Catalysis by organic solids. Stereoselective Diels-Alder reactions promoted by microporous molecular crystals having an extensive hydrogen- bonded network

Article abstract of DOI:10.1021/ja964198s

Anthracenebisresorcinol derivative 1 as an organic network material shows a novel catalysis in the solid state for the acrolein-cyclohexadiene Diels-Alder reaction. The suggested mechanism involves a catalytic cycle composed of sorption of the reactants in the cavities of polycrystalline host 1, preorganized intracavity reaction, and desorption of the product. The host also promotes stereoselective intracavity reactions for alkyl acrylates and cyclohexadiene but, in this case, not in a catalytic manner. Relevance of the present system as a functional organic analog of zeolites is discussed in light of the kinetics of respective elementary processes and the effects of pulverization of the catalyst thereupon as well as X-ray crystal structures.

Full text of DOI:10.1021/ja964198s

J. Am. Chem. Soc. 1997, 119, 4117-4122
4117
Catalysis by Organic Solids. Stereoselective Diels-Alder
Reactions Promoted by Microporous Molecular Crystals Having
an Extensive Hydrogen-Bonded Network
Ken Endo,† Takashi Koike,† Tomoya Sawaki,† Osamu Hayashida,†,‡
Hideki Masuda,§,‡ and Yasuhiro Aoyama*,†,‡
Contribution from the Institute for Fundamental Research of Organic Chemistry,
Kyushu UniVersity, Hakozaki, Higashi-Ku, Fukuoka 812-81, Japan, CREST, Japan Science and
Technology Corporation (JST), and Department of Applied Chemistry, Nagoya Institute of
Technology, Gokiso-Cho, Showa-Ku, Nagoya 466, Japan
ReceiVed December 5, 1996X
Abstract: Anthracenebisresorcinol derivative 1 as an organic network material shows a novel catalysis in the solid
state for the acrolein-cyclohexadiene Diels-Alder reaction. The suggested mechanism involves a catalytic cycle
composed of sorption of the reactants in the cavities of polycrystalline host 1, preorganized intracavity reaction, and
desorption of the product. The host also promotes stereoselective intracavity reactions for alkyl acrylates and
cyclohexadiene but, in this case, not in a catalytic manner. Relevance of the present system as a functional organic
analog of zeolites is discussed in light of the kinetics of respective elementary processes and the effects of pulverization
of the catalyst thereupon as well as X-ray crystal structures.
Introduction
Scheme 1
Solid-state reactions constitute an important area of selective
organic reactions.1 There are also many examples of lattice
inclusion compounds,2 in which the host affects the rates and
stereochemistry of the reactions of included guest molecules.1
To the best of our knowledge, however, organic solids have
seldom been used as catalysts.3 In fact, solid or heterogeneous
catalysts including zeolites4 have so far been exclusively
inorganic materials. We are particularly interested in the
construction of organic network materials as functional analogs
of zeolites.5 Organic molecules show a remarkable diversity
not only in molecular structures but also in network dimen-
sionalities and topologies.6 The cavity sizes7 and functional
catalytic groups are also designable in principle.
molecules can be bound simultaneously in the cavity as in
adduct 1‚2(CH3CO2CH2CH3)‚2(C6H6).9 The present work is
concerned with Diels-Alder reactions (eq 1 in Scheme 1)
between enclathrated acrolein (2a) or an acrylic ester (2b-d)
We have recently shown that anthracenebisresorcinol deriva-
tive 1 (Scheme 1) forms a hydrogen-bonded network whose
cavities are capable of crystalline-phase guest-addition, -removal,
and -exchange.8 Interestingly, small polar and apolar guest
(5) For selected examples, see: (a) Barrer, R. M.; Shanson, V. H. J.
Chem. Soc., Chem. Commun. 1976, 333-334. (b) Ermer, O. J. Am. Chem.
Soc. 1988, 110, 3747-3754. (c) Weber, E.; Pollex, R.; Czugler, M. J. Org.
Chem. 1992, 57, 4068-4070. (d) Copp, S. B.; Subramanian, S.; Zaworotko,
M. J. Ibid. 1992, 114, 8719-8720. (e) Reddy, D. S.; Craig, D. C.; Rae, A.
D.; Desiraju, G. R. J. Chem. Soc., Chem. Commun. 1993, 1737-1739. (f)
Wang, X.; Simard, M.; Wuest, J. D. J. Am. Chem. Soc. 1994, 116, 12119-
12120. (g) Venkataraman, D.; Lee, S.; Zhang, J.; Moore, J. S. Nature 1994,
371, 591-593. (h) Reddy, D. S.; Craig, D. C.; Desiraju, G. R. J. Am. Chem.
Soc. 1996, 118, 4090-4093.
† Kyushu University.
‡ CREST.
§ Nagoya Institute for Technology.
X Abstract published in AdVance ACS Abstracts, April 1, 1997.
(1) (a) Schmidt, G. M. J. Pure Appl. Chem. 1971, 27, 647-678. (b)
Green, B. S.; Lahav, M.; Rabinovioch, D. Acc. Chem. Res. 1979, 12, 191-
197. (c) Addadi, L.; Ariel, S.; Lahav, M.; Leiserowitz, L.; Popovitz-Biro,
R.; Tang, C. P. Chem. Phys. Solids Their Surf. 1980, 8, 202-244. (d)
Scheffer, J. R. Acc. Chem. Res. 1980, 13, 283-290. (e) Desiraju, G. R.,
Ed.; Organic Solid State Chemistry; Elsevier: Amsterdam, 1987. (f)
Ramamurthy, V.; Venkatesan, K. Chem. ReV. 1987, 87, 433-481. (g) Toda,
F.; Top. Curr. Chem. 1988, 149, 211-238. (h) Toda, F. Synlett 1993, 303-
312. (i) Toda, F. Acc. Chem. Res. 1995, 28, 480-486.
(6) (a) For a list of recent publications on the formation of ordered crystal
structures involving hydrogen-bonded 1D, 2D, and 3D motifs, see: Aoyama,
Y.; Endo, K.; Anzai, T.; Yamaguchi, Y.; Sawaki, T.; Kobayashi, K.;
Kanehisa, N.; Hashimoto, H.; Kai, Y.; Masuda, H. J. Am. Chem. Soc. 1996,
118, 5562-5571. Also see: (b) MacDonald, J. C.; Whitesides, G. M. Chem.
ReV. 1994, 94, 2383-2420. (c) Valiyaveettil, S.; Enkelmann, V.; Mu¨lln,
K. J. Chem. Soc., Chem. Commun. 1994, 2097-2098. (d) Hosseini, M.
W.; Ruppert, R.; Schaeffer, P.; De Cian, A.; Kyritsakas, N.; Fischer, J.
Ibid. 1994, 2135-2136. For examples of (infinite) metal-coodination
networks, see: (e) Abrahams, B. F.; Hoskins, B. F.; Liu, J.; Robson, R. J.
Am. Chem. Soc. 1991, 113, 3045-3051. (f) Reference 3c. (g) Reference
5d. (h) Stang, P. J.; Cao, D. H.; Saito, S.; Arif, A. M. Ibid. 1995, 117,
6273-6283. (i) Fujita, M.; Kwon, Y. J.; Sasaki, O.; Yamaguchi, K.; Ogura,
K. Ibid. 1995, 117, 7287-7288.
(2) For a list of recent publications on lattice inclusion compounds, see:
Aoyama, Y.; Endo, K.; Kobayashi, K.; Masuda, H. Supramolec. Chem.
1995, 4, 229-241.
(3) For recent publications very briefly referring to apparently catalytic
behaviors of organic crystals, see: (a) Toda, F.; Tanaka, K.; Sekikawa, A.
J. Chem. Soc., Chem. Commun. 1987, 279-280. (b) Soai, K.; Watanabe,
M. Ibid. 1990, 43-44. (c) Fujita, M.; Kwon, Y. J.; Washuzu, S.; Ogura, K.
Ibid. 1994, 116, 1151-1152.
(4) (a) Ho¨lderich, W.; Hesse, M.; Na¨umann, F. Angew. Chem., Int. Ed.
Engl. 1988, 27, 226-246. (b) Suib, S. L. Chem. ReV. 1993, 93, 803-826.
(c) Breck, D. W. Zeolite Molecular SieVes, Structure, Chemistry, and Use;
John Wiley & Sons: New York, 1974.
(7) Cf., Endo, K.; Ezuhara, T.; Koyanagi, M.; Masuda, H.; Aoyama, Y.
J. Am. Chem. Soc. 1997, 119, 499-505.
S0002-7863(96)04198-4 CCC: $14.00 © 1997 American Chemical Society

4118 J. Am. Chem. Soc., Vol. 119, No. 18, 1997
Endo et al.
Figure 2. Correlations between rate constants and normalized amounts
of catalysts: nonpulverized host 1 in the solid state (solid line),
pulverized host 1 in the solid state (dashed line), and soluble resorcinol
(dotted line). Runs a-g refer to those in Figure 1.
Figure 1. Time courses of the reactions of dienophile 2a (1.5 mmol)
and diene 3 (30 mmol) (2a/3 ) 20) at 25 °C in the presence and absence
of host 1 as either nonpulverized pieces (runs b-d) or pulverized
powders (runs e-g): 1/2a ) 0 (a), 0.017 (b and e), 0.033 (f), 0.083 (c
and g), 0.17 (d).
substrate 2a used. Clearly, the facilitated reactions go beyond
the amounts of the host up to the completion. These results
indicate that host 1 functions as a catalyst with a turnover rate
constant of kcat ) 0.20 h-1 for nonpulverized host or 0.33 h-1
for pulverized host; it is obtained as the slope, (mole of 4a
formed)/((mole of 1 used)h), of the linear correlation between
rate constants and amounts of host 1 (Figure 2).
as a dienophile and 1,3-cyclohexadiene (3) in place of ethyl
acetate and benzene, respectively. We report here that host 1
promotes highly stereoselective intracavity reactions either
catalytically for aldehyde dienophile 2a or stoichiometrically
for ester dienophiles 2b-d.10-12
Results and Discussion
It is easy to demonstrate that the catalysis is exerted by the
solid host. The solubility of host 1 in the present 1:20 mixture
of 2a and 3 is unmeasurably small. If the solubilized host, if
any, were the catalyst, runs b-g (Figure 1) having the same
concentration of the host at saturation should have the same
rates. This is not the case. The rates of the reactions show a
linear dependence on the amounts of insoluble host. It is also
confirmed that a supernatant solution separated from insoluble
host 1 by filtration exhibits practically the same rate as the
reference run (run a). The catalytic activity of solubilized host
can be evaluated by using a 2a-rich solution, e.g., a 10:1
mixture of 2a and 3, in which host 1 is significantly soluble.
Treatment, in a similar manner as above (Figures 1 and 2), of
the kinetic data with respect to limiting substrate 3 in the absence
Catalytic Acrolein-Diene and Stoichiometric Acrylate-
Diene Reactions. The liquid-phase reaction of acrolein (2a)
and cyclohexadiene (3) (eq 1 of Scheme 1) is very slow at 25
°C but is accelerated by host 1 in the solid state, giving rise to
a 95:5 mixture of the endo and exo products 4aendo and 4aexo
.
Figure 1 shows the time courses of the reactions in the absence
(a) and presence (b-g) of a catalytic amount (1/2a ) 0.017 (b
and e), 0.033 (f), 0.083 (c and g), 0.17 (d)) of the host as either
nonpulverized pieces (approximate size, 1 × 1 × 1 mm3) (runs
b-d shown in solid lines) or pulverized powders (runs e-g
shown in dashed lines) sitting in a 1:20 mixture of dienophile
2a and diene 3. Host 1 is hardly soluble in the present 3-rich
solution. Nevertheless, the reactions are faster in the presence
of increasing amounts of the host. There is in fact a linear
dependence of the rate constants13 of the reactions and the
amounts of nonpulverized (solid line) or pulverized (dashed line)
host as shown in Figure 2, where the amounts of product 4a
formed and host 1 used are normalized with that of limiting
and presence of varying amounts of the host reveals that kcat
)
0.02 h-1 as the slope of a linear correlation between rate
constants, (4a formed)/((3 used)h), and (1 used)/(3 used).14 Thus,
the homogeneous catalysis (kcat ) 0.02 h-1) by the solubilized
host is less pronounced as compared with heterogeneous
catalysis (kcat ) 0.20 or 0.33 h-1) by the same host in the solid
state. Evaluation of the homogeneous catalytic activity of
resorcinol in the 1:20 (2a to 3) mixture leads to a similar
conclusion: kcat ) 0.027 h-1 (dotted line in Figure 2).14
Furthermore, it is important to note that resorcinol shows
practically no catalysis in the solid state,15 in marked contrast
to host 1. Another interesting observation is that the present
reaction is subject to inhibition by an otherwise inert compound
such as methyl benzoate (vide infra).
(8) (a) Reference 2. (b) Endo, K.; Sawaki, T.; Koyanagi, M.; Kobayashi,
K.; Masuda, H.; Aoyama, Y. J. Am. Chem. Soc. 1995, 117, 8341-8352.
(c) Reference 6a.
(9) Kobayashi, K.; Endo, K.; Aoyama, Y.; Masuda, H. Tetrahedron Lett.
1993, 34, 7929-2932.
(10) Diels-Alder reactions have been demonstrated to occur in the solid
state by cocrystallization of diene and dienophile: Kishan, K. V. R.;
Desiraju, G. R. J. Org. Chem. 1987, 52, 4641-4644. Also see: Sergeev,
G. B.; Komarov, V. S.; Zvonov, A. V. Dokl. Akad. Nauk SSSR 1983, 270,
139-142.
(11) For Diels-Alder reactions in the host-guest systems, see: (a)
Walter, C. J.; Anderson, H. L.; Sanders, J. K. M. J. Chem. Soc., Chem.
Commun. 1993, 458-460 (high exo-selectivity). (b) Hirst, S. C.; Hamilton,
A. D. J. Am. Chem. Soc. 1991, 113, 382-383 (intramolecular reaction).
(c) Sternbach, D. D.; Rossana, D. M. Ibid. 1982, 104, 5853-5854
(intramolecular reaction). (d) Rideout, D. C.; Breslow, R. Ibid. 1980, 102,
7816-7817 (hydrophobic acceleration). For other pericyclic reactions, see;
(e) Mock, W. C.; Irra, T. A.; Wepsiec, J. P.; Adhya, M. J. Org. Chem.
1989, 54, 5302-5308. (f) Ueno, A.; Moriwaki, F.; Iwama, Y.; Suzuki, I.;
Osa, T.; Ohta, T.; Nozoe, S. J. Am. Chem. Soc. 1991, 113, 7034-7036.
(12) Solid-state photodimerization of olefins is well-known and is also
relevant to the present reactions (reference 1). Also see: Granagura, K.;
Ramasubbu, N.; Verkatesan, K.; Ramamurthy, V. J. Org. Chem. 1985, 50,
2337-2346.
To our surprise, change in dienophiles from acrolein (2a) to
ethyl acrylate (2b) results in a dramatic change in the overall
features. Host 1, which turns out to be a catalyst in the acrolein
system, shows almost no catalysis in the acrylate-diene
reaction.
Sorption of Reactants. Recrystallization at room temper-
ature of the host from a mixture of acrolein (2a) and cyclo-
(14) This treatment of homogeneous kinetic data is only for the purpose
of comparison with heterogeneous kinetic data; it is difficult to define
concentration for an insoluble catalyst.
(15) Resorcinol is soluble in the present reaction mixture (2a/3 ) 1:20)
up to the amount of resorcinol/2a ) 0.33, where the pseudo-first-order rate
constant of the reaction is ∼0.01 h-1 (Figure 2). Further addition of
resorcinol results in its precipitation without causing any notable increase
in the rate.
(13) The slope of the yield-time plots (Figure 1) at an early stage of
the reaction corresponds to the pseudo-first-order rate constant, the accurate
value of which can also be obtained by the standard pseudo-first-order
treatment of the kinetic data.

Catalysis by Organic Solids
J. Am. Chem. Soc., Vol. 119, No. 18, 1997 4119
resorcinol rings, dienophile 2b, and diene 3 are shown in red,
black, green, and pink, respectively, and hydrogen bonds are
in light blue. (1) Host 1 forms a hydrogen-bonded (O-H‚‚‚O-
H) network to give a 2D sheet (Figure 4a), where neighboring
anthracene columns c and c′ are not equivalent in respect to
the orientation of included guest molecules. (2) In contrast to
those in related adducts,6a,8,9 two neighboring anthracene rings
along a column are not parallel so that the cavities roofed and
bottomed there are not cylindrical but bowl-shaped, as shown
by a side view (Figure 4b and its explanation 4c) of two adjacent
cavities in column c. Two dienophile (2b) molecules, being
roughly parallel with nearby anthracene rings, are bound in each
cavity via host-guest hydrogen-bonding (O-H‚‚‚O-H‚‚‚OdC)
with their vinyl (CdC) groups pointing to the wider or open
side. (3) Diene 3 occupies the space left in this side in such a
way as to be sandwiched by the vinyl groups of the ester guests
(side views of Figure 4b,c and top view (d) with its explanation
(e) where front (top) and rear (bottom) molecules of 2b are
shown in thick and thin lines, respectively). (4) The guest-
binding mode in the cavities in column c′ is similar but the
orientation of dienophile 2b is opposite; the ethoxy groups
instead of the vinyl point to the open side and sandwich diene
3 (illustrations c′ and e′, where double bonds of cyclohexadiene
can not be fixed because of disorder).
Intracavity Reactions. Good evidence for the intracavity
Diels-Alder reactions comes from solid-gas experiments,
where host 1 is allowed to be in contact with the vapors of the
guests at 25 °C in a sealed vessel.8 The initially formed host-
dienophile-diene ternary adducts of an approximate composi-
tion of 1‚2(2a)‚2(3) or 1‚2(2b)‚3 are ultimately converted to
the corresponding product adducts 1‚2(4a) and 1‚2b‚4b. The
reactant adducts obtained by recrystallization or solid-liquid
complexation behave similarly. Thus, the crystalline-phase
intracavity reactions are complete on the stoichiometric basis,
leaving 1 mol of excess dienophile 2b intact for the acrylate
adduct.
The stereoisomer ratios for the intracavity reactions increase
with increasing steric bulkiness of the substituents (R) in
dienophiles H2CdCH(CdO)R; 4endo/4exo ) 95:5, 98:2, 99:1,
and (>99):(<1) when R ) H (a), OCH2CH3 (b), O(CH2)3CH3
(c), and OCH2CH(CH3)2, respectively. For comparison, the
selectivities of reference runs in the absence of host 1 at reflux
temperatures are 4endo/4exo = 87:13 and slightly higher (∼9:1)
at 25 °C and are practically independent of substituents (R).
The crystal structure for the reactant adduct 1‚2(2b)‚3 is not
very suggestive of the stereoselectivity. Almost nothing is
known about the actual reaction geometry in cavity c′. In cavity
c, an ene-diene proximity (∼3.8 Å; Figure 4b) is assured but
their orientation is not suited for reaction (Figure 4e). Rotation
of diene 3 by either 60° or 120° brings the diene moiety into
an orientation suitable for reaction with the dienophile, expect-
edly giving rise to the endo (4bendo) or the exo (4bexo) product,
respectively.
Figure 3. Correlations between rate constants and normalized amounts
of methyl benzoate (MB) as an inhibitor under otherwise identical
conditions for runs c, d, and g.
hexadiene (3) affords a quaternary adduct 1‚x(2a)‚x(3)‚y(4a)
containing Diels-Alder product 4a in addition to both of the
reactants. The composition (x and y) changes (2 > x > 0 and
0 < y < 2) with duration of recrystallization, while keeping (x
+ y) ) 2; the total 1/dienophile/diene ratio is maintained at
1:2:2. Recrystallization even at -23 °C fails to give single
crystals of adduct 1‚2(2a)‚2(3) completely free from product
4a. The crystal structure of a similar 1:2:2 adduct 1‚2(CH3-
CO2CH2CH3)‚2(C6H6) has been determined.9 The two ester
molecules are hydrogen-bonded to the host in each cavity, and
the two benzene molecules occupy the void space left, resulting
in a close ester-benzene proximity.
Crude adduct 1‚2(2a)‚2(3) can be also obtained by dipping
the host in a relatively 2a-rich solution, e.g., a 1:1 mixture of
2a and 3, although the resulting adduct subsequently dissolves
in the solution. This is why we are forced to carry out the
catalytic reactions by using a highly hydrocarbon-rich solution.
Actually, when dipped in a 1:20 (2a to 3) mixture, host 1 rapidly
(in seconds to minutes) forms an adduct of an apparent
composition of 1‚2a‚3(3). As the reaction proceeds, product
4a accumulates in the cavity; the composition of the adduct at
60% conversion of the catalytic reaction is 2a/3/4a = 1:1:2.
These results strongly suggest that the internal cavities of
host 1 provide the sites of reaction. This is further supported
by the competitive inhibition of the reactions by methyl benzoate
(MB), which is a potential guest of the present host.8a,b Figures
I, II, and III (Supporting Information) show the time courses of
the reactions in the absence and presence of MB at an inhibitor/
reactant ratio of MB/2a ) 0.5 or 1 under otherwise identical
conditions for runs c, d, and g. Figure 3 shows how the rate
constants (c, d, and g) decrease with increasing amounts of
soluble MB; inhibition is ∼50% (c′, d′, and g′) or 70-80%
(c′′, d′′, and g′′) at MB/2a ) 0.5 or 1, respectively. Independent
measurements indicate that host 1 preferentially binds inhibitor
MB rather than reactant 2a in a ratio of MB/2a ) 3 or 8 when
the composition in the solution phase is MB/2a ) 0.5 or 1,
respectively.
Because of technical problems,18 it is difficult to follow the
kinetics of intracavity reactions accurately. Only an approximate
half-life of τ ) 2 h (corresponding to the first-order rate constant
of k1cav ) 0.3 h-1 ) can be estimated for the acrolein-diene (2a-
3) reaction. The acrylate-diene (2b-3) reaction, which takes
longer, turns out to be biphasic. The relatively slow reaction
When ethyl acrylate (2b) is used in place of acrolein (2a),
host 1 affords a 1:2:1 adduct 1‚2(2b)‚3 upon recrystallization
or solid-state complexation.16 The acrylate adduct is kinetically
more stable than the acrolein adduct. The former retains single-
crystallinity at low temperatures to allow X-ray diffraction
studies at 230 K. The crystal structure17 is summarized as
follows in reference to Figure 4, where anthracene rings,
cav
with τ ) 20 h (k1 ) 0.03 h-1 ) up to ∼50% conversion is
followed by a much faster reaction in the second stage. The
biphasic nature may be understood on the basis of the crystal
(16) Recrystallization of host 1 from a mixture of ethyl acrylate (2b)
and benzene affords a similar 1:2:1 adduct 1‚2(2b)‚(C6H6).
(17) P21/n, a ) 23.804(6) Å, b ) 13.814(4) Å, c ) 24.314(3) Å, â )
109.34(1)°, V ) 7544.1 Å3, Z ) 4, dcalcd ) 1.19 g/cm3, number of unique
reflections ) 9728 (2θ e 51.4°), number of reflections used ) 4600 (I0 <
3σ(I0 )), R ) 0.062, and Rw ) 0.070.
(18) One of the problems, especially for the acrolein system, is
spontaneous desorption of included cyclohexadiene from the isolated
adducts. Thus, the adducts undergoing intracavity reactions should be kept
under the vapor of the diene.

4120 J. Am. Chem. Soc., Vol. 119, No. 18, 1997
Endo et al.
Figure 4. Crystal structure of adduct 1‚2(2b)‚3 at 230 K; two neighboring columns c and c′ in a 2D sheet constructed by a hydrogen-bonded
network (a), side view of two adjacent cavities of column c (b) and its explanation (c), top view of a cavity of column c (d) and its explanation (e),
and guest-binding mode (in schematic form) in cavities of column c′ (c′ and e′ where double bonds of cyclohexadiene can not be fixed because of
disorder). The anthracene rings, resorcinol rings, dienophile 2b, and diene 3 are shown in red, black, green, and pink, respectively, and hydrogen
bonds are represented in light blue.

Catalysis by Organic Solids
J. Am. Chem. Soc., Vol. 119, No. 18, 1997 4121
structure. There are two types of the cavities, i.e., those along
columns c and c′. The former (cavity c) is reactive, since diene
3 is sandwiched by the vinyl groups of dienophile 2b (Figure
4). In the latter (cavity c′), it is the ethoxy groups of the
dienophile that sits above and below the six-membered ring of
the diene (Figure 4c′,e′). A drastic change in the crystal
structure might take place when the reaction in cavity c is
complete.
components when product adduct 1‚4b‚2b is dipped in a 1:20
(2b to 3) mixture. The different behaviors of the two adducts
is not due to the difference in incoming dienophiles 2a vs 2b.
When dipped in a solution of methyl acrylate (MA) and diene
3 under otherwise identical conditions, adduct 1‚2(4a) undergoes
facile 4a/(MA + 3) exchange, while adduct 1‚4b‚2b shows
practically no tendency of 4b/(MB + 3) exchange.
Guest-exchange may be controlled by both steric and
electronic factors. The acrolein product 4a as an aldehyde is
less bulky and less basic than the acrylate product 4b as an
ester; basicity may be an important factor that governs host-
guest hydrogen-bonding. It might be easier for smaller guests
less strongly hydrogen-bonded to the host to leave the cavity
and diffuse in the solid. At the present, however, knowledge
is too scarce to allow a convincing explanation for the
remarkable difference in the desorption abilities of the acrolein
and acrylate products 4a and 4b. We have to learn more about
the general mechanism of lattice diffusion.25 Another decisive
factor is how host-guest hydrogen-bonding changes, in geom-
etry and strength, upon reaction.
The approximate kinetic parameters estimated above are not
very dependent on whether the starting adducts are pulverized
or not. They may be compared with the pseudo-first-order rate
uncat
constants of k1
) 1.3 × 10-3 h-1 (τ ) 500 h) and e2 ×
10-4 h-1 (τ g 3000 h), respectively, for 2a and 2b, in the
uncatalyzed liquid-phase 2a-3 and 2b-3 reference reactions
under pseudo-first-order conditions, i.e., for the 1:20 dienophile-
diene mixtures (cf. run a in Figure 1). There is little doubt that
the facilitation of the intracavity reactions is primarily due to
the proximity of dienophile and diene simultaneously bound in
a cavity.
The proximity effects19 are often expressed in terms of
cav
effective concentration factors (m) by using the equation k1
Catalytic Mechanism. Combination of the elementary
processes leads to a simplified catalytic mechanism of the
acrolein-cyclohexadiene reaction. It involves (a) initial binding
of reactants (in seconds to minutes), (b) intracavity reaction (in
hours), and (c) product/reactant exchange (biphasic; in seconds
to minutes and hours for the faster and slower phases,
respectively). The relative time scales of respective steps
suggest that step b must be rate-determining, in accord with
the kinetics of catalytic reactions.26 Step c can also be partially
rate-limiting; catalyst 1 indeed exists as quarternary adducts
incorporating both product and reactants in the course of the
catalysis (vide supra). It is also evident that host 1 fails to
catalyze the acrylate-diene (2b-3) reaction due to product
inhibition.
cav
) k2m, where k1 is the observed first-order rate constant in
respect to one reactant, e.g., dienophile in the present case, of
a preorganized bimolecular reaction, m is the effective molarity
of the reaction partner, i.e., diene, and k2 is the second-order
rate constant for the reference reaction in homogeneous solution.
Putting experimental values of k1cav ) 0.3 h-1 and k2 ) 1.3 ×
10-4 M-1 h-1 and k1cav ) 0.03 h-1 and k2 e 2 × 10-5 M-1 h-1
for the acrolein-diene and acrylate-diene reactions,20 the
effective molarities of the enclathrated cyclohexadiene can be
calculated as m ) 2000 and g 1500 M for the acrolein and
acrylate systems, respectively. For comparison, 1,3-cyclohexa-
diene as a neat liquid has a molarity of ∼10 M. Thus,
geometrical factors play a decisive role. The proximity effects,
however, may not be the sole origin of rate enhancement.
Diels-Alder reactions are well-known to be accelerated by
acids, especially Lewis acids, which activate enone-type dieno-
philes via coordination.21 A slight but distinct catalysis by
soluble resorcinol is noted above. The fixed hydrogen-bonding
between host and dienophile may provide a potential source of
acid catalysis in the present intracavity reactions.22,23
Desorption of Products. It is in the fate of intracavity
Diels-Alder products that the acrolein and the acrylate systems
differ essentially. When the acrolein product adduct 1‚2(4a)
in powders is dipped in a 1:20 (2a to 3) mixture, product 4a
included can be exchanged with reactants 2a and 3. This
process is biphasic. About 50% of 4a bound is rapidly
exchanged in seconds to minutes, while desorption of the
remaining 50% takes hours. The adduct in nonpulverized pieces
exhibits a similar overall feature.24 In marked contrast, the
acrylate-diene product 4b resists to be exchanged by its
The surface often plays important or essential roles in
heterogeneous catalyses, where catalytic activities usually show
dramatic size-dependencies. The biphasic nature of step c above
may suggest a depth-dependence of the catalytic abilities of
internal cavities.24 It is also possible that only cavities near
the surface actually take part in the catalytic processes. The
point, however, is that there is no evidence to believe that the
present catalysis is simply a surface phenomenon. (1) The
elementary processes show no essential dependence on whether
the host or its adduct is pulverized or not. The effect of
pulverization on the overall catalytic (turnover) rate constants
is only by a factor of 1.7 (kcat ) 0.20 and 0.33 h-1 for
nonpulverized and pulverized host, respectively; Figure 2). Thus,
the size effect is quite modest at best. (2) The surface
mechanism seems to fail to give a good explanation why acrylate
product 4b, and not the acrolein product 4a, formed on the
surface should continue to stay there to deactivate the catalytic
sites. The stereoselectivity (4aendo/4aexo ) 95:5) of the catalytic
acrolein-diene reaction is the same as that of the single-turnover
intracavity reaction (4aendo/4aexo ) 95:5). This, coupled with
(19) Storm, D. R.; Koshland, D. E., Jr. J. Am. Chem. Soc. 1972, 94,
5805-5814.
(20) k2 ) k1uncat/[3], where [3] = 10 M for the 1:20 mixture of dienophile
2a or 2b and diene 3.
(21) (a) Kagan, B. F.; Riant, O. Chem. ReV. 1992, 92, 1007-1019. (b)
Pindur, U.; Luiz, G.; Otto, C. Ibid. 1993, 93, 741-761.
(22) There is no doubt that acrolein incorporated is hydrogen-bonded to
the host, as evidenced by a large (30 cm-1) shift to lower wavenumber in
νCdO for acrolein upon adduct formation.
(25) The mechanism of guest-exchange has not yet been established. It
may even involve local dissolution of the host followed by its recrystalli-
zation. For related work, see references 5f and 8b.
(23) There are numerous examples of facilitated solid-state reactions via
hydrogen bonding (reference 1). For the effect of hydrogen bonding on the
tautomeric composition, see: Herbstein, F. H.; Kapon, M.; Reisner, G. M.;
Lehman, M. S.; Kress, R. B.; Wilson, R. B.; Shiau, W. I.; Duesler, E. N.;
Paul, I. C.; Curtin, D. Y. Proc. R. Soc. London 1985, A399, 295-319.
(24) The size-insensitivity of the desorption properties might be inter-
preted as suggesting that the biphasic nature is due to some kind of phase
transition in the crystal structure upon guest desorption or exchange and
not due to its dependence on the depth of the cavity from the surface.
(26) The first-order expression of intracavity reaction is (mole of 2a
consumed)/h ) k1cav(mole of 2a included). If the intracavity reaction is
rate-determining, the rate expression for the catalytic reaction would be
(mole of 4a formed)/h ) (mole of 2a consumed)/h ) nk1cav(mole of host
1), where n is the number of included guest molecules in the cavity. Under
the conditions of catalytic reactions using a 1:20 (2a to 3) mixture, n = 1,
as stated in the text. This indicates that kcat ) k1cav. The experimental values
are kcat ) 0.20 h-1 (for nonpulverized host) or 0.33 h-1 (for pulverized
host) and k1 ) 0.3 h-1
.
cav

4122 J. Am. Chem. Soc., Vol. 119, No. 18, 1997
Endo et al.
verized pieces prepared as above or pulverized powders. The reactions
were monitored by occasional sampling and analysis (1H NMR and/or
gas chromatography) of the liquid phase. The reproducibility of the
rate constants is within 7 or 10% for runs with pulverized or
nonpulverized host, respectively.
Solid-state guest-binding and guest-exchange studies were carried
in a similar manner as reported for related systems.8a,b Pieces of reactant
adduct 1‚2(2a)‚2(3) or 1‚2(2b)‚3 were placed in the vapor phase of a
liquid mixture of the reactants, or directly in the liquid phase for adduct
1‚2(2b)‚3, in a sealed tube at 25 °C. A few pieces were picked up at
appropriate time intervals, dissolved in DMSO-d6, and analyzed by 1H
NMR for the progress of the intracavity reactions.
Host 1 could be recrystallized either (a) by allowing the vapor of
diene 3 to slowly diffuse into a solution of 1 in dienophile 2a or 2b at
room temperature or (b) by cooling a solution of 1 in a dienophile-
diene mixture at -23 °C. Single crystals free from Diels-Alder
products were obtained only for the acrylate-diene (2b-3) system in
method b.
X-ray Structure Determination. A crystal of 1‚2(2b)‚3 with
dimensions of 0.2 × 0.3 × 0.35 mm was mounted in a sealed capillary.
Diffraction data were collected on a Rigaku RAXIS II imaging plate
area detector using graphite-monochromated Mo KR radiation (λ )
0.710 73 Å) at 230 K. The accurate unit cell parameters used for
refinement were determined by least-squares calculations on the setting
angles for 25 reflections with 2θ ) 20-25° that were collected on an
Enraf Nonius CAD4 diffractometer using the same radiation under the
same conditions. Reflection data were corrected for both Lorenz and
polarization effects. The linear absorption coefficient µ for Mo KR
radiation is 0.76 cm-1, and absorption correction was not applied.
The structure was solved by a combination of the direct method
using SIR92 program and the Fourier techniques using DIRDIF92
program. All of the non-hydrogen atoms were refined anisotropically
by full-matrix least-squares calculations. Some hydrogen atoms were
located from difference Fourier maps and refined isotropically. The
rest, except for one hydrogen atom on each methylene carbon of
cyclohexadiene, were introduced at calculated positions and fixed. Each
refinement was continued until all shifts were smaller than one-third
of the standard deviations of the parameters involved. The weighting
scheme w-1 ) (σ2(Fo) + (0.0015Fo)2) was employed for the crystal.
The final difference Fourier maps did not show any significant features.
These calculations were performed on an INDY workstation by using
the teXsan crystallographic software package of Molecular Structure
Corporation. The crystal structure was visualized by using the Cerius2
set of computer programs of Molecular Simulations Incorporated, which
were also used for determining the bond lengths, bond angles, dihedral
angles, and many other intermolecular distances. The location of the
double bonds in cyclohexadiene incorporated in column c was based
on the bond lengths, bond angles, and dihedral angles thus determined.
On the other hand, cyclohexadiene in column c′ looked like benzene,
probably due to disorder; it was not possible to fix its orientation.
the competitive inhibition effects of methyl benzoate, is also in
a strong support of the cavity mechanism.
Concluding Remarks
The internal cavities of host 1 play essential roles in the
present systems. (1) The two reactant molecules of different
polarity characters are assembled in the same cavity.27 (2) The
preorganized intracavity reactions exhibit remarkable rate
enhancements as well as high stereoselectivities. (3) The
products either leave the cavities or stay therein, resulting in
turnover or deactivation of the catalyst. These are also what
are characteristic of zeolitic4 as well as enzymatic catalyses.28
The present work thus suggests a potential utility of this type
of microporous crystals as functional organic analogs of zeolites
and also as a new type of enzyme mimics, where a decisive
factor is how product inhibition is minimized.29 In addition to
the specific points pertaining to the present systems, there are
a number of general concerns. One is the types of reactions.
In principle, product/reactant exchange would be facilitated
when the products have lower hydrogen-bonding abilities than
the reactants for either steric or electronic reasons. An extreme
case would be such reactions as decarboxylation, where the
hydrogen-bonding sites are lost upon reaction. Another is
molecularity. Associative bimolecular reactions including the
present Diels-Alder reactions yield one molecule of product,
whose liberation at the sacrifice of fixing two reactant molecules
upon product/reactant exchange in the cavity must be entropi-
cally very unfavorable. Dissociative unimolecular reactions
yielding two or more product molecules would be much more
favorable in this respect. A third point of interest is a particular
use of water as a solvent of the liquid phase. Water may weaken
host-guest hydrogen-bonding and hence give rise to a higher
degree of reversibility in the host-guest complexation. This
may also facilitate product/reactant exchange.30
Experimental Section
General Procedures. Compound 1, prepared as before,2 was
recrystallized from a mixture of ethyl acetate and benzene.9 Single
crystals (approximate size, 1 × 1 × 1 mm3) of adduct 1‚2(CH3CO2-
CH2CH3)‚2(C6H6) obtained were heated at 200 °C in vacuo for 10 h to
give pieces of apohost 1, which apparently retained the shape and size
of the original single crystals. Pulverization of the apohost was effected
by using a mortar and a pestle.
Catalytic reactions were carried out under nitrogen at 25 °C by using
a 1:20 mixture of dienophile 2a or 2b (1.5 mmol) and diene 3 (30
mmol) in the absence and presence of the apohost as either nonpul-
(27) For examples of templated bimolecular or multimolecular reactions,
see: (a) Diederich, F.; Lutter, H.-D. J. Am. Chem. Soc. 1989, 111, 8438-
8446. (b) Anderson, S.; Anderson, H.-L.; Sanders, J. K. M. Acc. Chem.
Res. 1993, 26, 469-475. (c) Kelly, T. R.; Bridger, G. J.; Zhao, C. J. Am.
Chem. Soc. 1990, 112, 8024-8034. (d) Nowick, J. S.; Feng, O.; Tjirijua,
P.; Ballester, P.; Rebek, J., Jr. Ibid. 1991, 113, 8831-8839.
Acknowledgment. This work was supported by grant-in-
aids from the Ministry of Education, Science, and Culture, Japan,
including that for COE Research “Design and Control of
Advanced Molecular Assembly Systems” (no. 08CE2005) and
also by CREST (Core Research for Evolutional Science and
Technology) of Japan Science and Technology Corporation.
(28) For the enzymes catalyzing the Diels-Alder reactions, see: Oikawa,
H.; Katayama, K.; Suzuki, Y.; Ichihara, A. J. Chem. Soc., Chem. Commun.
1995, 1321-1322. For the catalytic antibodies for the Diels-Alder reactions,
see: (a) Hilvert, D.; Hill, K. W.; Nared, K. D.; Auditon, M.-J. M. J. Am.
Chem. Soc. 1989, 111, 9261-9262. (b) Braisted, A. C.; Schultz, P. G. Ibid.
1990, 112, 7430-7431.
Supporting Information Available: Crystal parameters,
tables of crystal structural data with thermal ellipsoid figure
including atomic coordinates, Biso/Beq, anisotropic displacement
parameters, bond lengths, bond angles, torsion angles, and non-
bonded contacts out to 3.60 Å for adducts 1‚2(2b)‚3, and figures
showing the time courses of the reactions in the absence and
presence of methyl benzoate as an inhibitor (21 pages). See
any current masthead page for ordering and Internet access
instructions.
(29) Cf., Mackay, L. G.; Wylie, R. S.; Sanders, J. K. M. J. Am. Chem.
Soc. 1994, 116, 3141-3142.
(30) When dipped in an aqueous solution of an appropriate guest such
as cyclohexanol, cyclohexanediol, and alkyl acetate, apohost 1 binds two
molecules of the guest together with 3-5 molecules of water (Aoyama,
Y.; Imai, Y.; Endo, K.; Kobayashi, K. Tetrahedron 1995, 51, 343-352).
The guest-binding turns out to be reversible. For example, exchange between
ethyl acetate and propyl acetate readily takes place when the ethyl acetate
adduct is dipped in an aqueous solution of propyl acetate (Aoyama, Y.;
Oue, Y.; Sawaki, T.; Endo, K. unpublished results).
JA964198S