摘要:

Solvent Effects in the Kinetics of Fast Proton-transfer Reactions : Diffusion, Preliminary Complexes, and Steric Factors BY GRAHAM D. BURFOOT, EDWARD F. CALDIN* AND HOWARD GOODMAN University Chemical Laboratory, Canterbury, Kent Received 27th June, 1973 The kinetics of the reaction of tri-n-butylamine with the substituted phenol tetrabromophenol- phthalein ethyl ester (Bromophthalein Magenta E) in various solvents have been investigated, by a microwave temperaturejump method. The rate constants are inversely correlated with the viscos- ities of the solvents, and are 15 to 30 times less than the values calculated for dXusion control, Bromophthalein Magenta E shows close similarities with other substituted phenols (e.g. 2.4-dinitro- phenol and Bromophenol Blue) in its reactions with tertiary aliphatic amines.All the data on these reactions are satisfactorily correlated by a theory which envisages diffusion-controlled formation of weakly-bound complexes, during whose life-time the reactant molecules can rotate into positions where the active centres are adjacent so that reaction can proceed. Alternative models are con- sidered and found unsatisfactory. In an earlier paper results were reported for the kinetics and thermodynamics of the reactions of tertiary amines with tetrabromophenolphthalein ethyl ester (Bromophthalein Magenta E) in chlorobenzene, and it was concluded that the rate- determining step is not the proton-transfer but is probably the formation of a hydrogen- bonded complex, according to the scheme represented in eqn (1) : AH+B+AH... B+A-. . . HB+ where A H and B represent the free acid (I) and base molecules respectively, A H . . . B is a hydrogen-bonded intermediate (11), (1) kl2 k23 k21 k32 Br 0 (11) and A-. . . HB+ is an intimate ion pair (III). The overall forward rate constants are high and approach the value predicted by the Smoluchowski theory of diffusion- controlled bimolecular reactions. For encounters between spherical particles of similar radius, in a homogeneous fluid, this theory gives : where D is the sum of the translational diffusion coefficients of the reacting species (DA+ DB), and o is the sum of the effective radii (rA+rB). If we substitute for DA 105 kD = 4xoD (2)106 KINETICS OF FAST PROTON-TRANSFERS and DB the values given by the Stokes-Einstein equation for spherical molecules comparable in size with the solvent molec~les,~ we obtain for kD where rA = rB : and for the corresponding enthalpy of activation : DA = kT/4nyrA and DB = kT/4nyrB, (3) kD = 4 RT/y (4) AH; = B where q = qo exp(B/RT). (5) For reactions in chlorobenzene at 25"C, the value of kD calculated from eqn (4) is 1.35 x 1Olo dm3 mol-1 s-l.The observed rate constants for the reactions of Bromo- phenol Magenta E with a series of tertiary aliphatic amines are of the order of 10-100 times less than this, and vary with the mine in a way that indicates that the steric effects of the carbon chains are important. The value of AH: calculated from eqn (5) is 2.1 kcal mol-1 (8.8 kJ mol-l) ; the observed values for most of the amines are smaller than this.These divergencies from the simple theory of diffusion control merit further examination. In the present paper we consider further the reaction of Bromophthalein Magenta E with tri-n-butylamine in a series of solvents. The experimental investigation is parallel to that carried out by Ivin, McGarvey, Simmons and Small on the analogous reaction of 2,4-dinitrophenol,4 whose kinetics show marked similarities to those of the present EXPERIMENTAL MATERIALS Tetrabromophenolphthalein ethyl ester and tri-n-butylamine were purified as described in an earlier paper.' Solvents were reagent-grade, and were washed with NaHC03 and water to remove acid, dried with CaS04 and kept over molecular sieve; refractive indices agreed with literature values within 0.0002. APPARATUS AND PROCEDURE Rate constants were determined by means of the microwave temperature-jump apparatus as described in the earlier paper.l The sensitivity of the detection system was improved by a factor of about 4 by using a more sensitive photomultiplier (RCA 2055) and by raising the intensity of the tungsten lamp for about 2 s by means of a voltage-pulsing circuit.A polythene cell was used; the temperature rise was about 0.1"C. The temperature was controlled to +O.l"C by circulating the reaction solution through a coil in a thermostat. Each temperature-jump was replicated, normally 6-10 times, and the relaxation times z were determined by the method of Crooks, Tregloan and Zetter ; the standard deviations were 5-10 %. Rate constants for the forward and backward reactions, kf and kb, were derived as described previously by plotting r1 against (K([AHj+ [B])+ 1) or ([AHj+[B]+K-'). Equilibrium measurements were made on a Hitachi Perkin-Elmer 1 39 spectrophotometer. RESULTS Equilibrium constants at various temperatures are given in table 1, and rate constants in table 2.In table 3 we summarise the values of the equilibrium constants and rate constants at 25"C, and the thermodynamic and activation parameters, computed by a weighted-least-squares program. In fig. 1 we show the plot of the reciprocal of the forward rate constant (l/kJ against the viscosity (y) of the solvent at 25°C. The slope is (1.42+0.05)10-3 mol J-l, the intercept is (7.0+0.5)10-10 mol dm-3 s, and the ratio slope/intercept is thus (0.20+0.014) x lo7 dm3 J-' s.The data for the analogous reaction of 2,4-dinitrophenol are plotted for comparison.G . D. BURFOOT, E. F. CALDIN AND €3. GOODMAN 107 TABLE 1 .-EQUILIBRIUM CONSTANTS FOR REACTION OF BROMOPHTHALEIN MAGENTA E WITH TRI-n-BUTYLAMINE IN VARIOUS SOLVENTS solvent iodobenzene bromobenzeiie n-pen t yl chloride n-butyl chloride 10-4e/dm3 mol-1 cm-1 3.86j10.01 3.69 -J= 0.03 3.96k0.01 3.90+ 0.1 temp./ O C 25.0 16.70 19.85 24.40 29.20 33.90 15.6 19.5 24.7 29.3 34.1 16.8 25.2 32.7 40.7 no. of solns 5 5 5 5 5 5 5 5 5 5 5 4 5 5 5 1OWHlol mol dm-3 2.35 2.8-4.2 2.8-4.2 2.8-4.1 2.7-4.1 4.6-9.3 4.6-9.2 4.6-9.2 4.6-9.1 4.5-9.1 2.8-4.2 2.98 2.95 1.95 1.93 1 OS[Blol mol d m - 3 2.2-1 1 5.8-14 5.8-14 5.8-14 5.8-14 5.7-14 5.5-16 5.5-16 5.5-16 5.4-16 5.4-16 2.6-6.5 5.6-7.7 2.5-7.6 2.5-7.5 10-4Kldm3 mol-I 3.16k0.30 9.8+ 1.3 6.3 1 jl 0.25 3.94k0.17 2.43 + 0.08 1.62 & 0.10 2.132 0.08 3.0rfr: 0.14 L32+ 0.03 0.922 0.01 0.62_+0.01 3.3 6 k 0.1 7 1.60+ 0.01 0.95+ 0.02 0.49j10.01 [AHlo = initial conc.of Bromophthalein Magenta E ; [B], = initial conc. of tri-n-butylamine ; E = extinction coefficient at 568 nm at 25°C ; K = equilibrium constant ; uncertainties are standard deviations. TABLE 2.-RATE CONSTANTS FOR REACTION OF BROMOPHTHALEIN MAGENTA E WITH TRI-ll- BUTYLAMINE IN VARIOUS SOLVENTS 10-4~1 temp./ dm3 no. of 105[AH]0/ 105[BI0/ K([AHlo+ IO-Sk,/ 10-4k,i "C mol-1 solns mol dm-3 mol dm-3 [Blo) 1O%/s dm3 mol-1 s-1 s-1 iodobenzene 25.0 3.16 5 1.57 2.5-15 1.9-5.5 47-16 3.52f0.09 1.115rt0.03 bromobenzene 3.6 37.6 4 1.37 0.9-5.4 4.7-17 352-68 3.1450.5 0.084f0.001 11.2 15.8 4 2.68 1.8-7.0 4.0-8.7 124-40 3.77f0.6 0.24k0.04 20.4 5.9 4 2.78 3.5-14 3.0-8.4 42-16 4.5650.07 0.77*0.01 25.0 3.7 5 2.49 2.6-21 2.2-8.1 38-10 4.4550.10 1.20&0.03 34.0 1.55 4 2.67 3.5-10 1.7-2.6 16-10 5.62fO.12 3.6350.09 n-pentyl chloride -8 22.2 5 1.0 1.0-4.1 2.2-7.4 172-63 3.84f0.4 0.17f0.01 0 11.3 5 1.0 1.0-4.1 1.4-3.9 140-56 3.79*0.4 0.3450.03 10 4.85 7 1 1.3-7.6 1.2-3.4 68-21 4.59f0.9 0.95&0.2 17 2.68 8 1 1.3-7.5 0.5-1.9 27-15 6.3350.3 2.3650.1 25 1.36 7 1 2.5-10.0 0.4-1.3 13-8 7.08rt0.5 5.2&0.4 n-butyl chloride - 8 33.9 10 0.6-2.6 0.6-3.5 2.1-8.3 255-71 4.61 50.4 0.136&0.013 0 15.1 4 1.31 0.6-3.4 1.8-4.1 110-80 5.03&0.03 0.33350.002 7 7.76 5 1.27 1.1-5.6 1.2-3.8 65-32 5.2150.23 0.67f0.03 16 3.47 4 1.3 2.0-7.0 0.8-2.2 32-18 5.9rt0.02 1.71*0.005 25 7.60 4 1.2 2.3-8.0 0.5-1.3 13-9 7.8rt0.25 4.9*0.15 32.5 0.902 4 1.3 2.9-9.8 0.3-0.9 9-6 8.6rt1.6 9.5f1.8 [AH10 - initial conc.of Bromophthalein Magenta E ; [BID = initiaI conc. of tri-n-butylamine; K = equilibrium constant calculated from data in table 1 ; z = reIaxation time ; kf = rate constant for forward reaction ; k , = rate con- stant for reverse reaction. DISCUSSION The salient experimental features of the kinetics of the reactions of the aliphatic amines with Bromophthalein Magenta E in chlorobenzene may be summarised as follows. (i) The forward rate constants kf are high, ranging from 5 x 10s dm3 mo1-l s-l for tri-n-butylamine to 2 x lo9 dm3 mol-1 s-l for trimethy1amine.l These values are smaller than kD (eqn (4)) by a factor of 7 to 30.108 KINETICS OF FAST PROTON-TRANSFERS TABLE 3.-THERMODYNAMIC AND KINETIC DATA FOR THE REACTION OF BROMOPHTHALE I N MAGENTA E WITH TRI-Il-BUTYLAMINE IN VARIOUS SOLVENTS 10-4K/dm3 mol-1 AG"/kcal mol-1 AH"/kcal mol-1 ASo/cal K- 1 mol- 1 10-8kt/dm3 mol-1 s-1 AG f pca~ mol-1 AH%/kcal mol-1 AS? /cal K-1 mol-1 1 0-4kb/s- 1 A G t /kcaI mol-1 AH% /kcal mol-1 AS $ /cal K- 1 mol- 1 103q/p0i~e kt lk, lO-9kD/dm3 mol-1 s-1 AH$ /kcal mol-1 (AH3 -AH?)/kcal mol-1 PhI 3.1610.3 - 6.1 1 10.05 3.5110.09 5.78 30.02 1.115 k0.03 11.8910.02 15.6 6.5 0.05 2.8 - PhBr 3.7 f 0.15 -6.25k0.03 - 17.6f0.4 -38.712 4.44 1 0.12 5.64 10.02 2.0f0.3 -9.6fl 1.20i0.03 11.86f0.02 20.5 10.3 +28.8 f 1.0 10.6 9.6 0.05 +0.3&0.3 2.5 PhCl* 3.17f0.05 -6.15k0.01 - 16.3f0.3 -34.1f1 5.210.7 5.53h0.01 1.5f0.5 -13.tjf2 1.56h0.2 11.69 fO.O1 17.9 f0.4 +20.9fl 7.5 13.5 0.04 2.1 +0.6&0.5 PenCl 1.36f0.01 -5.6010.02 - 14.810.2 -30.910.6 7.1 f0.5 5.3610.04 3.010.6 8.of0.6 5.4*0.4 10.96 f 0.05 15.8 f 0.6 f16.lf2.1 5.47 0.04 2.0 - 1.0&0.6 18 BunCl 1.60k0.03 -5.71 AO.01 - 14.33r0.35 -28.7f 1.1 7.810.25 5.31 10.03 1.210.4 -13.7fl 4.92kO.4 1 1.02 k0.03 14.6 k0.4 +12.310.6 4.27 0.03 1.8 +0.6*0.4 24 K, kf, kb, AG and 9 at 25°C ; 9 = viscosity; kD and AH$ calculated from eqn (4).1 cal I 4.184 J. * ref. (1). I I to 20 lo4 T/kg m-l s-l FIG. 1 .-Plot of 1 /kf against viscosity (7) for the reactions of tri-n-butylamine with (a) Bromophthalein Magenta E at 25°C (e) and (b) 2,4-dinitrophenol at 22°C (0). (ii) They are broadly similar to the values of kf for the reactions of the same amines with two other substituted phenols, namely 2,4-dinitrophenol and Bromo- phenol Blue lo; for a given amine the values lie within a factor of 2 (table 4).The variations in kf along the series of amines are much the same for each acid ; so also are those of the backward rate constant kb and the equilibrium constant K (table 4).G . D. BURFOOT, E . F . CALDIN AND H . GOODMAN 109 (iii) The values of kf do not correlate with the spectroscopically-determined overall equilibrium constant K, nor with the values of K,, estimated in the previous paper (scheme 1).l (iv) The values of kf vary with the length of the carbon chains of the amines in a way that strongly suggests a steric (space-filling) effect of the carbon atoms beyond the second.l. * TABLE 4.--COMPARISON OF KINETIC DATA FOR REACTIONS OF MINES WITH 2,4-DINITROPHENOL, BROMOPHENOL BLUE, AND BROMOPHTHALEIN MAGENTA E, IN CHLOROBENZENE AT ca.25°C l0glO(K/dm3 mol-1) 10g10(k,/dn13 mol-* s-1) 1oglo(kb/S-1) DNP BPB BME DNP BPB BME DNP BPB BME trimethylamine 3.44 3.53 3.84 39.3 29.1 9.28 >5.9 25.6 5-43 triethylamine 4.28 4.53 5.08 9.08 9.08 9.20 4.81 4.56 4.13 tri-n-propylamine 3.59 3.69 4.32 8.67 8.75 8.90 5.09 5.06 4.57 tri-n-butylamine 3.74 3.86 4.50 8.5 8.74 8.72 4.86 4.88 4.19 2,4,6-trimethyl- coilidine - 5.96 2.34 - 7.8 8.97 - 1.8 6.63 AH$ I kcal mol-1 DNP BME 0.7 2.0 3.0 -0.9* 1.5 0.5 DNP = 2,4-dinitrophenol; K at 25"C, k at 20-24°C (ref. (7)). BPB = Bromophenol Blue ; K and k at 25°C (ref. (10)) BME = Bromophthalein Magenta E ; K and k at 25°C (ref.(1)). * ref. (4). These facts and other considerations given below suggest that, for all three of the substituted phenols mentioned in paragraph (ii) above, the reactions with tertiary aliphatic amines in chlorobenzene proceed by a mechanism in which the proton- transfer is not the rate-limiting step ; it may be preceded by the rate-limiting formation of a hydrogen bond : k12 K23 AH+B+AH . . . B+A- . . . HBf. k2 1 fast It is possible that the hydrogen-bonded species is not formed as a true intermediate in all these reactions, i.e. that there is no potential energy barrier between the reactants and the ion-pair. All that is required for reaction is that the two molecules should be in a favourable orientation in which they can proceed to form a hydrogen-bonded complex, or to form an ion-pair, without an energy-barrier.The essential point in our interpretation is that the proton-transfer step in the forward direction is not a rate-limiting step. In what follows, the formula AH. . . B may represent either a hydrogen-bonded intermediate or a pair of molecules correctly oriented for proton- transfer. VARIATION OF RATE CONSTANT WITH SOLVENT The values of kf in various solvents may now be considered. The facts are as follows. (a) The values of kf for the reactions of the three phenols with various amines in various solvents (tables 3 and 4) are smaller, by a factor up to 30, than the values calculated from simple diffusion theory for the rate constant k, of a diffusion- controlled reaction. The values of kJk, in various solvents for the reactions of Bromophthalein Magenta E are given in table 3.(b) The values of AH? for the reactions of Bromophthalein Magenta E with various amines in chlorobenzene (table 6 of ref. (1)) mostly do not agree with the value calculated from eqn (4) and the temperature-dependence of q, namely AH; = * There is an error in fig. 2 of ref. (7); the point representing log kf for quinuclidine is shown one unit too low.110 KINETICS OF FAST PROTON-TRANSFERS 2.1 kcal mol-1 (8.8 kJ mol-l). The values of A H f for the reaction with tri-n- butylamine in various solvents (table 3) do not differ greatly from AH; ; the differ- ence may be positive or negative. For the reaction of 2,4-dinitrophenol with tri-n- butylamine in chlorobenzene,2 the value of AH: is slightly negative (- 0.9 & 0.4 kcal mol-', or - 3.5 +_ 1.5 kJ mol-l).(c) When different solvents are used, the values of kf for a given reaction show a good inverse correlation with the viscosity. This has been observed for the reactions of tri-n-butylamine both with 2,4-dini trophenol and with Bromophthalein Magenta E (this work). The plots of l / k f against q, shown in fig. 1, are good straight lines. They do not agree, however, with eqn (4); they show an intercept, and the slope differs greatly from 1 /4 RT. An intercept on the plot of k , against y is to be expected if in place of Smoluchow- ski's treatment we use Noyes' appr~ach,~ in which it is assumed that at the critical separation of the molecular centres (0) the concentration of the approaching molecule drops not to zero but to a finite value determined by a rate constant k , which would obtain if diffusion effects were entirely absent.This ko would in the simplest case be equal to the collision number 2, but in general may be put equal to p Z , where p is a collision efficiency l1 which may be less than unity by reason of a geometrical steric factor or an activation energy or both. This approach takes account of both the diffusional and the chemical restrictions on the rate of reaction, and results in the equation for the forward rate constant : where kD = 4noD as before (eqn (2)). D (eqn (3)), this becomes : The intercept on a plot of k-1 against y would thus beinterpreted ask-ol. For the reaction of Bromophthalein Magenta E with tri-n-butylamine, OUT results give 1 /(intercept) = (1.4 +,O.1) x lo9 dm3 mol-l s-I ; for the reaction of 2,4-dinitrophenol,4 the values is (3 _f2) x lo9 dm3 mol-1 s-l. The slope of the plot of l/k against y should according to eqn (6a) be 1/4 RT, which has the value 1 . 0 ~ niol J-l. Howver, the slope for the reaction of Bromophthalein Magenta E (fig. 1) is 14 times this value, and for the reaction of 2,4-dinitrophenol the factor is 28. It is clear that the two-step model is inadequate, even though the factor p has been introduced to take account of steric and activation factors; this factor appeus in the intercept, not in the slope. k-' = k,l+k,l ( 6 ) k , ' i- y/4RT. ( 6 4 In terms of the Stokes-Einstein expression for ic-1 = THEORY OF WEAK INTERMEDIATE COMPLEX All the observations can be explained, however, by a reaction scheme in which the formation of a weak intermediate complex, which we denote by (AH,B), precedes the formation of the hydrogen-bonded species and ion-pair : IC 1 k2 k3 AH+B+AH,B+AH .. . B s A - . . . HBf. (7) k- 1 k-2 k-3 A similar scheme has been treated by Ivin, et aL8 (see case 3 of their paper). We apply the steady-state approximation to the species AH,B and A H . . . B. Provided that the formation of the ion-pair is fast conipared with the reverse step (k3 + k--3), we obtain for the forward rate constant We may assunie that k3 kf = klk2k3/[k--l(k3 + k--2) +k,k3]. (8) k+. since the activation energy for breaking the hydrogenG . D. BURFOOT, E. F. CALDIN AND H. GOODMAN 111 bond will be around 7 kcal mol-1 (30 kJ mol-I), whereas that for proton-transfer within the hydrogen bond is likely to be small.Applying Noyes' relation (eqn (6)) to the first step, Substituting for kl in eqn (9), and putting k, = 4 RT/q (eqn (4)), we obtain : A plot of kT1 against y will then be linear, provided that (k-1/k2) is independent of the solvent, with slope = [(k-,/k,) + 1114 RT; intercept = [(k-,/k,) + l]/ko ; and (slope/intercept) = ko/4 RT. The unexpectedly large slopes of the plots in fig. 1 may now be satisfactorily interpreted in terms of the three-step scheme (7) with (k-' /k2) 9 1. For the reactions of tri-n-butylamine with Bromophthalein Magenta E and 2,4-dinitrophenolY the slopes give (k-l/k2) as 13 and 27 respectively. The intercepts give for ko the values (2.0k0.14) x 1O1O and (9+6) x 1O1O dm3 mol-1 s-l respectively.Eqn (8) then reduces to : k f = k,k,/(k-, +k,). (9) k,' = ko1+ki1. (10) (11) k i = k , ' [(k- l/kz) + 11 + (q/4RT)[(k- l/k2) + 11. The expression for kf (eqn (9)) gives, since (k-,/k,)> 10, the approximation : kf 2: k,k2/k-,. (12) Since k , varies linearly with viscosity (fig. 1) and so does k i according to our model (eqn (lo)), we assume in order to satisfy eqn (9) and (12) that k,,/k2 is effectively independent of viscosity. We can now give a fuller description of the model that fits the kinetic scheme (7) so far, and test it against the rest of the observations. We assume that the intermed- iate complex (AH,B), formed by diffusion-controlled encounter, is held together for a short time by dipole-dipole or dispersion forces.The lifetime of the complex is therefore not determined by the properties of the solvent cage, as in a simple en- counter, but is prolonged by reason of the attractive forces between the reactant molecules. Such forces are relatively non-directional, and during the life of the complex the molecules can rotate; there is thus a chance that before the complex dissociates the molecules will rotate into an orientation favourable to the formation of the hydrogen bond (AH. . . B) or ion-pair A- . . . HBf, which requires a fairly specific direction. The chance of this occurring before the complex dissociates will d.epend on the steric properties of the molecules, on their rates of rotation, and on the lifetime of the complex. The lifetime will depend on the strength of the attractive forces, and therefore on such properties of the reactant molecules as their chain- length.The chance of reactions taking place during this lifetime will also be affected if energy of activation is needed to open an internal hydrogen bond, or to bring about desolvation. (We must assume, however, that the chance does not depend on properties of the solvent, such as viscosity, since for given reactants ( k l / k 2 ) at 25°C appears to be independent of the solvent.) The effect of varying the chain-length of the base will be different for different acids, since it depends on the lifetime of the complex and therefore on the intermolecular forces. Steric factors and molecular attractions must both be taken into account. VARIATION OF BASE A N D ACID I N A GIVEN SOLVENT The expression for the forward rate constant obtained by combining eqn (9) and (10) is : lcf' = (ki1+k,')[(k-1/k2)+1].(13)112 KINETICS OF FAST PROTON-TRANSFERS The observed rate constants for a given reaction in a given solvent will therefore diffkr from kD by a factor depending on ko and on ( ~ t . - . ~ / k ~ ) . When the base is changed, the value of kf for a series of aliphatic amines with a given acid decreases with increase of chain-length.'. 7 9 * This may be attributed to the effect on k, ; increase of chain length will decrease the effective solid angle avail- able at the reaction centre, and may also affect the rate of rotation of the amine within the complex. (Of the other parameters in eqn (13), ko may be expected to be much the same for all the aliphatic amines; k-l may be affected by dipole-dipole forces, which are difficult to estimate but work in the same direction as the changes in k2, and by dispersion forces, which would influence k, in the wrong direction.) The observed lack of correlation between kf and the overall equilibrium constant K, or KI2 for the formation of a hydrogen-bonded complex,' is also to be expected ; none of the factors in the expression for kf (eqn (13)) has any connection with proton- transfer, nor with the strength of the hydrogen bond.We can find an upper limit for k, as follows. We have assumed (to explain the linear plot of k f l against q) that k-l (unlike k , ) is much less than the diffusion- controlled value l2 (ca. loll s-l); thus we may put k-.l< lo9 s-l.The value of (kV1/k2) for the reaction of Bromophthalein Magenta E with tri-n-butylamine is 13, so to an order of magnitude k2 < los s-l. This rate constant represents the frequency with which the reactive sites come into favourable positions and react, and may be compared with reciprocal rotation times of the order of 1O'O s-l. A factor of the order of 100 would therefore account for the results; for the corresponding reaction of 2,4-dinitrophenol, the figure is about double. These figures seem not unreasonable, since they include the effect of the geometrical steric factors and also any effects due to desolvation and rupture of an internal hydrogen bond. When the acid is changed, keeping the mine fixed, the observed rate constants for the three acids are broadly similar (table 4), suggesting that the values of (k-,/k2) do not greatly differ.We should expect k2 to decrease with increasing size of the acid molecule, both because rotation will be slower and because the reactive site will occupy a smaller fraction of the surface. It appears therefore that there is some compensating decrease of k-l, possibly by reason of increased dispersion forces. Other factors, however, such as internal hydrogen bonding (see below) may also be important . THE VALUE OF p , AND THE NUMBER OF COLLISIONS PER ENCOUNTER As we have seen, the values of ko at 25°C are (9+6) x 1O1O for 2,4-dinitrophenol and (2.040.14) x 1O1O for Bromophthalein Magenta E. The relation ko = Zp then gives, if we assume 2-1 x loll dm3 mol-l s-l, the corresponding values of the collision efficiency p at 25°C for the formation of the preliminary complex as about 1 and about 0.2 respectively, for the two acids.Values near unity are indeed to be expected, since the orientation of the molecules is not critical ; and since 2,4-dinitro- phenol is much the smaller molecule, the values are in the order that might be expected. Alternatively, we may regard the experimental values of ko as lower limits for the collision number 2; they are compatible with calculations indicating that the value of 2 in solution is of the order of lo1' dm3 11101-1 s-l, not greatly less than the value calculated from kinetic theory for collisions in the gas phase. Reference to table 3 shows, however, that the value of ko for Bromophthalein Magenta E lies within the range of values of kD, whereas we should expect kD to be at least an order of magnitude less than 2, on account of the evidence that in an encounter there are multiple collisions.l4 This suggests that 2 is somewhat larger than 1 x lo1 dm3 mol-l s-l.G. D. BURFOOT, E. F . CALDlN AND H. GOODMAN 113 ENTHALPIES OF ACTIVATION From the approximate relation (12), we obtain : AH: = AH:-- (AH? -AH,*). For AH:, eqn (10) gives : (14) The value of AH: must thus lie between AH: and AH:. If, in the expression ko = Zp,p is a purely geometrical factor, as is likely for the formation of a weak complex in the absence of desolvation effects, then AH: = 0. Combining eqn (14) and (15), we then obtain : AH: = AH: ~ -(AH?, -AH:).(k,:ok,) It is thus not surprising that observed values of AH: differ from the calculated values of AH: (table 3). The first term on the right in eqn (16) may be evaluated for the reactions of tri-n-butylamine with 2,4-dinitrophenol in chlorobenzene and with Bromophthalein Magenta E in various solvents. The results are shown in table 5. TABLE 5.-ANALYSIS OF ENTHALPY OF ACTIVATION FOR REACTIONS OF TRI-n-BUTYLAMINE WITH DNP AND BME (ko/ko+k~)4Hg / (AH- ?-AH? 1 acid solvent ko/(ko+kD) 1-1 kcal mol-1 DNP chloro benzene 0.9+ 0.6 2.1k1.2 + 3.1 + 0.9 BME chloro benzene 0.60+0.07 1.3 + 0.2 -0.2k0.5 br orno benzene 0.68+ 0.07 1.7+ 0.2 - 0.3 + 0.4 n-pentyl chloride 0.52+ 0.07 1 .Ok 0.2 - 2.0+ 0.6 n-butyl chloride 0.462 0.07 0.8k0.2 - 0.4+ 0.4 For symbols see text. DNF = 2,4-dinitrophenol.BME = Bromophthalein Magenta E. For the reaction of 2,4-dinitrophenol with tri-n-butylamine in ~hlorobenzene,~ the difference between AH: and AH; (3.5k0.4 kcal mol-l) is almost entirely due to the second term (AH-: -AH; = 3.1 k0.4 kcal mol-l). Since AH; cannot be negative, it follows that AH-: is at least 3 kcal mol-1 (12 kJ mol-') ; thus the break-up of the complex (AH,B) is associated with an appreciable activation energy, as our model requires. The energy necessary to separate the reactant molecules against the attractive forces, whether dipole-dipole or dispersion,l may well amount to several kcal mol-l. The negative value of AH: is thus explained, in terms of AH--: ; it does not appear to be possible to explain it without a three-step model such as that proposed.For the corresponding reaction of tri-n-butylamine with Bromophthalein Magenta E, the position is less clear. The contributions to AH: from the term (AH-: - A H ; ) are small or positive in all the solvents (table 5). If AH-: is appreciable, as our model requires, and as it appears to be in the reaction of 2,4-dinitrophenol, then AH: must also be appreciable, perhaps 3-6 kcal mol-1 (12-25 kJ mol-l). Such an energy-barrier can account for the deviation of kz from the rotation-controlled value of around 10"114 KINETICS OF FAST PROTON-TRANSFERS s-l. It might be partly due to an intern21 hydrogen bond, for whose existence there is some evidence (see ref. (16)-(19) in ref. (l)), and partly to des~lvation.'~ The changes of AH: in the series of amines (table 4) cannot be interpreted with certainty, since we do not know ko for these reactions.If the changes are due to variations of AHfl, the increase from trimethylamine to tri-n-propylamine can be attributed to increased dispersion forces, but the interpretation of the decrease with tri-n-butylamine is not obvious. ALTERNATIVE EXPLANATIONS We now consider some possible alternative explanations of the observations, particularly of the fact that the forward rate constants are an order of magnitude less than kD, so that the slopes of the plots in fig. 1 are much larger than the value 114 RT predicted by diffusion theory along with the Stokes-Einstein relation. GEOMETRICAL STERIC FACTORS I N A LSTEP PROCESS WITHOUT AN INTER- We have noted that a geometric factor will appear because the reaction centres are accessible only over a limited solid angle.It has been found possible to reproduce the rate constants of lo7 to lo9 dm3 mol-1 s-I found for a number of proton-transfer reactions in water by simply multiplying the value of k, by a steric factor y which is calculated from the solid angles of the reactive centres and gives the chance of a favourable orientation on collision.18 However, in an encounter there are repeated collisions, and the chance of a collision with correct orientation will not be equal to y. This treatment moreover fails to predict the intercepts on the plots of kfl against viscosity. Again, the rate constant for the reaction of Bromophthalein Magenta E is 40 % greater than that for the reaction of 2,4-dinitrophenol with the same base (tri-n-butylamine) in the same solvent, whereas the geometric factor alone would lead to a markedly smaller value.We have already noted that a geometric factor can be successfully introduced only in a 3-step scheme. MEDIATE COMPLEX CORRECTIONS TO THE STOKES-EINSTEIN EQUATION Although the Stokes-Einstein equation appears to be soundly based as regards its general form,19 the measured diffusion coefficients for small molecules are often larger than it predicts 20-22 and would give higher values for k,; an effect that is in the wrong direction to explain our observations. We have tested the equation for our reagents by measuring diffusion coefficients in bromobenzene at 20°C, by means of a diaphragm cell 23 with iodine as reference substance. The values obtained are close to those calculated from eqn (3) with values of r estimated by the methods of Ed- ward 22 ; they are as follows : Bromophthalein Magenta E, observed (0.43 f 0.01) x cm2 s-l, calculated 0.74 x Hence for (DA+ DB) the experimental value is 1.23 x cm2 s-l, differing by only It is possible to modify the calculated value of kD by using models of the liquid that are physically more realistic than the homogeneous-fluid model.21* 24 Co- operative short-range motions may reduce diffusion coefficients below the Stokes- Einstein values.It appears, however, that no plausible microscopic model can change by as much as a factor of 2 the value of kD for a species of known size and diffusion c~efficient.~~ cm2 s-l, calculated 0.62 x ; tri-n-butylamine, observed (0.80&0.03) x cm2 s-I, while the calculated value is 1.36 x 10 %.G .D . BURFOOT, E. F. CALDIN AND H. GOODMAN 115 INTERNAL HYDROGEN BONDING We have taken account of the possibility that the need to break an internal hydro- gen bond in the second step will affect k2 and AH;. We must however consider the possibility that there is a preliminary equilibrium between the internally-hydrogen- bonded form AH and a form AH‘ in which the bond is broken, followed by diffusion- controlled encounter and reaction of AH’ with B : KA kD AH+AH’; AH’+B-+A- . . . HB+. The observed rate constant would be k, = KAkD according to the Smoluchowski equa- tion, or k, = KA[kDkO/(kD+kO)] according to the Noyes equation. This would not account for the observed variations of k, and AH, with the amine B, and could not account for the low or even negative values observed for AH:, since the expression for AH: is in which all the quantities are positive.SOLVENT POLARITY The correlation of k, with solvent viscosity might conceal a more fundamental correlation with some other property such as polarity, measured empirically by parameters such as ZZ6 We have determined the 2 values and find no correlation with kf. For the solvents in the same order as in table 3, they are respectively (within k0.4) 56.5, 57.1, 57.1, 56.4 and 53.9. Thus iodobenzene and n-pentyl chloride, which have very different values of k, have the same values of 2 ; and n-butyl chloride, in which k, is much the same as in n-pentyl chloride, has a diffient value of 2.DESOLVATION If it were assumed that desolvation had to precede the formation of (AH,B) in the three-step model, or A H . . . B in the two-step model, the rate constant would be reduced by a factor equal to the equilibrium constant for desolvation. It is most unlikely, however, that this equilibrium constant would be nearly the same for all the solvents investigated, as the linear plots of k f l against viscosity would require, especially since the solvents are known to differ in their ability to form donor-acceptor complexes and hydrogen-bonded complexes. Desolvation may well be required, but it is more likely to occur in the second step and so to influence k, and AH;. (There is some evidence for it in the kinetics of the formation of a donor-acceptor complex in chlorobenzene.’) RATE-LIMITING ROTATION, RESTRICTED BY SOLVENT VISCOSITY If the rotation of the reactant molecules were restricted by viscous resistance of the adjacent solvent, the result would be that kf would depend on the viscosity.* Suppose equilibrium is set up in the first step, while the second step is rate-limiting. Then and (since for neutral molecules Kl may be taken 27 as (4/3)71r3N0, which is independ- ent of temperature) kf = Klk2 (1 8) AH:= AH<+AH,f = AH:. (19) Suppose for simplicity that the reactant molecules are spheres of volume V, and116 KINETICS OF FAST PROTON-TRANSFERS that the reaction sites occupy effective fractions xAH and xB of their surfaces. The mean time of rotation given by Debye’s equation is and with calculated molecular volumes 22 this is about 2 x 10-lo s.There are then the following possibilities. (i) If the second step requires appreciable activation energy, for desolvation and rupture of an internal hydrogen bond, then k2 is controlled by this activation energy and not by the rate of rotation, so that the viscosity-depend- ence of kf cannot be explained. (ii) If the second step is completed whenever the reaction cites become adjacent, in a time short compared with z, then whence, using eqn (5), These predictions cannot be reconciled with the results. Steric effects, and a linear plot of k; against q, are predicted by eqn (21), but it does not account for the intercept on the plot. According to eqn (22), the value of AH: should be dependent only on the solvent, and should be equal to AH:.This is at variance with the fact that for the reaction of Bromophthalein Magenta E with mines AH: in the series of amines is not constant. It is also incompatible with the negative value observed with 2,4- dinitr~phenol.~ The same objections apply if there is a fast preliminary desolvation step, so that there is an equilibrium concentration of desolvated molecules which alone take part in the rate-limiting second step. If the equilibrium constant for the desolvation is Ks, this gives whence Z, = (3V/kT)q (20) ki = Kik2 = KI(XAHXB)(~T/~ V)/q AH: = AH,’ = B = AH:. (21) (22) ki = K1Ksk2 (23) AH; = AH:+B. (24) Eqn (24) allows different values for AH: for different systems, but not values less than B, as observed. CONCLUSION We conclude that the most likely interpretation of the results is the three-step scheme (eqn (7)), in which the initial formation of the loose complex AH,B is diffusion- controlled while its dissociation is not, and the subsequent rotation step is rate- limiting but not mainly by reason of viscosity effects. The forward rate constant is lower than the diffusion-controlled value because the reactant molecules have to move into a suitable mutual orientation within the loose complex first formed.The chance of this happening within the lifetime of the complex depends both upon the con- figurations of the reactant molecules and upon the intermolecular forces between them. We acknowledge an S.R.C. research grant to H. G. and a C.A.P.S. studentship sponsored by Shell Research Centre, Thornton, to G.B., and thank Dr. B. H. Robinson and Dr. J. J. McGarvey for helpful discussions. E. F. Caldin, J. E. Crooks and D. O’Donnell, J.C.S. Faraday I, 1973, 69, 1O00. R. M. Noyes, Progr. Reaction Kinetics, 1961, 1, 129. E. Mchughlin, Trans. Faraday SOC., 1959,55, 28. K. J. Ivin, J. J. McGarvey, E. L. Simmons and R. Small, Trans. Faradzy SOC., 1971, 67, 104.G. D. BURFOOT, E . F. CALDIN AND H. GOODMAN 117 K. J. Ivh, J. J. McGarvey and E. L. Simmons, Trans. Faraday SOC., 1971,67,97. E. F. Caldin and J. E. Crooks, J. Chem. SOC., 1967,959. K. J. Ivin, J. J. McGarvey, E. L. Simmons and R. Small, J.C.S. Fwaaby I, 1973,69,1016. J. E. Crooks, P. A. Tregloan and M. S. Zetter, J. Phys. E, 1970,3,73. lo J. E. Crooks, P. Sheridan and D. O'Donnell, J. Chem. SOC. B, 1970, 1285. S. R. Logan, Trans. Faraday SOC., 1967,63, 1712. l2 M. Eigen, W. Kruse, G. Maass and L. De Maeyer, Progr. Reaction Kinetics, 1964, 2, 287 ; R. M. Fuoss, J. Amer. Chem. SOC., 1958,80,5059. l3 R. P. Bell, Trans. Faraday SOC., 1937, 33,496 ; J. Chem. SOC., 1943, 629. l4 see, for example, A. M. North, The Collision Theory of Chemical Reactions in Liquids (Methuen, London, 1964), and references in E. F. Caldin, Fast Reactions in Solution (Blackwell, Oxford, l5 E. A. Moelwyn-Hughes, Kinetics of Reactions in Solution (Oxford, 1943, pp. 88, 206; E. S. l6 E. Grunwald and E. K. Ralph, J. Amer. Chem. Soe., 1967,89,4405. 17kE. F. Caldin, J. E. Crooks, D. O'Donnell, D. Smith and S. Toner, J.C.S. Faraday I, 1972, 68, Is A. Weller, Progr. Reaction Kinetics, 1961, 1, 187. l9 C. Longuet-Higgins, Nuovo Cimento Suppl. 1, 1955, 2, 140. 2o W. C. Ware, J. Chem. Phys., 1962,66,455. 21 A. D. Osborne, H. J. V. Tyrrell and M. Zaman, Trans. Faraday SOC., 1964, 60, 395. 22 J. T. Edward, J. Chem. Ed., 1970,47,261. 23 J. C. Gage, Trans. F a r d y SOC., 1948,44,253. 24 C. A. Fehder and P. L. Erneis, J. Amer. Chem. SOC., 1970,92,2246. 25 R. M. Noyes and H. Raman, J. Amr. Chem. SOC., 1958,80,2410. 26 E. M. Kosower, Physical OrgaHic Chemistry (Wiley, New York, 1968), pp. 293 seq. 27 J. E. Prue, J. Chem. SOC., 1965,7534; Ionic @uiibriu (Pergamon, Oxford, 1966), p. 92. ' E. F. Caldin, J. E. Crooks and D. O'Donnell, J.C.S. Faraday I, 1973,69,993. 1964), pp. 282-285. Amis, Kinetics of Chemical Change in Solution (Macmillan, New York, 1949), p. 183. 849.

ISSN:0300-9599    

DOI:10.1039/F19747000105

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

P . B . AYSCOUGH AND K . MACH 119 We reported previously the e.s.r. spectra of the fluoralkyl radicals CF3, CHF,, CH2F, CHFCOO-, CF,COO-, CF3CF, and CF3CF2CF2 obtained by the photolysis of polycrystalline Pb(1V) fluoroalkanoates. These showed the characteristic features described above for fluorine nuclei in the a-position and displayed additional structure in the wing peaks which could be attributed to interaction of F and H nuclei in the B-position. Although no detailed interpretation of the central features was attempted it seems reasonable, in view of the simplicity of the chemical behaviour of these systems, to use these spectra as the basis for identification of similar radicals in the y-irradiated carboxylic acid and salts. One of us has described in detail the e.s.r.spectra observed in a number of y- irradiated fluorinated carboxylic acids and their sodium salts at cryogenic tempera- tures.l0’ l1 The main features of the spectra observed at 77 K in the acids RCF,- COOW were attributed to perfluoroalkyl radicals RCF, which were replaced COM- pletely or in part following thermal annealing by secondary radicals of a type which suggested loss of fluorine from the parent acid. In the sodium salts some secondary radicals were present at the temperature of the y-irradiation but again the primary radicals disappeared at a lower temperature than the secondary radicals. These observations were interpreted in terms of fluorine abstraction by the primary per- fluoroalkyl radicals and a tentative mechanism was put forward. However, some further observations and a re-examination of central regions of the spectra, together with a report on the products of the radiolysis of trifluoroacetic acid,12 have thrown further light on the mechanism of the radiolysis.EXPERIMENTAL The procedure for y-irradiation at 77 K, for recording the e.s.r. spectra and for thermal annealing of samples has been fully described.l3? l4 In this work the total dose of 6oCo y-rays was typically 2 to 4 x 1020 eV ~ r n - ~ . Photolysis of the y-irradiated samples was made using a high-pressure mercury lamp giving radiation mainly of wavelengths 300-400 nm. All the fluorinated carboxylic acies were supplied by Koch-Light Laboratories and were used without further purification. The sodium salts were prepared by controlled neutralisation of aqueous solutions of the acids and then recrystallised.The pure acids freeze to a poly- crystalline state but aqueous solutions containing 30-50 % by volume of water generally give transparent glasses which were more suitable for photolysis. All samples were y-irradiated in Spectrosil high-purity quartz tubes after freeze-pumping to a pressure < mmHg. The very narrow peak at g = 2.00 in the e.s.r. spectrum attributed to radiation damage in the quartz is readily recognisable. The very narrow peaks with 50.2 mT separation also observed in all samples at 77 K are caused by hydrogen atoms trapped in the quartz. RESULTS Six fluorinated ca.rboxylic acids and their sodium salts were examined after y- irradiation at 77 K and during subsequent thermal and photochemical annealing. In some cases frozen aqueous solutions of the acids were also examined as an aid to the interpretation of the e.s.r.spectra and in one case, trifluoroacetic acid, the anhyd- ride was studied. we used the characteristic wing features corresponding to the summation CAr(max) for radicals containing one, two or three a-fluorine atoms to identify the trapped radicals. In this way we have established that in all the acids RCF,COOH and salts RCF,COONa studied, primary fluoroalkyl radicals RCF, are present immediately after irradiation at 77 K. In the sodium salts additional radicals with one less a-fluorine than the parent compound predom- inate at 77 K, and species of this kind assume great importance in the irradiated Following established procedures 5 9 6,120 E.S.R.OF 7-IRRADIATED RCFZCOOH acids after thermal annealing. Although the central portions of the spectra are not easily interpreted in detail because of the complexity of the spectra in polycrystalline samples, there are clear indications of the presence of at least one further type of radical in the pure acids and in the frozen aqueous solutions. We believe that these are radical-anions RCF,cOOH-, similar to those already reported in many carboxylic acids and esters at cryogenic temperatures. 3-1 Chachaty and Shiotani reached the same conclusion in relation to CF3COOH, CF3CF2COOH and (CF2COOH)2, though we believe that the hyperfine parameters attributed to the radical-anions in these systems are incorrect. Srygley and Gordy l8 also suggested the presence of radical-anions in the ammonium salt CF3COONH4 though again we cannot agree with the spectral assignments quoted to support this hypothesis.TRIFLUOROACETIC ACID The behaviour of this compound on y-irradiation appears to be typical of fully fluorinated acids and it has been examined in the greatest detail. We have studied the anhydride, the pure polycrystalline acid and frozen aqueous solutions of the acid, each subjected to thermal and photochemical annealing. THE ANHYDRIDE The spectra observed in the pure polycrystalline anhydride at 77K and after bleaching with visible light at 77 K are showii in fig. 1. The spectra ire conveniently discussed in three sections (i) the weak outermost peaks with positive and negative bands separated by 70-75mT, (ii) an intermediate region with complex structure separated by 25-30 mT, (iii) the much more intense central region which has a pro- nounced doublet structure with splittings of 5-9 mT. Interpretation is greatly helped by the work of Maruani and co-workers ’* who used the values for the principal components of the fluorine hyperfine tensor derived for CF3 radicals in single crystal CF3CONH2 to compute the spectra in an inert random matrix.The tensor is nearly axially symmetric with A f ~ 2 5 . 3 and A f ~ 9 mT : thus one expects pronounced discontinuities to appear in the polycrystalline spectra at magnetic fields corresponding approximately to $-+A!, ++Af;, ,+AT and +$AT, the perpendicular features being much more intense. Second-order transitions will be superimposed on the *+ transitions (approximately 1.8 mT for the perpendicular transitions and 0.5 mT for the parallel transitions), and there is some g-anisotropy, so the central part of the polycrystalline spectra will be somewhat unsymmetrical.l Maruani et al. found excellent agreement between experimental spectra for CF3 radicals in a krypton matrix at 4.2 K and computed spectra assuming randomly oriented, non-rotating, non-inverting radicals. We have observed essentially similar spectra for CF, radicals in u.v.-irradiated (CF3C00)4Pb at 77 K (see fig. l(c)). It is apparent that the main features of the spectrum shown in fig. l(b) are also well described by this model which predicts major features at & 37.9, f 12.6, + 13.5 and k4.5 mT referred to the centre of the spectrum.The outermost peaks in fig. l(b) are separated by 76.0 mT while the peaks in the intermediate region are separated by about 27 mT. The central region shows peaks at -6.0, -4.3, +5.0 and +6.0 mT corresponding to the $AF, transitions with second-order splitting and g-anisotropy superimposed. The very close similarity between fig. l(b) and l(c) leaves no doubt that the only radicals present in y-irradiated (CF3C0)20 at 77 K after photobleaching are CF3 radicals (the narrow peaks at k25.1 mT are caused by H atoms in the quartz of the sample tube). The diffaence between fig. l(a) and l(b) is clearly due to the presence of a strong doublet of doublets with splittings 4.1 and 1.7 mT which can beP . B . AYSCOUGH AND K . MACH 121 removed by photoylsis at 77K without leaving any other paramagnetic species. This is a characteristic of the radical-anions found in y-irradiated carboxylic acids and esters I3-l6 and greatly facilitates the interpretation of these complex spectra.We shall consider the interpretation of the hyperfine parameters in more detail after reporting the results for the pure acid and the aqueous samples. * I I I 1 1 31 0 330 350 field strength/mT FIG. l.--(a) Pure (CF3CO)20 after y-irradiation at 77 K, (b) the same sample after 15 min photolysis at 77 K, (c) (CF3C00)4Pb after 10 min photolysis at 77 K. (Radical anion peaks are marked *.) AQUEOUS TRIFLUOROACETIC ACID When an approximately 50 % by volume solution of CF,COOH is y-irradiated at 77 K the initial spectrum observed is that shown in fig.2(a). The similarity to fig. l(a) is apparent (the outer and intermediate regions are shown at higher gain in fig. 2) but the central doublet is less well resolved and gives a hyperfine splitting of about 4.6 mT (cf. 4.1 mT in the anhydride). In addition some small peaks at +_ 19.8 mT are attributable to a small concentration of cF,COOH radicals. Photolysis at 77 K leads to the removal of the central doubIet, as in the anhydride, and the appearance of a rather narrow singlet at the centre of the spectrum (see fig. 2(b)-(e). The latter is attributed to the central peak (MI = 0 transitions) of the triplet122 E . S . R . OF Y-IRRADIATED RCFZCOOH spectrum of rotating cF,COOH radicals, seen more clearly in the sodium salts (see e.g. fig. 4 of ref.(10)). On the evidence of the outer peaks at 19.8 mT there is no increase in concentration of cF,COOH radicals during photobleaching so we must conclude that the central peak in the unbleached sample is sufficiently broad at 77 K 310 330 350 FIG. 2.-(u) 6 mol dm-3 CF3COOH/Hz0 after y-irradiation at 77 K, (6) the same sample after 30 s photolysis, measured after 10 min at 77 K, (c) the same sample after 200 s photolysis, measured at 77 K immediately after photolysis, (d) the same sample after 20 min at 77 K, (e) the same sample after 20 min photolysis at 77 K. (Peaks marked .T belong to cFzCOOH radicals, those marked * belong to CF3cOOH- radical anions.)P . B . AYSCOUGH AND K . MACH 123 as to be obscured by the very much stronger doublet of the radical-anion. During the photobleaching there are also some minor changes in the spectrum of CF3 radicals, shown most clearly by the broadening and shifting of the outermost peaks.This effect is to some extent reversible since the final spectrum after 25 min photolysis (fig. 2(e)) has an overall width almost identical with that measured initially (fig. 2(a)). The whole spectrum disappears at about 170 K without further change in structure. PURE POLYCRYSTALLINE TRIFLUOROACETIC ACID The main features of the spectrum shown in fig. 3(a) and observed immediately after y-irradiation at 77 K are similar to those of fig. 2(a) and may be attributed to CF3 and cF3cOOH- mainly, with traces of CF,COOH radicals. The radical-anion doublet splitting is about 5.2 mT and the outer peaks of the CF3 spectrum are broad, with separation 75-79 mT.Photobleaching for 15 min removes the radical-anion, leaves the cF,COOH spectrum unchanged but improves the resolution of the outer- most CF3 radical features (fig. 3(b)). Thermal annealing to 130 K further improves resolution (fig. 3(c)) and after heating to 180 K and recording at 77 K (fig. 3(4) the outermost CF3 peaks are clearly doublets, slightly asymmetric, with separation 4.5 and 5.0 mT. In the spectrum the features in the intermediate region are also doubled ir the same way and some additional similar structure appears in the central region, I I I I 1 31 0 330 350 field strength/mT FIG. 3.-(a) Pure CF,COOH after y-irradiation at 77 K, (6) after 15 min photolysis at 77 K, (c) the same sample after heating to 130 K, measured at 77 K, ( d ) the same sample after heating to 180 K, measured at 77 K.(Peaks marked * belong to CF3c00H- radical anions.)124 E.S.R. OF 7-IRRADIATED RCF2COOH suggesting interaction with an additional spin-3 nucleus. The magnitude of the interaction suggests that this is most likely to be fluorine from a neighbouring CF3COOH molecule, probably sited along the axis of the carbon 2pz orbital of the nearly planar CF3 radical. The less well resolved structure visible in fig. 3(b) and 3(c) no doubt indicates similar interaction with neighbouring fluorine atoms in less well defined positions : thermal annealing permits relaxation to the preferred configuration of fig. 3(d). SODIUM TRIFLUOROACETATE The e.s.r. spectrum observed in y-irradiated sodium trifluoroacetate has been discussed previously." It consists of features very similar to those of CF3 radicals in the CF3COOH/H20 matrices though the centre of the spectrum is obscured by the very much larger signal from cF,COO- radicals.g* l1 The CF3 radicals disappear on heating to 160 K leaving only cF2COO- radicals which are stable at room temp- erature.The temperature dependence of the fluorine parameters in these radicals has been described elsewhere.'' There is no trace at any temperature of the doublet spectrum attributed in the irradiated acids to the radical-anion. THE R A DI CA L - A NION CF3COOH- In their study of the radiolysis of CF3COOH Chachaty and Shiotani l 7 attribute the double doublet (a, = 4.8 mT, a2 = 1.6 mT) observed at 77 K to the radical anion CF3cOOH- but they also associate the 4.8 mT doublet features separated by 30 mT with this species, i.e.they interpret the central double doublet as the MI = 0 transitions of a species with two equivalent fluorine atoms ( A F ~ 15 m") and a third fluorine atom with AT = 4.8 mT and All <line width (- 1.5 mT), further split by interaction with a proton (a = 1.6 mT). The wing features then correspond to the MI = 1 transitions of two equivalent fluorine atoms doubled by the proton interaction. We believe that this interpretation is incorrect for the following reasons : (1) the hyperfine coupling of 15 mT is unacceptably high for the 8-fluorine atoms in CFJcOOH- (Chachaty and Shiotani '' quote the relationship a; = -2.0+8.0 cos2 8 where 8 is the dihedral angle between the carbon 2pz orbital and the C-F bond for a c-C-F species and the anisotropy of /?-fluorine atoms is much less than that of a-fluorine atoms), (2) it is very difficult to visualise the kind of motional averaging of the 8-F hyperfine coupling which will result in the paramdas quoted, even if a maximum value of 15 mT were permitted, (3) the wing features at about +_ 15 mT are more readily assigned to CF3 radicals as outlined earlier and are present in numerous systems when the radical-anions have disappeared (see e.g.fig. 3(d)), (4) these wing features are singlets in CF,COOH/H20 matrices and in (CF,CO),O (see fig. 1 and 2), i.e. they do not reflect the central doublet splitting of the radical anion in all cases. We believe that a much simpler interpretation is possible which is compatible with the relatively isotropic nature of the #I-F hyperfine interaction and the relationship ar = -2.0+8.0 cos2 6.20 We suggest that the main doublet splitting (4.1 mT in (CF3C0)20, 5.2 mT in CF,COOH, 4.6 mT in CF,COOH/H,O) is attributable to the interaction of one 8-fluorine and the minor splitting (about 1.6 mT in all samples) to a second B-fluorine.The coupling with the third fluorine is less than the linewidth (1-1.5 mT). This interpretation is reasonable for a species cF3COOH- with one C-F bond nearly parallel to the 2p orbital of the carbon atom in the carboxyl group, the precise angle depending on the matrix, with a fractional unpaired spin 0.7-0.8 on the carboxyl carbon atom. Such an interpretation is very similar to that generally accepted for the radical anions of aliphatic carboxylic acids and esters.' 3-1P .B . AYSCOUGH AND K . MACH 125 PERFLUOROPROPANOIC ACID The spectra of the radicals trapped in y-irradiated CzFSCOOH have been described earlier lo so'only a brief summary is needed here. The initial spectra at 77 K can be interpreted in terms of the radicals CF3CF2 and CFJeFCOOH in the manner used to describe the spectra of c F 3 and CF,COOH radicals earlier. It is not certain whether radical-anions are present or not since there are no major unidentifiable features in the central region : no photolyses were carried out. The CF3cF, radicals are characterised by positive and negative wing features typical of RCF, radicals in polycrystalline media as in u.v.-irradiated (C2FSC00)4Pb.9 The wing peaks are doublets with 5.8 mT separation indicating, we believe, interaction with one of the three p-fluorines in a fixed CF3 group, the other two having hyperfine couplings less than the line width.(Note the similarity in hyperfine coupling to that of CF3COOH- described earlier.) The CF,CFCOOH radicals are characterised by wing peaks with separation between 22.6 and 19.9 mT depending on temperature and matrix, typical of RCFCOOH or RCFCOO- radicals. The wing peaks in the pure and aqueous acid are quartets with an average splitting of about 1.45 mT, rather lower than that for the p-fluorines of a rotating CF3 group (about 2 mT). In samples of CF3CF2COOH/H20 between 173 and 190 K, at which temperature the radicals disappear, these quartets are clearly binomial : in the sodium salt above 113 K they are also very well resolved with relative intensities 1 : 3 : 3 : 1 and peak-to-peak separation of 1.8 mT.In the sodium salt these radicals are stable to at least 270 K. Our interpretation again differs in some respects from that of Chachaty and Shiotani but is fully supported by detailed studies of varying temperature on the stability and mobility of the radicals. It is important to emphasise that the C2F5 radicals are lost at a much lower temperature than the secondary radicals in the pure acid, aqueous acid and sodium salt, so that there is always a temperature range in which the revers- ible changes in motional restriction of the CF3 group in CF3CFCOOH radicals can be studied separately from the irreversible changes which occur when the C2F5 radicals are lost.PERFLUOROBUTANOIC ACID The radicals CF3CF2cF2 and CF,CF,cFCOOH are readily identified in the e.s.r. spectra recorded immediately after y-irradiation at 77 K. For the C3F7 radicals the wing features separated by 45.6mT are again doublets but the doublet splitting is only 2.4 mT compared with 5.8 mT in C2F,. The overall width is typical of RCF, radicals and the doublet splitting is reasonable for one p-fluorine obeying the cos2 8 relationship quoted earlier ; the second fluorine has a hyperfine splitting less than the line width. In the sodium salt the corresponding splittings are 44.0 and 2.2 mT but in these samples at 77 K the secondary radicals CF3CF2cFCOOH are in great excess. These latter radicals show characteristic wing features separated by 22.4 mT in the acid and 21.8 mT in the sodium salt at 77 K.In the acid they are doublets with a splitting of 5.0 mT, attributed to interaction with one #l-fluorine as in C3F7 radicals, but in the sodium salt there are additional peaks suggesting that more than one configuration is possible. This suggestion is supported by a number of thermally reversible changes in the wing structure reported earlier.1o The new evidence relating to the mechanism of the radiolysis is concerned with photolysis of the y-irradiated samples. It is seen in fig. 4(a) that the central region of the spectrum has the appearance of a broad doublet of about 5 mT splitting super- imposed on a broad single peak. When heated to 143 K and recorded at 77 K the central peaks have diminished in size and the wing features attributed to CF3CF2- CFCOOH radicals have increased, while those belonging to C3F7 radicals are126 E .S . R . OF 7-IRRADIATED RCF2COOH unchanged (fig. 4(b)). When another sample was photolysed for 10 min at 77 K the central doublet was removed (fig. 4(c)). Subsequent heating to 143 K produced no further change, i.e. no additional CF,CF,cFCOOH radicals were formed. This is convincing evidence that the central doublet arises from the radical anion C3F,- COOH- and that this species is the precursor of CF,CF,CFCOOH radicals by means of the reaction CF3CF2CF2c00H--, CF,CF$FCOOH + F-. 31 0 320 330 340 field strength/mT FIG. 4.-(a) C3F7COOH after y-irradiation at 77 K, (6) the same sample after heating to 143 K, measured at 77 K, (c) as (a) followed by 10 min photolysis at 77 M.(Peaks marked * belong to radical anions, those marked t belong to RR’CF radicals.) PERFLUOROOCTANOIC ACID The behaviour of this acid is essentially the same as that of perfluorobutanoic acid, i.e. the main species present at 77 K are RcF, radicals and the parent radical- anion which again gives rise to a pair of broad lines with splitting about 5 mT. These peaks can be removed by photolysis at 77 K. Thermal annealing to 143 K results in the formation of secondary radicals RRcF in the non-photolysed samples but not in samples which have been subjected to U.V. irradiation.P . B . AYSCOUGH AND K . MACH 127 PERFLUOROSUCCINIC ACID As reported earlier the main radicals seen in y-irradiated perfluorosuccinic acid ((CF,COOH),) at 77K are cF2CF2COOH and a lesser amount of HOOCCF2- CFCOOH.In the sodium salt the yields of the corresponding species cF2CF2COO- and -OOCCF,cFCOO- are approximately equal. Thus the general pattern of behaviour is similar to that of the monocarboxylic acids. However, all these radicals are exceptionally stable so that both RcF, and RR’cF radicals are present still at 373 K. As a result of the extended temperature range available for study one ob- serves complex, mainly reversible, changes in the wing features of both RcF, and RR’cF radicals on thermal annealing which have been attributed to changes in the extent of hindrance of rotational motion brought about by relaxation of the matrix.l0 Unfortunately the complications brought about by these effects obscure changes which might be related to chemical effects and this particular study tells us little about the mechanism of the radiolysis.No photolytic studies were carried out. 310 320 330 340 field strength/mT FIG. 5.-(n) (CF2),(COOH)2/H20 after y-irradiation at 77 K, (b) the same sample after heating to 143 K, measured at 77 K, (c) as (a) followed by 10 min photolysis at 77 K. (Peaks marked * belong to radical anions, those marked f belong to RR’CF radicals.)128 E.S.R. OF ?-IRRADIATED RCF,COOH PERFLUOROGLUTARIC ACID In contrast to perflurosuccinic acid there is clear evidence in perfluoroglutaric acid ([CF,]3(COOH)2) for the presence of the parent radical-anion in the y-irradiated pure and aqueous acid samples at 77 K.The main features of fig. 5(a) are 1 : 2 : 1 triplets in the wings, corresponding to the parallel features of RCF, radicals, separated by 44.4 mT (triplet splitting is 3.8 mT) and a broad poorly resolved central triplet similar to that seen in C3F,COOH. A small yield of RR'cF radicals is indicated by the presence of wing features separated by 21.8 mT. Many more RR'cF radicals are formed when the sample is heated to 143 K (fig. 5(b)). However, if the thermal annealing is preceeded by 10 min photolysis at 77 K the central peaks change in such a way as to indicate the removal of an approximately 5 mT doublet, and the subsequent thermal annealing brings about no increase in the yield of R R c F radicals. DISCUSSION On the basis of the work reported in this and earlier papers 9-11 it is possible to make some generalisations about the radiolysis of fluorinated carboxylic acids which are compatible with studies of radiolysis products l 2 and with the known behaviour of fluoroalkyl radicals.(1) Although it is not generally possible to interpret every feature of the e.s.r. spectra observed in polycrystalline fluorinated compounds we can be reasonably certain that all the spectra observed in these studies can be attributed to three different kinds of radical. Thus for the species RCF,COOH (R = F, CF3, CF3CF2, C6F13, CF,COOH, CF2CF2COOH) the only paramagnetic species observed at 77K or higher temperatures are radical anions RCF,cOOH-, fluoroalkyl radicals RCF, and secondary radicals formed by loss of fluorine, generally RCFCOOH.(2) The radical anions are characterised by e.s.r. spectra which are broad doublets of roughly 5 mT splitting, resulting from hyperfine interaction with one p-fluorine in a fixed orientation ; sometimes additional poorly resolved structure is seen which may be attributed to interaction with a second fluorine atom. Photolysis of samples containing radical-anions results in removal of the radical-anions without apparently replacement by any other paramagnetic species. We attribute this behaviour to photoejection from the radical-anion of either a hydrogen atom or an electron (which subsequently forms a hydrogen atom when it encounters any protonated species in the matrix). Hydrogen atoms are not trapped at 77 K, nor can they abstract fluorine, so they presumably combine.In contrast, the thermal decomposition of RCF2- COOH- appears to result in the loss of F- : this accounts for the presence of small amounts of RCFCOOH radicals in most samples at 77 K and large amounts in all samples when raised to rather higher temperatures. In the cases of trifluoroacetic, perfluorobutanoic, perfluorooctanoic and perfluoroglutaric acids there is clear evi- dence of the conversion of RCF,COOH- to RCFCOOH on thermal annealing. (3) Fluoroalkyl radicals RCF, are always present in the y-irradiated samples at 77 K. They are less rigidly trapped than the secondary radicals and disappear at lower temperature than the latter when samples are heated. The temperature range in which they are lost is, coincidentally, approximately the same as that in which the radical-anions are converted to secondary radicals giving rise to an earlier suggestion that the fluoroalkyl radicals abstracted fluorine atoms from the parent acids.That this is not so is shown by our own work on the behaviour of fluoroalkyl radicals in u.v.-irradiated lead perfluoroalkan~ates,~ by the absence of CF4 in the products from the radiolysis of CF3COOH l 2 and is confirmed by the present work which showsP . B . AYSCOUGH AND K . MACH 129 that when the radical-anions are removed by photolysis the thermal annealing of RCF, radicals is not accompanied by any increase in the RR'CF spectra. (4) Relative yields of RcF, and RCFCOOH radicals cannot be measured with any accuracy in polycrystalline samples because the spectra overlap extensively but gross differences in yield can be discerned. These have been described in more detail in the earlier papers lo* l1 but they appear to have little correlation with the structure of the acid and we are unable to draw any useful conclusion from these results.(5) As in the case of aliphatic carboxylic acids 1 3 9 l4 there is no direct e.s.r. evidence for the fate of the positive species RCF,cOOH+ which we presume to be the product of the primary ionisation. However, we use the analogy of other carboxylic acids to propose the sequence of reactions RCF2COOH++ RCFzCOOH-+RCF2COOH~ +RCF2COO-+RCF2 + C02 which leads to the fluoroalkyl radicals seen in all samples. This mechanism is supported by the evidence of Betts and Cherniak l 2 who found that C 0 2 was the main gaseous product from the radiolysis of CF,COOH (liquid or solid) and that the only fluorocarbon gaseous products were C2F6 and CF3H.(6) We have found no evidence of the presence of fluoroacyl radicals RCF2c0 which, by analogy with other carboxylic acids, might be intermediate radicals. However, the complexity and asymmetry of the central regions of all the spectra recorded to 77K was such that we would have great difficulty in identifying any contribution from such species. M. T. Rogers and L. D. Kispert, J. Chem. Phys., 1967,46,3193. R. J. Lontz and W. Gordy, J. Chem. Phys., 1962,37, 1357. F. G. Herring, W. C. Lin and M. R. Mustafa, Canad. J. Chem., 1970, 48,447. M. Iwasaki, K. Toriyama and B. Eda, J. Chem. Phys., 1965,42,63. M. Iwasaki, J.Chem. Phys., 1966,45,991. M. Iwasaki, S. Noda and K. Toriyama, Mol. Phys., 1970, 18,201. 'I J. Maruani, C. A. McDowell, H. Nakajima and P. Raghunathan, Mol. Phys., 1968, 14, 349. * J. Maruani, J. A. R. Coope and C. A. McDowell, Mol. Phys., 1970,18, 165. P. B. Ayscough, J. Machova and K. Mach, J.C.S. Faraday I, 1973, 69, 750. l o K. Mach, Coll. Czech. Chem. Comm., 1972,37,663. l 1 K. Mach, Coll. Czech. Chem. Comm., 1972, 37, 923. l2 J. Betts and E. A. Cherniak, Canad. J. Chem., 1971,49, 3389. l3 P. B. Ayscough, K. Mach, J. P. Oversby and A. K. Roy, Trans. Faraday SOC., 1971, 67, 360. l4 P. B. Ayscough and J. P. Oversby, Trans. Faraday SOC., 1971, 67, 1635. P. B. Ayscough and J. P. Oversby, J.C.S. Faraday I, 1972, 68, 1153. l6 Y. Nakajima, S. Sat0 and S. Shida, Bull.Chem. SOC. Japan, 1969,42,2132. C. Chachaty and M. Shiotani, J. Chim. phys., 1971, 66, 300. l 8 F. D. Srygley and W. Gordy, J. Chem. Phys., 1967,46, 2245. l9 see e.g. R. J. Lontz, J. Chem. Phys., 1966, 45, 1339. 2o C. Chachaty, A. Forchioni and M. Shiotani, J. Chem., 1970, 48,435. 1-5P . B . AYSCOUGH AND K . MACH 129 that when the radical-anions are removed by photolysis the thermal annealing of RCF, radicals is not accompanied by any increase in the RR'CF spectra. (4) Relative yields of RcF, and RCFCOOH radicals cannot be measured with any accuracy in polycrystalline samples because the spectra overlap extensively but gross differences in yield can be discerned. These have been described in more detail in the earlier papers lo* l1 but they appear to have little correlation with the structure of the acid and we are unable to draw any useful conclusion from these results.(5) As in the case of aliphatic carboxylic acids 1 3 9 l4 there is no direct e.s.r. evidence for the fate of the positive species RCF,cOOH+ which we presume to be the product of the primary ionisation. However, we use the analogy of other carboxylic acids to propose the sequence of reactions RCF2COOH++ RCFzCOOH-+RCF2COOH~ +RCF2COO-+RCF2 + C02 which leads to the fluoroalkyl radicals seen in all samples. This mechanism is supported by the evidence of Betts and Cherniak l 2 who found that C 0 2 was the main gaseous product from the radiolysis of CF,COOH (liquid or solid) and that the only fluorocarbon gaseous products were C2F6 and CF3H. (6) We have found no evidence of the presence of fluoroacyl radicals RCF2c0 which, by analogy with other carboxylic acids, might be intermediate radicals.However, the complexity and asymmetry of the central regions of all the spectra recorded to 77K was such that we would have great difficulty in identifying any contribution from such species. M. T. Rogers and L. D. Kispert, J. Chem. Phys., 1967,46,3193. R. J. Lontz and W. Gordy, J. Chem. Phys., 1962,37, 1357. F. G. Herring, W. C. Lin and M. R. Mustafa, Canad. J. Chem., 1970, 48,447. M. Iwasaki, K. Toriyama and B. Eda, J. Chem. Phys., 1965,42,63. M. Iwasaki, J. Chem. Phys., 1966,45,991. M. Iwasaki, S. Noda and K. Toriyama, Mol. Phys., 1970, 18,201. 'I J. Maruani, C. A. McDowell, H. Nakajima and P. Raghunathan, Mol.Phys., 1968, 14, 349. * J. Maruani, J. A. R. Coope and C. A. McDowell, Mol. Phys., 1970,18, 165. P. B. Ayscough, J. Machova and K. Mach, J.C.S. Faraday I, 1973, 69, 750. l o K. Mach, Coll. Czech. Chem. Comm., 1972,37,663. l 1 K. Mach, Coll. Czech. Chem. Comm., 1972, 37, 923. l2 J. Betts and E. A. Cherniak, Canad. J. Chem., 1971,49, 3389. l3 P. B. Ayscough, K. Mach, J. P. Oversby and A. K. Roy, Trans. Faraday SOC., 1971, 67, 360. l4 P. B. Ayscough and J. P. Oversby, Trans. Faraday SOC., 1971, 67, 1635. P. B. Ayscough and J. P. Oversby, J.C.S. Faraday I, 1972, 68, 1153. l6 Y. Nakajima, S. Sat0 and S. Shida, Bull. Chem. SOC. Japan, 1969,42,2132. C. Chachaty and M. Shiotani, J. Chim. phys., 1971, 66, 300. l 8 F. D. Srygley and W. Gordy, J. Chem. Phys., 1967,46, 2245.l9 see e.g. R. J. Lontz, J. Chem. Phys., 1966, 45, 1339. 2o C. Chachaty, A. Forchioni and M. Shiotani, J. Chem., 1970, 48,435. 1-5 Electron Spin Resonance Studies of Elementary Processes in Radiation- and Photo-chemistry Part 1 Z.-f.-Fluorinated Carboxylic Acids at Cryogenic Temperatures BY PETER B. AYSCOUGH'~ Department of Physical Chemistry, University of Leeds, Leeds LS2 9JT KAREL MACH J. Heyrovsky Institute of Physical Chemistry and Electrochemistry, Prague, Czechoslovakia AND Received 5th July, 1973 Some information concerning the mechanism of the radiolysis of perfhorinated carboxylic acids is presented from an examination of the e.s.r. spectra of radicals trapped in polycrystalline samples of the acids between 77 and 300 K. Three types of radicals have been identified iny-irradiated samples of RCF,COOH (R = F, CF3, CF3CF2, C6HI3, CF2COOH, CF2CF2COOH).These are (1) parent radical anions RCF&OOH-, (2) primary fluoralkyl radicals RcF2, and (3) secondary radicals RCFCOOH. The radical anions lose F- on thermal annealing to yield additional RCFCOOH radicals, but when samples containing RCF,COOH- are u.v.-irradiated at 77 K these species disappear without the appearance of any identifiable radical product, as do the corresponding species in non- fluorinated carboxylic acids. The RCF, radicals are believed to originate from the cationic species formed in the initial ionization following proton transfer and elimination of C02. One of the reasons for the scarcity of applications of e.s.r. to the study of the radiolysis of fluorinated compounds is the very great complexity of the e.s.r.spectra obtained in the solid state. This is a consequence of the considerable anisotropy in the fluorine hyperfine tensors which have been established for the radicals CF,, CF,CONH, etc. in single cry~tals.~-~ The simplification which results from the use of single crystals cannot, unfortunately, be taken advantage of in the case of the fluorinated carboxylic acids and their salts which are the subjects of the present study. The basis for identification of radicals containing one, two or three a-fluorine atoms is the existence of characteristic features on the outer edges of the spectra corresponding to the maximum values of the total a-fluorine hyperfine coupling tensor X&(max).Thus, according to Iwasaki et aZ.,4-6 radicals with one a-fluorine show, in the f3st derivative spectra, positive and negative absorption peaks separated by about 20 mT, corresponding to the extremities of the overall spectrum. Those with two a-fluorines give spectra with these outermost features separated by 40-45 mT, while CF3 radicals have an overall width of about 75 mT, depending on matrix and temperature. The much more intense central structure in all of these spectra is interpreted only with great difficulty because the other principal components of the a-fluorine hyperfine coupling tensor are small (about 2mT), may be positive or negative, and are similar in magnitude to the fluorine nuclear Zeeman interaction at typical magnetic fields. Nevertheless some computations have been made and agreement with experimental spectra is sati~factory.~.t Part 11, P. B. Ayscough and K. J. Olsen, J.C.S. Furuduy I, 1972, 68, 1635. 118P . B . AYSCOUGH AND K . MACH 119 We reported previously the e.s.r. spectra of the fluoralkyl radicals CF3, CHF,, CH2F, CHFCOO-, CF,COO-, CF3CF, and CF3CF2CF2 obtained by the photolysis of polycrystalline Pb(1V) fluoroalkanoates. These showed the characteristic features described above for fluorine nuclei in the a-position and displayed additional structure in the wing peaks which could be attributed to interaction of F and H nuclei in the B-position. Although no detailed interpretation of the central features was attempted it seems reasonable, in view of the simplicity of the chemical behaviour of these systems, to use these spectra as the basis for identification of similar radicals in the y-irradiated carboxylic acid and salts.One of us has described in detail the e.s.r. spectra observed in a number of y- irradiated fluorinated carboxylic acids and their sodium salts at cryogenic tempera- tures.l0’ l1 The main features of the spectra observed at 77 K in the acids RCF,- COOW were attributed to perfluoroalkyl radicals RCF, which were replaced COM- pletely or in part following thermal annealing by secondary radicals of a type which suggested loss of fluorine from the parent acid. In the sodium salts some secondary radicals were present at the temperature of the y-irradiation but again the primary radicals disappeared at a lower temperature than the secondary radicals.These observations were interpreted in terms of fluorine abstraction by the primary per- fluoroalkyl radicals and a tentative mechanism was put forward. However, some further observations and a re-examination of central regions of the spectra, together with a report on the products of the radiolysis of trifluoroacetic acid,12 have thrown further light on the mechanism of the radiolysis. EXPERIMENTAL The procedure for y-irradiation at 77 K, for recording the e.s.r. spectra and for thermal annealing of samples has been fully described.l3? l4 In this work the total dose of 6oCo y-rays was typically 2 to 4 x 1020 eV ~ r n - ~ . Photolysis of the y-irradiated samples was made using a high-pressure mercury lamp giving radiation mainly of wavelengths 300-400 nm. All the fluorinated carboxylic acies were supplied by Koch-Light Laboratories and were used without further purification.The sodium salts were prepared by controlled neutralisation of aqueous solutions of the acids and then recrystallised. The pure acids freeze to a poly- crystalline state but aqueous solutions containing 30-50 % by volume of water generally give transparent glasses which were more suitable for photolysis. All samples were y-irradiated in Spectrosil high-purity quartz tubes after freeze-pumping to a pressure < mmHg. The very narrow peak at g = 2.00 in the e.s.r. spectrum attributed to radiation damage in the quartz is readily recognisable. The very narrow peaks with 50.2 mT separation also observed in all samples at 77 K are caused by hydrogen atoms trapped in the quartz.RESULTS Six fluorinated ca.rboxylic acids and their sodium salts were examined after y- irradiation at 77 K and during subsequent thermal and photochemical annealing. In some cases frozen aqueous solutions of the acids were also examined as an aid to the interpretation of the e.s.r. spectra and in one case, trifluoroacetic acid, the anhyd- ride was studied. we used the characteristic wing features corresponding to the summation CAr(max) for radicals containing one, two or three a-fluorine atoms to identify the trapped radicals. In this way we have established that in all the acids RCF,COOH and salts RCF,COONa studied, primary fluoroalkyl radicals RCF, are present immediately after irradiation at 77 K. In the sodium salts additional radicals with one less a-fluorine than the parent compound predom- inate at 77 K, and species of this kind assume great importance in the irradiated Following established procedures 5 9 6,120 E.S.R.OF 7-IRRADIATED RCFZCOOH acids after thermal annealing. Although the central portions of the spectra are not easily interpreted in detail because of the complexity of the spectra in polycrystalline samples, there are clear indications of the presence of at least one further type of radical in the pure acids and in the frozen aqueous solutions. We believe that these are radical-anions RCF,cOOH-, similar to those already reported in many carboxylic acids and esters at cryogenic temperatures. 3-1 Chachaty and Shiotani reached the same conclusion in relation to CF3COOH, CF3CF2COOH and (CF2COOH)2, though we believe that the hyperfine parameters attributed to the radical-anions in these systems are incorrect.Srygley and Gordy l8 also suggested the presence of radical-anions in the ammonium salt CF3COONH4 though again we cannot agree with the spectral assignments quoted to support this hypothesis. TRIFLUOROACETIC ACID The behaviour of this compound on y-irradiation appears to be typical of fully fluorinated acids and it has been examined in the greatest detail. We have studied the anhydride, the pure polycrystalline acid and frozen aqueous solutions of the acid, each subjected to thermal and photochemical annealing. THE ANHYDRIDE The spectra observed in the pure polycrystalline anhydride at 77K and after bleaching with visible light at 77 K are showii in fig.1. The spectra ire conveniently discussed in three sections (i) the weak outermost peaks with positive and negative bands separated by 70-75mT, (ii) an intermediate region with complex structure separated by 25-30 mT, (iii) the much more intense central region which has a pro- nounced doublet structure with splittings of 5-9 mT. Interpretation is greatly helped by the work of Maruani and co-workers ’* who used the values for the principal components of the fluorine hyperfine tensor derived for CF3 radicals in single crystal CF3CONH2 to compute the spectra in an inert random matrix. The tensor is nearly axially symmetric with A f ~ 2 5 . 3 and A f ~ 9 mT : thus one expects pronounced discontinuities to appear in the polycrystalline spectra at magnetic fields corresponding approximately to $-+A!, ++Af;, ,+AT and +$AT, the perpendicular features being much more intense.Second-order transitions will be superimposed on the *+ transitions (approximately 1.8 mT for the perpendicular transitions and 0.5 mT for the parallel transitions), and there is some g-anisotropy, so the central part of the polycrystalline spectra will be somewhat unsymmetrical. l Maruani et al. found excellent agreement between experimental spectra for CF3 radicals in a krypton matrix at 4.2 K and computed spectra assuming randomly oriented, non-rotating, non-inverting radicals. We have observed essentially similar spectra for CF, radicals in u.v.-irradiated (CF3C00)4Pb at 77 K (see fig. l(c)). It is apparent that the main features of the spectrum shown in fig. l(b) are also well described by this model which predicts major features at & 37.9, f 12.6, + 13.5 and k4.5 mT referred to the centre of the spectrum. The outermost peaks in fig.l(b) are separated by 76.0 mT while the peaks in the intermediate region are separated by about 27 mT. The central region shows peaks at -6.0, -4.3, +5.0 and +6.0 mT corresponding to the $AF, transitions with second-order splitting and g-anisotropy superimposed. The very close similarity between fig. l(b) and l(c) leaves no doubt that the only radicals present in y-irradiated (CF3C0)20 at 77 K after photobleaching are CF3 radicals (the narrow peaks at k25.1 mT are caused by H atoms in the quartz of the sample tube). The diffaence between fig.l(a) and l(b) is clearly due to the presence of a strong doublet of doublets with splittings 4.1 and 1.7 mT which can beP . B . AYSCOUGH AND K . MACH 121 removed by photoylsis at 77K without leaving any other paramagnetic species. This is a characteristic of the radical-anions found in y-irradiated carboxylic acids and esters I3-l6 and greatly facilitates the interpretation of these complex spectra. We shall consider the interpretation of the hyperfine parameters in more detail after reporting the results for the pure acid and the aqueous samples. * I I I 1 1 31 0 330 350 field strength/mT FIG. l.--(a) Pure (CF3CO)20 after y-irradiation at 77 K, (b) the same sample after 15 min photolysis at 77 K, (c) (CF3C00)4Pb after 10 min photolysis at 77 K.(Radical anion peaks are marked *.) AQUEOUS TRIFLUOROACETIC ACID When an approximately 50 % by volume solution of CF,COOH is y-irradiated at 77 K the initial spectrum observed is that shown in fig. 2(a). The similarity to fig. l(a) is apparent (the outer and intermediate regions are shown at higher gain in fig. 2) but the central doublet is less well resolved and gives a hyperfine splitting of about 4.6 mT (cf. 4.1 mT in the anhydride). In addition some small peaks at +_ 19.8 mT are attributable to a small concentration of cF,COOH radicals. Photolysis at 77 K leads to the removal of the central doubIet, as in the anhydride, and the appearance of a rather narrow singlet at the centre of the spectrum (see fig. 2(b)-(e). The latter is attributed to the central peak (MI = 0 transitions) of the triplet122 E .S . R . OF Y-IRRADIATED RCFZCOOH spectrum of rotating cF,COOH radicals, seen more clearly in the sodium salts (see e.g. fig. 4 of ref. (10)). On the evidence of the outer peaks at 19.8 mT there is no increase in concentration of cF,COOH radicals during photobleaching so we must conclude that the central peak in the unbleached sample is sufficiently broad at 77 K 310 330 350 FIG. 2.-(u) 6 mol dm-3 CF3COOH/Hz0 after y-irradiation at 77 K, (6) the same sample after 30 s photolysis, measured after 10 min at 77 K, (c) the same sample after 200 s photolysis, measured at 77 K immediately after photolysis, (d) the same sample after 20 min at 77 K, (e) the same sample after 20 min photolysis at 77 K. (Peaks marked .T belong to cFzCOOH radicals, those marked * belong to CF3cOOH- radical anions.)P .B . AYSCOUGH AND K . MACH 123 as to be obscured by the very much stronger doublet of the radical-anion. During the photobleaching there are also some minor changes in the spectrum of CF3 radicals, shown most clearly by the broadening and shifting of the outermost peaks. This effect is to some extent reversible since the final spectrum after 25 min photolysis (fig. 2(e)) has an overall width almost identical with that measured initially (fig. 2(a)). The whole spectrum disappears at about 170 K without further change in structure. PURE POLYCRYSTALLINE TRIFLUOROACETIC ACID The main features of the spectrum shown in fig. 3(a) and observed immediately after y-irradiation at 77 K are similar to those of fig.2(a) and may be attributed to CF3 and cF3cOOH- mainly, with traces of CF,COOH radicals. The radical-anion doublet splitting is about 5.2 mT and the outer peaks of the CF3 spectrum are broad, with separation 75-79 mT. Photobleaching for 15 min removes the radical-anion, leaves the cF,COOH spectrum unchanged but improves the resolution of the outer- most CF3 radical features (fig. 3(b)). Thermal annealing to 130 K further improves resolution (fig. 3(c)) and after heating to 180 K and recording at 77 K (fig. 3(4) the outermost CF3 peaks are clearly doublets, slightly asymmetric, with separation 4.5 and 5.0 mT. In the spectrum the features in the intermediate region are also doubled ir the same way and some additional similar structure appears in the central region, I I I I 1 31 0 330 350 field strength/mT FIG.3.-(a) Pure CF,COOH after y-irradiation at 77 K, (6) after 15 min photolysis at 77 K, (c) the same sample after heating to 130 K, measured at 77 K, ( d ) the same sample after heating to 180 K, measured at 77 K. (Peaks marked * belong to CF3c00H- radical anions.)124 E.S.R. OF 7-IRRADIATED RCF2COOH suggesting interaction with an additional spin-3 nucleus. The magnitude of the interaction suggests that this is most likely to be fluorine from a neighbouring CF3COOH molecule, probably sited along the axis of the carbon 2pz orbital of the nearly planar CF3 radical. The less well resolved structure visible in fig. 3(b) and 3(c) no doubt indicates similar interaction with neighbouring fluorine atoms in less well defined positions : thermal annealing permits relaxation to the preferred configuration of fig.3(d). SODIUM TRIFLUOROACETATE The e.s.r. spectrum observed in y-irradiated sodium trifluoroacetate has been discussed previously." It consists of features very similar to those of CF3 radicals in the CF3COOH/H20 matrices though the centre of the spectrum is obscured by the very much larger signal from cF,COO- radicals.g* l1 The CF3 radicals disappear on heating to 160 K leaving only cF2COO- radicals which are stable at room temp- erature. The temperature dependence of the fluorine parameters in these radicals has been described elsewhere.'' There is no trace at any temperature of the doublet spectrum attributed in the irradiated acids to the radical-anion.THE R A DI CA L - A NION CF3COOH- In their study of the radiolysis of CF3COOH Chachaty and Shiotani l 7 attribute the double doublet (a, = 4.8 mT, a2 = 1.6 mT) observed at 77 K to the radical anion CF3cOOH- but they also associate the 4.8 mT doublet features separated by 30 mT with this species, i.e. they interpret the central double doublet as the MI = 0 transitions of a species with two equivalent fluorine atoms ( A F ~ 15 m") and a third fluorine atom with AT = 4.8 mT and All <line width (- 1.5 mT), further split by interaction with a proton (a = 1.6 mT). The wing features then correspond to the MI = 1 transitions of two equivalent fluorine atoms doubled by the proton interaction. We believe that this interpretation is incorrect for the following reasons : (1) the hyperfine coupling of 15 mT is unacceptably high for the 8-fluorine atoms in CFJcOOH- (Chachaty and Shiotani '' quote the relationship a; = -2.0+8.0 cos2 8 where 8 is the dihedral angle between the carbon 2pz orbital and the C-F bond for a c-C-F species and the anisotropy of /?-fluorine atoms is much less than that of a-fluorine atoms), (2) it is very difficult to visualise the kind of motional averaging of the 8-F hyperfine coupling which will result in the paramdas quoted, even if a maximum value of 15 mT were permitted, (3) the wing features at about +_ 15 mT are more readily assigned to CF3 radicals as outlined earlier and are present in numerous systems when the radical-anions have disappeared (see e.g.fig. 3(d)), (4) these wing features are singlets in CF,COOH/H20 matrices and in (CF,CO),O (see fig.1 and 2), i.e. they do not reflect the central doublet splitting of the radical anion in all cases. We believe that a much simpler interpretation is possible which is compatible with the relatively isotropic nature of the #I-F hyperfine interaction and the relationship ar = -2.0+8.0 cos2 6.20 We suggest that the main doublet splitting (4.1 mT in (CF3C0)20, 5.2 mT in CF,COOH, 4.6 mT in CF,COOH/H,O) is attributable to the interaction of one 8-fluorine and the minor splitting (about 1.6 mT in all samples) to a second B-fluorine. The coupling with the third fluorine is less than the linewidth (1-1.5 mT). This interpretation is reasonable for a species cF3COOH- with one C-F bond nearly parallel to the 2p orbital of the carbon atom in the carboxyl group, the precise angle depending on the matrix, with a fractional unpaired spin 0.7-0.8 on the carboxyl carbon atom.Such an interpretation is very similar to that generally accepted for the radical anions of aliphatic carboxylic acids and esters.' 3-1P . B . AYSCOUGH AND K . MACH 125 PERFLUOROPROPANOIC ACID The spectra of the radicals trapped in y-irradiated CzFSCOOH have been described earlier lo so'only a brief summary is needed here. The initial spectra at 77 K can be interpreted in terms of the radicals CF3CF2 and CFJeFCOOH in the manner used to describe the spectra of c F 3 and CF,COOH radicals earlier. It is not certain whether radical-anions are present or not since there are no major unidentifiable features in the central region : no photolyses were carried out.The CF3cF, radicals are characterised by positive and negative wing features typical of RCF, radicals in polycrystalline media as in u.v.-irradiated (C2FSC00)4Pb.9 The wing peaks are doublets with 5.8 mT separation indicating, we believe, interaction with one of the three p-fluorines in a fixed CF3 group, the other two having hyperfine couplings less than the line width. (Note the similarity in hyperfine coupling to that of CF3COOH- described earlier.) The CF,CFCOOH radicals are characterised by wing peaks with separation between 22.6 and 19.9 mT depending on temperature and matrix, typical of RCFCOOH or RCFCOO- radicals. The wing peaks in the pure and aqueous acid are quartets with an average splitting of about 1.45 mT, rather lower than that for the p-fluorines of a rotating CF3 group (about 2 mT).In samples of CF3CF2COOH/H20 between 173 and 190 K, at which temperature the radicals disappear, these quartets are clearly binomial : in the sodium salt above 113 K they are also very well resolved with relative intensities 1 : 3 : 3 : 1 and peak-to-peak separation of 1.8 mT. In the sodium salt these radicals are stable to at least 270 K. Our interpretation again differs in some respects from that of Chachaty and Shiotani but is fully supported by detailed studies of varying temperature on the stability and mobility of the radicals. It is important to emphasise that the C2F5 radicals are lost at a much lower temperature than the secondary radicals in the pure acid, aqueous acid and sodium salt, so that there is always a temperature range in which the revers- ible changes in motional restriction of the CF3 group in CF3CFCOOH radicals can be studied separately from the irreversible changes which occur when the C2F5 radicals are lost.PERFLUOROBUTANOIC ACID The radicals CF3CF2cF2 and CF,CF,cFCOOH are readily identified in the e.s.r. spectra recorded immediately after y-irradiation at 77 K. For the C3F7 radicals the wing features separated by 45.6mT are again doublets but the doublet splitting is only 2.4 mT compared with 5.8 mT in C2F,. The overall width is typical of RCF, radicals and the doublet splitting is reasonable for one p-fluorine obeying the cos2 8 relationship quoted earlier ; the second fluorine has a hyperfine splitting less than the line width.In the sodium salt the corresponding splittings are 44.0 and 2.2 mT but in these samples at 77 K the secondary radicals CF3CF2cFCOOH are in great excess. These latter radicals show characteristic wing features separated by 22.4 mT in the acid and 21.8 mT in the sodium salt at 77 K. In the acid they are doublets with a splitting of 5.0 mT, attributed to interaction with one #l-fluorine as in C3F7 radicals, but in the sodium salt there are additional peaks suggesting that more than one configuration is possible. This suggestion is supported by a number of thermally reversible changes in the wing structure reported earlier.1o The new evidence relating to the mechanism of the radiolysis is concerned with photolysis of the y-irradiated samples. It is seen in fig.4(a) that the central region of the spectrum has the appearance of a broad doublet of about 5 mT splitting super- imposed on a broad single peak. When heated to 143 K and recorded at 77 K the central peaks have diminished in size and the wing features attributed to CF3CF2- CFCOOH radicals have increased, while those belonging to C3F7 radicals are126 E . S . R . OF 7-IRRADIATED RCF2COOH unchanged (fig. 4(b)). When another sample was photolysed for 10 min at 77 K the central doublet was removed (fig. 4(c)). Subsequent heating to 143 K produced no further change, i.e. no additional CF,CF,cFCOOH radicals were formed. This is convincing evidence that the central doublet arises from the radical anion C3F,- COOH- and that this species is the precursor of CF,CF,CFCOOH radicals by means of the reaction CF3CF2CF2c00H--, CF,CF$FCOOH + F-.31 0 320 330 340 field strength/mT FIG. 4.-(a) C3F7COOH after y-irradiation at 77 K, (6) the same sample after heating to 143 K, measured at 77 K, (c) as (a) followed by 10 min photolysis at 77 M. (Peaks marked * belong to radical anions, those marked t belong to RR’CF radicals.) PERFLUOROOCTANOIC ACID The behaviour of this acid is essentially the same as that of perfluorobutanoic acid, i.e. the main species present at 77 K are RcF, radicals and the parent radical- anion which again gives rise to a pair of broad lines with splitting about 5 mT. These peaks can be removed by photolysis at 77 K. Thermal annealing to 143 K results in the formation of secondary radicals RRcF in the non-photolysed samples but not in samples which have been subjected to U.V.irradiation.P . B . AYSCOUGH AND K . MACH 127 PERFLUOROSUCCINIC ACID As reported earlier the main radicals seen in y-irradiated perfluorosuccinic acid ((CF,COOH),) at 77K are cF2CF2COOH and a lesser amount of HOOCCF2- CFCOOH. In the sodium salt the yields of the corresponding species cF2CF2COO- and -OOCCF,cFCOO- are approximately equal. Thus the general pattern of behaviour is similar to that of the monocarboxylic acids. However, all these radicals are exceptionally stable so that both RcF, and RR’cF radicals are present still at 373 K. As a result of the extended temperature range available for study one ob- serves complex, mainly reversible, changes in the wing features of both RcF, and RR’cF radicals on thermal annealing which have been attributed to changes in the extent of hindrance of rotational motion brought about by relaxation of the matrix.l0 Unfortunately the complications brought about by these effects obscure changes which might be related to chemical effects and this particular study tells us little about the mechanism of the radiolysis.No photolytic studies were carried out. 310 320 330 340 field strength/mT FIG. 5.-(n) (CF2),(COOH)2/H20 after y-irradiation at 77 K, (b) the same sample after heating to 143 K, measured at 77 K, (c) as (a) followed by 10 min photolysis at 77 K. (Peaks marked * belong to radical anions, those marked f belong to RR’CF radicals.)128 E.S.R. OF ?-IRRADIATED RCF,COOH PERFLUOROGLUTARIC ACID In contrast to perflurosuccinic acid there is clear evidence in perfluoroglutaric acid ([CF,]3(COOH)2) for the presence of the parent radical-anion in the y-irradiated pure and aqueous acid samples at 77 K.The main features of fig. 5(a) are 1 : 2 : 1 triplets in the wings, corresponding to the parallel features of RCF, radicals, separated by 44.4 mT (triplet splitting is 3.8 mT) and a broad poorly resolved central triplet similar to that seen in C3F,COOH. A small yield of RR'cF radicals is indicated by the presence of wing features separated by 21.8 mT. Many more RR'cF radicals are formed when the sample is heated to 143 K (fig. 5(b)). However, if the thermal annealing is preceeded by 10 min photolysis at 77 K the central peaks change in such a way as to indicate the removal of an approximately 5 mT doublet, and the subsequent thermal annealing brings about no increase in the yield of R R c F radicals.DISCUSSION On the basis of the work reported in this and earlier papers 9-11 it is possible to make some generalisations about the radiolysis of fluorinated carboxylic acids which are compatible with studies of radiolysis products l 2 and with the known behaviour of fluoroalkyl radicals. (1) Although it is not generally possible to interpret every feature of the e.s.r. spectra observed in polycrystalline fluorinated compounds we can be reasonably certain that all the spectra observed in these studies can be attributed to three different kinds of radical. Thus for the species RCF,COOH (R = F, CF3, CF3CF2, C6F13, CF,COOH, CF2CF2COOH) the only paramagnetic species observed at 77K or higher temperatures are radical anions RCF,cOOH-, fluoroalkyl radicals RCF, and secondary radicals formed by loss of fluorine, generally RCFCOOH. (2) The radical anions are characterised by e.s.r.spectra which are broad doublets of roughly 5 mT splitting, resulting from hyperfine interaction with one p-fluorine in a fixed orientation ; sometimes additional poorly resolved structure is seen which may be attributed to interaction with a second fluorine atom. Photolysis of samples containing radical-anions results in removal of the radical-anions without apparently replacement by any other paramagnetic species. We attribute this behaviour to photoejection from the radical-anion of either a hydrogen atom or an electron (which subsequently forms a hydrogen atom when it encounters any protonated species in the matrix).Hydrogen atoms are not trapped at 77 K, nor can they abstract fluorine, so they presumably combine. In contrast, the thermal decomposition of RCF2- COOH- appears to result in the loss of F- : this accounts for the presence of small amounts of RCFCOOH radicals in most samples at 77 K and large amounts in all samples when raised to rather higher temperatures. In the cases of trifluoroacetic, perfluorobutanoic, perfluorooctanoic and perfluoroglutaric acids there is clear evi- dence of the conversion of RCF,COOH- to RCFCOOH on thermal annealing. (3) Fluoroalkyl radicals RCF, are always present in the y-irradiated samples at 77 K. They are less rigidly trapped than the secondary radicals and disappear at lower temperature than the latter when samples are heated.The temperature range in which they are lost is, coincidentally, approximately the same as that in which the radical-anions are converted to secondary radicals giving rise to an earlier suggestion that the fluoroalkyl radicals abstracted fluorine atoms from the parent acids. That this is not so is shown by our own work on the behaviour of fluoroalkyl radicals in u.v.-irradiated lead perfluoroalkan~ates,~ by the absence of CF4 in the products from the radiolysis of CF3COOH l 2 and is confirmed by the present work which showsP . B . AYSCOUGH AND K . MACH 129 that when the radical-anions are removed by photolysis the thermal annealing of RCF, radicals is not accompanied by any increase in the RR'CF spectra. (4) Relative yields of RcF, and RCFCOOH radicals cannot be measured with any accuracy in polycrystalline samples because the spectra overlap extensively but gross differences in yield can be discerned. These have been described in more detail in the earlier papers lo* l1 but they appear to have little correlation with the structure of the acid and we are unable to draw any useful conclusion from these results. (5) As in the case of aliphatic carboxylic acids 1 3 9 l4 there is no direct e.s.r. evidence for the fate of the positive species RCF,cOOH+ which we presume to be the product of the primary ionisation. However, we use the analogy of other carboxylic acids to propose the sequence of reactions RCF2COOH++ RCFzCOOH-+RCF2COOH~ +RCF2COO-+RCF2 + C02 which leads to the fluoroalkyl radicals seen in all samples. This mechanism is supported by the evidence of Betts and Cherniak l 2 who found that C 0 2 was the main gaseous product from the radiolysis of CF,COOH (liquid or solid) and that the only fluorocarbon gaseous products were C2F6 and CF3H. (6) We have found no evidence of the presence of fluoroacyl radicals RCF2c0 which, by analogy with other carboxylic acids, might be intermediate radicals. However, the complexity and asymmetry of the central regions of all the spectra recorded to 77K was such that we would have great difficulty in identifying any contribution from such species. M. T. Rogers and L. D. Kispert, J. Chem. Phys., 1967,46,3193. R. J. Lontz and W. Gordy, J. Chem. Phys., 1962,37, 1357. F. G. Herring, W. C. Lin and M. R. Mustafa, Canad. J. Chem., 1970, 48,447. M. Iwasaki, K. Toriyama and B. Eda, J. Chem. Phys., 1965,42,63. M. Iwasaki, J. Chem. Phys., 1966,45,991. M. Iwasaki, S. Noda and K. Toriyama, Mol. Phys., 1970, 18,201. 'I J. Maruani, C. A. McDowell, H. Nakajima and P. Raghunathan, Mol. Phys., 1968, 14, 349. * J. Maruani, J. A. R. Coope and C. A. McDowell, Mol. Phys., 1970,18, 165. P. B. Ayscough, J. Machova and K. Mach, J.C.S. Faraday I, 1973, 69, 750. l o K. Mach, Coll. Czech. Chem. Comm., 1972,37,663. l 1 K. Mach, Coll. Czech. Chem. Comm., 1972, 37, 923. l2 J. Betts and E. A. Cherniak, Canad. J. Chem., 1971,49, 3389. l3 P. B. Ayscough, K. Mach, J. P. Oversby and A. K. Roy, Trans. Faraday SOC., 1971, 67, 360. l4 P. B. Ayscough and J. P. Oversby, Trans. Faraday SOC., 1971, 67, 1635. P. B. Ayscough and J. P. Oversby, J.C.S. Faraday I, 1972, 68, 1153. l6 Y. Nakajima, S. Sat0 and S. Shida, Bull. Chem. SOC. Japan, 1969,42,2132. C. Chachaty and M. Shiotani, J. Chim. phys., 1971, 66, 300. l 8 F. D. Srygley and W. Gordy, J. Chem. Phys., 1967,46, 2245. l9 see e.g. R. J. Lontz, J. Chem. Phys., 1966, 45, 1339. 2o C. Chachaty, A. Forchioni and M. Shiotani, J. Chem., 1970, 48,435. 1-5

ISSN:0300-9599    

DOI:10.1039/F19747000118

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Series Methods in the Analysis of Kinetic Data for Chemical Sys terns BY PAUL A. ADAMS,”~ AND JOHN G. SHEPPARD Dept. of Chemistry, University of Rhodesia, P.O. Box MP 167, Salisbury, Rhodesia Receiued 3 1 st July, 1973 Application of series methods to the evaluation of rate coefficients has been considered for reac- tions in which explicit integration of the system of dependent differential equations is impossible The approximation of explicit integrals for such systems is considered, and a detailed study of the system : k i k2 Xl+XZ xi +x2+x3 carried out using simulation techniques. Methods of analysing rate data for the above system are considered in detail, and it is shown that series methods provide precise methods for the estimation of both kl and k2 in the above system.Estimation of rate coefficients characterising individual steps in chemical reaction networks has recently received considerable attention. Methods discussed concern mainly the extraction of rate coefficients from experi- mental data on reaction networks where no explicit solution of the system of depend- ent differential equations is possible. In many cases, even where such explicit solutions are possible, the form of the solution is such that experimental data cannot be easily analysed using the explicit functions, to give rate coefficienk6 Since many comparatively simple networks are included in the above statement, rate coefficient extraction is a general and important problem facing the experimental kineticist. Little attention seems to have been paid to series methods as a tool for rate coefficient evaluation.This is somewhat surprising since series methods are simple to apply (even to complex reaction networks), and, excepting cases to be discussed, are generally applicable. In this paper we consider three aspects of series methods applied to chemical kinetics problems : (1) the derivation of the series solution ; (2) the application of the series solution ; (3) the use of series solutions in the approximation of explicit integrals in chemical kinetics. These aspects will be considered in the main by the application of series methods to the system (1) for which no explicit representation of XI, X2, or XJ, as functions of time, is possible k i kz X P X 2 XI +XZ+XS. (1) The application of series methods to the analysis of rate data for complex reaction t present address : Dept.of Chemistry, University College London, 20, Gordon St., London WCl HOAJ. 130P. A . ADAMS A N D J . G . SHEPPARD 131 systems seems to have been first carried out by Ingold in his study of the hydrolysis of diesters.' More recently Darvey et al. have applied series methods to the determina- tion of the number of intermediates in multi-intermediate enzymic systems.* It is interesting to note that the kinetic determinant method of Erofeev provides an identical method to that of Darvey et al. for the analysis of multi-intermediate system^.^ The estimation of apparent first order rate constants in kinetic systems using poly- nomial methods, has recently been considered by Maltby and Johnson.lO The method used by Darvey et al.to derive the series solutions is standard and simple to apply, it will therefore be used in this study in preference to the successive approximation method due to Ingold which is difficult to apply to complex systems, and represents time as a power series in concentration rather than the more useful reverse. THE METHOD OF SERIES ANAYLSIS The set of differential rate equations arising from (1) are given in (2), the operator " D " denoting the operation d/dt, t representing time. DXlW = -k,X,(t) -kzX1(t)X2(t) DX,(t) = kiXi(t) -k2Xl(t)Xz(t) = DXi(t) + 2kiXi(t) DX3(0 = k2Xl(OX2(t). (2) It is now assumed that the concentration of the species X i (in the case considered i = 1-3) can be represented by the Stirling (Maclaurin) form of Taylor's expansion, i.e.I1 +...D ~ X X O ) ~ ~ ~ 3 x ~ ( o ) t 3 + 3! Xi(t) = Xi(0)+ DXi(0)t-t- 2! (3) The conditions under which the expansion (3) is justified will be discussed later. In order to express the concentration of all chemical species in the form (3), we must evaluate X,(O), DXi(0), D2X,(0), etc. to give as many terms as axe required. In the present case only the series for X,(t), and X2(t), will be evaluated, however, evaluation of the series for X,(t) is simple and obvious. Using (2) we can write the nth differential coefficients of X1, X2, and X3 in the general form (4) D"Xl(0) = - klD"-'X1(0) - k2Dn-'~1(0)X2(O)] D"X2(0) = D"Xl(0) +2klD"-lX1(0) (4) D"X3(0) = k2D"-1Bl(0)X2(0)]. The nth differential coefficients of the products [X,(O)X,(O)] are obtained using Leibnitz's theorem (9, D'[ Xl(0)X2(O)] = r!s! fl DrX1(0) D"X2(0) the summation in (5) being extended to all positive integral values of r and s inclusive of zero, which satisfy (r+s) = n.The coefficients in the series (3) are evaluated in the following manner for increasing n : n = 0 Xl(0) = Xy X,(O) = 0 DX,(O) = - klX; - k2X:X; = -klX; DXZ(0) = k,Xi n = 1132 SERlES METHODS I N KINETIC ANALYSIS n = 2 D2X1(0) = -klDX1(0)-Ccz[DX1(0)+ DX,(O)] = klXy D2X2(0) = -k:Xy D3X1(0) = - k l D2X1(0)-k2[ D2X1(0)+2 DXl(0) DX,(O)+ D2X2(0)] t~ = 3 = - k:Xy+2k:k2(Xy)2 D ~ X ~ ( O ) = k:xy+2k:k,(~;)~ etC. its the term in t 4 in eqn (6) and (7) The series solutions for the concentrations of X1 and X2 with time are given as far k;X yt (k:X - 2 k k2( X;1)2)t X,(t) = X;-k,X;t+-- + 2! 3! ( k f X; - 4k:kz(X:)2 - 8k:k2( X;1)2)t4 4! kIX;t2 (k:X; +2k:k2(X;)2)t3 - X,(t) = k,X,”t- ~ + 2! 3! (k:X,O +4k:k;(X;)’ +4k:k2(X;)’)t4 4! -.- (7) The process can of course be continued as far as one wishes, however, the rapid increase in the complexity of the coefficients makes extension past the first four or five terms impractical. APPLICATION OF THE SERIES RESULT Rate coefficients can be extracted from experimental data in principle by fitting polynomials to the data and evaluating the rate coefficients by direct comparison of the experimental and derived series. This is the procedure suggested by Maltby and Johnson lo; however, although of value, it may in certain circumstances not be the best method available. Consider the series (6) and (7) obtained for system (l), it can easily be seen that a plot of X2(t)/t against time will give a curve of initia1 slope - k ;Xy/2, and intercept at zero time equal to k,X;, allowing the estimation of both k, and Xy.Moreover, as will be seen in the next section, it is possible to estimate kl/k2 for the system (1) using the final concentration attained by Xz, therefore for this sys- tem, polynomial regression is not the best method for the analysis of experimental data. This is not to say that polynomial regression is of no value in the analysis of data by series methods. Application of series methods to the system studied by Ingold,’ and subsequent comparison of the coefficients of polynomials fitted to the experi- mental data with the Coefficients of the series solution polynomials, gives rate con- stants within 1 % of the values obtained by Ingold for the hydrolysis of diethyl sucqinate.Thus polynomial regression, combined with a term by term comparison of the experimental polynomial and the derived time series polynomial, can lead to accurate evaluation of individual rate constants. This procedure has been used successfully by the authors to obtain the three rate constants in the acid-catalysed hydrolysis of glycerol triacetate, measuring only the appearance of acetic acid with time. In essence therefore, each case must be considered on its own merits regarding the application of the derived series solutions, indeed the series approximation to systemP . A . ADAMS AND J . G . SHEPPARD 133 (1) provides a further, much more powerful method of data analysis for that system than the two methods just discussed.This will now be considered. APPROXIMATION OF EXPLICIT SOLUTIONS TO RATE EQUATIONS It can be shown for system (1) that k,, k2, X , ( t ) , and X , ( t ) , are related by expres- sion @ ) . I 3 a - X A t ) X2(t)+2u In ~ = X , ( t ) - X: a where a = kl/k2. Subtraction of the two series (6), and (7), gives : 2k;Xyt2 2k:X;lt3 2! 3! X,(t)- X , ( t ) = Xi-2k,X,"t+ ~ - -~ + (9) 2(k:X,"- 4k:kz(X;)2)t4 4 ! -+... To a close approximation, the series (9) can be represented by (10) thus (8) becomes, on substitution of (10) x,(t) - x 2 ( t ) 2 2X~e-~l' -Xi (10) Eqn (1 1) is an explicit relationship between the concentration of X2 and time, the " goodness " of the approximation depending on how closely eqn (10) represents the concentration diffaence X,(t) - X,(t).In order to check the closeness of the approximation (lo), system (I) was simulated using a general numerical integration program developed at the University of Rhod- esia. Three sets of rate constants were used to cover the three cases kl > k2, k , = k,, and k, <k2. The appropriate plots of concentration against time are shown in fig. 1 , and illustrate the fact that at XI(?) equal to zero (i.e. complete reaction) X 2 ( t ) approaches a limiting value.14 Fig. 2 shows the first order plot of In (Xl(t)- X z ( t ) + X y ) against time as suggested by eqn (10). In all cases the function has been x2 1 -*-'I,------ - x3 0 5 . 0 10.0 time/s (4 time/s (6) 1-c O.! I 1x1 \ time/s (c) FIG.1.-(u) kl = 1.0 s-I, kz = 0.1 dm3 mol-I s-' : limiting value X&) = 0.9115 when X,(t) = 0.0. ( b ) kl = 1.0 s-', k2 = 1.0 dm3 mol-' s-' : limiting value X&) = 0.5361 when X , ( t ) = 0.0. (c) kl = 1.0 s-', k2 = 10.0 dm3 mol-' s-' : limiting value X 2 ( t ) = 0.099 59 when X , ( i ) = 0.0.134 SERIES METHODS IN KINETIC ANAL,YSTS plotted to > 60 % of the limiting X , ( t ) concentration. The excellent linearity of the three plots strongly suggests that (1 0) approximates the concentration difference X , ( t ) - X , ( t > , very closely for all ratios kl/k2. The deviation of the function from the predicted value is found in all three cases to be less than 0.5 % at 60 74 reaction, and less than 2 % at 80 % reaction. Y - O I timels (4 \ timels (b) timels (c) FIG.2.-Validity of eqn (10). Solid line corresponds to the relationship Y = ln(2)- t, Y = ln(X,(r) -X&)+X;). Final point shown corresponds to (a) 63 % reaction ; (b) 60 % reaction ; (c) 62 % reaction. The pairs of rate constants employed are in the same order as in fig. 1. 1 I e 0 0 c Xz(t)lim(Xl(t)/a-+o) FIG. 3.-Relationship between a and the limiting concentration of X 2 ( t ) for mechanism (1) : X; = 1 .oo. From eqn (10) the slope of the lines shown in fig. 2(a)-(c) should be equal to - k , allowing estimation of both kl and k2 for system (l), since the ratio kl/k2 can be obtained from the limiting value of X 2 ( t ) when XI = 0.0. The relationship between kl /k2 and the limiting value of X 2 ( t ) is calculated from theP . A. ADAMS AND J .G . SHEPPARD rearranged form of eqn (8) shown below. The value of X , ( t ) (8), in order to give (12). a - Mt) X2(t)/a+2 In ~ = - X:/a. a 135 is set equal to zero in The only unknown in eqn (12) is a, and we can therefore solve by the method of successive approximations to obtain the values of a corresponding to various limiting values of X,(t) at known values of Xy. Fig. 3 shows the plot of a against Xz(t)li,(X,(t)l,.+O) with X; normalised to unity. DISCUSSION The validity of the method of series analysis applied to chemical reaction net- works rests on the conditions under which the expansion (3) is valid. The three conditions under which the Stirling form of Taylor’s Theorem fails are : (1) that Xi(0), or one of its differential coefficients, becomes infinite between the values of the var- iable considered; (2) that X,(O), or one of its differential coefficients, becomes dis- continuous between the same values; (3) that the series does not approach a finite limit, i.e.the remainder cannot be made to vanish as n approaches infinity. Since all chemical reactions must approach a limiting value, governed if nothing more than by the amounts of reactant available, condition three presents no problem. However, certain chemical reactions show both infinite rates and discontinuities, attempted application of series methods to the analysis of these systems must therefore be invalid. An example of the case in point, is the variation of reaction rate with respect to pressure for a mixture of oxygen and hydrogen at constant temperat~re,’~ infinite rates and discontinuities being exhibited.In the great majority of cases, however, series analysis is possible and the analysis of systems involving reactions of any order can be carried out using the generalised Leibnitz Theorem (13), where summation is carried out so that (r+s+ t + . . .) = n, for all positive integral values of r, s, t, etc., including zero. From the series derived, rate constants can be evaluated if data are available on the concentration against time variation of the reactants and products. With modern techniques such data is normally available, and therefore series methods provide a general method of rate constant evaluation applicable to a large number of chemical systems. We are indebted to Eng P. Ridler of the University of Rhodesia for allowing us to use his general numerical integration program in order to carry out the simulation studies.B. Saville, J. Phys. Chem., 1971, 75, 2215. I. D. Gay, J. Phys. Chem., 1971, 75,1610. G. R. Howe, J. Phys. Chem., 1971, 75,1319. J. C. H. Chen and W. D. Huntsman, J. Phys. Chem., 1971,75,430. G. E. P. Box, W. G. Hunter, J. F. MacGregor and J. Erjavec, Technometrics, 1973,15, No. 1,33. Jen-Yuan Chien, J. Amer. Chem. Soc., 1948, 70, 2256. C. K. Ingold, J. Chem. Sac., 1931,2173. B. V. Erofeev, Zhur. fiz. Khim., 1950, 24,721. * I. G. Darvey, S. J. Prokhovnik and J. F. Williams, J. Theor. Biol., 1966, 13,96.136 SERIES METHODS IN KINETIC ANALYSIS l o M. C. Maltby and W. D. Johnson, Austral. J. Chem., 1971, 24,2417. l 2 P. A. Adams and J.G. Sheppard, unpublished results. l3 S. W. Benson, J. Chem. Phys., 1952, 20, 1605. l4 Comprehensive Chemical Kinetics, ed. C . H. Bamford and C. F. H. Tipper (Elsevier, 1969), l 5 H. W. Thompson and C. N. Hinshelwood, Pruc. Roy. SOC. A, 1929,122,610. J. Edwards, A Treatise on the Diferential Calculus (Macmillan, 1898, 81. vol. 2, p. 57. Series Methods in the Analysis of Kinetic Data for Chemical Sys terns BY PAUL A. ADAMS,”~ AND JOHN G. SHEPPARD Dept. of Chemistry, University of Rhodesia, P.O. Box MP 167, Salisbury, Rhodesia Receiued 3 1 st July, 1973 Application of series methods to the evaluation of rate coefficients has been considered for reac- tions in which explicit integration of the system of dependent differential equations is impossible The approximation of explicit integrals for such systems is considered, and a detailed study of the system : k i k2 Xl+XZ xi +x2+x3 carried out using simulation techniques.Methods of analysing rate data for the above system are considered in detail, and it is shown that series methods provide precise methods for the estimation of both kl and k2 in the above system. Estimation of rate coefficients characterising individual steps in chemical reaction networks has recently received considerable attention. Methods discussed concern mainly the extraction of rate coefficients from experi- mental data on reaction networks where no explicit solution of the system of depend- ent differential equations is possible. In many cases, even where such explicit solutions are possible, the form of the solution is such that experimental data cannot be easily analysed using the explicit functions, to give rate coefficienk6 Since many comparatively simple networks are included in the above statement, rate coefficient extraction is a general and important problem facing the experimental kineticist.Little attention seems to have been paid to series methods as a tool for rate coefficient evaluation. This is somewhat surprising since series methods are simple to apply (even to complex reaction networks), and, excepting cases to be discussed, are generally applicable. In this paper we consider three aspects of series methods applied to chemical kinetics problems : (1) the derivation of the series solution ; (2) the application of the series solution ; (3) the use of series solutions in the approximation of explicit integrals in chemical kinetics.These aspects will be considered in the main by the application of series methods to the system (1) for which no explicit representation of XI, X2, or XJ, as functions of time, is possible k i kz X P X 2 XI +XZ+XS. (1) The application of series methods to the analysis of rate data for complex reaction t present address : Dept. of Chemistry, University College London, 20, Gordon St., London WCl HOAJ. 130P. A . ADAMS A N D J . G . SHEPPARD 131 systems seems to have been first carried out by Ingold in his study of the hydrolysis of diesters.' More recently Darvey et al. have applied series methods to the determina- tion of the number of intermediates in multi-intermediate enzymic systems.* It is interesting to note that the kinetic determinant method of Erofeev provides an identical method to that of Darvey et al.for the analysis of multi-intermediate system^.^ The estimation of apparent first order rate constants in kinetic systems using poly- nomial methods, has recently been considered by Maltby and Johnson.lO The method used by Darvey et al. to derive the series solutions is standard and simple to apply, it will therefore be used in this study in preference to the successive approximation method due to Ingold which is difficult to apply to complex systems, and represents time as a power series in concentration rather than the more useful reverse. THE METHOD OF SERIES ANAYLSIS The set of differential rate equations arising from (1) are given in (2), the operator " D " denoting the operation d/dt, t representing time.DXlW = -k,X,(t) -kzX1(t)X2(t) DX,(t) = kiXi(t) -k2Xl(t)Xz(t) = DXi(t) + 2kiXi(t) DX3(0 = k2Xl(OX2(t). (2) It is now assumed that the concentration of the species X i (in the case considered i = 1-3) can be represented by the Stirling (Maclaurin) form of Taylor's expansion, i.e.I1 +... D ~ X X O ) ~ ~ ~ 3 x ~ ( o ) t 3 + 3! Xi(t) = Xi(0)+ DXi(0)t-t- 2! (3) The conditions under which the expansion (3) is justified will be discussed later. In order to express the concentration of all chemical species in the form (3), we must evaluate X,(O), DXi(0), D2X,(0), etc. to give as many terms as axe required. In the present case only the series for X,(t), and X2(t), will be evaluated, however, evaluation of the series for X,(t) is simple and obvious. Using (2) we can write the nth differential coefficients of X1, X2, and X3 in the general form (4) D"Xl(0) = - klD"-'X1(0) - k2Dn-'~1(0)X2(O)] D"X2(0) = D"Xl(0) +2klD"-lX1(0) (4) D"X3(0) = k2D"-1Bl(0)X2(0)].The nth differential coefficients of the products [X,(O)X,(O)] are obtained using Leibnitz's theorem (9, D'[ Xl(0)X2(O)] = r!s! fl DrX1(0) D"X2(0) the summation in (5) being extended to all positive integral values of r and s inclusive of zero, which satisfy (r+s) = n. The coefficients in the series (3) are evaluated in the following manner for increasing n : n = 0 Xl(0) = Xy X,(O) = 0 DX,(O) = - klX; - k2X:X; = -klX; DXZ(0) = k,Xi n = 1132 SERlES METHODS I N KINETIC ANALYSIS n = 2 D2X1(0) = -klDX1(0)-Ccz[DX1(0)+ DX,(O)] = klXy D2X2(0) = -k:Xy D3X1(0) = - k l D2X1(0)-k2[ D2X1(0)+2 DXl(0) DX,(O)+ D2X2(0)] t~ = 3 = - k:Xy+2k:k2(Xy)2 D ~ X ~ ( O ) = k:xy+2k:k,(~;)~ etC.its the term in t 4 in eqn (6) and (7) The series solutions for the concentrations of X1 and X2 with time are given as far k;X yt (k:X - 2 k k2( X;1)2)t X,(t) = X;-k,X;t+-- + 2! 3! ( k f X; - 4k:kz(X:)2 - 8k:k2( X;1)2)t4 4! kIX;t2 (k:X; +2k:k2(X;)2)t3 - X,(t) = k,X,”t- ~ + 2! 3! (k:X,O +4k:k;(X;)’ +4k:k2(X;)’)t4 4! -.- (7) The process can of course be continued as far as one wishes, however, the rapid increase in the complexity of the coefficients makes extension past the first four or five terms impractical. APPLICATION OF THE SERIES RESULT Rate coefficients can be extracted from experimental data in principle by fitting polynomials to the data and evaluating the rate coefficients by direct comparison of the experimental and derived series.This is the procedure suggested by Maltby and Johnson lo; however, although of value, it may in certain circumstances not be the best method available. Consider the series (6) and (7) obtained for system (l), it can easily be seen that a plot of X2(t)/t against time will give a curve of initia1 slope - k ;Xy/2, and intercept at zero time equal to k,X;, allowing the estimation of both k, and Xy. Moreover, as will be seen in the next section, it is possible to estimate kl/k2 for the system (1) using the final concentration attained by Xz, therefore for this sys- tem, polynomial regression is not the best method for the analysis of experimental data.This is not to say that polynomial regression is of no value in the analysis of data by series methods. Application of series methods to the system studied by Ingold,’ and subsequent comparison of the coefficients of polynomials fitted to the experi- mental data with the Coefficients of the series solution polynomials, gives rate con- stants within 1 % of the values obtained by Ingold for the hydrolysis of diethyl sucqinate. Thus polynomial regression, combined with a term by term comparison of the experimental polynomial and the derived time series polynomial, can lead to accurate evaluation of individual rate constants. This procedure has been used successfully by the authors to obtain the three rate constants in the acid-catalysed hydrolysis of glycerol triacetate, measuring only the appearance of acetic acid with time.In essence therefore, each case must be considered on its own merits regarding the application of the derived series solutions, indeed the series approximation to systemP . A . ADAMS AND J . G . SHEPPARD 133 (1) provides a further, much more powerful method of data analysis for that system than the two methods just discussed. This will now be considered. APPROXIMATION OF EXPLICIT SOLUTIONS TO RATE EQUATIONS It can be shown for system (1) that k,, k2, X , ( t ) , and X , ( t ) , are related by expres- sion @ ) . I 3 a - X A t ) X2(t)+2u In ~ = X , ( t ) - X: a where a = kl/k2. Subtraction of the two series (6), and (7), gives : 2k;Xyt2 2k:X;lt3 2! 3! X,(t)- X , ( t ) = Xi-2k,X,"t+ ~ - -~ + (9) 2(k:X,"- 4k:kz(X;)2)t4 4 ! -+...To a close approximation, the series (9) can be represented by (10) thus (8) becomes, on substitution of (10) x,(t) - x 2 ( t ) 2 2X~e-~l' -Xi (10) Eqn (1 1) is an explicit relationship between the concentration of X2 and time, the " goodness " of the approximation depending on how closely eqn (10) represents the concentration diffaence X,(t) - X,(t). In order to check the closeness of the approximation (lo), system (I) was simulated using a general numerical integration program developed at the University of Rhod- esia. Three sets of rate constants were used to cover the three cases kl > k2, k , = k,, and k, <k2. The appropriate plots of concentration against time are shown in fig.1 , and illustrate the fact that at XI(?) equal to zero (i.e. complete reaction) X 2 ( t ) approaches a limiting value.14 Fig. 2 shows the first order plot of In (Xl(t)- X z ( t ) + X y ) against time as suggested by eqn (10). In all cases the function has been x2 1 -*-'I,------ - x3 0 5 . 0 10.0 time/s (4 time/s (6) 1-c O.! I 1x1 \ time/s (c) FIG. 1.-(u) kl = 1.0 s-I, kz = 0.1 dm3 mol-I s-' : limiting value X&) = 0.9115 when X,(t) = 0.0. ( b ) kl = 1.0 s-', k2 = 1.0 dm3 mol-' s-' : limiting value X&) = 0.5361 when X , ( t ) = 0.0. (c) kl = 1.0 s-', k2 = 10.0 dm3 mol-' s-' : limiting value X 2 ( t ) = 0.099 59 when X , ( i ) = 0.0.134 SERIES METHODS IN KINETIC ANAL,YSTS plotted to > 60 % of the limiting X , ( t ) concentration. The excellent linearity of the three plots strongly suggests that (1 0) approximates the concentration difference X , ( t ) - X , ( t > , very closely for all ratios kl/k2.The deviation of the function from the predicted value is found in all three cases to be less than 0.5 % at 60 74 reaction, and less than 2 % at 80 % reaction. Y - O I timels (4 \ timels (b) timels (c) FIG. 2.-Validity of eqn (10). Solid line corresponds to the relationship Y = ln(2)- t, Y = ln(X,(r) -X&)+X;). Final point shown corresponds to (a) 63 % reaction ; (b) 60 % reaction ; (c) 62 % reaction. The pairs of rate constants employed are in the same order as in fig. 1. 1 I e 0 0 c Xz(t)lim(Xl(t)/a-+o) FIG. 3.-Relationship between a and the limiting concentration of X 2 ( t ) for mechanism (1) : X; = 1 .oo.From eqn (10) the slope of the lines shown in fig. 2(a)-(c) should be equal to - k , allowing estimation of both kl and k2 for system (l), since the ratio kl/k2 can be obtained from the limiting value of X 2 ( t ) when XI = 0.0. The relationship between kl /k2 and the limiting value of X 2 ( t ) is calculated from theP . A. ADAMS AND J . G . SHEPPARD rearranged form of eqn (8) shown below. The value of X , ( t ) (8), in order to give (12). a - Mt) X2(t)/a+2 In ~ = - X:/a. a 135 is set equal to zero in The only unknown in eqn (12) is a, and we can therefore solve by the method of successive approximations to obtain the values of a corresponding to various limiting values of X,(t) at known values of Xy. Fig. 3 shows the plot of a against Xz(t)li,(X,(t)l,.+O) with X; normalised to unity.DISCUSSION The validity of the method of series analysis applied to chemical reaction net- works rests on the conditions under which the expansion (3) is valid. The three conditions under which the Stirling form of Taylor’s Theorem fails are : (1) that Xi(0), or one of its differential coefficients, becomes infinite between the values of the var- iable considered; (2) that X,(O), or one of its differential coefficients, becomes dis- continuous between the same values; (3) that the series does not approach a finite limit, i.e. the remainder cannot be made to vanish as n approaches infinity. Since all chemical reactions must approach a limiting value, governed if nothing more than by the amounts of reactant available, condition three presents no problem.However, certain chemical reactions show both infinite rates and discontinuities, attempted application of series methods to the analysis of these systems must therefore be invalid. An example of the case in point, is the variation of reaction rate with respect to pressure for a mixture of oxygen and hydrogen at constant temperat~re,’~ infinite rates and discontinuities being exhibited. In the great majority of cases, however, series analysis is possible and the analysis of systems involving reactions of any order can be carried out using the generalised Leibnitz Theorem (13), where summation is carried out so that (r+s+ t + . . .) = n, for all positive integral values of r, s, t, etc., including zero. From the series derived, rate constants can be evaluated if data are available on the concentration against time variation of the reactants and products. With modern techniques such data is normally available, and therefore series methods provide a general method of rate constant evaluation applicable to a large number of chemical systems. We are indebted to Eng P. Ridler of the University of Rhodesia for allowing us to use his general numerical integration program in order to carry out the simulation studies. B. Saville, J. Phys. Chem., 1971, 75, 2215. I. D. Gay, J. Phys. Chem., 1971, 75,1610. G. R. Howe, J. Phys. Chem., 1971, 75,1319. J. C. H. Chen and W. D. Huntsman, J. Phys. Chem., 1971,75,430. G. E. P. Box, W. G. Hunter, J. F. MacGregor and J. Erjavec, Technometrics, 1973,15, No. 1,33. Jen-Yuan Chien, J. Amer. Chem. Soc., 1948, 70, 2256. C. K. Ingold, J. Chem. Sac., 1931,2173. B. V. Erofeev, Zhur. fiz. Khim., 1950, 24,721. * I. G. Darvey, S. J. Prokhovnik and J. F. Williams, J. Theor. Biol., 1966, 13,96.136 SERIES METHODS IN KINETIC ANALYSIS l o M. C. Maltby and W. D. Johnson, Austral. J. Chem., 1971, 24,2417. l 2 P. A. Adams and J. G. Sheppard, unpublished results. l3 S. W. Benson, J. Chem. Phys., 1952, 20, 1605. l4 Comprehensive Chemical Kinetics, ed. C . H. Bamford and C. F. H. Tipper (Elsevier, 1969), l 5 H. W. Thompson and C. N. Hinshelwood, Pruc. Roy. SOC. A, 1929,122,610. J. Edwards, A Treatise on the Diferential Calculus (Macmillan, 1898, 81. vol. 2, p. 57.

ISSN:0300-9599    

DOI:10.1039/F19747000130

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Study of the Oxidation of Molybdenum Surfaces by Energy Loss Spectroscopy Combined with Auger Electron Spectroscopy B Y TOMOJI KAWAI, KIMIO KUNIMORI, TAMOTSU KONDOW, TAKAHARU ONISHI AND KENZI TAMARU" Department of Chemistry, The University of Tokyo, Hongo, Bunkyo-Ku, Tokyo, Japan Received 3rd April, 1973 The oxidation of molybdenum as an evaporated film and of its annealed surface was investigated by Energy Loss Spectroscopy (ELS) combined with Auger Electron Spectroscopy (AES). The 10 eV and the lower energy loss peaks towards 5 eV were identified as the clean surface and the oxidized layer surface plasma loss peaks, respectively. The behaviours of these surface plasma loss spectra revealed that the oxidation of the evaporated film surface proceeded uniformly, rapidly and without any induction period under vacuum of lo-' Torr, whereas that of the annealed sample proceeded more slowly through " oxidized patches ", exhibiting a considerable induction period under the same vacuum conditions.These phenomena were interpreted with reference to the surface conditions and structures studied by the AES and ELS ; the fresh evaporated surface is clean and rough, whereas the annealed surface is smoothed by the heat-treatment and contaminated by sulphur and carbon which partially cover the active surfaces and give rise to the patchy surface oxidation. Energy analysis of scattered electrons has recently been shown to be an extremely useful tool for the study of solid ~urfaces,l-~ especially when various techniques are combined, for instance, low energy electron diffraction and Auger electron spectro- scopy (AES).5* Energy loss spectroscopy (ELS) is of particular interest since it is suitable for revealing the electronic structures of the surfaces of solids and adsorb- a t e ~ .~ ~ AES, on the other hand, reveals the atomic identities existing on the sur- faces. Therefore, simultaneous measurements of ELS and AES may provide a deeper understanding of the surface properties and reactions than is possible with either one alone. On a metal surface, a surface mode of plasma excitation (surface plasmon) exists and can be observed by ELS, preferably in a reflection geometry. The surface mode appears at a lower frequency than the corresponding bulk mode, and shifts and/or reduces its intensity as the surface becomes contaminated by foreign ~pecies.~ In one case, the energy loss peak due to the surface plasmon on the clean surface shifts in the direction of lower energy, approaching a definite position of energy loss as the contamination proceeds. In another case, this initial peak gradually vanishes while another peak appears at the final energy position of the previous case.7 In the oxidation of aluminium, the surface plasma loss peak adopted the former behaviour under high vacuum condition ( Torr), it behaved as in the latter case.2 These phenomena cannot be explained by the different collision number of the gaseous molecules on the surface, but could be closely connected with the mechanism of the surface oxidation processes itself, so 137 T ~ r r ) , ~ whereas under lower vacuum138 OXIDATION OF MO that it can be investigated through the analysis of the behaviours of the surface plasmon loss peaks with reference to the surface conditions.In this paper, the oxidation process of a molybdenum film has been studied by varying its surface properties, taking advantage of the combined ELS and AES techniques. H 60 40 20 I mass number FIG. 1 . 4 0 ) The block diagram of the electron spectrometer for the ELS and the AES. This is a modified Simpson-Kuyatt type with 180" hemispherical energy selector whose optical circle has 25 mm radius. (b) Mass spectrum of the residual gas during the measurements.T . KAWAI, K . KUNIMORI, T. KONDOW, T . ONISHI, K . TAMARU 139 EXPERIMENTAL A modified Simpson-Kuyatt type electron spectrometer was constructed for the measure- ments of ELS and AES (fig.l(a)). The monochromator, as well as the analyzer, was a 180" hemispherical electrostatic lens pair, average radius of 2.5 cm, equipped with retarding electrostatic lenses. The energy resolution (AEG 500 meV) was high and constant enough, throughout the energy range of the measurements, to obtain the true shapes of the energy loss spectra and the Auger electron spectra. The scattered electrons, after being velocity selected, were detected by a single electron counting method. The collision chamber and the main chamber of the spectrometer were differentially pumped. This permitted one to introduce foreign gases into the collision chamber without disturbing the spectrometer. An evaporated Mo film was prepared and annealed in the collision chamber without exposure to air.Molybdenum (H. Cross Co. Ltd. 99.96 % purity) was evaporated onto a Mo foil which was fixed on a sample holder in the centre of the chamber. The Mo foil was chemically etched, heated to 1400°C in uacuo, then alternatively exposed to oxygen and hydrogen and finally checked for surface impurities by AES examination. Both chambers were evacuated by oil diffusion and rotary pumps. In order to prevent these chambers from being contam- inated, a foreline alumina trap was used as well as an ordinary liquid N2 trap. The pressure was measured by Bayard-Alpert gauge and was of the order 1 x lo-' Torr during the measure- ments. A NEVA quadrupole mass spectrometer was used to mass-analyze the residual gases (H20>90 %) (fig.l(b)). The operating parameters were as follows : energy of the primary electron beam, 380 eV for the ELS and 1.5 keV for the AES, primary beam current, 2x lo-* A for the ELS and 3 FA for the AES ; the incident and scattering angles, 27" each. RESULTS AND DISCUSSION 1. OXIDATION OF A FRESHLY EVAPORATED M O FILM When Mo was instantaneously evaporated on a Mo foil under a pressure of 1.2 x Torr, the energy loss spectra exhibited distinct time-dependent changes as is demonstrated in fig. 2. The Mo foil itself had two loss peaks at 23 eV and 5 eV before the evaporation. In terms of a free electron model based on the assumption of 6 free electrons (4a5(5s)l per atom, the 23 eV peak is assigned to be a bulk plasmon e~citation.~. l o After evaporation two peaks appeared at 10 eV and 23 eV the 5 eV loss peak having disappeared.This 23 eV peak did not change during the measure- ments. On the other hand, the 10 eV peak shifted with time towards low energies and finally reached a position of 5 eV energy loss. It took about 30 min to accomplish the complete shift. The rate of the shift was rather faster at the beginning, without an induction period, decreasing gradually with time. The Auger spectra taken immediately after the evaporation exhibited peaks due to molybdenum and oxygen atoms together with very weak carbon peaks which became evident in the latter half of the measurements. Just after the evaporation the amount of surface carbon was less than a few percent of monolayer coverage, judged from the relative intensities of carbon and molybdenum and the change in their intensity.The exact amount of the surface oxygen at the initial stage of the oxidation was difficult to estimate because of the rapid oxidation rate, but the extrapolated value for the oxygen was estimated to be nearly the same as that for carbon. Considerable growth of a 5 15 eV oxygen peak with time indicates that " swface oxidation " proceeds, as shown in fig. 3(a). It is highly likely that the M o surface is oxidized by water, which was one of the major residual gases, as indicated by mass analysis of background gases in the vacuum chambers. Stern and Ferrell have theoretically predicted that the loss peak of surface plasmon undergoes a red-shift depending on the thickness of the oxidized layer formed on the140 OXIDATION OF M O 6) 38min 5) 30min 4 ) Z m i n 3) 12min 2) 6 min 1 ) 3 m i n after e va po rat i D n 8cfore evaporation 1 1 I 1 0 Ib 'io 3b 9 energy losslev FIG.2.-The variation of the characteristic energy loss spectra with time, for the freshly evaporated molybdenum surface. S ( b ) 0 1 v 160 ZbO 240 3b0 500 550 kinetic energy/eV FIG. 3.-Auger electron spectra from a molybdenum surface. (a) Freshly evaporated surface. (6) After anneaiing in hydrogen at 1.5 x Sulphur and carbon are ob- Torr, looO°C, for 20 min. served on this annealed surface.T . KAWAI, K . KUNIMORI, T . KONDOW, T . ONISHI, K . TAMARU 141 metal, provided that the layer is formed uniformly.7* l1 * It can, accordingly, be concluded that the time-dependent peak is attributable to surface plasmon excitation.The thickness of the layer was estimated from the equation of Stern and Ferrell, assuming its validity, and was plotted as a function of reaction time (see fig. 5). In conclusion, the thickness of the oxidized layer increased with time in a uniform manner over the whole surface on this evaporated sample, similar to the oxidation of an A1 film under high vacuum c~nditions.~ If the assignment of the 23 eV peak is correct, the surface plasma loss on the clean surface should be 23/,/2 = 16.5 eV on the basis of the free electron model. This theoretical prediction does not agree with our experimental results. If one or more interband transitions occur with the energy close to that of the surface plasmon excitation expected from the free electron model, the observed surface plasmon loss may deviate from the theoretical prediction of the free electron model.2. OXIDATION OF THE FILM ANNEALED IN HYDROGEN ATMOSPHERE The evaporated film, whose surface plasmon loss was 5eV, was annealed in hydrogen at 1.5 x Torr pressure and a temperature of 1000°C for 20 min. The loss spectra of the annealed specimen were examined at different times after the annealing (see fig. 4). The other experimental conditions were identical to the case of the evaporated film. The behaviours and the shapes of the surface plasmon were quite different from those of the evaporated sample. Immediately after the treat- ment, the 10 eV loss peak was found as well as that at 23 eV. As the surface oxidation proceeded, a new peak appeared at 5 eV and coexisted with the peak at 10 eV whose intensity was gradually decreasing.These two peaks did not shift with time but their intensities changed. The new peak at 5 eV corresponds to the 5 eV peak observed in the experiment on the freshly evaporated film, i.e., the loss peak by the surface plasma excitation of the Mo surfzce covered by a sufficient amount of the oxidized layer. This phenomenon shows that both the oxidized part and non-oxidized part coexist during the course of the oxidation, and suggests that the oxide nuclei grow at the surface layer, the oxidation proceeding at the boundary between the metal and its oxide. The AES of this specimen demonstrated the appearance of sulphur and carbon atoms immediately after the annealing as is shown in fig.3(6). Seemingly, sulphur and carbon had diffused from the inside onto the surface of the specimen during the annealing procedure. Considerable parts of this annealed surface are covered by carbon and sulphur judging from the reduction of the intensity of the molybdenum. The ratio of C and S peak areas to Mo AES peak area were 2.1 and 3.2 respectively. In addition, the relative intensity of the surface to the bulk loss on the fresh evaporated film was approximately 25 times larger than on the annealed one. This means that the surface area of the evaporated film may be larger than that of the annealed one, that is, the surface of the evaporated one may be rougher, though the contaminants may reduce their relative intensity to some extent.12* l 3 3" E+tanhkD 2&+(1 +cZ) tanh kD us = where us is the frequency of the surface plasmon when a surface oxidized layer of dielectric constant E is formed to thickness D ; up is the frequency of the bulk plasmon (fiup = 23 eV for Mo) ; k repre- sents the wave number of the surface wave excited in the solid by the incident electron and is obtained from the momentum and the energy conservation law ; k = mus cos Bltiko where 8 is the incident and scattering angle (27") and ko is the wave number of the incident electron.142 OXIDATION OF M O In fig.6, graph (a) and (b), the percentage ratio of the 5 eV peak to the sum of the 5 and the 10 eV one was plotted against time (t) and t 2 . There is an induction period of about 10 min. In the initial stage of the reaction, the ratio was proportional to t 2 but above a ratio of 50 :d it was expressed in terms of aft2 +/I, where a' and are I t 10) After one day 9) 58min 8) 52min 7) 46min 6 ) 38min 5) 32min 4) 25 min 3) 19 min 2 ) 13 min I ) 4 min a f t e r l e a t i n g up in H2, 1 .5 ~ mmHg 20 min 0 10 20 30 energy losslev FIG. 4.-The variation of the characteristic energy loss spectra with time, for the molybdenum surface annealed in hydrogen at 1.5 x Torr, 1000°C, for 20 min. FIG. 5.-The /O 10 20 30 40 50 elapsed timelniin thickness of the oxidized layer as a function of elapsed time after evayora tion.T. KAWAI, K . KUNIMORI, T . KONDOW, T . ONISHI, K . TAMARU 143 time-independent constants. These phenomena can be explained qualitatively as follows. Assuming that the diameter of the circular patches increases with a constant velocity and the size of every patch is approximately the same, the ratio may be roughly equal to ((n.nE2)/So)t2 before these patches overlap.Here n is the number of the active points, So is the total surface area and E represents the average growth velocity of the diameter. After the overlapping, the effective n may decrease. There- fore the slope is smaller than that at the initial stage. t P) elapsed timelmin 5 + square of the elapsed timelmin FIG. 6.-The percentage ratio of the 5 eV peak to the sum of the 5 and 10 eV ones as a function of (a) elapsed time after annealing in hydrogen in hydrogen and (b) square of the elapsed time. The different behaviours in the ELS and the AES on these different surfaces is of interest in the view of the correlation between the surface reaction and the surface properties.Considerations on the rates of the oxidation at the early stage (fig. 5(a), fig. 6(a)) revealed that the uniform oxidation of the Mo surface starts rapidly just after the evaporation, which suggests a large collisional cross section for the reaction between gases and the evaporated surface. The induction period of about 10 min on the annealed surface, on the other hand, implies that the oxidation scarcely proceeded at the beginning and after about 10min began to develop rapidly. The oxidation cross section is very small at the beginning. As to the surface properties, AES demonstrated that the fresh evaporated surface is clean and is scarcely contaminated by other elements, whereas the annealed surface is contaminated by considerable amounts of carbon and sulphur.Another difference between the films would be the geometric structure of the surfaces, i.e., the roughness factor of the evaporated surface is larger than that of the annealed one. Taking these facts into consideration, the different behaviours in Mo oxidation are interpreted as follows ; on the fresh evaporated surface which is rough and active,144 OXIDATION OF M O the H20 in the ambient gas attacks the surface, colliding with the whole surface area, and the oxidation proceeds uniformly into the inside of the film. On the annealed surface, which is partially covered by carbon or sulphur and not so rough as the evaporated film, on the other hand, the oxidation can start only at a limited number of reactive points, and the oxidized part develops its area through the boundaries of the two phases, forming the " oxidized patches ".The authors are grateful to Prof. Kozo Kuchitsu of the University of Tokyo and Dr. Katsuya Nakayama of the Electrotechnical Laboratory for valuable discussions. L. A. Harris, J. Appl. Phys., 1968, 39, 1419. C. J. Powell and J. B. Swan, Phys. Rev., 1960, 118, 640. E. J. Sheibner and L. N . Tharp, Surface Sci., 1967, 8, 427. D. Edwards, Jr., and F. M. Propst, J. Chem. Phys., 1971,55, 5175. R. E. Weber and W. T. Peria, J. Appl. Phys., 1967, 38,4355. P. W. Palmberg and T. N. Rhodin, J. Appl. Phys., 1968, 39,2425. H . Raether, Springer Tracts in Modern Physics, 1965, 38, 84. H. Iback, J. Vac.Sci. Technol., 1972, 9, 713. C. Kunz, Z. Phys., 1966, 196, 311. lo G. J. Dooley and T. W. Haas, .I. Chem. Phys., 1970,52,993. E. A. Stern and R. A. Ferrell, Phys. Rev., 1960, 120, 130. l2 C. J. Powell, Phys. Rev., 1968, 175, 972. l 3 J. W. Swaine and R. C. Plumb, J. Appl. Phys., 1962, 33, 2378. Study of the Oxidation of Molybdenum Surfaces by Energy Loss Spectroscopy Combined with Auger Electron Spectroscopy B Y TOMOJI KAWAI, KIMIO KUNIMORI, TAMOTSU KONDOW, TAKAHARU ONISHI AND KENZI TAMARU" Department of Chemistry, The University of Tokyo, Hongo, Bunkyo-Ku, Tokyo, Japan Received 3rd April, 1973 The oxidation of molybdenum as an evaporated film and of its annealed surface was investigated by Energy Loss Spectroscopy (ELS) combined with Auger Electron Spectroscopy (AES).The 10 eV and the lower energy loss peaks towards 5 eV were identified as the clean surface and the oxidized layer surface plasma loss peaks, respectively. The behaviours of these surface plasma loss spectra revealed that the oxidation of the evaporated film surface proceeded uniformly, rapidly and without any induction period under vacuum of lo-' Torr, whereas that of the annealed sample proceeded more slowly through " oxidized patches ", exhibiting a considerable induction period under the same vacuum conditions. These phenomena were interpreted with reference to the surface conditions and structures studied by the AES and ELS ; the fresh evaporated surface is clean and rough, whereas the annealed surface is smoothed by the heat-treatment and contaminated by sulphur and carbon which partially cover the active surfaces and give rise to the patchy surface oxidation.Energy analysis of scattered electrons has recently been shown to be an extremely useful tool for the study of solid ~urfaces,l-~ especially when various techniques are combined, for instance, low energy electron diffraction and Auger electron spectro- scopy (AES).5* Energy loss spectroscopy (ELS) is of particular interest since it is suitable for revealing the electronic structures of the surfaces of solids and adsorb- a t e ~ . ~ ~ AES, on the other hand, reveals the atomic identities existing on the sur- faces. Therefore, simultaneous measurements of ELS and AES may provide a deeper understanding of the surface properties and reactions than is possible with either one alone.On a metal surface, a surface mode of plasma excitation (surface plasmon) exists and can be observed by ELS, preferably in a reflection geometry. The surface mode appears at a lower frequency than the corresponding bulk mode, and shifts and/or reduces its intensity as the surface becomes contaminated by foreign ~pecies.~ In one case, the energy loss peak due to the surface plasmon on the clean surface shifts in the direction of lower energy, approaching a definite position of energy loss as the contamination proceeds. In another case, this initial peak gradually vanishes while another peak appears at the final energy position of the previous case.7 In the oxidation of aluminium, the surface plasma loss peak adopted the former behaviour under high vacuum condition ( Torr), it behaved as in the latter case.2 These phenomena cannot be explained by the different collision number of the gaseous molecules on the surface, but could be closely connected with the mechanism of the surface oxidation processes itself, so 137 T ~ r r ) , ~ whereas under lower vacuum138 OXIDATION OF MO that it can be investigated through the analysis of the behaviours of the surface plasmon loss peaks with reference to the surface conditions.In this paper, the oxidation process of a molybdenum film has been studied by varying its surface properties, taking advantage of the combined ELS and AES techniques. H 60 40 20 I mass number FIG. 1 . 4 0 ) The block diagram of the electron spectrometer for the ELS and the AES.This is a modified Simpson-Kuyatt type with 180" hemispherical energy selector whose optical circle has 25 mm radius. (b) Mass spectrum of the residual gas during the measurements.T . KAWAI, K . KUNIMORI, T. KONDOW, T . ONISHI, K . TAMARU 139 EXPERIMENTAL A modified Simpson-Kuyatt type electron spectrometer was constructed for the measure- ments of ELS and AES (fig. l(a)). The monochromator, as well as the analyzer, was a 180" hemispherical electrostatic lens pair, average radius of 2.5 cm, equipped with retarding electrostatic lenses. The energy resolution (AEG 500 meV) was high and constant enough, throughout the energy range of the measurements, to obtain the true shapes of the energy loss spectra and the Auger electron spectra. The scattered electrons, after being velocity selected, were detected by a single electron counting method.The collision chamber and the main chamber of the spectrometer were differentially pumped. This permitted one to introduce foreign gases into the collision chamber without disturbing the spectrometer. An evaporated Mo film was prepared and annealed in the collision chamber without exposure to air. Molybdenum (H. Cross Co. Ltd. 99.96 % purity) was evaporated onto a Mo foil which was fixed on a sample holder in the centre of the chamber. The Mo foil was chemically etched, heated to 1400°C in uacuo, then alternatively exposed to oxygen and hydrogen and finally checked for surface impurities by AES examination. Both chambers were evacuated by oil diffusion and rotary pumps. In order to prevent these chambers from being contam- inated, a foreline alumina trap was used as well as an ordinary liquid N2 trap.The pressure was measured by Bayard-Alpert gauge and was of the order 1 x lo-' Torr during the measure- ments. A NEVA quadrupole mass spectrometer was used to mass-analyze the residual gases (H20>90 %) (fig. l(b)). The operating parameters were as follows : energy of the primary electron beam, 380 eV for the ELS and 1.5 keV for the AES, primary beam current, 2x lo-* A for the ELS and 3 FA for the AES ; the incident and scattering angles, 27" each. RESULTS AND DISCUSSION 1. OXIDATION OF A FRESHLY EVAPORATED M O FILM When Mo was instantaneously evaporated on a Mo foil under a pressure of 1.2 x Torr, the energy loss spectra exhibited distinct time-dependent changes as is demonstrated in fig.2. The Mo foil itself had two loss peaks at 23 eV and 5 eV before the evaporation. In terms of a free electron model based on the assumption of 6 free electrons (4a5(5s)l per atom, the 23 eV peak is assigned to be a bulk plasmon e~citation.~. l o After evaporation two peaks appeared at 10 eV and 23 eV the 5 eV loss peak having disappeared. This 23 eV peak did not change during the measure- ments. On the other hand, the 10 eV peak shifted with time towards low energies and finally reached a position of 5 eV energy loss. It took about 30 min to accomplish the complete shift. The rate of the shift was rather faster at the beginning, without an induction period, decreasing gradually with time. The Auger spectra taken immediately after the evaporation exhibited peaks due to molybdenum and oxygen atoms together with very weak carbon peaks which became evident in the latter half of the measurements.Just after the evaporation the amount of surface carbon was less than a few percent of monolayer coverage, judged from the relative intensities of carbon and molybdenum and the change in their intensity. The exact amount of the surface oxygen at the initial stage of the oxidation was difficult to estimate because of the rapid oxidation rate, but the extrapolated value for the oxygen was estimated to be nearly the same as that for carbon. Considerable growth of a 5 15 eV oxygen peak with time indicates that " swface oxidation " proceeds, as shown in fig. 3(a). It is highly likely that the M o surface is oxidized by water, which was one of the major residual gases, as indicated by mass analysis of background gases in the vacuum chambers.Stern and Ferrell have theoretically predicted that the loss peak of surface plasmon undergoes a red-shift depending on the thickness of the oxidized layer formed on the140 OXIDATION OF M O 6) 38min 5) 30min 4 ) Z m i n 3) 12min 2) 6 min 1 ) 3 m i n after e va po rat i D n 8cfore evaporation 1 1 I 1 0 Ib 'io 3b 9 energy losslev FIG. 2.-The variation of the characteristic energy loss spectra with time, for the freshly evaporated molybdenum surface. S ( b ) 0 1 v 160 ZbO 240 3b0 500 550 kinetic energy/eV FIG. 3.-Auger electron spectra from a molybdenum surface. (a) Freshly evaporated surface. (6) After anneaiing in hydrogen at 1.5 x Sulphur and carbon are ob- Torr, looO°C, for 20 min.served on this annealed surface.T . KAWAI, K . KUNIMORI, T . KONDOW, T . ONISHI, K . TAMARU 141 metal, provided that the layer is formed uniformly.7* l1 * It can, accordingly, be concluded that the time-dependent peak is attributable to surface plasmon excitation. The thickness of the layer was estimated from the equation of Stern and Ferrell, assuming its validity, and was plotted as a function of reaction time (see fig. 5). In conclusion, the thickness of the oxidized layer increased with time in a uniform manner over the whole surface on this evaporated sample, similar to the oxidation of an A1 film under high vacuum c~nditions.~ If the assignment of the 23 eV peak is correct, the surface plasma loss on the clean surface should be 23/,/2 = 16.5 eV on the basis of the free electron model.This theoretical prediction does not agree with our experimental results. If one or more interband transitions occur with the energy close to that of the surface plasmon excitation expected from the free electron model, the observed surface plasmon loss may deviate from the theoretical prediction of the free electron model. 2. OXIDATION OF THE FILM ANNEALED IN HYDROGEN ATMOSPHERE The evaporated film, whose surface plasmon loss was 5eV, was annealed in hydrogen at 1.5 x Torr pressure and a temperature of 1000°C for 20 min. The loss spectra of the annealed specimen were examined at different times after the annealing (see fig. 4). The other experimental conditions were identical to the case of the evaporated film. The behaviours and the shapes of the surface plasmon were quite different from those of the evaporated sample.Immediately after the treat- ment, the 10 eV loss peak was found as well as that at 23 eV. As the surface oxidation proceeded, a new peak appeared at 5 eV and coexisted with the peak at 10 eV whose intensity was gradually decreasing. These two peaks did not shift with time but their intensities changed. The new peak at 5 eV corresponds to the 5 eV peak observed in the experiment on the freshly evaporated film, i.e., the loss peak by the surface plasma excitation of the Mo surfzce covered by a sufficient amount of the oxidized layer. This phenomenon shows that both the oxidized part and non-oxidized part coexist during the course of the oxidation, and suggests that the oxide nuclei grow at the surface layer, the oxidation proceeding at the boundary between the metal and its oxide.The AES of this specimen demonstrated the appearance of sulphur and carbon atoms immediately after the annealing as is shown in fig. 3(6). Seemingly, sulphur and carbon had diffused from the inside onto the surface of the specimen during the annealing procedure. Considerable parts of this annealed surface are covered by carbon and sulphur judging from the reduction of the intensity of the molybdenum. The ratio of C and S peak areas to Mo AES peak area were 2.1 and 3.2 respectively. In addition, the relative intensity of the surface to the bulk loss on the fresh evaporated film was approximately 25 times larger than on the annealed one.This means that the surface area of the evaporated film may be larger than that of the annealed one, that is, the surface of the evaporated one may be rougher, though the contaminants may reduce their relative intensity to some extent.12* l 3 3" E+tanhkD 2&+(1 +cZ) tanh kD us = where us is the frequency of the surface plasmon when a surface oxidized layer of dielectric constant E is formed to thickness D ; up is the frequency of the bulk plasmon (fiup = 23 eV for Mo) ; k repre- sents the wave number of the surface wave excited in the solid by the incident electron and is obtained from the momentum and the energy conservation law ; k = mus cos Bltiko where 8 is the incident and scattering angle (27") and ko is the wave number of the incident electron.142 OXIDATION OF M O In fig.6, graph (a) and (b), the percentage ratio of the 5 eV peak to the sum of the 5 and the 10 eV one was plotted against time (t) and t 2 . There is an induction period of about 10 min. In the initial stage of the reaction, the ratio was proportional to t 2 but above a ratio of 50 :d it was expressed in terms of aft2 +/I, where a' and are I t 10) After one day 9) 58min 8) 52min 7) 46min 6 ) 38min 5) 32min 4) 25 min 3) 19 min 2 ) 13 min I ) 4 min a f t e r l e a t i n g up in H2, 1 . 5 ~ mmHg 20 min 0 10 20 30 energy losslev FIG. 4.-The variation of the characteristic energy loss spectra with time, for the molybdenum surface annealed in hydrogen at 1.5 x Torr, 1000°C, for 20 min.FIG. 5.-The /O 10 20 30 40 50 elapsed timelniin thickness of the oxidized layer as a function of elapsed time after evayora tion.T. KAWAI, K . KUNIMORI, T . KONDOW, T . ONISHI, K . TAMARU 143 time-independent constants. These phenomena can be explained qualitatively as follows. Assuming that the diameter of the circular patches increases with a constant velocity and the size of every patch is approximately the same, the ratio may be roughly equal to ((n.nE2)/So)t2 before these patches overlap. Here n is the number of the active points, So is the total surface area and E represents the average growth velocity of the diameter. After the overlapping, the effective n may decrease. There- fore the slope is smaller than that at the initial stage. t P) elapsed timelmin 5 + square of the elapsed timelmin FIG.6.-The percentage ratio of the 5 eV peak to the sum of the 5 and 10 eV ones as a function of (a) elapsed time after annealing in hydrogen in hydrogen and (b) square of the elapsed time. The different behaviours in the ELS and the AES on these different surfaces is of interest in the view of the correlation between the surface reaction and the surface properties. Considerations on the rates of the oxidation at the early stage (fig. 5(a), fig. 6(a)) revealed that the uniform oxidation of the Mo surface starts rapidly just after the evaporation, which suggests a large collisional cross section for the reaction between gases and the evaporated surface. The induction period of about 10 min on the annealed surface, on the other hand, implies that the oxidation scarcely proceeded at the beginning and after about 10min began to develop rapidly.The oxidation cross section is very small at the beginning. As to the surface properties, AES demonstrated that the fresh evaporated surface is clean and is scarcely contaminated by other elements, whereas the annealed surface is contaminated by considerable amounts of carbon and sulphur. Another difference between the films would be the geometric structure of the surfaces, i.e., the roughness factor of the evaporated surface is larger than that of the annealed one. Taking these facts into consideration, the different behaviours in Mo oxidation are interpreted as follows ; on the fresh evaporated surface which is rough and active,144 OXIDATION OF M O the H20 in the ambient gas attacks the surface, colliding with the whole surface area, and the oxidation proceeds uniformly into the inside of the film. On the annealed surface, which is partially covered by carbon or sulphur and not so rough as the evaporated film, on the other hand, the oxidation can start only at a limited number of reactive points, and the oxidized part develops its area through the boundaries of the two phases, forming the " oxidized patches ". The authors are grateful to Prof. Kozo Kuchitsu of the University of Tokyo and Dr. Katsuya Nakayama of the Electrotechnical Laboratory for valuable discussions. L. A. Harris, J. Appl. Phys., 1968, 39, 1419. C. J. Powell and J. B. Swan, Phys. Rev., 1960, 118, 640. E. J. Sheibner and L. N . Tharp, Surface Sci., 1967, 8, 427. D. Edwards, Jr., and F. M. Propst, J. Chem. Phys., 1971,55, 5175. R. E. Weber and W. T. Peria, J. Appl. Phys., 1967, 38,4355. P. W. Palmberg and T. N. Rhodin, J. Appl. Phys., 1968, 39,2425. H . Raether, Springer Tracts in Modern Physics, 1965, 38, 84. H. Iback, J. Vac. Sci. Technol., 1972, 9, 713. C. Kunz, Z. Phys., 1966, 196, 311. lo G. J. Dooley and T. W. Haas, .I. Chem. Phys., 1970,52,993. E. A. Stern and R. A. Ferrell, Phys. Rev., 1960, 120, 130. l2 C. J. Powell, Phys. Rev., 1968, 175, 972. l 3 J. W. Swaine and R. C. Plumb, J. Appl. Phys., 1962, 33, 2378.

ISSN:0300-9599    

DOI:10.1039/F19747000137

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Pulse Radiolysis Study of the Reactions of Hydrated Electrons with Naphthalene, Phenanthrene, Biphenyl and Fluorene in Aqueous Micellar Solutions BY JANOS H. FENDLER Department of Chemistry, Texas A & M University, College Station, Texas 77843 HUGH A. GILLIS AND NORMAN V. KLASSEN * Division of Physics, National Research Council of Canada, Ottawa, Canada K1A OSI Received 14th May, 1973 The reaction of hydrated electrons with naphthalene, phenanthrene, biphenyl and fluorene to form the respective aromatic radical anions has been studied in aqueous solutions (1 % methanol, pH 12.1) in the presence and absence of micellar hexadecyltrimethylammonium bromide (CTAB). The reaction rates are 3-9 times faster in the presence of micellar CTAB suggesting that CTAB will, in general, enhance this type of reaction.An absorption spectrum of the free fluorene radical anion was obtained. In the presence of CTAB the aromatic radical anions decay to second transients which are probably the hydrogen adducts of the aromatic molecules. Investigations of radiation-induced reactions in aqueous micellar systems may lead to a better understanding of radiation biological processes involving macro- molecules which often exist as aggregates and/or tertiary structures. lm4 Micellar systems probably provide better approximations of the microenvironment of the bind- ing and reaction sites in macromolecules than does pure water. Furthermore, for surfactants having a 16-18 carbon chain the concentration at which monomeric surfactants aggregate to form micelles, i.e., the critical micelle concentration (CMC), is generally low, between and M,5 thereby obviating direct radiation effects and scavenging from the spurs.The rate constant for the addition of the hydrated electron (e? to benzene has been shown to decrease by a factor of 3 in the presence of aniomc micellar sodium dodecyl sulphate (NaLS). Conversely, cationic micellar hexadecyltrimethylaon- ium bromide (CTAB) increased the rate constant for the same reaction by a factor of We have used the method of pulse radiolysis to investigate the generality of these micellar effects on the reaction of e i with aromatic molecules to form the aromatic radical anions. We also report spectroscopic data for the radical anions and the hydrogen atom adducts formed by protonation of the radical anions.EXPERIMENTAL The preparation and purification of the surfactants have been described.6 Surface tension determinations of each of the surfactants were carried out using a du Nouy tensio- meter. No minima in the (surface tension, surfactant concentration) plots were observed, indicating the absence of impurities. The CMC of CTAB, NaLS and polyoxyethylene (15) 145146 PULSE RADIOLYSIS nonylphenol (Igepal CO-730) are 9 . 2 ~ M, 8.1 x M and M respectively.2 Reagent grade biphenyl, fluorene, naphthalene and phenanthrene were used without further purification to make up micellar solutions. The water was doubly distilled and further purified by passing the water vapour, mixed with oxygen, through a quartz furnace at 750°C. All solutions contained 1 % reagent grade methanol (as a hydroxyl radical scavenger) by volume and were adjusted to the appropriate alkaline pH with sodium hydroxide.The solutions were degassed under vacuum by 4-6 successive freeze-pump-thaw cycles. For the experiments in which the reaction of e i with the aromatic solutes was studied in the absence of micelles, 99.999 % naphthalene, biphenyl and phenanthrene and zone refined fluorene (James Hinton Co.) were used. The concentrations of aromatic solute (Ar) which could be used were limited by the low solubility of these compounds in water.' In these experiments the solutions were degassed by a stream of argon. To reduce removal of the aromatic solute from the solution by volatilization, the stream of argon was first saturated with the aromatic compound by passage through a prebubbler containing a large quantity of an identical solution. Nevertheless, it was difficult to ensure that oxygen was completely removed from the solution without some loss of the aromatic solute.The linear accelerator at the National Research Council was used to produce 20-300 ns pulses of 35 MeV electrons, the dose per pulse ranging from 3 x 10I6 to 2 x 1017 eV 8-l. The pulse radiolysis apparatus has been described.8 The wavelength settings of the Bausch and Lomb high intensity monochromators were calibrated in the u.-v. and visible with the spectral lines of a low-pressure Hg arc. The bandwidths used to determine narrow absorp- tion maxima were narrow enough to ensure that the measured values of GElmax were, at most, 10 % lower than the true values. Dosimetry was carried out by pulsing aqueous solutions of 5 x M potassium thiocyanate saturated with N20, and GE~~~(CNS)F = 4.2 x lo4 M-l cm-l (100 eV)-l was RESULTS AND DISCUSSION Naphthalene, biphenyl, fluorene and phenanthrene were more soluble in aqueous micellar CTAB, NaLS and Igepal CO-730 than in water, which indicates their solubil- ization lo by the micelles.TABLE 1 .-HYDRATED ELECTRON LIFETIMES AND k(e4 + Ar) IN AQUEOUS SOLUTIONS a k(&fAr)x 10-10, M-1 s-1 surfact ant t3 for esp calculated from calculated from hydrocarbon decaylps eG decay Ar7 formation 2.0 5 . 0 ~ M CTAB - 5.0 x M CTAB 7.9 x M biphenyl 0.040 2.1 2.2 5 . 0 ~ loV2 M CTAB 3 . 2 ~ M biphenyl 0.094 2.2 2.3 5 . 0 ~ M CTAB 5 . 0 ~ M naphthalene 0.383 2.9 5 .0 ~ lod2 M CTAB 5 . 0 ~ M naphthalene 0.046 2.9 3.0 5.0 x M CTAB 2.7 x M fluorene 0.113 2.2 2.7 5 . 0 ~ M CTAB 5 . 0 ~ M phenanthrene 0.037 3.6 4.0 - 18-60 1 . 8 ~ M naphthalene 0.6+0.1 0.5k0.1 0 . 9 ~ M naphthalene 1.2+0.2 0.5k0.1 2.7 x M phenanthrene 3+ 1 0.4k0.2 3 . 7 ~ M biphenyl 2.1 + 0.3 0.7+ 0.1 1.8 x M biphenyl 3.5+ 1 0.8+ 0.2 2 . 2 ~ M fluorene 5+ 1 0.5k0.2 1.1 x M fluorene 6 5 2 0.6k0.3 a All solutions were pH 12.1, contained 1 % methanol and were at ambient temperature. b These values have been corrected for the decay of e& in the absence of aromatic solute,J . H . FENDLER, H . A . GILLIS AND N . V . KLASSEN 147 The lifetime of e 4 following a pulse was determined in the presence and in the absence of naphthalene, phenanthrene, biphenyl and fluorene in aqueous solutions by following the decay of e& at - 700 nm.In the presence of the aromatic solubilizates, e,; reacts to form the radical anion (Ar') ; Absorption due to eG decayed exponentially for at least 80 % of its decay. The half-lives of e& measured from first order plots are given in table 1. In addition, decay rates of e& were measured for the following micellar solutions containing 1 % methanol : NaLS, NaLS +biphenyl, NaLS +naphthalene, NaLS + phenanthrene, Igepal CO-730, Igepal CO-730 +biphenyl. As found previously with benzene,2 the rate constant for reaction (1) is 3-9 times higher in the presence of micellar CTAB but is lowered by micellar NaLS or Igepal CO-730. The ensuing discussion of micellar solutions is limited to the spectral and kinetic data obtained in the presence of CTAB in which the build-up of transient absorption due to Ar', parallel with e i decay, was clearly observed.e, + Ar+Ar'. (1) SPECTRA Reaction of e& with naphthalene, phenanthrene, biphenyl and fluorene in micellar CTAB results in the consecutive formation of two transient absorbing species whose spectra are shown in fig. 1. The appearance of the first transients parallels the decay of e;. The first transients decay in tens of microseconds to form the second transients which decay in milliseconds, a decay too slow to measure accurately because of ripple in the output of the Xenon arc analyzing lamp. The spectra of the first transients were measured after the decay of eG. Spectra of the second transients were measured after the decay of Ar'.Comparison with published spectra reveals the first transients to be aromatic radical anions formed by reaction (1). This assignment was substantiated by the absence of the radical anion in pulse irradiated solutions of biphenyl, 5 x M CTAB containing 5 % acetone, a known electron scavenger. The spectra of the first transients are due largely to free Ar', since alkali metal countercations were not present. Nor do the spectra have contributions from aromatic cations or triplets. The contribution of H atom adducts to the absorption at Amax of the first transients is small, probably less than 10 %, with the possible exception of phenanthrene. There are insufficient data to correct the spectra for H adducts. A small blank correction has been made to obtain the spectra in fig.1. The blank was the transient absorption due to -CH20H formed by the reaction of OH with methanol. The correction was made using the spectrum of *CH20H reported for pH 12 by Simic et al. (a) NAPHTHALENE RADICAL ANION The spectrum of the first transient in naphthalene solutions (fig. 1A) has maxima at 324, 365, 775 and 850 nm. The sharp maximum at 324 nm is shown more clearly in the inset. The absorption spectrum of the naphthalene radical anion (Naph') has been determined in 2-methyltetrahydrofuran (MTHF) glass at - 180°C l2 and in tetrahydrofuran (THF) at 25"C.l 3* l4 These published spectra coincide, in detail, with the first transient shown in fig. IA. Under the conditions for the data in fig. 1A the conversion of H atoms to eG is small since k(H+OH-) = 1.8 x lo7 M-l s-l , l5 k(H+methanol) = 1.6 x lo6 M-' s-l,16 and k(H+naphthalene) is probably -5 x lo9 M-l s-l .G(H) = 0.6 and 83 %148 PULSE RA I ' I I I I 1 I 0 ' 300 400 500 600 700 800 900 LbO 300 400 500 600 700 800 wavelength/nm FIG. 1.-Absorption spectra of the aromatic radical anions and their decay products in pulse irrad- iated aqueous solutions (pH 12.1) containing aromatic solubilizate, 5 x M CTAB and 1 % methanol. The spectra of the radical anions were measured at their maximum absorbance. The decay products were measured in solutions which had already received 1OI8 eV g-l. A : 5 x M naphthalene solution ; 0 represents Naph' whose major peak is enlarged in the inset with bars representing bandwidths ; A represents the decay product (probably NaphH-) measured 150 ps after the pulse. B : 5 x M phenanthrene solution ; 0 represents Phen- ; A repre sents the decay product (probably PhenH-) measured 200 ps after the pulse.C: 7.9 x M biphenyl solution ; 0 represents Biph- ; A represents the decay product (probably BiphH-) measured 620 ps after the pulse. D : 5.3 x M fluorene solution ; 0 represents Fluo' ; A represents the decay products (see text) measured 100 ps after the pulse.J . H. FENDLER, H . A . GILLIS AND N . V . KLASSEN 149 of H will react with naphthalene. It will be shown later that the contribution of H atom adducts to the absorbance of the first transient at 324 nm is probably 10 % or less. By neglecting any absorbance due to H atom adducts and by taking G(Naph') = G(eJ = 2.8, we calculate from our measurements of G&324 = 4.05 x lo4 and GE85o = 6.9 X lo3 (fig.IA), that &324 = 1.5 x lo4 M-' cm-l and &850 = 2.5 x lo3 M-l cm-l. By comparison, in THF at 25°C the extinction coefficients of Naph' are 1.60 x lo4 M-l cm-l at A, = 323 nm l4 and 2.46 x lo3 M-l cm-l at A, = 820 nm. (b) PHENANTHRENE RADICAL ANION The absorption spectrum of the first transient in phenanthrene solutions (fig. 1B) has a peak at 420 nm, a shoulder at 450 nm, a broad peak centred at 650 nm and peaks at 975 and 1125 nm. This transient is believed to be the phenanthrene radical anion (Phen-), based on the close resemblance of its spectrum to the spectrum of Phen' in THF.17 By neglecting the possible contribution of H adducts to the absorbance of the first transient at 420 and 975 run and by taking G(Phen') = 2.8 we calculate from G~420 = 2.6 x lo4 and G E ~ ~ ~ = 6.7 x lo3 (fig.1B) that ~ 4 2 0 = 9.3 x lo3 M-' cm-l and E~~~ = 2 . 4 ~ lo3 M-l cm-l. These are very similar to the extinction coefficients of Phen' in THF which are 8.8 x lo3 M-l cm-l at A, = 415 nm and 2.1 x lo3 M-l cm-1 at A,,, = 943 n111.l~ (C) BIPHENYL RADICAL ANION The absorption spectrum of the first transient in biphenyl solutions (fig. 1C) has A, = 405 and 635 nm and corresponds to the spectrum of the biphenyl radical anion (Biph') found in THF,13* l4 and in ethanol.18 If the products of H atom reactions do not contribute significantly to the absorption at 405 and 635 nm, we can take G(Biph') = 2.8 and calculate from G E ~ ~ ~ = 7.3 x lo4 and G&635 = 2.4 x lo4 (fig.1C) that &405 = 2.6 x lo4 M-I cm-1 and &635 = 8.6 x lo3 M-' cm-1 . These values are smaller than the extinction coefficients in THF, which are 3.8 x lo4 M-l cm-l at A,,, = 400 nm l4 and 1.25 x lo4 M-I cm-l at Amax = 630 nm I 3 but the ratios of the values of E,,, are identical. (d) FLUORENE RADICAL ANION The absorption spectrum of the f%st transient in fluorene solutions has A, = 395 and 695 nm with shoulders at - 380 and - 660 nm. The same maxima are attributed to the fluorene radical anion (Fluo').lg9 2o Thus, fig. 1D appears to be the first reported spectrum of Flu0 -i. without large interfering absorptions between 3 10 and 500 nm and it confirms the suggestion I9 that A, = 397 nm is probably due to the free radical anion. By neglecting any possible contribution of H adducts to the absorption at A, and by taking G(F1uo') = 2.8 we calculate from G&695 = 2.8 x lo4 (fig.1D) that &695 = 1.0 x lo4 M-l cm-l compared to &695 = 8.8 x lo3 M-l cm-' at A, = 695 nm in MTHF at 77 K 2 0 (e) SECOND TRANSIENTS The second transients in CTAB solutions of naphthalene, phenanthrene and biphenyl have maxima at 330,400 and 320 nm respectively. Their spectra in solutions which have already received 10l8 eV g-1 are shown in fig. 1. The values of GE, but not the spectral shapes, were markedly dose dependent below an accumulated dose of 10l8 eV g-l. Apparently ArT decays by more than one mechanism. The decay of the second transients was too slow to be followed accurately with our apparatus.150 PULSE RADIOLYSIS The second transients in naphthalene, phenanthrene and biphenyl solutions are probably the hydrogen adducts, NaphHe, PhenH., and BiphH- respectively, formed by protonation of the radical anions : At-' +H,O+ArH-+OH- (2) This is likely to be a major reaction of Ar? as revealed by other investigations such as a study of the protonation of Naph' by water in dimethoxyethane.21 Absorptions at 330 nm similar to the second transient in fig.1A have been observed in pulse irradiated solutions of naphthalene in several organic solvents and tentatively assigned to the hydrogen adduct of 23 In order to assign the second transient in fig. 1 A to NaphH- with more certainty, a saturated solution of naphthalene (-2.5 x M)containing 0.1 M t-butanol and adjusted to pH 1.3 with perchloric acid was pulsed and the spectrum of the resulting transient measured.Under these conditions, e4 is converted into H atoms which then react with naphthalene to produce NaphH-. Fig. 2 shows the resulting spectrum ; extinction coefficients were calculated by taking G(H) = 3.4. NaphH- is seen to have a sharp maximum at 330 nm ( E = Also shown in fig. 2 are the data points from the second transient in fig. 1A normalized to the NaphH. spectrum at 330 nm. The coincidence is excellent and confirms the assignment of the second transient in fig. 1A to NaphH.. 9.8 x lo3 M-l cm-I ) and a smaller maximum at 380 nm (E = 1.3 x lo3 M-' cm-' >. u IC 8 E n I 0 7+ X w 4 2 A 400 500 ~~~ 300 I - I i P -/ hJ . H .FENDLER, H . A . GILLIS A N D N . V . RLASSEN 151 From E~~~ = 9.8 x lo3 M-l cm-l, G(NaphH0) is calculated to be 1.5 under the conditions for which the second transient is shown in fig. 1A. A G(NaphH0) of 1.5 is much greater than can be accounted for by the reaction of the H atoms produced in water. Presumably a large part of the NaphH. of fig. 1A is produced by reaction (2). Since the completion of this work, Wallace and Thomas 24 have reported on the reactions of elq and H atoms with biphenyl in micellar NaLS and CTAB. They report that Biph' in NaLS undergoes slow hydrolytic decay to give a long lived transient with A, = 310 nm. This probably corresponds to the second transient in fig. 1C for which we report A, = 320 nm and which we tentatively identify as BiphHe.However, Wallace and Thomas 24 have produced BiphH- by H atom addition to biphenyl in NaLS and report a spectrum with maxima at 365 and - 305 nm. Similarly, Arai and Dorfman * have attributed a maximum at 360 nm to BiphH. in ethanol, and Sawai and Hamill 2 5 have assigned maxima at 320 and 363 rn to BiyhH. in a y-irradiated 95 % methanol, 5 % water glass at 77 K. Thus, the fact that the second transient in fig. 1C has only a single maximum at 320 nm must be explained. The answer may lie in the suggestion of Sawai and Hamill 2 5 that the 320 and 363 nm bands might represent different isomers of BiphH.. Perhaps the proton- ation of Biph? in CTAB leads only to the isomer or isomers which have a band at 320 nm. The second transient in fluorene solutions (fig. 1D) has A, = 312 and 370 nm.The 370 nm peak is probably due to the free fluorenyl carbanion for which Amax in THF is believed to be 374 nm.26 The maximum at 312 nm might be due to the hydrogen adduct of fluorene (FluoH.) by analogy with the BiphH- maximum at 320 nm. KINETICS (a) e&+Ar In order to determine whether the presence of CTAB enhances the rates of reaction (1) it was necessary to determine these rates in the absence of CTAB since they have not been published except for the rate with naphthalene. For k(e4 + naphthalene) two widely different values, 0.31 x lo9, ref. (27) and 5.4 x lo9 M-l s-l, ref. (28) have been reported. The rates k(e& + Ar) were determined by following the decay of e i at wavelengths in the range 675-800nm. In each case, the particular wavelength was chosen to minimize the interference due to the build-up of product absorption.The solutions were pH 12.1 and contained 1 % methanol. Of biphenyl, fluorene, phenanthrene and naphthalene, only naphthalene has a solubility in water greater than M. The decay of e& in the most dilute solutions was quite sensitive to impurities such as oxygen. The rate constants are listed in table 1 and all fall within the range (6 4) x lo9 M-1 s-1 . In each case, the decay of e& was found to be exponential when absorption by the products was taken into account. The value of 5 x lo9 M-l s-l deter- mined in the present study for k(e& + naphthalene) agrees with the larger 28 of the the two published values. (b) e& + Ar/CTAB The rate constants for reaction (1) have been determined in the presence of CTAB by following both the exponential decay of eG and the exponential formation of Ar' .The second order rate constants (table 1) were then calculated by taking the concentra- tion of Ar as equal to its bulk concentration. The second order rate constants so calculated did not change with changes in the concentration of Ar. This implies that152 PULSE RADIOLYSIS each molecule of Ar maintains the same reactivity towards e& even at Ar concentra- tions high enough that many micelles contain more than one molecule of Ar. The aggregation number of CTAB is 61,5 so that in solutions of 5 x M CTAB the concentration of micelles is 8 x Values of the second order rate constants based on the decay of eG and on the formation of Ar' are in good agreement with one another.This agreement is illustrated in fig. 3 in whch the decay and growth of transient absorptions at 700 and 407 MI in a biphenyl/CTAB solution are shown. The overall decay of absorption at 700 nm is composed of the decay of e& and the formation of Biph'. The plot in fig. 3 begins 40 ns after the start of the 20 ns pulse when the absorption at 407 nm is largely due to Biph'. The value k(ei + biphenyl) = 2.2 x 1O1O M-l s-l in table 1 is in good agreement with the value of 2 . 4 ~ lozo M-l s-l reported by Wallace and Thomas. 24 M. 0.10 a I 1 I , 40 60 80 I00 tinie/ns FIG. 3.-Decay of eai and formation of Biph- in a pulse irradiated aqueous solution of 7.94 x M biphenyl, 5 x 0 represents (OD-ODco)700 nm measured from the upper trace in the inset oscilloscope tracing.0 represents (OD, - OD)407 nm measured from the lower trace in the inset. OD, is the optical density 300 ns after the start of the pulse. The abscissa represents time after the start of the pulse. It is a coincidental result of the choice of wave- M CTAB, 1 % methanol at pH 12.1. lengths and bandwidths that one straight line fits both sets of points. For naphthalene, phenanthrene, biphenyl and fluorene, reaction (1) is apparently 3-9 times faster in the presence of micellar CTAB. Thus it is likely that the previous observation of a larger rate constant for electron addition to benzene in micellar CTAB is a general phenomenon. It has been suggested that the net positive charge on the CTAB micellar surface promotes the motion of e& to the m i ~ e l l e .~ ~ It has also been suggested that electrostatic interactions between the n-system of benzene and the net positive charge on the CTAB surface render benzene more susceptible to nucleophilic attack by e&. ArT was found to decay exponentially in CTAB solutions with rate constants in the range 1 x 104-5 x lo4 s-l. Below 400 nm a complex decay of absorbance occurs in fluorene solutions, probably the result of the simultaneous protonation of Fluo- , formation of the fluorenyl anion and protonation of the fluorenyl anion. (C) ELECTRON TRANSFER PROCESSES We attempted to investigate micellar effects on electron transfer processes. In Biph' + Phenanthrene + Biphenyl + Phen particular we chose the systemJ . H .FENDLER, H . A . GILLIS AND N . V . KLASSEN 153 which has been studied in isopropan01.~~ The relatively large E at 1025 nm of Phen' as compared with Biph? made investigation of this pair feasible. A typical solution contained 2.3 x M biphenyl, 5.0 x M phenanthrene, 5.0 x 10-2 M CTAB and 1 % methanol at pH 12.1. A pulse width of 20 ns was used. It was expected that e& would add mostly to biphenyl and that subsequent electron transfer to phenanthrene would occur. Contrary to expectation, absorption at 1025 nm due to Phenl grew in completely during the decay time of e& as did absorption at 408 nm, apparently due to both Biph' and Phen". The decay of the absorption at 408 nm occurred on a much longer timescale. In this experiment a considerable fraction of the micelles contained both phenanthrene and biphenyl.The results imply that, in such micelles, either the attack of eG is directed initially to phenanthrene or the half-life for electron transfer is less than 20 ns. J. H. F. thanks the U.S. Atomic Energy Commission for partial support. J. H. Fendler and L. K. Patterson, J. Phys. Chem., 1970, 74,4608. K. M. Bansal, L. K. Patterson, E. J. Fendler and J. H. Fendler, Int. J. Radiat. Phys. Chem., 1971, 3, 321. J. H. Fendler, E. J. Fendler, G. Bogan, L. K. Patterson and K. M. Bansal, Chem. Comm., 1972, 14. L. K. Patterson, K. M. Bansai, G. Bogan, G. A. Infante, E. J. Fendler and J. H. Fendler, J. Amer. Chem. SOC., 1972,94,9028 ; J. H. Fendler, G. W. Bogan, E. J. Fendler, G. A. Infante and P. Jirathana, Reaction Kinetics in Micelles and Membranes, ed.E. H. Cordes (Plenum Press, N.Y., 1973), p. 53. E. J. Fendler and J. H. Fendler, Adu. Phys. Org. Chem., 1970, 8, 271. M. Casilio, E. J. Fendler and J. H. Fendler, J. Chem. SOC. B., 1971, 1377. Solubilities of Inorganic and Organic Compounds, ed. H. Stephen and T. Stephen (Pergamon, New York, 1963, 1964), vol. 1 and vol. 2. T. K. Cooper, D. C. Walker, H. A. Gillis and N. V. Klassen, Canad. J. Chem., 1973,51,2195 ; J. W. Purdie, H. A. Gillis and N. V. Klassen, Canad. J. Chem., 1973, 51, 3132. G. E. Adams, J. W. Boag, J. Currant and B. D. Michael, Pulse Radiolysis, ed. M. Ebert (Academic Press, London, 1965), p. 117. lo P. H. Elworthy, A. T. Florence and C. B. Macfarlane, Solubilization by Surface Active Agents and its Application to Chemistry and the BioIogical Sciences (Chapman and Hall, London, 1968).M. Simic, P. Neta and E. Hayon, J. Phys. Chem., 1969, 73, 3794. G. J. Hoijtink and P. J. Zandstra, Mol. Phys., 1960, 3, 371. l 3 J. Jagur-Grodzinski, M. Feld, S. L. Yang and M. Szwarc, J, Phys. Chem., 1965, 69, 628. I4 P. Chang, R. V. Slates and M. Szwarc, J. Phys. Chem., 1966,70, 3180. l 5 M. S. Matheson and J. Rabani, J. Phys. Chem., 1965, 69, 1324. l6 P. Neta and R. H. Schuler, J. Phys. Chem., 1972, 76, 2673. l 7 P. Balk, G. J. Hoijtink and J. W. H. Schreurs, Rec. Trau. chim., 1957, 76, 813. S. Arai and L. M. Dorfman, J. Chem. Phys., 1964, 41, 2190. D. Casson and B. J. Tabner, J. Chem. Soc. B, 1969, 888. 'O J. Wendenburg and H. Mockel, 2. Naturforsclz., 1968, 23b, 1171. S. Bank and B. Bockrath, J.Amer. Chern. Soc., 1971, 93, 430. '' F. S. Dainton, T. Morrow, G. A. Salmon and G. F. Thompson, Proc. Roy. SOC. A, 1972,328, 457; T. J. Kemp and J. P. Roberts, Trans. Faraday Soc., 1968, 64, 2106; F. S. Dainton, T. J. Kemp, G. A. Salmon and J. P. Keene, Nature, 1964, 203, 1050. 23 F. S. Dainton, J. P. Keene, T. J. Kemp, G. A. Salmon and J. Teply, Proc. Chem. SOC., 1964,265. 24 S . C. Wallace and J. K. Thomas, Rad. Res., 1973, 54, 49. 2 5 T. Sawai and W. H. Hamill, J. Phys. Chem., 1969,73, 2750. 26 T. E. Hogen-Esch and J. Smid, J. Amer. Chem. Soc., 1966, 88, 307. '' E. J. Hart, S. Gordon and J. K. Thomas, J. Phys. Chem., 1964, 68, 1271. '' M. Anbar and E. J. Hart, J. Amer. Chem. SOC., 1964,86, 5633. 29 S. Arai, D. A. Grev and L. M. Dorfman, J. Chem. Phys., 1967,46,2572. Pulse Radiolysis Study of the Reactions of Hydrated Electrons with Naphthalene, Phenanthrene, Biphenyl and Fluorene in Aqueous Micellar Solutions BY JANOS H.FENDLER Department of Chemistry, Texas A & M University, College Station, Texas 77843 HUGH A. GILLIS AND NORMAN V. KLASSEN * Division of Physics, National Research Council of Canada, Ottawa, Canada K1A OSI Received 14th May, 1973 The reaction of hydrated electrons with naphthalene, phenanthrene, biphenyl and fluorene to form the respective aromatic radical anions has been studied in aqueous solutions (1 % methanol, pH 12.1) in the presence and absence of micellar hexadecyltrimethylammonium bromide (CTAB). The reaction rates are 3-9 times faster in the presence of micellar CTAB suggesting that CTAB will, in general, enhance this type of reaction.An absorption spectrum of the free fluorene radical anion was obtained. In the presence of CTAB the aromatic radical anions decay to second transients which are probably the hydrogen adducts of the aromatic molecules. Investigations of radiation-induced reactions in aqueous micellar systems may lead to a better understanding of radiation biological processes involving macro- molecules which often exist as aggregates and/or tertiary structures. lm4 Micellar systems probably provide better approximations of the microenvironment of the bind- ing and reaction sites in macromolecules than does pure water. Furthermore, for surfactants having a 16-18 carbon chain the concentration at which monomeric surfactants aggregate to form micelles, i.e., the critical micelle concentration (CMC), is generally low, between and M,5 thereby obviating direct radiation effects and scavenging from the spurs.The rate constant for the addition of the hydrated electron (e? to benzene has been shown to decrease by a factor of 3 in the presence of aniomc micellar sodium dodecyl sulphate (NaLS). Conversely, cationic micellar hexadecyltrimethylaon- ium bromide (CTAB) increased the rate constant for the same reaction by a factor of We have used the method of pulse radiolysis to investigate the generality of these micellar effects on the reaction of e i with aromatic molecules to form the aromatic radical anions. We also report spectroscopic data for the radical anions and the hydrogen atom adducts formed by protonation of the radical anions.EXPERIMENTAL The preparation and purification of the surfactants have been described.6 Surface tension determinations of each of the surfactants were carried out using a du Nouy tensio- meter. No minima in the (surface tension, surfactant concentration) plots were observed, indicating the absence of impurities. The CMC of CTAB, NaLS and polyoxyethylene (15) 145146 PULSE RADIOLYSIS nonylphenol (Igepal CO-730) are 9 . 2 ~ M, 8.1 x M and M respectively.2 Reagent grade biphenyl, fluorene, naphthalene and phenanthrene were used without further purification to make up micellar solutions. The water was doubly distilled and further purified by passing the water vapour, mixed with oxygen, through a quartz furnace at 750°C. All solutions contained 1 % reagent grade methanol (as a hydroxyl radical scavenger) by volume and were adjusted to the appropriate alkaline pH with sodium hydroxide.The solutions were degassed under vacuum by 4-6 successive freeze-pump-thaw cycles. For the experiments in which the reaction of e i with the aromatic solutes was studied in the absence of micelles, 99.999 % naphthalene, biphenyl and phenanthrene and zone refined fluorene (James Hinton Co.) were used. The concentrations of aromatic solute (Ar) which could be used were limited by the low solubility of these compounds in water.' In these experiments the solutions were degassed by a stream of argon. To reduce removal of the aromatic solute from the solution by volatilization, the stream of argon was first saturated with the aromatic compound by passage through a prebubbler containing a large quantity of an identical solution.Nevertheless, it was difficult to ensure that oxygen was completely removed from the solution without some loss of the aromatic solute. The linear accelerator at the National Research Council was used to produce 20-300 ns pulses of 35 MeV electrons, the dose per pulse ranging from 3 x 10I6 to 2 x 1017 eV 8-l. The pulse radiolysis apparatus has been described.8 The wavelength settings of the Bausch and Lomb high intensity monochromators were calibrated in the u.-v. and visible with the spectral lines of a low-pressure Hg arc. The bandwidths used to determine narrow absorp- tion maxima were narrow enough to ensure that the measured values of GElmax were, at most, 10 % lower than the true values.Dosimetry was carried out by pulsing aqueous solutions of 5 x M potassium thiocyanate saturated with N20, and GE~~~(CNS)F = 4.2 x lo4 M-l cm-l (100 eV)-l was RESULTS AND DISCUSSION Naphthalene, biphenyl, fluorene and phenanthrene were more soluble in aqueous micellar CTAB, NaLS and Igepal CO-730 than in water, which indicates their solubil- ization lo by the micelles. TABLE 1 .-HYDRATED ELECTRON LIFETIMES AND k(e4 + Ar) IN AQUEOUS SOLUTIONS a k(&fAr)x 10-10, M-1 s-1 surfact ant t3 for esp calculated from calculated from hydrocarbon decaylps eG decay Ar7 formation 2.0 5 . 0 ~ M CTAB - 5.0 x M CTAB 7.9 x M biphenyl 0.040 2.1 2.2 5 . 0 ~ loV2 M CTAB 3 . 2 ~ M biphenyl 0.094 2.2 2.3 5 . 0 ~ M CTAB 5 .0 ~ M naphthalene 0.383 2.9 5 . 0 ~ lod2 M CTAB 5 . 0 ~ M naphthalene 0.046 2.9 3.0 5.0 x M CTAB 2.7 x M fluorene 0.113 2.2 2.7 5 . 0 ~ M CTAB 5 . 0 ~ M phenanthrene 0.037 3.6 4.0 - 18-60 1 . 8 ~ M naphthalene 0.6+0.1 0.5k0.1 0 . 9 ~ M naphthalene 1.2+0.2 0.5k0.1 2.7 x M phenanthrene 3+ 1 0.4k0.2 3 . 7 ~ M biphenyl 2.1 + 0.3 0.7+ 0.1 1.8 x M biphenyl 3.5+ 1 0.8+ 0.2 2 . 2 ~ M fluorene 5+ 1 0.5k0.2 1.1 x M fluorene 6 5 2 0.6k0.3 a All solutions were pH 12.1, contained 1 % methanol and were at ambient temperature. b These values have been corrected for the decay of e& in the absence of aromatic solute,J . H . FENDLER, H . A . GILLIS AND N . V . KLASSEN 147 The lifetime of e 4 following a pulse was determined in the presence and in the absence of naphthalene, phenanthrene, biphenyl and fluorene in aqueous solutions by following the decay of e& at - 700 nm. In the presence of the aromatic solubilizates, e,; reacts to form the radical anion (Ar') ; Absorption due to eG decayed exponentially for at least 80 % of its decay.The half-lives of e& measured from first order plots are given in table 1. In addition, decay rates of e& were measured for the following micellar solutions containing 1 % methanol : NaLS, NaLS +biphenyl, NaLS +naphthalene, NaLS + phenanthrene, Igepal CO-730, Igepal CO-730 +biphenyl. As found previously with benzene,2 the rate constant for reaction (1) is 3-9 times higher in the presence of micellar CTAB but is lowered by micellar NaLS or Igepal CO-730. The ensuing discussion of micellar solutions is limited to the spectral and kinetic data obtained in the presence of CTAB in which the build-up of transient absorption due to Ar', parallel with e i decay, was clearly observed. e, + Ar+Ar'.(1) SPECTRA Reaction of e& with naphthalene, phenanthrene, biphenyl and fluorene in micellar CTAB results in the consecutive formation of two transient absorbing species whose spectra are shown in fig. 1. The appearance of the first transients parallels the decay of e;. The first transients decay in tens of microseconds to form the second transients which decay in milliseconds, a decay too slow to measure accurately because of ripple in the output of the Xenon arc analyzing lamp. The spectra of the first transients were measured after the decay of eG.Spectra of the second transients were measured after the decay of Ar'. Comparison with published spectra reveals the first transients to be aromatic radical anions formed by reaction (1). This assignment was substantiated by the absence of the radical anion in pulse irradiated solutions of biphenyl, 5 x M CTAB containing 5 % acetone, a known electron scavenger. The spectra of the first transients are due largely to free Ar', since alkali metal countercations were not present. Nor do the spectra have contributions from aromatic cations or triplets. The contribution of H atom adducts to the absorption at Amax of the first transients is small, probably less than 10 %, with the possible exception of phenanthrene. There are insufficient data to correct the spectra for H adducts.A small blank correction has been made to obtain the spectra in fig. 1. The blank was the transient absorption due to -CH20H formed by the reaction of OH with methanol. The correction was made using the spectrum of *CH20H reported for pH 12 by Simic et al. (a) NAPHTHALENE RADICAL ANION The spectrum of the first transient in naphthalene solutions (fig. 1A) has maxima at 324, 365, 775 and 850 nm. The sharp maximum at 324 nm is shown more clearly in the inset. The absorption spectrum of the naphthalene radical anion (Naph') has been determined in 2-methyltetrahydrofuran (MTHF) glass at - 180°C l2 and in tetrahydrofuran (THF) at 25"C.l 3* l4 These published spectra coincide, in detail, with the first transient shown in fig. IA. Under the conditions for the data in fig.1A the conversion of H atoms to eG is small since k(H+OH-) = 1.8 x lo7 M-l s-l , l5 k(H+methanol) = 1.6 x lo6 M-' s-l,16 and k(H+naphthalene) is probably -5 x lo9 M-l s-l . G(H) = 0.6 and 83 %148 PULSE RA I ' I I I I 1 I 0 ' 300 400 500 600 700 800 900 LbO 300 400 500 600 700 800 wavelength/nm FIG. 1.-Absorption spectra of the aromatic radical anions and their decay products in pulse irrad- iated aqueous solutions (pH 12.1) containing aromatic solubilizate, 5 x M CTAB and 1 % methanol. The spectra of the radical anions were measured at their maximum absorbance. The decay products were measured in solutions which had already received 1OI8 eV g-l. A : 5 x M naphthalene solution ; 0 represents Naph' whose major peak is enlarged in the inset with bars representing bandwidths ; A represents the decay product (probably NaphH-) measured 150 ps after the pulse.B : 5 x M phenanthrene solution ; 0 represents Phen- ; A repre sents the decay product (probably PhenH-) measured 200 ps after the pulse. C: 7.9 x M biphenyl solution ; 0 represents Biph- ; A represents the decay product (probably BiphH-) measured 620 ps after the pulse. D : 5.3 x M fluorene solution ; 0 represents Fluo' ; A represents the decay products (see text) measured 100 ps after the pulse.J . H. FENDLER, H . A . GILLIS AND N . V . KLASSEN 149 of H will react with naphthalene. It will be shown later that the contribution of H atom adducts to the absorbance of the first transient at 324 nm is probably 10 % or less. By neglecting any absorbance due to H atom adducts and by taking G(Naph') = G(eJ = 2.8, we calculate from our measurements of G&324 = 4.05 x lo4 and GE85o = 6.9 X lo3 (fig. IA), that &324 = 1.5 x lo4 M-' cm-l and &850 = 2.5 x lo3 M-l cm-l.By comparison, in THF at 25°C the extinction coefficients of Naph' are 1.60 x lo4 M-l cm-l at A, = 323 nm l4 and 2.46 x lo3 M-l cm-l at A, = 820 nm. (b) PHENANTHRENE RADICAL ANION The absorption spectrum of the first transient in phenanthrene solutions (fig. 1B) has a peak at 420 nm, a shoulder at 450 nm, a broad peak centred at 650 nm and peaks at 975 and 1125 nm. This transient is believed to be the phenanthrene radical anion (Phen-), based on the close resemblance of its spectrum to the spectrum of Phen' in THF.17 By neglecting the possible contribution of H adducts to the absorbance of the first transient at 420 and 975 run and by taking G(Phen') = 2.8 we calculate from G~420 = 2.6 x lo4 and G E ~ ~ ~ = 6.7 x lo3 (fig.1B) that ~ 4 2 0 = 9.3 x lo3 M-' cm-l and E~~~ = 2 . 4 ~ lo3 M-l cm-l. These are very similar to the extinction coefficients of Phen' in THF which are 8.8 x lo3 M-l cm-l at A, = 415 nm and 2.1 x lo3 M-l cm-1 at A,,, = 943 n111.l~ (C) BIPHENYL RADICAL ANION The absorption spectrum of the first transient in biphenyl solutions (fig. 1C) has A, = 405 and 635 nm and corresponds to the spectrum of the biphenyl radical anion (Biph') found in THF,13* l4 and in ethanol.18 If the products of H atom reactions do not contribute significantly to the absorption at 405 and 635 nm, we can take G(Biph') = 2.8 and calculate from G E ~ ~ ~ = 7.3 x lo4 and G&635 = 2.4 x lo4 (fig.1C) that &405 = 2.6 x lo4 M-I cm-1 and &635 = 8.6 x lo3 M-' cm-1 . These values are smaller than the extinction coefficients in THF, which are 3.8 x lo4 M-l cm-l at A,,, = 400 nm l4 and 1.25 x lo4 M-I cm-l at Amax = 630 nm I 3 but the ratios of the values of E,,, are identical. (d) FLUORENE RADICAL ANION The absorption spectrum of the f%st transient in fluorene solutions has A, = 395 and 695 nm with shoulders at - 380 and - 660 nm. The same maxima are attributed to the fluorene radical anion (Fluo').lg9 2o Thus, fig. 1D appears to be the first reported spectrum of Flu0 -i. without large interfering absorptions between 3 10 and 500 nm and it confirms the suggestion I9 that A, = 397 nm is probably due to the free radical anion.By neglecting any possible contribution of H adducts to the absorption at A, and by taking G(F1uo') = 2.8 we calculate from G&695 = 2.8 x lo4 (fig. 1D) that &695 = 1.0 x lo4 M-l cm-l compared to &695 = 8.8 x lo3 M-l cm-' at A, = 695 nm in MTHF at 77 K 2 0 (e) SECOND TRANSIENTS The second transients in CTAB solutions of naphthalene, phenanthrene and biphenyl have maxima at 330,400 and 320 nm respectively. Their spectra in solutions which have already received 10l8 eV g-1 are shown in fig. 1. The values of GE, but not the spectral shapes, were markedly dose dependent below an accumulated dose of 10l8 eV g-l. Apparently ArT decays by more than one mechanism. The decay of the second transients was too slow to be followed accurately with our apparatus.150 PULSE RADIOLYSIS The second transients in naphthalene, phenanthrene and biphenyl solutions are probably the hydrogen adducts, NaphHe, PhenH., and BiphH- respectively, formed by protonation of the radical anions : At-' +H,O+ArH-+OH- (2) This is likely to be a major reaction of Ar? as revealed by other investigations such as a study of the protonation of Naph' by water in dimethoxyethane.21 Absorptions at 330 nm similar to the second transient in fig.1A have been observed in pulse irradiated solutions of naphthalene in several organic solvents and tentatively assigned to the hydrogen adduct of 23 In order to assign the second transient in fig. 1 A to NaphH- with more certainty, a saturated solution of naphthalene (-2.5 x M)containing 0.1 M t-butanol and adjusted to pH 1.3 with perchloric acid was pulsed and the spectrum of the resulting transient measured.Under these conditions, e4 is converted into H atoms which then react with naphthalene to produce NaphH-. Fig. 2 shows the resulting spectrum ; extinction coefficients were calculated by taking G(H) = 3.4. NaphH- is seen to have a sharp maximum at 330 nm ( E = Also shown in fig. 2 are the data points from the second transient in fig. 1A normalized to the NaphH. spectrum at 330 nm. The coincidence is excellent and confirms the assignment of the second transient in fig. 1A to NaphH.. 9.8 x lo3 M-l cm-I ) and a smaller maximum at 380 nm (E = 1.3 x lo3 M-' cm-' >. u IC 8 E n I 0 7+ X w 4 2 A 400 500 ~~~ 300 I - I i P -/ hJ .H . FENDLER, H . A . GILLIS A N D N . V . RLASSEN 151 From E~~~ = 9.8 x lo3 M-l cm-l, G(NaphH0) is calculated to be 1.5 under the conditions for which the second transient is shown in fig. 1A. A G(NaphH0) of 1.5 is much greater than can be accounted for by the reaction of the H atoms produced in water. Presumably a large part of the NaphH. of fig. 1A is produced by reaction (2). Since the completion of this work, Wallace and Thomas 24 have reported on the reactions of elq and H atoms with biphenyl in micellar NaLS and CTAB. They report that Biph' in NaLS undergoes slow hydrolytic decay to give a long lived transient with A, = 310 nm. This probably corresponds to the second transient in fig. 1C for which we report A, = 320 nm and which we tentatively identify as BiphHe.However, Wallace and Thomas 24 have produced BiphH- by H atom addition to biphenyl in NaLS and report a spectrum with maxima at 365 and - 305 nm. Similarly, Arai and Dorfman * have attributed a maximum at 360 nm to BiphH. in ethanol, and Sawai and Hamill 2 5 have assigned maxima at 320 and 363 rn to BiyhH. in a y-irradiated 95 % methanol, 5 % water glass at 77 K. Thus, the fact that the second transient in fig. 1C has only a single maximum at 320 nm must be explained. The answer may lie in the suggestion of Sawai and Hamill 2 5 that the 320 and 363 nm bands might represent different isomers of BiphH.. Perhaps the proton- ation of Biph? in CTAB leads only to the isomer or isomers which have a band at 320 nm. The second transient in fluorene solutions (fig.1D) has A, = 312 and 370 nm. The 370 nm peak is probably due to the free fluorenyl carbanion for which Amax in THF is believed to be 374 nm.26 The maximum at 312 nm might be due to the hydrogen adduct of fluorene (FluoH.) by analogy with the BiphH- maximum at 320 nm. KINETICS (a) e&+Ar In order to determine whether the presence of CTAB enhances the rates of reaction (1) it was necessary to determine these rates in the absence of CTAB since they have not been published except for the rate with naphthalene. For k(e4 + naphthalene) two widely different values, 0.31 x lo9, ref. (27) and 5.4 x lo9 M-l s-l, ref. (28) have been reported. The rates k(e& + Ar) were determined by following the decay of e i at wavelengths in the range 675-800nm.In each case, the particular wavelength was chosen to minimize the interference due to the build-up of product absorption. The solutions were pH 12.1 and contained 1 % methanol. Of biphenyl, fluorene, phenanthrene and naphthalene, only naphthalene has a solubility in water greater than M. The decay of e& in the most dilute solutions was quite sensitive to impurities such as oxygen. The rate constants are listed in table 1 and all fall within the range (6 4) x lo9 M-1 s-1 . In each case, the decay of e& was found to be exponential when absorption by the products was taken into account. The value of 5 x lo9 M-l s-l deter- mined in the present study for k(e& + naphthalene) agrees with the larger 28 of the the two published values. (b) e& + Ar/CTAB The rate constants for reaction (1) have been determined in the presence of CTAB by following both the exponential decay of eG and the exponential formation of Ar' .The second order rate constants (table 1) were then calculated by taking the concentra- tion of Ar as equal to its bulk concentration. The second order rate constants so calculated did not change with changes in the concentration of Ar. This implies that152 PULSE RADIOLYSIS each molecule of Ar maintains the same reactivity towards e& even at Ar concentra- tions high enough that many micelles contain more than one molecule of Ar. The aggregation number of CTAB is 61,5 so that in solutions of 5 x M CTAB the concentration of micelles is 8 x Values of the second order rate constants based on the decay of eG and on the formation of Ar' are in good agreement with one another.This agreement is illustrated in fig. 3 in whch the decay and growth of transient absorptions at 700 and 407 MI in a biphenyl/CTAB solution are shown. The overall decay of absorption at 700 nm is composed of the decay of e& and the formation of Biph'. The plot in fig. 3 begins 40 ns after the start of the 20 ns pulse when the absorption at 407 nm is largely due to Biph'. The value k(ei + biphenyl) = 2.2 x 1O1O M-l s-l in table 1 is in good agreement with the value of 2 . 4 ~ lozo M-l s-l reported by Wallace and Thomas. 24 M. 0.10 a I 1 I , 40 60 80 I00 tinie/ns FIG. 3.-Decay of eai and formation of Biph- in a pulse irradiated aqueous solution of 7.94 x M biphenyl, 5 x 0 represents (OD-ODco)700 nm measured from the upper trace in the inset oscilloscope tracing. 0 represents (OD, - OD)407 nm measured from the lower trace in the inset. OD, is the optical density 300 ns after the start of the pulse.The abscissa represents time after the start of the pulse. It is a coincidental result of the choice of wave- M CTAB, 1 % methanol at pH 12.1. lengths and bandwidths that one straight line fits both sets of points. For naphthalene, phenanthrene, biphenyl and fluorene, reaction (1) is apparently 3-9 times faster in the presence of micellar CTAB. Thus it is likely that the previous observation of a larger rate constant for electron addition to benzene in micellar CTAB is a general phenomenon. It has been suggested that the net positive charge on the CTAB micellar surface promotes the motion of e& to the m i ~ e l l e .~ ~ It has also been suggested that electrostatic interactions between the n-system of benzene and the net positive charge on the CTAB surface render benzene more susceptible to nucleophilic attack by e&. ArT was found to decay exponentially in CTAB solutions with rate constants in the range 1 x 104-5 x lo4 s-l. Below 400 nm a complex decay of absorbance occurs in fluorene solutions, probably the result of the simultaneous protonation of Fluo- , formation of the fluorenyl anion and protonation of the fluorenyl anion. (C) ELECTRON TRANSFER PROCESSES We attempted to investigate micellar effects on electron transfer processes. In Biph' + Phenanthrene + Biphenyl + Phen particular we chose the systemJ .H . FENDLER, H . A . GILLIS AND N . V . KLASSEN 153 which has been studied in isopropan01.~~ The relatively large E at 1025 nm of Phen' as compared with Biph? made investigation of this pair feasible. A typical solution contained 2.3 x M biphenyl, 5.0 x M phenanthrene, 5.0 x 10-2 M CTAB and 1 % methanol at pH 12.1. A pulse width of 20 ns was used. It was expected that e& would add mostly to biphenyl and that subsequent electron transfer to phenanthrene would occur. Contrary to expectation, absorption at 1025 nm due to Phenl grew in completely during the decay time of e& as did absorption at 408 nm, apparently due to both Biph' and Phen". The decay of the absorption at 408 nm occurred on a much longer timescale. In this experiment a considerable fraction of the micelles contained both phenanthrene and biphenyl.The results imply that, in such micelles, either the attack of eG is directed initially to phenanthrene or the half-life for electron transfer is less than 20 ns. J. H. F. thanks the U.S. Atomic Energy Commission for partial support. J. H. Fendler and L. K. Patterson, J. Phys. Chem., 1970, 74,4608. K. M. Bansal, L. K. Patterson, E. J. Fendler and J. H. Fendler, Int. J. Radiat. Phys. Chem., 1971, 3, 321. J. H. Fendler, E. J. Fendler, G. Bogan, L. K. Patterson and K. M. Bansal, Chem. Comm., 1972, 14. L. K. Patterson, K. M. Bansai, G. Bogan, G. A. Infante, E. J. Fendler and J. H. Fendler, J. Amer. Chem. SOC., 1972,94,9028 ; J. H. Fendler, G. W. Bogan, E. J. Fendler, G. A. Infante and P. Jirathana, Reaction Kinetics in Micelles and Membranes, ed. E. H. Cordes (Plenum Press, N.Y., 1973), p. 53. E. J. Fendler and J. H. Fendler, Adu. Phys. Org. Chem., 1970, 8, 271. M. Casilio, E. J. Fendler and J. H. Fendler, J. Chem. SOC. B., 1971, 1377. Solubilities of Inorganic and Organic Compounds, ed. H. Stephen and T. Stephen (Pergamon, New York, 1963, 1964), vol. 1 and vol. 2. T. K. Cooper, D. C. Walker, H. A. Gillis and N. V. Klassen, Canad. J. Chem., 1973,51,2195 ; J. W. Purdie, H. A. Gillis and N. V. Klassen, Canad. J. Chem., 1973, 51, 3132. G. E. Adams, J. W. Boag, J. Currant and B. D. Michael, Pulse Radiolysis, ed. M. Ebert (Academic Press, London, 1965), p. 117. lo P. H. Elworthy, A. T. Florence and C. B. Macfarlane, Solubilization by Surface Active Agents and its Application to Chemistry and the BioIogical Sciences (Chapman and Hall, London, 1968). M. Simic, P. Neta and E. Hayon, J. Phys. Chem., 1969, 73, 3794. G. J. Hoijtink and P. J. Zandstra, Mol. Phys., 1960, 3, 371. l 3 J. Jagur-Grodzinski, M. Feld, S. L. Yang and M. Szwarc, J, Phys. Chem., 1965, 69, 628. I4 P. Chang, R. V. Slates and M. Szwarc, J. Phys. Chem., 1966,70, 3180. l 5 M. S. Matheson and J. Rabani, J. Phys. Chem., 1965, 69, 1324. l6 P. Neta and R. H. Schuler, J. Phys. Chem., 1972, 76, 2673. l 7 P. Balk, G. J. Hoijtink and J. W. H. Schreurs, Rec. Trau. chim., 1957, 76, 813. S. Arai and L. M. Dorfman, J. Chem. Phys., 1964, 41, 2190. D. Casson and B. J. Tabner, J. Chem. Soc. B, 1969, 888. 'O J. Wendenburg and H. Mockel, 2. Naturforsclz., 1968, 23b, 1171. S. Bank and B. Bockrath, J. Amer. Chern. Soc., 1971, 93, 430. '' F. S. Dainton, T. Morrow, G. A. Salmon and G. F. Thompson, Proc. Roy. SOC. A, 1972,328, 457; T. J. Kemp and J. P. Roberts, Trans. Faraday Soc., 1968, 64, 2106; F. S. Dainton, T. J. Kemp, G. A. Salmon and J. P. Keene, Nature, 1964, 203, 1050. 23 F. S. Dainton, J. P. Keene, T. J. Kemp, G. A. Salmon and J. Teply, Proc. Chem. SOC., 1964,265. 24 S . C. Wallace and J. K. Thomas, Rad. Res., 1973, 54, 49. 2 5 T. Sawai and W. H. Hamill, J. Phys. Chem., 1969,73, 2750. 26 T. E. Hogen-Esch and J. Smid, J. Amer. Chem. Soc., 1966, 88, 307. '' E. J. Hart, S. Gordon and J. K. Thomas, J. Phys. Chem., 1964, 68, 1271. '' M. Anbar and E. J. Hart, J. Amer. Chem. SOC., 1964,86, 5633. 29 S. Arai, D. A. Grev and L. M. Dorfman, J. Chem. Phys., 1967,46,2572.

ISSN:0300-9599    

DOI:10.1039/F19747000145

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Phase Structure, Nuclear Magnetic Resonance and Rheological Proper ties of Vi scoelas tic Sodium Dodecyl Sulphate and Trimethylammoniuni Bromide Mixtures BY CHRISTOPHER A. BARKER, DOROTHY SAUL, GORDON J. T. TIDDY,* BARBARA A. WHEELER AND EDWIN WILLIS Unilever Research Port Sunlight Laboratory, Port Sunlight, Wirral, Cheshire L62 4XN Received 5th July, 1973 The phase diagram of octyltrimethylammonium bromide and sodium dodecyl sulphate (SDS) in water ( > 9 0 %) is reported at 298 K ; the phase rsgions observed are liquid, two liquid, and liquid plus liquid crystal. Liquids with viscoelastic properties occur close to the SDS rich liquid boundary, and a correlation is observed between the surfactant n.m.r. linewidths and the rheological properties of these liquids. The most likely explanation of these effects is one involving the occurrence of both cylindrical and spherical micelles, but the possibility that the viscoelasticity is due to the presence of a microemulsion can not be excluded.In the course of our investigation of the rheological properties of mixtures of different surface active agents, viscoelastic effects were observed with mixtures of anionic and cationic surfactants and of zwitterionic/anionic surfactants. These effects occurred within narrow composition ranges and it was suspected that this behaviour might be related to the proximity of these compositions to phase bound- aries. Furthermore, the observation of a two liquid coexistence region in the aqueous corner of the sodium dodecyl sulphate (SDS)/octyltrimethylammonium bromide (C,TAB)/water system prompted interest in a more detailed investigation of the phase structure of this system.In this paper the results of a joint phase diagram and rheological investigation of the SDS/C,TAB/wata system in the water rich region are reported. The relationship between the marked changes in macroscopic viscosity with composition and the motions of the surfactant molecules have also been investigated using high resolution and spin echo n.m.r. techniques. EXPERIMENTAL MATERIALS The SDS was B.D.H. specially purified grade and CsTAB was obtained from Palmer Research Laboratories. Both materials were used without further purification. DzO used to make up n.m.r. samples was 99.7 % pure (p Scientific Ltd.) and water was twice distilled before use. PHASE STUDIES Samples, prepared by the method previously described,2 were left for at least a week at 298 K before measurements were made.Phase structures were determined using a polarising microscope (Reichert) fitted with a hot stage. X-ray diffraction patterns were obtained at 298 K using Rigaku-Denki low angle scattering equipment. Several samples in the two isotropic liquid region were separated by ultracentrifugation (MSE Superspeed 50) and their compositions were analysed. 154C. BARKER, D . SAUL, G . TIDDY, B . WHEELER, E . WILLIS 155 N.m.r. measurements were made on solutions in D20 using a Bruker Physik 3228 spin echo spectrometer at 60 MHz, a Varian 220 MHz high resolution spectrometer and a Perkin Elmer R12 spectrometer (60 MHz). RHEOLOGICAL MEASUREMENTS Measurements at 298 K were made using the Haake Rotovisko concentric cylinder viscometer and the Weissenberg rheogoniometer (model R18).The Haake was used with the NV cup and bob assembly which covered a range of Newtonian shear rates between 15.86 and 2570 s - I . The rheogoniometer was fitted with 7.5 cm diameter plates and was used in the steady mode (constant speed) to measure vicsosities and normal forces (first normal stress difference) over a range of shear rates between 7 and 702s-l. Measurements of dynamic viscosity (11') and dynamic rigidity (G') were made using the instrument in the oscillatory mode at frequencies between 0.025 and 25 Hz. Full details of the experimental techniques and data processing procedures are given elsew Samples for rheological measurements were prepared by mixing solutions of the separate surfactants in varying proportions, the smaller volume always being added to the larger.Samples were prepared by : (i) varying compositions at a total surfactant concentration of' 2 % by weight ; (ii) varying concentration of total surfactaiit with constant proportion The composition range below 2 % total surfactant concentration where the mixtures showed viscoelastic properties was determined by a titration techniquz . The composition was noted when the mixtures first became viscoelastic judged visually by observing recoil after swirling, when the mixtures became turbid and finally where viscoelasticity vanished. Of SDS/CgTAB. RESULTS PHASE STUDIES The phase diagrams of the SDS/C,TAB/water system are given in fig.1 and 2. In fig. 1 the area of two coexisting liquids (2L) above 1 % surfactant concentration is illustrated. Also shown in fig. 1 are the compositions where viscoelastic properties FIG. 1 .-Phase diagram of sodium dodecyl sulphate and octyltrimethylammonium bromide with :>98 % water at 298 K. - , Phase boundary ; - - -, viscoelastic region ; - . - ., turbid region.156 VISCOELASTIC MICELLAR SOLUTIONS were observed and where the mixtures had a turbid appearance. Fig. 2 shows the existence of a second 2L region in addition to the liquid crystalline (LC) plus isotropic liquid region. Broken lines illustrate boundaries not accurately determined. No 2L+LC three phase region was observed but the presence of a small amount of a second isotropic liquid phase in the L + LC samples cannot be excluded.SDS 90- 10 A FIG. 2.-Phase diagram of sodium dodecyl sulphate and octyltrimethylammonium bromide with >90 % water at 298 K. - , Phase boundary ; x- x , boundaries not accurately determined. In the 2L region, analyses (% water, and C, H, N, Br, Na elemental analysis) demonstrated that the top layer was surfactant rich while the lower layer contained a larger concentration of NaBr. X-ray studies indicated that the liquid crystal phase was lamellar (i.e. diffraction line spacings were in the ratio 1 : 3). RHEOLOGICAL STUDIES Measurements were made with the Haake Rotovisko on mixtures of 2 % total surfactant concentration (section AA, fig. 2), and on mixtures with a constant ratio of SDS/C8TAB (section BB) over the concentration range 0.5-10 %.Commencing from the SDS axis (and increasing the C8TAB concentration) solutions containing 2 % surfactant varied from optically clear with water-like Newtonian viscosity, to thick clear viscoelastic gels which were shear thinning. On further increasing the proportion of CBTAB the solutions became turbid but still remained viscoelastic, finally at 2 % of C8TAB, solutions were clear and water-like. The rheological data was analysed according to the power law z = rcy where z = shear stress (dyn cm-2), f = shear rate (s-I), n = power law index, K = consistency (dyn cm-2 9). The compositions, visual appearance, the range of shear rates used and values of IC and n for the 2 % samples are given in table 1, and those for samples with constantC.BARKER, D. SAUL, G . TIDDY, B . WHEELER, E. WILLIS 157 TABLE ~.-RHEOLOGLCAL DATA FOR 2 % SDS/C2TAB SOLUTIONS no. of visual appearance range of shear K n points [S$Sg [C8TABl/ wt. % 1.50 0.50 clear, thin 285-2568 0.01 1.05 5 1.428 0.572 clear, elastic 48-2577 0.16 0.87 8 1.360 0.640 clear, elastic, thick 16.4-2663 5.40 0.36 10 1 . 2 7 2 0.728 clear, elastic, thick 16.2-2626 2.73 0.48 10 1.250 0.750 turbid, elastic, thin 16.2-2632 0.13 0.82 10 1.200 0.800 turbid, elastic, thin 15.9-2586 0.14 0.76 10 TABLE 2.-RHEOLOGICAL DATA FOR SOLUTIONS OF SDS/CsTAB WITH CONCENTRATION RATIO (BY WT.) OF 15 : 7 total conc./ wt. % 10 7.5 5.0 3.2 2.0 0.5 visual appearance range of shear turbid, thin, slightly elastic 15.9-2590 turbid, thin, slightly elastic 15.9-2576 clear gel 15.9-143 clear, elastic, thick 15.9-1537 clear, elastic, thick 16.4-2663 clear, elastic 47.8-2580 { 301-2709 K t1 0.84 0.72 0.15 0.90 1.30 0.92 50.30 0.27 45.02 0.098 5.4 0.36 0.09 0.85 no.of points 10 10 5 5 9 10 7 0.1 f 10 100 1000 shear rate Is-' FIG. 3.-Apparent viscosity against shear rate for 5.0 % SDS + C8TAB. concentration/mol % FIG. 4.-First normal stress difference against concentration for samples along line BB of fig. 2.158 VISCOELASTIC MICELLAR SOLUTIONS ratios of SDS to CsTAB (15 : 7 by wt.) in table 2. At 5.0 % concentration rather unusual behaviour is observed. When the apparent viscosity is plotted against shear rate the data falls on two lines, with greater shear thinning occurring at shear rates above about 300 s-l (see fig.3). Measurements on the Weissenberg rheogoniometcr were confined to systems of constant SDS : CsTAB molar ratio (concentrations 0.1-5.0 %) corresponding to the 100- ri I 9 R -a =z u 10- 1.0 1 1 2 3 4 5 concentration/mol % against concentration. FIG. 5.-Dynamic rigidity (G) measured at 10 Hz (0, left hand scale) and 1 Hz ( x right hand scale) concentrationlmol % FIG. 6.-Dynamic viscosity (9') against concentration. 0, A, 10 Hz; X , 1 Hz. (A, points measured with steady shear, p = 10 s-'.)C . BARKER, D. SAUL, G. TIDDY, B . WHEELER, E. WILLIS 159 section BB on fig. 2. The data obtained from the oscillatory measurements were in excellent agreement with those obtained under steady shear, i.e. plots of y' and 2G' against angular frequency were superimposable on the plots of q and first normal stress d-ifference (NSD) against shear rate respectively.Full details of the rheological work on these systems will be published elsewhere. Fig. 4-6 illustrate respectively how the first NSD, dynamic rigidity and both q and y' vary with concentration along the section BB. MAGNETIC RESONANCE STUDIES In order to investigate surfactant molecular motions, proton n.m.r. relaxation time measurements ( Tl, spin lattice relaxation time ; T2, spin-spin relaxation time) were made on a series of SDS/C8TAB samples in D20 containing approximately 5 % (by wt. -0.173 mol ~ l m - ~ ) of surfactant with a variation of composition from pure SDS to pure CsTAB solutions. Both T1 and T2 relaxation curves were non-exponen- tial, and thus were determined by several different relaxation times.Tl values were of the order of 0.5-1.0 s while T2 values were in the range 0.5-0.025 s. Relaxation times were shortest for samples with high viscosities, i.e. for samples close to the 2L region in the SDS rich area of the phase diagram. The wide variation in T2 values prompted the use of high resolution n.m.r., because proton resonaiice from different groups can be distinguished with this technique. T2 measured by spin echo is related to the line width of the high resolution resonance (Av+) by the relationship given in eqn (2) : Ay+ = nT,'. (1) TABLE 3.-LINE WIDTHS (AV)g a OF RESONANCES IN S D S / C 8 T A B / w ~ ~ ~ ~ SAMPLES AT 298 K composition/ SDS C+TAB 58TAB (mol % SDS) OCH2 NCHz WCH3)3 PCH2 [CWn 91.7 88.5 85.1 78 .O 74.0 73.3 72.4 70.0 66.0 15.0 5.8 2.0 2.2 2.7 3.5 22.5 22 22 26.5 17 7 3 2.0 2.0 2.0 4.5 > 22 > 22 >22 N 22 3 2.5 1.6 1.6 1.8 1.8 7.0 7.5 7.7 7.3 5.5 1.4 1.5 16 17 16 25 >40 > 40 >40 >40 N 40 24 24 11.5 12.0 13.0 13.0 33 31.5 36 34 34 23 23 a Spectra recorded at 220 MHz, Av+ in Hz, values for NCH, refer to line width at half height minus the contribution from indirect coupling ( J ) .In the high resolution n.m.r. spectra little or no differential broadening of reson- ances was observed for CsTAB rich solutions or for solutions with compositions close to the SDS/water axis. Differential peak broadening was observed for samples in the viscoelastic region with the line width being of the magnitude expected from the T, measurement; the values are shown in table 3 and fig.7. It can be seen that a dramatic change in line widths occurs at ca. 77 mol % SDS, and that the change is greater at the head group than at the terminal methyl group. The residual HOD resonance had a line-width determined by magnet inhomogeneity in all the samples.I60 VISCOELASTIC MICELLAR SOLUTIONS 1 I 1 I I 100 90 80 70 60 mol % SDS FIG. 7.-Variation of n.m.r. linewidths with SDS/CaTAB ratio. 0, OCHz ; A, CHzCHJ ; x, N(CH3)J. Changes in n.m.r. relaxation times are related to changes in translational and rotational molecular motions? Motions with correlation times (z,) of the order of 10-l' s (normal liquid motions) result in equal values for Tl and T2 ( N 1 s). A decrease in molecular motions reduces TI and T2 from about 1 s to about 20 ms where Tl has a minimum value (7, N s).For slower motions Tl increases again while T2 becomes very small (T,<10-100 P S ) . ~ In the normal micellar liquids, SDS and CsTAB micelles have Tl N T2 = 1 s, showing fast molecular motion (z, 21 10-l O - lo-" s). If the addition of CsTAB to SDS micelles caused a gradual reduction in molecular motion then Tl and T, would be reduced by similar amounts until T1 passed its minimum value of ca. 20 ms. This is not observed, since TI is always in the range 0.5-1 s while T2 changes over the range 0.5-0.025s. Therefore, it can be concluded that there is a discontinuous change in motion at concentration sea. 77 mol % SDS, with some surfactant molecules having z,> - s. This can be explained by either : (a) the presence of two types of micelle, one spherical with rapid motion (2, < 10-lo s) and the second having restricted mobility (zc> s) ; or (b) a change in micellar type to a large cylindrical (or other shaped) micelle which has regions of both fast and restricted molecular motion.Further, since only a single high resolution n.m.r. spectrum is observed, the rate of molecular exchange bdween the different species or environments must be faster than lo4 s-l. The n.m.r. line widths are dominated by the contributions from surfactant molecules with 2,) - lo-* s. It should be noted that although some molecular motions may be restricted (such as chain bending) overall rotation of the molecule about the long axis is likely to be rapid (2," 10-lo) as is the case in liquid crystalline phases. A comparison of SDS and CsTAB a-CH2 group line widths with those of terminalC. BARKER, D.SAUL, G . TIDDY, B . WHEELER, E. WILLIS 161 methyl groups indicates that in the restricted regions the head group is more immobil- ised than the interior of the micelle. Also, the CsTAB a-CH2 line widths are broader than those of the SDS a-CH, indicating that either the CsTAB chain is more re- stricted than the SDS chain or that proportionately more C8TAB molecules are pres- ent in restricted regions than SDS molecules. The NMe, resonance is relatively sharp because rotation about the N-Me bond is rapid, and takes place at varying angles to the N-CH2 group rotational axis. Theoretical calculations have shown that even with relatively slow rotational rates two axes of rotation can cause dramatic narrowing of resonance bands.7 DISCUSSION our results show that in the SDS/csTAB/Water system, an increase in Viscosity with shear thinning and viscoelastic behaviour occurs only at compositions which are close to the SDS rich side of the two liquid coexistence region.Viscoelastic behaviour was observed in solutions containing as little as 0.1 % surfactant but the existence of two isotropic liquids could not be detected under the microscope at concentrations below about 1 % (see fig. 1). The solutions become visibly turbid just before the two liquid phases were discernible under the microscope and the viscosity reached its maximum value at the SDS rich side of the line defining the onset of turbidity.The n.m.r. results demonstrate that the surfactant molecules in the viscoelastic solutions exist in both " restricted " and " liquid-like " environments. It is difficult to envisage how both of these could arise from a single type of micellar species although this possibility cannot be completely excluded. A more likely situation is that there are two,types of micellesexistinginequilibrium, normal spherical micelles and restricted micelles which axe probably cylindrical and of the type proposed by Pilpel for surfactant/electrolyte/water systems. The motion of surfactant molecules in a restricted cylindrical micelle would involve lateral diffusion along the surface of the micelle, together with chain rotational and bending motions. If a diffusion coefficient of cm2 s-l is postulated, then the time required for diffusion of the surfactant molecule around the circular cross section of a cylindrical micelle is ca.s (from D = Z2/6t). This is the same as or longer than the longest z, limit deduced from the n.m.r. measurement, and thus the existence of a cylindrical micelle with slow overall rotation about its long axis would be sufficient to account for the n.m.r. line widths. The alkyl chains could be assumed to undergo chain bending motions (involving internal rotations about CH,-CH2 bonds) and this would contribute towards the narrower lines observed for the CH2 and CH, groups further away from the head group. The presence of cylindrical micelles would account for the rheological properties of the system. Under static conditions the cylinders would be randomly oriented by Brownian motion.In flow the cylinders will become aligned with the amount of order increasing with shear rate and the resistance to flow decreases with increasing order, giving shear thinning behaviour. Whcn flow ceases, the cylinders again resume a random configuration and some of the energy used to align them is released, i.e. viscoelastic behaviour. However, due to the proximity of the viscoelastic mixtures to the two liquid phase boundary we cannot exclude the possibility that the rheological properties are due to the presence of a micro-emulsion. The onset of turbidity is probably due to the presence of emulsion droplets which are large enough to scatter light but are too small to be seen in transmitted light by the microscope used in the phase studies.Further, before the turbidity becomes apparent to the eye there may be very small emulsion 1-61 62 VISCOELASTIC MICELLAR SOLUTIONS droplets which are interacting to form a network throughout the solution leading to shear thinning and viscoelastic properties. The emulsion (of a small amount of a surfactant rich phase in a sodium bromide rich continuous phase) would contain the restricted surfactant molecules in one of the phases, but these would not be responsible for the viscoelasticity. An attempt to differentiate between the two mechanisms by separation of a turbid viscoelastic solution into the two component liquids by ultra-centrifugation proved to be inconclusive. The sample separated into a small fraction of an upper surfactant rich layer which was viscous and shear thinning but not elastic, and a much larger volume of an aqueous layer with low viscosity (- 1-2 cP).The only feature in the n.m.r. spectra of either layer was the sharp resonance from HDO. (The concentration of surfactant in the aqueous layer may have been insufficient for the observation of an n.m.r. spectrum.) However, it would seem that the restricted micellar environment observed by n.m.r., is a property of the surfactant rich layer. Nevertheless, visco- elasticity was only observed where the system was in the dispersed state. Thus we are unable to unequivocally distinguish whether the rheological properties are due to cylindrical micelles or to interacting micro-emulsion droplets. However, in other mixed surfactant systems which do not have a two liquid coexistence region the viscoelastic behaviour is undoubtedly due to the presence of " restricted " cylindri- cal micelles.D. H. Chen and D. G. Hall, KolloidZ. 2. Polymere, 1973, 251, 41. G. J. T. Tiddy, J.C.S. Faraday I, 1972, 68, 608. K. Walters, Basic Concepts and Formulae for Rheogoniometer (Sangamo Controls Ltd., 1968). K. Walters and R. A. Kemp, Deformation and Flow of High Polymers, ed. R. Wetton and R. Whorlow (McMillan, London, 1968), chap. 18. A. R. Eastwood, B. Yates and H. A. Barnes, Rheologica Acta, 1973, in press. N. Bloembergen, E. M. Purcell and R. V. Pound, Phys. Rev., 1948, 73, 679. N. Pilpel, J. Colloid Sci., 1954, 9, 285. D. Saul, G. J. T. Tiddy, B. A. Wheeler, P. A. Wheeler and E.Willis, J.C.S, Faraduy I, 1974,70, 163. ' J. G. Powles and H. S. Gutowsky, J. Chem. Phys., 1953, 21, 1695. Phase Structure, Nuclear Magnetic Resonance and Rheological Proper ties of Vi scoelas tic Sodium Dodecyl Sulphate and Trimethylammoniuni Bromide Mixtures BY CHRISTOPHER A. BARKER, DOROTHY SAUL, GORDON J. T. TIDDY,* BARBARA A. WHEELER AND EDWIN WILLIS Unilever Research Port Sunlight Laboratory, Port Sunlight, Wirral, Cheshire L62 4XN Received 5th July, 1973 The phase diagram of octyltrimethylammonium bromide and sodium dodecyl sulphate (SDS) in water ( > 9 0 %) is reported at 298 K ; the phase rsgions observed are liquid, two liquid, and liquid plus liquid crystal. Liquids with viscoelastic properties occur close to the SDS rich liquid boundary, and a correlation is observed between the surfactant n.m.r.linewidths and the rheological properties of these liquids. The most likely explanation of these effects is one involving the occurrence of both cylindrical and spherical micelles, but the possibility that the viscoelasticity is due to the presence of a microemulsion can not be excluded. In the course of our investigation of the rheological properties of mixtures of different surface active agents, viscoelastic effects were observed with mixtures of anionic and cationic surfactants and of zwitterionic/anionic surfactants. These effects occurred within narrow composition ranges and it was suspected that this behaviour might be related to the proximity of these compositions to phase bound- aries. Furthermore, the observation of a two liquid coexistence region in the aqueous corner of the sodium dodecyl sulphate (SDS)/octyltrimethylammonium bromide (C,TAB)/water system prompted interest in a more detailed investigation of the phase structure of this system. In this paper the results of a joint phase diagram and rheological investigation of the SDS/C,TAB/wata system in the water rich region are reported.The relationship between the marked changes in macroscopic viscosity with composition and the motions of the surfactant molecules have also been investigated using high resolution and spin echo n.m.r. techniques. EXPERIMENTAL MATERIALS The SDS was B.D.H. specially purified grade and CsTAB was obtained from Palmer Research Laboratories. Both materials were used without further purification.DzO used to make up n.m.r. samples was 99.7 % pure (p Scientific Ltd.) and water was twice distilled before use. PHASE STUDIES Samples, prepared by the method previously described,2 were left for at least a week at 298 K before measurements were made. Phase structures were determined using a polarising microscope (Reichert) fitted with a hot stage. X-ray diffraction patterns were obtained at 298 K using Rigaku-Denki low angle scattering equipment. Several samples in the two isotropic liquid region were separated by ultracentrifugation (MSE Superspeed 50) and their compositions were analysed. 154C. BARKER, D . SAUL, G . TIDDY, B . WHEELER, E . WILLIS 155 N.m.r. measurements were made on solutions in D20 using a Bruker Physik 3228 spin echo spectrometer at 60 MHz, a Varian 220 MHz high resolution spectrometer and a Perkin Elmer R12 spectrometer (60 MHz).RHEOLOGICAL MEASUREMENTS Measurements at 298 K were made using the Haake Rotovisko concentric cylinder viscometer and the Weissenberg rheogoniometer (model R18). The Haake was used with the NV cup and bob assembly which covered a range of Newtonian shear rates between 15.86 and 2570 s - I . The rheogoniometer was fitted with 7.5 cm diameter plates and was used in the steady mode (constant speed) to measure vicsosities and normal forces (first normal stress difference) over a range of shear rates between 7 and 702s-l. Measurements of dynamic viscosity (11') and dynamic rigidity (G') were made using the instrument in the oscillatory mode at frequencies between 0.025 and 25 Hz.Full details of the experimental techniques and data processing procedures are given elsew Samples for rheological measurements were prepared by mixing solutions of the separate surfactants in varying proportions, the smaller volume always being added to the larger. Samples were prepared by : (i) varying compositions at a total surfactant concentration of' 2 % by weight ; (ii) varying concentration of total surfactaiit with constant proportion The composition range below 2 % total surfactant concentration where the mixtures showed viscoelastic properties was determined by a titration techniquz . The composition was noted when the mixtures first became viscoelastic judged visually by observing recoil after swirling, when the mixtures became turbid and finally where viscoelasticity vanished.Of SDS/CgTAB. RESULTS PHASE STUDIES The phase diagrams of the SDS/C,TAB/water system are given in fig. 1 and 2. In fig. 1 the area of two coexisting liquids (2L) above 1 % surfactant concentration is illustrated. Also shown in fig. 1 are the compositions where viscoelastic properties FIG. 1 .-Phase diagram of sodium dodecyl sulphate and octyltrimethylammonium bromide with :>98 % water at 298 K. - , Phase boundary ; - - -, viscoelastic region ; - . - ., turbid region.156 VISCOELASTIC MICELLAR SOLUTIONS were observed and where the mixtures had a turbid appearance. Fig. 2 shows the existence of a second 2L region in addition to the liquid crystalline (LC) plus isotropic liquid region. Broken lines illustrate boundaries not accurately determined.No 2L+LC three phase region was observed but the presence of a small amount of a second isotropic liquid phase in the L + LC samples cannot be excluded. SDS 90- 10 A FIG. 2.-Phase diagram of sodium dodecyl sulphate and octyltrimethylammonium bromide with >90 % water at 298 K. - , Phase boundary ; x- x , boundaries not accurately determined. In the 2L region, analyses (% water, and C, H, N, Br, Na elemental analysis) demonstrated that the top layer was surfactant rich while the lower layer contained a larger concentration of NaBr. X-ray studies indicated that the liquid crystal phase was lamellar (i.e. diffraction line spacings were in the ratio 1 : 3). RHEOLOGICAL STUDIES Measurements were made with the Haake Rotovisko on mixtures of 2 % total surfactant concentration (section AA, fig.2), and on mixtures with a constant ratio of SDS/C8TAB (section BB) over the concentration range 0.5-10 %. Commencing from the SDS axis (and increasing the C8TAB concentration) solutions containing 2 % surfactant varied from optically clear with water-like Newtonian viscosity, to thick clear viscoelastic gels which were shear thinning. On further increasing the proportion of CBTAB the solutions became turbid but still remained viscoelastic, finally at 2 % of C8TAB, solutions were clear and water-like. The rheological data was analysed according to the power law z = rcy where z = shear stress (dyn cm-2), f = shear rate (s-I), n = power law index, K = consistency (dyn cm-2 9). The compositions, visual appearance, the range of shear rates used and values of IC and n for the 2 % samples are given in table 1, and those for samples with constantC. BARKER, D.SAUL, G . TIDDY, B . WHEELER, E. WILLIS 157 TABLE ~.-RHEOLOGLCAL DATA FOR 2 % SDS/C2TAB SOLUTIONS no. of visual appearance range of shear K n points [S$Sg [C8TABl/ wt. % 1.50 0.50 clear, thin 285-2568 0.01 1.05 5 1.428 0.572 clear, elastic 48-2577 0.16 0.87 8 1.360 0.640 clear, elastic, thick 16.4-2663 5.40 0.36 10 1 . 2 7 2 0.728 clear, elastic, thick 16.2-2626 2.73 0.48 10 1.250 0.750 turbid, elastic, thin 16.2-2632 0.13 0.82 10 1.200 0.800 turbid, elastic, thin 15.9-2586 0.14 0.76 10 TABLE 2.-RHEOLOGICAL DATA FOR SOLUTIONS OF SDS/CsTAB WITH CONCENTRATION RATIO (BY WT.) OF 15 : 7 total conc./ wt.% 10 7.5 5.0 3.2 2.0 0.5 visual appearance range of shear turbid, thin, slightly elastic 15.9-2590 turbid, thin, slightly elastic 15.9-2576 clear gel 15.9-143 clear, elastic, thick 15.9-1537 clear, elastic, thick 16.4-2663 clear, elastic 47.8-2580 { 301-2709 K t1 0.84 0.72 0.15 0.90 1.30 0.92 50.30 0.27 45.02 0.098 5.4 0.36 0.09 0.85 no. of points 10 10 5 5 9 10 7 0.1 f 10 100 1000 shear rate Is-' FIG. 3.-Apparent viscosity against shear rate for 5.0 % SDS + C8TAB. concentration/mol % FIG. 4.-First normal stress difference against concentration for samples along line BB of fig. 2.158 VISCOELASTIC MICELLAR SOLUTIONS ratios of SDS to CsTAB (15 : 7 by wt.) in table 2. At 5.0 % concentration rather unusual behaviour is observed. When the apparent viscosity is plotted against shear rate the data falls on two lines, with greater shear thinning occurring at shear rates above about 300 s-l (see fig.3). Measurements on the Weissenberg rheogoniometcr were confined to systems of constant SDS : CsTAB molar ratio (concentrations 0.1-5.0 %) corresponding to the 100- ri I 9 R -a =z u 10- 1.0 1 1 2 3 4 5 concentration/mol % against concentration. FIG. 5.-Dynamic rigidity (G) measured at 10 Hz (0, left hand scale) and 1 Hz ( x right hand scale) concentrationlmol % FIG. 6.-Dynamic viscosity (9') against concentration. 0, A, 10 Hz; X , 1 Hz. (A, points measured with steady shear, p = 10 s-'.)C . BARKER, D. SAUL, G. TIDDY, B . WHEELER, E. WILLIS 159 section BB on fig. 2. The data obtained from the oscillatory measurements were in excellent agreement with those obtained under steady shear, i.e.plots of y' and 2G' against angular frequency were superimposable on the plots of q and first normal stress d-ifference (NSD) against shear rate respectively. Full details of the rheological work on these systems will be published elsewhere. Fig. 4-6 illustrate respectively how the first NSD, dynamic rigidity and both q and y' vary with concentration along the section BB. MAGNETIC RESONANCE STUDIES In order to investigate surfactant molecular motions, proton n.m.r. relaxation time measurements ( Tl, spin lattice relaxation time ; T2, spin-spin relaxation time) were made on a series of SDS/C8TAB samples in D20 containing approximately 5 % (by wt. -0.173 mol ~ l m - ~ ) of surfactant with a variation of composition from pure SDS to pure CsTAB solutions.Both T1 and T2 relaxation curves were non-exponen- tial, and thus were determined by several different relaxation times. Tl values were of the order of 0.5-1.0 s while T2 values were in the range 0.5-0.025 s. Relaxation times were shortest for samples with high viscosities, i.e. for samples close to the 2L region in the SDS rich area of the phase diagram. The wide variation in T2 values prompted the use of high resolution n.m.r., because proton resonaiice from different groups can be distinguished with this technique. T2 measured by spin echo is related to the line width of the high resolution resonance (Av+) by the relationship given in eqn (2) : Ay+ = nT,'. (1) TABLE 3.-LINE WIDTHS (AV)g a OF RESONANCES IN S D S / C 8 T A B / w ~ ~ ~ ~ SAMPLES AT 298 K composition/ SDS C+TAB 58TAB (mol % SDS) OCH2 NCHz WCH3)3 PCH2 [CWn 91.7 88.5 85.1 78 .O 74.0 73.3 72.4 70.0 66.0 15.0 5.8 2.0 2.2 2.7 3.5 22.5 22 22 26.5 17 7 3 2.0 2.0 2.0 4.5 > 22 > 22 >22 N 22 3 2.5 1.6 1.6 1.8 1.8 7.0 7.5 7.7 7.3 5.5 1.4 1.5 16 17 16 25 >40 > 40 >40 >40 N 40 24 24 11.5 12.0 13.0 13.0 33 31.5 36 34 34 23 23 a Spectra recorded at 220 MHz, Av+ in Hz, values for NCH, refer to line width at half height minus the contribution from indirect coupling ( J ) .In the high resolution n.m.r. spectra little or no differential broadening of reson- ances was observed for CsTAB rich solutions or for solutions with compositions close to the SDS/water axis. Differential peak broadening was observed for samples in the viscoelastic region with the line width being of the magnitude expected from the T, measurement; the values are shown in table 3 and fig.7. It can be seen that a dramatic change in line widths occurs at ca. 77 mol % SDS, and that the change is greater at the head group than at the terminal methyl group. The residual HOD resonance had a line-width determined by magnet inhomogeneity in all the samples.I60 VISCOELASTIC MICELLAR SOLUTIONS 1 I 1 I I 100 90 80 70 60 mol % SDS FIG. 7.-Variation of n.m.r. linewidths with SDS/CaTAB ratio. 0, OCHz ; A, CHzCHJ ; x, N(CH3)J. Changes in n.m.r. relaxation times are related to changes in translational and rotational molecular motions? Motions with correlation times (z,) of the order of 10-l' s (normal liquid motions) result in equal values for Tl and T2 ( N 1 s).A decrease in molecular motions reduces TI and T2 from about 1 s to about 20 ms where Tl has a minimum value (7, N s). For slower motions Tl increases again while T2 becomes very small (T,<10-100 P S ) . ~ In the normal micellar liquids, SDS and CsTAB micelles have Tl N T2 = 1 s, showing fast molecular motion (z, 21 10-l O - lo-" s). If the addition of CsTAB to SDS micelles caused a gradual reduction in molecular motion then Tl and T, would be reduced by similar amounts until T1 passed its minimum value of ca. 20 ms. This is not observed, since TI is always in the range 0.5-1 s while T2 changes over the range 0.5-0.025s. Therefore, it can be concluded that there is a discontinuous change in motion at concentration sea.77 mol % SDS, with some surfactant molecules having z,> - s. This can be explained by either : (a) the presence of two types of micelle, one spherical with rapid motion (2, < 10-lo s) and the second having restricted mobility (zc> s) ; or (b) a change in micellar type to a large cylindrical (or other shaped) micelle which has regions of both fast and restricted molecular motion. Further, since only a single high resolution n.m.r. spectrum is observed, the rate of molecular exchange bdween the different species or environments must be faster than lo4 s-l. The n.m.r. line widths are dominated by the contributions from surfactant molecules with 2,) - lo-* s. It should be noted that although some molecular motions may be restricted (such as chain bending) overall rotation of the molecule about the long axis is likely to be rapid (2," 10-lo) as is the case in liquid crystalline phases.A comparison of SDS and CsTAB a-CH2 group line widths with those of terminalC. BARKER, D. SAUL, G . TIDDY, B . WHEELER, E. WILLIS 161 methyl groups indicates that in the restricted regions the head group is more immobil- ised than the interior of the micelle. Also, the CsTAB a-CH2 line widths are broader than those of the SDS a-CH, indicating that either the CsTAB chain is more re- stricted than the SDS chain or that proportionately more C8TAB molecules are pres- ent in restricted regions than SDS molecules. The NMe, resonance is relatively sharp because rotation about the N-Me bond is rapid, and takes place at varying angles to the N-CH2 group rotational axis.Theoretical calculations have shown that even with relatively slow rotational rates two axes of rotation can cause dramatic narrowing of resonance bands.7 DISCUSSION our results show that in the SDS/csTAB/Water system, an increase in Viscosity with shear thinning and viscoelastic behaviour occurs only at compositions which are close to the SDS rich side of the two liquid coexistence region. Viscoelastic behaviour was observed in solutions containing as little as 0.1 % surfactant but the existence of two isotropic liquids could not be detected under the microscope at concentrations below about 1 % (see fig. 1). The solutions become visibly turbid just before the two liquid phases were discernible under the microscope and the viscosity reached its maximum value at the SDS rich side of the line defining the onset of turbidity.The n.m.r. results demonstrate that the surfactant molecules in the viscoelastic solutions exist in both " restricted " and " liquid-like " environments. It is difficult to envisage how both of these could arise from a single type of micellar species although this possibility cannot be completely excluded. A more likely situation is that there are two,types of micellesexistinginequilibrium, normal spherical micelles and restricted micelles which axe probably cylindrical and of the type proposed by Pilpel for surfactant/electrolyte/water systems. The motion of surfactant molecules in a restricted cylindrical micelle would involve lateral diffusion along the surface of the micelle, together with chain rotational and bending motions.If a diffusion coefficient of cm2 s-l is postulated, then the time required for diffusion of the surfactant molecule around the circular cross section of a cylindrical micelle is ca. s (from D = Z2/6t). This is the same as or longer than the longest z, limit deduced from the n.m.r. measurement, and thus the existence of a cylindrical micelle with slow overall rotation about its long axis would be sufficient to account for the n.m.r. line widths. The alkyl chains could be assumed to undergo chain bending motions (involving internal rotations about CH,-CH2 bonds) and this would contribute towards the narrower lines observed for the CH2 and CH, groups further away from the head group.The presence of cylindrical micelles would account for the rheological properties of the system. Under static conditions the cylinders would be randomly oriented by Brownian motion. In flow the cylinders will become aligned with the amount of order increasing with shear rate and the resistance to flow decreases with increasing order, giving shear thinning behaviour. Whcn flow ceases, the cylinders again resume a random configuration and some of the energy used to align them is released, i.e. viscoelastic behaviour. However, due to the proximity of the viscoelastic mixtures to the two liquid phase boundary we cannot exclude the possibility that the rheological properties are due to the presence of a micro-emulsion. The onset of turbidity is probably due to the presence of emulsion droplets which are large enough to scatter light but are too small to be seen in transmitted light by the microscope used in the phase studies.Further, before the turbidity becomes apparent to the eye there may be very small emulsion 1-61 62 VISCOELASTIC MICELLAR SOLUTIONS droplets which are interacting to form a network throughout the solution leading to shear thinning and viscoelastic properties. The emulsion (of a small amount of a surfactant rich phase in a sodium bromide rich continuous phase) would contain the restricted surfactant molecules in one of the phases, but these would not be responsible for the viscoelasticity. An attempt to differentiate between the two mechanisms by separation of a turbid viscoelastic solution into the two component liquids by ultra-centrifugation proved to be inconclusive. The sample separated into a small fraction of an upper surfactant rich layer which was viscous and shear thinning but not elastic, and a much larger volume of an aqueous layer with low viscosity (- 1-2 cP). The only feature in the n.m.r. spectra of either layer was the sharp resonance from HDO. (The concentration of surfactant in the aqueous layer may have been insufficient for the observation of an n.m.r. spectrum.) However, it would seem that the restricted micellar environment observed by n.m.r., is a property of the surfactant rich layer. Nevertheless, visco- elasticity was only observed where the system was in the dispersed state. Thus we are unable to unequivocally distinguish whether the rheological properties are due to cylindrical micelles or to interacting micro-emulsion droplets. However, in other mixed surfactant systems which do not have a two liquid coexistence region the viscoelastic behaviour is undoubtedly due to the presence of " restricted " cylindri- cal micelles. D. H. Chen and D. G. Hall, KolloidZ. 2. Polymere, 1973, 251, 41. G. J. T. Tiddy, J.C.S. Faraday I, 1972, 68, 608. K. Walters, Basic Concepts and Formulae for Rheogoniometer (Sangamo Controls Ltd., 1968). K. Walters and R. A. Kemp, Deformation and Flow of High Polymers, ed. R. Wetton and R. Whorlow (McMillan, London, 1968), chap. 18. A. R. Eastwood, B. Yates and H. A. Barnes, Rheologica Acta, 1973, in press. N. Bloembergen, E. M. Purcell and R. V. Pound, Phys. Rev., 1948, 73, 679. N. Pilpel, J. Colloid Sci., 1954, 9, 285. D. Saul, G. J. T. Tiddy, B. A. Wheeler, P. A. Wheeler and E. Willis, J.C.S, Faraduy I, 1974,70, 163. ' J. G. Powles and H. S. Gutowsky, J. Chem. Phys., 1953, 21, 1695.

ISSN:0300-9599    

DOI:10.1039/F19747000154

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Phase Structure and Rheological Properties of a Mixed Zwit terionic/ Anionic Surfact ant Sys tern BY DOROTHY SAUL, GORDON J. T. TIDDY, BARBARA A. WHEELER, PHILLIP A. WHEELER* AND EDWIN WILLIS Unilever Research, Port Sunlight Laboratory, Port Sunlight, Wirral, Cheshire L62 4XN Received 5th July, 1973 The phase diagram of the mixed zwitterionic/anionic surfactant system hexadecyldimethyl- ammoniopropanesulphonate/sodium dodecyl sulphate/water has been determined in the aqueous region (> 90.0 % water) by optical microscopy and low angle X-ray scattering ; the phases observed were an isotropic surfactant solution and a hexagonal liquid crystalline phase. Some aqueous solutions were found to be viscoelastic and the composition boundaries of solutions with these properties were parallel to the phase boundaries.N.m.r. was used additionally to study the structure of the viscoelastic solutions and the results are interpreted using a model which involves the existence of both normal spherical niicelles and cylindrical micelles in equilibrium. In this laboratory the rheology and phase equilibria of various types of nixed surfactant solutions have been studied. Notably a mixed cationic/anionic surfactant system was investigated with particular emphasis on the aqueous region of the three component (cationic/anionic/water) phase diagram. In this system viscoelastic behaviour was observed for compositions close to the anionic side of a two liquid coexistence phase boundary and such samples showed line broadening in the high resolution nuclear magnetic resonance (n.m.r.) spectra.The n.m.r. data were interpreted as showing the existence of two types of micellar species, a normal spher- ical micelle with rapid motion and a second type with restricted molecular mobility, which is asymmetric, probably cylindrical. In this present study the phase diagram of the mixed zwitterionic/anionic surfactant system hexadecyldimethylammoniopropanesulphonate (HDPS)(I)/sodium dodecyl sulphate (SDS)/water has been determined in the aqueous region (> 90 % water) by optical microscopy and low angle X-ray scattering. The relationship between phase boundaries and the extent of viscoelasticity in the system has been examined. Addi- tionally, both high resolution and pulsed n.m.r. were used to study the structure of the viscoelastic solutions.CH3 I I CH3 C16H33Nf-CH2CH2CH20SO; (1) EXPERIMENTAL MATERIALS HDPS was prepared by the method of Clunie et aL3 and after recrystallisation from acetone+isopropanol was >99 % pure by elemental analysis and exhibited no minimum in the surface tension against log concentration plot (c.m.c. = 4.5 x mol dmV3 at 163164 VISCOELASTIC MIXED SURFACTANT SOLUTIONS 308 K). SDS was B.D.H. specially pure grade with c.m.c. = 8.3 x mol dm-j (at 298 K) and was used without further purification. DzO was p Chemicals 99.7 % pure grade. HzO was deionised and distilled. MEASUREMENTS Since HDPS is not soluble to the extent of 10 % by weight at temperatures below - 303 K all the experiments were done at 308 K. Mixtures containing up to 10 % total surfactant were prepared by adding water to the solids and leaving them to equilibrate for 2-6 days at 308 K.Phase studies were performed using a Reichert-Neopan polarising microscope fitted with a hot stage, and low angle X-ray measurements were obtained using Rigaku-Denki equip- ment. Rheological measurements under conditions of steady shear were obtained using a Haake Rotovisko viscometer fitted with the NV cup and bob assembly, and under conditions of both steady and oscillatory shear using a Weissenberg rheogoniometer 4* fitted with a 5.0 cm diameter cone and plate. The range of compositions where mixtures showed viscoelastic properties was determined visually by the titration technique described previously.2 The order of addition of the separate surfactant solutions was found to be important. The boundary of the viscoelastic region on the SDS side was approached from the SDS axis and vice versa on the HDPS side.Nuclear magnetic resonance measurements were made using a Bruker B-KR 322s 4-62 MHz pulse spectrometer and a Perkin-Elmer R12A 60 MHz high resolution spectro- meter. RESULTS AND DISCUSSION 1. OPTICAL OBSERVATIONS AND X - R A Y ANALYSIS The phase diagram of the aqueous region is shown in fig. 1. Samples in the liquid crystal (LC) and liquid+ LC regions were optically transparent when viewed through the microscope with unpolarised light but striations in incipient geometric texture (Rosevear’s classification 6, indicative of hexagonal phase structure were visible when the samples were viewed through crossed polarising lenses.SDS FIG. 1.-The dilute region of the HDPS/SDS/water phase diagram.TABLE l.-X-RAY SPACINGS AND STRUCTURE PARAMETERS FOR THE LIQUID CRYSTAL PHASE IN HDPS/SDS SYSTEM sample 6 % HDPS 4 % SDS 90 % H2O 7 % HDPS 3 % SDS 90 % H2O 10.5 % HDPS 4.5 % SDS 85 % H2O 14 % HDPS 6 % SDS 80 % H2O observed spacings dolnm (do/t/3)/nm (d0P)lnm 10.82 6.05 5.36 k0.15 k0.06 j--0.05 10.82 6.71 5.51 k0.15 k0.07 k0.06 9.95 5.66 4.75 10.14 k0.06 k0.05 8.81 4.94 4.32 k0.12 k0.05 kO.04 (do/ +/ 7 )/nm 3.98 & 0.03 4.23 & 0.03 3.57 rt 0.03 3.43 0.03 volume fraction of surfactan t Y O 0.108 k 0.002 0.108 & 0.002 0.172 k0.15 0.244 f 0.004 lattice parameter+ dpbm 12.29 k0.15 12.89 k0.15 11.18 k0.003 10.10 k0.15 * calculated according to the method described in ref. (9) cylinder diameter+ dclm 4.23 - + 0.10 4.45 & 0.10 4.87 - +0.12 5.24 k0.13 water spacing* dinterlm 8.06 0.25 8.44 f 0.25 6.31 f 0.27 4.86 0.28 surface hydrophilic area per group SL/nrn2 0.58 & 0.04 0.57 & 0.04 0.56 & 0.04 0.55 & 0.04166 VISCOELASTIC MIXED SURFACTANT SOLUTIONS X-ray diffraction data were obtained for composition ratios of HDPS : SDS of 7 : 3 and 6 : 4 at 10 % total concentration and for the ratio 7 : 3 at concentrations of 15 and 20 %.The spacings (table 1) show that the LC has hexagonal phase structure i.e. hexagonally packed cylinders. Table 1 also shows values of the lattice parameters d,,, the diameter of the surfactant cylinders do the resulting inter cylinder water spacing tiinter and the surface area available for each polar head group, S.For a weight ratio of HDPS : SDS of 7 : 3 (mole ratio 1.7 : 1) the cylinder diameter increases from 4.55 nm to 5.25 nm when the concentration is increased from 10 to 20 %. This 16 % increase in diameter is probably due to an increase in ‘‘ trans ” chain conformations, or to a decrease in the aqueous content of the head group environment. It is noteable that for the potassium oleate/water system Ekwall et aL7 found that doubling the potassium oleate concentration in the hexagonal mesophase region increased the cylinder diameter by less than 1 %. The average area per polar head group was, within experimental error, invariant with concentration. 15- 10- pc 1 F 5- 2. VISCOELASTIC PROPERTIES The boundary of viscoelastic behaviour determined by observing recoil in swirled solutions is shown in fig.1 and extends beyond the mesophase regions. All samples within the LC and liquid+LC boundaries were viscoelastic. As mentioned earlier the method of solution preparation had an important bearing on the viscoelastic behaviour. Dilution of a 5 % mixture with a weight ratio of HDPS : SDS of 7 : 3 to a concentration of 1 % gave a viscoelastic solution, however, the viscoelasticity slowly decayed, finally disappearing after several days. A 1 % mixture of identical composition prepared by mixing separate 1 % solutions of the surfactants did not show viscoelastic properties. The lowest concentration at which viscoelasticity was observed on mixing separate solutions was about 2 %. 2ol 1:- ISOTROPIC PHASE 5 K L+LC + LC - ~ L + L C X I R O ~ I C YHDPSi 4 a 3 A 4 R S D S 9 8 7 6 5 4 3 2 1 0 composition FIG.2.-Viscosity, 9, against composition for HDPS/SDS samples at 10.0 % total surfactant con- centration at shear rate = 17.7 s-’ : 0, original results ; A, repeated results.D. SAUL, G . TIDDY, B . WHEELER, P . WHEELER AND E. WILLIS 167 50 3. RHEOLOGICAL MEASUREMENTS Samples whose compositions lay within the LC and liquid + LC regions were shear thinning, viscoelastic and exhibited normal forces (i.e. when sheared they exerted a force in the direction normal to the direction of flow). Values of the first normal stress difference (a,) were calculated from measurements of normal force. Plots of viscosity (q) and of o1 against shear rate (9) were obtained from measurements made under steady shear conditions.The plots of viscosity against composition across sections of the phase diagram illustrated in fig. 2-6 were obtained from the former. - ISOTROPIC "T 0 b composition FIG. 3.-Viscosity, 77, against composition for HDPS/SDS samples at 10.0 % total surfactant con- centration : 0, shear rate = 1.12 s-' ; A, shear rate = 7.06 s-l. Plots of dynamic viscosity (q') and of dynamic rigidity (G') against angular fre- quency (0) and against 2w were obtained from measurements made under oscillating shear. The plots of q against 9 and of q' against w or 2w and those of crl against w or 2co and those of c1 against 9 and of G' against w or 2w were expected to superimpose. For the 1 % and 5 % concentrations the position of the maximum in the viscosity against composition curves is independent of shear rate, however, at a concentration of 8 % the maximum shifts to a higher HDPS : SDS ratio as the shear rate is increased.At 10 % concentration three maxima were observed, that corresponding to the LC region of the phase diagram was independent of shear rate and reproducible. The other two maxima which roughly correspond to the liquid + LC/isotropic solution phase boundaries on each side of the LC region were only apparent at the higher shear rates and were difficult to reproduce. The oscillatory and the steady shear data were168 VISCOELASTIC MIXED SURFACTANT SOLUTIONS not superimposable in these regions which together with the difficulties of repro- ducibility suggests that the equilibrium of the systems was being disturbed by shearing.Pulsed n.m.r. measurements on a solution of composition 3.5 % HDPS, 1.5 % SDS, 95 % D20 gave the value for the spin-lattice relaxation time, TI of ca. 5 x 10-l s and a spin-spin relaxation time, T2 of ca. 2 x s. High resolution spectra of this system showed broad lines (10-20 Hz) whereas a sample of concentration 1 %, prepared by mixing 1 % solutions of the individual surfactants showed sharp reson- ances. A sample of identical composition to the latter, prepared by diluting the viscoelastic 5 % sample initially gave broad n.m.r. lines which gradually became sharper over a period of several days. composition FIG. 4.-Viscosity, 7 against composition for HDPs/SDS samples at 8.0 % total surfactant concen- tration; 0, shear rate = 2.81 s-' ; A, shear rate = 7.06 s-l; 0, shear rate = 17.74 s-l.N.m.r. measurements were also made on samples of 5 % concentration but of varying composition. Fig. 5 illustrates how n.m.r. line width, viscosity and first normal stress difference vary with composition. It is noticeable that each of the curves exhibits a maximum at approximately the same composition, indicating that the entities responsible for the rheological properties are also responsible for the broadening of the n.m.r. lines. Also the plot of linewidth of the hydrocarbon chain -CH2-- peak against conceptration (fig. 7) shows a definite break in the region of 2.0-2.5 % surfactant above which the slope is linear. The discontinuity appears to represent a c.m,c. for the micelles with restricted molecular motion and the positionD.SAUL, G. TIDDY, B . WHEELER, P. WHEELER AND E . WILLIS 169 of the break corresponds to the minimum concentration at which viscoelasticity is observed. The slow decay of viscoelasticity on dilution from 5.0 % total surfactant shows that the units responsible for the viscoelastic behaviour have breakdown times which are much longer than the time scales usually associated with micellar equilibria. composition hydrocarbon chain (0) against composition of 5 % HDPs/SDS solutions. FIG. 5.-Viscosity (a), first normal stress difference ( x ) at = 2.81 s-I and n.m.r. line width of the I I I I 1 4.0 728 1 .o HDPS 3.2 3.4 3.6 SDS 1 8 16 1 *4 composition (%) FIG. 6.-Viscosity, 7, against composition for HDPSlSDS samples at 1.0 % total surfactant con- centration : 0, shear rate = 2570 s-' ; A, shear rate = 1235 s-' ; 0, shear rate = 428 s-l.170 VISCOELASTIC MIXED SURFACTANT SOLUTIONS 2og O' ; 4 i k 4 k-" % HDPSISDS FIG.7.-N.m.r. line width (hydrocarbon chain -CH2-- peak) against surfactant concentration of HDPS/SDS mixtures (ratio HDPS : SDS = 3.5 : 1.5). The n.m.r. spin-lattice (TI) and spin-spin (Tz) relaxation times recorded for a visco- elastic 5.0 % surfactant solution were not equal as would be the case for normal liquid motions which have correlation times (2,) of the order of 10-l' s. Non-equality of TI and T2 is observed for slower molecular motions but for any particular system there are fixed pairs of values of Tl and T2 which correspond to a given correlation time. The experimental results found for this system do not correspond to such a pair and by similar arguments to those used for the cationic/anionic/water system more than one micellar species may be present one of which has slow molecular motion (correlation time s).It seems likely that the micelles with restricted molecular motion are cylindrical and could be regarded as precursors of the hexagonal LC phase. Such micelles would be expected to give rise to the observed rheological properties. Further, the slow decay of viscoelasticity on dilution is probably associated with the breakdown of the cylindrical micelles present at 5 % concentration into the spherical micelles which are present in 1 % concentrations at equilibrium. Studies of the kinetics of this process are continuing.H. A. Barnes, A. R. Eastwood and B. Yates, RheoZogica Acta, in press. C. A. Barker, D. Saul, G. J. T. Tiddy, B. A. Wheeler, E. Willis, J.C.S. Faraday I, 1974,70,154. J. S. Clunie, J. M. Corkhill, T. F. Goodman and C. P. Ogden, Trans. Faraday Soc., 1967, 63, 505. K. Walters, Basic Concepts and Formulae for the Rheogoniometer (Sangamo Controls Ltd., 1969). K. Walters and R. A. Kemp, Deformation and Flow of High Polymers, ed. R. Wetton and R. Whorlow (McMillan, London, 1968), chap. 18, p. 237. F. B. Rosevear, J. Amer. Oil Chem. SOC., 1954, 31, 635 (fig. 24). ' P. Ekwall, L. Maodell and K. Fontell, J. CoZZoid Interface Sci., 1969, 31, 508. Phase Structure and Rheological Properties of a Mixed Zwit terionic/ Anionic Surfact ant Sys tern BY DOROTHY SAUL, GORDON J.T. TIDDY, BARBARA A. WHEELER, PHILLIP A. WHEELER* AND EDWIN WILLIS Unilever Research, Port Sunlight Laboratory, Port Sunlight, Wirral, Cheshire L62 4XN Received 5th July, 1973 The phase diagram of the mixed zwitterionic/anionic surfactant system hexadecyldimethyl- ammoniopropanesulphonate/sodium dodecyl sulphate/water has been determined in the aqueous region (> 90.0 % water) by optical microscopy and low angle X-ray scattering ; the phases observed were an isotropic surfactant solution and a hexagonal liquid crystalline phase. Some aqueous solutions were found to be viscoelastic and the composition boundaries of solutions with these properties were parallel to the phase boundaries. N.m.r. was used additionally to study the structure of the viscoelastic solutions and the results are interpreted using a model which involves the existence of both normal spherical niicelles and cylindrical micelles in equilibrium.In this laboratory the rheology and phase equilibria of various types of nixed surfactant solutions have been studied. Notably a mixed cationic/anionic surfactant system was investigated with particular emphasis on the aqueous region of the three component (cationic/anionic/water) phase diagram. In this system viscoelastic behaviour was observed for compositions close to the anionic side of a two liquid coexistence phase boundary and such samples showed line broadening in the high resolution nuclear magnetic resonance (n.m.r.) spectra. The n.m.r. data were interpreted as showing the existence of two types of micellar species, a normal spher- ical micelle with rapid motion and a second type with restricted molecular mobility, which is asymmetric, probably cylindrical.In this present study the phase diagram of the mixed zwitterionic/anionic surfactant system hexadecyldimethylammoniopropanesulphonate (HDPS)(I)/sodium dodecyl sulphate (SDS)/water has been determined in the aqueous region (> 90 % water) by optical microscopy and low angle X-ray scattering. The relationship between phase boundaries and the extent of viscoelasticity in the system has been examined. Addi- tionally, both high resolution and pulsed n.m.r. were used to study the structure of the viscoelastic solutions. CH3 I I CH3 C16H33Nf-CH2CH2CH20SO; (1) EXPERIMENTAL MATERIALS HDPS was prepared by the method of Clunie et aL3 and after recrystallisation from acetone+isopropanol was >99 % pure by elemental analysis and exhibited no minimum in the surface tension against log concentration plot (c.m.c.= 4.5 x mol dmV3 at 163164 VISCOELASTIC MIXED SURFACTANT SOLUTIONS 308 K). SDS was B.D.H. specially pure grade with c.m.c. = 8.3 x mol dm-j (at 298 K) and was used without further purification. DzO was p Chemicals 99.7 % pure grade. HzO was deionised and distilled. MEASUREMENTS Since HDPS is not soluble to the extent of 10 % by weight at temperatures below - 303 K all the experiments were done at 308 K. Mixtures containing up to 10 % total surfactant were prepared by adding water to the solids and leaving them to equilibrate for 2-6 days at 308 K.Phase studies were performed using a Reichert-Neopan polarising microscope fitted with a hot stage, and low angle X-ray measurements were obtained using Rigaku-Denki equip- ment. Rheological measurements under conditions of steady shear were obtained using a Haake Rotovisko viscometer fitted with the NV cup and bob assembly, and under conditions of both steady and oscillatory shear using a Weissenberg rheogoniometer 4* fitted with a 5.0 cm diameter cone and plate. The range of compositions where mixtures showed viscoelastic properties was determined visually by the titration technique described previously.2 The order of addition of the separate surfactant solutions was found to be important. The boundary of the viscoelastic region on the SDS side was approached from the SDS axis and vice versa on the HDPS side.Nuclear magnetic resonance measurements were made using a Bruker B-KR 322s 4-62 MHz pulse spectrometer and a Perkin-Elmer R12A 60 MHz high resolution spectro- meter. RESULTS AND DISCUSSION 1. OPTICAL OBSERVATIONS AND X - R A Y ANALYSIS The phase diagram of the aqueous region is shown in fig. 1. Samples in the liquid crystal (LC) and liquid+ LC regions were optically transparent when viewed through the microscope with unpolarised light but striations in incipient geometric texture (Rosevear’s classification 6, indicative of hexagonal phase structure were visible when the samples were viewed through crossed polarising lenses. SDS FIG. 1.-The dilute region of the HDPS/SDS/water phase diagram.TABLE l.-X-RAY SPACINGS AND STRUCTURE PARAMETERS FOR THE LIQUID CRYSTAL PHASE IN HDPS/SDS SYSTEM sample 6 % HDPS 4 % SDS 90 % H2O 7 % HDPS 3 % SDS 90 % H2O 10.5 % HDPS 4.5 % SDS 85 % H2O 14 % HDPS 6 % SDS 80 % H2O observed spacings dolnm (do/t/3)/nm (d0P)lnm 10.82 6.05 5.36 k0.15 k0.06 j--0.05 10.82 6.71 5.51 k0.15 k0.07 k0.06 9.95 5.66 4.75 10.14 k0.06 k0.05 8.81 4.94 4.32 k0.12 k0.05 kO.04 (do/ +/ 7 )/nm 3.98 & 0.03 4.23 & 0.03 3.57 rt 0.03 3.43 0.03 volume fraction of surfactan t Y O 0.108 k 0.002 0.108 & 0.002 0.172 k0.15 0.244 f 0.004 lattice parameter+ dpbm 12.29 k0.15 12.89 k0.15 11.18 k0.003 10.10 k0.15 * calculated according to the method described in ref.(9) cylinder diameter+ dclm 4.23 - + 0.10 4.45 & 0.10 4.87 - +0.12 5.24 k0.13 water spacing* dinterlm 8.06 0.25 8.44 f 0.25 6.31 f 0.27 4.86 0.28 surface hydrophilic area per group SL/nrn2 0.58 & 0.04 0.57 & 0.04 0.56 & 0.04 0.55 & 0.04166 VISCOELASTIC MIXED SURFACTANT SOLUTIONS X-ray diffraction data were obtained for composition ratios of HDPS : SDS of 7 : 3 and 6 : 4 at 10 % total concentration and for the ratio 7 : 3 at concentrations of 15 and 20 %.The spacings (table 1) show that the LC has hexagonal phase structure i.e. hexagonally packed cylinders. Table 1 also shows values of the lattice parameters d,,, the diameter of the surfactant cylinders do the resulting inter cylinder water spacing tiinter and the surface area available for each polar head group, S. For a weight ratio of HDPS : SDS of 7 : 3 (mole ratio 1.7 : 1) the cylinder diameter increases from 4.55 nm to 5.25 nm when the concentration is increased from 10 to 20 %.This 16 % increase in diameter is probably due to an increase in ‘‘ trans ” chain conformations, or to a decrease in the aqueous content of the head group environment. It is noteable that for the potassium oleate/water system Ekwall et aL7 found that doubling the potassium oleate concentration in the hexagonal mesophase region increased the cylinder diameter by less than 1 %. The average area per polar head group was, within experimental error, invariant with concentration. 15- 10- pc 1 F 5- 2. VISCOELASTIC PROPERTIES The boundary of viscoelastic behaviour determined by observing recoil in swirled solutions is shown in fig. 1 and extends beyond the mesophase regions. All samples within the LC and liquid+LC boundaries were viscoelastic.As mentioned earlier the method of solution preparation had an important bearing on the viscoelastic behaviour. Dilution of a 5 % mixture with a weight ratio of HDPS : SDS of 7 : 3 to a concentration of 1 % gave a viscoelastic solution, however, the viscoelasticity slowly decayed, finally disappearing after several days. A 1 % mixture of identical composition prepared by mixing separate 1 % solutions of the surfactants did not show viscoelastic properties. The lowest concentration at which viscoelasticity was observed on mixing separate solutions was about 2 %. 2ol 1:- ISOTROPIC PHASE 5 K L+LC + LC - ~ L + L C X I R O ~ I C YHDPSi 4 a 3 A 4 R S D S 9 8 7 6 5 4 3 2 1 0 composition FIG. 2.-Viscosity, 9, against composition for HDPS/SDS samples at 10.0 % total surfactant con- centration at shear rate = 17.7 s-’ : 0, original results ; A, repeated results.D.SAUL, G . TIDDY, B . WHEELER, P . WHEELER AND E. WILLIS 167 50 3. RHEOLOGICAL MEASUREMENTS Samples whose compositions lay within the LC and liquid + LC regions were shear thinning, viscoelastic and exhibited normal forces (i.e. when sheared they exerted a force in the direction normal to the direction of flow). Values of the first normal stress difference (a,) were calculated from measurements of normal force. Plots of viscosity (q) and of o1 against shear rate (9) were obtained from measurements made under steady shear conditions. The plots of viscosity against composition across sections of the phase diagram illustrated in fig.2-6 were obtained from the former. - ISOTROPIC "T 0 b composition FIG. 3.-Viscosity, 77, against composition for HDPS/SDS samples at 10.0 % total surfactant con- centration : 0, shear rate = 1.12 s-' ; A, shear rate = 7.06 s-l. Plots of dynamic viscosity (q') and of dynamic rigidity (G') against angular fre- quency (0) and against 2w were obtained from measurements made under oscillating shear. The plots of q against 9 and of q' against w or 2w and those of crl against w or 2co and those of c1 against 9 and of G' against w or 2w were expected to superimpose. For the 1 % and 5 % concentrations the position of the maximum in the viscosity against composition curves is independent of shear rate, however, at a concentration of 8 % the maximum shifts to a higher HDPS : SDS ratio as the shear rate is increased.At 10 % concentration three maxima were observed, that corresponding to the LC region of the phase diagram was independent of shear rate and reproducible. The other two maxima which roughly correspond to the liquid + LC/isotropic solution phase boundaries on each side of the LC region were only apparent at the higher shear rates and were difficult to reproduce. The oscillatory and the steady shear data were168 VISCOELASTIC MIXED SURFACTANT SOLUTIONS not superimposable in these regions which together with the difficulties of repro- ducibility suggests that the equilibrium of the systems was being disturbed by shearing. Pulsed n.m.r. measurements on a solution of composition 3.5 % HDPS, 1.5 % SDS, 95 % D20 gave the value for the spin-lattice relaxation time, TI of ca.5 x 10-l s and a spin-spin relaxation time, T2 of ca. 2 x s. High resolution spectra of this system showed broad lines (10-20 Hz) whereas a sample of concentration 1 %, prepared by mixing 1 % solutions of the individual surfactants showed sharp reson- ances. A sample of identical composition to the latter, prepared by diluting the viscoelastic 5 % sample initially gave broad n.m.r. lines which gradually became sharper over a period of several days. composition FIG. 4.-Viscosity, 7 against composition for HDPs/SDS samples at 8.0 % total surfactant concen- tration; 0, shear rate = 2.81 s-' ; A, shear rate = 7.06 s-l; 0, shear rate = 17.74 s-l. N.m.r. measurements were also made on samples of 5 % concentration but of varying composition.Fig. 5 illustrates how n.m.r. line width, viscosity and first normal stress difference vary with composition. It is noticeable that each of the curves exhibits a maximum at approximately the same composition, indicating that the entities responsible for the rheological properties are also responsible for the broadening of the n.m.r. lines. Also the plot of linewidth of the hydrocarbon chain -CH2-- peak against conceptration (fig. 7) shows a definite break in the region of 2.0-2.5 % surfactant above which the slope is linear. The discontinuity appears to represent a c.m,c. for the micelles with restricted molecular motion and the positionD. SAUL, G. TIDDY, B . WHEELER, P. WHEELER AND E . WILLIS 169 of the break corresponds to the minimum concentration at which viscoelasticity is observed.The slow decay of viscoelasticity on dilution from 5.0 % total surfactant shows that the units responsible for the viscoelastic behaviour have breakdown times which are much longer than the time scales usually associated with micellar equilibria. composition hydrocarbon chain (0) against composition of 5 % HDPs/SDS solutions. FIG. 5.-Viscosity (a), first normal stress difference ( x ) at = 2.81 s-I and n.m.r. line width of the I I I I 1 4.0 728 1 .o HDPS 3.2 3.4 3.6 SDS 1 8 16 1 *4 composition (%) FIG. 6.-Viscosity, 7, against composition for HDPSlSDS samples at 1.0 % total surfactant con- centration : 0, shear rate = 2570 s-' ; A, shear rate = 1235 s-' ; 0, shear rate = 428 s-l.170 VISCOELASTIC MIXED SURFACTANT SOLUTIONS 2og O' ; 4 i k 4 k-" % HDPSISDS FIG.7.-N.m.r. line width (hydrocarbon chain -CH2-- peak) against surfactant concentration of HDPS/SDS mixtures (ratio HDPS : SDS = 3.5 : 1.5). The n.m.r. spin-lattice (TI) and spin-spin (Tz) relaxation times recorded for a visco- elastic 5.0 % surfactant solution were not equal as would be the case for normal liquid motions which have correlation times (2,) of the order of 10-l' s. Non-equality of TI and T2 is observed for slower molecular motions but for any particular system there are fixed pairs of values of Tl and T2 which correspond to a given correlation time. The experimental results found for this system do not correspond to such a pair and by similar arguments to those used for the cationic/anionic/water system more than one micellar species may be present one of which has slow molecular motion (correlation time s). It seems likely that the micelles with restricted molecular motion are cylindrical and could be regarded as precursors of the hexagonal LC phase. Such micelles would be expected to give rise to the observed rheological properties. Further, the slow decay of viscoelasticity on dilution is probably associated with the breakdown of the cylindrical micelles present at 5 % concentration into the spherical micelles which are present in 1 % concentrations at equilibrium. Studies of the kinetics of this process are continuing. H. A. Barnes, A. R. Eastwood and B. Yates, RheoZogica Acta, in press. C. A. Barker, D. Saul, G. J. T. Tiddy, B. A. Wheeler, E. Willis, J.C.S. Faraday I, 1974,70,154. J. S. Clunie, J. M. Corkhill, T. F. Goodman and C. P. Ogden, Trans. Faraday Soc., 1967, 63, 505. K. Walters, Basic Concepts and Formulae for the Rheogoniometer (Sangamo Controls Ltd., 1969). K. Walters and R. A. Kemp, Deformation and Flow of High Polymers, ed. R. Wetton and R. Whorlow (McMillan, London, 1968), chap. 18, p. 237. F. B. Rosevear, J. Amer. Oil Chem. SOC., 1954, 31, 635 (fig. 24). ' P. Ekwall, L. Maodell and K. Fontell, J. CoZZoid Interface Sci., 1969, 31, 508.

ISSN:0300-9599    

DOI:10.1039/F19747000163

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Upper and Lower Critical Solution Temperatures in the 3- Polystyrene(3) Cosolvent System Acetone( 1) + Diethyl Ether(2) BY JOHN M. G. COWE* AND IAIN J. MCEWEN Chemistry Department, University of Stirling, Stirling, Scotland Received 1 1 th July, 1973 Phase diagrams for different molecular weight fractions of polystyrene in the mixed solvent system acetone+ diethyl ether have been obtained. The system shows a variation of upper and lower critical solution temperatures with solvent composition which indicates the cosolvent nature of the mixed solvent. No solvent composition is found which can dissolve high molecular weight poly- styrene (M> lo6). The separation of critical solution temperatures can be predicted by the Prigogine theory of polymer solution thermodynamics, but not the absolute values.Phase separation, which occurs when a polymer solution is cooled below the well known upper critical solution temperature (UCST), has been the subject of extensive study; UCST phenomena being widely employed as the basis of polymer fraction- ation. The opposite effect, that of phase separation on heating a polymer solution, has received much less attention. This second separation of a polymer solution into two liquid phases is associated with the large difference in thermal expansions of polymer and solvent and should, in principle, be observed for all non-polar polymer +solvent systems near the critical temperature of the solvent. The minimum in the coexistence curve occurring at these temperatures is termed the lower critical solution temperature (LCST).The polymer + solvent systems polystyrene + acetone and polystyrene + diethyl ether have been studied by Patterson and coworkers.2 Both solvents are compara- tively " poor " in that they dissolve only low molecular weight polystyrene fractions. The effect of this is to cause the UCST and LCST to be separated by a relatively small temperature range. As the molecular weight of the polymer is increased, the range of complete polymer-solvent miscibility decreases, i.e., the UCST is raised and the LCST is lowered. At a certain molecular weight the UCST and LCST coalesce, so that the two regions of immiscibility merge, giving an " hour-glass " shaped phase diagram. Mixtures of relatively poor solvents can, in some cases, produce enhanced solvent power.When this happens the mixed solvent is said to exhibit a synergistic effect. The enhancement of solubility may be detected by a study of the limiting viscosity number or the second virial coefficient measured as a function of solvent composition at a particular ternperat~re.~ Alternatively, a more extensive study of the solubility of a polymer in the mixed solvent may be carried out which illustrates more fully the behaviour as a function of both temperature and solvent composition. In this paper we have examined the phase equilibria for several polystyrene fractions (component 3) in the mixed solvent acetone (component l)+diethyl ether (component 2). The variation of UCST and LCST with solvent composition reflects the cosolvent behav- iour of acetone +ether mixtures and establishes the limits of solution in the system.171172 CRITICAL SOLUTION TEMPERATURE An attempt to predict both UCST and LCST using the Prigogine-Flory theory of polymer solution^,^ as developed by Patterson and coworker^,^ has been made and the results assessed. EXPERIMENTAL The polystyrene samples used were obtained from the Pressure Chemical Co. who quote M,/M, ratios of less than 1-06. The solvents used were best grade and were fractionally distilled before use. UCST and LCST were estimated from cloud-point curves. These were determined optically in thick-walled Pyrex tubes (i.d. = 5 mm) essentially as described previously.6 For a true monodisperse sample the maxima and minima of the cloud point curves can be taken to represent, respectively, the UCST and LCST for the polymer.It is known that polydispersity displaces the critical point to higher polymer concentration ’ ; however, no indication of displacement, which would result in a point of inflexion in the cloud-point curve, was observed throughout this study. Accordingly we have taken the CST as the turning points of the coexistence curves. RESULTS Typical cloud-point curves for the system are shown in fig. 1. In pure acetone, polystyrene fraction M, = 20 400 gives an hour-glass phase diagram which is almost identical to that obtained by Patterson and coworkers for a 19 800 fraction (fig. lA, curve a). At temperatures and compositions inside the hour-glass a two phase region exists where acetone and polystyrene are immiscible. On addition of ether to solutions of polystyrene in acetone ($1 = 0.95) the phase diagram opens out to form two regions of immiscibility separated by a continuous one phase region; both UCST and LCST can now be identified.The effect of increasing the ether content of the mixed solvent in the range 41 = 0.91 to 41 = 0.67 leads to further separation of the UCST and LCST as shown by curves c and din fig. 1A and 1B. This behaviour is also observed in the single solvent acetone but by using polymer fractions of decreasing molecular weight. At compositions beyond $1 -0.67 there is a tendency for the critical temperatures to begin converging once again. The cloud-point curves for polystyrene M, = 20 400 in pure ether and at 41 = 0.05 are shown in fig. lB, curves e and$ With the next fraction in the series, polystyrene M, = 37 0o0, the UCST and LCST are separated by a single phase region when the solvent composition lies between C#I~ = 0.82 and 0.05.Fig. 1C shows the acetone-rich end of the composition range. At 41 = 0.82 the UCST and LCST have just separated while at 41 = 0.83 the regions of immiscibility have merged to give an hour-glass diagram. The dashed line in this diagram represents an estimate. The two phase nature of this region was established by heating a solution of composition 43 = 0.13 to 320 K ; a homogeneous solution was not obtained. With increasing ether content the cloud-point curves for fraction M,,, = 37 O00 are further separated as can be seen from ourves c and d of fig. 1C whose corresponding UCST have decreased to 250 and 203 K respectively.As fractions of higher molecular weight are used the solvent composition range over which solution can be effected is further decreased. Cloud point curves for polystyrene fraction Mw = 860 000 are shown in fig. 1D. The highest fraction studied (Mw = 2 x lo6) could not be dissolved under any of the possible conditions suggested by fig. 2 in which the CST for all the polystyrene fractions are plotted as a function of solvent composition. (It should be noted that this is not a true phase diagram as the volume fraction of the polymer at the critical temperatures is a functionJ . M. G . COWIE AND I . J . MCEWEN 173 I- l , l 3 I 1 L 43 (C) 0.0 0. I 0.2 0.3 FIG. 1.-A-D. Typical cloud-point curves for the system acetone+ether+polystyrene.+1 =volume fraction acetone in the solvent mixture, 43 = volume fraction polymer in solution. A & B : fraction 20 400; (a) dl = 1.00, (6) 41 = 0.95, (c) +1 = 0.91, (d) +1 = 0.67, (e) 41 = 0.00, (f) +1 = 0.05. C : fraction 37 000; (a) = 0.83, (b) = 0.82, (c) +1 = 0.71, (d) +l = 0.50. D : fraction 860000; (a) (bl = 0.30, (6) 41 == 0.33, (c) $1 = 0.38.174 CRITICAL SOLUTION TEMPERATURE of both M, and solvent composition.) The resulting set of contours represents the variation of CST with solvent composition and defines the regions of solubility of polystyrene in the liquid mixtures. It can be seen that high molecular weights will not dissolve in any composition of the mixed solvent. 0 0.5 I #2 FIG. 2.-Plots of UCST and LCST against 42 for all fractions.d2 = volume fraction ether in solvent mixture. (a) Fraction 20 400, (6) fraction 37 OOO, (c) fraction 110 OOO, ( d ) fraction 267 000, (e) estimated curve for fraction 411 000, (f) fraction 860 000. FIG. 3.-UCST and LCST, taken from sections on fig. 2, as a function of r-*. (a) q51 = 0.35, (b) #1 = 0.50, (c) #1 = 0.20. Dashed lines calculated and fitted as described in text. This molecular weight dependence of the polymer solubility in the mixed solvent is illustrated more clearly in fig. 3, where cross-sections at constant solvent composition are plotted as a function of polymer chain length. Table 1 contains the values of UCST and LCST for all the fractions studied, the fraction molecular weights and the solvent compositions.J . M . G. COWIE AND I . J .MCEWEN 175 TABLE 1.-UCST AND LCST FOR POLYSTYRENE FRACTIONS IN THE MIXED SOLVENT ACETONE+ DIETHYL ETHER fraction Mw dl LCST/K UCST/K - - 1 .oo 0.95 367 310 0.91 375 292 0.83 383 261 0.67 386 219 0.20 362 187 0.09 345 197 0.05 333 210 0.00 316 230 20400 0.40 379 189 0.82 326 312 0.71 358 250 37000 0.50 364 203 0.20 341 193 0.05 297 239 fraction Ma 41 LCST/K UCST/K 0.17 295 227 0.30 322 207 llOOO0 0.40 327 213 0.50 328 230 0.60 3 14 266 0.24 290 23 1 0.30 304 222 267000 0.40 3 10 225 0.47 307 239 0.50 301 - 411000 0.31 291 233 0.30 268 248 86OOOO 0.33 275 245 0.38 274 252 DISCUSSION To date, only one publication has dealt with both UCST and LCST in a quasi- ternary system.8 The system studied was polystyrene disrolved in solvent + non- solvent mixtures and points of similarity can be seen.Hour-glass shaped phase diagrams were obtained and the separation of UCST and LCST increased with improving solvent power, The system acetone( 1) + ether(2) +polystyrene(3), although approximating to a solvent + non-solvent system at low molecular weights, is essen- tially one of two poor solvents. The behaviour of the cloud point curves bears some resemblance to that reported by Wolf et aL8 but is due to the cosolvent nature of the mixtures. The cosolvent nature and the separation of UCST and LCST with im- proving solvent power is illustrated in fig. 2. Synergism is indicated by the increase in solubility with any mixture compared with that of the pure components. The " best " solvent mixture, which we may define as the one which will dissolve the highest molecular weight fraction, appears to occur at approximately Analysis of existing vapour pressure data for the binary liquid system acetone+ ether shows that, at mole fraction 0.5 and at 303 K, AGE = 452 Jmol-l, i.e., acetone and ether are moderately incompatible as evidenced by the positive excess free energy of mixing.The cosolvent action which lowers the UCST also leads to a less easily explained raising of the LCST above that expected for a linear interpolation between the two components. The effect can be explained qualitatively if one postulates that the introduction of a polymer chain into the binary solvent environment serves to bind the acetone and ether molecules more strongly by acting as a bridge between these relatively incompatible species. This would have the effect of reducing the expected rate of expansion of the binary liquid pair relative to that of the polymer, thereby raising the LCST. This is in keeping with the nature of cosolvent systems already studied ; qualitatively, the cosolvent effect of acetone +ether mixtures can be ex- plained by a preference for (1-2-3) contacts over (1-2), (1-3) or (2-3) contacts.2 It is interesting to note that high molecular weight polystyrene is soluble, in the appropriate mixture of acetone and ether, only at temperatures well below room = 0.34.176 CRITICAL SOLUTION TEMPERATURE temperature.This somewhat surprising result indicates that use of a mixed solvent for polymer studies should be approached with some caution unless the phase rela- tions for the system have previously been determined.Since attempts to dissolve a polystyrene fraction of MW = 2 x lo6 under the conditions defined by the area inside the contour for the 860 OOO fraction (fig. 2) were not successful, it was concluded that no single mixture of acetone and ether will act as a theta solvent for polystyrene. This is confirmed by plotting, in fig. 3, UCST and LCST against r-i for three sections, each at a constant but different solvent composi- tion, from fig. 2. Here r has been taken as the degree of polymerisation and not the ratio of molar volumes. The data form a curve at each solvent composition such that at high values of r the UCST and LCST coalesce. The curves define the solubil- ity, at a particular q51, as a function of chain length.The single phase region lies within the bounds of the curve, outside this a homogeneous solution cannot form. This behaviour has already been reported for the system secondary cellulose acetate + acetone. The simpler, two parameter, theories of polymer solutions l o do not predict LCST nor do they allow curvature in (CST, r-3) plots. The three parameter Prigogine theory,4* which considers changes in free volume of the Components, has been applied to systems showing LCST. As yet no refinement is available which takes account of quasi-ternary systems. We have made the assumption that the theory, as it stands, can be applied to the liquid mixtures if these are treated as single liquids and have derived the required " average " parameters using the ideal mixing rule.Patterson and Delmas have shown," using the Prigogine theory, that at the point of critical miscibility Here 3c, is the number of external degrees of freedom of the solvent molecule, z is a measure of the difference in free volume of the components of the mixture and is obtained from expansion data, while v2 represents the difference in the chemical nature of the components. v17 the reduced volume of the solvent, may also be ob- tained from expansion data. By solving eqn (1) the variation of CST with molecular weight may be predicted. Values of c1 and T~ for both acetone and ether have been calculated by Patterson et aZ.* from the equation of state data of Flory and Eichinger.12 These values, and those of the temperature reduction parameters, TT, are shown in table 2.We define an average parameter for the solvent as where xi is the parameter for the pure solvent and 4i is the volume fraction. Values of c1v2 are obtained by a fitting technique ; the value of v2 is adjusted SO that the coalescence point of UCST and LCST, predicted by (I), is the same as that indicated by the experimental curves in fig. 3. These procedures are fully described elsewhere. The results of the theoretical calculation are shown by the dashed curves in fig. 3. In each case, the theory has successfully predicted the general shape of the experi- mental curves and has almost exactly reproduced the separation of UCST and LCST. The major weakness is that the theory fails to predict the absolute values of the CST correctly. In order to match the theoretical and experimental curves it is necessary to displace the temperature axes.This procedure has been adopted for other systems.2* Eqn (1) applies at zero or negligible pressure. The cloud point j i = 41x1 +42x2 (2)J. M. G . COWIE AND I . J. MCEWEN 177 curves obtained here are at the vapour pressure of the solvent which may be as high as 10 atm in some cases. However the effect of pressure, which raises LCST and lowers UCST,13 cannot account for the discrepancies between theory and experiment. It may be that some improvement in the absolute predictive power of the theory for the critical temperatures could be obtained if accurate solvent expansion factors at high temperatures were known. No such data are presently available. TABLE 2.-PARAMETERS USED IN THE CALCULATION OF CST system Tf /K C1T2 clY2 x 103 polystyrene+ acetone 7 4349 0.156 17.74 polystyrene+ acetone+ ether, polystyrene+ ether t 4056 0.229 5.55 41 = 0.50 4203 0.190 8.51 5 41 = 0.35 4155 0.202 6.73 5 41 = 0.20 4113 0.21 3 6.06 3 t taken from ref.(2) ; fitted as described in text and in ref. (6). The authors wish to thank S.R.C. for financial support to one of us (I. J. McE.). D. Patterson, Macromolecules, 1969, 2, 672. K. S. Siow, G. Delmas and D. Patterson, Macromolecules, 1972, 5,29. J. M. G. Code and J. T. McCrindle, European Polymer J., 1972,8,1185. (a) I. Prigogine (with the collaboration of V. Mathot and A. Bellman$), The Molecular Theory of Solutions (North-Holland, Amsterdam, 1957) ; (b) P. J. Flow, Disc. F'uduy Soc., 1970,49, 7.D. Patterson, J. Polymer Sci. C, 1969, 16, 3379. J. M. G. Cowie, A. Maconnachie and R. J. Ramon, Macromolecules, 1971, 4, 57. B. A. Wolf, J. W. Breitenbach and H. Senftl, J. Polymer Sci. C, 1970, 31, 345. J. Sameshita, J. Amer. Chem. Soc., 1918, 40, 1482. 6. Delmas and D. Patterson, IUPAC Symposium Macromolecular Chemistry, Toronto, 1968. ' R. Koningsveld, L. A. KIeintjens and A. R. Shultz, J. Polymer Sci. A-2, 1970, 8, 1261. lo H. Tompa, Polymer Solutions (Butterworths, London, 1956). l 2 B. E. Eichinger and P. J. Flory, Trans. Furaday SOC., 1968,68,2035. l3 D. Patterson, Pure Appl. Chem., 1972, 31, 133. Upper and Lower Critical Solution Temperatures in the 3- Polystyrene(3) Cosolvent System Acetone( 1) + Diethyl Ether(2) BY JOHN M. G. COWE* AND IAIN J.MCEWEN Chemistry Department, University of Stirling, Stirling, Scotland Received 1 1 th July, 1973 Phase diagrams for different molecular weight fractions of polystyrene in the mixed solvent system acetone+ diethyl ether have been obtained. The system shows a variation of upper and lower critical solution temperatures with solvent composition which indicates the cosolvent nature of the mixed solvent. No solvent composition is found which can dissolve high molecular weight poly- styrene (M> lo6). The separation of critical solution temperatures can be predicted by the Prigogine theory of polymer solution thermodynamics, but not the absolute values. Phase separation, which occurs when a polymer solution is cooled below the well known upper critical solution temperature (UCST), has been the subject of extensive study; UCST phenomena being widely employed as the basis of polymer fraction- ation.The opposite effect, that of phase separation on heating a polymer solution, has received much less attention. This second separation of a polymer solution into two liquid phases is associated with the large difference in thermal expansions of polymer and solvent and should, in principle, be observed for all non-polar polymer +solvent systems near the critical temperature of the solvent. The minimum in the coexistence curve occurring at these temperatures is termed the lower critical solution temperature (LCST). The polymer + solvent systems polystyrene + acetone and polystyrene + diethyl ether have been studied by Patterson and coworkers.2 Both solvents are compara- tively " poor " in that they dissolve only low molecular weight polystyrene fractions.The effect of this is to cause the UCST and LCST to be separated by a relatively small temperature range. As the molecular weight of the polymer is increased, the range of complete polymer-solvent miscibility decreases, i.e., the UCST is raised and the LCST is lowered. At a certain molecular weight the UCST and LCST coalesce, so that the two regions of immiscibility merge, giving an " hour-glass " shaped phase diagram. Mixtures of relatively poor solvents can, in some cases, produce enhanced solvent power. When this happens the mixed solvent is said to exhibit a synergistic effect. The enhancement of solubility may be detected by a study of the limiting viscosity number or the second virial coefficient measured as a function of solvent composition at a particular ternperat~re.~ Alternatively, a more extensive study of the solubility of a polymer in the mixed solvent may be carried out which illustrates more fully the behaviour as a function of both temperature and solvent composition.In this paper we have examined the phase equilibria for several polystyrene fractions (component 3) in the mixed solvent acetone (component l)+diethyl ether (component 2). The variation of UCST and LCST with solvent composition reflects the cosolvent behav- iour of acetone +ether mixtures and establishes the limits of solution in the system. 171172 CRITICAL SOLUTION TEMPERATURE An attempt to predict both UCST and LCST using the Prigogine-Flory theory of polymer solution^,^ as developed by Patterson and coworker^,^ has been made and the results assessed.EXPERIMENTAL The polystyrene samples used were obtained from the Pressure Chemical Co. who quote M,/M, ratios of less than 1-06. The solvents used were best grade and were fractionally distilled before use. UCST and LCST were estimated from cloud-point curves. These were determined optically in thick-walled Pyrex tubes (i.d. = 5 mm) essentially as described previously.6 For a true monodisperse sample the maxima and minima of the cloud point curves can be taken to represent, respectively, the UCST and LCST for the polymer. It is known that polydispersity displaces the critical point to higher polymer concentration ’ ; however, no indication of displacement, which would result in a point of inflexion in the cloud-point curve, was observed throughout this study.Accordingly we have taken the CST as the turning points of the coexistence curves. RESULTS Typical cloud-point curves for the system are shown in fig. 1. In pure acetone, polystyrene fraction M, = 20 400 gives an hour-glass phase diagram which is almost identical to that obtained by Patterson and coworkers for a 19 800 fraction (fig. lA, curve a). At temperatures and compositions inside the hour-glass a two phase region exists where acetone and polystyrene are immiscible. On addition of ether to solutions of polystyrene in acetone ($1 = 0.95) the phase diagram opens out to form two regions of immiscibility separated by a continuous one phase region; both UCST and LCST can now be identified.The effect of increasing the ether content of the mixed solvent in the range 41 = 0.91 to 41 = 0.67 leads to further separation of the UCST and LCST as shown by curves c and din fig. 1A and 1B. This behaviour is also observed in the single solvent acetone but by using polymer fractions of decreasing molecular weight. At compositions beyond $1 -0.67 there is a tendency for the critical temperatures to begin converging once again. The cloud-point curves for polystyrene M, = 20 400 in pure ether and at 41 = 0.05 are shown in fig. lB, curves e and$ With the next fraction in the series, polystyrene M, = 37 0o0, the UCST and LCST are separated by a single phase region when the solvent composition lies between C#I~ = 0.82 and 0.05.Fig. 1C shows the acetone-rich end of the composition range. At 41 = 0.82 the UCST and LCST have just separated while at 41 = 0.83 the regions of immiscibility have merged to give an hour-glass diagram. The dashed line in this diagram represents an estimate. The two phase nature of this region was established by heating a solution of composition 43 = 0.13 to 320 K ; a homogeneous solution was not obtained. With increasing ether content the cloud-point curves for fraction M,,, = 37 O00 are further separated as can be seen from ourves c and d of fig. 1C whose corresponding UCST have decreased to 250 and 203 K respectively. As fractions of higher molecular weight are used the solvent composition range over which solution can be effected is further decreased.Cloud point curves for polystyrene fraction Mw = 860 000 are shown in fig. 1D. The highest fraction studied (Mw = 2 x lo6) could not be dissolved under any of the possible conditions suggested by fig. 2 in which the CST for all the polystyrene fractions are plotted as a function of solvent composition. (It should be noted that this is not a true phase diagram as the volume fraction of the polymer at the critical temperatures is a functionJ . M. G . COWIE AND I . J . MCEWEN 173 I- l , l 3 I 1 L 43 (C) 0.0 0. I 0.2 0.3 FIG. 1.-A-D. Typical cloud-point curves for the system acetone+ether+polystyrene. +1 =volume fraction acetone in the solvent mixture, 43 = volume fraction polymer in solution. A & B : fraction 20 400; (a) dl = 1.00, (6) 41 = 0.95, (c) +1 = 0.91, (d) +1 = 0.67, (e) 41 = 0.00, (f) +1 = 0.05.C : fraction 37 000; (a) = 0.83, (b) = 0.82, (c) +1 = 0.71, (d) +l = 0.50. D : fraction 860000; (a) (bl = 0.30, (6) 41 == 0.33, (c) $1 = 0.38.174 CRITICAL SOLUTION TEMPERATURE of both M, and solvent composition.) The resulting set of contours represents the variation of CST with solvent composition and defines the regions of solubility of polystyrene in the liquid mixtures. It can be seen that high molecular weights will not dissolve in any composition of the mixed solvent. 0 0.5 I #2 FIG. 2.-Plots of UCST and LCST against 42 for all fractions. d2 = volume fraction ether in solvent mixture. (a) Fraction 20 400, (6) fraction 37 OOO, (c) fraction 110 OOO, ( d ) fraction 267 000, (e) estimated curve for fraction 411 000, (f) fraction 860 000. FIG.3.-UCST and LCST, taken from sections on fig. 2, as a function of r-*. (a) q51 = 0.35, (b) #1 = 0.50, (c) #1 = 0.20. Dashed lines calculated and fitted as described in text. This molecular weight dependence of the polymer solubility in the mixed solvent is illustrated more clearly in fig. 3, where cross-sections at constant solvent composition are plotted as a function of polymer chain length. Table 1 contains the values of UCST and LCST for all the fractions studied, the fraction molecular weights and the solvent compositions.J . M . G. COWIE AND I . J . MCEWEN 175 TABLE 1.-UCST AND LCST FOR POLYSTYRENE FRACTIONS IN THE MIXED SOLVENT ACETONE+ DIETHYL ETHER fraction Mw dl LCST/K UCST/K - - 1 .oo 0.95 367 310 0.91 375 292 0.83 383 261 0.67 386 219 0.20 362 187 0.09 345 197 0.05 333 210 0.00 316 230 20400 0.40 379 189 0.82 326 312 0.71 358 250 37000 0.50 364 203 0.20 341 193 0.05 297 239 fraction Ma 41 LCST/K UCST/K 0.17 295 227 0.30 322 207 llOOO0 0.40 327 213 0.50 328 230 0.60 3 14 266 0.24 290 23 1 0.30 304 222 267000 0.40 3 10 225 0.47 307 239 0.50 301 - 411000 0.31 291 233 0.30 268 248 86OOOO 0.33 275 245 0.38 274 252 DISCUSSION To date, only one publication has dealt with both UCST and LCST in a quasi- ternary system.8 The system studied was polystyrene disrolved in solvent + non- solvent mixtures and points of similarity can be seen.Hour-glass shaped phase diagrams were obtained and the separation of UCST and LCST increased with improving solvent power, The system acetone( 1) + ether(2) +polystyrene(3), although approximating to a solvent + non-solvent system at low molecular weights, is essen- tially one of two poor solvents.The behaviour of the cloud point curves bears some resemblance to that reported by Wolf et aL8 but is due to the cosolvent nature of the mixtures. The cosolvent nature and the separation of UCST and LCST with im- proving solvent power is illustrated in fig. 2. Synergism is indicated by the increase in solubility with any mixture compared with that of the pure components. The " best " solvent mixture, which we may define as the one which will dissolve the highest molecular weight fraction, appears to occur at approximately Analysis of existing vapour pressure data for the binary liquid system acetone+ ether shows that, at mole fraction 0.5 and at 303 K, AGE = 452 Jmol-l, i.e., acetone and ether are moderately incompatible as evidenced by the positive excess free energy of mixing.The cosolvent action which lowers the UCST also leads to a less easily explained raising of the LCST above that expected for a linear interpolation between the two components. The effect can be explained qualitatively if one postulates that the introduction of a polymer chain into the binary solvent environment serves to bind the acetone and ether molecules more strongly by acting as a bridge between these relatively incompatible species. This would have the effect of reducing the expected rate of expansion of the binary liquid pair relative to that of the polymer, thereby raising the LCST.This is in keeping with the nature of cosolvent systems already studied ; qualitatively, the cosolvent effect of acetone +ether mixtures can be ex- plained by a preference for (1-2-3) contacts over (1-2), (1-3) or (2-3) contacts.2 It is interesting to note that high molecular weight polystyrene is soluble, in the appropriate mixture of acetone and ether, only at temperatures well below room = 0.34.176 CRITICAL SOLUTION TEMPERATURE temperature. This somewhat surprising result indicates that use of a mixed solvent for polymer studies should be approached with some caution unless the phase rela- tions for the system have previously been determined. Since attempts to dissolve a polystyrene fraction of MW = 2 x lo6 under the conditions defined by the area inside the contour for the 860 OOO fraction (fig.2) were not successful, it was concluded that no single mixture of acetone and ether will act as a theta solvent for polystyrene. This is confirmed by plotting, in fig. 3, UCST and LCST against r-i for three sections, each at a constant but different solvent composi- tion, from fig. 2. Here r has been taken as the degree of polymerisation and not the ratio of molar volumes. The data form a curve at each solvent composition such that at high values of r the UCST and LCST coalesce. The curves define the solubil- ity, at a particular q51, as a function of chain length. The single phase region lies within the bounds of the curve, outside this a homogeneous solution cannot form. This behaviour has already been reported for the system secondary cellulose acetate + acetone.The simpler, two parameter, theories of polymer solutions l o do not predict LCST nor do they allow curvature in (CST, r-3) plots. The three parameter Prigogine theory,4* which considers changes in free volume of the Components, has been applied to systems showing LCST. As yet no refinement is available which takes account of quasi-ternary systems. We have made the assumption that the theory, as it stands, can be applied to the liquid mixtures if these are treated as single liquids and have derived the required " average " parameters using the ideal mixing rule. Patterson and Delmas have shown," using the Prigogine theory, that at the point of critical miscibility Here 3c, is the number of external degrees of freedom of the solvent molecule, z is a measure of the difference in free volume of the components of the mixture and is obtained from expansion data, while v2 represents the difference in the chemical nature of the components.v17 the reduced volume of the solvent, may also be ob- tained from expansion data. By solving eqn (1) the variation of CST with molecular weight may be predicted. Values of c1 and T~ for both acetone and ether have been calculated by Patterson et aZ.* from the equation of state data of Flory and Eichinger.12 These values, and those of the temperature reduction parameters, TT, are shown in table 2. We define an average parameter for the solvent as where xi is the parameter for the pure solvent and 4i is the volume fraction.Values of c1v2 are obtained by a fitting technique ; the value of v2 is adjusted SO that the coalescence point of UCST and LCST, predicted by (I), is the same as that indicated by the experimental curves in fig. 3. These procedures are fully described elsewhere. The results of the theoretical calculation are shown by the dashed curves in fig. 3. In each case, the theory has successfully predicted the general shape of the experi- mental curves and has almost exactly reproduced the separation of UCST and LCST. The major weakness is that the theory fails to predict the absolute values of the CST correctly. In order to match the theoretical and experimental curves it is necessary to displace the temperature axes. This procedure has been adopted for other systems.2* Eqn (1) applies at zero or negligible pressure.The cloud point j i = 41x1 +42x2 (2)J. M. G . COWIE AND I . J. MCEWEN 177 curves obtained here are at the vapour pressure of the solvent which may be as high as 10 atm in some cases. However the effect of pressure, which raises LCST and lowers UCST,13 cannot account for the discrepancies between theory and experiment. It may be that some improvement in the absolute predictive power of the theory for the critical temperatures could be obtained if accurate solvent expansion factors at high temperatures were known. No such data are presently available. TABLE 2.-PARAMETERS USED IN THE CALCULATION OF CST system Tf /K C1T2 clY2 x 103 polystyrene+ acetone 7 4349 0.156 17.74 polystyrene+ acetone+ ether, polystyrene+ ether t 4056 0.229 5.55 41 = 0.50 4203 0.190 8.51 5 41 = 0.35 4155 0.202 6.73 5 41 = 0.20 4113 0.21 3 6.06 3 t taken from ref. (2) ; fitted as described in text and in ref. (6). The authors wish to thank S.R.C. for financial support to one of us (I. J. McE.). D. Patterson, Macromolecules, 1969, 2, 672. K. S. Siow, G. Delmas and D. Patterson, Macromolecules, 1972, 5,29. J. M. G. Code and J. T. McCrindle, European Polymer J., 1972,8,1185. (a) I. Prigogine (with the collaboration of V. Mathot and A. Bellman$), The Molecular Theory of Solutions (North-Holland, Amsterdam, 1957) ; (b) P. J. Flow, Disc. F'uduy Soc., 1970,49, 7. D. Patterson, J. Polymer Sci. C, 1969, 16, 3379. J. M. G. Cowie, A. Maconnachie and R. J. Ramon, Macromolecules, 1971, 4, 57. B. A. Wolf, J. W. Breitenbach and H. Senftl, J. Polymer Sci. C, 1970, 31, 345. J. Sameshita, J. Amer. Chem. Soc., 1918, 40, 1482. 6. Delmas and D. Patterson, IUPAC Symposium Macromolecular Chemistry, Toronto, 1968. ' R. Koningsveld, L. A. KIeintjens and A. R. Shultz, J. Polymer Sci. A-2, 1970, 8, 1261. lo H. Tompa, Polymer Solutions (Butterworths, London, 1956). l 2 B. E. Eichinger and P. J. Flory, Trans. Furaday SOC., 1968,68,2035. l3 D. Patterson, Pure Appl. Chem., 1972, 31, 133.

ISSN:0300-9599    

DOI:10.1039/F19747000171

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Reaction of Oxygen Atoms with Carbonyl Compounds Part 2.-Acetaldehyde BY G. P. R. MACK AND B. A. THRUSH* University of Cambridge, Department of Physical Chemistry, Lensfield Road, Cambridge. CB2 1EP Received 27th July, 1973 The reaction of excess atomic oxygen with acetaldehyde in a discharge-flow system has been studied at temperatures between 195 and 573 K. Kinetic studies and product analyses establish the mechanism to be O+CHjCHO -+ CHjCO+OH (1) 0fOH-t Oz+H (2) (9) O+CH3 -+ HzCO+H (7) followed by reaction between oxygen atoms and formaldehyde. At 573 K, the reduction of the carbon dioxide yield on increasing the pressure from 1 to 10 Torr is caused by unimolecular decomposi- tion of acetyl radicals. At 300 K, kl = (2.88 k 0.3) x 1 0 I 1 cm3 mol-I s-l. O+ CHSCO + CH3 + COZ CvetanoviE studied the reaction of acetaldehyde with oxygen atoms produced by He deduced that the initial the mercury-sensitised decomposition of nitrous oxide.step is hydrogen abstraction This was challenged by Avramenko and colleagues 2-4 who maintained that insertion to yield CH,COOH and CH20HCH0 and molecular fragmentaion to CHzCO + H,O and HzCO + H2 + CO were the predominant processes. Subsequent studies, notably by Avery and CvetanoviE have confirmed that (1) is the sole initial step. In discharge flow systems, where atomic oxygen is the dominant reactive species, hydroxyl radicals formed in (1) are rapidly removed by (2) where k2 = 3 x 10l3 cm3 mol-1 s-l (ref. (6)). Oxygen atoms also react rapidly with the acetyl radicals produced, but there is disagreement as to the mechanism.Cadle and power^,^ who used gas chromatography, found substantial yields of meth- ane and carbon dioxide and suggested the scheme O+CH3C0 -+ CH,C02 (3) 0 + CH3CH0 + OH + CH3C0 + 63 W mol-l. (1) O+OH + O2+H+71 kJ mol-1 CH3CO2 + CH3 +CO, O+CH, + products (4) CH3 + CH3CH0 + CH4 + CH3C0 + 77 kJ mol-I. (5) Preliminary results of Niki, McKnight and Weinstock who used a time-of-flight mass spectrometer suggest that ketene is a major intermediate being formed in the reaction 0 + CH3C0 -+ CH2C0 + OH + 249 kJ mol-l. (6) 178G . P . R . MACK AND B . A . THgUSH 179 Since the reaction 0 + CH3 + H2C0 + H + 286 kJ mol-1 (7) is known to be rapid (k7 >, 2 x 1013 cm3 mol-l s-l) the following overall stoichiometries for each 0 + CH3CH0 initial step if the mechanism is (1) then there occur (3) (4) (7) and (2) twice, or if (1) is followed by (6) and (2) twice.Distinguishing between these mechanisms is not straightforward because any ketene formed reacts as rapidly with oxygen atoms as does acetaldehyde,1° whereas formaldehyde reacts more slowly. Comparison of stable product analyses and of kinetic studies and measurements of the stoichiometry of hydrogen atom production using e.p.r. and chemiluminescence for the three reactions are needed to elucidate the mechanism. It is thus established that reaction (6) is negligible by comparison with the secondary reactions give 4 0 + CHSCHO + HZCO + C02 + 0 2 + 2H 2 0 + CH3CH0 3 CH2C0 + O2 + 2H (1) (11) (3). EXPERIMENTAL The reaction between oxygen atoms and acetaldehyde was studied in the two discharge A ow systems described in Part 1 ; one was required for product analyses by gas chromato- graphy in which the oxygen atom concentrations were also measured using the air afterglow, and in the other apparatus 0 and H atom concentrations were measured by gas phase e.p.r.As previously, atomic oxygen in a nitrogen carrier was generated by titrating active nitrogen with nitric oxide. Acetaldehyde was dried over calcium sulphate for 24 h after acetic acid had been removed with sodium bicarbonate, it was then fractionally distilled in vacua For accurate measurements of flow rates, acetaldehyde vapour was normally diluted with nitro- gen and the viscosities of these mixtures calculated assuming a Lennard-Jones potential. The L-J parameters for acetaldehyde, c = 0.451 nm and c/k = 355 K were computed from viscosity measurements.Checks showed that the experimental and calculated capillary constants for these mixtures agree within 1 %. In some experiments, the non-condensible products were retained by lining the trap with 30/60 mesh silica gel l3 which had been heated to 340 K to eliminate traces of CO, COZ and CH4. Substantial amounts of the nitrogen carrier were also adsorbed during trapping and desorbed with the carbon monoxide during heating which made the analyses difficult. The permanent gases were separated on a 2 m 13X molecular sieve column in series with a 2m Porapak Q column. The g.1.c. analyses showed that formaldehyde was a product, particularly at intermediate reaction times.A trap was therefore prepared so that the main gas flow impinged directly on a 10ml portion of 0.1 % chromotropic acid in concentrated sulphuric acid. After a known time, 0.5 in1 water was added and the solution stood for 30 min before formaldehyde was determined by comparing the absorbance at 580 nm with standard solutions l4 using a Unicam SP800 spectrophotometer. This method is highly specific and is unaffected by a 100 fold excess of acetaldehyde. All error limits are one standard deviation. RESULTS AND DISCUSSION OXYGEN ATOM REMOVAL The overall stoichiometry was measured by extrapolating the number of oxygen atoms consumed per acetaldehyde molecule to zero acetaldehyde concentration as described in Part 1. These plots were normally linear, but occasionally curved up- wards slightly at low acetaldehyde concentrations.At low acetaldehyde concentra-180 0-ATOM+ CARBONYL REACTIONS 0.2 0.4 0.6 tions, a correction is needed for incomplete reaction. This was made using the rate equation given below (In); linear plots were then consistently obtained, as for example in fig. 1. For medium reaction times of 30-50ms, corresponding to reaction (1) being essentially complete with the experimental value of kl fi 3 x 10" cm3 mol-' s-l, the stoichiometry of oxygen atom removal was 4.3 f 0.3 but this rose to a maximum of about 7 at long reaction times and also at higher temperatures, indicating the occurrence of a slower subsequent reaction. At 195 K this stoichiometry was 4.0kO.l at medium reaction times. J 1.6 10' O[CH3CHOIo /mol ~ m - ~ FIG.1.-Stoichiometries of oxygen atom consumption. Reaction time = 24.1 ms : [010 = 1.9 x 10-lo mol ~ m - ~ . Experimental, 0 : corrected for incomplete reaction, n = 4, 0 ; n = 5, A. HYDROGEN ATOM PRODUCTION The production of hydrogen atoms was investigated using e.p.r. The yields decreased rapidly with increasing reactant flow ; the hydrogen atom concentration passed through a maximum which depended on initial oxygen atom concentration and reaction time but generally corresponded to the acetaldehyde flow for which [O]^ 3 ~ 1 . 6.0 b g 2.0 1 010[CH3CHO]o /mol ~ m - ~ FIG. 2.-Stoichiometry of hydrogen atom production. 0 Short reaction time (4.5 ms) lower reactant scale : A Medium reaction time (53 ms) upper reactant scale. The number of oxygen atoms consumed per hydrogen atom produced was ob- tained by extrapolation at short and long reaction times.As illustrated by fig. 2, the intercepts were reproducible over the full extent of reaction giving ([O], -[O])/[H] = 2.0kO.l.G. P. R. MACK AND B . A . THRUSH 181 Since the corresponding experimental stoichiometries for hydrogen atom pro- duction in the O+H2C0 and O+CH2C0 reactions lo* *l are both 2, this finding supports I as the initial path, a view which is consistent with increased oxygen atom consumption at long reaction times O+H,CO being about three times slower than 0 + CHSCHO. PRODUCT ANALYSES Despite the difficulties of recovering carbon monoxide when the products were trapped on silica gel, reasonable carbon balances were obtained, demonstrating that methane was not produced in significant quantities (table 1) in contrast to the experi- ments of Cadle and Power~.~ As the final carboniferous products are predominantly CO and C 0 2 , the carbon dioxide yields were measured by the simpler technique of direct condensation in the trap.These measurements (table 2) show that 1.30+_0.04 molecules of C02 are produced per acetaldehyde molecule reacted in a large excess of atomic oxygen both at and below room temperature. The C 0 2 yield falls at higher temperatures. TABLE 1 total total total pressure/ reaction 1O10[0]0/ [O]O/ COz per CO per C& per % carbon Tom temp. /K timelms mol crn--' [CH~CHOJO CH3CHO CH3CHO CH3CHO recovered 4.5 300 420 1.3 32.6 1.30 0.62 0.00 96 4.0 300 340 1.5 21.4 1.32 0.84 0.02 109 2.8 300 370 1.3 18.5 1.26 0.64 0.01 95 1.0 473 >loo 1.1 15.0 1.26 0.50 0.01 88 TABLE 2 pressure/Torr temp./K 1.70 195 1.90 300 2.04 300 3.25 300 3.48 300 5.70 300 1.98 473 4.96 473 1.10 573 1 .lo 573 3.56 573 10.9 573 11.8 573 reaction timelms 165 174 221 330 320 150 125 220 205 210 144 200 205 10'0tO10l [Olol mol cm-3 [CH~CHOIO 2.9 30 1.7 25 2.5 24 1.8 11 1.8 21 1.8 9 2.7 22 2.1 12 1.8 29 1.8 38 2.9 28 3.5 27 3.3 34 total pmol CH3CHO/ 30.0 25.4 29.5 102.5 33.5 29.4 47.2 25.9 10.0 4.2 12.8 10.8 10.9 total a 1 CrmOl 38.4 33.3 37.4 119.5 46.0 39.3 58.1 32.3 10.8 4.3 13.0 8.8 8.6 yield of coz per CH3CHO 1.28 1.32 1.26 1.18 1.34 1.34 1.24 1.34 1.08 1.02 1.02 0.81 0.80 As in the 0 + H2C0 system,l it was confirmed that the production of C02 in the reaction OH + CO 3 C02 + H + 105 kJ mol-l (8) and by the heterogeneous reaction of oxygen atoms with CO was completely negli- gible.Short reaction times were avoided because trial experiments established that heterogeneous reaction between atomic oxygen and unreacted acetaldehyde on the silver foil used to quench atomic oxygen yielded more than 1.4 COz molecules per acetaldehyde reacted. When acetaldehyde was present in sufficient excess to give182 0-ATOM+ CARBONYL REACTIONS essentially complete consumption of atomic oxygen, in addition to COY C02, H20 and H2C0 the products included small amounts of ketene and traces of other car- bony1 compounds including biacetyl. Formaldehyde yields were measured under conditions of almost complete reaction of acetaldehyde with excess atomic reaction. In table 3, these yields are compared with those calculated using the measured stoichiometry (4) and rate coefficient kl = 2.88 x 10l1 cm3 mol-1 s-l for the O+CH,CHO reaction, assuming that it proceeds entirely by mechanism I and the rate constant for the O+H2C0 reaction is 9 x 1O1O cm3 mol-l s-l.ll The corresponding ketene yields are calculated on the same basis assuming that I1 is the sole mechanism and the rate constant for O+CH2C0 is 3.43 x loll cm3 mol-' s-l.TABLE 3 yield calc. react ion [0]0/ HCHO per HCHO per pressure/Torr timelms I%%'/ [CH~CHOIO CH3CHO CH3CHO 2.38 1.89 1.90 1.91 2.14 2.19 1.88 1.91 61 76 77 81 104 106 21 9 224 1.7 2.4 2.3 1.9 2.0 1.8 2.9 2.9 17 23 26 21 9 9 30 25 0.23 0.30 0.17 0.26 0.16 0.17 0.01 0.01 0.54 0.33 0.34 0.41 0.40 0.43 0.01 0.01 calc.CHzCO per 0.15 0.03 0.03 0.06 0.06 0.07 0.0 0.0 CH3CHO calc. unreacted CH3CHO 0.07 0.01 0.01 0.02 0.02 0.03 0.0 0.0 In the following paper,lo it is shown that the reaction of atomic oxygen with ketene does not yield formaldehyde and under comparable conditions gives 0.64 molecules of C02 per ketene reacted. It is therefore clear from the results in table 3 that formation of formaldehyde by mechanism I is the major path in the 0 + CH3CH0 reaction. As both acetaldehyde and ketene lo are almost completely destroyed on the silver foil in the presence of atomic oxygen, the most obvious explanation of the irregular shortfall of the measured formaldehyde yields compared with the calculated ones is due to oxidation of formaldehyde on the silver foil by atomic oxygen. We therefore conclude that the mechanism of the 0 + CH3CH0 is predominantly (1) (9) 0 + CHSCHO -+ OH + CH3CO 0 + CH3C0 -+ CH3 + C02 + 433 kJ mol-l O+CH3 + H2CO+H followed by the slower reaction between atomic oxygen and formaldehyde. O+OH -+ 02+H RATE COEFFICIENT The rates of oxygen atom decay were measured by both air afterglow and e.p.r.using short reaction times. The decays were not strictly logarithmic due to significant acetaldehyde consumption and the results were interpreted using the integrated second-order rate equation where nx = [O], - [O] and n = stoichiometry of oxygen atom removal.S . P . R. MACK AND B . A . THRUGH 183 1010[CH3CHO]o/mol ~ m - ~ 24.1ms; [010 = 1.9x10-10molcm-3. 0 , n = 4 ; O , n = 5 ; A , n = 6. FIG. 3.-Kinetics of oxygen atom consumption for various stoichiometries.Reaction time = Plots of the left hand side of eqn (111) against [CH3CHO], gave good straight lines for n = 4, 5 and 6 as shown in fig. 3 ; with n<4, such plots were curved. The data from such plots are given in table 4, where the observed stoichiometry of oxygen atom consumption 0, = ([0], - [O])/[CH,CHO], is compared with values calculated for stoichiometries n = 4 and n = 5. Their agreement with experiment is consistent with the experimentally measured n = 4.3k0.3. Since mechanism (I) corresponds to n = 4, this value was used to calculate kl in table 4, the mean result from 36 experiments being k, = (2.88k0.30) x loll cm3 mol-l s-l at 300 K. TABLE 4 pressure1 Torr 0.69 0.80 0.80 1.05 1.05 1.05 2.1 1 2.1 1 2.15 2.15 2.20 2.28 2.97 4.99 4.99 5.13 10' 'ki I- calc.Oc reaction 101010]0/ cm3 mol-1 s time/ms molcm-3 n = 4 expt. Oc n = 4 n = 5 8.73 6.90 8.70 4.66 6.94 9.78 8.88 12.80 6.40 14.90 7.18 24.60 24.30 7.50 12.66 7.35 1.70 1.70 1.70 5.10 3.70 3.70 2.60 1.60 1 .oo 1 .oo 2.30 1.86 1.80 1.74 1.74 1.28 3.63 3.80 3.33 2.99 2.19 2.37 2.83 2.67 3.74 3.88 2.87 2.98 2.39 2.13 2.1 1 2.04 1.69 1.25 1.75 2.20 2.20 2.80 1.22 1.44 1.10 2.25 1 S O 2.80 2.80 1.10 0.65 - 1.39 1.15 1.39 1.98 2.09 2.59 1.34 1.78 - - 1.51 2.93 2.87 1.25 0.95 - 1.40 1.15 1.40 2.05 2.17 2.76 1.35 1.83 - - 1.53 3.19 3.1 1 1.26 0.94 - This agrees well with Cadle and Power's value of (2.4k0.7) x loll cm3 mol-l s-l at 300 K, and (2.3k0.4) x loll cm3 rnol-i s-l from CventanoviE's ratio of the rate constant to that for 0 + C2H4 combined with recent values 16-18 for the latter rate coefficient.The value kl = 1.1 x 10l1 cm3 mol-1 s-l reported by Niki, McKnight and Weinstock is substantially lower.184 0-ATOM+ CARBONYL REACTIONS CARBON DIOXIDE YIELDS In the mechanism presented so far, a yield of one carbon dioxide molecule per acetaldehyde oxidised is predicted as a result of the O+CH3C0 reaction. The additional CO, observed is accounted for by the reaction of atomic oxygen with the formaldehyde formed. In Part 1 0 + H,CO -+ OH + HCO + 71 kJ mol-1 (10) 0 + HCO + OH + CO + 350 kJ mol-' (1 1) 0 + HCO -+ H + CO, + 455 kJ mol-I (12) the mechanism was shown to be H + HCO + H, + CO + 358 kJ mol-l O+OH -+ H + 0 2 where k , , : k12 : k13 = 0.54 : 0.46 : 4.0.Complete oxidation of formaldehyde in a great excess of atomic oxygen yields 0.46 molecules of C02, but the rapid reaction (13) of the hydrogen atoms formed in the preliminary reaction between oxygen atoms and acetaldehyde at the typical initial ratio of 20 : 1 would reduce the C02 yield to 0.3. Bearing in mind that COz can be formed in the oxidation on the silver foil of any remaining formaldehyde, the average total C02 yield of 1.3 at and below room temperature is well explained. Benson and O'Neal l9 have shown that the acetyl radical undergoes thermal unimolecular decomposition above 500 K. CH3CO + M + CH3CO* + M (14, - 14) CH3CO* + CH3 + CO. (1 5 ) As acetyl radicals react directly with atomic oxygen to yield carbon dioxide whereas methyl gives formaldehyde which gives a lower yield of carbon dioxide, thermal decomposition of the acetyl radical can explain the lower CO, yield at 573 K and its decrease from 1.06 at 1.1 Torr to 0.80 at 11.6 Torr total pressure.In this pressure range, Benson and O'Neal's data find the decomposition of acetyl radicals to be second order with k,, = 3 x lOI4 exp( - 50 kJ mol-l/RT) cm3 mol-, s-l. If it is assumed that N2 has a collisional activation efficiency of 0.3 in this reaction and that the reaction 0 + CHjCO + CH3 + CO2 has the same rate constant as that for oxygen atoms with the isoelectronic species NO, where k16 = 6 x 1OI2 cm3 mol-I s-I at 573 K,20 then the calculated C02 yield at 573 K from reaction (9) falls from 0.88 at 1.1 Torr to 0.58 at 11.6 Torr. When 0.3 is added to these numbers to allow for C o t production from formaldehyde the agree- ment with the experimental measurements is remarkably good considering the assump- tion made as to the rate constant of reaction (9).(9) O+N02 3 NO+O, (16) REACTIONS OF H, OH, ETC. The reaction H+ CHSCHO -+ H2 + CHjCO was studied by McKnight, Niki and Weinstock 21 who obtained a rate constant of 2 x 10" cm3 mol-' s-, for the fully deuterated system at 300 K ; this is much tooG . P. R. MACK AND B . A . THRUSH 185 They estimate that the subse- slow to compete with reaction (1) for acetaldehyde. quent reaction H + CH3CO -+ H2 + CH2CO is at least one hundred times faster, this reaction may compete with (9) at high ratios of acetaldehyde to atomic oxygen, thus accounting for the detection of ketene under such conditions.(18) Under such conditions, the rapid reaction OH + CH3CHO -+ H2O + CH3CO (19) must also be considered since k19 = 9 x 10l2 cm3 mol-' s-l at 300 K 22 ; this is almost one third the rate constant of reaction (2) and it is necessary to work with [0]0< 10[CH3CHOIo if the hydroxyl formed in reaction (1) is not to react signifi- cantly with acetaldehyde. Methyl radicals formed in reactions (9) and (15) will be removed predominantly by atomic oxygen. The reaction CH3 + CH3CHO + CH4 + CH3CO being negligible since k,> lo6 k5 at 300 K.99 23 The combination of methyl radicals with hydrogen atoms to form methane is heterogeneous in flow pressures 2 4 9 25 and will also not compete with (7) when [O], % [CH,CHO]. (5) CONCLUSIONS The experiments described here clearly establish that the reaction of excess atomic oxygen with acetaldehyde in a discharge-flow system yields formaldehyde through the intermediacy of the acetyl and methyl radicals (mechanism I above).The subsequent oxidation of formaldehyde occurs by the mechanism given in Part 1.'' The fall in carbon dioxide yield with increasing total pressure at the highest temperature is consistent with previous data on the thermal unimolecular decomposi- tion of the acetyl radical l9 which is its main precursor. The oxygen atom is normally regarded as an electrophilic reagent,26 which is consistent with its attack on the acetyl radical at the carbon atom where the unpaired electron is largely localised. As it might be argued that the path to CH3 +CO, is preferred energetically, being 433 kJ mol-' exothermic compared with 249 kJ mol-l for hydrogen abstraction to yield OH+CH,CO, it is interesting to note that the reaction of nitrogen atoms with the isoelectronic molecule NO2 yields predominantly 0 +N20 rather than NO + NO which is 151 kJ mol-' more ex other mi^.^' In contrast, oxygen atoms abstract hydrogen from acetaldehyde (which is 63 kJ mol-' exothermic) rather than adding at the double bond and then rupturing to CH3 + C02 + H which is 113 kJ mol-' exothermic or to CH, + C02 which would be 551 kJ mol-1 exothermic but awkward sterically and spin forbidden overall. As activation energies are observed both in the spin allowed addition of hydrogen atoms to olefins and in the overall spin-forbidden addition of O(3P) to olefins, the absence of a reaction initiated by oxygen atom addition to acetaldehyde could arise from steric factors or from a spin forbidden step in the formation of the initial adduct.We thank the Ministry of Defence for an E.M. Research Contract and Dr. L. Phillips of E.R.D.E. for useful discussions. R. J. CvetanoviE, Canad. J. Chem., 1956, 34, 775. L. I. Avramenko and R. V. Kolesnikova, Izziest. Akad. Nauk S.S.S.R., 1961, 1231. L. I. Amamenko, R. V. Kolesnikova and M. F. Sarokina, Izvest. Akad. Nauk S.S.S. R., 1961, 1005.186 0-ATOM+ CARBONYL REACTIONS L. I. Avramenko and R. V. Lorentso, Zhur. Fiz. Khim., 1952, 26, 1084. H. E. Avery and R. J. CvetanoviE, J. Chem. Phys., 1965, 43, 3727. M. A. A. Clyne and B. A. Thrush, Proc. Roy. SOC. A, 1963, 275,544.H. Niki, C. McKnight and B. Weinstock, Abs. 154th Annual Meeting Amer. Chem. Soc., 1967, v110. H. Niki, E. E. Daby and B. Weinstock, J. Chem. Phys., 1968,48, 5729. lo G. P. R. Mack and B. A. Thrush, J.C.S. Furuduy I, 1974, 70, 187. G. P. R. Mack and B. A. Thrush, J.C.S. Furuduy I, 1973, 69,208. l 2 P. M. Craven and J. D. Lambert, Proc. Roy. SOC. A, 1951,205,442. l 3 N. Demchuk and H. D. Gesser, Cunud. J. Chem., 1963,41, 1645. l4 A. P. Altshuller, L. J. Lang and A. F. Wartburg, Int. J. Air Water Pollution, 1962, 6, 381. l5 A. P. Altshuller, D. L. Miller and S. F. Sleva, Anal. Chem., 1961, 33, 621. l6 J. M. Brown and B. A. Thrush, Trum. Furuduy SOC., 1967, 63, 630. l7 H. Niki, E. E. Daby and B. Weinstock, 12th Symp. Combustion Inst., 1969, p. 277. R. Atkinson and R.J. CvetanoviE, J. Chem. Phys., 1971, 55, 659. l9 H. E. O’Neal and S. W. Benson, J. Chem. Phys., 1962, 36,2196. 2o A. A. Westenberg and N. de Haas, J. Chem. Phys., 1969,50,707. 21 C. McKnight, H. Niki and B. Weinstock, J. Chem. Phys., 1967, 47, 5219. 22 E. D. Morris, D. H. Stedman and H. Niki, J.?Amer. Chem. Soc., 1971, 93, 5370. 23 J. A. Kerr and J. G. Calvert, J. Phys. Chem., 1965, 69, 1022. 24 R. W. Carr, I. D. Gay, G. P. Glass and H. Niki, J. Chem. Phys., 1968, 49, 846. 2 5 P. B. Davies, B. A. Thrush and A. F. Tuck, Trans. Faruduy Soc., 1970, 66, 686. 26 R. J. CvetanoviE, Adu. Photochent., 1963, 1, 115. 27 M. A. A. Clyne and B. A. Thrush, Trans. Fuvaduy SOC., 1961, 57, 69 ; L. F. Phillips and H. I. ’ R. D. Cadle and J. W. Powers, J. Phys. Chem., 1967,71, 1702.Schiff, J. Chem. Phys., 1965, 42, 3173. Reaction of Oxygen Atoms with Carbonyl Compounds Part 2.-Acetaldehyde BY G. P. R. MACK AND B. A. THRUSH* University of Cambridge, Department of Physical Chemistry, Lensfield Road, Cambridge. CB2 1EP Received 27th July, 1973 The reaction of excess atomic oxygen with acetaldehyde in a discharge-flow system has been studied at temperatures between 195 and 573 K. Kinetic studies and product analyses establish the mechanism to be O+CHjCHO -+ CHjCO+OH (1) 0fOH-t Oz+H (2) (9) O+CH3 -+ HzCO+H (7) followed by reaction between oxygen atoms and formaldehyde. At 573 K, the reduction of the carbon dioxide yield on increasing the pressure from 1 to 10 Torr is caused by unimolecular decomposi- tion of acetyl radicals. At 300 K, kl = (2.88 k 0.3) x 1 0 I 1 cm3 mol-I s-l.O+ CHSCO + CH3 + COZ CvetanoviE studied the reaction of acetaldehyde with oxygen atoms produced by He deduced that the initial the mercury-sensitised decomposition of nitrous oxide. step is hydrogen abstraction This was challenged by Avramenko and colleagues 2-4 who maintained that insertion to yield CH,COOH and CH20HCH0 and molecular fragmentaion to CHzCO + H,O and HzCO + H2 + CO were the predominant processes. Subsequent studies, notably by Avery and CvetanoviE have confirmed that (1) is the sole initial step. In discharge flow systems, where atomic oxygen is the dominant reactive species, hydroxyl radicals formed in (1) are rapidly removed by (2) where k2 = 3 x 10l3 cm3 mol-1 s-l (ref. (6)). Oxygen atoms also react rapidly with the acetyl radicals produced, but there is disagreement as to the mechanism.Cadle and power^,^ who used gas chromatography, found substantial yields of meth- ane and carbon dioxide and suggested the scheme O+CH3C0 -+ CH,C02 (3) 0 + CH3CH0 + OH + CH3C0 + 63 W mol-l. (1) O+OH + O2+H+71 kJ mol-1 CH3CO2 + CH3 +CO, O+CH, + products (4) CH3 + CH3CH0 + CH4 + CH3C0 + 77 kJ mol-I. (5) Preliminary results of Niki, McKnight and Weinstock who used a time-of-flight mass spectrometer suggest that ketene is a major intermediate being formed in the reaction 0 + CH3C0 -+ CH2C0 + OH + 249 kJ mol-l. (6) 178G . P . R . MACK AND B . A . THgUSH 179 Since the reaction 0 + CH3 + H2C0 + H + 286 kJ mol-1 (7) is known to be rapid (k7 >, 2 x 1013 cm3 mol-l s-l) the following overall stoichiometries for each 0 + CH3CH0 initial step if the mechanism is (1) then there occur (3) (4) (7) and (2) twice, or if (1) is followed by (6) and (2) twice.Distinguishing between these mechanisms is not straightforward because any ketene formed reacts as rapidly with oxygen atoms as does acetaldehyde,1° whereas formaldehyde reacts more slowly. Comparison of stable product analyses and of kinetic studies and measurements of the stoichiometry of hydrogen atom production using e.p.r. and chemiluminescence for the three reactions are needed to elucidate the mechanism. It is thus established that reaction (6) is negligible by comparison with the secondary reactions give 4 0 + CHSCHO + HZCO + C02 + 0 2 + 2H 2 0 + CH3CH0 3 CH2C0 + O2 + 2H (1) (11) (3).EXPERIMENTAL The reaction between oxygen atoms and acetaldehyde was studied in the two discharge A ow systems described in Part 1 ; one was required for product analyses by gas chromato- graphy in which the oxygen atom concentrations were also measured using the air afterglow, and in the other apparatus 0 and H atom concentrations were measured by gas phase e.p.r. As previously, atomic oxygen in a nitrogen carrier was generated by titrating active nitrogen with nitric oxide. Acetaldehyde was dried over calcium sulphate for 24 h after acetic acid had been removed with sodium bicarbonate, it was then fractionally distilled in vacua For accurate measurements of flow rates, acetaldehyde vapour was normally diluted with nitro- gen and the viscosities of these mixtures calculated assuming a Lennard-Jones potential.The L-J parameters for acetaldehyde, c = 0.451 nm and c/k = 355 K were computed from viscosity measurements. Checks showed that the experimental and calculated capillary constants for these mixtures agree within 1 %. In some experiments, the non-condensible products were retained by lining the trap with 30/60 mesh silica gel l3 which had been heated to 340 K to eliminate traces of CO, COZ and CH4. Substantial amounts of the nitrogen carrier were also adsorbed during trapping and desorbed with the carbon monoxide during heating which made the analyses difficult. The permanent gases were separated on a 2 m 13X molecular sieve column in series with a 2m Porapak Q column. The g.1.c. analyses showed that formaldehyde was a product, particularly at intermediate reaction times.A trap was therefore prepared so that the main gas flow impinged directly on a 10ml portion of 0.1 % chromotropic acid in concentrated sulphuric acid. After a known time, 0.5 in1 water was added and the solution stood for 30 min before formaldehyde was determined by comparing the absorbance at 580 nm with standard solutions l4 using a Unicam SP800 spectrophotometer. This method is highly specific and is unaffected by a 100 fold excess of acetaldehyde. All error limits are one standard deviation. RESULTS AND DISCUSSION OXYGEN ATOM REMOVAL The overall stoichiometry was measured by extrapolating the number of oxygen atoms consumed per acetaldehyde molecule to zero acetaldehyde concentration as described in Part 1.These plots were normally linear, but occasionally curved up- wards slightly at low acetaldehyde concentrations. At low acetaldehyde concentra-180 0-ATOM+ CARBONYL REACTIONS 0.2 0.4 0.6 tions, a correction is needed for incomplete reaction. This was made using the rate equation given below (In); linear plots were then consistently obtained, as for example in fig. 1. For medium reaction times of 30-50ms, corresponding to reaction (1) being essentially complete with the experimental value of kl fi 3 x 10" cm3 mol-' s-l, the stoichiometry of oxygen atom removal was 4.3 f 0.3 but this rose to a maximum of about 7 at long reaction times and also at higher temperatures, indicating the occurrence of a slower subsequent reaction. At 195 K this stoichiometry was 4.0kO.l at medium reaction times.J 1.6 10' O[CH3CHOIo /mol ~ m - ~ FIG. 1.-Stoichiometries of oxygen atom consumption. Reaction time = 24.1 ms : [010 = 1.9 x 10-lo mol ~ m - ~ . Experimental, 0 : corrected for incomplete reaction, n = 4, 0 ; n = 5, A. HYDROGEN ATOM PRODUCTION The production of hydrogen atoms was investigated using e.p.r. The yields decreased rapidly with increasing reactant flow ; the hydrogen atom concentration passed through a maximum which depended on initial oxygen atom concentration and reaction time but generally corresponded to the acetaldehyde flow for which [O]^ 3 ~ 1 . 6.0 b g 2.0 1 010[CH3CHO]o /mol ~ m - ~ FIG. 2.-Stoichiometry of hydrogen atom production. 0 Short reaction time (4.5 ms) lower reactant scale : A Medium reaction time (53 ms) upper reactant scale.The number of oxygen atoms consumed per hydrogen atom produced was ob- tained by extrapolation at short and long reaction times. As illustrated by fig. 2, the intercepts were reproducible over the full extent of reaction giving ([O], -[O])/[H] = 2.0kO.l.G. P. R. MACK AND B . A . THRUSH 181 Since the corresponding experimental stoichiometries for hydrogen atom pro- duction in the O+H2C0 and O+CH2C0 reactions lo* *l are both 2, this finding supports I as the initial path, a view which is consistent with increased oxygen atom consumption at long reaction times O+H,CO being about three times slower than 0 + CHSCHO. PRODUCT ANALYSES Despite the difficulties of recovering carbon monoxide when the products were trapped on silica gel, reasonable carbon balances were obtained, demonstrating that methane was not produced in significant quantities (table 1) in contrast to the experi- ments of Cadle and Power~.~ As the final carboniferous products are predominantly CO and C 0 2 , the carbon dioxide yields were measured by the simpler technique of direct condensation in the trap.These measurements (table 2) show that 1.30+_0.04 molecules of C02 are produced per acetaldehyde molecule reacted in a large excess of atomic oxygen both at and below room temperature. The C 0 2 yield falls at higher temperatures. TABLE 1 total total total pressure/ reaction 1O10[0]0/ [O]O/ COz per CO per C& per % carbon Tom temp. /K timelms mol crn--' [CH~CHOJO CH3CHO CH3CHO CH3CHO recovered 4.5 300 420 1.3 32.6 1.30 0.62 0.00 96 4.0 300 340 1.5 21.4 1.32 0.84 0.02 109 2.8 300 370 1.3 18.5 1.26 0.64 0.01 95 1.0 473 >loo 1.1 15.0 1.26 0.50 0.01 88 TABLE 2 pressure/Torr temp./K 1.70 195 1.90 300 2.04 300 3.25 300 3.48 300 5.70 300 1.98 473 4.96 473 1.10 573 1 .lo 573 3.56 573 10.9 573 11.8 573 reaction timelms 165 174 221 330 320 150 125 220 205 210 144 200 205 10'0tO10l [Olol mol cm-3 [CH~CHOIO 2.9 30 1.7 25 2.5 24 1.8 11 1.8 21 1.8 9 2.7 22 2.1 12 1.8 29 1.8 38 2.9 28 3.5 27 3.3 34 total pmol CH3CHO/ 30.0 25.4 29.5 102.5 33.5 29.4 47.2 25.9 10.0 4.2 12.8 10.8 10.9 total a 1 CrmOl 38.4 33.3 37.4 119.5 46.0 39.3 58.1 32.3 10.8 4.3 13.0 8.8 8.6 yield of coz per CH3CHO 1.28 1.32 1.26 1.18 1.34 1.34 1.24 1.34 1.08 1.02 1.02 0.81 0.80 As in the 0 + H2C0 system,l it was confirmed that the production of C02 in the reaction OH + CO 3 C02 + H + 105 kJ mol-l (8) and by the heterogeneous reaction of oxygen atoms with CO was completely negli- gible.Short reaction times were avoided because trial experiments established that heterogeneous reaction between atomic oxygen and unreacted acetaldehyde on the silver foil used to quench atomic oxygen yielded more than 1.4 COz molecules per acetaldehyde reacted. When acetaldehyde was present in sufficient excess to give182 0-ATOM+ CARBONYL REACTIONS essentially complete consumption of atomic oxygen, in addition to COY C02, H20 and H2C0 the products included small amounts of ketene and traces of other car- bony1 compounds including biacetyl. Formaldehyde yields were measured under conditions of almost complete reaction of acetaldehyde with excess atomic reaction.In table 3, these yields are compared with those calculated using the measured stoichiometry (4) and rate coefficient kl = 2.88 x 10l1 cm3 mol-1 s-l for the O+CH,CHO reaction, assuming that it proceeds entirely by mechanism I and the rate constant for the O+H2C0 reaction is 9 x 1O1O cm3 mol-l s-l.ll The corresponding ketene yields are calculated on the same basis assuming that I1 is the sole mechanism and the rate constant for O+CH2C0 is 3.43 x loll cm3 mol-' s-l. TABLE 3 yield calc. react ion [0]0/ HCHO per HCHO per pressure/Torr timelms I%%'/ [CH~CHOIO CH3CHO CH3CHO 2.38 1.89 1.90 1.91 2.14 2.19 1.88 1.91 61 76 77 81 104 106 21 9 224 1.7 2.4 2.3 1.9 2.0 1.8 2.9 2.9 17 23 26 21 9 9 30 25 0.23 0.30 0.17 0.26 0.16 0.17 0.01 0.01 0.54 0.33 0.34 0.41 0.40 0.43 0.01 0.01 calc.CHzCO per 0.15 0.03 0.03 0.06 0.06 0.07 0.0 0.0 CH3CHO calc. unreacted CH3CHO 0.07 0.01 0.01 0.02 0.02 0.03 0.0 0.0 In the following paper,lo it is shown that the reaction of atomic oxygen with ketene does not yield formaldehyde and under comparable conditions gives 0.64 molecules of C02 per ketene reacted. It is therefore clear from the results in table 3 that formation of formaldehyde by mechanism I is the major path in the 0 + CH3CH0 reaction. As both acetaldehyde and ketene lo are almost completely destroyed on the silver foil in the presence of atomic oxygen, the most obvious explanation of the irregular shortfall of the measured formaldehyde yields compared with the calculated ones is due to oxidation of formaldehyde on the silver foil by atomic oxygen.We therefore conclude that the mechanism of the 0 + CH3CH0 is predominantly (1) (9) 0 + CHSCHO -+ OH + CH3CO 0 + CH3C0 -+ CH3 + C02 + 433 kJ mol-l O+CH3 + H2CO+H followed by the slower reaction between atomic oxygen and formaldehyde. O+OH -+ 02+H RATE COEFFICIENT The rates of oxygen atom decay were measured by both air afterglow and e.p.r. using short reaction times. The decays were not strictly logarithmic due to significant acetaldehyde consumption and the results were interpreted using the integrated second-order rate equation where nx = [O], - [O] and n = stoichiometry of oxygen atom removal.S . P . R. MACK AND B . A . THRUGH 183 1010[CH3CHO]o/mol ~ m - ~ 24.1ms; [010 = 1.9x10-10molcm-3.0 , n = 4 ; O , n = 5 ; A , n = 6. FIG. 3.-Kinetics of oxygen atom consumption for various stoichiometries. Reaction time = Plots of the left hand side of eqn (111) against [CH3CHO], gave good straight lines for n = 4, 5 and 6 as shown in fig. 3 ; with n<4, such plots were curved. The data from such plots are given in table 4, where the observed stoichiometry of oxygen atom consumption 0, = ([0], - [O])/[CH,CHO], is compared with values calculated for stoichiometries n = 4 and n = 5. Their agreement with experiment is consistent with the experimentally measured n = 4.3k0.3. Since mechanism (I) corresponds to n = 4, this value was used to calculate kl in table 4, the mean result from 36 experiments being k, = (2.88k0.30) x loll cm3 mol-l s-l at 300 K.TABLE 4 pressure1 Torr 0.69 0.80 0.80 1.05 1.05 1.05 2.1 1 2.1 1 2.15 2.15 2.20 2.28 2.97 4.99 4.99 5.13 10' 'ki I- calc. Oc reaction 101010]0/ cm3 mol-1 s time/ms molcm-3 n = 4 expt. Oc n = 4 n = 5 8.73 6.90 8.70 4.66 6.94 9.78 8.88 12.80 6.40 14.90 7.18 24.60 24.30 7.50 12.66 7.35 1.70 1.70 1.70 5.10 3.70 3.70 2.60 1.60 1 .oo 1 .oo 2.30 1.86 1.80 1.74 1.74 1.28 3.63 3.80 3.33 2.99 2.19 2.37 2.83 2.67 3.74 3.88 2.87 2.98 2.39 2.13 2.1 1 2.04 1.69 1.25 1.75 2.20 2.20 2.80 1.22 1.44 1.10 2.25 1 S O 2.80 2.80 1.10 0.65 - 1.39 1.15 1.39 1.98 2.09 2.59 1.34 1.78 - - 1.51 2.93 2.87 1.25 0.95 - 1.40 1.15 1.40 2.05 2.17 2.76 1.35 1.83 - - 1.53 3.19 3.1 1 1.26 0.94 - This agrees well with Cadle and Power's value of (2.4k0.7) x loll cm3 mol-l s-l at 300 K, and (2.3k0.4) x loll cm3 rnol-i s-l from CventanoviE's ratio of the rate constant to that for 0 + C2H4 combined with recent values 16-18 for the latter rate coefficient. The value kl = 1.1 x 10l1 cm3 mol-1 s-l reported by Niki, McKnight and Weinstock is substantially lower.184 0-ATOM+ CARBONYL REACTIONS CARBON DIOXIDE YIELDS In the mechanism presented so far, a yield of one carbon dioxide molecule per acetaldehyde oxidised is predicted as a result of the O+CH3C0 reaction.The additional CO, observed is accounted for by the reaction of atomic oxygen with the formaldehyde formed. In Part 1 0 + H,CO -+ OH + HCO + 71 kJ mol-1 (10) 0 + HCO + OH + CO + 350 kJ mol-' (1 1) 0 + HCO -+ H + CO, + 455 kJ mol-I (12) the mechanism was shown to be H + HCO + H, + CO + 358 kJ mol-l O+OH -+ H + 0 2 where k , , : k12 : k13 = 0.54 : 0.46 : 4.0.Complete oxidation of formaldehyde in a great excess of atomic oxygen yields 0.46 molecules of C02, but the rapid reaction (13) of the hydrogen atoms formed in the preliminary reaction between oxygen atoms and acetaldehyde at the typical initial ratio of 20 : 1 would reduce the C02 yield to 0.3. Bearing in mind that COz can be formed in the oxidation on the silver foil of any remaining formaldehyde, the average total C02 yield of 1.3 at and below room temperature is well explained. Benson and O'Neal l9 have shown that the acetyl radical undergoes thermal unimolecular decomposition above 500 K. CH3CO + M + CH3CO* + M (14, - 14) CH3CO* + CH3 + CO. (1 5 ) As acetyl radicals react directly with atomic oxygen to yield carbon dioxide whereas methyl gives formaldehyde which gives a lower yield of carbon dioxide, thermal decomposition of the acetyl radical can explain the lower CO, yield at 573 K and its decrease from 1.06 at 1.1 Torr to 0.80 at 11.6 Torr total pressure.In this pressure range, Benson and O'Neal's data find the decomposition of acetyl radicals to be second order with k,, = 3 x lOI4 exp( - 50 kJ mol-l/RT) cm3 mol-, s-l. If it is assumed that N2 has a collisional activation efficiency of 0.3 in this reaction and that the reaction 0 + CHjCO + CH3 + CO2 has the same rate constant as that for oxygen atoms with the isoelectronic species NO, where k16 = 6 x 1OI2 cm3 mol-I s-I at 573 K,20 then the calculated C02 yield at 573 K from reaction (9) falls from 0.88 at 1.1 Torr to 0.58 at 11.6 Torr.When 0.3 is added to these numbers to allow for C o t production from formaldehyde the agree- ment with the experimental measurements is remarkably good considering the assump- tion made as to the rate constant of reaction (9). (9) O+N02 3 NO+O, (16) REACTIONS OF H, OH, ETC. The reaction H+ CHSCHO -+ H2 + CHjCO was studied by McKnight, Niki and Weinstock 21 who obtained a rate constant of 2 x 10" cm3 mol-' s-, for the fully deuterated system at 300 K ; this is much tooG . P. R. MACK AND B . A . THRUSH 185 They estimate that the subse- slow to compete with reaction (1) for acetaldehyde. quent reaction H + CH3CO -+ H2 + CH2CO is at least one hundred times faster, this reaction may compete with (9) at high ratios of acetaldehyde to atomic oxygen, thus accounting for the detection of ketene under such conditions.(18) Under such conditions, the rapid reaction OH + CH3CHO -+ H2O + CH3CO (19) must also be considered since k19 = 9 x 10l2 cm3 mol-' s-l at 300 K 22 ; this is almost one third the rate constant of reaction (2) and it is necessary to work with [0]0< 10[CH3CHOIo if the hydroxyl formed in reaction (1) is not to react signifi- cantly with acetaldehyde. Methyl radicals formed in reactions (9) and (15) will be removed predominantly by atomic oxygen. The reaction CH3 + CH3CHO + CH4 + CH3CO being negligible since k,> lo6 k5 at 300 K.99 23 The combination of methyl radicals with hydrogen atoms to form methane is heterogeneous in flow pressures 2 4 9 25 and will also not compete with (7) when [O], % [CH,CHO]. (5) CONCLUSIONS The experiments described here clearly establish that the reaction of excess atomic oxygen with acetaldehyde in a discharge-flow system yields formaldehyde through the intermediacy of the acetyl and methyl radicals (mechanism I above).The subsequent oxidation of formaldehyde occurs by the mechanism given in Part 1.'' The fall in carbon dioxide yield with increasing total pressure at the highest temperature is consistent with previous data on the thermal unimolecular decomposi- tion of the acetyl radical l9 which is its main precursor. The oxygen atom is normally regarded as an electrophilic reagent,26 which is consistent with its attack on the acetyl radical at the carbon atom where the unpaired electron is largely localised.As it might be argued that the path to CH3 +CO, is preferred energetically, being 433 kJ mol-' exothermic compared with 249 kJ mol-l for hydrogen abstraction to yield OH+CH,CO, it is interesting to note that the reaction of nitrogen atoms with the isoelectronic molecule NO2 yields predominantly 0 +N20 rather than NO + NO which is 151 kJ mol-' more ex other mi^.^' In contrast, oxygen atoms abstract hydrogen from acetaldehyde (which is 63 kJ mol-' exothermic) rather than adding at the double bond and then rupturing to CH3 + C02 + H which is 113 kJ mol-' exothermic or to CH, + C02 which would be 551 kJ mol-1 exothermic but awkward sterically and spin forbidden overall. As activation energies are observed both in the spin allowed addition of hydrogen atoms to olefins and in the overall spin-forbidden addition of O(3P) to olefins, the absence of a reaction initiated by oxygen atom addition to acetaldehyde could arise from steric factors or from a spin forbidden step in the formation of the initial adduct.We thank the Ministry of Defence for an E.M. Research Contract and Dr. L. Phillips of E.R.D.E. for useful discussions. R. J. CvetanoviE, Canad. J. Chem., 1956, 34, 775. L. I. Avramenko and R. V. Kolesnikova, Izziest. Akad. Nauk S.S.S.R., 1961, 1231. L. I. Amamenko, R. V. Kolesnikova and M. F. Sarokina, Izvest. Akad. Nauk S.S.S. R., 1961, 1005.186 0-ATOM+ CARBONYL REACTIONS L. I. Avramenko and R. V. Lorentso, Zhur. Fiz. Khim., 1952, 26, 1084. H. E. Avery and R. J. CvetanoviE, J. Chem. Phys., 1965, 43, 3727. M. A. A. Clyne and B. A. Thrush, Proc. Roy. SOC. A, 1963, 275,544. H. Niki, C. McKnight and B. Weinstock, Abs. 154th Annual Meeting Amer. Chem. Soc., 1967, v110. H. Niki, E. E. Daby and B. Weinstock, J. Chem. Phys., 1968,48, 5729. lo G. P. R. Mack and B. A. Thrush, J.C.S. Furuduy I, 1974, 70, 187. G. P. R. Mack and B. A. Thrush, J.C.S. Furuduy I, 1973, 69,208. l 2 P. M. Craven and J. D. Lambert, Proc. Roy. SOC. A, 1951,205,442. l 3 N. Demchuk and H. D. Gesser, Cunud. J. Chem., 1963,41, 1645. l4 A. P. Altshuller, L. J. Lang and A. F. Wartburg, Int. J. Air Water Pollution, 1962, 6, 381. l5 A. P. Altshuller, D. L. Miller and S. F. Sleva, Anal. Chem., 1961, 33, 621. l6 J. M. Brown and B. A. Thrush, Trum. Furuduy SOC., 1967, 63, 630. l7 H. Niki, E. E. Daby and B. Weinstock, 12th Symp. Combustion Inst., 1969, p. 277. R. Atkinson and R. J. CvetanoviE, J. Chem. Phys., 1971, 55, 659. l9 H. E. O’Neal and S. W. Benson, J. Chem. Phys., 1962, 36,2196. 2o A. A. Westenberg and N. de Haas, J. Chem. Phys., 1969,50,707. 21 C. McKnight, H. Niki and B. Weinstock, J. Chem. Phys., 1967, 47, 5219. 22 E. D. Morris, D. H. Stedman and H. Niki, J.?Amer. Chem. Soc., 1971, 93, 5370. 23 J. A. Kerr and J. G. Calvert, J. Phys. Chem., 1965, 69, 1022. 24 R. W. Carr, I. D. Gay, G. P. Glass and H. Niki, J. Chem. Phys., 1968, 49, 846. 2 5 P. B. Davies, B. A. Thrush and A. F. Tuck, Trans. Faruduy Soc., 1970, 66, 686. 26 R. J. CvetanoviE, Adu. Photochent., 1963, 1, 115. 27 M. A. A. Clyne and B. A. Thrush, Trans. Fuvaduy SOC., 1961, 57, 69 ; L. F. Phillips and H. I. ’ R. D. Cadle and J. W. Powers, J. Phys. Chem., 1967,71, 1702. Schiff, J. Chem. Phys., 1965, 42, 3173.

ISSN:0300-9599    

DOI:10.1039/F19747000178

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Reaction of Oxygen Atoms with Carbonyl Compounds Part 3.-Ketene BY G. P. R. MACK AND B. A. THRUSH* University of Cambridge, Department of Physical Chemistry, Lensfield Road, Cambridge CB2 1EP Received 27th July, 1973 The reaction between excess atomic oxygen and ketene has been studied in a discharge flow system by product analysis and kinetically using e.p.r. and chemiluminescznce. This rate constant for the initial step, which yields predominantly HCO+ HCO was found to be (3.4 k 0.3) x 10' ' cm3 mol-' s-' at 293 K. The subsequent reactions are O+HCO -+ OH+CO O+HCO -+ H+COz H+HCO-t H2+CO OfOH -+ 02+H. Can, Gay, Glass and Niki used a mass-spectrometer to study the reaction of ketene with atomic oxygen in a discharge-flow system. They deduced that the initial step was addition of an oxygen atom to ketene, predominantly at the olefinic carbon atom ; the observed products (CO, C02, H20, H2C0, H2, H) could be explained by rapid decomposition of the adduct followed by reaction of its fragments with atomic oxygen and with ketene.We have investigated the 0 + CH2C0 reaction in the apparatus previously used for the O+CH3CH0 and O+H2C0 reactions 2* ; and show that the initial step yields predominantly HCO + HCO formed by the isomerisation of the initial adduct to a vibrationally excited glyoxal molecule which then decomposes. Our results do not support the view that ketene is an important intermediate in the reaction of excess atomic oxygen with acetaldehyde. EXPERIMENTAL The reaction was studied in the discharge flow systems previously de~cribed.~ Atomic concentrations were measured by e.p.r.spectroscopy or the air afterglow and the reaction products were analysed by gas chromatography. All experiments were carried out at 293 K. Ketene was prepared by the thermal decomposition of acetic anhydride at 820K. Acetone and unreacted anhydride were removed at 195 K in a trap packed with glass wool. The ketene was purified by bulb-to-bulb distillation and a residual impurity of ethane was estimated by gas chromatography to be 5-8 %. All errors quoted are one standard deviation. RESULTS AND DISCUSSION STOICHIOMETRY OF THE REACTION The number of oxygen atoms consumed per ketene molecule added and per hydrogen atom formed were measured by e.p.r. The values were plotted against initial ketene concentration and extrapolated to zero reactant concentration to obtain the limiting stoichiometries of reaction.For both medium (cu. 30ms) and long 187188 0-ATOM + CARBONYL REACTIONS (ca. 100 ms) reaction times good straight lines were obtained; these plots yielded 3.7k0.2 oxygen atoms consumed per ketene molecule at long reaction times. The ratio of oxygen atoms consumed to hydrogen atoms produced was obtained by similar extrapolation to zero ketene concentration as shown in fig. 1. It fell slightly from 2.1 +O. 1 at short reaction times (ca. 5 ms) to 1.8 f: 0.1 at medium reaction times. The hydrogen atom yield per reactant molecule dropped significantly as the reactant flow increased, but this effect was less than in the O+CH,CH02 and O+H2C03 reactions, so that for large ketene flows [HI> [O].The maximum concentration of hydrogen atoms generally occurred for [O] - 3[H]. - OO 1 2 3 G 1 010[CH2CO]o~mol ~ m - ~ FIG. 1.-Stoichiometry of hydrogen atom production at short reaction time (4.46 ms). [010 = 2.52 x mol STABLE PRODUCTS The analyses for carbon dioxide (table 1) show that 0.65+0.03 molecules of C02 per CH2C0 are formed at long reaction times. For shorter reaction times, where the reaction was not complete before quenching on the silver foil, the C02 yields are higher. Small amounts of formaldehyde and water could be detected under such conditions, but only traces of ketene suggesting that it was largely destroyed on the silver foil. TABLE 1 .-CARBON DIOXIDE YIELDS total total yield pmol pmol C W CHzCO r010/ CHiCO/ co21 total pressure/ reaction 1O~~~OIol Torr timelms mol cm-3 [CH~COIO 1.85 122.0 4.6 29.8 31.5 20.0 0.64 1.85 122.0 4.2 21.5 39.4 26.8 0.68 1.85 122.0 4.1 25 .O 39.2 24.2 0.62 2.25 21.9 2.46 10.0 39.2 30.8 0.79 1.93 11.8 1.31 25 .O 30.1 22.7 0.75 KINETICS OF OVERALL REACTION As established by Carr et d.,l the reaction is first order in [O] and in [CH,CO].The rate coefficient was determined from the oxygen atom decays at fixed short reaction times when varying amounts of ketene were added, the integrated second- order rate equation was used in the form where nx = [O], - [O] and n is the stoichiometry of oxygen atom consumption. To obtain large decays, these runs were carried to comparable initial oxygen atom and ketene concentrations, and the stoichiometries measured with excess oxygenG . P .R . MACK A N D k. A . THRUSH 189 atoms cannot be substituted directly. Apart from the initial points, values of n = 2, 3 and 4 gave good straight lines (fig. 2) and it is not possible to deduce the correct value of n for high ratios of [CH,CO], to [O], on this basis. It is shown below that n = 3.0k0.2 under these conditions ; putting n = 3 yields a value of kl = (3.42$. 0.09) x 10l1 cm3 mol-l s-l at 293 K which is independent of total pressure over the range 0.9 to 2.7 Torr (table 2). The steeper oxygen atom decay with small [CH,CO], arises from a higher effective stoichiometry of oxygen atom consumption under such conditions. 1010[CH2C8]o/mol ~ r n - ~ [010 = 2.52 x mol crne3. 0, n = 2 ; A, n = 3 ; 0, n = 4. FIG. 2.-Kinetics of oxygen atom consumption for various stoichiometries. Reaction time = 4.46 ms.TABLE 2.-DETERMINATION OF THE RATE COEFFICIENTS OF 0-k KETENE run K11 K12 K2 K13 K4 K3 K5 K1 reaction t helms 3.27 5.24 4.46 6.42 5.68 6.57 8.08 4.91 1010t010/ mol cm-3 2.12 2.12 2.52 2.12 2.46 2.66 2.42 4.55 1010[CH2CO]o range/ mol cm-3 0.2-5.9 0.6-3.6 0.27-4.1 0.58-2.56 0.86-3.85 1.03-3.5 0.77-2.53 0.78-7.9 kl/crn3 mol-1 s-1 for n = 3 3.12 3.57 3.40 3.81 3.19 3.28 3.53 2.78 REACTION MECHANISM Although the C-H bond energy in ketene has not been determined, there is no reason to suppose that it is any lower than in ethylene which hits a slightly longer C-H bond (1.085 A as against 1.079 A). Thus abstraction of an H atom by 0 would be at least 25 kJ mol-1 endothermic giving a Boltzmann factor of less than The initial step must, as with ethylene, involve the addition of an oxygen atom followed by fragmentation of the excited adduct.Carr et al.' consider the following processes at 293 K even if the reverse process has no activation energy. 0 + CH,CO+CH,CO~-CCO + H20 + 46 kJ mol-1 (14 -+H,CO + CO + 408 kJ mol-1 (I@ -+CH, + C02 + 197 kJ mol-1 (1 4 +HCO + H + CO + 56 kJ mol-1 (14 -+HCO+HCO+ 138 kJ mol-l. ( 1 4190 0-ATOM + CARBONYL REACTIONS Apart from its inherent implausibility being a weakly exothermic four-centre reaction, the lack of water formation in the presence of excess atomic oxygen excludes any large contribution from reaction (la). With reaction (lb) as initial step, atomic hydrogen is produced only by the subsequent 0 + H2C0 reaction which is four times slower than (1).In the presence of excess atomic oxygen (1 c) would be followed by giving a stoichiometry of which agrees with neither of the measured stoichiometries. rapidly with ketene [CH,CO], which is not observed. be (Id) and (le) with carbon dioxide being formed in one of the subsequent steps O+CH2 -+ CO+2H+316 kJ mol-' (2) 2 0 + CH2C0 -+ CO + C02 + 2H Furthermore, CH2 reacts and this would give a reduced hydrogen atom yield at high Thus, none of these three processes can be important, and the dominant paths must 0 + HCO -+ OH + CO + 350 kJ mob1 0 + HCO -+ H + C02 + 455 kJ mol-' (3) (4) H+ HCO -+ H2 + CO + 358 kJ mol-' O+OH 3 H + 0 2 +71 kJ niol-l (5) (6) where k3 : k4 : k, = 0.46 : 0.54 : 4.0 and k6 = 3 x l O I 3 cm3 mol-' s - ' .~ The initial ratios [O] to [CH,CO] used for the C02 analyses were such that the relation derived in Part 1 would predict CO, yields per CH2C0 molecule of 0.4 and 0.8 respectively for paths ( I d ) and (le) plus their subsequent reactions and 0.4 for path (lb) followed by complete oxidation of the formaldehyde formed. However, there was no evidence of a decrease in the stoichiometry of hydrogen atom production at short reaction times as would be required if hydrogen atom production depended on the slow O+H2C0 reaction. Reactions (Id) and (le) plus subsequent steps correspond respectively to 2.46 and 3.92 oxygen atoms consumed per ketone molecule to yield 2 hydrogen atoms. Thus the product analyses and stoichiometries show that (le) is the dominant initial process, probably with a contribution of about 25 % from The reac- tion of OH radicals with ketene has not been studied, but its rate constant is probably close to the value of 10l2 cm3 inol-I s-l found by Morris and Niki for OH+C2H4, and to the similar rate constant of O+C2H4,6 as there is close parallelism in the reaction rates of 0 and OH with ~ l e f i n s .~ The high rate constants of OH + CH3CH0 and OH+H,CO are almost certainly due to abstraction at weak aldehydic C-H bonds which are absent in ketene. In the 0 + CH2C0 reaction any OH formed will be removed almost exclusively by reaction (6). (Id). Reactions of ketene with other species are almost certainly negligible. The reaction H+CH,CO-,CH,CO~'+CH,+CO+ 122 kJ mol-i (7) has a rate constant O+CH,CO, and its importance is further reduced by the rapid subsequent step (8) which consumes 0 and regenerates H.At the higher initial ratios of [CH2CO] to [O] used in the kinetic studies, the first effect of the increase in [H]/[O] ratio as the reaction proceeds on the observed stoichiometry will, therefore, not involve the initial step but affect competition between H and 0 for HCO, since k , = 4(k3 +k4). of 8 x 1O1O em3 mol-l s-l, one quarter that deduced here for 0 + CH3 -+ H2C0 + H + 286 kJ mol-I (8)G . P. R . MACK A N D B . A . THRUSH 191 The relative values of k3, k4 and k, quoted above were combined with the average values of [O] and [HI to calculate the stoichiometries of oxygen atom removal (n) at different points on the kinetic plots used to calculate kl. Values for a typical run are shown in table 3.It can be seen that n = 3.0k0.2 over the region used to determine the rate coefficient, but at low ketene additions, where few H atoms are formed but the decrease in [O] is too small to measure accurately, the stoichiometry of oxygen atom consumption is higher, as illustrated by the steeper initial decrease in [O] in fig. 2. Applying this value of n to all the kinetic runs yields kl = (3.4k0.3) x loll cm3 mol-1 s-l at 293 K. TABLE 3 .-CALCULATION OF EFFECTIVE STOICHIOMETRY FOR EACH EXPERIMENTAL POINT IN RUN K1 (all concentrations in lo-’* cm3 mol-’ s-l) 0.784 2.96 1.60 0.534 3.67 0.267 3.58 1.791 1.97 2.58 0.765 2.99 0.383 3.27 2.708 1.33 3.22 0.810 2.46 0.41 5 3.10 3.397 1.06 3.49 0.801 2.20 0.401 3.1 1 4.133 0.86 3.70 0.760 1.97 0.380 2.99 5.392 0.60 3.96 0.686 1.65 0.343 2.93 6.146 0.51 4.04 0.687 1.52 0.344 2.87 7.839 0.36 4.19 0.610 1.28 0.315 2.83 “WCOlo [Ole,ptl [OIreacted [Hlexptl <O> (H> n This is somewhat lower than the value of kl = 5.3 x loll cm3 mol-1 s-l from the mass spectrometric study of the ketene decay by Carr et aZ.l ; however, these workers found iz = 2.0k0.5 and the observed rates of oxygen atom consumption in the two systems agree well.Their observations that the C02 from the l80 + CH2C160 reaction contained 20 % each of C1*02 and C1602, and their finding of large yields of water not detected in our experiments suggest that their results were affected by heterogeneous reactions of ketene in the sampling zone. CONCLUSIONS Our results are consistent with reaction (le) to yield HCO+HCO being the dom- inant primary step in the O+CH2C0 reaction. The short-fall in the COz yield suggests a smaller contribution from (Id) which yields HCO+H+CO.This latter process is to be expected to accompany (le) since the appearance of at least 60 % of the energy released in (Ie) as internal energy of either HCO fragment would lead to its dissociation. At low pressures CH3 + CHO are the products of the reaction of oxygen atoms with eth~lene.~ This behaviour corresponds to intersystem crossing and isomerisa- lion of the initial adduct to yield a highly vibrationally excited molecule of the C2H40 isomer with the lowest heat of formation, followed by its unimolecular decomposition. This arises because the steady state concentration of an isomer at any particular (high) energy is proportional to the density of its internal energy levels at that energy, and as the level density rises sharply with internal energy the most stable isomer predominates.Another example is the formation of CH2 + CO in the 0 + C2H2 reaction where the intermediate ketene molecule can be trapped in rigid matrices.ll On this basis the products of the 0 + CH2C0 reaction at low pressures should correspond to the rup- ture of the weakest bond in glyoxal which is the most stable C2H202 isomer. This192 O-ATOM + CARBONYL REACTIONS process would yield two HCO radicals in agreement with our conclusion that process (le) predominates, and that the observed products come from the subsequent rac- tions (3), (4), (5) and (6).The high stoichiometries of oxygen atom consumption and the CO, yields obs- erved here are not consistent with ketene being a significant intermediate in the reac- tion of excess atomic oxygen with acetaldehyde, reported in the preceding paper, since this would require higher stoichiometries of oxygen atom consumption and lower CO, yields than are observed in the 0 + CH3CH0 reaction. We thank the Ministry of Defence for an E.M. Research Contract and Dr. L. Phillips of E.R.D.E. for helpful discussions. R. W. Carr, I. D. Gay, G. P. Glass and H. Niki, J. Chem. Phys., 1968, 49, 846. G. P. R. Mack and B. A. Thrush, J.C.S. Firaday I, 1974,70, 178. G. P. R. Mack and B. A. Thrush, J.C.S. Fwahy I, 1973,69,208. H. Niki, C. McKnight and B. Weinstock, Abs. 154th Ann. Mtg.Amer. Chem. Soc., 1967, V110. C. B. Moore and G. C. Pimentel, J. Chem. Phys., 1963,38,2816. J. M . Brown and B. A. Thrush, Trans. Faraday Suc., 1967, 63, 630. ' W. Braun, A. M. Bass and M. Pilling, J. Chem. Phys., 1970, 52, 5131. M. A. A. Clyne and B. A. Thrush, Proc. Roy. SOC. A, 1963, 275, 544. E. D. Morris and H. Niki, J. Phys. Chem., 1971, 75, 3640. I. Haller and G. C. Pimentel, J. Amer. Chem. Soc., 1962, 84, 2855. l o R. J. CvetanoviE, J. Chem. Phys., 1955, 23, 1375. Reaction of Oxygen Atoms with Carbonyl Compounds Part 3.-Ketene BY G. P. R. MACK AND B. A. THRUSH* University of Cambridge, Department of Physical Chemistry, Lensfield Road, Cambridge CB2 1EP Received 27th July, 1973 The reaction between excess atomic oxygen and ketene has been studied in a discharge flow system by product analysis and kinetically using e.p.r.and chemiluminescznce. This rate constant for the initial step, which yields predominantly HCO+ HCO was found to be (3.4 k 0.3) x 10' ' cm3 mol-' s-' at 293 K. The subsequent reactions are O+HCO -+ OH+CO O+HCO -+ H+COz H+HCO-t H2+CO OfOH -+ 02+H. Can, Gay, Glass and Niki used a mass-spectrometer to study the reaction of ketene with atomic oxygen in a discharge-flow system. They deduced that the initial step was addition of an oxygen atom to ketene, predominantly at the olefinic carbon atom ; the observed products (CO, C02, H20, H2C0, H2, H) could be explained by rapid decomposition of the adduct followed by reaction of its fragments with atomic oxygen and with ketene. We have investigated the 0 + CH2C0 reaction in the apparatus previously used for the O+CH3CH0 and O+H2C0 reactions 2* ; and show that the initial step yields predominantly HCO + HCO formed by the isomerisation of the initial adduct to a vibrationally excited glyoxal molecule which then decomposes.Our results do not support the view that ketene is an important intermediate in the reaction of excess atomic oxygen with acetaldehyde. EXPERIMENTAL The reaction was studied in the discharge flow systems previously de~cribed.~ Atomic concentrations were measured by e.p.r. spectroscopy or the air afterglow and the reaction products were analysed by gas chromatography. All experiments were carried out at 293 K. Ketene was prepared by the thermal decomposition of acetic anhydride at 820K. Acetone and unreacted anhydride were removed at 195 K in a trap packed with glass wool.The ketene was purified by bulb-to-bulb distillation and a residual impurity of ethane was estimated by gas chromatography to be 5-8 %. All errors quoted are one standard deviation. RESULTS AND DISCUSSION STOICHIOMETRY OF THE REACTION The number of oxygen atoms consumed per ketene molecule added and per hydrogen atom formed were measured by e.p.r. The values were plotted against initial ketene concentration and extrapolated to zero reactant concentration to obtain the limiting stoichiometries of reaction. For both medium (cu. 30ms) and long 187188 0-ATOM + CARBONYL REACTIONS (ca. 100 ms) reaction times good straight lines were obtained; these plots yielded 3.7k0.2 oxygen atoms consumed per ketene molecule at long reaction times.The ratio of oxygen atoms consumed to hydrogen atoms produced was obtained by similar extrapolation to zero ketene concentration as shown in fig. 1. It fell slightly from 2.1 +O. 1 at short reaction times (ca. 5 ms) to 1.8 f: 0.1 at medium reaction times. The hydrogen atom yield per reactant molecule dropped significantly as the reactant flow increased, but this effect was less than in the O+CH,CH02 and O+H2C03 reactions, so that for large ketene flows [HI> [O]. The maximum concentration of hydrogen atoms generally occurred for [O] - 3[H]. - OO 1 2 3 G 1 010[CH2CO]o~mol ~ m - ~ FIG. 1.-Stoichiometry of hydrogen atom production at short reaction time (4.46 ms). [010 = 2.52 x mol STABLE PRODUCTS The analyses for carbon dioxide (table 1) show that 0.65+0.03 molecules of C02 per CH2C0 are formed at long reaction times.For shorter reaction times, where the reaction was not complete before quenching on the silver foil, the C02 yields are higher. Small amounts of formaldehyde and water could be detected under such conditions, but only traces of ketene suggesting that it was largely destroyed on the silver foil. TABLE 1 .-CARBON DIOXIDE YIELDS total total yield pmol pmol C W CHzCO r010/ CHiCO/ co21 total pressure/ reaction 1O~~~OIol Torr timelms mol cm-3 [CH~COIO 1.85 122.0 4.6 29.8 31.5 20.0 0.64 1.85 122.0 4.2 21.5 39.4 26.8 0.68 1.85 122.0 4.1 25 .O 39.2 24.2 0.62 2.25 21.9 2.46 10.0 39.2 30.8 0.79 1.93 11.8 1.31 25 .O 30.1 22.7 0.75 KINETICS OF OVERALL REACTION As established by Carr et d.,l the reaction is first order in [O] and in [CH,CO]. The rate coefficient was determined from the oxygen atom decays at fixed short reaction times when varying amounts of ketene were added, the integrated second- order rate equation was used in the form where nx = [O], - [O] and n is the stoichiometry of oxygen atom consumption.To obtain large decays, these runs were carried to comparable initial oxygen atom and ketene concentrations, and the stoichiometries measured with excess oxygenG . P . R . MACK A N D k. A . THRUSH 189 atoms cannot be substituted directly. Apart from the initial points, values of n = 2, 3 and 4 gave good straight lines (fig. 2) and it is not possible to deduce the correct value of n for high ratios of [CH,CO], to [O], on this basis.It is shown below that n = 3.0k0.2 under these conditions ; putting n = 3 yields a value of kl = (3.42$. 0.09) x 10l1 cm3 mol-l s-l at 293 K which is independent of total pressure over the range 0.9 to 2.7 Torr (table 2). The steeper oxygen atom decay with small [CH,CO], arises from a higher effective stoichiometry of oxygen atom consumption under such conditions. 1010[CH2C8]o/mol ~ r n - ~ [010 = 2.52 x mol crne3. 0, n = 2 ; A, n = 3 ; 0, n = 4. FIG. 2.-Kinetics of oxygen atom consumption for various stoichiometries. Reaction time = 4.46 ms. TABLE 2.-DETERMINATION OF THE RATE COEFFICIENTS OF 0-k KETENE run K11 K12 K2 K13 K4 K3 K5 K1 reaction t helms 3.27 5.24 4.46 6.42 5.68 6.57 8.08 4.91 1010t010/ mol cm-3 2.12 2.12 2.52 2.12 2.46 2.66 2.42 4.55 1010[CH2CO]o range/ mol cm-3 0.2-5.9 0.6-3.6 0.27-4.1 0.58-2.56 0.86-3.85 1.03-3.5 0.77-2.53 0.78-7.9 kl/crn3 mol-1 s-1 for n = 3 3.12 3.57 3.40 3.81 3.19 3.28 3.53 2.78 REACTION MECHANISM Although the C-H bond energy in ketene has not been determined, there is no reason to suppose that it is any lower than in ethylene which hits a slightly longer C-H bond (1.085 A as against 1.079 A).Thus abstraction of an H atom by 0 would be at least 25 kJ mol-1 endothermic giving a Boltzmann factor of less than The initial step must, as with ethylene, involve the addition of an oxygen atom followed by fragmentation of the excited adduct. Carr et al.' consider the following processes at 293 K even if the reverse process has no activation energy. 0 + CH,CO+CH,CO~-CCO + H20 + 46 kJ mol-1 (14 -+H,CO + CO + 408 kJ mol-1 (I@ -+CH, + C02 + 197 kJ mol-1 (1 4 +HCO + H + CO + 56 kJ mol-1 (14 -+HCO+HCO+ 138 kJ mol-l.( 1 4190 0-ATOM + CARBONYL REACTIONS Apart from its inherent implausibility being a weakly exothermic four-centre reaction, the lack of water formation in the presence of excess atomic oxygen excludes any large contribution from reaction (la). With reaction (lb) as initial step, atomic hydrogen is produced only by the subsequent 0 + H2C0 reaction which is four times slower than (1). In the presence of excess atomic oxygen (1 c) would be followed by giving a stoichiometry of which agrees with neither of the measured stoichiometries. rapidly with ketene [CH,CO], which is not observed. be (Id) and (le) with carbon dioxide being formed in one of the subsequent steps O+CH2 -+ CO+2H+316 kJ mol-' (2) 2 0 + CH2C0 -+ CO + C02 + 2H Furthermore, CH2 reacts and this would give a reduced hydrogen atom yield at high Thus, none of these three processes can be important, and the dominant paths must 0 + HCO -+ OH + CO + 350 kJ mob1 0 + HCO -+ H + C02 + 455 kJ mol-' (3) (4) H+ HCO -+ H2 + CO + 358 kJ mol-' O+OH 3 H + 0 2 +71 kJ niol-l (5) (6) where k3 : k4 : k, = 0.46 : 0.54 : 4.0 and k6 = 3 x l O I 3 cm3 mol-' s - ' . ~ The initial ratios [O] to [CH,CO] used for the C02 analyses were such that the relation derived in Part 1 would predict CO, yields per CH2C0 molecule of 0.4 and 0.8 respectively for paths ( I d ) and (le) plus their subsequent reactions and 0.4 for path (lb) followed by complete oxidation of the formaldehyde formed.However, there was no evidence of a decrease in the stoichiometry of hydrogen atom production at short reaction times as would be required if hydrogen atom production depended on the slow O+H2C0 reaction. Reactions (Id) and (le) plus subsequent steps correspond respectively to 2.46 and 3.92 oxygen atoms consumed per ketone molecule to yield 2 hydrogen atoms. Thus the product analyses and stoichiometries show that (le) is the dominant initial process, probably with a contribution of about 25 % from The reac- tion of OH radicals with ketene has not been studied, but its rate constant is probably close to the value of 10l2 cm3 inol-I s-l found by Morris and Niki for OH+C2H4, and to the similar rate constant of O+C2H4,6 as there is close parallelism in the reaction rates of 0 and OH with ~ l e f i n s .~ The high rate constants of OH + CH3CH0 and OH+H,CO are almost certainly due to abstraction at weak aldehydic C-H bonds which are absent in ketene. In the 0 + CH2C0 reaction any OH formed will be removed almost exclusively by reaction (6). (Id). Reactions of ketene with other species are almost certainly negligible. The reaction H+CH,CO-,CH,CO~'+CH,+CO+ 122 kJ mol-i (7) has a rate constant O+CH,CO, and its importance is further reduced by the rapid subsequent step (8) which consumes 0 and regenerates H. At the higher initial ratios of [CH2CO] to [O] used in the kinetic studies, the first effect of the increase in [H]/[O] ratio as the reaction proceeds on the observed stoichiometry will, therefore, not involve the initial step but affect competition between H and 0 for HCO, since k , = 4(k3 +k4).of 8 x 1O1O em3 mol-l s-l, one quarter that deduced here for 0 + CH3 -+ H2C0 + H + 286 kJ mol-I (8)G . P. R . MACK A N D B . A . THRUSH 191 The relative values of k3, k4 and k, quoted above were combined with the average values of [O] and [HI to calculate the stoichiometries of oxygen atom removal (n) at different points on the kinetic plots used to calculate kl. Values for a typical run are shown in table 3. It can be seen that n = 3.0k0.2 over the region used to determine the rate coefficient, but at low ketene additions, where few H atoms are formed but the decrease in [O] is too small to measure accurately, the stoichiometry of oxygen atom consumption is higher, as illustrated by the steeper initial decrease in [O] in fig.2. Applying this value of n to all the kinetic runs yields kl = (3.4k0.3) x loll cm3 mol-1 s-l at 293 K. TABLE 3 .-CALCULATION OF EFFECTIVE STOICHIOMETRY FOR EACH EXPERIMENTAL POINT IN RUN K1 (all concentrations in lo-’* cm3 mol-’ s-l) 0.784 2.96 1.60 0.534 3.67 0.267 3.58 1.791 1.97 2.58 0.765 2.99 0.383 3.27 2.708 1.33 3.22 0.810 2.46 0.41 5 3.10 3.397 1.06 3.49 0.801 2.20 0.401 3.1 1 4.133 0.86 3.70 0.760 1.97 0.380 2.99 5.392 0.60 3.96 0.686 1.65 0.343 2.93 6.146 0.51 4.04 0.687 1.52 0.344 2.87 7.839 0.36 4.19 0.610 1.28 0.315 2.83 “WCOlo [Ole,ptl [OIreacted [Hlexptl <O> (H> n This is somewhat lower than the value of kl = 5.3 x loll cm3 mol-1 s-l from the mass spectrometric study of the ketene decay by Carr et aZ.l ; however, these workers found iz = 2.0k0.5 and the observed rates of oxygen atom consumption in the two systems agree well.Their observations that the C02 from the l80 + CH2C160 reaction contained 20 % each of C1*02 and C1602, and their finding of large yields of water not detected in our experiments suggest that their results were affected by heterogeneous reactions of ketene in the sampling zone. CONCLUSIONS Our results are consistent with reaction (le) to yield HCO+HCO being the dom- inant primary step in the O+CH2C0 reaction. The short-fall in the COz yield suggests a smaller contribution from (Id) which yields HCO+H+CO. This latter process is to be expected to accompany (le) since the appearance of at least 60 % of the energy released in (Ie) as internal energy of either HCO fragment would lead to its dissociation.At low pressures CH3 + CHO are the products of the reaction of oxygen atoms with eth~lene.~ This behaviour corresponds to intersystem crossing and isomerisa- lion of the initial adduct to yield a highly vibrationally excited molecule of the C2H40 isomer with the lowest heat of formation, followed by its unimolecular decomposition. This arises because the steady state concentration of an isomer at any particular (high) energy is proportional to the density of its internal energy levels at that energy, and as the level density rises sharply with internal energy the most stable isomer predominates. Another example is the formation of CH2 + CO in the 0 + C2H2 reaction where the intermediate ketene molecule can be trapped in rigid matrices.ll On this basis the products of the 0 + CH2C0 reaction at low pressures should correspond to the rup- ture of the weakest bond in glyoxal which is the most stable C2H202 isomer. This192 O-ATOM + CARBONYL REACTIONS process would yield two HCO radicals in agreement with our conclusion that process (le) predominates, and that the observed products come from the subsequent rac- tions (3), (4), (5) and (6). The high stoichiometries of oxygen atom consumption and the CO, yields obs- erved here are not consistent with ketene being a significant intermediate in the reac- tion of excess atomic oxygen with acetaldehyde, reported in the preceding paper, since this would require higher stoichiometries of oxygen atom consumption and lower CO, yields than are observed in the 0 + CH3CH0 reaction. We thank the Ministry of Defence for an E.M. Research Contract and Dr. L. Phillips of E.R.D.E. for helpful discussions. R. W. Carr, I. D. Gay, G. P. Glass and H. Niki, J. Chem. Phys., 1968, 49, 846. G. P. R. Mack and B. A. Thrush, J.C.S. Firaday I, 1974,70, 178. G. P. R. Mack and B. A. Thrush, J.C.S. Fwahy I, 1973,69,208. H. Niki, C. McKnight and B. Weinstock, Abs. 154th Ann. Mtg. Amer. Chem. Soc., 1967, V110. C. B. Moore and G. C. Pimentel, J. Chem. Phys., 1963,38,2816. J. M . Brown and B. A. Thrush, Trans. Faraday Suc., 1967, 63, 630. ' W. Braun, A. M. Bass and M. Pilling, J. Chem. Phys., 1970, 52, 5131. M. A. A. Clyne and B. A. Thrush, Proc. Roy. SOC. A, 1963, 275, 544. E. D. Morris and H. Niki, J. Phys. Chem., 1971, 75, 3640. I. Haller and G. C. Pimentel, J. Amer. Chem. Soc., 1962, 84, 2855. l o R. J. CvetanoviE, J. Chem. Phys., 1955, 23, 1375.

ISSN:0300-9599    

DOI:10.1039/F19747000187

出版商:RSC    

年代:1974

数据来源: RSC