Templation in the Formation of Carceplexes

Article abstract of DOI:10.1021/jo9803582

The formation of carceplexes and hemicarceplexes involves an assembly process whereby a molecular template is entrapped between two bowl-shaped molecules. The process can be highly selective such that a million-fold range in template recognition has been

Full text of DOI:10.1021/jo9803582

J . Org. Chem. 1998, 63, 4103-4110
4103
Tem p la tion in th e F or m a tion of Ca r cep lexes
Robert G. Chapman and J ohn C. Sherman*
Department of Chemistry, University of British Columbia, Vancouver, BC, Canada V6T 1Z1
Received February 26, 1998
The formation of carceplexes and hemicarceplexes involves an assembly process whereby a molecular
template is entrapped between two bowl-shaped molecules. The process can be highly selective
such that a million-fold range in template recognition has been realized. The relative template
abilities, or template ratios, for 34 molecules in forming carceplex 2‚guest is described. Species
with one, two, and three bridges linking the two “bowls” are reported and were used to delineate
the guest-determining step of the reaction. The effect of base and solvent on the reaction are also
described, and a discussion of the reaction mechanism is presented.
In tr od u ction
appears to work only if a suitable template molecule is
present, and the abilities of template molecules to form
this carceplex were found to vary by 106-fold. This
remarkable selectivity was underscored by the determi-
nation of a 75% yield of carceplex 2a ‚pyrazine (the best
template) in the presence of only a stoichiometric amount
of pyrazine in N-methylpyrrolidinone (NMP, the poorest
template)6 as solvent (i.e., 104-fold excess NMP).5 What
drives this reaction?
Molecular recognition is integral to myriad functions,
from biological processes to molecular devices.1 An
understanding of the noncovalent interactions that gov-
ern the recognition process is paramount to the elucida-
tion of newly discovered natural systems and to the
design and modification of de novo systems. An ideal
model to study such noncovalent interactions is one that
is easily and fully characterizable and shows high
selectivity in its recognition. One of the most striking
examples of such guest selectivity by a designed host
system is where guest molecules function as templates
in the formation of carceplexes, which are closed surface
compounds that incarcerate smaller molecules or ions.2a
The creation of carceplexes and their relations, hemicar-
ceplexes, has proliferated recently, yet the mechanism
of their formation is still poorly understood.2 The forma-
tion of such species is of wide interest, as the assembly
process that is responsible involves not only sophisticated
molecular recognition but also molecular encapsulation
and an unusually sensitive templation process.3
In a preliminary report, the extremely large gap in
template abilities between pyrazine and NMP was filled
in by 22 other guests-as-templates that had intermediary
template abilities.5 The templating ability of one guest
or template molecule versus another were called template
ratios. The template ratios were determined by competi-
tion experiments whereby two guests are present during
carceplex formation and the product ratios carceplex
2‚guest 1:carceplex 2‚guest 2, were determined from
1
integration of H NMR spectra to generate a template
ratio; thus, template ratios and product ratios are syn-
onomous in our system.5 What is the significance of
template ratios? Since the formation of the OCH2O
bridges between the two bowls is irreversible, there must
be a key bridge that locks the guests within the capsule
irreversibly, with respect to the time scale of the experi-
ment. This particular bridge formation would be con-
sidered the guest-determining step (GDS),6 the step
beyond which the guest no longer exchanges under the
reaction conditions. Regarding the guest/template mol-
ecules, what comes before or after the GDS is irrelevant
to their ultimate fate. That is, the ratio of template
molecules entrapped in the carceplex is determined solely
by their competition at the GDS. Thus, just as a product
ratio represents the relative rate of a product-determin-
ing step,6 a template ratio reflects the relative rate of
reaction of the GDS for two different guest molecules.
More generally, for a series of molecules the template
ratios reflect the relative rates of the GDS for that series
of guest molecules. This step is 106 times faster in the
presence of pyrazine than in the presence of NMP.
In the formation of carceplex 2‚guest, two bowl-shaped
molecules (tetrol 1) wrap around a guest/template mol-
ecule and the two “bowls” get sewn together by the
irreversible formation of four OCH2O bridges (Scheme
1).4 In this process, seven molecules come together and
eight new covalent bonds form. Yet, the reaction is
highly efficient, with yields as high as 87%.5 The reaction
(1) Comprehensive Supramolecular Chemistry; Lehn, J .-M., Atwood,
J . L., Davies, J . E. D., MacNicol, D. D., Vo¨gtle, F., Eds.; Pergamon:
New York, 1996.
(2) (a) Cram, D. J .; Karbach, S.; Kim, Y. H.; Baczynskyj, L.;
Kalleymeyn, G. W. J . Am. Chem. Soc. 1985, 107, 2575-2576. (b)
Cram, D. J .; Cram, J . M. Container Molecules and Their Guests, from
the series Monographs in Supramolecular Chemistry; Stoddart, J . F.,
Ed.; Royal Society of Chemistry: Cambridge, 1994. (c) Sherman, J .
C. Tetrahedron 1995, 51, 3395-3422. (d) Sherman, J . C. In Large Ring
Compounds; Semlyen, J . A., Ed.; J ohn Wiley & Sons, Ltd.: London,
1996; pp 507-524. (e) Helgeson, R. C.; Knobler, C. B.; Cram, D. J . J .
Am. Chem. Soc. 1997, 119, 3229-3244. (f) Helgeson, R. C.; Paek, K.;
Knobler, C. B.; Maverick, E. F.; Cram, D. J . J . Am. Chem. Soc. 1996,
118, 5590-5604. (g) van Wageningen, A. M. A.; Timmerman, P.; van
Duynhoven, J . P. M.; Verboom, W.; van Veggel, F. C. J . M.; Reindoudt,
D. N. Chem. Eur. J . 1997, 3, 639-654.
(3) Chapman, R. G.; Sherman, J . C. Tetrahedron 1997, 53, 15911-
15946, and references therein.
(4) Sherman, J . C.; Knobler, C. B.; Cram, D. J . J . Am. Chem. Soc.
1991, 113, 2194-2204.
(5) Chapman, R. G.; Chopra, N.; Cochien, E. D.; Sherman, J . C. J .
Am. Chem. Soc. 1994, 116, 369-370.
(6) Abbreviations: NMP, N-methylpyrrolidinone; GDS, guest-
determining step (the GDS is also the product-determining step; we
use GDS because our focus is on the guest or template effect in the
assembly process); DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; DMA,
dimethylacetamide; CPK, Corey-Pauling-Koltun space filling models;
∆δd, the difference in chemical shifts, usually between free and
entrapped guests; MALDI, matrix-assisted laser desorption ionization.
S0022-3263(98)00358-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 05/15/1998

4104 J . Org. Chem., Vol. 63, No. 12, 1998
Chapman and Sherman
Sch em e 1. F or m a tion of Ca r cep lex 2‚Gu est a n d Com p lex 3‚Gu est fr om Tetr ol 1
Ta ble 1. Su ita ble Tem p la te Molecu les. Th e Boxed Molecu les F or m ed Ca r cep lex 2a ‚Gu est bu t Wer e Ch a r a cter ized a s
Mixtu r es
Alternatively, the transition state for the GDS is 8.3 kcal/
mol lower in the presence of pyrazine than in the
presence of NMP at 300 K.7
Resu lts a n d Discu ssion
1. Deter m in a tion of Tem p la te Ra tios. The first
step in investigating the assembly process to form
carceplex 2‚guest is to screen for successful template
molecules. The choice of reaction solvent is critical, for
a polar solvent facilitates the bridging reactions and one
ideally wants a noncompetitive template molecule as
solvent. The relatively large dipolar aprotic solvent NMP
was chosen as the bulk solvent for screening potential
template molecules for such reasons. Thus, 70 potential
template molecules were screened by incorporating them
as cosolvents (∼5 mol % based on NMP) into the reaction
mixtures, and the reactions were run over 2 days at 60
°C. Of these 70 template molecules, 40 molecules were
found to be suitable templates for the formation of
carceplex 2a ‚guest (Table 1) while 30 molecules were
unsuitable templates (Table 2). Guests that did not lead
to formation of carceplex 2a ‚guest were classified (Table
2) in terms of their size, basicity, or polarity, as these
properties are the most likely causes for their unsuit-
ability as templates.9
But what drives the reaction? To answer this question
directly, one must examine the transition state for the
GDS. As this is not readily available, a complex formed
between two bowls and a guest molecule, complex 3‚guest,
was explored.8 It was determined that the guest selec-
tivity of this complex correlated with the template ratios.
Thus, complex 3‚guest represents a good transition state
model for the GDS in the formation of carceplex 2‚guest.8
This paper reports the determination of the GDS, the
characterization of four new intermediates containing
1-3 interbowl bridges, and the effect of base and solvent
on the reaction. In addition, the template ratios for 10
new guests are reported. Finally, a discussion of the
reaction mechanism is provided.
(7) The 8.3 kcal/mol difference in activation energy was calculated
using the following equation: ∆∆Gq ) -RT ln(kpyrazine/kNMP), where T
) 300 K, kNMP is the rate constant for formation of carceplex 2a ‚NMP,
and kpyrazine is the rate constant for formation of carceplex 2a ‚pyrazine;
kpyrazine/kNMP ) 106.
(8) Chapman, R. G.; Sherman, J . C. J . Am. Chem. Soc. 1995, 117,
9081-9082.
The diverse range of incorporated guest molecules
(Table 1) prompted us to investigate their templating
abilities further. Thus, competition experiments were

Templation in the Formation of Carceplexes
J . Org. Chem., Vol. 63, No. 12, 1998 4105
Ta ble 2. Un su ita ble Gu ests
Ta ble 3. Ca r cep lex 2a ‚Gu est Yield s a n d Com p etition
performed on 34 of the successful template molecules to
generate template ratios (Table 3).5,9-11 Each of the
template ratios were referenced to the poorest template
(NMP) which was arbitrarily given a value of 1.
2. In ter p r eta tion of Tem p la te Ra tios. The data
from Table 2 and Table 3 demonstrate some definite
trends, most notably in guest size. For instance, the
largest suitable template, NMP, is the only template to
have as many as seven non-hydrogen atoms. For the
unsuitable templates, the majority are simply too large
for the interior of the carceplex. It is fairly remarkable
just how selective the carceplex is toward the size of its
guest molecule. For example, benzene is a reasonably
Exp er im en ts
template
guest
pyrazine
methyl acetate
1,4-dioxane
dimethyl sulfide
ethyl methyl sulfide
dimethyl carbonate
DMSO
1,3-dioxolane
2-butanone
pyridine
dimethyl sulphone
1,4-thioxane
2,3-dihydrofuran
furan
tetrahydrofuran
pyridazine
acetone
thiophene
1,3-dithiolane
(()-2-butanol
benzene
2-propanol
pyrrole
tetrahydrothiophene
1,3-dioxane
acetamide
yield, %a
ratiob
condnsb
1
2
3
4
5
6
7
8
9
87
75
68
52
67
52
63
64
75
46
60
55c
38
54
50
30
51
23
43
47
43
74
73
34
45
26
24
35
38
17
14d
15
4
1000000
470000
290000
180000
130000
73000
70000
38000
37000
34000
19000
14000
13000
12000
12000
8600
6700
5800
4400
2800
2400
1500
1000
410
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
B
B
B
B
B
B
B
B
B
B
10
12
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
(9) In the presence of pyrazine and the four secondary amines,
carceplex 2‚pyrazine formed, confirming that the amines are poor
templates and were not merely degrading CH2BrCl. The cause of their
unsuitability is unknown and may be due to their solvation by NMP,
not necessarily their basicity. There are also several carceplexes not
included in Table 1 that were characterized only by 1H NMR spec-
troscopy that deserve mention. Imidazole, 1,2,4-triazole, and cyclo-
pentadiene all gave carceplexes 2a ‚guest in < 5% yields. Also,
methanol gave a mixture of carceplexes 2a ‚guest in < 5% yield; 1H
NMR spectra of this material indicated that the majority was carceplex
2a ‚NMP and carceplex 2a ‚CH2BrCl, but there were also signals at
-0.58 and -0.62 ppm that could not be assigned and may be due to
entrapped methanol. None of these guests were included in the
competition experiments.
200
160
100
73
61
45
21
20
(10) A number of the carceplexes were initially isolated as mixtures
with carceplex 2a ‚NMP. Such mixtures were obtained using the
following guests: acetamide, trioxane, acetonitrile, ethanol, ethyl
acetate, dimethylacetamide, and dimethylformamide. These mixtures,
however, were readily separated via silica gel chromatography to give
carceplex 2a ‚NMP and the respective carceplex 2a ‚guest. There were
also a number of guest molecules that led to an inseparable mixture
of carceplexes: CHCl3, diethyl ether, and tetrahydropyran furnished
an inseparable mixture of carceplexes 2a ‚guest and 2a ‚CH2BrCl.
Generally, a guest molecule is a very poor template when the small
amount of CH2BrCl present in the reaction (1 mmol of CH2BrCl
compared to 50 mmol of guest) effectively competes with it. The
carceplex 2a ‚CH2BrCl impurity in principle can be avoided by use of
trioxane
acetonitrile
ethanol
ethyl acetate
diethyl ether
dimethylacetamide
dimethylformamide
NMP
7
1
5e
a
b
Yield refers to the reaction run with only one guest. Condi-
tions A: 1 mol % guests, 2 days, 60 °C. Conditions B: 5 mol %
guests, 1 day ambient temperature, 2 days, 60 °C. c Characterized
as a mixture of carceplex 2a ‚1,4-thioxane (86%) and carceplex
a
larger bridging material. Although both dibromomethane and
diiodomethane are suitable for the bridging reaction, their template
ability surpasses that of the poor guests; therefore, they were not
suitable for this purpose. Methylene ditosylate is a bridging material
that is too large to fit in the interior of a carceplex. Indeed, this
bridging material was successfully used in the formation of carceplex
2a ‚pyrazine and could, in principle, be used to create pure carceplexes
using poorer templates. Nevertheless, carceplexes 2a ‚CHCl3,
2a ‚tetrahydropyran, and 2a ‚diethyl ether were characterized as
mixtures by 1H NMR and matrix-assisted laser desorption ionization
mass spectrometry. Of these guests, only diethyl ether was included
in our competition experiments because it only yielded a small impurity
(<4%) of carceplex 2a ‚CH2BrCl. The dihalomethanes were omitted
from the competition experiments since they act as bridging reagents.
The reactions carried out with 1,4-thioxane as guest resulted in an
inseparable mixture of carceplexes 2a ‚1,4-thioxane and 2a ‚1,4-dioxane
in a 7:1 ratio (1,4-thioxane contains ∼1% impurity of 1,4-dioxane). The
superior templating ability of 1,4-dioxane over 1,4-thioxane results in
a 14% yield of carceplex 2a ‚1,4-dioxane in the presence of a vast excess
of 1,4-thioxane.
d
2a ‚1,4-dioxane (14%). Characterized as a mixture of carceplex
2a ‚diethyl ether (96%) and carceplex 2a ‚bromochloro-
methane (4%). e Reaction run in neat NMP.
good template molecule, but fluorobenzene is unsuitable;
carceplex formation is apparently sensitive to the sub-
stitution of a hydrogen (van der Waals radius ) 1.2 Å)12
for a fluorine (van der Waals radius ) 1.47 Å),12 although
this differentiation could also be electronic in nature.
Furthermore, 1,4-dioxane is 21 times better than the
slightly larger 1,4-thioxane; moreover, 1,4-dithiane (both
oxygens replaced by sulfur) is an altogether unsuitable
(11) The experimental procedure and a sample calculation for the
generation of template ratios is given in the Supporting Information.
(12) Carey, F. A.; Sundberg, R. J . Advanced Organic Chemistry, 3rd
ed.; Plenum: New York, 1990; p 120.

4106 J . Org. Chem., Vol. 63, No. 12, 1998
Chapman and Sherman
Ta ble 4. Disa p p ea r a n ce of Tetr ol 1a a t 25 °C
template. The formation of carceplexes is very sensitive
to the addition of a single methylene. For example,
methyl acetate is 104 times better than ethyl acetate, and
1,3-dioxolane is 190 times better than 1,3-dioxane. The
addition of an oxygen can also create differences in
template abilities. For example, dimethyl sulfide is 2.5
times better than DMSO which is itself 3.7 times better
than dimethyl sulfone. Substantial differences in tem-
plate abilities exist between pyrazine, pyridine, and
benzene, which only differ by the substitution of a
nitrogen(s) for a C-H(s). Pyrazine is 29 times better as
a template than pyridine which is itself 14 times better
than benzene; the introduction of an extra hydrogen
apparently causes significant steric strain in the transi-
tion state of the GDS, although, again, electronic effects
may also be important.
% tetrol remaining
guest ) pyrazine
% tetrol remaining
guest ) NMP
time (h)
6
12
48
13
0
0
83
80
10
Sch em e 2. Sch em a tic Rep r esen ta tion of
Ca r cep lex In ter m ed ia tes (R ) CH3)
Polarity of the template molecules becomes important
only at the extremes. Although the polarity of the
suitable templates spans a diverse range from DMSO (ꢀr
) 46)13 to benzene (ꢀr ) 2),13 both highly polar molecules
such as water and highly apolar molecules such as
cyclopentane are unsuitable templates. High symmetry
tends to increase the templating power of guests as is
evident with 1,4-dioxane which is 1400 times better than
1,3-dioxane. Similarly, pyrazine (1,4-diazabenzene) is
120 times better than pyridazine (1,2-diazabenzene).
Cyclic guests are much better than acyclic guests. For
example, 1,4-dioxane is the third best template while
dimethoxyethane is an unsuitable template. Further-
more, THF is 570 times better than diethyl ether.
Another general trend that can be gleaned from Table
3 is that the yields usually increase with increasing
template abilities of the guest molecules. For example,
the best guest (pyrazine) has the highest yield and the
poorest guests (NMP and DMF) have the lowest yields.
Yet, there are many exceptions to this trend. Therefore,
the yield alone is not strictly indicative of the template
ability of the guest molecules for formation of carceplex
2a ‚guest as has been explained earlier.5
There was no correlation found between our template
ratios (Table 3) and solvent parameters such as acceptor
number, dielectric constant,13 dipole moment,13 ET,13 or
Hildebrand’s δ.14 Graphs of ln(template ratios) versus
these five solvent parameters result in correlations of r2
< 0.14. Therefore, the polarity of the template does not
appear to be a dominant factor in the template abilities
of the guest molecules. Additionally, there were no
apparent similarities between template ratios and the
theoretical MM2 calculated energies for carceplex
2a ‚guest.5 More sophisticated theoretical studies of
carceplex 2‚guest and complex 3‚guest have been per-
formed and yield a reproduction of the general trend
observed experimentally.15
search for the GDS is described below; it entailed a closer
analysis of the carceplex reaction in general.
(a ) Isola tion of Rea ction In ter m ed ia tes. Is the
selectivity observed for the formation of carceplex 2‚guest
the result of the formation of the first interbowl (O-CH2-
O) bridge? To address this question, we looked at the
rate of formation of the first bridge by following the
disappearance of the starting material (tetrol 1a ) using
both our best guest (pyrazine) and our poorest guest
(NMP) under identical reaction conditions. We found
that the disappearance of tetrol is much faster in the
presence of 1 mol % pyrazine in NMP than in neat NMP
(Table 4). This indicates that, relative to NMP, pyrazine
accelerates the formation of the first bridge in the
carceplex reaction. However, these findings do not
provide information about the actual GDS. We therefore
focused our attention on the isolation of reaction inter-
mediates (Scheme 2), with the intention of performing
competition reactions starting from each intermediate.
We expected that the corresponding template ratios
might not agree with the template ratios in Table 3 (i.e.,
starting from tetrol 1a ). If this were the case, we would
conclude that that particular intermediate is already past
the GDS on the carceplex reaction pathway.
Initially, we attempted to isolate the intermediates by
performing the reaction under normal conditions except
that the reaction was stopped prematurely. Unfortu-
nately, unreacted tetrol 1a and the final product, carce-
plex 2a ‚guest, predominated. No intermediates were
isolated. This is noteworthy, as it shows that the reaction
is highly cooperative. To isolate carceplex reaction
intermediates and in a general search for clues to what
drives this reaction, we performed the carceplex reaction
in various solvents using various bases.
3. Th e Gu est-Deter m in in g Step (GDS). To delin-
eate the driving forces for formation of carceplex 2‚guest
and the ensuing template effect, one should ideally look
at the transition state of the GDS and compare different
templates at that point. Which step is the GDS? The
(13) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry,
2nd ed.; VCH: New York, 1988, see Chapter 2 and the Appendix.
(14) Hildebrand, J . H.; Prausnitz, J . M.; Scott, R. L. Regular and
Related Solutions; Van Nostrand Reinhold Co.: New York, 1970; pp
27, 213-215.
(15) Nakamura, K.; Sheu, C.; Keating, A. E.; Houk, K. N.; Sherman,
J . C.; Chapman, R. G.; J orgensen, W. L. J . Am. Chem. Soc. 1997, 119,
4321-4322.
(b) Effect of Ba se on th e F or m a tion of Ca r cep lex
2‚Gu est. What base strength is suitable for the forma-
tion of carceplex 2‚guest? Do metal cations play a
significant role in the formation of carceplexes by forming
salt bridges between two tetrol molecules prior to cova-
lent bond formation? To answer these questions, a series

Templation in the Formation of Carceplexes
J . Org. Chem., Vol. 63, No. 12, 1998 4107
Ta ble 5. Effect of Ba se on F or m a tion of Ca r cep lex
trace amounts of carceplex reaction intermediates were
detected. It is interesting to note that the highly apolar
solvent cyclohexane led to carceplex formation.
2a ‚Gu est
equiv
of base
% yield of
2a ‚pyrazine
a
base
pKa
Of the solvents employed, nitrobenzene proved to be
the most successful because it produced carceplexes in
good yields and was itself not a competitive guest. When
tetrol 1b (pendent group ) methyl),16 2.1 equiv of DBU,
CH2I2, and 5 mol % DMSO as guest were stirred at 60
°C in nitrobenzene for 2 h, a mixture of carceplex
intermediates resulted (Scheme 2). This reaction yielded
27% monobridged 4, 16% A,B-bis-bridged 5, and 18%
recovered tetrol 1b. Other reaction products included
A,C-bis-bridged 6‚DMSO, tris-bridged 7‚DMSO, and car-
ceplex 2b‚DMSO. Intermediates 6 and 7 were formed
as mixtures with a small percentage of CH2I2 as guest
and therefore could not be fully characterized. It was,
however, possible to isolate the A,C-bis-bridged 6‚DMSO
and tris-bridged 7‚DMSO species from the reaction of
tetrol 1b with CH2BrCl and potassium carbonate in
DMSO (at ambient temperature for 18 h), in 4.6% and
22% yields, respectively; in addition, 7.6% of carceplex
2b‚DMSO was isolated under these conditions.
pyrazine
NaOAc
pyridine
NaHCO3
morpholine
triethylamine
K2CO3
DBU
LiC5H5
KOH
60
30
30
30
30
30
30
60
30
200
30
30
0.65
5
5.25
6.35
8.21
10
10
12
15
0, 1a recovered
0, 1a recovered
0, 1a recovered
78
0, 1a recovered
0, 1a recovered
80
60
5
80
60
15.7
19
40
t-BuOK
NaH
0, 1a recovered
a
pKa of conjugate acid in H2O.
Ta ble 6. Solven ts for th e Ca r cep lex Rea ction Usin g
DBU a s th e Ba se
time
(days)
% yield of
2a ‚guest
solvent
guest
acetonea,b
acetonea,d
THFa,b
acetone
acetone
THF
2
8
2
10c
65c
0c
(d ) Deter m in a tion of th e GDS. Intermediates A,C-
bis-bridged 6‚DMSO and tris-bridged 7‚DMSO each
contain a tightly encapsulated molecule of DMSO which
remained encapsulated after workup. The stability of
A,C-bis-bridged 6‚DMSO and tris-bridged 7‚DMSO was
THFa,d
THF
8
20c
30c
0e
benzenea,d
benzene
pyrazine
pyrazine
MeOAc
pyrazine
pyrazine
pyrazine
10
10
5
6
8
a,d
CHCl3
toluened,f
25
60
0
0
70
cyclohexaned,f
xylenesd,f
1
tested by recording their H NMR spectra in deuterated
d,f
CHCl2CHCl2
2
2
nitrobenzene as a function of time and temperature. No
noticeable changes in their 1H NMR spectra were ob-
served, even at temperatures as high as 120 °C for 48 h!
Therefore, the activation energy for egress of DMSO from
the holes in 6‚DMSO and 7‚DMSO is large; thus, these
compounds can be regarded as carceplexes. Neverthe-
less, we further examined 6‚DMSO and 7‚DMSO to be
sure that they are past the GDS on the carceplex reaction
pathway. We thus performed competition experiments
with 6‚DMSO and 7‚DMSO in the presence of pyrazine,
a superior template molecule, to furnish the correspond-
ing carceplex 2b‚guest. In both cases, these reactions
resulted in the isolation of only carceplex 2b‚DMSO,
which shows that no guest exchange occurs with these
intermediates; they are past the GDS on the carceplex
reaction pathway.
A,B-Bis-bridged 5 and monobridged 4 were isolated
with no detectable encapsulated guest. Examination of
these two intermediates by CPK models6 indicates that
they have significant openings that allow for easy guest
escape. Does the guest-determining step for carceplex
formation occur after the formation of these intermedi-
ates? We performed competition experiments using these
two intermediates and tetrol 1b at ambient temperature
using six of the templates from Table 3 (results from
Table 3 were obtained at 60 °C using tetrol 1a ). The
results of these competition studies are presented in
Table 7. The template ratios between tetrol 1b and
monobridged 4 correlate well (a log-log plot of the
template ratios gave r2 ) 1.0) while the template ratios
nitrobenzened,f
a
Reaction mixture was refluxed. b Twelve equivalents of CH2BrCl
were added at the start. c A trace of intermediates was formed.
Ten equivalents of CH2BrCl were added per day. e Tetrol 1a was
d
recovered. f Reaction mixture was stirred at 60 °C.
of standardized reactions were run in the presence of
various bases in NMP using pyrazine as the template
molecule and bromochloromethane as the bridging re-
agent. The results are summarized in Table 5 and
indicate that weak bases such as pyrazine, NaOAc, and
pyridine are too weak to lead to carceplex formation,
whereas exceedingly strong bases such as NaH are also
unsuitable. The suitable bases range in pKa from 6 to
19 (pKa values of conjugate acids in H2O). Although the
amine bases triethylamine and morpholine fall in the
correct pKa range, they are not suitable for formation of
carceplex 2‚guest. This may be the result of triethy-
lamine and morpholine being too weakly basic in NMP,
the solvent used for the reaction.
The successful formation of carceplex 2a ‚pyrazine with
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as base indi-
cates that metal salt bridges are not necessary for
carceplex formation. Moreover, this discovery is particu-
larly noteworthy because it greatly expands the number
of potential solvent systems that can be explored for
carceplex formation due to the greater solubility of DBU
over M2CO3 (M ) K or Cs) in apolar organic solvents.
(c) Effect of Solven t on th e F or m a tion of Ca r ce-
p lex 2‚Gu est. Although the above experiments indicate
that the formation of carceplex 2‚guest is sensitive to the
base employed, they did not provide any evidence about
carceplex reaction intermediates. We therefore explored
a variety of solvents systems for running the carceplex
reaction using DBU as the base (Table 6). Generally, the
reaction to form carceplexes proceeds slower in these low-
polarity solvents than in NMP. Nonetheless, when
acetone, THF, and benzene were used as bulk solvents,
(16) The preparation of carceplex 2b‚guest (i.e., R ) CH3) was
realized partway through this work (see ref 17). The advantages of
the methyl “feet” over the phenethyl feet are manifold, including the
following: (1) The 1H NMR spectra are simpler with methyl feet and
create a larger open window to observe entrapped guest. (2) Although
solubility was feared to be a problem, carceplex 2b‚guest is actually
more soluble in chloroform than carceplex 2a ‚guest. (3) Crystals grown
from methyl-footed carceplexes are far more stable, and their X-ray
crystal structures are easier to solve.

4108 J . Org. Chem., Vol. 63, No. 12, 1998
Chapman and Sherman
Ta ble 7. Tem p la te Ra tios Sta r tin g fr om 1b, 4, a n d 5
relate with the template ratios reported here, and thus,
the complex is a good transition state model for the GDS
in the formation of carceplex 2‚guest.8
from
from
from
guest
tetrol 1b
monobridged 4
A,B-bis-bridged 5
pyrazine
1,4-dioxane
DMSO
pyridine
acetone
860.0
177.0
18.8
13.5
1.8
1130.0
245.0
21.0
17.0
2.0
40.0
5.4
17.0
19.0
3.0
Con clu sion s
The mechanism for formation of carceplex 2‚guest as
we understand it to date is as follows: Complex 3‚guest
forms, followed by formation of the first bridge, with
guests still in fast exchange. The formation of the second
bridge, either A,B or A,C, is then the GDS. The host
species formed during the transition state of this step is
highly rigid and binds strongly with high selectivity to
guests such as pyrazine. The most stable complexes (i.e.,
those containing the best template molecule) form the
second bridge the fastest, thereby entrapping more of
that guest under the reaction conditions. Subsequent
bridging leads to product.
benzene
1.0
1.0
1.0
for tetrol 1b and A,B-bis-bridged 5 show a poor correla-
tion (a log-log plot of the template ratios yields r2 ) 0.7).
Combining the results for the monobridged 4 and A,B-
bis-bridged 5 with those of the A,C-bis-bridged 6‚DMSO
and the tris-bridged 7‚DMSO, we conclude that the
formation of the second OCH2O bridge in the carceplex
reaction (either A,B-bis or A,C-bis) is the GDS. After
formation of this second OCH2O bridge, the guest is
irreversibly entrapped under the reaction conditions, and
subsequent bridging leads to the corresponding product
(carceplex 2‚guest). Therefore, the formation of the
second bridge is 106 times faster in the presence of
pyrazine than in the presence of NMP. We suggest that
the poor correlation (r2 ) 0.7) for the template ratios
determined for the formation of carceplex 2b‚guest from
A,B-bis-bridged 5 versus tetrol 1b is a result of guest
exchange being slower than formation of the third
OCH2O bridge (thus, the normal template ratios are
skewed by the guest’s in-rate). Indeed, in experiments
to be reported shortly,18 charged complexes corresponding
to A,B-bis-bridged 5‚guest were found to exchange guests
over a period of days, which is far slower than bridge
formation. Interestingly, neutral complexes correspond-
ing to A,B-bis-bridged 5‚guest exchange their guests over
a period of minutes; thus, neutral A,B-bis-bridged 5 is
isolated without guest.
5. P r ea ssocia tion of Bow ls P r ior to Br id ge F or -
m a tion . Investigation into the template requirements
for the formation of carceplex 2‚guest led to a similar
investigation into the template requirements for the
formation of a hemicarceplex where one of the four
(ArOCH2OAr) linkages is replaced by (ArH, HAr), thus
leaving a hole in the side of the capsule.19 The reaction
to form this hemicarceplex is statistically more demand-
ing than the reaction to form carceplex 2‚guest because
there is the potential for misalignment during formation
of the first bridge between two starting triol molecules.20
Template ratios were determined, and a good correlation
was found with the template ratios measured for the
formation of carceplex 2a ‚guest (r ) 0.97).19 Therefore,
the driving forces are similar for the two processes. The
favorable interactions between the template and the
forming host compounds overwhelm any disparity that
might be caused by the lower symmetry of the hemicar-
ceplex. Also, much greater than statistical yields were
observed, which suggests that hydrogen bonding may
play an integral role in aligning the bowls prior to
covalent bond formation. Indeed, in complex 3‚guest, two
molecules of tetrol 1 are linked via charged hydrogen
bonds. The guest selectivities of complex 3‚guest cor-
The carceplex reaction discussed here is an excellent
model for the study of noncovalent interactions between
molecules because small changes in guest size, shape,
and/or electronic properties results in vast differences in
incarceration propensities. For example, the substitution
of a nitrogen of pyrazine with a methine group to give
pyridine disrupts the complementarity of pyrazine with
the forming cavity as illustrated by the 29-fold decrease
in the template ratio. Substitution of both nitrogens of
pyrazine for two methines to give benzene results in a
420-fold decrease in the template effect.
Pyrazine’s favorable noncovalent interactions with the
forming carceplex lowers the Gibbs free energy of activa-
tion for this step by 8.3 kcal/mol relative to NMP at 300
K. The ability of the forming carceplex to recognize
subtle changes in guest properties ultimately led to a
million-fold difference in template ratios between the
poorest template (NMP) and the best template (pyrazine).
The formation of the second OCH2O interbowl bridge was
found to be the step responsible for the selectivity (GDS)
observed in the formation of carceplex 2‚guest. Poorer
guests/templates such as NMP may provide fewer favor-
able noncovalent interactions and more unfavorable
interactions in the transition state of the GDS, thus
leading to slower formation of the second bridge. Guests
such as pyrazine provide many favorable noncovalent
interactions that stabilize the transition state for forma-
tion of the second OCH2O bridge.
Better guests generally afford higher yields of carceplex
2‚guest. Poorer guests most likely allow more polymer-
ization to occur from the monobridged species. In addi-
tion, some guests may cause low yields because they can
react with the bridging material (e.g., dimethyl sulfide,
pyridine, thiophene). Some guests may lower yields by
distortion of the host after the GDS and, thus, inhibit
final bridging (e.g., benzene, DMA, NMP). These ten-
dencies may lower the yields somewhat, but they will
have only a small effect on the template ratios.5
This study of 34 carceplexes has provided valuable
insight into the nature of noncovalent interactions.
Current studies include a detailed analysis of complex
3‚guest and related species as well as investigations into
the template effect in systems that contain larger cavities
and those that cannot form complex 3‚guest. Hopefully,
the information gained from these studies will enable us
to create more complex assemblies such as very large
carceplexes and polymeric capsules.
(17) Fraser, J . R.; Borecka, B.; Trotter, J .; Sherman, J . C. J . Org.
Chem. 1995, 60, 1207-13.
(18) Chapman, R. G.; J .; Sherman, J . C. Unpublished results.
(19) Chopra, N.; Sherman, J . C. Supramol. Chem. 1995, 5, 31-7.
(20) A derivative of compound 7 where the pendent groups are
phenethyls has been reported: Kurdistani, S. K.; Robbins, T. A.; Cram.
D. J . J . Chem. Soc., Chem Commun. 1995, 1259-1260.

Templation in the Formation of Carceplexes
J . Org. Chem., Vol. 63, No. 12, 1998 4109
gel gravity column (20 g) and eluted with CHCl3/hexanes (3:
1), affording 2a ‚(()-2-butanol as a white solid which was
recrystallized from CHCl3/EtOAc and dried at 110 °C (0.1
mmHg) for 24 h (51 mg, 47%): mp >250 °C; 1H NMR (CDCl3,
400 MHz) δ 7.12-7.24 (m, 40H, Ha, Hb, Hc, Ha′, Hb′, and Hc′),
6.73 (s, 4H, Hd or Hd′), 6.71 (s, 4H, Hd or Hd′), 6.58 (s, 8H, He),
6.16 (d, J ) 7.3 Hz, 8H, Hf and Hf ′), 4.89 (t, J ) 7.9 Hz, 8H,
Hg and Hg′), 4.46 (d, J ) 7.3 Hz, 4H, Hh or Hh′), 4.37 (d, J )
7.3 Hz, 4H, Hh or Hh′), 2.65 (m, 16H, Hi and Hi′), 2.44 (m, 16H,
Hj and Hj′), 0.91 (br, 1H, CH3CHOHCH2CH3), -0.93 (br, 2H,
CH3CHOHCHxHyCH3), -1.32 (br, 1H, CH3CHOHCHxHyCH3),
-3.27 (d, J ) 6.0 Hz, 3H, CH3CHOHCH2CH3), -3.47 (t, J )
7.3 Hz, 3H, CH3CHOHCH2CH3); MS (MALDI) m/z (rel inten-
sity) 2180 ((M‚CH3CHOHCH2CH3 + Na+)+; 100) calcd for
Exp er im en ta l Section
Gen er a l Exp er im en ta l. See also ref 17. Nitrobenzene
and NMP were stirred over BaO for 24 h, distilled under
reduced pressure, and stored under N2 over 4 Å molecular
sieves prior to use. All other commercially available reagents
were used as purchased without further purification unless
stated otherwise. Desorption chemical ionization (DCI) and
liquid secondary ion mass spectrometry (LSIMS) mass spectra
were recorded on a Kratos Concept II HQ, and matrix assisted
laser desorption ionization (MALDI) mass spectra were re-
corded on a VG Tofspec in reflectron mode with 3,5-dihydroxy-
benzoic acid used as the matrix. Melting points were mea-
sured on a Mel-Temp II apparatus. 1H NMR spectra were
recorded on a Bruker WH-400 spectrometer in CDCl3 at
ambient temperature using the residual 1H as a reference (7.24
ppm) unless noted otherwise. For characterization of carce-
plexes 2a ‚DMF, 2a ‚DMA, and 2a ‚DMSO, see ref 4. For
characterization of carceplex 2b‚pyrazine, see ref 17.
C136H122O25‚Na+ ) 2179. Anal. Calcd for C136H122O25
: C,
75.75; H, 5.70. Found: C, 75.35; H, 5.70.
Mon obr id ged 4 a n d A,B-Bis-Br id ged 5. A mixture of
tetrol 1b (656 mg, 1.0 mmol), DBU (0.314 mL, 2.1 mmol),
DMSO (5.0 mL, 70 mmol), and CH2I2 (0.80 mL, 10.0 mmol) in
nitrobenzene (100 mL) was stirred at 60 °C for 2 h. The
reaction mixture was concentrated in vacuo, water (50 mL)
was added, and the slurry was acidified with 2 M HCl. The
slurry was extracted with EtOAc (3 × 100 mL), and the
combined organic extracts were washed with saturated aque-
ous NaHCO3 (50 mL) and brine (50 mL) and dried over
anhydrous MgSO4. Silica gel (5 g) was added to the organic
solution, and the solvent was removed in vacuo. The silica
gel-absorbed sample was dry loaded onto a silica gel gravity
column (400 g) and eluted with MeOH/CHCl3 (1:9). Reverse
phase TLC with acetone/H2O (4:1) as eluent was used to follow
the progress of the column chromatography. The products
were recrystallized from benzene/ether/acetone/hexane and
dried at 110 °C (0.1 mmHg) for 24 h affording 1b (120 mg,
18.3%, recovered starting material), 5 (180 mg, 27.2%), and 4
(105 mg, 15.7%) as white solids. Compound 4 was character-
ized as follows: mp >250 °C; 1H NMR (CDCl3, 400 MHz) δ
6.98 (s, 2H, Ha), 6.73 (s, 2H, Hb), 6.72 (s, 4H, Hc), 5.94 (d, J )
6.8 Hz, 4H, He or He′), 5.65 (d, J ) 7.0 Hz, 4H, He or He′), 5.41
(s, 2H, Hd), 5.33 (br, 6H, Hh and Hh′), 4.91 (m, 8H, Hf and Hf ′),
4.45 (d, J ) 6.8 Hz, 4H, Hg or Hg′), 4.33 (d, J ) 7.0 Hz, 4H, Hg
or Hg′), 1.70 (m, 24H, CH3); MS (DCI, ammonia) m/z (rel
2a ‚P yr a zin e. A mixture of tetrol 1a (102 mg, 0.10 mmol),
pyrazine (420 mg, 5.3 mmol), K2CO3 (1.4 g, 10 mmol), and
CH2BrCl (65 mL, 1.0 mmol) in NMP (50 mL) was stirred at
60 °C for 24 h. An additional 1.0 mmol of CH2BrCl was added,
and the reaction was stirred for an additional 24 h at 60 °C.
The reaction mixture was concentrated in vacuo, water (50
mL) was added, and the slurry was acidified with 2 M HCl.
The slurry was extracted with CHCl3 (3 × 60 mL), and the
combined organic extracts were washed with saturated aque-
ous NaHCO3 (30 mL) and brine (30 mL) and dried over
anhydrous MgSO4. Silica gel (0.5 g) was added to the CHCl3
solution, and the solvent was removed in vacuo. The silica
gel-absorbed sample was dry loaded onto a silica gel gravity
column (20 g) and eluted with CHCl3/hexanes (3:1), affording
2a ‚pyrazine as a white solid which was recrystallized from
CHCl3/EtOAc and dried at 110 °C (0.1 mmHg) for 24 h (97
1
mg, 87%): mp >250 °C; H NMR (CDCl3, 400 MHz) δ 7.13-
7.24 (m, 40H, Ha, Hb, and Hc), 6.93 (s, 8H, Hd), 6.47 (s, 8H,
He), 6.02 (d, J ) 7.5 Hz, 8H, Hf), 4.90 (t, J ) 7.9 Hz, 8H, Hg),
4.26 (d, J ) 7.5 Hz, 8H, Hh), 4.07 (s, 4H, C4H4N2), 2.66 (m,
16H, Hi), 2.51 (m, 16H, Hj); MS (DCI, isobutane) m/z (rel
intensity) 2070 ((M - CH2 + 2H)+; 100), 2082 (M+; 80), 2162
((M‚C4H4N2)+; 90), 2204 ((M‚C4H4N2 + CH(CH3)2)+; 25). Anal.
Calcd for C136H116O24N2: C, 75.54; H, 5.41; N, 1.30. Found:
C, 75.41; H, 5.35; N, 1.23.
intensity) 1342 ((M
73H64O24‚H2O: C, 65.27; H, 4.95. Found: C, 65.28; H, 5.08.
+
NH4)+; 100). Anal. Calcd for
C
1
Compound 5 was characterized as follows: mp >250 °C; H
NMR (CDCl3, 400 MHz) δ 6.82 (s, 4H, Ha), 6.72 (s, 4H, Hb),
6.62 (d, J ) 6.2 Hz, 2H, Hc or Hd), 6.41 (br, 2H, Hc or Hd), 6.19
(d, J ) 7.4 Hz, 2H, He′ or He′′), 6.09 (d, J ) 7.1 Hz, 4H, He),
5.91 (d, J ) 6.8 Hz, 2H, He′ or He′′), 5.42 (s, 4H, OH), 5.05 (q,
J ) 7.4 Hz, 2H, Hf ′ or Hf ′′), 4.89 (m, 6H, Hf and (Hf ′ or Hf ′′)),
4.61 (br, 2H, Hg′ or Hg′′), 4.59 (m, 6H, Hg and (Hg′ or Hg′′)), 1.72
(d, J ) 7.4 Hz, 6H, CH3), 1.69 (d, J ) 7.4 Hz, 6H, CH3), 1.66
(d, J ) 7.4 Hz, 12H, CH3); MS (DCI, ammonia) m/z (rel
2a ‚(()-2-Bu t a n ol. A mixture of tetrol 1a (102 mg, 0.10
mmol), (()-2-butanol (2.5 mL, 27 mmol), K2CO3 (1.4 g, 10
mmol), and CH2BrCl (65 mL, 1.0 mmol) in NMP (50 mL) was
stirred at room temperature for 24 h. An additional 1.0 mmol
of CH2BrCl was added, and the reaction was stirred at 60 °C
for an additional 48 h. The reaction mixture was concentrated
in vacuo, water (50 mL) was added, and the slurry was
acidified with 2 M HCl. The slurry was extracted with CHCl3
(3 × 60 mL), and the combined organic extracts were washed
with saturated aqueous NaHCO3 (30 mL) and brine (30 mL)
and dried over anhydrous MgSO4. Silica gel (0.5 g) was added
to the CHCl3 solution, and the solvent was removed in vacuo.
The silica gel-absorbed sample was dry loaded onto a silica
intensity) 1354 ((M
74H64O24‚2H2O: C, 64.72; H, 4.99. Found: C, 64.75; H, 4.68.
6‚DMSO a n d 7‚DMSO. A mixture of tetrol 1b (1.0 g, 1.52
+
NH4)+; 100). Anal. Calcd for
C
mmol), K2CO3 (6.0 g, 43.4 mmol), and CH2BrCl (1.0 mL, 15.4
mmol) in DMSO (125 mL) was stirred at room temperature
for 18 h. The reaction mixture was concentrated in vacuo,
water (50 mL) was added, and the slurry was acidified with 2
M HCl. The slurry was extracted with CHCl3/EtOAc 2:1 (3 ×

4110 J . Org. Chem., Vol. 63, No. 12, 1998
Chapman and Sherman
for C76H70O25S‚1.5H2O: C, 63.28; H, 5.10. Found: C, 63.25;
H, 4.94. 7‚DMSO was characterized as follows:20 mp >250 °C;
1H NMR (CDCl3, 400 MHz) δ 6.91 (s, 4H, Ha), 6.88 (s, 2H, Hb),
6.68 (s, 2H, Hc), 6.64 (d, J ) 6.4 Hz, 2H, Hd or He), 6.52 (s, 2H,
Hf), 6.45 (br, 2H, Hd or He), 6.20 (d, J ) 7.6 Hz, 4H, Hg or Hg′),
5.97 (d, J ) 7.2 Hz, 4H, Hg or Hg′), 5.70 (s, 2H, OH), 5.00 (q,
J ) 7.5 Hz, 4H, Hh or Hh′), 4.92 (q, J ) 7.5 Hz, 4H, Hh or Hh′),
4.63 (br, 4H, Hi or Hi′), 4.19 (br, 4H, Hi or Hi′), 1.71 (d, J ) 7.5
Hz, 12H, CH3), 1.68 (d, J ) 7.5 Hz, 12H, CH3), -1.22 (s, 6H,
(CH3)2SO); MS (MALDI) m/z (rel intensity) 1451 ((M‚(CH3)2SO
+ Na+)+; 100), calcd for C77H70O25S‚Na+ ) 1450. Anal. Calcd
for C77H70O25S‚1H2O: C, 63.98; H, 5.02. Found: C, 64.16; H,
4.92.
Ack n ow led gm en t. We thank NSERC of Canada for
financial support. R.G.C. thanks NSERC and UBC for
graduate fellowships. Acknowledgment is made to the
donors of the Petroleum Research Fund, administered
by the American Chemical Society, for partial support
of this research.
150 mL), and the combined organic extracts were washed with
saturated aqueous NaHCO3 (50 mL) and brine (50 mL) and
dried over anhydrous MgSO4. Silica gel (5 g) was added to
the organic solution, and the solvent was removed in vacuo.
The silica gel-absorbed sample was dry loaded onto a silica
gel gravity column (400 g) and eluted with 3% MeOH in CHCl3.
The products were recrystallized from ether/acetone/hexane
and dried at 110 °C (0.1 mmHg) for 24 h affording 6‚DMSO
(50 mg, 4.6%), 7‚DMSO (240 mg, 22.1%), and 2b‚DMSO (83
mg, 7.6%) as white solids. 6‚DMSO was characterized as
Su p p or tin g In for m a tion Ava ila ble: Preparation and
characterization of 40 carceplexes (2a ‚guest and 2b‚guest) and
procedures for determining template ratios (20 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
1
follows: mp >250 °C; H NMR (CDCl3, 400 MHz) δ 6.91 (s,
4H, Ha), 6.69 (s, 4H, Hb), 6.56 (s, 4H, Hc), 6.04 (d, J ) 7.2 Hz,
8H, Hd), 5.74 (s, 4H, OH), 4.93 (q, J ) 7.3 Hz, 8H, He), 4.31
(br, 8H, Hf), 1.69 (d, J ) 7.3 Hz, 24H, CH3), -1.15 (s, 6H,
(CH3)2SO); MS (MALDI) m/z (rel intensity) 1438 ((M‚(CH3)2SO
+ Na+)+; 100), calcd for C76H70O25S‚Na+ ) 1438. Anal. Calcd
J O9803582