A Nucleophilic Substitution Reaction Performed in Different Types of

Jun 16, 2004 - The rate was independent of the microstructure of the microemulsion; however, the rate was very dependent on the type of surfactant use...
4 downloads 10 Views 172KB Size
Download PDF
Langmuir 2004, 20, 6107-6115

6107

A Nucleophilic Substitution Reaction Performed in Different Types of Self-Assembly Structures Maria Ha¨ger,† Ulf Olsson,‡ and Krister Holmberg*,§ Institute for Surface Chemistry, P.O. Box 5607, SE-114 86 Stockholm, Sweden, Physical Chemistry 1, Lund University, P.O. Box 124, 221 00 Lund, Sweden, and Chalmers University of Technology, Applied Surface Chemistry, SE-412 96 Go¨ teborg, Sweden Received December 15, 2003. In Final Form: April 23, 2004 A nucleophilic substitution reaction between 4-tert-butylbenzyl bromide and potassium iodide has been performed in oil-in-water microemulsions based on various C12Em surfactants, i.e., dodecyl ethoxylate with m number of oxyethylene units. The reaction kinetics was compared with the kinetics of reactions performed in other self-assembly structures based on very similar surfactants and in homogeneous liquids. The reaction was fastest in the micellar system, intermediate in rate in the microemulsions, and most sluggish in the liquid crystalline phase. Reaction in a Winsor I system, i.e., a two-phase system comprising an oil-in-water microemulsion in equilibrium with excess oil, was equally fast as reaction in a one-phase microemulsion. The reactions in microemulsion were surprisingly fast compared to reaction in homogeneous, protic liquids such as methanol and ethanol. The rate was independent of the microstructure of the microemulsion; however, the rate was very dependent on the type of surfactant used. When the C12Em surfactant was replaced by a sugar-based surfactant, octyl glucoside, the reaction was much more sluggish. The high reactivity in microemulsions based on C12Em surfactants is belived to be due to a favorable microenvironment in the reaction zone. The reaction is likely to occur within the surfactant palisade layer, where the water activity is relatively low and where the attacking species, the iodide ion, is poorly hydrated and, hence, more nucleophlic than in a protic solvent such as water or methanol. Sugar surfactants become more hydrated than alcohol ethoxylates and the lower reactivity in the microemulsion based on the sugar surfactant is probably due to a higher water activity in the reaction zone.

Introduction Organic synthesis is not always straightforward. Incompatible reactants, e.g., an inorganic salt and a lipophilic organic compound, often have problems in coming into the intimate contact needed for a reaction to occur. There are ways to overcome this problem, such as using aprotic, polar solvents such as DMF and DMSO. However, these solvents are toxic and have high boiling points, which makes them difficult to remove after completed reaction. They are therefore not suitable for use in large scale operations. An alternative is to use two-phase (wateroil) systems, where the reaction can take place at the interface between the two phases. To enhance the reaction vigorous stirring and/or addition of a phase transfer agent can be used. A phase transfer agent is usually a lipophilic quaternary ammonium salt (Q salt) or a crown ether. These compounds have the ability to transfer anionic reactants from the water phase into the organic phase where the nucleophile becomes highly reactive due to low degree of solvation and to relatively loose ion pairing with the Q salt or crown ether. Microemulsions have also been found to be useful as media for organic reactions.1-8 A microemulsion is a thermodynamically stable system containing oil, water, †

Institute for Surface Chemistry. Lund University. § Chalmers University of Technology. ‡

(1) Schwuger, M.; Stichdorn K.; Schoma¨cker R. Chem. Rev. 1995, 95, 849. (2) Holmberg, K. Adv. Colloid Interface Sci. 1994, 51, 137. (3) Mackay, R. A.; Wang, J. Tetrahedron Lett. 1992, 33, 6415. (4) Schoma¨cker, R. Nachr. Chem., Tech. Lab. 1992, 40, 1344. (5) Sjo¨blom, J.; Lindberg, R.; Friberg, S. E. Adv. Colloid Interface Sci. 1996, 95, 125. (6) Holmberg, K.; Ha¨ger, M. In Adsorption and Aggregation of Surfactants in Solution; Mittal K. L., Shah D. O., Eds.; Marcel Dekker: New York, 2002; p 327.

and surfactant. It may consist of droplets, either water in oil or oil in water, or it may have a structure where both the aqueous and the oil domains are continuous. Regardless of the structure, the surfactant is situated as a monolayer between the water and oil domains. The type of structure formed depends among other things on temperature, salt concentration, cosurfactant concentration and the relative amount of the microemulsion components. These ternary systems usually have an ability to dissolve both polar and nonpolar compounds. The interest in microemulsions in preparative organic synthesis relates to the extremely large interfacial area between the oil and water domains, at which reactants can collide and react. Microemulsions are also highly dynamic systems with interfaces disintegrating and reforming at the time scale of milliseconds, which might also promote the kinetics of the reaction. Other factors that may affect the reaction kinetics are surfactant concentration, local reactant concentration, reactant distribution between and within the different subvolumes in the microemulsion, choice of surfactant, temperature, etc. To optimize the conditions of a specific organic synthesis in a microemulsion, it is important to have a good understanding of the reaction kinetics and a proper knowledge about which parameters influence the reaction rate. A common approach has been to describe the kinetics according to the so-called pseudophase model, in which the reaction volume is divided into subvolumes.9,10 However, this model is not very suitable for quantitative (7) Ha¨ger, M.; Currie, F.; Holmberg, K. In Colloid Chemistry. Topics in Current Chemistry; Antonietti, M., Ed.; Springer: Heidelberg 2003; p 53. (8) Holmberg, K. Curr. Opin. Colloid Interface Sci. 2003, 8, 187. (9) Schoma¨cker, R.; Stickdorn, K.; Knoche, W. J. Chem. Soc., Faraday Trans. 1991, 87, 847. (10) Oh, S.-G.; Kizling, J.; Holmberg, K. Colloids Surf., A 1995, 97, 169.

10.1021/la030434i CCC: $27.50 © 2004 American Chemical Society Published on Web 06/16/2004

6108

Langmuir, Vol. 20, No. 15, 2004

Ha¨ ger et al.

Scheme 1. Nucleophilic Substitution Reaction between Various Benzyl Bromides and KIa

a

R is H, CH3, CH(CH3)2, or C(CH3)3.

calculations since it is difficult to determine the partition coefficients of the reactants between the different subvolumes. It is also difficult to measure the rate constants within the subvolumes. In this work we have studied the reaction kinetics of a nucleophilic substitution reaction performed in oil-inwater microemulsions based on different fatty alcohol ethoxylates or a sugar surfactant. The effects of surfactant concentration and of aggregate size and shape were investigated, and a new interfacial kinetic model is proposed. The reaction was also performed in a micellar system and in a liquid crystalline phase with a composition very similar to that of a microemulsion, and the rates in the different media were compared. For comparison, the reaction was also performed in polar organic solvents, i.e., microscopically homogeneous systems. Experimental Section

volume ratio between the surfactant and the oil solution was kept constant at 0.819 in all formulations, and the weight fraction of oil and surfactant, φ, was 0.2, unless otherwise mentioned. Thereby, the volume fraction of oil based on oil and aqueous phase was 0.12, and the volume fraction of surfactant based on total volume was 0.11. Two quaternary microemulsions were also used. These consisted of D2O, octane, octanol, and C12E8 or C8G1. In both systems, the volume fraction of oil based on oil and aqueous phase was 0.2 and the volume fraction of the surfactant and cosurfactant based on total volume was 0.11 and 0.03, respectively. For comparison, the reaction was performed in both protic and aprotic solvents: d-MeOH, d-EtOH, d-DMF, and d-DMSO. The standard reaction was carried out in the oil-in-water microemulsion based on D2O, C12E5, and decane. All reactions, except those performed in the lamellar and Winsor I systems, were carried out at 23 °C, and the reaction kinetics was analyzed by in situ 1H NMR, using a 600 MHz Varian Inova spectrometer. The temperature in the NMR probe was controlled by an external thermostat with an accuracy of (0.1 °C. The reactions were monitored for 3 h, and spectra were recorded every 5 min. The decrease of the substrate and the increase of the product were obtained from the integrals of the singlet signals from the methylene groups at δ 4.42 ppm (-CH2Br) and δ 4.39 ppm (-CH2I), respectively. The reactions in the lamellar and the Winsor I systems were performed in a thermostated water bath. The reaction was quenched by CDCl3 after different time intervals. The aqueous phase was removed after 5 min of centrifugation at 3000 rpm, and the organic phase was dried by Mg2SO4. After filtration the organic phase was analyzed by 1H NMR. NMR Diffusometry. The pulsed field nuclear magnetic resonance technique was used to determine the different selfdiffusion coefficients of the components in the D2O-C12E4/C12E5decane system at T ) 23 °C by monitoring the 1H signal on a Varian 500 MHz spectrometer, equipped with a field gradient pulse probe unit. The experiments were performed with a Hahn echo pulse sequence, where the gradient strength was varied between 0.3 and 4.8 T/m in 16 experimental steps and ∆ was 100 ms.

Materials. The nonionic surfactants used, tetra(ethylene glycol) monododecyl ether (C12E4), penta(ethylene glycol) monododecyl ether (C12E5), hexa(ethylene glycol) monododecyl ether (C12E6), and octa(ethylene glycol) monododecyl ether (C12E8), were all Nikko Chemicals products. The sugar surfactant, n-octyl-βD-glucopyranoside (C8G1), was from Anatrace, Inc., and the hydrocarbons, octane and decane, were from Sigma. All the deuterated solvents, d-dimethyl sulfoxide (d-DMSO), d-dimethylformamide (d-DMF), d-methanol (d-MeOH), d-ethanol (dEtOH), and deuterium oxide (D2O), where all from Aldrich. The reagents, potassium iodide (KI) and all the benzyl bromides, benzyl bromide (BB), 4-methylbenzyl bromide (4-MBB), 4-isopropylbenzyl bromide (4-IPBB), and 4-tert-butylbenzyl bromide (4-TBBB), were purchased from Merck and Aldrich. The two alcohols used, 1-pentanol (C5OH) and 1-heptanol (C7OH), were both obtained from Fluka. The chemicals used had a purity of g99%, except for the benzyl bromides (97% or 98%), and all of them were used without any further purification. Phase Diagram Determination. Samples for the phase diagrams were prepared by titration of the aqueous phase into stock solutions containing the organic solution and the surfactant(s). Screw-capped glass tubes with stirrer were used in all the experiments. The samples were equilibrated and studied in a thermostated water bath by visual light and crossed polarizers. The lower temperature boundary of the microemulsion region was observed by increasing the temperature step by step 1 °C at the time. The equilibrium is reached slowly when decreasing the temperature, while it occurs much more rapidly, within a few minutes, when increasing the temperature.11-14 The other phase boundaries were detected both by increasing and by decreasing the temperature. The concentrations of all components were controlled by weight with an accuracy of (0.001 g, and the total amount of each sample was 1 g. Reaction Procedure. The aqueous solution contained KI in D2O, and the oil solution contained the benzyl bromide in decane or octane. The reactants were always used in a mole ratio of 1:1. Ternary self-assembly structures containing D2O, decane, and a surfactant or a combination of surfactants (C12E4/C12E5, C12E5, C12E6, and C12E5/CnOH) were used as reaction medium. The

Reaction. The reaction investigated, a nucleophilic substitution reaction between benzyl bromide or a 4-alkylbenzyl bromide and potassium iodide, is illustrated in Scheme 1. Phase Behavior. The ternary system C12E5-decaneD2O has previously been characterized by Olsson and coworkers.15-19 Most of these studies were made at a constant volume ratio between surfactant and oil of 0.819, and the total weight fraction of oil and surfactant was given as φ. Figure 1 shows the phase behavior at φ ) 0.2 in the absence and in the presence of 4-TBBB. As shown, an oil-in-water microemulsion, a so-called L1 phase, exists between 23.5 and 30.5 °C when no substrate has been added. The droplets are known to be spherical at the lower phase boundary and grow as the temperature is increased.20

(11) Morris, J.; Olsson, U.; Wennerstro¨m, H. Langmuir 1997, 13, 606. (12) Wennerstro¨m, H.; Morris, J.; Olsson, U. Langmuir 1997, 13, 6972. (13) Egelhaaf, S.; Olsson, U.; Schurtenberger, P.; Morris, J.; Wennerstro¨m, H. Phys. Rev. E 1999, 60, 5681. (14) Evilevitch, A.; Olsson, U.; Jo¨nsson, B.; Wennerstro¨m, H. Langmuir 2000, 16, 8755.

(15) Olsson, U.; Schurtenberger, P. Langmuir 1993, 9, 3389. (16) Leaver, M. S.; Olsson, U. Langmuir 1994, 10, 3449. (17) Fukuda, K.; Olsson, U.; Wu¨rz, U. Langmuir 1994, 10, 3222. (18) Olsson, U.; Wennerstro¨m, H. Adv. Colloid Interface Sci. 1994, 49, 113. (19) Olsson, U.; Schurtenberger, P. Prog. Colloid Sci. 1997, 104, 157. (20) Bagger-Jo¨rgensen, H.; Olsson, U.; Mortensen, K. Langmuir 1997, 13, 1413.

Results

Nucleophilic Substitution Reactions in Microemulsions

Figure 1. Partial phase diagram of the ternary D2O-C12E5decane system in the presence of 4-TBBB. The volume ratio between surfactant and oil is 0.82 and the total weight fraction of oil + surfactant is 0.20. Raising the temperature gives a transition from an oil-in-water microemulsion in equilibrium with excess oil (L1 + O region) to an oil-in-water microemulsion (L1 region) and on to a lamellar liquid crystalline phase (LR region).

Figure 2. Partial phase diagram of the ternary D2O-C12E5decane system in the presence of KI. The volume ratio between surfactant and oil is 0.82, and the total weight fraction of oil + surfactant is 0.20.

Langmuir, Vol. 20, No. 15, 2004 6109

Figure 3. Partial phase diagram of the ternary D2O-C12E5decane system in the presence of 1-pentanol (b) and 1-heptanol (0), given in mass fraction. The volume ratio between surfactant and oil is 0.82, and the total weight fraction of oil + surfactant + cosurfactant is 0.20. The arrows show the reaction composition for the different alcohols.

Figure 4. Partial phase diagram of the quaternary D2O-C8G1octanol-octane system at 25 °C, given in mass fraction. At low amounts of octanol, a microemulsion coexists with excess oil (2). At higher concentrations of octanol, a microemulsion coexists with excess water (2). Between the low and high concentrations, a narrow microemulsion coexists with excess of both water and oil (3). At higher surfactant amount, the latter region is transformed into an oil-in-water microemulsion (1).

Close to the upper phase boundary the structure of the microemulsion becomes bicontinuous. The area per surfactant headgroup is almost invariant (47 Å2) for φ > 0.05 within the temperature range of the L1 phase.20-22 Increasing φ up to 0.4 gives similar trends as for φ ) 0.2, and the boundaries lie at approximately the same temperatures.23,24 Addition of the hydrophobic 4-TBBB leads to a shift of the L1 region to lower temperatures. On the other hand, adding increasing amounts of KI moves the L1 region to higher temperatures, which is illustrated in Figure 2. Addition of small amounts of either 1-pentanol or 1-heptanol leads to a shift of the L1 region to lower temperatures (Figure 3). The lower temperature boundary was not detected for 1-heptanol due to problems of regulating the temperature below 18 °C. Different self-assembly structures were characterized at 23 °C when varying the number of oxyethylene groups

of the nonionic surfactant C12Em. A microemulsion was obtained for ratios of C12E5 to C12E4 between 10:0 and 8:2. A ratio of 1:1 gave a lamellar crystalline phase. C12E6 gave an L1 phase in equilibrium with excess oil, a socalled Winsor I system. The excess oil phase was approximately 80% of the total amount of oil. The phase behavior of the D2O-C8G1-octanol-octane system has been studied in detail by Sottman and coworkers.25-27 In Figure 4 a phase diagram at 25 °C is shown of the system in the absence of reactants where the volume fraction of oil based on oil and aqueous phase was kept constant at 0.2. As seen, a very small L1 region is observed. The oil droplets of the L1 phase have shown to be relatively spherical at this composition.25 Adding the

(21) Leaver, M. S.; Olsson, U.; Wennerstro¨m, H. Strey, R. J. Phys. II 1994, 4, 515. (22) Leaver, M.; Furo, I.; Olsson, U. Langmuir 1995, 11, 1524. (23) Leaver, M. S.; Olsson, U.; Wennerstro¨m, H.; Strey, R.; Wu¨rz, U. J. Chem. Soc., Faraday Trans. 1994, 91, 4269. (24) Olsson, U.; Wu¨rz, U. J. Phys. Chem. 1993, 97, 5435.

(25) Reimer, J.; So¨derman, O.; Sottman, T.; Kluge, K.; Strey, R. Langmuir, in press. (26) Sottman, T.; Kluge, K.; Strey, R.; Reimer, J.; So¨derman, O. Langmuir 2002, 18, 3058. (27) Kluge, K.; Stubenrauch, C.; Sottman, T.; Strey, R. Tenside, Surfactants, Deterg. 2001, 38, 30.

6110

Langmuir, Vol. 20, No. 15, 2004

Figure 5. Effect of solvent on the rate of reaction between 4-TBBB and KI at 23 °C. The reaction media used are D2OC12E5-decane microemulsion (O), MeOH (4), EtOH (0), DMF (]), and DMSO (×).

reactants, either the benzyl bromide or the KI, did not affect the phase behavior at the composition where the reaction takes place. A similar L1 region was found when the sugar surfactant was replaced by the alcohol ethoxylate C12E8, otherwise keeping the composition the same. Reaction Kinetics. The reaction profiles in different reaction media are shown in Figure 5. In the microemulsion-based reactions the reaction mixture was equilibrated at the given temperature in the NMR probe, and the first NMR spectrum was taken at approximately 7 min. The substrate concentration was measured as a function of time, and an exponential decay of the substrate concentration was observed. No hydrolysis product, i.e., alkylbenzyl alcohol, was observed. The reaction was also performed in various polar liquids that were either protic or aprotic solvents. As shown in Figure 5, the reaction in the aprotic solvents, d-DMF and d-DMSO, had already reached equilibrium at the first measurement, giving a yield of approximately 75%. Thus, the initial reaction rate could not be determined, but it is obvious that the rate is orders of magnitude higher than the rate in the protic solvents. As seen, the decay in the protic solvents, d-MeOH and d-EtOH, is relatively tardy, indicating a lower reaction rate. The reaction is slightly slower in d-MeOH than in d-EtOH. Comparing the rates obtained in the protic solvents with that found in the D2O-C12En-decane microemulsion, it can be seen that the reactions in the former are only slightly faster than the reaction in the latter. The effect of the amount of surfactant on the reaction rate was studied in the D2O-C12E5-decane system by varying the total fraction of oil and surfactant while keeping the ratio between the two constant. This study can be made in different ways with respect to the concentrations of reactants when changing the total fraction of oil and surfactant. The overall concentration of the reactants may be kept constant or the concentration of one of the reactants in the bulk phase may be kept constant. In this work either the overall concentration of the reactants or the substrate concentration in the oil phase was kept constant. The results turned out to be the same. As shown in Figure 6, there is no clear difference in reaction rate when the volume ratio between surfactant and oil is kept constant while the weight fraction of surfactant and oil is increased. However, to be able to predict the actual effect of the surfactant concentration, the local concentrations in the subvolumes has to be considered. This issue will be discussed later.

Ha¨ ger et al.

Figure 6. Effect of surfactant amount on the rate of reaction between 4-TBBB and KI at 23 °C in D2O-C12E5-decane microemulsion. The volume ratio between surfactant and oil is 0.82, and the weight fraction of surfactant and oil is 0.10 (0), 0.20 (O), 0.30 (]), and 0.40 (4). The concentrations of the reactants were adjusted so that their mole ratio was always 1:1.

Figure 7. Reaction profile for the nucleophilic substitution reaction at 23 °C in (a) methanol and (b) D2O-C12E5-decane microemulsion between KI and BB (]), 4-MBB (×), 4-IPBB (0), and 4-TBBB (O). The concentrations of the reactants were adjusted so that their mole ratio was always 1:1.

To investigate the effect on the reaction rate of the solubility characteristics of the substrate, the reaction was performed with various benzyl bromides, as is illustrated in Scheme 1. In d-MeOH the reactivity order was 4-TBBB ) 4-IPBB > 4-MBB . BB (Figure 7a). Carrying out the reactions in the D2O-C12E5-decane microemulsion gave a totally different reaction pattern, as can be seen in Figure 7b. Conversion to the benzyl iodide is faster for the less lipophilic substrates, i.e., BB

Nucleophilic Substitution Reactions in Microemulsions

Langmuir, Vol. 20, No. 15, 2004 6111

Table 1. Self-Diffusion Coefficients of D2O (DH2O), C12E4/C12E5 (DC12Em), and Decane (Ddecane) of the Components in the D2O-C12E4/C12E5-Decane Systems microemulsion

mC12E4/ (mC12E4 + mC12E5)

DH2O (10-9 m2/s)

DC12Em (10-12 m2/s)

Ddecane (10-12 m2/s)

D2O-C12E5-decane D2O-C12E4-C12E5-decane D2O-C12E4-C12E5-decane D2O-C12E4-C12E5-decane

0.1 0.15 0.2

1.5 1.5 1.5 1.5

15.2 7.9 6.2 5.0

13.6 7.9 6.3 4.6

Figure 9. Reaction profiles for the reaction between 4-TBBB and KI at 23 °C in D2O-C12E5-decane microemulsion (O), D2OC12E8-octanol-octane microemulsion (]) and D2O-C8G1octanol-octane microemulsion (0). The concentrations of the reactants were adjusted so that their mole ratio was always 1:1.

Figure 8. Reaction profiles for the reaction between 4-TBBB and KI at 23 °C performed in (a) D2O-C12En-decane microemulsions, 10:0 (O), 9:1 (0), 8.5:1.5 (]), 8:2 (×), and (b) lamellar phase (0), L1 microemulsion (O), L1 micellar system (4), Winsor I system without stirring (]), and Winsor I system with stirring (×). The volume ratio between surfactant(s) and oil is 0.819. The concentrations of the reactants were adjusted so that their mole ratio was always 1:1.

and 4-MBB. The reactions with the more lipophilic substrates 4-IPBB and 4-TBBB were considerably more sluggish. Reactions with 4-TBBB were performed in microemulsions based on varying ratios of C12E5 and C12E4. Selfdiffusion NMR showed that the diffusion rate of the surfactant decreased with increasing content of C12E4, indicating a growth of the oil droplets (Table 1). As illustrated in Figure 8a, the change in droplet size did not affect the reaction rate to any appreciable extent as long as the system remained an oil-in-water microemulsion. Going to a 1:1 molar ratio of C12E5 and C12E4 means a transition into a lamellar liquid crystalline phase. As can be seen in Figure 8b, the reaction rate was somewhat lower in this phase than in the L1 microemulsion. The figure also shows that reaction in a Winsor I system, i.e., a two-phase system comprising an oil-in-water microemulsion in equilibrium with oil, is equally fast as the reaction in the one-phase microemulsion. Stirring the Winsor system is obviously not necessary. The figure also

shows that reaction in a pure micellar system, an aqueous C12E5 solution with 4-TBBB solubilized in the micelles, was more rapid than the reaction in the microemulsion. The reaction with 4-TBBB as substrate was also studied in a D2O-C8G1-octanol-octane microemulsion, i.e., a microemulsion in which the alcohol ethoxylate had been replaced by a sugar-based surfactant and where a cosurfactant (octanol) had been added in order to obtain a microemulsion. The reaction in this system, which was also an oil-in-water microemulsion, was significantly slower than the reaction in the D2O-C12E5-decane system (Figure 9). To ascertain that the reduction in reaction rate is not due to the presence of the cosurfactant, an oil-in-water microemulsion based on D2O, C12E8, octanol, and octane was formulated. Thus, the composition of this microemulsion was the same as that of the C8G1-based microemulsion except for the type of nonionic surfactant used. As can be seen, the reaction rates in the microemulsions based on the C12En surfactants are approximately the same and much higher than the rate in the microemulsion based on the C8G1 system. Figure 10 shows results from a more systematic investigation of the effect of addition of a medium chain alcohol on the reaction rate. As can be seen, the presence of alcohol does not much influence the reaction rate. Discussion In principle, the reaction shown in Scheme 1 may be either a unimolecular or a bimolecular substitution reaction, i.e., an SN1 or an SN2 reaction. The former is a two-step reaction, and the first step, ionization of the substrate, is the rate-determining step. Since the slow step involves only the substrate, the rate depends solely on the concentration of that. The latter is a one-step reaction and both the substrate and the nucleophile are involved in the rate-determining step. To determine the mechanism of the reaction between 4-TBBB and KI, it

6112

Langmuir, Vol. 20, No. 15, 2004

Ha¨ ger et al.

Figure 10. Effect of medium chain alcohols on the reaction between 4-TBBB and KI at 23 °C in D2O-C12E5-decane microemulsion (O). The volume ratio between surfactant + cosurfactant and oil is 0.819, and the weight ratio between C12E5 and alcohol is 96:4 (1-pentanol) (0), 92:8 (1-pentanol) (]), and 92:8 (1-heptanol) (×). The concentrations of the reactants were adjusted so that their mole ratio was always 1:1.

was performed in a series of organic solvents. It is known that the effect of the solvent is very different for the two types of reaction pathways.28-30 For an SN1 reaction with a neutral substrate, a polar solvent increases the rate since there is a greater charge in the transition state than in the starting substrate. The increase is often more pronounced for a protic than for an aprotic polar solvent since the leaving group can form hydrogen bonds with the latter. In contrast, for an SN2 reaction between a neutral substrate and an anionic nucleophile, the reaction is hindered by polar solvents because the initial charge is spread out in the transition state. For such SN2 reactions there is also a considerable difference between protic and aprotic polar solvents. The nucleophile is more solvated in a protic solvent while the transition state is more solvated in an aprotic solvent. Thus, an SN2 reaction involving a neutral substrate and an anionic nucleophile is generally retarded by polar solvents and more so by a protic than an aprotic polar solvent. The reaction between 4-TBBB and KI was performed in four different polar solvents, two protic (d-MeOH and d-EtOH, with dielectric constants of 33 and 25, respectively), and two aprotic (d-DMSO and d-DMF, with dielectric constants of 47 and 38, respectively). The reaction rates in the aprotic solvents were orders of magnitudes faster than in the protic solvents. The reaction was somewhat faster in ethanol than in the more polar methanol. These solvent effects clearly show that the reaction is of the SN2 type. The rate of a bimolecular nucleophilic substitution reaction performed in a homogeneous medium can be expressed as

1 dnj ) kCACB r)V dt

Figure 11. Reverse substrate concentration against reaction time for the reaction between 4-TBBB and KI at 23 °C in D2OC12E5-decane microemulsion. The slope of the curve gives the rate constant for an SN2 reaction when the reactants are used in stoichiometric amounts.

of KI and 4-TBBB, respectively. The rate constants, calculated as the slope from the straight line obtained when plotting the reverse concentration of substrate against reaction time (see Figure 11), are compiled in Table 2. The reaction media containing self-assembly structures are not homogeneous at the molecular level in that they consist of microscopic domains of water and oil separated by a surfactant film. The reaction may occur in either of the two domains as well as at the interface. However, since the solubility of KI in decane and of 4-TBBB in D2O is extremely low, the reaction can be assumed to be a purely interfacial reaction, i.e., no reaction occurs in the bulk phases. For an interfacial reaction in a microemulsion, the rate can be described as

ri ) -

1 dnBi 1 dnAi )) kiCAiCBi Vs dt Vs dt

where Vs is the surfactant volume and ki the rate constant at the interface. CAi and CBi are the concentrations at the interface of KI and 4-TBBB, respectively. The volume fraction of surfactant can be written as

φs )

(28) Reichardt, C. Solvents and solvent effect in Organic Chemistry, 2nd ed.; VCH: New York, 1988. (29) Bentley, T. W.; Schleyer, P. R. Adv. Phys. Org. Chem. 1977, 14, 1. (30) Parker, A. J. Chem. Rev. 1969, 69, 1.

Vs Vw + Vs + Vo

(3)

where Vw, Vs, and Vo are the volumes of water, surfactant, and oil, respectively. From (1) and (2) it follows that

r ) riφs

(4)

kCACB ) φskiCAiCBi

(5)

and

(1)

where V is the total volume of the reaction mixture and nj is the amount of component j (mole), k is the rate constant, and CA and CB, in our case, are the concentrations

(2)

where φs is the volume fraction of surfactant. The partition coefficient of the reactants between bulk and interface in the microemulsion is given as

KAwi )

CAw CAi

and

KBoi )

CBo CBi

(6)

Nucleophilic Substitution Reactions in Microemulsions

Langmuir, Vol. 20, No. 15, 2004 6113

Table 2. Rate Constants for the Reaction between 4-TBBB and KI in Various Self-Assembly Structures and Homogeneous Liquids microemulsion D2O-C12E5-decane (φ ) 0.1) D2O-C12E5-decane (φ ) 0.2) D2O-C12E5-decane (φ ) 0.3) D2O-C12E5-decane (φ ) 0.4) D2O-C12E5-C5OH-decane (4:96 C5OH/C12E5) D2O-C12E5-C5OH-decane (8:92 C5OH/C12E5) D2O-C12E5-C7OH-decane (2:98 C7OH/C12E5) D2O-C12E5-decane (10:0 C12E4/C12E5) D2O-C12E4-C12E5-decane (1:9 C12E4/C12E5) D2O-C12E4-C12E5-decane (1.5:8.5 C12E4/C12E5) D2O-C12E4-C12E5-decane (2:8 C12E4/C12E5) D2O-C12E6-decane (no stirring) D2O-C12E6-decane (stirring) D2O-C12E8-octanol-octane D2O-C8G1-octanol-octane micelle D2O-C12E5 lamellar phase D2O-C12E4-C12E5-decane (1:1 C12E4/C12E5) homogeneous solvent d-MeOH d-EtOH d-DMF std-DMSO

C4-TBBB (mM)

k (dm3/mol‚s)

9.03 8.76 8.47 8.32 8.74 8.70 8.72 8.68 8.72 8.74 8.75 8.76 8.76 10.49 11.35

0.0041 0.0044 0.0045 0.0044 0.0041 0.0033 0.0040 0.0038 0.0037 0.0035 0.0034 0.0050 0.0055 0.0031 0.0008

9.98

0.0110

8.74

0.0025

9.6 9.6 9.6 9.6

0.0086 0.020

and the interface concentrations can be expressed as

CAi )

CA φs + KAwiφw

(7)

C Bi )

CB φs + KBoiφo

(8)

where φw and φo are the volume fractions of water and oil, respectively, calculated on the total volume. Substitution of (7) and (8) in (5) leads to

k ) ki

φs (φs + φwKAwi)(φs + φoKBoi)

(9)

Previous work has shown that the concentration gradient in the water phase of a similar system is small.31 Assuming no concentration gradient for the reactants between bulk and interfacial zone, i.e., that KAwi and KBoi are equal to 1, gives

k ) ki

φs (φs + φw)(φs + φo)

(10)

Plotting k(φs + φw)(φs + φo) against φs results in the curve shown in Figure 12. As expected, the rate constant increases with increasing volume fraction of surfactant. These results are in agreement with an observation made for a similar interfacial reaction model where the total interfacial area was calculated from known values of the headgroup area of the nonionic surfactant.32 The assumption that there is only a small concentration difference for KI between bulk water and interface is supported by recent studies of iodide binding to the interface. 127I NMR investigations, as well as quadruple splitting experiments performed by 2H NMR, indicated certain, but not very (31) Alexey, K.; Olsson, U.; Wennerstro¨m, H. J. Phys. Chem. 1995, 99, 6220. (32) Tjandra, D.; Lade, M.; Wagner, O,; Schoma¨cker, R. Chem. Eng. Technol. 1998, 21, 666.

Figure 12. Relation of the rate constant, k, together with the local concentration of reactants at the interface expressed as (φs + φo)(φs + φw) and the volume fraction of surfactant, φs, for the reaction between 4-TBBB and KI at 23 °C in D2O-C12E5decane microemulsion. 4-TBBB and KI were used in a mole ratio of 1:1.

marked, accumulation of iodide at the oil-water interface of the microemulsion. These results are reported in a separate communication.33 The solubility characteristics of the substrate are important. A negligible solubility in the aqueous phase is a prerequisite for the reaction to occur entirely at the interface. If the substrate, i.e., the benzyl bromide, is soluble in water to an appreciable extent, a bulk reaction in the water domain will accompany the reaction at the interface. To determine the importance of a possible competing bulk reaction, four different benzyl bromides were reacted with KI using either d-MeOH or the oil-inwater microemulsion based on D2O, C12E5, and decane as reaction medium. The relative reaction rates in d-MeOH were the expected ones with the benzyl bromides carrying a larger alkyl group in the 4-position being the most reactive followed by the 4-methyl-substituted benzyl bromide. An alkyl group in the para position will increase the electrophilicity of the benzylic methylene carbon and the inductive effect will be larger for an isopropyl and a (33) Ha¨ger, M.; Currie, F.; Holmberg, K. Colloid Surf., A, in press.

6114

Langmuir, Vol. 20, No. 15, 2004

Ha¨ ger et al.

tert-butyl group than for a methyl group. One may note that the order is opposite the so-called Baker-Nathan order, where a methyl substituent on the aromatic ring brings about a stronger acceleration of a nucleophilic substitution at a benzylic carbon than a tert-butyl or an isopropyl substituent. The Baker-Nathan effect is solventspecific and is believed to be due to differences in solvation of the different alkyl groups.34 Obviously, the BakerNathan effect, i.e., an extra acceleration with a methyl substituent, is not seen in this case. The relative rates were very different when the reactions were carried out in microemulsion. Now the methylsubstituted substrate reacts fastest followed by the unsubstituted benzyl bromide. The tert-butyl- and the isopropyl-substituted substrates again react with the same rate. These results are interpreted as follows. The methyl substituted and, even more, the unsubstituted benzyl bromides have non-negligible water solubility. For these substrates reaction in the bulk water domain occurs in parallel to the interfacial reaction and the observed rate is the sum of the two processes. The tert-butyl- and the isopropyl-substituted benzyl bromides, on the other hand, react only at the interface. The rate for a reaction carried out entirely in the aqueous phase can be written as

rw ) -

dnj 1 1 dnj )) kwCAwCBw (11) Vw dt V(1 - φos) dt

where kw is the rate constant in the aqueous phase and CAw and CBw are the concentrations of KI and 4-TBBB, respectively, in the aqueous domain. For simplicity the microemulsion is in the following divided into two subvolumes instead of three. The surfactant volume is split between the oil and aqueous subvolumes. (It is assumed that the concentration of reactants is the same in the bulk as in the interfacial zone.) The combined volume fraction of oil and half of the surfactant is given by

φos )

Vo +

Vs 2

V

(12)

From (1) and (11) it follows that

kCACB ) kw(1 - φos)CAwCBw

(13)

Since the surfactant is split into the bulk oil and water domains and since the reactant concentrations are assumed to be the same in the bulk and at the interface, the partition coefficient between the bulk domains can be written as

KBow )

CBo CBw

(14)

The concentrations in the aqueous domain can be rewritten as

CAw ) CBw )

CA 1 - φos CB

1 + φos(KBow - 1)

(15)

(16)

Substitution of (15) and (16) in (13) gives (34) Schubert, W. M.; Sweeney, W. A. J. Org. Chem. 1956, 21, 119.

k ) kw

(1 - φos) (1 - φos)(1 + φos(KBow - 1)) kw

)

1 (17) 1 + φos(KBow - 1)

Since the value for KBow can be expected to be .1, the expression is reduced to

k ) kw

1 φosKBow

(18)

Equation 18 indicates that the rate constant is inversely proportional to φos, i.e., to the oil + surfactant volume fraction. The fact that the rate constant for the reaction between KI and 4-TBBB in the microemulsion is almost independent of the amount of added surfactant and that the reaction rate is not much lower in this medium than in d-MeOH indicate that the reaction does not occur in the aqueous bulk phase. This confirms the assumption that the reaction between 4-TBBB and KI can be regarded as an interfacial reaction. To investigate the effect of aggregation size and shape on the reaction rate, the reaction between 4-TBBB and KI was performed in microemulsions with varying ratios of C12E5 to C12E4. Self-diffusion NMR was used to measure the self-diffusion coefficients of the components of the system. Table 1 shows that increasing the relative amount of C12E4 leads to a decrease of the observed diffusion coefficient of both surfactant and oil, which indicates an increase of the hydrodynamic radius. Since surfactant and oil diffuse with the same diffusion rate, one can confirm that the aggregates are discrete for the systems studied. However, the results from the kinetics experiments clearly showed that the rates were independent of the microstructure as long as the composition remained within the L1 domain. When C12E6 was used as surfactant, a Winsor I system, i.e., an oil-in-water microemulsion in equilibrium with oil, was obtained. The reaction was equally fast in this two-phase system as in a one-phase microemulsion. It was also shown that it was not necessary to stir the Winsor system. The rate with and without stirring was the same. The observation that a Winsor system is just as effective a reaction medium as a one-phase microemulsion has been observed before, for another nucleophilic substitution reaction.35 The fact that a Winsor I system can be used instead of a one-phase microemulsion is practically important. Formulation of a one-phase microemulsion is often a problem, particularly when one wants a high loading of reactants into the oil and water domains, and one may end up with various types of twoor three-phase systems. Evidently, such systems may be just as useful as reaction media, as long as one of the phases is a microemulsion. The excess phase (or phases) can be regarded as reservoirs for the reactant (or reactants) while the reaction occurs at the oil-water interface of the microemulsion phase. The reaction between 4-TBBB and KI was also performed in a micellar system and in a lamellar liquid crystalline phase. Compared to the reaction in microemulsion, the rate was higher in the micellar system and lower in the liquid crystalline phase. The higher rate in the micellar system, which was based on C12E5, can be explained by the fact that all the substrate is situated in the interfacial region. (There is no hydrophobic core in (35) Bode, G.; Lade, M.; Schoma¨cker, R. Chem. Eng. Technol. 2000, 23, 405.

Nucleophilic Substitution Reactions in Microemulsions

these micelles.) The higher dynamics of a micelle than of a microemulsion droplet may also be advantageous. An important drawback of using a micellar system is that micelles have a very limited solubilization capacity. Such reactions will have to be made with very low concentrations of reactants. The lower rate of reaction in the lamellar liquid crystalline phase, which was based on a 1:1 mixture of C12E5 and C12E4, may be due to slower dynamics of such a system compared to a microemulsion based on the same (or very similar) surfactant. It is interesting that very different rate constants were obtained when the reaction between 4-TBBB and KI was carried out in the very similar oil-in-water microemulsions D2O-C12E8-octanol-octane and D2O-C8G1-octanoloctane. The sugar-based surfactant is obviously much less suitable for the purpose. This difference may be explained by a difference in dielectrical constant in the interfacial zone. Micelles of sugar-based surfactants have been shown to have a higher dielectrical constant than micelles of alcohol ethoxylates.36 It is likely that the same holds true for interfaces of microemulsions. As discussed above, this type of SN2 reaction proceeds slower the more polar the reaction medium. The more polar environment within the interfacial zone for the microemulsion based on the sugarcontaining surfactant than for the microemulsion based on alcohol ethoxylate will then lead to a reduction in reaction rate. An interesting experiment for the future would be to perform an SN1 reaction in these two microemulsions. SN1 reactions are accelerated by polar solvents; thus, such a reaction should proceed faster in the D2O-C8G1-octanol-octane system than in the D2OC12E8-octanol-octane microemulsion. The polarity of the interfacial zone is believed to be decisive of the reaction rate, as explained above. Another factor that might contribute to the observed differences in reactivity between different systems is that of the dynamics of the system. For all types of surfactant aggregates there is a rapid exchange of surfactants between the aggregate and the bulk. As a general rule, such an exchange is faster the smaller the aggregate. The observed reaction order, i.e., micelles > microemulsion droplet > liquid crystals, corresponds with the trend of exchange rate of the surfactants that make up the aggregate.37 We plan to investigate and compare the dynamics of the microemulsions D2O-C12E8-octanoloctane and D2O-C8G1-octanol-octane. If the alcohol ethoxylate-based microemulsion has a more rapid exchange dynamics, it may be an explanation, complementary to that of polarity of the interfacial zone, to the higher (36) Drummond, C. J.; Grieser, F.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1 1989, 85, 521. (37) Evans, D. F.; Wennerstro¨m, H. The Colloidal Domain, 2nd ed.; Wiley-VCH: New York, 1999.

Langmuir, Vol. 20, No. 15, 2004 6115

rate of reaction in that system than in the microemulsion based on the sugar-containing surfactant. The rate of reaction in a properly formulated microemulsion is fast compared to reaction in a homogeneous protic liquid. As seen in Table 2, the rate constant in EtOH is 0.020 dm3/mol‚s and in MeOH 0.0086 dm3/mol‚s. One may expect that the rate constant would have been considerably smaller in water than in MeOH (if water would have been a possible solvent, which is not the case) because this type of SN2 reaction runs slower the more polar the solvent, as was discussed above. The rate constant in the D2O-C12E5-decane microemulsion with φ ) 0.1 is 0.0041 dm3/mol‚s. However, since the reaction in the microemulsion is assumed to occur only inside the surfactant palisade layer, the interfacial rate constant is a more relevant parameter. The interfacial rate constant in the microemulsion is 0.0071 dm3/mol‚s, as given by the slope of the curve in Figure 12. Hence, the interfacial rate constant in the microemulsion is of the same order as in MeOH but smaller than in EtOH. The relatively large value of the interfacial rate constant for reaction in the microemulsion probably reflects the low water activity inside the surfactant palisade layer. Conclusions By using solvents of different polarity as reaction media, we have shown that the reaction between 4-TBBB and KI is a nucleophilic substitution of SN2 type. The reaction was also performed in microemulsions of different composition, in a micellar system and in a liquid crystalline system. The reaction was fastest in the micellar system, intermediate in rate in the microemulsions, and more sluggish in the liquid crystalline phase. The most important finding was that the reaction ran nearly as fast in a properly formulated microemulsion as in a protic solvent. The reaction rate in microemulsion was found to depend on the type of surfactant used in the formulation. A microemulsion based on an alcohol ethoxylate gave a much faster reaction than a microemulsion based on a sugar surfactant. This is probably due to differences in water activity in the interfacial zone. The sugar-based surfactant will become more hydrated than the alcohol ethoxylates; thus, the dielectrical constant will be higher for the former surfactant. A more polar reaction medium is known to retard this type of SN2 reaction. Acknowledgment. We thank the Swedish Foundation for Strategic Research through its Colloid and Interface Technology Program for financing M.H. M.H. also thanks the Institute for Surface Chemistry AB for financial support. Furthermore, we thank the Swedish NMR Center for granting spectrometer time and Patrik Jarvoll for help with the self-diffusion measurements. LA030434I