J . Chem. SOC., Faraday Trans. I , 1982, 78, 3379-3382 Catalytic Deamination by Reversed-flow Gas Chromatography BY MICHAEL KOTINOPOULOS, GEORGE KARAISKAKIS AND NICHOLAS A. KATSANOS* Physical Chemistry Laboratory, University of Patras, Patras, Greece Received 22nd March, 1982 Rate constants and activation parameters for the deamination of 1 -aminopropane, 2-aminopropane and aminocyclohexane over 13X molecular sieve have been determined using the technique of reversed-flow gas chromatography. In aminocyclohexane the fraction of the surface which is catalytically active has been estimated, and this is found to increase with increasing temperature. Reversed-flow gas chromatography (r.f.g.c.) is a new differential method for studying the detailed kinetics of surface catalysed reactionsly and other slow physical processes, such as diffu~ion.~ The method has been used successfully to study the dehydration of alcohols over 13X molecular sieve and y-aluminium ~ x i d e .~ The activation parameters determined were found to agree with those determined by other techniques. Moreover, the fraction of catalytically active surface sites was determined and found equal to the fractional conversion of the reactant to products. In the present paper we used the r.f.g.c. technique to study another class of organic reactions, namely catalytic deaminations to alkenes. The catalyst was molecular sieve 1 3X and the reactants 1 -aminopropane, 2-aminopropane and aminocyclohexane, to form unsaturated hydrocarbons as main products. The r.f.g.c. method is very simple. It uses a conventional gas chromatograph and a column consisting of two lengths I' and 1 in series containing the catalyst.The reactant is introduced as a pulse between these two lengths, and after a certain time the direction of the carrier-gas flow is repeatedly reversed. This gives rise to extra peaks in the chromatographic trace, whose height or area under the curve depends on the exact time of each flow reversal. The analytical mathematical form of this dependence, which is characteristic of the mechanistic scheme of the reaction, permits the calculation of rate constants from the experimental data. EXPERIMENTAL MATERIALS Molecular sieve 13X, 80-100 mesh (Applied Science Laboratories) was used as the catalytic surface. 1 -aminopropane (puriss grade), 2-aminopropane (puriss grade), aminocyclohexane (purum grade) and cyclohexene (puriss grade) were obtained from Fluka AG.Propene was from Matheson (G. P. grade, 99.7% purity) and benzene was from Merck AG (Uvasol). With 1 -aminopropane as reactant nitrogen was used as carrier gas, while with 2-aminopropane and aminocyclohexane the carrier gas was helium. Both gases were products of Linde, Athens, (99.99% purity). 33793380 CATALYTIC DEAMINATIONS APPARATUS AND PROCEDURE The experimental set-up and the procedure followed in the r.f.g.c. method have already been reported.2 The lengths l' + I of the chromatographic column (glass, I.D. 4 mm) containing the catalyst were 3 + 108, 0.5 + 8 and 1.3 +46 cm for reactants 1-aminopropane, 2-aminopropane and aminocyclohexane, respectively. The conditioning of the columns was conducted in situ by holding them at 693 K for 19 h, under carrier-gas flow (0.76 cm3 s-l, corrected at column temperature). During the kinetic runs the flow rate was in the region 0.65-0.87 cm3 s-l.For each amine, 1 mm3 of liquid, using a microsyringe, was injected onto the column length 1 through an injector placed at the junction of the two columns, and with the carrier gas flowing in direction F (forward). Plots and calculations were made on a Hewlett-Packard 9825 A desk-top computer connected to a 9872 B plotter. RESULTS AND DISCUSSION The main products of the deamination reactions studied were identified experi- mentally by comparing their retention times with those of pure substances, under the same conditions.These main products were propene in the deaminations of 1-aminopropane and 2-aminopropane, and cyclohexene in the case of aminocyclohexane. The kinetics of the deamination reactions studied conform to a simple first-order decomposition of the adsorbed amine to give the adsorbed product(s). Therefore, the same equation as for the dehydration of alcohols4 describes the kinetic law f = W b P (ktR) - 11 exp (- k4,J (1) wherefis the area under peaks F and R (where F indicates forward and R reversed with respect to the direction of the gas flow), rn the mass of amine injected and g its 5 - 4 Y .- c 0) .3 Y - 2 3 s E - 2 1 0 1 2 3 4 t t o t l 1 0 3 s FIG. 1.-Plots of eqn ( 1 ) for the deamination of aminocyclohexane to cyclohexene over 13X molecular sieve at 597 K: 0, R peaks; A, F peaks.M. KOTINOPOULOS, G.KARAISKAKIS AND N. C. KATSANOS 338 1 TABLE RATE CONSTANTS FROM R PEAKS (kR) AND FROM F PEAKS (kF) FOR THE The fvalues in the rate constants are standard errors. DEAMINATION OF THREE AMINES OVER 13x MOLECULAR SIEVE AT VARIOUS TEMPERATURES 1 -aminopropane 624 644 65 1 657 2-aminopropane 580 592 61 1 623 aminocyclohexane 574 586 597 604 626 3.9 f 0.6 5.8 f0.5 7 f l 7.8 f 0.3 1.54f0.02 3.0 fO.l 7.0f0.1 12.4 f 0.3 2.3 f 0.3 3.3 f0.3 7.3 f 0.3 10.9k0.8 35f1 4.3 +0.3 6.6f0.3 8.0 f0.5 7.9 f 0.2 1.64 f 0.04 3.0k0.2 6.73 f 0.06 12.1 f0.2 2.22 & 0.08 3.7 f0.3 7.4 f 0.3 10.3 k0.5 32+ 1 - 10 - 14 - 14 - 1.3 - 6.5 0.0 3.9 2.4 3.5 -1.4 5.5 8.6 - 12 TABLE 2.-ACTIVATION ENERGIES (E,) AND ENTROPIES (AS’) FOR THE DEAMINATION OF THREE AMINES OVER 13x MOLECULAR SIEVE EJkJ mol-l -ASz/J K-l mol-l amine R peaks F peaks R peaks F peaks 1 -aminopropane 71 f 2 66f9 212f3 218f 17 2-aminopropane 144f3 138f3 84+6 94&6 aminocyclohexane 162 f 1 1 156f5 51 f 17 59f8 fraction held on reactive surface sites, k the rate constant, ttot the total length of time until the last reversal of the gas flow, and t , the retention time of the product in the direction opposite to the flow.An example of plotting lnfagainst ttot, according to eqn(l), is shown in fig. 1 for both R and F peaks. The deviation from linearity in all plots was small, as judged from t-tests of significance on the relevant coefficients of regression. The probability of the various t values being exceeded was < 1 % in all cases.From the slopes of the above-mentioned plots the values of k were calculated and are collected in table 1. They are denoted as k , (from R peaks) or kF (from F peaks). The last column of this table gives the percentage difference between the two rate constants, showing a fairly good agreement between the k , and k , values in most cases, as predicted by the theory. Some relatively large differencies may be due to accidental errors. Eqn (1) has been derived2 on the basis that the mechanistic scheme of the reaction is fast k fast A+S $ A-S -+ D-S + D+S3382 CATALYTIC DEAMINATIONS where A is the gaseous reactant, D the gaseous product, and A-S and D-S the respective adsorbed species on active centres S. Deamination of aminocyclohexane over y-A120, at 525 K has been shown2 to proceed via the formation of an adsorbed intermediate B-S between A-S and D-S.The fact that the present results conform to a simple first-order law is most probably due to very fast formation or disappearance of the intermediate B-S, caused either by the different catalytic surface and/or by the higher temperatures used here. The energies and entropies of activation for the deamination reactions, calculated from conventional Arrhenius plots, are given in table 2. The differences between activation parameters determined from R peaks and from F peaks lie Yjithin the limits of experimental error, showing that secondary reactions of the detected product or irreversible adsorption of it are negligible. This is because the lengths Z and I' of the column responsible for the R and F peaks, respectively, are very different (by a factor of 16-36, cf.the Experimental section). No comparison of our results with literature values can be made, since to the best of our knowledge no deaminations over zeolites have been studied previously. As in the dehydration of alcohol^,^ one can find the absolute value for the pre-exponential factor in eqn (I), mg[exp (ktR)- 13, if the response of the detecting system is known. Then, knowing m and tR from experiment, and k from the gradient of eqn (l), the fraction g can be calculated. The mean values determined from F and R peaks for the deamination of aminocyclohexane are 0.24, 0.34, 0.46 and 0.76 at temperatures 574, 586, 604 and 626 K, respectively. In aminopropanes the absolute value of g could not be calculated, because the response of the flame ionization detector to the product propene could not be found accurately. It was possible, however, to ascertain that the g value in this case does not change significantly with temperature, as in the case of aminocyclohexane above.It was shown previously4 that the g values give the fractional conversion of the reactant to product. Thus, an increasing conversion of aminocyclohexane to cyclohexene with increasing working temperature is observed here. This is not due to irreversible adsorption of the product on the solid catalyst, since the g values calculated from the F and R peaks are not significantly different, in spite of the fact that these two types of peaks are due to two different column lengths (Z' and I ) containing very different amounts of catalyst. It is hoped that the present paper, together with previous studies1* 2 * will help to introduce workers in various fields of heterogeneous catalysis to the technique of reversed-flow gas chromatography, which is a new tool for studying the kinetics of surface reactions. We thank Mrs Margaret Barkoula for assistance. N. A. Katsanos and I. Georgiadou, J. Chem. SOC., Chem. Commun., 1980, 242 and 640. N. A. Katsanos, J. Chem. SOC., Faraday Trans. I , 1982, 78, 1051. N. A. Katsanos and G. Karaiskakis, J. Chromatogr., 1982, 237, I . G. Karaiskakis, N. A. Katsanos, 1. Georgiadou and A. Lycoughiotis, J. Chem. SOC., Faraday Trans. I , 1982, 78, 2017. (PAPER 2/496)