ADSORPTION IN MORDENITE features of the adsorption potential surfaces within Hmordenite, would be as accurately predicted by use of the simple Lennard-Jones potential, by summations including lattice atoms up t o about only 10 distance from the adsorbed molecule, and that inclusion of only nearest neighbors in summations may for many purposes be adequate. Although these approaches differ vastly in complexity and required computer time, their predictions are far more coincident, at least in
1647 the present instances, than those of procedures using the alternate formulas for interaction parameters. It is concluded that the use of the 6-12 potential in summations extending only moderately beyond nearest lattice atom neighbors, with scaling between Kirkwood-Muller and London parameters, and perhaps beyond, left as an empirical degree of frcedom, offers the most expedient means of seeking generalities characteristic of adsorption in zeolites.
Adsorption in Mordenite. 11. Gas Chromatographic Measurement of Limiting Heats of Adsorption of Nonpolar Molecules by Guillermo D. Mayorga and Donald L. Peterson" California State College, Hayward, Hayuard, California 94648 (Received October 18, 1971) Publication costs assisted by California State College, Hayward
A comparison between heats of adsorption derived from chromatographic retention-volume measurements and values predicted on the basis of adsorption potential calculations shows the latter to be quite successful in the case of atomic adsorbates when the interaction parameters are evaluated by the London formula. The poor agreement in the case of nonpolar molecular adsorbates reflects an inadequacy of treating the molecules as spherical, possibly even in the case of methane.
The availability of contour maps of adsorption potentials of a series of nonpolar molecules in H-mordenitel makes possible the prediction of their heats, free energies, and entropies of adsorption. For comparisons of the predicted heats, available experimental data2 were augmented by a series of chromatographic measurements of adsorbabilities of these compounds at several temperatures, from which heats of adsorption at vanishing coverage were estimated. A further purpose of the reported measurements is to augment a prior demonstration* of the utility of mordenite as a gas chromatographic packing material. The unidimensionality of mordenite's pore structure raises some doubts about whether equilibrium will be reached, and the operating assumption of chromatography realized, on the time scale of practicable chromatographic measurements.
Experimental Section A Varian Aerograph Model 90-P3 chromatograph, which comes equipped with a flow control system that operates under 65 psi pressure of carrier gas, was modified with the aid of a Moore Flow Control and a precision Whitey needle valve t o allow use of a lower
pressure (10 psi or less), and precise control of the flow rate. A modified 0.25411. Swagelock tee was attached to the injection port with one of the free sides of the tee serving as the new injection port. The free side of the tee was connected through a valve to a mercury manometer. The inlet pressure was read to within 0.005 cm with the aid of a cathetometer. The carrier gas was helium; this and the argon, methane, ethane, propane, and n-butane used in the experiments mere instrument grade gases from Matheson Co. The neon and krypton were high-purity grade gases from the General Electric and Air Reduction Go., respectively. The H-mordenite was produced by exchanging Na-mordenite (Zeolon Lot HB-gE, Code SR4710) thrice with 1 M "*NO3 over a period of 3 days. Between exchanges the sample was washed with deionized water. After the final exchange the (1) G . D. Mayorga and D. L. Peterson, J . Phys. Chem., 76, 1641
(1972). (2) R. M.Barrer and D. L. Peterson, Proc. Roy. Soc., Ser. A , 280, 466 (1964). (3) K. Torii, M. Asaka, and H. Yamazaki, Kogyo Kagaku Zasshi, 72, 661,664 (1969).
The Journal of Physical Chemistry, Vol. 76, No. 11, 1078
1648
G. D. MAYORGA AND D. L. PETERSON
sample was washed until no sodium could be detected in the effluent. Retention of crystallinity was verified by X-ray powder photography. Flame photometric analysis showed the product to contain 0.02% wt sodium or about one sodium atom in each 35 unit cells. The column packing was prepared by pelletizing the ammonium mordenite and then grinding and screening the pellets to obtain a 40-60 mesh fraction. This was packed in a column 50.5 cm in length and 0.47 cm in internal diameter. Dehydrated mordenite in the hydrogen form was produced by passing helium through the column at 350" and the nominal flow rate for 24 hr; in this form the packing weighed 4.73 g. Retention volumes were measured at temperatures between 30 and 350" using a helium flow rate of ca. 30 ml/min, under which conditions the number of theoretical plates was between 80 and 300. The ratio of the column inlet pressure, Pi, to that at the outlet, Po,varied between 1.2 and 1.4. The flow rate, to, was measured at the column exit pressure and a t room temperature, T,, with the aid of a conventional soap-bubble flowmeter in which the vapor pressure of water was P,. The corrected rctention volume, VR', referred to dry carrier gas at the column temperature, T,, is given by
in which tR and tv are the observed retention times for the compound studied and for a nonadsorbed component, respectively, and
is the correction factor of James and Martin4 for the variable pressure within the column. Where c is the molar concentration of a solute in the carrier gas, q is the number of moles of solute per unit mass of adsorbent, and m is the mass of the adsorbent in the column, the fundamental relationship of gas chromatography is VR" = -q m C
Adsorption coefficients are more commonly expressed as volume of gas at STP adsorbed per gram adsorbentatmosphere pressure, usp, than as q/c. For an ideal gas at standard temperature and pressure, us,
cc(STP) . ( , t m ) Tom
= 273*16 -
Independence of v,, on carrier gas flow rate was shown in the instance of propane at 276'. Successive reductions of flow rate of as much as 40% of the nominal rate produced random variations in vsp not exceeding 3%. The Journal of Physical Chemistry, Vol. 76, hro. 11, 1978
j
C2H6
P
12
:I
1.
.4
-,f
-.4
Figure 1. Variation of the specific retention volume with t,he column temperature for argon, krypton, and n-alkanes.
Results and Discussion Logarithmic plots of observed values of us, are shown in Figure 1. Values of the limiting heat of adsorption
are contained in Table I. In the two instances where a comparison with other measurements is possible, the agreement is viewed as satisfactory, particularly when the lower Na-ion content of the present sample is taken into account. A comparison with predictions based on adsorption potential calculations was made with the aid of statistical considerations which lead,j for the present cases, to
AH"
= pmin
RT
+T
Table I also contains predicted values of AH" found from the minimum values of adsorption potential surfaces computed with the aid of two alternate formulas, those of Kirkwood and Muller6 and of London,? for interaction parameters. The calculations treat the paraffin moIecules as spherical. Thus, the inclusion of ethane, and more particularly, of propane and n-butane, serves mainly to illustrate the magnitude of errors introduced by this assumption. The calculated values of AHo shown in Table I utilize the Lennard-Jones 6-12 potential. I n the case of summations using the Kirkwood-Muller formula, values computed by the 6-8-10-12 potential exceeded those of the 6-12 by about 10%. Inclusion of the 8-10 terms in the London (4) A. T. James and A . P. Martin, Biochem. J., 50, 679 (1952). (5) T. L. Hill, J . Chem. Phys., 17, 520 (1949). (6) J. G. Kirkwood, Phys. Z.,33, 57 (1932), and H. R. Muller, Proc. Roy. SOC.,Ser. A, 154, 624 (1936). (7) F. London, 2.Phys. Chem. ( L e i p t i g ) , B11, 222 (1930).
ADSORPTION I N MORDENITE
1649
summation may be expected to produce a similar effect. This would bring the predicted heats using this formula very nearly into quantitative agreement with the observed values for the rare gases. That the value predicted by this method for methane, even allowing for the effect of these terms, is smaller than observed may be an indication that even methane is inadequately treated as spherical for these purposes.
-12-
2z :
-
JO-
2
:
-I
q
-
-8-
z
-6:
8 ! -
Table I : Comparison of Observed and Predicted Limiting Heats of Adsorption in H-Mordenite
Argon Krypton Methane Ethane Propane n-But ane
b
A H o , kcsl/mol
7
Moleoule
8 :
y------Cdod---KirkwoodMuller 6-12
-----
I-
Obsd----
London 6-12
3.7 4.9 4.4 6.4 3.1 2.7
8.4 11.6 6.4 11.3 6.4 6.2
Present work
4.2& 0.3 5 . 1 d= 0 . 2 6 . 1 & 0.5 8.0 d= 0 . 3 10.0 f 0 . 7 12.0 d= 0 . 3
Ref 2
+ 2.0n kcal/mol
in the hydrogen form. Limiting heats of light normal paraffins on seolite Ca-X, for example, follow approximately a very tsimilar linear relationn - A H o = 4.5
+ 2.2n kcal/mol
The two caseu share the interesting feature of a large (negative) A H o intercept, a property not shown, a t least in the case of Ca-X,l0 when heats at appreciable coverages are used. Limiting heats for light paraffins on neolite Ca-A provide another example, for which the reported vduesl1 are approximated by -AHo
=
1.7
+ 2.3n kcal/mol
While more examples are needed to make clear the influence of channel structure and cation composition on the parameters of these relations, it would not be too surprising to find that the slope is a relatively invariant quantity. Thus, the predicted values for so different an adsorbent as graphite (with which observed values are in good agreement) are given by a line of quite similar slope, albeit much smaller intercept 1 2
-AHo = 0.85
+ 1.88%kcal/mol
--
B -*--
-
-
4.6 5.5
Figure 2 shows the observed linear dependence of limiting heat of adsorption for the normal paraffins on their carbon number n
-.AHo = 4.0
-9-
0
"
I
I
I
I
2
3
4
Thus, limiting heats of adsorption of at least light n-alkanes in mordenite appear not to be unique: the discontinuity predicted by the assumption of spherical shapes is, as expected, wholly artificial. However, it
Conclusion It appears that the use of the London formula in extended summations of interactions with lattice atoms, and of a polarizability for the oxygen atoms of the lattice2 which reflects their actual state of chemical combination, leads to eminently satisfactory predictions of the heats of adsorption of atomic adsorbates. The agreement in these instances is even improved by the inclusion of the inverse eighth and tenth terms in the dispersion potential. If this success can be taken seriously, then the need of accounting for molecular geometry, even in the case of spherical tops like methanc is at once apparent. The predictions even of limiting adsorption coeffi(8) P. E. Eberly, J . Phys. Chem., 67, 2404 (1963). (9) H.W.Habgood, Can. J . Chem., 42, 2340 (1964). (10) R. M. Barrer and J. W.Sutherland, Proc. Roy. Soc., Ser. A , 237, 439 (1956). (11) A. V. Kiselev, Y . V. Khrapova, and K. D. Shcherbakova, Petrol. Chem. USSR, 2 , 558 (1963). (12) N.N.Avgul, A. A. Isirikyan, A . V. Kiselev, I. A. Lygina, and D. P. Poshkus, Bull. Acad. Sci. USSR, Div. Chem. Sci. (Eng. Transl.), 11, 1334 (1957). The Journal of Physical Chemistry, Vol. 76, No. 11, 1072
RAFIK0. LOUTFY AND RAOUF 0. LOUTFY
1650 cients, since they depend exponentially on the adsorption potentials, seems a futile task indeed until such time as adsorption heats can be predicted with more
confidence. It is hoped that the extension of the sorts of comparisons made here to other zeolites and adsorbates will make this possible.
Correlations between the Electrochemical and Spectroscopic Behavior of Some Benzophenones and Thiobenzophenonesl by Rafik 0. Loutfy” and Raouf 0. Loutfy University of Western Ontario, Photochemistry Unit,London 79, Ontario, Canada
(Received September 87, 1971)
Publication costs assisted by the Photochemistry Unit, University of Western Ontario
The electrochemistryof a number of benzophenones and the corresponding thioketones have been investigated, using ac polarographic technique. The half-wave potentials (E112)s)were found to be linearly related to the n, T* triplet energies. This relation enabled an estimate to be made of the position of the n, T* triplet energies of those compounds where n, T * triplet states were not observed spectroscopically. A linear dependence between the half-wave potentials of the carbonyl and the corresponding thiocarbonyl has also been observed. The usual Hammett correlation was obtained for both series of compounds. The heterogeneous rate constants for several of these compounds have been determined by Randel’s method of analyzing the ac data. A reversible (Kappg > 1 x 10-2 cm/sec) one-electron reduction was found for these compounds. The heterogeneous rate constants have been compared with the hyperfine splitting constants and a fair prediction of the variation in the experimental rate constant with structure was found.
Introduction It has been well established that polarographic half-wave potentials are linearly related to the calculated energy of the lowest vacant molecular orbital, both for the unsaturated2and aroma ti^^,^ hydrocarbons. This linear dependence has also been extended to the energy of the lowest excited statea6 The electrochemical reduction of benzophenones, in aqueous media, has been studied by Zuman, et a l e The substituent effects on E,,,’s were determined using Hammett plots. Recently the electrolytic reduction of a number of derivatives of benzophenone037 and thiobenzophenones has been studied in aprotic solvents. It has been shown that these compoundsO-sundergo a one-electron reduction to form a stable anion radical in dimethylformamide and acetonitrile. However, there have been no kinetic studies reported of the electron transfer for these compounds. I n the present paper, the results of a kinetic study of the electroreduction of benzophenone, 4,4’-dimethoxybenzophenone, and bis(dimethy1amino) benzophenone and the corresponding thio ketones are reported. A correlation between the n, n* triplet energies and the El/,’sare presented and discussed. The Journal of Physical Chemistry, Vol. Y6,No. 11, 19YB
Experimental Section All experiments were carried out in acetonitrile (AN) containing 0.05 M tetraethylammonium perchlorate (TEAP) . The solvent purification procedure is described in detail in ref 9. The accessible potential range of the solvent was +2.7 to -2.1 V on Pt and +0.6 to -2.8 V on dropping mercury electrode vs. (Ag~AgCl~O.1 M TEAP in AN/10.05 M TEAP in AN). TEAP was recrystallized from hot water and dried for (1) Publication No. 26 from the Photochemistry Unit. (2) A. Maccok, Nature (London), 163, 178 (1949). (3) (a) A. Streitwieser, Jr., “Molecular Orbital Theory for Organic Chemists,” Wiley, New York, N . Y., 1961, pp 173-185; (b) M. E. Peover in “Electroanalytic Chemistry,” Vol. 11, A. J. Bard, Ed., Marcel Dekker, New York, N. Y., 1967, Chapter 1. (4) H. B. Mark, Jr., Rec. Chem. Progr., 29, 217 (1968). (5) (a) E . 8. Pysh and N. C. Yang, J. Amer. Chem. Soc., 8 5 , 2124 (1963); (b) A. Maaaenga, D . Lomnitz, J. Villegas, and C. J. Palowczyk, Tetrahedron Lett., 21, 1665 (1969). (6) P. Zuman, 0. Exner, R. F. Rekber, and W. Nauta, Collect. Czech. Chem. Commun., 3 3 , 3213 (1968). (7) J . M. Saveant and L. Nadjo, J. Electroanal. Chem., 30, 41 (1971). (8) (a) L. Lunazai, G. Maccagnani, G. Mazzanti, and G. Plaucci, J. Chem. Soc. B , 162 (1971); (b) R . M . Elofson, F . F . Gadallah, and L. A. Gadallah, Can. J. Chem., 47,3979 (1969). (9) W. R. Fawcett, P. A. Forte, R . 0. Loutfy, and J. M. Prokipcsk, ibid., 50, 263 (1972).
