Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Org Biomol Chem. Author manuscript; available in PMC 2010 November 2.
Published in final edited form as:
PMCID: PMC2970626

Radical Pairs with Rotational Fluidity in the Photochemical Reaction of Acetophenone and Cyclohexane in the Zeolite NaY: A 13C CPMAS NMR and Product Analysis Study


The photochemical reaction of acetophenone and cyclohexane in the zeolite NaY occurs by combination of the geminate radical pairs to give products that reveal a significant amount of rotational fluidity, which was also documented by intermolecular nuclear dipolar interactions measurements using cross polarization 13C NMR (CPMAS) experiments.


During the last few years, a number of organized assemblies have been examined as reaction media to control the product distribution of photochemically generated transients, such as excited states and reactive intermediates.i Examples include molecular crystals, zeolites, inclusion complexes (both in the solid state and in solution), liquid crystals, micelles, Langmuir-Blodget films, and others. The effect of these media varies widely and depends primarily on their influence on the motion of the reactants, generally restricting some reaction pathways and sometimes opening new ones.ii Not surprisingly, the largest effects as compared to reactions in solution are seen with reactions caried out in solids. In one extreme, close-packed molecular crystals restrict molecular translation, rotation, and large amplitude conformational motions, so that chemical reactivity is limited to least motion pathways.iii Alternatively, photochemical reactions carried out in zeolites take advantage of most molecular degrees of freedom, in measures that depend on the sizes and shapes of the reactants as compared to those of the zeolite cavities and channels.iv, v The relation between molecular fitting and reaction control has been deduced by product analysis in the MX and MY zeolites by changing the sizes of the cation M,vi and by adding molecular “spectators”vii and chiral co-adsorbates.viii A few years ago, while studying the reaction between benzophenone-d10 and cyclohexane in NaX we demonstrated that intermolecular {1H}-13C cross polarization in 13C CPMASix NMR experiments could be used as a measure of close proximity and restricted mobility.x A relatively static diphenylketyl-cyclohexyl radical pair (R=Ph, Scheme 1) formed by intermolecular hydrogen abstraction in the orientation depicted in Scheme 1A led to 1,1-diphenyl-cyclohexylcarbinol as the only product (as indicated by the half arrows). In this paper we describe an analogous study with acetophenone and cyclohexane in NaY (R=Me, Scheme 1). We report how the smaller reactant experiences greater mobility, so that the radical pair formed by hydrogen transfer can rotate and explore additional reaction pathways.xi In addition to the product formed from orientation A, products were also obtained from the bonding pathways indicated by orientations B and C. The increased mobility of the reactants was also documented by intermolecular {1H}-13C cross polarization measurements, which revealed a weaker dipole-dipole interaction.

Scheme 1
Alternative forms of C-C bond formation allowed by in-cage rotational motion of the ketyl-cyclohexyl radical pair in the zeolite NaY.


Photochemical Experiments

Acetophenone (50 μl) was irradiated using a 400W medium pressure mercury lamp (<300 nm) in dry cyclohexane for 2 h. For photoreactions within NaY zeolite, 50 μl of acetophenone and 3 ml of cyclohexane were mixed in a Pyrex test tube. To this solution was added 200 mg of NaY activated for 3 h 425 °C and the resulting suspension stirred for 6 h under an argon atmosphere. The supernatant cyclohexane layer was removed, tested for the removal of acetophenone and the zeolite washed three times with 3 ml fresh cyclohexane. Samples were irradiated as cyclohexane slurries for 1–6 hours in Pyrex test tubes (λ<300 nm) with the 400W medium pressure mercury lamp. Products were extracted by stirring with ethyl acetate for 6 h. Product analyses were carried by GC and GC-MS with retention times and fragmentation patterns calibrated with those of authentic samples. The identities of products 2,xii 3axiii and 3bxiv and 4xv were confirmed by comparing spectral data with that previously reported in the literature.

Solid State NMR Measurements

The 13C NMR spectra of all zeolite samples were carried out at 75.47 MHz on a Bruker Avance instrument. The 13C CPMAS experiments were run with the TOSS sequencexvi in order to suppress the spinning sidebands. Adamantane was employed as an external chemical shift reference and the magic angle was adjusted with KBr. The 90° pulse width was 4.25 μs. Adequate cross polarization was achieved using 1.5 ms as contact time in the case of natural abundance samples and 5 ms in the case of acetophenone-d8. A rotation rate of 10 kHz was used for all samples. Zeolite samples for 13C-CP-MAS studies were prepared with acetophenone labeled with 13C at the carbonyl and methyl carbon to shorten the acquisition times.

Results and Discussion

Zeolites X and Y are microcrystalline aluminosilicates made up of [SiO4]4− and [AlO4]5− tetrahedra that form an extended three-dimensional network with an architecture that results in relatively large cavities known as supercages (Fig. 1). With an internal diameter of ca. 12 Å, each supercage is connected to four others by openings that have ca. 8 Å diameter, so that medium-size organic molecules are able to diffuse through the three-dimensional network. The rate of diffusion, rotation, and conformational motions of the molecules included therein depend on their size and shape, as well as the specific chemical interactions with zeolite walls. Given that each aluminum center carries a negative charge, the structure of the zeolites requires an equivalent number of charge-compensating cations, which are largely responsible for the specific binding interactions with the organic adsorbates. Na-exchanged Y zeolites (NaY) such as the one used in this study have a low Al content, and a smaller number of cations than the analogus X-zeolites. While samples of dried NaY may be considered relatively inert, they are able remove acetophenone from a cyclohexane solution in as much as 25% of their own dry weight, clearly suggesting a very favorable binding interaction.

Fig. 1
Structure of the Faujasite X and Y framework illustrating the sites (I–III) of the charge-compensating cations.

Photochemical Results

Our observations on the solution and zeolite photochemical reactivity of acetophenone and cyclohexane are summarized in Scheme 2. Photochemical excitation in solution results in formation of pinacol 4 and several low molecular weight products, including bicyclohexane and cyclohexene. It is well known that the reaction occurs by intermolecular H-abstraction from the rapidly formed triplet excited state to form a triplet radical pair that separates to become free radicals.xvii The results suggest that the kinetics of termination for the ketyl and cyclohexenyl radicals free radical are very different as no coupling between them is observed in solution.xviii In contrast, the reaction in NaY is characterized by a very efficient cage effect. While 1-cyclohexyl-1-phenylethanol 2 is formed by coupling the cyclohexyl and acetophenone-derived radicals at the ketyl carbon, 2′-cyclohexylacetophenone 3a and 4′-cyclohexylacetophenone 3b are formed by coupling at the ortho- and para-positions, respectively. It is interesting that pinacol 4 and bicyclohexyl were completely absent (by GCMS) when the photoreaction was conducted in the zeolite.

Scheme 2
Photochemical reactivity of acetophenone and cyclohexane in (a) dilute solution, and (b) co-adsorbed in the zeolite NaY.

It has been previously established that radical combination is the dominant reaction pathway for ketyl and alkyl geminate radical pairs generated by intermolecular hydrogen abstraction with the zeolite supercage.xix,x Some examples include the reactions of benzaldehyde, p-methylbenzaldehyde, acetophenone, and benzophenone with toluene,xix and benzophenone with cyclohexane.x All these reactant pairs favor the coupling of the corresponding alkyl and ketyl radicals to give a tertiary alcohol. Based on this precedent, the absence of pinacol 4 and the formation of the tertiary alcohol 2 from acetophenone and cyclohexane within NaY come as no surprise. The results suggest: (1) that having two reacting acetophenone molecules in the same supercage is very unlikely, (2) that reacting ketones share the supercages with cyclohexane molecules and, (3) that the resulting radical pairs are unable to separate and become free radicals. However, the formation of 2′-cyclohexylacetophenone 3a, and 4′-cyclohexylacetophenone 3b suggest that radical pairs are able to explore several orientations for the ortho- and para-coupling reactions to occur (Scheme 1).

It should be noted that formation of 3a and 3b requires ketonization and oxidation of the initially formed coupling products. The initial coupling at the ortho- and para-positions followed by ketonization of resulting enols leads to 2,4- and 1,4-cyclohexadienone intermediates 3a-H2 and 3b-H2 (Scheme 2). While no evidence could be obtained for these intermediates by solid state 13C CPMAS NMR (vide infra), control experiments with samples of commercial 1,4-cyclohexadiene in NaY showed that oxidation and aromatization takes place spontaneously within the time scale of the photochemical reaction and analysis. The results from a series of runs under various experimental conditions are shown in Table 1. Notably, photochemical experiments carried out for 6 h with areated, O2-purged, and Ar-puged samples showed relatively modest differences, suggesting that chemisorbed oxygen may be the oxidant.xx Given that the spin density at the ortho- and para-positions are relatively similar,xxi the yields of 3a and 3b suggest that ortho-coupling is sterically hindered as compared to para-coupling. One may also speculate that ortho-coupling should be kinetically disfavored with respect to coupling with the nearby ketyl position. Although the uncertainity of our measurement is relatively high at low conversion values (17%), the relative yields of 3 (a and b) as compared to 2 are relatively constant for conversion values of 56% and 95%. Combination at the ketyl radical center to form 2 accounts for ca. 42% and combination at the ortho- and para-carbons to form 3a and 3b accounts for ca. 48% of the products formed.

Table 1
Photoproduct distribution from acetophenone and cyclohexane in solution and in the zeolite NaY.

Solid State 13C-CPMAS NMR Studies

Reaction progress was documented in situ by solid state NMR measurements by comparing the spectra before and after several irradiation periods. Samples were prepared with acetophenone labeled with 13C at the carbonyl and methyl carbons in order to have sufficient signal intensity and chemical resolution to detect the starting material, potential intermediates, and final products. The 13C-CPMAS spectra of acetophenone and cyclohexane coadsorbed in NaY acquired before photolysis, and after 2.0 and 5.5 hours of UV irradiation are illustrated in Fig. 2. The spectrum of acetophenone and cyclohexane before photolysis exhibits 13C peaks at 27 ppm and 208 ppm. Independent measurements with samples of the two components (Fig 2) confirmed that high field peak belongs to the overlapping signals of cyclohexane and the methyl group of acetophenone (Fig. 1a). The signal at 208 ppm corresponds to the carbonyl carbon and the natural abundance aromatic signals are not visible in this experiment. Evidence of reaction is clearly visible after 2.0 h of irradiation (Fig. 2b) with a signal belonging to the tertiary alcohol of 2-cyclohexyl-2-phenylethanol 2 clearly visible at 78 ppm along with some broadening of the methyl and carbonyl signals. Further irradiation increased the intensity of the carbinol carbon in 2 and led to changes in the carbonyl region.

Fig. 2
13C CPMAS NMR spectra (75.47 MHz) of 13C CO- and Me-labelled acetophenone and cyclohexane co-adsorbed in zeolite NaY (a) before photolysis, (b) after 2 hours of photolysis (17% conversion), (c) after 5.5 hours of photolysis (40% conversion).

A decrease in the intensity of the signal at 208 ppm was accompanied by the appearance of a broad peak at ca. 216 ppm, which is consistent with the formation of 4′-cyclohexylacetophenone 3b along with a small amount of the its ortho-isomer 3a. Deconvolution of the carbonyl signals in terms of two Lorentzian functions revealed that the reactant carbonyl had broaden from 516 Hz before photolysis (Fig. 2a) to 944 Hz after 5.5 h irradiation (Fig. 2c). This broadening is indicative of an increased heterogeneity, which leads us to suggest that molecular motion within the zeolite cage is more restricted once the products are formed. We analyzed the peak at 27 ppm in the spectrum obtained after 5.5 h of irradiation using the program dmfitxxii as a superposition of lines in order to account for the two major products (2 and 3a) and the two reactants. The cyclohexane peak in Fig. 3b can be fit to a Lorentzian with a shift of δ=27.56 ppm and a line width of 34.95 Hz. Keeping those parameters for cyclohexane constant, and modeling the peak around 27 ppm in Fig. 2a (corresponding to cyclohexane and acetophenone) as a superposition of two Lorentzians, we obtained a shift of δ=25.91 and width of 197 Hz for acetophenone. We found that the experimental spectrum could not be reproduced by these values as well as the literature values of the chemical shift product 3b (δ=26.55) constant. A satisfactory fit was obtained by fitting co-added cyclohexane and acetophenone with a line width of 299 ppm, and the two products with line widths of 533 and 234 Hz, respectively. The estimated errors in this model are relatively small (±0.01–0.2 ppm and ±8–30 Hz) and the spectrum is relatively well simulated. While highly tentative, this analysis suggests a nearly 10-fold increase in the line width of cyclohexane and a 50% increase in the linewdith of the methyl group of acetophenone, as expected for a highly anisotropic sample. In further agreement with this conclusion we determine that the approximate line width of the static spectra increases from 486 Hz before photolysis to 782 Hz after 40% reaction.

Fig 3
13C CPMAS spectra (75.47 MHz) of (a) acetophenone and cyclohexane co-adsorbed in zeolite NaY obtained with a contact time of 1.5 ms and 7200 scans, (b) acetophenone-d8 and cyclohexane co-adsorbed in zeolite NaY obtained with a contact time of 5 ms and ...

The relative rigidity and mobility of the two reactants within the cavity of the zeolite NaY was probed in a qualitative manner with 13C CP-MAS NMR measurements using acetophenone-d8 and natural abundance cyclohexane. The cross polarization (CP) experiment involves the selective excitation of the 1H nuclei followed by dipolar magnetization transfer to the less abundant 13C. A reasonable analogy may be drawn between the dipolar interactions in the nuclear CP experiment and those that dictate the Forster type energy transfer mechanism (FRET).ix,x The rate and efficiency of cross polarization rely on the strength of the 1H-13C dipolar coupling, which depends on their distances and orientations. The most interesting difference between the magnetic and electronic experiments is that a dismal spectral overlap between 1H and 13C, with resonance frequencies at 300 and 75 MHz (in our instrument), is converted into a perfect frequency match at fields that correspond to ca. 40 kHz by using the spin-locking RF pulse established by Hartman and Hahn.xxiii Under these conditions, isoenergetic 1H-13C coupled transitions transfer the magnetization from the abundant 1H to the rare 13C over time scales that vary from 0.1 to 10 ms. While most 13C CP-MAS experiments involve intramolecular magnetization transfer, our experiment relies on the transfer of magnetization from one molecule to the other.

As a reference point we first acquired a 13C CPMAS spectrum of natural abundance acetophenone and cyclohexane coadsorbed in zeolite NaY. A relatively noisy spectrum was obtained in 7200 scan with a cross polarization time (contact time) of 1.5 ms (Fig. 3b). The peak at 27 ppm is attributed to the methyl group of acetophenone and the equivalent carbons of cyclohexane. Although the two signals overlap, they have significantly different line widths. Two broad signals at 127 ppm and 145 ppm correspond to the aromatic carbons but the signal of the carbonyl carbon is not visible with the unlabeled samples. The spectrum in Fig. 3b was obtained with a sample prepared with acetophenone-d8 and natural abundance cyclohexane using a cross polarization time of 5 ms and 31900 scans. Since no signals from acetophenone were observed despite the longer contact time and the greater number of transients, we conclude that there is no significant dipolar coupling between acetophenone and cyclohexane, and that intramolecular dipolar coupling in the cyclocarbon is very weak. Given that the two molecules share the same supercages for the reaction to take place, we conclude that the local environment must be highly fluid so that the lifetime of a given acetophenone-cyclohexane pair must be sufficiently long for hydrogen transfer in the excited state (ca. >10−9 s), but shorter than the timescale needed for magnetization transfer, which is ca. 10−3 to 10−2 seconds. A great deal of rotational freedom is also suggested by the sharp line width of cyclohexane in Fig. 3 and by the low intensity of its signal, even after 31900 scans, as expected for a very inefficient intramolecular cross polarization.

When the photochemical and CPMAS results obtained with acetophenone are compared to those previously determined for benzophenone one may conclude that a relatively small difference in the volume of the two ketones translates in substantial fluidity changes in the zeolite. To support this, one can estimate the average number of reactant molecules that can fit within a supercage by considering their van der Waals volume with respect the volume of the supercage. Themolecular volumes (Vm) of benzophenone, acetophenone and cyclohexane can be estimated with a van der Waals increment approach first suggested by Bondixxiv with the values later refined by Gavezzotti.xxv The method assigns van der Waals volume to different molecular components (methyls, methylenes, aromatic methines, carbonyls, hydroxyls, halides, etc), which are added together to approximate the total volume of the molecule.xxiv,xxv Values (VM) of 177 Å3 and 120 Å3 are obtained for benzophenone and acetophenone, respectively, and a value of 101 Å3 for cyclohexane. The average free volume of the supercages of NaY is reportedxxvi to be VSC=839 Å3. To estimate the number of molecules that may be accommodated within a supercage, we assume the known packing coefficient of benzene in NaY as an upper limit. The packing coefficient, CK, was defined by Kitaigorodskixxvii as the ratio between the filled volume, based on the van der Waals molecular volume, and the total available volume. In the case on benzene, it has been determined that, on average, there are 5.4 benzene molecules per supercage.xxii With a van der Waals volume of 85.4 Å3 per benzene molecule, the packing coefficient of benzene becomes, CK=(5 × 85.4)/839 = 0.55. Assuming the same packing coefficient for the reactants in our study one may expect that supercage occupancies will occur with a number of molecules that have a total added volume VTOT ≈ VSC * CK = 462 Å3. While this volume is surpassed slightly with one benzophenone and three cyclohexane molecules for a VTOT=478 Å3, an analogous occupancy in the case of acetophenone would occur with a VTOT=421.4 Å3, which may help explain the significantly higher molecular and radical pair rotational fluidities.


Photochemical reactions in zeolites take place under conditions where the time scales for molecular and intermolecular degrees of freedom are strongly reorganized as compared to solution and most other reaction media. While slow diffusion allows for reactants and product to enter and exit the zeolite interior, bimolecular reactions take advantage of relatively long-lived cage effects. While reactions derived from random encounters of free radicals are efficiently prevented, the options available to the geminate radical pair depend on how the latter fits within the rigid environment of the zeolite supercages. Intermolecular cross polarization experiments indicate that acetophenone and cyclohexane have a significantly enhanced rotational freedom as compared with that of benzophenone and cyclohexane. The greater rotational fluidity allows the radical pair derived from the former reactant couple to explore reaction pathways that are not available to the latter.


This research was supported by NIH grant MARC U*STAR 2T34GM008415-16, Mount St. Mary’s College Professional Development Grant #1311-11, and NSF grant CHE0551938. We thank Drs. R. Taylor and J. Strouse from UCLA for help and advise.


i. Ramamurthy V, editor. Photochemistry in organized & Constrained Media. New York: VCH; 1991.
ii. (a) Turro NJ, Garcia-Garibay MA. In: Photochemistry in Organized and Constrained Media. Ramamurthy V, editor. New York: 1992. (b) Weiss RG, Ramamurthy V, Hammond GS. Acc Chem Res. 1993;26:530. (c) Hembury GA, Borovkov VV, Inoue Y. Chem Rev. 2008;108:1. [PubMed]
iii. (a) Schmidt GMJ. In: Solid State Photochemistry. Ginsburg D, editor. Verlag Chemie; 1976. (b) Cohen MD. Angew Chem Int Ed Engl. 1975;14:386. (c) Keating AE, Garcia-Garibay MA. In: Molecular and Supramolecular Photochemistry. Ramamurthy V, Schanze K, editors. New York: 1998. (d) Garcia-Garibay MA. Acc Chem Res. 2003;36:491. [PubMed]
iv. (a) Gessner F, Olea A, Lobaugh JH, Johnston LJ, Scaiano JC. J Org Chem. 1989;54:259. (b) Ramamurthy V, Corbin DR, Johnston LJ. J Am Chem Soc. 1992;114:3870. (c) Sykora M, Kincaid JR, Dutta PK, Castagnola NB. J Phys Chem B. 1999;103(2):309. (d) Incavo Joseph A, Dutta PK. J Phys Chem. 1990;94(7):3075.
v. Yoon KB, Kochi JK. J Am Chem Soc. 1989;111(3):1128–1130. (b) Yoon KB, Huh TJ, Corbin DR, Kochi JK. J Phys Chem. 1993;97:6492–6499.
vi. Turro NJ, Zhang Z. In: Photochemistry on Solid Surfaces. Anpo M, Matsuura T, editors. Amsterdam: 1989.
vii. Scaiano JC, Kaila M, Corrent S. J Phys Chem B. 1997;101:8564.
viii. (a) Joy A, Ramamurthy V. Chem Eur J. 2000;6:1287. [PubMed] (b) Sivaguru J, Natarajan A, Kaanumalle LS, Shailaja J, Uppili S, Joy A, Ramamurthy V. Acc Chem Res. 2003;36:509. [PubMed]
ix. (a) Schaefer J, Stejskal EO. Top Carb-13 NMR Spec. 1979;3:283. (b) Parmer JF, Dickinson LC, Chien CW, Porter RS. Macromolecules. 1987;20:2308. (c) Gobbi GC, Silvestri R, Russel TP, Lyerla JR, Fleming WW, Nishi T. J Polym Sci Part C: Polym Lett. 1987;25:61.
x. Cizmeciyan D, Sonnichsen LB, Garcia-Garibay MA. J Am Chem Soc. 1997;119:184.
xi. Garcia-Garibay MA. Curr Op Sol St Mat Sc. 1998;3/4:399.
xii. Dosa PI, Fu GC. J Am Chem Soc. 1998;120:445–446.
xiii. National Institute of Advanced Industrial Science and Technology. SDBS NO 11527. [accessed Aug 26, 2008].
xiv. Cahiez G, Luart D, Lecomte F. Org Lett. 2004;6:4395. [PubMed]
xv. National Institute of Advanced Industrial Science and Technology. SDBS NO 21853. [accessed Aug 26, 2008].
xvi. Dixon WT, Shaefer J, Sefcik MD, McKay RA. J Mag Res. 1982;49:341.
xvii. (a) Wagner PJ, Park BS. Org Photochem. 1991;11:227. (b) Scaiano JC. J Photochem. 1973/74;2:81.
xviii. Babcock BW, Dimmel DR, Graves J, David P. An analogous observation was reported. J Org Chem. 1981;46:736.
xix. Lei X, Turro NJ. J Photochem Photobio A. 1992;69:53.
xx. Garcia-Garibay MA, Lei X, Turro NJ. J Am Chem Soc. 1992;114:2749.
xxi. Benson HG, Hudson A. Mol Phys. 1970:185.
xxii. Massiot D, Fayon F, Capron M, King I, Le Calvé S, Alonso B, Durand JO, Bujoli B, Gan Z, Hoatson G. Mag Res Chem. 2002;40:70.
xxiii. Hartmann SR, Hahn EL. Phys Rev. 1962;128:2042.
xxiv. Bondi A. J Phys Chem. 1964;68:441.
xxv. Gavezzotti A. J Am Chem Soc. 1990;105:5220.
xxvi. Ramamurthy V, Corbin DR, Eaton DF. J Org Chem. 1990;55:5269.
xxvii. Kitaigorodskii AI. Molecular Crystals and Molecules. Academic Press; 1973.