NOTES
436
Vol. 63
and
suggest that it might be used as an alternate method for initiating rapid reactions, but with the important differences outlined below. (1) There is only slight dependence of blackfor an infinite cylindrical absorber of radius r suspended along the axis of a cylindrical transparent body absorption on wave length. I n the usual flash photolyses, often only the ultraviolet tail of the iiisulator of radius ri. flash emission is used for excitation because many Inserting the values F = 1.32 cal. cm. -2 flash-', of the molecules of interest absorb only in the blue At = 2.5 X sec., K 2= 4.0 X cal. cm.-' set.-' deg.-', u = 1.36 X 10-l2cal. see. deg.-4, pt or ultraviolet regions of the spectrum. The flash = 2.25 g. ~ m . - ~ cpl , = 0.45 cal. g.-I deg.-l, el = lamp output,g however, extends from the ultra0.997, rl = 0.05 cm. and TO = 298'K. we obtain violet into the near infrared wave length regions. the bell-shaped curves shown in Fig. 4 for tempera- With the heterogeneous heating technique, much ture rises of spheres and cylinders of carbon em- more of the flash lamp output can be converted bedded in a matrix with the thermal conductivity into chemical excitation. (2) There is a threshold flash energy for the of a liquid hydrocarbon. If, the matrix is a gas a t pyrolytic excitation (Fig. 2) which is not seen in low pressure, the lower thermal conductivity causes true photochemical reactions, except perhaps with the temperature to approach 5000°2ba t radii below very short flash times, or unless part of the disloU4cm. These are the temperatures necessary to vaporize the carbon and tungsten mentioned cretely absorbed radiation is converted to heat in earlier. The temperature rises in gases are seen the secondary reactions, which could add an to be in the same range as those produced by shock intensity-dependent pyrolysis to the original photolysis. tubes.6 (3) The pyrolytic process does not depend on Note from equations 1 and 3 that temperature the wave length of absorption of the reagents. rise increases with increasing flux F, and with decreasing flash time At. We have not studied these This permits initiation in weakly or non-absorbing variables independently as yet, because in our systems, and, if kinetic spectroscopic measurements circuit both depend on capacitance per flash, our are to be made on the reaction intermediates, it is only circuit variable. Our several millisecond flash possible to select parent reagents whose spectra do times are rather long compared to other flash cir- not interfere with those of the labile species. It is cuitsle-18 so that probably we have not reached the also likely that the flash heating technique will permit study of intermediates which do not occur highest temperatures possible. in photochemical reactions. Recently flash photolysisa has been used ex(4) The flash heating technique has a hot, very tensively to initiate rapid photochemical reactions, reactive surface present. Such a heterogeneous including some which are essentially p y r o l y s e ~ . * ~ - ~ ~ The short reaction times and the vigor of the reac- energy transfer medium will complicate many retions produced by heterogeneous flash heating actions, just as with homogeneous sensitizers.20-21 This may be a serious disadvantage in some cases, (16) J. H. Callomon and D. A. Ramsay, Can. J . Phus., 3 6 , 129 but perhaps of great interest in the study of surface(1957). labile species interactions. (17) 8. Claesson and L. Lindqviat. Arkiv Kemi, la, 1 (1958). We are deeply indebted to our many colleagues (18) H. Fischer, J . Opt. Soc. Am., 47, 981 (1957). (19) K. Rnox, R. G. W. Norrish and G. Porter, J . Chem. Sac., 1477 for technical advice and assistance in the work re(1952). ported here. We wish especially to thank Mrs. (20) R. G. W. Norrish and G. Porter, Proc. Roy. Soc. (London), M. Y . Hellman, Mrs. E. Tvetenstrand and Mr. AalO, 439 (1952). N. A. Kuebler for performing many of the experi(21) R. G. W. Norrish, G. Porter and B. A. Thrush, ibid., AM7, 423 (1955). ments.
L
I
i
L
NOTES ADSORPTION CHARACTERISTICS OF HOMOIONIC BENTONITES BY W. H. SLABAUGH Contribution from the Department of Chemistry, OreoonlState Cotteoe. C o w a l l i s , Oreoon Received August & 1968
A number of investigators'-6 have examined the (1) S. B. Hendricka, R. A. Nelson and L. T. Alexander, J . Am' Chem. Soc.. 62, 1457 (1940). (2) R . W . Moonev. A. G. Kienan and L. A. Wood, iba'd.. 74, 1371
(1952). (3) A. C. Zettlemoyer, C. J. Young and J . NAL. 69, 902 (1955).
J. Chessick, THISJOUR-
adsorption and desorption of water vapor by clay materials. The present, work is an attempt to point out further the importance of the exchangeable cation on these adsorption characteristics, and to propose a two step procesg in the adsorption of water vapor. Experimental Materials.-Homoionic bentonites were prepared in an ion-exchange column in a manner described previously.6 (4) J. R . Goates and C. V. Hatch, Soil Sci., 76, 275 (1983); 77, 313 (1954). (5) 13. D. Orchestron, ibid., 76, 453 (1953). (6) W. 11. Slabaugh, THISJOUBNAL, 69, 1022 (1955).
March, 1959 The bentonite was a Wyoming type, supplied by the Baroid Division of the National Lead Company. This study was bascd primarily upon the performance of ordinary freeze-dried clays, with an occasional reference to oven-dried clays which usually showed quite similar Lehwior, particularly with the univalent cation types. There was an obvious advantage in studying the freesedried samples because of their rapid response to vapor pressure changes, particularly upon out-gassing, where ovendried clays showed a tendency to “explode” when vapor was drawn off rapidly. Equilibration time was usually about one hour on adsorption and three or more hours on desorption. Apparatus .-An eight-place gravimetric adsorption device, epploying Cu-Be alloy springs thermostated at 30.0 =k 0.1 , was used to observe the adsorption behavior of eight samples concurrently. There was an over-all accuracy of measurement, accounting for temperature control, spring drift, and indeterminate sources of error, of no more than 0.4% error in the weight measurements.
Results and Discussion I n all, adsorption characteristics of fourteen homoionic clays were examined by observing their interaction with water vapor at three temperatures. These clays fell into three groups, those similar to sodium bentonite (Fig. l ) , those similar to calcium bentonite (Fig. 2), and those that produced a smooth, conventional S-shaped isotherm. The first group-sodium, potassium and magnesium clays-adsorb water vapor in two distinct steps. For example, sodium bentonite over the temperature range 14 to 37.5” adsorbed water vapor linearly up to 28 mg. per gram of clay. At this point the slope of the isotherm changed abruptly, and water vapor was further adsorbed linearly. The desorption of water vapor from the monovalent cation clays was far from reversible, as noted in the desorption curve a t 14” in Fig. l . Calcium bentonite was unique and alone in the second group. The three adsorption isotherms for calcium bentonite in Fig. 2 show that, even though some kind of phase change or discontinuity appears a t lower temperatures, there is apparently smooth adsorption a t the highest temperature (37.5’). Call’ discovered a similar two-stage process in the swelling of calcium bentonite that was freeze-dried, the two stages corresponding to two different 001 spacings of the lattice. I n another study, to be reported in a later paper, there is evidence that a t about 31” there is a notable change in the surface behavior of calcium bentonite. Consequently, it is not surprising to observe a change in the shape of the adsorption isotherm a t temperatures above this transition point. A third group of clays showed S-shaped adsorption isotherms similar to the calcium bentonite curve a t 37”. These include H, Li, Sn, Ba, Zn, Cd, Fe and Ni bentonites, and none of them show any obvious indication of discontinuities up to 0.8 relative vapor pressure. In spite of the lack of discontinuities in many of these isotherms, many of them, upon analysis through the BET equation, reveal two distinct values for Vm, the volume of water vapor required to form a monolayer. Following the concept of Dole8 the two values for V,, as derived from the two slopes and the two corresponding intercepts for the two legs of each curve shown in Fig. 3, are con(7) F. Call, NatuTe, 172, 126 (1953). ( 8 ) M. Dole, J . Chsm. phys., 16, 20 (1948).
NOTES
437
100 I
I /37.5.
I
-.50 M
8 4
Na-Bentonite
0
02
0.1
0.3
0.4
0.5
,
0.6
0.7
0.8
PIPa. Fig. 1.-Adsorption
isotherms of sodium bentonite a t three different temperatures.
0.3 0.4 0.5 0.6 0.7 0 B PIPO. Fig. 2.-Adsorption isotherms of calcium bentonite a t three different temperatures. 0
0.2
0.1
r0 I
/
37.5a
I
Na- Bentonite
0
0.1
0.2
0.3
0.40
Ca-Bentonite
0.1
0.2
,
0.3
-
0.4
PIPO. Fig. 3.-BET plots for water vapor on sodium and calcium bentonite a t three different temperatures.
sidered to be two groups of molecules adsorbed a t unique energies. For instance, at 27.0’ .the fir& value of V , for sodium bentonite gives an area of 86 rn.”g. and the second value of V , corresponds to 188 m.2/g. For calcium bentonite the areas are 299 m.2/g. and 462 m.2/g. for its two V , values. Occasionally, the BET plots gave relationships from which only one V , value could be determined.
NOTES
438
.-$
14,0001
e
O
1 Ca- Bentonite
12,000.
: 10,0001
t? I+
0 20 40 60 80 100 120 Mg. HaO adsorbed/g. bentonite. Fig. 4.-Isosterio heats of adsorption of water vapor by sodium and calcium bentonites.
Ordinarily, isosteric heats are calculated from a pair of reversible isotherms at two temperatures. In this instance, however, there was hysteresis in the adsorption-desorption process, particularly with the monoionic clays, and the isosteric heats obtained in this manner may not be strictly valid. I n spite of this, isosteric heats of adsorption, cslculated from data collected a t 14.0 and 27.0" for the adsorption process, not only agree well with the heats of adsorption derived from heat of immersion studies,* but they further agree with the proposal that there is a discontinuity in the adsorption process. Figure 4 summarizes the isosteric heats for sodium and calcium bentonite, with points A and B corresponding t o the discontinuities A and B in Figs. 1 and 2 for sodium and calcium bentsonitelrespectively. Table I lists the isosteric heats a t the two V , values obtained from BET plots. These conclusions are further substantiated by the fact that similar calculations on the desorption isotherms give almost the same results. TABLE I ISOSTERIC HEATSOF ADSORPTION CORRESPONDING TO THE Two VALUESOF Vm DERIVEDFROM BET PLOTS, IN KCAL. PER MOLE OF ADSORBED WATER,AND BET AREASWITH ARGON Cation type clay
Li Na (oven-dried) Na (freeze-dried)
K H Mg Ca Ba
Sn Zn Cd Cn Fe
Ni
AX1
14.0 12.5 11.4 11.1 12.9 15.6 14.6 14.9 13.8 15.1 13.0 16.4 12.8 9.8
AH8
.. 10.0 10.6 10.0 11.3 12.5 11.0 12.6 12.4 11.6 11.9 13.5
.. ..
Area, m.l/g.
22.4 10.3 13.9 29.6
Vol. 63
ported that the chromic acid oxidation of isopropyl alcohol is slightly but significantly catalyzed by pyridine. The reaction was carried out in aqueous solution at 25" with a pyridine-pyridinium ion buffer and about 0.005 M (H+). Recently2RoEek and KrupiEka have repeated this work and find that the catalysis by pyridine is less than that previously' reported, and in fact is too small to be of importance. When we use the exact experimental conditions described by the respective authors, we can (at least approximately) repeat both sets of data. However, a reconsideration of these experimental conditions shows that, in conformity with RoEek and KrupiEka's claim, the catalysis by pyridine is very small. Roc'ek and KrupiEka added crystalline pyridinium perchlorate to their solution of alcohol, perchloric acid and chromic acid. We have confirmed that their procedure is satisfactory. I n the earlier experiments, * however, the acidities of the solutions containing sodium perchlorate on the one hand and pyridinium perchlorate on the other were matched by adjusting the solutions until a spectrophotometric determination of indicator color was the same for the two solutions. This method involved the assumption that, a t constant ionic strength, the effects of the two salts (sodium and pyridinium perchlorates) on the activity coefficients of the dye were the same. Reinvestigation has shown that this assumption is not correct. Experimental
Materials.-Thymolsulfonphthalein (Eastman Kodak) of m.p. 212-214' (with dec.) showed only a single spot in
paper chromatography.3
4-Chloro-2,6-dinitrophenol4 was
TABLE I SALTEFFECT O N THYMOLSULFONPHTHALEIN molar, T = 25". Indicator, 3.45 x HC104, M
pH0.b
ds~rnp'
Indicator alone 0 00473 2.36 0.227 NaClO4, 0.195 M ,00473 2.35 .224 PyHC104, 0.195 M .00473 2.33 .192 PgHClO4, 0.195M .00608 2.22 .223 N-MePyC104, 0.195M .00473 2.35 .186 N-MePyC104, 0.195 M .00629 2.24 ,224 Measured with a, a Average of three measurements. glass electrode and a Beckman model G pH meter.
TABLE I1 EFFECT ON 4CHLORO-2,6-DINITROPHENOL molar, T = 25' Indicator, 5.22 X
SALT
HClO4, M
21.0 33.6 26.4 35.0
PYRIDINE CATALYSIS OF THE OXIDATION OF ISOPROPYL ALCOHOL BY CHROMIC ACID BY F. H. WESTHEIMERA N D Y.W. CHANG Contribution from the Mallinckrodt Labo~atoriesof Halvard Universe'tu Cambridge, Maes. Received September 0, 1068
In 1951, Holloway, Cohen and Westheimer' re-
pH'*b
drrarnr"
0,00473 2.36 0.081 .00473 2.36 .086 .00473 2.35 .lo4 .00625 2.19 .086 N-MePyC104, 0.195 M ,00473 2.37 .I08 ,00702 2.18 .087 Measured with a Average of three measurements. glass electrode and a Beckman model G pH meter. Indicator alone NaC104, 0.195 M PyHC104, 0.195 M
purified by recrystallization of the potassium salt, and then Merck reagent grade perof the phenol; m.p. 81-82'. ( 1 ) F. Holloway, M. Cohen and F. H. Westheimer, J. Am. Chrm. Soe., 78, 65 (1951). (2) J. R o h k and J. Krupifka, Coll. Czech. Chem. Communicaliona,
in press: J. Rofek, private oommunioation. (3) G. T. Franglen, Nature, 176, 144 (1955). (4) M. Dubois, 2. Chem., 10, 206 (1867).
?
e
