Preparation and characterization of bicyclic amide acetals and monothioacetals

Article abstract of DOI:10.1139/v03-204

We have prepared and characterized new examples of 5-substituted-1-aza-4,6- dioxabicyclo[3.3.0]octanes (bicyclic amide acetals) and examples of the new heterocyclic system, 5-substituted-1-aza-4-oxa-6-thiabicyclo[3.3.0]octanes (bicyclic amide monothioacetals). Detailed analysis of NMR coupling constants and X-ray structure determination for an example of each class of compound established the stereochemistry and conformation of these ring systems.

Full text of DOI:10.1139/v03-204

268
Preparation and characterization of bicyclic amide
acetals and monothioacetals1
J. Peter Guthrie, Roger T. Gallant, and Michael C. Jennings
Abstract: We have prepared and characterized new examples of 5-substituted-1-aza-4,6-dioxabicyclo[3.3.0]octanes
(bicyclic amide acetals) and examples of the new heterocyclic system, 5-substituted-1-aza-4-oxa-6-thiabicyclo[3.3.0]oc-
tanes (bicyclic amide monothioacetals). Detailed analysis of NMR coupling constants and X-ray structure determination
for an example of each class of compound established the stereochemistry and conformation of these ring systems.
Key words: bicyclic amide acetals, preparation, coupling constants, X-ray structure.
Résumé : On a préparé et caractérisé de nouveaux exemples de 1-aza-4,6-dioxabicyclo[3.3.0]octanes substitués en posi-
tion 5 (amide acétals bicycliques) et des exemples du nouveau système hétérocyclique 1-aza-4-oxa-6-thiabicyclo[3.3.0]oc-
tanes substitués en position 5 (amide monothioacétals bicycliques). Une étude détaillée des constantes de couplage de
la RMN et la détermination de la structure, par diffraction des rayons X, d’un exemple de chacune de ces classes de
composés permet d’établir la stéréochimie et la conformation de ces composés cycliques.
Mots clés : amide acétals bicycliques, préparation, constantes de couplage, structure par diffraction des rayons X.
[Traduit par la Rédaction] Guthrie et al. 278
Introduction
[2]
We have been interested in determining the energies of
tetrahedral intermediates in acyl transfer reactions by an in-
direct thermochemical approach (1–7). Attempts to extend
this approach from simple amides (2) to amides with various
substituents on the α-carbon ran into frustrating problems of
purification. Amide acetals could be made but could not be
purified. In an attempt to overcome this problem we have
turned to bicyclic versions of these compounds. Our plan is
to determine the heats of hydrolysis and, thus, the heats of
formation of the bicyclic acetals and monothioacetals and
proceed from these to, first, the heats of formation of the
acyclic acetals and monothioacetals, using computational
methods to determine the heats of reaction for the hypotheti-
cal hydrogenations shown in reactions [1] and [2], and
thence, to the free energies of formation of the analogous
tetrahedral intermediates.
By methods that we have reported (8) and that involve a
hypothetical hydrolysis, free energies of the acyclic amide
acetals can be used to calculate the free energies of the cor-
responding amide hemiacetals and amide hydrates.
[3]
[1]
Similarly, the free energies of the acyclic amide
monothioacetals can be used to calculate the free energies of
the corresponding amide hemithioacetals.
[4]
Received 9 July 2003. Published on the NRC Research Press
Web site at http://canjchem.nrc.ca on 4 February 2004.
J.P. Guthrie,2 R.T. Gallant, and M.C. Jennings.
Department of Chemistry, University of Western Ontario,
London, ON N6A 5B7, Canada.
The energies of these adducts can be combined with free
energies of the amides to permit calculation of the free energy
changes for addition reactions of amides, reactions that are
quite unfavorable and difficult to determine experimentally.
1This article is part of a Special Issue dedicated to Professor
Ed Piers.
2Corresponding author (e-mail: peter.guthrie@uwo.ca).
Can. J. Chem. 82: 268–278 (2004)
doi: 10.1139/V03-204
© 2004 NRC Canada

Guthrie et al.
269
Scheme 1.
[5]
X = OH, OCH3, SCH3.
The first stage in this project was to prepare and charac-
terize the bicyclic compounds, only a few of which were
known. Calorimetric studies of these compounds are com-
plete and computational studies to complement and in some
cases to supplement them are nearly complete. The rest of
our studies on these compounds will be reported in due
course.
best samples were only 94% and 90% pure, with the
remainder being the corresponding oxazoline. Details of
these preparations and of the preparations of other com-
pounds used in this work are found in the experimental sec-
tion.
Results
The compounds examined in this work are shown in
Scheme 1. The original plan had been to prepare compounds
with a range of electron-withdrawing substituents, but diffi-
culties were encountered in some of these attempted prepa-
rations, and then it became clear that the rate of hydrolysis
was very sensitive to electron withdrawal, such that strong
electron-withdrawing groups made hydrolysis too slow for
the calorimetric measurements we planned to carry out.
Although some of the bicyclic amide acetals are known
compounds, most are novel, and preparations range from
very easy to quite difficult.
The standard methods for preparation of compounds 1,
substituted 1-aza-4,6-dioxabicyclo[3.3.0]octanes, involve ei-
ther reaction of diethanolamine with the corresponding
nitrile, using a small amount of sodium as catalyst, or reac-
tion of diethanolamine with the corresponding acyclic amide
acetal. Similar methods allowed preparation of the novel
compounds 2, substituted 1-aza-4-oxa-6-thiabicyclo[3.3.0]oc-
tanes, by using 2-(2-mercaptoethyl)aminoethanol. The com-
pounds 2 were thermally sensitive and could not be obtained
by the standard reaction of a nitrile with 2-(2-mercapto-
ethyl)aminoethanol at 80 °C; such conditions led only to the
oxazoline corresponding to the nitrile. Reaction at room
temperature, though slow, led to the desired compounds. At-
tempted distillation of the crude 5-substituted-1-aza-4-oxa-6-
thiabicyclo[3.3.0]octane at temperatures over 110 °C also
led to breakdown to the corresponding oxazoline. A reason-
able explanation is relatively facile cleavage of ethylene sul-
fide, e.g., by
The oxazolidinooxazolidine system in the bicyclic amide
acetals has two identical five-membered rings — each with
four distinctive proton environments — leading to an ABXY
pattern, with complicated splitting. The assignment of the
1
methylene signals was based on the H, 13C, gCOSY, and
gHSQC spectra. The proton NMR spectra of the 5-R-1-aza-
4,6-dioxabicyclo[3.3.0]octanes are resolved well enough to
allow a direct graphical calculation of the true chemical
shifts and coupling constants. These calculated values were
verified with the use of ACD Labs H NMR spin simulation
software (9), as shown in Fig. 1. The worst deviation in
chemical shift between the simulated and experimental spec-
tra was approximately 0.002 parts per million (ppm). The
chemical shifts and coupling constants resulting from this
analysis are summarized in Tables 1 and 2. Analysis of the
coupling constants by means of the Karplus–Conroy equa-
tion (10–12), using generic values of J° = 8.5 Hz and J180
=
9.5 Hz (as recommended when no better estimates are avail-
able (13)) led to the conclusion that the geometry of the ring
systems for 5-(4-nitrophenyl)- 1-aza-4,6-dioxabicyclo[3.3.0]oc-
tane is as shown in
The logic of this assignment is as follows: (i) If the AB
and XY protons were eclipsed (whether A with X and B
3
3
with Y or A with Y and B with X), then JAX + JAY
=
3JBX + JBY, but this is not the case; (ii) If the protons are
staggered, then one of A and B will be anti to one of X and
Y and gauche to the other, and the other of A and B will be
gauche to both X and Y. Since J is largest for protons that
are anti, the largest of the J coupling constants must be for
the pair of protons that are most nearly anti. This is JAX
3
The 5-substituted-1-aza-4-oxa-6-thiabicyclo[3.3.0]octanes
were higher boiling than the corresponding oxazolines and
so purification by fractional distillation was possible if the
temperature could be kept low enough to avoid thermal frag-
mentation. For the aryl-substituted compounds this was not
possible, and pure product could be obtained only when the
5-substituted-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane could
be crystallized, as for 2 with R = phenyl and 4-nitrophenyl.
In the case of 2 with R = 4-chlorophenyl and 4-
methoxyphenyl, crystallization was not achieved, and the
3
3
;
3
3
(iii) If A is approximately anti to X, then JAY ≈ JBX, as is
found; if instead A and Y were approximately anti, then
3
3JAX ≈ JBY, which is not the case.
An X-ray structure was obtained for 5-(4-nitrophenyl)-1-
© 2004 NRC Canada

270
Fig. 1. Experimental and simulated H NMR spectra for 5-isopropyl-1-aza-4,6-dioxabicyclo[3.3.0]octane.
Can. J. Chem. Vol. 82, 2004
1
Table 2. Coupling constants for the 5-R-1-aza-4,6-
dioxabicyclo[3.3.0]octanes (Hz).
Table 1. Chemical shifts for the 5-R-1-aza-4,6-dioxabicyclo-
[3.3.0]octanes in CDCl3 (ppm) (Varian Mercury 400).
R
2JAB
3JAX
3JAY
3JBX
3JBY
2JXY
R
A
B
X
Y
H
Me
Et
i-Pr
–8.20
–8.20
–8.10
–8.30
–8.00
8.40
8.35
8.25
8.70
7.85
6.04
6.32
6.28
6.28
6.24
6.83
6.67
6.77
6.63
6.50
4.30
4.30
4.10
3.55
4.55
–10.90
–10.95
–10.90
–11.00
–10.70
H
Me
Et
i-Pr
3.838
3.865
3.904
3.895
3.949
3.796
3.790
3.813
3.816
3.870
3.199
3.210
3.219
3.219
3.275
2.924
2.909
2.947
2.959
2.966
MeOCH2
MeOCH2
Ph
p-NO2-Ph
–8.15
–8.35
8.00
8.30
6.56
6.40
6.90
6.63
4.35
4.25
–10.80
–11.00
Ph
p-NO2-Ph
4.119
4.115
3.998
3.992
3.358
3.353
3.094
3.127
p-Cl-Ph
CF3
–8.20
–7.95
8.15
7.13
6.56
6.60
6.87
6.50
4.30
5.85
–11.05
–10.88
p-Cl-Ph
CF3
4.094
4.093
3.968
4.012
3.324
3.339
3.079
3.028
Cl2CH
–8.10
7.65
6.54
6.80
4.73
–10.88
Cl2CH
4.095
4.055
3.468
3.062
aza-4,6-dioxabicyclo[3.3.0]octane. This structure, Fig. 2,
confirms that the bicyclic amide acetals have the five-
membered rings cis-fused, as Anteunis et al. had concluded
(14) based on analysis of the H NMR spectra. The dihedral
angles were close to, though not identical with, those de-
duced from the Karplus–Conroy analysis; the worst devia-
tion was for the A–X dihedral, 146° rather than 162°. Exam-
ination of bond lengths in the X-ray structure showed no
sign of an anomeric effect, which would be expected to
cause a shortening of the quaternary carbon–nitrogen bond
and a lengthening of at least one of the quaternary carbon–
oxygen bonds.
1
© 2004 NRC Canada

Guthrie et al.
271
Table 3. MS peak listing of the 5-R-1-aza-4,6-dioxabicyclo[3.3.0]octanes.
R
[M]+
[M – R]+
[M – CH2O]+
[RCO]+
[C3H4NO]+
[C3H6N]+
85(100)
99(100)
113(41)
127(l)
H
Me
Et
i-Pr
115(20)
129(12)
143(4)
157(5)
158(l)
114(8)
—
—
57(52)
71(26)
—
70(3)
56(25)
56(34)
56(20)
56(17)
56(7)
114(12)
114(100)
114(16)
114(100)
70(12)
70(34)
70(8)
MeOCH2
129(2)
70(30)
CF3
183(9)
197(l)
114(100)
114(100)
153(49)
167(4)
97(3)
—
70(43)
70(29)
56(60)
—
C12CH
Ph
p-NO2-Ph
191(12)
236(8)
114(10)
114(8)
161(61)
206(100)
105(100)
150(67)
70(6)
70(6)
56(19)
—
p-Cl-Ph
225(8)
114(3)
195(37)
139(78)
70(3)
—
Note: Values in table are m/z (relative intensity (%)).
Fig. 2. Crystal structure of 5-(4-nitrophenyl)-1-aza-4,6-
dioxabicyclo[3.3.0]octane.
In the alkyl-substituted bicyclic amide acetals, the substi-
tuent elimination route appears to become more prevalent as
R becomes larger. When the substituent was an aryl group,
cleavage of the ring was favored, yielding dominant [M –
CH2O]+ and acyl cation fragments.
The thiazolidinooxazolidine ring system has a more com-
plicated 1H NMR splitting pattern, with two overlapping
ABXY patterns. The pattern for the thiazolidine ring will be
labeled A′B′X′Y′; the primes are used to indicate the analogy
with the unprimed peak in the oxazolidine ring. This is not
the usual convention, where single primes are used to indi-
cate protons that are chemically but not magnetically equiva-
lent, which is not the case here. All the protons are different,
and the pattern, though complicated, can be analyzed as in
the oxazolidinooxazolidine case. The assignment of the
1
methylene signals was based on the H, 13C, gCOSY, and
gHSQC spectra. The 13C signal for the methylene at lowest
field was assigned to CH2-O; the 13C signal for the methy-
lene at highest field was assigned to CH2S. The combination
of gCOSY and gHSQC spectra allowed the remaining as-
signments. Simulated spectra were in good agreement with
the observed spectra, as shown in Fig. 3. The chemical shifts
and coupling constants are found in Tables 4 and 5.
For most of the 5-substituted-1-aza-4-oxa-6-thiabicyclo-
[3.3.0]octanes, the same logic as for the dioxa case can be
used to deduce the dihedral angles for the oxazolidino ring
from the Karplus–Conroy curve, but for three compounds
(R = H, ethyl, and methoxymethyl), all four 3J values are ap-
3
proximately equal, which leads to a contradiction. If JAX
+
3JAY ≈ JBX + JBY, then one might conclude that the hydro-
gens were eclipsed, but if so the dihedral angles would be 0°
and 120°, leading to distinctly different coupling constants.
We are forced to conclude that in fact the structure is partly
staggered, and that the Karplus–Conroy curve, especially
with generic parameter values, is only an approximate guide.
The structures are, in all cases, likely to be staggered but
closer to eclipsed than to fully staggered geometry. The ap-
proximate nature of the dihedral angles deduced from the
Karplus–Conroy curve is shown by comparison with the
X-ray structure for R = 4-nitrophenyl, where the AY and BX
dihedrals are found to be 27.45° and 26.62°, respectively,
and not 15° and 15°, as deduced from the coupling con-
stants. Likewise the AX dihedral is 144.5°, according to the
X-ray analysis, rather than 157°, as deduced from coupling
constants. The Karplus–Conroy curve is a useful guide but is
not exact. The coupling constants for the thiazolidino ring
3
3
The mass spectra of these compounds were analyzed for
characteristic fragmentation patterns. There were differences
in fragmentation pattern for 5-aryl- vs. 5-alkyl-substituted
1-aza-4,6-dioxabicyclo[3.3.0]octanes. These results are col-
lected in Table 3. The molecular ion peak for each bicyclic
amide acetal was present with low relative intensity. Two
primary fragmentation routes appear to be important based
on the electron impact studies of these compounds, as fol-
lows: cleavage of the ring and elimination of the substituent.
© 2004 NRC Canada

272
Can. J. Chem. Vol. 82, 2004
1
Fig. 3. Experimental and simulated H NMR spectra for 5-(4-nitrophenyl)-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane.
were not all the same, and the logic used for the dioxa case
led to a simpler analysis. Analysis of the coupling constants
for 6-(4-nitrophenyl)-1-aza-6-thia-4-oxabicyclo[3.3.0]octane
by the Karplus–Conroy equation leads to the geometries shown:
gles are in approximate, though not exact, agreement with
the values from the Karplus–Conroy analysis.
The X-ray structure revealed an N-C-S anomeric effect;
the quaternary carbon–nitrogen bond is shorter relative to
the CH2—N bonds (1.43 Å vs. 1.46 Å or 1.50 Å), and the
quaternary carbon–sulfur bond is longer than the CH2—S
bond (1.93 Å vs. 1.80 Å). The X-ray crystal structure data
are found in Tables S-1 to S-10.3
The mass spectra of the 5-substituted 1-aza-6-thia-4-
oxabicyclo[3.3.0]octanes showed more complex fragmenta-
tion patterns, as one might expect for the less-symmetrical
structures. Once again there were differences in preferred
fragmentation pattern for 5-aryl- vs. 5-alkyl-substituted com-
pounds. These results are found in Table 6. The molecular
ion is present in greater amount in the sulfur-containing
bicyclic acetals. As for the dioxa series, elimination of the
An X-ray structure was determined for 6-(4-nitrophenyl)-
1-aza-6-thia-4-oxabicyclo[3.3.0]octane; see Fig. 4. Again,
the five-membered rings are cis-fused, and the dihedral an-
3 Supplementary data may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Council
Canada, Ottawa, ON K1A 0S2, Canada (http://www.nrc.ca/cisti/irm/unpub_e.shtml for information on ordering electronically). CCDC
214681 and 214682 contain the supplementary data for this paper. These data can be obtained, free of charge, via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge, U.K.; fax
+44 1223 336033; or deposit@ccdc.cam.ac.uk).
© 2004 NRC Canada

Guthrie et al.
273
Table 4. Chemical shifts (ppm) for 5-R-1-aza-4-oxa-6-thiabicyclo[3.3.0]octanes in CDCl3 (ppm) (Varian Inova 600).
R
A
B
X
Y
A′
B′
X′
Y′
H
Me
Et
i-Pr
MeOCH2
Ph
p-NO2-Ph
p-Cl-Ph
p-MeO-Ph
4.104
4.078
4.053
4.026
4.026
4.225
4.217
4.207
4.213
3.947
3.967
3.918
3.884
3.930
4.029
4.001
4.005
4.026
3.176
3.230
3.200
3.206
3.186
3.353
3.344
3.327
3.332
2.798
2.813
2.816
2.832
2.777
3.030
3.058
3.018
3.009
3.122
3.302
3.233
3.183
3.137
3.346
3.344
3.335
3.345
2.830
2.921
2.824
2.802
2.745
3.070
3.099
3.069
3.057
3.500
3.533
3.509
3.505
3.466
3.589
3.623
3.567
3.552
3.054
3.068
3.013
3.026
3.002
3.211
3.232
3.172
3.158
Table 5. Coupling constants for 5-R-1-aza-4-oxa-6-thiabicyclo [3.3.0] octanes (Hz).
R
2JAB
3JAX
3JAY
3JBX
3JBY
2JXY
2JA′B′
3JA′X′
3JA′Y′
3JB′X′
3JB′Y′
2JX′Y′
H
Me
Et
i-Pr
–7.80
–8.03
–7.13
–7.65
–7.73
7.58
8.60
7.43
7.35
7.35
6.45
7.74
7.56
7.50
6.60
6.69
7.55
7.15
7.25
7.00
7.65
5.70
7.50
5.57
7.32
–9.23
–9.15
–8.33
–9.45
–9.23
–10.20
–10.50
–10.20
–10.43
–10.43
4.89
5.57
5.76
5.23
5.40
10.20
10.95
10.95
10.88
10.88
2.35
1.20
0.00
1.20
1.28
5.60
6.15
5.80
5.55
6.00
–12.23
–12.90
–12.60
–12.60
–12.83
MeOCH2
Ph
p-NO2-Ph
–7.73
–7.73
7.58
7.73
7.74
7.68
7.60
7.65
5.31
5.14
–9.68
–9.98
–10.13
–9.98
5.49
5.52
9.98
9.68
2.50
3.00
5.90
5.80
–12.30
–12.45
p-Cl-Ph
p-MeO-Ph
–7.80
–7.80
7.50
7.50
7.68
7.80
7.40
7.40
5.06
5.23
–9.90
–9.60
–10.05
–9.48
5.49
5.49
9.75
9.87
2.50
2.45
6.00
5.95
–12.15
–12.83
substituent R on C5 becomes more important as R becomes
larger, and ring cleavage is favored when R is an aryl group.
Fig. 4. Crystal structure of 5-(4-nitrophenyl)-1-aza-4-oxa-6-
thiabicyclo[3.3.0]octane.
Discussion
As we had hoped, the bicyclic amide acetals and mono-
thioacetals were much easier to purify than the acyclic am-
ide dimethyl acetals and the amide dialkyl monothioacetals,
although the bicyclic compounds are still rather sensitive. A
range of these compounds has now been prepared, though it
proved difficult to prepare even the bicyclic acetals of
amides with electron-withdrawing substituents. Although no
kinetics were carried out, it is clear that the rate of hydroly-
sis of these compounds is very sensitive to the nature of the
substituent on the 5-carbon, as would be expected, given that
this carbon becomes the carbon of an iminium ion during the
hydrolysis. It is interesting that the hydrolysis is facile in
basic solution, while the corresponding cations (from
protonation or alkylation on nitrogen) can be studied in solu-
tion (14) or even recrystallized (15). We presume that this is
a consequence of the facile reclosure of the cation, resulting
from C–N cleavage and the more effective competition by
hydroxide to trap the cation from C–O cleavage.
[6]
[7]
© 2004 NRC Canada

274
Can. J. Chem. Vol. 82, 2004
Table 6. MS peak listing of the 5-R-1-aza-4-oxa-6-thiabicyclo[3.3.0]octanes.
R
[M]+
[M – R]+
[M – CH2S]+
[M – CH3S]+
[M – C2H3S]+
85(70)
99(10)
113(12)
127(4)
84(21)
98(3)
112(23)
126(22)
72(68)
86(31)
100(46)
114(87)
H
Me
Et
131(66)
145(41)
173(49)
173(100)
130(4)
130(6)
130(19)
130(40)
i-Pr
MeOCH2
175(68)
130(82)
129(l)
128(7)
116(20)
Ph
p-NO2-Ph
207(13)
252(21)
130(l)
130(2)
161(3)
206(4)
160(21)
205(20)
148(10)
193(15)
p-Cl-Ph
p-MeO-Ph
241(18)
237(2)
130(l)
130(l)
195(2)
—
194(12)
190(l)
182(9)
178(l)
R
[M – C2H4S]+
[M – C2H3SO]+
[M – C3H6SO]+
[RCO]+
[C3H4NO]+
[C3H6N]+
71(21)
85(13)
99(14)
113(13)
115(3)
56(100)
70(11)
84(41)
98(42)
100(64)
—
—
—
57(85)
71(48)
73(2)
70(15)
70(11)
70(19)
70(29)
70(1)
H
Me
Et
i-Pr
56(100)
56(100)
56(100)
56(92)
55(36)
69(21)
83(10)
85(9)
MeOCH2
56(30)
Ph
p-NO2-Ph
147(12)
192(8)
132(l)
177(2)
117(32)
162(15)
105(100)
150(100)
70(2)
70(3)
—
—
p-Cl-Ph
p-MeO-Ph
181(20)
177(94)
166(l)
162(1)
151(45)
147(100)
139(100)
135(35)
70(3)
70(1)
—
—
We have demonstrated by NMR and X-ray structure deter-
mination that both the 5-substituted 1-aza-4,6-dioxabi-
cyclo[3.3.0]octanes and the 5-substituted 1-aza-6-thia-4-
oxabicyclo[3.3.0]octanes are cis fused, with the ring methy-
lene groups approximately staggered. Examination of the
structure of 5-(4-nitrophenyl)-1-aza-4,6-dioxabicyclo[3.3.0]-
octane showed no sign of changes in bond lengths attribut-
able to an anomeric effect. All four C—O bonds were essen-
tially the same length, 1.41, 1.41, 1.41, and 1.40 Å, as were
all three C—N bonds, 1.47, 1.46, and 1.46 Å. Thus, the
bonds that cannot be part of an anomeric effect are the same
length as those that, in principle, might be. Consideration of
the expected position of the nitrogen lone pair showed that
this was not lined up to overlap with the σ* orbital of the
C5—O bond. Since the anomeric effect is attributed to this
kind of overlap, the absence of bond length changes be-
comes reasonable. By contrast, the structure of 5-(4-nitro-
phenyl)-substituted 1-aza-6-thia-4-oxabicyclo[3.3.0]octane
shows definite changes in bond length in the directions ex-
pected for an anomeric effect, namely, shortening of the
C5—N bond (1.43 Å vs. 1.46 Å or 1.50 Å) and lengthening
of the C5—S bond (1.80 Å vs. 1.93 Å). There is even a
shortening of the C5—O bond (1.40 Å vs. 1.44 Å). Examina-
tion of the structure suggests a dihedral angle between the N
lone pair and the C—S bond of 140°. Apparently, this angle,
though well short of 180°, is enough to allow for overlap of
the lone pair and σ* antibonding orbitals. The longer C—S
bonds, relative to C—O bonds, force a loss of symmetry in
the bicyclic monothioacetal, and this allows partial overlap
of the sort leading to an anomeric effect.
Varian Gemini 300 (1H at 299.9 MHz, 13C at 75.4 MHz), a
Varian Mercury 400 (1H at 400.1 MHz, 13C at 100.6 MHz),
a Varian Inova 600 (1H at 599.7 MHz), or a Bruker AMX-
500 (University of Ottawa, Ont.) spectrometer (1H at
500.1 MHz, 13C at 125.8 MHz). Mass spectra were recorded
on a Finnigan MAT 8200 high-resolution spectrometer.
Melting points were recorded on a Gallenkamp melting
point apparatus.
Data were collected on a Nonius Kappa-CCD diffracto-
meter (either at the University of Western Ontario or the
University of Toronto) using COLLECT software (16). For
both compounds, the unit cell parameters were calculated
and refined from the full data set. Crystal cell refinement
and data reduction were carried out using the Nonius
DENZO package (17). The data were scaled using
SCALEPACK (18). The SHELXTL 5.1 (19) program pack-
age was used to solve the structure by direct methods, fol-
lowed by successive difference Fouriers. The molecule was
very well behaved. All non-hydrogen atoms were refined
with anisotropic thermal parameters. The hydrogen atoms
were calculated geometrically and were riding on their re-
spective carbon atoms.
Crystals of 5-(p-nitrophenyl)-1-aza-4,6-dioxabicyclo-
[3.3.0]octane were grown from diethyl ether. A colourless,
thin plate was mounted on a glass fibre. Data were collected
at low temperature (–73 °C). The crystal data and refinement
parameters for C11H12N2O4 are listed in Table S-1.3 Inter-
atomic distances and angles are listed in Table S-3. The re-
flection data and systematic absences were consistent with
the monoclinic space group P21/c. The largest residual elec-
tron density peak (0.155 e/A3) was not associated with any-
thing of significance. Full-matrix least-squares refinement
Experimental
2
on F gave R1 = 4.39 and wR2 = 10.43 at convergence.
General methods
Crystals of 5-(p-nitrophenyl)-1-aza-4-oxa-6-thiabicyclo-
[3.3.0]octane were grown from hexanes. A colourless, thin
plate was mounted on a glass fibre and sent to the University
Nuclear magnetic resonance spectra were recorded on a
Varian Gemini 200 (1H at 200.1 MHz, 13C at 50.3 MHz), a
© 2004 NRC Canada

Guthrie et al.
275
of Toronto. Data were collected at low temperature (–123 °C).
The crystal data and refinement parameters for C11H12N2O3S
are listed in Table S-6. Interatomic distances and angles are
listed in Table S-8. The reflection data were consistent with
the triclinic space group P1. The largest residual electron
density peak (0.724 e/A3) was not associated with anything
of significance. Full-matrix least-squares refinement on F2
gave R1 = 7.94 and wR2 = 16.88 at convergence.
The hexanes were removed by rotary evaporation. The resi-
due was doubly distilled at reduced pressure to give a clear
colourless liquid.
In this way were prepared the following five compounds.
5-Methyl-1-aza-4,6-dioxabicyclo[3.3.0]octane
25.5% yield; bp 52 °C (5.0 torr) (lit. (15) value bp 65°
(11 torr)). 1H NMR (400 MHz, CDCl3) δ: 1.49 (3H, s), 2.88–
2.96 (2H, m), 3.18–3.26 (2H, m), 3.76–3.83 (2H, m), 3.83–
3.90 (2H, m). 13C NMR (100 MHz, CDCl3) δ: 24.0, 53.0,
64.8, 123.1. Exact mass calcd. for C6H11NO2: 129.0790;
found: 129.0791.
Materials
1-Aza-4,6-dioxabicyclo [3.3.0] octane
1-Aza-4,6-dioxabicyclo [3.3.0] octane was prepared from
diethanolamine (Caledon) and N,N-dimethylformamide
dimethyl acetal (Aldrich) following Arnold and Kornilov
(20) in 80.9% yield; bp 68 °C (7.0 torr (1 torr =
5-Ethyl-1-aza-4,6-dioxabicyclo[3.3.0] octane
17.4% yield; bp 81 °C (9.5 torr) (lit. (15) value bp 72 °C
1
(12 torr)). H NMR (400 MHz, CDCl3) δ: 0.96 (3H, t), 1.78
1
133.322 Pa)) (lit. (20) value bp 65 to 66 °C (5 torr)). H
(2H, q), 2.92–3.00 (2H, m), 3.18–3.27 (2H, m), 3.78–3.85
(2H, m), 3.87–3.95 (2H, m). 13C NMR (100 MHz, CDCl3) δ:
8.7, 30.6, 53.6, 65.0, 125.1. Exact mass calcd. for
C7H13NO2: 143.0946; found: 143.0944.
NMR (400 MHz, CDCl3)4 δ: 2.90–2.96 (2H, m), 3.16–3.25
(2H, m), 3.77–3.87 (4H, m), 5.80 (1H, s). 13C NMR
(100 MHz, CDCl3) δ: 52.53, 64.39, 114.91. Exact mass
calcd. for C5H9NO2: 115.0633; found: 115.0633.
In the same way were prepared the following two com-
pounds.
5-Isopropyl-1-aza-4,6-dioxabicyclo[3.3.0] octane
1
31.2% yield; bp 86 °C (10 torr). H NMR (400 MHz,
CDCl3) δ: 0.97 (3H, s), 0.98 (3H, s), 1.96 (1H, sept), 2.92–
2.98 (2H, m), 3.17–3.25 (2H, m), 3.77–3.83 (2H, m), 3.85–
3.92 (2H, m). 13C NMR (100 MHz, CDCl3) δ: 17.4, 35.8,
54.3, 65.2, 126.8. Exact mass calcd. for C8H15NO2:
157.1103; found: 157.1098.
1-Aza-4-oxa-6-thiabicyclo[3.3.0]octane
75.8% yield; bp 64 °C (0.050 torr). H NMR (400 MHz,
CDCl3) δ: 2.78–2.85 (2H, m), 3.03–3.08 (1H, m), 3.10–3.14
(1H, m), 3.15–3.20 (1H, m), 3.48–3.52 (1H, m),.3.92–3.97
(1H,q), 4.08–4.13 (1H,q). 13C NMR (100 MHz, CDCl3) δ:
31.08, 49.59, 55.46, 65.37, 106.96. Exact mass calcd. for
C5H9NOS: 131.0405; found: 131.0398.
1
5-Methoxymethyl-1-aza-4,6-dioxabicyclo[3.3.0] octane
1
10.1% yield; bp 105 °C (8 torr). H NMR (400 MHz,
CDCl3) δ: 2.94–3.00 (2H, m), 3.24–3.31 (2H, m), 3.41
(3H, s), 3.47 (2H, s), 3.84–3.90 (2H, m), 3.92–3.98 (2H, m).
13C NMR (100 MHz, CDCl3) δ: 53.0, 59.6, 65.2, 73.9,
122.1. Exact mass calcd. for C7H13NO3: 159.0895; found:
159.0890.
5-Methyl-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
Prepared from from N,N-dimethylamide dimethyl acetal
(21); 90.2% yield; bp 58 °C (0.025 torr). 1H NMR
(400 MHz, CDCl3) δ: 1.81 (3H, s), 2.78–2.85 (1H, m), 2.90–
2.95 (1H, m), 3.02–3.12 (1H, m), 3.20–3.35 (2H, m), 3.50–
3.58 (1H, m), 3.92–4.02 (1H, q), 4.03–4.12 (1H, q). 13C
NMR (100 MHz, CDCl3) δ: 31.0, 33.8, 50.5, 55.8, 65.7,
119.2. Exact mass calcd. for C6H11NOS: 145.0561; found:
145.0557.
5-Phenyl-1-aza-4,6-dioxabicyclo[3.3.0] octane
17.3% yield; bp 150 °C (5.0 torr) (lit. (15) value bp 79 to
1
80 °C (0.03 torr)). H NMR (400 MHz, CDCl3) δ: 3.05–3.13
(2H, m), 3.30–3.40 (2H, m), 3.96–4.03 (2H, m), 4.09–4.16
(2H, m), 7.32 (3H, t), 7.61 (2H, d). 13C NMR (100 MHz,
CDCl3) δ: 53.0, 65.7, 123.2, 125.9, 128.1, 128.4, 140.1. Ex-
act mass calcd. for C11H13NO2: 191.0946; found: 191.0943.
General preparation of 5-substituted-1-aza-4,6-
dioxabicyclo[3.3.0]octanes
Diethanolamine (Caledon) (25.0 g, 0.238 mol) and a sliver
of sodium metal were placed in a two-necked, round-
bottomed flask fitted with a condenser. The system was
sealed with the use of septa. An exhaust needle was placed
in the septum at the top of the condenser, while a stream of
nitrogen was fed through the septum on the neck of the
round-bottomed flask. A positive pressure of nitrogen was
maintained throughout the remainder of the reaction. A mag-
netic stirrer and gentle heating with an oil bath (~50 °C)
helped dissolve the sodium metal. Once the sodium had
completely dissolved, the appropriate nitrile (0.239 mol) was
added to the reaction flask with a syringe. The temperature
of the oil bath was raised to 100 °C, and the reaction pro-
ceeded for 40 h at that temperature. Upon cooling, the reac-
tion mixture was placed in a separatory funnel, and the
product was extracted with multiple portions of hexanes.
5-(p-Nitrophenyl)-1-aza-4,6-dioxabicyclo[3.3.0]octane
The experimental procedure followed that of the synthesis
of 5-methyl-1-aza-4,6-dioxabicyclo[3.3.0]octane with some
modifications. A sliver of sodium metal was dissolved in
warm diethanolamine (Caledon) (25.1 g, 0.238 mol), as
usual. However, solid p-nitrobenzonitrile (Aldrich) (35.0 g,
0.236 mol) was dissolved in 1,4-dioxane (Caledon) prior to
addition to the reaction flask. The reaction proceeded for
48 h at 80 °C. Upon removal of the 1,4-dioxane by rotary
evaporation, a yellow solid precipitated. The solid was
recrystallized from diethyl ether, giving shiny, pale yellow
1
crystals (15.8 g, 28.3% yield); mp 90 to 91 °C. H NMR
(400 MHz, CDCl3) δ: 3.08–3.14 (2H, m), 3.28–3.38 (2H, m),
3.95–4.00 (2H, m), 4.05–4.14 (2H, m), 7.77 (2H, d), 8.19
(2H, d). 13C NMR (100 MHz, CDCl3) δ: 53.8, 66.5, 122.4,
4 Abbreviations used: m, multiplet; s, singlet, d, doublet; t, triplet; q, quartet.
© 2004 NRC Canada

276
Can. J. Chem. Vol. 82, 2004
123.9, 127.7, 147.5, 148.1. Exact mass calcd. for
C11H12N2O4: 236.0797; found: 236.0807.
modifications. Diethanolamine (Caledon) (12.5 g, 0.119
mol) and 75 mL of 1,4-dioxane (Caledon) were placed in the
reaction flask with no catalyst. Caution: The standard reac-
tion of 1:1 amine:nitrile with sodium catalyst in the absence
of solvent resulted in an explosion. Dichloroacetonitrile
(Aldrich) (10.0 mL, 0.125 mol) was added slowly to the re-
action flask with a syringe. Once the nitrile was completely
added, the oil bath was heated gently until it reached 50 °C.
The evolution of ammonia appeared to be very rapid, so
heating was stopped. The reaction continued at room tem-
perature for 48 h. The 1,4-dioxane solvent was removed by
rotary evaporation, leaving a thick residue. It was placed in
a separatory funnel for extraction with hexanes. A white
solid precipitated in the separatory funnel when hexanes
were added. Recrystallization with hexanes resulted in white
shiny flakes (5.65 g, 24.0% yield); mp 62 to 63 °C. 1H NMR
(400 MHz, CDCl3) δ: 3.02–3.10 (2H, m), 3.42–3.51 (2H, m),
4.02–4.14 (4H, m), 5.75 (1H, s). 13C NMR (100 MHz,
CDCl3) δ: 54.3, 67.2, 74.0, 123.4. Exact mass calcd. for
C6H9NO2Cl2: 197.0009; found:. 197.0007.
5-(p-Chlorophenyl)-1-aza-4,6-dioxabicyclo[3.3.0]octane
Solid p-chlorobenzonitrile (Aldrich) (20.0 g, 0.145 mol)
was placed in a round-bottomed flask. Enough diethano-
lamine (Caledon) was added to the reaction flask to com-
pletely cover the nitrile. This excess of diethanolamine
helped reduce the amount of p-chlorobenzonitrile that sub-
limed onto the upper walls of the reaction flask during the
reaction. The round-bottorned flask was fitted with a con-
denser and kept under an N2 atmosphere. The oil bath was
heated to 50 °C and the reaction proceeded for 4 days. The
reaction flask was allowed to cool before addition of hex-
anes to extract the title compound. Removal of hexanes by
rotary evaporation resulted in a white solid that contained
the target compound and unreacted nitrile. Recrystallization
could not remove all of the nitrile. The white solid was
placed in a sublimation apparatus, to take advantage of the
willingness of the nitrile to sublime. Reduced pressure
(0.025 torr) and very gentle heating removed the nitrile ex-
clusively. The remaining solid was recrystallized in hexanes
5-Ethyl-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
1
With some heating, a sliver of sodium metal (Aldrich) was
dissolved in freshly prepared 2-(2-mercaptoethyl)-amino-
ethanol (20.0 g, 0.165 mol) in a round-bottomed flask.
Propionitrile (Aldrich) (25 mL, 0.328 mol) was added to the
reaction flask and allowed to stir for 2 weeks at room tem-
perature under an N2 atmosphere. It was found that an ex-
cess of nitrile made the consistency of the final reaction
mixture easier to handle. An extraction with diethyl ether re-
moved the title compound and its corresponding oxazoline.
The mixture was doubly distilled at reduced pressure to give
a clear colorless fraction (3.5 g, 13.3% yield); bp 50 °C
to give white crystals (2.9 g, 8.9% yield); mp 43–45 °C. H
NMR (400 MHz, CDCl3) δ: 3.05–3.12 (2H, m), 3.29–3.37
(2H, m), 3.94–4.00 (2H, m), 4.06–4.13 (2H, m), 7.30 (2H,
d), 7.53 (2H, d). 13C NMR (100 MHz, CDCl3) δ: 53.0, 65.8,
122.8, 127.6, 128.1, 134.3, 138.8. Exact mass calcd. for
C11H12NO2Cl: 225.0557; found: 225.0557.
5-Trifluoromethyl-1-aza-4,6-dioxabicyclo[3.3.0]octane
A sliver of sodium metal was dissolved in diethanolamine
(Caledon) (28.7 g, 0.273 mol) and placed in a 300 cc Parr
Instruments bomb. The bomb was placed into a dry ice – ac-
etone bath to lower the temperature of the bomb and its con-
tents to –78 °C. At that temperature, the contents of the
bomb solidified. Freshly prepared, condensed trifluoroace-
tonitrile (prepared from trifluoroacetamide (22)) (~30 mL,
~0.316 mol) was poured into the chilled bomb. The bomb
was then immediately sealed. Undoubtedly some trifluoro-
acetonitrile was lost during this crude transfer. The bomb
was placed into its temperature controller and set to 70 °C.
The temperature inside the bomb rose dramatically, requir-
ing cooling of the bomb. Once the temperature was brought
under control, the bomb was adjusted to 70 °C where it re-
mained for 48 h. At the end of the reaction period, the bomb
was flushed with nitrogen before opening to ensure removal
of any remaining trifluoroacetonitrile. Inside the bomb was
an orange viscous liquid that was mainly unreacted di-
ethanolamine. Extraction with hexanes resulted in pure 5-
trifluoromethyl-1-aza-4,6-dioxabicyclo[3.3.0]octane. The liquid
crystallized on its own after sitting on the bench for
2 weeks. Recrystallization from hexanes resulted in translu-
cent crystals (0.25 g, 0.5% yield). 1H NMR (400 MHz, CDCl3)
δ: 3.01–3.09 (2H, m), 3.31–3.40 (2H, m), 3.99–4.06 (2H, m),
4.07–4.15 (2H, m). 19F NMR (300 MHz, CFCl3) δ: 83.0. 13C
NMR (100 MHz, CDCl3) δ: 52.5, 66.7, 120.5, 123.3. Exact
mass calcd. for C6H8NO2F3: 183.0505; found: 183.0503.
1
(0.025 torr). H NMR (400 MHz, CDCl3) δ: 1.05 (3H, t),
1.99 (2H, m), 2.80–2.88 (2H, m), 2.99–3.09 (1H, m), 3.19–
3.29 (2H, m), 3.48–3.57 (1H, m), 3.92 (1H, q), 4.05 (1H, q).
13C NMR (100 MHz, CDCl3) δ: 9.9, 33.0, 37.0, 51.0, 56.0,
65.5, 123.3. Exact mass calcd. for C7H13NOS: 159.0718;
found: 159.0717.
In the same way were prepared the following three com-
pounds.
5-Isopropyl-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
1
14.3% yield; bp 53 °C (0.025 torr). H NMR (400 MHz,
CDCl3) δ: 1.00 (3H, d), 1.10 (3H, d), 2.06 (1H, sept), 2.78–
2.88 (2H, m), 2.98–3.08 (1H, m), 3.14–3.25 (2H, m), 3.47–
3.55 (1H, m), 3.90 (1H, q), 4.03 (1H, q). 13C NMR
(100 MHz, CDCl3) δ: 18.1, 18.5, 32.8, 40.2, 52.0, 56.9, 65.8,
126.4. Exact mass calcd. for C8H15NOS: 173.0874; found:
173.0878.
5-Methoxymethyl-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
1
19.0% yield; bp 73 °C (0.025 torr). H NMR (400 MHz,
CDCl3) δ: 2.75–2.86 (2H, m), 3.00–3.12 (1H, m), 3.15–3.27
(2H, m), 3.40 (3H, s), 3.48–3.55 (1H, m), 3.58 (2H, s), 3.98
(1H, q), 4.08 (1H, q). 13C NMR (100 MHz, CDCl3) δ: 32.0,
51.0, 56.0, 60.0, 66.0, 76.5, 120.0. Exact mass calcd. for
C7H13NO2S: 175.0667; found: 175.0671.
5-Dichloromethyl-1-aza-4,6-dioxabicyclo[3.3.0] octane
The experimental procedure followed that of the synthesis
of 5-methyl-1-aza-4,6-dioxabicyclo[3.3.0]octane, with some
5-Phenyl-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
The product mixture, after extraction, was distilled at re-
duced pressure only to remove the by-product, 2-phenyl-
© 2004 NRC Canada

Guthrie et al.
277
oxazoline (~80 °C at 0.25 torr), leaving relatively pure
bicyclic amide acetal in the still-pot. The contents of the
still-pot were dissolved in diethyl ether and placed in a
freezer. After 5 days, a solid had precipitated and was
recrystallized, giving white shiny flakes (12.6% yield); mp
46–48 °C. 1H NMR (500 MHz, CDCl3) δ: 2.96–3.05
(2H, m), 3.14–3.20 (1H, m), 3.28–3.34 (2H, m), 3.53–3.57
(1H, m), 4.00 (1H, q), 4.19 (1H, q), 7.27 (1H, t), 7.32 (2H,
t), 7.63 (2H, d). 13C NMR (125 MHz, CDCl3) δ: 33.74,
50.77, 56.37, 65.16, 120.00, 127.86, 142.92. Exact mass
calcd. for C11H13NOS: 207.0718; found: 207.0723.
5-(p-Methoxyphenyl)-1-aza-4-oxa-6-thiabicyclo[3.3.0]oc-
tane
The experimental procedure followed that of the synthesis
of 5-(p-chlorophenyl)-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane,
using freshly prepared 2-(2-mercaptoethyl)-aminoethanol
(22.5 g, 0.186 mol) and p-methoxybenzonitrile (Aldrich)
(25.7 g, 0.193 mol). After sublimation of the product mix-
ture, the remaining colorless liquid was determined to be
90% title compound and 10% p-methoxyphenyl-oxazoline
1
(3.4 g, 7.7% yield). H NMR (400 MHz, CDCl3) δ: 2.96–
3.07 (2H, m), 3.07–3.20 (1H, m), 3.27–3.37 (2H, m), 3.50–
3.60 (1H, m), 3.78 (3H, s), 4.02 (1H, q), 4.20 (1H, q), 6.82
(2H, d), 7.55 (2H, d). 13C NMR (100 MHz, CDCl3) δ: 34.0,
51.5, 55.5, 56.5, 66.0, 120.4, 127.5, 130.0. Exact mass calcd.
for C12H15NO2S: 237.0824; found: 237.0825.
5-(p-Nitrophenyl)-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
With some heating, a sliver of sodium metal was dis-
solved in freshly prepared 2-(2-mercaptoethyl)-aminoethanol
(20.0 g, 0.165 mol) in a round-bottomed flask. In a separate
flask, solid p-nitrobenzonitrile (Aldrich) (24.5 g, 0.165 mol)
was dissolved in 1,4-dioxane (Caledon). The nitrile solution
was then added to the reaction flask. The reaction mixture
was allowed to stir for 2 weeks at room temperature under
an N2 atmosphere. At the end of the reaction period, rotary
evaporation was used to remove the 1,4-dioxane solvent. Di-
ethyl ether was used to extract the title compound from the
residue. The ethereal solution was placed in the freezer to
facilitate crystallization. Within a short period of time, a yel-
low solid formed, which was collected and recrystallized
from ether to give shiny, pale yellow flakes (5.4 g, 13.0%
2-(2-Mercaptoethyl)aminoethanol
This procedure was based on a preparation of N-(2-
methoxyethyl)-N-(2-mercaptoethyl)amine (23). Ethanolamine
(Aldrich) (25.0 g, 0.409 mol) and 200 mL of 1,4-dioxane
(Caledon) were added to a two-necked, round-bottomed
flask. A condenser and a dropping funnel, containing
roughly 25 mL of 1,4-dioxane, were connected to the round-
bottomed flask. Septa were used to seal the system. An ex-
haust needle was placed in one septum, while nitrogen was
introduced through the other. The contents were warmed to a
temperature just below that at which refluxing occurred. At
this point, ethylene sulfide (Aldrich) (25.0 g, 0.416 mol) was
transferred with the use of a cannula to the dropping funnel
that contained dioxane. Once transfer was completed, the
dropping funnel was opened to allow the dioxane solution
containing ethylene sulfide to be added dropwise. On com-
pletion of the addition, the reaction mixture was allowed to
reflux for 6 h. The reaction mixture was allowed to cool be-
fore removal of the dioxane by rotary evaporation. The resi-
due was immediately distilled under reduced pressure. If the
residue was not distilled immediately (for instance, if it were
distilled the following day), a noticeable reduction in yield
was observed. The distillate collected was a clear, colorless
liquid and, on occasion, a white solid when returned to at-
mospheric pressure (23.7 g, 47.8% yield); bp 50 °C (0.050
torr) (lit. (24) value bp 66–69 °C (0.03 torr)). 1H NMR
(400 MHz, CDCl3) δ: 1.97 (3H, br s), 2.66 (2H, t), 2.76 (2H,
t), 2.81 (2H, t), 3.63 (2H, t). 13C NMR (100 MHz, CDCl3) δ:
25.0, 50.4, 51.6, 60.9. Exact mass calcd. for C4H11NOS:
121.0561; found: 121.0558.
1
yield); mp 75 to 76 °C. H NMR (400 MHz, CDCl3) δ:
3.05–3.15 (2H, m), 3.20–3.29 (1H, m), 3.33–3.40 (2H, m),
3.61–3.68 (1H, m), 4.02 (1H, q), 4.23 (1H, q), 7.75 (2H, d),
8.18 (2H, d). 13C NMR (100 MHz, CDCl3) δ: 34.8, 51.8,
57.8, 65.8, 119.0, 124.0, 127.0, 147.9, 150.2. Exact mass
calcd. for C11H12N2O3S: 252.0569; found: 252.0565.
5-(p-Chlorophenyl)-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane
The experimental procedure followed that of the synthesis
of 5-(p-nitrophenyl)-1-aza-4-oxa-6-thiabicyclo[3.3.0]octane,
using freshly prepared 2-(2-mercaptoethyl)-aminoethanol
(20.0 g, 0.165 mol) and p-chlorobenzonitrile (Aldrich)
(25.1 g, 0.183 mol) dissolved in 1,4-dioxane. When the ethe-
real solution from the extraction was placed in the freezer,
white needle-like crystals formed. These were later con-
firmed to be p-chlorophenyl-oxazoline. Several attempts at
recrystallization successively removed less and less
oxazoline. The product mixture still contained oxazoline.
Since the oxazoline had a lower boiling point than the
bicyclic amide monothioacetal, sublimation was used in a fi-
nal attempt to remove the oxazoline. The sublimation was
carried out under reduced pressure (0.025 torr) with very
gentle heating. Before long, the cold finger of the sublima-
tion apparatus was coated with a white solid. The remaining
colorless liquid in the sublimation apparatus was determined
to be 94% pure by NMR (4.2 g, 10.5% yield). Repeated at-
tempts at sublimation provided no improvement in the purity
Acknowledgments
We thank the Natural Sciences and Engineering Research
Council of Canada for financial support of this work.
References
1. J.P. Guthrie. J. Am. Chem. Soc. 95, 6999 (1973).
2. J.P. Guthrie. J. Am. Chem. Soc. 96, 3608 (1974).
3. J.P. Guthrie. Can. J. Chem. 54, 202 (1976).
4. J.P. Guthrie. Can. J. Chem. 55, 3562 (1977).
5. J.P. Guthrie. J. Am. Chem. Soc. 100, 5892 (1978).
6. J.P. Guthrie and P.A. Cullimore. Can. J. Chem. 58, 1281
(1980).
1
of the title compound. H NMR (400 MHz, CDCl3) δ: 2.98–
3.11 (2H, m), 3.14–3.22 (1H, m), 3.28–3.38 (2H, m), 3.54–
3.61 (1H, m), 4.02 (1H, q), 4.21 (1H, q), 7.28 (2H, d), 7.55
(2H, d). 13C NMR (100 MHz, CDCl3) δ: 34.01, 51.0, 56.6,
65.5, 119.4, 127.2, 128.3, 133.8, 141.7. Exact mass calcd.
for C11H12ClNOS: 241.0328; found: 241.0326.
© 2004 NRC Canada

278
7. J.P. Guthrie, J. Barker, P.A. Cullimore, J. Lu, and D.C. Pike.
Can. J. Chem. Vol. 82, 2004
17. Nonius. DENZO package [computer program]. Delft, The
Netherlands, 1997.
Can. J. Chem. 71, 2109 (1993).
8. J.P. Guthrie. Can. J. Chem. 53, 898 (1975).
9. ACD Labs. H NMR spin simulation software [computer pro-
gram]. Version 3.5. Advanced Chemistry Development, Inc.,
Toronto, Ont. 1998.
18. Nonius. SCALEPACK [computer program]. Delft, The Nether-
lands, 1997.
19. G.M. Sheldrick. SHELXTL 5.1. Bruker-Axn Inc., Madison,
Wis. 1997.
10. D.L. Pavia, G.M. Lampman, and G.S. Kriz. Introduction to
spectroscopy. Saunders College Publishing, Orlando. 1996.
11. M. Karplus. J. Am. Chem. Soc. 85, 2870 (1963).
12. H. Conroy. Adv. Org. Chem. 2, 265 (1960).
13. D. Williams and I. Fleming. Spectroscopic methods in organic
chemistry. Wiley, New York. 1989. p. 92.
20. Z. Arnold and M. Kornilov. Collect. Czech. Chem. Commun.
29, 645 (1964).
21. S. Hanessian and E. Moralioglu. Can. J. Chem. 50, 233
(1972).
22. H. Gilman and R. Jones. J. Am. Chem. Soc. 65, 1458 (1943).
23. J. Haefele and R. Broge. Proc. Sci. Sect. Toilet Goods Assoc.
32, 52 (1959).
14. M. Anteunis, J. Gelan, and R. Van Cauwenberghe. Org. Magn.
Reson. 6, 362 (1974).
24. R. Weinman, M. Gollis, J. James, and A. Pomponi. J. Org.
15. R. Feinauer. Synthesis, 16 (1971).
Chem. 27, 4222 (1962).
16. Nonius. COLLECT software [computer program]. Delft, The
Netherlands. 1998.
© 2004 NRC Canada