Competing sulfonylation and phosphonylation following rearrangement of an O-sulfonyl-N-phosphinoylhydroxylamine with tert-butylamine: Demonstration of a phosphonamidic-sulfonic anhydride intermediate and 18O-labelling evidence on how it may be formed

Article abstract of DOI:10.1039/b305510h

Reaction of R(Ph)P(O)NHOSO₂Bn (R = PhMeCH) with Bu ^tNH₂ in CH₂Cl₂ gives a mixture of RP(O)(NHPh)NHBu^t and Bu^tNHSO₂Bn, the proportion of the sulfonamide increasing steadily (14.6% to 32.9%) as the concentration of amine is reduced (8.0 to 0.2 mol dm^-3). Apparently the phenyl and sulfonate groups in the substrate become transposed, giving a phosphonamidic-sulfonic anhydride RP(O)(NHPh)OSO₂Bn which then reacts at the phosphorus or sulfur atom to give the final products; an authentic sample of the anhydride gives similar mixtures of products. Substrate labelled with ¹⁸O in the sulfonyl position (57 mol% one ¹⁸O atom) gives sulfonamide containing most but not all of the label (43.7 mol% one ¹⁸O atom with 2.0 mol dm^-3 amine). This implies partial equilibration of the three sulfonate oxygen atoms during rearrangement, or after the anhydride intermediate has been formed.

Full text of DOI:10.1039/b305510h

View Article Online / Journal Homepage / Table of Contents for this issue
Competing sulfonylation and phosphonylation following
rearrangement of an O-sulfonyl-N-phosphinoylhydroxylamine with
tert-butylamine: demonstration of a phosphonamidic-sulfonic
18
anhydride intermediate and O-labelling evidence on how it may
be formed1
Martin J. P. Harger*
Department of Chemistry, The University, Leicester, UK LE1 7RH
Received 16th May 2003, Accepted 5th August 2003
First published as an Advance Article on the web 28th August 2003
Reaction of R(Ph)P(O)NHOSO2Bn (R = PhMeCH) with ButNH2 in CH2Cl2 gives a mixture of RP(O)(NHPh)-
NHBut and ButNHSO2Bn, the proportion of the sulfonamide increasing steadily (14.6% to 32.9%) as the
concentration of amine is reduced (8.0 to 0.2 mol dmϪ3). Apparently the phenyl and sulfonate groups in the substrate
become transposed, giving a phosphonamidic-sulfonic anhydride RP(O)(NHPh)OSO2Bn which then reacts at the
phosphorus or sulfur atom to give the final products; an authentic sample of the anhydride gives similar mixtures of
products. Substrate labelled with 18O in the sulfonyl position (57 mol% one 18O atom) gives sulfonamide containing
most but not all of the label (43.7 mol% one 18O atom with 2.0 mol dmϪ3 amine). This implies partial equilibration of
the three sulfonate oxygen atoms during rearrangement, or after the anhydride intermediate has been formed.
diamide 4, but attack at the sulfur atom would give a sulfonic
amide (RЈNHMs) instead. In practise products corresponding
Introduction
N-Phosphinoylhydroxylamines [R2P(O)NHOH] are the phos-
phorus analogues of hydroxamic acids and when suitably activ-
to attack at sulfur have not been observed. That may be of little
significance, however, since it seems that PV-sulfonic mixed
ated they undergo rearrangement on treatment with alkoxides
anhydrides often react exclusively at the phosphorus atom.11,12
or aliphatic amines.2,3 Thus, for example, the O-methylsulfonate
The observation of products corresponding to nucleophilic
attack at sulfur would obviously afford important support for a
1 (R = Ph) gives the phosphonic diamide 4 with RЈNH2, a
phenyl group having migrated from phosphorus to nitrogen.2
phosphonamidic-sulfonic anhydride. Also, it would open the
Benzyl groups will also migrate,4 and even simple alkyl groups,5
way to an exploration of how the anhydride is formed, using
but they do so only reluctantly and the unsymmetrical substrate
substrate labelled with 18O in one of the sulfonate positions.
1 (R = alkyl or benzyl) gives exclusively the product 4 of phenyl
As regards the possibility of attack at sulfur, alkylsulfonyl
compounds (RCH2SO2X; X = leaving group) can react with
migration (Scheme 1).3,6
nucleophiles both by direct (associative) displacement of the
leaving group and by a dissociative elimination–addition (EA)
mechanism with
a
sulfene intermediate (RCH᎐SO ).13
᎐
Increased acidity of the Cα–H bond (e.g. RCH2SO2X; R = 2Ph)
should accelerate the EA pathway, and in doing so might make
reaction at the sulfur atom of a phosphonamidic-sulfonic
anhydride more competitive with reaction at phosphorus.
Results and discussion
Evidence for a phosphonamidic-sulfonic anhydride intermediate
Scheme 1
The benzylsulfonyl derivative 5 [δP 30.0; δH 4.56 (2H, s)] was
prepared from Ph2P(O)NHOH using benzylsulfonyl chloride.
On treatment with ButNH2 (2 mol dmϪ3) it gave overwhelm-
ingly the normal phosphonic diamide rearrangement product 4
(R = Ph, RЈ = But). A trace of the sulfonic amide ButNHSO2-
CH2Ph was detected by comparison (1H NMR, GLC, MS) with
an authentic sample,14 but as it amounted to р1% of the total
product it is of little mechanistic significance.
Rearrangement may proceed via a monomeric metaphos-
phonimidate 2, analogous to the isocyanate formed in a Lossen
rearrangement,7 and some features of the reactions do accord
well with a reactive three coordinate PV intermediate.8 Other
aspects, however, are difficult to reconcile with such an inter-
mediate, at least as the sole product-forming species.8–10 One
possible alternative or competing pathway involves attack of
the nucleophile (RЈNH2) prior to migration of the phenyl
group, forming a five coordinate phosphorane. However, the
modest sensitivity to steric effects in both the substrate (R =
Me, Et, Pri) and the nucleophile (RЈ = Me, Pri, But) argues
against such an associative mechanism.8 Another possibility is
that the phenyl and sulfonate groups in the substrate become
transposed, resulting in a phosphonamidic-sulfonic mixed
anhydride 3 (Scheme 1). Nucleophilic attack at the phosphorus
atom of the mixed anhydride would then give the phosphonic
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 3 9 0 – 3 3 9 5
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
3390

View Article Online
Scheme 2
Table 1 Proportion (mol% by GLCa) of sulfonamide 12 in the amide
To further encourage attack at sulfur we considered replacing
b
product (11 ϩ 12) obtained from 7 or 8 with ButNH2 in CH2Cl2
one of the P-phenyl groups in the substrate by a bulky alkyl
or benzyl group. Such groups have much lower migratory
aptitudes than phenyl,3,6 so any phosphonamidic-sulfonic
anhydride that is formed in the rearrangement will retain the
bulky group on the phosphorus atom. Steric factors may then
retard attack of the nucleophile at phosphorus and give attack
at sulfur more chance to compete. The N-phosphinoyl-
hydroxylamine 6 has an α-methylbenzyl group on the phos-
phorus atom. This is bulky but it is also chiral, so 6 exists as
diastereoisomers and was previously prepared for stereo-
chemical studies.9 A mixture of the diastereoisomers of 6 (δP
41.9 and 40.1; ratio ∼3 : 1) was now treated briefly with PhCH2-
SO2Cl in ice-cold pyridine to give the benzylsulfonyl derivative
7 (75%) as a mixture of diastereoisomers, δP(CDCl3) 42.6
and 41.0 (ratio 3.3 : 1 after crystallisation). The two benzylic
protons in each of the diastereoisomers of 7 are themselves
diastereotopic, and they appear in the 1H NMR spectrum as an
AB quartet centred at δH 4.60 (major diastereoisomer) or 4.19.
With MeNH2 (2 mol dmϪ3) in CH2Cl2 the only substantial
phosphorus-containing product obtained from the benzyl-
sulfonate 7 was the expected phosphonic diamide 4 (R =
PhMeCH, RЈ = Me) (diastereoisomers, δP 27.7 and 27.1), corre-
Sulfonamide 12 (mol%)
[ButNH2]/mol dmϪ3
Reactant 7
Reactant 8
0.2
0.4
0.8
1.4
2.0
4.0
8.0
32.9
27.0
22.0
19.0
17.3
14.7
14.6
34.7
27.0
24.9
20.9
19.3
15.3
16.0
a GLC peak areas corrected for different responses of detector for
products 11 and 12. b For 7 with neat ButNH2, 14.0 mol% 12 was
produced.
the extent of its involvement: the phosphonic diamide (but
not the sulfonamide) can also be formed directly from the
metaphosphonimidate 2 (Scheme 1, step aЈ) without involve-
ment of the anhydride 3, and this could well be the major
pathway. At lower concentrations of amine a smaller propor-
tion of the metaphosphonimidate might be expected to go
directly to the phosphonic diamide; more would form the phos-
phonamidic-sulfonic anhydride (Scheme 1, step c) and thence,
to some extent, the sulfonamide instead of the phosphonic
diamide, so that the proportion of the sulfonamide in the prod-
uct would be increased. As a test of this the reaction of the
benzylsulfonate 7 with ButNH2 was repeated using a wide range
of amine concentrations (in CH2Cl2) and the relative yields of
sulfonamide 12 and phosphonic diamide 11 were determined by
GLC. The results (Table 1) show that the proportion of the
sulfonamide (mol%) in the amide product (11 ϩ 12) does
indeed increase steadily on dilution, from 14% in neat ButNH2
to 32.9% with 0.2 mol dmϪ3 amine.
The symmetrical phosphonamidic anhydride 15 was a
byproduct in these reactions (31P NMR), especially at low
amine concentrations. That is an inevitable consequence of
nucleophilic attack at the sulfur atom of the phosphonamidic-
sulfonic mixed anhydride 8, since the displaced phosphon-
amidate anion 13 (the leaving group) is itself an effective
nucleophile and will attack the mixed anhydride (at the P
atom) in competition with the amine. The absolute yields of
the amide products 11 and 12 will be reduced but, because
the symmetrical anhydride does not react with ButNH2
under the conditions employed, their relative yields will not be
affected.
1
sponding to migration of the phenyl group. However, the H
NMR spectrum of the reaction mixture suggested a small but
significant amount of the sulfonamide MeNHSO2CH2Ph [δH
4.26 (s) and 2.70 (d, JHH 5, NHMe); ca. 5%], and this was
confirmed by GLC (by comparison with an authentic sample15)
and MS (Mϩ 185) after isolation by TLC.
With ButNH2 (2 mol dmϪ3) the corresponding phosphonic
diamide 11 (diastereoisomers, δP 22.7 and 22.4) was again the
dominant product (see Scheme 2) but now the 31P NMR spec-
trum of the reaction mixture also contained a minor peak at δP
19.0 (8%), corresponding to the phosphonamidate anion 13
(isolated and characterised as the methyl phosphonamidate 14)
and several small peaks δP 26–23 (10% of total phosphorus),
apparently due to the phosphonamidic anhydride 15 (several
diastereoisomers). The anion 13 is the expected byproduct of
nucleophilic attack at the sulfur atom of the phosphonamidic-
sulfonic mixed anhydride 8, while the phosphonamidic
anhydride 15 would result from subsequent attack of the anion
13 at the phosphorus atom of 8. More conclusive evidence for
involvement of the phosphonamidic-sulfonic anhydride was the
formation of the sulfonamide 12 [δH 4.23 (s) and 1.34 (s,
NHBut)] in ca. 17% yield; it was isolated chromatographically
and fully characterised.
To extract full mechanistic value from these results we would
need precise knowledge of the behaviour of the phosphon-
amidic-sulfonic anhydride 8 with ButNH2, i.e. reaction at phos-
phorus vs. reaction at sulfur. Fortunately we were able to obtain
an authentic sample of 8 (diastereoisomers, δP 29.15 and 28.7),
contaminated with only a little of the symmetrical anhydride
15, by heating the phosphonamidic chloride 9 with silver
benzylsulfonate in MeCN (Scheme 2).16
Extent of involvement of the phosphonamidic-sulfonic anhydride
The foregoing observations clearly implicate a phosphonamidic-
sulfonic anhydride in the rearrangement but they do not reveal
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 3 9 0 – 3 3 9 5
3391

View Article Online
Scheme 3
Table 2 18O Content (mol%) of products from rearrangement of
As expected, this anhydride gave a mixture of the phosphonic
diamide 11 and the sulfonamide 12 when treated with ButNH2
(2.0 mol dmϪ3) in CH2Cl2. Surprisingly, however, the propor-
tion of the sulfonamide (19.3 mol%) was only slightly greater
than for the rearrangement of 7 under the same conditions
(17.3%). Still more surprising was the effect of varying the
amine concentration. Not only did the relative yield of the sul-
fonamide change, showing that the behaviour of the anhydride
8 is not constant, but the changes paralleled closely those
seen in the rearrangement (Table 1). Indeed, given that the
diastereoisomer composition of any phosphonamidic-sulfonic
anhydride formed by rearrangement is unlikely to be the same
as that of the authentic sample, and that the diastereoisomers
may differ somewhat in their behaviour with ButNH2, the dis-
crepancies between the two sets of results (Table 1) can prob-
ably be discounted. Certainly there is not the anticipated
divergence between the two sets of results at high amine con-
centrations. The implication is clear and mechanistically very
important: most of the rearrangement of 7, and possibly all of
it, proceeds via the phosphonamidic-sulfonic anhydride even
at the highest concentrations of amine. Little if any of the
rearrangement product is formed directly from the meta-
phosphonimidate. It is not difficult to reconcile this with path
a in Scheme 1 when the amine concentration is low: the
metaphosphonimidate 2 might reasonably be formed without
any amine in its solvation shell, and recombine with the
sulfonate anion (step c) before it can diffuse away and react with
the amine. But at the highest amine concentrations, and
especially in neat ButNH2, the metaphosphonimidate will be
formed in contact with amine molecules. One of these will
surely be as well placed as the sulfonate anion to form a bond to
phosphorus, yet apparently it does not do so. Unless the meta-
phosphonimidate has a marked preference for combination
with the sulfonate anion, path a becomes untenable. The
alternative is a concerted rearrangement of the conjugate base
of the substrate (Scheme 1, path b). Then the sulfonate
group would migrate to phosphorus at the same time as the
phenyl group migrates to nitrogen, via a transition state
such as 16 or 17, and there would be no opportunity for prod-
ucts to be formed without the intervention of the phos-
phonamidic-sulfonic anhydride. Experiments with 18O-labelled
substrate would, we hoped, confirm the identity of the sulfonyl-
ating agent and also shed light on the mechanism of the
rearrangement.
18O-sulfonyl-N-phosphinoylhydroxylamine 8 with ButNH2
[ButNH2]/mol dmϪ3
8.0
2.0
0.4
57.7
Sulfonate anion 10
57.6
0.0
49.1
8.4
57.6
0.0
43.7
13.7
Phosphonic diamide 11
Sulfonamide 12
0.0
41.4
Phosphonamidate anion 13
(16.3)a
a Not measured directly; value shown is 18O content of (12 ϩ 13)
unaccounted for by 12 [18O in (12 ϩ 13) = 18O in (10 ϩ 11)].
enriched with one 18O atom in the SO2 group (FAB mass
spectrometry; negligible double-labelled material).
The rearrangement of 18O-labelled 7 was examined using
2.0 mol dmϪ3 ButNH2 in CH2Cl2. The phosphonic diamide 11
(mixture of diastereoisomers) and the sulfonamide 12 (ca. 17%)
were isolated directly by TLC but the (water-soluble) salts 10
and 13 were first converted into the methyl esters by treatment
with diazomethane. The amount of 18O in each of the products
was determined by mass spectrometry (EI generally but CI for
12).† For the products resulting from nucleophilic attack at the P
atom of the phosphonamidic-sulfonic anhydride 8 the results
were as expected: all of the label was located in the sulfonate
anion 10 (methyl ester: 57.6% one 18O atom) and none in the
phosphonic diamide 11. Of more concern are the products that
result from attack at the sulfur atom, since it is these that will
reflect the distribution of the 18O label between the bridging and
non-bridging sulfonate positions in the anhydride 8. The sul-
fonamide 12 was found to contain 76% of the available label
(43.7% one 18O atom) and the phosphonamidate anion 13 24%
(methyl ester 14: 13.7% one 18O atom). This implies a 24 : 76
distribution of label between the bridging and non-bridging
positions of the phosphonamidic-sulfonic anhydride 8.
If the anhydride 8 was formed by concerted rearrangement
of the conjugate base of the substrate (i.e. 7 less the NH proton)
the expected distribution of the label would be 50 : 50 for a
transition state such as 16 or 0 : 100 for a transition state such as
17. The measured distribution (24 : 76) is equally far removed
from either. The experimentally observed value does, however,
relate to the anhydride as it is when it forms the product (10 ϩ
11 or 12 ϩ 13), and some scrambling of the label might occur
after the anhydride has been formed but before it is converted
into product. Like a phosphonamidic chloride,17 the phos-
phonamidic-sulfonic anhydride will tend to react at phosphorus
by a stepwise preassociative elimination–addition mechanism.18
The reverse of the base-induced elimination shown in Scheme 3
is therefore likely to be important, especially at lower amine
(nucleophile) concentrations, so there could well be some equi-
libration of bridge and non-bridge labelled anhydrides. At
higher concentrations of amine there is a greater probability
that the nucleophile will already be in place (preassociation)
when the elimination occurs, ready to take the metaphos-
phonimidate on to product; return to the anhydride should be
less important and scrambling of the 18O label less extensive. As
a test of this the rearrangement of the labelled substrate 7 was
Mechanism of rearrangement to the phosphonamidic-sulfonic
anhydride
Isotopically labelled PhCH2SO2Cl was prepared from the
unlabelled compound by hydrolysis with H218O–pyridine
followed by treatment of the hydrolysis product with oxalyl
chloride. Reaction of labelled PhCH2SO2Cl with the N-phos-
phinoylhydroxylamine 6 afforded the sulfonyl derivative 7 (5 : 1
mixture of diastereoisomers) having ca. 57% of the molecules
† In all mass spectrometric measurements of 18O enrichment, allowance
was made for 34S and natural abundance 18O by comparing directly the
spectrum of the enriched material with that of a sample of natural
abundance.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 3 9 0 – 3 3 9 5
3392

View Article Online
repeated using other amine concentrations (Table 2). The dis-
tribution of the label (bridge vs. non-bridge) in the anhydride,
as inferred from examination of the reaction products, was 15 :
85 when the ButNH2 concentration was increased to 8.0 mol
The yield of sulfonamide relative to phosphonic diamide
declines as the concentration of the amine is increased, but not
because less anhydride is formed: it is the balance between the
competing reactions at the sulfur and phosphorus atoms of the
anhydride that changes, and over a wide range of amine con-
centrations the product mixtures obtained from the hydroxyl-
amine derivative 7 parallel closely those obtained from an
authentic sample of the anhydride 8.‡ Regardless of the concen-
tration of amine, most and possibly all of the rearrangement of
7 must, it seems, proceed via the anhydride 8. The same is likely
to be true for all O-sulfonyl-N-phosphinoylhydroxylamines, but
it is only in exceptional cases that attack at sulfur competes
substantially with attack at phosphorus and makes manifest the
involvement of the phosphonamidic-sulfonic anhydride.
dmϪ3 and 28 : 72 when it was reduced to 0.4 mol dmϪ3
.
Scrambling will, of course, tend to even out the distribution
of the label, but not beyond the statistical value of 33 : 67. If the
rearrangement is concerted, the transition state cannot be as in
16: an initial 50 : 50 distribution of label could never give rise to
an experimental (average) value of 24 : 76. The transition state
could conceivably be like 17, although there would have to be
extensive scrambling in the anhydride, even with 2.0 mol dmϪ3
amine, for the initial 0 : 100 distribution of label to result in the
observed 24 : 76 (average) value.
Superficially the results of the labelling experiments might
seem to rule out entirely a non-concerted route to the phos-
phonamidic-sulfonic anhydride (Scheme 1, step a ϩ step c). The
sulfonate group would in that case become completely detached
from the nitrogen atom before it begins to bond to phosphorus,
and the 18O label must, in the limit, end up evenly distributed
between the bridging position of the anhydride and the two
non-bridging positions (ratio 33 : 67). Initially, however, the
sulfonate oxygen atom (unlabelled) released from the N–O
bond will not be equivalent to the other two because of differ-
ences in association (with ButNH3ϩ) or solvation (by ButNH2 or
CH2Cl2). If the new bond to phosphorus is formed before equi-
libration of the oxygen atoms is complete, and makes preferen-
tial use of this (unlabelled) oxygen atom, the anhydride will be
formed with less than a third of the available 18O label located in
the bridging position. The observed result, 24% of the available
18O in the bridge, could then be rationalised by a fairly modest
bias for incorporation of the unlabelled sulfonate oxygen in the
new P–O bond, or a strong bias with some subsequent scram-
bling. With regard to the latter possibility, an exceptionally high
reactivity for one of the oxygen atoms of the sulfonate anion
might also resolve the dilemma noted earlier, viz. the failure
of ButNH2, even at the highest concentration, to suppress
anhydride formation (Scheme 1, step c) by intercepting the
metaphosphonimidate and taking it directly to the phosphonic
diamide product (step aЈ). There are several examples of
racemisation and isomerisation in which a carboxylate19 or
sulfonate20 anion is released and recaptured with only partial
equilibration of the oxygen atoms, and at least one example of
a more profound rearrangement (migration from N to C) in
which the sulfonate oxygen atom released by bond-breaking is
preferentially employed in subsequent bond making.21 In all of
these, however, the anion recombines with a cation, not an
uncharged species like a metaphosphonimidate. Nonetheless,
it does seem quite possible that rearrangement to the
phosphonamidic-sulfonic anhydride is non-concerted, notwith-
standing the incomplete equilibration of the sulfonate oxygen
atoms.
Experimental
Mps were determined using a Kofler hot-stage apparatus and
are uncorrected. H NMR spectra were recorded on a Bruker
1
ARX 250 spectrometer (Me4Si internal standard; coupling con-
stants J given in Hz) and 31P NMR spectra (1H decoupled) were
recorded on the same instrument at 101.3 MHz (positive chem-
ical shifts downfield from 85% H3PO4). Mass spectra were
recorded on a Kratos Concept spectrometer in EI (70 eV), CI
(NH3) or FAB (m-nitrobenzyl alcohol matrix) mode, or a
Perkin Elmer TurboMass GC-MS instrument (in EI mode). IR
spectra were recorded as Nujol mulls on a Perkin Elmer 298
spectrometer. GLC analyses were performed using a Philips PU
4500 chromatograph (flame-ionisation detector) fitted with an
OV 1701 wide bore capillary column (1 µm film; 15 m × 0.53
mm) using helium as carrier gas (19 ml minϪ1); peak areas were
measured with a Spectra-Physics SP4270 integrator. Amines
were dried over and distilled from KOH, and CH2Cl2 was
distilled from CaH2. Light petroleum refers to the fraction bp
60–80 ЊC and ether to diethyl ether.
O-Benzylsulfonyl-N-(diphenylphosphinoyl)hydroxylamine 5
While cooling constantly in ice, pyridine (0.4 ml) was added to
N-(diphenylphosphinoyl) hydroxylamine2 (117 mg, 0.5 mmol),
followed immediately by PhCH2SO2Cl (114 mg, 0.6 mmol). The
mixture was shaken for 8 min. Much of the pyridine was evap-
orated in vacuo (no heating) and iced water was added, giving
an oil that soon solidified. Thorough washing of the solid with
MeOH–H2O (2 : 1) and drying at 35 ЊC at 0.1 mmHg afforded
the sulfonate 5 (161 mg, 83%); δP(CDCl3) 30.0; δH(CDCl3) 8.29
(1H, d, JPH 7.5, NH), 8.0–7.3 (15H, Ph) and 4.56 (2H, s,
SO2CH2Ph); m/z (FAB) 388 [(M ϩ H)ϩ, 100] and 201 (30). A
sample crystallised from MeOH had mp 132–133 ЊC (decomp.)
(Found: C, 58.9; H, 4.7; N, 3.4. C19H18NO4PS requires C, 58.9;
H, 4.7; N, 3.6%).
O-Benzylsulfonyl-N-[phenyl(1-phenylethyl)phosphinoyl]-
hydroxylamine 7
The N-phosphinoylhydroxylamine 69 (mixture of diastereo-
isomers) (104 mg, 0.4 mmol) was sulfonylated as above to give,
after crystallisation from CH2Cl2–ether and drying (35 ЊC, 0.1
mmHg), the sulfonate 7 (124 mg, 75%), mp 75–78 ЊC; δP(CDCl3)
42.6 and 41.0 (ratio 3.3 : 1; diastereoisomers); δH(CDCl3) major
diastereoisomer: 8.30 (1H, d, JPH 7, NH), 8.0–7.1 (Ph groups),
4.60 (2H, AB; δA 4.66, δB 4.54, JAB 14; PhCH2SO2), 3.74 (1H,
dq, JPH 14.5, JHH 7.5, PhMeCH ) and 1.74 (3H, dd, JPH 16.5,
JHH 7.5, PhMeCH); minor diastereoisomer: 7.72 (1H, d, JPH
Conclusion
The sulfonamide 12 is a substantial product in the reaction of
the O-benzylsulfonyl derivative of the N-phosphinoylhydroxyl-
amine 6 with ButNH2. It is formed at the expense of the normal
phosphonic diamide rearrangement product 11, but not by
direct transfer of the sulfonyl group; substrate 7 specifically
labelled with 18O in the sulfonyl (SO2) group produces sul-
fonamide containing considerably less isotope. Rather, trans-
position of the phenyl and sulfonate groups in 7 gives the phos-
phonamidic-sulfonic mixed anhydride 8 and it is this that
sulfonylates the amine. Whether the anhydride is formed con-
certedly after deprotonation (NH ) of the substrate, or non-
concertedly via a metaphosphonimidate (Scheme 1, step a ϩ
step c), is not clear; if it is concerted, however, the 18O-labelling
experiments show that the transition state must resemble 17
rather than 16.
‡ For substitution by a preassociative stepwise elimination–addition
mechanism the amine will be involved in the rate-limiting elimination
stage as both base and nucleophile; the order in amine will therefore be
greater than for a simple (no preassociation) stepwise elimination–
addition mechanism. If attack at sulfur is less dependent on preassoci-
ation than attack at phosphorus it will benefit less from any increase in
the concentration of amine.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 3 9 0 – 3 3 9 5
3393

View Article Online
4.5, NH), 8.0–7.1 (Ph groups), 4.19 (2H, AB; δA 4.27, δB 4.12,
JAB 14), 3.64 (1H, dq, JPH 12, JHH 7.5, PhMeCH ) and 1.54 (3H,
dd, JPH 17.5, JHH 7.5, PhMeCH); m/z (FAB) 416 [(M ϩ H)ϩ,
100%] (Found: C, 60.8; H, 5.6; N, 3.2. C21H22NO4PS requires C,
60.7; H, 5.5; N, 3.4%).
PhCH2SO2NHMe, crystallised from CHCl3–light petroleum,
mp 110 ЊC (lit.,15 109–110 ЊC); δH(CDCl3) 7.40 (5H, s), 4.26 (2H,
s), 4.04 (br s, NH) and 2.71 (3H, d, JHH 5, NHMe); m/z (EI) 185
(Mϩ, 3%) and 91 (100).
PhCH2SO2NHBut, crystallised from CH2Cl2–light petrol-
eum, mp 107–108 ЊC (lit.,14 107.5–109.5 ЊC); δH(CDCl3) 7.35
(5H, s), 4.22 (2H, s), 4.11 (s, NH) and 1.35 (9H, s); m/z (EI) 227
(Mϩ, р1%); m/z (CI) 245 [(M ϩ NH4)ϩ, 35%], 228 [(M ϩ H)ϩ,
10], 148 (55) and 91 (100).
O-Benzyl[18O]sulfonyl-N-[phenyl(1-phenylethyl)phosphinoyl]-
hydroxylamine
(a) A solution of H218O (22 mg, 1.1 mmol) in pyridine (200 µl)
was stirred and cooled in ice while PhCH2SO2Cl (191 mg,
1.0 mmol) was added. After 10 min the mixture was moved to a
60 ЊC bath and heated for 15–20 min. The pyridinium sulfonate
was converted into the tert-butylammonium sulfonate by
adding CH2Cl2 (2.5 ml) and, in portions, ButNH2 (146 mg,
2.0 mmol), evaporating the volatile material in vacuo, and treat-
ing the residue with CH2Cl2 and ButNH2 again. Trituration
with ether gave the solid tert-butylammonium salt. This salt
was suspended in CH2Cl2 (2.5 ml) and stirred with oxalyl chlor-
ide (254 mg, 2.0 mmol) and a catalytic quantity of DMF (2 µl).
After 0.5 h the volume was reduced and ether was added. The
precipitate (ButNH3Cl) was removed and the filtrate was con-
centrated. The residue was extracted with a little warm ether
and the extract was diluted with light petroleum to give crystals
of benzyl[18O]sulfonyl chloride (145 mg, 75%), mp 90–91 ЊC;
δH(CDCl3) 7.47 (5H, s) and 4.87 (2H, s); m/z (EI) 190, 192, 194
(Mϩ, 5%; ratio 1.9 : 2.9 : 1) and 91 (100).
PhCH2SO2OMe, crystallised from ether–light petroleum, mp
59–61 ЊC (lit.,22 60–62 ЊC); δH(CDCl3) 7.41 (5H, s), 4.36 (2H, s)
and 3.75 (3H, s); m/z (EI) 186 (Mϩ, 15%) and 91 (100).
Samples of the following compounds were available from
previous work.
The phosphonic diamide 4 (R = Ph, RЈ = But),2 δP(CH2Cl2)
12.1.
The phosphonic diamide 4 (R = PhMeCH, RЈ = Me),9
δP(CH2Cl2) 26.2 and 25.7 (diastereoisomers); m/z (EI) 274 (Mϩ,
40%) and 169 (Mϩ Ϫ CHMePh, 100).
The phosphonic diamide 11,9 δP(CDCl3) 22.5 and 22.75
(diastereoisomers); m/z (EI) 316 (Mϩ, 25%), 211 (Mϩ
CHMePh, 45) and 155 (Mϩ Ϫ CHMePh Ϫ C4H8, 100).
Ϫ
The methyl phosphonamidate 14,10 δP(CDCl3) 31.0 and 30.8
(diastereoisomers); m/z (EI) 275 (Mϩ, 100%).
The phosphonamidic anhydride 15,10 δP(CDCl3) 25.65–24.23
(several diastereoisomers; 5 principal peaks); m/z (EI) 504 (Mϩ,
20%) and 105 (PhMeCHϩ, 100).
(b) The N-phosphinoylhydroxylamine 6 was treated with
benzyl[18O]sulfonyl chloride using the method previously
employed for the unlabelled material (see above). The resulting
18O-labelled sulfonate 7, δP(CDCl3) 42.6 and 41.0 (ratio 5.0 : 1;
diastereoisomers), m/z (FAB) 416 and 418 [(M ϩ H)ϩ], had 57%
of the molecules singly labelled and <1% doubly labelled.
Rearrangement reactions of O-benzylsulfonyl-N-phosphinoyl-
hydroxylamines
(a) The substrate 5 (0.05 mmol) was added to a 2.0 mol dmϪ3
solution of ButNH2 in CH2Cl2. After 20 min at 25–30 ЊC the
volatile material was evaporated. The 31P NMR spectrum of the
crude product consisted of a single peak, δP(CDCl3) 12.7, and
the 1H NMR spectrum (CDCl3) indicated that the yield of
PhCH2SO2NHBut (δH 4.22 for the authentic sample) was only
0.5% relative to PhCH2SO3Ϫ (δH 4.03). The NMR solution was
diluted, washed with water and examined by GLC, which indi-
cated a trace amount of PhCH2SO2NHBut (tR 1.0 min at
210 ЊC) in addition to the dominant product (tR 7.2 min).
Crystallisation from ether–light petroleum afforded the phos-
phonic diamide 4 (R = Ph, RЈ = But), mp 176–178 ЊC (lit.,2 176–
178 ЊC), IR and 1H NMR spectra as previously described,2 and
the mother liquor afforded (TLC) a trace of PhCH2SO2NHBut;
m/z (CI) 245 [(M ϩ NH4)ϩ, 60%] and 228 [(M ϩ H)ϩ, 20].
(b) The substrate 7 (mixture of diastereoisomers) (0.02
mmol) was added to a 2.0 mol dmϪ3 solution of MeNH2 in
CH2Cl2 (0.3 ml). After 20 min the mixture was concentrated,
diluted with CH2Cl2, and washed with a very small volume of
water; the phosphonic diamide 4 (R = PhMeCH, RЈ = Me) was
isolated as a mixture of diastereoisomers; δP(CDCl3) 27.6 and
Phosphonamidic-sulfonic anhydride 8
(a) Benzylsulfonyl chloride (300 mg) was hydrolysed by heating
in H2O–EtOH (1 : 9) (2 ml) at 80 ЊC for 2 h. Volatile material
was evaporated and the residue was extracted with water (1 ml).
The aqueous extract was washed with CH2Cl2 (2 × 0.3 ml) and
was then kept in vacuo until solid benzylsulfonic acid was
obtained. The sulfonic acid (ca. 1.6 mmol) was added to a
stirred suspension of Ag2O (232 mg, 1.0 mmol) in MeCN
(3 ml). After 10 min the mixture was filtered through Celite, the
filtrate was concentrated, and the residual solid was washed
repeatedly with CH2Cl2 to give silver benzylsulfonate which was
dried in vacuo over P2O5.
(b) With careful exclusion of moisture MeCN (0.35 ml) was
added to PhCH2SO3Ag (70 mg, 0.25 mmol) and N-phenyl-P-(1-
phenylethyl)phosphonamidic chloride 9 (mixture of diastereo-
isomers)9 (56 mg, 0.20 mmol) in a septum-capped tube. The
mixture was maintained at 50–55 ЊC overnight (16 h) giving the
phosphonamidic-sulfonic anhydride 8 contaminated with some
of the symmetrical phosphonamidic anhydride 15 [10% of total
phosphorus (31P NMR); 5 mol%]. The MeCN was evaporated
in vacuo via a needle inserted through the septum and then
CH2Cl2 (0.5 ml) was added. The mixture was shaken and was
left to stand until the silver salts had settled. Portions of the
solution of 8 were removed by syringe and used immediately;
δH(CDCl3) (mixture of diastereoisomers) 7.5–6.9 (Ph groups),
5.87 and 5.40 (both d, JPH 8, NH), 4.67 (s) and 4.34 (AB quar-
tet; δA 4.45, δB 4.23, JAB 14.5) (SO2CH2Ph), 3.585 and 3.525
(both dq, JPH 21, JHH 7.5, PhMeCH ), 1.64 and 1.595 (both dd,
JPH 20, JHH 7.5, PhMeCH); δP(CDCl3) 29.15 and 28.7 (ratio ca.
1 : 1; diastereoisomers) [impurity 15, 26.1–25.4 (several
diastereoisomers)].
1
26.85 (ratio 3 : 2). The H NMR spectrum was as previously
reported9 with small additional peaks attributable to PhCH2-
SO2NHMe (5%); δH(CDCl3) 4.26 (s, PhCH2) and 2.70 (d, JHH 5,
NHMe). The presence of the sulfonamide was confirmed by
GLC, tR 8.4 min at 150 ЊC. TLC (silica, ether) afforded separate
samples of the phosphonic diamide, Rf 0.1, m/z (EI) 274 (Mϩ,
50%) and 169 (Mϩ Ϫ CHMePh, 100), and the sulfonamide, Rf
0.4, m/z (CI) 203 [(M ϩ NH4)ϩ, 100%] and 186 [(M ϩ H)ϩ, 10].
(c) The substrate 7 (3.3 : 1 diastereoisomer mixture)
(0.24 mmol) was added to a 2.0 mol dmϪ3 solution of ButNH2
in CH2Cl2 (4.8 ml). After 30 min at ca. 30 ЊC the volatile
material was evaporated and the crude product was examined
spectroscopically; δP(CDCl3) 22.8 and 22.7 (80%, ratio 2 : 1;
diastereoisomers of 11) and 18.95 (8%; anion 13) (also several
peaks 25.9–23.3); δH(CDCl3) included 4.23 (PhCH2SO2NHBut)
and 4.07 (PhCH2SO3Ϫ) in a 1 : 5 ratio. The crude product was
partitioned between CH2Cl2 and water. Preparative TLC (silica,
ether) of the organic portion gave the phosphonic diamide 11
Authentic samples used for analysis of products of rearrangement
reactions
The following were prepared from benzylsulfonyl chloride and
the appropriate amine or MeOH–Et3N.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 3 9 0 – 3 3 9 5
3394

View Article Online
(mixture of diastereoisomers), Rf 0.2, crystallised from light
petroleum, mp 121–123 ЊC (lit.,9 120–124 ЊC); δP(CDCl3) 22.65
and 22.4 (ratio 3 : 2); m/z (EI) 316 (Mϩ, 60%); IR and 1H NMR
spectra as previously reported.9 The sulfonamide 12 was also
obtained, Rf 0.6, crystallised from light petroleum, mp 107–108
ЊC (lit.,14 107.5–109.5 ЊC); m/z (CI) 245 [(M ϩ NH4)ϩ, 100%]
and 228 [(M ϩ H)ϩ, 30]; δH(CDCl3) 7.35 (5H, s), 4.24 (2H, s),
3.93 (s, NH) and 1.36 (9H, s).
did not react with ButNH2 under the conditions employed
(amine concentration and reaction time) for the reactions of 7
and 8.
Acknowledgements
Thanks are due to Dr G. Eaton for recording mass spectra and
Professor P. M. Cullis for valuable discussions.
The aqueous portion was freeze dried, methylated (CH2N2),
and washed with water. GLC (T 160 ЊC for 1 min then 8 ЊC
minϪ1 to 240 ЊC) indicated a mixture of PhCH2SO2OMe (tR
2.1 min) (major) and PhMeCHP(O)(NHPh)OMe (tR 8.4 and
9.0 min; diastereoisomers) by comparison with authentic
samples. Preparative TLC (silica, ether) afforded samples of the
methyl sulfonate, Rf 0.5, mp 60–60.5 ЊC, m/z (EI) 186 (Mϩ, 15%)
and 91 (100), and the methyl phosphonamidate, Rf 0.15, m/z
(EI) 275 (Mϩ, 100%), δP(CDCl3) 30.85 and 30.5 (diastereo-
isomers); the 1H NMR spectra of these were as for the
authentic samples.
References
1 For a preliminary communication of some of the 18O-labelling
results see: M. J. P. Harger, Chem. Commun., 1997, 403.
2 M. J. P. Harger, J. Chem. Soc., Perkin Trans. 1, 1983, 2699.
3 M. J. P. Harger and A. Smith, J. Chem. Soc., Perkin Trans. 1, 1987,
683.
4 M. J. P. Harger and A. Smith, J. Chem. Soc., Perkin Trans. 1, 1990,
1447.
5 M. J. P. Harger and P. A. Shimmin, Tetrahedron, 1992, 48, 7539.
6 M. J. P. Harger and A. Smith, J. Chem. Res. (S), 1991, 358.
7 E. S. Wallis and R. D. Dripps, J. Am. Chem. Soc., 1933, 55, 1701;
H. L. Yale, Chem. Rev., 1943, 33, 209; L. Bauer and O. Exner,
Angew. Chem., Int. Ed. Engl., 1974, 13, 376.
8 M. J. P. Harger and A. Smith, J. Chem. Soc., Perkin Trans. 1, 1990,
2507.
9 M. J. P. Harger and R. Sreedharan-Menon, J. Chem. Soc.,
Perkin Trans. 1, 1994, 3261.
10 M. J. P. Harger and R. Sreedharan-Menon, J. Chem. Soc., Perkin
Trans. 2, 1999, 159 . A very low concentration of ButNH2 had to be
used in this study; the conclusions might not be valid under more
normal conditions.
11 W. Dabkowski, Z. Skrzypczynski, J. Michalski, N. Piel, L. W.
McLaughlin and F. Cramer, Nucleic Acids Res., 1984, 12, 9123;
W. Dabkowski, J. Michalski and Z. Skrzypczynski, Chem. Ber.,
1985, 118, 1809.
(d) The reaction of 18O-labelled substrate 7 (5.0 : 1
diastereoisomer mixture) with 2.0 mol dmϪ3 ButNH2 in CH2Cl2
was conducted as in (c) above (on one-tenth scale). The 31P
NMR signal for the phosphonamidate anion 11 now consisted
of two peaks, δP 19.46 (16O) and 19.42 (18O) (ratio 88 : 12). The
18O content of the products (after methylation of the anions 10
and 13) was determined by mass spectrometry (EI generally but
CI for the sulfonamide 12) on isolated samples (TLC). The
isolated sample of the methyl phosphonamidate 14 contained
some impurities and GC-MS was used for the analysis.
Comparable experiments were also carried out using 0.4 and
8.0 mol dmϪ3 solutions of ButNH2 in CH2Cl2.
Comparison of reactions of O-sulfonyl-N-phosphinoyl-
hydroxylamine 7 and phosphonamidic-sulfonic anhydride 8
12 J. Michalski, W. Dabkowski and J. Wasiak, Pol. J. Chem., 1992, 66,
879.
13 J. F. King, Acc. Chem. Res., 1975, 8, 10; J. F. King, J. Y. L. Lam and
S. Skonieczny, J. Am. Chem. Soc., 1992, 114, 1743; J. F. King and
R. Rathore in The Chemistry of Sulfonic Acids and Esters and their
Derivatives, ed. S. Patai and Z. Rappoport, Wiley, Chichester, 1991,
ch. 17; S. N. Lyashchuk, Y. G. Skrypnik and V. P. Besrodnyi,
J. Chem. Soc., Perkin Trans. 2, 1993, 1153.
The substrate 7 or 8 in CH2Cl2 was added to a solution of
ButNH2 (4–160 equiv.) in CH2Cl2 to give a reaction mixture of
the required concentration in amine (Table 1) and 0.05 mol
dmϪ3 in substrate. The reaction of 7 was also carried out in neat
ButNH2. When the reaction was complete (2–90 min) the vola-
tile material was evaporated and the product mixture was dis-
solved in CH2Cl2 and washed with water. It was then analysed
in duplicate by GLC (T 150 ЊC for 1 min then 8 ЊC minϪ1 to
220 ЊC) and the peak areas of the sulfonamide 12 (tR 4.6 min)
and the phosphonic diamide 11 (diastereoisomers, tR 10.6 and
11.2 min) were determined by integration. In some cases the
mixture was also examined by 1H NMR spectroscopy (CDCl3),
using NBut signals to deduce the relative amounts of the sul-
fonamide 12 (δH 1.36) and phosphonic diamide 11 (δH 1.305 and
1.19). Comparison of the GLC and NMR measurements indi-
cated that the GLC detector response for the sulfonamide was
0.75 molϪ1 relative to the response for the phosphonic diamide;
the GLC results shown in Table 1 have been corrected to allow
for this. Control experiments showed that the symmetrical
phosphonamidic anhydride 15, a byproduct of the reactions,
14 J. C. Sheehan, U. Zoller and D. Ben-Ishai, J. Org. Chem., 1974, 39,
1817.
15 T. B. Johnson and J. A. Ambler, J. Am. Chem. Soc., 1914, 36, 372.
16 c.f. J. Wasiak and J. Michalski, Tetrahedron Lett., 1994, 35, 9473.
17 S. Freeman and M. J. P. Harger, J. Chem. Soc., Perkin Trans. 2, 1988,
81; S. Freeman and M. J. P. Harger, J. Chem. Soc., Perkin Trans. 1,
1988, 2737.
18 W. P. Jencks, Acc. Chem. Res., 1980, 13, 161; W. P. Jencks, Chem. Soc.
Rev., 1981, 10, 345.
19 H. L. Goering, M. M. Pombo and K. D. McMichael, J. Am. Chem.
Soc., 1963, 85, 965; H. L. Goering and K. Humski, J. Org. Chem.,
1975, 40, 920.
20 H. L. Goering and B. E. Jones, J. Am. Chem. Soc., 1980, 102, 1628
and references cited therein. Y. Yukawa, H. Morisaki, K. Tsuji,
S.-G. Kim and T. Ando, Tetrahedron Lett., 1981, 22, 5187.
21 U. Gessner, A. Heesing, L. Keller and W. K. Homann, Chem. Ber.,
1982, 115, 2865 (see especially the rearrangement in methanol).
22 W. E. Truce and R. W. Campbell, J. Am. Chem. Soc., 1966, 88, 3599.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 3 9 0 – 3 3 9 5
3395