Preparation of Organic Nanoparticles Using Microemulsions: Their

Jan 5, 2007 - For example, galenic pharmacy, a branch of pharmacy, includes the study of liposoluble drugs for transdermal use (in a form of self-stic...
0 downloads 0 Views 599KB Size
Download PDF
Langmuir 2007, 23, 1965-1973

1965

Preparation of Organic Nanoparticles Using Microemulsions: Their Potential Use in Transdermal Delivery C. Destre´e,*,† J. Ghijsen,‡ and J. B.Nagy† Laboratoire de RMN, Faculte´ s UniVersitaires Notre-Dame de la Paix, and LISE, Faculte´ s UniVersitaires Notre-Dame de la Paix, 61 Rue de Bruxelles, B-5000 Namur, Belgium ReceiVed December 23, 2005. In Final Form: September 7, 2006 Organic nanoparticles of cholesterol and retinol have been synthesized in various AOT (Aerosol OT; sodium bis(2-ethylhexyl) sulfosuccinate)/heptane/water microemulsions by direct precipitation of the active principle in the aqueous cores. The nanoparticles are observed by transmission electron microscopy (TEM) using the adsorption of a contrasting agent, such as iodine vapor. The size of the nanoparticles can be influenced, in principle, by the concentration of the organic molecules and the diameter of the water cores, which is related to the ratio R ) [H2O]/[surfactant]. The particles remain stable for several months. The average diameter of the cholesterol nanoparticles varies between 3.0 and 7.0 nm, while that of retinol varies between 4.0 and 10 nm. The average size of the cholesterol nanoparticles does not change much either as a function of the ratio R or as a function of the concentration of cholesterol. The constant size of the nanoparticles can be explained by the thermodynamic stabilization of a preferential size of the particles. Chloroform is used to carry the active principle into the aqueous cores. Retinol molecules form J-complexes composed of two or three molecules, as detected by UV-visible spectroscopy.

1. Introduction The pharmaceutical response of an organism to a drug is linked directly to the drug concentration at the site where it is supposed to act efficiently. Therefore, large amounts of drugs have to be administrated to the patient in order to reach the desired level because the drug is distributed through the whole body in relation with its physicochemical properties. As a consequence, undesirable side effects also appear due to the high amounts of drugs reaching healthy tissues and organs. One possible solution to circumvent these difficulties is the vectorization of the drug or active principle. Instead of introducing the drug in a pure form, it will be associated to a vector that will carry it to the specific target.1 This way, liposomes, nanospheres, and nanocapsules have already been used to carry the active principle to the organs and tissues and monoclonal antibodies and glycoproteins have been used for transport to the cells. The nanospheres and nanocapsules are prepared by polymerizing the dispersed monomers and/or by dispersing the preformed polymers.2 The vectors of the first generation are pharmaceutical systems that are to be introduced, in an intravascular manner, close to the target organ. The microparticles containing the antitumoral agent are administered by intra-arterial catheterization. Two types of microparticles are prepared: the microcapsules formed by a polymeric wall, delimitating a reservoir containing the drug, and the microspheres formed by a matrix where the antitumoral agent is dispersed.2 The vectors of the second generation have sizes smaller than 1 µm. The vectorization can be passive or active. In the first case, the drug released is acting following its physicochemical properties. In active vectorization, the distribution of the drug is controlled by extra corporal means (e.g., magnetization). The colloidal vectors have to satisfy three objectives: (a) protect against chemical, enzymatic, or immunological inactivation; (b) * To whom correspondence [email protected]. † Laboratoire de RMN. ‡ LISE.

should

be

addressed.

(1) Birrenbach, G.; Speiser, P. P. J. Pharm. Sci. 1976, 65, 1763.

E-mail:

improve the transport of the drug and its penetration into the target cells; and (c) increase the specific action of the drug by modifying its concentration at the target cells.2 The principal passive vectors are the liposomes3 formed by phospholipidic concentric layers. The liposomes and the polymeric vectors may contain either hydrophilic or hydrophobic molecules. Their penetration is easy, and finally, they are phagocytosed by the liver. Their drawback is, however, due to their in vivo instability. Polymeric nanoparticles have been used since 1976.1 The nanoparticles are either nanocapsules or nanospheres4 that are identical to the microparticles except for their size. The nanospheres are prepared by the direct polymerization of a monomer in an emulsion or in inverse micelles or dissolved in a nonsolvent for the polymer (water and surfactant). This way, nanospheres of polyacrylamide,1,5 polyalkylmethacrylate,6 and polyalkylcyanoacrylate3,4 were prepared. It is also possible to disperse the preformed polymer to obtain the nanospheres. Generally, they are prepared from natural macromolecules (e.g., albumin and gelatin).7,8 The nanocapsules are prepared by methods similar to those of the nanospheres, either by direct polymerization5,9 or by dispersing the preformed polymer, such as the poly(D,L-lactide).10 The active vectorization includes an exterior intervention on the colloidal system. That way, vesicles having a phase transition temperature (Tc) higher than the physiological temperature were used. By increasing the local temperature up to Tc (i.e., local (2) Couvreur, P.; Dubernet, C.; Puisieux, F. Eur. J. Pharm. Biopharm. 1995, 41, 2. (3) Benoit, J. P.; Couvreur, P.; Devissaguet, J. P.; Fessi, H.; Puisieux, F.; Roblot-Treupelle, L. J. Pharm. Belg. 1986, 41, 319. (4) Couvreur, P.; Kante, B.; Roland, M.; Guiot, P.; Bauduin, P.; Speiser, P. J. Pharm. Pharmacol. 1979, 31, 331. (5) Alle´mann, E.; Gurny, R.; Doelker, E. Eur. J. Pharm. Biopharm. 1993, 39, 173. (6) Kreuter, J. Pharm. Acta HelV. 1978, 53, 2. (7) Marty, J. J.; Oppenheim, R. C.; Speiser, P. Pharm. Acta HelV. 1978, 53, 17. (8) Oppenheim, R. C. Int. J. Pharm. 1981, 8, 217. (9) Khouri, N.; Fessi, H.; Roblot-Treupelle, L.; Devissaguet, J. P.; Puisieux, F. Pharm. Acta HelV. 1986, 61, 274. (10) Fessi, H.; Puisieux, F.; Devissaguet, J. P.; Ammoury, N.; Benita, S. Int. J. Pharm. 1989, 55, 1.

10.1021/la0534726 CCC: $37.00 © 2007 American Chemical Society Published on Web 01/05/2007

1966 Langmuir, Vol. 23, No. 4, 2007

hyperthermy), the drug could be liberated at the target cells.11 Magnetite (Fe3O4) nanoparticles were adsorbed, together with the dotorubicilline drug, into nanospheres. These nanospheres could then be driven by an external magnetic field to the target cells.5,12 The vectors of the third generation include, in addition, a specific function of recognition of the target cells.3 The general method of precipitation of organic molecules was examined recently.13 Special batch precipitations were applied for pharmaceutical preparations.14 A particularly simple solvent process was invented to produce microparticles of probucol stabilized by sodium dodecyl sulfate and polyvinylpyrrolidone.15 Cholesterol, one of the drugs used in this study, is a lipid that stems directly from nutrients rich in cholesterol (e.g., of animal origin) or is synthesized by the liver. It takes part in the composition of the cell membrane, and it is necessary for the synthesis of various hormones and of vitamin D. It is also part of the bile salt indispensable for the digestion of the lipids. Retinol (or vitamin A) has hydrating properties, cicatrizing capacity, protecting ability against UV radiation, healing properties (e.g., measles), and regulation activities (e.g., sebaceous and sudoripar glands). It also reduces acne, inhibiting the formation of comedo. In addition, retinol is able to transform precancerous cells into normal ones. Currently, many pharmaceutically active principles are delivered in the cutaneous manner. For example, galenic pharmacy, a branch of pharmacy, includes the study of liposoluble drugs for transdermal use (in a form of self-sticking patch). In that respect, the preparation of nanoparticles will facilitate the penetration of the drugs through the stratum corneum. To facilitate the transport of the drugs, they can be prepared in the form of nanoparticles. An original method consists of preparing them by using microemulsions. Indeed, inorganic nanoparticles have been prepared since the first propositions made by S. Friberg and the late F. Gault. These ideas were followed by a rapid increase in original research works related to the preparation of metal and metal boride nanoparticles.16-22 The precursor metallic salts were reduced using NaBH4, N2H4, H2, and solvated electrons.22 Halides, sulfides, selenides, and tellurides were prepared by precipitating the corresponding metal salts.21,23 Very recently, J-type aggregates of the pseudoisocyanines and other photographic sensitizing agents could be deposited onto nanosized silver halide particles.24 In this paper, the preparation of organic nanoparticles of cholesterol and retinol (vitamin A) is described as obtained by (11) See, e.g.: O ¨ zer, A. Y.; Farivar, M.; Hincal, A. A. Eur. J. Pharm. Biopharm. 1993, 39, 97. (12) Widder, K.; Flouret, G.; Senyei, A. J. Pharm. Sci. 1979, 68, 79. (13) Texter, J. In Reaction and Synthesis in Surfactant Systems; Texter, J., Ed.; Surfactant Science Series; Marcel Dekker: New York, 2001; Vol. 100, p 577. (14) Violante, M. R.; Fisher, H. W. U.S. Patent 4,997,454, March 5, 1991. (15) Frank, S.; Lo¨froth, J. E.; Bostanian, L. U.S. Patent 5,780,062, July, 14, 1998. (16) Boutonnet, M.; Kizling, J.; Stenius, P.; Maire, G. Colloids Surf. 1982, 5, 209. (17) B.Nagy, J.; Gourgue, A.; Derouane, E. G. Stud. Surf. Sci. Catal. 1983, 16, 193. (18) B.Nagy, J.; Derouane, E. G.; Gourgue, A.; Lufimpadio, N.; Ravet, I.; Verfaillie, J.-P. In Surfactants in Solution, Modern Aspects; Mittal, K. L., Ed.; Plenum Press: New York, 1989; Vol. 10, p 1. (19) B.Nagy, J.; Claerbout, A. In Surfactants in Solution; Mittal, K. L., Shah, D. O., Eds.; Plenum Press: New York, 1991; Vol. 11, p 363. (20) Fendler, J. H. Chem. ReV. 1987, 87, 877. (21) B.Nagy, J.; Barette, D.; Fonseca, A.; Jeunieau, L.; Monnoyer, Ph.; Piedigrosso, P.; Ravet-Bodart, I.; Verfaillie, J.-P.; Wathelet, A. In Nanoparticles in Solids and Solutions; Fendler, J. H., Dekany, I., Eds.; Kluwer: Dordrecht, 1996; p 71. (22) B.Nagy, J. In Handbook of Microemulsion Science and Technology; Kumar, P., Mittal, K. L., Eds.; Marcel Dekker: New York, 1999; p 499. (23) Monnoyer, Ph.; Fonseca, A.; B.Nagy, J. Colloids Surf., A 1995, 100, 233. (24) Jeunieau, L.; B.Nagy, J. Colloids Surf., A 1999, 151, 419.

Destre´ e et al.

direct precipitation of these molecules in AOT/heptane/water microemulsions (AOT ) sodium bis(2-ethylhexyl) sulfosuccinate). The aim of this work is to prepare nanoparticles of pharmaceutical interest. Nanoparticles are used nowadays in some formulations for their capacity for drug vectorization. Using nanoparticles of drugs has double interests; their use could reduce the amount of drug delivered and, hence, reduce its secondary effects. 2. Experimental Section 2.1. Description of the Microemulsion. A water-in-oil microemulsion is a thermodynamically stable, isotropic, and optically transparent dispersion of two immiscible liquids stabilized by a surfactant. The important properties are governed essentially by the water-to-surfactant molar ratio R ) [H2O]/[surfactant]. This factor is linearly correlated with the size of the water droplets. The AOT/heptane/water ternary diagram is illustrated in Figure 1.25,26 It can be seen that the microemulsion exists in a large domain (L2 in Figure 1). AOT (or sodium bis(2-ethylhexyl) sulfosuccinate) is an anionic surfactant. 2.2. Materials. The two molecules cholesterol and retinol are represented in Figure 1. Their origins, together with those of all the substances studied, are as follows: cholesterol (95%, Sigma-Aldrich); retinol (95%, Sigma Aldrich); heptane (99%, HPLC grade, SigmaAldrich); AOT (Sigma-Aldrich); MilliQ water (18.2 mΩ/cm); chloroform (99%, stabilized with 0.75% ethanol, Jansen Chemica); iodine (resublimated, Riedel-de-Hae¨n); silica gel (60, HPLC grade, Merck). 2.3. General Method for Organic Nanoparticle Preparation. The synthesis of organic nanoparticles consists of precipitating them directly in the water cores of the microemulsion.27 All of the syntheses are performed at room temperature. The experimental procedure is as follows: First, dissolve AOT in heptane (0.12M) and then add the desired amount of water following the R ) [H2O]/[surfactant] ratio. After the addition of water, treat the mixture under ultrasound until a limpid solution is obtained and then add the solution of the active principle (cholesterol or retinol) in chloroform drop-by-drop or by using a syringe. Finally, treat the system under ultrasound for 15 min. Treatments with ultrasound are used to give the necessary energy to the system to speed up the formation of the micelles. 2.4. Revelation of the Organic Nanoparticles. The organic nanoparticles prepared in the microemulsion are deposited on a metallic grid (3 mm diameter) that has a rhodium face and a copper face. The rhodium face is covered by a vinyl polymer, Formvar, to facilitate the adhesion of the particles on the grid. Iodine was essentially used as a contrasting agent. Iodine, adsorbed on silica, is placed in a closed system, and the grids are put into HPLC tubes to allow iodine to be deposited by sublimation. The TEM (transmission electron microscopy) pictures were taken on a Philips Tecnaı¨ 10 electron microscope. Finally, the photographs were analyzed by Adobe Photoshop and Global Lab systems that allowed one to automatically measure the size of the particles and to construct the histograms.

3. Results and Discussion 3.1. Preparation of Organic Nanoparticles. The nanoparticles of cholesterol and retinol synthesized in the AOT/heptane/water microemulsions are illustrated in Figure 2. One can see that the size of the nanoparticles is found in a narrow size distribution. Indeed, Figure 2 also shows the histograms of the nanoparticles, with a narrow size distribution for both cholesterol and retinol. (25) Rouvie`re, J.; Couret, J.-M.; Lindheimer, M.; Dejardin, J.-L.; Marrony, R. J. Chem. Phys. 1979, 76, 289. (26) Cabos, C.; Debord, P. J. Appl. Crystallogr. 1979, 12, 502. (27) Jeunieau, L.; Debuigne, F.; B.Nagy, J. In Reaction and Synthesis in Surfactant Systems; Texter, J., Ed.; Surfactant Science Series; Marcel Dekker: New York, 2001; Vol. 100, p 609.

Organic Nanoparticles from Microemulsions

Langmuir, Vol. 23, No. 4, 2007 1967

Figure 1. Ternary diagram of an AOT/heptane/water system (the large domain L2 shows the existence of the microemulsion) and the molecular structures of cholesterol and retinol.

Figure 2. Photographs of (a) cholesterol and (b) retinol nanoparticles synthesized in AOT/heptane water microemulsions and the corresponding histograms of (a) cholesterol and (b) retinol nanoparticles.

It is also amazing that the size varies only in the ranges of 3.0 and 7.0 nm for cholesterol and 4.0 and 10.0 nm for retinol. Parts a and b of Figure 3 illustrate the variation of the diameter of the particles as a function of R for cholesterol and retinol. For almost all concentrations studied, a minimum in the diameter of the cholesterol particles is observed at R ) 8 (Figure 3a). Conversely, no minimum is observed for the variation of the size as a function of the concentration of cholesterol and retinol (Figure 3c and d). The experimental points of retinol are rather scattered, and no

clear-cut trend can be derived. One should note, however, that, for the variation of the size as a function of R, the total amount of active principle added increases with increasing R, as the volume of the chloroform solution is equal in each case to that of water in the microemulsion. To gain a deeper insight into the mechanism of nanoparticle formation, the number of nanoparticles per water core has been computed (Figure 4) as follows. The volume of the water core is given by

1968 Langmuir, Vol. 23, No. 4, 2007

Destre´ e et al.

4 Vwc ) πr3 3 where r is the radius (in nm) of the water core computed from the correlation28

r ) 0.18R + 0.45 The number of water cores is obtained by dividing the total amount of water (Vt) by the volume of a water core:

Nwc )

Vt Vwc

In this hypothesis, the amount of water in the organic phase and in the surfactant layer is neglected. On the other hand, the volume of a nanoparticle is obtained from the TEM data, Vnp ) 4/3πr3np. If one can suppose that the density of the nanoparticle is equal to that of the macroscopic material (this hypothesis could be, however, questioned), the mass of a single nanoparticle is equal to mnp ) Vnpd, with d ) 1.052 for cholesterol29 and d ) 1.02 for retinol.30 Finally, the number of nanoparticles is given by

Nnp )

mt mnp

where mt is the total mass of the active principle introduced into the microemulsion. In that respect, the very small amount of substances dissolved in water and in the organic phase is neglected. The number of nanoparticles per water core can be calculated, but it is important to note that this number is rather small. In fact, the Nnp/wc values vary from 0.01 to 0.03 for cholesterol and from 0.0003 to 0.19 for retinol. The number of nanoparticles per water core is given by

Nnp/wc )

Nnp Nwc

Figure 4 illustrates the variation of Nnp/wc as a function of R for various concentrations of the active principle. To study the influence of R on Nnp/wc, one has to take into account that the amount of the active principle increases with increasing water content. To eliminate the influence of the concentration of the active principle, the Nnp/wc values have to be normalized to the same concentration of the active principle. This is carried out by multiplying Nnp/wc by the ratio V(R ) 4)/V(R ) x), where V(R ) 4) is the corresponding volume for R ) x. (Note that, for the sake of simplicity, the ratio of the volumes has been dropped in the figures.) The variation of Nnp/wc as a function of R could shed some light on the presence of a minimum in the nanoparticle diameter as a function of R (Figure 3a) for cholesterol. The precipitation of the active principle occurs in the inner water cores; hence, these water droplets can be considered as nanoreactors. For R ) 4, the number of nanoparticles per water core is small (∼0.01); hence, the size of the nanoparticles becomes bigger due to the small number of nuclei formed. As R increases, the number of nuclei formed increases (R ) 8); hence, the size of the nanoparticles decreases, for the same total concentration of the active principle. When R increases further (R ) 12), the number (28) Eastoe, J.; Robinson, B. H.; Visser, A. J. W. G.; Steatler, D. C. J. Chem. Soc., Faraday Trans. 1991, 87, 1899. (29) Merck Index, 12th ed.; Budavari, S., Ed.; Merck Research Laboratories Division of Merck & Co, Inc.: New Jersey, 1996; p 369. (30) Landolt-Bo¨rnstein, New Series; Springer: Berlin, 1985; Vol. 106, p 166.

Figure 3. Variation of the diameter of the nanoparticles as a function of R: (a) cholesterol and (b) retinol. Variation of the diameter of the nanoparticles as a function of concentration: (c) cholesterol and (d) retinol.

of nuclei decreases, resulting in a larger nanoparticle size. For R values higher than 12, no further change occurs (Figure 3a). One further point remains to be elucidated, and this is the influence of the volume of chloroform itself. Indeed, in the abovereported experiments, the volume of chloroform solution is equal to the volume of water. This volume changes, of course, with the R values. To check the influence of the chloroform, experiments have been carried out where the same volume of

Organic Nanoparticles from Microemulsions

Langmuir, Vol. 23, No. 4, 2007 1969

Figure 4. Variation of the number of nanoparticles formed per water core as a function of R for various concentrations of the active principle: (a) cholesterol and (b) retinol.

Figure 5. Variation of the diameter of the cholesterol nanoparticles as a function of R for various concentrations; the volume of added chloroform was maintained as constant (0.3 mL).

chloroform was used in all cases. Figure 5 shows unambiguously that, in these cases, the size of the nanoparticles remains constant regardless of the parameters used. The presence of chloroform around the nanoparticle will be shown by 2H NMR analysis of CDCl3. To rationalize these data, one should analyze the possible mechanisms of nanoparticle formation. 3.2. Mechanism of Nanoparticle Synthesis in Microemulsions. The aqueous droplets continuously collide, coalesce, and break apart, resulting in a continuous exchange of solution content. In fact, the half-life of the exchange reaction between the droplets is of the order of 10-3-10-2 s.31,32 Two models have been proposed to explain the variation of the size of the particles with the precursor concentration and with the size of the aqueous droplets. The first is based on the LaMer diagram33,34 which has been proposed to explain the precipitation in an aqueous medium and, thus, is not specific to the microemulsion. This diagram (Figure 6) illustrates the variation of the concentration with time during a precipitation reaction and is based on the principle that the nucleation is the limiting step in the precipitation reaction. In the first step, the concentration increases continuously with increasing time. As the concentration reaches the critical supersaturation value, nucleation occurs. This leads to a decrease of the concentration. Between the concentrations C*max and C*min, nucleation occurs. Later, the decrease of the concentration is due to the growth of (31) Fletcher, P. D. I.; Howe, A. M.; Robinson, B. H. J. Chem. Soc., Faraday Trans. 1 1987, 83, 985. (32) Atik, S. S.; Thomas, J. K. Chem. Phys. Lett. 1981, 79, 351. (33) LaMer, V. K.; Dinegarn, R. H. J. Am. Chem. Soc. 1950, 72, 4847. (34) Sugimoto, T. AdV. Colloid Interface Sci. 1987, 28, 65.

Figure 6. LaMer diagram for the precipitation reactions.

the particles by diffusion. This growth occurs until the concentration reaches the solubility value. This model, i.e., that nucleation occurs in the first part of the reaction and later only growth of the particles occurs, has been applied to the microemulsion medium. If this model is followed, the size of the particles will increase continuously with the concentration of the precursor, or a minimum in the variation of the size with the concentration can also be expected. This stems from the fact that the number of nuclei is constant and the increase of concentration leads to an increase in the size of the particles. The second model is based on the thermodynamic stabilization of the particles. In this model, the particles are thermodynamically stabilized by the surfactant. The size of the particles stays constant when the precursor concentration and the size of the aqueous droplets vary. In this case, the nucleation occurs continuously during the formation of the nanoparticles. The thermodynamic stabilization of a certain size of the nanoparticle is nicely demonstrated by the formation of ZrO2 in aqueous media. Primary particles ( crystalline needles) are formed first that coagulate under the influence of carboxylic acids of various lengths, leading to the formation of monodisperse nanoparticles, the size of which is governed by the nature and concentration of the carboxylic acid.35

1970 Langmuir, Vol. 23, No. 4, 2007

Destre´ e et al.

control) or when the nutrient is exhausted (in the case when the LaMer diagram is followed) (5, Figure 7). The nanoparticle is stabilized by a layer of surfactant, the polar headgroups of the surfactant are hydrated, and a part of the vector solvent is also included at the interface (see below). 3.3. Properties of the Nanoparticles. Retinol molecules form a J-type complex by aggregation, because the excitonic band in the UV-visible region of the spectrum is bathochromic with respect to the monomer absorption band (Figure 8). From the variation of the maxima of the absorption band and the bandwidths of the absorption band, the number of monomers per J-complex can be computed.24,36-38 Figure 9 illustrates the variation of ∆νmax ) (νmax)monomer - (νmax)J-complex as a function of the bandwidth at half-height [∆ν1/2(∆max)monomer ) 30.770 cm-1 and (∆ν1/2)monomer ) 6020 cm-1]. The correlation is quite linear, and the bathochromic effect increases with decreasing bandwidth, which is characteristic of the excitonic band. Indeed, the larger the excitonic domain, the faster the excitation transfer from one molecule to the other and the smaller the bandwidth will be. If a bandwidth of ∼2000 cm-1 is estimated for the J-complex, ν∞ can be estimated to be 23.770 cm-1 for a J-complex of infinite length. From the bandwidth at half-height, the number of monomers in the J-complex (N) is computed as

(∆ν1/2)monomer (∆ν1/2)J-complex

) N1/2

The so-obtained values vary between 1.5 and 2.8, meaning that the J-complex is formed between two and three molecules. From the variation of νmax, based on a linear aggregate, the N value can also be computed as Figure 7. Mechanism of the formation of organic nanoparticles in microemulsions.

These two models are limiting models; the LaMer diagram does not take into account the stabilization of the particles by the surfactant, and the thermodynamic stabilization model does not take into account that the nucleation of the particles is more difficult than the growth by diffusion. Obviously, when the particle size passes through a minimum as a function of R, as in the case of cholesterol, the model of LaMer is playing a role; that is, the slow nucleation governs the number of the first nuclei, which grow until the exhaustion of the nutrient. On the other hand, when the size of the nanoparticles remains constant as a function of either R or the initial concentration of the organic molecules, the thermodynamic control is favoring a certain size of the nanoparticles. Finally, a model can explain the formation of organic nanoparticles by direct precipitation in the inner water cores of the microemulsion (Figure 7). The active principle dissolved in the good vector solvent (chloroform) is transported through the organic medium and the interface formed by the surfactants (2, Figure 7 ). The nucleus formed is a molecule of either cholesterol or retinol stabilized at the interface by the interaction with the surfactant molecules (3, Figure 7). The fast exchange (10-3-10-2 s) between the water cores brings other organic molecules, and the nanoparticle is formed (4, Figure 7). The growth can come either from collision between the micelles or from isolated molecules coming through the solution outside the micelles. Finally, the size of the nanoparticle exceeds that of the initial water core and the particle growth stops when a certain size is reached (in the thermodynamic (35) Lerot, L.; Legrand, F.; De Bruycker, P. J. Mater. Sci. 1991, 26, 2353.

cos

[

]

(νmax)monomer - (νmax )J-complex π ) (N + 1) (νmax)monomer - ν∞

where ν∞ is the νmax of a J-complex of infinite length. Note that this method also yields N values varying between two and three molecules. As a conclusion, the nanoparticles of retinol containing between 100 and 1000 molecules are formed by J-aggregates separated by molecules the orientation of which cannot fit the J-complex. Indeed, the more probable structure of the J-complex is a ladder form for the trimer and for the dimer in which the retinol molecules have to maintain the same direction of the dipole moment. In fact, as the excitation wavelength is much larger than the size of the aggregate, only those molecules which are in phase can be excited, leading to a bathochromic shift:

In both aggregates, the stabilization can occur when the monomers are in a shifted position. One point that could be further developed concerns the formation of the first nucleus. It is possible that the first nucleus could be just the J-complex, the presence of which should be detected using fast kinetic methods. 3.4. Characterization of Cholesterol Nanoparticles by 2H NMR. First, the influence of the temperature on the state of the (36) Jeunieau, L. Ph.D. Thesis, University of Namur, 1999. (37) Knapp, E. W. Chem. Phys. 1984, 85, 73. (38) Kobayashi, T., Ed.; J-Aggregates; World Scientific: Singapore, 1996.

Organic Nanoparticles from Microemulsions

Langmuir, Vol. 23, No. 4, 2007 1971

Figure 8. UV-visible spectra of (a) monomeric retinol in heptane (5 × 10-5 M) and (b) a J-type complex obtained in the AOT/heptane/water microemulsion (R ) 12; C ) 20 g/L). References: heptane for the monomer and the microemulsion for the J-complex.

Figure 9. Variation of the difference between the absorption maxima of the monomer and the J-complex, ∆νmax ) (νmax)monomer (νmax)J-complex (in cm-1), with the bandwidth at half-height, ∆ν1/2 (in cm-1).

water in the cores of the empty microemulsions is studied. The study of the water in the microemulsions containing only the solvent and also in the microemulsions containing nanoparticles is carried out. The R factor () [H2O]/[AOT]) of the studied microemulsions is equal to 4 and 1. This factor is related to the size of the aqueous core. The reference is pure deuterated water. The 2H NMR spectra of different microemulsions are represented in Figure 10 for the empty microemulsions (R ) 1; (a)), the microemulsion containing chloroform (b), and the microemulsion with cholesterol nanoparticles (c). Two types of NMR lines are observed at room temperature. The first one is located at -0.86 ppm. This stems from the free molecules of water and from the water molecules bound to the surfactant. Hydrogen bonds characterize these states of water, and hence, the chemical shift occurs at lower fields. For lower R values, the shift corresponds to the bound water molecules

which interact with the surfactant polar group and with the counterion Na+. The second and third lines are due to the water molecules present as monomers or dimers. These are located at higher fields (-3.65 and -4.04 ppm, respectively), because only a small amount of hydrogen bonds are present. The water molecules also interact with the surfactant. These molecules are trapped at the interface. Generally, three kinds of water may exist in a microemulsion medium: “bulk” water in the center of the water core; “bound” water, which interacts with the hydrophilic part of the surfactant molecule; and “trapped” water, which is trapped in the interface in the form of monomers or dimers.39 Bulk water molecules are normally not present for R values below 6-10, where all the water molecules are structured because of their interaction with Na+ counterions and the strong dipole of the AOT polar group.40 In this case, where the ratio R ) [H2O]/[AOT] is equal to 1, only two kinds of water molecules should be expected. Therefore, it is assumed that the two NMR lines observed here correspond to bound water and to trapped water. To check this assumption, the same experiment was done for higher R values. The chemical shift increases with the R value until it reaches approximately that of the pure deuterated water (used as reference), while the line width at half-height decreases with increasing R values (Figure 11). This variation has already been observed40 and is the result of a fast exchange (faster than 2 × 1010 s-1) between the bulk water and the bound water. At low R values, the observed chemical shift comes from the variation of the number of hydrogen bonds in which the water molecules are involved. In fact, the water molecules adsorbed at the interface (or solvating the Na+ ions) form fewer hydrogen bonds, provoking a high-field chemical shift. This decreasing number of hydrogen bonds has previously been shown by Wong et al. using 1H NMR experiments.41 The microemulsions are stable as a function of temperature. When the temperature decreases, the intensity of the line decreases and it broadens until it finally vanishes (Figure 10). The solidification of the water occurs at 253 K in the empty microemulsions and at 243 K in the other two cases, while pure water solidifies at 273 K. These values cannot be directly compared with those obtained in microemulsions with low R values. Indeed, the decrease of the line of free water stems from two factors: the solidification of the water and the increase of the bound water. There is no significant shift of the NMR lines. (39) Jain, T. K.; Varshney, M.; Maitra, A. J. Phys. Chem. 1989, 93, 7409. (40) Hauser, H.; Gaering, G.; Pande, A.; Luisi, P. L. J. Phys. Chem. 1989, 93, 7869. (41) Wong, M.; Thomas, J. K.; Novak, T. J. Am. Chem. Soc. 1977, 99, 4730.

1972 Langmuir, Vol. 23, No. 4, 2007

Destre´ e et al.

Figure 10. 2H NMR spectra of water in microemulsions (R ) 1) [(a) empty; (b) with solvent (chloroform); and (c) with nanoparticles of cholesterol] as a function of temperature.

Figure 11. (a) Variation of the 2H chemical shift of D2O as a function of the R factor. (b) Variation of the line width as a function of the R factor.

The decrease in the temperature of water solidification has been extensively studied in emulsions where the influence of the size of the water pocket as well as the composition and the nature of precipitates was emphasized.42-45 The other two lines of the trapped water do not significantly change: their location (light shift toward low fields) and width stay constant. T1 measurements (longitudinal relaxation time) have also been carried out. T1 is greater for these trapped molecules, indicating a smaller mobility. Indeed, as T1 increases when the temperature decreases for trapped molecules, we can conclude that an increase of T1 corresponds to an increase of correlation time (a smaller mobility). The comparison of the two types of trapped water molecules shows that the monomers are situated at a higher field (-4.04 ppm) and the dimers at a lower field (-3.65 ppm). Indeed, for monomers, the shift is toward higher fields because of the decrease of the hydrogen bonds and the relaxation time is greater because the water molecules are less mobile. The nature of the “free-hydrogen bound water” seems to be similar in microemulsions containing the solvent or the nanoparticles if the R values are low: log(ν1/2) (ν1/2 represents the width at half-height) increases as a function of 1/T. For the trapped water molecules, log(ν1/2) does not significantly change as a function of the temperature. For R ) l, the variation of log(T1) as a function of 1/T is similar in all cases, i.e., for empty (42) Clausse, D. Encyclopedia of Emulsion Technology; Becher, P., Ed.; Marcel Dekker: New York, 1987; Vol. 2, p 77. (43) Clausse, D.; Bouabdillah, D.; Cochet, N.; Luquet, M. P.; Pulvin, S. Pure Appl. Chem. 1991, 63, 1491. (44) Franks, F. In Characterization of Proteins; Franks, F., Ed.; Humana Press: Totowa, NJ, 1998; p 95. (45) Franks, F.; Murase, N. Pure Appl. Chem. 1992, 64, 1667.

microemulsions, for microemulsions containing the solvent, and, finally, for microemulsions with nanoparticles. Indeed, the water molecules are still structured at the interface and are not perturbed by the presence of the solvent (or the particle). From these results, it can be seen that the presence of the nanoparticles or the solvent in the microemulsions does not significantly influence the nature of the water. Hence, the nanoparticles should be located in the organic phase and not in the aqueous cores. The other reason for localizing the nanoparticles in the organic phase is that the diameter of the nanoparticles (5.0 nm) is greater than the aqueous cores size (R ) l; d ) 1.26 nm) and it is, thus, impossible to locate the nanoparticles inside the aqueous cores. On the other hand, some water could be adsorbed on the nanoparticles. Another hypothesis is that the solvent could be in the organic phase or adsorbed on the nanoparticles but not in the aqueous cores. To check the veracity of this hypothesis, deuterated CDCl3 is used in the following experiments. The aim is to determine the location of the chloroform solvent. Different spectra have been compared: pure CDCl3, CDCl3 in heptane, and CDCl3 in heptane/AOT solution (Figure 12). The first line stands for the associated CDCl3 at 2.40 ppm (pure CDCl3) and at 2.66 ppm for CDCl3 in heptane or in heptane/ AOT solution. The two small lines at -3.18 and -3.58 ppm (higher fields), in the case of CDCl3 in heptane and in heptane/ AOT solution, stand for CDCl3 dimers or monomers. In the microemulsions, the lines are situated at the same positions . The attributes of the lines are also based on T1 for CDCl3: the line situated at 2.7 ppm has a T1 equal, e.g., to 251 ms (at T ) 293 K and for R ) 1), and the lines at higher fields have a T1 equal to 682 ms (for low temperatures). These last lines at higher fields are due to monomers or dimers because they are less mobile. These molecules can be situated in the heptane phase, or they can be adsorbed on the exterior of the AOT molecules, i.e., at the interface. The line at lower fields has almost the same chemical shift as those obtained in heptane/AOT solution. Thus, the CDCl3 molecules should solvate the polar head group of AOT and should be in contact with the aqueous cores. A study of the influence of temperature on CDCl3 in microemulsions (empty or with nanoparticles) was also carried out. Figure 12 shows the different spectra. The high-intensity lines (due to associated CDCl3) decrease as a function of temperature in the two cases. In presence of nanoparticles, the intensity is smaller and the line is broader at 233 K, in comparison with those without nanoparticles. The hypothesis is that CDCl3 would be adsorbed on the nanoparticles. Indeed, the width of the lines is inversely related to the T2* (transversal relaxation time). Thus, when the width increases,

Organic Nanoparticles from Microemulsions

Langmuir, Vol. 23, No. 4, 2007 1973

Figure 12. CDCl3 spectra of microemulsions (R ) 1) [(a) with cholesterol nanoparticles and (b) without nanoparticles] as a function of temperature.

Figure 13. General picture of the stabilization of the nanoparticles prepared from the microemulsion.

1H NMR measurements have been performed to characterize a possible reaction between cholesterol and iodine. Cholesterol was dissolved in CDCl3 (∼6.6 × 10-2 M), and the 1H NMR spectrum showed that the ratio of the olefinic protons to the total number of protons is 2.21%, a value quite close to the theoretical one, 2.17%. When solid cholesterol was exposed to iodine for 6 h, the ratio found was 2.18%. This result unambiguously shows that iodine is only physically adsorbed on the nanoparticles and that no chemical reactions (iodine addition) occurred between the cholesterol molecule and the iodine. The interaction between cholesterol and iodine could be of an n f σ* or π f σ* type, taking into account the interaction either with the double bond or with the OH group of cholesterol.

4. Conclusions the relaxation time decreases and the molecules are more frozen. The other two lines at higher fields (due to monomers or dimers) do not significantly change as a function of the temperature. When nanoparticles are present, T1 is greater than it is when the microemulsion is empty, especially for R ) 4. The CDCl3 molecules should be more frozen in the presence of nanoparticles as a consequence of their adsorption. For R ) 1, all molecules are already frozen and there is no significant change for T1. As a conclusion, the chemical shift is independent of the presence of the nanoparticles. Some water is adsorbed on the nanoparticles, in addition to the surfactants. Second, NMR measurements of deuterated chloroform show that it is adsorbed on the nanoparticles. The NMR spectra allow one to conclude that the nanoparticles are essentially stabilized in the organic phase with the solvent and surfactant molecules (water, chloroform, and AOT) adsorbed on them (Figure 13). 3.5. Interaction between the Molecules of the Nanoparticles and the Contrasting Agent. In previous publications,27 we have reported the use of iodine as a contrasting agent for TEM measurements. However, no investigation has been carried out to study the interaction between the molecules composing the nanoparticles and the contrasting agent.

Organic nanoparticles can be prepared by direct precipitation in the water cores that behave as nanoreactors. The formation of quasi-monodisperse nanoparticles stems either from the thermodynamic stabilization of a certain size or from the LaMer diagram where a slow nucleation is followed by a fast particle growth. The final nanoparticles have a size greater than that of the water core. These nanoparticles are stabilized by a layer of surfactant molecules, which also include water of stabilization of the polar headgroups as well as some organic solvent that was used as a vector for the active principle. Retinol molecules are associated in dimers and trimers that form a J-complex, the UV-visible absorption of which is bathochromic with respect to the monomer absorption. Finally, we hope the active principles in nanoparticle form could be much more easily adsorbed by the cutaneous layer and, hence, facilitate the pharmaceutical action. Acknowledgment. C. Destre´e gratefully acknowledges financial support from FRIA, Belgium. LA0534726