experimental scatter of replicate determinations of either kind; the same conclusion obtains for the experiments on ethyl iodide and n-propyl iodide (vide infra).
TABLE I11 PARAMETERS OF THE ARRHENIUS AND EYRINQEQUATIONS FOR REACTION OF CYANIDEIONWITH ALKYL IODIDES
TABLE I SPECIFICRATE CONSTANTS FOR THE REACTION BETWEEN CYANIDE IONA N D METHYLIODIDE (kr i av. devn., 1. ,mole-%seo.-l) X 10' CN -
from estimation of I-
log Aa
koal. mole-:
AX*, koal. mole-1
10.8 11.5 11.7 11.2
19.2 19.4 20.4 19.9
18 6 18.8 19.8 19.3
Eexp,
Iodide
t, oc.
Vol. 64
W. D. JOHNSTON, R. R. HEIKESAND J. PETROLO
1720
Solvent
CHs CHa CzHs n-C3HT
W Aq.EtOH Aq.EtOH Aq.EtOH
AS'ma,
e.u.
-11.1 -7.5 -6.9 -9.3
Av.
Discussion The values recorded in Table I11 are in general + 11.4 correFpondence with activation parameters reported 19.5 for similar ion-molecule and in good 31 .O ngreemen t with the Arrhenius parameters reported 45.8 recently by NIarshall and Moelwyn-Hughes12 for 58.3 the methyl iodide cyanization in aqueous solution. For methyl iodide cyanization the difference of B. Solvent: 50% aqueous ethanol 4.95 zk 0.05 4.89 f 0.14 4.91 f0.14 about 3.5 e.u. in the entropies of activation in the 11.4 two solvent systems is in the direction and of the 37.1 f 0 . 4 37.8 f 0 . 3 38.4 f 0.3 31.0 magnitude which would be expected for a simple 129.2 f 5 . 6 125.5 f 5 . 5 127.4 f 5 . 7 43.0 dielectric effect.13J4 For a cyanide ion radius of The results obtained for the reaction of cyanide 1.05 A.,l6J6and using values for the dielectric conion with ethyl and n-propyl iodides are summarized stants of water-ethanol mixtures interpolated from in Table 11. The Arrhenius and Eyring equation the data of Akerlof," the effects of solvent on IC2, attributed solely to variation in the entropy of activaparameters are collected in Table 111. tion, correspond to a radiusofor the activated complex of approximately 4 A., which seems quite TABLE I1 reasonable. SPECIFICRATE CONSTANTS FOR THE REACTION BETWEEN Acknowledgment.-This research was supported CYANIDE IONA N D ETHYL AND n-PRopyL IODIDES IN 50% by the U. S. Atomic Energy Commission. A. Solvent: water 1.08 0.11 3.04 f 0.26 10.5 f 0 . 3 40.3 f 0.8 137.1 f 6.9
AQUEOUS ETHANOL
Av.
(8) Estimates of imprecision: log A , &0.3; Eexp and A H * , f 0.3 kcal. mole-1: AS*iss. zt0.8 e.u. (9) T. I. Crowell and L. P. Hammett, J. Am. Chem. Soc., 70, 3444
1.17f0.09 8.94f0.26 49.2 f 3 . 3
(1948). (10) P. M. Dunbar and L. P. Hammett, abid., 72, 109 (1950). (11) J. Hine and W. H. Brader. Jr., tbid., 75, 3964 (1953). (12) R. W. Marshall and E. A. Moelmyn-Hughes, J. Chem. SOC. 2640 (1959).
av. devn. 1. ,mole-' sec.-l) X 10' from estimation of CN 1-
(kl &
1,
"C.
11.4 31.0 43.2
A. Ethyl iodide 1.12 f 0.09 1.22 f 0.09 8.62 f 0 . 2 8 9.13 f .I5 50.4 f 6 . 2 48.4 f .8
11.4 31.0 43.2
n-Propyl iodide 0.91 f 0.06 0.91 f 0.06 7.10 f 0.48 6.70 f 0.35 27.0 f 2.4 27.9 f 2 . 4
(13) K. J. Laidler and H. Eyring, Ann. N. Y . Acad. Sce'., 59, 303
(1940).
B
0.91 f0.06 6.90f0.42 27.9 1 2 . 4
(14) (15) (16) (17)
E. D. Hughes and C. K. Ingold, J . Chem. Soc., 244 (1935). J. M. Bijvoet and J. A. Lely, Rec. trau. cham.. 59, 908 (1940). G. J. Verweel and J. M. Bijvoet, Z. Kmst., 100, 201 (1938). G. Akerlof, J . Am. Chem. Soc.. 54, 4125 (1932).
THE PREPARATION OF FINE POWDER HEXAGOXAL Fe?C AND ITS COERCIVE FORCE BYW. D. JOHNSTON, R. R. HEIKESAND J. PETROLO Westinghouse ReseaTch Laboratories, Pittsburgh 56, Pennsylciania Received May 16, 1060
The relatively rare hexagonal form of FezC has been prepared in fine owder form by the reaction of a gaseous mixture of HI and CO with Raney iron at 240'. The occurrence of the hexagonay form appears t o require the exclusion of Fea04. If FesO, is present, the carbide is found in the more common Hagg form. The coercive force of the fine powder hexagonal Fe& prepared by this method is approximately 800 oe.
Introduction The compound Fe2C is known to exist in two crystalline forms. The more common form is referred to as the Hagg carbide which may be indexed On either Or orthorhombic axes' The formula for this carbide has been given as both Fe2C and The other form is relatively
rare and is a distinctly different hexagonal structure in which the iron atoms are believed to be arranged in hexagonal close packing2 This form is referred to as the hexagonal carbide. The conditions for (1) K.H. Jack, Proc. Rov. Soc. (London). 8195, 56 (1948). (2) L. J. E. Hofer, E. M. Cohnand W. C. Peebles, J . A m . Chem. Soc.. 11, 189 (1949).
PREPARATION OF FINEPOWDER HEXAGONAL IRON CARBIDE
Nov., 1960
1721
formation of the hexagonal carbide are quite ob- mett. In order that u l t r a h e Fe2C would result, the iron scure. It has been suggested that iron catalysts starting material necessarily had to be ultrafine, -200 A. Fine iron prepared by the Raney process was used in containing copper12or iron catalysts promoted by a this work. The preparation and magnetic properties of mixture of KzO and &os3 or partial carbidinga this material have been described previously.8 Initial carburization experiments involved the handling result in the hexagonal form. It has also been shown that the hexagonal form transforms appreci- of the pyrophoric iron owder in a nitrogen-filled dry box. procedure was suise uently abandoned since i t was ably to the Hagg form a t temperatures in excess of This difficult to ensure that the j r y box remained free of traces of 300 O . oxygen. Instead of using a glove box it was possible to Little magnetic data are available for these ma- transfer Raney Fe to the Pyrex carburization vessel through terials. Hofer et ul. 2,4,5 have given 247 and 350- a ground joint by pipetting the iron powder under pyridine. This transfer technique prevents oxidation of the pyrophoric 380' as the Curie points for the Hagg and hexagonal iron powder by the air. The use of pyridine as a protective forms respectively. Hofer5 also gives data from agent for Ranep iron has been previously described.8 which the average values of the saturation magne- The pyridine could be quantitatively removed from the tization per gram of compound, Q, may be calculated. iron by continuous evacuation first a t room temperature finally a t 100'. (The grease used in the two sto cocks These values are 145 for hexagonal FezC 130 for and and the ground joint of this apparatus was Dow &rning Hagg FezC and 139 for cementite Fe3C. Since high vacuum silicone grease which interacted only slightly both forms of Fe2C are crystallographically uni- with yridine vapor.) By this technique the carburization axial, it is probable that they have high values of vessercould be brought to constant weight in order to dethe weight of the iron sample (generally 2-4 g.). the magnetocrystalline anisotropy constant K . termine At the end of the Carburization run a second weighing perSince coercive force is directly proportional to K mitted the carbon uptake to be calculated. Carbiding by the relationship I H , = 0.96 K/ad where IH, is gasee were passed over the iron sample contained in an enthe coercive force and d is the density, a large value larged part of the vessel which was contained in a wellregulated furnace. A thermocouple placed in the well of of coercive force might be expected for single do- the carburization vessel permitted the sample temperature main spherical particles. Although the values of K to be measured during the course of a run. Products could for these materials are unknown, it is possible to get be removed from the carburization vessel a t the end of a some feeling for the magnitude of coercive force that run by evacuating the vessel and then introducing a protective cover of pyridine by way of a separatory funnel might be expected if one uses the reported value which was connected to the vessel by one of the stopcocks. K,, for Fe3C (1.18 X lo6 e r g s / ~ m . ~ ) The ~ . value In this way, contact with the air was completely eliminated obtained for IH, using this approximation is of the and the sample could be easily transferred by ipetting as order of 1000 to 1200 oe for the various iron car- previously described. Before carburization a\ gas lines were evacuated and then flushed with the appropriate gases bides. to prevent any possibility of oxidation of the sample. Since the iron carbides are thermally unstable Instrument-grade propane gas purified over hot copper materials they cannot be prepared by methods gave little or no carbide formation a t temperatures of which involve elevated temperatures. Further, for 240-275' and for times as long as 9 hours in contrast to the findings of Emmett8 which suggests that an oxide supported the production of the fine particle materials (-200 iron catalyst may be considerably more reactive than bulk A. diameter) lyhich might yield high values of coer- Raney iron. Purified CO gave carbide formation but the cive force, temperature must be kept to a minimum. products were always contaminated with Fe804. (Both Hofer, Cohn and Peebles2 have studied FezC in tank CO and CO formed from C02 and graphite at 1200' used. These gases were purified by passing through connection with catalysis in the Fisher Tropsch w.ere Ascarite and magnesium perchlorate.) Finally, pure synthesis. In this work they developed a method samples of Fe& were prepared by carbiding a t 240' with a for preparing both FezC and FeaC. Their prepara- stream of CO and hydrogen in which the ratio CO/H2 varied tions can be described in terms of the r e a c t i o n ~ . ~ ~from ~ ~ ~ J to */l. (A mixture of Hz and CO has reviously
+
240 '
2Fe 2CO +Fe2C Fe2C Fe +Fe3C
+
+ COZ
I n preparing Fe3C Cohn and Hofer interrupted the first reaction a t 2/3 completion where Fe and FezC were present in equal amounts. They then annealed this mixture in vacuo and allowed the second reaction to take place. The preparation of FezC has also been described by Emmetta whose interest was also in catalysis. Emmett preferred to carburize iron-containing catalysts with a hydrocarbon such as propane. He states that with a hydrocarbon there is no oxidation of the product, no exotherm in the carburization reaction (as has been noted for a CO carburization) and no free carbon precipitated. Preparation Techniques.-At the outset of this work attempts were made to obtain Fe2C in ultrafine particles by applying the methods of either Hofer and Cohn or Em(3) H. H. Podgurski, J. T. Kummer, T. W. DeWitt and P. H. Emmett, J. Am. Chem. Soc., 72, 5382 (1950). (4) E. M. Cohn and L. J. E. Hofer, ibid., '73, 4662 (1950). (5) L. J. E. Hofer and E. M. Cohn. ibid., 81, 1576 (1959). (6) P. Blum and R. Pauthenet, Compt. rend., 2ST, 1501 (1953). (7) L.J. E. Hofer and E. hf. Cohn. J . Chsm. Phys., 18,766 (1950).
been used by Trillat and Oketonig and Michel an{ Bernierlo for the preparation of cementite from iron a t temperatures in excess of 500°.) Little or no oxide was found in the Xray patterns of the Fe2C products. In addition more carbon was taken up by the iron when hydrogen was used in addition to CO. In this case however a side reaction also occurred yielding small amounts of a yellow hquid. This liquid has not been investigated but it is probably either a mixture of hydrocarbons or iron carbonyl.
Results In Fig. 1 are graphed data obtained by carbiding with CO HP. The carbon uptake indicated as z in FezCz,was determined from the weight gain of the sample. Values of 1.0 and 0.9 in the 2 parameter correspond to the two formulas proposed, FezC and FezoC9.' Invariably the carbiding reaction proceeded rapidly at first accompanied by a 40-60' exotherm and then proceeded more slowly to complete carburization in about 6 hours as indicated on the graph. X-Ray diffraction patterns obtained on these materials provided qualitative confirmation of the weight gain data. In samples no. 15 and 16
+
(8) W.D.Johnston, R. R. Heikes and J. Petrolo, J . Am. Chem. SOC. '79, 5388 (1957). (9) J. J. Trillst and 8. Oketoni. Comp. rend., 2SS,51 (1951).
(10) A. Michel and R. Bernier, Reu. Met., 46, 821 (1949).
W. D. JOHNSTON, R. R. HEIHESAND J. PETROLO
1722
Vol. 64
used was '/s instead of the usual 6 / ~ or 3/1. The satisfactory agreement of the data indicated that the CO/H2 ratio was not critical. Sample No. 17 provided the only discrepancy between the weight gain data and the X-ray observations. This sample was carburized a t 200' instead of the usual 240'. The weight uptake was high in this case but according to the X-ray pattern only partial carbide forma6001 tion occurred. Thus we conclude that at 200' part of the weight gain of the sample was due to the deposition of gross amounts of either iron carbonyl or the formation of high molecular weight hydrocarbons and that a higher temperature is required 1.1 for the formation of pure iron carbide. It was surprising to find that the carbide was formed almost entirely in the relatively rare hexagonal form in the experiments where hydrogen was used with the carbon monoxide. In reactions where hydrogen was not added to prevent oxide contamination, the amount of the hexagonal form decreased and the more common Hagg form in0 creased in amount. It thus appears that the preN al LL dominant effect in the stabilization of the hexagonal C .14 240 VI - carbide is the absence of Fe304,and it is conceivable AX that there may be little or no connection with the 15 240 611 - presence of copper or A 1 2 0 3 KzO. .3 ._ 16 240 3/1 The saturation magnetization of sample no. 13 .2 # 17 200 3/1 was measured and found to be 127.5 c.g.s./g. at 18 240 Y3 -78.5' and 122.4 c.g.s./g. at - 18' which are somewhat lower than the 145 c.g.s./g. measured by I I I I I I I 2 3 4 5 6 Hofer6a t 50'. Hours Reacted with CO. Figure 1 shows that the coercive force of these Fig. 1.-Carburization time us. iHo and carbon uptake. samples is relatively high indicating that fine particles have been obtained. The graph also shows that I H , increases rapidly with carbide formation. The sample prepared a t 200' has a relatively low 900I-I value of IH, as would be expected since as mentioned above the X-ray pattern indicates that only -0 w 800 partial carbide formation occurred. It should be pointed out that the values of IH, for 0 700 earlier samples, in which oxide contamination occurred, are much lower than those shown here. E 6 0 0 For eight earlier samples the values for 1H, range from 450-635. These data emphasize that for high 0, 500 e purity samples hydrogen is required in the carbiding gas. I2 400 Vacuum annealing of partially carburized mixW > 'E 300 tures of Fe and FezC successfully gave Fed2 at 400'. This reaction has been observed to start at tem8 u 200 peratures as low as 260°.2*4JJ In order to note the change in I H , as this reaction progressed, annealing 0 lA . I 1 I I 1 I I I / 100 200 300 400 500 experiments were performed on samples no. 14, 15 and 16 by measuring the coercive force of the speciTemperature ("C). Fig. 2.-Temperature us. coercive force; samples 14, 15, mens as a function of annealing time and tempera16, of Fe Fez C. Each sample annealed for 0.5 hr. a t ture. These three samples were used since they each temperature indicated from 275 to 500". spanned the optimum composit,ion for the preparation of Fe&. In all cases the value of IH, remained weak iron lines were visible in accord with the essentially constant t,o 325' and t,hen dropped partial carbon uptake due to insufficient reaction gradually as shown in Fig. 2. This may be due to a time. In samples no. 13, 14 and 18 where the number of effects among which are particle growth weight gains approached that for a complete reac- or an unfavorable difference in either critical particle tion, only iron carbide lines appeared. Sample no. size or magnetocrystalline anisotropy between the 18 is of particular interest since the CO/H2 ratio two materials.
r
I
I
I
I
1
1
c
-4
X
+
i
C
v)
E
v
+
