Inverse Nonionic Microemulsion Studied by Means of 1H, 13C, and

University of Venice, via Torino 155b, 30170 Venice, Italy. Langmuir , 2009, 25 (5), pp 3224–3231. DOI: 10.1021/la803826c. Publication Date (Web...
1 downloads 4 Views 796KB Size
Download PDF
3224

Langmuir 2009, 25, 3224-3231

Inverse Nonionic Microemulsion Studied by Means of 1H, 13C, and PGSTE NMR during Silica Nanoparticle Synthesis Fioretta Asaro,*,† Alvise Benedetti,‡ Nina Savko,† and Giorgio Pellizer† Department of Chemical Sciences, UniVersity of Trieste, Via L. Giorgieri 1, 34127 Trieste, Italy, and Department of Physical Chemistry and INSTM, Ca’ Foscari UniVersity of Venice, Via Torino 155b, 30170 Venice, Italy ReceiVed NoVember 18, 2008. ReVised Manuscript ReceiVed December 29, 2008 The soluble species present in the reaction mixture that leads to silica nanoparticle production through the base catalyzed hydrolysis of tetraethyl orthosilicate (TEOS) and the successive condensation were investigated in situ, under the actual synthesis conditions, by means of 1H, 13C, and 29Si NMR spectroscopy. The two former nuclei, owing to higher sensitivity and their presence both in the reacting species and in the constituents of the W/O microemulsion (cyclohexane-igepal-CA-520-concentrated ammonia solution) afforded insight into the inverse microemulsion and allowed us to assess the kinetic rate of the hydrolysis step. It was verified that the microemulsion microstructure is maintained during the reaction. The characterization of the final nanoparticles was carried out by means of transmission electron microscopy (TEM). Special attention was paid to the reaction medium, and an extended assignment of the 1 H and 13C resonances of the surfactant headgroup is reported together with the discussion of the changes they undergo due to the environmental modifications induced by transition from cyclohexane solution to W/O microemulsion and further to NH3 containing W/O microemulsion. The self-diffusion coefficient measurements revealed that NH3 exchanges among the inverse micelles diffusing through cyclohexane and confirmed that the preferred localization for ethanol, a byproduct of the reaction, is the bulk oil.

Introduction Nanosized inorganic particles constitute a field of rapidly growing interest, and their tailored synthesis is currently subject of intense study. These particles may show unique properties that are not shared by bulk materials and therefore find a palette of innovative applications. Here, we focused our attention on silica nanoparticles that have many diversified usages, for example, as diagnostic means, whether loaded with suited fluorophores,1-3 or in drug4,5 and even gene delivery,5,6 offering the advantage over polymer nanoparticles of an easy introduction of functional groups such as amines, thiols, and carboxyls by modification of the surface hydroxyls.7 There are basically two synthetic approaches relying on the NH3 catalyzed hydrolysis of silicon alkoxides and polymerization of the hydrolysis products, namely the sol-gel synthesis carried out in low-molecular-weight alcohol and water solution, known as Sto¨ber synthesis, and that carried out in reverse microemulsion, where the compartmentalization due to the reverse micelles is exploited to influence the kinetics of tetraethyl orthosilicate (TEOS) hydrolysis and silica nanoparticle growth.8-10 The * Corresponding author. Telephone: +39 040 558 3951. Fax: +39 040 558 3903. E-mail: [email protected]. † University of Trieste. ‡ Ca’ Foscari University of Venice.

(1) Burns, A.; Ow, H.; Wiesner, U. Chem. Soc. ReV. 2006, 35, 1028–1042. (2) Ye, Z.; Tan, M.; Wang, G.; Yuan, J. J. Mater. Chem. 2004, 14, 851–856. (3) Bringley, J. F.; Penner, T. L.; Wang, R.; Harder, J. F.; Harrison, W. J.; Buonemani, L. J. Colloid Interface Sci. 2008, 320, 132–139. (4) Finnie, S. K.; Barlett, J. R.; Barbe`, C. J. A.; Kong, L. Langmuir 2007, 23, 3017–3024. (5) Barbe´, C.; Bartlett, J.; Kong, L.; Finnie, K.; Lin, H. Q.; Larkin, M.; Calleja, S.; Bush, A.; Calleja, G. AdV. Mater. 2004, 16, 1959–1966. (6) Luo, D.; Saltzman, W. M. Gene Ther. 2006, 13, 585–586. (7) Bagwe, R. P.; Yang, C.; Hilliard, L. R.; Tan, W. Langmuir 2004, 20, 8336–8342. (8) Arriagada, F. J.; Osseo-Asare, K. Colloids Surf. 1990, 50, 321–339. (9) Arriagada, F. J.; Osseo-Asare, K. Colloids Surf., A 1999, 154, 311–326. (10) Arriagada, F. J.; Osseo-Asare, K. J. Colloid Interface Sci. 1999, 211, 210–220.

previous studies, focused on the latter synthetic route, concerned mainly the reactive chemical species. They successfully determined the reaction rate and its dependence on TEOS, ammonia, and surfactant concentrations11,12 and the influence of the surfactant’s molecular structure12 and described how to control the size of the final product through the rationale choice of the hydrocarbon solvent.13 Furthermore small-angle X-ray scattering (SAXS) measurements shed light on the formation of solid nanoparticles.4 Only quite recently, quantitative time-resolved SAXS investigations afforded detailed information about the evolution of the microemulsion as well.14 A necessary precaution before correlating the feature of the product with the microstructure of the microemulsion is to establish the changes the microemulsion undergoes during the reaction. They may be quite drastic, like the transition to a bicontinuous structure and, eventually, the demixing observed in a system prepared with a cationic surfactant, a medium chain alcohol as cosurfactant and hexane, induced by the EtOH produced during the synthesis of silica.15 Here, we report high resolution NMR results and pulsed field gradient stimulated echo (PGSTE) self-diffusion coefficient determination relevant to the soluble species, both to the constituents of the microemulsion and to those taking part to reaction, which, viewed overall, corresponds to

Si(OCH2CH3)4+2H2O f SiO2+4CH3CH2OH We followed the quantitative evolution of TEOS and ethanol (EtOH) and assessed their preferential distribution in the various (11) Chang, C.-L.; Fogler, H. S. AIChE J. 1996, 42, 3153–3163. (12) Chang, C.-L.; Fogler, H. S. Langmuir 1997, 13, 3295–3307. (13) Jin, Y.; Lohstreter, S.; Pierce, D. T.; Parisien, J.; Wu, M.; Hall, C., III; Xiaojun Zhao, J. Chem. Mater. 2008, 20, 4411–4419. (14) Riello, P.; Mattiazzi, M.; Pedersen, J. S.; Benedetti, A. Langmuir 2008, 24, 5225–5228. (15) Venditti, F.; Angelico, R.; Palazzo, G.; Colafemmina, G.; Ceglie, A.; Lopez, F. Langmuir 2007, 23, 10063–10068.

10.1021/la803826c CCC: $40.75  2009 American Chemical Society Published on Web 02/02/2009

NMR Study of an InVerse Nonionic Microemulsion

Langmuir, Vol. 25, No. 5, 2009 3225

environments provided by this, on the mesoscale heterogeneous, reaction medium.

Experimental Section Materials. Igepal CA-520 (5 polyoxyethylene iso-octylphenyl ether) and igepal CO-520 (5 polyoxyethylene nonylphenyl ether), tetraethyl orthosilicate (TEOS) with 98% purity, cyclohexane, and aqueous ammonia (NH3) solution (29.6% wt) were all purchased from Sigma Aldrich and used without purification. Preparation of Samples. Microemulsions were prepared by dissolving the surfactant (igepal CA-520) in cyclohexane at 0.1 M concentration. In a NMR tube, by a microsyringe, was added to a weighted aliquot of the solution the required volume of (i) water, to reach a H2O/surfactant molar ratio R ) 4.4, referred to as “H2O microemulsion”, and (ii) concentrated aqueous NH3 (29.6% wt), to reach R ) 3.0, referred to as “NH3 microemulsion”. The “NH3 microemulsion” system constituted also the reaction medium for silica synthesis. The reaction was initiated by adding the needed TEOS amount to obtain a H2O/TEOS molar ratio h ) 7.5, referred as “reaction mixture”. The dilution of the system is high; therefore, no macroscopic phase separation was detected, not even after complete consumption of TEOS. NMR Measurements. The NMR spectra were recorded on a JEOL Eclipse 400 (9.4 T) NMR spectrometer operating at 399.78 MHz for 1H, 100.53 MHz for 13C, and 79.42 MHz for 29Si. All spectra were acquired at room temperature without field-frequency lock. For 1H NMR, four scans were acquired with 20° pulses, to avoid radiation damping due to the presence of the fully protonated solvent, employing a spectral width of 3.6 kHz over 8K complex points, interleaved by 12.3 s. The 29Si spectra were acquired by insensitive nuclei enhanced by polarization transfer (INEPT),16,17 exploiting the scalar coupling to methylene protons, 3J(29Si, 1H) ) 3.7 Hz,18 with a spectral width of 4 kHz over 8K complex points, accumulating 512 scans with a recycle time of 16 s. For 13C, 3300 scans were acquired employing 45° pulses and a 19.12 kHz spectral width over 32K complex data points. The raw data were zero filled either two or four times prior to Fourier transform (FT), and the 29Si ones were processed as absolute value. All the chemical shifts are referred to tetramethylsilane (TMS). Cyclohexane, which resonates at 1.444 ppm from TMS in the 1H spectrum and at 27.69 ppm in the 13C spectrum, was used as internal reference for 1H and 13C chemical shifts. PulsedFieldGradientStimulatedEcho(PGSTE)Measurements. The 1H NMR diffusion measurements were carried out at 25 °C on a Varian 500 NMR spectrometer (11.74 T) operating at 500 MHz for 1H, equipped with a model L650 Highland Technology pulsedfield gradient amplifier (10 A) and a standard 5 mm indirect detection, pulsed field gradient (PFG) probe. An enhanced stimulated echo pulse sequence with spin lock19 was employed, with 15 different z-gradient strengths, Gz, between 0.5 and 50 G/cm, a pulsed gradient duration, δ, of 2 ms, and a diffusion interval, ∆, chosen in the range 50-100 ms. The gradients were calibrated on the value of D ) 1.90 × 10-9 m2 s-1 for 1H in D2O (99.9%).19 Solvent suppression was accomplished by presaturation. The data were processed as diffusion ordered spectroscopy (DOSY)20 spectra by means of the relevant routine of the Varian VNMRJ software, version 2.2C, which fits the echo intensities21 against q2(∆ - δ/3) on the basis of the Stejskal-Tanner equation

Transmission Electron Microscopy (TEM). TEM images were taken at 300 kV with a Jeol 3010 instrument with an ultrahigh resolution (UHR) pole-piece (0.17 nm point resolution at Scherzer defocus), equipped with a Gatan slow-scan CCD camera (model 794) and an Oxford Instrument EDS microanalysis detector (model 6636). The powder samples were dispersed in isopropyl alcohol by sonication for approximately 3 min, followed by deposition onto a copper grid covered with a holey carbon film. The nanoparticles had been isolated from the reaction mixture 1 week after the reaction beginning. The microemulsion was broken by addition of acetone, the lower phase was collected, the solvent was evaporated, and the solid was washed three times with dichloromethane.

Results 1

where q ) γΗGzδ, with γH being the 1H magnetogyric ratio, A and A0 are signal intensities in the presence and absence of Gz, respectively, and D is the self-diffusion coefficient.

H NMR Spectra. Cyclohexane is a good solvent to the end of NMR studies because it originates just one signal in both 1H and 13C spectra, with either no or very scarce overlap with the signals of the other molecules of the system, and it can be used for referencing purposes. Moreover, it may be eliminated by solvent suppression, actually carried out during the diffusion measurements by 1H NMR. The 1H NMR measurements provide information about the very reaction medium through the surfactant spectrum. The signals of the aromatic protons appeared at the highest frequency end of the spectrum and the protons of the tail at the lowest one, with that of the tert-butyl methyl groups being the tallest and most shielded (Figure 1). However, the most interesting resonances are those of the polyoxyethylene headgroup in the central part of the spectrum, as they are highly sensitive to the phase change. Four of them resonated well apart from one another and from all the other CH2 groups that originated the envelope of “internal” protons, between 3.50 and 3.65 ppm, in cyclohexane (Figure 2). The “internal” protons refer neither to the “first” (see the nomenclature in ref 22) oxyethylene unit bonded to the phenyl ring nor to the “terminal” one, which bears the OH. The assignment, reported in Figure 2, was carried out by means of 1 H-1H correlated spectroscopy (HH COSY) and 1H-13C correlated spectroscopy (HC COSY). It is in agreement with that reported in literature23 for the protons of Triton X-100, a higher homologue of igepal CA-520, having a 9-10 oxyethylene unit headgroup and the same hydrocarbon moiety. It was further confirmed by the high sensitivity to the environment displayed by the methylene groups of the terminal oxyethylene unit (positions ψ and ω). It is interesting to follow the changes in the 1 H spectral patterns of the headgroup on going from a 0.1 M solution in cyclohexane to the microemulsions obtained by adding either water or concentrated ammonia, on the basis of the phase diagrams.10,11 Indeed, upon formation of the microemulsion, the main signal of the polyoxyethylene moiety, due to hydration, shifted to higher frequencies and widened the covered chemical shift range (Figure 2). The effect was more pronounced for water than for concentrated ammonia, with the volumes of the aqueous phases being equal, probably because in the latter case the very high solute concentration affects the hydrogen bonding network of water. A shift to higher frequencies of the signal of the ψ CH2 group was detected, too. Spectral changes occurred also in the alkyl region, namely, a broadening of the signals of the tail accompanied by a slight shift to lower frequencies. On the other hand, during the reaction progress, only few subtle changes of the surfactant signals could be appreciated, in particular a broadening of the signal of the ψ CH2 group.

(16) Morris, G. A.; Freeman, R. J. Am. Chem. Soc. 1979, 101, 760–762. (17) Doddrell, D. M.; Pegg, D. T.; Brooks, W.; Bendall, M. R. J. Am. Chem. Soc. 1981, 103, 727–728. (18) Kinrade, S. D.; Maa, K. J.; Schach, A. S.; Sloan, T. D.; Knight, C. T. G. J. Chem. Soc., Dalton Trans. 1999, 18, 3149–3150. (19) Antalek, B. Concepts Magn. Reson. 2002, 14, 225–258.

(20) Morris, K. F.; Johnson, C. S., Jr J. Am. Chem. Soc. 1993, 115, 4291–4299. (21) Pelta, M. D.; Barjat, H.; Morris, G. A.; Davis, A. L.; Hammond, S. J. Magn. Reson. Chem. 1998, 36, 704–714. (22) Podo, F.; Ray, A.; Nemethy, G. J. Am. Chem. Soc. 1973, 95, 6164–6171. (23) Yuan, H.; Miao, X.; Zhao, S.; Shen, L.; Yu, J.; Du, Y. Magn. Reson. Chem. 2001, 39, 33–37.

[

(

A ) A0 exp -Dq2 ∆ -

δ 3

)]

(1)

3226 Langmuir, Vol. 25, No. 5, 2009

Asaro et al.

Figure 1. 1H NMR spectrum of igepal CA-520 in CDCl3.

Figure 2. Region of the 1H NMR spectrum corresponding to the signals of the exchangeable protons (/) and of the surfactant headgroups, with the relevant assignment, for the samples: (A) 0.1 M solution of igepal CA-520 in cyclohexane, (B) “NH3 microemulsion”, and (C) “H2O microemulsion”.

A very interesting 1H signal, diagnostic of both the system and the reaction advancement, is the broad singlet originated by the exchangeable protons of the NH and OH groups of ammonia, water, surfactant, and EtOH. In the “H2O microemulsion” (Figure 2C), it appeared at 4.5-4.6 ppm, a value close to the one of bulk water, indicating water pools with a well-defined network of hydrogen bonds,24,25 while for the “NH3 microemulsion” its position was intermediate (Figure 2B) and moved to higher frequency upon water addition. During the reaction, it broadened remarkably (Figure 3) due to a new species containing slower exchanging protons. The broadening of the signal, already at early reaction times, was accompanied by a fast height decrease, since its integrated intensity should not change because the protons (24) Hoffmann, M. H.; Conradi, M. S. J. Am. Chem. Soc. 1997, 119, 3811– 3817. (25) Hoffmann, M. M.; Bennett, M. E.; Fox, J. D.; Wyman, D. P. J. Colloid Interface Sci. 2005, 287, 712–716.

of the water molecules consumed by the overall reaction are returned as EtOH alcoholic protons. 1 H is the best suited nucleus to follow TEOS hydrolysis owing to its high sensitivity, which allows the immediate detection of proton containing species, even when present in tiny amounts and at short accumulation times, for example, in the present case, less than 1 min. The protons of the methylene groups of TEOS resonated centered at 3.77 ppm, partially overlapped with the β CH2 signal of the surfactant headgroup, while the quartet of the methylene of the produced EtOH was completely buried in the broad envelope of the “internal” protons (Figure 3). The triplets of the methyl groups of both TEOS and EtOH appeared at lower frequencies than the cyclohexane signal, at about 1.17 and 1.14 ppm, respectively, and the reaction advancement could be easily followed quantitatively through these 1H signals (Figure 4).

NMR Study of an InVerse Nonionic Microemulsion

Langmuir, Vol. 25, No. 5, 2009 3227

Figure 3. Region of the 1H NMR spectrum corresponding to the signals of the exchangeable protons (/) and of the surfactant headgroup protons for the samples: “reaction mixture” (A) 0.5 h and (B) 37 h after the reaction beginning.

Figure 4. 1H NMR signals of the methyl groups of TEOS (at higher frequency, b) and EtOH (at lower frequency, O) from the “reaction mixture”: (A) 0.5 h and (B) 37 h after the reaction beginning. 13 C NMR Spectra. In the 13C NMR spectrum, the CH2 and CH3 carbons of both TEOS and EtOH resonated well apart from each other (the carbons of the CH2 groups at 59.27 and 57.95 ppm and those of the CH3 groups at 18.45 and 18.78 ppm for TEOS and EtOH, respectively) and also from any other peak; therefore, each of them can be exploited efficiently in the kinetic study (Figure 5). The carbons of the igepal headgroup, analogously to the corresponding protons, demonstrated to be extremely sensitive to the environment. On changing from cyclohexane solution to “H2O microemulsion”, the singlet of the ω position, at 62.1 ppm, moved to lower frequencies by 0.25 ppm and split into five signals, with the effect being slightly less intense for the “NH3 microemulsion” (Figure 6). A similar behavior was shown by the signals between 72.0 and 71.0 ppm, which correspond to the majority of the polyoxyethylene moiety’s carbons (Figure 7). The carbons of the “first” oxyethylene unit bonded to the phenyl ring, positions R and β, as also the methyl carbons of the tail, shifted slightly to higher frequency, by about 0.07 ppm.

Figure 5. 13C NMR signals of the methylene (at higher frequency) and methyl (at lower frequency) carbons of TEOS and EtOH from the “reaction mixture”: 0.5 h (s) and 37 h (- - -) after the reaction beginning.

The 13C resonances of the surfactant were assigned on the basis of the literature concerning Triton X-10022,23 and poly(ethylene oxide) (PEO),26 and with the help of HH and HC COSY spectra. Furthermore, the assignment of the headgroup carbons

3228 Langmuir, Vol. 25, No. 5, 2009

Asaro et al.

Figure 6. Region of the 13C NMR spectrum corresponding to the signals of the ω carbon for the samples: (A) 0.1 M solution of igepal CA-520 in cyclohexane, (B) “NH3 microemulsion”, and (C) “H2O microemulsion”.

Figure 7. Region of the 13C NMR spectrum corresponding to the signals of the headgroup carbons for the samples: (A) 0.1 M solution of igepal CA-520 in cyclohexane and (B) “reaction mixture” 0.5 h after the reaction beginning.

was supported by the shift experienced by the relevant signals upon changing from cyclohexane solution to microemulsion, with the highest variation being displayed by those of the “terminal” oxyethylene unit, the one bearing the OH group, which participates also as a donor in hydrogen bonding with water, and the least by the nuclei of the “first” unit, the one next to the phenyl ring. 29 Si NMR Spectra. In the present case, the 29Si NMR spectra were acquired through INEPT, a very efficient technique to increase the signal intensity of heteroatoms coupled to protons, especially in the case of nuclei, such as 29Si, with negative magnetogyric ratios,16,17 to counteract the low TEOS concentration, 0.04 M, at the beginning of the reaction. In this way, the signal from silicon species containing organic residues is selectively detected. The spectra, accumulated for only about 2 hours to allow the perception of the TEOS consumption kinetics, (26) Breen, J.; van Duijn, D.; de Bleijser, J.; Leyte, J. C. Ber. Bunsen Ges. 1986, 90, 1112–1122.

displayed only the signal of the starting reagent, decreasing in time. Signals of partially hydrolyzed species did not appear even in an overnight directly detected 29Si NMR spectrum, run in the presence of Cr(acac)3 as relaxation agent, which showed, in addition to that of TEOS, only the broader signal due to the glass of the NMR tube and of the probe head.9 Diffusion Measurements. The diffusion coefficient values obtained for the surfactant, TEOS, EtOH, and the exchangeable protons from the “H2O microemulsion” and the “reaction mixture” (“NH3 microemulsion” with TEOS addition), soon and 65 h after the reaction beginning, are reported in Table 1. In all the samples examined, the echo decay did not show any deviation from a unique exponential curve. This means that the obtained diffusion coefficient values are mean values, averaged over the different situations experienced by the relevant molecules. For example, in the “reaction mixture” after 65 h, the surfactant is in exchange among the bulk oil, the surface of the reverse (27) Walderhaug, H.; Johannessen, E. J. Solution Chem. 2006, 35, 979–989.

NMR Study of an InVerse Nonionic Microemulsion

Langmuir, Vol. 25, No. 5, 2009 3229

Table 1. Self-Diffusion Coefficientsa Reported in 10-10 m2 s-1 Measured at 25 °C sample b

0.1 M igepal in cyclohexane “H2O microemulsion”b “reaction mixture” at t ) 0c “reaction mixture” at t ) 65 hc EtOH in cyclohexane at 20 °Cd a

igepal t-Bu

exchangeable protons

3.38 ( 0.02 2.11 ( 0.02 2.40 ( 0.02 2.57 ( 0.02

0.97 ( 0.01 3.34 ( 0.01

Plus/minus the standard error from the fitting process.

b

∆ ) 100 ms. c ∆ ) 75 ms.

micelles, and that of silica nanoparticles. Moreover, the residence time in every environment is much shorter than the diffusion interval, ∆. If it were longer, a multiexponential decay would be observed, as found in the case of the highly concentrated igepal CA-520-n-heptane-water ternary system.28 The diffusion coefficient values for the surfactant molecule were obtained from the signal of the tert-butyl protons that resonate at 0.71 ppm (Figure 1), as it is the most intense and therefore the most reliable. The data were confirmed by the aromatic signals and also by the CH2 groups of both the “first” and the “terminal” oxyethylene units that resonated quite isolated. The diffusion coefficient values for TEOS and EtOH were obtained from the decays of the tallest peaks of the relevant methyl’s triplet, with the other two peaks of each multiplet giving values in close agreement. On the contrary, the methylenic 1H signals could not be employed due to their overlap with the resonances of the surfactant’s headgroup. In the “H2O microemulsion” and the starting “reaction mixture”, the diffusion coefficient for the signal of the exchangeable protons was easily obtained. On the contrary, for the latter sample analyzed 65 h after TEOS addition, the time chosen in order to have a significant presence of silica nanoparticles, its determination was hampered by the severe broadening which led to a partial overlap with the resonances of the surfactant β CH2 group. TEM. The TEM micrographs showed that the final product is highly uniform particles with a diameter of about 40 nm (Figure 8), in line with the results reported in the literature for synthesis carried out in an igepal-CO-520-concentrated ammonia-cyclohexane solution with a comparable H2O/igepal molar ratio.9

Discussion Igepal CA-520 was employed in most experiments, instead of igepal CO-520, the surfactant of choice for this synthetic route.8-10,12,14 It is preferable from the perspective of the NMR experiments since its iso-octyl tail originates only three signals in the 1H spectrum and four in the 13C one, whereas the latter gives rise to a myriad of signals of low intensity, due to the variable branching of the nonyl residue. However, the reaction was carried out in the presence of igepal CO-520 as well, and remarkable changes were not detected, probably because the only difference is the alkyl moiety of the tail, while the more important headgroup was unchanged. The trend of TEOS methyl 1H signal intensities versus time could be nicely fitted according to the well-established first order kinetic equation with respect to TEOS concentration (Figure 9).11,12

dTEOS ) -k[TEOS][OH-] dt

(2)

The value of the hydrolysis constant11 kh ) k[OH-] ) 0.024 h , from the data fit, is in the range of those reported in the literature for TEOS hydrolysis in microemulsion, obtained by IR -1

(28) Fleischer, G.; Gra¨tz, K.; Ka¨rger, J.; Meyer, H. W.; Quitzsch, K. J. Colloid Interface Sci. 1997, 190, 9–16.

d

TEOS

7.7 ( 0.1 8.3 ( 0.1

EtOH

8.87 ( 0.07 9.36 ( 0.04

Reference 27.

spectroscopy,11,12 and for the overall process of SiO2 production, determined by SAXS.14 TEOS hydrolysis is the rate determining step in the microemulsion system11 and governs also the final size of the particles due to the successive stages of nucleation and growth. The determination of the kh value was supported also by the good fit of the trend of the EtOH methyl signal intensities that obey the following law

[EtOH(t)] ) 4[TEOS(t ) 0)]{1 - exp(-kht)}

(3)

derived from eq 2 and from the stoichiometry of the reaction. The intensities of the 13C signals of TEOS and EtOH lie on the curves of the corresponding 1H data (Figure 9). On the contrary, the intensities of the 29Si signals are not appropriate for exhaustive kinetic study due to the poor S/N ratio. 29 Si NMR spectroscopy greatly contributed to the understanding of the Sto¨ber synthesis,29 where much higher concentrations of the starting tetraalkyl orthosilicate are used. It detected the monohydrolyzed monomer as the most abundant reaction intermediate,29 at variance with the present system where no intermediate species were observed neither in the 29Si NMR spectra nor in the 1H and 13C NMR spectra. More detailed information about the microscopic dynamics of the Sto¨ber reaction had been obtained by means of 29Si NMR spectroscopy from a system initially subjected to an acid treatment so that variously hydrolyzed monomers and dimers were present before addition of NH3. In this way, it was assessed that the nucleation rate is limited by the hydrolysis of the singly hydrolyzed monomer and that the doubly hydrolyzed product, which phase separates, can be considered as the very first nucleus.30 In the case of Sto¨ber synthesis, the aggregation model, that postulates that the particle size is governed by competition between the processes of nucleation and aggregation, emerged as best suited.30 The increase of the number of particles in time, which indicates that continuous nucleation occurs during the reaction period,29 and the monotonic increase of particle size with water concentration in the presence of electrolyte, which favors higher aggregation rates, are consistent with it.30 Great caution must be exerted in transferring these results to microemulsion systems due to the compartmentalization provided by the reverse micelles. They affect qualitatively and kinetically the intermediate steps, which can be intra- and/or intermicellar and depend on the water to surfactant molar ratio (R) and NH3 concentration.9,10 Intermicellar exchange becomes easier approaching the stability boundary of the microemulsion phase by increasing R and NH3 concentration. At low R, due to the small amount of “free” water, the TEOS hydrolysis is slow and fewer nuclei are formed, and consequently, larger particles are produced.9,10 For the reaction medium, we chose an R value near the minimum of the curve of the particle size versus R at high ammonia concentration. Indeed, the particles obtained are small (d ∼ 40 nm). Their uniformity suggests that the events of (29) Green, D. L.; Jayasundara, S.; Lam, Y.-F.; Harris, M. T. J. Non-Cryst. Solids 2003, 315, 166–179. (30) Lee, K.; Look, J.-L.; Harris, M. T.; McCormick, A. V. J. Colloid Interface Sci. 1997, 194, 78–88.

3230 Langmuir, Vol. 25, No. 5, 2009

Figure 8. TEM micrograph of the SiO2 nanoparticles.

Figure 9. Intensities versus reaction time of the 1H TEOS (O) and EtOH (0) methyl signal and 13C TEOS (2) and EtOH ([) methylene signal. The curves are the fitting of the intensities of 1H methyl signals carried out by means of eq 2 for TEOS and eq 3 for EtOH.

nucleation and growth are well separated in time for this microemulsion composition. SAXS studies confirmed that nucleation takes place during a limited period of time; in fact, the number density of the particles became constant after the first 10 h.14 Diffusion coefficients are commonly exploited to characterize microemulsions, for example, to determine the connectivity of the phase and the size of water droplets.31,32 In the case of the “H2O microemulsion”, the diffusion coefficient of the exchangeable proton signal, which corresponds mainly to water, lower than that of the surfactant, indicates the presence of discrete droplets, as postulated by previous studies.12 Its value, employing the Stokes-Einstein relation (D ) kT/(3πηdh)) and the cyclohexane literature viscosity constant η) 0.9 mPa · s,27 indicates that the hydrodynamic diameter (dh) of reverse micelles should be about 5 nm, in line with SAXS findings.14 The dramatic increase of the diffusion coefficient value of the exchangeable proton signal in the “reaction mixture” (Table 1) can be rationalized by the intermicellar diffusion of NH3 through the bulk oil, in line with the not negligible oil solubility of NH3. The transition to a bicontinuous system can be ruled out, since the diffusion coefficient of the surfactant has not changed as dramatically as that of the exchangeable protons. Its slight increase with respect to “H2O microemulsion” rather reflects a reduction of the droplet size.12 (31) Lindman, B.; Stilbs, P. In Microemulsions “Structure and Dynamics”; Friberg, S. E., Bothorel, P., Eds.; CRC Press: Boca Raton, FL, 1987; Chapter 5. (32) Furo´, I. J. Mol. Liq. 2005, 117, 117–137.

Asaro et al.

Unfortunately, the self-diffusion coefficient value of the surfactant molecules cannot provide detailed information about the evolution of the microemulsion structure during the reaction and about the presence and growth of SiO2 nanoparticles, which, on the contrary, is achieved in SAXS investigations.14 In fact, 65 h after the reaction beginning, that is in an advanced stage, it has just slightly increased. It must be recalled that it is a mean value, since, in the inverse microemulsion, the surfactant molecules are in fast exchange between the outer layer of the reverse micelles and the bulk oil, where they diffuse easily as monomeric species. During the reaction, the size of the reverse micelles decreases slightly owing to H2O consumption, but there is a new site corresponding to the covering layer of hydrated silica nanoparticles.4 The presence of the nanoparticles, greater in size than the reverse micelles, should decrease the surfactant’s diffusion not only by binding but also by acting as obstacles; however, their hindering effect results in being negligible probably because they are, by far, less numerous than the micelles.9 The surfactant molecules that are adsorbed on the silica surface both keep the particles in solution by steric stabilization and provide an impediment to incoming reacting species. However, the confinement is not stiff, since during the particle growth further surfactant molecules may be adsorbed. The high self-diffusion coefficient of the surfactant in the microemulsions with respect to that of the exchangeable protons in “H2O microemulsion”, corresponding to the self-diffusion coefficient of the inverse micelles, indicates that a lot of surfactant is dissolved in the bulk, which may act as a surfactant’s reservoir. The diffusion coefficient of EtOH, higher than the surfactant one, indicates that the partitioning of the reaction byproduct, between the water pool and the oil, occurs with preference for the latter, as previously suggested by Riello et al.,14 and in agreement with the localization of the alcohol determined in a cyclohexane-AOT-water inverse microemulsion always by means of PGSE-NMR.27 An estimation of EtOH partitioning between the two environments can be attempted. The molar partition of EtOH in the aqueous phase (p) can be calculated from the measured self-diffusion coefficient (Dobs) according to eq 4, considering fast exchange of EtOH among the various situations during the diffusion interval:

p)

Dfree-Dobs Dfree-Dmic

(4)

where Dfree is the self-diffusion coefficient for EtOH in cyclohexane27 and Dmic is approximated by the value of the selfdiffusion coefficient measured for the exchangeable proton in the “H2O microemulsion”, since that of the “NH3 microemulsion” is too high (higher than the surfactant’s) to be attributed mainly to micellar diffusion. In this way, a p value of 0.06 is obtained. Conversely, calculating p by means of the partition equilibrium constant (Kc) value reported for EtOH in the inverse micelles of an AOT-water-cyclohexane system, employing as Voil and Vwater the volumes of cyclohexane and NH3 solution, respectively, according to eq 5

Kc )

( 1 -p p )( V

Voil water

)

(5)

a value of p ) 0.10 was obtained. It is in good agreement with the previous one, considering the coarse approximations. Moreover, Kc may depend significantly on the surfactant and, above all, on the composition of the aqueous phase. Further information on the microemulsion was afforded by the 1 H and 13C spectra of the surfactant molecules. The 1H spectra

NMR Study of an InVerse Nonionic Microemulsion

Langmuir, Vol. 25, No. 5, 2009 3231

Figure 10. 13C NMR spectrum of igepal CA-520 in CDCl3: signals of the aromatic quaternary carbons.

evolve from the solution in an organic solvent, such as cyclohexane or CDCl3, to the microemulsion in a way that parallels the evolution due to hydration of both Triton X-10022 and poly(ethylene oxide) (PEO) alcohol surfactants.25 The commercial alkyl phenyl ethoxylated surfactants are mixtures of oligomers whose ethylene oxide number (EON) varies in agreement with a Poisson distribution.33 The higher number of peaks in the 13C spectra of the microemulsion systems is due to the resolution of the signals originated by the surfactant molecules having different EONs; a nice example is provided by the terminal OH bearing carbon (position ω) (Figure 6). On the other hand, in CDCl3 solution, the electronic effect of the PEO chain length is appreciable, at the operating fields of 9.4 T, only for the carbons of the phenyl ring in position 1, that is, the one bearing the polyoxyethylene chain, and 4, in para to the former (Figure 10), and the shift differences are very small, in the order of a few hundredths ppm. The ability of water to resolve the NMR signals of different oligomers and the spectral changes originated by phase transition must be related to the water-polyoxyethylene interaction. Several NMR and Raman spectroscopic studies have shown that polyoxyethylene (PEO) is a very flexible polymer with a high degree of internal motion26 in solution. It has been proposed that hydration and the average conformational state are interdependent.34,35 In water solution, even in the case of very short polyoxyethylene chains, the probability of finding the two methylenic groups of one unit in gauche conformation is rather high36 and for longer sequences the helical arrangement of the solid state is partially retained,37,38 at variance with the melt, where it is present as a random coil. The PEO 7/2 helix consists of a trans-gauche-trans sequence for each O-C-C-O repetitive unit. NMR studies carried out on Triton X-100 micelles in water revealed a significant contribution of the coiled conformation for the headgroups, probably favored by packing requirements at the micellar surface, which also modulates the penetration of water in the polar headgroup region, whereas the (33) Siegel, M. M.; Tsao, R.; Oppenheimer, S.; Chang, T. T. Anal. Chem. 1990, 62, 322–327. (34) Hey, M. J.; Ilett, S. M.; Davidson, G. J. Chem. Soc., Faraday Trans. 1995, 91, 3897–3900. (35) Karlstro¨m, G. J. Phys. Chem. 1985, 89, 4962–4964. (36) Goutev, N.; Nickolov, Z. S.; Georgiev, G.; Matsuura, H. J. Chem. Soc., Faraday Trans. 1997, 93, 3167–3171. (37) Connor, T. M.; McLauchlan, K. A. J. Phys. Chem. 1965, 69, 1888–1893. (38) Derkaoui, N.; Said, S.; Grohens, Y.; Olier, R.; Privat, M. J. Colloid Interface Sci. 2007, 305, 330–338.

polyoxyethylene chains of the monomeric surfactant in water are more elongated.39 Moreover, the EON determines the partitioning of polyoxyethylene alkylphenyl ether surfactants between oil and water in microemulsion systems.40,41 The small changes of the 1H and 13C spectral patterns of the surfactant occurring during the reaction, with shifts of the opposite sign with respect to those induced by the formation of the microemulsion, are probably the result of the consumption of water that decreases the hydration of the surfactant headgroup. The 1H, 13 C, and PGSTE NMR results confirm that the microemulsion does not undergo drastic changes due to the formation of the product and, above all, of the byproduct EtOH.

Conclusions 1

H has confirmed to be the most efficient nucleus to follow TEOS hydrolysis during the process of base catalyzed SiO2 nanoparticle production in inverse microemulsion, with the increasing signal of the byproduct EtOH exploitable as a control. 1 H and 13C provide information in molecular detail on soluble species, thus complementing the results of techniques such as scattering methods sensitive to objects of nanometric size in order to obtain a thorough picture of the system. Insight into the compartmentalization provided by the inverse microemulsion is afforded by the PGSTE-NMR results. They have demonstrated that NH3 diffuses through the bulk oil and have confirmed that the latter constitutes the preferential environment for EtOH. Thus, it is advantageous to work in a very diluted system, where the particles are far apart and most EtOH is oil dissolved, in order to maintain the stability of the reaction medium. Acknowledgment. Italian MIUR (PRIN_2006039789_005) is gratefully acknowledged for financial support and Fondazione CRTrieste for the purchase of the Varian 500 NMR spectrometer. We thank D. Cristofori for the TEM micrographs. Supporting Information Available: Partial 1H NMR spectrum of the “NH3 microemulsion” with addition of EtOH. This material is available free of charge via the Internet at http://pubs.acs.org. LA803826C (39) Yuan, H.-Z.; Cheng, G.-Z.; Zhao, S.; Miao, X.-J.; Yu, J.-Y.; Shen, L.-F.; Du, Y.-R. Langmuir 2000, 16, 3030–3035. (40) Ma´rquez, N.; Bravo, B.; Cha´vez, G.; Ysambertt, F.; Salager, J. L. Anal. Chim. Acta 2002, 452, 129–141. (41) Ma´rquez, N.; Bravo, B.; Ysambertt, F.; Cha´vez, G.; Subero, N.; Salager, J. L. Anal. Chim. Acta 2003, 477, 293–303.