Intramolecular nucleophilic displacement of halogen by phosphinate and thiophosphinate anions: Relative rates of formation of five- And six-membered rings 1

Article abstract of DOI:10.1039/a900623k

Intramolecular nucleophilic substitution transforms the phosphinate anions XCH₂CH₂(CH₂)_nCH₂(Ph)P(O)O (n = 0, 1; X = Br, Cl) (Et₃NH⁺ salts; CH₂Cl₂ solution) into cyclic phosphinate esters 14 (n - 0,1); unusually the five-membered ring product (n = 0) is formed only 4.3 (X = Br) or 5.7 (X = Cl) times faster than the six (n = 1). The analogous cyclisation of the thiophosphinate anions ClCH₂CH₂(CH₂)_nCH ₂(Ph)P(S)CT (n = 0,1) gives the products 16 (n = 0, 1) with the sulfur atom in the ring; the five-membered ring is now formed 30 times faster than the six, still a rather modest rate advantage.

Full text of DOI:10.1039/a900623k

View Article Online / Journal Homepage / Table of Contents for this issue
Intramolecular nucleophilic displacement of halogen by
phosphinate and thiophosphinate anions: relative rates of
formation of five- and six-membered rings1
Amirah Chaudhry, Martin J. P. Harger,* Philippa Shuff and Alison Thompson
Department of Chemistry, The University, Leicester, UK LE1 7RH
Received (in Cambridge) 22nd January 1999, Accepted 30th March 1999
Intramolecular nucleophilic substitution transforms the phosphinate anions XCH2CH2(CH2)nCH2(Ph)P(O)OϪ
(n = 0, 1; X = Br, Cl) (Et3NHϩ salts; CH2Cl2 solution) into cyclic phosphinate esters 14 (n = 0, 1); unusually the
five-membered ring product (n = 0) is formed only 4.3 (X = Br) or 5.7 (X = Cl) times faster than the six (n = 1). The
analogous cyclisation of the thiophosphinate anions ClCH2CH2(CH2)nCH2(Ph)P(S)OϪ (n = 0, 1) gives the products
16 (n = 0, 1) with the sulfur atom in the ring; the five-membered ring is now formed 30 times faster than the six, still
a rather modest rate advantage.
The carboxylate anion (RCO2Ϫ) is one of the most important
and most studied functional groups in reactions involving
neighbouring group participation2 or intramolecular catalysis.3
By contrast little is known about the phosphorus analogue, the
phosphinate anion (R2PO2Ϫ). Phosphinate is less basic than
carboxylate (e.g. Et2PO2H, pKa 3.29),4 and presumably less
nucleophilic,5 but even the phosphonate dianion (RPO32Ϫ) (e.g.
EtPO3H2, pKa1 2.45; pKa2 7.85)6 has not been studied exten-
sively. That the dianion can provide catalysis seems clear since
the phosphono-substituted carboxylic ester 1 (n = 0 or 1; Ar =
p-nitrophenyl) is hydrolysed much faster than unsubstituted
CH3CO2Ar.7,8 The mechanism (Scheme 1) is thought to involve
1 (Scheme 1) the rate with n = 0 is only 1.5 times greater than
with n = 1.8 Phosphonate, it seems, may be an exception to the
rule. On the other hand, it could be that the hydrolysis of 1 is
inappropriate as a basis for generalisation. The case for the
phosphonate group acting as a nucleophile rather than a base is
persuasive, albeit that the anhydride 3 has not been detected;7,8
less certain is whether the measured rate (release of p-nitro-
phenoxide) is a reliable indicator of the rate of cyclisation (k1).
If return of the tetrahedral intermediate 2 (kϪ1) were (much)
faster with n = 0, relative to the rate-limiting breakdown to
anhydride 3 (k2), then the actual cyclisation (k1) might also be
(much) faster when n = 0 even though the release of p-nitro-
phenoxide is not.
O–
Generalisation ideally requires a reaction in which the prod-
uct of cyclisation can be observed directly and its formation is
the rate determining step. Also, to allow better comparison with
carboxylate, it would be preferable to have phosphinate anion
as the nucleophile instead of phosphonate dianion. Intra-
molecular alkylation (SN2) should be ideal, especially as the
corresponding cyclisation of halogeno carboxylates 6 to lac-
tones 7 is well known (Scheme 2).15 With phosphinate, however,
OAr
CO2Ar
k1
(CH2)n
(CH2)n
O
k–1
2–
PO3
P
O
–
O
1
2
k2 (–OAr–)
O
–
CO2
PO3
CH2X
H2O
(CH2)n
(CH2)n
O
(CH2)n
O
(CH2)n
(–X–)
2–
–
P
CO2
O
O
–
O
7
6
3
Scheme 2
Scheme 1
there is little encouragement to be had from intermolecular
reactions. Phosphinate and related monoanions have been
alkylated, using reactive alkyl halides (PhCH2X, MeX) and sil-
ver salts or crown ether catalysis,16 but the reverse is more often
encountered, i.e. dealkylation (especially demethylation) of PV
esters by SN2 attack of halide ion.17 It was therefore with some
uncertainty that we decided to explore the cyclisation of halo-
geno phosphinates and, for comparison, thiophosphinates.
the neighbouring phosphonate dianion acting as a nucleophile
and forming a cyclic anhydride intermediate 3 with a five (n = 0)
or six (n = 1) membered ring. As a rule cyclisations involving
functional groups separated by a saturated alkyl chain form
five-membered rings more quickly than six by a factor of
around 102.9–13 Typical examples are the intramolecular SN2
reaction of the aminoalkyl bromide 4, which forms the cyclic
CH2Br
CO2Ph
(CH2)n
(CH2)n
Results and discussion
Preparation of substrates
–
CH2NH2
CO2
5
4
Following literature precedent the 3-bromopropylphosphinate
ester 8 (n = 0, X = Br) was prepared by heating PhP(OEt)2 with
1,3-dibromopropane (Arbusov reaction).18 The same method
using 1,4-dibromobutane proved equally satisfactory for the
4-bromobutyl ester 8 (n = 1, X = Br). With a five-fold excess of
ammonium salt 100 times faster when n = 0 than when n = 1,
and the hydrolysis of the ester 5 (n = 0 or 1), which is 140 times
faster when the cyclic anhydride intermediate is five membered
rather than six.14 It is remarkable, then, that in the hydrolysis of
J. Chem. Soc., Perkin Trans. 1, 1999, 1347–1352
1347

View Article Online
Table 1 13C NMR spectra (CDCl3, 75.5 MHz)
δC (JPC/Hz)
Alkyl
Phenyl
ω-C
β-C ϩ γ-C
α-C
C-1
C-2 ϩ C-3a
C-4
Ph
33.7 (18.5)
25.3 (2)
29.2 (100)
131.4 (132)
131.0 (10)
128.4 (13)
132.1
Br(CH2)3P(O)
Br(CH2)4P(O)
Cl(CH2)3P(O)
Cl(CH2)4P(O)
Cl(CH2)3P(S)
Cl(CH2)4P(S)
OH
Ph
32.7
20.5 (3)
33.2 (15.5)
29.4 (99)
131.7 (133)
130.9 (10)
128.3 (13)
131.9
OH
Ph
45.0 (18)
44.1
25.2 (3)
27.9 (95)
29.6 (100)
34.4 (77)
36.2 (76)
131.5 (132)
131.9 (131)
133.4 (103)
133.6 (103)
131.0 (10)
128.4 (13)
132.1 (3)
132.0 (2.5)
132.1
OH
Ph
19.3 (3.5)
33.1 (13.5)
131.0 (10)
128.3 (13)
OH
Ph
44.7 (19.5)
44.0
25.8
130.5 (11.5)
128.4 (13)
OH
Ph
20.0 (2.5)
32.8 (17)
130.6 (11)
128.4 (13)
132.1 (3)
OH
O
P
Ph
70.4 (5)
24.3
25.6 (82)
26.3 (89)
36.2 (65)
31.6 (71)
130.7 (133)
130.9 (134)
133.3 (103)
132.7 (102)
131.2 (11)
128.4 (13)
132.4 (3)
O
O
66.7 (6)
26.5 (6)
20.0 (8)
130.7 (10)
128.3 (13)
132.2 (2.5)
132.2 (2.5)
132.4 (3)
P
Ph
O
P
O
36.6 (5.5)
27.6 (2.5)b
27.5 (2)
131.5 (11)
128.5 (13)
S
Ph
O
27.3 (5)b
21.9 (7)
130.4 (10)
128.5 (12.5)
P
Ph
S
a Which signal is due to C-2 and which to C-3 is not known. b The assignments of the signals δC 27.6 and 27.3 should possibly be reversed.
the dibromoalkane there were no obvious signs of further reac-
tion between the products 8 and PhP(OEt)2. The esters 8 were
not distilled (decomposition18) but the phosphinic acids 9
obtained on hydrolysis (48% HBr at 80–85 ЊC) (Scheme 3) were
introduced but, because of the high nucleophilicity of thioate
anions, it must be done in a way that produces the free acid, not
its conjugate base, if the risk of premature cyclisation is to be
averted. The phosphinic acids 9 were therefore converted into
the phosphinic chlorides 10 (X = Cl, Br) (Scheme 3) and the
P᎐O groups transformed into P᎐S by exchange with P S
CH2X
CH2X
CH2X
᎐
᎐
4
10
(COCl)2
HX–H2O
heat
(dioxane solution; DMF catalyst; reflux). The phosphinothioic
chlorides 11 were then hydrolysed in the absence of base. Thio-
phosphinyl compounds are much less reactive than their phos-
phinyl counterparts19 and hydrolysis (5% H2O in acetone)
required several hours to reach completion (31P NMR spec-
troscopy). The phosphinothioic acids were obtained as oils that
were difficult to purify. The chloro compounds 12 (X = Cl) were
sufficiently pure (n = 0, 90%; n = 1, >95%) for full spectroscopic
characterisation but the identity of the more reactive of the
bromo compounds (n = 0) could only be inferred from its sub-
sequent behaviour.
(CH2)n
(CH2)n
(CH2)n
O
O
P
O
P
P
Ph Cl
Ph OEt
Ph OH
10
8
9
P4S10
CH2X
CH2X
H2O
(CH2)n
(CH2)n
S
S
P
P
Ph Cl
Ph OH
Two details in the 13C NMR spectra of the phosphinic acids 9
and phosphinothioic O-acids 12 are noteworthy (Table 1): 1JPC
is ~25% smaller for the P᎐S compounds [sp3 C, 75 Hz (P᎐S) cf.
12
11
Scheme 3
᎐
᎐
100 Hz (P᎐O); sp2 C, 100 Hz (P᎐S) cf. 130 Hz (P᎐O)] and for
᎐
᎐
᎐
2
3
crystalline and quite easy to purify. The chloroalkyl esters 8
(n = 0, 1; X = Cl) were prepared in the same way from
PhP(OEt)2 and ClCH2CH2(CH2)nCH2Br and were hydrolysed
(37% HCl at 120 ЊC) to give the phosphinic acids 9 (X = Cl). In
principle the alternative esters 8 (X = Br) could have been
formed in the Arbusov reaction, by displacement of chlorine
rather than bromine, but in practise this was not a problem.
Obtaining the phosphinothioic O-acids 12 (X = Cl, Br)
presents something of a challenge. Sulfur obviously has to be
both types JPC (0–3.5 Hz) is much smaller than JPC (13.5–17
Hz). Similar features have been noted in the spectra of Bu3PO
and Bu3PS.20
Cyclisation reactions
The bromo phosphinic acids (δP ~ 44) in CH2Cl2 were converted
with Et3N into the phosphinate anions 13 (n = 0, 1; X = Br)
(δP ~ 27) and these passed slowly, but cleanly and completely, to
1348
J. Chem. Soc., Perkin Trans. 1, 1999, 1347–1352

View Article Online
Table 2 Rates of cyclisation of halogeno phosphinates 13 and thio-
a product δP 57 (n = 0) or 37 (n = 1). Relative to the acyclic
phosphinate ester 8 (δP ~ 44) the product chemical shift is 13
ppm downfield (n = 0) or 7 ppm upfield (n = 1). Such a differ-
ence is reasonable for the five- or six-membered cyclic phos-
phinate 14 (n = 0 or 1) (Scheme 4), given that a comparable
phosphinates 15 in CH2Cl2 at 35 ЊC
k/10Ϫ6 sϪ1
Ring size
13 (X = Br)
13 (X = Cl)
15 (X = Cl)
CH2X
5 (n = 0)
6 (n = 1)
178
41
2.5
0.44
230
7.8
(CH2)n
(CH2)n
O
O
O
(–X–)
–
P
P
Ph
14
Ph
13
O
(n = 0, X = Cl) was only 90% pure, however, it seems probable
that these were a consequence of the impurities rather than
byproducts of the cyclisation. One of the minor products had a
mass spectrum (Mϩ 214) suggestive of the cyclic ester 18 (n = 0)
CH2X
S
having both a P᎐S group and sulfur in the ring. This points to a
(CH2)n
(CH2)n
᎐
S
(–X–)
potential problem in our method of making the phosphino-
thioic O-acids (Scheme 3). If the phosphinothioic chloride 11 is
not adequately purified before hydrolysis any traces of P4S10 (or
P4S10 Ϫ nOn) will generate H2S which can react with 11 to give
the dithiophosphinic acid (12 with SH in place of OH).
–
P
P
Ph
16
O
Ph
O
15
Scheme 4
dependence of δP on ring size (bond angle) is seen with cyclic
phosphate esters.21 The products were isolated and purified and
their structures 14 (n = 0) and 14 (n = 1) confirmed spectro-
scopically. They are both known compounds having been
obtained previously in other ways.18,22
The chloro phosphinothioic acids (δP ~ 86) were similarly
transformed into the thiophosphinate anions 15 (n = 0, 1;
X = Cl) (δP ~ 63) and thence into products having δP 72 (n = 0)
or 39 (n = 1). In principle a thiophosphinate anion can act as a
nucleophile through the O or the S atom although in inter-
molecular reactions with alkyl halides (SN2) it is the S atom that
is always alkylated.23 The chemical shift of the product δP 39 is
Rates of cyclisation
The cyclisations of the bromo phosphinates 13 (X = Br) and
chloro thiophosphinates 15 (X = Cl) (Scheme 4) in CH2Cl2 at
35 ЊC were monitored by 31P NMR spectroscopy.‡ In each case
nine or ten spectra were recorded at regular intervals extending
to 85–90% completion (3–72 h). The chloro phosphinates 13
(X = Cl) were examined in the same way except that in one case
(n = 1) reaction was so slow it was followed only to 75% comple-
tion (40 days). From the relative peak areas of substrate and
product in each spectrum first order plots [log (a Ϫ x) vs. t] were
constructed. These were approximately linear and the slopes of
the lines afforded values of the rate constant k ( 8%) (Table 2).
The bromo thiophosphinates 15 (X = Br) were not included in
the study because of the very high reactivity of one of them
(n = 0) and our inability to obtain the acid reasonably pure
(>75%). Nonetheless, it was apparent from preliminary experi-
ments that they cyclise with half-lives of about 1 min and 40
min at 25 ЊC, and at 35 ЊC they will presumably react about
twice as quickly. Three aspects of the results (Table 2) require
comment.
too small for a P᎐S compound 17 so by implication it must be
᎐
(CH2)n
O
S
(CH2)n
S
S
P
P
Ph
Ph
18
17
the cyclic P᎐O compound 16 (n = 1) (Scheme 4) with endocyclic
᎐
sulfur. This was confirmed by the spectra of the isolated
Leaving group. For the phosphinates 13 (oxygen nucleophile)
the bromo compounds cyclise faster than the chloro com-
pounds by a factor approaching 102 [Br/Cl rate ratio: 71 for five-
ring formation (n = 0); 93 for six-ring formation (n = 1)]. For the
thiophosphinates 15 (sulfur nucleophile) the picture is broadly
similar although without better data for the bromo compounds
a precise comparison is not possible. Differences in reactivity of
this magnitude are much as expected for alkyl bromides and
chlorides undergoing SN2 reaction with anionic nucleophiles.24
material, notably νmax = 1190 cmϪ1 (P᎐O) and δ 3.38 and 2.89
᎐
H
(diastereotopic CH2S protons). The value of δP 72 for the other
product is less clear cut: it is larger than we would anticipate for
a P᎐O compound, even a five-membered cyclic one [contrast
᎐
δP 57 for 14 (n = 0)], but small for a P᎐S group in a five-
᎐
membered ring. From other features [νmax = 1190 cmϪ1 and δH
3.53 and 3.29 (CH S)], however, it is clear that this too is a P᎐O
᎐
compound 16 (n =20) with sulfur in the ring. Why δP should be
so large, and so different from the value for the six-membered
ring (∆δP = 33 ppm), is not clear; superficially it points to an
even greater contraction of the bond angle at phosphorus than
is usual for a five-membered ring and is present in 14 (n = 0).†
Ring size. With oxygen as the nucleophile the rate ratio for
formation of the five- and six-membered rings (k5/k6) is 4.3 for
the bromo phosphinates and 5.7 for the chloro compounds.
With sulfur as the nucleophile this ratio (k5/k6) is 30 for the
chloro thiophosphinates (and seemingly not dissimilar for the
1
For the cyclic esters 14 and 16, JPC is still 20% smaller for the
sulfur-containing compounds (Table 1) even though there is
now a defined P᎐O group and just a single bond to sulfur.
᎐
bromo compounds).
A 30-fold preference for the five-
Whereas 16 (n = 1) was the only product obtained from the
thiophosphinate 15 (n = 1, X = Cl), several minor products (2–
3% each) accompanied 16 (n = 0). Given that the substrate 12
membered ring is possibly not anomalous although it is cer-
tainly at the low end of expectation. The five-fold preference
seen with the phosphinates undoubtedly is anomalous however,
especially when compared with the corresponding cyclisation
of the carboxylates 6 (X = Br); there the k5/k6 ratio is 100 (Naϩ
† We are grateful to a referee for pointing out that δP 72 is not an
unreasonable chemical shift for the five-membered cyclic compound 16
(n = 0). It is within 20 ppm of the value of δP for the acyclic analogue
PhEtP(O)SEt (δP 53.3; M. Mikolajczyk and J. Luczak, J. Org. Chem.,
1978, 43, 2132) and a downfield shift of 15–20 ppm for a five-membered
ring is not uncommon. Also, it is within 35 ppm of the value of δP for
the six-membered cyclic analogue 16 (n = 1) (δP 39) and an upfield shift
of 5–15 ppm, relative to the acyclic case, is not uncommon for a
six-membered ring.
‡ The anions 13 and 15 were generated from the corresponding
acids using 1.33 equiv. Et3N. The 31P chemical shifts suggest that the
phosphinic acids are ~95% ionised under these conditions and the
phosphinothioic acids 100% (see Experimental). No allowance has
been made for incomplete ionisation in deducing the values of the rate
constant k (Table 2).
J. Chem. Soc., Perkin Trans. 1, 1999, 1347–1352
1349

View Article Online
salts in 99% DMSO).15 Of possible significance in these cyclis-
ations is the setting of the nucleophilic O atom, a tetrahedral
centre in a phosphinate but a trigonal centre in a carboxylate.
However, the alkoxides 19, like the phosphinates, have the
are uncorrected. 1H NMR spectra were recorded at 90 MHz on
a Varian EM 390 spectrometer or at 250 or 300 MHz on Bruker
ARX 250 or AM 300 instruments (Me4Si internal standard;
coupling constants J given in Hz); 13C NMR spectra were
recorded on the AM 300 at 75.5 MHz. 31P NMR spectra (1H
decoupled) were recorded at 36.2 MHz on a JEOL JNM-
FX90Q spectrometer (positive chemical shifts downfield from
85% H3PO4). Mass spectra were obtained in EI mode unless
otherwise indicated on a Kratos Concept spectrometer. CH2Cl2
was distilled from CaH2. Light petroleum refers to the fraction
bp 60–80 ЊC and ether to diethyl ether. Diethyl phenylphos-
phonite was prepared from PhPCl2 by reaction with EtOH–
pyridine in light petroleum and was purified by distillation, bp
60–62 ЊC at 0.05 mmHg.
CH2Cl
SO2Cl
(CH2)n
(CH2)n
CH2O–
19
CH2OH
20
nucleophilic oxygen attached to a tetrahedral centre and for
them25 the k5/k6 ratio of 200 is even larger than for the carb-
oxylates.
Nucleophile. As would be expected for an SN2 process, the
thiophosphinates, with sulfur as the nucleophile, cyclise faster
than the phosphinates, with oxygen.26 For the chloro com-
pounds the S/O rate ratio (kS/kO) is 92 for five-membered ring
formation and 18 for six (with the bromo compounds the
picture seems rather similar). Intuitively the larger value may
be thought normal, given that oxygen does not compete with
sulfur in the reactions of thiophosphinate anions with alkyl
halides.23 Model reactions suggest otherwise, however. The
intermolecular alkylation of Ph2P(S)OϪ with 1-bromopropane
occurs exclusively at the S atom [≤1% Ph2P(S)OPr] as expected,
but it is only ca. 22 times faster than the alkylation of
Ph2P(O)OϪ [t1/2 8.3 h and 185 h respectively at 35 ЊC for Et3NHϩ
salts with 1.0 mol dmϪ3 PrBr in CH2Cl2]. It seems, therefore,
that the smaller kS/kO ratio for the six-membered ring formation
is actually normal and that the value for five-membered ring
formation is anomalously large. The anomaly could result from
an unusually large value for kS but an unusually small kO seems
more likely.
Chloroalkyl(phenyl)phosphinic acids 9 (X ؍ Cl)
(a) 1-Bromo-3-chloropropane (23.6 g, 150 mmol) was stirred
vigorously at 150 ЊC (bath temp.) in a N2 atmosphere and
diethyl phenylphosphonite (5.94 g, 30 mmol) was added drop-
wise during 0.5 h. Heating was continued until 31P NMR spec-
troscopy showed that the phosphonite (δP 158.5) had all been
consumed (0.8 h). Volatile material was removed under reduced
pressure and the residue was distilled to give ethyl 3-chloro-
propyl(phenyl)phosphinate 8 (n = 0, X = Cl) (5.46 g, 74%), bp
120 ЊC (oven temp.) at 0.2 mmHg, δP(CH2Cl2) 42.7, still con-
taminated with some of the low-boiling byproduct (δP 45.3;
5%). This material was used in (b). A sample purified by redistil-
lation had δP(CDCl3) 43.6; δH(CDCl3, 90 MHz) 7.9–7.5 (5 H,
m), 3.95 (2 H, m, OCH2Me), 3.51 (2 H, br t, JHH 6, ClCH2), 2.3–
1.8 (4 H, m) and 1.33 (3 H, t, JHH 7, OCH2CH3); νmax(film)/cmϪ1
1220 (P᎐O); m/z 248, 246 (Mϩ, 4%) and 141 (Mϩ Ϫ C H Ϫ
᎐
2
4
C3H6Cl, 100) (Found: Mϩ, 246.0576. C11H1635ClO2P requires
M, 246.0576).
The same method starting from 1-bromo-4-chlorobutane
afforded ethyl 4-chlorobutyl(phenyl)phosphinate
8 (n = 1,
Conclusion
X = Cl) (70%), bp 125 ЊC (oven temp.) at 0.05 mmHg;
δP(CDCl3) 44.2; δH(CDCl3, 90 MHz) 7.9–7.4 (5 H, m), 3.95
(2 H, m), 3.47 (2 H, br t, JHH 6), 2.25–1.5 (6 H, m) and 1.31 (3
Two anomalies become apparent when the rates of cyclisation
of the halogeno phosphinates 13 and thiophosphinates 15 are
analysed: the ratio kO-5/kO-6 is unusually small and kS-5/kO-5 is
unusually large. The common factor in these is kO-5 and both of
the anomalies would disappear if kO-5 were larger. The conclu-
sion seems clear: intramolecular nucleophilic attack by a phos-
phinate anion is less favourable than expected when the product
is a five-membered ring. To a lesser extent the same may be true
for thiophosphinate (we have no data for thiocarboxylate with
which to make comparison) but the problem is certainly less
pronounced when the five-membered cyclic transition state con-
tains sulfur in place of oxygen. Longer bonds (P–S and S–C)
and a more accommodating bond angle (P–S–C) apparently
relieve the transition state of some strain.27 That there is strain
in the O-5 transition state seems certain,28 and the exceptionally
high reactivity (P–O bond cleavage) of five-membered cyclic
phosphonate (and phosphate) esters is most likely due in large
part to ring strain.29 Interestingly, King and Rathore have also
found an anomalously small kO-5/kO-6 ratio (2.2) for the spon-
taneous hydrolysis of the hydroxy sulfonyl chlorides 20 (n = 0,
1).30 Here, of course, the sulfonyl group is not the nucleophile
but again the five-membered cyclic product (a sultone) shows
exceptionally high reactivity attributable to ring strain.31 The
anomalies observed with the halogeno phosphinates are not
large but they are significant, in part because they reinforce the
conclusions drawn from the study of intramolecular nucleo-
philic catalysis by phosphonate dianion and in part because
intramolecular SN2 is a fundamentally important class of
reaction.
H, t, JHH 7); νmax(film)/cmϪ1 1215 (P᎐O); m/z 262, 260 (Mϩ, 1%),
᎐
225 (Mϩ Ϫ Cl, 100) and 141 (Mϩ Ϫ C2H4 Ϫ C4H8Cl, 60)
(Found: Mϩ, 260.0733. C12H1835ClO2P requires M, 260.0733).
(b) Ethyl 3-chloropropyl(phenyl)phosphinate
8
(n = 0,
X = Cl) (5.4 g, 21.9 mmol) was stirred and heated (bath temp.
120 ЊC) with concentrated (37%) hydrochloric acid (50 ml) for
4.5 h. When cool the aqueous layer was decanted from the oil
and was extracted with CH2Cl2; the extract was combined with
the oil and the resulting solution was washed with water, dried
(MgSO4) and concentrated to a solid. Crystallisation twice
from EtOAc–light petroleum afforded 3-chloropropyl(phenyl)-
phosphinic acid 9 (n = 0, X = Cl) (2.11 g, 44%), mp 82–83 ЊC;
δP(CDCl3) 42.9; δH(CDCl3, 250 MHz) 12.53 (1 H, s), 7.8–7.35
(5 H, m), 3.45 (2 H, t, JHH 6, ClCH2) and 2.05–1.85 (4 H, m);
νmax(Nujol)/cmϪ1 2620, 2200 and 1710 (all br, OH); m/z (CI)
221, 219 (M ϩ Hϩ, 10%) and 183 (M ϩ Hϩ Ϫ HCl, 100)
(Found: C, 49.4; H, 5.5. C9H12ClO2P requires C, 49.4; H, 5.3%).
In the same way ethyl 4-chlorobutyl(phenyl)phosphinate 8
(n = 1, X = Cl) gave 4-chlorobutyl(phenyl)phosphinic acid
(31%), mp 52–53 ЊC; δP(CDCl3) 43.7; δH(CDCl3, 250 MHz)
12.60 (1 H, s), 7.8–7.35 (5 H, m), 3.41 (2 H, t, JHH 6.5) and 1.9–
1.5 (6 H, m); m/z (CI) 235, 233 (M ϩ Hϩ, 75%) and 197
(M ϩ Hϩ Ϫ HCl, 100) (Found: C, 50.9; H, 5.95. C10H14ClO2P
requires C, 51.6; H, 6.1%. Found: Mϩ, 232.0420. C10H1435ClO2P
requires M, 232.0420).
Bromoalkyl(phenyl)phosphinic acids 9 (X ؍ Br)
Ethyl bromoalkyl(phenyl)phosphinates 8 (X = Br) were pre-
pared from 1,3-dibromopropane and 1,4-dibromobutane with
PhP(OEt)2 as in (a) above but the crude products were not dis-
tilled (decomposition18); rather, they were hydrolysed as in (b)
Experimental
Mps were determined using a Kofler hot-stage apparatus and
1350
J. Chem. Soc., Perkin Trans. 1, 1999, 1347–1352

View Article Online
above, but with concentrated (48%) hydrobromic acid [4.5 h at
80–85 ЊC (bath temp.)], to give the phosphinic acids:
Cyclisation reactions
(a) A mixture of 4-bromobutyl(phenyl)phosphinic acid 9 (n = 1,
X = Br) (111 mg, 0.40 mmol) and Et3N (54 mg, 0.53 mmol) in
CH2Cl2 (1.6 ml) maintained at 35 ЊC for 41 h gave a single
product (δP 37.0). The solvent was evaporated and the product
was separated from Et3NHBr by repeated extraction of the
residue with ether. Distillation gave 2-phenyl-1,2-oxaphos-
phinane 2-oxide 14 (n = 1) (75 mg, 95%), bp 150 ЊC (oven temp.)
at 0.2 mmHg, as an oil that solidified; crystallised from ether,
mp 84–85 ЊC (lit.,22 86 ЊC); δP(CDCl3) 37.0; δH(CDCl3, 300
MHz) 7.85–7.75 (2 H, m), 7.55–7.40 (3 H, m), 4.54 and 4.19
(both 1 H, m; CH2O) and 2.3–1.75 (6 H, m); νmax(Nujol)/cmϪ1
3-Bromopropyl(phenyl)phosphinic acid 9 (n = 0, X = Br),
crystallised from aqueous MeOH then from EtOAc–light
petroleum, mp 79–80 ЊC; δP(CDCl3) 42.6; δH(CDCl3, 300 MHz)
13.05 (1 H, s), 7.85–7.4 (5 H, m), 3.34 (2 H, br t, JHH 6, BrCH2)
and 2.1–1.9 (4 H, m); νmax(Nujol)/cmϪ1 2650, 2300 and 1710 (all
br, OH); m/z (ϪFAB) 263, 261 [(M Ϫ H)Ϫ, 15%] and 81, 79
(BrϪ, 100) (Found: C, 41.1; H, 4.6. C9H12BrO2P requires C,
41.4; H, 4.4%).
4-Bromobutyl(phenyl)phosphinic acid 9 (n = 1, X = Br), crys-
tallised from aqueous EtOH then from EtOAc–light petroleum,
mp 78–80 ЊC (lit.,32 76–77 ЊC); δP(CDCl3) 43.8; δH(CDCl3, 300
MHz) 13.3 (1 H, s), 7.85–7.35 (5 H, m), 3.28 (2 H, t, JHH 6.5),
1.9–1.7 (4 H, m) and 1.7–1.5 (2 H, m); m/z (ϪFAB) 277, 275
[(M Ϫ H)Ϫ, 15%] and 81, 79 (BrϪ, 100).
1220 (P᎐O); m/z 196 (Mϩ, 100%), 195 (25), 142 (Mϩ Ϫ C H ,
᎐
4
6
90), 141 (Mϩ Ϫ C4H7, 35) and 77 (35).
The corresponding reaction of 3-bromopropyl(phenyl)-
phosphinic acid 9 (n = 0, X = Br), reaction time 16 h, gave a
single product (δP 57.0) which was isolated in the same way.
Distillation gave 2-phenyl-1,2-oxaphospholane 2-oxide 14
(n = 0), bp 150 ЊC (oven temp.) at 0.3 mmHg (lit.,18 157 ЊC at 0.7
mmHg); δP(CDCl3) 58.0; δH(CDCl3, 300 MHz) 7.80–7.70 (2 H,
m), 7.60–7.45 (3 H, m), 4.56 and 4.31 (both 1 H, m; CH2O) and
Chloroalkyl(phenyl)phosphinothioic O-acids 12 (X ؍ Cl)
Oxalyl chloride (0.76 g, 6.0 mmol) was added in portions to a
stirred solution of 4-chlorobutyl(phenyl)phosphinic acid 9
(n = 1, X = Cl) (0.70 g, 3.0 mmol) in CH2Cl2 (7 ml). After 0.5 h
the phosphinic chloride 10 (n = 1, X = Cl) (δP 57.1) was isolated
by evaporation of volatile material and was dissolved in
dioxane (4.5 ml). The solution was stirred with P4S10 (333 mg,
0.75 mmol) and a catalytic amount of DMF (3 mg) at 120 ЊC
(bath temp.) for 2 h, when 31P NMR spectroscopy showed reac-
tion (O/S exchange) to be complete. The solvent was evaporated
and the residue was pumped in vacuo. Extraction of the residue
with CH2Cl2 afforded 4-chlorobutyl(phenyl)phosphinothioic
chloride 11 (n = 1, X = Cl), δP(CDCl3) 89.9; δH(CDCl3, 90 MHz)
8.2–7.3 (5 H, m), 3.51 (2 H, br t, JHH 6), 2.7–2.3 (2 H, m) and
2.1–1.5 (4 H, m); m/z 270, 268, 266 (Mϩ, 50%), 233, 231
(Mϩ Ϫ Cl, 85), 178, 176 (Mϩ Ϫ C4H7Cl, 100), 177, 175
(Mϩ Ϫ C4H8Cl, 40) and 145, 143 (Mϩ Ϫ C4H8Cl Ϫ S, 70). This
was dissolved in acetone (10 ml) containing water (540 mg,
30 mmol) and hydrolysis (δP 91.2→81.2) was allowed to
continue overnight (60% complete in 2.3 h at 28 ЊC). Volatile
material was evaporated. The crude product was purified by
extraction from ether (8 ml) into ice-cold aqueous NaOH (5
mmol in 9 ml H2O), immediate acidification of the extract
(7 mmol HCl in 3.5 ml H2O), and back-extraction into ether (10
ml, 2 × 5 ml), giving 4-chlorobutyl(phenyl)phosphinothioic
O-acid 12 (n = 1, X = Cl) (0.66 g, 88%) as an oil, δP(CDCl3) 86.2;
δH(CDCl3, 300 MHz) 7.95–7.8 (2 H, m), 7.6–7.4 (4 H, m;
includes OH), 3.49 (2 H, t, JHH 6.5), 2.25–2.05 (2 H, m) and
1.85–1.6 (4 H, m); νmax(film)/cmϪ1 1110 and 915; m/z 250, 248
(Mϩ, 30%), 213 (Mϩ Ϫ Cl, 100), 158 (Mϩ Ϫ C4H7Cl, 45), 157
(Mϩ Ϫ C4H8Cl, 50) and 125 (Mϩ Ϫ C4H8Cl Ϫ S, 60) (Found:
Mϩ, 248.0191. C10H1435ClOPS requires M, 248.0192).
2.5–1.9 (4 H, m); νmax(film)/cmϪ1 1215 (P᎐O); m/z 182 (Mϩ,
᎐
60%), 181 (25), 141 (Mϩ Ϫ C3H5, 100) and 77 (40).
(b)
A mixture of 4-chlorobutyl(phenyl)phosphinothioic
O-acid 12 (n = 1, X = Cl) (103 mg, 0.42 mmol) and Et3N (54 mg,
0.53 mmol) in CH2Cl2 (1.6 ml) was sealed in a glass ampoule.
After 8 days at ca. 37 ЊC the solution contained a single sub-
stantial product (δP 39.0) (97%). The solvent was evaporated
and the residue was extracted repeatedly with ether and was
then partitioned between CH2Cl2 and water. The combined
organic portions were chromatographed on silica gel (column:
30 mm × 9 mm), eluting with ether then with EtOAc. The later
fractions afforded 2-phenyl-1,2-thiaphosphinane 2-oxide 16
(n = 1) (68 mg, 77%), bp 150 ЊC (oven temp.) at 0.1 mmHg,
which slowly solidified; δP(CDCl3) 40.5; δH(CDCl3, 250 MHz)
7.95–7.8 (2 H, m), 7.6–7.45 (3 H, m), 3.38 and 2.89 (both 1 H,
m; CH2S) and 2.5–1.8 (6 H, m); νmax(Nujol)/cmϪ1 1190 (P᎐O);
᎐
m/z 212 (Mϩ, 100%), 184 (20), 179 (15), 158 (Mϩ Ϫ C4H6, 40),
157 (Mϩ Ϫ C4H7, 40), 125 (Mϩ Ϫ C4H7S, 50) and 77 (20). A
sample crystallised from ether had mp 74–75 ЊC (Found: C,
56.3; H, 5.9; Mϩ, 212.0425. C10H13OPS requires C, 56.6; H,
6.2%; M, 212.0425).
The corresponding reaction of 3-chloropropyl(phenyl)-
phosphinothioic O-acid 12 (n = 0, X = Cl) (90% pure), reaction
time 8 h at 35 ЊC and overnight at room temperature, gave one
principal product (δP 71.7) (~90%) and several minor products
(δP 85.6, 84.5, 83.6, 63.0; each 2–3%). The principal product was
isolated by extraction into ether and chromatography on silica
gel (elution with EtOAc; some minor products eluted first using
ether); it was identified as 2-phenyl-1,2-thiaphospholane
2-oxide 16 (n = 0), bp 150 ЊC (oven temp.) at 0.1 mmHg;
δP(CDCl3) 73.4; δH(CDCl3, 300 MHz) 8.0–7.85 (2 H, m), 7.6–
7.45 (3 H, m), 3.53 and 3.29 (both 1 H, m; CH2S) and 2.6–2.1
(4 H, m); m/z 198 (Mϩ, 95%), 157 (Mϩ Ϫ C3H5, 100) and 77 (40)
(Found: Mϩ, 198.0268. C9H11OPS requires M, 198.0268). The
oil did not solidify but when dissolved in ether crystals mp 102–
In the same way 3-chloropropyl(phenyl)phosphinic acid 9
(n = 0, X = Cl) was converted into 3-chloropropyl(phenyl)-
phosphinothioic O-acid 12 (n = 0, X = Cl), a waxy solid,
δP(CDCl3) 85.5 [impurity δP 90 (broad), 10%]; δH(CDCl3, 300
MHz) 8.13 (1 H, s, OH), 7.9–7.8 (2 H, m), 7.6–7.4 (3 H, m), 3.53
(2 H, t, JHH 6.5), 2.4–2.15 (2 H, m) and 2.15–1.9 (2 H, m);
νmax(Nujol)/cmϪ1 1110 and 910; m/z 236, 234 (Mϩ, 15%), 199
(Mϩ Ϫ Cl, 20), 198 (Mϩ Ϫ HCl, 75) and 157 (Mϩ Ϫ C3H6Cl,
100) (Found: Mϩ, 234.0035. C9H1235ClOPS requires M,
234.0035).
103.5 ЊC were formed, νmax(Nujol)/cmϪ1 1190 (P᎐O).
᎐
One of the minor products was obtained pure (GLC) by TLC
(silica gel; Rf 0.3 with 1:1 ether–light petroleum); the mass
spectrum suggested 2-phenyl-1,2-thiaphospholane 2-sulfide 18
(n = 0), m/z 214 (Mϩ, 100%), 173 (Mϩ Ϫ C3H5, 50), 105 (Mϩ Ϫ
S Ϫ Ph, 35) and 63 (40).
Bromoalkyl(phenyl)phosphinothioic O-acids 12 (X ؍ Br)
Using the same method as for the chloro compounds the bromo
phosphinic acids 9 (n = 1, 2; X = Br) were converted into the
bromo phosphinothioic acids 12 (n = 1, 2; X = Br). The high
reactivity of the anions (cyclisation) precluded purification by
extraction into aqueous NaOH and the acids (especially n = 0)
were not obtained in a sufficiently pure state for proper charac-
terisation. The crude acids were cyclised by addition of base
(ButNH2) to solutions in CH2Cl2.
Rate studies
The cyclisation reactions of the phosphinic acids 9 (n = 0, 1;
X = Br, Cl) and phosphinothioic O-acids 12 (n = 0, 1; X = Cl)
(0.075 mmol) in CH2Cl2 (0.3 ml) containing Et3N (0.10 mmol)
were carried out in sealed 5 mm NMR tubes supported in 10
mm tubes containing D2O (NMR lock). The probe of the
J. Chem. Soc., Perkin Trans. 1, 1999, 1347–1352
1351

View Article Online
3 W. P. Jencks, Catalysis in Chemistry and Enzymology, McGraw-Hill,
New York, 1969; M. L. Bender, Mechanisms of Homogeneous
Catalysis from Protons to Proteins, Wiley-Interscience, New York,
1971.
4 P. C. Crofts and G. M. Kosolapoff, J. Am. Chem. Soc., 1953, 75,
3379.
5 R. N. McDonald, A. K. Chowdhury, W. Y. Gung and K. D. DeWitt,
in Nucleophilicity, ed. J. M. Harris and S. P. McManus, American
Chemical Society, Washington, 1987, p. 90.
NMR spectrometer was maintained at 35 ЊC. Except for the
faster reactions the sample was removed from the probe and
placed in a water bath maintained at 35 ЊC between spectra. As
detailed in Results and discussion, 31P NMR spectra (1H
decoupled) were recorded at regular intervals and the inform-
ation so obtained was used to deduce the values of the rate
constant k for cyclisation (Table 2).
The extent to which the acids were ionised under the condi-
tions of reaction was deduced as follows. For each of the phos-
phinic acids 9 the 31P chemical shift was monitored as Et3N was
added in portions [5 or 6 × 0.35 equiv. then 3–5 × 0.7 equiv.] to
a CH2Cl2 solution. Relative to the free acid the shift to higher
field (ca. δP 44→26) was 17.5–17.8 ppm with 3–4 equiv. Et3N
(100% ionisation) and no further change occurred with addi-
tional Et3N. Under the conditions used in the rate study (1.33
equiv. Et3N) the shift to higher field (16.8–17.4 ppm) was ca.
95% of the maximum, indicating ca. 95% ionisation initially. In
general there was a small further shift (≤1 ppm) to higher field
as reaction progressed and the excess of the amine (relative to
the remaining substrate) became greater. Exceptionally in the
very slow reaction of 9 (n = 1, X = Cl) there was a gradual shift
to lower field (1 ppm at t = 28 days), probably because reaction
between the solvent (CH2Cl2) and Et3N gradually reduced the
excess of the amine. For the phosphinothioic O-acids 12 the
shift to higher field (ca. δP 86→63) was 23.3–23.4 ppm and this
maximum was achieved even with 1.33 equiv. Et3N, indicating
100% ionisation under the conditions of reaction.
6 K. H. Worms and M. Schmidt-Dunker, in Organic Phosphorus
Compounds, ed. G. M. Kosolapoff and L. Maier, Wiley-Interscience,
New York, 1976, vol. 7, p. 110.
7 S. L. Shames and L. D. Byers, J. Am. Chem. Soc., 1981, 103, 6177.
8 Y.-K. Li and L. Byers, J. Chem. Res. (S), 1993, 26.
9 C. J. M. Stirling, Tetrahedron, 1985, 41, 1613 and references cited
therein.
10 M. A. Casadei, C. Galli and L. Mandolini, J. Am. Chem. Soc., 1984,
106, 1051 and references cited therein.
11 F. G. Bordwell and W. T. Brannen, J. Am. Chem. Soc., 1964, 86,
4645.
12 D. F. DeTar and W. Brooks, J. Org. Chem., 1978, 43, 2245.
13 D. F. DeTar and N. P. Luthra, J. Am. Chem. Soc., 1980, 102, 4505.
14 E. Gaetjens and H. Morawetz, J. Am. Chem. Soc., 1960, 82, 5328.
15 G. Illuminati and L. Mandolini, Acc. Chem. Res., 1981, 14, 95.
16 Z. E. Golubski, Synthesis, 1980, 632; T. Furuta, H. Torigai,
T. Osawa and M. Iwamura, J. Chem. Soc., Perkin Trans. 1, 1993,
3139; J. Gigg and R. Gigg, Carbohydr. Res., 1996, 287, 263; but see
also D. V. Patel, K. Rielly-Gauvin and D. E. Ryono, Tetrahedron
Lett., 1990, 31, 5591.
17 R. Engel, in A Handbook of Organophosphorus Chemistry, ed. R.
Engel, Dekker, New York, 1992, p. 175. See also A. Holý, Synthesis,
1998, 381.
18 R. B. Wetzel and G. L. Kenyon, J. Org. Chem., 1974, 39, 1531.
19 A. A. Neimysheva, V. I. Savchik, M. V. Ermolaeva and I. L.
Knunyants, Bull. Acad. Sci. USSR, 1968, 2104; V. A. Baranskii,
G. D. Eliseeva and N. A. Sukhorukova, J. Gen. Chem. USSR (Engl.
Transl.), 1988, 58, 687.
20 T. A. Albright, W. J. Freeman and E. E. Schweizer, J. Org. Chem.,
1975, 40, 3437.
21 D. Gorenstein, J. Am. Chem. Soc., 1975, 97, 898.
22 S. Kobayashi, M. Tokunoh and T. Saegusa, Macromolecules, 1986,
19, 466.
23 R. S. Edmundson, in The Chemistry of Organophosphorus Com-
pounds, ed. F. R. Hartley, Wiley, Chichester, 1996, vol. 4, p. 433;
P. C. Crofts, in Organic Phosphorus Compounds, ed. G. M.
Kosolapoff and L. Maier, Wiley-Interscience, New York, 1973,
vol. 6, p. 44.
24 N. S. Isaacs, Physical Organic Chemistry, Longman, Harlow, 1987,
p. 385; A. J. Parker, Chem. Rev., 1969, 69, 1.
25 A. J. Kirby, Adv. Phys. Org. Chem., 1980, 17, 183.
26 F. G. Bordwell, T. A. Cripe and D. L. Hughes, in Nucleophilicity,
ed. J. M. Harris and S. P. McManus, American Chemical Society,
Washington, 1987, ch. 9.
Model reactions
Diphenylphosphinothioic O-acid (7 mg, 0.03 mmol) was dis-
solved in a 1.0 mol dmϪ3 solution of 1-bromopropane in
CH2Cl2 (0.5 ml), Et3N (0.04 mmol) was added, and the solution
was maintained at 35 ЊC. Examination by 31P NMR spec-
troscopy showed reaction (δP 54.8→41.5) to be 27% complete
at t = 3.85 h and 39% at t = 5.9 h (t1/2 ~ 8.3 h). When com-
plete (72 h) the product was isolated and characterised as
S-propyl diphenylphosphinothioate, δP(CDCl3) 43.4; δH(CDCl3,
300 MHz) 7.95–7.83 (4 H, m), 7.60–7.43 (6 H, m), 2.78 (2 H, dt,
JPH 11, JHH 7.5), 1.66 (2 H, sextet, JHH 7.5) and 0.94 (3 H, t,
JHH 7.5); νmax(film)/cmϪ1 1200 (P᎐O); m/z 276 (Mϩ, 30%), 234
᎐
(Mϩ Ϫ C3H6, 40), 202 (Mϩ Ϫ SC3H6, 100) and 201 (Mϩ Ϫ
SC3H7, 100). No product δP 60–110 was observed (≤1%) indi-
cating that no Ph2P(S)OPr was formed.
A similar experiment with diphenylphosphinic acid showed
reaction (δP 17.9→29.9) to be 25% complete at t = 70 h, 48%
at t = 178 h (t1/2 ~ 185 h). The product was isolated and con-
firmed to be propyl diphenylphosphinate, δP(CDCl3) 31.3;
δH(CDCl3, 250 MHz) 7.9–7.8 (4 H, m), 7.6–7.4 (6 H, m), 3.99
(2 H, dt, JPH ~ JHH ~ 7), 1.75 (2 H, sextet, JHH 7) and 0.98 (3 H,
27 A. S. Pell and G. Pilcher, Trans. Faraday Soc., 1965, 61, 71.
28 J.-C. Yang and D. G. Gorenstein, Tetrahedron, 1987, 43, 479.
29 G. R. J. Thatcher and R. Kluger, Adv. Phys. Org. Chem., 1989, 25,
99 and references cited therein; R. Kluger and S. D. Taylor,
J. Am. Chem. Soc., 1990, 112, 6669. For an alternative view, see
A. Dejaegere and M. Karplus, J. Am. Chem. Soc., 1993, 115, 5316.
30 J. F. King and R. Rathore, Phosphorus Sulphur, 1987, 33, 165.
31 E. T. Kaiser, Acc. Chem. Res., 1970, 3, 145.
t, JHH 7); νmax(melt)/cmϪ1 1220 (P᎐O); m/z 260 (Mϩ, 20%), 219
᎐
(Mϩ Ϫ C3H5, 100), 217 (Mϩ Ϫ C3H7, 45) and 201 (40).
32 D. G. Hewitt and G. L. Newland, Aust. J. Chem., 1977, 30, 579.
References
1 Preliminary communication: A. Chaudhry, M. J. P. Harger, P. Shuff
and A. Thompson, J. Chem. Soc., Chem. Commun., 1995, 83.
2 B. Capon, Quart. Rev., 1964, 18, 45.
Paper 9/00623K
1352
J. Chem. Soc., Perkin Trans. 1, 1999, 1347–1352