Progress toward a rationally designed, chemically powered rotary molecular motor

Article abstract of DOI:10.1021/ja066044a

Building on prototype 1, which achieves 120° of phosgene-powered unidirectional rotation to rotamer 6 (see Figure 5 in the full article), 7 was designed to accomplish repeated unidirectional rotation (see Scheme 7). Compound 7 contains an amino group on each blade of the triptycene and a 4-(dimethylamino)pyridine (DMAP) unit to selectively deliver phosgene (or its equivalent) to the amine in the "firing position". The synthesis of 7 is described: the key constructive steps are a benzyne addition to an anthracene to generate the triptycene, a stilbene photocyclization to construct the helicene, and a Stille coupling to incorporate the DMAP unit. The DMAP unit was shown to regioselectively relay 1,1′-carbonyldiimidazole (but not phosgene) to the proximal amino group, as designed, but rotation of the triptycene does not occur. Extensive attempts to troubleshoot the problem led to the conclusion that the requisite intramolecular urethane formation, as demonstrated in the prototype (1 → 4), does not occur with 7 (to give 85) or 97 (to give 100). We speculate that either (i) hydrogen bonding between the hydroxypropyl group and functionality present in 7 but absent from 1 or (ii) a Buergi-Dunitz (or similar) interaction involving the DMAP (see 106) prevents achievement of a conformation conducive to intramolecular urethane formation.

Full text of DOI:10.1021/ja066044a

Published on Web 12/16/2006
Progress toward a Rationally Designed, Chemically Powered
Rotary Molecular Motor
T. Ross Kelly,* Xiaolu Cai, Fehmi Damkaci, Sreeletha B. Panicker, Bin Tu,
Simon M. Bushell, Ivan Cornella, Matthew J. Piggott, Richard Salives,
Marta Cavero, Yajun Zhao, and Serge Jasmin
Contribution from the E. F. Merkert Chemistry Center, Department of Chemistry, Boston
College, Chestnut Hill, Massachusetts 02467
Received August 19, 2006; E-mail: ross.kelly@bc.edu
Abstract: Building on prototype 1, which achieves 120° of phosgene-powered unidirectional rotation to
rotamer 6 (see Figure 5 in the full article), 7 was designed to accomplish repeated unidirectional rotation
(see Scheme 7). Compound 7 contains an amino group on each blade of the triptycene and a
4-(dimethylamino)pyridine (DMAP) unit to selectively deliver phosgene (or its equivalent) to the amine in
the “firing position”. The synthesis of 7 is described: the key constructive steps are a benzyne addition to
an anthracene to generate the triptycene, a stilbene photocyclization to construct the helicene, and a Stille
coupling to incorporate the DMAP unit. The DMAP unit was shown to regioselectively relay 1,1′-
carbonyldiimidazole (but not phosgene) to the proximal amino group, as designed, but rotation of the
triptycene does not occur. Extensive attempts to troubleshoot the problem led to the conclusion that the
requisite intramolecular urethane formation, as demonstrated in the prototype (1 f 4), does not occur with
7 (to give 85) or 97 (to give 100). We speculate that either (i) hydrogen bonding between the hydroxypropyl
group and functionality present in 7 but absent from 1 or (ii) a Bu¨rgi-Dunitz (or similar) interaction involving
the DMAP (see 106) prevents achievement of a conformation conducive to intramolecular urethane
formation.
The construction of motors of ever smaller sizes has
fascinated scientists for decades. In 1959, the Nobel laureate
physicist Richard Feynman posted1 a $1000 reward for the first
“operating electric motor [that] is only 1/64 inch cube.” The
reward was collected within a year (see Figure 1). More recently,
microfabrication using photolithography techniques has led
to motors whose diameters are about that of a human hair
(Figure 2).2
The motors illustrated in Figures 1 and 2 are rotary motors.
Barring major developments in subatomic physics, the ultimate
in miniaturized motors, rotary or otherwise, would be molecular-
scale motors. Nature4 has evolved a number of complex
molecular systems that function as rotary motors, including F1-
ATPase (Figure 3)5 and flagella (Figure 4).6 Both F1-ATPase
(1) Feynman, R. P. There is plenty of room at the bottom. In Miniaturization;
Gilbert, H. D., Ed.; Reinhold: New York, 1961; Chapter 16, pp 282-296.
Figure 1. Bottom right: William McLellan’s creation that collected
(2) Howe, R. T.; Muller, R. S.; Gabriel, K. J.; Trimmer, W. S. N. IEEE
Feynman’s reward of $1000 for the first “operating electric motor [that] is
Spectrum 1990, 27, 29-35 and cover. See also: Wise, K. D.; Najafi, K.
only 1/64 inch cube.” The “scale bar” at the top is the head of a pin.
[Reprinted with permission; see ref 3.]
Science 1991, 254, 1335-1342.
(3) Figure 1: Photo courtesy of the Archives, California Institute of Technology.
(4) For an overview, see: Berg, J. M.; Tymoczko, J. L.; Stryer, L. Molecular
Motors. Biochemistry, 6th ed.; W. H. Freeman: New York, 2006; Chapter
34.
and flagella are chemically powered motors: F1-ATPase is
driven by hydrolysis of ATP, while flagella are fueled by a
proton gradient.
(5) (a) Boyer, P. D. Angew. Chem., Int. Ed. 1998, 37, 2297-2307. (b) Walker,
J. E. Angew. Chem., Int. Ed. 1998, 37, 2308-2319. For a review, see:
Noji, H. In Molecular Motors; Schliwa, M., Ed.; Wiley-VCH: Weinheim,
2003; pp 141-152.
In 1999, we achieved a proof of principle of the first rationally
designed, chemically powered rotary molecular motor.9 In the
same year, Koumura, Feringa, and colleagues reported10 a light-
driven rotary molecular motor. In 2005, Feringa and co-workers
accomplished11 a chemically powered rotary molecular motor
(6) Berg, H. C. Annu. ReV. Biochem. 2003, 72, 19-54.
(7) Figure 3: Graphic downloaded with permission from the website
www.mech.northwestern.edu/courses/389.S02/intro.html. See also: Wang,
H.; Oster, G. Nature 1998, 396, 279-282.
(8) Figure 4: Graphic reproduced with permission from Berg, J. M.; Tymoczko,
J. L.; Stryer, L. Biochemistry, 5th ed.; W. H. Freeman: New York, 2002;
p 968.
9
376
J. AM. CHEM. SOC. 2007, 129, 376-386
10.1021/ja066044a CCC: $37.00 © 2007 American Chemical Society

Toward a Chemically Powered Rotary Molecular Motor
A R T I C L E S
Figure 4. Schematic representation of a flagellum. [Reprinted with
permission; see ref 8. Copyright 2002 W. H. Freeman and Co.]
Figure 2. An electric motor created by using photolithography possessing
a 100 µm diameter rotor. [Photo courtesy of R. S. Muller, University of
California, Berkeley. Reprinted with permission; see ref 2.]
While the demonstration that the prototype behaved as
designed was gratifying, much still remained to be addressed.
In particular, to advance 1 to a continuously rotating motor, as
summarized conceptually in Figure 6, four things need to be
achieved:
(i) an amino group must be incorporated on each blade of
the triptycene;
(ii) a means for selectively delivering phosgene (or its
equivalent) to the amine in the “firing position” must be devised;
(iii) a phosgene-fueled, 120° rotation of the triptycene must
be brought about by formation of an intramolecular urethane;
and
(iv) the remains of the phosgene must be removed by cleavage
of the urethane to allow subsequent repetition of the three
preceding steps.
We now report the achievement of the first two objectives:
construction of a triaminotriptycene assembly and incorporation
of a 4-(dimethylamino)pyridine (DMAP) unit to selectively
deliver the phosgene equivalent to the proximal amino group.
Those achievements were accomplished with 7 (Figure 7).15
Figure 3. The naturally occurring rotary motor F1-ATPase. [Reprinted by
permission from Macmillan Publishers Ltd.: Nature, copyright 1998; see
ref 7.]
(13) (a) Leigh, D. A.; Wong, J. K. Y.; Dehez, F.; Zerbeto, F. Nature 2003, 424,
174-179. (b) Hernandez, J. V.; Kay, E. R.; Leigh, D. A. Science 2004,
306, 1532-1537. (c) Hawthorne, M. F.; Zink, J. I.; Skelton, J. M.; Bayer,
M. J.; Liu, C.; Livshits, E.; Baer, R.; Neuhauser, D. Science 2004, 303,
1849-1851. (d) Magnera, T. F.; Michl, J. Altitudinal Surface-Mounted
Molecular Rotors. In Molecular Machines; Kelly, T. R., Ed.; Topics in
Current Chemistry 262; Springer: Berlin, 2005; pp 63-97. (e) Fujita, T.;
Kuwahara, S.; Harada, N. Eur. J. Org. Chem. 2005, 4533-4543. (f)
Kuwahara, S.; Fujita, T.; Harada, N. Eur. J. Org. Chem. 2005, 4544-
4556. (g) Khuong, T.-A. V.; Nunez, J. E.; Godinez, C. E.; Garcia-Garibay,
M. A. Acc. Chem. Res. 2006, 39, 413-422. (h) Wang, L; Hampel, F.;
Gladysz, J. A. Angew. Chem., Int. Ed. 2006, 45, 4372-4375. For some
earlier work, see: (i) Iwamura, H.; Mislow, K. Acc. Chem. Res. 1988, 21,
175-182. (j) Kelly, T. R.; Bowyer, M. C.; Bhaskar, K. V.; Bebbington,
D.; Garcia, A.; Lang, F.; Kim, M. H.; Jette, M. P. J. Am. Chem. Soc. 1994,
116, 3657-3658. (k) Bedard, T. C.; Moore, J. S. J. Am. Chem. Soc. 1995,
117, 10662-10671. (l) Kelly, T. R.; Sestelo, J. P.; Tellitu, I. J. Org. Chem.
1998, 63, 3655-3665. (m) Joachim, C.; Gimzewski, J. K. In Molecular
Machines and Motors; Sauvage, J.-P., Ed.; Structure & Bonding 99;
Springer-Verlag: Berlin, 2001; pp 1-18. (n) Raehm, L.; Sauvage, J.-P. In
Molecular Machines and Motors; Sauvage, J.-P., Ed.; Structure & Bonding
99; Springer-Verlag: Berlin, 2001; pp 55-78.
capable of repeated rotation, and Branchaud et al.12 have report-
ed related findings. A number of other groups13 have proffered
complementary approaches to achieving rotary motion on a molec-
ular scale, and, as numerous recent reviews attest,14 much work
has also been devoted to constructing other molecular devices.
The operation of our prototype is illustrated in Figure 5; the
system is fueled by phosgene and accomplishes a unidirectional,
120° clockwise rotation.
(9) (a) Kelly, T. R.; De Silva, H.; Silva, R. A. Nature 1999, 400, 150-152.
(b) Kelly, T. R.; Silva, R. A.; De Silva, H.; Jasmin, S.; Zhao, Y. J. Am.
Chem. Soc. 2000, 122, 6935-6949. (c) Kelly, T. R. Acc. Chem. Res. 2001,
34, 514-522. (d) Kelly, T. R.; Sestelo, J. P. In Molecular Machines and
Motors; Sauvage, J.-P., Ed.; Structure & Bonding 99; Springer-Verlag:
Berlin, 2001; pp 19-53.
(10) Koumura, N.; Zijlstra, R. W. J.; van Delden, R. A.; Harada, N.; Feringa,
B. L. Nature 1999, 401, 152-155. For more recent, related work, see: (a)
Vicario, J.; Walko, M.; Meetsma, A.; Feringa, B. L. J. Am. Chem. Soc.
2006, 128, 5127-5135. (b) Eelkema, R.; Pollard, M. M.; Vicario, J.;
Katsonis, N.; Ramon, B. S.; Bastiaansen, C. W. M.; Broer, D. J.; Feringa,
B. L. Nature 2006, 440, 163 and references therein.
(11) van Delden, R. A.; ter Wiel, M. J.; Pollard, M. M.; Vicario, J.; Koumura,
N.; Feringa, B. L. Nature 2005, 437, 1337-1340.
(12) (a) Dahl, B. J.; Branchaud, B. P. Tetrahedron Lett. 2004, 45, 9599-9602.
(b) Lin, Y.; Dahl, B. J.; Branchaud, B. P. Tetrahedron Lett. 2005, 46, 8359-
8362.
(14) (a) Balzani, V.; Credi, A.; Raymo, F. M.; Stoddart, J. F. Angew. Chem.,
Int. Ed. 2000, 39, 3348-3391. (b) Molecular Machines and Motors;
Sauvage, J.-P., Ed.; Structure & Bonding 99; Springer-Verlag: Berlin, 2001.
(c) Stoddart, J. F., Guest Ed. Molecular Machines Special Issue. Acc. Chem.
Res. 2001, 34, 409-522. (d) Balzani, V.; Venturi, M.; Credi, A. Molecular
DeVices and Machines: A Journey into the Nanoworld; Wiley-VCH:
Weinheim, 2003. (e) Molecular Motors; Schliwa, M., Ed.; Wiley: Wein-
heim, 2003. (f) Molecular Machines; Kelly, T. R., Ed.; Topics in Current
Chemistry 262; Springer: Berlin, 2005; pp 1-227. (g) Kottas, G. S.; Clarke,
L. I.; Horinek, D.; Michl, J. Chem. ReV. 2005, 105, 1281-1376. (h) Kelly,
T. R. Angew. Chem., Int. Ed. 2005, 44, 4124-4127.
(15) Note: The work on 7 did not lend itself to preliminary communications,
so there are no earlier papers.
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 377

A R T I C L E S
Kelly et al.
Figure 5. Sequence of events in the chemically powered rotation of 1 to 6.
acid chloride or anhydride), followed by attack of the nucleo-
phile on the N-acylpyridinium salt. The acylation of amines with
acyl chlorides is usually so fast that it does not require the
addition of a catalyst. However, it has been documented19 that
the DMAP-catalyzed reaction of m-chloroaniline with benzoyl
chloride is about 106 times faster than the noncatalyzed reaction,
suggesting a decidedly higher reactivity of acid chlorides toward
DMAP than toward anilines. We thus envisioned16 that inclusion
of a suitably positioned DMAP group in our design of a
molecular motor would allow for selective intramolecular
delivery of phosgene to only the amino group in the firing
position, i.e., the one situated proximate to the DMAP moiety
(see bold arrow in Figure 7). The design of 7, i.e., the placement
of the DMAP unit, was arrived at using Spartan-based (pBP/
DN**//AM1) molecular modeling.20 Modeling of 7 shows that
once the DMAP unit has reacted with the phosgene, the carbonyl
carbon of the resulting acylated species (acylpyridinium ion) is
in very close proximity to the nearby aniline for conformations
that are close in energy to the ground-state conformation by
rotation around bond a and/or low-energy rotation around bond
b in Figure 7 (on the basis of earlier work,21 the barrier to full
rotation around bond b is estimated to be 20-25 kcal/mol).
Figure 6. Schematic for a continually rotating molecular motor involving
selective (and repeated) delivery of Cl2CdO to the amino group in the
“firing” position and cleavage of the urethane only after each 120° of rotation
has occurred.
Figure 7. Proposed repeatedly rotating motor.
Model studies with 8 had previously16 demonstrated that the
presence of the DMAP unit in 8 directs monoacylation
exclusively to the adjacent amino group. In the case of the non-
pyridine control (9), a mixture of both possible monoacylation
products was obtained.
Synthesis. Given the foregoing rationale, the synthesis of 7
was undertaken. The synthesis of 7 was patterned as closely as
(17) For reviews, see: (a) Sheinkman, A. K.; Suminov, S. I.; Kost, A. N. Russ.
Chem. ReV. 1973, 42, 642-661. (b) Hassner, A.; Krespki, L. R.; Alexanian,
V. Tetrahedron 1978, 34, 2069-2076. (c) Ho¨fle, G.; Steglich, W.;
Vorbru¨ggen, H. Angew. Chem., Int. Ed. Engl. 1978, 17, 569-583. (d)
Ragnarsson, U.; Grehn, L. Acc. Chem. Res. 1998, 31, 494-501. (e) Spivey,
A.; Maddaford, A.; Redgrave, A. J. Org. Prep. Proced. Int. 2000, 32, 333-
365. (f) Murugan, R.; Scriven, E. F. V. Aldrichim. Acta 2003, 36, 21-27.
(g) Spivey, A. C.; Arseniyadis, S. Angew. Chem., Int. Ed. 2004, 43, 5436-
5441.
(18) For some recent publications involving DMAP derivatives, see, inter alia:
(a) Vedejs, E.; Chen, X. J. Am. Chem. Soc. 1996, 118, 1809-1810. (b)
Kawabata, T.; Nagato, M.; Takasu, K.; Fuji, K. J. Am. Chem. Soc. 1997,
119, 3169-3170. (c) Sammakia, T.; Hurley, T. B. J. Org. Chem. 1999,
64, 4652-4664. (d) Fu, G. C. Acc. Chem. Res. 2000, 33, 412-420. (e)
Spivey, A. C.; Charboneau, P.; Fekner, T.; Hochmuth, D. H.; Maddaford,
A.; Madalier-Jugroot, C.; Redgrave, A.; Whitehead, M. A. J. Org. Chem.
2001, 66, 7394-7401. (f) Cupperly, D.; Gros, P.; Fort, Y. J. Org. Chem.
2002, 67, 238-241. (g) Xu, S. J.; Held, I.; Kempf, H. M.; Steglich, W.;
Zipse, H. Chem. Eur. J. 2005, 11, 4751-4757.
A DMAP unit was enlisted as a delivery vehicle because
DMAP and other 4-(dialkylamino)pyridines are frequently
employed as catalysts for the acylation of sterically hindered
alcohols and phenols.17,18 The process usually takes place via
the reaction of DMAP with the acylating agent (typically an
(19) Litvinenko, L. M.; Kirichenko, A. I. Dokl. Akad. Nauk SSSR 1967, 176,
97-100; Dokl. Chem. Engl. Transl. 1967, 763-766.
(20) Software used: Spartan, Version 4.0; Wavefunction, Inc.: Irvine, CA, 1999
(for Silicon Graphics computers).
(16) Kelly, T. R.; Cavero, M. Org. Lett. 2002, 4, 2653-2656.
9
378 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Toward a Chemically Powered Rotary Molecular Motor
A R T I C L E S
Scheme 1
Scheme 2
possible after the synthesis9,22 of prototype 1 (Scheme 1), with
the hope that the fewer deviations from the earlier route (Scheme
1), the better the chances for success. The synthesis of 7
commenced (Scheme 2) with the known preparation23 of
diaminoanthraquinone 22 from commercially available anthrone
(20), and the two amines were acetylated. The anthraquinone
unit was then converted24 to an anthracene (24), which, under
carefully controlled conditions, can be brominated to 25 in high
yield.25 The regiochemistry of the bromination is that anticipated
on the basis of the reinforcing directing effects of the two
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 379

A R T I C L E S
Kelly et al.
acetamido groups in the electrophilic aromatic substitution.
Suzuki coupling26,27 of bromide 25 with boronic acid 34
provided aldehyde 26. Boronic acid 34, isolated as the trimeric
anhydride 34a, was prepared from the known28 bromotolual-
dehyde 32 as shown in eq 1.
nature of a substituent would have such a profound impact on
the reactivity of species as inherently reactive as benzynes are.
Wittig reaction between aldehyde 30 and the ylide derived
from phosphonium salt 45 gave stilbene 31 as a 10:1 mixture
of E/Z isomers. Phosphonium salt 45 was obtained by the
sequence summarized in Scheme 3. The known naphthol 41
was prepared by the indicated route developed by Boger.31
Reaction of the naphthol 41 with triflic anhydride gave triflate
42. Stille coupling of triflate 42 with DMAP-stannane 4716
generated 43. Reduction of the ester in 43 to the benzylic alcohol
44 and reaction of the latter with 2.5 equiv of triphenylphosphine
hydrobromide afforded phosphonium salt 45 directly.
Scheme 3
Because the ensuing benzyne addition proceeds in higher
yield if the amide NH’s and the aldehyde are protected, those
groups were converted to their Boc and dioxolane derivatives,
respectively, to give 27 (Scheme 2). Reaction of 27 with
4-nitrobenzyne,29 prepared by diazotization of the corresponding
5-nitroanthranylic acid, gave an approximately 1:1 regioisomeric
mixture of the two triptycenes 28 and 29 in a combined yield
of 78%. The two regioisomers could be separated by careful
chromatography. Assignment of the regiochemistry to the two
benzyne adducts was accomplished after reduction30 of the nitro
group and acetylation of the resulting amine. Partial cleavage
of the Boc and acetal groups occurs during this sequence, and
it is completed by treatment with trifluoroacetic acid and water.
Assigning the structure of the desired regioisomer (30) at that
1
stage was then elementary because the H NMR spectrum of
30 reveals the three-fold symmetry of the triptycene unit. It
warrants mention that attempts to go directly from 26 to 30
using 4-acetamidobenzyne (36, from diazotization of 5-aceta-
midoanthranylic acid) in place of nitrobenzyne 35 failed. Even
after a sample of authentic 30 was available (via 29) to help
search for 30 in the crude 26 + 36 product, no significant
amount of 30 could be detected. It was a surprise to us that the
(21) See ref 9 and (a) Kelly, T. R.; Tellitu, I.; Sestelo, J. P. Angew. Chem., Int.
Ed. Engl. 1997, 36, 1866-1868. (b) Kelly, T. R.; Sestelo, J. P.; Tellitu, I.
J. Org. Chem. 1998, 63, 3655-3665. (c) Davis, A. P. Angew. Chem., Int.
Ed. 1998, 37, 909-910.
(22) The purpose of the methyl group is to block stilbene photocyclization from
occurring on the otherwise less sterically encumbered position para to the
triptycene.
(23) Perry, P. J.; Reszka, A. P.; Wood, A. A.; Read, M. A.; Gowan, S. A.;
Dosanjh, H. S.; Trent, J. O.; Jenkins, T. C.; Kelland, L. R.; Neidle, S. J.
Med. Chem. 1998, 41, 4873-4889.
(24) Rogers, M. E.; Averill, B. A. J. Org. Chem. 1986, 51, 3308-3314.
(25) Kelly, T. R.; Bridger, G. J.; Zhao, C. J. Am. Chem. Soc. 1990, 112, 8024-
8034.
With stilbene 31 in hand, we were optimistic that the synthesis
of the final target (7) would soon be achieved, because the three
remaining tasks (Scheme 4), stilbene photocyclization,32 hy-
droxypropyl installation, and acetamide hydrolysis, had strong
precedent in the synthesis of the prototype (Scheme 1).
Unfortunately, our hopes for a swift conclusion were soon
dashed, as we were unable to accomplish the photocyclization.
Extremely extensive efforts that examined solvent, temperature,
filters, lamps, and pH (to protonate the DMAP), all failed to
generate any detectable amount of desired product. Comparison
of the unsuccessful photocyclization of 31 to 48 with the
previously successful photocyclization of 16 to 17 implicated
(26) (a) Miyaura, N.; Yanagi, T.; Suzuki, A. Synth. Commun. 1981, 11, 513-
519. (b) Sato, M.; Miyaura, N.; Suzuki, A. Chem. Lett. 1989, 1405-1408.
(c) Sasaki, M.; Fuwa, H. Synlett 2004, 1851-1874. (d) Kotha, S.; Lahiri,
K.; Kashinath, D. Tetrahedron 2002, 58, 9633-9695.
(27) Guillier, F.; Nivoliers, F.; Godard, A.; Marsais, F.; Que´guiner, G.; Siddiqui,
M. A.; Snieckus, V. J. Org. Chem. 1995, 60, 292-296.
(28) Goodman, J. P.; St. Pyrek, J. J. Heterocycl. Chem. 1993, 30, 1645-1651.
(29) Logullo, F. M.; Seitz, A. H.; Friedman, L. Organic Syntheses; Wiley &
Sons: New York, 1973; Collect. Vol. V, pp 54-59.
(30) Makosza, M.; Winiarski, J. J. Org. Chem. 1984, 49, 1494-1499.
9
380 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Toward a Chemically Powered Rotary Molecular Motor
A R T I C L E S
Scheme 4
plings, but we hoped that recent advances in the couplings of
aryl chlorides, notably due to Fu33 and Buchwald,34 might save
the day. Chloride 56, unlike bromide 53 (which suffered rapid
decomposition upon photolysis), proved a competent substrate
for photocyclization. Consequently, chlorostilbene 62 was
prepared (eq 3) as a 10:1 E/Z mixture by a Wittig reaction
between the ylide derived from phosphonium salt 63 and
aldehyde 30.
The synthesis of 63 is summarized in Scheme 5. The
aforementioned unsaturated acid 39 was hydrogenated to 64 over
Pd/C. The latter was converted with oxalyl chloride to the
corresponding acid chloride 65, which, without isolation,
underwent AlCl3-induced intramolecular Friedel-Crafts acy-
lation to give tetralone 66. Semmler-Wolff rearrangement35 of
the derived oxime 67 then gave naphthylamine 68, which was
converted to chloride 69 through application36 of a Sandmeyer
two possible culprits: the presence of a DMAP or the two extra
acetamides on the triptycene. A control photolysis experiment
with 50, which lacks the DMAP but retains the three acetamides,
led to the formation of the desired photoproduct 51 (eq 2),
revealing that it was apparently the presence of the DMAP in
31 that was responsible for the difficulty.
Scheme 5
Protracted efforts (not detailed here) to identify a functional
group that would allow both the photocyclization and the
subsequent incorporation of the DMAP were conducted. Sub-
strates that were synthesized and examined are 52-56. Triflate
52 and bromide 53 failed in the photocylization; ethers 54 and
55 underwent the photocyclization adequately (20-25% yield)
to 57 and 58 but could not be advanced to 59 (59 appears
unexpectedly unstable; attempts to convert 58 to 59 in the
presence of a triflating agent (Tf2O) in order to trap 59 as 60
failed).
(31) Boger, D. L.; McKie, J. A.; Cai, H.; Cacciari, B.; Baraldi, P. G. J. Org.
Chem. 1996, 61, 1710-1729.
(32) For a review, see: Mallory, F. B.; Mallory, C. W. Org. React. 1984, 30,
1-456.
As a last resort, we examined the photocyclization of
chlorostilbene 56 to chlorohelicene 61. Aryl chlorides have
historically been inferior partners in palladium-catalyzed cou-
(33) Littke, A. F.; Schwarz, L.; Fu, G. C. J. Am. Chem. Soc. 2002, 124, 6343-
6348.
(34) Milne, J. E.; Buchwald, S. L. J. Am. Chem. Soc. 2004, 126, 13028-13032.
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 381

A R T I C L E S
Kelly et al.
Table 1. Yields of Coupling Reactions of Chlorides 78 and 75
reaction. A methoxymethyl (MOM)-protected version of the
hydroxypropyl side chain was installed by cleavage37 of the
methyl ether and alkylation of the resulting phenol with bromide
7038 to give 71. The ester group in 71 was reduced, and the
resulting benzylic alcohol 72 was converted to the bromide 74
by way39 of unisolated mesylate 73. Reaction of bromide 74
with triphenylphosphine then gave phosphonium salt 63.
with Pyridine Derivatives
Scheme 6
attempted optimization, the yield was only about 7%. Nonethe-
less, despite extreme limitations in material, we elected to push
forward to 7, as shown in Scheme 6, since only two deprotection
steps remained; they were in due course reduced to practice.41
Notwithstanding the limited quantities of 7 available (during
the course of the several months devoted to the following
studies, the yield42 of the Stille coupling of 75 with 47 to give
76 was raised to 16%), it was still possible to evaluate whether
7 functioned as designed. The plan (or at least the desire) was
that the sequence of events summarized in Scheme 7 would
happen. Specifically, the hope was that the DMAP unit would
capture a phosgene molecule to give 81 and pass it to the
proximal triptycylamino group, giving 82. If subsequent events
proceeded in analogy to Figure 5, then unidirectional rotation
would ensue, but in a manner that could be repeated ad
infinitum.
Addition of phosgene followed by triethylamine to a solution
of 7 in CDCl3 according to the reaction conditions that
accomplished the unidirectional rotation of the prototype (Figure
5) instead led to near-instantaneous precipitation of the vast
majority of the material. That precipitation was originally
attributed to insolubility of amine hydrochloride salts, but the
same problem persisted when tetrahydrofuran (THF) was used
as solvent. Mass spectrometry (MS) of the crude reaction
mixture revealed that it contained virtually no volatiles, even
under conditions where amine hydrochloride salts of 7, 85, 86,
etc. would be volatilized. The problem was not insolubility, but
polymerization, presumably due to intermolecular urea forma-
tion. Intermolecular urea formation did not complicate the study
of the prototype (Figure 5), although, a priori, it loomed as a
possible concern.
As expected (Scheme 6) on the basis of the successful
conversion of 56 to 61, photocyclization of 62 afforded helicene
75 (26% yield). But the continuing inferiority of aryl chlorides
(compared to aryl bromides) in some demanding palladium-
catalyzed coupling reactions was soon manifested.40 Model
studies with chloromethoxynaphthalene 78 as a surrogate for
75 were promising (Table 1) with DMAP-stannane 47 using
Fu’s33 conditions, and with commercially available 2-pyridylzinc
bromide using Buchwald’s34 protocol for the Negishi coupling.
Unfortunately, despite the very generous assistance of Professors
Fu, Buchwald, and others (see Acknowledgments), extension
of the model studies to the real system was disappointing. The
Negishi coupling of 80 with 75 gave no identifiable 76, an
outcome presaged by the difference in yields of the reactions
of 78 with 79 and 80. In contrast to the 75% yield of the reaction
of 78 and 47 with Fu’s catalyst, the yield in the real system (75
+ 47 f 76) plummeted to a single digit. After 4 months of
Dilution of the initial concentration of 7 by a factor of 100
(which should diminish the rate of a bimolecular reaction by a
factor of 10 000) and lowering the reaction temperature to -78
°C effectively suppressed polymerization (at least during the
(35) For a review, see: (a) Conky, R. T.; Ghosh, S. In Mechanisms of Molecular
Migrations; Thyagarajan, B. S., Ed.; Interscience: New York, 1971; Vol.
4, pp 197-308. See also: (b) Garst, M. E.; Cox, D. D.; Harper, R. W.;
Kemp, D. S. J. Org. Chem. 1975, 40, 1169-1170.
(41) Strictly speaking, since 7 is chiral, it would be necessary to resolve 7 to
have a system capable of truly unidirectional rotation. But since enantiomers
in an achiral environment give identical NMR, MS, and IR spectra, carrying
out the resolution would not have increased the information generated and,
therefore, would not have warranted the investment of effort required to
develop a resolution of 7.
(42) Yields of palladium-catalyzed couplings are reported to often improve
substantially when the couplings are conducted in a microwave reactor
[(a) Larhed, M.; Moberg, C.; Hallberg, A. Acc. Chem. Res. 2002, 35, 717-
727. (b) Kappe, C. O. Angew. Chem., Int. Ed. 2004, 43, 6250-6284 and
references therein]. Due to the paucity of 75, we were not able to extensively
study the effect of microwave-promoted reactions, but in our one attempt
(a CEM Discovery microwave reactor was utilized), the microwave-
promoted reaction gave results inferior to those obtained under optimized
non-microwave conditions.
(36) (a) Doyle, M. P.; Siegfried, B.; Dellaria, J. F. J. Org. Chem. 1977, 42,
2426-2431. (b) Cadogan, J. I.; Roy, D. A. J. Chem. Soc. C 1966, 14,
1249-1250.
(37) Brooks, P. R.; Wirtz, M. C.; Vetelino, M. G.; Rescek, D. M.; Woodworth,
G. F.; Morgan, B. P.; Coe, J. W. J. Org. Chem. 1999, 64, 9719-9721.
(38) Sato, T.; Kumagawa, T.; Sugimoto, A.; Yamakawa, K. Bull. Chem. Soc.
Jpn. 1987, 60, 301-310.
(39) Andrus, M. B.; Hicken, E. J.; Meredith, E. L.; Simmons, B. L.; Cannon, J.
F. Org. Lett. 2003, 5, 3859-3862.
(40) For a recent review of the Stille reaction, see: Espinet, P.; Echavarren, A.
M. Angew. Chem., Int. Ed. 2004, 43, 4704-4734.
9
382 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Toward a Chemically Powered Rotary Molecular Motor
A R T I C L E S
Scheme 7
initial reaction period). Because of the conditions of high dilution
we were forced to employ, MS was used to monitor reaction
progress. Operationally, that was accomplished by quenching
aliquots of the reaction mixture into methanol and recording
the mass spectrum of the resulting mixture. But a new problem
surfaced: The mass spectrum showed not only a peak for the
expected monomethylurethane 87, but also smaller, although
still significant peaks for bi- (88) and triurethanes (89), even
significant amounts of bi- and triurethanes when only 1 equiv
of phosgene was used strongly implied that the DMAP was not
performing its intended role of capturing and delivering ones
and only onesphosgene molecule. It appeared that the phosgene
was too reactive toward the triptycyl amines to give the DMAP
time to function as desired.
A less reactive surrogate for phosgene was sought, and 1,1′-
carbonyldiimidazole was chosen. Reaction between 7 and 1,1′-
carbonyldiimidazole only gives monoacylation (monitored as
before by methanol quench/mass spectrometry). As judged by
1H NMR, the methanol quench led to the clean formation of a
single monomethylurethane. That finding suggests that only one
of the triptycylamines is being converted to the corresponding
imidazolyl urea, but it does not establish which amine is being
acylated.
The structure of the methanol-quenched product from monoa-
cylation of 7 with 1,1′-carbonydiimidazole was assigned un-
ambiguously as 90 by careful and extensive 2-D NMR studies
on both 90 and 7. The 1H NMR chemical shift assignments for
when only 1 equiv of phosgene was used. MS is not able to
determine whether the monourethane was derived from 82 (or
83) or from “phosgenylation” of one of the other two triptycene-
based amino groups, or whether the peak represented a mixture
of all three possible monourethanes. But the formation of
the protons in 90 and 7 are given in the Supporting Information.
A long-range correlation, between Ha on the helicene and
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 383

A R T I C L E S
Kelly et al.
protons on two blades of the triptycene in both 90 and 7, allowed
for the identification of which blade of the triptycene in 90 bore
the methyl urethane. It is the one blade that does not exhibit a
long-range correlation with Ha; that one blade corresponds to
the blade proximate to the DMAP. Comparison of the chemical
shifts in 90 and 7 (see Supporting Information) indicates that it
is only in the triptycene blade proximate to the DMAP that there
1
is a significant difference in the H chemical shifts for 90 and
7. We conclude that the strategy of using the DMAP to
selectiVely deliVer the 1,1′-carbonyldiimidazole to only one of
the three aminotriptycene blades to giVe 91 has succeeded. All
that remains is for 91 to convert, a` la eq 4, to isocyanate 83 to
trigger (Scheme 7) unidirectional rotation to 86 as in the
prototype.
Figure 8. 13C NMR spectrum of the crude product of the reaction between
7 and 99% 13C-carbonyl-labeled 1,1′-carbonyldiimidazole. The peak at δ
144 Hz is from excess 13C-carbonyl-labeled 1,1′-carbonyldiimidazole. The
peak at δ 149 Hz corresponds to 13C-carbonyl-labeled 91 (see text).
carbonyldiimidazole was prepared and substituted for the 1,1′-
carbonyldiimidazole used to prepare 91.
Since the natural abundance of 13C is approximately 1%, the
carbonyl-derived 13C-labeled (from 99% 13C-phosgene) peak
for the yet-undetermined 83, 85, or 91 stands out in the 13C
NMR spectrum of the unknown (Figure 8) like a telephone pole
in a recently mown hayfield. The chemical shift of the peak (δ
149 Hz) is substantially downfield of the chemical shift range
(ca. δ 125 Hz) for N-aryl isocyanates, but well within the range
expected for the carbonyl carbon in 85 and 91.45 Heteronuclear
multiple bond correlation (HMBC) NMR spectroscopy served
to distinguish between 85 and 91 because it shows a strong
correlation between the 13C carbonyl carbon and two imidazole
hydrogens (and none to a tether methylene), thereby establishing
that the species in hand was 91.
As mentioned above, Staab’s documentation of the equilib-
rium in eq 5 had led us to anticipate a similar equilibrium
between imidazole urea 91 and isocyanate 83 plus imidazole
(eq 4). But apparently that equilibrium was not being established,
because the expected cascade of events (Scheme 7) that would
be unleashed upon the formation of 83 did not occur. Attempts
to activate the imidazole by protonation46 by addition of one or
several equivalents of methanesulfonic acid, or metal ion47
(Hg2+) coordination [several equivalents of Hg(triflate)2 were
added in case of competing coordination by other nitrogens in
91], did not provoke any reaction. Not surprisingly, attempts
to methylate48 the imidazole with methyl triflate to give 94 led
to a complex mixture of products. Efforts to promote the
conversion of 91 to 83 by using the phosphazene base known
as P1-t-Bu-tris(tetramethylene)49 and the guanidine base 2-tert-
butyl-1,1,3,3-tetramethylguanidine50 also failed. The thermal
instability of 91 severely constrained other options: a CDCl3
Imidazolyl ureas of primary anilines are reported by Staab43
to be in equilibrium with the corresponding isocyanates (eq 5).
If (i) that equilibrium occurs here and (ii) the 7-based system
behaves in analogy to the prototype (Figure 5), then the sequence
of events in Scheme 7 should unfold. But it does not. In the
prototype (Figure 5), so-called prebarrier urethane 4 forms so
fast upon addition of phosgene and triethylamine to 1 that the
1
formation of isocyanate 2 cannot be detected by H NMR.
Rather, 4 is the first detectable intermediate, and it is the rotation
of 4 to 5 that is the rate-determining step. But in the present
case of 7, MS of the methanol-quenched reaction mixture shows
no significant peak for either 85 or 86 (being isomers, 85 and
86 would give molecular ions of the same mass).
In seeking to troubleshoot the situation, it was essential to
identify the entity actually present in solution after the DMAP-
achieved delivery of the 1,1′-carbonyldiimidazole to the adjacent
aminotriptycene blade. MS indicated that, prior to the methanol
quench (to give 90), the molecule had a mass of 760, consistent
with isocyanate 83, the isomeric intramolecular urethane 85 (or
86), or possibly 91, which might fragment to 83 in the mass
spectrometer. While appropriate infrared spectroscopy measure-
ments might have provided the answer, we enlisted 13C NMR
labeling studies. 13C-Labeling of the carbonyl carbon of 1,1′-
carbonyldiimidazole has been reported.44 Using commercially
available 99% 13C-labeled phosgene, 13C-carbonyl labeled 1,1′-
(45) Evident from examination of 13C NMR spectra of N-aryl isocyanates,
urethanes, and ureas in The Aldrich Library of 13C & 1H FT NMR Spectra,
1st ed.; Pouchet, C. J., Behnke, J., Eds.; Aldrich Chemical Co., Inc.:
Milwaukee, WI, 1993; Vols. 1-3.
(46) (a) Staab, H. A.; Wendel, K.; Datta, A. P. Justus Liebigs Ann. Chem. 1966,
694, 78-85. (b) Oakenfull, D. G.; Jencks, W. P. J. Am. Chem. Soc. 1971,
93, 178-188. (c) Oakenfull, D. G.; Salvesen, K.; Jencks, W. P. J. Am.
Chem. Soc. 1971, 93, 188-194.
(47) Of common metal ions, mercuric ion has the strongest affinity for imidazole
[(a) Brooks, P.; Davidson, N. J. Am. Chem. Soc. 1960, 82, 2118-2123.
(b) Smith, R. M.; Martell, A. E. Critical Stability Constants; Plenum
Press: New York and London, 1975; Vol. II, pp 144-145, 208]. Ley has
used zinc-ion coordination to activate N-acylimidazoles to hydrolysis (Ford,
M. J.; Ley, S. V. Synlett 1990, 255-256), but at somewhat elevated
temperatures.
(48) Ulibarri, U.; Choret, N.; Bigg, D. C. H. Synthesis 1996, 1286-1288.
(49) Schwesinger, R.; Hasenfratz, C.; Schlemper, H.; Walz, L.; Peters, E. M.;
Peters, K.; von Schnering, H. G. Angew. Chem., Int. Ed. Engl. 1993, 32,
1361-1363. Reagent obtained from Fluka.
(50) Barton, D. H. R.; Chen, M.; Jaszberenyi, J. C.; Taylor, D. K. Org. Synth.
1997, 74, 101-107. Reagent obtained from Fluka.
(43) Staab, H. A. Angew. Chem., Int. Ed. Engl. 1962, 1, 351-367.
(44) Nelson, V. C. J. Labelled Compd. Radiopharm. 1996, 38, 713-723.
9
384 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007

Toward a Chemically Powered Rotary Molecular Motor
A R T I C L E S
solution of 91 is stable for ca. 3 days at -25 °C but decomposes
overnight at room temperature. Attempts to prepare 94 directly
by use of Rapoport’s51 1,1′-carbonylbis(3-methylimidazolium)
ditriflate (95) were thwarted because the very reactive 95
behaves toward 7 like the unselective phosgene rather than like
1,1′-carbonyldiimidazole itself. Oxalyldiimidazole (96) was also
examined as a possible fuel but failed (as judged by MS, giving
monoacylation but no further reaction).
and stirring the crude residue first for 4 h with a mixture of
water and CH2Cl2 (which, as evidenced by MS, cleaved the
imidazolylurea in 98 to give 99) and then with aqueous THF
(which hydrolyzed the sole trifluoroacetyl ester in 99) gave
monoamine 97.
With 97 in hand, we now had a species where we could, as
with the prototype, use phosgene/triethylamine as fuel. The
system thus differs from the prototype by only a single
variable: the precise structure of the motor molecule. If our
failure to achieve rotation of 7 with 1,1′-carbonyldiimidazole
and oxalyldiimidazole was a consequence of a difference in the
fuel, then reaction of 97 with phosgene/triethylamine should
result in unidirectional rotation (via 100 and 101). On the other
hand, if the problem is with the precise structure of the motor
molecule, then use of the same fuel (phosgene/triethylamine)
as works with the prototype will still not produce rotation.
Two explanations for our failure to implement Scheme 7
presented themselves. One possibility was that, despite the
seemingly apt precedent of Figure 5 for Scheme 7, there was
some unrecognized design flaw that doomed Scheme 7 to
failure. The other possibility was that our problems were due
to limitations inherent in the fuels examined.
To resolve the question, we sought to re-examine phosgene,
the fuel that was successful in powering the prototype (Figure
5), under circumstances where lack of selectivity for monoa-
cylation would not be a problem. The hope was to use a
molecule such as 97, where the two anilines we wished to keep
uninvolved are removed from the action by acylation. It might
be regarded as far-fetched to imagine that 7 could be selectively
converted to 97. But in 91, the one aniline we do not seek to
protect is already masked. As events unfolded, exploratory
experiments involving reaction of in situ-generated 91 with
ostensibly 2 equiv of trifluoroacetic anhydride (TFAA, as a
solution in CDCl3) gave, as established by mass spectrometric
monitoring of the reaction, a mixture of mono-, bis-, and
tristrifluoroacetyl derivatives of 91. The formation of a tristri-
fluoroacetate, presumably 98, was an unexpected complication,
but perhaps a result of inaccurate measurements of the amount
of 7 or TFAA resulting from conducting the exploratory,
moisture-sensitive experiment on submilligram amounts of 91
due to the paucity of 7. Fortunately, a simple solution presented
itself: use excess TFAA to convert all of 91 to 98.
With the opportunity for achieving a clear-cut answer finally
secured, 97 was treated with phosgene/triethylamine exactly as
was done with the prototype. Quenching of an aliquot of the
reaction mixture into methanol and mass spectrometric exami-
nation showed a strong molecular ion peak for monomethy-
lurethane 102, indicating that the reaction of 97 with phosgene
to give 103 (and/or 104) was definitely successful. But the mass
spectrum showed no peak attributable to intramolecular urethane
100 or 101. (The mass spectrum of the reaction mixture prior
to a methanol quench also showed no peak for 100 or 101,
excluding the possibility that the desired intramolecular ure-
thanes were formed initially but then were being cleaved to 97
or 105 in the methanol quench.) In short, despite the seemingly
compelling precedent provided by the prototype system (Figure
5), the completely developed versions do not rotate.
Trifluoroacetate esters are exceptionally labile to hydrolysis.52
Evaporation of volatiles from 98 (to remove residual TFAA
that might trifluoroacetylate the unmasked aniline NH2 in 99)
(51) Saha, A. K.; Schultz, P.; Rapoport, H. J. Am. Chem. Soc. 1989, 111, 4856-
4859. This paper reports a 1H NMR spectrum for 95 of δ 8.86 (s, 2H),
7.44 (m, 2H), 7.16 (m, 2H), 4.00 (s, 6H), but the solvent is not explicitly
indicated. The general section says that NMR spectra were recorded in
CDCl3 unless otherwise indicated. In our hands, 95 is not sufficiently soluble
in CDCl3 to permit recording of a 1H NMR spectrum. The 1H NMR
spectrum for 95 in CD3CN shows δ 9.18 (s, 2H), 7.98 (m, 2H), 7.72 (m,
2H), 4.00 (s, 6H), in agreement with a spectrum of 95 in CD3CN kindly
provided by Dr. S. Sabesan (see: Sabesan, S. Tetrahedron Lett. 1997, 38,
3127-3130).
(52) Cramer, F.; Ba¨r, H. P.; Rhaese, H. J.; Sa¨nger, W.; Scheit, K. H.; Schneider,
G.; Tennigkeit, J. Tetrahedron Lett. 1963, 16, 1039-1042.
9
J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007 385

A R T I C L E S
Kelly et al.
achievements to report: (i) success in the construction of 7, a
challenging synthetic target in and of itself, and (ii) the
demonstration that the DMAP-based delivery strategy works.
We also believe that a relatively easy solution exists for
transforming 7 from a near miss to a fully functional, repeatedly
rotating, chemically powered molecular motor. But a confluence
of circumstances, including the principal investigator’s (PI’s)
decision 4 years ago not to submit new grant applications, the
ending of currently funded grants, and the consequent planned
gradual winding down of his research program, render it unlikely
that further work on this program will be accomplished at
Boston College. It has been an exciting time and, as indicated
in the lists of coauthors (present and past) and the Acknowledg-
ments, the PI is grateful to many for helping make it happen.
Acknowledgment. We are grateful to the National Institutes
of Health (grant no. GM56262) for support of this work. We
thank Professors Greg Fu33 (MIT), Steve Buchwald34 (MIT),
John Verkade57 (Iowa State), and Antonio Echavarren40 (Institut
Catala` d’Investigacio´ Qu´ımica, Tarragona, Spain) for exchange
of information about palladium-catalyzed couplings, Professors
Sally Mallory32 (U. Penn) and Frank Mallory32 (Bryn Mawr)
for discussions of stilbene photocyclizations, and Dr. Subra-
maniam Sabesan51 (Dupont) for information regarding 95. We
also thank Dr. Eric L. Elliott for countless mass spectra, and
Justin Pine, Meaghan O’Malley, Agata Sajkiewicz, Jack Beierle,
Matt Demers, and Agnes Bak for assistance with the preparation
of starting materials. Finally, we acknowledge numerous Boston
College colleagues (faculty, postdocs, students and staff) for
helpful suggestions and assistance, and Dr. John Schwab
(NIGMS) for his interest.
We are surprised that the fully elaborated systems 7 and 9753
are not successful. It is true that models are only models, but
Figure 5 appears to be a compelling model, especially given
that the DMAP unit functions as designed. We suggest two
explanations for why 7 and 97 do not behave as motors. Perhaps
in 7 and 97, the hydroxypropyl group adopts a conformation
different from that in the prototype, possibly due to hydrogen-
bonding interactions with the DMAP or the added substituents
on the other two triptycene blades. Alternatively, rotation around
the triptycene/helicene bond in 7 and 97 may be so constrained
[perhaps by hydrogen bonding or a Bu¨rgi-Dunitz (or similar)
interaction54 as in 106] that access to a rotationally excited
rotamer (compare 3 in Figure 5) that brings the hydroxyl group
on the propyl side chain within reach of the isocyanate is not
possible.55,56
Supporting Information Available: General experimental
procedures, details of preparations/characterizations of all
compounds, and copies of 1H and 13C NMR spectra. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA066044A
(56) We were unable to establish exactly what 97 is converted to (“X”) upon
addition of phosgene and triethylamine. Possibilities for X include 103,
104, 106, i, and ii. Use of 99% 13C-labeled phosgene in place of normal
phosgene and 13C NMR spectroscopy might have distinguished among those
possibilities. But X, whatever it is, completely decomposes in less than 30
min at room temperature (see Experimental Section for preparation of 102).
Given the very limited quantities of 97 available, it would not have been
possible to record a 13C NMR spectrum of 13C-labeled X before it
decomposed.
While we would, of course, have preferred to see Scheme 7
unfold as planned, we still believe we have two significant
(53) Had 97 rotated, cleavage of the two relatively labile trifluoroacetamidess
which can be achieved with methanol containing triethylaminesand
urethane cleavage as with 101 f 105 would also have given a repeatedly
rotating molecular motor system, although a somewhat more convoluted
one.
(54) Bu¨rgi, H. B.; Dunitz, J. D. Acc. Chem. Res. 1983, 16, 153-161.
(55) It has been suggested that use of excess phosgene may disrupt the Bu¨rgi-
Dunitz interaction, but excess phosgene does not change the outcome. MS
indicated that the alcohol is not phosgenylated by excess phosgene under
the conditions examined.
(57) Su, W.; Urgaonkar, S.; McLaughlin, P. A.; Verkade, J. G. J. Am. Chem.
Soc. 2004, 126, 16433-16439.
9
386 J. AM. CHEM. SOC. VOL. 129, NO. 2, 2007