Langmuir 2004, 20, 3835-3837
3835
Oxidation of Self-Organized Nonionic Surfactants Fredrik Currie, Martin Andersson, and Krister Holmberg* Chalmers University of Technology, Applied Surface Chemistry, SE-412 96 Go¨ teborg, Sweden Received January 5, 2004. In Final Form: February 23, 2004 Nonionic surfactants containing a polyoxyethylene headgroup are known to slowly undergo oxidative degradation when exposed to air. The oxidation, which starts by abstraction of a hydrogen atom from a methylene group in R-position to an ether oxygen, is accelerated by metal ions. Silver ion mediated oxidation of a technical grade surfactant of this type, Brij 30, was investigated in two types of self-assembled systems, a water-in-oil microemulsion and a liquid crystalline phase. It was found that in both systems the longer homologues, i.e., the surfactant homologues that carry a longer polyoxyethylene chain, decompose faster than the shorter homologues. This trend was found to be more pronounced when the surfactant is present in a liquid crystal than in a microemulsion. The difference is explained in terms of differences in accessibility of the polyoxyethylene chains to the silver ions.
Introduction It is well-known that alcohol ethoxylates are susceptible to oxidation when exposed to air.1-5 This is not surprising since such species are polyethers and ethers are known to undergo autoxidation via formation of hydroperoxides at the methylene groups adjacent to the ether bond.6 The process leads to chain cleavage with different types of aldehydes being the main scission products. The oxidation of fatty alcohol ethoxylates is normally a slow process, and it is only after long storage time that the oxidation becomes a technical problem, i.e., causes a change in the physical-chemical properties of a formulation. The oxidative breakdown is believed to primarily occur from the hydroxyl end of the surfactant molecule, which means that the chain scission that occurs after hydroperoxide formation results in a gradual shortening of the polyoxyethylene chain. This, in turn, means that the surfactant will become more hydrophobic upon storage. The effect of the autoxidation on the cloud point development has been studied in some detail. It was found, for instance, that the cloud points of 1% aqueous solutions of hexa(ethylene glycol) monododecyl ether (C12E6) and penta(ethylene glycol) monododecyl ether (C12E5) dropped from 60 to 31 °C and from 32 to 15 °C, respectively, on standing in closed bottles at 40 °C for 170 days.7 The gradual change in physical-chemical properties of solutions of surfactants containing polyoxyethylene chains may cause formulation problems. However, the dermatological impact of the oxidative degradation is probably a more severe concern. It has been found that the surface active aldehydes formed are more biologically active than * Corresponding author. Tel: +46-31-772-2969. Fax: +46-31160062. E-mail:
[email protected]. (1) Donbrow, M. In Nonionic Surfactants: Physical Chemistry; Surfactant Science Series 23; Schick, M. J., Ed.; Marcel Dekker: New York, 1987; p 1011. (2) Holmberg, K.; Jo¨nsson, B.; Kronberg, B.; Lindman, B. Surfactants and Polymers in Aqueous Solution, 2nd ed.; Marcel Dekker: Chichester, 2003; p 25. (3) Porter, M. R. Handbook of Surfactants, 2nd ed.; Chapman & Hall: Glascow, 1994; p 100. (4) Talmage, S. S. Environmental and Human Safety of Major Surfactants: Alcohol Etoxylates and Alkylphenol Ethoxylates; CRC Press: Boca Raton, FL, 1994; p 35. (5) Bodin, A.; Linnerborg, M.; Nilsson, J. L. G.; Karlberg, A.-T. Chem. Res. Toxicol. 2003, 16, 575. (6) March, J. Advanced Organic Chemistry, 3rd ed.; Wiley-Interscience: New York, 1985: p 633. (7) Blute, I.; Svensson, M.; Holmberg, K.; Bergh, M.; Karlberg, A.-T. Colloids Surf., A 1999, 150, 105.
Figure 1. Mechanism for oxidative degradation of fatty alcohol ethoxylates.
the starting surfactant. For instance, the C12E5-aldehyde, i.e., C12E5 with the terminal hydroxymethyl group replaced by an aldehyde group (C12H25(OCH2CH2)4OCH2CHO), shows contact allergenic activity, which the parent surfactant does not do. The autoxidation also leads to the formation of a large number of low molecular weight aldehydes, hydroxyaldehydes, and hydroperoxides, all of which may have a non-negligible skin toxicity and allergenic activity.5,8,9 Thus, autoxidation of this class of nonionic surfactants is a practical problem in applications that involve exposure to the skin. The initial step in the oxidation of the nonionic surfactant is the formation of a free radical in R-position to the ether oxygen. This radical is formed by hydrogen abstraction, which in autoxidation is brought about by oxygen but which may also be mediated by easily oxidized metal ions. It is likely that for nonionic surfactants used in technical formulations minor amounts of transition metal ions are important in accelerating the oxidation process. Also noble metal salts speed up the process. It has recently been found that metallic silver in the form of nanoparticles is slowly formed when a silver salt is dissolved together with a fatty alcohol ethoxylate. A concomitant formation of aldehyde groups, identified by 1 H NMR, supported the view that the surfactant acted as the reducing agent.10-12 The reaction is shown in Figure 1. The surfactant-mediated reduction of silver ions to silver has been taken advantage of in making silver nanofibers using surfactant liquid crystals with a reversed hexagonal geometry as template.12 (8) Bergh, M.; Shao, L. P.; Hagelthorn, G.; Ga¨fvert, E.; Nilsson, J. L. G.; Karlberg, A.-T. J. Pharm. Sci. 1998, 87, 276. (9) Bodin. A.; Shao, L. P.; Nilsson, J. L. G.; Karlberg, A.-T. Contact Dermatitis 2001, 44, 207. (10) Andersson, M.; Palmqvist, A. E. C.; Pedersen, J. S. To be submitted. (11) Andersson, M.; Ha¨relind Ingelsten, H.; Palmqvist, A.; Skoglundh, M.; Holmberg, K. In Self-Assembled Systems; Robinson, B. H., Ed.; IOS: Oxford, 2003; p 44.
10.1021/la0499665 CCC: $27.50 © 2004 American Chemical Society Published on Web 04/09/2004
3836
Langmuir, Vol. 20, No. 10, 2004
Figure 2. Homologue distribution of Brij 30.
Letters
Figure 3. Change in relative amounts of homologues during reaction in microemulsion.
Surfactants are almost always present in self-assembled structures, and an ordered alignment of the surfactant molecules may affect the mode with which the metalcatalyzed oxidation occurs. If the surfactant molecules are aligned in a tightly packed palisade layer, the metal ion may have difficulties in penetrating deeply into this region. This would mean that hydrogen abstraction, leading to the R-oxygen stabilized radical, will not readily occur from the methylene groups in the interior of the polyoxyethylene chain. If that is the case, oxidation at the terminal oxyethylene group will be favored. The aim of the present work was to look into this matter. Experimental Section The surfactant used was Brij 30 from Aldrich. Deuterium oxide was from Glaser AG and had a purity of 99.8%. Methanol, of HPLC grade, was received from Riedel-de-Hae¨n. Silver nitrate of 99.995% purity and heptane of 99% purity were from Aldrich. All chemicals were used as received. All microemulsion reactions were performed at 20 °C. To monitor the reaction and analyze the product mixture, a Varian 3400 GC and a Hewlett-Packard 3396 series II integrator were used. The column was 25 m long, 0.2 mm in inner diameter, and filled with cross-linked methyl silicon. Silver nitrate (0.004 g) was dissolved in deuterium oxide (1.996 g). To an 8 mL glass vial Brij 30 (0.80 g, 20 wt %), heptane (3.00 g, 75 wt %) and the silver nitrate solution (0.20 g, 5 wt %) were added. The water-in-oil microemulsion was formed instantly. After the sample was mixed, the glass vial was immediately protected from light by alumina foil. One microliter of the mixture was injected into the gas chromatography (GC) column at different times during approximately 20 h. All liquid crystalline phase reactions were performed at 20 °C. For monitoring and analysis, the same instrument was used as described for the microemulsion reactions. Brij 30 (2.40 g, 60 wt %) and the silver nitrate solution described above (1.60 g, 40 wt %) were mixed in an 8 mL glass vial. The glass vial was immediately wrapped in alumina foil to protect it from light. With a spatula, samples were transferred from this vial into another vial where they were dissolved in 1 mL of methanol. A sample of 1 µL was immediately injected into the GC column. The reaction was monitored during approximately 20 h.
Results and Discussion Figure 2 shows the homologue distribution of the nonionic surfactant used, Brij 30, which is a technical grade C12E4. The curve has the normal appearance.13 If all hydroxyl groups, i.e., those of the starting alcohol and of the glycol ethers formed, would have had the same reactivity, a Poisson distribution of oligomers would have been obtained. The right-hand side of the curve follows (12) Andersson, M.; Alfredsson, V.; Kjellin, P.; Palmqvist, A. E. C. Nano Lett. 2002, 2, 1403. (13) Holmberg, K.; Jo¨nsson, B.; Kronberg, B.; Lindman, B. Surfactants and Polymers in Aqueous Solution, 2nd ed.; Marcel Dekker: Chichester, 2003; p 17.
Figure 4. Change in relative amounts of homologues during reaction in liquid crystalline phase.
the Poisson distribution. Since the starting alcohol is slightly less acidic than the glycol ethers, its deprotonation is disfavored, leading to a lower probability for reaction with ethylene oxide.14 This is the reason for the high percentage of unreacted fatty alcohol in the commercial product. Two self-organized systems were investigated, one water-in-oil microemulsion and one lamellar liquid crystalline phase. The former was a pseudotertiary system composed of Brij 30, heptane, and an aqueous solution of silver nitrate. The second system contained only surfactant and the aqueous silver nitrate solution. A large molar excess of surfactant to silver ions was used in both systems. The reasons for picking these two systems were that it is well-known that microemulsions and liquid crystalline phases differ in dynamics, i.e., in the residence time of a surfactant at the interface. Both systems are commonly used in various applications, which means that it is a practically relevant comparison. Figure 3 shows the change in the relative amounts of the different homologues with time for the microemulsion. As can be seen, the relative amounts of longer homologues decrease and those of shorter homologues increase. This indicates that in the microemulsion the silver ion mediated oxidation preferentially occurs with the longer homologues. Figure 4 illustrates the change in the relative amounts of the homologues with time when the oxidation is performed in the liquid crystalline phase. The figure shows the same general appearance as for the microemulsion, but the trend is more pronounced. The steeper slope of the curves, positive for the shorter homologues and negative for the longer ones, indicates that the difference in reactivity between the homologues is larger when oxidation takes place in the liquid crystalline phase than in the microemulsion. (14) Weibull, B.; Nycander, B. Acta Chem. Scand. 1954, 8, 847.
Letters
The results from the two sets of experiments tell us two things. First, silver ion mediated degradation of the nonionic surfactant occurs more readily with the longer homologues than with the shorter ones. Second, the longer ethoxylates are more susceptible to degradationsin relative termssin the liquid crystalline phase than in the microemulsion. The observation that the longer homologues are more susceptible to oxidation than their shorter counterparts is expected. First of all, they contain more methylene groups adjacent to ether oxygens; thus, simple statistics will make them more likely candidates for hydrogen abstraction by the metal ion. In addition, the terminal oxyethylene groups of the longer homologues protrude into the aqueous domain where they are likely to be particularly accessible for attack by the silver ions. The finding that the difference in reactivity between the longer and the shorter homologues is more accentuated in the liquid crystalline phase than in the microemulsion is not a priori expected. In both systems the surfactants align with their polyoxyethylene chains pointing toward the water domain where the metal ions are located. In the microemulsion the surfactants form a monolayer with the hydrophobic tail exposed to the hydrocarbon subvolume. In the liquid crystalline phase the surfactants align in bilayers with the hydrophobic tails facing each other. We propose that the smaller difference in reactivity between longer and shorter homologues seen in the microemulsion is due to the fact that the microemulsion (but not the liquid crystal) is a highly dynamic system.15 The interface of a microemulsion constantly disintegrates and re-forms, and a single surfactant will therefore be much less protected by its neighbors than in the case of a liquid crystalline phase. Thus, protrusion of the terminal oxyethylene groups of the longer homologues into the aqueous domain will not be so important since there are ample opportunities for the silver ions to interact with the entire oxyethylene chain at the instances when the surfactant is not an integral part of an aligned palisade layer. Still, (15) Wormuth, K.; Lade, O.; Lade, M.; Schoma¨cker, R. In Handbook of Applied Surface and Colloid Chemistry; Holmberg, K., Ed.; Wiley: Chichester, 2002; Vol. 2, p 55.
Langmuir, Vol. 20, No. 10, 2004 3837
statistics will favor the longer homologues because these contain more oxyethylene groups, which may account for the fact that also in the microemulsion case there is a gradual increase in ratio between shorter and longer homologues. One must be aware of the fact that the method used does not account for all oxidative degradation processes of alcohol ethoxylates. For instance, abstraction of a hydrogen from the R-methylene group on the hydrocarbon side of the ether oxygen originating from the fatty alcohol is known to be an important event in the oxidative breakdown and the corresponding hydroperoxide has been identified and is believed to be formed in relatively large amounts.5 This degradation route, which is likely to occur also in the systems used in this work, will reduce the amount of all homologues but will not affect the ratio between them. In this work dideuterium oxide was used instead of water to form the desired surfactant phases. Going from water to dideuterium oxide means minor changes in the physical-chemical properties of the systems, e.g., a small shift in cloud point. Since the work was based on NMR, which called for the use of D2O, all experiments were done with this solvent. There is no doubt that the results obtained are relevant also for surfactants in H2O. Conclusions Brij 30 undergoes oxidative degradation in the presence of a silver salt. The longer homologues, i.e., molecules with longer polyoxyethylene chains, are degraded more rapidly than the shorter homologues. This difference is more pronounced when the surfactant has self-assembled into a lyotropic liquid crystal than when it is in the form of a microemulsion. Acknowledgment. We acknowledge the Swedish Foundation for Strategic Research through its Colloid and Interface Technology Program for financial support. We are also grateful to Mr. Hans Oskarsson for help with the GC experiments. LA0499665
