360 953 290 974 220 1024 168 1112 144 674 CO 762 CO 849 CO 953 CO 974 19 1024 CO 1112 7.5 674 CM 762 CM 849 CO 953 CO 974 CO 1024 23 1112 11 674 CO 762 (») 849 co 953 116 974 49 1024 34 1112 20 Area of grain inM 2 550 CO 1070 CO 10 920 15 200 31 640 350 CO 2810 3410 4490 5950 6940 CO («) («) CO 52 630 '") 104 850 (") CO CO CO CO 42 o*> 87 oin CO CO CO 8660 20 410 29 411 49 140 a No satisfactory count could be made. On account of this uncertainty, therefore, there are given in Figs. 1 to 11, inclusive, typical micrographs of the specimens to show the structural condition which resulted from the treatment which the material received. They also illustrate the difficulties encountered in making an accurate grain-size determination, as well as the structural features other than grain size which affect the properties of the material. In Figs. 12 and 13 are given typical micrographs of the steels heated at the highest tempera- 562 Scientific Papers of the Bureau of Standards ivoi.z6 lure, highly enough magnified to demonstrate the effect of the rate of cooling upon the structure of the steel; that is, the inner structure of the grains. The specimens of series 1 (A-i, .... E-i and A-n .... E— i 1 ), which were heated to a temperature slightly below the Ac t transformation retain the original grain size of the material. The specimens heated just above this transformation show a decrease in grain size (in those specimens in which the grains are clearly defined). This is best seen in the materials which were cooled in the air. In all of the five types of steels the increase in grain size is very slight for the lower temperatures. Not until the Ac 3 transformation has occurred and the steel is in the gamma phase, is the increase in the grain size very marked. In some of the steels this occurs upon heating when the temperature of the Ac 3 transformation is reached. In two, however (0.46 and 0.70 per cent carbon) , no appreciable increase in grain size occurs until the temperature is considerably above that of the Ac 3 transformation (approximately 200 ). The marked increase for these specimens occurs also within a relatively very narrow range of temperature, 950 to 975 C, approximately. In Fig. 14 there is given a portion of the iron-carbon constitu- tional diagram, in which has been indicated, for purposes of refer- ence, the range of temperature in which the grain size increases most rapidly in each of the steels used. IV. HARDNESS DETERMINATIONS Two methods were used for determining the mechanical hard- ness of the material. The specimen used for the determination of the microstructure was large enough (1 by r, 1 : inches face) to permit all the determinations of hardness to be made upon the one piece. The two forms of the Brinell hardness-testing set used will be referred to as the " standard " and the " micro-Brinell " set, respectively. For the former the usual hydraulic tvpe was used, and a load of 500 kg was applied to the specimen by means of a ball 10 mm in diameter for 30 seconds. For the second set of determinations a small "dead-weight" type of Brinell appa- ratus was used. The instrument (Fig. 15), which was developed by the Ordnance Department of the United States Army, was loaned for the purpose. A load of 15 kg for 30 seconds upon a ball one-sixteenth inch in diameter was used. The details of construction and manipulation have been described elsewhere. 10 111 S. I. Goodale and R M. Banks, Development of Brinell Harness Tests ..n Thin Sheet Brass. Proc. A. S. T M , 19, l'l II. p. 757; 1919. Rawdon "1 Jvneno-G'l\ Hardness and Grain Size of Steels 563 . -*"■ ','■■ .-- .■.-„>-. ',V-\ *r .-' ,' 34-9° ; j ' ■ V* 1 * 1 *V-XV-..~ . '« j* 9SJ' if '.-•' . . /■■-■' S74" 1 •>-: V ''■"■ , f ^ \ v ' '' f » ■ * ^ /OZ4° y Fig. 1. — Micros tructure of o.oy per cent carbon steel (steel A) after annealing and cooling in air; X/j (reduced from Y.ioo) Each specimen was heated six hours at temperature (°CJ shown. Etched with 2 per cent alcoholic nitric acid 1988°— 20 2 564 Scientific Papers of the Bureau of Standards [Voi.16 • ] :" ■■■ - i -' J ' . ' ■" •' ' > , 674.° * * * f ■ >; £S3° . Ik 7<£2° 974-° ■ * ' F ■- '" **0? '< ;'*> , - .-1 V % ' - ■•,-Jr /0Z4-° ■a * Y w //2tf° Fig. 2. — Microstructure of coy pet cent carbon steel {steel A) after annealing and cooling in the furnace; >'~j {reduced from ■ 100) Each specimen was heated six hours at temperature (° C) shown. Etched with 2 per cent alcoholic nitric acid Rawdon J imc no-Gil] Hardness and Grain Size of Steels 565 76Z° 843" 97^-" )ML i ' i'-J^ :«^r «ki^ V;^ T \V> ^HP ^- : ///-a* Fig. 3. — Microstructure of 0.19 per cent carbon steel (steel B) after annealing and cooling in air; X/\J (reduced from Xioo) Each specimen was heated six hours at temperature (° C) shown. Etched with 2 per cent alcoholic nitric acid 5 66 Scientific Papers of the Bureau of Standards [Voi.16 W^ Fig. ^.—Microstructureofo.igpei cent carbon steel (steel B) after annealing and cooling infurnace; ■ 75 (reduced from Xioo) Each specimen was heated mx hours .it temperature (° C) shown. Etched with 2 per cent alcoholic nitric acid Rau-dott 1 Jimeno-Gil] Hardness and Grain Size of Steels 567 Fig. 5. — Microstructure of 0.46 per cent carbon steel(steel C) after annealing and cooling in air; - 75 {reducedfrom ■ too) Each specimen was heated six hours at temperature (°C) shown. Etched with 2 per cent alcoholic nitric acid 5 68 Scientific Papers of the Bureau of Standards [Voi.i6 Fig. 6. — Microstructure 0/0.46 per cent carbon steel (steel C) after annealing and cooling in furnace; X?5 (reduced from y.100) Each specimen was heated six hours at temperature (° C) shown. Etched with 2 per cent alcoholic nitric acid Rawdon Jimeno-Gil\ Hardness and Grain Size of Steels 569 p IG - —MurosttHChire of 0.70 per cent carbon steel (steel D) after annealing and cooling in air; >'}y (reduced ] 'ram Xioo) Each specimen was heated six hours at temperature (° C) shown. Etched with 2 per cent alcoholic nitric acid 57° Scientific Papers of the Bureau of Standards [Vol. 16 Fig. S. — Microstracture ofo ?o per cent carbon steel {steel D) after annealing and cooling in furnace; ■ f 5 (reduced from ■ too) Each specimen was heated six hours at temperature (" C) shown. Etched with 2 per cent alcoholic nitric acid Rawdon Jimeno-Gil] Hardness and Grain Size of Steels 57 1 Fig. q. — Microstruciure of 1 .12 per cent carbon steel (steel E) after annealing and cooling in air; ■ 75 (reduced from Xioo) Each specimen was heated six hours at temperature (°C) shown. Etched with 2 per cent alcoholic nitric acid 1988°— 20 3 57 2 Scientific Papers 0} the Bureau of Standards iva.16 Fig. 10. — Microslructure of 1.12 pet cent carbon steel (steel E) after annealing and cooling in furnace; X.75 {reduced from X.IOO) Each specimen was heated six hours at temperature I 'C) shown. Etched with 2 percent alcohulic nitric acid. Rawdon Jimeno-Gili Hardness and Grain Size of Steels 573 r 5 -; SPf* ^•'./■i.*.^ ;stA.^-: ; 3_ ft / v* • ■ H 1%' KB c ■> yam - i Fig. ii.— Microstructure of the five types of steel used, after maximum grain refinement. Magnification of micrographt at left, X75 (reduced from Xioo); at right X375 (reduced from X5 00 ) A specimen of each of the steels was quenched in water after being heated 35 minutes at 754° C, it was then heated for 6 hours at 674° C. Etched with 2 per cent alcoholic nitric acid 574 Scientific Papers of the Bureau oj Standards [Voi.16 The Brinell hardness number was calculated by means of the ordinary formula: N Pressure P area of spherical indentation irtD in which t=}4 (D - -\JD 2 — d 2 ), D and d being the diameter of the ball and of its indentation, respectively. Although the loads Fig. i: — Effect of rate of cooling upon the structure of steels B, C, and h. ; ■ yj 5 {reduced from Y500) These were cooled in air after a period of heating of six hours at 1 1 u°C. Etched with 2 per cent alcoholic nitric acid. (Compare Fig. n) used for the two Brinell sets were not in exactly the same ratio as the sizes of the two balls, the hardness numbers obtained by the two are in very fair agreement. The results of the determinations of hardness are summarized in Tables 4 and 5. The comparative hardness of the various jimfZcu] Hardness and Grain Size of Steels 575 steels in different conditions of structure are best shown, however, in the graphical results of Figs. 16 to 19, inclusive. In plotting the diagrams the temperatures at which the various steels were -■:,-'.■ ■;•■*- Fig. 13. — Effect of rate of cooling upon the structure of steels B, I ', and E; X375 [reduced from ■ These were cooled in the furnace after a period of heating of six hours at 1112° C. Etched with 2 per cent alcoholic nitric acid. (Compare Fig- 1 2- ) Note the character of the pcarUte and the width of the ferrite envelopes) treated in order to develop the different grain size, rather than the numerical measures of the grain-size determinations, were used as abscissas. 576 Scientific Papers of the Bureau of Standards [Vol. $N V ® ® ®® »\. • IS <• ® ® • • I ® ® ® ® ® ® • ®® ® • ^ • X • « • • • l ^ § «0 n3 occurred in a period of six hours has been Indicated Fig. 15. — Apparatus for obtaining micro-Brinell hardness numbt Rawdon Jimeno-Gilj Hardness and Grain Size of Steels 577 TABLE 4. — Micro-Brinell Hardness Numbers of Air-Cooled and Furnace-Cooled Specimens a Specimen Hardness number Specimen furnace cooled Hardness number air cooled Maximum Minimum Average '< Maximum Minimum Average & A-l. . 83 76 79 78 85 82 86 72 82 88 103 105 102 103 87 90 153 162 173 177 185 164 177 169 194 199 201 213 193 200 151 212 225 259 270 236 260 A-ll 76 75 83 73 83 70 70 91 87 87 85 93 79 80 150 139 145 146 164 135 148 167 164 175 165 170 163 170 140 165 182 194 191 191 195 75 95 150 164 156 69 71 77 70 76 61 66 88 84 83 83 81 72 75 148 132 140 142 157 130 138 160 156 166 161 148 152 163 135 162 179 186 185 ISO 191 74 91 144 159 153 A 2 81 88 83 88 74 84 91 107 107 106 105 89 93 157 164 174 180 191 167 185 171 198 202 204 219 202 208 155 218 228 267 279 240 264 77 82 81 SI 71 81 86 99 103 95 101 85 88 144 161 170 172 180 161 165 166 186 194 197 208 191 195 148 208 220 256 262 231 257 A-22 73 A 3 A-33 81 i-4 A-44 2 A-4 2 A 44 78 1 A 5 l-A-55 65 AS A-55 68 B 1 B-ll 89 E 2 B-22 85 B 3 B-33 85 B-4 B-44 84 2-B-44 87 IBS 1 B 55 75 B-5 B 55 77 C-l C 11 152 C 2 136 C 3 C-33 143 C-4 C-44 144 2-C-44 161 1 C 5 l-C-55 132 C 5 C-55 146 D 1 D 11 164 D 2 D 22 100 D 3 D 33 171 D 4 D-44 164 2 D 4 2-D^t4 159 1-D 5 l-D-55 155 D 5 D-55 165 El E-ll 137 E-2 E-22 164 E-3 E-33 180 E^» E-44 190 2 E-4 2-E-44 189 1 E 5 l-E-55 184 E 5 E-55 193 A-66 74 B-66 93 C-66 147 D-66 162 E-66 155 15 kg load applied to one-sixteenth-inch ball for 30 seconds. b Average of 6 determinations. 578 Scientific Papers of the. Bureau of Standards TABLE 5.— Standard Brinell Hardness Numbers" [Vol. 16 Specimen, air cooled Hard- ness number'' Specimen, furnace cooled Hard- ness number '' Specimen, air cooled Hard- ness number'' Specimen, furnace cooled Hard- ness number'' A-l 75 78 81 81 81 72 80 89 95 95 96 98 83 93 124 149 160 161 170 159 176 A-ll - 73 72 7! 69 65 62 64 85 82 76 76 74 73 76 140 122 129 129 143 120 130 D-l 149 182 185 188 198 184 194 130 198 212 237 258 227 232 D 11 A-2 A-22 D 2 D-22 A-3 A-33 D-3 A-4 A-44 D-4... 2 A 4 2-A-44 l-A-55 A-55 2-D-44 l-D-55 D-55 153 139 l-A-5 A-5 l-D-5 D-5 B-l B-ll E 1 E 11 B-2 B-22 E 2 .... B-3 B-33 B-44 E 3 E-33 167 B-4 E-4 2-B-4 2-B-44 l-B-55 B-55 i 2 E 4 2-E-44 l-E-55 E-55 l-B-5 1 E 5 B-5 E-5 C-l C-ll C-22 C-2 B 66 C-3 C 33... . C-4 C-44 2-C-44 l-C-55 C-55 2-C-4 E 66 . l-C-5 C-5 " vo kg load applied to 1,-. mm ball for 30 seconds. ■'' Averagi oi . readings, V. FACTORS AFFECTING HARDNESS From a comparison of the results of the determinations of hard- ness as summarized in Figs. 16 to 19 with the micrographs showing the structure of the materials resulting from the different treat- ments, it is evident that no simple direct relation between grain size and Brinell hardness exists for carbon steels as does for some alloys; for instance, alpha brass. Grain size is a matter of minor importance with respect to hardness, as compared with some of the other factors involved. Rawdoil "1 Jimeno-Gilj Hardness and Grain Size of Steels 579 700 600 300 /OOO //OO'C Fig. i6. — Micro-Brinell hardness of the five steels aftet grain growth Each specimen was heated six hours at the temperature indicated and cooled in the air 700 300 SOO 1000 I / OO'C Fig. 17. — Micro-Brinell hardness of the jive steels aftei grain growth Kach specimen was heated six hours at the temperature indicated and couled in the furnace 5 So Scientific Papers of the Bureau of Standards [\\i.i6 s H a) A vfl 6-, W a •a u aj U) w a ^H w 6 pq u *« 2 < i ,«8 *S d w 0" V CO Stanc ard Brine hard ness rg rg Micro - Brinell hard- ness eg fM nj ivi rsi ON f ■ Mil, B 1-. 03 ft) O) 00 VO £> Oi co *o *- r* t- ■* O t> (SgSS g-O * < ffer- nce r>a _ _- in m io Q ™ Q" Oi LO Stand ard Brinel hard- ness O K o> Micro- Brinell hard- ness - 1 1-1 4> 1 « I M 1-. a in ra ft> ft) ^ r- _| o in - \o o m 00 00 o m to o oi Pe cent diff enc > < i m in m S d Q w 0) Stand ard Brinel hard- ness cro inel rd- Sn-« c is ^ T m eg on f^ Per- centag differ ence a -L. co a >D "> ■* 4 m ■5 > —. CO o ID m ^ £ " + """' + » S fl oi q° Stand ard Brinel hard- ness CT\ £ o> J. to ■3-9 S S ■"" ■""' 1-1 1-1 , g^a CO o o fl a\ o in r- lO •*■ — ' *• < en Ti ro fM eg rg + ^. en in r— ^ £> m i£> CO «n o lO 00 C^ 00 Ol (M CI rg + •S (M m + 00 rg \D tn O <\1 m ^ «M ^. N T <-•: t- r- CO in ID r- O rg «*• co '-' o '- , m ro 00 < ^f co On 00 m rg ^ " in ^ cn „. o\ c- tj- ^ C7i o oo eg Oi en > < rg »H -■ M p) ^ r, + r1 „ m M ^. fV] _ „ Rawdon "I Jimeno-Gil I Hardness and Grain Size of Steels 581 7 00 800 900 W //OO'C Fig. 18. — "Standard" Brinell hardness of tin- Jive steels a/I, 1 grain growth Bach specimen was heated six hours at the temperature indicated and cooled in the air 700 300 900 1000 //00°C Fig. 19. — "Standard" Brinell hardness 0/ the Jive steels a/ter grain growth Each specimen was heated six hours at the temperature indicated and cooled in the furnace 582 Scientific Papers of the Bureau of Standards [Voi.16 The purpose of determining the Brinell hardness by the two methods previously described was to show, if possible, the hard- ness of the grains as individuals or in small aggregates as com- pared with the average hardness of the material when measured in the usual manner. The results obtained by Portevin " on coarsely grained cast metals and alloys indicates that the hard- ness of individual crystals is not a constant value even for the same specimen, but it depends upon the relation of the surface bearing the Brinell impression to the internal orientation of the crystal. For convenience in reference the results of the determi- nations of hardness by the two methods are summarized in Table 6. With but very few exceptions (8 per cent of the total number of determinations), the micro-Brinell hardness is somewhat higher than that obtained in the usual manner. This is in agreement with the results obtained by Goodale and Banks 12 upon cartridge brass, although the results here obtained by the two methods are in much closer agreement than those referred to. The fact that the hardness as obtained by the use of the small testing set is somewhat higher than that determined by the usual type of appa- ratus upon the same material is not necessarily to be interpreted as 1 icing due to a characteristic behavior of single grains (or small aggregates of grains) as compared with the combined effect of a large number of grains. It is, at least in part, to be attributed to a difference in behavior of the two types of apparatus, charac- teristic of the two methods. The average difference between the "micro-" and the "standard" Brinell hardness number is some- what greater for the steels of higher carbon content than for those low in carbon, particularly in the air-cooled specimens. It is impossible, however, to trace any clear and definite relationship between the differences existing between the two sets of hardness numbers and the grain size of the corresponding materials. In Fig. 20 are given micrographs to illustrate the fact that no appreciable and systematic difference in hardness measured by the micro-Brinell testing apparatus as related to the number and size of grains covered In- the impression could be detected. Im- pressions lying entirely within the limits of a single grain were found to lie essentially of the same size as those taken on the same specimen, but overlapping several crystals. The fact that the pearlite grain is not uniform throughout but consists of clearly 11 A. Portevin, Hardness Tests on individual Crystals, Rev de -Mel . 12, p. 94; 1915. Also, The me- ( hanii ;il Anisotropy of Coarse-Grained Metals and Alloys ami the Brinell Test, C. R.. 160, p. 344; 191 5. '- Proi A. S T, M . 1!), Pt. II, p. 757; 1919. Rawdon Jimeno-GUl Hardness and Grain Size of Steels 583 defined divisions which may vary very considerably among them- selves in size and in orientation is probably largely responsible for this. The micrographs of Fig. 20 also illustrate well the fact that FlG. 20. — Relation of the impression of the ball in the micro-Brinell test to Ike grain of the metal of steel C, heated six hours at 1112° C and air-cooled; X/5 (reduced from \ioo) (a) — At junction of [our grains; (b) — At junction of two crams; (c) — Within one grain, near boundary; (d) — Entirely within oue grain. All impressions indicate the same hardness number. Etched with 2 per cent alcoholic nitric acid solution the pearlite grains, particularly when large, must each be con- sidered as an aggregate rather than as a single unit in their properties. The influence of the very pronounced increase in grain size which occurs in the steel when in the gamma condition upon the 584 Scientific Papers of the Bureau of Standards ivoi.16 hardness is indicated in all of the diagrams of Figs. 16 to 19, in- clusive. A sharp drop in the hardness curve corresponds to this rapid increase of grain size. The curves also illustrate other features of interest concerning the hardness. The five steels naturally divide themselves into two classes. The two of the lower carbon content (0.07 and 0.19 per cent) are quite similar in their behavior; when in the annealed state (that is, cooled in the furnace) the tendency is for the hardness to remain rather uniform, aside from the drop which occurs as a result of the in- crease in grain size. The effect of cooling the material more rapidly — that is, in air — is to accentuate this drop and to raise the hardness generally throughout. It is of particular interest that the maximum temperature at which the specimen is heated, provided this does not exceed that at which the pronounced in- crease in grain size occurs, is quite negligible as to the degree of hardness produced upon air cooling. The three steels of higher carbon content (0.46, 0.70, and 1.12 per cent) form a group with some striking characteristics in common. The general tendency is for all of these materials to increase in hardness upon anneal- ing; cooling in air accentuates this tendency to a very marked degree. The drop in hardness corresponding to the pronounced increase in grain size which occurs is noticeable in all the speci- mens; in spite of this drop, however, the general tendency for an increase in hardness with increasing temperature of annealing is still marked. The maximum temperature to which the speci- men is heated before cooling either in the furnace or in the air is a factor of importance in determining the hardness. This is due, without doubt, to the larger percentage of carbon in the specimens comprising this group as compared with those of the former one. The rate at which the steel is cooled after the long period of heating is a factor of prime importance in determining its hardness. Figs. 12 and 13 show the structural condition of the hardening constituent (pearlite or sorbite) due to the rate at which the material is cooled. Although the specimens which were cooled in the furnace have been given the same designation — that is, "furnace cooled " — it is probable that the rate of cooling was not always the same, but varied somewhat according to the maximum temperature to which the furnace was heated. It may be inferred from the micrographs (Figs. 9 and 10) that a slight decarburization occurred in specimen E at the high tern- £w?ca] Hardness and Grain Size of Steels 585 peratures, in spite of the precautions taken. Such a slight de- carburization, however, would only decrease the slope of the curve slightly and not affect the general results obtained. The increase in the size of the aggregates of ferrite in the speci- mens heated at the higher temperatures appears to have but little effect upon the micro-Brinell hardness determinations. A slightly greater tendency toward scattering at the higher temper- atures may be observed in the plotted data of the curves, but this is all. The pronounced tendency shown by all the steels toward an increase in hardness at the highest temperature used as com- pared with the same steel heated at the next lower temperature is very striking (Figs 16 to 19). Each specimen shows a marked rise in hardness immediately following the pronounced drop, previously attributed to the increased grain size. The material, although "overheated," has not been " burnt," so the increase of hardness can not be attributed to this cause. It is, however, approaching a condition analogous to that resulting from casting, and evidently this pronounced tendency for an increase in hard- ness is due to some condition similar to that within a casting, which does not reveal itself in the structure alone. While it is realized that the factors discussed do not account entirely for the phenomena observed as regards Brinell hardness of the materials studied, still it is evident that the conclusion is warranted that grain size is a factor of minor importance. The rate of cooling, together with the accompanying structural and other changes due to the transformations which occur within the metal, are of far greater import. There is no simple relation be- tween grain size and Brinell hardness, as in the metals and alloys of simpler structure. VI. GRAIN GROWTH UPON ANNEALING AFTER COLD WORKING 13 The second method for producing coarsely grained material, previously referred to on page 559, consisted in heating specimens which had been deformed by cold working. The results obtained are included here for their suggestiveness rather than as a com- plete study of the subject of grain growth after strain. The rapid recrystallization of iron and mild steel, by which extremely coarse crystals can be produced upon annealing after permanent 13 The authors are indebted to R. W. Woodward for aid in making the mechanical tests discussed in this section. 586 Scioitific Papers of the Bureau of Standards [Vol.16 deformation of the structure, is well know and has formed the subject of several extensive researches. Among these may be mentioned that of Sauveur, 11 who, following the suggestions of Charpy 15 and Le Chatelier, 10 has shown that rapid recrystalliza- ti( m occurs in iron only when the deformation has been carried to a certain "critical" degree. The most extensive study of the subject has been made by Chappell, 17 who differs in his conclu- sions from Sauveur. The specimens used for the production of the coarsely granular areas were somewhat similar to those of Chappell, who employed Fremont's device for obtaining differential stress within the same bar. 18 These were tapered tension specimens of the size and form shown in Fig. 21. As materials, two bars of each of the composi- tions A, C, and D (Table 1) were used; these were annealed for two hours at 650 C and cooled in the furnace. This was for the purpose of removing the effect of the cold working which the sur- face of the specimen received during the necessary machine work and which might affect the subsequent recrvstallization. This U M U I* I* I* 1° Fig. 21. — Tension specimen used for differentially straining the steel before annealing effect may be of considerable magnitude, as has been previouslv shown. 19 The specimens, after polishing the sloping surface suitably for microscopic examination, were stressed in tension until rupture occurred; one observer closely watched the pol- ished specimen and the stress was recorded as soon as a roughening of the surface appeared at each of the marked points (Fig. 21). In this way the necessary stress required to slightly deform the specimen and give rise to the diagonal flow marks, or " Luder's lines," could be measured rather accurately. The maximum stress carried by the bar at each of the marked subdivisions was also calculated from the measured maximum stress borne by the specimen. After microscopic examination of the surface of the strained bars for the occurrence and distribution of slip bands as 11 Albert Sauveur. Crystalline Growth of Ferrite Below Its Thermal Critical Range, Proc. Int. Asso. Test. Mat., oth Cong . '-; 1912. 15 Georges Charpy, Sur la Maladie de l'Ecrouissage, Rev, de Met., Memoirs, *. p. 655; 1910. le Henri Le Chatelier, Notes de Metallographie, Rev. de Met,. 8, p. 367; 1911. 17 C Chappell, Reerystallization of Deformed Iron, J. Iron and Steel Inst., 89, No. 1. p. 460; 1914. " Knmont. Mesure de la Limite Elastique des M^taux, Bull. Soe. Encouragement Ind. Nat., itb. Pt. II, p. 363; 1903. Iy H. S. Rawdon. B. S. Tech. Papers, No. 60. jiHww-ai] Hardness and Grain Size of Steels 587 distinct from the Luder's lines, the specimens were annealed for a period of six hours at 686° C (680-692) ; that is, slightly below the critical range. One bar, A-i , C-i , and D-i , of each set was cooled in the air, the others, A-i 1 , C-i 1 , and D-i 1 , were allowed to cool in the furnace. It has recently been shown by Hanson 20 that the time required for the recrystallization of strained metals to occur, at least for the soft metals, aluminum, zinc, etc., is very short. A period of a few minutes at the chosen temperature is sufficient. The object of the long period of annealing was to remove the hard- ness due to the cold work, as well as to permit the maximum grain growth to occur. It was hoped that in this way specimens which varied only in grain size along the length of the bar might be obtained. The results of the hardness measurements of the strained-and- annealed specimens, together with the data of the tensional stressing, are summarized in Table 7. The appearance of some of the bars after recrystallization is shown in Fig. 23. In the case of the low-carbon steel only (A-i and A-n, 0.07 per cent carbon) was the grain found to increase in size during the treatment given. In the specimens of higher carbon content (0.40 and 0.70 per cent) no appreciable change could be detected. This confirms Chappell's observations 2l in this respect. The hardness measurements of specimens A-i and A-u are in general conformity with those of the bars of which the grain was coarsened by heat alone (Tables 4 and 5) ; that is, a pronounced increase in grain size is accompanied by a lowering of the Brinell hardness. In the other specimens (C and D) the hardening effect of the straining was removed by the prolonged heating, and the speci- men showed no pronounced differences in hardness along its length. In some cases the end of the bar which was most severely strained was found to be somewhat softer than the other, although no perceptible increase has occurred in the size of the grain in such specimens. Although the steels were heated at a temperature considerably below that of the A t transformation, all the speci- mens cooled in air were noticeably harder than those allowed to cool in the furnace. The results of the tension test throw some additional light upon the conditions necessary for the rapid recrystallization of soft steel upon annealing after strain. -" D. Hanson. Rapid Recrystallization in Deformed Nonlerrous Metals. J. Inst. Metals, '20, No. 2, p. 141; 1918. 21 J. Iron and Steel Inst., 89, No. 1. p. 460; 1914, 588 m ;'*3 w , ^HnHHngKj rf .H 3 -A* ' t? 4- i 'i *«'*. M ; ? ■/€">■■*- -£*> 5" ■ Fig. 22.— ( ondition of the surface of tapered tension specimen of o.oy pei cent carbon steel .1 a/'tef straining; : LIBRARY OF CONGRESS