Synthesis and conformational analysis of saturated cis-and trans-1,3,2-benzodiazaphosphinine 2-oxides

Article abstract of DOI:10.1002/ejoc.200500668

By cyclization of N-unsubstituted, and N¹- or N ³-methyl-substituted cis- and trans-2-(aminomethyl)cyclohexylamines with phenylphosphonic dichloride, phenyl dichlorophosphate and bis(2-chloroethyl)phosphoramidic dichloride, P epimeri

Full text of DOI:10.1002/ejoc.200500668

FULL PAPER
DOI: 10.1002/ejoc.200500668
Synthesis and Conformational Analysis of Saturated cis-and trans-1,3,2-
Benzodiazaphosphinine 2-Oxides
Zita Zalán,[a,b] Henri Kivelä,[a] László Lázár,[a,b] Ferenc Fülöp,*[b] and Kalevi Pihlaja*[a]
Keywords: Phosphorus heterocycles / Fused-ring systems / Structure elucidation / Configuration determination / Confor-
mation analysis
By cyclization of N-unsubstituted, and N1- or N3-methyl-sub-
stituted cis- and trans-2-(aminomethyl)cyclohexylamines
with phenylphosphonic dichloride, phenyl dichlorophos-
phate and bis(2-chloroethyl)phosphoramidic dichloride, P
epimeric diastereomers a and b of the corresponding 1,3-un-
substituted and 1- or 3-methyl-substituted 2-phenyl-, 2-
phenoxy- and 2-[bis(2-chloroethyl)amino]decahydro-1,3,2-
benzodiazaphosphinine 2-oxides have been synthesized.
The stereochemistry and conformations of the prepared satu-
fused derivatives with an equatorial phenoxy group in a hy-
pothetical double-chair conformation (14a, 16a, 18a) were
observed to adopt nonchair heteroring conformations with a
pseudoaxial P–OPh bond instead. The corresponding b epi-
mers, as well as the trans-fused 2-phenyl and 2-[bis(2-chloro-
ethyl)amino] derivatives (both P epimers) were found to re-
tain double-chair conformations. In the cis-fused series, the
position of the N-in/N-out equilibrium was dependent on the
steric and stereoelectronic requirements of the phosphorus
rated 1,3,2-benzodiazaphosphinine 2-oxides were deter- substituents (and thus on the phosphorus configuration) as
mined by 1H, 13C and 31P NMR spectroscopy aided by DFT
geometry optimizations and J-coupling-constant calculations
for selected structures. The 2-phenoxy-substituted trans-
well as on the substituents at the ring nitrogen atoms.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
hydes,[8] α-alkylations of P-alkyl derivatives[9] and re-
ductions of ketones.[10] Stereochemical studies on some mo-
nocyclic and tetrahydroisoquinoline-condensed 1,3,2-diaza-
phosphinane 2-oxide model compounds have led to the
conclusion that, in contrast to the conformationally diverse
1,3,2-oxazaphosphinane analogues, these compounds could
be characterized by chair or twisted-chair conforma-
tions.[4,11]
As a continuation of our previous studies on cyclohex-
ane-condensed phosphorus-containing 1,2,3-heterocy-
cles,[12,13] our aim in this work was to prepare P epimeric
diastereomers of cis- and trans-1,2,3,4,4a,5,6,7,8,8a-decahy-
dro-1,3,2-benzodiazaphosphinine 2-oxides bearing substitu-
ents on the heteroatoms for the purpose of investigating the
effects of the presence of these substituents and the relative
configuration of the phosphorus atom on the predominant
conformation of the saturated bicyclic system.
Introduction
As a consequence of their valuable pharmacological ef-
fects and synthetic utility, 1,3,2-oxazaphosphinane deriva-
tives have attracted considerable interest. This ring system
is found in alkylating anticancer drugs (cyclophosphamide,
ifosfamide), numerous derivatives of which have been syn-
thesized to determine their structure–activity relation-
ships.[1] Alkylation of phosphorus-stabilized carbanions de-
rived from 1,3,2-oxazaphosphinane 2-oxides have proved to
be an efficient method for the diastereoselective construc-
tion of carbon–carbon bonds.[2,3]
However, the analogous 1,3,2-diazaphosphinane ring
system has been less thoroughly studied either from a phar-
macological or stereochemical point of view.[4] The 1,3,2-
N,N,P analogue of cyclophosphamide was found to have
no significant anticancer activity.[5] The synthetic impor-
tance of 1,3,2-N,N,P heterocycles increased when 1,3,2-di-
azaphosphinane phosphoramides proved to be effective
auxiliaries with which to induce stereoselective carbon–car-
bon or carbon–hydrogen bond-forming reactions, for exam-
ple, asymmetric aldol reactions,[6,7] allylations of alde-
Results and Discussion
Synthesis
[a] Department of Chemistry, University of Turku,
20014 Turku, Finland
The methods applied earlier to the synthesis of 1,3,2-
N,N,P heterocycles were based on the ring-closure of the
corresponding diamines with the appropriate phosphorus-
containing fragments.[4,14] The cis and trans cyclohexane
1,3-diamines 3a,b and 4a,b required for the cyclization reac-
tions were prepared by lithium aluminium hydride re-
E-mail: kpihlaja@utu.fi
[b] Institute of Pharmaceutical Chemistry, University of Szeged,
6701 Szeged, POB 427, Hungary
E-mail: fulop@pharm.u-szeged.hu
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2145

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
duction of the corresponding cis- and trans-2-aminocyclo- raphy. The ratios of the diastereomers formed were deter-
hexanecarboxamides (1a,b and 2a,b) (Scheme 1).[15,16] Di- mined from the NMR spectra of the crude products (see
amines 6a,b, bearing a methyl substituent on the nitrogen Exp. Sect.).
atom adjacent to the cyclohexane ring, were synthesized by
analogous reductions of the urethanes 5a,b derived from
carboxamides 1a,b.
Characterization of Structures
The solution-state structures and conformations of the
prepared compounds 7–22a,b in CDCl3 were characterized
1
by H, 13C and 31P NMR spectroscopic methods at 25 °C.
A limited conformational search followed by J-coupling-
constant calculations on the geometry-optimized conforma-
tions was also undertaken for the epimeric pair 10a and 10b
by using DFT methods.
1
The H and 13C NMR signals (see Tables 1, 2, 3, 4, 5, 6,
and 7) were assigned with the help of standard 2D homo-
and heteronuclear correlation methods (cf. Exp. Sect.). The
signals from (diastereotopic) geminal methylene protons
could be individually assigned based on their chemical
shifts and/or J coupling constants: In the trans-fused series,
axial protons (Hax) in the carbocycle always resonate at
lower frequencies than their equatorial counterparts (Heq)
and display wider, better-resolved multiplet patterns as a
result of large vicinal JHax,Hax coupling constants. Likewise,
the J coupling between 4-Hax and 4a-H is larger than that
between 4-Heq and 4a-H (10.5–11.5 and 3.7–4.9 Hz, respec-
Scheme 1. Reagents and conditions: (i) LiAlH4, THF, reflux, 12 h,
51% (3a), 55% (3b); see ref.[16] (4a,b); (ii) ClCOOEt, NaHCO3,
toluene, H2O, room temp., 1 h, 75% (5a), 98% (5b); (iii) LiAlH4,
THF, reflux, 3 h, 43% (6a), 36% (6b). R = H: 1, 3; R = Me: 2, 4;
cis: a; trans: b.
The cyclization reactions of the cis and trans diamines tively). For compounds in which the heteroring prefers a
3a,b, 4a,b and 6a,b with phenylphosphonic dichloride, chair conformation (Figure 1), 4-Heq can also be readily as-
phenyl dichlorophosphate and bis(2-chloroethyl)phos- signed because of its large J coupling to the phosphorus
phoramidic dichloride were performed at ambient tempera- atom. Even though the carbocyclic Hax signals did not
ture to give 1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodi- overlap with the Heq signals, there was often significant
azaphosphinine 2-oxides 7–22 (Scheme 2 and Scheme 3). In overlap amongst the protons in similar stereopositions (e.g.,
all cases, two diastereomers (a and b) differing in the config- between 6-Hax and 7-Hax); unless an accurate spin simula-
uration of the phosphorus atom relative to the carbon tion was obtained these were only assigned as a group to
atoms at the annelation positions were formed which, ex- the corresponding δ range. The C-6 and C-7 carbon signals
cept for 20a,b, could be separated by column chromatog- of the trans-fused derivatives could be assigned based on a
Scheme 2. Reagents and conditions: (i) Cl2POPh/Et3N, THF, column chromatography, 8–26%; (ii) Cl2PO(OPh)/Et3N, THF, column
chromatography, 22–36%; (iii) Cl2PO[N(CH2CH2Cl)2]/Et3N, THF, column chromatography, 15–29%. R1 = R2 = H: 3a, 7, 13, 19; R1 =
Me, R2 = H: 4a, 9, 15, 21; R1 = H, R2 = Me: 6a, 11, 17, 22.
2146
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
Scheme 3. Reagents and conditions: (i) Cl2POPh/Et3N, THF, column chromatography, 10–21%; (ii) Cl2PO(OPh)/Et3N, THF, column
chromatography, 15–33%; (iii) Cl2PO[N(CH2CH2Cl)2]/Et3N, THF, column chromatography. R1 = R2 = H: 3b, 8, 14, 20; R1 = Me, R2 =
H: 4b, 10, 16; R1 = H, R2 = Me: 6b, 12, 18.
W-type coupling (1.2–3.3 Hz) of the latter to the phospho- constants were required. Also, interpolating the NMR pa-
rus atom, resolved in the 1D 13C NMR spectrum, even rameters of readily assigned compounds 13a (assumed to
when the small difference in their chemical shifts made the be pure N-in, vide infra) and 17b (assumed to be pure N-
assignment by 2D methods uncertain. The compounds in out) to different population-weighted average values was
the cis-fused series can undergo ring-inversion, which is fast helpful in the assignment of the carbocyclic part of the cis
enough on the NMR timescale at 25 °C to yield population- derivatives (the “population” could be varied until the in-
averaged signals. Since the axial and equatorial positions of terpolated NMR parameters were consistent with the ob-
ring substituents are exchanged upon ring-inversion, we served ones).
have instead denoted the diastereotopic hydrogen atoms in
For each product, the NMR spectra were consistent with
the cis series as Hx and Hy: Hx atoms are axial in the limit- the constitution, connectivity and the type of ring-fusion
ing N-in invertomer and become equatorial in the N-out (cis or trans) of the expected structure. For example, trans-
form (Figure 2) and vice versa for Hy. In the cases in which and cis-fused compounds displayed large (9.4–10.7 Hz) and
3
one of the ring-inverted forms was strongly preferred over small (2.7–5.0 Hz) values of J4a–H,8a-H, respectively, and
4
the other, the Hx and Hy protons could be readily assigned only in the former series was the W-type JP,C-7 coupling
on the basis of their J-coupling patterns in the same way as (1.2–3.3 Hz) resolved; the ipso-carbon atom of the aromatic
with the trans-fused compounds. For compounds notably ring in the 2-phenoxy-substituted compounds was deshi-
populating both forms, a successful spin simulation and a elded with respect to that of phenyl-substituted compounds
more careful examination of the resulting vicinal coupling (151.4–152.4 and 131.9–134.7 ppm, respectively) and the C-
1
Table 1. Selected H NMR chemical shifts [ppm] and JH,H and JP,H coupling constants [Hz] for the trans-fused compounds at 25 °C in
CDCl3 (δTMS = 0.00 ppm).
[a]
[b]
1-R2
3-R1
2-Z
4ax
4eq
4a
8a
4ax,4eq
4ax,4a 4ax,P
4eq,4a 4eq,P
4a,8a
8a,P
8a
H
H
H
H
CH3
CH3
H
H
H
H
H
CH3
CH3
H
H
H
H
CH3
CH3
H
H
H
Ph
Ph
Ph
Ph
Ph
Ph
OPh
OPh
OPh
OPh
OPh
OPh
3.16
2.73
3.05
2.94
3.18
2.93
2.91
2.95
2.85
2.92
3.00
2.91
3.15
3.18
2.92
3.05
3.01
3.08
3.16
3.10
2.94
2.90
3.08
2.98
3.07
3.12
1.61
1.55
1.70
1.83
1.71
1.85
1.59
1.44
1.67
1.57
1.97
1.58
1.43
1.37
3.20
2.70
3.16
2.89
2.97
2.90
2.98
2.96
2.98
2.86
2.68
2.82
3.07
2.74
–11.3
–12.9
–11.0
–11.7
–11.7
–13.2
–10.8
–12.9
–10.6
–11.7
–10.9
–13.8
–11.4
–13.0
11.2
11.0
10.8
11.2
11.3
11.1
11.1
11.3
10.5
11.2
11.5
11.5
11.4
11.1
3.9
2.6
3.1
2.8
3.2
1.5
10.6
1.3
9.8
1.4
5.0
1.3
2.7
3.8
4.4
3.9
4.6
4.7
3.6
3.8
4.9
4.0
4.9
4.3
3.7
4.0
4.1
4.3
21.5
23.8
20.2
16.9
23.5
24.2
11.3
28.1
9.4
26.2
16.6
28.2
26.1
21.3
10.0
10.1
9.4
2.1
–0.6
2.1
1.0
1.9
1.5
3.9
1.3
3.6
1.1
7.2
1.1
0.6
0.9
8b
10a
10b
12a
12b
14a
14b
16a
16b
18a
18b
20a
20b
10.1
10.1
10.1
10.3
10.0
10.1
10.1
10.6
9.9
H
CH3
CH3
H
[c]
NR2[c] 3.08
10.7
10.0
H
H
NR2 2.78
[a] δ4-Hax abbreviated to 4ax and similarly for others. [b] 2J4-Hax,4-Heq abbreviated to 4ax,4eq and similarly for others. [c] R = CH2CH2Cl.
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2147

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
1
Table 2. Selected H NMR chemical shifts [ppm] for the cis-fused compounds at 25 °C in CDCl3 (δTMS = 0.00 ppm).
1-R2
3-R1
2-Z
4-Hx
4-Hy
4a-H
5-Hx
5-Hy
6-Hx
6-Hy
7-Hx
7-Hy
8-Hx
8-Hy
8a-H
7a
H
H
H
H
CH3
CH3
H
H
H
H
H
CH3
CH3
H
H
H
H
CH3
CH3
H
H
H
Ph
Ph
Ph
Ph
Ph
Ph
OPh
OPh
OPh
OPh
OPh
OPh
NR2
NR2
NR2
NR2
NR2
NR2
3.15
3.57
3.27
3.47
3.09
3.25
3.40
3.03
3.35
2.85
3.24
2.88
3.20
3.49
3.03
3.34
2.94
3.26
3.12
3.06
3.05
2.83
3.55
3.34
3.01
3.49
2.78
3.43
2.98
3.51
3.07
2.99
3.15
2.73
3.63
3.18
1.52
1.87
1.89
1.87
2.30
2.39
1.52
2.18
1.62
2.32
1.77
2.23
1.54
1.70
1.99
1.68
2.23
2.19
2.10
2.01
2.13
2.10
1.83
1.82
2.00
1.67
2.01
1.63
1.98
1.63
1.85
1.37
1.45
1.47
1.48
1.58
1.54
1.32
1.53
1.39
1.52
1.35
1.55
1.37
1.31
1.37
1.34
1.35
1.47
1.46
1.32
1.42
1.28
1.41
1.38
1.46
1.33
1.76
1.72
1.70
1.78
1.46
1.44
1.80
1.27
1.76
1.31
1.70
1.22
1.71
1.63
1.61
1.65
1.54
1.79
1.69
1.55
1.73
1.40
1.72
1.54
1.82
1.48
1.56
1.61
1.52
1.81
1.61
1.47
1.43
1.42
1.43
1.27
1.29
1.48
1.26
1.49
1.22
1.39
1.19
1.48
1.62
1.68
1.65
1.65
2.00
1.80
1.64
1.70
1.58
1.63
1.48
2.04
1.65
1.73
1.63
1.74
1.65
1.78
1.66
1.65
1.96
1.68
1.87
2.01
1.71
1.72
3.45
3.86
3.61
3.85
3.11
3.22
3.71
3.41
3.64
3.28
3.30
3.06
3.44
3.77
3.36[d]
3.77
2.97
3.14
7b
9a
9b
11a
11b
13a
13b
15a
15b
17a
17b
19a
19b
21a
21b
22a
22b
H
CH3
CH3
H
H
H
H
CH3
CH3
H
1.85
1.40
1.33
1.73
1.44
1.64
1.59
[a]
[b]
[c]
[b]
[c]
[a]
[a]
CH3
CH3
H
1.83
1.66
1.75
1.41
1.56
1.49
1.30
1.45
1.42
1.77
1.31
1.49
1.46
1.19
1.33
1.60
2.09
1.69
1.60
1.60
1.72
H
)
[a] δ = 1.67–1.81 ppm (3 H; 5-Hx, 8-Hx, 8-Hy). [b] δ = 1.45–1.54 ppm (2 H; 5-Hy, 6-Hy . [c] δ = 1.30–1.39 ppm (2 H; 6-Hx, 7-Hy). [d] Under
the signals from R = CH2CH2Cl, detected by HSQC.
Table 3. 13C NMR chemical shifts [ppm] for the trans-fused compounds at 25 °C in CDCl3 (δTMS = 0.0 ppm).
1-R2 3-R1 2-Z
4
4a
5
6
7
8
8a
R2
R1
i-C
o-C
m-C
p-C
8a
H
H
H
H
CH3
CH3
H
H
H
H
H
CH3
CH3
H
H
H
H
CH3
CH3
H
H
H
Ph
Ph
Ph
Ph
Ph
Ph
OPh
OPh
OPh
OPh
OPh
OPh
47.0
48.3
56.0
57.6
46.1
47.5
47.6
48.4
56.4
57.5
47.4
47.4
41.7
43.1
42.2
42.1
43.0
42.7
39.9
42.6
40.0
42.4
40.4
42.8
41.3
43.1
28.7
28.3
28.7
28.7
29.5
29.1
29.2
28.0
29.2
28.3
29.9
28.7
28.2
28.5
25.4
25.3
25.3
25.1
25.3
25.1
25.2
25.3
25.0
25.1
25.3
25.0
25.3
25.3
24.5
24.8
24.6
24.7
24.9
25.0
24.2
24.6
24.1
24.9
24.5
24.8
24.3
24.9
35.2
34.8
34.7
34.5
30.7
30.8
35.2
34.5
35.1
33.8
30.7
30.1
35.2
34.6
55.5
57.4
54.7
56.9
62.5
64.3
56.5
57.7
56.6
57.1
64.8
63.6
56.4
57.3
–
–
–
–
30.9
30.5
–
–
–
–
–
35.3
35.3
–
–
–
–
34.9
35.5
–
–
–
134.2 132.3 128.1 131.4
133.9 131.0 128.5 131.3
132.5 132.8 128.2 131.5
134.5 130.9 128.4 131.0
133.9 132.3 128.2 131.3
134.5 130.9 128.5 131.0
151.7 121.1 129.4 124.2
151.4 120.4 129.6 124.3
151.8 121.1 129.4 124.2
151.7 120.4 129.5 124.1
151.8 120.7 129.4 124.0
151.7 120.4 129.5 124.2
8b
10a
10b
12a
12b
14a
14b
16a
16b
18a
18b
20a
20b
H
–
CH3
CH3
H
32.0
30.3
–
NR2 47.9
NR2 48.3
48.3[a] 42.8[b]
48.7[a] 42.4[b]
–
–
–
–
H
H
–
–
[a] N(CH2CH2Cl)2. [b] N(CH2CH2Cl)2.
Table 4. 13C NMR chemical shifts [ppm] for the cis-fused compounds at 25 °C in CDCl3 (δTMS = 0.0 ppm).
1-R2 3-R1 2-Z
4
4a
5
6
7
8
8a
R2
R1
i-C
o-C
m-C
p-C
7a
H
H
H
H
CH3
CH3
H
H
H
H
H
CH3
CH3
H
H
H
H
CH3
CH3
H
H
H
Ph
Ph
Ph
Ph
Ph
Ph
47.0
45.4
55.5
55.2
42.3
42.7
47.8
42.9
56.9
52.1
45.7
42.2
36.1
36.0
36.7
36.6
36.1
36.7
35.7
36.0
36.9
35.5
37.1
37.3
35.5
35.5
35.8
36.2
35.6
35.9
24.1
25.4
25.6
25.5
28.2
28.0
23.3
28.3
24.2
28.3
25.0
29.0
24.3
25.1
24.8
24.5
25.3
21.8
22.4
25.5
21.5
25.4
21.3
24.6
21.2
24.7
19.7
20.4
20.4
19.8
24.0
23.5
19.1
24.6
19.2
24.9
20.2
25.4
20.2
32.9
33.3
32.5
32.8
27.3
27.2
32.7
30.7
31.5
29.9
28.3
24.4
32.1
33.3
31.9
32.9
26.0
27.7
52.1
50.3
52.3
49.2
61.4
60.1
52.2
54.3
51.8
54.1
59.1
62.6
52.3
51.2
53.3
50.6
62.5
60.4
–
–
–
–
32.7
33.0
–
–
–
–
–
35.5
35.8
–
–
–
–
35.6
35.4
–
–
–
133.9 131.0 128.5 131.2
133.6 132.2 128.2 131.7
134.7 131.1 128.3 130.9
131.9 132.9 128.2 131.7
134.6 132.0 128.2 130.9
134.4 131.9 128.3 131.1
151.4 120.4 129.6 124.3
151.6 120.5 129.5 124.2
151.7 120.4 129.5 124.1
152.0 120.4 129.4 123.9
151.9 120.5 129.4 124.0
151.9 120.6 129.4 124.0
7b
9a
9b
11a
11b
13a
13b
15a
15b
17a
17b
19a
19b
21a
21b
22a
22b
OPh
OPh
OPh
OPh
OPh
OPh
NR2 46.7
NR2 46.6
NR2 53.5
NR2 56.3
NR2 41.9
NR2 43.7
H
–
CH3
CH3
H
H
H
H
CH3
CH3
30.7
33.2
–
–
–
–
32.4
32.3
48.6[a] 42.4[b]
48.3[a] 42.9[b]
49.4[a] 42.7[b]
49.5[a] 43.0[b]
49.8[a] 43.0[b]
49.5[a] 42.9[b]
–
–
–
–
–
–
–
–
–
–
–
–
H
25.0[c] 25.0[c] 20.0
–
CH3
CH3
H
26.8
25.3
28.8
27.3
22.9
25.5
21.1
22.6
22.3
19.5
25.2
22.9
35.4
35.0
–
H
–
[a] N(CH2CH2Cl)2. [b] N(CH2CH2Cl)2. [c] Two adjacent signals at δ = 25.03 and 25.05 ppm; individual assignment not obtained.
2148
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
Table 5. 31P NMR chemical shifts [ppm] and JP,C coupling constants [Hz] for the trans-fused compounds at 25 °C in CDCl3 [δP(85%
H3PO4) = 0.0 ppm].
δP
P,4
4.6
P,4a
P,7
P,8
P,8a
P,R2
P,R1
P,i-C
P,o-C
P,m-C
P,p-C
8a
8b
16.0
18.7
19.6
16.3
20.4
18.0
7.2
8.0
8.3
9.5
8.8
3.1
6.4
1.7
5.0
3.6
6.0
11.2
4.7
9.2
3.6
12.2
4.7
2.3
8.3
2.6
2.7
1.9
2.2
1.2
1.4
2.7
3.3
2.8
3.2
1.5
2.1
2.8
2.8
9.9
9.3
10.1
9.2
6.1
4.9
6.7
11.4
7.3
12.0
2.1
4.2
2.7
4.1
3.2
–
–
–
–
3.4
4.3
–
–
–
–
–
3.6
4.0
–
–
–
–
163.6
152.2
163.1
151.2
164.6
150.5
8.0
7.9
8.7
9.1
7.6
9.8
10.3
9.7
9.8
9.6
9.6
4.5
4.7
4.4
4.5
4.7
13.8
13.7
13.8
13.3
13.7
13.3
1.2
3.0
3.0
2.9
2.8
2.9
3.0
1.5
1.0
1.3
1.4
1.2
1.0
–
2.8
[a]
10a
10b
12a
12b
14a
14b
16a
16b
18a
18b
20a
20b
[a]
3.9
2.8
2.0
3.7
1.5
[a]
1.7
3.7
1.8
3.6
1.9
1.6
3.4
3.2
0.8
[a]
2.2
2.9
[a]
[a]
–
0.9
1.5
3.4
3.3
2.9
4.3
–
–
–
–
1.1
[a]
[a]
10.8
15.6
14.3
8.8
11.0
9.2
9.1
4.6
[a,c]
–
–
4.7[b]
4.2[b]
–
–
2.9[c]
–
[a] Not resolved. [b] N(CH2CH2Cl)2. [c] N(CH2CH2Cl)2.
Table 6. 31P NMR chemical shifts [ppm] and JP,C coupling constants [Hz] for the cis-fused compounds at 25 °C in CDCl3 [δP(85% H3PO4)
= 0.0 ppm].
δP
P,4
2.7
P,4a
P,8
P,8a
P,R2
P,R1
P,i-C
P,o-C
P,m-C
P,p-C
7a
19.0
17.6
16.1
21.2
15.5
16.5
8.0
7.0
8.9
8.4
10.2
8.4
14.4
16.8
14.5
19.6
14.6
15.8
5.4
3.4
3.5
1.5
4.6
7.3
3.8
6.7
2.8
5.5
6.0
6.3
6.2
3.1
4.6
1.7
3.5
7.9
8.2
8.8
7.2
9.9
1.0
2.8
2.7
4.2
3.5
–
–
–
–
5.4
4.6
–
–
–
–
–
4.1
3.1
–
–
–
152.8
162.0
153.2
162.6
162.2
161.0
7.8
7.8
9.2
8.6
9.2
10.2
9.9
9.7
9.4
10.0
9.7
4.8
4.8
4.6
4.5
4.6
4.6
13.7
13.8
13.7
13.6
13.8
2.7
2.9
3.0
2.9
2.7
7b
4.1
[a]
9a
[a]
9b
4.1
[a]
11a
11b
13a
13b
15a
15b
17a
17b
19a
19b
21a
21b
22a
22b
3.4
3.2
3.2
[a]
13.8
2.8
[a]
[a]
11.2
3.4
2.8
3.7
[a]
[a]
[a]
[a]
[a]
[a]
3.2
–
1.0
[a]
[a]
11.5
0.9
[a]
[a]
[a]
[a]
3.2
–
[a]
[a]
1.8
3.1
2.8
7.8
–
–
–
–
2.8
Ͻ 1
–
0.9
1.6
–
–
–
–
–
–
[a]
[a]
1.9
–
–
–
–
2.9
2.6
8.7
7.6
10.7
4.9
3.0
3.0
3.5
2.8
4.3[b]
5.0[b]
4.6[b]
4.6[b]
4.6[b]
4.7[b]
2.7[c]
–
–
–
–
–
–
[a][c]
2.9
[a]
[a][c]
[a][c]
[a][c]
[a][c]
[a]
11.7
[a]
2.7
1.0
[a]
[a]
4.4
–
[a] Not resolved. [b] N(CH2CH2Cl)2. [c] N(CH2CH2Cl)2.
Table 7. Selected JH,H and JP,H coupling constants [Hz] for the cis-fused compounds at 25 °C in CDCl3.
4x,4y
4x,P
4y,4a
4y,P
4a,5x
5x,6x
5y,6y
6x,7x
6y,7y
7x,8x
7y,8y
8x,P
8y,8a
8a,P
7a
–13.3
–12.2
–11.8
–11.5
–11.6
–12.1
–13.3
–13.2
–11.9
–11.8
–13.2
–13.6
–13.3
–11.9
–11.8
–11.6
–11.5
–11.5
4.2
7.7
6.4
4.4
16.0
12.9
1.7
25.2
2.2
22.7
2.9
28.0
6.3
3.2
4.3
4.5
3.2
10.2
9.4
2.2
11.0
2.4
11.0
4.3
12.3
3.7
3.5
7.6
2.6
11.7
8.2
23.1
20.3
14.7
21.7
8.3
10.9
27.9
5.9
26.1
3.6
24.6
1.6
20.2
25.2
8.2
25.5
3.5
11.7
10.7
10.1
11.5
5.0
5.0
12.5
3.5
12.4
3.6
10.5
2.3
11.2
11.3
7.8
11.9
2.8
12.2
11.3
10.9
12.2
4.5
6.1
13.3
4.3
13.1
3.9
11.0
2.6
4.3
5.0
5.5
4.5
11.5
10.7
3.0
12.4
3.1
12.5
4.8
13.8
4.8
12.1
11.3
10.7
12.2
5.1
5.6
13.2
4.6
12.9
4.2
11.0
3.2
4.0
4.6
5.2
3.6
11.2
10.2
3.5
11.5
3.4
11.9
5.1
13.3
4.0
12.8
11.8
11.3
12.4
5.4
6.2
13.9
4.1
13.6
4.5
11.7
3.0
3.9
5.2
5.3
4.3
11.3
10.6
2.4
11.8
2.9
11.9
5.6
13.2
5.0
5.7
4.6
4.9
3.8
4.2
5.1
3.6
10.3
9.7
3.3
11.1
3.2
11.5
5.1
2.6
4.7
5.6
4.2
14.0
8.3
1.5
21.9
1.5
22.9
3.4
7b
9a
9b
5.0
[a]
11a
11b
13a
13b
15a
15b
17a
17b
19a
19b
21a
21b
22a
22b
[a]
7.6
[a]
6.9
[a]
4.2
[a]
12.3
3.7
22.8
4.4
11.4
12.2
11.4
5.8
5.2
14.9
2.8
24.1
8.0
11.9
4.5
11.8
4.4
12.1
4.2
6.1
3.9
3.0
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
13.7[b]
2.5
12.6
2.9
6.9
3.6
13.5
9.4
12.6
3.2
6.9
3.4
12.9
9.1
12.9
3.6
7.3
3.7
12.9
9.4
6.6
3.3
11.9
8.5
[a]
20.6
6.4
17.4
6.3
1.5
1
[a] Not resolved. [b] Poor H spin simulation due to the presence of impurities in the sample [the b epimer and degradation product(s)].
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2149

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
Figure 1. Conformations of the trans-fused 1,3,2-diazaphosphinanes. R1, R2 = H, CH3; Z = Ph, OPh, N(CH2CH2Cl)2. For both (racemic)
epimers, the figure is sketched for the enantiomer with the (4aR,8aS) configuration.
4 and C-8a carbon atoms were deshielded by around 6– the deshielding effect of the 1,3-syn diaxial P=O bond.[18,19]
10 ppm upon methyl substitution of the adjacent ring nitro- On the other hand, the P configurations are not directly
gen atoms (N3 and N1, respectively).
evident from the relative phosphorus chemical shifts
(Table 5) within an epimeric pair.
Relative Configurations at Phosphorus
trans-Fused Compounds with Z = OPh (14,16,18a,b)
The epimers a and b differ from each other by the config-
uration at the phosphorus atom relative to the bridge-head
carbons C4a and C8a (Figure 1). The configurations were
assigned by NMR methods, the approach depending on the
type of ring-fusion (cis/trans) and the P substituent (Z).
Owing to the well-known axial preference of P–OR and
P–OAr bonds in saturated 1,3,2-diheterophosphinane 2-ox-
ides,[13,17] often attributed to the (generalized) anomeric ef-
fect, epimers a are expected to populate nonchair heteroring
conformations with a pseudoaxial phenoxy group whereas
trans-Fused Compounds with Z = Ph (8, 10, 12a,b)
epimers b should retain a double-chair conformation. The
3
The epimers with an axial Ph substituent (b) have smaller epimers b thus have larger JP,4-Heq values than epimers a
1JP,i-C values than the equatorial epimers (a) (150.5–152.2 (26.2–28.2 and 9.4–16.6 Hz, respectively, Table 1) since in
and 163.1–164.6 Hz, respectively, Table 5), a trend that has the former the P2–N3–C4–Heq torsion angle is close to 180°
been reported in the literature, for example, for P–C and P– which results in a maximal value of the corresponding
N bonds.[17] Epimers b, but not a, also display NOESY coupling constant as per the well-known Karplus relation-
cross peaks at (o-H,4-Hax) and (o-H,8a-H) confirming the ships.[17] In the a series, the 1,4-syn axial arrangement of
1,3-syn diaxial nature of 4-Hax, 8a-H and Ph. Consistently 4a-H and P-OPh in the nonchair conformations (Figure 1)
with this, in epimers a the protons 4-Hax and 8a-H resonate results in deshielding of 4a-H and shielding of C-4a with
at higher frequencies than those in epimers b as a result of respect to series b (Table 1 and Table 3).
2150
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
Figure 2. Conformations of the cis-fused 1,3,2-diazaphosphinanes. R1, R2 = H, CH3; Z = Ph, OPh, N(CH2CH2Cl)2; R = CH2CH2Cl.
For both (racemic) epimers, the figure is sketched for the enantiomer with the (4aR,8aR) configuration. C: chair heteroring; SB: (unspeci-
fied) skew-boat heteroring. The SB conformations shown are chosen for representative purposes only.
trans-Fused Compounds with Z = N(CH2CH2Cl)2 (20a,b)
dominant rotameric state of the phenyl group is similar to
that of conformation A of 10a as depicted in Figure 3).
Such an NOE was not seen for 11b. Also, the axial P=O
bond in the N-out form of 11a results in a slight deshielding
of the 4-Hy proton of epimer a with respect to that of b.
Compounds 13a, 15a and 17a (Z = OPh) have much larger
populations of the N-in conformation than the correspond-
ing b epimers (Table 8) as a result of the axial tendency of
OPh.[13,17] On the other hand, compounds 21a and 22a [Z
= N(CH2CH2Cl)2] have larger N-out populations than the
corresponding b epimers since the steric demands of an
N(CH2CH2Cl)2 substituent outweigh the stereoelectronic
ones, driving the substituent towards an equatorial posi-
tion.[12] Both 19a and 19b are predominantly N-in, the N-
The axial P=O, equatorial P–Nexo configuration of 20a
may be inferred from the observations that the 4-Hax and
8a-H signals of 20a are deshielded by around 0.3 ppm with
respect to those of 20b (Table 1) as a result of the deshield-
ing effect[18,19] of an axial P=O bond and that the |1JP,N-exo
|
value of 20a is larger than that of 20b, as expected[17] (only
1
for the former epimer was the passive JP,N-exo coupling re-
1
solved in the f1 dimension of the H{15N}-HMBC spec-
trum at 20 Hz FID resolution). Also, the selective 1D
NOESY spectrum of 20b showed that the protons 4-Hax
and 8a-H are spatially close to the N(CH2CH2Cl)2 protons.
Similar NOEs were not observed for 20a.
cis-Fused Compounds
in populations being equal within experimental error (85
7
and 87 3%, respectively). The P configuration of these
compounds was verified by selective 1D NOESY measure-
ments [in the case of 19a, 8a-H and N(CH2CH2Cl)2 protons
received a NOE from 4-Hx] and by the fact that 19b has
deshielded 4-Hx and 8a-H resonances in comparison to
those of 19a (Table 2) due to its 1,3-syn diaxial P=O.
In order to determine the relative P configuration in cis-
fused compounds, it was necessary to know the position
of the N-in/N-out conformational equilibrium (cf. Figure 2).
Fortunately, this could be determined independently (vide
infra, Table 8). With the knowledge of the dominant con-
formation it was then possible to deduce the relative P con-
figuration from NMR criteria similar to those used for
trans-fused compounds. Thus 7a and 9a (Z = Ph, mostly
Conformations
N-in) display an NOESY cross-peak at (o-H,8a-H) and have
The conformational processes present in the compounds
1
a characteristic “axial” JP,i-C value (152.8 and 153.2 Hz, studied were observed to be fast on the NMR timescale at
respectively, Table 6), whilst 7b and 9b (mostly N-in as well) 25 °C in CDCl3. Thus, only population-weighted averages
1
1
have “equatorial” values of JP,i-C (162.0 and 162.6 Hz, of the H, 13C and 31P NMR signals were observed. The
respectively) and have deshielded 4-Hx and 8a-H resonances feasible conformations arise from nitrogen-inversion, P-
due to a 1,3-syn diaxial P=O. Compounds 11a and 11b (Z substituent rotamerism and the different conformational
= Ph) are mostly N-out; the cis relationship of Ph and 4a- states of the heteroring (e.g., chair or skew-boat). Addition-
H in 11a is evident from the observed NOE between 4a-H ally, in the case of the cis-fused compounds simultaneous
and the ortho-protons of Ph (which is possible if the pre- ring-inversion of both rings resulting in an exchange be-
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2151

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
Figure 3. The optimized chair-like heteroring conformations for 10a and 10b and their relative Gibbs energies, P,4-Heq torsion angles and
3
2
Fermi contact coupling constants JP,4-Heq and J4-Hax,4-Heq, as obtained from DFT calculations. –R,R– = –(CH2)4–.
tween the N-in and N-out forms is a prominent conforma- ascribed to the N3-methyl substituent effect without taking
tional process (Figure 2). In the following, the carbocyclic nonchair conformations into consideration. Similarly, com-
part is always assumed to adopt a chair conformation. This pounds 12a and 12b (R2 = CH3) clearly exhibit smaller val-
3
assumption is confirmed for the trans-fused compounds by ues of JP,C-8 (6.1 and 4.9 Hz, respectively) than 8,10a,b
the expected values of the vicinal JH,H coupling constants in (9.2–10.1 Hz, R2 = H) due to N-methyl substitution in the
the carbocycle (characteristic “diequatorial” and “diaxial” corresponding coupling pathway. The chair heteroring con-
values) whenever these values could be extracted as well as formation for 12a,b is nevertheless evident from the other
3
for the cis-fused compounds which notably populate only relevant J values.
one of the ring-inverted forms.
The substituents at N1 and N3 can be either equatorial
or axial and the positions are interconvertible through ni-
trogen inversion. It is known that the absolute value of a
geminal 2JHCH coupling constant decreases when one of the
C–H bonds is placed antiperiplanar to an electron lone pair
on a directly bonded heteroatom.[20] For the trans-fused P-
trans-Fused Compounds with Z = Ph
From the results for analogous saturated 3,1,2-benzoxaza-
phosphinine 2-oxides,[13] we expect that the steric require-
ments of an axial P-phenyl substituent in trans-fused deriv-
atives are not strong enough to drive the heteroring into
nonchair conformations in appreciable amounts. Thus both
a and b epimers of compounds 8, 10 and 12 are assumed
to predominantly populate chair heteroring conformations.
Consistently with this, these derivatives generally show large
2
phenyl-substituted compounds, the observed J4-Hax,4-Heq
values (Table 1) range between –11.0 Hz for 10a (R1 = CH3)
and –13.2 Hz for 12b (R1 = H) which indicates that in the
former the N3 substituent is mostly equatorial and in the
latter axial (Figure 1 and Figure 3). In general, these values
indicate that the N3 substituent in trans-fused compounds
with an equatorial P=O bond (b) adopt an axial position
more readily than the other epimer (a). When R1 is hydro-
gen, its position is clearly axial in b and equatorial in a,
whereas the position of an N3-methyl group is mostly equa-
torial also in b, as expected from the larger steric size of the
methyl group.
3
3
3
3J4-Hax,4a-H, JP,4-Heq and JP,C-8 and small JP,4-Hax values
(Table 1 and Table 5) as expected for double-chair confor-
mations from Karplus-like dependencies of J on the rel-
evant torsion angle. As already noted, the JP,i-C values
(Table 5) clearly indicate that the phenyl substituent is equa-
torial in the case of epimers a and axial in the case of epi-
mers b, which is also in agreement with a predominant chair
heteroring conformation. The 3JP,4-Heq value for compound
10b (R1 = CH3) is surprisingly small (16.9 Hz) in compari-
son to that of the other compounds in this subset (20.2–
24.2 Hz). Based on the DFT coupling-constant calculations
3
1
DFT Computations for Epimers 10a and 10b
In order to gain more insight into the conformational
on the optimized chair heteroring conformations of 10a and behavior of trans-fused P-Ph-substituted compounds and
10b (vide infra), such a difference can, however, be fully how this behavior is reflected in the NMR parameters, DFT
2152
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
computations on the P epimeric pair 10a and 10b were per-
The effect of the axial or equatorial position of the CH3
2
formed as follows. In the limited conformational search substituent on the geminal J4-Hax,4-Heq coupling constant
only chair conformations of both rings were considered. As was also studied by calculating the Fermi contact contri-
the starting point, the previously optimized double-chair bution to it for representative conformations of 10a and 10b
conformations of 3-oxa-analogues (compounds 8a and 8b (Figure 3). As expected (vide supra), the absolute values in
in ref.[13]) were used. The O3 atom was replaced by N–CH3, the case of the equatorial CH3 are smaller (a: –9.9 Hz, b:
and all the possible combinations of axial and equatorial –10.6 Hz) than those of the axial CH3 (a: –13.2 Hz, b:
2
N1 and N3 substituents with two alignments for the phenyl –13.6 Hz). This, in part, validates the use of J4-Hax,4-Heq as
ring [τ(O,P,i-C,o-C) = 0 or 90°] were constructed, yielding an indicator of the stereoposition of the N3 substituent.
a total of 2×2×2 = 8 starting structures for both 10a and
The P–Cipso bond in conformations F, G, H and J (Fig-
10b. These structures were then subjected to geometry opti- ure 3) is longer than that in A–E by around 0.01 Å (0.02 Å
mization at the B3LYP level of theory using a locally dense in the case of G). This is as expected for a hyperconjugative
basis set (cf. Exp. Sect.) which resulted in five unique opti- n(N) Ǟ σ*(P–C) interaction in the former structures in
mized geometries for both epimers (Figure 3, 10aA–E, which at least one of the nitrogen lone pairs is antiper-
10bF–J). The subsequent vibrational analyses showed that iplanar to the P–C bond. The lengthening is in agreement
1
each of these structures corresponds to a true local mini- with the smaller observed JP,i-C value for 10b than for 10a.
3
mum in electronic energy. The coupling constants JP,4-Heq
,
trans-Fused Compounds with Z = OPh
3JP,4-Hax and in some cases also 2J4-Hax,4-Heq were then com-
puted at the spin-unrestricted UB3LYP/cc-pVTZ level of
theory by the FPT method (cf. Exp. Sect.). Some of the
results are summarized in Figure 3 and the Cartesian coor-
dinates and energies of the optimized geometries are pro-
vided in the Supporting information.
Owing to the favored (pseudo)axial P–OPh, (pseudo)
equatorial P=O arrangement of phosphorus substituents
(vide supra), the P-OPh-substituted epimers b are expected
to retain chair heteroring conformations. This is readily
3
confirmed, for example, by the large observed JP,4-Heq and
3
3JP,C-8 and small JP,4-Hax values for 14b, 16b and 18b
The calculated relative free energies for the different con-
formations of 10a suggest that, in this case, the equatorial
position of the N3-methyl group is much more favorable
than the axial one. The axial methyl substituent is likely to
interact sterically with the hydrogen atom at C4a and also
forces the phenyl ring to assume a sterically more-hindered
rotameric state (possibly with reduced conjugation with the
P=O bond). On the other hand, both N3-methyl stereoposi-
tions in 10b are of comparable energy. For this epimer, a
destabilizing gauche interaction between the methyl and
phenyl substituents is removed when the N3-methyl changes
from an equatorial to an axial position. A stabilizing di-
pole–dipole interaction between the N3–CH3 and P=O
bonds is also possible for both methyl positions in contrast
to 10a in which such an interaction should prefer only the
CH3-equatorial structure. The N3 geometry in the “CH3-
equatorial” structures of 10b is actually almost planar
whereas in those of 10a it is clearly tetrahedral (the sum of
the bond angles is 356.7° for the lowest energy structure of
10b and 343.2° for that of 10a). These N3 geometries are
likely favored by both the repulsive CH3/Ph gauche interac-
tion and the stabilizing N3–CH3/P=O dipole–dipole inter-
action. The near planar N3 geometry in the lowest energy
structures of 10b is concomitant with the flattening of the
heteroring and the distortion of the P,4-Heq torsion angle
away from the ideal antiperiplanar value. This is reflected
(Table 1 and Table 5). The J values in the corresponding a
series, in contrast, clearly indicate that the axial preference
of OPh forces the heteroring of these epimers to populate
nonchair conformations. By taking into consideration the
constraints imposed by the trans-fused carbocycle and the
requirement that OPh be pseudoaxial, as well as the fact
that 4-Hax should be trans-diaxial with 4a-H in any notably
3
populated conformation (the observed J4-Hax,4a-H values
are similar to those of the b series), we conclude that the
viable boat/skew heteroring conformations in the a series
are B2,4a and 1S4a (Figure 1). In the B2,4a conformation, the
3
3JP,4-Hax and JP,4-Heq values should be nearly equal owing
to their comparable torsion angles. This value is estimated
to be around 9 Hz from an approximate Karplus-type equa-
3
tion [Eq. (1); parametrized to reproduce the observed JP,4-
3
and JP,4-Heq values for 14b at τ = (–)60 and 180°]. An
Hax
inspection of the torsion angles in the idealized B2,4a and
1S4a conformations suggests that the change from B2,4a to
3
1S4a should be accompanied by an increase in JP,4-Heq and
3JP,8a-H and a decrease in 3JP,4-Hax and 3JP,C-8. The observed
values of these coupling constants (Table 1 and Table 5)
thus show that the predominant heteroring conformation
of 14a and 16a is close to B2,4a, whilst that of 18a resembles
1S4a. In our previous DFT computational and NMR study
of 3-oxa-analogues[13] we found a similar shift from B2,4a
1
3
towards S4a conformations upon increasing the steric size
in the calculated values of JP,4-Heq, as expected from the
3
3
of the N1 substituent. The JP,4-Hax and JP,4-Heq values of
18a (5.0 and 16.6 Hz, respectively) alone could alternatively
mean that considerable amounts of a chair conformation
is present in this case, but this possibility is excluded by
Karplus relationships, resulting in a population-weighted
value of 18.4 Hz for 10b as opposed to 23.3 Hz for 10a (at
298.15 K, assuming a Boltzmann distribution). This is in
good qualitative agreement with the observed difference in
this coupling constant between these epimers (Table 1) and
confirms that nonchair conformations need not be consid-
ered in order to explain the “small” observed value of
3JP,4-Heq for 10b.
3
3
comparison of the values of JP,8a-H and JP,C-8 with those
of 18b (which is a double-chair conformation).
3JPNCH [Hz] = 20.5×cos2 (τ) – 7.6×cos(τ)
(1)
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2153

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
trans-Fused Compounds with Z = N(CH2CH2Cl)2 (20a,b)
ling constants are very sensitive to the position of the N-in/
N-out equilibrium as a result of the large change in the di-
hedral angle between the coupled protons upon ring-inver-
sion, but their values in the limiting N-in and N-out forms
are expected to be insensitive to heteroring substitution and
conformation. The 13C chemical-shift difference δC-6 – δC-7
is positive in N-in and negative in N-out due to γ-gauche
shielding[22] of C-7 in N-in and C-6 in N-out. The values of
PI(N-in) and PI(N-out) were obtained from compounds 13a
and 17b, respectively. These compounds display extreme
values for the parameters PI within the cis-fused series and
The P-N(CH2CH2Cl)2 substituent in similar compounds
is known to prefer an equatorial position for steric
reasons.[12,21] Consistent with this, the heteroring conforma-
tion of the epimer 20a can be readily verified as a chair
from the J-coupling-constant criteria similar to those used
above. As in the a series of P-phenyl-substituted com-
pounds, its N3 hydrogen atom is predominantly equatorial
(2J4-Hax,4-Heq = –11.4 Hz). The epimer 20b is also predomi-
nantly in a double-chair conformation, but displays slightly
3
3
3
smaller JP,4-Heq and JP,C-8 and slightly larger JP,4-Hax val-
ues than 20a. This could indicate a small contribution from
nonchair heteroring conformations or simply reflect some
distortion of torsion angles in the chair conformation in
response to the sterically demanding axial P substituent. In
compound 20b the N3 hydrogen atom preferentially adopts
an axial position (2J4-Hax,4-Heq = –13.0 Hz), as is generally
observed in this study for compounds with an equatorial
P=O bond.
3
3
their JH,H values are also similar to the JHax,Hax and
3JHeq,Heq values of the trans-fused compounds.
The x(N-in) values (Table 8) obtained demonstrate re-
markable effects of the P and N substituents and the P con-
figuration on the N-in/N-out equilibrium. The substitution
of the N1 hydrogen atom by a methyl group is accompanied
by a considerable shift towards the N-out form when the
other substituents and the P configuration are kept con-
stant (e.g., 7a Ǟ 11a, 7b Ǟ 11b, 13a Ǟ 17a). The increase
in the steric size of the N1 substituent thus drives the N1
atom into an equatorial position with respect to the fused
carbocycle, as previously observed for the 3-oxa-ana-
logues.[13] The effect of P substitution depends both on the
type of substituent and the P configuration. The axial pref-
erence of a P-phenoxy group is evidently the strongest sin-
gle factor affecting the N-in/N-out equilibrium within the
studied range of N1, P2 and N3 substituents, as seen by the
preference of the N-in form in the a epimers and the N-out
form in the b epimers of compounds 13, 15 and 17. Com-
bined with the effect from the N1 substituent, this allows
the above use of 13a and 17b as model compounds for the
N-in and N-out forms to be rationalized. In contrast, P-
phenyl and -N(CH2CH2Cl)2 substituents prefer an equato-
rial position for steric reasons, as already mentioned. Thus,
the N-in population in the b epimers of 9, 11, 21 and 22 is
larger than in the respective a epimers. By comparing the
x(N-in) values within, for example, the 9a/21a or 11a/22a
pair [the same N1 and N3 substitution, but the first mem-
ber has phenyl and the second N(CH2CH2Cl)2 as the P sub-
stituent], it is obvious that the phenyl group is sterically
less demanding than N(CH2CH2Cl)2 in these compounds.
Finally, the effect of the N3 substituent seems to comprise
both steric and stereoelectronic factors. When the N3 sub-
stituent is hydrogen, the ring-inverted form with an equato-
rial P=O bond is stabilized as a result of a dipole–dipole
interaction (or even an unconventional hydrogen bond) be-
tween N–H and P=O. These bonds can become more paral-
lel through an increase in the planarity of the N3 nitrogen
atom if P=O is equatorial. Thus, in the a series, the replace-
ment of the N3-methyl by a hydrogen atom (9a Ǟ 7a, 21a
cis-Fused Compounds
In the cis-fused compounds, interconversion between the
N-in and N-out forms (Figure 2) is possible. The N1 atom
is axial with respect to the fused carbocycle in the former,
and equatorial in the latter. The observed NMR parameters
are population-weighted averages of the parameters of the
limiting forms, which can be expressed by Equation (2),
where P can be, for example, a chemical shift or a J coup-
ling constant, or a linear combination thereof, and x(N-in)
and x(N-out) are the fractional populations of the limiting
forms. The parameters P(N-in) and P(N-out) can them-
selves be averages of several heteroring conformations, for
example. We will use Equation (2) in two ways. First, we
will select a set of NMR parameters PI for which PI(obs)
is sensitive (only) to the position of the N-in/N-out equilib-
rium and for which both PI(N-in) and PI(N-out) can be reli-
ably estimated. We will then use these in Equation (2) to
yield the fractional population of the N-in form for each
compound. Secondly, we consider certain parameters PII
for which PII(N-in) and PII(N-out) are sensitive to the heter-
oring conformation of the corresponding form and calcu-
late the value of PII in one of the limiting forms by estimat-
ing the value in the other form.
P(obs) = x(N-in)×P(N-in) + x(N-out)×P(N-out)
(2)
Position of the N-in/N-out Equilibrium in the cis-Fused
Compounds
The x(N-in) values for the cis-fused compounds were cal- Ǟ 19a) results in an increased population of the N-in form,
culated from Equation (2) by using the following 10 NMR but a decreased one in the b series (9b Ǟ 7b, 21b Ǟ 19b).
3
parameters for PI: JHx,Hx coupling constants in the car- Also, compound 7a (Z = Ph, R1 = H) displays a larger
3
bocycle (three different, cf. Figure 2 and Table 7), JHy,Hy population of the N-in form (0.91) than 7b (0.82) despite
3
(three different), J4-Hy,4a-H
, ,
3J4a-H,5-Hx 3J8-Hy,8a-H and the (weak) equatorial preference of Ph. In the case of P-
δC-6 – δC-7 (Table 4). The mean values of the 10 calculations phenoxy substitution, however, an increase in the steric size
for each compound are given in Table 8. The selected coup- of the N3 substituent seems to favor the N-out form regard-
2154
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
Table 8. The N-in/N-out equilibrium for the cis-fused compounds at 25 °C in CDCl3. N-in mol fractions x(N-in) and predominant heteror-
ing conformations (chair or skew-boat) in the limiting forms with some relevant coupling constants [Hz].
x(N-in)[a]
PII(obs)
PII(N-in)
3JP,4-Hx
PII(N-out)
3JP,4-Hx
3
1
[b]
[c]
3JP,4-Hx JP,4-Hy JP,C
3JP,4-Hy
1JP,C
3JP,4-Hy
1JP,C
Z = Ph
7a
9a
11a
7b
9b
0.91 0.03
0.77 0.03
0.21 0.03
0.82 0.05
0.90 0.05
0.28 0.03
4.2
6.4
16.0
7.7
4.4
23.1
14.7
8.3
20.3
21.7
10.9
152.8
153.2
162.2
162.0
162.6
161.0
C
2.5 1.3
25.0 1.4
151.7 1.1
C*
C*
21.5
20.2
3.9
3.1
163.6
163.1
C
2.3 1.8
18.1 1.7
150.2 1.6
[d]
[d]
[d]
[d]
C*
23.5
3.2
164.6
[d]
[d]
[d]
[d]
C*
C*
C*
3.9
3.1
3.2
21.5
20.2
23.5
163.6
163.1
164.6
[d]
[d]
[d]
[d]
11b
12.9
SB
16.8 1.8
6.0 2.0
159.6 1.5
Z = OPh
[f]
[f]
[f]
[f]
[f]
[f]
13a[e]
15a
17a
1.00
1.7
2.2
2.9
25.2
22.7
28.0
27.9
26.1
24.6
5.9
3.6
1.6
–
–
–
–
–
–
C*
C*
1.3
1.4
28.1
26.2
–
–
–
–
–
–
–
–
–
–
–
–
0.99 0.02
0.79 0.04
0.13 0.03
0.11 0.02
0.00
C*
1.3
28.2
SB
C*
C*
C*
8.9 9.5
28.1
26.2
11.2 11.3
[d]
[d]
[d]
13b
1.3
1.4
1.3
[d]
[f]
[d]
[f]
[d]
[f]
15b
17b[e]
28.2
Z = NR2
19a
21a[g]
22a
19b
21b
22b
0.85 0.07
0.49 0.04
0.03 0.02
0.87 0.03
0.95 0.04
0.39 0.02
6.3
20.2
8.2
3.5
25.2
25.4
17.4
–
–
–
–
–
–
C
2.8 2.8
23.2 2.5
–
–
–
–
–
–
C*
C*
26.1
2.7
–
–
–
–
–
–
14.9
24.1
5.2
2.8
8.0
SB
5.2 4.1
13.9 3.5
26.1[h]
2.7[h]
[d]
[d]
[d]
C*
26.1
2.7
[d]
[d]
[d]
C*
C*
C*
2.7
2.7[h]
2.7
26.1
26.1[h]
26.1
[d]
[d]
[d]
SB
11.4 2.1
11.8 2.3
[a] Arithmetic mean standard deviation of 10 calculations of x(N-in) using Equation (2) with the parameter set PI (see text). [b] Predomi-
nant heteroring conformation in the N-in form. C: chair, SB: skew-boat (unspecified). An asterisk indicates an assumed chair conforma-
tion and its PII values are taken from the corresponding trans derivative (see text). Values for nonasterisked conformations are derived
from Equation (2) [errors are from a total differential; assuming ∆P(obs) 0.5 Hz and ∆P(C*) 1.5 Hz]. [c] Predominant heteroring
conformation in the N-out form. [d] Uncertain owing to the small population of the corresponding form. [e] Reference compounds for
the PI(N-in) and PI(N-out) values in Equation (2). [f] Zero population within experimental error. [g] x(N-in) calculated from J4y,4a, J4a,5x
and (δC-6–δC-7) only. [h] Values from 20a are used.
less of the P configuration: in these compounds the axial tion of the corresponding form was less than around 20%,
preference of OPh likely forces the P=O bond to be equato- some general conclusions can be drawn. The results for the
rial in both the N-in and N-out forms (vide infra).
P-OPh derivatives are consistent with the assumption that
conformations with a (pseudo)axial phenoxy group are al-
most exclusively populated. That is, the epimers a of 13, 15
and 17 should adopt a double-chair conformation in the N-
in form and a skew-boat heteroring conformation(s) in the
N-out form, and vice versa for the epimers b (Figure 2).
The presence of nonchair conformations could, indeed, be
verified for the N-out form of 17a (Table 8), whilst the N-
in/N-out equilibrium of the other phenoxy-substituted de-
rivatives was too strongly biased for the verification of non-
chair conformations. Nonchair heteroring conformations
for derivatives with a “steric” P substituent (Ph, NR2) are
reasonable only for the N-in-a and N-out-b structures. In
the former, the steric environment of the P substituent is
similar to that of the trans-fused b epimers (two 1,3-syn
diaxial hydrogen atoms) and, by analogy, it is no surprise
that N-in-7a, -9a and -19a are observed to adopt double-
chair conformations with an axial P substituent (Table 8,
Figure 2). Remarkably, skew-boat conformations seem to
contribute to the N-in-21a form (R2 = H, R1 = CH3, Z =
NR2). This can be explained by the loss of one stabilizing
N–H/P=O dipole–dipole interaction in the double-chair
conformation in comparison to N-in-19a (R1 = R2 = H). In
the N-out-b structures with a steric P substituent the prefer-
ence for nonchair conformations should be greater than
Heteroring Conformations in the Limiting N-in and N-out
Forms
We assume that the heteroring in the limiting forms is
chair whenever it results in the adoption of the preferred
position for the P substituent [axial for OPh, equatorial for
Ph and N(CH2CH2Cl)2]. Thus, with the P-OPh derivatives,
the N-in form of the a epimers and the N-out form of the b
epimers are expected to adopt a double-chair conformation
and vice versa for the “steric” P substituents. This assump-
tion is validated, for example, by the fact that the observed
3JP,4-Hx and 3JP,4-Hy values are consistent with a chair struc-
ture for compounds in which the assumed double-chair
structure is almost exclusively populated (13a, 17b etc.).
To analyse the “nonassumed” heteroring conformations,
two to three NMR PII parameters were chosen: JP,4-Hx
,
3
3JP,4-Hy and, for the P-phenyl-substituted derivatives,
1JP,i-C. The values for the assumed double-chair structures
were obtained from the corresponding trans-fused com-
pounds (with similar substitution and stereoposition of the
P substituent), from which the values for the other ring-
inverted form were calculated using Equation (2). The re-
sults are summarised in Table 8. Although this calculation
gave unacceptable error limits in cases in which the popula-
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2155

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
until a good match with the observed spectrum was obtained, fol-
lowed by total line-shape fitting (T mode) to yield the reported δH,
JH,H and JP,H values. RMS errors of the iteratively solved 1H NMR
spectra together with an example of a simulated vs. observed spec-
trum may be found in the Supporting information.
that in the trans-fused b epimers due to the presence of a
1,3-syn diaxial methylene group, C(8)H2, in a chair confor-
mation. Consistent with this, a notable contribution from
nonchair conformations to N-out-11b and -22b is feasible
based on the PII(N-out) values in Table 8. For the former,
this is most clearly seen from the large JP,C(N-out) value
(159.6 Hz) which implies that the P–C bond is predomi-
nantly pseudoequatorial.
1
Computations: Computations were performed with the Gaussian 98
electronic structure program.[25] Geometry optimizations and sub-
sequent vibrational analyses (1 bar, 298.15 K, scaling factor 0.9804)
[26]
were performed at the B3LYP level of theory[27–29] by using a
locally dense basis set: the –(CH2)4– part of the molecule was de-
fined by the basis set 3-21G, heteroatoms by 6-31+G(d,p) and all
the other atoms by 6-31G(d,p). Thermal free-energy values and the
number of imaginary frequencies for each optimized structure were
obtained from the vibrational analyses (no imaginary frequencies
were found, confirming that the geometries obtained corresponded
to energy minima). The Fermi contact (FC) contribution to J-
coupling constants was calculated by the finite perturbation theory
(FPT) method,[30,31] at the spin-unrestricted UB3LYP/cc-pVTZ
level, by applying a perturbation of 0.01 au to the phosphorus atom
(3JP,H) or to one of the 4-H protons (2JH,H) with the FIELD key-
word of Gaussian 98. Other contributions to the J coupling con-
Conclusions
The preferred conformations of the prepared saturated
1,3,2-benzodiazaphosphinine 2-oxides display a clear de-
pendence on the nature of the N and P substituents and the
P configuration. The P-phenoxy substituent has a strong
stereoelectronic axial preference resulting in population of
skew-boat heteroring conformations whenever the axial po-
sition cannot be obtained in a chair structure. The N-in/N-
out equilibrium in the cis-fused derivatives is also strongly
biased towards the form in which a double-chair conforma- stants were omitted in this study. The calculated JP,H coupling con-
stants were scaled by using the calibration equation of Tähtinen et
al.[31] (Figure 3 in ref.[31]). Both in the geometry optimizations and
the FC calculations, tight SCF convergence criteria were used.
tion and an axial position for OPh can be simultaneously
attained. The P-phenyl and -bis(2-chloroethyl)amino sub-
stituents generally display a steric equatorial preference, the
latter more so than the former, as is evident from the posi-
tion of the N-in/N-out equilibrium in the cis-fused deriva-
tives. Nevertheless, the trans-fused derivatives studied here
retain chair conformations in the case of axial Ph or
N(CH2CH2Cl)2 substitution. Methyl substitution of the N1
atom in the cis-fused series drives the N-in/N-out equilib-
rium towards the N-out form and in the trans-fused P-OPh-
substituted a series it drives the predominant heteroring
General Procedures: Melting points were recorded on a Kofler hot
plate microscope apparatus and are uncorrected. Elemental analy-
ses were performed with a Perkin–Elmer 2400 CHNS elemental
analyzer. Mass spectra were recorded on a Finnigan MAT 95S and
a ZabSpecETOF instrument using electron-impact ionization.
Chemicals were generally of the highest purity. Silica gel 60 (0.040–
0.063 mm) was used for column chromatography. Merck Kieselgel
60F254 plates were used for TLC.
Carboxamides 1a, 1b, 2a and 2b[15] and the diamines 4a and 4b[16]
were prepared according to literature procedures.
1
conformation from B2,4a towards S4a. A N3-hydrogen sub-
stituent seemingly has a stabilizing effect on conformations
with an equatorial P=O bond.
cis- and trans-2-(Aminomethyl)cyclohexylamine (3a and 3b): cis- or
trans-2-Aminocyclohexanecarboxamide (1a or 1b) (7.11 g,
50 mmol), respectively, was added in small portions to a stirred
and cooled suspension of LiAlH4 (11.39 g, 300 mmol) in dry THF
(400 mL). The mixture was stirred and refluxed for 12 h and then
cooled and the excess LiAlH4 was decomposed by the addition of
a mixture of water (22.8 mL) and THF (50 mL). The inorganic
salts were filtered off and washed with EtOAc (3×100 mL). The
combined organic filtrate and washings were dried (Na2SO4) and
evaporated under reduced pressure to give an oily product. The
crude diamines were purified by distillation.
Experimental Section
NMR Spectroscopy: NMR spectra were recorded with JEOL JNM-
LA400 and JNM-A500 and Bruker Avance 400 and 500 NMR
spectrometers at 25 °C using CDCl3 as solvent. The following
1
NMR spectra were acquired for each compound: H, 13C with 1H-
decoupling , 31P with 1H-decoupling, dqf-COSY and 1H{13C}-
1
HSQC or 13C{1H}-HETCOR. In addition, H NMR with 31P-de-
coupling, NOE difference, 1D and/or 2D NOESY, 1H{15N}-
HMBC, 1H{13C}-HMBC and/or 13C{1H}-COLOC experiments
were acquired when deemed necessary. Typical acquisition and pro-
cessing parameters are provided in the Supporting Information.
The 1H and 13C chemical shifts are referenced to internal tet-
ramethylsilane (δTMS = 0.00 ppm), and the 31P shifts are referenced
to the 1H resonance of internal TMS according to the unified shift
scale[23] recommended by IUPAC (secondary reference 85%
H3PO4: ΞP = 40.480742, δP = 0.0 ppm). 1H chemical shifts and
JH,H and JP,H coupling constants were extracted by an iterative
analysis of the 1H NMR spectra using the PERCH NMR soft-
ware.[24] The initial trial parameters for the iterative analysis were
3a: Colorless oil; yield: 3.27 g (51%); b.p. 87–88 °C (13 Torr). The
1H NMR spectroscopic data of the product correspond to the lit-
erature[32] data of the enantiomerically pure isomer.
3b: Colorless oil; yield: 3.53 g (55%); b.p. 67–69 °C (14 Torr). The
1H NMR spectroscopic data of the product correspond to the lit-
erature[32] data of the enantiomerically pure isomer.
cis-2-(Ethoxycarbonylamino)cyclohexanecarboxamide (5a): Ethyl
chloroformate (6.51 g, 60 mmol) was added to a stirred mixture
of carboxamide 1a (7.11 g, 50 mmol), toluene (150 mL), NaHCO3
(6.30 g, 75 mmol) and H2O (150 mL) and the mixture was stirred
at room temperature for 1 h. The organic layer was separated and
the aqueous layer was extracted with EtOAc (3×75 mL). The com-
bined extracts were dried (Na2SO4) and evaporated to yield a crys-
talline product.
1
obtained, as far as possible, manually from the H NMR spectra
and were complemented by use of “chemical intuition” and results
from earlier successful analyses. These trial parameters were refined
by using the software in integral-transform-fitting mode (D mode)
2156
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
5a: White crystals; yield: 8.04 g (75%). Compound 5a was used in
the next step without further purification.
9a: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 64.04, H 7.79, N 10.86. MS: m/z = 264 [M]+.
trans-2-(Ethoxycarbonylamino)cyclohexanecarboxamide
(5b):
9b: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.25, H 8.20, N 10.79. MS: m/z = 264 [M]+.
NaHCO3 (5.04 g, 60 mmol), H2O (80 mL) and ethyl chloroformate
(5.21 g, 48 mmol) were added to a stirred mixture of carboxamide
1b (5.69 g, 40 mmol) and toluene (80 mL) and the mixture was
stirred at room temperature for 1 h. The precipitated white crystals
were filtered off and washed with H2O and Et2O to yield a crystal-
line product.
rac-(2R,4aR,8aS)-
and
rac-(2S,4aR,8aS)-3-Methyl-2-phenyl-
1,2,3,4,4a,5,6,7,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Oxide
(10a and 10b)
10a: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.98, H 8.20, N 10.97. MS: m/z = 264 [M]+.
5b: White crystals; yield: 8.40 g (98%). Compound 5b was used in
the next step without further purification.
10b: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.89, H 7.83, N 11.01. MS: m/z = 264 [M]+.
cis- and trans-2-(Methylamino)cyclohexylmethylamine (6a and 6b):
The corresponding carboxamide (5a or 5b) (7.93 g, 37 mmol) was
added in small portions to a stirred and cooled suspension of Li-
AlH4 (4.21 g, 111 mmol) in dry THF (200 mL). The mixture was
stirred and refluxed for 3 h and then cooled and the excess LiAlH4
was decomposed by the addition of a mixture of water (8.4 mL)
and THF (40 mL). The inorganic salts were filtered off and washed
with EtOAc (3×75 mL). The combined organic filtrate and wash-
ings were dried (Na2SO4) and evaporated under reduced pressure
to give an oily product. The crude diamines were purified by distil-
lation.
rac-(2S,4aR,8aR)-
and
rac-(2R,4aR,8aR)-1-Methyl-2-phenyl-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (11a and 11b)
11a: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.47, H 8.24, N 10.44. MS: m/z = 264 [M]+.
11b: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.38, H 7.75, N 10.47. MS: m/z = 264 [M]+.
rac-(2S,4aR,8aS)-
and
rac-(2R,4aR,8aS)-1-Methyl-2-phenyl-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (12a and 12b)
6a: Colorless oil; yield: 2.26 g (43%); b.p. 77–78 °C (15 Torr). The
1H NMR spectroscopic data of the product correspond to the lit-
erature[33] data.
12a: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.89, H 8.27, N 10.42. MS: m/z = 264 [M]+.
6b: Colorless oil; yield: 1.89 g (36%); b.p. 70–71 °C (14 Torr). The
1H NMR spectroscopic data of the product correspond to the lit-
erature[34] data.
12b: C14H21N2OP (264.30): calcd. C 63.62, H 8.01, N 10.60; found
C 63.43, H 7.72, N 11.10. MS: m/z = 264 [M]+.
rac-(2R,4aR,8aR)-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (13a and 13b)
and
rac-(2S,4aR,8aR)-2-Phenoxy-
General Method for Ring-Closure Reactions: A solution of the ap-
propriate phosphorus-containing reagent [phenylphosphonic di-
chloride, phenyl dichlorophosphate or bis(2-chloroethyl)phos-
phoramidic dichloride, 1 equiv.] in dry THF (50 mL) was added
dropwise to a stirred solution of the appropriate diamine (3a,b, 4a,b
and 6a,b, 7 mmol) and triethylamine (2 equiv.) in dry THF
(100 mL) at room temperature. The reaction mixture was stirred for
48 h at room temperature and then filtered to remove triethylamine
hydrochloride. The filtrate was evaporated to dryness. From the
crude product, epimers a and b were obtained in pure form as
13a: C13H19N2O2P (266.28): calcd. C 58.64, H 7.19, N 10.52; found
C 58.50, H 7.38, N 10.23. MS: m/z = 266 [M]+.
13b: C13H19N2O2P (266.28): calcd. C 58.64, H 7.19, N 10.52; found
C 58.43, H 6.93, N 10.76. MS: m/z = 266 [M]+.
rac-(2R,4aR,8aS)-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (14a and 14b)
and
rac-(2S,4aR,8aS)-2-Phenoxy-
1
described in Table 9. H, 13C and 31P NMR spectroscopic data of
products 7–22a,b are given in Tables 1, 2, 3, 4, 5, 6, and 7 and in
the Supporting Information.
14a: C13H19N2O2P (266.28): calcd. C 58.64, H 7.19, N 10.52; found
C 58.86, H 7.38, N 10.43. MS: m/z = 266 [M]+.
rac-(2S,4aR,8aR)-
and
rac-(2R,4aR,8aR)-2-Phenyl-
14b: C13H19N2O2P (266.28): calcd. C 58.64, H 7.19, N 10.52; found
C 58.79, H 6.98, N 10.40. MS: m/z = 266 [M]+.
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ide (7a and 7b)
rac-(2S,4aR,8aR)- and rac-(2R,4aR,8aR)-3-Methyl-2-phenoxy-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (15a and 15b)
7a: C13H19N2OP (250.28): calcd. C 62.39, H 7.65, N 11.19; found
C 62.07, H 7.36, N 10.82. MS: m/z = 250 [M]+.
7b: C13H19N2OP (250.28): calcd. C 62.39, H 7.65, N 11.19; found
C 62.12, H 7.74, N 11.32. MS: m/z = 250 [M]+.
15a: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 59.68, H 7.40, N 9.91. MS: m/z = 280 [M]+.
rac-(2S,4aR,8aS)-
and
rac-(2R,4aR,8aS)-2-Phenyl-
15b: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 59.74, H 7.68, N 10.15. MS: m/z = 280 [M]+.
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ide (8a and 8b)
rac-(2S,4aR,8aS)- and rac-(2R,4aR,8aS)-3-Methyl-2-phenoxy-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (16a and 16b):
8a: C13H19N2OP (250.28): calcd. C 62.39, H 7.65, N 11.19; found
C 62.51, H 7.40, N 10.79. MS: m/z = 250 [M]+.
8b: C13H19N2OP (250.28): calcd. C 62.39, H 7.65, N 11.19; found
C 62.19, H 7.57, N 10.64. MS: m/z = 250 [M]+.
16a: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 59.71, H 7.69, N 9.72. MS: m/z = 280 [M]+.
rac-(2R,4aR,8aR)-
and
rac-(2S,4aR,8aR)-3-Methyl-2-phenyl-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (9a and 9b)
16b: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 59.78, H 7.33, N 10.12. MS: m/z = 280 [M]+.
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2157

Z. Zalán, H. Kivelä, L. Lázár, F. Fülöp, K. Pihlaja
FULL PAPER
Table 9. Purification of the crude products 7–22 of the ring-closure reactions.[a]
a:b[b]
A
B
C
Yield [g] (%)
D[c]
M.p. [°C]
[d]
7
8:2
a
7a
n-hexane
n-hexane
Et2O
Et2O
Et2O
0.19 (11)
0.14 (8)
–
–
–
155–157
187–190
180–184
174–176
87–89
7b
8
67:33
68:32
36:64
13:87
55:45
52:48
50:50
55:45
49:51
50:50
50:50
62:38
9:1
b
a
b
a
b
a
b
a
b
a
b
a
8a
0.18 (10)
0.33 (19)
0.48 (26)
0.22 (12)
0.19 (10)
0.39 (21)
0.22 (12)
0.19 (10)
0.28 (15)
0.19 (10)
0.41 (22)
0.41 (22)
0.41 (22)
0.47 (25)
0.57 (29)
0.65 (33)
0.63 (32)
0.65 (33)
0.71 (36)
0.61 (31)
0.57 (29)
0.29 (15)
0.64 (29)
0.57 (26)
–
8b
EtOAc
iPr2O
mixt.
–
–
–
–
–
9
15:1
9:1
9a
9b
Et2O
172–174
157–160
194–196
146–149
155–159
175–178
162–165
136–138
129–131
155–158
134–135
114–115
109–111
132–134
138–139
102–103
123–124
98–100
10
11
12
13
14
15
16
17
18
19
10a[e]
10b
11a
11b
12a
12b
13a
13b
14a
14b
15a
15b
16a
16b
17a
17b
18a
18b
19a
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
Et2O
10:1
9:1
–
10:1
9:1
mixt.
mixt.
EtOAc
EtOAc
iPr2O
mixt.
mixt.
mixt.
mixt.
mixt.
EtOAc
–
–
–
–
–
–
–
–
Et2O
Et2O
20:1
20:1
20:1
15:1
9:1
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
Et2O
n-hexane
n-hexane
–
n-hexane
n-hexane
Et2O
170–174
139–140.5
133–136
–
19b
[d]
[f]
20
21
7:1
7:3
–
a
43:57
21a
21b
22a
22b
0.34 (15)
0.39 (17)
0.44 (19)
0.55 (24)
foam
153–156
119–121
129–132
22
50:50
9:1
a
Et2O
[a] Crude products containing both P epimers (with ratio a:b) were purified by column chromatography using ethyl acetate/methanol as
eluent (solvent ratio = A). The more mobile (B) and less mobile diastereomers were crystallized by evaporation from their respective
fractions and were filtered from solvent C to yield the diastereomers in pure form. For recrystallization, if necessary, solvent D was used.
1
[b] As determined by H NMR spectroscopy (31P NMR for 9). [c] mixt. = diisopropyl ether/ethyl acetate (3:1). [d] Not obtained owing
1
to the presence of side-products and/or poor solubility in CDCl3. [e] Contained 10b as a minor impurity (ca. 4% by H NMR). [f] Ob-
tained only as a 76:24 mixture (a:b) of epimers after column chromatography.
rac-(2R,4aR,8aR)- and rac-(2S,4aR,8aR)-1-Methyl-2-phenoxy-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (17a and 17b)
rac-(2S,4aR,8aS)-
and
rac-(2R,4aR,8aS)-2-[Bis(2-chloroethyl)
amino]-1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine
2-Oxides (20a and 20b):[35] This compound was obtained as a 76:24
mixture of 20a and 20b (by 1H NMR) of the epimers.
C11H22Cl2N3OP (314.19): calcd. C 42.05, H 7.06, N 13.37; found
C 42.41, H 6.79, N 13.53. MS: m/z = 313 [M]+.
17a: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 59.69, H 7.68, N 9.78. MS: m/z = 280 [M]+.
17b: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 60.17, H 7.38, N 10.08. MS: m/z = 280 [M]+.
rac-(2R,4aR,8aR)- and rac-(2S,4aR,8aR)-2-[Bis(2-chloroethyl)
amino]-3-methyl-1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiaza-
phosphinine 2-Oxides (21a and 21b)[35]
rac-(2R,4aR,8aS)- and rac-(2S,4aR,8aS)-1-Methyl-2-phenoxy-
1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine 2-Ox-
ides (18a and 18b)
21a: C12H24Cl2N3OP (328.22): calcd. C 43.91, H 7.37, N 12.80;
found C 44.11, H 7.57, N 13.06. MS: m/z = 327 [M]+.
18a: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 60.24, H 7.32, N 9.67. MS: m/z = 280 [M]+.
21b: C12H24Cl2N3OP (328.22): calcd. C 43.91, H 7.37, N 12.80;
found C 43.74, H 7.22, N 12.99. MS: m/z = 327 [M]+.
18b: C14H21N2O2P (280.30): calcd. C 59.99, H 7.55, N 9.99; found
C 60.20, H 7.71, N 10.13. MS: m/z = 280 [M]+.
rac-(2R,4aR,8aR)- and rac-(2S,4aR,8aR)-2-[Bis(2-chloroethyl)
amino]-1-methyl-1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiaza-
phosphinine 2-Oxides (22a and 22b)[35]
rac-(2S,4aR,8aR)- and rac-(2R,4aR,8aR)-2-[Bis(2-chloroethyl)
amino]-1,2,3,4,4a,5,6,7,8,8a-decahydro-1,3,2-benzodiazaphosphinine
2-Oxides (19a and 19b)[35]
22a: C12H24Cl2N3OP (328.22): calcd. C 43.91, H 7.37, N 12.80;
found C 44.18, H 7.48, N 12.72. MS: m/z = 327 [M]+.
19a: C11H22Cl2N3OP (314.19): calcd. C 42.05, H 7.06, N 13.37;
found C 41.81, H 6.87, N 13.53. MS: m/z = 313 [M]+.
22b: C12H24Cl2N3OP (328.22): calcd. C 43.91, H 7.37, N 12.80;
found C 44.06, H 7.52, N 13.13. MS: m/z = 327 [M]+.
19b: C11H22Cl2N3OP (314.19): calcd. C 42.05, H 7.06, N 13.37;
found C 42.21, H 6.80, N 13.04. MS: m/z = 313 [M]+.
Supporting Information (for details see the footnote on the first
page of this article): 1H, 13C and 31P NMR chemical shifts and
2158
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 2145–2159

Studies on Benzodiazaphosphinine Oxides
FULL PAPER
[20] V. M. S. Gil, W. von Philipsborn, Magn. Reson. Chem. 1989,
27, 409–430.
[21] J.-C. Yang, D. O. Shah, N. U. M. Rao, W. A. Freeman, G. Sos-
novsky, D. G. Gorenstein, Tetrahedron 1988, 44, 6305–6314.
[22] K. Pihlaja, E. Kleinpeter, Carbon-13 NMR Chemical Shifts in
Structural and Stereochemical Analysis, VCH Publishers, New
York, 1994, p. 58.
JH,H, JP,H and JP,C coupling constants for compounds 7–22a,b.
Typical acquisition and processing parameters for the 1D and 2D
NMR spectra. Cartesian coordinates and energies for the DFT op-
timized geometries of 10a and 10b.
[23] R. K. Harris, E. D. Becker, S. M. Cabral de Menezes, R. Good-
fellow, P. Granger, Pure Appl. Chem. 2001, 73, 1795–1818.
[24] See for example: R. Laatikainen, M. Niemitz, U. Weber, J. Sun-
delin, T. Hassinen, J. Vepsäläinen, J. Magn. Reson., Ser. A
1996, 120, 1–10; home page: http://www.perchsolutions.com.
[25] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Mont-
gomery Jr, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas,
J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson,
P. Y. Ayala, Q. Cui, K. Morokuma, P. Salvador, J. J. Dannen-
berg, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Ste-
fanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gom-
perts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. G.
Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M.
Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian 98, revision
A.11, Gaussian, Inc., Pittsburgh PA, 2001.
[26] J. B. Foresman, Æ. Frisch, Exploring Chemistry with Electronic
Structure Methods, 2nd ed., Gaussian Inc., Pittsburgh PA,
1996, p. 64.
[27] A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100.
[28] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[29] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[30] V. G. Malkin, O. L. Malkina, D. R. Salahub, Chem. Phys. Lett.
1994, 221, 91–99.
Acknowledgments
Z. Z. is grateful to the Centre for International Mobility (CIMO),
Finland for a grant. Dr. Petri Tähtinen is thanked for technical
assistance with the DFT calculations.
[1]
[2]
[3]
[4]
[5]
S. M. Ludeman in Biomedical Chemistry (Ed.: P. F. Torrence),
Wiley-Interscience, New York, 2000, pp. 163–187.
K. Afarinkia, A. J. Twist, H. Yu, J. Org. Chem. 2004, 69, 6500–
6503.
R. Pedrosa, A. Maestro, A. Pérez-Encabo, R. Raliegos, Synlett
2004, 1300–1302.
Z. Zalán, T. A. Martinek, L. Lázár, F. Fülöp, Tetrahedron
2003, 59, 9117–9125, and references cited therein.
O. M. Friedman, E. Boger, V. Grubliauskas, H. Sommer, J.
Med. Chem. 1963, 6, 50–58.
[6]
[7]
[8]
S. E. Denmark, S. Fujimori, Org. Lett. 2002, 4, 3473–3476.
S. E. Denmark, S. Fujimori, Org. Lett. 2002, 4, 3477–3480.
S. E. Denmark, D. M. Coe, N. E. Pratt, B. D. Griedel, J. Org.
Chem. 1994, 59, 6161–6163.
[9]
S. E. Denmark, J.-H. Kim, Can. J. Chem. 2000, 78, 673–688.
[10] B. Burns, N. P. King, H. Tye, J. R. Studley, M. Gamble, M.
Wills, J. Chem. Soc., Perkin Trans. 1 1998, 1027–1038.
[11] S. E. Denmark, R. L. Dorow, J. Am. Chem. Soc. 1990, 112,
864–866.
[12] T. Viljanen, P. Tähtinen, K. Pihlaja, F. Fülöp, J. Org. Chem.
1998, 63, 618–627.
[31] P. Tähtinen, A. Bagno, A. Koch, K. Pihlaja, Eur. J. Org. Chem.
2004, 4921–4930.
[13] H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K.
Pihlaja, Eur. J. Org. Chem. 2005, 1189–1200.
[14] S. E. Denmark, X. Su, Y. Nishigaichi, D. M. Coe, K.-T. Wong,
S. B. D. Winter, J. Y. Choi, J. Org. Chem. 1999, 64, 1958–1967.
[15] K. Pihlaja, F. Fülöp, J. Mattinen, G. Bernáth, Acta Chem.
Scand., Ser. B 1987, 41, 228–231.
[32] K. Okamoto, M. Noji, Y. Kidani, Bull. Chem. Soc. Jpn. 1981,
54, 713–718.
[33] W. L. F. Armarego, P. A. Reece, J. Chem. Soc., Perkin Trans. 1
1974, 2313–2319.
[34] W. L. F. Armarego, T. Kobayashi, J. Chem. Soc. (C) 1971,
238–245.
[16] L. Lázár, A. Göblyös, T. A. Martinek, F. Fülöp, J. Org. Chem.
2002, 67, 4734–4741.
[35] In this study, compounds 19–22a,b are named as 2-[bis(2-chlo-
roethyl)amino]-substituted 1,3,2-benzodiazaphosphinine 2-ox-
ides to provide uniformity with the nomenclature for the 2-
phenyl and 2-phenoxy derivatives. The correct IUPAC name,
for example, for 19a would be rac-(2S,4aR,8aR)-N,N-bis(2-
chloroethyl)decahydro-1,3,2-benzodiazaphosphinin-2-amine 2-
oxide.
[17] D. G. Gorenstein, Prog. Nucl. Magn. Reson. Spectrosc. 1983,
16, 1–98.
[18] G. S. Bajwa, S. Chandrasekaran, J. H. Hargis, A. E. Sopchik,
D. Blatter, W. G. Bentrude, J. Am. Chem. Soc. 1982, 104, 6385–
6392.
[19] K. Taira, K. Lai, D. G. Gorenstein, Tetrahedron 1986, 42, 229–
Received: September 1, 2005
Published Online: March 1, 2006
238.
Eur. J. Org. Chem. 2006, 2145–2159
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2159