Quantitative analyses of mixtures of 2-deuterio-1-vinylcyclobutanes

Article abstract of DOI:10.1021/jo901663v

(Chemical Equation Presented) Making distinctions between two stereoisomers characterized by diastereotopic deuterium atoms can ordinarily be achieved using standard NMR spectroscopic methods. Mixtures of stereoisomers having both diastereotopic and enant

Full text of DOI:10.1021/jo901663v

pubs.acs.org/joc
Quantitative Analyses of Mixtures of 2-Deuterio-1-vinylcyclobutanes
John E. Baldwin,*,† David J. Kiemle,‡ and Alexey P. Kostikov†
‡
†Department of Chemistry, Syracuse University, Syracuse, New York 13244, and Department of Chemistry,
State University of New York College of Environmental Science and Forestry, Syracuse, New York 13210
jbaldwin@syr.edu
Received August 1, 2009
Making distinctions between two stereoisomers characterized by diastereotopic deuterium atoms can
ordinarily be achieved using standard NMR spectroscopic methods. Mixtures of stereoisomers
having both diastereotopic and enantiotopic deuterium labels, however, may be difficult to analyze
quantitatively. The present work introduces a simple way to gain quantitative analyses of mixtures of
the four stereoisomeric 2-deuterio-1-vinylcyclobutanes, an essential prerequisite to establishing the
stereochemical characteristics of the thermal stereomutations of vinylcyclobutane and its structural
isomerizations to cyclohexene. The unconventional NMR method introduced and validated in this
work will likely prove convenient and generally applicable to related stereochemical investigations.
Introduction
labeled with larger substituents such as methyl and cyano,
structures more susceptible to analyses through chiral capil-
lary GC or other well established methods.1 The deuterium-
labeled vinylcyclobutanes t-(1S,2S)-1, c-(1S,2R)-1, t-(1R,2R)-1,
and c-(1R,2S)-1 could lead to uncovering the stereochemical
and kinetic characteristics of the thermal reactions of the parent
hydrocarbon, vinylcyclobutane, and thereby provide the best
comparisons with theory-based efforts to probe the mechanistic
subtleties at play. Thorough stereochemical and kinetic work on
the parent hydrocarbon, labeled only with deuterium substi-
tuents, would facilitate attaining these objectives.
Projected kinetic studies based on stereoisomers of deu-
terium-labeled vinylcyclobutanes 1 prompted a search for
adequate quantitative analytical methods. An applicable
analytical capacity would need to track the four stereo-
isomers during gas-phase thermal reactions as they evolved
from a single structure toward equilibrium among all four.
The thermal stereomutations could be followed quantita-
tively even as each individual vinylcyclobutane isomer reac-
ted as well through competitive fragmentations and struc-
tural isomerizations.
One anticipates the involvement of short-lived diradical
intermediates of considerable conformational flexibility and
dynamic complexities separating each specific chiral isomer
of vinylcyclobutane and its formation of three isomeric
vinylcyclobutanes and a mixture of seven d2-labeled cyclo-
hexenes. The time-dependent mixtures of vinylcyclobutanes
and cyclohexenes would define the stereochemical landscape
and reflect the dynamic alternatives.
Theory-based attempts to model the thermal chemistry of
vinylcyclobutanes and the anticipated fragmentations,
stereomutations, and [1,3] sigmatropic carbon shifts might
well determine relevant potential energy conformational
contours and enable calculations of reaction dynamics over
the transition region. Ambitious initial efforts toward these
objectives have been sketched, but they have not yet been
well advanced.2 An experimentally grounded description
of the mechanistic paths involved could complement and
Similar thermal reactions have been followed in stereoche-
mical and kinetic detail for vinylcyclobutanes stereochemically
(1) (a) Baldwin, J. E.; Burrell, R. C. J. Am. Chem. Soc. 2001, 123, 6718–
6719. (b) Baldwin, J. E.; Burrell, R. C. J. Am. Chem. Soc. 2003, 125, 15869–
15877. (c) Baldwin, J. E.; Burrell, R. C. J. Phys. Chem. A 2003, 107, 10069–
ꢀ
10073. (d) Baldwin, J. E.; Fede, J.-M. J. Am. Chem. Soc. 2006, 128,
5608–5609. (e) Doering, W. von E.; Cheng, X.; Lee, K.; Lin, Z. J. Am. Chem.
Soc. 2002, 124, 11642–11652. (f) Cheng, X. Ph.D. Dissertation, Harvard
University, 1989; Diss. Abstr. Int. B 1990, 50, 3472.
(2) Northrop, B. H.; Houk, K. N. J. Org. Chem. 2006, 71, 3–13.
7866 J. Org. Chem. 2009, 74, 7866–7872
Published on Web 09/16/2009
DOI: 10.1021/jo901663v
r
2009 American Chemical Society

Baldwin et al.
JOCArticle
illuminate theoretically derived understandings of the com-
plex reactions involved, but adequate experimental data are
simply lacking.
have identical 13C NMR spectral absorptions. Mixtures of the
two could not be analyzed quantitatively by standard NMR
methods; mixtures in any proportions would appear to be
identical.
A novel method to ascertain relative concentrations of
each of the four 3,4-d2-cyclohexenes in mixtures including
the three 3,6-d2-cyclohexenes has been found and validated,
thus opening a path toward uncovering the stereochemical
and dynamic characteristics of thermal reactions of vinyl-
cyclobutane.3 The second analytical requisite, a suitable
method to determine quantitatively the time-dependent rela-
tive concentrations of the four isomers of 1, remained an
elusive pursuit.
A simple conversion of d0-1 to cyclobutanecarboxyalde-
hyde through an oxidative step and further through a con-
densation with (R)-R-methylbenzylamine would make the
former pro-R and pro-S carbons in pro-R-13C-d0-2 and pro-
S-13C-d0-2 structures diastereomeric. The carbon-13 labels
would experience different regions of the chirotopical struc-
tural space and could be distinguished through 13C NMR
spectroscopy. The diastereomers could in principle register a
significant chemical shift difference, Δδ, thanks to the chiro-
topical influence of the imine function.
A variety of imaginable techniques, ranging from vibra-
tional circular dichroism infrared spectroscopy to estab-
lished NMR spectroscopic techniques applied to enantio-
meric structures4,5 involving chiral lanthanide shift reagents,
chiral solvents, and chiral structural modifications prior to
NMR analyses, as well as more sophisticated NMR appro-
aches exploiting chiral liquid crystal solvents, were consid-
ered. Most were attempted, and each proved unsuccessful.
All of these options work well in many instances but not
always. Our attempts to apply chiral liquid crystal solvents
were particularly frustrating, insofar as confidence in the
method and appreciation for its impressive capabilities
dependent on chemical shift anisotropies, dipolar couplings,
and quadrupolar splittings6-8 were not matched in our
laboratory with sufficiently dexterous skills. The technical
necessities required for preparing the lyotropic system poly-
γ-benzyl-L-glutamate of the right average molecular weight
in CH2Cl2 or another solvent, repeated centrifugations,
∼20 times, of samples in sealed dual-head tubes in both
directions to achieve complete dissolution of polymer and
dispersion of the sample throughout the gel, setting suitable
transversal relaxation times and an optimal spinning rate, or
none at all, and other experimental prerequisites were simply
not realized.
This phenomenon is well-known: linking a suitable chir-
otopical structure with a possible mixture of two enantio-
mers can often reflect the enantiomeric excess of the sample.9
The imines d0-2 were prepared and analyzed in CDCl3 at
75.5 MHz; the 13C NMR spectrum confirmed a substantial
Δδ value separating pro-R and pro-S atom absorptions
(68 ppb). Other NMR spectra of d0-2 solutions recorded
using 500 and 600 MHz proton frequency spectrometers and
at different concentrations gave similar but not identical
Δδ values. The Δδ distinction thus demonstrated would
contribute as well for d-labeled structural analogues such
as the isomers of 2.
Disappointing attempts to apply reported analytical
methods were followed in due course by imagining, imple-
menting, and validating a simple technique for gaining
quantitative analyses of mixtures of the isomers of 1.
Results and Discussion
The new approach was attained through a structural
modulation of the vinylcyclobutane system to provide a
predictable basis for the required analyses. Unlabeled vinyl-
cyclobutane (d0-1) is prochiral, and the carbons C(2)
and C(4) are distinct: the two structural alternatives, pro-
R-13C-d0-1 and pro-S-13C-d0-1, would be enantiomers and yet
The distinctive CHD carbons in each of the four stereo-
isomers of 2 would reflect chemical shift contributions from
the chirotopical influence of the imine function and also from
stereochemically distinctive deuterium perturbations of car-
bon-13 δ values. Thus, all four isomers of 2 might be
analyzed quantitatively by 13C{1H,2H} NMR spectroscopy
quite easily.
(3) Baldwin, J. E.; Kiemle, D. J.; Kostikov, A. P. J. Org. Chem. 2009, 74,
3866–3874.
(4) Parker, D. Chem. Rev. 1991, 91, 1441–1457.
(5) Wenzel, T. J.; Wilcox, J. D. Chirality 2003, 15, 256–270.
(6) (a) Sarfati, M.; Courtieu, J.; Lesot, P. Chem. Commun. 2000, 1113–
1114. (b) Sarfati, M.; Lesot, P.; Merlet, D.; Courtieu, J. Chem. Commun.
2000, 2069–2081. (c) Courtieu, J.; Lesot, P. ; Meddour, A.; Merlet, D.; Aroulanda,
C. In Encyclopedia of Nuclear Magnentic Resonance; Grant, D. M., Harris,
R. K., Eds.; Wiley: Chichester, 2002; Vol. 9, pp 497-505. (d) O'Hagan, D.; Goss,
R. J. M.; Meddour, A.; Courtieu, J. J. Am. Chem. Soc. 2003, 125, 379–387.
(e) Lesot, P.; Sarfati, M.; Courtieu, J. Chem. Eur. J. 2003, 9, 1724–1745.
(f) Lafon, O.; Berdague, P.; Lesot, P. Phys. Chem. Chem. Phys. 2004, 6, 1080–
1084.
Whether pro-R-13C-d0-2 or pro-S-13C-d0-2 would have the
more downfield carbon-13 absorption could not be pre-
dicted, but the chirotopical influence contributing to all
ꢀ
(7) Gouriou, L.; Lloyd-Jones, G. C.; Vyskocil, S.; Kocovsky, P.
J. Organomet. Chem. 2003, 687, 526–537.
(8) Potrzebowski, M. J.; Jeziorna, A.; Kazmierski, S. Concepts Magn.
Reson. 2008, 32A, 201–218.
(9) Inter alia: (a) Weix, D. W.; Dreher, S. D.; Katz, T. J. J. J. Am. Chem.
ꢀ
Soc. 2000, 122, 10027–10032. (b) Perez-Fuertes, Y.; Kelly, A. M.; Fossey,
J. S.; Powell, M. E.; Bull, S. D.; James, T. D. Nature Protocols 2008, 3, 210–214.
J. Org. Chem. Vol. 74, No. 20, 2009 7867

Baldwin et al.
JOCArticle
SCHEME 1. Preparation of Imines 2a
aReagents and conditions: (a) (1R)-(-)-menthol, DCC, DMAP,
CH2Cl2; (b) SOCl2; (c) BrCCl3, DMAP, 2-mercaptopyridine 1-oxide
sodium salt; (d) Bu3SnD, InCl3, BEt3, THF; 55% over three steps;
(e) LiAlH4; (f) PCC, CH2Cl2; (g) (R)-(þ)-R-methylbenzylamine.
FIGURE 1. 13C{1H,2H} absorptions of a mixture of isomers 2 in
the CHD region.
described by Gunther,10 Berger,11 and others.12 For “frozen”
€
13C absorptions for isomers of 2 could be counted on to be
consistent. Both this phenomenon and stereochemical-de-
pendent deuterium perturbations of 13C chemical shifts
could be anticipated. Experimental evidence, then, could
lead to secure assignments of 13C absorptions to specific
isomers of 2.
A sample of a mixture of all four d-labeled isomers of 2 was
made so that the hypothetical path to a successful stereo-
chemical analysis might be surveyed. The synthetic steps
leading to the desired uneven mixture of the four isomeric
structures of 2 are outlined in Scheme 1.
The formation of the monoester from the trans diacid
(a) took place with a very modest diastereoselective outcome.
The three-step reaction sequence (b-d) afforded mixtures
favoring trans-bromo and trans-deuterio substituents at
C(2), for in each case a cyclobutyl radical was involved as
an intermediate and reacted preferentially to give trans
diastereomers from atom-donor coreactants, BrCCl3 leading
to the bromides and Bu3SnD to 2-d-cyclobutanecarboxylic
esters.
chair-conformation cyclohexane, for instance, ax D and
eq D substituents lead to upfield 13C 1Δ chemical shift
perturbations (over one bond) of 444.9 and 398.4 ppb,
respectively,10 a significant difference of 46.5 ppb. A quali-
tatively similar distinction would be expected for the isomers
of 2. A structure having a pseudoaxial D would be expected
to have a larger upfield perturbation than a pseudoequator-
ial D isomer.
All CHD carbons in isomers of 2 were well upfield from
absorptions associated with CH2 methylene carbons; the
observed 3Δ deuterium perturbations of chemical shifts were
appreciably smaller, as expected.10 The downfield peaks are
3
contributed by 13C(4)H2 nuclei experiencing Δ perturba-
tions of chemical shifts provided by deuterium substituents
at C(2)HD as well as differential shielding from chirotopical
influences of the imine function. For 4-d-tert-butylcyclohex-
anes the axial and equatorial deuterium substituents con-
tribute upfield shifts of δ by 14.7 and 37.7 ppb, respectively, a
difference of 23 ppb. In isomers characterized by a pseudoe-
3
quatorial versus a pseudoaxial C(2)HD deuterium the Δ
The mixture of menthyl esters was converted to the imines
2 as outlined (e-g) without stereochemical modifications.
When the mixture of isomers 2 was studied by 13C{1H,2H}
NMR spectroscopy with broadband decoupling of both
proton and deuterium spins,3 a pattern of four absorptions
for CHD carbons was observed, and the peaks were easily
assigned, provisionally. The spectral features of importance
portrayed in the upfield CHD four-peak pattern observed
(Figure 1) gave relative intensities left to right of 37, 19, 33,
and 11%. The absorptions in the four-line pattern at δ 25.232
and 25.180, the first and third, together accounting for 70%
of relative intensity, were tentatively assigned to trans iso-
mers. The less intense absorptions at δ 25.209 and 25.157
were ascribed to the cis isomers. If any one peak could be
surely assigned to a specific stereoisomer of 2, all of the
assignments would follow. That one initial assignment could
not be made unambiguously without further information.
The chemical shifts of CHD methylene carbons flanking
the imine function in isomers of 2 would be influenced by
deuterium perturbations of carbon-13 δ values, a stereoche-
mically very informative and long-established effect well
difference at 13C(4)H2 would be expected to have similar net
upfield δ shifts. The first and second and the third and fourth
absorptions in Figure 2 are separated by 24 ppb. The
spacings between absorptions in Figure 2 for 13C(2)HD
peaks match those shown in Figure 1, though the figures
record slightly different chemical shift reference settings.
To outline the logic of the assignments, one might assume
temporarily that the most intense absorption at δ 25.232 seen
in Figure 1 may be assigned to isomer t-(1S,2S)-2. Then the
third absorption would correspond to t-(1R,2R)-2. The
differences between the two trans and the two cis isomers
(10) (a) Aydin, R.; G€unther, H. Angew. Chem. 1981, 93, 1000-1002;
Angew. Chem., Int. Ed. Engl. 1981, 20, 985-986. (b) Aydin, R.; Wesener,
J. R.; G€unther, H.; Santillan, R. L.; Garibay, M.-E.; Joseph-Nathan, P. J. Org.
Chem. 1984, 49, 3845–3847.
(11) (a) Berger, S. In Encyclopedia of Nuclear Magnetic Resonance; Grant,
D. M., Harris, R. K., Eds.; Wiley: Chichester, 1996; Vol. 2, pp 1168-1172.
(b) Novak, P., Drazenvikic-Topic, Smrecki, V.; Meic, Z. In New Advances in
Analytical Chemistry; Atta-ur-Rahman, Ed.; Harwood Academic Publishers:
Singapore, 2000; pp 135-168.
(12) O’Leary, D. J.; Allis, D. G.; Hudson, B. S.; James, S.; Morgera,
K. B.; Baldwin, J. E. J. Am. Chem. Soc. 2008, 130, 13659–13663.
7868 J. Org. Chem. Vol. 74, No. 20, 2009

Baldwin et al.
JOCArticle
c-(1S,2R)-2, shifted upfield by the trans to cis change (an
“eq D” to an “ax D” substituent change) while the relatively
deshielded position of the pro-S-13C(2)HD group remained
constant. The last absorption, the weakest and the most
upfield of the four, would be assigned to c-(1R,2S)-2; it
would be shifted upfield by an “ax D” and by being located
at the position most shielded preferentially by the chiroto-
pical imine function, as in pro-R-13C-d0-2.
Thus, the proof of principle of this approach to gaining
quantitative analytical assessments of mixtures of the set of
four deuterium-labeled stereoisomers 1 and of 2 was secured,
but the specific structural assignments were based only on a
tentative assumption. A rigorous assignment of absolute
stereochemistry linking specific isomers of 2 with specific
13C absorptions required additional experimental evidence.
Two reference compounds at hand13 were used to prepare
mixtures of different ratios of isomers of 2. A racemic sample
of benzyl c-2-d-cyclobutanecarboxylate (c-(1S,2R)-3 and
c-(1R,2S)-3) was converted through a LAH reduction, a
PCC oxidation, and a condensation with (R)-R-methylben-
zylamine to give a mixture of c-(1S,2R)-2 and c-(1R,2S)-2.
The precursor racemic ester was examined by 2H NMR and
found to be a mixture of 94.2% cis (δ = 2.329)/5.8% trans
(δ = 2.222) isomers.
FIGURE 2. 13C{1H,2H} absorptions observed for a mixture of the
four isomers of 2 in the upfield C(2)HD and the downfield C(4)H2
regions.
The sample of the imine isomers of 2 derived from the
racemic cis ester exhibited the expected strong 13C{1H,2H}
NMR absorptions (Figure 3) for the second and fourth peaks
found in Figure 1 and in the upfield pattern in Figure 2. The
Δδ between the two cis peaks was 51 ppb; the δ values were
25.232, 25.208, 25.180, and 25.157 ppm. The relative peak
areas were 7, 41, 7, and 45%.
c-(1R,2S)-2-Benzyloxycarbonylcyclobutanecarboxylic acid
of high ee was secured expeditiously following Bolm and co-
workers through a highly asymmetric addition of benzyl
alcohol to cis-1,2-cyclobutanedicarboxylic anhydride, facili-
tated by quinidine at -55 °C .14 The (1R,2S) acid ester was
2
employed to prepare t-(1S,2S)-3;13 according to H NMR
spectroscopy, the t-(1S,2S)-3 made through a sequence of
steps was 5.3% cis (δ = 2.334) and 94.7% trans (δ = 2.225).
FIGURE 3. Pattern for 13C{1H,2H} NMR absorptions for C(2)HD
carbons for all four isomers of 2 prepared from a racemic mixture of
c-(1S,2R)-3 and c-(1R,2S)-3; the isomers of 2 from left to right are
t-(1S,2S)-2, c-(1S,2R)-2, t-(1R,2R)-2, and c-(1R,2S)-2.
in Figure 1 were identical, 52 ppb, a consequence of the
difference in chemical shift influences contributed by the
chirotopical influence of the imine substituent. The shifts
between the first two absorptions at δ 25.232 and 25.209 and
the pair at 25.180 and 25.157 were both 23 ppb, favoring the
upfield chemical shift perturbation by the pseudoaxial sub-
stituent, as in the cyclohexane prototype. Thus, the second
absorption, at 25.209 ppm, might be tentatively assigned to
The imine t-(1S,2S)-2 was prepared from the stereoche-
mically related ester t-(1S,2S)-3, and it together with other
(13) Baldwin, J. E.; Kostikov, A. P. Synthetic work in progress.
(14) (a) Bolm, C.; Schiffers, I.; Dinter, C. L.; Gerlach, A. J. Org. Chem.
2000, 65, 6984–6991. (b) Bolm, C.; Schiffers, I.; Atodiresei, I.; Hackenberger,
C. P. R. Tetrahedron: Asymmetry 2003, 14, 3455–3467. (c) Bolm, C.;
Atodiresei, L.; Schiffers, I. Org. Synth. 2005, 82, 120–125.
J. Org. Chem. Vol. 74, No. 20, 2009 7869

Baldwin et al.
JOCArticle
TABLE 1. Upfield Perturbations of 13C(2) Chemical Shifts for the
Four Stereoisomers of 2-d-1-Cyclobutanecarboxyaldehyde
(R)-r-Methylbenzylimines 2
13C(2)-D Eq 13C(2)-D Ax pro-S-13
C
pro-R-13C sum (rel
isomer
(rel ppb)
(rel ppb)
(rel ppb) (rel ppb) ppb)a
t-(1S,2S)-2
c-(1S,2R)-2
t-(1R,2R)-2
c-(1R,2S)-2
0
0
0
0
23
52
75
23
23
0
52
52
aThe relative chemical shifts for these absorptions in Figures 1, 2, and
3, left to right, are 0, 23 or 24, 52, and 75 ppb.
using a tabular scheme. The chemical shift perturbations
associated with the two options available for each influence
are shown in Table 1. The two structural aspects each provid-
ing two possible upfield chemical shift perturbations for
13C(2)HD in stereoisomers of 2 combine to determine the four
distinct chemical shifts exhibited in 13C{1H,2H} spectra.
The contributions of these chemical shift perturbations
recorded by 13C{1H,2H} NMR spectroscopy are not strictly
comparable to those affording stereochemically informative
spectra for mixtures of isomeric 3,4-d2-cyclohexenes in the
presence of 3,6-d2-cyclohexenes.3 In both cases, structural
adjustments were employed to convert hydrocarbons poten-
tially formed through thermal reactions of deuterium-la-
beled vinylcyclobutanes to deuterium-labeled cycloalkanes
retaining all the relevant stereochemical information en-
coded in thermal reaction products, and the cycloalkane
derivatives in each case were conformationally restricted so
that stereochemical options at CHD elements within a ring
could be readily identified, yet the chemical shift perturba-
tions in the two instances depended on different structural
influences.
In the 3,4-d2-cyclohexenes and the derivatives employed to
gain access to the stereochemical information required, the
chemical shift perturbations were provided by 1Δ CHD and
2Δ CHD contributions combining in four possible ways,
characteristic of the four possible stereochemical options.
Since the information could be read by scanning four
13C absorptions well separated from absorptions associated
with isomers derived from 3,6-d2-cyclohexene precursors,
they could not conflagrate the requisite data deleteriously.
In analyses of mixtures of isomeric 2-d-1-cyclobutanecar-
boxyaldehyde (R)-R-methylbenzylimines 2 derived from
deuterium-labeled vinylcyclobutanes, the 13C{1H,2H} NMR
analytical method depended on two types of chemical shift
perturbations, each with two options, thus leading to four
combinations each characteristic of a specific stereoisomer of
2. The pseudoaxial versus pseudoequatorial D substituents at
C(2)HD provided one alternative; the choice between pro-R-
versus pro-S-13C(2) atoms provided the second.
FIGURE 4. 13C{1H,2H} absorptions of a mixture of the four iso-
mers of 2 rich in t-(1S,2S)-2, in the upfield CHD and the downfield
CH2 regions.
isomers of 2 were identified by 13C{1H,2H} NMR (Figure 4).
The dominant peak, peak 1 from left to right in the upfield
pattern, confirmed the t-(1S,2S)-2 absolute stereochemistry
of its structure.
The 13C{1H,2H} NMR spectrum obtained for t-(1S,2S)-2
together with the other three isomers identified it as the first
absorption from the left of the four-peak upfield 13C(2)HD
pattern. Reductions of absolute stereochemical control dur-
ing the preparation of t-(1S,2S)-2 from t-(1S,2S)-3 were
anticipated to be modest, though some could well have
intervened. But the absolute stereochemical finding was
unambiguous. With the absolute stereochemistry assign-
ment for peak 1 secured, all other assignments followed.
The second peak was upfield by 23 ppb, a switch from an
“eq D” to an “ax D” substituent, so the absorption is for
c-(1S,2R)-2. The third and fourth peaks are further upfield,
for the chirotopical influence of the imine function comes
into play as an additional factor; the third and fourth
absorptions are each upfield from the first and second peaks
respectively by 52 ( 1 ppb. The chirotopical influence favors
stronger shielding for pro-R-13C(2)HD than for pro-S-13C-
(2)HD. Isomer c-(1R,2S)-2 has the most upfield 13C{1H,2H}
absorption, for it has an “ax D” substituent in the magnetic
environment which provides the best shielding by the chiral
imine function, at pro-R-13C(2)HD. The four absorptions,
from left to right, are for t-(1S,2S)-2, c-(1S,2R)-2, t-(1R,2R)-
2, and c-(1R,2S)-2, as most clearly shown in Figure 1.
The downfield pattern in Figure 2, left to right, may be
assigned to c-(1R,2S)-2, t-(1R,2R)-2, c-(1S,2R)-2, and
t-(1S,2S)-2. Peak 4 is t-(1S,2S)-2, obviously, from the domi-
nant absorption characteristic of the authentic sample
(Figure 4). Peak 3 is downfield from peak 4 by 24 ppb,
corresponding to the change of stereochemistry at C(2), from
a pseudoequatorial D to a pseudoaxial D for the cis isomer
c-(1S,2R)-2. Peak 1 (c-(1R,2S)-2) is similarly downfield from
peak 2 by 24 ppb (Figure 2). The gaps between the two cis and
the two trans isomers (52 and 53 ppb, Figure 2) reflect the
relatively deshielded position of pro-S-C(4)H2 in peaks 1 and
2 versus pro-R-C(4)H2 dispositions in peaks 3 and 4. The
Conclusions
This exploitation of 13C{1H,2H} NMR spectroscopy will
make possible realistic attempts to characterize the kinetics
and stereochemistry of thermal stereomutations of d-labeled
vinylcyclobutanes 1, the instigating goal leading to this cap-
ability, but it is likely to be of greater importance elsewhere.
For hydrocarbons, at least, the method will almost surely
prove more practically useful than analyses based on chiral
lanthanide chemical shift reagents, though these shift reagents
are sometimes serviceable in other structural environments.
3
same Δ perturbation difference separating peak 2 from 1,
3
and peak 4 from 3, follows from the larger upfield Δ δ
perturbation afforded by a pseudoequatorial D substituent.
Another way to present the chemical shift perturbation
contributions by pseudoaxial versus pseudoequatorial D
substituents at C(2)HD, and the relative perturbations associ-
ated with pro-R- versus pro-S-13C atoms, can be summarized
7870 J. Org. Chem. Vol. 74, No. 20, 2009

Baldwin et al.
JOCArticle
The specific chiral perturbing substituent introduced to
modify vinylcyclobutane worked well, but obviously a great
variety of other chiral structural modifications could be con-
sidered and utilized and possibly proved to be better. The
method being introduced is not dependent on modulating a
structure through introducing a specific chiral functional
group. It depends on interrogating the stereochemical char-
acter of mixtures of deuterium-labeled compounds by intro-
ducing and exploiting a chirotopical influence on NMR
shielding characteristics while taking advantage simulta-
neously of stereochemically sensitive deuterium perturbations
on 13C chemical shifts. The absorptions registered by 13C-
{1H,2H} NMR spectroscopy provided useful analytical data
quantifying relative concentrations of the stereoisomers of 1
and of 2 with considerable ease. Other applications of this
novel method will surely be exercised and found to be valuable.
of the four stereoisomeric menthyl 2-bromocyclobutanecarbox-
ylates (4.11 g, 13 mmol) in 65% in overall yield. 1H NMR
(CDCl3): δ 0.78 (m, 3H), 0.93 (m, 6H), 0.9-2.6 (m, 13H), 3.43
(m, 1H), 4.62 (m, 1H), 4.72 (m, 1H). 13C NMR: complicated
spectrum for the mixture of isomers.
Tributyltin deuteride (4.4 g, 15 mmol) was added dropwise
to the solution of menthyl 2-bromocyclobutanecarboxylates
(4.11 g, 13 mmol) and indium(III) chloride (155 mg, 0.7 mmol)
in THF (125 mL). The reaction mixture was stirred for 3 h at rt,
after which time no conversion of the starting material was
observed. Triethylborane (0.5 mmol solution in THF) was then
added, and the reaction mixture was stirred overnight at rt. The
following morning, the mixture was heated to reflux for 2 h, all
volatiles were removed under reduced pressure, and the residue
was redissolved in ether (100 mL). The ethereal solution was
washed with 1 M HCl (50 mL) and then with brine (50 mL), dried
over Na2SO4, filtered, and concentrated in vacuo. Purification on
silica gel afforded the four stereoisomeric (R)-menthyl 2-deuterio-
cyclobutanecarboxylates in 92% yield (2.87 g, 12 mmol). 1H NMR
(CDCl3): δ 0.78 (d, J = 6.9 Hz, 3H), 0.91 (m, 6H), 2.3-0.9
(m, 13H), 3.12 (q, 1H), 4.69 (td, 1H). 13C NMR: δ 16.7, 18.6, 21.1,
22.4, 23.9, 26.7, 31.8, 34.7, 38.7, 41.3, 47.5, 74.1, 175.5.
To a suspension of LiAlH4 (27 mg, 0.7 mmol) in dry ether
(5 mL) was added menthyl 2-deuteriocyclobutylcarboxylates
(120 mg, 0.5 mmol), and the reaction mixture was stirred for 1 h
at rt. It was then carefully quenched and washed with 2 M HCl
(2 ꢀ 3 mL), and the combined aqueous layer was extracted with
ether (2 ꢀ 2 mL). The combined organic phase was washed with
saturated sodium bicarbonate (3 mL), dried over MgSO4, and
filtered; the filtrate was evaporated under atmospheric pressure
at ∼40-50 °C.
Experimental Section
All 13C{1H,2H} spectra were acquired at 30 °C with a 600 MHz
1H (150.9 MHz 13C frequency) NMR spectrometer, acquired by
1
2
decoupling both H and H simultaneously using inverse-gated
WALTZ16 decoupling sequences to obtain completely decoupled
13C absorptions.3
Condensation of Cyclobutanecarboxyaldehyde with (R)-(þ)-r-
Methylbenzylamine. Cyclobutylmethanol (86 mg, 1 mmol) was
added to a suspension of PCC (237 mg, 1.1 mmol) and Celite
(350 mg) in CH2Cl2 (3 mL). The resulting black mixture was
stirred for 4 h and then passed through silica gel packed in a 5-cm
Pasteur pipet and eluted by 3 mL of CH2Cl2. (R)-(þ)-R-
Methylbenzylamine (121 mg, 1 mmol) was added to the eluate
followed by about 100 mg of MgSO4. The reaction mixture was
allowed to stand for 1 h and then filtered, concentrated in vacuo,
and redissolved in CDCl3. The imine product was analyzed
without further purification, since weak signals characteristic of
unreacted amine did not interfere with the regions of interest. 1H
NMR (CDCl3): δ 1.50 (d, 3H), 2.25-1.80 (m, 6H), 3.15 (quintet
of d, 1H), 4.30 (q, 1H), 7.34 (m, 5H), 7.80 (d, 1H). 13C NMR: δ
19.1, 24.9, 25.615, 25.683, 40.2, 69.6, 125.9, 126.9, 128.6, 145.4,
166.5. The CH2 absorptions for pro-R-13C and pro-S-13C atoms
in the d0-2 imine at δ 25.615 and 25.683 were confirmed through
DEFT spectrum editing, though they could not be assigned.
Isomeric t-(1S,2S)-2, c-(1S,2R)-2, t-(1R,2R)-2, and c-(1R,2S)-
2 (R)-(þ)-R-Methylbenzylimines. To a solution of racemic trans-
cyclobutane-1,2-dicarboxylic acid (2.88 g, 20 mmol), DMAP
(∼ 50 mg), and (-)-menthol (3.18 g, 20.4 mmol) in CH2Cl2
(100 mL) was added DCC (4.54 g, 22 mmol) at 0 °C. The reac-
tion mixture was stirred for 3 h and then filtered. The filtrate was
evaporated in vacuo; the residue was redissolved in ether and
filtered through a thin layer of silica gel. Ether was removed
in vacuo to provide diastereomers of crude mono-menthyl
trans-cyclobutane-1,2-dicarboxylic acids contaminated with a
small amount of dimenthyl esters.
Thionyl chloride (3.7 mL, 50 mmol) was added to the crude
monomenthyl esters (5.64 g, 20 mmol), and the reaction mixture
was heated at reflux for 3 h. Excess SOCl2 was removed under
reduced pressure; the resulting acyl chloride product was used
without further purification. A 100-mL round-bottomed flask
was charged with mercaptopyridine N-oxide sodium salt (3 g,
20 mmol), DMAP (∼50 mg), and bromotrichloromethane (50 mL).
The resulting suspension was heated at reflux for 2 h, and then the
acyl chloride from the SOCl2 reaction was added dropwise. The
reaction mixture was heated at reflux overnight, cooled to rt, and
concentrated under reduced pressure. The residue was redissolved
in ether (100 mL); the ethereal solution was washed with 1 M HCl
(2 ꢀ 50 mL), dried over Na2SO4, filtered, and concentrated in
vacuo. Purification of the residue on silica gel afforded a mixture
The residue was added to a suspension of PCC (237 mg,
1.1 mmol) and Celite (350 mg) in CH2Cl2 (3 mL). The resulting
black mixture was stirred for 4 h and then passed through silica
gel and condensed with (R)-(þ)-R-methylbenzylamine (121 mg,
1 mmol) as detailed above. The reaction mixture, containing
imines derived from the four stereoisomeric 2-d-cyclobutane-
carboxaldehydes was allowed to stand for 1 h and then filtered,
concentrated in vacuo, and redissolved in CDCl3. The mixture
of imines was analyzed by 13C{1H,2H} NMR for absorptions
characteristic of the chiral amine-derived imine; the unreacted
amine did not interfere with the NMR chemical shift region of
interest. The isomers t-(1S,2S)-2, c-(1S,2R)-2, t-(1R,2R)-2, and
c-(1R,2S)-2 were present in the relative concentrations 37, 19,
33, and 11% (Figure 1). See also Figure 2.
cis-2-d-Cyclobutanecarboxylates c-(1S,2R)-3 and c-(1R,2S)-
3, and trans-2-d-Cyclobutanecarboxylate t-(1S,2S)-3 were avail-
able from another synthetic project.13 The 1H, 2H, and 13C
NMR spectra for the racemic benzyl cis-2-d-cyclobutanecar-
boxylates c-(1S,2R)-3 and c-(1R,2S)-3 and for benzyl trans-
(1S,2S)-2-d-cyclobutanecarboxylate t-(1S,2S)-3 are included
in the Supporting Information. The unlabeled benzyl ester of
cyclobutanecarboxylic acid is a known and fully characterized
compound.15 The racemic sample of c-(1S,2R)-3 and c-(1R,2S)-
3 was converted following the reactions used above to prepare a
mixture of t-(1S,2S)-2, c-(1S,2R)-2, t-(1R,2R)-2, and c-(1R,2S)-
2 in the relative concentrations of 7, 41, 7, and 45%. The 13C
{1H,2H} spectrum of these absorptions is shown in Figure 3. The
sample of t-(1S,2S)-3 was converted following the reactions
used above to prepare a mixture of t-(1S,2S)-2, c-(1S,2R)-2,
t-(1R,2R)-2, and c-(1R,2S)-2 rich in t-(1S,2S)-2. The most
downfield absorption of the C(2)HD pattern was dominant
(15) (a) Bender, D. M.; Peterson, J. A.; McCarthy, J. R.; Gunaydin, H.;
Takano, Y.; Houk, K. N. Org. Lett. 2008, 10, 509–511. (b) Hale, J. J.; Lynch,
C. L.; Caldwell, C. G.; Willoughby, C. A.; Kim, D.; Shen, D.-M.; Mills, S. G.;
Chapman, K. T.; Chen, L.; Gentry, A.; Maccoss, M.; Konteatis, Z. D. US
2002094989 A1, 103 pp.
J. Org. Chem. Vol. 74, No. 20, 2009 7871

Baldwin et al.
JOCArticle
(Figure 4). Another mixture of the isomers of 2 rich in t-(1S,2S)-
2 is shown in Figure S5 in the Supporting Information.
Supporting Information Available: NMR spectral data
for pro-R-13C-d0-2 and pro-S-13C-d0-2, 2-d-labeled stereo-
isomers of benzyl cyclobutanecarboxylate ((c-(1S,2R)-3 and
c-(1R,2S)-3, and t-(1S,2S)-3), and a 54, 8, 12, and 26% mixture
of t-(1S,2S)-2, c-(1S,2R)-2, t-(1R,2R)-2, and c-(1R,2S)-2. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Acknowledgment. We thank the National Science Foun-
dation (CHE-0240104 and CHE-0514376) for financial
ꢀ
support of this work and Dr. Jean-Marie Fede for valuable
contributions to the preliminary stages of this study.
7872 J. Org. Chem. Vol. 74, No. 20, 2009