Journal of Sol-Gel Science and Technology 28, 15–18, 2003 c 2003 Kluwer Academic Publishers. Manufactured in The Netherlands.
Interrogation of Microporous Silica by Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry JAMIE L. COHEN AND JAMES A. COX∗ Department of Chemistry and Biochemistry, Miami University, Oxford, OH 45056, USA
[email protected]
Received July 24, 2002; Accepted January 31, 2003
Abstract. Microporous sol-gels prepared from a tetramethyl orthosilicate precursor with 5-hydroxytryptophan, 5-HTPP, as a dopant were investigated by matrix-assisted laser desorption/ionization mass spectrometry (MALDI). Spectra of 5-HTPP and its oxidation products were obtained when (a) the matrix, α-hydroxycinnamic acid, was included in the sol, (b) thin wafers of silica were used, and (c) the surface of the gel was abraded after mounting the sol-gel on the MALDI sample puck so that laser reflection was minimized. The methodology permitted elucidation of the 5-HTPP oxidation pathway. In microporous silica a low molecular product, the dione, was obtained whereas in a mesoporous matrix, dimeric products were formed. Keywords: mass spectrometry, microporous silica, mesoporous silica, indole oxidation, reaction pathway
Introduction The ability to perform reactions in the pore volume of sol-gel materials is one factor in their potential applicability to a wide range of chemical studies. To date, the scope of these applications has been limited. The use of sol-gels as hosts for enzymes and other proteins dominate the reported applications. The protection against denaturization of the enzyme by sample components, the preservation of selectivity and activity of enzymes that results from free motion in the host phase, and the general stabilization of proteins are among the attributes that make encapsulation in sol-gels attractive [1–4]. In the reported studies, interrogation of reaction pathways in these porous solids was not the focus; however, sol-gels have characteristics suggesting that they will have application as media for chemical reactions in the pore volume. Of particular importance is the ability to control mean pore size over a wide range. Only a few studies have been reported on the influence of sol-gel matrices on reaction pathways. The pho∗ To
whom all correspondence should be addressed.
tochemistry of UO2+ 2 in microporous silica was hypothesized to yield UIV via reduction of UO+ 2 by ethanol radicals rather than to form it by the bimolecular disproportionation of UO+ 2 , which is the pathway in homogeneous solution [5]. The electrochemical reduction of UO2+ 2 in a silica sol-gel showed a matrix effect that was demonstrated as the catalysis of the disproportionation of UO+ 2 by ion-exchange interaction of this intermediate with the negative sites on silica [6]. A related report showed a matrix effect on the photoconversion of trans-4-methoxy-4 -(2-hydroxyethoxy)-azobenzene to the cis form [7]. Extending the applications of sol-gels as reaction media will require increasing the scope of interrogation methods. Because of the favorable optical properties of the most widely used of these materials, silica, spectroscopic methods have been employed extensively [8]. The presence of electrolyte in the pore volume of nonconductive forms such as silica and the intrinsic conductivity of mixed-valence forms such as vanadia have permitted electrochemical interrogation of these materials and components therein [9, 10]. For study of coupling reactions, mass spectrometry is well suited. We
16
Cohen and Cox
developed matrix-assisted laser desorption/ionization mass spectrometry (MALDI) methodology for the elucidation of components in mesoporous silica [11]. The present report is an extension of that study to a more challenging matrix, microporous silica.
Experimental Unless otherwise stated, all chemicals were ACS Reagent Grade from Fisher Scientific (Fair Lawn, NJ). The 5-hydroxy-L-tryptophan (5-HTPP) and the Triton X-114 were from Sigma (St. Louis, MO). The following were obtained from Aldrich (Milwaukee, WI): tetramethyl orthosilicate (TMOS), 98% purity; α-cyano-4-hydroxycinnamic acid (CHCA), 97% purity; and 2,5-dihydroxybenzoate (DHB), 99% purity. House-distilled water that was further purified with a Barnstead NANOpure II system was used in all cases. Microporous silica was prepared from a solution comprising 1.0 mL of each of the following (added in the listed order): methanol, 0.05 M H2 SO4 , 0.2 M KCl, 0.4 mM 5-HTPP, and TMOS. In certain experiments, 17.5 mg of CHCA was included. Because the CHCA did not completely dissolve, it was added before the TMOS. In addition, the solution was stirred for 1 h and filtered before the TMOS was added. The mesoporous silica was prepared from a mixture of 0.030 g of Triton X-114, 3.0 mL of methanol, 1.5 mL of 0.05 M H2 SO4 , 1.5 mL of 1.0 mM 5-HTPP, and 3.0 mL of TMOS. These mixtures were stirred for 1 h prior to casting samples. The MALDI experiments were performed with a Bruker Reflex III instrument with a time-of-flight mass analyzer (Billerica, MA). The accelerating voltage was 28.5 kV, and the effective flight path with a reflectron mass analyzer was 290 cm. A nitrogen gas laser (Laser Scientific, Inc., Franklin, MA) operated at 5 mW (4 ns pulse duration) with an output at 337.1 nm was used. The matrices (CHCA and DHB) provided the calibration lines for the spectra. Samples were mounted with double-sided tape, taking care to not have any of the tape exposed.
Results and Discussion The study was initiated as a means of diagnosing whether the oxidation of indoles in a silica sol-gel followed a pathway that depended on pore size. The hypothesis was that the products in mesoporous silica
Figure 1. MALDI mass spectrum of 5-HTPP in mesoporous silica after air oxidation for three weeks. The sample is a monolith that is held to the target puck by double-sided tape. The matrix and external calibration standard is DHB. The y-axis label, I, is absolute intensity.
will include dimers and other oligomers, thereby emulating the liquid phase pathway [12], whereas in microporous silica, monomeric species such as diones were expected. Based on our previous work [11], MALDI was tested as the method of product identification. Consistent with our previous report [11], a mass spectrum taken of mesoporous silica that included CHCA provided evidence for dimerization upon oxidation of 5-HTPP. Indeed, even when the MALDI matrix was added after gelation, the mass spectrum of the oxidation product was obtained. For example, Fig. 1 is the spectrum of a mesoporous sol-gel that was obtained after the 5-HTPP dopant was air-oxidized for three weeks. The DHB matrix was pipetted onto the surface of the monolith 48 h prior to obtaining the spectrum. The peak at m/z 451.3 differs by only 0.15% from an expected product, the dimer of 5-HTPP with the addition of an oxygen atom and the elimination of two hydrogen atoms. This agreement is significant in that an external calibration standard, DHB, was employed. In our previous work, the calibration compound (matrix) was included in the sol. Of importance to general studies of reactions in sol-gel materials is that a signal was obtained for the doped chemical system. When the experiment was repeated with microporous silica, mass spectra were not obtained with either externally added DHB or internally doped CHCA. Neither 5-HTPP nor its oxidation products were observed. Negative results also were obtained when a Nd:YAG laser operated at 355 nm was used. The monoliths were cracked prior to use and the inner portion exposed to the laser beam to assure that the problem was not exclusion of the 5-HTPP from the outer surface of the monolith during the sol-gel processing. The contrast between the results from mesoporous and microporous silica suggested that sampling was primarily from the
Interrogation of Microporous Silica by MALDI
interior rather than the surface of the sample. In this regard, interrogation of the surface is not expected to be dependent on pore size. Subsequent experiments were performed exclusively on sol-gels that were doped with the CHCA matrix prior to processing. To obtain signals from dopants in microporous silica, two changes were made in preparing the samples. First, the geometry was altered so that the size of the sol-gel matched the 2-mm diameter target zones on the MALDI sampler holder, a step that also resulted in thin, somewhat-curved wafers. Because of the significant radius of curvature, the thickness cannot be measured, but the maximum was about 0.08 cm. These structures were formed on a glass slide that was first pretreated R with Rain-X (Blue Coral-Slick 50, Ltd., Cleveland, OH), an anti-wetting agent that is widely available in retail stores. Clean, dry microscope slides were coated with this solution, dried, polished with a soft tissue, and coated a second time. Using a microliter pipet, several droplets (ca. 50 µL as a mean volume) of sol were spotted onto the treated glass surface and dried under ambient laboratory conditions. In this manner, multiple samples were prepared with a range of diameters. In contrast to the use of untreated glass, the sol-gels did not crack even after three weeks of aging. Because they did not stick to the surface, they were readily removed from the glass slide and mounted on the MALDI sample puck where they were held with double-sided tape, as previously described. The factor(s) responsible for the influence of the sample geometry on the signal generation step was not identified. The second step in the sample treatment that was used to obtain a MALDI signal from microporous silica was to roughen the surface so that reflection of the radiation from the laser was decreased. Scraping the surface of the mounted sol-gels with a spatula, a process that also cracked the samples, was sufficient. The mass spectrum that was obtained after three weeks of air oxidation of the 5-HTPP is shown in Fig. 2. Known peaks for the matrix at m/z 190.2 and 379.2 served as calibration points. Consistent with the hypothesis regarding influence of pore size on the reaction products, peaks for the dimer and for the dimer with addition of oxygen (and concomitant loss of two hydrogen atoms) at m/z 437 and 451, respectively, were absent. The formation of dimeric product at a concentration below the detection limit of the method was not precluded by the data in Fig. 2; however, sensitivity of the silica background peak in the m/z 500 region was comparable to the analogous peak in Fig. 1. A more
17
Figure 2. MALDI mass spectrum of 5-HTPP in microporous silica after air oxidation for three weeks. The sample is a 2-mm (dia.) curved wafer (maximum thickness, ca. 0.08 cm) with the surface abraded to minimize reflection of the laser. The matrix and internal calibration standard is CHCA. The y-axis label, I, is absolute intensity.
important point is that the signal at m/z 234.9 agrees to within 0.04% of the predicted position of the protonated dione, which is the expected product when the dimerization is blocked. Conclusion MALDI methodology that was successful for identifying dopants in mesoporous silica was not suited for the interrogation of microporous silica. By using a combination of thin sample wafers, abrasion of the surface to attenuate reflection loss of the laser radiation, and inclusion of the MALDI matrix in the precursor sol, mass spectra were obtained for dopants in microporous silica. The utility of the resulting methodology to study chemical reactions in the pore volume of the sol-gel was illustrated by investigation of the air oxidation of an indole, 5-HTPP. In the microporous solid, the dione was formed whereas in mesoporous silica, dimeric products were observed. Thus, in contrast to our previous study on reactions in silica [6], ion exchange between the backbone and a cationic intermediate did not perturb the pathway that was predicted on the basis of pore size relative to potential products. Acknowledgments This work was supported in part by donors to the Petroleum Research Foundation of the American Chemical Society through grant PRF 33863 AC5. Partial support was received by J.L.C. from the Howard Hughes Medical Institute through a grant to Miami
18
Cohen and Cox
University and from the Undergraduate Summer Scholars program at Miami University.
References 1. D. Avnir, S. Braun, and M. Ottolenghi, in Supramolecular Architecture: Synthetic Control in Thin Films and Solids, edited by T. Bien (American Chemical Society Symposium Series 499, ACS, Washington, DC, 1992). 2. D. Avnir, S. Braun, O. Lev, and M. Ottolenghi, Chem. Mater. 6, 1605 (1994). 3. B.C. Dave, B. Dunn, J.S. Valentine, and J.I. Zink, Anal. Chem. 66, 1120A (1994). 4. L.M. Ellerby, C.R. Nishida, F. Nishida, S.A. Yamanaka, B. Dunn, M. El-Sayed, J.S. Valentine, and J.I. Zink, Science 255, 1113 (1992).
5. S. Dai, D.H. Metcalf, G.D. Del Cul, and L.M. Toth, Inorg. Chem. 35, 7786 (1996). 6. M.E. Tess and J.A. Cox, J. Electroanal. Chem. 457, 163 (1998). 7. M. Ueda, H.-B. Kim, T. Ikeda, and K. Ichimura, Chem. Mater. 4, 1229 (1992). 8. C.J. Brinker and G.W. Scherer, Sol-Gel Chemistry: The Physics and Chemistry of Sol-Gel Processing (Academic Press, New York, NY, 1990). 9. K.S. Alber and J.A. Cox, Mikrochim. Acta 127, 131 (1997). 10. O. Lev, Z. Wu, S. Bharathi, V. Glezer, A. Modestov, J. Gun, L. Rabinovich, and S. Sampath, Chem. Mater. 9, 2354 (1997). 11. J.B. Laughlin, C.J. Cassady, and J.A. Cox, Rapid Commun. Mass Spectrom. 9, 1505 (1997). 12. K.A. Humphries, M.Z. Wrona, and G. Dryhurst, J. Electroanal. Chem. 346, 377 (1993).