223 The Effect of Pressure on Reactions in Solution I—Sodium Ethoxide and Ethyl Iodide to 3000 kg/cm2 II—Pyridine and Ethyl Iodide to 8500 kg/cm2 By R. O. Gibson, E. W. Fawcett, and M. W. Perrin (Communicated by C. N. Hinshelwood, F.R.S.—Received February 6, 1935) Introduction Previous work has shown that pressure has an accelerating effect on a large number of chemical reactions in the liquid phase.* The rates of the reaction between pyridine and cetyl halides have been studied at different temperatures and pressures,f but it was not found possible to interpret the results by means of the ordinary reaction velocity equations, and no conclusions could be drawn as to the mechanism of the pressure effect. In making a further attempt to determine this mechanism, it seemed desirable to take reactions which have already been studied at atmospheric pressure and are known to follow the simple velocity equations, and to investigate the effect of pressure on the velocity con stants. The effect of temperature on the velocity constant of a reaction k, can be expressed by the Arrhenius expression k = Ae-E/RT.
224 R. O. Gibson, E. W. Fawcett, and M. W. Perrin types of reaction, and, if so, where this difference lies, reactions have been chosen for study which are typical of the two classes, namely:— (1) C2H5ONa + C2H5I -> C2H5OC2H5 + Nal and / i (2) C5H5N + C2H5I ->■ C5H5N ' x c 2H5 The first reaction is best carried out in ethyl alcohol, and the original plan was to study the second in the same solvent. It was found, however, that alcohol was not a suitable solvent for the second reaction and it has therefore been studied in detail in acetone. As one of the characteristics of reactions of the slow type is the pronounced effect of the nature of the solvent on the rate of reaction, an attempt was made to study the second reaction in hexane, but this solvent was also found not to be very suitable.
On Reactions in Solution 225 tap of the mercury reservoir, which was placed in the position shown. When a sample had been taken, the lead A was removed for cleaning and the top of the vessel closed by the ground joint, B. The glass vessel was placed in a thermostat controlled at the desired temperature to 0 • 10 C. Experiments above the boiling point of the solvent were carried out by sealing portions of the solution in a number of glass tubes, which were placed in the thermostat and taken out for analysis at intervals. As the B.
226 R. O. Gibson, E. W. Fawcett, and M. W. Perrin over mercury in a glass bell, A, of about 120 cm3 capacity, which was attached by means of a packing gland to the high-pressure side of a valve, V, built into the head of the steel pressure vessel as shown. Pressure was transmitted to the reaction mixture by a light lubricating oil through the lead, P, which was con nected to a Cailletet-type screw press, f n V/\ and measured with a Bourdon gauge, frequently checked against a calibrated ■j pressure balance. Before starting an ex 1 1 periment, the reaction mixture was placed in a tube in the glass cylinder, C, and forced into the bell, A, through the nickel tube, B, by compressed air applied at D. The valve, V, was then closed and the oil pressure applied . The second lead of the valve, V, was normally closed by a valve (not shown), and this lead enabled the tube B to be cleaned after use. When required, samples were drawn off through B, by opening the valve Y, during which time the pressure in the vessel was main tained by means of the screw press. When the reaction was studied at temperatures above the boiling-point of the solvent at atmospheric pressure, a cylinder was placed inside the tube C for the sample, and air pressure applied at D, in order to prevent evaporation in the leads. The pressure vessel was placed in a thermostat controlled to 0-1° at the required tem perature.
On Reactions in Solution 227 enclosed in a small glass tube inverted over mercury in a glass or steel liner, which fitted into the reaction vessel, R, as shown. The remaining space in R and the high-pressure cylinder, C, were filled with a light lubricating oil. An oil pressure, p ,was applied through the lead, L, and intensified by the piston B, to a high-pressure P, in C. The packing for the high-pressure end of the piston was a rubber bung, D,* which was renewed every second or third run. The high-pressure P was estimated in two ways. In the first a Bourdon gauge, reading to 4000 kg/cm2, and calibrated to 2500 kg/cm2, was connected to C, and the pressure readings compared with those of a calibrated gauge connected to the low-pressure side, L. A linear relationship, reproducible with different bungs, was obtained between these readings for rising pressure and this was extrapolated to the pres sures used. In the second method the pressure P was calculated from the ratio of the areas of the piston, the pressure on the low-pressure gauge and, on lowering the pressure, the reading at which the piston began to move, as indicated by the tell-tale E viewed through a catheto- meter. For a rising pressure, pA — — F, where A and a are the areas of the low- and high-pressure sides of the piston respectively, and F the frictional force. If p'is the value on lowering the pressure at which the tell-tale is observed to move, then A = -f F, whence p = JLj-JL — 2 ' With p = 500 kg/cm2, the first method gave a value of P — 8450, and the second P = 8420 kg/cm2. The reaction vessel, R, was totally im mersed in a thermostat controlled to 0*1°. The reactants were left for a known time under pressure and then removed for analysis.
On Reactions in Solution 229 ethoxide and ethyl iodide solutions, adding, if necessary, a known volume of alcohol to give the initial concentrations required. About 100 cc of the mixture were introduced into the high-pressure apparatus, fig. 2, as described above, and a further 100 cc of the same mixture put into a stoppered flask, in the same thermostat as the pressure apparatus, to serve as the control experiment at atmospheric pressure. The first samples were taken after 30-60 minutes and the following samples at suitable intervals. The reaction was followed by pipetting 10 cc from the samples as soon as possible after sampling (the temperature being assumed that of the thermostat, 15-30°) into 25 cc of standard HC1 and titrating the excess acid with NaOH, using phenolphthalein as an indicator. The quantity of ethoxide remaining in the solution was obtained as described above for the standardization of the ethoxide solutions.
230 R. O. Gibson, E. W. Fawcett, and M. W. Perrin ON ^D (<4 O o>on O^N ft O O cN oo m O oo oo ctf j s i « N+ MI I O o+ C+<N 1-HI £ O w c3 £+lO | roimn- d(rovN^-o Mo,r^ oh o((oNN ff(iM o o o o o o o uo 6 2 •Bo is, VO mIT) ON Tf ^ ON VO <1M0 W) •o « u, n© ov o ^ in-H i-(nh | CNS h<N O O O O o © © ut ©oO<NoN * goo«s o o o o o o o T3 <L> § p a 3 3 £ K X j o O —' —< ^ ^ <M 0 o o o o o o o U 53 + O o © b o o o o o u C 3 I < A e 5 £ VO in in ^ ^ ^ O. ft ♦1 p3H 8 -2 q,«ft "d G X z O Oc j W W wo Wo o o o U HoI—I cj o U<M rEH z 3 op 0O0 1 ^ cn vo h oo *n rs| uc bsalc % T+—1 y—l l Ol | O1 c+o N1 w Oc ( P4 W 5H3 r<N- v(No V^O t O On Ph CooN oo toO Qo T©-H m© O a 6 © © b o o mH CJ Z Mb£> <H Oco On VOO chn n no m T-H t ^ co Zo 1 NO TOf tO ©o ,o"H or-H O O © U © b o b o o >H H U z S OwiJ «1i—3; c<O3 *o£00o 53. OOEX) O"oT °oOnn 9oOnN ooONo CooToV > *3 38 !^= tc •§ b o b o o o U E w 3 I-) G OioO ON 0ir0> Oio OO O^ HC cC<*Dn> Tf ON <N (N m© m© £ 1—< *—H.
On Reactions in Solution 231 shows the effect of temperature on the velocity constants at the two pres sures with an initial concentration of sodium ethoxide of approximately 0-1 mols/litre. The actual initial concentrations, corrected to the temperatures and pressures of the experiments, are given in columns 2 and 6. The values of the velocity constants at 1 and 2980 kg/cm2 are given in terms of 1 litre at t° and 1 kg/cm2 per gm mol/minute in columns 3 and 7, whilst column 8 gives the values at 2980 kg/cm2 corrected for the com pressibility of the solvent. The ratios of these corrected values to the corresponding values of the control experiments at 1 kg/cm2 are given in column 11.
232 R. O. Gibson, E. W. Fawcett, and M. W. Perrin The values of the constants A and E of the Arrhenius expression for the variation of k with temperature have been calculated from the results given in columns 3 and 8 by the method of least squares and the following expressions obtained:— at X kg/cm2 k= 1 -28 x 1013 <?-20>74°/RT and at 2980 kg/cm2 kv = 2-23 X 1013 where R = 1 -985 cals/gm mol, and T is the absolute temperature. The values of k calculated from these expressions are given in columns 4 and 9 of Table I, together with the percentage deviations from the observed values, in order to show the accuracy of the experiments. It is seen that the “ activation energy,” E, is practically unchanged by pressure, and that the pressure effect seems entirely due to an increase in the “ collision term,” A.