Effects of surface adsorption and confinement on the photochemical selectivity of previtamin D3 adsorbed within porous sol-gel derived alumina

Article abstract of DOI:10.1021/ja9804767

It has been demonstrated that porous sol-gel derived (SGD) alumina can function as an effective microorganized medium in which to perform size- or shape-selective organic transformations. For the photolysis of 7- dehydrocholesterol (7D) encapsulated within SGD-alumina, it was found that the adsorption characteristics of 7D and the pore size of the SGD-alumina greatly influenced the photochemical selectivity of the cis-trans isomerization of previtamin D₃ to tachysterol₃. An increase in the ratio of previtamin D₃ (P₃) to tachysterol₃ (T₃) was obtained on photolysis of 7D adsorbed within SGD-alumina with a pore size of 10-20 ?. This result is attributed to a confinement effect that inhibits the cis-trans isomerization of P₃ to T₃. In addition to the confinement effect on the photochemical selectivity, it was found that the adsorption characteristics of 7D to the surface of SGD-alumina also play a critical role in the photochemical selectivity. The results indicate that confinement in addition to chemisorption via the 3-hydroxyl group of 7D to the surface of activated SGD- alumina is necessary to optimize the photochemical selectivity. The presented results also provide direct experimental evidence to demonstrate that adsorption solely to a metal oxide surface via the 3-hydroxyl group is not sufficient to inhibit the cis-trans isomerization of P₃ to D₃.

Full text of DOI:10.1021/ja9804767

VOLUME 121, NUMBER 21
J UNE 2, 1999
© Copyright 1999 by the
American Chemical Society
Effects of Surface Adsorption and Confinement on the Photochemical
Selectivity of Previtamin D3 Adsorbed within Porous Sol-Gel
Derived Alumina
Forrest S. Schultz*,† and Marc A. Anderson‡
Contribution from the Department of Chemistry and Water Chemistry Program, UniVersity of
WisconsinsMadison, Madison, Wisconsin 53706
ReceiVed February 11, 1998
Abstract: It has been demonstrated that porous sol-gel derived (SGD) alumina can function as an effective
microorganized medium in which to perform size- or shape-selective organic transformations. For the photolysis
of 7-dehydrocholesterol (7D) encapsulated within SGD-alumina, it was found that the adsorption characteristics
of 7D and the pore size of the SGD-alumina greatly influenced the photochemical selectivity of the cis-trans
isomerization of previtamin D3 to tachysterol3. An increase in the ratio of previtamin D3 (P3) to tachysterol3
(T3) was obtained on photolysis of 7D adsorbed within SGD-alumina with a pore size of 10-20 Å. This result
is attributed to a confinement effect that inhibits the cis-trans isomerization of P3 to T3. In addition to the
confinement effect on the photochemical selectivity, it was found that the adsorption characteristics of 7D to
the surface of SGD-alumina also play a critical role in the photochemical selectivity. The results indicate that
confinement in addition to chemisorption via the 3-hydroxyl group of 7D to the surface of activated SGD-
alumina is necessary to optimize the photochemical selectivity. The presented results also provide direct
experimental evidence to demonstrate that adsorption solely to a metal oxide surface via the 3-hydroxyl group
is not sufficient to inhibit the cis-trans isomerization of P3 to D3.
Introduction
molecular crystals,4 liquid crystals,5 zeolites,6 clays,7 cyclodex-
trins,8 organic polymers,9 nonporous silica and alumina,10 and
semiconductors11 are a few representative examples of micro-
organized media. Studies of these systems have produced
encouraging results that demonstrate the utility of carrying out
Many investigators have invested significant efforts to develop
systems of microorganized media to increase the selectivity of
chemical reactions. With these systems, chemical selectivity has
been achieved by confining a reactive chemical species within
an organized structure that can limit the conformational flex-
ibility of reactants or transition states leading to different
products. Micelles,1 vesicles,2 adsorbed mono- and multilayers,3
(3) Spooner, S. P.; Whitten, D. G. In Photochemistry in Organized and
Constrained Media; Ramamurthy, V., Ed.; VCH: New York, 1991; Chapter
15 and references therein.
(4) (a) Scheffer, J. R.; Pokkuluri, P. R. Chapter 5. In Photochemistry in
Organized and Constrained Media; Ramamurthy, V., Ed.; VCH: New York,
1991. (b) Zimmerman, H. E.; Zuraw, M. J. J. Am. Chem. Soc. 1989, 111,
2358. (c) Zimmerman, H. E.; Zuraw, M. J. J. Am. Chem. Soc. 1989, 111,
7974. (d) Ramamurthy, V.; Venkatesan, K. Chem. ReV. 1987, 87, 433-
481. (e) Zimmerman, H. E.; Zhu, Z. J. Am. Chem. Soc. 1995, 117, 5245-
5262.
† Department of Chemistry. Current address: Department of Chemistry,
University of Wisconsin-Stout, Menomonie, WI 54751.
‡ Water Chemistry Program.
(1) (a) Thomas, J. K.Chem. Rev. 1980, 80, 283. (b) Turro, N. J.; Cox,
G. S. Paczkowsky, M. A. Top. Curr. Chem. 1985, 129, 57. (c) de Mayo,
P.; Sydnes, L. K. J. Chem. Soc., Chem. Commun. 1989, 994.
(2) (a) Takagi, K.; Sawaki, Y.; Crit. Rev. Biochem. Mol. Biol. 1993, 28,
323-367. (b) Suddaby, B. R.; Brown, P. E.; Russell, J. C.; Whitten, D. G.
J. Am. Chem. Soc. 1985, 107, 5609-5617.
(5) Hrovat, D.; Lieu, J. H.; Turro, N. J.; Wiess, J. D. G. J. Am. Chem.
Soc. 1984, 106, 7033.
10.1021/ja9804767 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/14/1999

4934 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Schultz and Anderson
reactions in microorganized media. In many instances, the
chemoselectivity, regioselectivity, and/or stereoselectivity of
chemical reactions can be enhanced.
Unfortunately, many of these systems suffer from various
shortcomings that limit their ultimate practicality. Sol-gel
derived (SGD) alumina exhibits many characteristics that make
it an excellent material to explore as a microorganized medium
to affect the selectivity of shape- and/or size-selective organic
reactions. Thermal stability, chemical stability, mechanical
stability, optical transparency, narrow pore size distribution,
controllable pore size, and ease of thin film and monolith
preparation are all characteristics of SGD-alumina that distin-
guish it from other microorganized media systems that, quite
often, lack one or more of these desirable features.
In recent years, the synthesis and characterization procedures
for SGD-alumina materials have been well-established. This
work has examined the various aspects of hydrolysis, gel
formation, and sintering conditions for the preparation of
supported and unsupported alumina materials with controlled
pore sizes and pore size distributions.12 Although the results of
this work have established many intricacies of SGD-alumina
processing, these studies have focused on the application of
SGD-alumina for separations and catalysis. However, the recent
development of porous SGD-alumina as an effective separation
medium and catalyst has produced a material with definitive
pore sizes and shapes that more precisely confine organic
molecules than alumina materials of the past. Earlier publications
have presented the photochemistry and photophysics of adsorbed
organic species on alumina and silica; however, the alumina
and silica used in these studies are nonporous materials in which
any effects are attributed to surface adsorption, rather than
confinement within pores.10 We report in this paper the first
application of SGD-alumina as a microorganized medium in
Figure 1. Photochemical and thermal isomerizations of D3 isomers.
which surface adsorption and confinement within pores affects
the selectivity of a chemical reaction.
To examine the effectiveness of SGD-alumina as a micro-
organized medium, we studied the photochemical selectivity of
7-dehydrocholesterol (7D) encapsulated within the porous
structure of SGD-alumina. Figure 1 shows the series of
isomerizations that occur during the photochemical production
of vitamin D (D3) from 7D. The formation of D3 begins with a
photochemically allowed conrotatory electrocyclic ring opening
of 7D yielding previtamin D3 (P3). P3 is susceptible to
conrotatory electrocyclic ring closure to 7D, conrotatory elec-
trocyclic ring closure to lumisterol3 (L3), or cis-trans isomer-
ization to tachysterol3 (T3). D3 is produced from P3 via a
thermally allowed [1,7] sigmatropic hydrogen shift. Photolysis
of 7D establishes a quasi-photostationary equilibrium between
7D, P3, T3, and L3. Comprehensive studies have established the
solution-phase photochemistry of 7D and the related isomers.13
A key characteristic of the solution-phase photochemistry of
7D is the high selectivity toward T3 at many photolysis
wavelengths.14 A tremendous amount of research effort has been
directed toward understanding the photochemical selectivity of
the vitamin D system and ultimately improving the selectivity
toward previtamin D3.
(6) (a) Lem, G.; Kaprinidis, N. A.; Schuster, D. I.; Ghatlia, N. D.; Turro,
N. J.J. Am. Chem. Soc. 1993, 115, 7009. (b) Ramamurthy, V.; Eaton, D.
F.; Casper, J. V. Acc Chem. Res. 1992, 25, 299. (c) Ramamurthy, V. In
Photochemistry in Organized and Constrained Media; Ramamurthy, V.,
Ed.; VCH: New York, 1991; pp 429-493. (d) Ramamurthy, V.; Corbin,
D. R.; Turro, N. J.; Zhang, Z.; Garcia-Garibay, M. A. J. Org. Chem. 1991,
56, 255-261. (e) Turro, N. J.; Cheng, C. C.; Abrams, L. Corbin, D. R. J.
Am Chem. Soc. 1987, 109, 2449-2456.
(7) (a) Ogawa, M.; Kuroda, K. Chem. ReV. 1995, 95, 399-438. (b)
Thomas, J. K. Acc. Chem. Res. 1988, 21, 275-280. (c) Usami, H.; Takagi,
K.; Sawaki, Y. Chem. Lett. 1992, 1405-1408.
(8) (a) Ramamurthy, V.; Eaton, D. F. Acc. Chem. Res. 1988, 21, 300.
(b) Moorthy, J. N.; Venkatesan, K.; Wiess, R. G. J. Org. Chem. 1991, 56,
6957.
(9) (a) O’Brien, M. E.; Zimmerman, H. E. J. Org. Chem. 1994, 59,
1809-1816. (b) Horie, K.; Mita, I. AdV. Polym. Aci. 1989, 88, 77. (c) Guillet,
J. Chapter 5. Polymer Photophysics and Photochemistry; Cambridge
University Press: Cambridge, England, 1985.
(10) (a) Liu, Y. S.; de Mayo, P.; Ware, W. R. J. Phys. Chem. 1993, 97,
5995-6001. (b) Lednev, I. K.; Mathivanan, N.; Johnston, L. J. J. Phys.
Chem. 1994, 98, 11444-11451. (c) Thomas, J. K. Chem. ReV. 1993, 93,
301-320. (d) Ford, W. E.; Kamat, P. V. J. Phys. Chem. 1989, 93, 6423-
6428. (e) Johnston, L. J. Chapter 8. In Photochemistry in Organized and
Constrained Media; Ramamurthy, V., Ed.; VCH: New York, 1991.
(11) (a) Al-Ekabi, H. In Photochemistry in Organized and Constrained
Media; Ramamurthy, V., Ed.; VCH: New York, 1991; pp 495-534. (b)
Fox, M. A. Acc. Chem. Res. 1983, 16, 314-321.
(12) (a) Yoldas, B. E. Am. Ceram. Soc. Bull. 1975, 54, 286-288. (b)
Yoldas, B. E. Am. Ceram. Soc. Bull. 1975, 54, 289-290. (c) Yoldas, B. E.
J. Mater. Sci. 1975, 10, 1856-1860. (d) Leenaars, A. F. M.; Burggraaf, A.
J. J. Colloid Interface Sci. 1985, 105, 27-40. (e) Leenaars, A. F. M.; Keizer,
K.; Burggraaf, A. J. J. Mater. Sci. 1984, 19, 1077-1088. (f) Hsieh, H. P.;
Bhave, R. R.; Fleming, H. L. J. Membr. Sci. 1988, 39, 221-241. (g) Okubo,
T.; Watanabe, M.; Kusakabe, K.; Morooka, S. J. Mater. Sci. 1990, 25,
4822-4827. (h) Uhlhorn, R. J. R.; Huis In't Veld, M. H. B. J.; Keizer, K.;
Burggraaf, A. J. J. Mater. Sci. 1992, 27, 527-537. (i) Vendange, V.;
Columban, Ph. J. Mater. Res. 1996, 11, 518-528. (j) Chane-Ching, J.-Y.;
Klein, L. C. J. Am. Ceram. Soc. 1988, 71, 86-90. (k) Gieselmann, M. J.
Ph.D. Thesis, University of WisconsinsMadison, 1991.
In attempts to control the selectivity of D3 formation, several
studies have examined the photochemistry of 7D in association
with various microorganized media. These studies have included
confinement of 7D within epidermal layers and synthetic
phospholipid bilayers, adsorption to silica, and via micellar
formation. Although various results have pointed toward an
attenuation of T3 formation, substantial discrepancies exist in
the photochemical product distributions.15-19 The ratio of P3/
(13) (a) Havinga, E.; Sanders, M.; Pot, J. Fortchr. Chem. Org. Natur-
stoffen 1969, 29, 131-157. (b) Verloop, A.; Koevoet, A. L.; Van
Moorselaar, R.; Havinga, E. Recueil 1959, 78, 1004-1016.
(14) (a) Havinga, E.; Jacobs, H.; Gielen, J. Tetrahedron Lett. 1981, 40,
4013-4016. (b) Dauben, W.; Phillips, R. J. Am. Chem. Soc. 1982, 104,
5780-5781.
(15) Holick, M. F.; MacLaughlin, J. A.; Doppelt, S. H. Science 1981,
211, 590-593.
(16) (a) Moriarity, R. M.; Schwartz, R. N.; Lee, C.; Curtis, V. J. Am.
Chem. Soc. 1980, 102, 4257-4259. (b) Yamamoto, J. K.; Borch, R. F.
Biochemistry 1985, 24, 3338-3344.
(17) Terenetskaya, I. P.; Perminova, O. G.; Yeremenko, A. M. J. Mol.
Struct. 1990, 219, 359-364.
(18) Holick, M. F.; Richtand, N. M.; McNeill, S. C.; Holick, S. A.;
Frommer, J. E.; Henley, J. W.; Potts, J. T., Jr. Biochemistry 1979, 18, 1003-
1008.

Photochemical SelectiVity of PreVitamin D3
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 4935
microscope. Infrared (IR) spectra were recorded on a Nicolet Magna-
IR 750 Series II infrared spectrometer. Loading concentrations of the
SGD-alumina with 7D were determined with a Perkin-Elmer 2400 CHN
elemental analyzer.
Porous SGD-Alumina: Boehmite (γ-AlOOH) Sol.22 To 775 mL
(43.1 mol) of purified water at 80 °C was added 187 mL of 2 M ATSB
(0.374 mol) in 2-butanol. After refluxing and stirring for 2 h, 17 mL
of 1.6 M nitric acid (0.0272 mol) was added. After another 2 h of
refluxing and stirring, the condenser was removed and the 2-butanol
was allowed to distill away. Once the distillation temperature reached
100 °C and the odor of 2-butanol was no longer present, the condenser
was replaced and refluxing was continued overnight. The resulting sol
was filtered sequentially with a Whatman GF/F glass fiber filter and a
0.45 µm Membra-Fil membrane filter (Fisher Scientific). Xerogels were
formed in evaporating dishes constructed of optical quality acrylic
plastic. The dishes were placed in desiccator boxes containing
magnesium chloride hexahydrate as a drying agent. After 2 days of
drying, the xerogels were sintered at 2 °C/min to 400, 700, or 950 °C.
The resulting porous plates were 0.5 mm thick. Thin films of porous
alumina were prepared by dip-coating with the alumina sol. By
concentrating the freshly prepared sol by a factor of 3, 2 µm thick
films of SGD-alumina were obtained on Pyrex or quartz glass supports.
Film thickness was measured by profilometry and ultraviolet interfer-
ence fringe patterns.
SGD-alumina plates (ca. 3 cm × 3 cm) were utilized for the
photolysis experiments. The pore size of the SGD-alumina was varied
by sintering the plates at 400, 700, and 950 °C to achieve slitlike pores
of average widths of <20, <44, and 69 Å, respectively. Activation of
the SGD-alumina involves removal of surface-adsorbed water at
elevated temperatures in an atmosphere of air to generate a partially
dehydroxylated surface in which Lewis acidic aluminum sites are
exposed. For the incorporation of 7D, the plates were placed in a
solution containing 7D dissolved in chloroform. In those experiments
in which the plates were activated at 300 °C for 3 h, the plates were
added directly from a furnace to the solution. After adsorption
equilibrium was reached, the plates were removed from the solution,
rinsed with fresh solvent, and evacuated for 4 h at 1 Torr. The plates
were then placed in an immersion well photochemical apparatus (Ace
Glass, Inc.). A Hanovia 400 W medium-pressure lamp was used for
the photolysis experiments. A Vycor filter was used to isolate
wavelengths greater than 240 nm.
Figure 2. Prominent rotational conformations of P3.
T3 is highly dependent on the photolysis conditions; therefore,
it is difficult if not completely inaccurate to compare results of
one study to another without extreme attention given to
experimental detail. Achievement of a photostationary-state,
excitation wavelength, and attenuation of shorter wavelengths
by the microorganized media are aspects of these results that
are not definitive. In studies of human epidermal photolysis,
careful attention was given to the irradiation wavelengths, and
any effects on the product distribution were attributed to
attenuation of shorter wavelengths by the stratum corneum rather
than confinement effects.20
It is generally an accepted characteristic of the photochemistry
of P3 that the formation of T3 arises predominately from the
cZt rotamer, while the cZc rotamer leads to the formation of
D3 (see Figure 2). A recent computational study has addressed
the photolysis selectivity for 7D adsorbed on silica on the basis
of a substituent effect at position C3.21 In this study the
conformational selectivity between the cZc and cZt rotamer of
P3 is skewed toward the cZc rotamer by a bulky substituent at
C3. These data suggests that adsorption of 7D to a silica surface
via the 3-hydroxyl group would enhance the photochemical
selectivity toward P3. We will provide direct experimental
evidence in this paper that demonstrates that neither H-bonding
nor chemisorption via the 3-hydroxyl group to a SGD-alumina
surface are sufficient to inhibit T3 formation and affect the
photochemical selectivity.
The results of the research presented in this paper have two-
fold importance: (1) This research demonstrates the first report
of porous SGD-alumina functioning as an effective microor-
ganized medium to alter the selectivity of an organic transfor-
mation that is confined within the porous structure of SGD-
alumina. (2) The photochemical selectivity of the cis-trans
isomerization of P3 to T3 can be greatly altered by the confining
structure of SGD-alumina in conjunction with chemisorption.
Adsorption alone via the 3-hydroxyl to the surface of a metal
oxide such as SGD-alumina is not sufficient to inhibit the cis-
trans isomerization of P3 to T3.
The photolysis product distributions were determined with a Waters
HPLC system equipped with a UV photodiode array detector. A Zorbax
Sil column (4.6 mm i.d. × 25 cm) was used with a mobile phase of
0.4% 2-propanol in hexane with 0.05% water. UV spectral character-
istics, retention time, and independent samples established peak
identities. Since the desorption process required refluxing in ethanol,
it was assumed that the formation of any D3 was a result of thermal
isomerization from P3. The percentage of P3 in the photolysis mixtures
was taken as the summation of P3 and D3 amounts as determined by
HPLC.
Experimental Section
Results and Discussion
Aluminum tri-sec-butoxide (ATSB) (97%) and 2-butanol (99%) were
purchased from Aldrich Chemical Co. and used as received. 7-Dehy-
drocholesterol (Aldrich, 98%) was recrystallized from acetone. Purified
water was obtained from a Barnstead NANOpore UV ultrapure water
system. All solvents were of spectroscopic grade and were distilled
from calcium hydride. Nitrogen adsorption/desorption analyses were
performed with a Micromeritics ASAP 2010 instrument. Ultraviolet-
visible spectra were obtained with a Hewlett-Packard 8452A spectro-
photometer. Sol particles were sized by laser light scattering with a
Brookhaven Instruments Corp. BI-2030AT digital correlator and BI-
200SM goniometer equipped with a Lexal model 85 argon laser. Atomic
force microscopy (AFM) images were obtained with a Digital Instru-
ments Nanoscope III AFM/LFM instrument. Transmission electron
microscopy (TEM) images were obtained with a JEOL 200CX
Characterization of the Particles and Pores of SGD-
Alumina. The particle size distribution of the boehmite (γ-
AlOOH) sols was determined by dynamic light scattering
techniques. Figure 3 shows the size distribution of particles
based on the relative number of particles obtained with the
γ-AlOOH sol. From this figure it can be seen that the sol is
composed of particles with diameters in the range 5-8 nm. The
AFM images of the surface of SGD-alumina sintered at 400
°C are shown in Figure 4. The important features of these AFM
images are the particle size and shape. From these images it
can be seen that the particles are flat and oblong in shape, with
dimensions of 6 × 10 nm across the face of the particle and a
thickness of ca. 2 nm. The TEM micrograph of SGD-alumina
is shown in Figure 5. This micrograph reaffirms the particle
(19) Terenetskaya, I. P.; Perminova, O. G.; Yeremenko, A. M. J. Mol.
Struct. 1992, 267, 93-98.
(20) MacLaughlin, J. A.; Anderson, R. R.; Holick, M. F. Science 1982,
216, 1001-1003.
(22) Gieselmann, M. J. Ph.D. Thesis, University of WisconsinsMadison,
1991.
(21) Dmitrenko, O.; Reischl, W. Monatsch. Chem. 1996, 127, 445-453.

4936 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Schultz and Anderson
Figure 3. Relative number distribution of particle sizes in γ-A1OOH
sol.
Figure 6. Pore size distribution for SGD-alumina sintered at 400, 700,
and 950 °C based on the Kelvin equation and the method of BJH.
alumina was characterized by utilizing data from nitrogen
adsorption/desorption isotherms. Table 1 summarizes the char-
acteristics of the porous structure obtained for SGD-alumina
sintered at 400, 700, and 950 °C. The pore sizes were calculated
from Brunauer, Emmett, and Teller (BET) desorption data using
the Kelvin equation. These results reveal the relation between
sintering temperature and pore size. Figure 6 presents the pore
size distribution obtained from the Kelvin equation for SGD-
alumina sintered at 400, 700, and 950 °C in which the
corresponding average pore diameters are 36, 44, and 69 Å,
respectively. The pore size distributions were obtained with the
method of Barrett, Joyner, and Halenda (BJH) by utilizing the
desorption data of the isotherm.23 An important consideration
with the pore widths determined for the SGD-alumina is the
validity of the Kelvin equation at the pore dimensions present
within the SGD-alumina of this study.24 Studies have addressed
this breakdown in the Kelvin equation and have quantified the
error propagated through the calculations to derive a pore size
distribution.25 At pore diameters of less than 50 Å, the calculated
values are inflated by more than 50%. Therefore, the reported
pore size distributions in Figure 6 may be substantially elevated,
especially the distribution for the 400 °C sintered sample that
exhibits the smallest pore dimensions of the three samples. These
pores are calculated to have a width of 36 Å, a value that may
be inflated by more than 50%. On the basis of the approximate
dimensions of the platelike particles as determined by AFM of
6 × 10 × 2 nm, a slitlike pore width of 20 Å would be expected.
The calculated average pore width for the 700 °C sintered SGD-
alumina of 44 Å also falls in the region of uncertainty for the
Kelvin equation. Although the uncertainty in this value is not
as great as that for the 400 °C sintered SGD-alumina, the
average pore width for the 700 °C sintered SGD-alumina would
be expected to be much less than 44 Å.
Figure 4. AFM image of SGD-alumina surface sintered at 400 °C.
White bar ) 6 nm.
In the past few years, several computational methods have
been developed that address the difficulties in obtaining pore
size distributions of microporous materials. Horvath and Kawa-
zoe26 and Saito and Foley27 have developed mean field methods
that apply an average potential field across the adsorbate
molecules in the pore for slit pores and cylindrical pores,
respectively. The Horvath-Kawazoe method may provide a
more accurate distribution for smaller slit-pore materials.
Application of this model to the nitrogen sorption data reports
a pore width of 10 Å. Therefore, for the SGD-alumina sintered
(23) Barrett, E. P.; Joyner, L. G.; Halenda, P. H. J. Am. Chem. Soc.
1951, 73, 373.
(24) Deleted in proof.
Figure 5. TEM micrograph of SGD-alumina sintered at 400 °C.
(25) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity,
2nd ed.; Academic: New York, 1982; p 152-154.
(26) Horvath, G.; Kawazoe, K. J. Chem. Eng. Jpn. 1983, 16, 470-475.
(27) Saito, A.; Foley, H. C. AIChE J. 1991, 37, 429-436.
size as observed with the AFM image. From the TEM
micrograph, the packing of ca. 6 nm particles can be observed.
Nitrogen Sorption Studies. The porous structure of SGD-

Photochemical SelectiVity of PreVitamin D3
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 4937
Figure 8. UV spectra of 7D in various media: (a) ethanol, (b) diethyl
ether, (c) hexane, and (d) SGD-alumina.
Figure 7. Optical transmission of SGD-alumina plates 0.5 mm thick
sintered at (a) 400 °C, (b) 700 °C, and (c) 950 °C.
within the SGD-alumina samples of different pore dimensions
indicates that 7D adsorbs with a different geometric relation to
the SGD-alumina surface. Relating the adsorption quantity of
7D to the surface area of the SGD-alumina produces the
following surface area occupation values. For SGD-alumina
sintered at 400, 700, and 950 °C, the surface area occupation
values for 7D were 146, 80, and 70 Å2 per 7D molecule,
respectively. Molecular dimensions obtained from MM228 for
7D indicate that the cross-sectional areas for a plane perpen-
dicular to the steroidal ring structure and a plane parallel to the
ring structure are 40 and 120 Å2, respectively. These values
associated with the surface area occupation values indicate that
the adsorption geometry of 7D is of a parallel nature for SGD-
alumina sintered at 400 °C, while the adsorption geometry is
of a perpendicular nature for SGD-alumina sintered at 700 and
950 °C (see Figures 13 and 14).
Table 1. Characteristics of the Porous Structure of SGD-Alumina
Sintered at Various Temperatures
sintering
temp (°C)
pore diameter (Å)
Kelvin equation
BET surface
area (m2/g)
porosity
(%)
400
700
950
36
44
69
251
175
87
47.7
47.7
41.5
Table 2. Final Adsorption Concentrations of 7D Adsorbed in
Activated and Nonactivated SGD-Alumina Sintered at 400, 700, and
950 °C
g of 7D/g of SGD-alumina
sintering temp (°C)
no activation
activated at 300 °C
400
700
950
0.11
0.14
0.08
0.12
0.15
0.08
The surface area requirement for adsorption indicates that
the steroidal ring structure is parallel to the surface of the SGD-
alumina that has been sintered at 400 °C. This type of adsorption
geometry would also be expected for 7D adsorbed in a slitlike
pore in which the pore width is narrow enough to prevent the
perpendicular adsorption geometry as observed in the wider
pores of the 700 and 950 °C sintered SGD-alumina. Earlier it
was shown that the pore width for SGD-alumina sintered at
400 °C was in the range 10-20 Å. The longitudinal dimension
of 7D is 17 Å as determined from MM2 calculations. These
relative dimensions indicate that 7D would be required to adsorb
in a more parallel manner for SGD-alumina sintered at 400 °C.
Electronic Interactions with Surface of SGD-Alumina. It
has been reported in many instances that the photochemical
reactivity and selectivity of organic molecules adsorbed on
alumina or silica surfaces is affected by electronic interactions
with the oxide surface.10 Figure 8 shows the UV absorption
spectra for 7D dissolved in various organic solvents and
encapsulated within activated SGD-alumina. The absorbance
maxima are at 262, 272, 282, and 294 nm for all of the systems.
The difference in maximum intensities observed at lower
wavelengths for the SGD-alumina system is attributed to light
scattering by the porous structure of SGD-alumina. These UV
data indicate that electronic interactions between the diene
moiety of 7D and the surface of SGD-alumina are not prevalent.
Surface Adsorption Characteristics. Figure 9 shows the
diffuse-reflectance IR spectra of 7D in the region of 2000-
4000 cm-1. Spectrum a results from 7D that has been encap-
sulated within nonactivated SGD-alumina. The OH stretching
frequency occurs at 3471 cm-1. Spectrum b results from a
mixture of powdered SGD-alumina that has not been activated
and powdered 7D. The OH stretching frequency occurs at 3495
cm-1. As a comparison spectrum, spectrum c is a diffuse
at 400 °C, the pore size distribution can be expected to be in
the range 10-20 Å.
Optical Characteristics of SGD-Alumina. Figure 7 shows
the UV-visible transmittance spectra for SGD-alumina sintered
at 400, 700, and 950 °C. All of the materials exhibit excellent
optical transparency through a broad range of wavelengths. As
discussed above, one method of increasing the pore size in
alumina is to increase the sintering temperature. From this figure
it can be seen that, as the pore size increases with increased
sintering temperatures, the low-wavelength cutoff exhibits a shift
to longer wavelengths. This resulting shift is attributed to
increased Rayleigh scattering from the larger pores and particles
within the SGD-alumina structure.
Adsorption of 7-Dehydrocholesterol in SGD-Alumina.
Table 2 presents the equilibrium adsorption concentration of
7D adsorbed on SGD-alumina sintered at 400, 700, and 950
°C. The specific surface areas for the SGD-alumina samples
that were sintered at 400, 700, and 950 °C are 251, 175, and 87
m2/g, respectively, as presented above. The final adsorption
values for 7D indicate that adsorption is not directly related to
the specific surface area of the SGD-alumina sample. We believe
that this type of adsorption may arise if the pore dimensions in
the 400 °C sample are too small to provide access for 7D. It is
also possible that the adsorption of 7D in small pores occupies
a greater amount of surface area than adsorption in larger pores.
For smaller or narrower pores, the adsorption of 7D may be
envisioned to occur in such a manner in which the cholesterol
ring structure is essentially parallel to the surface of the platelike
particle. For larger, or wider, pores, the surface adsorption of
7D may occur in a manner in which the cholesterol ring structure
is perpendicular to the surface of the platelike particle. Calcula-
tion of the surface area occupied by each 7D molecule adsorbed
(28) Allinger, N. L. J. Comput. Chem. 1993, 14, 755-768.

4938 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Schultz and Anderson
Figure 10. Percent composition of isomers on photolysis of 7D in
hexane with Vycor filter: 7D (b), T3 (0), P3 ([), and L3 (O).
Figure 9. Diffuse-reflectance IR spectra of (a) 7D adsorbed on SGD-
alumina, not activated, (b) 7D mixed with SGD-alumina, (c) 7D KBr
pellet, and (d) 7D adsorbed on SGD-alumina, activated.
Table 3. Percent Compositions Resulting from the Photolysis of
7D Encapsulated in SGD-alumina of Different Pore Dimensions and
Activation Conditions
reflectance measurement of a KBr pellet of 7D. The OH
stretching frequency for this spectrum remains at 3495 cm-1
.
% composition
Spectrum d results from 7D that has been encapsulated from
solution into activated SGD-alumina. The OH stretching band
has essentially been eliminated.
SGD-alumina
% loading
of 7D
sintering temp (°C) activation
7D P3
52 35
95 2.6 1.9 0.2 1.4
63 20 15 1.4
T3 L3 P3/T3
hexane
6
7
1.5
400
400
400
400
700
700
700
700
700
950
950
950
950
ambient
ambient
300 °C
300 °C
ambient
ambient
300 °C
300 °C
300 °C
300 °C
300 °C
ambient
ambient
9.4
3.5
8.9
1.3
11
3.7
1.4
6.2
12
0.7
6.9
6.9
1.9
The surface characteristics of alumina are significant in the
binding that occurs between 7D and the SGD-alumina. Reported
spectroscopic studies have indicated that activated alumina can
cleave the heteroatom-H bond of alcohols and of primary and
secondary amines with concomitant attachment of the alkoxide
or amide groups to the alumina surface.29 The diffuse-reflectance
IR spectra shown in Figure 9 indicate that the 3-hydroxyl group
of 7D is chemisorbed onto a coordinatively unsaturated Lewis
acidic aluminum site on the surface of the activated SGD-
alumina. However, on the SGD-alumina that has not been
activated, the 3-hydroxyl group is hydrogen bonded to surface
hydroxyl groups of the SGD-alumina. It has been shown that
the surface hydroxyl groups of alumina serve as hydrogen bond
donors rather than hydrogen bond acceptors.30
2
96 3.4 0.3 0.2 11
11 62 24 2.2 2.5
85 7.5 4.9 0.9 1.5
69 18 12 2.2 1.5
57 26 15 4.2 1.7
70 18 11 1.1 1.6
95
3
1.3 0.3 2.3
40 36 20 4.3 1.8
84 10
76 15
5.0 0.8 2.0
7.8 1.2 1.9
34 38 23 4.9 1.7
Desorption Characteristics. It was found that the desorption
characteristics of 7D and its photoproducts were solvent- and
temperature-dependent as expected for sorption processes. The
greatest percent recovery was obtained in refluxing ethanol, in
which the percent recovery after photolysis was essentially
quantitative. Any unrecovered material is believed to be the
result of the formation of the highly insoluble bischolestadienols.
UV-visible spectral analysis of the plates after photolysis and
extraction exhibits data characteristic of those reported for the
bischolestadienols.31
Photolysis of 7D Encapsulated within SGD-Alumina. The
effect on the product distribution obtained from the photolysis
of encapsulated 7D in SGD-alumina was determined in relation
to the product distribution obtained from the photolysis of 7D
in hexane solution. The solution photochemistry of 7D in hexane
with broad-band irradiation >240 nm is shown in Figure 10.
The isomeric composition at the photostationary state as
tabulated in Table 3 is 6% 7D, 52% P3, 35% T3, and 7% L3.
Figures 11 and 12 show the percent composition of isomers
as a function of time for the photolysis of 7D encapsulated in
SGD-alumina that has been sintered at 400 °C. In this photolysis
experiment, the loading level of 7D was 9.4% by mass and the
SGD-alumina was not activated. These conditions represent the
limiting conditions concerning the time required for attainment
Figure 11. Percent composition of isomers on photolysis of 7D
encapsulated in non-activated SGD-alumina sintered at 400 °C: 7D
(b), T3 (0), P3 ([), and L3 (O).
of a photostationary state. From these figures it can be seen
that the photostationary state is obtained after ca. 4 h of
irradiation. For other systems in which the SGD-alumina was
activated or the loading level was decreased, the achievement
of a photostationary state occurred at shorter irradiation times.
For subsequent SGD-alumina systems, the time required for
achieving a photostationary state were independently deter-
mined. For systems at low loading levels, photolysis times of
only several minutes were necessary for the achievement of a
photostationary state. Table 3 summarizes the percent composi-
tion of the photostationary state obtained for the photolysis of
7D encapsulated in SGD-alumina of different pore dimensions,
loading percentage, and activation condition.
(29) (a) DeCanio, E. C.; Nero, V. P.; Bruno, J. W. J. Catal. 1992, 135,
444-457. (b) Parera, J. M.; Figoli, N. S. J. Catal. 1969, 14, 303-304.
(30) Breton, G. W.; Daus, K. A.; Kropp, P. J. J. Org. Chem. 1992, 57,
6646-6649.
(31) Boomsma, F.; Jacobs, H. J. C.; Havinga, E.; Van Der Gen, A.
Recueil 1973, 92, 1361-1367.
We believe that the confinement effect on the product
distribution from the photolysis of 7D is not solely a result of

Photochemical SelectiVity of PreVitamin D3
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 4939
dimensions of the 400 °C sample are not small enough to inhibit
the rotation of the A ring system, but the pore dimensions are
small enough to inhibit the rotation of the C/D ring system.
For the 700 °C and 950 °C SGD-alumina, the pore dimensions
of these materials are not capable of imposing sufficient
confinement on P3 to prevent the cis-trans isomerization to
T3.
At higher loading levels it is believed that 7D adsorbs to the
surface of activated SGD-alumina sintered at 700 and 950 °C
in a Langmuir-Blodgett manner. This arrangement at high
loading levels provides sufficient confinement to inhibit the cis-
trans isomerization to T3. The source of confinement results
from the proximity of surface-adsorbed molecules. The chemi-
sorption through the 3-hydroxy group is important since it
provides a source of restriction that does not exist in the
nonactivated systems or the solid-state crystal system. At lower
loading levels sufficient distance exists between the adsorbed
molecules to allow the cis-trans isomerization to T3.
For the 400 and 700 °C SGD-alumina that were not activated
before adsorption, the P3/T3 ratio was essentially unchanged
from that obtained with the solution photochemistry. The
increase in the ratio with the nonactivated 950 °C SGD-alumina
is attributed to excitation at longer wavelengths which greatly
favors the formation of P3.14 Earlier it was shown that the larger
pore 950 °C SGD-alumina exhibited a lower wavelength cutoff
of ca. 300 nm.
Figure 12. Percent composition of isomers on photolysis of 7D
encapsulated in nonactivated SGD-alumina sintered at 400 °C: 7D (b),
T3 (0), P3 ([), and L3 (O).
In terms of conformational control, we believe that the
formation of the tZc rotamer of P3 is highly unfavorable for P3
that is chemisorbed to the surface of SGD-alumina. Although
the distribution of rotamers is highly shifted toward the cZc
rotamer, we believe that the formation of T3 occurs from the
cZc conformation as demonstrated by those systems in which
confinement is minimal. In a recent study of low-temperature
photochemistry of P3, it was noted that T3 could arise from either
conformer.32 Our experimental evidence strongly supports this
statement.
Figure 13. Steric interactions arising from the transition-state con-
formation required for cis-trans isomerization of P3 to T3 in SGD-
alumina sintered at 400 °C and activated.
Solid-State Photolysis of 7D. It has been demonstrated by
many researchers that the solid-state photochemistry of organic
crystals can greatly affect product distributions when compared
to the corresponding results of solution photochemistry.4 These
results are attributed to the crystal packing structure that forms
a microorganized medium that restricts the conformational
flexibility of the molecules during reaction. If the selectivity of
the photochemistry of 7D is affected by confinement within a
microorganized medium, then it is quite possible that the solid-
state photochemistry of 7D would show deviations from the
solution photochemical results. Since the photolysis of 7D is
performed in the absence of solvent, the solid-state photoreac-
tivity of 7D was also examined. Figure 15 shows the percent
composition obtained for the photolysis of a film of 7D with
light of wavelengths >240 nm. The films were prepared by
dip-coating a glass tube in a concentrated solution of 7D
dissolved in chloroform. The isomeric compositions are shown
in Table 4. Since T3 and P3 are oils, on formation of these
products any confinement integrity of the crystal structure would
be expected to be greatly diminished. Since the product
distribution on photolysis of 7D is directly dependent on the
photochemistry of P3, it may not be appropriate to consider this
solid-state photochemistry. However, since the oil of P3 is a
condensed phase, it is quite possible that intrinsic confinement
effects may affect the product distribution.
Figure 14. Steric interactions arising between co-adsorbed molecules
in SGD-alumina sintered at 700 and 950 °C at the transition state
required for cis-trans isomerization of P3 to T3.
adsorption or chemisorption of 7D to the SGD-alumina surface.
For activated SGD-alumina sintered at 400 °C, we believe that
the pore dimensions are small enough to inhibit the cis-trans
isomerization of P3 that is chemisorbed to the SGD-alumina
surface (see Figure 13). Rotation about the C6-C7 bond can
be accomplished by rotation of the A-ring or the C/D ring system
or by a combination of rotation of these two systems. The
anchoring of the A-ring by chemisorption at the 3-hydroxyl of
7D to the surface of the SGD-alumina requires rotation about
the C6-C7 bond to be performed by the much larger C/D ring
system of the cholesterol molecule. For 7D, which is hydrogen-
bonded to the surface of the nonactivated SGD-alumina sintered
at 400 °C, the surface adsorption is not strong enough to prevent
rotation about the C6-C7 bond by the smaller A ring system
of P3. From these results it can be concluded that the pore
Although the results of the photolysis of 7D in the solid state
indicate that different mechanistic concerns may affect the
(32) Mu¨ller, A. M.; Lochbrunner, S.; Schmid, W. E.; Fuss, W. Angew.
Chem., Int. Ed. Engl. 1992, 31, 505-507.

4940 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Schultz and Anderson
T3. Larger pore SGD-alumina sintered at 700 and 950 °C, with
pore dimensions >20 Å, can also inhibit the cis-trans isomer-
ization of P3 to T3 when the loading level of 7D is sufficiently
high enough for coadsorbed molecules to impose an additional
source of confinement. Although surface adsorption is expected
to occur predominantly via the cZc rotamer of P3, the presented
results demonstrate that adsorption of 7D via the 3-hydroxyl in
the absence of additional confinement is not sufficient to inhibit
the cis-trans isomerization of P3 to T3.
Although SGD-alumina does possess many of the features
required of a microorganized medium, some limitations do exist.
In the case of photochemical reactions, “self-filtering” of light
is a prominent limitation for highly concentrated systems. Also,
scale-up of this methodology is at present a limitation. However,
we expect that the flexibility in processing methods of sol-gel
derived materials will allow the construction of flow through
reactors that utilize thin films of the ceramic material. In the
case of alumina, the surface can be catalytically active toward
a variety of molecules and reactions. This activity can create
other products, but in properly chosen systems, the activity can
be utilized to promote desirable reactions. The size and shape
of confining pores is obviously important to control size-
selective reactions. We have found that the pore shape of SGD-
alumina matches well to the space required to control the cis-
trans isomerization of P3 to T3. Yet we do not expect any and
all size-selective reactions to be affected by SGD-alumina.
Current studies are exploring other classes of organic reactions
with a variety of space requirements.
Figure 15. Photolysis of 7-dehydrocholesterol film with Vycor filter:
7D (b), T3 (0), P3 ([), and L3 (O).
Table 4. Percent Composition of Photolysis of 7D Film
% composition
photolysis medium
7D
P3
T3
L3
P3/T3
hexane
7D thin film
6
64
52
17
35
17
7
2.4
1.5
1.0
product distribution, it can be concluded that the observed
selectivity with the SGD-alumina system is not a direct
consequence of solid-state or condensed-phase photochemistry.
The SGD-alumina provides additional conformational flexibility
control on 7D and the isomers derived from it.
Conclusion
Acknowledgment. We thank the EPA through the Water
Resources Center for funding of this project (Electrically Biased
Photocatalytic Ceramic Membranes for Treatment of Impaired
Waters). We thank Dr. Isabel Tejedor for her help with SGD-
alumina production and IR interpretation.
The results presented in this paper demonstrate the first report
of the ability of SGD-alumina to function as an effective
microorganized media to alter the selectivity of an organic
transformation. We have shown that the pore dimensions of
activated SGD-alumina sintered at 400 °C, with pore dimensions
of 10-20 Å, can inhibit the cis-trans isomerization of P3 to
JA9804767