Sept., 1955
MAGNETIC SUSCEPTIBILITY OF THE SYSTEM Ni0-AI2O3
ion in the catalyRt surface, as in the model of an acid site proposed by M. W. Tamele (ref. 1). G. A. MILLs.-Quinoline chemisorption studies have shown only about one-tenth the acidity values of those reported by the butylamine titration procedure indicating a range of acid strengths for Si02-Al203, as indeed is discussed by the Author. Yet it is stated that "no appreciable difference in total acidities was found when different indicators were used." How is this accounted for? Is it a consequence of the long times used in the titrations, as is the case with sodium hydroxide?
831
0. Jo"soN.-The low values of acidity obtained by quinoline chemisorption may be due to the large size of the quinoline molecule and to the higher tem erature used as well as to the effect of acid strength propose$ by Dr. Mills. Recent work in our laboratories has established that there is a range of acid strengths on acid catalysts, but the present data in the pK range of 1.5 to 5 did not show this effect. The long time used in the titration simply assures that equilibrium is reached and does not represent an irreversible reaction with the catalyst as in the case of sodium hydroxide.
MAGNETIC SUSCEPTIBILITY STUDIES IN THE DUAL HYDROUS OXIDE SYSTEM : Ni0-Al2O3 BY W. 0. MILLIGAN AND JAMES T. RICHARDSON' The Rice Instilute, Houston, Texas Received February 86. 1066
A comprehensive investigation of the magnetic susceptibility of the system: Ni0-A120p, has been conducted on 105 samples, prepared a t every 5 mole yo,and heat-treated at temperature levels of 300,400, 500, 600 and 700". The magnetic measurements were made a t room temperature. The resulting susceptibility-composition curve for the 300" temperature level exhibits three regions of interest,. I n region (a) from 100-80 mole % NiO, the susceptibility curve drops rapidly as the ' NiO; alumina content increases; in region (b) the curve exhibits a broad maximum a t a composition of 60 to 80 mole % and in region (c) the susceptibility increases re ularly as the amount of alumina is increased. The following somewhat speculative explanation is offered. I n region (a$, the initial higher susceptibility of the pure NiO is attributed to the decreased magnetic interaction, resulting from a decrease in the number of next nearest magnetic neighbors, as a consequence of the very small crystallite size, and the plate-like mor hology of the aggregates of crystallites. I n region (b), the maximum in the susceptibility curve results from the existence ofa marked change in the structure of the alumina gel, previously known from X-ray studies t o occur in the zone of mutual protection against crystallization. Such zones are likewise detectable in numerous multiple hydrous oxide systems by means of X-ray and electron diffraction studies (amorphous-like zone), and adsorption measurements (enhance differential and integral isoteric heats of adsorption and enhanced specific adsorptive capacity). In region (c) the regularly increasing magnetic susceptibility results from the increase in degree of dispersion of the NiO within the alumina gel structure. At higher temperatures of heat-treatment, the susceptibility of the NiO increases regularly throughout the entire com osition range. Here the composition zone of mutual protection has narrowed or vanished. At the low concentrations ornickel, the susceptibility curves and the magnetic moments for NiO increase toward a limiting value of 3.2 bohr magnetons in accordance with the previous work of Selwood.
Introduction I n a previous report2 from this Laboratory, it was found that dual hydrous oxide gels in the system NiO-AI2O1 exhibit a mutual protective action against crystallization. The observed composition zone of mutual protection occurred a t about 50-70 mole yo nickelous oxide, in samples heat-treated at a temperature of 500". The zone of protection narrows and finally vanishes as the temperature of heat-treatment is increased, as do multiple zones of protection which have been observed in numerous other multiple hydrous oxide svstems.a At the 1000" temDerature level. X-rav L ~ d e n c edemonstrated the -formation of nick& spinel, Ni0.A1203. The magnetic properties of the system Ni0-AI2O3 have been studied recently bv Selwood and Hill4 in the region of low nickel oxide concentrations. The paramagnetism of the nickel increased markedly when small amounts of nickel oxide were dispersed over a large excess of alumina. Current investigationss in this Laboratory show that very small crystals of pure nickel oxide, prepared by the controlled thermal decomposition of hydrous nickel(1) Humble Oil & Refining Co. Fellow at The Rice Institute. 19521955. (2) W. 0. Milligan and L. Merten, THIEJOURNAL, 60, 465 (1946). (3) W. 0. Milligan, ibid., 66, 497 (1961). (4) P. W. Selwood and F. N. Hill, J . Am. Chsm. Soc., '71, 2522 (1949). (5) W. 0. Milligan and James T. Richardson, t o be published.
ous hydroxide, do not exhibit an antiferromagnetic NBel point a t 250°, as do massive crystals previously studied by others. Traces of sodium, or possibly silica, irreversibly adsorbed a t the time of precipitation of the gel, stabilize the cubic, paramagnetic form of NiO, whereas the unstabilized oxide is antiferromagnetic and rhombohedral below the NBel point (about 250"). Inasmuch as Selwood's careful investigations did not include the high concentration range of nickel oxide, and in view of our results obtained for highly dispersed nickel oxide free of alumina, it was considered desirable to extend our previous magnetic studies to the entire system NiO-A1203. It is the purpose of this paper to report the results of a systematic magnetic susceptibility investigation of this system for samples heat-treated a t several temperature levels. Experimental Preparation of Samples.-Dual hydrous oxide gels were prepared by precipitation from varying volumes of 0.5 M nickelous nitrate and 0.5 M (with respect to A120a)aluminum nitrate solutions, using as the precipitant an amount of 3.77 Af sodium hydroxide solution (carbonate-free) such that the final pH value was 7.50. The required amounts of alkali were determined by preliminary titration on aliquots of the metal ion solution mixtures. The compositions of the gels were adjusted to give a series of twenty-one hydrous oxide gels a t every 5 mole %. The gels were washed with distilled water by decantation in five-gallon containers, followed by ten washings in a
832
W. 0. MILLIGAN AND JAMES T. RICHARDSON
which the susceptibility of the NiO alone is calculated on the assumption that the susceptibility of the alumina is independent of the concentration. In the inset in the lower portion of Fig. 1, there is given the magnetic moment in bohr magnetons of the nickel ion as a function of concentration. The magnetic moment, HefI, was computed for concentrations of nickel below 15 weight %, from the relationship pelt = 2 . 8 3 d T T where xm is the molar susceptibility, and T is the absolute temperature. This expression is assumed to be valid a t the low concentrations, where the nickel ions are sufficiently dispersed so that the ideal Curie law holds, in accordance with the views of Selwood.4 X-Ray Diffraction Patterns.-X-Ray diffraction patterns were obtained for many of the samples, employing standard powder methods. These results will not be reported in detail here, inasmuch as they agree with previous work in this Laboratory.e
w* 8C
-1
2 u. 0 f
60
a
UI
K 0 W
"0
40
r(
X
*c 2
-c
20
m
W 0
a
'"0
.-iL 0
5
60
a 0 W
m
2
I
Discussion It will be noted in the lower portion of Fig. 1,
LL 0
K
Vol. 59
40
X
: c
-I
20
t-
P
W
'" u
a u)
0
, WT d N i , 20
40
60
80
100
WEIGHT PERCENT 1110.
Fig. 1.-Upper portion, magnetic susceptibilities per gram of dry sample as a function of composition; lower portion, magnetic susceptibility per gram of NiO as a function of composition. Samples heat treated a t A, 300"; B, 400'; C, 500"; D, 600'; E, 700'. Inset, magnetic moment of nickel ion as a function of low nickel concentration. pentrifuge, until the supernatant liquid was free of nitrate ions. The samples were filtered, air-dried and ground to a uniform powder. Separate portions of the air-dried samples were heated for two hours in a stream of dry nitrogen a t 300, 400, 500, 600 and 700". These samples were then allowed to come to equilibrium with the moisture in the atmosphere, in order that the subsequent transfers could be made without detectable loss or gain of adsorbed water. The series of twenty-one gels heated at 500' was analyzed for nickel content, using the standard gravimetric dimethylglyoxime method. The water content was determined from the loss in weight on prolonged ignition to constant weight a t 1000°. The alumina content was determined by difference, and the mole % of NiO and A1208 were calculated on the dry basis. The series of gels derived from the original air-dried gels a t the remaining four temperatures of heat-treatment were analyzed for water content, as just described, but the mole ratios of NiO to A1203 were assumed to be independent of the temperature of heat-treatment. Magnetic Susceptibility Measurements.-Magnetic susceptibility measurements were conducted in an apparatus already described,e employing techniques reported elsewhere.6 The samples were loaded into the magnetic apparatus concurrently with the weighing of the aliquots for chemical analysis, in order that the water contents of the samples would be accurately known. The measured magnetic Susceptibilities were converted to a basis of one gram of dry sample, assuming that the diamagnetic susceptibilities of adsorbed water, alumina and the diamagnetic contribution of the nickel were all additive. These data are presented in the upper portion of Fig. 1. I n the lower ortion of Fig. 1, there is given an interpretation of the &ta, in (6) W. 0. Milligan and H. B . Whiteliurst, Rev. Sci. Inatr., 23, 618 (1952).
where the magnetic susceptibility of the nickel oxide is plotted as a function of composition, that the curve corresponding to the 300" heat-treatment level exhibits three regions of interest. In the first region a t high concentrations of nickel oxide, the susceptibility decreases rapidly as the amount of alumina in the gels increases. In previous work6 the high susceptibility of the pure nickel oxide was attributed t o extremely small size of the crystallites, and to the plate-like morphology of the aggregates (secondary particle) of the crystallites. The effect of reduction in crystal size and the morphology of the aggregates on the ratio of specific surface to specific volume (and hence the average number of next nearest magnetic neighbors) on the magnetic interaction is detailed elsewhere.s The observed reduction in the susceptibility as a consequence of the addition of alumina could be explained on the basis of an increase in crystal size, or to change in the plate-like morphology. X-Ray diffraction studies show that the addition of small amounts of alumina retards the development of the original hexagonal nickelous hydroxide plates, and the samples of gels containing alumina consist of spheroidal particles. Detailed X-ray diffraction, electron diffraction and electron microscopic studies have not been made on the structure and morphology of the original airdried hydrous oxides or hydroxides, but it is evident that the plate-like form of the nickelous hydroxide is modified, As a hypothesis, on the basis of the presently known information, the reduction of the susceptibility of the nickel oxide on addition of alumina is attributed to change in the morphology of the original nickelous hydroxide, which is reflected in the samples heat-treated a t 300") so as to give spheroidal aggregates, rather than platelike aggregates. Inasmuch as the size of the crystallites in the pure NiO a t a heat-treatment level of 300" is of the order of 50-100 A., one would expect little further reduction in size as the alumina is added. The second region of interest is the maximum in the susceptibility curve in the composition range 60 to 80 mole % nickel oxide. This composition region corresponds to the zone of mutual protection already reported2 for the system, wherein the samples are essentially amorphous t o
..
Sept., 1955
MAGNETIC SUSCEPTIBILITY OF THE SYSTEM Ni0-A120a
X-rays. The maximum in the susceptibility curve for NiO in this composition region is considered to reflect the enhanced dispersion of the NiO within the gel structure of the alumina. In the third region of interest, the susceptibility of the NiO increases in a regular manner, because of the increasing dispersion as the amount of alumina increases. This regular change would have continued toward the higher concentrations of NiO, had not the zone of protection occurred, wherein the properties of the alumina gels were markedly changed. The one or more zones of mutual protection encountered in numerous multiple hydrous oxide systems were first detected by X-ray diffraction methods.' I n 1948 it was predicted' that such zones of protection may correspond to regions of maximum surface, and which, therefore, may exhibit enhanced adsorptive and other surface properties. Enhanced differential and integral isoteric heats of adsorption18enhanced specific adsorptive capacity8 and a marked increase in the number of crystallites per aggregateeJOhave been found in the zones of mutual protection occurring in the system BeO-InzOa. In view of these results, it is not surprising that magnetic susceptibility measurements can detect in a zone of protection an enhanced dispersion of NiO in alumina, the increased dispersion making the Ni ions more mag(7) H. B. Weiser, W. 0. Milligan and G. A. Mills, TRIEJOURNAL, S2, 942 (1948).
(8) W. 0. Milligan and C. R. Adams, ibid., 67, 885 (1953). (9) C. R.A d a m and W. 0. Milligan, ibid., 68,219 (1954). (10) C. R. Adams and W. 0. Milligan, ibid., 68, 817 (1954).
833
netically dilute, thereby decreasing the average number of next nearest magnetic neighbors, and thus the magnetic interaction. In the samples heat-treated at successively higher temperatures, the zone of mutual protection as indicated by X-ray studies2 narrows and vanishes, and inspection of Fig. 1 shows that the magnetic susceptibility of the NiO does not show any irregularity in this region. I n an earlier investigation,2 X-ray diffraction patterns established the existence of a definite nickel spinel, NiO.A1203, after heat-treatment of the original air-dried gel at lOOO", the new crystalline phase not being detectable a t lower heattreatments. It was considered possible that magnetic susceptibility measurements might be able to detect the formation or incipient formation of the spinel at lower temperatures of heat-treatment. However, no indication has been found of any irregularity in the susceptibility at a mole ratio of NiO to of unity. In the inset of Fig. 1, for very low nickel concentrations, the magnetic moment approaches a limiting value of 3.2 bohr magnetons. These results agree almost exactly with the previous results of S e l ~ o o dand , ~ hence will not be discussed in detail here. We are grateful to the Humble Oil & Refining Company for making possible this investigation by establishing a fellowship program a t The Rice Institute, and for making available facilities for the magnetic measurements and for chemical analyses. We wish to thank Dr. Charles F. Squire and Dr. L. W. Vernon for helpful discussions.
