Supramolecular Steric Effects as the Means of Making Reactive Carbon Radicals Persistent. Quantitative Characterization of the External Surface of MFI Zeolites through a Persistent Radical Probe and a Langmuir Adsorption Isothez

Article abstract of DOI:10.1021/jo991298i

The photochemistry of tetraphenylacetone (1) adsorbed on the external surface of a MFI zeolite (the sodium form of LZ-105) has been investigated in combination with computational chemistry, surface area measurements, EPR analysis, and classical adsorption isotherms. All of the methods are consistent with a supramolecular structural model in which 1 is first adsorbed strongly through intercalation of a single benzene ring into a hole on the LZ-105 external surface (site I) followed by a weaker binding to the external framework between the holes (site II) until a monolayer of 1 is formed. From both computational and surface area measurements, it is estimated that the site I holes on the external surface will be filled at ca. 0.3-0.5 wt %/wt loading of 1/LZ-105, which corresponds to 6.5 × 10¹⁸ (ca. 10^-5 mol) of holes or molecules of 1 adsorbed in holes per gram of zeolite. The supramolecular composition of ca. 0.3-0.5% of 1 on LZ-105 characterizes a "break point" for the photochemistry and the EPR measurements, since it represents the value for saturation of the site I holes with 1. These conclusions are supported quantitatively by experimental isotherms of the adsorption of 1 on LZ-105. Photolysis of 1 intercalated in the site I holes causes fragmentation into two isomeric supramolecular diphenylmethyl (DPM) radicals, one (DMP)_in which is adsorbed into the internal surface and becomes strongly persistent (half-life of many weeks) and the other (DMP)_ex which diffuses on the external surface and rapidly dimerizes (less than a few minutes) to produce the radical-radical combination product tetraphenylethane (2). Photolysis of 1 adsorbed on the solid external surface produces two supramolecularly equivalent DPM radicals (DMP)_ex that diffuse on the external surface and rapidly dimerize to produce 2, and do not produce persistent DPM radicals. * To whom correspondence should be addressed.

Full text of DOI:10.1021/jo991298i

J . Org. Chem. 2000, 65, 1319-1330
1319
Su p r a m olecu la r Ster ic Effects a s th e Mea n s of Ma k in g Rea ctive
Ca r bon Ra d ica ls P er sisten t. Qu a n tita tive Ch a r a cter iza tion of th e
Exter n a l Su r fa ce of MF I Zeolites th r ou gh a P er sisten t Ra d ica l
P r obe a n d a La n gm u ir Ad sor p tion Isoth er m †
Takashi Hirano,‡,§ Wei Li,‡ Lloyd Abrams,| Paul J . Krusic,| M. Francesca Ottaviani, and
Nicholas J . Turro*,‡
Department of Chemistry, Columbia University, New York, New York 10027, E. I. duPont de Nemours
and Co., Central Research Department, Experimental Station, Wilmington, Delaware 19880, and
University of Urbino, Institute of Chemical Sciences, Piazza Risorgimento, 6, 61029 Urbino, Italy
Received August 17, 1999
The photochemistry of tetraphenylacetone (1) adsorbed on the external surface of a MFI zeolite
(the sodium form of LZ-105) has been investigated in combination with computational chemistry,
surface area measurements, EPR analysis, and classical adsorption isotherms. All of the methods
are consistent with a supramolecular structural model in which 1 is first adsorbed strongly through
intercalation of a single benzene ring into a hole on the LZ-105 external surface (site I) followed by
a weaker binding to the external framework between the holes (site II) until a monolayer of 1 is
formed. From both computational and surface area measurements, it is estimated that the site I
holes on the external surface will be filled at ca. 0.3-0.5 wt %/wt loading of 1/LZ-105, which
corresponds to 6.5 × 1018 (ca. 10-5 mol) of holes or molecules of 1 adsorbed in holes per gram of
zeolite. The supramolecular composition of ca. 0.3-0.5% of 1 on LZ-105 characterizes a “break
point” for the photochemistry and the EPR measurements, since it represents the value for
saturation of the site I holes with 1. These conclusions are supported quantitatively by experimental
isotherms of the adsorption of 1 on LZ-105. Photolysis of 1 intercalated in the site I holes causes
fragmentation into two isomeric supramolecular diphenylmethyl (DPM) radicals, one (DMP)in which
is adsorbed into the internal surface and becomes strongly persistent (half-life of many weeks) and
the other (DMP)ex which diffuses on the external surface and rapidly dimerizes (less than a few
minutes) to produce the radical-radical combination product tetraphenylethane (2). Photolysis of
1 adsorbed on the solid external surface produces two supramolecularly equivalent DPM radicals
(DMP)ex that diffuse on the external surface and rapidly dimerize to produce 2, and do not produce
persistent DPM radicals.
In tr od u ction
such as methyl radical CH3• under the same conditions,
i.e., a kinetic definition characterizes radical persistence.
Using these criteria, TPM may be characterized as being
both “stable” and “persistent”.
Ra d ica l Sta bility a n d Ra d ica l P er sisten ce. The
discovery of triphenylmethyl (TPM) by Gomberg in 1900
launched the field of “carbon-centered” free radical
chemistry.1 TPM exists in measurable equilibrium with
its dimer, which allows for its direct observation by EPR
spectroscopy in fluid solution at room temperature. This
feature of TPM has caused chemists to classify TPM as
a “stable” radical. Ingold has differentiated the concept
of “stability” and “persistence” of carbon-centered radi-
cals.2 “Stability” is the term recommended for a radical
R• that is associated with a R-H bond strength that is
less than a comparable C-H bond strength in a compa-
rable alkane, i.e., a thermodynamic definition character-
izes radical stability. On the other hand, “persistence” is
suggested for a radical R• when the R• has a lifetime
significantly greater than that of a comparable radical
In ordinary solutions, the persistence of a carbon-
centered radical is determined by physical processes such
as diffusion, as well as chemical processes such as the
rate of radical-radical reactions. As a first and good
approximation, the persistence of a carbon-centered
radical in fluid solution at room temperature in an inert,
deaerated, solvent is determined only by the rate of
radical-radical reactions (combination and dispropor-
tionation), and these are often diffusion-controlled bimo-
lecular reactions. Ingold2 has argued that the persistence
of most carbon-centered radicals is primarily a conse-
quence of intramolecular steric factors that slow radical-
radical reactions. Indeed, in achieving persistence through
steric interactions, severely twisted triaryl radicals, such
as TPM, sacrifice most of the resonance stabilization
energy due to electron delocalization.
* To whom correspondence should be addressed.
† E. I. duPont de Nemours and Co. contribution no. 7931.
‡ Columbia University.
Ingold’s ideas concerning intramolecular steric effect
have proven to be effective in the synthetic design of
persistent carbon-centered radicals.2,3 In this report we
employ the ideas4 of supramolecular chemistry for the
§ Present address: The University of Electro-Communications,
Department of Applied Physics and Chemistry, Chofu, Tokyo 182,
J apan.
| E. I. duPont de Nemours and Co.
University of Urbino.
(1) Gomberg, M. Ber. 1900, 33, 3150-3163; J . Am. Chem. Soc. 1900,
22, 757-771.
(3) McBride, J . M. Tetrahedron 1974, 30, 2009-2022.
(4) Lehn, J . M. Supramolecular Chemistry; VCH: New York, 1995.
(2) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1976, 9, 13-19.
10.1021/jo991298i CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/06/2000

1320 J . Org. Chem., Vol. 65, No. 5, 2000
Hirano et al.
synthetic design of carbon-centered radicals that are
rendered persistent by intermolecular, rather than intra-
molecular factors, i.e., supramolecular steric effects. The
first goal of our studies is to design supramolecular
persistence for benzyl and related radicals that are very
reactive in solution at room temperature (i.e., those which
manifest diffusion-controlled, radical-radical reactivity
in “molecular” solutions). In addition, we seek conditions
for which radical diffusion is allowed but for which
supramolecular steric effects greatly inhibit the rate of
radical-radical reactions, thereby creating radical per-
sistence.
Zeolites a s Su p r a m olecu la r En vir on m en ts for
Rea ctive Ca r bon -Cen ter ed Ra d ica ls. Zeolites are
fascinating materials for industrial purposes such as
molecular sieves and catalysts.5-8 From the standpoint
of supramolecular chemistry,4 zeolites possess a well-
organized, crystalline, “hard matter” microenvironment
for performing chemical reactions, namely the rigid
aluminosilicate external and internal framework. How-
ever, in addition to the “hard” walls provided by the
zeolite framework, zeolites possess “soft” void space (i.e.,
holes, channels, intersections) that is available for the
adsorption of molecules possessing appropriate steric
size/shape characteristics. In this study we employ as a
supramolecular host the sodium form of LZ-105, a zeolite
possessing the MFI topology. The latter topology pos-
sesses an interconnected three-dimensional void space
consisting of pores and channels of ca. 5.5 Å diameter
(Figure 1a,b), and intersections of ca. 9 Å diameter.9,10
As supramolecular guests, we have derivatives of diben-
zyl ketones which, upon photolysis, produce benzyl that
fits in the “soft” holes of the external surface and can
diffuse within the internal channels and intersections of
the internal framework at room temperature.11-17
An important requirement for persistence of a mobile
guest radical in a supramolecular environment is that it
does not possess a significant chemical reactivity with
F igu r e 1. Schematic representation of the channel structure
in a crystal and the hole structure of the [001] external surface
of the MFI zeolite (top), and the molecular dimension of 1 (a),
DP M (b), and two DP M radicals (c).
Sch em e 1
(5) Davis, M. E.; Lobo, R. F. Chem. Mater. 1992, 4, 756-768.
(6) Meier, W. M.; Olson, D. H. Atlas of Zeolite Structure Types;
Butterworth-Heinemann: London, 1992.
(7) Breck, D. W. Zeolite Molecular Sieves: Structure, Chemistry, and
Use; Wiley and Sons: London, 1974.
the host. This is the case for MFI zeolites as hosts, since
the framework of these zeolites consists of strong SiOSi
and SiOAl bonds. Furthermore, if the cation that com-
pensates the negative charge associated with the Al atom
is sodium, there generally is not a Lewis acid effect for
hydrocarbon radicals, although at some point the electron-
donating capability of a radical may become significant
and cations may result.
A Mod el for Ad sor p tion of Or ga n ic Molecu les
on to MF I Zeolites. A working model for adsorption of
an organic molecule, M, onto a MFI zeolite is shown
schematically in Scheme 1. In the first step of adsorption,
a molecule Ms is adsorbed from solution (the solvent,
isooctane, was selected since it is not adsorbed into the
internal framework channel network) onto one of two
sites of the external surface: either a hole (Mh ) or the
external framework (Mf). The holes and external frame-
work are termed the “external surface”. The model
(8) Szostak, R. Molecular Sieves: Principle of Synthesis and Iden-
tification; Van Nostrand Reinhold: New York, 1989.
(9) (a) Kokotailo, G. T.; Lawton, S. L.; Olson, D. H.; Meier, W. M.
Nature (London) 1978, 272, 437-438. (b) Olson, D. H.; Kokotailo, G.
T.; Lawton, S. L.; Meier, W. M. J . Phys. Chem. 1981, 85, 2238-2243.
(10) Flanigen, E. M.; Bennet, J . M.; Grose, R. W.; Cohen, J . P.;
Patton, R. L.; Kirchner, R. M.; Smith, J . V. Nature (London) 1978, 272,
512-516.
(11) (a) Turro, N. J .; Cheng, C.-C.; Mahler, W. J . Am. Chem. Soc.
1984, 106, 5022-5023. (b) Turro, N. J .; Cheng, C.-C.; Abrams, L.;
Corbin, D. R. J . Am. Chem. Soc. 1987, 109, 2449-2456.
(12) Reviews: (a) Turro, N. J . Pure Appl. Chem. 1986, 58, 1219-
1228. (b) Ramamurthy, V.; Turro, N. J . J . Incl. Phenom. Mol. Recognit.
Chem. 1995, 21, 239-282.
(13) Photochemistry of dibenzyl ketone: (a) Engel, P. S. J . Am.
Chem. Soc. 1970, 92, 6074-6076. (b) Robbins, W. K.; Eastman, R. H.J .
Am. Chem. Soc. 1970, 92, 6076-6077. (c) Robbins, W. K.; Eastman,
R. H.J . Am. Chem. Soc. 1970, 92, 6077-6079.
(14) Photochemistry of tetraphenyl acetone: (a) Quinkert, G.; Opitz,
K.; Weirdorff, W. W.; Weinlich, J . Tetrahedron Lett. 1963, 1863-1868.
(b) Quinkert, G. Pure Appl. Chem. 1964, 9, 607-621.
(15) (a) Turro, N. J .; Gould, I. R.; Baretz, B. H. J . Phys. Chem. 1983,
87, 531-532. (b) Gould, I. R.; Baretz, B. H.; Turro, N. J . J . Phys. Chem.
1987, 91, 925-929.
assumes that absorption is dynamic (Figure 1, k1, k-1
,
(16) Lunazzi. L.; Ingold, K. U.; Scaiano, J . C. J . Phys. Chem. 1983,
87, 529-530.
(17) (a) Turro, N. J .; Lei, X.; Li, W.; McDermott, A.; Abrams, L.;
Ottaviani, M. F.; Beard, H. S. Chem. Commun. 1998, 695-696. (b)
Turro, N. J .; McDermott, A.; Lei, X.; Li, W.; Abrams, L.; Ottaviani, M.
F.; Beard, H. S.; Houk, K. N.; Beno, B. R.; Lee, P. S. Chem. Commun.
1998, 697-698.
the rates of adsorption at the sites on the surface need
not be equal) and reaches an equilibrium between
adsorption at the two sites and the solution in contact
with the zeolite. The internal channels and intersections
(Scheme 1) are termed collectively as the internal surface.

Supramolecular Steric Effects on Carbon Radicals
J . Org. Chem., Vol. 65, No. 5, 2000 1321
If its size/shape characteristics are appropriate, the
molecule Mh will eventually diffuse (k2) into the internal
surface, leading to a molecule that is adsorbed in the
internal channels (Mc) or intersections (Mi). Molecules
that initially adsorb on the external framework (Mf) must
first diffuse and be adsorbed into a hole in order to be
adsorbed into the internal surface i.e., Mf f Mh f Mi
(or Mc). It is important to note that Mf, Mh , Mc, and Mi
each corresponds to distinct supramolecular structures.
Since these distinct supramolecular structures correspond
to the same supramolecular composition, but differ with
respect to the manner in which the guest is “connected
to the host”, these structures are supramolecular isomers.
Estim a tion of th e Moles of Holes on th e Exter n a l
Su r fa ce of MF I Zeolites. An important structural issue
is whether the holes (site I) or the external framework
(site II) bind a guest molecule M more strongly. It is of
interest to estimate the total number of holes on the
external surface in order to anticipate, experimentally,
the loading at which the holes will be filled by M, i.e., to
determine the maximum moles of holes, Mh , that are
possible per gram of zeolite. The moles of holes may be
computed quantitatively, and bounds can be placed on
the number of molecules adsorbed in the holes on the
external framework from the estimated or measured
surface area of a MFI zeolite and knowledge of the
crystalline parameters of MFI zeolites as follows.
loading is expected to correspond to a composition at
which the supramolecular structure changes from Mh /
LZ-105 to Mf/LZ-105, where the notation refers to dif-
ferent supramolecular isomers.
Molecu la r P h otoch em istr y of 1,1,3,3-Tetr a p h en yl
Aceton e (1) a n d 1,1,3-Tr ip h en yl Aceton e (3). The
photochemistry of dibenzyl ketone (DBK) derivatives has
proven to be useful for characterizing the supramolecular
structures and dynamics of molecular adsorbents and
radicals produced by photolysis.11,12 The molecular pho-
tochemistry of DBK and its derivatives consists of
adsorption of light to form a singlet, S1, followed by
intersystem crossing and R-cleavage from a triplet T1 to
produce two benzyl radical derivatives (after rapid elimi-
nation of CO).13-16 The molecular photochemistry of the
benzyl radicals in fluid solution consists, nearly exclu-
sively, of diffusion-controlled radical combination reac-
tions with a cage effect close to zero, i.e., essentially no
geminate pair reactions occur. In contrast, the supramo-
lecular photochemistry11,12 of DBK in zeolites shows
substantial geminate cage effects, which can be directly
related to simple ideas of “supramolecular steric effects”
imposed by the supramolecular “cage” produced by the
“hard walls” of the zeolite framework.
1,1,3,3,-Tetraphenylacetone, 1, and 1,1,3-triphenyl-
acetone, 3, were selected as precursors of diphenylmethyl
radicals (DPM). These molecules were selected because
their size/shape characteristics are much larger (Figure
1) than the characteristic size/shape of the holes on the
external surface. 1 is photolyzed to yield 1,1,2,2-tetraphen-
ylethane (2), quantitatively (eq 1), and 3 is photolyzed
to yield (eq 2) a statistical mixture of 2 (25%), 4 (50%),
and 5 (25%) expected from random coupling of the DPM
and benzyl (BZ) radicals produced by photolysis. The
molecular photochemistry of both 1 and 3 in nonviscous
solution consists of formation of radicals that undergo
quantitative random radical-radical coupling.
Scanning electron microscope analysis of LZ-105 re-
veals that the zeolite consists of crystals (particles) whose
size is of the order of ca. 0.3 µm. An order of magnitude
estimation of the surface area of 1.0 g of the sample can
be made if it is assumed that the shape of the particles
is cubic. The volume of a 0.3 µm crystal is ca. 2.7 × 10-20
m3. From the volume of a single particle and the crystal
density of LZ-105 which is 1.8 g/mL () 1.8 × 106 g m-3),
1.0 g of LZ-105 will consist of 2.1 × 1013 particles whose
external surface area is estimated to be ca. 11 m2 g-1
.
The external surface area of samples of LZ-105 was
measured directly by mercury porosimetry independently
and yielded a value for the surface area of 15 m2 g-1
.
Thus, both the estimated and experimental values of the
external surface area are completely consistent with one
another and provide confidence that a reliable surface
area can be associated with the samples of LZ-105.
The number of holes on the LZ-105 external surface
may now be estimated from knowledge of the surface area
and from the dimensions of the unit cell of silicalite as a
model MFI zeolite.9 The size of the unit cell of silicalite
is 20.4 Å × 20.4 Å × 13.8 Å, and the surface area of a
unit cell is calculated to be 1960 Å2. There are eight holes
per unit cell, four of which are the entrance to the
straight channels (Figure 1), and the other four are the
entrance to the sinusoidal channels (Figure 1). Therefore,
the number density is 8 holes/1960 Å2, or 1 hole per 245
Å2 () 2.45 × 10-18 m2). The value 245 Å2 is defined as a
“unit surface area” of a hole that is available to an
adsorbed molecule. By using this value and the measured
external surface area (15 m2 g-1), the number of holes
on the external surface of this sample of LZ-105 is
computed to be 6.5 × 1018 holes g-1(ca. 10-5 mol). The
fraction of external surface corresponding to site I holes
is ca. 20-30%, the remaining 70-80% being the site II
external framework. We are finally in a position to
compute the adsorbed amount of M (molecular weight
362) to be ca. 0.4 mg per 100 mg of LZ-105 (0.3-0.5 wt
%/wt loading), when all the holes were filled by M. This
1 and 3 represent molecules, M (Scheme 1), that cannot
diffuse into the internal surface. Thus, 1 and 3 can only
be adsorbed at site I (holes) and site II (framework). We
shall now show that the siting of 1 and 3 has profound

1322 J . Org. Chem., Vol. 65, No. 5, 2000
Hirano et al.
Ta ble 1. P r od u ct An a lysis of th e P h otor ea ction of 1
Ad sor bed on LZ-105
Sch em e 2
yield/% (conversion yield/%)
loading, irradiation conversion,
%
time/min
%
2
Ph2CO
0.10
0.10
0.30
0.30
0.51
0.50
0.70
0.70
1.10
1.10
5
10
5
10
5
10
5
10
5
36
38
22
30
21
33
26
35
27
38
16 (43)
13 (35)
10 (44)
13 (43)
10 (45)
13 (40)
14 (52)
18 (50)
15 (56)
23 (62)
2
3
1
1
trace
trace
1
trace
1
10
trace
and the photolysis of 3 in n-hexane gave tetraphenyl-
ethane (2), triphenylethane (4), and diphenylethane (5)
in 19, 33, and 17% yields (essentially a statistical
coupling), respectively.
The photoreaction of 1/zeolites has already been in-
vestigated by laser flash photolysis in the zeolite systems
of NaX and silicalite, as well as silica gel. DPM is
produced and is transient with decay times ranging over
many orders of magnitude.18 The products reported were
2 and benzophenone.
and decisive effects on the photochemistry of these
different supramolecular structures.
Ap p lica tion of th e Mod el to th e P h otoch em istr y
of 1. From the molecular dimensions of 1 and previous
results,17 it is expected (Scheme 2) that 1 will first adsorb
in the site I holes (1h ) of the external surface of LZ-105
to form a supramolecular complex, 1h /LZ-105. From the
considerations discussed above, the holes will be filled
at ca. 0.3-0.5% loading. Further adsorption beyond this
composition will occur at the site II external framework
to form a new supramolecular structure, 1f/LZ-105, until
a monolayer is formed. Since the supramolecular struc-
ture of 1/LZ-105 is expected to change at ca. 0.3-0.5%
loading of 1, it is of interest to determine if the photo-
chemistry of 1/LZ-105 was a function of loading and also
to determine if some special photochemical effects oc-
curred at the point expected for the completion of the
filling of the holes with 1.
From computer simulations, the 1h/LZ-105 species
possess a structure in which one benzene ring is inter-
calated in a hole and the remainder of the molecule
extends in a direction away from the surface. Photolysis
of 1h/ LZ-105 will produce two distinct supramolecular,
isomeric diphenylmethyl radicals (DPM), one intercalated
in the hole (DPMh) and the other adsorbed on the
external framework (DPMf). The DPMh radicals are
predisposed to diffusion into the internal surface, whereas
the DPMf radicals are disposed to diffusion on the
external framework. The radicals on the external frame-
work are expected to undergo dimerization to form
tetraphenylethane (2). However, the radicals entering the
internal surface are expected to be persistent, since the
largest space in the internal surface corresponds to the
intersections possessing a diameter of ca. 9 Å, whereas
the smaller molecular cross section of 2 is ca. 11 Å (Figure
1). We shall show below that the photochemistry of 1 and
of triphenylacetone, 3, support these supramolecular
structural considerations. Furthermore, we shall show
that EPR and classic adsorption isotherms quantitatively
confirm the conclusion that the holes are first filled by 1
and 3 and that the holes are completely filled at ca. 0.3-
0.5% loading of 1 and 3.
Samples of 1/LZ-105 and 3/LZ-105 were prepared by
adsorption of 1 (or 3) onto LZ-105 from an isooctane
solution, followed by removal of the bulk of the solvent
by drying with a stream of argon gas followed by vacuum
degassing at ca. 5 × 10-5 Torr to remove the remainder
of the isooctane in order to prepare a “dry” sample.
After exposing the photolyzed 1/LZ-105 to air, the
products were extracted by soaking overnight with THF
containing methanol (5% v/v) and then were analyzed by
GC. The yields of the products are summarized in Table
1. Tetramethylethane 2 was obtained as the main
product in ca. 40-45% yield, in addition to a trace
amount of benzophenone. The yield of 2 was independent
of the percentage of loading up to ca. 0.5%. Above 0.5%
loading of 1 on LZ-105, the yield of 2 increased to 50-
60%.
In the case of the photoreaction of 3/LZ-105 (0.5%
loading), followed by exposure to air and extraction with
THF containing methanol (5% v/v) GC analysis, 2, 4, and
5 were the major detectable products and were formed
in yields of 4, 13, and 5%, respectively.
P h otolysis of 1/LZ-105 a n d 3/LZ-105. EP R An a ly-
sis. Irradiated samples of 1/LZ-105 or 3/LZ-105 exhibited
strong EPR spectra (Figure 2). In the case of irradiated
1/LZ-105 the EPR spectrum (Figure 2a) could be confi-
dently assigned to the diphenylmethyl radical (DPM)
through simulation (Figure 2d) with hyperfine coupling
constants, aR ) 14.8 G, ao ) 3.7 G, am ) 1.1 G, ap ) 4.1
G (Figure 2d) comparable to reported values.19,20 The EPR
signals of DPM were persistent for many days under
vacuum at room temperature, and sharpened slightly
with time (Figure 2b), indicating that the radicals moved
from the internal surface sites of lower mobility (e.g., the
channels of the internal framework) to sites of higher
mobility (e.g., the intersections of the internal frame-
(18) (a) Kelly, G.; Willsher, C. J .; Wilkinson, F.; Netto-Ferreira, J .
C.; Weir, A. O. D.; J ohnston, L. J .; Scaiano, J . C. Can. J . Chem. 1990,
68, 812-819. (b) J ohnston, L. J .; Scaiano, J . C.; Shi, J .-L.; Siebrand,
W.; Zerbetto, F. J . Phys. Chem. 1991, 95, 10018-10024.
(19) Bassindale, A. R.; Bowles, A. J .; Hudson, A.; J ackson, R. A.;
Schreiner, K.; Berndt, A. Tetrahedron Lett. 1973, 3185-3186.
(20) Livingston, R.; Zeldes, H.; Conradi, M. S. J . Am. Chem. Soc.
1979, 101, 4312-4319.
Resu lts
As expected from literature reports (see eqs 1 and 2),
photolysis of 1 in solution yielded 2 in quantitative yield,

Supramolecular Steric Effects on Carbon Radicals
J . Org. Chem., Vol. 65, No. 5, 2000 1323
F igu r e 3. EPR spectra (a, b, c) of the product of the
photoreaction of 3 adsorbed on LZ-105. Spectra a and b of the
radicals observed under degassed conditions 5 min and 5 h
after irradiation of 3 for 10 min, respectively. Spectrum c was
obtained by subtracting b from a. The simulated spectrum d
of benzyl radical was calculated by using the parameters in
the text.
F igu r e 2. EPR spectra (a, b, c) of the products of the
photoreaction of 1 adsorbed on LZ-105. Spectra a and b of DPM
were observed under degassed conditions 5 min and 5 h after
irradiation of 1 for 10 min, respectively. Spectrum c was
observed 1 min after exposing the DPM sample to air. The
simulated spectrum d was calculated by using the parameters
in the text.
From the time dependence of the conversion of Figure
3a to Figure 3b, it is estimated that the benzyl radicals
produced from photolysis of 3/LZ-105 possess a half-
lifetime of ca. 40 min at room temperature, a time scale
for which DPM is completely persistent. From a quan-
titative evaluation of the EPR signal intensities of DPM
produced from photolysis of 1 and 3, it was estimated
that the amount of DPM generated from 3 on LZ-105 was
approximately half of the amount of DPM from 1 on LZ-
105.
Tim e-Dep en d en ce of th e Gen er a tion of P er sis-
ten t DP M fr om th e P h otor ea ction of Tetr a p h en yl-
a ceton e on LZ-105 a s a F u n ction of Loa d in g. The
above product and EPR analyses provide evidence that
the working model for the photolysis of 1/LZ-105 and
3/LZ-105 (Schemes 1 and 2) is valid and that the holes
are filled first and are saturated with 1 and 3 at ca. 0.3-
0.5% loading. Further support for the model is available
from a quantitative investigation of the time dependence
of the EPR signal intensity of DPM radicals produced
by photolysis of 1.
Figure 4 shows the time dependence of the generation
of DPM as a function of the percentage of loading of 1 on
LZ-105. Figure 5 shows that the EPR signal intensities
of the persistent DPM obtained by irradiation for 10, 30,
and 90 min correlated with the percentage of loading of
1 on LZ-105. However, a clear break in the sensitivity of
the signal intensity to loading occurs at ca. 0.3-0.5%
loading. For example, in the case of 0.1% loading, the
yield of DPM reached the maximum by irradiation for
40 min and gradually decreased with longer irradiation
over 40 min. Above 0.2% loading, the maximum amount
of DPM was obtained by irradiation around 90 min.
Importantly, the amount of the persistent DPM linearly
increased with an increase in the loading amount of 1
below 0.3-0.5% loading (Figure 5). Significantly, above
0.3-0.5% loading, the amount of the persistent DPM
increased only slightly with an increase in the percentage
of loading of 1. The results of 90 min irradiation shown
work). Upon exposure of the sample to air, the signals of
DPM immediately disappeared and were replaced with
a new spectrum possessing a larger g-factor and aniso-
tropy (Figure 2). This new species is assigned to a peroxy
radical (Figure 2c).21 The EPR signal of the peroxy radical
gradually decreased at room temperature. Formation of
the peroxy radical is reversible: if the sample is subjected
to a vacuum, the spectrum of the DPM radical reappears,
replacing the spectrum of the peroxy radical.
An irradiated sample of 3/LZ-105 under degassed
conditions also exhibits strong EPR signals (Figure 3).
For 1, except for a small amount of broadening, the
spectral features of the EPR spectrum were constant with
time. However, the EPR spectrum produced from the
photolysis of 3 changed significantly with time. After 5
h, the EPR spectrum (Figure 3b) matched the spectra of
DPM obtained by the photoreaction of 1/LZ-105 (Figure
2a,b). Evidently, one (or more) radicals in addition to
DPM is produced by photolysis of 3; the EPR of this other
radical can be detected along with that of DPM, but the
radical is not as persistent as DPM. The obvious candi-
date for a transient radical is the benzyl radical (BZ),
which is expected to be produced (eq 2) along with DPM
in the photolysis of 3. Indeed, subtraction of the spectrum
in Figure 3b (5 h after photolysis) from Figure 3a (5 min
after photolysis) yields the spectrum shown in Figure 3c.
Figure 3d shows the spectrum of BZ, simulated from the
reported hyperfine coupling constants; aR) 16.4 G, ao)
6.2 G, am) 1.8 G, ap) 5.2 G (Figure 3d).20,22-24 Agreement
between the simulated (Figure 3d) and subtracted spec-
trum (Figure 3c), together with expectation of the radical
products from photolysis of 3 leave no doubt that Figure
3a represents a mixture of BZ and DPM radicals.
(21) Schlick, S.; Kevan, K. J . Phys. Chem. 1979, 83, 3424-3429.
(22) Dixon, W. T.; Norman, R. O. C. J . Chem Soc. 1964, 4857-4860.
(23) Fischer, H. Z. Naturforsch. A 1965, 20, 488-489.
(24) Neta, P.; Schuler, R. H. J . Phys. Chem. 1973, 77, 1368-1370.

1324 J . Org. Chem., Vol. 65, No. 5, 2000
Hirano et al.
F igu r e 4. Time-course of the generation of DPM by the
photoreaction of 1 adsorbed on LZ-105 with various loading
amounts.
F igu r e 6. Time-course of the adsorbed amount of 1 on 100
mg of LZ-105 from an isooctane solution of 1 (initial concentra-
tion: 1.0 mg of 1 in 3 mL).
F igu r e 5. The EPR signal intensity of DPM generated by
irradiation of 1/LZ-105 for 10 (a), 30 (b), and (c) 90 min vs
loading percentage of 1 on LZ-105.
in Figure 5c shows that the slope above 0.3-0.5% loading
was steeper than the others. These results are consistent
with the working model: the supramolecular structure
of 1/LZ-105 changes from 1h/LZ-105 to 1f/LZ105 at
loading of ca. 0.3-0.5%, the point at which the holes on
the external surface are becoming completely filled. After
this point, further adsorption of TPM begins to occur on
the solid framework between the holes. It may be
concluded at this point that photolysis of 1f/LZ-105
produces DPMf that diffuse and combine to form 2 on
the external surface rather than be adsorbed into the
internal surface to form persistent DPM radicals.
La n gm u ir Isoth er m s for th e Ad sor p tion of Tet-
r a p h en yla ceton e (1) Ad sor bed on LZ-105. Having
obtained photochemical and EPR evidence for the work-
ing model, we add further support through Langmuir
isotherm analysis of the adsorption of 1, 3, and other
model compounds on LZ-105.
F igu r e 7. Langmuir isotherm analysis of 1 adsorbed on the
LZ-105 surface from an isooctane solution. A (y axis) is the
amount (mg) of 1 adsorbed on 100 mg of LZ-105 at equilibrium,
and C (x axis) is the amount (mg) of 1 left in a 3 mL solution.
Curves a and b were calculated by using the Langmuir
isotherm eqs 1 and 2 in the text, respectively.
reached equilibrium in ca. 40 min. The Langmuir iso-
therm for adsorption of 1 on LZ-105 consists of a plot
(Figure 7) of the amount of 1 remaining in solution (x-
axis) versus the amount of 1 adsorbed on 100 mg of LZ-
105 at equilibrium (y-axis). An attempt (Figure 7a) was
made to fit the experimental data (Figure 7 b) to the one
site Langmuir isotherm (eq 3)
Adsorption of the molecules on the LZ-105 external and
internal surfaces may be evaluated quantitatively by a
Langmuir isotherm analysis.25 In a typical experiment,
1 was adsorbed on 100 mg of LZ-105 from an isooctane
solution of 1 (1.0 mg of 1 in 3 mL of isooctane) at room
temperature. The time dependence of the adsorption of
1 is shown in Figure 6. Under these conditions, the
adsorption of 1 on LZ-105 from the isooctane solution
A ) MkC/(1 + kC)
(3)
where A is the amount (mg) of adsorbate adsorbed on
100 mg of LZ-105 at equilibrium, M is the maximum
amount adsorbed on 100 mg of LZ-105 at equilibrium, k
is the Langmuir constant that is related to an equilibrium
constant, and C is the amount (mg) of adsorbate left in 3
mL of solution at equilibrium. It is clear from Figure 7a,
that a single site model does not give a good fit to the
data.
(25) Oscik, J .; Cooper, I. L. Adsorption; Ellis Horwood Ltd, Chich-
ester, 1982.

Supramolecular Steric Effects on Carbon Radicals
J . Org. Chem., Vol. 65, No. 5, 2000 1325
Ta ble 2. La n gm u ir Isoth er m An a lysis of th e Ad sor p tion
of 1, 3, 6, a n d 7 on LZ-105 fr om Isoocta n e Solu tion s by
Usin g th e Tw o-Ad sor p tion Site Mod el
A: Calculated by Using the Value of the Adsorbed
Amount (mg) on 100 mg of LZ-105
site I
site II
ketone
MI, mg
kI
MII, mg
kII
1
3
6
7
0.33 ( 0.03
0.4 ( 0.2
0.40 ( 0.02
0.25 ( 0.03
300 ( 100
200 ( 400
350 ( 50
120 ( 50
0.50 ( 0.05
0.6 ( 0.2
0.59 ( 0.02
0.49 ( 0.07
1.6 ( 0.7
1 ( 2
2.1 ( 0.3
0.8 ( 0.4
B: Calculated by Using the Value of Adsorbed
Amount (mol) per “Unit Surface Area” of LZ-105
site I
site II
F igu r e 8. Langmuir isotherm analysis of 3, 6, and 7 adsorbed
on the LZ-105 surface from an isooctane solution.
ketone MI′, mol
kI′
MI′, mol
kII′
1
3
6
7
0.93 ( 0.01 (10.8 ( 0.6) × 107 1.52 ( 0.02 (4.1 ( 0.3) × 105
1.74 ( 0.01 (8.0 ( 0.2) × 107 2.51 ( 0.01 (5.1 ( 0.1) × 105
1.01 ( 0.02 (3.0 ( 0.2) × 107 1.91 ( 0.02 (2.2 ( 0.1) × 105
This finding led us to expect that correlations could be
found between the molecular structure of adsorbed mol-
ecules and the supramolecular binding parameters ex-
tractable from a Langmuir analysis. Although 3 and
dibenzyl ketone derivatives 6 and 7 have larger molecular
cross sections (ca. 6-7 Å) than the diameters of the
channels of LZ-105, as for 1, the benzene (ca 5 Å) portion
of the molecular structures of 3, 6, and 7 can fit into the
holes on the external surface. Thus, it is expected that
3, 6, and 7 will possess Langmuir isotherms that fit the
two-site model for adsorption on the external surface.
1.3 ( 0.1
(7 ( 2) × 107
1.90 ( 0.07 (3.2 ( 0.9) × 105
However, since the paradigm of Schemes 1 and 2
suggests that there should be two sites of differing
binding energies (the holes and the solid surface), it is
natural to expect that a two-site model, rather than a
one-site model is required to fit the experimental data.
Therefore, we modified the Langmuir isotherm to take
this inference into account and applied the data to eq 4
A ) MIkIC/(1 + kIC) + MIIkIIC/(1 + kIIC)
(4)
where sites I and II have the maximum amounts ad-
sorbed on LZ-105 at equilibrium MI and MII and the
Langmuir constants kI and kII, respectively. Figure 7b
shows that, as expected from the two-site model, eq 4
provides an excellent fit to the experimental data. The
values of MI, MII, kI, and kII were obtained from the fitting
curve b as summarized in Table 2A and, according to
Scheme 1, provide quantitative information on the bind-
ing strengths and fractions of 1 molecules bound in the
two sites, i.e., the supramolecular structure of the external
surface of 1/ LZ-105. We are now in a position to test this
expectation quantitatively and to then correlate the
results with the photochemical and EPR results.
From Langmuir analysis, the binding constant, kI, of
site I is a hundred times larger than kII of site II,
indicating that site I has a much stronger adsorption
affinity for 1 than site II on the LZ-105 external frame-
work surface. According to the working model, site I is
the holes on the external surface and site II is the solid
framework between the holes. From Langmuir analysis,
the maximum amount of 1 adsorbed on site I of the LZ-
105 surface is ca. 0.3 mg per 100 mg of LZ-105, which
corresponds to 0.3-0.5% loading, experimentally consis-
tent with the loading expected for filling the holes from
both computation and the results of photochemical and
EPR analysis. Thus, a quantitative working model of
Schemes 1 and 2 are supported by excellent agreement
between data from several independent sources: theo-
retical computations, photochemical product analysis,
EPR analysis, and classical isotherm analysis.
Indeed, isotherms for adsorption of 3, 6, and 7 were
well fitted by eq 4 (Figure 8). Adsorption constants
extracted from eq 4 are summarized in Table 2A. The
binding constants for 3, 6, and 7 were similar to those of
1, supporting evidence for the existence of two adsorption
sites which are structurally similar for this class of DBK
derivatives: site I (holes) which have ca. 100 times
stronger adsorption ability for the benzene moiety of each
molecule than does site II (solid surface). The constants
MI, MII, kI, and kII were also described as the values MI′,
MII′, kI′, and kII′. The data in Table 2B are also expressed
in terms of a “unit surface area” as a unit of 245 Å2, which
is the area of a hole.
Tim e Dep en d en ce of th e Ad sor p tion of Dip h en -
ylm eth a n e a n d Rela ted Str u ctu r es on LZ-105. Ad-
sorption of diphenylmethane, 8, on LZ-105 was employed
as an isosteric model for DPM in Langmuir studies. In
addition, o-xylene (9), p-xylene (10), triphenylmethane
(11), acetophenone (12), and 2′-methylacetophenone (13)
were studied as model compounds to investigate specific
molecular structural features of adsorption on LZ-105.
p-Xylene and acetophenone were employed as model
small molecules (hydrocarbon and ketone) which should
be readily adsorbed into the internal surface of LZ-
105.9,26-28 On the other hand, o-xylene, triphenylmethane,
and 2′-methylacetophenone were used as model sterically
bulky molecules, in the supramolecular sense, which
cannot be adsorbed into the internal surface of LZ-105,
La n gm u ir Isoth er m s for th e Ad sor p tion of De-
r iva tives of Tetr a p h en yla ceton e a n d Rela ted Str u c-
tu r es Ad sor bed on LZ-105. Given the above results, it
appears that classical Langmuir analysis is capable of
providing information on the supramolecular structure
of molecules adsorbed on the external surface of LZ-105.
(26) van Koningsveld, H.; Tuinstra, F.; van Bekkum, H.; J ansen, J .
C. Acta Crystallogr., Sect. B 1989, B45, 423-431.
(27) Fyfe, C. A.; Strobl, H.; Kokotailo, G. T.; Kennedy, G. J .; Barlow,
G. E. J . Am. Chem. Soc. 1988, 110, 3373-3380.
(28) Reischman, P. T.; Schmitt, K. D.; Olson, D. H. J . Phys. Chem.
1988, 92, 5165-5169.

1326 J . Org. Chem., Vol. 65, No. 5, 2000
Hirano et al.
F igu r e 9. Adsorption time-course of diphenylmethane and its related compounds on 100 mg of LZ-105 from an isooctane solution.
The time-course (A) for o-xylene (9), triphenylmethane (11), and 2′-methylacetophenone (13) (initial concentration: 2.1 mg in 3
mL); the time-course (B) for diphenylmethane (8) (initial concentration: 2.1 mg in 3 mL); and the time-course (C) for p-xylene
(10) and acetophenone (12) (initial concentration: 10.0 mg in 3 mL).
Sch em e 3
Ta ble 3. La n gm u ir Isoth er m An a lysis of th e Ad sor p tion
of Dip h en ylm eth a n e-Rela ted Com p ou n d s on 100 m g of
LZ-105 fr om Isoocta n e Solu tion sa
LZ-105 from an isooctane solution. The following qualita-
tive features are apparent from Table 3. (1) Molecules
whose steric cross sections allow access to the internal
surface (p-xylene and acetophenone) showed an M value
over 4 mg on 100 mg of LZ-105, a large amount of
adsorption consistent with entry into the extensive
internal surface. (2) Molecules whose steric cross sections
are too bulky to allow access to the internal surface (o-
xylene, triphenylmethane, and 2′-methylacetophenone)
showed an M value below 0.6 mg on 100 mg of LZ-105, a
value consistent with the formation of a monolayer on
the external surface which was estimated and measured
to have a surface area of ca. 15 m2/g. (3) Carbonyl-
containing compounds tend to have a larger M value than
that of a corresponding hydrocarbon compound of a
similar molecular size, e.g., the M value of acetophenone
is larger than that of p-xylene. (4) The maximum ad-
sorbed amount (1.4 mg/100 mg of LZ-105) of diphenyl-
methane observed in Figure 9c is much larger than the
M value of 2′-methylacetophenone, although diphenyl-
methane has no carbonyl group, consistent with internal
adsorption.
compound
M, mg
k
triphenylmethane
p-xylene
0.19 ( 0.01
4.6 ( 0.2
0.05
4 ( 1
130 ( 30
b
o-xylene
acetophenone
2′-methylacetophenone
9.5 ( 0.3
0.59 ( 0.01
90 ( 20
2.4 ( 0.2
a
Constants M and k were calculated by using the Langmuir
equation 3.
but which contain a benzene ring that can be adsorbed
in the holes on the external surface. The adsorption of
8-13, as a function of time, on LZ-105 (100 mg) from
isooctane solutions at room temperature is shown in
Figure 9. For the results in Figure 9, the initial concen-
trations of diphenylmethane, o-xylene, triphenylmethane,
and 2′-methylacetophenone were 2.1 mg/3 mL, respec-
tively, and for p-xylene and acetophenone, the initial
concentrations were 10.0 mg/3 mL, respectively. In the
case of o-xylene, triphenylmethane, and 2′-methylac-
etophenone, adsorption reached an equilibrium in 10 min
(Figure 9a). In the case of p-xylene and acetophenone, a
fast adsorption was observed in 10 min followed by a
slower adsorption, with equilibrium being achieved in ca.
2 h (Figure 9c). Diphenylmethane also showed a fast
adsorption process in 10 min, but then adsorption
increased relatively slowly and linearly achieving a
maximum amount in 10 h (Figure 9b).
Adsorption properties of compounds 8-13 were also
analyzed using a Langmuir isotherm analysis. In con-
trast to the results for 1, 3, 6, and 7, for which eq 4 was
required for a good fit, the isotherms for the other
molecules, with the exception of o-xylene, were fitted by
using eq 3, and the M and k values were summarized in
Table 3. The Langmuir constants k of o-xylene could not
be determined, because of its low adsorption ability on
Discu ssion
For 1 and 3 the results of computations, surface area
measurements, photochemical product analysis, EPR
analysis and adsorption isotherms are all consistent with
the expectations of Schemes 1 and 2. It is assumed that
these ketones are initially adsorbed in the holes (Mh /LZ-
105) of the external surface until ca. 0.3-0.5% loading
and then on the external framework (Mf/LZ-105), until
a monolayer is formed at ca. 1% loading, and that
photolysis of these distinct supramolecular isomers pro-
duces different supramolecular DPM radicals, DMPh and
DPMf.
Scheme 3 depicts the situation schematically. The
photoreaction of 1 generates two DPM radicals on the

Supramolecular Steric Effects on Carbon Radicals
J . Org. Chem., Vol. 65, No. 5, 2000 1327
Sch em e 4
external surface; one of the radicals intercalated in a hole
on the surface (DMPh) is predisposed to diffuse into the
internal surface, and the second radical of the pair
(DPMf) becomes detached from its partner and diffuses
relatively freely on the external surface to form 2 in ca.
40-45% yield. This is reasonable, since at ca. 0.3-0.5%
loading, most of the holes on the external surface are
likely to remain filled by Mh until very high conversions
are achieved. Since there is no evidence for a rapid
disappearance of DPM on the time scale of the EPR
experiments (minutes), we conclude that the combination
reactions of DPM on the external surface are completed
in the time scale of minutes. The time scale of adsorption
of diphenylmethane, an isosteric analogue of DPM, is
relatively slow (Figure 9b), supporting the conclusion that
any radicals produced on the external surface are much
more likely to undergo radical-radical reactions on the
external surface, which provides no significant steric
constraints to diffusion or radical-radical coupling.
The DPMh radicals that diffuse essentially quantita-
tively into the internal surface are strongly persistent
and exhibit a readily measured EPR spectrum (Figures
2 and 3). Addition of air to the sample results in rapid
oxidation by oxygen to produce peroxy radicals that are
readily detectable by EPR (Figure 2).
persistence demands that any radical-radical reactions
are strongly inhibited when the DPM radicals are ad-
sorbed in the internal surface. Two mechanisms can
produce persistence: (1) immobilization and spatial
separation of the radicals once adsorbed on the internal
surface, and (2) mobility of the radicals, but inhibition
of radical-radical reactions due to a supramolecular
steric effect as the radicals approach each other.
Photolysis of 3 produces a DPM radical and a benzyl
radical. From a quantitative analysis of the EPR results,
it was concluded that there is no difference in the initial
amounts of DPM and benzyl radicals produced by pho-
tolysis of 3 adsorbed into the internal surface. These
results suggest that there is no significant radical size
selectivity for DPM or benzyl radicals with respect to the
probability of adsorption into the internal surface by
DPM or BZ (Scheme 4). This conclusion, in turn, suggests
that DPM radicals can diffuse in the internal surface and
are not immobilized. In addition, the relative sharpness
of the EPR spectrum of DPM is similar to that for other
radicals (such as the R-methyl benzyl radical17) which
have been shown unambiguously to undergo inhibited
but finite reaction on the internal surface. Thus, although
circumstantial, the evidence favors a certain mobility of
the DPM radicals.
The following lines of reasoning provide further sup-
port for the proposed models of Schemes 1-3. The
product analysis of the photoreaction of 1 adsorbed on
LZ-105 showed that up to ca. 0.5% loading the yield of 2
is essentially constant at ca. 40-45% and is independent
of irradiation time. This result indicates that approxi-
mately half of the DPM generated by the photoreaction
of 1 leads to the radical coupling reaction to produce 2,
and the remaining DPM diffuses to the internal surface
and becomes persistent (Scheme 4). This conclusion is
further supported by quantitative EPR analysis which
demonstrated that DPM is produced in ca. 50% yield.
Above the point at which all holes are filled, the yield of
2 is expected to increase. Indeed, there is an increase in
2 from ca. 40-45% to ca. 50-60% above ca. 0.5% loading,
the composition at which the holes on the external
surface are filled with 1.
Product analysis of the photoreaction of 3/LZ-105
showed that 2, 4, and 5 are formed in the ratio of 2:4:5
(4:13:5), which is comparable to the product ratio ob-
tained by the photoreaction of 3 in n-hexane. From this
we conclude that the low yield of radical coupling
products was mainly formed on the external surface of
LZ-105 where free diffusion, analogous to solution, can
occur.
The time-course of the generation of DPM on LZ-105
(Figure 4) was monitored by EPR and provides kinetic
evidence supporting the model. Dependence of the yield
of DPM produced by photolysis of 1/LZ-105 on the
percentage of loading of 1 indicates that the supramo-
lecular structure of 1/LZ-105 changes at ca. 0.5% loading
(Figure 5). This demonstrates a correlation between the
site location of 1 adsorbed on the LZ-105 surface and the
yield of persistent DPM occurring until ca, 0.5% loading,
the composition that corresponds to the loading at which
the holes on the external surface are filled. Above ca.
0.5% loading, the holes are all filled by 1h /LZ-105, and
the excess of 1 beyond 0.5% will be adsorbed on the
external surface (Scheme 4) to form 1f/LZ-105, where both
Ba sis for P er sisten ce of DP M/LZ-105. Im m obili-
za tion or Ster ic Effect on Ra d ica l-Ra d ica l Rea c-
tion s? DPM radicals that are adsorbed into the internal
surface are remarkably persistent, demonstrated by the
fact that the EPR spectrum of 1 lasts for weeks. This

1328 J . Org. Chem., Vol. 65, No. 5, 2000
Hirano et al.
Sch em e 5
DPMs generated from one molecule of 1 have sufficient
mobility to lead to the radical coupling reaction. In the
case of 90 min irradiation (Figure 5c), the slope above
0.3-0.5% loading is steeper than that for shorter irradia-
tion times. This result indicates that under “high” loading
(>0.5%) and long irradiation conditions (hours), more
persistent DPM radicals can be generated than the
number of holes on the LZ-105 surface because each
channel can contain more than one DPMi. The explana-
tion of this result is that the persistent DPM radicals
generated at an early time of irradiation diffuse away
from their initial site in a hole into the internal surface,
generating a hole that is available for adsorption by DPM
produced by photolysis of 1 adsorbed on the solid external
surface (Scheme 4).
in Table 2b. The normalized values of MI′ for 1 and 3 are
approximately one, strongly suggesting that one molecule
each of 1 and 3 are adsorbed in the holes in a 1 to 1
stoichiometry on the external LZ-105 surface. In sum-
mary, we have convincing evidence from the photochem-
istry and Langmuir analysis that the bulky molecules
are first strongly adsorbed in the holes (site I). After the
holes are filled at ca. 0.3-0.5% loading, these molecules
are more weakly adsorbed on site II, the solid external
surface between the holes as shown in the paradigm of
Schemes 1 and 2.
Su m m a r y a n d Con clu sion s
Com p a r ison of th e P h otoch em ica l Resu lts w ith
th e La n gm u ir Isoth er m s. We will now show that the
results of computation, photochemical, and EPR all
correlate with the results of the Langmuir isotherm
analysis and the evaluation of the surface area of the LZ-
105 particles. The isotherms of 1 and its related com-
pounds 3, 6, and 7 could be fit by a two-site Langmuir
isotherm analysis. The Langmuir constant (kI) of site I
is about one hundred times larger than that of site II
(Table 2a), indicating that 1 binds much more strongly
to site I than to site II. The maximum adsorption amount
(MI) of 1 at site I of LZ-105 (100 mg) was estimated to be
0.3 mg, i.e., 0.3-0.5% loading by weight. The same value
of ca. 0.3-0.5% loading matches the value of the maxi-
mum adsorption in the hole obtained by the time-course
dependence (Figure 5) of the generation of DPM from the
photoreaction of 1/LZ-105. Therefore, from computations,
photochemical and EPR analysis, and Langmuir analy-
ses, we can confidently assign site I to a hole on the
external surface, and site II to the solid external surface.
From the value of the estimated and measured exter-
nal surface area of LZ-105 (15 m2 g-1), the number of
holes available for adsorption is estimated to be sufficient
to ca. 0.4 mg per 100 mg of LZ-105, in excellent agree-
ment with the estimation of the loading at which all the
holes are filled with 1 from the photochemical results,
and further supporting the assignment of site I as a hole
on the external surface. The amount of adsorbed amounts
of 1, 3, 6, and 7 in Table 2a is given in wt/wt units (mg
per 100 mg of LZ-105). These units are not useful in
expressing the fact that the system involves surface area
as a critical dimension for expressing the number density
of holes. To emphasize the importance of the surface area,
the data in Table 2b is expressed in terms of another unit,
mole per unit surface area () 245 Å2). The surface-
normalized constants MI′, MII′, kI′, and kII′ are summarized
To characterize the external surface of a MFI zeolite,
we have applied computations, photochemical probes,
Langmuir isotherms, and EPR to support the paradigms
shown in Schemes 1-5. From both computations and
experiment, it is concluded that the holes on the external
surface provide sites for up to ca. 0.3-0.5% loading of 1.
The filling of the holes constitutes stage I of Scheme 1
and produces a supramolecular structure that can be
termed 1h/(LZ-105)ex. During this stage, the photoreaction
of 1 yields two DPM radicals that may be termed 1h/(LZ-
105)ex and 1f/(LZ-105)ex, i.e., one radical is produced in
the hole on the external surface and the other is produced
on the solid external surface.
For stage II, after the holes are filled, further addition
of 1 results in adsorption on the solid external framework
surface between the holes. Photoreaction of the molecules
on the solid external framework surface also yields two
DPMs, but in this case both of the radicals have mobility
and rapidly undergo radical coupling on the external
surface. As photolysis proceeds to higher conversion, and
some holes become vacant, one of the two freely diffusing
radicals produced from photolysis of 1 adsorbed on site
II may diffuse into a vacant hole and become persistent.
The maximum adsorption amounts of dibenzyl ketone
derivatives 1, 3, 6, and 7 of stage II are estimated by the
addition of MI and MII of Table 2. The values of MI + MII
are in the range of 0.8-1 mg per 100 mg of LZ-105,
consistent with the formation of a monolayer that is
independent of structure. Thus, these results strengthen
the paradigm and demonstrate, through a novel method,
that the persistent radical probe and Langmuir isotherm
analysis are practically useful for characterizing the
external surface of MFI zeolites.
Completely persistent DPM at room temperature
under degassed conditions was obtained by the photore-

Supramolecular Steric Effects on Carbon Radicals
J . Org. Chem., Vol. 65, No. 5, 2000 1329
(200 MHz, CDCl3) δ 3.79 (s, 2 H), 5.22 (s, 1 H) and 7.10-7.35
(m, 15 H); IR (KBr) 1709 cm-1; UV λmax (ꢀ) 300 (319) and 260
(698) nm in isooctane.
action of 1 and 3 on the external surface of LZ-105, and
the generation of DPM was easily observed by EPR. The
reason for the persistence of DPM is explained by the
transfer of DPM from a hole to the channel structure of
LZ-105. The channel system of LZ-105 provides the
supramolecular steric effect on DPM, in which DPM
maintains its mobility. The amount of DPM photochemi-
cally generated from 1 is dependent on the coverage of
the holes, which are filled by 0.3-0.5% loading of 1. The
Langmuir isotherm analyses of dibenzyl ketone deriva-
tives show the existence of two adsorption sites on the
external surface of LZ-105, with the adsorption property
of one site being stronger than that of the other. The
strong adsorption site is assigned as the hole, and the
other site is assigned as the external framework surface
between the holes. The Langmuir isotherm analysis also
supports the fact that the holes are filled up by 0.3-0.5%
loading of 1, and the entire surface is covered by ca. 1%
loading of 1, quantitatively. Characterization of the LZ-
105 particles also supports the coverage of the holes by
1 as the first adsorption step. These results strengthen
the paradigm of Scheme 1 as the adsorption event of an
organic molecule on the external surface of MFI zeolites.
P h otor ea ction of 1 in n -Hexa n e. A solution of 1 (101 mg,
0.279 mmol) in n-hexane (650 mL) was irradiated by using a
450 W medium-pressure mercury lamp for 30 min under Ar.
After evaporating the solvent, the yellow residue was purified
by using TLC (silica gel, CH2Cl2), to give 93.4 mg (100%) of
tetraphenylethane 2 as a colorless power, 2; colorless cubes
(from ethyl acetate); mp. 207-207.5 °C (lit.32 211-212 °C);
1
EIMS m/z 334 (M+, 2), 167 (100); H NMR (200 MHz, CDCl3)
δ 4.77 (s, 2H) and 6.95-7.20 (m, 20H).
P r ep a r a tion a n d P h otor ea ction of 1 Ad sor bed on LZ-
105. From the UV absorption measurement of the supernatant
of 1 and LZ-105 mixture in isooctane, it was established that
an adsorption of 1 on LZ-105 reached equilibrium in 40 min
as shown in Figure 6. To a stock solution of 1 with a known
concentration in isooctane was added 100 mg or 200 mg of LZ-
105, and the suspension was stirred for 1 h at room temper-
ature. Excess isooctane was evaporated with an Ar stream.
The zeolite sample was transferred to a quartz vessel followed
by evacuation with a vacuum line to ca. 5 × 10-5 Torr
overnight. The evacuated sample was irradiated by a medium-
pressure lamp (450 W) being tumbled for an appropriate time.
After irradiation, the sample was exposed to air, and the
products were extracted with 2-3 mL of THF containing
methanol (5% v/v) overnight. The suspension was filtered and
evaporated to give a residue containing 1 and products. The
residue was dissolved in 1-3 mL of ethyl acetate, and
2-ethylnaphthalene (1 µL/1 mL of ethyl acetate) was added to
the solution as an internal standard for GC. For GC analysis,
the individual response factors of 1, 2, and benzophenone to
2-ethylnaphthalene were determined under the following
conditions (injector; 280 °C, oven temperature; from 150 to 250
°C). Recovery of 1 and the yield of 2 and benzophenone were
obtained by using their response factors as shown in Table 1.
For characterization of a radical with EPR, the samples
were prepared in a quartz tube with a branch quartz tube for
EPR measurement. The evacuated sample was irradiated
using a lamp with tumbling for an appropriate time and was
put in the EPR spectrometer’s cavity at room temperature.
After the EPR measurement, the procedure of irradiation and
EPR measurement was repeated for tracing the generation of
DPM.
Exp er im en ta l Section
Gen er a l Meth od s. LZ-105 (Si/Al ) ca. 20) obtained from
Union Carbide Co. was calcined at 500 °C in air more than 24
h before use. Isooctane of spectrophotometric grade (Aldrich)
was used for adsorption experiments. EPR spectra were
measured at room temperature. 1H NMR spectra were re-
corded at 200 MHz. UV absorption spectra were obtained on
a spectrophotometer. GC analyses were carried out using a
chromatograph equipped with a flame ionization detector, a
data station, and a capillary column (25 m × 0.2 mm).
P r ep a r a tion of Tetr a p h en yla ceton e 1. To a solution of
diphenylmethane (5.0 g, 30.0 mmol) in anhydrous THF (80
mL) was added 20 mL of 1.6 M n-butyllithium (32 mmol) in
n-hexane at room temperature under Ar.29 After stirring the
red-colored mixture for 24 h at room temperature, the solution
was cooled to -80 °C. Diphenylacetyl chloride (3.0 g, 11.5
mmol) in anhydrous THF (10 mL) was added to the cooled
solution, and the reaction mixture was stirred at room-
P h otor ea ction of 3 in n -Hexa n e a n d on LZ-105. A
solution of 3 (1.77 × 10-3 mol L-1) in n-hexane (3 mL) was
irradiated using a 450 W medium-pressure mercury lamp for
1 min under degassed conditions. After evaporating the
solvent, the residue was dissolved in 1 mL of ethyl acetate,
and 1 µL of 2-ethylnaphthalene was added to the solution as
an internal standard for GC. Before GC analysis, the indi-
vidual response factors of triphenylacetone 3, ethane deriva-
tives 2, 4, and 5, and benzophenone to 2-ethylnaphthalene
were determined. The recovery of 3 and the yields of products
were obtained by GC with their response factors.
In the case of the photoreaction of 3 adsorbed on LZ-105,
the samples of 3 adsorbed on LZ-105 were prepared in a
manner similar to that of 1. A zeolite sample of 3 in a quartz
vessel under vacuum was photolyzed using a medium-pressure
mercury lamp (450 W) and tumbling the vessel for 5 min. After
exposing the sample to air, products were extracted with 3
mL of THF containing methanol (5% v/v) and were analyzed
by GC using 2-ethylnaphthalene as an internal standard.
Ad sor p tion Tim e-Cou r se of 1 a n d Its Rela ted Com -
p ou n d s on LZ-105 fr om a n Isoocta n e Solu tion . To a 3 mL
solution of 1 and its related compounds in isooctane with a
definite initial concentration, 100 mg of LZ-105 was added,
and the suspension was stirred at room temperature. After
an appropriate time, the mixture was centrifuged for 1 min,
and the UV absorption of the supernatant was measured for
determining the amount of 1 and its related compounds left
in the solution.
temperature overnight. The reaction was quenched in
a
saturated NaCl aqueous solution (50 mL), and the product was
extracted with ether (400 mL). The ether layer was washed
with brine and dried with Na2SO4. After removing the solvent,
the residue was recrystallized from methanol/ethyl acetate,
to give 2.5 g (54%) of 1 as a colorless powder; mp. 135-135.5
°C (lit.30 133-134 °C); EIMS m/z 362 (M+, 3), 195 (7), 167 (100);
1H NMR (200 MHz, CDCl3) δ 5.25 (s, 2 H) and 7.10-7.36 (m,
20 H); IR (KBr) 1709 cm-1; UV λmax (ꢀ) 298 (338) and 260 (1010)
nm in CH3CN, 300 (334) and 260 (955) nm in isooctane.
P r ep a r a tion of Tr ip h en yla ceton e 3. Diphenylmethyl-
lithium was prepared from diphenylmethane (5.0 g, 30.0 mmol)
and n-butyllithium (32 mmol). To a solution of diphenylmeth-
yllithium at -80 °C was added phenylacetyl chloride (2.0 g,
13 mmol) in anhydrous THF (8 mL) , and the mixture was
stirred at room-temperature overnight. The reaction was
quenched in a saturated NaCl aqueous solution (50 mL), and
the product was extracted with ether (300 mL). The ether layer
was washed with brine and dried with Na2SO4. After the
solvent was removed, the residue was purified by a silica gel
column chromatography and recrystallization from methanol,
to give 0.96 g (26%) of 331 as a colorless powder: mp 81-82
°C; EIMS m/z 286 (M+, 8), 167 (100), 165 (25), 91 (19); 1H NMR
(29) (a) Wright, B. B.; Platz, M. S. J . Am. Chem. Soc. 1984, 106,
4175-4180. (b) Zieger, H. E.; Angres, I.; Mathisen, D. J . Am. Chem.
Soc. 1976, 98, 2580-2585.
La n gm u ir Isoth er m An a lysis of 1 a n d Its Rela ted
Com p ou n d s on LZ-105. LZ-105 (100 mg) was added to a 3
mL solution of 1 or its related compounds with various
(30) Dean, D. O.; Dickinson, W. B.; Quayle, O. R.; Lester, C. T. J .
Am. Chem. Soc. 1950, 72, 1740-1745.
(31) Hou, Z.; Takamine, K.; Aoki, O.; Shiraishi, H.; Fujiwara, Y.;
Taniguchi, H. J . Org. Chem. 1988, 53, 6077-6084.
(32) Olah, G. A.; Surya Prakash, G. K. Synthesis 1978, 397-398.

1330 J . Org. Chem., Vol. 65, No. 5, 2000
Hirano et al.
concentrations in isooctane, and the suspension was stirred
at room temperature until the adsorption reached equilibrium.
Most of the compounds, except p-xylene and acetophenone,
were adsorbed for 1 h; p-xylene and acetophenone were
adsorbed for 3 h. The suspension was centrifuged for 5 min.
The absorption spectra of the supernatant and the initial
solution were measured, and the amount of compound ad-
sorbed on LZ-105 at equilibrium was determined from the
absorption difference between the supernatant and the initial
solution. Correlation between the amount (mg) of adsorbate
left in a 3 mL solution and the amount (mg) of adsorbate
adsorbed on 100 mg of LZ-105 at equilibrium was fitted using
eqs 1 or 2 on a Macintosh computer employing the fitting
routine of the graphic program Igor Pro, Version 3.12 for
Macintosh (Wave Metrics, Inc.).
Ack n ow led gm en t. The authors thank the National
Science Foundation for grant (NSF CHE-93-13102). This
work was supported in part by the National Science
Foundation and the Department of Energy under Grant
No. NSF CHE-98-10367 to the Environmental Molecu-
lar Sciences Institute (EMSI) at Columbia University.
The authors also thank Dr. Xuegong Lei for helpful
discussions. The excellent technical assistance and
suggestions of Mr. Zhiqiang Liu in preparing the cover
graphic are gratefully acknowledged.
J O991298I