Phospho-transfer catalysis on the asymmetric hydrophosphonylation of aldehydes

Article abstract of DOI:10.1016/S0022-328X(97)00176-9

We report here a precise, in situ ³¹P{¹H}-NMR method of assaying enantiopurity of α-hydroxyphosphonate esters, the products of the carbonyl hydrophosphonylation (Pudovik) reaction. This method is based upon a diazaphospholidine chiral derivatising agent (CDA) which satisfies all of the criteria for a precise assay; (i) derivatisation of α-hydroxyphosphonate esters is both rapid and clean, (ii) kinetic resolution is absent and (iii) ³¹P{¹H} chemical shift dispersions are excellent (> 5ppm). Calibration of this assay has been achieved by cross-referencing the ³¹P{¹H}-NMR signals obtained for the CDA-derivatised ester of (MeO)₂PC=O)CHPh(OH) to optical rotation measurements from scalemic material obtained upon lipase catalysed hydrolysis (F-AP 15, Rhizopus oryzae) of (MeO)₂P(=O)CHPh(OAc). Analysis of NMR chemical shift and coupling parameters for a closely related series of derivatised α-hydroxyphosphonate esters support further configuration assignments on the basis of inference. We report also on the configurational stability of α-hydroxyphosphonate esters in the presence of acids, organonitrogen bases and metal salts. ²H-labelling and carbonyl crossover experiments reveal that low levels of epimerisation (< 2%) at the alpha-carbon atom (C_α) of α-hydroxyphosphonate esters is possible under certain conditions of catalysis and within certain limits. A design strategy for the construction of catalyst systems in the Pudovik reaction is outlined based upon a combination of Lewis acidic (E) and Lewis basic (N) sites. Four types of catalyst are outlined, members of two distinct Classes I and II according to the nature of the acid and base sites, along with our investigations of representative examples of each Class. A variety of Class I.1 systems based on β-amino alcohols (one hydrogen bonding E site and one organonitrogen N site), have been assayed in the model reaction between (MeO)₂P(O)H and PhCHO. Results suggest that catalysis of the Pudovik reaction is clean and efficient in certain cases but that catalytic activity is strongly dependent upon the nature of the basic (N) nitrogen centre. Moreover, only low levels (< 10%) of enantioselectivity are afforded by all amino alcohols assayed. Achiral variants of Class 1.2 catalysts (multiple hydrogen bonding E and/or N sites) have been examined to model carbonyl and H-phosphonate binding; an amphoteric receptor based on a pyridine dicarboxamide scaffold has been synthesised and shown to bind benzaldehyde > 50% more strongly (K₁₁ 0.53 mol^-1 dm³) than dimethyl-H-phosphonate (K₁₁ 0.34 mol^-1 dm³, 298 K) and to catalyse the hydrophosphonylation reaction between these two substrates with a second order rate constant comparable to that of triethylamine (both k₂ 5.9 × 10^-2 mol^-1 dm³ h^-1, 293 K). However, one of the major limitations of this model is that competitive product inhibition dominates after some 15 turnovers (75% completion). Model studies reveal that hydrophosphonylation catalysis via a nitrogen Lewis base is accelerated up to 10-fold upon the introduction of [Zn(OSO₂CF₃)₂] as co-catalyst. Consequently, Class II.1 systems employ metal salts [Zn(OSO₂CF₃)₂] as Lewis acidic E sites and chiral co-catalysts capable of binding to the metal and also acting as Lewis basic N sites. Such systems catalyse the addition of (MeO)₂P(O)H to PhCHO cleanly with modest turnover numbers (< 10 turnovers; 50% completion) but with enhanced enantio selectivity over Class I catalysts (< 40% e.e.). However, competitive product inhibition is still problematic. Class II.2 systems are related to Class II.1 but possess directly coordinated E and N sites with more basic N functions and consequently are far more active as catalysts than the other classes. This increased catalytic activity is exemplified by one of the simplest achiral members, diethylzinc which catalyses the 100% chemo- and regioselective addition of (MeO)₂P(O)H to PhCHO to afford (MeO)₂P(O)CHPh(OH) with an average turnover rate (over a 1 h reaction time at 298 K) of 115 h^-1 compared to ca. 1 h^-1 for NEt₃ under analogous conditions. Chiral variants are proposed.

Full text of DOI:10.1016/S0022-328X(97)00176-9

Ž
.
Journal of Organometallic Chemistry 550 1998 29–57
Phospho-transfer catalysis
On the asymmetric hydrophosphonylation of aldehydes
1
a
a
a
a
Stephen R. Davies , Michael C. Mitchell , Christopher P. Cain , Paul G. Devitt ,
b
a,)
Roger J. Taylor , Terence P. Kee
School of Chemistry, UniÕersity of Leeds, Leeds LS2 9JT, UK
a
b
Ciba Central Research, Ciba Geigy PLC, Hulley Road, Macclesfield, Cheshire SK10 2NX, UK
Received 13 December 1996
Abstract
We report here a precise, in situ 31P H -NMR method of assaying enantiopurity of a-hydroxyphosphonate esters, the products of the
Ä1
4
Ž
.
Ž
.
carbonyl hydrophosphonylation Pudovik reaction. This method is based upon a diazaphospholidine chiral derivatising agent CDA
Ž . Ž .
which satisfies all of the criteria for a precise assay; i derivatisation of a-hydroxyphosphonate esters is both rapid and clean, ii kinetic
31
Ž
.
Ä1
4
Ž
.
resolution is absent and iii P H chemical shift dispersions are excellent )5ppm . Calibration of this assay has been achieved by
31
Ä1
cross-referencing the P H -NMR signals obtained for the CDA-derivatised ester of MeO 2P 5O CHPh OH to optical rotation
4
Ž
.
Ž
.
Ž
.
Ž
.
measurements from scalemic material obtained upon lipase catalysed hydrolysis F–AP 15, Rhizopus oryzae of
Ž
.
Ž
.
Ž
.
MeO 2P 5O CHPh OAc . Analysis of NMR chemical shift and coupling parameters for a closely related series of derivatised
a-hydroxyphosphonate esters support further configuration assignments on the basis of inference. We report also on the configurational
stability of a-hydroxyphosphonate esters in the presence of acids, organonitrogen bases and metal salts. 2H-labelling and carbonyl
Ž
.
Ž
.
crossover experiments reveal that low levels of epimerisation -2% at the alpha-carbon atom Ca of a-hydroxyphosphonate esters is
possible under certain conditions of catalysis and within certain limits. A design strategy for the construction of catalyst systems in the
Ž
.
Ž .
Pudovik reaction is outlined based upon a combination of Lewis acidic E and Lewis basic N sites. Four types of catalyst are outlined,
members of two distinct Classes I and II according to the nature of the acid and base sites, along with our investigations of representative
Ž
examples of each Class. A variety of Class I.1 systems based on b-amino alcohols one hydrogen bonding E site and one organonitrogen
Ž .
.
Ž
.
N site , have been assayed in the model reaction between MeO 2P O H and PhCHO. Results suggest that catalysis of the Pudovik
Ž
.
reaction is clean and efficient in certain cases but that catalytic activity is strongly dependent upon the nature of the basic
N
nitrogen
Ž
.
centre. Moreover, only low levels -10% of enantioselectivity are afforded by all amino alcohols assayed. Achiral variants of Class I.2
Ž
.
catalysts multiple hydrogen bonding E andror N sites have been examined to model carbonyl and H-phosphonate binding; an
amphoteric receptor based on a pyridine dicarboxamide scaffold has been synthesised and shown to bind benzaldehyde )50% more
strongly K11 0.53 moly dm3 than dimethyl-H-phosphonate K11 0.34 moly dm , 298 K and to catalyse the hydrophosphonylation
1
1
3
Ž
.
Ž
.
reaction between these two substrates with a second order rate constant comparable to that of triethylamine both k2 5.9=10y moly1
2
Ž
dm3 hy1, 293 K . However, one of the major limitations of this model is that competitive product inhibition dominates after some 15
.
Ž
.
turnovers 75% completion . Model studies reveal that hydrophosphonylation catalysis via a nitrogen Lewis base is accelerated up to
. x . x
w
Ž
w
Ž
10-fold upon the introduction of Zn OSO2CF3 as co-catalyst. Consequently, Class II.1 systems employ metal salts Zn OSO2CF3 as
2
2
Lewis acidic E sites and chiral co-catalysts capable of binding to the metal and also acting as Lewis basic N sites. Such systems catalyse
Ž .
Ž
.
Ž
.
the addition of MeO 2P O H to PhCHO cleanly with modest turnover numbers -10 turnovers; 50% completion but with enhanced
Ž
.
enantioselectivity over Class I catalysts -40% e.e. . However, competitive product inhibition is still problematic. Class II.2 systems are
related to Class II.1 but possess directly coordinated E and N sites with more basic N functions and consequently are far more active as
catalysts than the other classes. This increased catalytic activity is exemplified by one of the simplest achiral members, diethylzinc which
Ž
.
Ž .
Ž
.
Ž .
Ž
.
catalyses the 100% chemo- and regioselective addition of MeO 2P O H to PhCHO to afford MeO 2P O CHPh OH with an average
)
Corresponding author.
1
It is a great pleasure to dedicate this article to Professor Ken Wade on the occasion of his 65th birthday in recognition of his many
outstanding contributions to main group chemistry. One of the authors considers himself honoured to have been taught, guided and deeply
influenced by Kens’ approach to science. Long may that approach continue.
0022-328Xr98r$19.00 q 1998 Elsevier Science S.A. All rights reserved.

(
)
30
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
y
1
turnover rate over a 1 h reaction time at 298 K of 115 h compared to ca. 1 hy1 for NEt3 under analogous conditions. Chiral variants
Ž
.
are proposed. q 1998 Elsevier Science S.A.
Keywords: Enantiopurity assay; Chiral derivatising agent; Absolute configuration; Enantioselective; Pudovik reaction; Configurational stability;
Amphoteric catalysts; Metal complexes
1. Introduction
Phospho-transfer catalysis refers to the process of
w
Ž
.
Ž .x
transferring the phospho function RO P O or a
2
derivative thereof, from one chemical site to another
Ž
. w x
mediated by a distinct catalytic species Scheme 1 1 .
One representative example of such a transformation
is the kinase and phosphatase catalysed phosphorylation
w
x
w x
of proteins involving P–O bond formation 2 , a fun-
damental process in the regulation of biological systems
the significance of which was recently demonstrated by
the award of the 1992 Nobel prize for medicine or
physiology to Edmond Fisher and Edwin Krebs.
Scheme 1. Phospho-transfer catalysis.
w
x
and modified oligonucleotides 4,10 . In many of these
applications stereochemistry plays a major role in deter-
A somewhat less well investigated relative of phos-
phorylation is that of phosphonylation, where instead of
forming a new P–O bond, the new linkage formed is a
w
x
mining the physiological properties of the system 10 .
One of the main themes within our programme of
phosphorus chemistry is the development of catalytic
enantioselective variants of the Pudovik reaction. Our
strategy towards this objective is based on addressing
w
x
w
x
Ž
.
phosphorus–carbon P–C bond XsC in Scheme 1 .
Whereas phosphorylation results in phosphate esters
Ž
.
XsO, Scheme 1 , phosphonylation results in phos-
Ž
.
Ž .
phonate esters XsC, Scheme 1 ; both classes of
which play profoundly important roles in biological and
medicinal chemistry. Phosphate-diesters for example
form the backbone of genetic materials DNA and RNA
and control signal transduction through the central ner-
four distinct problems; 1 being able to assay enantiop-
urity of the product a-functionalised phosphonate esters
Ž .
readily, conveniently and precisely, 2 the determina-
tion of absolute configuration within a-functionalised
Ž .
phosphonate esters, 3 the establishment of configura-
w x
vous system 3 whilst phosphonate esters, especially
those with secondary functionality in the X group, find
widespread application as mimics of phosphate esters
tional stability a-functionalised phosphonate esters un-
Ž .
der various conditions and 4 the development of a
catalyst design strategy for the Pudovik reaction. In this
paper we describe our solutions to problems 1–3 and
the foundations of our strategy for hydrophosphonyla-
tion catalyst design in order to address problem 4.
Portions of this work have been released in preliminary
w x
4 .
The addition of a dialkyl-H-phosphonate ester to an
aldehyde to afford an a-hydroxyphosphonate ester, the
Pudovik reaction Scheme 2 , is a phosphonylation pro-
Ž
.
w x
w x
form 11 .
cess which has been known for many years 5 , but only
relatively recently has interest been focused on the
development of asymmetric variants. This in turn has
grown from a need to access functionalised phosphonate
esters and phosphonic acids readily, cleanly, efficiently
and with close control over stereochemistry since func-
2. Results and discussion
2.1. Problem 1. Assaying enantiopurity of a-hydroxy-
phosphonate esters
Ž
tionalised phosphonate derivatives especially those with
.
functionality in the alpha position have been shown to
possess highly desirable physiological properties as, for
example, substitutes for amino carboxylic acids, the
2.1.1. Synthesis and application of a chiral deriÕatising
agent
Among the various methods available for the deter-
mination of enantioselectivity in scalemic compounds,
w x
w x
design of phosphono-peptides 6 , antibiotics 7 , antivi-
w x
w
x
ral agents 8 , enzyme inhibitors 6,9 , agrochemicals
Ž
.
Scheme 2. XsO, NR Rsalkyl, aryl .

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
31
Scheme 3. Some examples of organophosphorus derivatising agents.
31
Ä1
4
the analysis of diastereoisomers by NMR spectroscopy
and where analysis via P H -NMR spectroscopy re-
w
x
has a number of attractive features 12 . The technique
requires the use of a suitable chiral derivatising agent
veals chemical shift dispersions to be invariably higher
1
Ž
than those observed in H-NMR spectroscopy Scheme
Ž
.
. w x
3 13 . In addition to the non-racemic derivatives illus-
CDA to convert the enantiomers of a scalemic com-
pound to a pair of diastereoisomers which are then
differentiated on the basis of their NMR properties.
A number of considerations influence the choice of
Ž .
CDA in any specific application: i the CDA should be
Ž .
readily accessible and convenient to manipulate, ii
trated in Scheme 3, a number of achiral organophospho-
rus reagents have been found to be useful derivatising
w
x
agents 14 , operating on the basis of Horeau’s principle
w x
15 .
Several methods have been employed to assay enan-
tiopurity in a-hydroxyphosphonate esters, including,
reaction with the substrate should be facile and should
Ž
.
w x
proceed without kinetic resolution or racemisation, iii
Mosher and other chiral acids 16 a,b,e,f, circular
w x w x
w x
NMR spectra should provide a chemical shift dispersion
which permits effective base-line separation and accu-
dichroism 16 e, HPLC 16 g, peptide derivatives 16 c
w
x
and ammonium salts 16 d. Since for our purposes we
required a rapid flexible assay, we have selected a
Ž .
rate integration, iv due consideration should be given
Ž .
to data collection and processing, v the use of the
non-racemic phosphorochloridite CDA 2 derived from
X
w
Ž .
x
CDA should help in the assignment of absolute configu-
the chiral diamine N,N -bis 1- S -phenylethyl -1,2-eth-
Ž .
Ž .
rations and vi the CDA should be inexpensive.
ylenediamine 1 Scheme 4 based on the phosphorotri-
w
x
Arguably the most important criteria are those related
to the efficiency of derivatisation without racemisation
and accurate integration by ensuring adequate chemical
shift dispersion. For the assay of alcohols, amines,
thiols and chiral alkenes several CDAs based on
organophosphorus esters or acid halides have been used
over the last ten years which derivatise substrates cleanly
amidite CDA 3 reported recently by Feringa 17 for the
determination of enantioselectivity in a range of chiral
alcohols, thiols and amines. Indeed, the use of chiral
C2-symmetric diamine based auxiliaries in the design of
chiral organophosphorus derivatising agents has been
w
x
pioneered by Alexakis and co-workers 13 e,f.
We found that although CDA 2 is less stable towards
Ž .
Ž
.
Ž
. Ž .
Ž
.
Ž
.
Ž
. Ž .
Scheme 4. i 0.5 1,2-Cl2 CH2 2 , 1008C, 16 h 60% ; ii P NMe2 3, C6 H6, 96 h; iii PCl3, 2 NEt3, C7 H8, 258C, 16 h 76% ; iv H2O, NEt3
Ž
. Ž .
94% ; v Esters 6, CDCl3, 258C.

(
)
32
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
31
Ä1
4
Ž
.
Ž
Fig. 1.
P
.
H -NMR of 5f; in ppm. Asunreacted 2; BsP III resonances of 5fx and 5fy ; Csexpansion of B data collected as described in
Ž .
Section 4 ; DsP V resonances of 5fx and 5fy ; Es4 caused by trace hydrolysis of 2.
Ž
hydrolysis than 3 solutions of 2 in pre-dried CDCl3
It has been reported that the use of a strong base such
as triethylamine in conjunction with phosphorochlo-
ridite CDAs can lead to side reactions in the derivatisa-
w
x
begin to hydrolyse to the H-phosphonate ester 4 18
within one week at ca. 58C under an atmosphere of
dinitrogen the former is significantly more reactive
towards a-hydroxyphosphonate esters seconds versus
.
w
x
tion of some alcohols 13 e. In trial reactions with CDA
2, we have found that both triethylamine and the milder
Ž
.
several hours respectively at ambient temperature .
N-methylimidazole result in essentially no difference to
31
Ä1
4
Moreover, we find that CDA 2 is also a more facile
derivatising agent than commercially available Mosher
acid chloride. Thus, a CDCl3 solution comprising 2 and
triethylamine at concentrations of 0.125 M and 0.25 M
respectively reacts rapidly and quantitatively with a
range of racemic a-hydroxyphosphonate esters of the
measured P H integrations.
2.1.2. Chemical shift dispersion and diastereoisomer
resolution
Given the significantly wider chemical shift range
1
afforded by the 31 P nucleus over that of H, the former
Ž
.
.
Ž
.
Ž
.
form
M eO 2 P 5O CHR OH
6
to afford
has distinct advantages in terms of being able to supply
effective chemical shift dispersion and hence the base-
line separation necessary for precise integration. More-
Ž
Ž .
Ž
phosphorus III –phosphorus V adducts 5 as a mixture
of diastereoisomers x and y see Fig. 1 and Table 1 for
selected data . Under our normal assay procedure see
Section 4 CDA 2 is present in excess over esters 6 see
.
Ž
Ž
Ž
.
over, the presence of both phosphorus III and phospho-
.
Ž .
rus V nuclei in the same molecule permits a choice of
.
Fig. 1 and the results displayed in Table 2 have been
assay label.
obtained under these conditions. However, even when 2
Ž
.
In all derivatives of the type 5 we find Table 2 , as
is present as the limiting reagent, integration of 31 P H -
NMR resonances reveals no evidence for kinetic resolu-
tion.
Ä1
4
other workers have previously, that the greater chemical
Table 2
Ž
.
Ž .
Ž
.
Table 1
Measured Enantioselectvities of Racemic MeO 2 P O CHR OH
Derivatised using CDA 2
31
Ž
.
Ž .
Ž
.
P-NMR Parameters for Phosphonate esters, MeO P O CHR OH
2
Derivatised using CDA 2
a
b
Ž
.
Ž . a
Ž
.
Ž .
% 5x y% 5y
R
Dd P III
Dd P V
E.e. c
a
a
Ž
5x
.
Ž
5y
.
Ž .
d P V
Ž .
5x; 5y
Ž
.
R
d P III
d P III
D J xIy
a C6 H5
5.54
5.60
4.76
5.40
4.76
4.28
5.48
0.01
0.09
0.00
0.12
0.03
0.00
0.01
0.00
0.01
0.04
0.02
50.1:49.9
50.3:49.7
50.1:49.8
50.0:50.0
49.7:50.3
49.7:50.3
49.5:50.5
50.2:49.8
49.9:50.1
49.8:50.2
50.4:49.6
0.2
0.6
0.3
0.0
0.6
0.6
1.0
0.4
0.2
0.4
0.8
Ž
.
Ž
.
b 1-C10 H7
c 2-C10 H7
d 2-BrC6 H4
e 3-BrC6 H4
f 4-BrC6 H4
g 4-MeC6 H4
h 4-MeOC6 H4 5.38
i 2-O2 NC6 H4
j 4-O2 NC6 H4
Ž
.
Ž
.
a C6 H5
b 1-C10 H7
c 2-C10 H7
127.75 16.8 122.21 14.2 23.55; 23.54 q2.6
Ž . Ž .
127.87 17.1 122.27 14.5 23.77; 23.68 q2.5
Ž
.
.
Ž
Ž
.
.
128.13 16.5 123.37 15.1 23.47; 23.47 q1.4
Ž
d 2-BrC6 H4 126.95 14.3 121.55 13.1 22.93; 22.81 q1.2
Ž . Ž .
128.78 16.3 124.02 14.6 22.74; 22.71 q1.7
e 3-BrC6 H4
f 4-BrC6 H4
Ž
.
Ž
.
128.68 16.0 124.40 15.1 22.83; 22.83 q0.9
Ž
.
Ž
.
g 4-MeC6 H4 127.64 17.1 122.16 14.6 23.78; 23.77 q2.5
1.72
3.11
Ž
.
Ž
.
h 4-MeOC6 H4 127.91 17.7 122.53 15.0 23.85; 23.85 q2.7
Ž
.
Ž
.
i 4-O2 NC6 H4 126.34 10.0 124.62 13.0 21.75; 21.74 y3.0
k 2-Ph2 PC6 H4 5.87
Ž
.
Ž
.
j 2-O2 NC6 H4 129.73 14.6 126.62 14.8 21.84; 21.80 y0.2
a In CDCl3; 298 K; ppm; 101.268 MHz.
b
b
Ž
.
Ž
.
k 2-Ph2 PC6 H4 128.61 17.4 122.74 17.1 23.44; 23.42 q0.3
b High frequency resonance first. Determined by automated electronic
integration of resonances that had essentially identical peak shapes
a In CDCl3; 300 K; ppm; 101.268 MHz; JPP in parentheses Hz ;
3
Ž
.
Ž
<
.
52.6 mmol solutions with respect to phosphonate.
for each diastereoisomer Fig. 1 .
E.e. of 5 determined as %5xy%5y .
b 5
c
Ž .
<
J
s3.0 Hz observed only to the P III nucleus.
PP

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
33
Ž
.
shift dispersion D dp is found between the
each case was found to be within 1%. Following enan-
tiopurity measurement, the 31 P-NMR spectra were re-
acquired over 26 316 Hz centred at 100 ppm 0.62 s
Ž
.
Ž
.
phosphorus III nuclei ca. 1.7–5.8 ppm rather than the
Ž .
Ž
.
Ž
phosphorus V nuclei -0.2 ppm ; indeed, the phos-
Ž
.
phorus III dispersions are significantly larger than those
reported for related derivatives with a variety of other
alcohols, amines and thiols which are commonly found
acquisition; 1.6 Hz digital resolution; 32 K data points;
1
.
0.3 s pulse delay and broad-band H decoupling to
confirm complete conversion of 6 to 5, the presence of
w
x
Ž .
in the range 0.1–2.0 ppm 17 .
excess CDA 2 and to obtain phosphorus V chemical
shifts Table 1 .
Ž
.
2.1.3. Data collection and analysis
In order to assay enantiopurity with high precision, it
is necessary to be able to integrate the phosphorus III
resonances of diastereoisomers 5 accurately. This in-
volves not only being able to achieve a sufficiently
large base-line separation but also taking account of
possible differences in nuclear Overhauser effects and
nuclear relaxation times.
One of the principal problems with accurate integra-
tion in 31 P-NMR spectroscopy is the potentially large
range of spin-lattice relaxation times T1 found for 31 P
nuclei which in turn necessitates the selection of a
suitable delay time between pulses to allow for com-
Ž
.
Ž
.
w
x
plete relaxation 19 . For the purposes of measuring
enantiopurity of a-hydroxyphosphonate esters, it is the
difference in spin-lattice relaxation times between the
Ž
.
All spectra have been recorded at 300 K on a Bruker
ARX 250 spectrometer operating at 101.268 MHz for
phosphorus. Initial data collection comprised a spectral
width of 3049 Hz, centred on the region of the phospho-
measured P III nuclei of the two diastereoisomers DT1
which governs the choice of delay time. Consequently,
Ž
.
we have measured T1 times for the P III nuclei of
5a–k at 300 K in CDCl3 solvent using the inversion
Ž
.
Ž
.
rus III resonances and digitised with 8192 points. A
recovery technique see Section 4 and the results are
reproduced in Table 3 and Fig. 2.
Ž
.
pulse width of 3.8 ms was used ca. 738 tip angle
followed by acquisition over 1.34 s, affording a digital
resolution of 0.74 Hz, with inverse-gated 1H decoupling
and a pulse delay of 3 s to ensure complete relaxation of
all 31 P nuclei. Under these conditions, the enantiomeric
As is evident from the data in Table 3, the T1 times
of the phosphorus III nuclei for both diastereoisomers
Ž
.
w
Ž
.x
are very similar and DT1 P III is ca. 0.1 s. Conse-
quently, it should be possible to perform analytically
sound enantiopurity measurements via 31 P-NMR inte-
Ž
Ž
.
excess e.e. of racemic a-hydroxyphosphonate esters in
31
Ä1
4 Ž . Ž
P H stacked inversion recovery experiment; Bsexpansion of P III resonance region see Section 4 for details of
Fig. 2. Asfull scale
.
acquisition parameters ; Csexponential plot of signal intensity versus delay time for the resonance at dP 122.3 ppm.

(
)
34
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
Table 3
Ž
those used in the enantiopurity determinations vide
Ž
.
Ž .
Ž
.
Relaxation Parameters for Phosphonate esters, MeO 2P O CHR OH
Derivatised using CDA 2
.
infra and therefore should reflect accurately the relative
T1 values in the reaction systems.
Ž
.Ž . a
Ž
.Ž . a
Ž . b
w x w x
T1 2 T1 4
R
T1P III x
T1P III y
T1P V
Under the above conditions of measurement, the T1
values have analytical value in helping to select suitable
delay times for the determination of enantioselectivities.
In support of this, examination of Table 3 reveals that
the common components of each sample, unreacted
CDA 2 and by-product H-phosphonate 4 18 have
reproducible T1 values thus attesting to the relative
accuracy of the determinations for esters 6. Interest-
ingly, the ester with the smallest relative molecular
a C6 H5
2.5
2.2
2.1
2.3
2.4
2.3
2.3
2.4
2.2
2.0
2.2
2.3
2.2
2.1
2.1
2.2
2.3
1.7
4.7
4.0
3.8
4.2
4.3
4.2
4.3
3.9
3.9
4.1
2.8
2.7
2.7
2.5
2.6
2.8
2.8
2.6
2.7
2.8
2.7
2.6
2.2
2.2
2.1
2.1
2.2
2.1
2.2
2.2
2.1
2.1
2.1
b 1-C10 H7
c 2-C10 H7
d 2-BrC6 H4
e 3-BrC6 H4
f 4-BrC6 H4
g 4-MeC6 H4
h 4-MeOC6 H4 2.2
i 4-O2 NC6 H4 2.2
j 2-O2 NC6 H4 2.3
k 2-Ph2 PC6 H4 1.8
w
x
Ž
.
Ž
.
Ž
.
mass, 6a RMMs513 has the largest P III T1 2.5 s
Ž
.
whereas the most massive ester 6k, 696 has the small-
a Determined by inversion recovery in CDCl3; 52.6 mmol; 300 K;
101.268 MHz; units in seconds; values are the average of both lines
Ž
.
est 1.8 s ; possibly reflecting differences in the molecu-
w
x
lar correlation times of each molecule in solution 20 .
Ž
.
of the P III doublets.
b Values given are the average T1 values of both diastereoisomers.
Furthermore, there is a small but definite trend in the T1
Ž
.
values of the phosphorus III nuclei with respect to the
31 P-NMR resonance such that the derivatives with the
higher frequency resonances possess slightly larger T1
values, although the differences are very small ca. 0.1
s . Nevertheless, subsequent analysis of coupling con-
stants and chemical shifts vide infra also reveals trends
which suggest that those derivatives with the same
pattern of coupling constants, chemical shifts and relax-
ation times possess the same relative stereochemistry
and hence the same absolute configuration at the alpha-
carbon site.
Ž
.x
.
.
grations of the P III resonances using a delay time of
ca. 5=DT1 P III , ca. 0.5 s. Indeed, comparison of
w
Ž
Ž
spectra collected with a 0.5 s delay time between pulses
.
Ž
.
as in Fig. 1 and a 3 s delay reveals no significant
Ž
.
difference in the integrated intensities.
Although the samples for T1 measurements were
Ž
.
prepared and stored prior to examination under an
oxygen-free atmosphere of dinitrogen, the samples were
not contained in flame-sealed NMR tubes but in normal
push-cap tubes bound with cling-film. Consequently,
over the course of the inversion recovery experiment
2.2. Problem 2. Determination of absolute configura-
tions of a-hydroxyphosphonate esters
Ž
.
ca. 10 h , oxygen from the air is likely to have entered
the tube. Consequently, we do not wish to place too
much emphasis on the absolute T1 values, which may
Ž
be slightly lower than true values indeed, film-sealed
Being able to determine enantioselectivities readily
is, however, only one half of the story. We also need to
be able to determine absolute configurations at the
a-carbon atoms Ca of the a-hydroxyphosphonate es-
ters. To do this unambiguously would require single-
crystal X-ray analyses of the diastereoisomers 5 in each
samples which were allowed to stand in the laboratory
atmosphere for up to 24 h prior to NMR examination
afforded T1 values which were lower by ca. 0.5 s .
However, the inversion recovery experiments reported
here have all been run under the same conditions as
.
Ž
.
w
x
Ž .
Fig. 3. View of esters down the Ca–O bond emphasising the conformations and different dihedral angles between phosphorus III and
X
w x
Ž
.
Ž
.
Ž .Ž
.
Ž
.
Ž
.
P CHAr OR functions. L large sP O OMe 2 , M medium sAr, S small sH.

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
35
Table 4
case examined which is, in practise, an unrealistic
prospect.
However, with such a closely related family of com-
1
Ž
.
Ž . .
Ž
H-NMR Parameters for Phosphonate esters, MeO P O CHR OH
Derivatised using CDA 2
2
a
a
b
b
b
b
5
dx
dy
2 Jx
2 Jy
3Jx
3Jy
D2 J d D3J d
pounds, it may be possible to suggest absolute configu-
ration assignments by cross-referencing of the phospho-
a
b
c
d
e
f
g
h
i
5.17
6.00
5.34
5.76
5.12
5.11
5.13
5.11
6.50
5.24
5.20
6.11
5.38
5.95
5.16
5.14
5.17
5.16
6.58
5.26
14.6 15.2 8.0
15.2 15.6 7.3
14.8 15.2 7.5
14.6 15.2 7.6
15.0 15.4 7.6
14.8 15.3 7.6
14.2 14.6 7.7
14.1 14.6 7.6
16.6 17.0 4.0
16.2 16.7 6.1
7.5
6.2
6.8
7.0
7.3
7.2
7.2
7.3
3.7
5.6
8.6
y0.6 q0.5
y0.4 q1.1
y0.4 q0.7
y0.6 q0.6
y0.4 q0.3
y0.5 q0.4
y0.4 q0.5
y0.5 q0.3
y0.4 q0.3
y0.5 q0.5
y0.2 q0.8
rus III 31 P chemical shifts reported here to optical
rotation measurements on authentic samples reported
Ž
.
w
x
previously by Hammerschmidt and co-workers 21 .
Fortunately, one a-hydroxyphosphonate ester is com-
mon to the series studied by both Hammerschmidt and
Ž
.
Ž
.
Ž
.
ourselves, namely MeO P 5O CHPh OH 6a, and we
2
have therefore worked on the basis of this as a correla-
tion point. Due to the availability of only the one
correlation point, it has been necessary to examine
carefully spectroscopic features within the series of
adducts 5a–k so that we may infer absolute configura-
tions in the remainder of the series with reasonable
confidence.
j
k
6.61c 6.75 c 14.8 15.2 9.4
a Chemical shift of the P C HR OH methine resonance for the
diastereoisomers which correspond to the higher frequency
w x
Ž
.
Ž
.
x
and
Ž .
Ž
Ž
.
lower frequency y phosphorus III resonance respectively.
b
. Ž3
Coupling between phosphorus III
.
Ž . Ž2
.
J
J
and phosphorus V
and the methine hydrogen for diastereoisomers 5x and 5y in Hz.
c 4
X
w x
Ž
.
J
of 10 Hz observed between the P C HR OR methine reso-
PH
nance and the Ph2 P phosphorus atom.
2.2.1. Correlating absolute configurations
d Difference in couplings Hz between diastereoisomers x and y,
Ž
.
3
Examination of JPP coupling constants in di-
Ž
.
xyy .
Ž
.
astereoisomers 5a–k Table 1 reveals consistently
larger values for the diastereoisomer with the higher
frequency 31 P-NMR shift diastereoisomer x in all
cases except those derived from 2-NO2 and 4-NO2
substituted benzaldehydes. Since, this coupling parame-
Ž
.
consistency of steric interactions resulting from the
Ž
.
same relative and thus in this case absolute configura-
tion at the alpha-carbon atoms in this family of com-
pounds.
w
ter is likely to be influenced significantly by the P—O
—C—P dihedral angle through a Karplus relationship
22 , we interpret this trend in terms of a similar confor-
mational preference for the corresponding diastereoiso-
mer throughout the series. Energetically more favoured
staggered conformational isomers for each epimer are
outlined in Fig. 3. The slightly larger JPP values of the
x isomers may indicate a greater preference for the
anti-periplanar conformer a in these isomers, where the
x
The empirical data collected in Table 4 suggest that
w
x
the diastereoisomers corresponding to the high fre-
X
Ž
.
w x
Ž
.
quency phosphorus III and low frequency P CHR OR
methine resonances of compounds 5 have the same
stereochemistry for each derivative which in turn sug-
gests that if we are able to assign absolute configura-
tions to any one of these adducts then, by inference, we
have assigned the remainder.
3
Ž .
larger L substituents occupy positions of maximum
w
x
separation 23 .
2.2.2. Calibration of the enantiopurity assay
1
Ä31
4
P -NMR spectra selec-
2
Ž
Ž
.
Ž
.
Examination of the
H
We have prepared scalemic MeO 2 P 5O -
3
.
Ž
.
tively decoupled of 5a–k reveals that JPH and JPH
CHPh OH using a slightly modified version of the
enzyme-mediated method reported by Hammerschmidt
Ž . Ž2
.
coupling constants between phosphorus V
J
and
Ž
. Ž3
.
w x
w
x
Ž
.
Ž
.
w
Ž .
x
phosphorus III J and P C HRO-, exhibit subtle yet
significant trends such that the diastereoisomers of 5
21 . Racemic MeO P 5O CHPh OC O Me is read-
2
Ž
.
Ž
.
Ž
.
ily obtained upon treatment of MeO P 5O CHPh OH
2
Ž
.
Ž
Ž
.
with the high frequency phosphorus III resonance ster-
eoisomer x correlate with the low frequency methine
resonance and that the low frequency methine resonance
consistently displays the lower JPH D Jxyy negative
and higher JPH D Jxyy positive by ca. 0.3–1.0 Hz
Table 4 and Fig. 4 . Such trends seem reasonable since
the JPH couplings, just as the JPP parameters de-
scribed above, are envisaged to be strongly dependent
upon the dihedral angle P III –O–C–H illustrated in
Fig. 3 which, in turn, will be a weighted average of all
possible conformations principally a, b and c . Conse-
quently, these trends in NMR chemical shift and cou-
pling parameters imply a consistency of molecular con-
formation for each diastereoisomer x and y due to a
6a with acetyl chloride and triethylamine Scheme 5 .
Enzymatic hydrolysis of MeO 2 P 5O CHPh-
.
Ž
.
Ž
.
w
Ž
.
Ž
x
OC 5O Me with lipase F—AP 15 affords scalemic
2
2
Ž
.
Ž
.
.
Ž
.
MeO P 5O CHPh OH in 33% yield after column
2
3
2
Ž
.
Ž
chromatography NMR analysis of the unpurified prod-
uct mixture reveals an extremely clean, essentially com-
plete 50% conversion . Polarimetry returns an optical
Ž
.
3
3
.
w x
Ž
.
rotation a of y30.48 cs0.5, acetone, 208C con-
D
w
Ž
.
x
sistent with the dominant enantiomer having the S
configuration and in agreement with the published pro-
Ž
.
w
x
cedure 18 . Compositional analysis of this scalemic
product using the CDA 2 reveals an e.e. of 83 1 %, in
Ž .
which it is the low frequency 31 P resonance which
dominates Scheme 5 . Consistently, hydrolysis of the
Ž
.

(
)
36
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
unconverted acetylphosphonate ester with MeOHrNEt3
Ž
since phosphonate esters are susceptible towards base-
catalysed transesterification, the presence of methanol
.
ensures that any exchange is degenerate affords
Ž
.
Ž
.
Ž
.
w x
MeO P 5O CHPh OH with an optical rotation a
2
D
Ž
.
Ž .
of q29.28 cs0.5, acetone and an e.e. of 81 1 %
CDA 2 in which it is the high frequency 31 P resonance
Ž
.
which dominates. Consequently, we may assign the
lower frequency resonance to the S enantiomer and the
higher frequency resonance as belonging to the R enan-
Ž
.
tiomer and by inference vide supra we propose the
same order of assignment in all the other benzaldehyde
derivatives described herein.
2.3. Problem 3. Configurational stability of a-hydroxy-
phosphonate esters
Ž
Since the catalysts used here are amphoteric vide
.
infra , it is necessary to confirm that a-hydroxyphos-
phonate esters are configurationally stable under condi-
tions commensurate with our catalytic protocols.
Racemisation may be envisaged to be possible through
two principal mechanisms as shown in Scheme 6; a
deprotonation at Ca followed by non-face-selective re-
X
w x
Ž
.
Fig. 4. P CHR OR methine hydrogen resonances of both di-
Ž .
Ž
.
Ž
astereoisomers x low frequency multiplet and y high frequency
multiplet of 5d and 5c.
A
Fully coupled 1H-NMR spectra of 5d
.
Ž .
1
Ž
.
Ž
. Ž . Ä31
left and 5c right . B
4
H
P spectra, decoupled selectively at the
Ž .
protonation of an intermediate prochiral enolate and b
1
Ž
. Ž .
Ž . Ä31
4
higher frequency phosphorus III x resonance. C
decoupled selectively at the lower frequency phosphorus III
H
P spectra,
. Ž .
Ž .
reversible aldehyde extrusion. Mechanism a is perhaps
Ž
y
X
w x)
Ä
w
Ž .
x
the less likely of the two but nevertheless, we have
p ro b e d th is p o s s ib ility b y tre a tin g
resonance.
diazaphospholidine.
P
s N, N -bis 1- S -phenylethyl -2-chloro-1,3,2-
31
Ž
.
Ž
.
Ž .
Ž
.
Ä1
4
P H -NMR spectra of derivatised fractions.
Scheme 5. Lipase F-AP 15 catalysed hydrolysis of rac- MeO 2 P O CHPh OAc and partial

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
37
Ž .
Ž
.
Scheme 6. Possible mechanisms for base-catalysed racemisation of a-hydroxyphosphonate esters. i Base 0.5 mol equiv. NEt3 ; CDCl3, D2O;
q
Ž .
Ž
.
Ž .
ii Base; iii aldehyde extrusion; iv RCHO,H .
Ž
.
Ž .
Ž
.
Ž
.
Ž
.
Ž .
Ž
.
Ž
. Ž
.
MeO P O CHPh OH with triethylamine 1 equiv. in
MeO P O CHPh OH , Zn OSO2CF3 10 mol% and
2
2
Ž 2
.
Ž
.
CDCl3 containing excess ca. 2–30 equivs. D2O and
testing for incorporation of deuterium at Ca. No such
incorporation was observed over a period of three weeks
at ambient temperature. The operation of reversible
extrusion mechanism b was probed by a crossover
experiment in which MeO P O CHPh OH dP 24.5
ppm and 4-BrC6 H4CHO 0.8 equiv. were mixed in
CDCl3 solvent in the presence of triethylamine
equiv. . After two weeks at ambient temperature ca.
208C , P H and H-NMR spectroscopy did reveal
D2O 12.5 equiv. in CDCl3 was allowed to stand under
an atmosphere of nitrogen for 17 days at ambient
temperature and although HrD exchange occurred at
Ž
.
the hydroxy site broad resonance at dH 3.2 , no incor-
Ž .
poration of deuterium at the alpha carbon site was
observed by H-NMR spectroscopy. Furthermore, when
the same experiment was performed in the presence of
NEt3 1 mol equiv. this result was unaffected. The
alternative mechanistic pathway, reversible aldehyde ex-
trusion, has been probed by treating
1
Ž
.
Ž .
Ž
. Ž
.
Ž2
.
Ž
Ž
.
1
.
Ž
31
1
.
Ä1
4
Ž
.
Ž .
Ž
.
Ž
evidence for the formation of the expected product of
crossover, MeO P O CH CHC6 H4 Br-4 OH
24.0 ppm although in only very low levels -2% ;
MeO P O CHPh OH and 4-BrC6 H4CHO 1 mol
2
Ž
.
Ž .
Ž
.Ž
.
Ž
.
Ž
.
Ž
.
10 mol% both in the
dP
.
equiv. with Zn OSO2CF3
2
2
.
Ž
Ž
.
absence and presence of NEt3 1 equiv. ; the former
protocol results in -2% aldehyde exchange after 17
days at 258C whilst the latter protocol produced ca. 10%
exchange after 7 days at 258C. We conclude that
racemisation via the expected mechanisms may poten-
tially be problematic under certain conditions but may
be minimised if i both Lewis acid and Lewis base
catalysts are used at low loading levels and ii hy-
drophosphonylation is performed on a timescale of hours
rather than days under ambient conditions.
Ž
.
addition of
Ž .
a
sample ca. 20 mg of pure
.Ž
Ž
.
Ž
.
MeO P O CH CHC6 H4 Br-4 OH to the final prod-
2
uct mixture confirmed the presence of this material in
the reaction mixture suggesting that although such a
mechanism is possible under the conditions applied it is
extremely slow. Consequently, this lack of evidence for
facile base-catalysed racemisation of a-hydroxyphos-
phonate esters suggests that the stereochemical integrity
of the Pudovik products is maintained under conditions
compatible with catalysis as long as the reaction times
Ž .
Ž .
Ž
.
are held to days and below rather than weeks. Thus,
since most of the catalytic runs reported here with
amino alcohols are complete on the timescale of hours,
we can be confident of operating under a regime in
which product racemisation is not a serious problem.
Since the most effective catalyst systems contain
metal ions vide infra , the possibility of metal-cata-
lysed racemisation of the product a-hydroxyphos-
phonate esters was examined. Such a process may be
facilitated by a similar overall scenario to Scheme 6 in
which binding of the a-hydroxyphosphonate ester to a
metal centre through a chelate ring should be favourable
2.4. Problem 4. Catalyst design in the asymmetric
PudoÕik reaction
2.4.1. Introduction
The Pudovik reaction may be run under several
w
x
w x
experimental protocols; thermal 25 , sonochemical 26 ,
Ž
.
w
x
w
x
w x
basic 27 , acidic 28 and metal-mediated 29 . With the
exception of the ultrasound catalysed process, which is
thought to proceed via a radical mechanism, the gener-
ally accepted mechanism for the Pudovik reaction in-
3
3
w
x
volves the attack of a s l phosphorus nucleophile 30
w
x
Ž
on the carbon atom of the C5X reagent XsO, S,
w
x
.
24 and should result in a lowering of the pKa of
adjacent hydrogen atoms. Thus, a solution comprising
CR2 , NR . In order to facilitate this reaction, it is
necessary to generate a lone-pair of electrons on phos-

(
)
38
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
phorus by a formal reduction of the s 4l5 H-phos-
a single metal centre acts as the Lewis acid due to
Ž
.
w
Ž .x w
x
phonate starting material Scheme 2 . Very few proto-
cols have combined this redox chemistry with the intro-
duction of asymmetry through chiral catalysis and those
that have, invariably involve complex andror aerobi-
cally-sensitive catalyst systems which afford modest
chelate ring formation Scheme 7 ix 24 . The problem
should be less severe with a bimetallic catalytic site but
is always likely to be problematic when the binding
profiles of both substrates and products are similar.
In Scheme 7 are highlighted a selection of interaction
motifs between reagents and products of the Pudovik
reaction and putative catalysts. The binding topology of
Ž
. w x
enantioselectivities e.e. values 31 .
Ž .
the catalyst sites is built around both Lewis acidic A
and Lewis basic B functionalities; they are amphoteric
Ž .
2.4.2. Catalyst design strategy
w
x
The binding and activation of carbonyl molecules
molecules 11 b. Consequently, different classes of cata-
lyst may be envisaged depending upon the nature and
disposition of A and B sites. These different classes
may both interact differently with substrates and reagents
and may suffer from differing degrees of competitive
w
x
may be achieved using either a single 32 or multiple
Lewis acidic binding site based on either hydrogen-
Ž
w
x
w x
bonding 33 , coordinate dative bonding 34 or a com-
w
x.
bination of the two 35 . Similarly, a Lewis acid should
be capable of binding the phosphoryl oxygen atom of an
H-phosphonate ester and lowering the pKa such that
w
Ž . Ž .x
product inhibition e.g. Scheme 7 ix – x .
We have selected two classes of hydrophosphonyla-
tion catalyst for examination, each class divided further
w
x
deprotonation should be significantly more facile 36 .
In a similar fashion, it is possible to bind and activate
both substrates simultaneously at the same or different
Lewis acid sites. This has the advantage of manipulat-
ing the Pudovik reaction towards an intramolecular
addition wherein the enthalpy of binding both substrates
offsets the unfavourable entropy involved in bringing
three species together in a single transition state. How-
ever, simultaneous activation also presents a potentially
serious problem in an addition process such as the
Pudovik reaction; the product a-hydroxyphosphonate
ester is also able to bind to the catalytic site thus leading
to competitive product inhibition. This problem may be
expected to be particularly serious in those cases where
Ž
into two sub-classes I.1, I.2 and II.1 and II.2 Scheme
8 . Class I.1 systems are based on a pair of binding
.
sites, one Lewis acidic and one Lewis basic, in which
the acidic site is a hydrogen-bond donor and the basic
Ž
.
site is a hydrogen-bond acceptor organo-nitrogen base .
Class I.2 systems are analogous but contain multiple
Ž
interaction sites e.g. two acid and one base site or one
acid and two base sites . Class II.1 and Class II.2
.
systems both contain a single acid and single base site,
the Lewis acid function being a metal ion and the base
function being either an organo-nitrogen, organo-oxygen
or organo-carbon group. The principal difference be-
tween the two classes is that in the former, the Lewis
Scheme 7. Possible interaction modes between substratesrproducts of the Pudovik reaction and putative catalytic sites.

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
39
Scheme 8. Four classes of amphoteric catalyst system for the hydrophosphonylation of carbonyls. Class I.1: One hydrogen-bonding Lewis acid
Ž .
Ž .
A and one hydrogen-bonding Lewis base B site. Class I.2: Multiple hydrogen-bonding Lewis acid andror Lewis base sites. Class II.1: One
metal ion Lewis acid and one Lewis base site, which may act also as a donor ligand to the metal. Class II.2: One metal ion Lewis acid and one
Lewis base site, the latter an oxygen, carbon or nitrogen function bound covalently and directly to the metal ion.
base site doubles as a chiral ligand whose interaction
with the Lewis acidic metal is reversible whilst in the
latter class, Lewis acid and Lewis base sites are directly
bonded covalently within a chiral environment Scheme
.
8 .
phosphonate to 2-nitrobenzaldehyde with an enantiose-
w
x
lectivity -28% 16 a. Since amino alcohols are exam-
ples of amphoteric Class I reagents containing Lewis
Ž
.
Ž
Ž .
basic nitrogen and hydrogen-bonding Lewis acidic
Ž
.
hydroxyl group functions, we envisage that both func-
tions play a part in facilitating this transformation.
Indeed, in their original report the authors comment that
esterification of the hydroxy group results in complete
loss of stereoselectivity. In our hands, we were unable
to achieve the same levels of both catalytic turnover and
2.4.3. Class I.1 catalysts
In 1983 Wynberg and co-workers reported that chiral
amino alcohols, such as the cinchona alkaloids quinine
and quinidine, catalysed the addition of dimethyl-H-
X
Ž
.
Fig. 5. Rsalkyl, aryl; BsBase. 8 quinine. 9a RsMe, 9b RsC6 H4CH2 , 9c Rs3,5- MeO 2C6 H2CH2. 10 RsMe. 11a RsR sMe; 11b
X
X
X
RsH, R si Pr; 11c RsH, R sn Bu; 11d RsH, R sEt.

(
)
40
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
Ž .
Ž
Ž
.
. Ž .
Ž .
Scheme 9. i RCHO RsC6 H5 a series; 3,5- MeO 2C6 H3 b series ; ii LiAlH4; iii 2-CHOC4 H3NH.
enantioselectivity of this reaction under the conditions
reported. However, changing solvent from toluene to
CH2Cl2 and moving from 2-nitrobenzaldehyde to un-
substituted benzaldehyde resulted in clean catalysis. In
addition, we have assayed a range of amphoteric chiral
amino alcohol systems as catalysts for the model addi-
tion reaction of dimethyl-H-phosphonate to benzal-
dehyde, the results of which are described below.
Ž .
Ž
. Ž . w
x
Ž
. Ž .
Scheme 10. i SeO2 , Ac2O, reflux 8 h; NaOH, H2O; steam distil 78% ; ii H3NOH Cl, EtOH, NaOAc, 48 h 258C 88% ; iii LiAlH4, Et2O,
Ž
. Ž Ž . Ž . Ž . Ž . Ž .
.
X
Ž .
3 h 258C; NaOH, H2O; iv EtO 2CsO;
v Recrystallise Et2O–hexane 1:2 vrv 31% ; vi NaOH, H2O; vii MeI xs , NaOH
X
X
q
i
Ž
. Ž
.
Ž
. Ž
.
Ž
.
RsR sMe ; viii R NH2 , H R s Pr, Bu, Et; RsH; 65–83% ; ix NaBH4, EtOH 42–58% .

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
41
Table 5
2.4.3.1. Syntheses of chiral amino alcohols. We have
selected a number of chiral amino alcohols, 8–11 as
shown in Fig. 5, which vary in several important fea-
Ž
.
Ž .
Rate and Selectivity Data in the Catalysed Addition of MeO 2 P O H
to PhCHO
Catalysta Solvent Yield Time ATRb
Ž .
E.e. c,d RrS d
Ž .
Ž .
Ž .
tures, i steric manifold, ii flexibility and iii the
number of hydrogen bonding interactions possible.
Amino alcohols 8–10 are based on a flexible, chiral
backbone where the nitrogen-bound substituent may be
varied. The camphor backbone of 11a–d provides a
more rigid framework which may, in turn, promote
more intimate contact between the amphoteric catalyst
and the active phosphonylating species. Amino alcohols
9–11 have been prepared by modifications of literature
Ž
.
%
h
Ž .
8
8
8
9a
9b
9c
10
THF
CH2Cl2 )95 96
mix e
)95 96
toluene )98 96
)95 96
)99
)99
)99
)102
1.4
1.2
)544
14
)348
)413
)413
y
21
39
2.9 2
S
S
S
y
y
y
S
S
S
y
y
y
S
S
S
y
R
S
S
Ž .
5.4 3
Ž .
4.5 2
0 f
y
y
toluene
toluene
2
2
144
168
Ž .
toluene )98 18
toluene 92
toluene )94 27
toluene )99 24
toluene )99 24
1.0 1
Ž .
11a
11b
11c
11d
23 h
23 i
23 j
23 k
24
648
8.3 7
Ž . g
1.8 5
0 f
Ž
.
w
x
procedures Schemes 9 and 10 37–43 .
0 f
y
toluene
0
24
2.4.3.2. Assaying Class I.1 catalysts in the asymmetric
PudoÕik reaction. Compounds 8–11 have been screened
Ž .
toluene 18
toluene 42
toluene 99
toluene -8
toluene )97 0.25
toluene )95 0.1
toluene )95 0.1
168
216
0.3
26 1
Ž .
42 1
Ž
.
as catalysts for the hydrophosphonylation at ca. 296 K
Ž .
66000
3
4 1
y
Ž
.
of benzaldehyde PhCHO using dimethyl phosphite
Ž .
)500
24 l
8 l
)77600m 6 1
w
Ž
.
x
Ž .
Ž .
MeO P O H , each at 2.0 M concentration with a
2
)190000m 9 1
Ž
.
catalyst concentration of 0.2 M concentration 10 mol%
11a l
)190000m 2 1
Ž .
unless noted otherwise. Solvent systems, reaction times,
yields, enantioselectivities and absolute configurations
are reproduced in Table 5 and complete details pre-
sented in Section 4.
a All catalytic runs were performed at 298 K with PhCHO and
Ž .
Ž
.
MeO 2 P O H substrates at 2.0 M in the solvents indicated and 10
mol% catalyst loading in each run except those with 11a and 23
catalysts which were 5 mol%.
One potential problem associated with the use of
secondary amino alcohol catalysts such as 11bId is
the possibility of oxazolidine formation between the
aldehyde and the catalyst, in a similar manner to estab-
b
Average turnover rate hy1 =10
q3
Ž
.
.
c
2
Ä
Ž
.
Ž
In %; e.s.d values determined using the equation Sn xi y x rn n
.41r2
x
Ž
.
y1
where x is the mean value of n at least 5 determinations
w
51 .
31
d Determined by
Ä1
4
P H -NMR spectroscopy using CDA 2.
Ž
.
lished reactions with ephedrine see Scheme 9 . Such a
side-reaction would give rise to a nitrogenous catalyst
with a significantly different structural profile to that
desired, specifically it would remove the hydroxyl hy-
drogen and consequently the possibility of further hy-
drogen bonding. Amino alcohol catalysts 11c and 11d
were indeed found to react with benzaldehyde generat-
e
Ž
.
CH2Cl2:toluene 1:1 vrv .
f No enantioselectivity observed by the assay procedure of CDA 2.
g Sample analysed after washing with toluene to remove the catalyst
Ž
.
see text .
No added Zn OSO2CF3 or NEt3.
h
Ž
.
2
i
Ž
.
Ž
.
Zn OSO2CF3
5 mol% .
5 mol% and NEt3 5 mol% .
5 mol% and NEt3 1 equiv. .
2
j
Ž
.
Ž
. Ž .
. Ž .
Zn OSO2CF3
2
.
k
Ž
Ž
Zn OSO2CF3
2
Ž
.
ing the corresponding oxazolidines 21a,b Scheme 12
l ZnEt2 at 5 mol%.
m These represent minimum values; actual ATR values may be
substantially higher.
during catalysis albeit in low yields. Treatment of amino
alcohol 11c with benzaldehyde in toluene solution at
Ž
808C furnished the corresponding R-oxazolidine con-
.
firmed by n.O.e. experiments 21 in ca. 95% diastereos-
electivity. This mixture was shown subsequently to be a
poor catalyst for the Pudovik reaction, less than 1% of
a-hydroxyphosphonate ester being detected after 48 h.
Since amino alcohol 11c affords 100% conversion to
a-hydroxyphosphonate ester within 24 h 258C , we can
be confident that oxazolidine 21 does not influence the
observed stereoselectivity of the catalytic run. We pre-
sume that this inability to catalyse reaction effectively is
connected to steric effects of the nitrogen base being
contained within a ring. Indeed, we have subsequently
exploited the lack of catalytic activity of polycyclic
and the corresponding amino alcohol through both NH
Ž
.
and OH hydrogen bonds possibly as in 7, Fig. 5 .
However, in order for this chelation to result in effec-
tive stereocontrol, it is necessary for the hydrogen-bond-
ing interactions to induce a pronounced conformational
preference in the alkoxy residues of the phosphito anion
which in turn controls the facial approach of a carbonyl
substrate. For this to be successful, three criteria must
Ž
.
Ž .
be met; i chelate ring formation via hydrogen bonding
should not be fluxional on the timescale of the phospho-
Ž .
nylation reaction, ii hydrogen bonding should occur
Ž
oxazolidines in Class II.1 and II.2 catalyst systems vide
infra .
preferentially at only one of the phosphite alkoxy groups,
.
Ž
.
failure to do so would lead to diastereoisomers, and iii
Our working hypothesis for Class I.1 catalysts is that
enantioselectivity results from intimate chemical inter-
action between the s 3l3 secondary phosphite tautomer
effective control over the conformations of the alkoxy
residues may be best achieved through sterically de-
manding substituents on the amino alcohol in close

(
)
42
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
proximity to the hydroxy function. We believe that the
most effective levels of stereocontrol in the Pudovik
reaction require that all three criteria be satisfied.
As revealed in Table 5, none of the amino alcohols
screened were found to afford significant stereoselectiv-
ity in the trial Pudovik reaction. Although it is difficult
to pin-point precise reasons for the poor selectivity
several comments can be made in the light of the above
criteria of our working hypothesis.
too far removed from the phosphorus atom to influence
aldehyde face control effectively. Consequently, we have
moved away from Class I.1 catalysts based on simple
chiral amphoteric amino alcohols and focused attention
on developing more potent systems.
2.4.4. Class I.2 catalysts
Whereas Class I.1 catalysts comprised essentially
two hydrogen-bond interactions, Class I.2 systems are
envisaged to have multiple interaction sites and, ulti-
mately, a more rigid binding interaction profile. How-
ever, in our initial investigations of Class I.2 type
catalysts we wished to lay the foundations for more
advanced systems by developing a model system capa-
Wynberg et al. report the highest enantioselectivities
e.e. 28% in the quinine and quinidine catalysed Pu-
dovik reaction with o-nitrobenzaldehyde whilst o-chlo-
Ž
.
w
x
w x
robenzaldehyde returns an e.e. of only 10% 16 a, 44 .
In light of this, an e.e. of 2–5% in the quinine catalysed
hydrophosphonylation of benzaldehyde is perhaps not
unreasonable considering the reduction in steric demand
at the reactive carbonyl carbon atom. These authors also
report that hydroxyl group binding is important since
acylation of this function in quinine results in zero
enantioselectivity and that enantioselectivity increases
as the size of the phosphorus-bound alkoxy substituent
Ž .
ble of i binding the carbonyl via hydrogen bonding
Ž .
better than the H-phosphonate ester, ii activating the
Ž
.
H-phosphonate and iii facilitating the hydrophospho-
nylation reaction. Considerations of enantioselectivity
were not addressed in this model study since we needed
to answer four important questions before tackling the
Ž .
more involved process of building in asymmetry. 1
Ž
.
R increases 7 in Fig. 5 consistent with a greater steric
effect through preferred orientations of these functions.
We had envisaged that, should hydroxyl hydrogen
bonding be important, solvent effects may influence
enantioselectivity. Specifically, moving from a hydro-
gen bond acceptor solvent such as THF to toluene or
CH2Cl2 should result in increased enantioselectivity.
Could we design a compound which would interact with
a planar Cs symmetric carbonyl such as benzaldehyde
more strongly than a pyramidal Cs symmetric substrate
Ž .
such as dimethyl-H-phosphonate; 2 would that com-
Ž .
pound also activate the H-phosphonate substrate; 3
would that compound act as a hydrophosphonylation
Ž .
catalyst and 4 is competitive product inhibition likely
Ž
.
This does appear to be the case Table 5 although the
increase is small. Two further scenarios may also be
envisaged to contribute to low enantioselectivities; the
to be a problem? Each of these questions is addressed
below.
w
x
Compound 22 is an example of a —B—A—B—
type molecule containing two basic dimethylamino
w
x
relatively weak nature of the hydrogen bonding 45
y1
Ž
.
Ž
.
each hydrogen bond is worth ca. 4–40 kJ mol
may
functions B-sites tethered to a pyridinedicarboxamide
framework. The pyridine nitrogen atom results in pre-
orientation of the carboxamide functions through hydro-
result in decomplexation and consequently a lessening
of the influence that the chiral functionalities can have
over the active site phosphorus atom and secondly, even
in chelate ring formation, the chiral functions may be
w
x
gen bonding 46 whilst still being available to hydrogen
bond to both carbonyl and H-phosphonate substrates
Ž .
i
Ž . Ž .
dimethyl-H-phosphonate; ii benzaldehyde and iii
Scheme 11. Amphoteric receptor 22. Possible modes of interaction and activation of
carbonyl and phosphito anion simultaneously.

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
43
Table 6
1H-NMR titration data for the interaction of amphoteric receptor 22
with dimethyl-H-phosphonate
a
b
b
w
x
wŽ
.
wŽ
.
2
c
Receptor
MeO
Ž . x
-
dH
DdH
1r MeO - 1rDdH
2
a
Ž . x
P O H
P O H
0.01
0.01
0.01
0.01
0.01
0.01
0.01
y
8.01
y
y
y
0.095
0.148
0.150
0.167
0.199
0.353
8.34 0.33
8.52 0.51
8.55 0.54
8.59 0.58
8.70 0.69
9.09 1.08
10.53
6.76
6.67
5.99
5.03
2.83
3.03
1.96
1.85
1.72
1.45
0.93
a Mol dmy3
b ppm.
.
c Moly1 dm3.
w
Ž .
Ž .
x
Scheme 11 i and ii . The carboxamide hydrogens
were identified in the 1H-NMR spectrum as a broadened
Ž
triplet resonance coupling to the adjacent methylene
hydrogens at dH 8.20 D1r2 17 Hz, 0.1 M, CDCl3,
298 K which diminished in intensity upon treatment
.
Ž
.
with D2O due to chemical exchange. Having synthe-
sised 22, we must address Problem 1; interaction stud-
ies.
Fig. 6. Double reciprocal plots of the 1:1 binding isotherms for the
interaction of amphoteric receptor 22 with ligand L in C6 D6 solvent
An interaction between the phosphonyl oxygen atom
Ž
. Ž .
Ž
.
Ž . Ž .
300 K .
a
Ls MeO 2 P O H and b
LsPhCHO. Data is anal-
Ž
.
Ž .
of MeO P O H and the carboxamide hydrogens of
2
Ä
w x 4
ysed according to the equation; 1rDs1r D11PK11P L q1rD11
o
1
receptor 22 is revealed from H-NMR titrations experi-
ments possibly as illustrated in Scheme 11 . Thus, the
Ž
.
see Section 4 for further details .
Ž
.
carboxamide hydrogen chemical shift is displaced from
y2
Ž
.
dH 8.01 C6 D6, 1=10
M to sequentially higher
frequency upon being treated with increasing amounts
of H-phosphonate Table 6 . Similar H-NMR titration
binding in enzyme systems may typically be several
orders of magnitude greater than this , it is clear that the
1
Ž
.
.
Ž
experiments with benzaldehyde as guest molecule pos-
interaction of receptor 22 with benzaldehyde is quantita-
.
sibly as shown in Scheme 11 reveal similar perturba-
tions in carboxamide chemical shifts Table 7 . Subse-
quent construction of the respective 1:1 isotherms, as
Ž
.
Ž .
tively similar to that with MeO P O H, although ca.
2
Ž
.
1.5 times larger in the case of the aldehyde. The signifi-
cance of such small differences in interaction constant
have yet to be investigated.
w
x
double reciprocal plots 47 , support 1:1 complex forma-
tion although in both cases, interaction of receptor and
We chose to tackle Problems 2 and 3 together since
active catalysis would entail activation of the H-phos-
guest is relatively weak, K11 P s0.34 mol dm3 and
y1
Ž .
y1
3
Ž .
Ž
.
K11 A s0.53 mol dm respectively Fig. 6a,b . Thus,
Ž
phonate reagent the Pudovik reaction does not proceed
Ž
although host–guest interactions are weak host–guest
at a measurable rate at ambient temperature in the
.
absence of 22 . As expected, compound 22 is an effec-
tive catalyst for the Pudovik reaction between PhCHO
Table 7
Ž
.
Ž .
Ž
.
and MeO P O H 2 M in each component in dry
2
1H-NMR titration data for the interaction of amphoteric receptor 22
with benzaldehyde
toluene solvent at 298 K under a dinitrogen atmosphere,
at a catalyst loading of 5 mol%, affording a-hydroxy-
a
a
b
b
c
w
x
w
x
w
x
Receptor
PhCHO
dH
DdH
1r PhCHO
1rDdH
Ž
.
Ž .
Ž
.
phosphonate ester MeO P O CHPh OH as the sole
2
0.01
0.01
0.01
0.01
0.01
0.01
0.01
y
8.01
y
y
y
product. Under these conditions, ca. 40% of 22 will be
0.065
0.118
0.129
0.143
0.147
0.238
8.37 0.36
8.65 0.64
8.72 0.71
8.78 0.77
8.82 0.81
9.16 1.15
15.38
8.47
7.75
6.99
6.80
4.20
2.78
1.56
1.41
1.30
1.24
0.87
w
x
bound by H-phosphonate and ca. 51% by aldehyde 48
but unfortunately, the interaction constants are not suffi-
ciently large to achieve catalyst saturation at suitable
concentration levels, higher concentrations lead to prod-
uct precipitation, so that our kinetic analysis described
below was not achieved under saturation conditions.
a Mol dmy3
b ppm.
.
Analysis of rate data according to second order kinet-
y2
moly1
c Moly1 dm3.
w
x
ics 49 affords a rate constant of 5.9=10

(
)
44
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
quently, under these conditions substrate binding via
pyridinedicarboxamide hydrogen bonding alone is in-
sufficient to facilitate the Pudovik reaction in the ab-
sence of a base. A cartoon reflecting how 22 may
catalyse hydrophosphonylation is outlined in Scheme
Ž
.
11 iii .
Problem 4, the possibility of competitive product
inhibition, may also be addressed through the kinetic
data above. Notice how the data in Fig. 7a support
linearity only over ca. 2 half-lives of reaction, corre-
sponding to 75% conversion or 15 catalytic turnovers.
When these data are extended beyond two half-lives as
in Fig. 7c, deviations from second order behaviour
become apparent as the reaction becomes retarded by a
factor greater than second order kinetics predicts. That
this behaviour does not appear in the corresponding
Ž
reaction catalysed by NEt3 Fig. 7b is plotted over three
.
half-lives suggests that deactivation of reaction is due
to competitive product inhibition caused by the binding
Ž
.
Ž .
Ž
.
of MeO P O CHPh OH to 22. Indeed, this proposal
2
1
is supported by H-NMR studies which show that the
carboxamide hydrogen resonance in 22 is perturbed
from dH 8.01 to 8.17 ppm in the presence of 1 mol
3
Ž
.
Ž .
Ž
.
equivalent of MeO P O CHPh OH
.
2
Thus, we have addressed four major problems in-
volving Class I.2 catalyst systems based on the
pyridinedicarboxamide architecture which lead us to
Ž .
conclude that; i effective catalysis is slow but clean
with hydrogen-bonded Lewis acid and nitrogenous
Ž .
Lewis base catalysts, ii competitive product inhibition
may be problematic.
2.4.5. Class II.1 and II.2 catalysts. Metallo-organic
systems
Fig. 7. Second order kinetic plots for the reaction between PhCHO
w x
Ž
.
Ž . w x
Ž
.
A and MeO 2 P O H P in toluene solvent 300 K , according to
Both Class I.1 and I.2 catalysts described above
contain hydrogen-bonding Lewis acid sites. One of the
major problems with this approach is that catalysis is
too sluggish to be performed at low temperature with
the result that the hydrogen bonding network possesses
insufficient binding and orientating capacity to influ-
ence stereoselectivity. We need to increase both sub-
strate–catalyst binding and the rate of catalysis. In this
regard an electrophilic metal ion is a far more potent
Lewis acid than a hydrogen bond. Indeed, the most
active hydrophosphonylation catalysts yet reported have
been based on rather exotic metallo-organic systems
w x x w x w x
w
w
x
the equation 1r P sk2Ptq1r Po P P o s A o s2.0 M; Catalyst s
Ž . Ž .
Ž
.
0.1 M.
a
Catalysts22 ca. 2 half-lives, ca. 15 turnovers ;
b
Ž
. Ž .
catalystsNEt3 ca. 3 half-lives, ca. 17 turnovers ; c catalysts22
Ž
.
ca. 3 half-lives, ca. 17 turnovers .
dmy3 hy1 which is comparable to that obtained when
Ž
.
using NEt3 as a catalyst 5 mol% under exactly analo-
gous conditions Fig. 7a,b 2. Neglecting any inherent
differences in basicity, since 22 contains two basic
functions per molecule, 22 is approximately half as
active a catalyst as is NEt3. Indeed, the presence of
amino groups in 22 is crucial for catalytic activity, since
a modified receptor, in which both dimethylamino
groups are replaced by methyl groups does not catalyse
the reaction at all under the above conditions. Conse-
Ž
.
w
x
31 . In seeking to develop a strategic approach to
catalyst design, we have started from the premise that,
just as in hydrogen-bonded Class I systems, both Lewis
acidic and Lewis basic functions are necessary compo-
nents. Indeed, we were supported in this premise by our
2 The same catalytic reaction also proceeds cleanly under aerobic
3 Unfortunately, solubility problems have prevented us from mea-
suring the interaction constant.
conditions, returning a second order rate constant of 4.7=10y2
moly1 dm3 hy1
.

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
45
n
Scheme 12. i BuLi, HC O NMe2 , Et2O 54% ; ii CH2OH 2 , Hq, C6 H6 72% ; iii NiCl2 PPh3 2 , Zn, NEt4 I, THF 30% ; iv HClaq
Ž .
Ž .
wŽ
Ž
X
. Ž . Ž
.
Ž
. Ž
.
Ž
.
Ž
. Ž .
X
q
Ž
. Ž .
. Ž
.
x Ž
.
Ž
.
77% ; v 11c, H , C7 H8 4R,4 R : 4S,4 S s24:1; 32% ; vi PhCHO, C7 H8 4R:4Ss24:1; 56% .
4
w
Ž
.
x
Ž
observation that whilst zinc triflate Zn O3SCF3
10
metal and the results of using 23 in the hydrophospho-
.
Ž
.
2Ž .
Ž
.
Ž . .
Ž
mol% failed to catalyse the addition of MeO P O H
nylation of PhCHO with MeO P O H reveal Table 5
that 23 is inactive as a catalyst 5 mol% in the absence
of a metal salt but catalysis results on the addition of a
2
2
Ž
.
to PhCHO in CH2Cl2 solvent under ambient conditions,
addition of a catalytic 10 mol% quantity of NEt3
resulted in 9.1 turnovers within 60 min. For compari-
son, in the absence of the metal salt, NEt3 alone 10
mol% resulted in 1.5 turnovers during the same period.
Ž
.
Ž
.
Ž
w
Ž
. x
catalytic 5 mol% quantity of Zn O3SCF3 to afford
2
Ž
Ž
.
Ž .
.
MeO P O CHPh OH with an e.e. of 26%, the highest
2
.
we had yet obtained. Further improvements in enantios-
electivity to 42% resulted from the co-addition of a
catalytic 5 mol% amount of NEt3 although selectivity
subsequently decreases upon a large increase in the
amount of NEt3 presumably due to competitive achiral
catalysis. However, turnover rates were extremely slow
w
Ž
.2 x
A similar result was observed also with Zn O3SCF3
Ž
.
Ž
.
and pyridine, where pyridine 5 mol% alone did not
catalyse phosphonylation over a period of 44 h but upon
w
Ž
.
2 x
Ž
.
addition of Zn O3SCF3
5 mol% , 11 turnovers were
achieved within 44 h. Clearly, the amphoteric systems
are some 6–10 times more active, under the above
conditions, than base alone.
We are examining two types of Class II catalyst
system, II.1 and II.2 which differ in the nature of
interaction between Lewis acid and Lewis base sites.
Our investigations of a particular Class II.1 catalyst are
described here along with preliminary observations on
more active, but achiral, Class II.2 systems.
Ž
.
Table 5 and examination of preliminary rate data
revealed that competitive inhibition was a problem.
Nevertheless, these results suggested to us that we had
both a viable strategy and chiral environment, we needed
to strengthen the interaction between chiral ligand and
metal so that a larger fraction of catalysis occurred
within a chiral environment. This has led us to consider
Class II.2 catalysts in which both Lewis acidic and
basic functions are directly and covalently coordinated
Scheme 8 . At least two factors suggest this to be a
fruitful strategy: 1 all of the previously published and
successful hydrophosphonylation catalysts belong to this
class of system 31 and 2 we have found that an
achiral member of the Class II.2 system, diethylzinc
Our Class II.1 catalysts are combinations of metal
Ž
.
Ž
.
salts Lewis acid and chiral nitrogen bases based on the
6,6X-difunctionalised-2,2X-bipyridine scaffold such as 23
Ž .
Ž
. w x
Scheme 12 11 c. Our initial, rather naive, thinking
w
x
Ž .
was to allow for the possibility of coordination complex
formation between the metal ion and chiral bipyridine
whilst still allowing for the bipyridine to function as a
Ž
.
base as well as ligand Scheme 8 . In this manner, a
w
x
proportion of P—C bond formation would take place
within a chiral environment whilst essentially all H-
phosphonate activation would be facilitated by a chiral
base. Indeed, H-NMR titration experiments suggest an
interaction between 23 and Zn O3SCF3
with the coordination of both nitrogen atoms to the
4 Ž .
a A more thorough examination of this solution equilibrium is
Ž .
b
1
currently in progress.
We envisage that product inhibition may
Ž .
result from the chelation of MeO 2 P O CHPh OH to Zn2q
a
Ž
.
Ž
.
w
Ž
.2 x
consistent
w
x
hypothesis which has ample literature precedent 24 and one which
we are currently investigating in more detail.

(
)
46
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
Scheme 13. Proposed mechanism for the hydrophosphonylation of aldehydes mediated by Class II.2 catalysts.
Ž
.
Lewis acidszinc and Lewis basesethyl is an ex-
tiopurity and absolute configuration in a-hydroxyphos-
phonate esters. We have also examined the configura-
tional stability of a-hydroxyphosphonate esters under
catalytic conditions such that catalyst compatibility can
be assessed readily. We have used a ‘bottom-up’ ap-
proach to catalyst design, starting with the most simple
chiral bases and amino alcohols Class I.1 systems ,
which are not particularly effective, moving on to
slightly more intricate amphoteric receptors which ex-
tremely potent and extremely selective hydrophospho-
nylation catalyst. For example, ZnEt2 catalyses the
100% chemoselective and regioselective addition of
Ž
Ž
Ž
.
Ž
.
M e O 2 P O H to P h C H O to a ffo rd
MeO P O CHPh OH with an average turnover rate
measured over 1 h in toluene solvent at 298 K of ca.
.
Ž .
Ž
.
2
.
Ž
.
115 hy1 compared to ca. 1 hy1 for NEt3 under analo-
gous conditions. A plausible mechanistic pathway is
outlined in Scheme 13.
Ž
.
ploit hydrogen-bonding Class I.2 systems . However,
w
x
In preliminary investigations, we have found that
mixtures of ZnEt2 with chiral amino alcohols such as 8,
11a and the commercially available oxazoline 24
our results support the findings of others 31 that Class
II systems will afford far more effective and stereoselec-
tive catalysts. To this end we have designed a new class
of chiral bipyridine exemplified by 23 which contains a
chiral function ‘proximal’ to the metal binding site and
which affords much improved stereoselectivities -
40% e.e. over simple chiral amino alcohols. Our cur-
rent efforts are directed towards i developing variants
Ž
.
Ž
.
Scheme 13 each produce active Table 5 hydrophos-
Ž
phonylation catalysts 5 mol% loading in ZnEt2 and
.
Ž
chiral ligand although, enantioselectivities are poor.
.
Nevertheless, we note that the presence of ZnEt2 ap-
pears to have resulted in an increased enantioselectivity
with quinine 8 to almost twice that in the absence of
metal as well as increasing turnover rates at least ten-
fold. We presume that one principal reason for the poor
stereoselectivities is, by analogy with Class I systems,
that the source of chirality in the auxiliaries used is too
Ž .
of 23 suitable for use as hydrophosphonylation co-cata-
lysts in conjunction with electropositive metal alkyls
Ž .
Ž .
and ii examining and manipulating the source s of
regio- and stereocontrol in such systems. The results of
these and related studies in phospho-transfer chemistry
will be the subject of future contributions.
w
x
remote from the active P–C bond-forming region.
3. Conclusion
4. Experimental
Our studies here represent a logical starting point for
examining catalyst design for the enantioselective hy-
drophosphonylation of carbonyls. We have developed
and calibrated an effective method of determining enan-
4.1. General
All reactions and manipulations were performed as
described previously 50 . NMR spectra were obtained
w
x

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
47
on JEOL FX90Q, JEOL FX100, Bruker ARX 250 MHz
and AM 400 instruments operating at 100.0 MHz,
250.133 MHz or 400.132 MHz for 1H, 100.614 MHz or
62.895 MHz for 13C and 36.2 MHz or 101.614 MHz for
31 P. Deuterated solvents were dried by flash filtration
4.3. Syntheses of a-hydroxyphosphonate esters
( )
(
)
(
)
MeO 2 P O CHR OH
The synthetic procedure employed, essentially the
same in each case, is described in detail only for
Ž
.
on a column of basic alumina Brockmann Grade I . All
spectra are referenced internally using either the resid-
Ž
.
compound Rs1-naphthyl . The purity of each com-
pound was assayed via 1H, 13C and 31 P-NMR spec-
troscopy prior to screening with CDA 2.
1
ual solvent resonance for H and 13C, TMS as dH and
dC s0 ppm or the methyl hydrogen resonance of toluene
Ž
.
Ž
internal reference as 2.11 ppm in C6 D6 with refer-
.
ence to TMSs0 ppm . 85% H3PO4 was used as
external reference for 31 P at zero ppm. All spectra are
reported at 298 K in either C6 D6 or CDCl3 unless
stated otherwise; the 13C and 31 P spectra being run
(
)
(
)
(
)
4.3.1. MeO 2 P 5O CH 1-C10 H7 OH
3
Ž
. Ž
LiN SiMe3 13.78 cm of a 1.0 M soln. in THF; 1
2
.
Eq. was added dropwise to a solution of dimethyl-H-
phosphonate 1.26 cm , 13.78 mmol in toluene 20
1
3
Ž
.
Ž
under conditions of broad-band H decoupling except
3
Ž
.
.
for the determinations of enantioselectivities vide infra .
Multiplicities are as; ddsdouble doublet, dsdoublet,
sssinglet, br-ssbroad singlet, msmultiplet, ts
triplet. High resolution mass spectrometry was per-
formed by the mass spectrometric service of the School
of Chemistry on a VG Autospec instrument operating in
the electron impact mode 70 eV . Preparative chro-
matography was carried out using Merck grade 9385,
230–400 mesh 40–63 mm silica gel. Optical rotations
were recorded on an Optical Activity AA 10 polarime-
ter operating at 589.44 nm. The compounds PCl3, NEt3,
cm , maintained at y788C in a dry-ice–acetone slush
bath and the mixture allowed to warm to room tempera-
ture over the course of 1 h. Upon cooling again to
3
Ž
y788C, a sample of 1-napthaldehyde 1.87 cm , 13.78
.
mmol was added, the mixture allowed to warm to room
temperature and then stirred for 2 h. After this time, a
3
Ž
.
Ž
.
saturated aqueous solution of NH4Cl 20 cm was
added and the resulting mixture stirred for 30 min. The
Ž
.
Ž
crude product was then extracted into toluene 3=15
3
.
cm , the organic layers combined and dried over
Na2 SO4, filtered, and the volatiles removed under re-
duced pressure to afford the crude product. Subsequent
Ž
.
Ž .
ClCH2CH2Cl, MeO P O H, N-methyl morpholine,
Ž .
2
20
Ä
w x
Ž .
y38.98 neat ; Lit:
25
Ž
.
S -a-phenylethylamine
a
recrystallization from pentane–toluene 2:1 vrv af-
4
Ž
.
Ä
w x
a
Ž
q44.88 cs10,
Ž
y398 Aldrich , q -camphor
forded the pure product as colourless crystals 0.46 g,
13% . dH 8.07–7.25 m, 7H, Ph-H , 5.88 d, 1H, JPH
11.6, PC H , 4.55 br-s, 1H, COH , 3.64 d, 3H, JPH
10.5, POC H3 , 3.52 d, 3H, JPH 10.4, POC H3 . dC
148.13 s, ArC , 133.59 s, ArC , 132.64 s, ArC ,
130.74 d, JPC 6.2, ArC , 128.87 d, JPC 3.3, ArC ,
2
.
4
.
Ž
.
Ž
EtOH ; Lit: q44.18 Aldrich and all carbonyl com-
3
.
Ž
.
Ž
pounds were purchased from commercial sources and
3
Ž
.
.
Ž
.
were either recrystallised solid carbonyls , chromato-
graphed on a short column of Brockmann Grade I basic
alumina ClCH2CH2Cl , distilled under nitrogen NEt3,
Ž
.
Ž
.
Ž
.
.
Ž
.
.
w
Ž
.
Ž
Ž
.
Ž
. x
Ž
.
Ž
.
Ž
liquid carbonyls, MeO P 5O H or used as received
128.73 s, ArC , 126.20 s, ArC , 125.62 d, JPC 5.6,
2
w
Ž
.
Ž
.
Ž
Ž
.
PCl3, LiN SiMe3 as a 1 M solution in THF, metal
salts . The lipase F–AP 15 Rhizopus oryzae was a gift
ArC , 125.37 d, JPC 3.3, ArC , 123.37 s, ArC , 67.10
2
1
2
x
(
)
Ž
.
.
Ž .
d, JPC 161.6, PCH , 53.86 d, JPC 7.1, POCH3 ,
53.55 d, JPC 7.5, POCH 3 . dp 24.58 s .
MeO 2 P 5O CH C6 H5 OH 6a.— 35% . dH 7.50–
7.21 m, 5H, Ph-H , 5.04 d, 1H, JPH 11.1, PC H ,
4.97 br-s, 1H, COH , 3.68 d, 3H, JPH 10.5, POC H3 ,
3.66 d, 3H, JPH 10.3, POC H3 . dC 136.43 Ar-Cipso ,
128.32, 127.01, 128.13 Ar-C , 70.49 d, JPC 159.8,
X
2
w
Ž .
Ž
.
from Amano Enzyme Europe Ltd. N,N -bis 1- S -phen-
ylethyl -1,2-ethylenediamine was prepared using an ap-
propriate literature procedure 17 .
x
(
)
Ž
(
)
(
)
Ž
.
2
w
x
.
Ž
.
3
Ž
Ž
.
Ž
.
3
.
Ž
.
1
Ž
.
Ž
X
{
[ ( )
]
4.2. Synthesis of N,N -bis 1- S -phenylethyl -2-chloro-
2
2
.
Ž
.
Ž
PCH , 53.94 d, JPC 7.2, POCH3 , 53.59 d, JPC 7.4,
Ž .
(
)
1,3,2-diazaphospholidine CDA 2
.
(
)
(
)
(
)
POCH3 . dp 25.04 s . MeO 2 P 5O CH 2-C10 H7 OH
Ž
.
Ž
.
Ž
Ž
6c.— 7% . dH 7.95–7.18 m, 7H, Ph-H , 5.23 d, 1H,
2
w
x
.
Ž
.
This was prepared as described previously 11 a.
JPH 11.2, PC H , 4.40 br-s, 1H, COH , 3.71 d, 3H,
3
3
20
w x
Ž . Ž
y38.98 cs1; CHCl3 ; Lit: y398 cs1;
.
Ž
.
a
JPH 9.1, POC H3 , 3.67 d, 3H, JPH 10.1, POC H3 .
dC quaternary carbons not observed , 133.87 s, ArC ,
133.17 s, ArC , 128.11 s, ArC , 127.67 s, ArC ,
126.20 s, ArC , 126.07 s, ArC , 124.77 d, JPC 4.3,
ArC , 70.80 d, JPC 159.6, PCH , 53.94 d, JPC 7.0,
POCH3 , 53.68 d, JPC 7.4, POCH3 . dp 24.15 s .
MeO 2 P 5O CH 2-BrC6 H4 OH 6d.— 42% . dH
7.79–7.14 m, 4H, Ph-H , 5.58 d, 1H, JPH 11.4,
.
w
x
Ž
.
Ž
Ž
.
Ž
Ž
.
CHCl3 14 PdH 7.47–7.12 m, 10 H, Ph-H , 4.22 s
broad, 2H, C H , 3.04 m, 4H, C H2 , 1.69 dd, 6H,
JHH 6.8, JPH 2.2, C H3 . dC 142.68 s, Ar-Cipso ,
.
Ž
.
Ž
Ž
Ž
.
Ž
.
.
3
4
.
Ž
.
.
Ž
.
Ž
Ž
1
2
Ž
.
Ž
.
.
Ž
.
128.63–127.02 Ar-C , 57.87 s broad, CH , 48.72
Ž .
2
Ž
.
Ž
.
.
Ž
.
Ž .
broad, CH2 , 22.72 broad, CH3 . dp 167.2 s . Found:
(
)
(
)
(
)
Ž
.
C, 64.9; H, 7.1; N, 8.3. C18 H22 N2 PCl requires C, 65.0;
H, 6.7; N, 8.4. Found: Mq, 332.1222. Calc. for
C18 H22 N2 PCl35: M, 332.1209.
2
Ž
.
Ž
3
.
Ž
.
Ž
PC H , 5.09 br-s, 1H, COH , 3.78 d, 3H, JPH 10.6,

(
)
48
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
3
3
.
Ž
.
Ž
.
Ž
.
POC H3 , 3.68 d, 3H, JPH 10.4, POC H3 . dC 136.47
br-s, 1H, COH , 3.56 d, 3H, JPH 10.6, POC H3 ,
3
Ž
.
Ž
.
Ž
.
Ž
.
Ž
s, ArC , 132.55 s, Ar-Cipso , 123.12 s, ArC , 129.60
s, ArC , 129.55 s, ArC , 127.62 s, Ar-C , 69.42 d,
3.49 d, 3H, JPH 10.4, POC H3 . dC 142.23 d, JPC
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
25.7, ArC , 137.47 d, JPC 9.7, ArC , 136.04 d, JPC
9.3, ArC , 134.89 s, ArC , 133.92 d, JPC 19.8, ArC ,
133.21 d, JPC 18.9, ArC , 129.65 s, ArC , 128.57 d,
JPC 8.4, ArC , 128.46 d, JPC 1.6, ArC , 128.37 d, JPC
4.5, ArC , 68.01 dd, JPC 160.9, JPC 31.7, PCH ,
53.59 d, JPC 7.2, POCH3 , 53.26 d, JPC 7.5,
POCH3 . dp 24.59 s, P5O , y18.18 s, PPh2 .
1
2
.
Ž
.
.
Ž
.
.
Ž
Ž
.
JPC 161.6, PCH , 54.09 d, JPC 7.1, POCH3 , 53.66
d, 2 J P C 7.2, PO C H
.
dp 23.98
Ž
(
.
Ž .
s .
Ž
.
Ž
3
)
(
)
(
)
Ž
.
.
Ž
.
Ž
MeO 2 P 5O CH 3-BrC6 H4 OH 6e.— 63% . dH
2
1
3
Ž
.
Ž
.
.
Ž
.
7.67–7.19 m, 4H, Ph-H , 5.03 d, 1H, JPH 9.2, PC H ,
3
2
2
Ž
Ž
.
Ž
.
Ž
.
Ž
5.03 br-s, 1H, COH , 3.73 d, 3H, JPH 10.5, POC H3 ,
3
.
Ž
.
Ž
.
Ž
.
3.71 d, 3H, JPH 10.4, POC H3 . dC 138.96, s, Ar-
.
Ž
.
Ž
.
Ž
Cipso , 131.14 s, ArC , 129.93 s, ArC , 129.79 s,
1
.
Ž
.
Ž
.
Ž
ArC , 125.64 s, ArC , 122.43 Ar-C , 69.86 d, JPC
160.3, PCH , 54.14 d, JPC 7.1, POCH3 , 53.68 d,
JPC 7.4, POCH3 . dp 23.46 s . MeO 2 P 5O CH 4-
BrC6 H4 OH 6f.— 41% . dH 7.49 d, 2H, JHH 8.0,
Ph-H , 7.35 d 2H, JHH 8.6, Ph-H , 5.01 d, 1H, JPH
11.3, PC H , 4.94 br-s, 1H, COH , 3.70 d, 3H, JPH
10.5, POC H3 , 3.69 d, 3H, JPH 10.4, POC H3 . dC
135.64 d, JPC 1.9, Ar-Cipso , 131.43 d, JPC 2.5, ArC ,
128.66 d, JPC 5.8, ArC , 122.12 d, JPC 4.0, ArC ,
69.90 d, JPC 160.6, PCH , 54.06 d, JPC 7.3,
POCH3 , 53.61 d, JPC 7.5, POCH3 . dp 23.46 s .
MeO 2 P 5O CH 4-MeC6 H4 OH 6g.— 37% . dH
7.37 d, 2H, JHH 8.1, Ph-H , 7.16 d, 2H, JHH 7.8,
Ph-H , 5.00 d, 1H, JPH 10.6, PC H , 4.48 br-s, 1H,
COH , 3.70 d, 3H, JPH 9.7, POC H3 , 3.65 d, 3H,
4.4. General procedure for the assay of enantiopurity of
a-hydroxyphosphonate esters using CDA
2
.
Ž
.
(
Ž
(
2
.
Ž .
(
)
)
3
)
Ž
.
Ž
A sample of the a-hydroxyphosphonate ester 6 under
investigation 52.6 mmol was weighed out into a small
vial quantities varied from ca. 11–21 mg depending
3
2
.
Ž
.
.
Ž
Ž
Ž
.
3
.
Ž
Ž
3
.
Ž
.
.
upon R in a nitrogen filled dry box. To this solid was
added 0.6 cm3 of a dry CDCl3 deoxygenated solution
Ž
Ž
.
Ž
.
.
.
Ž
Ž
.
comprising CDA 2 at 0.125 M and triethylamine at
0.25 M at room temperature which affords a molar
1
2
Ž
.
Ž
Ž
.
2
.
Ž
.
Ž .
ratio of 6:2:NEt3 of ca. 1:1.4:2.8. The mixture was
transferred to a 5 mm NMR tube, sealed under nitrogen
and shaken briskly to ensure complete dissolution. The
reaction mixture remained clear throughout, neither the
adducts nor the triethylammonium chloride precipitated
(
)
(
)
(
)
.
Ž
.
3
3
Ž
Ž
2
.
Ž
.
Ž
3
.
Ž
.
.
Ž
3
.
Ž
Ž
Ž
JPH 9.4, POC H3 , 2.34 s, 3H, C H3-Ar . dC 137.90 s,
from solution at this level of concentration. After 15
31
Ä1
4
Ž
.
.
.
Ž
.
Ž
.
ArC , 133.43 Ar-Cipso , 129.02 s, ArC , 126.97 s,
min the P H -NMR spectrum 101.268 MHz was
1
JPC
2
Ž
. Ž
Ž
.
ArC , 70.39 d,
160.3, PCH , 53.82 d, JPC 7.1,
collected 256 scans on a Bruker ARX 250 MHz
spectrometer over a spectral width of 3048.8 Hz centred
on the phosphorus III resonances, 3.8 ms pulse; 8192
data points; 1.344 s acquisition time and 0.74 Hz digital
resolution with a 3 s pulse delay between scans to
2
.
Ž
.
Ž
POCH3 , 53.52 d, JPC 7.3, POCH3 , 21.12 s, CH3-
Ž .
.
(
)
(
)
(
)
Ž
.
Ž
Ar . dp 24.47 s . MeO 2 P 5O CH 4-MeOC6 H4 OH
3
Ž
.
Ž
.
6h.— 39% . dH 7.41 d, 2H, JHH 8.9, Ph-H , 6.90
3
2
Ž
.
Ž
.
.
d, 2H, JHH 8.8, Ph-H , 4.98 d, 1H, JPH 10.2, PC H ,
Ž
.
Ž
.
Ž
4.21 br-s, 1H, COH , 3.80 s, 3H, C H3O-Ar , 3.71 d,
3H, JPH 10.5, POC H3 , 3.66 d, 3H, JPH 10.3,
POC H3 . dC 159.56 s, ArC , 128.42 Ar-Cipso , 128.42
s, ArC , 113.85 s, ArC , 70.16 d, JPC 161.5, PCH ,
55.21 s, CH3O-Ar , 53.79 d, JPC 7.1, POCH3 ,
53.56 d, JPC 7.3, POCH 3 . dp 24.48 s .
MeO 2 P 5O CH 4-O2 NC6 H4 OH 6i.— 62% . dH
8.23 d, 2H, JHH 8.7, Ph-H , 7.68 d, 2H, JHH 8.9,
Ph-H , 5.22 d, 1H, JPH 12.4, PC H , 5.04 br-s, 1H,
COH , 3.78 d, 3H, JPH 10.5, POC H3 , 3.76 d, 3H,
JPH 10.6, POC H3 . dC 147.62 s, ArC , 143.89 Ar-
Cipso , 127.62 s, ArC , 123.44 s, Ar-C , 69.93 d, JPC
allow relaxation and inverse-gated proton decoupling to
reduce nuclear Overhauser effects. Slight line-broad-
3
3
.
Ž
.
.
Ž
.
Ž
.
Ž
.
ening 2 Hz was applied to the FIDs to aid sensitivity
without compromising baseline separation. The FIDs
were Fourier transformed and phased automatically, the
same phase corrections being applied for each sample.
Automated electronic integration gives the relative pro-
portions of diastereoisomers as listed in Table 2 from
which enantioselectivities can be computed readily. In-
tegrations performed by manual-phasing were the same
within the overall error limit of 1%. Following this, the
spectrum was re-acquired over 26 316 Hz 0.62 s acqui-
sition; 1.6 Hz digital resolution; 0.3 s pulse delay and
broad-band H decoupling to ensure complete conver-
sion of 6 to 5, the presence of excess CDA 2, and to
1
Ž
Ž
.
Ž
.
.
2
Ž
.
Ž
2
Ž
.
Ž .
(
)
(
)
(
)
Ž
.
3
3
Ž
.
Ž
2
.
Ž
.
Ž
3
.
Ž
.
Ž
3
.
Ž
.
Ž
1
.
Ž
.
Ž
.
Ž
Ž
2
.
Ž
.
(
Ž
(
158.7, PCH , 54.42 d, JPC 7.1, POCH3 , 53.71 d,
2
1
.
Ž .
(
)
)
.
JPC 7.6, POCH3 . dp 23.46 s . MeO 2 P 5O CH 2-
)
Ž
.
Ž
O2 NC6 H4 OH 6j.— 22% . dH 8.03–7.41 m, 4H,
2
.
.
Ž
.
Ž
Ž .
Ph-H , 6.30 d, 1H, JPH 14.0, PC H , 5.98 br-s, 1H,
COH , 3.74 d, 3H, JPH 10.4, POC H3 , 3.72 d, 3H,
obtain the chemical shifts of the phosphorus V nuclei.
3
Ž
.
Ž
Due to the steady deterioration of CDA 2 over a
period of months when stored at room temperature
under a static atmosphere of dinitrogen, the concentra-
tion of H-phosphonate 4, formed due to hydrolysis,
increases. Whilst this in itself is not a problem and does
not lead to any side-reaction with the a-hydroxyphos-
phonate ester, an accompanying impurity is also pro-
duced which gives rise to a low intensity yet signifi-
3
.
Ž
.
Ž
JPH 10.6, POC H3 . dC 147.41 s, ArC , 132.65 Ar-
.
.
Ž
Ž
.
Ž
.
Ž
Cipso , 133.33 s, ArC , 128.89 s, ArC , 128.37 s,
1
.
Ž
.
.
ArC , 124.57 Ar-C , 65.37 d, JPC 161.9, PCH ,
2
2
Ž
Ž
54.48 d, JPC 7.3, POCH3 , 53.58 d, JPC 7.4,
Ž .
.
( ) ( ) (
dp 22.86 s . MeO 2 P 5 O CH 2-
POC H 3
.
)
Ž
.
Ž
Ph2 PC6 H4 OH 6k.— 47% . dH 7.96–7.08 m, 14H,
2
3
.
Ž
.
Ph-H , 6.19 dd, 1H, JPH 11.7, JPH 9.3, PC H , 4.05

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
49
Ž
cantly broadened resonance centred at dp 124 D1r2 ca.
through celite, the filtrate extracted into ethyl acetate
3=25 cm , the organic layers separated, combined
3
.
Ž
.
130 Hz , within the range of the integration window for
the assay. Since this impurity occurs only in samples of
2 which are several months old, we have consistently
prepared and used all samples of 2 within that period
thus eliminating the effects of this impurity.
and dried over MgSO4. Unreacted ester and hydrolysis
product a-hydroxy-a-phenylmethyldimethylphos-
phonate 6a were separated by column chromatography
on silica gel using the solvent system CH2Cl2–ethyl-
Ž
.
acetate 5:1 vrv . Acyl-6a elutes first followed by
Ž
.
4.5. InÕersion recoÕery experiments
scalemic 6a 0.072 g, 33% . The latter was analysed
subsequently by polarimetry and NMR spectroscopy
D
w x
Ž
.
Samples for inversion recovery experiments were
prepared in CDCl3 solvent in 5 mm NMR tubes under
an atmosphere of dinitrogen in exactly the same way as
described above for the measurements of enantiopurity.
Samples were deoxygenated by purging with dinitrogen
gas for ca. 30 s prior to sealing with cling-film. Spin-
lattice relaxation times T1 were measured at 300 K on
an ARX 250 MHz instrument using a standard Bruker
revealing a sy30.48 cs0.5, acetone and e.e.s
Ž .
Ž
83 1 % in favour of the S-enantiomer average of three
.
determinations respectively. Unhydrolysed acyl-6a was
converted subsequently to 6a by stirring with a mixture
of dry methanol 5 cm and triethylamine 1 cm for
24 h, following the method of Hammerschmidt 18 to
afford a further crop of scalemic 6a 0.068 g, 31% now
3
3
Ž
.
Ž
.
w
x
Ž
.
D
w x
Ž
enriched in the R-enantiomer. a sq29.28 cs0.5,
.
Ž .
Ž
pulse sequence package modified to provide inverse-
acetone and e.e.s81 1 % average of three determina-
1
Ž
.
gated H decoupling to lessen the effects of potential
tions .
.
n.O.e. differences and a delay time of 15 s between
pulse sequences to allow for complete relaxation. A
crude experiment first locates the approximate delay
times to achieve a null signal followed by selection of
delay times to provide at least four data points on either
side of this null signal. Signal intensity was plotted
against delay time and T1 values computed by curve
fitting using the standard Bruker WINNMR package
4.7. InÕestigations of the configurational stability of
a-hydroxyphosphonate esters
Ž
.
Ž .
Ž
. Ž
A CDCl3 solution of MeO P O CHPh OH 26.6
2
.
Ž
mg, 0.123 mmol was treated with NEt3 12.5 mg,
.
Ž
.
0.123 mmol and D2O ca. 35 mg, 1.75 mmol under an
atmosphere of dinitrogen. 1H-NMR spectra were ac-
quired at regular intervals over the course of 3 weeks
revealing no incorporation of deuterium at the alpha-
carbon site.
according to the Levenberg–Marguardt algorithm, It s
yt r T1
w
x
where It and Io represent signal
Io 1y2 Aexp
intensity after delay time t and with no delay respec-
tively and A is a variable parameter.
The possibility of reversible aldehyde extrusion was
31
probed by P H and 1H-NMR examination of a
Ä1
4
Ž
.
Ž .
Ž
.
4.6. Enzyme-catalysed hydrolysis of a-acylphosphonate
esters
CDCl3 solution containing MeO P O CHPh OH
21.6 mg, 0.1 mmol dp 24.5 and 4-BrC6 H4CHO
18.5 mg, 0.1 mmol chosen since the 4-halobenzalde-
2
Ž
Ž
. Ž
.
The procedure used to hydrolyse a-acyloxy-a-phen-
hydes have a similar reactivity to benzaldehyde and
31
Ž
.
Ž
.
.
ylmethyldimethylphosphonate MeO P 5O CHPh-
afford satisfactory P-NMR shift resolution and NEt3
2
w
Ž
.
x
Ž
.
OC 5O Me acyl-6a was a slight modification of that
38 mg, 0.38 mmol 0.5 equivs. . After 8 days at ambient
w
x
reported by Hammerschmidt 18 . a-Acyloxy-a-phenyl-
methyldimethylphosphonate was synthesised by the re-
action of 6a with acetyl chloride in the presence of
temperature, NMR spectroscopy revealed a resonance at
dp 24.0 consistent with formation of the crossover
Ž
.
Ž .
Ž
.Ž
. Ž
product MeO P O CH 4-BrC6 H4 OH confirmed by
2
w
x
triethylamine 18 .
A mixture of t-butylmethyl ether and hexane 4 cm ,
addition of a small quantity, 20 mg, of authentic com-
3
Ž
.
pound to this mixture . However, since this resonance
.
1:3 vrv was added to racemic a-acyloxy-a-phenyl-
constituted -2% of the product mixture, exchange is a
very slow process under these conditions.
Ž
methyldimethylphosphonate acyl-6a 0.26 g, 1.00
.
mmol followed by the addition of aqueous phosphate
The possibility of metal-catalysed racemisation, via
3
Ž
.
Ž . Ž .
HrD exchange and ii aldehyde extrusion
buffer solution pH 6.86, 15 cm and the mixture
stirred vigorously in a constant temperature water bath
maintained at 358C whilst 0.5 M aqueous NaOH was
added to adjust the pH of the mixture to 7. Lipase F–Ap
both
i
mechanisms, was explored in CDCl3 solution. A stan-
Ž .
Ž
.
Ž .
Ž
.
dard solution
A containing MeO P O CHPh OH
2
Ž
.
Ž
.
Ž
0.25 g, 1.16 mmol and Zn OSO2CF3
42 mg, 0.12
2
3
Ž
.
.
15 was added 0.1 g and the pH again adjusted to 7 by
addition of NaOH. The mixture was then stirred for 18
h whilst maintaining pH at 7 by manual titration with
NaOH. After the required time, reaction was quenched
mmol 10 mol% was made up to 10 cm in CDCl3
3
Ž .
under a dinitrogen atmosphere. i To a 0.86 cm vol-
ume of this standard solution in a 5 mm NMR tube was
Ž
.
added D2O 35 mg, 1.75 mmol at ambient temperature
and the tube shaken. After 17 days, 1H-NMR spec-
troscopy revealed HrD exchange at the hydroxy site
Ž
.
by the addition of aqueous HCl 2 M until the pH of
the mixture reached 4. The mixture was then filtered

(
)
50
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
Ž
Ž
but no exchange at the alpha-carbon site integration of
The mixture was then quenched by addition of water 2
3
cm3 followed by aqueous NaOH solution 3.0 cm ,
.
.
Ž
the alpha-CH resonance against the OMe resonance .
An identical experiment, performed in the presence of
NEt3 13.9 ml, 1 equiv. as an additional component,
.
15% wrv and the gelatinous mixture stirred for 1 h.
Ž
.
Filtration and removal of the volatiles under reduced
Ž
.
again revealed no HrD exchange at Ca within 7 days at
pressure afforded 14a as a clear liquid. 4.37 g, 81% .
3
3
Ž .
Ž
.
Ž
ambient temperature. ii An 0.86 cm aliquot of stan-
dH 7.34–7.00 m, 10H, Ph-H , 4.75 d, 1H, JHH 3.90,
PhC HO , 3.84 s, 2H, PhC H2 , 2.98 dq, 1H, JHH
3
Ž
.
Ž
.
Ž
dard solution A was treated with 4-BrC6 H4CHO 18.5
3
.
.
Ž
.
mg, 1 equiv. at ambient temperature and analysed by
6.50, C H Me , 0.85 d, 3H, JHH 6.53, CH Me . dC
1H and 31 P-NMR spectroscopy. The sample was un-
Ž
.
Ž
.
.
Ž
141.41 s, Ph-Cipso , 140.16 s, Ph-Cipso , 128.54 s,
.
.
Ž
Ž
.
.
Ž
Ž
changed after 24 h but after 14 days a small amount
-2% of C6 H5CHO was observed in the H-NMR
p re s u m a b ly d u e to e x tru s io n fro m
MeO P O CHPh OH . Repeating this experiment with
NEt3 1 equiv. as an additional component resulted in
slightly more crossover product being observed ca.
10% after 7 days at ambient temperature.
ArC , 128.08 s, ArC , 127.15 s, ArC , 127.05 s,
1
Ž
.
Ž
.
Ž
ArC , 126.13 s, ArC , 73.22 s, PhCHO 57.78 s,
.
Ž
.
Ž
.
CHMe , 51.28 s, Ph-CH2 , 14.68 s, CH Me . Found:
Mq,. Calc. for C16 H19 NO: M,.
Ž
.
Ž .
Ž
.
Ž2
.
Ž
(
)
4.10. Reaction of N-benzyl- 1R,2S -ephedrine with ben-
zaldehyde. Synthesis of 2RS, 4SS, 5RR -3-aza-3-benzyl-
.
(
)
4-methyl-1-oxa-2,5-diphenylcyclopentane 15a
(
)
4.8. Reaction of 1R,2S -norephedrine with benzal-
dehyde. Synthesis of 12a and 13a
Ž
.
Ž
To a solution of N-benzyl- 1R,2S -ephedrine 4.23
3
.
Ž
.
g, 17.54 mmol dissolved in CH2Cl2 30 cm , benzal-
3
Ž
.
Ž
Ž
.
A CH2Cl2 solution of 1R,2S -norephedrine 2.05 g,
dehyde 1.78 cm , 17.54 mmol was added and the
resulting mixture heated at reflux for 24 h. The solution
was filtered and the volatile materials removed under
reduced pressure to afford a sticky yellow solid, which
was shown to consist of two isomeric products which
were identified by H-NMR spectroscopy as isomeric
oxazolidines 15a in a 7:3 ratio. 5.35 g, 92% . Major
product: dH 8.20–7.16 m, 15H, Ph-H , 5.07 d, 1H,
JHH 8.6, PhC HO , 5.06 s, 1H, C2-H , 3.81 d, 2H,
3
.
Ž
.
13.58 mmol and benzaldehyde 1.38 cm , 13.58 mmol
˚
was stirred over 4 A molecular sieves at room tempera-
ture for 24 h. The resulting solution was filtered and the
volatiles removed under reduced pressure to afford a
sticky yellow solid, found to consist of a mixture of
1
Ž
.
Ž
.
three products 2.95 g, 91% isolated . The major com-
ponent 75% of mixture was assigned as the benzyli-
dine imine 12a whilst the two minor products 25%
Ž
.
Ž
.
.
Ž
3
Ž
.
.
Ž
Ž
2
2
.
Ž
.
were assigned as epimeric oxazolidines 13a in a 2:1
JHH 14.1, CH2 , 3.56 d, 2H, JHH 14.1, CH2 , 3.22
ratio on the basis of 1H-NMR spectroscopy. Major
3
3
Ž
.
Ž
dq, 1H, JHH 6.5, C H Me , 0.51 d, 3H, JHH 6.6,
Ž
Ž
.
Ž
Ž
.
Ž
.
Ž
.
product: dH 8.21 s, 1H, imine proton , 8.01–7.19 m,
10H, Ph-H , 4.81 d, 1H, JHH 4.38, PhC HO , 3.65 m,
1H, JHH 6.93, C H Me , 1.13 d, 1H, JHH 6.53,
CH Me . dC 160.63 s, imine C , 141.42–126.07 Ph-
C , 76.97 s, PhCHO , 71.10 s, CHMe , 16.23 s,
CH Me . Oxazolidines: dH 8.01–7.19 m, 10H, Ph-H ,
CH Me . dC 139.14–126.16 Ph-C , 96.88 s, C2 ,
3
.
.
Ž
.
.
Ž
.
Ž
.
82.19 s, PhCHO , 61.89 s, CHMe , 54.88 s, CH2 ,
17.23 s, CH Me . Minor Product: dH 8.20–7.16 m,
15H, Ph-H , 5.49 d, 1H, JHH 8.6, PhC HO , 5.06 s,
1H, C2-H , 3.81 d, 2H, JHH3 14.1, CH2 , 3.56 d,
JHH 14.1, CH2 , 3.55 m, 1H, JHH 6.5, C H ME , 0.51
3
3
.
Ž
Ž
Ž
3
.
Ž
.
Ž
Ž
.
.
Ž
Ž
.
Ž
2
.
Ž
.
.
Ž
.
Ž
2
.
Ž
.
.
Ž
.
3
3
Ž
.
Ž
Ž
.
Ž
.
6.03 and 5.58 s, 1H, C2-H , 5.08 d, 2H, JHH q7.75,
d, 3H, JHH 6.6, CH Me . dC 139.14–126.16 Ph-C ,
3
.
Ž
.
Ž
.
.
Ž
.
.
Ž
PhC HO , 3.82 and 3.65 m, 1H, JHH 6.93, C H Me ,
94.02 s, C2-carbon , 81.92 s, PhCHO , 57.56 s,
3
3
Ž
.
Ž
Ž
.
Ž
0.76 d, 3H, JHH 6.78, CH Me , 0.74 d, 3H, JHH
6.63, CH Me . dC 141.42 s, Ph-Cipso , 136.1 s, Ph-
Cipso , 130.79–126.07 Ph-C , 91.78 s, C2 , 81.67 and
81.28 s, PhCH , 58.48 and 56.63 s, CHMe , 16.17
CHMe , 49.48 s, CH2 , 8.43 s, CH Me . Found: C
.
Ž
.
.
Ž
82.5; H 7.3; N 4.1; C23 H23NO requires C 83.9; H 7.1;
qx
.
Ž
Ž
.
w
N 4.3. Found: M-H , 328.1680. Calc. for C23 H22 NO:
xq
Ž
.
Ž
.
w
M-H , 328.1701.
q
Ž
.
Ž
.
and 15.36 s, CH Me . mrz 239 M . Found: C 80.0;
H 7.3; N 5.7; C16 H17 NO requires C 80.3; H 7.2; N 5.9.
(
)
4.11. Synthesis of N,N-dibenzyl- 1R,2S -ephedrine 9a
3
(
)
4.9. Synthesis of N-benzyl- 1R,2S -ephedrine 14a
Ž
.
A THF 20 cm solution of the product mixture
Ž
.
from Experiment 4.11 5.02 g, 15.18 mmol was added
dropwise to a stirred suspension of LiAlH4 1.44 g,
37.95 mmol in THF 15 cm . Upon addition the
mixture was treated as described in Experiment 4.10 to
afford the crude product as a clear oil. 3.86 g, 77% .
Column chromatography of a portion of the above
Ž
.
Ž
3
Ž
A solution of N-benzylidine- 1R,2S -ephedrine 5.32
g, 22.23 mmol dissolved in THF solvent 20 cm was
added dropwise to a stirred suspension of LiAlH4 2.11
g, 55.58 mmol in THF 15 cm at ambient tempera-
ture. Upon complete addition, the turbid yellow solution
was refluxed under an atmosphere of dinitrogen for 18
h, after which time the solution had become colourless.
3
.
Ž
.
.
Ž
.
Ž
3
.
Ž
.
Ž
.
Ž
.
mixture 0.96 g employing a two-stage eluent system
of light petroleum 60–808C –Et2O 100 cm ; 4:1 vrv
3
Ž
.
Ž
.

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
51
3
Ž
.
Ž
(
) (
)
followed by light petroleum 60–808C –Et2O 200 cm ;
4.13. Synthesis of N- 3,5-dimethoxybenzyl - 1R,2S -
ephedrine 14b
.
3:2 vrv afforded pure amino alcohol 9a as a clear
Ž
.
viscous liquid 0.75 g, 78% recovery . dH 7.37–7.15
3
Ž
.
Ž
.
m, 15H, Ph-H , 4.73 d, 1H, JHH 6.23, PhC HO , 3.69
d, 1H, JHH 13.9, PhC H2 , 3.47 d, 1H, JHH 13.9,
PhC H2 3.08 q, 1H, JHH 6.65, C H Me , 2.52 s,
broad, 1H, -OH , 1.15 d, 3H, JHH 6.78, CH Me . dC
143.19 s, 1H, Ph-Cipso , 139.86 s, 1H, Ph-Cipso , 128.7
s, ArC , 128.56 s, ArC , 128.20 s, ArC , 128.01 s,
ArC , 127.25 s, ArC , 126.85 s, ArC , 126.72 s,
ArC , 75.68 s, PhCHO , 58.46 s, CHMe , 54.59 s,
Ž
.
A solution of N-3,5-dimethoxybenzylidine- 1R,2S -
2
2
Ž
.
Ž
Ž
.
Ž
ephedrine 3.57 g, 12.51 mmol dissolved in THF 20
3
.
Ž
.
Ž
cm3 was added dropwise with care to a stirred suspen-
.
3
.
Ž
.
3
Ž
.
Ž
.
sion of LiAlH4 1.19 g, 31.28 mmol in THF 15 cm .
Upon complete addition the solution was refluxed under
Ž
.
.
Ž
.
Ž
Ž
.
Ž
.
Ž
nitrogen for 24 h before quenching by the careful
.
.
Ž
.
Ž
.
Ž
3
Ž
.
addition of water 2 cm followed by aqueous NaOH
Ž
.
.
Ž
.
Ž
3
Ž
.
solution 3.0 cm , 15% wrv to give a gelatinous
mixture. Further stirring for 1 h, filtration and removal
of the volatiles materials under reduced pressure af-
forded compound 14b as a clear liquid which was found
q
.
Ž
PhCH2 , 9.11 s, CH Me . Found: M , 331.1935. Calc.
for C23 H25NO: M, 331.1936. a sy14.28 cs5.4,
D
w x
Ž
.
CH2Cl2 .
Ž
to be sufficiently pure for further reaction. 3.86 g,
4
.
Ž
.
Ž
77% . dH 7.58–7.16 m, 5H, Ph-H , 6.42 d, 2H, JHH
4
(
)
.
Ž
.
.
Ž
4.12. Reaction of 1R,2S -norephedrine with 3,5-di-
methoxybenzaldehyde. Synthesis of 12b and 13b
2.3, Hortho , 6.32 t, 1H, JHH 2.4, Hpara , 4.70 d, 1H,
3
.
Ž
Ž
JHH 4.0, PhC HO , 3.75 s, 1H, C H2 , 3.73 s, 6H,
3
.
Ž
.
OMe , 2.92 dq, 1H, JHH 6.6 and 4.0, C H Me , 0.82
3
Ž
Ž
.
Ž
.
d, 3H, JHH 6.5, CH Me . dC 142.60 s, Cipso , 141.38
Ž
.
Ž
.
1R,2S -Norephedrine 2.04 g, 13.47 mmol was
.
Ž
.
Ž
.
.
Ž
s, Cipso , 128.07 s, ArC , 127.06 s, ArC , 126.12 s,
3
˚
Ž
.
dissolved in CH2Cl2 30 cm and stirred over 4 A
molecular sieves at room temperature. 3,5-Dimethoxy-
.
Ž
.
Ž
Ž
ArC 105.95 s, Cortho , 104.50 s, Cpara , 73.36 s,
PhCHO , 57.72 s, CHMe , 55.29 s, OCH3 , 51.33 s,
CH2 , 13.84 s, CH3 . Found: C 71.8; H 7.7; N 4.4;
C18 H23NO3 requires C 71.7; H 7.7; N 4.7.
.
Ž
.
Ž
.
Ž
Ž
.
benzaldehyde 2.24 g, 13.47 mmol was added to this
solution and the mixture stirred for 24 h. The resulting
solution was then filtered and the volatiles removed
under reduced pressure to afford a cloudy viscous liq-
uid, shown to consist of a mixture of three isomeric
products by H-NMR spectroscopy. The major compo-
nent of this mixture was assigned to imine 12b and the
two minor components oxazolidine epimers 13b. 3.57
g, 89% . Major product: Imine — dH 8.20 s, 1H,
.
Ž
.
(
) (
)
4.14. Reaction of N- 3,5-dimethoxybenzyl - 1R,2S -
ephedrine with 3,5-dimethoxybenzaldehyde. Synthesis of
1
X
X
(
(
)
(
)
2RS,4SS,5RR -3-aza-3- 3 ,5 -dimethoxybenzyl -2-
3 ,5 -dimethoxyphenyl -4-methyl-1-oxa-5-phenylcyclop-
Ž
X
X
)
.
Ž
entane 15b
.
Ž
.
Ž
imine proton , 7.38–7.25 m, 5H, Ph-H , 6.89 d, 2H,
JHH 2.38, Hortho , 6.54 t, 1H, JHH 2.38, Hpara , 4.84
d, 1H, JHH 3.93, PhC HO , 3.83 s, 6H, OMe , 3.83
4
4
.
Ž
.
.
3
Ž
Ž
.
.
Ž
Ž
.
Ž
3
N-3,5-Dimethoxybenzyl- 1R,2S -ephedrine 2.83 g,
9.38 mmol was dissolved in ethanol 30 cm and
stirred over 4 A molecular sieves at room temperature.
A few drops of glacial acetic acid and 3,5-dimethoxy-
benzaldehyde 1.56 g, 13.58 mmol were added to this
solution and the mixture refluxed for 48 h. The resulting
solution was filtered, removal of the volatiles under
reduced pressure afforded a viscous yellow liquid, which
was found to consist of two compounds identified as
oxazolidine epimers 15b by 1H-NMR spectroscopy.
3
3
Ž
.
Ž
.
m, 1H, JHH 3.8, C H Me , 1.10 d, 3H, JHH 6.55,
˚
.
Ž
.
Ž
.
CH Me . dC 160.33 s, imine carbon , 141.02 s, Cipso
,
Ž
.
Ž
.
Ž
Ž
138.12 s, Cipso , 128.16–126.19 Ph-C , 105.95 s,
Cortho , 103.36 s, Cpara , 76.87 s, PhCHO 70.71 s,
CHMe , 55.48 s, OMe 15.836 s, CH Me . Minor
.
Ž
.
Ž
.
Ž
.
.
Ž
.
Ž
.
Ž
products. Oxazolidines — dH 7.39–7.25 m, 10H,
4
4
.
Ž
.
Ph-H , 6.85 dd, 4H, JHH 2.38 and JHH 0.75, Hortho ,
4
Ž
.
Ž
.
6.48 t, 1H, JHH 2.25, Hpara , 5.60 s, broad, 1H,
3
.
Ž
Ž
C2-H , 5.12 d, 1H, JHH 7.7, PhC HO , 3.83 m, 1H,
3
3
.
Ž
.
Ž
.
Ž
.
JHH 3.88, C H Me , 0.78 d, 3H, JHH 6.65, CH Me .
2.72 g, 96% . Major isomer: dH 7.42–7.24 m, Ph-H ,
6.88 d, 2H, JHH 2.3, Hortho , 6.45 t, 1H, JHH 2.3,
Hmeta , 5.15 d, 1H, JHH 7.5, PhC HO , 5.08 s, 1H,
C2-H , 3.73 m, 2H, JHH 7.0, CH2 , 3.65 s, 12H,
4
4
Ž
.
Ž
.
Ž
Ž
Ž
.
Ž
dC 141.02–126.19 Ph-C , 103.89 s, Cortho , 100.71 s,
Cpara , 91.80 s, C2 , 81.29 s, PhCHO , 58.52 s,
3
.
Ž
.
Ž
.
.
Ž
.
Ž
3
.
Ž
.
Ž
.
.
.
Ž
.
Ž
CHMe , 55.48 s, OMe , 15.83 s, CH Me . dH 7.39–
7.25 m, 10H, Ph-H , 6.75 dd, 4H, JHH 2.38 and 4JHH
4
3
Ž
.
Ž
Ž
.
OMe , 3.22 dq, 1H, JHH s6.6 and 7.5, C H Me , 0.65
4
3
.
Ž
.
Ž
Ž
.
Ž
.
0.75, Hortho , 6.42 t, 1H, JHH 2.55, Hpara , 6.02 s,
broad, 1H, C2-H , 3.64 dq, 1H, JHH 6.55 and JHH
0.7, C H Me , 0.79 d, 3H, JHH 6.68, CH Me . dC
141.01–126.19 s, Ph-C , 103.86 s, Cortho , 100.27 s,
Cpara , 91.67 s, C2 , 81.58 s, PhCHO , 56.58 s,
CHMe , 55.39 s, OMe , 15.37 s, CH Me . Found:
d, 3H, JHH 6.6, CH Me . dC 160.78 s, Cipso ,
3
4
.
Ž
Ž
.
Ž
.
Ž
128.02–126.18 Ph-C , 107.13 s, Cortho , 106.35 s,
Cortho , 101.13 s, Cpara , 98.86 s, Cpara , 96.73 s, C2 ,
3
.
Ž
.
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
.
Ž
Ž
Ž
Ž
.
Ž
.
Ž
.
82.16 s, PhCHO , 62.13 s, CHMe , 55.32 s, OMe ,
.
Ž
.
Ž
.
Ž
.
Ž
.
55.22 s, OMe , 55.01 s, CH2 , 17.03 s, CH Me .
.
Ž
.
Ž
.
Ž
.
Ž
Minor isomer: dH 7.42–7.24 m, Ph-H , 6.80 d, 2H,
JHH 2.3, Hortho , 6.30 t, 1H, JHH 2.3, Hmeta , 5.52 d,
4
4
Mq, 299.1510. Calc. for C18 H21NO3: M, 299.1521.
.
Ž
.
Ž

(
)
52
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
3
.
Ž
.
Ž
1H, JHH 7.5, PhC HO , 5.46 s, 1H, C2-H , 3.73 m,
solid which appeared, on the basis of NMR spec-
troscopy to consist of a single isomer of 16 1.3 g,
2H, 3JHH 7.0, C H2 , 3.65 s, 12H, OMe , 3.60 dq, 1H,
.
Ž
.
Ž
Ž
3
3
.
Ž
.
Ž
JHH s6.6 and 7.5, C H Me , 0.69 d, 3H, JHH 6.6,
90% . Recrystallisation from pentane–toluene 20:1
.
Ž
Ž
.
Ž
.
.
.
CH Me . dC 160.50 s, Cipso , 128.02–126.18 Ph-C ,
vrv at y358C afforded the title compound as white
.
Ž
.
Ž
Ž
.
Ž
.
106.42 s, Cortho , 105.79 s, Cortho , 101.04 s, Cpara ,
98.62 s, Cpara , 93.92 s, C2 , 81.83 s, PhCHO , 57.77
crystals 0.42 g . dH 8.70 s, broad, 1H, pyrrole NH ,
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
7.38–7.25 m, 5H, Ph-H , 6.85 m, 1H, pyrrole H ,
Ž
.
Ž
.
Ž
.
Ž
Ž
.
Ž
.
.
s, CHMe , 55.38 s, OMe , 55.27 s, OMe , 49.84 s,
6.37 m, 1H, pyrrole H , 6.20 m, 1H, pyrrole H , 5.11
3
q
.
Ž
.
Ž
Ž
.
Ž
CH2 , 9.03 s, CH Me . Found: M , 448.2117. Calc.
d, 1H, JHH 8.13, PhC HO , 4.83 s, 1H, C2-H , 2.94
3
.
.
Ž
for C27 H30 NO5: M, 448.2124.
dq, 1H, JHH 8.0 and 6.43, C H Me , 2.20 s, 1H,
3
.
Ž
Ž
NMe , 0.77 d, 3H, JHH 6.43, CH Me . dC 139.95 s,
(
)
.
Ž
.
.
Ž
.
4.15. Synthesis of N,N-bis- di-3,5-dimethoxybenzyl -
1R,2S -ephedrine 9b
Ph-Cipso , 128.44 s, Ph-Cipso , 127.94 s, Ph-C , 127.69
s, Ph-C , 127.63 s, Ph-C , 118.45 s, pyrrole-C ,
109.42 s, pyrrole-C , 108.45 s, pyrrole-C , 92.34 s,
(
)
Ž
.
Ž
.
Ž
.
Ž
Ž
.
.
Ž
Ž
Ž
.
Ž
.
.
Ž
.
Ž
A solution of 4S, 5R -2- 3,5-dimethoxybenzyl -3-
3,5-dimethoxyphenyl -4-methyl-5-phenyl oxazolidine
2.72 g, 9.39 mmol dissolved in THF 20 cm was
added dropwise to a stirred suspension of LiAlH4 0.8
g, 23.48 mmol in THF 15 cm , cautiously. Upon
complete addition the solution was refluxed under dini-
trogen for 24 h followed by quenching with the careful
C2 , 82.03 s, PhCHO , 63.70 s, CHMe , 35.84 s,
q
Ž
Ž
.
.
Ž
.
NMe , 14.96 s, CH Me . Found M , 242.1417.
C15 H18 N2O requires; Mq, 242.1419. Found: C 74.5; H
7.7; N 11.6; C15 H18 N2O requires C 74.4; H 7.5; N
11.6.
3
.
Ž
.
Ž
3
.
Ž
.
X
(
)
4.17. Synthesis of N- pyrrole-2 -methyl -N-methyl-
1R,2S -ephedrine 10
3
(
)
Ž
.
addition of water 2 cm followed by aqueous NaOH
3
Ž
.
solution 3.0 cm , 15% wrv to give a gelatinous
mixture. Further stirring for 1 h, filtration and removal
of the volatiles under reduced pressure afforded crude
Ž
.
A solution of 4S, 5R -2-pyrrole-N-methyl-4-methyl-
Ž
.
5-phenyl-oxazolidine 5.3 g, 21.9 mmol was dissolved
in THF solvent 20 cm and added dropwise to a
stirred suspension of LiAlH4 1.97 g, 54.8 mmol in
THF solution 20 cm over the course of ca. 20 min.
Upon complete addition the mixture was refluxed under
an atmosphere of dinitrogen for 36 h. The reaction was
3
Ž
.
Ž
.
9b as a clear viscous liquid 2.65 g, 63% . Further
purification of the compound was achieved by percola-
tion column chromatography on silica gel. Crude prod-
Ž
.
3
Ž
.
Ž
.
.
uct 0.5 g was suspended on a short column of silica
3
Ž
Ž
.
gel 5–7 g and eluted with toluene 100 cm followed
Ž
then quenched by the cautious addition of water 1.0
3
Ž
.
by toluene–EtOAc 25:2 vrv; 300 cm . The desired
cm3 followed by aqueous NaHCO3 solution 2.7 cm ,
3
.
Ž
product was eluted with the latter solvent system as a
3
.
Ž
.
15% wrv and a further aliquot of water 1.0 cm . The
gelatinous solution was stirred for 1 h, filtered and the
volatiles removed under reduced pressure to afford
compound 10 as a white solid. Recrystallisation from
Ž
.
Ž
Ž
.
clear oil 0.35 g, 69% recovery . dH 7.32–7.17 m, 5H,
Ph-H , 6.37 d, 4H, JHH 2.35, Hortho , 6.32 t, 2H,
JHH 2.33, Hpara , 4.72 d, 1H, JHH 6.7, PhC HO , 3.72
4
.
Ž
.
4
3
.
Ž
2
Ž
.
Ž
.
.
Ž
s, 12H, OMe , 3.64 d, 2H, JHH 13.83, C H2 , 3.40 d,
Ž
.
toluene–pentane 1:1 vrv afforded 10 as white crystals
3
2H, 2JHH 13.83, C H2 , 3.14 q, 1H, JHH 6.78, C H Me ,
Ž
.
Ž
.
Ž
.
3.38 g, 68% . dH 7.78 s, broad, 1H, pyrrole NH ,
3
Ž
.
Ž
Ž
.
1.18 d, 3H, JHH 6.71, CH Me . dC 160.65 s, Ph-Cipso
Ž
.
Ž
.
7.31–7.12 m, 5H, Ph-H , 6.45 m, 1H, pyrrole H ,
5.98 m, 1H, pyrrole H , 5.87 m, 1H, pyrrole H , 4.65
d, 1H, JHH 3.88 , 3.48 s, 2H, C H2 , 2.85 dq, 1H,
JHH 6.7, C H Me , 2.11 s, 3H, NMe , 0.98 d, 3H,
Ž
.
.
Ž
143.29 s, Ph-Cipso , 142.42 s, Ph-Cipso , 128.06 s,
Ž
.
Ž
.
.
Ž
.
.
Ž
.
Ž
ArC , 127.27 s, ArC , 126.88 s, ArC , 106.6 s,
3
Ž
.
Ž
Ž
.
Ž
.
Ž
Ž
.
.
Ž
Ž
Cortho , 99.06 s, Cpara , 76.05 s, PhCHO , 58.45 s,
CHMe , 55.20 s, OMe , 54.77 s, CH2 , 9.16 s,
3
.
.
Ž
.
.
Ž
.
Ž
3
.
Ž
.
Ž
JHH 6.7, CH Me . dC 143.45 s, Ph-Cipso , 129.60 s,
CH Me . Found: C 71.6; H 7.1; N 3.0; C27 H33NO5
.
Ž
.
Ž
.
Ž
Ph-Cipso , 128.14 s, Ph-C , 127.33 s, Ph-C , 126.43 s,
requires C 71.8; H 7.4; N 3.1. Found: Mq, 451.2376.
Calc. for C27 H33NO5: M, 451.2359.
.
Ž
.
Ž
.
Ph-C , 117.02 s, pyrrole-C , 107.70 s, pyrrole-C ,
Ž
.
Ž
.
Ž
Ž
106.51 s, pyrrole-C , 75.18 s, PhCHO , 62.76 s,
.
.
Ž
.
.
Ž
.
CHMe , 51.77 s, CH2 , 37.90 s, NMe , 9.29 s,
(
)
4.16. Reaction of 1R,2S -ephedrine with pyrrole-2-
carboxaldehyde. Synthesis of 2RS, 4SS,5RR -3-aza-3-,
4-dimethyl-1-oxa-5-phenyl-2- 2 -pyrrole cyclopentane
q
Ž
CH Me . mrz 243 M-H . Found: C 73.8; H 8.2; N
(
)
11.5; C15 H20 N2O requires C 73.7; H 8.3; N 11.5.
X
(
)
16
[
]
4.18. Systems based on the 1,7,7-trimethyl 2.2.1 heptane
scaffold 17–20
Ž
.
Ž
.
1R,2S -Ephedrine 0.98 g, 5.96 mmol was dis-
3
˚
Ž
.
Ž
.
solved in CH2Cl2 20 cm and stirred over 4 A
molecular sieves, pyrrole-2-carboxaldehyde 0.57 g, 5.96
mmol was added to the solution and the mixture stirred
for 24 h at room temperature. Filtration and removal of
the volatiles under reduced pressure afforded a white
1R,4S -2,3-bornanedione 17 was synthesised using
the published procedure of selenium dioxide oxidation
Ž
.
Ž
.
x
w
x
Ž
of 1R,4S -camphor 43 j. 83%, mp 199–2008C, Lit.
w
1988C 43 j; Found: C 72.3; H 8.7; C10 H14O2 requires
.
C 72.3; H 8.4 .

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
53
3
3
Ž
.
Ž
Keto-oximine 18 was synthesised as a mixture of
stereoisomers using the published method, room tem-
8.18 d, 2H, JHH 7.7, C5H3N-Hm , 7.86 t, 1H, JHH
3
.
Ž
.
7.7, C5H3N-Hp , 3.44 dt, 4H, JHH 5.8, HNC H2 ,
3
Ž
.
Ž
.
Ž
.
perature treatment of 1R,4S -2,3-bornanedione 17 with
2.41 t, 4H, JHH 6.0, C H2 NMe2 , 2.16 s, 12H, NMe2 .
Ž .
Ž
Ž
.
Ž
.
Ž
Ž
hydroxylamine hydrochloride and sodium acetate. 89%,
dC 0.1 M CDCl3 163.54 s, -C O NH , 148.83 s,
ww x x
Ž
.
Ž
.
.
.
Ž
.
mp 151–1568C, Lit. 41 l 1558C anti , 1178C syn ;
C5H3 N-Co , 138.70 s, C5H3 N-Cp , 124.67 s,
C5H3N-Cm , 58.20 s, CH2 NMe2 , 45.32 s, NMe2 ,
37.04 s, HNCH2 . Found: M , 307.2003. Calc. for
C15 H25N5O2: M, 307.2003.
Ž
.
Ž
.
Found: C 66.5; H 8.6; N 7.7; C10 H15NO2 requires C
q
.
w x
Ž
.
66.3; H 8.3; N 7.7 40 b.
Reduction of 18 with LiAlH4 in diethylether 41 b
w
x
w
x
w
x
affords a mixture of exo,exo and endo,endo
1R,4S -3-amino-2-hydroxy-1,7,7-trimethyl 2.2.1 heptane
19 in a ca. 85:15 ratio 70% . Separation and isolation
Ž
.
Ž
.
(
)
4.20. Binding studies with bis- N,N-dimethylethyl -2,6-
pyridinedicarboxamide
Ž
.
w
x
of isomerically pure exo,exo -19 was achieved by con-
Ž
version to oxazolidinone 20 Found: C 67.9; H 8.9; N
Binding constants for both dimethyl-H-phosphonate
.
7.3; C11H17 NO2 20 requires C 67.7; H 8.7; N 7.2 by
w
Ž .
P
x
w
Ž .x
K11
and benzaldehyde K11 A with amphoteric
reaction with diethylcarbonate followed by recrystallisa-
receptor 22 have been determined in C6 D6 solution at
298 K with a receptor concentration of 0.01 M and
increasingly larger molar equivalents of substrate. The
chemical shift dependence of the carboxamide hydrogen
resonance of amphoteric receptor 22 as a function of the
Ž
.
tion from hexane–ethyl acetate 2:1 vrv according to
Ž
the published procedure 30%; Found: C 70.1; H 11.3;
N 8.1; C10 H19 NO 19 requires C 71.0; H 11.2; N 8.3
43 f.
Alkylated amino alcohols 11a–d were prepared as
.
w
x
Ž
concentration of substrate dimethyl-H-phosphonate or
benzaldehyde was determined by 1H-NMR titration
w
x
described previously 41 h. 11a: Found: C 72.2; H 12.9;
N 7.0; C12 H23NO requires C 73.1; H 11.7; N 7.1. 11b:
Found: C 73.4; H 12.7; N 6.6; C13 H25NO requires C
73.9; H 11.8; N 6.6. 11c: Found: C 74.3; H 12.3; N 6.2;
C14 H27 NO requires C 74.6; H 12.0; N 6.2. 11d: Found:
C 72.8; H 12.0; N 7.0; C12 H23NO requires C 73.1; H
11.7; N 7.1.
.
with increasing molar concentrations of substrate. The
results of this analysis for both dimethyl-H-phosphonate
and benzaldehyde are collected in Tables 6 and 7. The
data in these tables have been analysed according to the
w
x
1:1 binding isotherm: 47
1
1
1
s
q
D
D11 PK11 P L
D11
(
)
4.19. Synthesis of bis- N,N-dimethylethyl -2,6-
pyridinedicarboxamide 22
Dsd22 obs yd pure and D
22 Ž
Ž
.
.
11
Ž
.
sd complex
.
2,6-Pyridinedicarboxylic acid 15.5 g, 7.6 mmol was
Ž
3
Ž
added to neat thionyl chloride 136 g, 83.2 cm , 1.14
mol and heated at reflux for 18 h to afford a clear pale
yellow oil. Excess thionyl chloride was removed in
vacuo to afford 2,6-pyridinedicarboxylchloride as a
white crystalline solid 18.7 g, 99% . A dichloromethane
yd22 pure both D values are in ppm .
. Ž
Ž
.
.
w x
w x
Since both L 4 22 and the measured binding
o
constants are small we are able to make the approxima-
w x w x
w x
Ž
.
tion, L s L where L is the concentration of uncom-
o Ž
Ž
plexed ligand either dimethyl-H-phosphonate or ben-
solution of 2,6-pyridinedicarboxylchloride 3.21 g, 15.7
mmol in 32 cm3 was added dropwise to a stirred
.
w x
.
zaldehyde and L is the total concentration of added
o
Ž
ligand. A double reciprocal plot of the data in Tables 6
solution of N,N-dimethylethylenediamine 2.76 g, 31.3
mmol and triethylamine 32.9 cm , 236 mmol in
dichloromethane 165 cm . The resulting clear orange
solution was stirred for 1 h after which time a slight
precipitate had developed. The dichloromethane solu-
tion was removed under reduced pressure to afford a
3
w x
and 7, 1rD versus 1r L as shown in Fig. 5, affords
.
Ž
.
o
3
Ž .
binding constants for organophosphorus Fig. 6a and
Ž
.
aldehyde Fig. 6b substrates of K11 P 0.34 mol dm3
y1
Ž
.
Ž .
and K11 A 0.53 mol dm3 respectively.
y1
Ž .
(
deep red viscous liquid. This liquid was extracted into
4.21. Kinetic studies on the bis- N,N-dimethylethyl-2,6-
pyridinedicarboxamide catalysed addition of dimethyl-
H-phosphonate to benzaldehyde
3
Ž
.
toluene 3=15 cm , filtered and all the volatile mate-
rials removed under reduced pressure to afford a red
liquid. This liquid was dissolved in the minimum amount
Ž
of dichloromethane and passed down three columns of
A mixture consisting of dimethyl-H-phosphonate 367
3
Ž
.
.
Ž
.
Florisil 25 g each , each time eluting with ca. 150 cm
of dichloromethane. All the aliquots were combined and
the volatiles removed under reduced pressure to afford
22 as a yellow, waxy solid. Recrystallisation from
ml, 4.0 mmol , toluene 227 ml and 0.20 M receptor 22
solution 1.0 cm , 0.2 mmol was placed in a 10 mm
NMR tube fitted with a 5 mm insert containing refer-
ence and lock solvent C6 D6 . To this solution was
added benzaldehyde 407 ml, 4.0 mmol so that the total
3
Ž
.
Ž
.
Ž
.
Ž
.
toluene afforded 22 as colourless crystals 1.32 g, 27% .
Ž
.
Ž
Ž .
.
dH 0.1 M CDCl3, 258C 8.20 br t, 2H,-C O NH- ,
volume of the reaction system was maintained at 2.0

(
)
54
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
cm3 The reaction was then followed at 298 K by
P H -NMR spectroscopy by monitoring the intensi-
Similar manipulations for the benzaldehyde system lead
to;
31
Ä1
4
Ž
.
ties of both dimethyl-H-phosphonate dp 11.8 and the
K11 A P A
Ž
.
o
Ž
product a-hydrnxybenzyldimethylphosphonate dp
RA s
5
Ž .
.
1qK11 A P A
24.3 . Spectra were recorded with a 2 s pulse delay and
Ž .
o
inverse gated 1H decoupling in order to minimise differ-
ences in 31 P T1 relation times and differential n.O.e.
effects dues to decoupling. The results were analysed
according to standard second order kinetics through the
Under the initial concentration conditions used in the
Ž .
w
x
w
Ž
.
x
above kinetic studies, 22 s0.1 M, MeO P O H s
2
w
x
PhCHO s2.0 M, ca. 40% of the catalyst will be
Ž
.
Ž .
bound to MeO P O H and ca. 51% bound to PhCHO
2
w x
w x
rate equation; 1r P s k2 P t q 1r P , where P s
o
at the start of reaction, a total of 90% complexation.
Ž
.
Ž .
w x
MeO P O H and P represents the initial concentra-
ow
x
w x
2
That this does not achieve complete initial saturation is
tion of P at time ts0 49 . A plot of 1r P versus time
y1
due to the low binding constants 0.34 mol dm3 and
Ž
Ž .
h affords the second order rate constant k2 s5.9=
PhCHO 0.53 moly1 dm3 respectively . As reaction
.
10y2 moly1 dm3 hy1 from the gradient Fig. 7a . The
above protocol was repeated under exactly the same
Ž
.
proceeds, the concentrations of both substrates will
decrease and consequently the degree of substrate com-
plexation will decrease but product complexation will
increase. If substrate starting concentrations in excess of
3 M were possible, then )51% of the catalyst would
be bound by phosphite and )59% by aldehyde, enough
to ensure saturation for at least the start of reaction.
However, at this concentration level product solubility
in toluene is problematic compromising any meaningful
kinetic data. We have also examined the same system
with benzaldehyde as reaction solvent. Under these
pseudo first-order conditions we achieve concentrations
conditions of temperature, solvent and concentrations
3
Ž
but replacing receptor 22 with triethylamine 1.0 cm of
.
a 0.20 M triethylamine solution, 0.2 mmol . Again
monitoring the progress of reaction via P H -NMR
spectroscopy and analysis through second order kinetics
w x w xo4
1r P sk2 Ptq1r P affords a second order rate
31
Ä1
x
Ä
constant k2 s5.9=10.2 moly1 dm3 hy1 Fig. 7b .
Unfortunately, we have been unable to perform satu-
ration kinetic experiments, where receptor 22 exists in a
bound form throughout the course of reaction. This is
due to a combination of the small binding constants
observed between 22 and the substrates and solubility of
both catalyst and product a-hydroxyphosphonate ester.
If the fraction of bound catalyst 22 is represented as RP
when bound to dimethyl-H-phosphonate and RA when
bound to benzaldehyde, Csamphoteric receptor cata-
lyst 22; P s dimethyl-H-phosphonate; A s benzal-
dehyde; CP s catalyst—H-phosphonate and CA s
catalyst—aldehyde complex; K11 s1:1 binding con-
stant.
Ž
.
Ž
of only 9.7 M in PhCHO and 0.1 M in phosphite pure
.
benzaldehyde is 9.8 M which unfortunately still does
not provide saturation kinetics 83% of catalyst will be
Ž
bound by aldehyde and only 3% by phosphite under
.
these conditions .
4.22. General screening procedure for catalysis assay
All reactions were performed in oven-dried 10 mm
NMR tubes fitted with a 5 mm NMR tube insert
containing C6 D6 to provide deuterium lock. The NMR
tube reaction vessel was loaded in either a dinitrogen-
CqP CP K11 P s0.34 moly1dm3
Ž .
CqA CA K11 A s0.53 moly1dm3
Ž
.
Ž
filled dry box or on a Schlenk line with PhCHO 407
Rp s CP r C and RA s CA r C
T
T
.
Ž
.
Ž .
Ž
ml, 0.425 g, 4.0 mmol and MeO P O H 367 ml,
0.440 g, 4.0 mmol followed by a measured volume of a
2
w x
where C is the total concentration of catalyst added.
.
T
standard solution of the catalyst to be tested made up in
K11 P s CP r C P P
Ž .
1
Ž .
Ž
the required solvent to a concentration of 0.4 M 1.0
w x
w x
w x w x
cm3 and the total volume of the reaction mixture then
now since P 4 C , the approximation P s P can
o
.
made up to 2.0 cm3 microsyringe with solvent. This
loading procedure affords a concentration profile within
be made to afford
Ž
.
K11 P s CP r C P P
Ž .
2
Ž .
o
Ž
. Ž .
.
Ž .
Ž
the reactor of PhCHO 2.0 M , MeO P O H 2.0 M
and catalyst 0.2 M . The time of initiation of reaction
was taken as that time when the catalyst was added.
2
w
x
w x
w x w x
w
x
Noting that; CP sRP P C and C s C y CP we
T
T
Ž
.
Ž .
can substitute in Eq. 2 to afford;
RP
Reaction progress was then monitored in situ by
31
K11 P s C T P
Ž .
P P
3
Ž .
Ä1
4
o
P H -NMR spectroscopy on a JEOL FX 90Q spec-
C
T y C PR
T 4
Ä
p
31
Ž
.
Ž
.
Ž .
trometer 36.2 MHz for P where both MeO P O H
2
this can be rearranged to afford;
Ž
.
Ž
.
Ž . .
Ž
. Ž
dp 11.7 and product MeO P O CHPh OH dp 24.9
2
can be distinguished readily. Upon completion of reac-
tion, the solvent was removed under reduced pressure
and the crude product mixture analysed using the enan-
K11 P P P
Ž .
o
RP s
4
Ž .
1qK11 P P P
Ž .
o

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
55
.
tiopurity assay procedure described in Experiment 4.4
ca. 11 mg of crude mixture are sampled with a stan-
Lannion, France; Placement Student , Anthony Cawley
Ž
Ž
.
Nuffield Foundation Bursar; 1994 and Matthew Glos-
Ž
.
dard CDA solution containing 2 at 0.125 M and NEt3 at
sop Nuffield Foundation Bursar; 1995 . Our grateful
thanks are extended also to Mr Simon Barratt for his
skilled technical assistance in helping to set up the
requisite T1 pulse programmes, Amano Enzyme Europe
Ltd. for a gift of lipase F—AP 15, Albright and Wilson
Ltd. for a gift of dimethyl-H-phosphonate, Dr Chris
Rayner and Mr Stephen Brand for valuable discussions
concerning the lipase experiments and Professor Ron
Grigg for access to the polarimeter.
.
0.25 M . We have found that the crude product often
appears as a sticky solid, possibly caused by retention
Ž
of aromatic solvents or the presence of unreacted al-
.
ways -5% starting materials or catalyst. Conse-
quently, it is necessary to ensure complete mixing of the
crude product and we find that sampling of at least four
fractions is a reasonable measure to be confident of a
homogeneous mixture and consequently a representative
assay. We found subsequently, that removal of the
catalyst from the crude product mixture by washing
with either toluene or pentane does not compromise the
enantiopurity of the product a-hydroxyphosphonate and
may therefore be used prior to measurements of enan-
tiopurity.
Due to the steady deterioration of CDA 2 over a
period of months when stored at room temperature
under a static atmosphere of dinitrogen, the concentra-
tion of H-phosphonate 4, formed due to trace hydroly-
sis, increases. Whilst this in itself is not a problem and
does not lead to any side-reaction with the a-hydroxy-
phosphonate ester, an accompanying impurity is also
produced which gives rise to a low intensity yet signifi-
References
w x
Ž
.
1
w x
2
T.P. Kee, Chem. Brit. 32 1996 41 and references therein.
L. Stryer, Biochemistry, 3rd ed., Freeman, New York, 1988.
A. Ullrich, J. Schlessinger, Cell 61 1990 203 and references
therein.
R.L. Hilderbrand Ed. , The Role of Phosphonates in Living
Systems, RCRC Press, Boca Raton, FL, 1983; R. Engel, Chem.
Rev. 77 1977 349; G.M. Blackburn, Chem. Ind. 1981 134;
M.S. Smyth, H. Ford Jr., T.R. Burke Jr., Tetrahedron Lett. 33
1992 4137; F. Hammerschmidt, Angew. Chem. Int. Ed. Engl.
33 1994 341; M. Kitamura, M. Tokunaga, R. Noyori, J. Am.
Chem. Soc. 117 1995 2931; E. de Clercq, A. Holy, I. Rosen-
burg, T. Sakuma, J. Balzarini, P.C. Maudgal, Nature 323 1986
w x
3
Ž
.
w x
4
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
cantly broadened resonance centred at dp 124 D1r2 ca.
464.
w x
5
Ž .
V.S. Abramov, Chem. Abstr. 46 1952 6140; L. Wozniak, J.
.
130 Hz , within the range of the integration window for
Ž
.
Chojnowski, Tetrahedron 45 1989 2465 and references therein.
For leading references see, J.G. Allen, F.R. Atherton, M.J. Hall,
C.H. Hassall, S.W. Holmes, R.W. Lambert, L.J. Nisbet, P.S.
Ringrose, Nature 272 1978 56; L. Cheng, C.A. Goodwin, M.F.
Scully, V.V. Kakkar, G. Claeson, Tetrahedron Lett. 32 1991
the assay. Since this impurity occurs only in samples of
2 which are over 4–5 months old, we have consistently
prepared and used all samples of 2 within that period
w x
6
Ž
.
5
Ž
.
and have thus eliminated the effects of this impurity .
Ž
.
7333; J. Med. Chem. 35 1992 3364; C-L.J. Wang, T.L.
Taylor, A.J. Mical, S. Spitz, T.M. Reilly, Tetrahedron Lett. 33
Ž
.
1992 7667; D. Green, G. Patel, S. Engendy, J.A. Baban, G.
Acknowledgements
Ž
.
Claeson, V.V. Kakkar, J. Deadman, Tetrahedron Lett. 34 1993
6917; Tetrahedron 50 1994 5099; R. Shankar, A.I. Scott,
Tetrahedron Lett. 34 1993 231; C. Chengzhi, D. Buechler, D.
Lowe, S. Urwyler, G. Shapiro, Tetrahedron 50 1994 5735; M.
Ž
.
Ž
.
Ž
.
The EPSRC, Ciba Central Research Macclesfield
and the Nuffield Foundation are thanked for their con-
tinuing generous support of our research programme in
phosphorus chemistry; including a Postdoctoral Fellow-
Ž
.
Nomizu, A. Otaka, T.R. Burke Jr., P.P. Roller, Tetrahedron, 50
Ž
.
Ž
.
1994 2691; O.M. Green, Tetrahedron Lett. 35 1994 8081; V.
Ojea, M. Ruiz, G. Shapiro, E. Pombo-Villar, Tetrahedron Lett.
Ž
.
Ž
.
Ž
ship S.R.D. , CASE studentship M.C.M. , Earmarked
Ž
.
35 1994 3273; M. Mahajna, E. Breuer, J. Chem. Soc. Perkin
Trans. I, 1994 1847; B. Yi, T.R. Burke Jr., Tetrahedron Lett.
36 1995 4733; D.B. Berkowitz, M.K. Smith, J. Org. Chem. 60
Ž
.
.
.
Ž
.
studentship P.G.D. , Quota studentship C.P.C. and
Ž
.
Ž
grants GRrH26789 and GRrK13523 to T.P.K. . We
thank also the many undergraduate students who have
contributed to the synthetic effort of this project over
the last two years; Hilary Walker, Jeremy Major, Smita
Thobani University of Birmingham , Tai Shek, Marie-
Laure Mengant Institut Universitaire Technology de
Ž
.
1995 1233; J-M. Campagne, J. Coste, P. Jouin, Tetrahedron
Ž
.
Lett. 36 1995 2079; J. Deadman, G. Claeson, M.F. Scully, J.
Ž
.
Enzym. Inhib. 9 1995 29; J. Deadman, S. Engendy, C.A.
Goodwin, D. Green, J.A. Baban, G. Patel, E. Skordalakes, N.
Chino, G. Claeson, V.V. Kakkar, M.F. Scully, J. Med. Chem.
Ž
.
Ž
.
38 1995 1511; V.V. Kakkar, M.F. Scully, J. Med. Chem. 38
Ž
Ž
.
1995 1511
w x
7
M. Kitamura, M. Tokunaga, R. Noyori, J. Am. Chem. Soc. 117
Ž
.
1995 2931 and references therein.
5 We have not investigated the nature of this impurity but since it
is most prevalent in old )4 months samples of 2 which contain
significant amounts of H-phosphonate 4, one possibility is a conden-
sation product between 2 and 4 to form a P—O—P species, a
reaction which is well precedented. This would give rise to dp
w x
8
L.J. Jennings, M. Macchia, A. Parkin, J. Chem. Soc. Perkin
Ž
.
Ž
.
Trans. I 1992 2197; D.M. Coe, S.M. Roberts, R. Storer, J.
Chem. Soc. Perkin Trans. I 1992 2695.
See for example, B. Lejczak, P. Kafarski, J. Zygmunt, Biochem-
istry 28 1989 3459; P.G. Schultz, Angew. Chem. Int. Ed.
Engl. 28 1989 1283; C. Meier, Angew. Chem. Int. Ed. Engl.
32 1993 1704; R. Hamilton, B.J. Walker, B. Walker, Tetrahe-
dron Lett. 34 1993 2847.
Ž
.
w x
9
w
x
Ž
Ž
.
.
Ž
.
signals in the correct region )100 ppm and also account for the
broadness of the signal by coupling to four quadrupolar 14 N nuclei as
Ž
.
part of a second order A2 XXXA2 spin system.
X
Ž
.

(
)
56
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
w
x
Ž
.
10 E. Uhlman, A. Peyman, Chem. Rev. 90 1990 543; M-J.
Perez-Perez, B. Doboszewski, J. Rozenski, P. Herdewijn, Tetra-
Secondly, Dd is sensitive to field effects of phenyl substituents
in the para position, plots of Dd against Taft sl parameters or
Ž
.
w
x
hedron: Asymmetry 6 1995 973; V.A. Solodenko, M.Y. Belik,
S.V. Galushko, V.P. Kukhar, E.V. Kozlova, D.A. Mironenko,
V.K. Svedas, Tetrahedron: Asymmetry 4 1993 1965 and refer-
ences therein.
Swain-Lupton F parameters 23b reveal that those substituents
with the larger field effects give rise to smaller Dd values. Plots
Ž
.
Ž
.
of Dd against resonance parameters of the Taft sR or Swain-
Ž .
Lupton R type do not reveal a recognisable correlation pattern.
b J. March, Advanced Organic Chemistry, Wiley Interscience,
w
x Ž .
Ž .
11 a P.G. Devitt, M.C. Mitchell, J.M. Weetman, R.J. Taylor, T.P.
Ž
.
Ž .
Kee, Tetrahedron: Asymmetry 6 1995 2039; b M.C. Mitchell,
New York, 1984, pp. 281–284.
Ž
.
Ž .
w
w
w
x
A. Cawley, T.P. Kee, Tetrahedron Lett. 36 1995 287; c S.R.
Davies, C.P. Cain, M.C. Mitchell, M.T. Glossop, M-L. Men-
gant, J. Duxbury, M. Thornton-Pett, R.J. Taylor, T.P. Kee, in
preparation.
24 A.W.G. Platt, D. Simpson, J. Fawcett, D.R. Russell, Inorg.
Ž
.
Chim. Acta. 223 1994 43.
x
25 M.P. Osipova, P.M. Lukin, V.A. Kukhtin, J. Gen. Chem. USSR
Ž
.
1982 392.
w
w
x
Ž
.
x
12 D. Parker, R.J. Taylor, Tetrahedron, 43 1987 5451; D. Parker,
26 C. Hubert, B. Oussaid, G. Etemad-Moghadam, M. Koenig, B.
Ž
.
Ž
.
Chem. Rev. 91 1991 1441; J.M. Seco, Sh. Latypov, E. Quinoa,
R. Riguera, Tetrahedron: Asymmetry 6 1995 107; R. Hulst,
R.M. Kellogg, B.L. Feringa, Recl. Trav. Chim. Pays Bas 115
Garrigues, Synthesis 1994 51; C. Hubert, A. Munoz, B. Gar-
Ž
.
Ž
.
rigues, J-L. Luche, J. Org. Chem. 60 1995 1488.
w
x
27 R. Huber, A. Knierzinger, J-P. Obrect, A. Vasella, Helvetica
Ž
.
Ž
.
1995 114.
Chim. Acta. 68 1985 1730; C. Yuan, D. Chen, Synthesis,
x Ž .
Ž
.
Ž
.
13 a R.C. Anderson, M.J. Shapiro, J. Org. Chem. 49 1984 1304;
1992 531; D.Y. Kim, K. Lee, D.Y. Oh, J. Chem. Soc. Perkin
Trans. I 1992 2451; Z. Li, S. Racha, L. Dan, H. EI-Subbagh,
E. Abushanab, J. Org. Chem. 58 1993 5779; M.M. Kabat,
Tetrahedron Lett. 34 1993 8543; C. Li, C. Yuan, Tetrahedron
Lett. 34 1993 1515; W.L. McEldoon, K. Lee, D.F. Wiemer,
Tetrahedron Lett. 34 1993 5843; K.M. Yager, C.M. Taylor,
A.B. Smith III, J. Am. Chem. Soc. 116 1994 9377.
Ž .
Ž
.
b J-M.B. Brunel, O. Pardigon, M. Maffei, G. Buono, Tetrahe-
dron: Asymmetry 3 1992 1243; c C.R. Johnson, R.C. Elliot,
T.D. Penning, J. Am. Chem. Soc. 106 1984 5019;
Alexakis, S. Mutti, P. Mangeney, J. Org. Chem. 57 1992
1224; e ; A. Alexakis, J.C. Frutos, S. Mutti, P. Mangeney, J.
Org. Chem. 59 1994 3326; f A. Alexakis, S. Mutti, J.F.
Normant, P. Mangeney, Tetrahedron: Asymmetry 1 1990 437.
Ž
.
Ž .
Ž
.
Ž
.
Ž .
Ž
Ž
.
d
A.
.
Ž
.
Ž .
Ž
.
Ž
.
Ž .
Ž
.
Ž
Ž
.
w
w
x
.
28 H-J. Ha, G.S. Nam, Synth. Commun. 22 1992 1143; H.
w
x
14 B.L. Feringa, A. Smaardijk, H. Wynberg, J. Am. Chem. Soc.
Takahashi, M. Yoshioka, N. Imai, K. Onimura, S. Kobayashi,
Ž
.
Ž
.
107 1985 4798; B.L. Feringa, A. Smaardijk, H. Wynberg, B.
Strijiveen, R.M. Kellogg, Tetrahedron Lett. 27 1986 997; B.
Strijiveen, B.L. Feringa, R.M. Kellogg, Tetrahedron, 43 1987
Synthesis, 1994 763; A. Bongini, R. Camerini, S. Hofman, M.
Ž
.
Ž
.
Panunzio, Tetrahedron Lett. 35 1994 8045.
29 F. Texier-Boullet, M. Lequitte, Tetrahedron Lett. 27 1986
Ž
.
x
Ž
.
Ž
.
123; C.J. Welch, Tetrahedron: Asymmetry, 2 1991 1127.
3515; T. Kauffmann, T. Abel, D. Schwarz, Chem. Ber. 126
w
w
x
Ž
.
Ž
.
15 J.P. Vigneron, M. Dhaenens, A. Horeau, Tetrahedron 29 1973
1993 453; T. Yokamatsu, Y. Yoshida, S. Shibuya, J. Org.
Ž
.
1055.
Chem. 59 1994 7930.
x Ž .
16
Ž
.
w
w
x
Ž
.
a
H. Wynberg, A.A. Smaardijk, Tetrahedron Lett. 24 1983
30 R.S. Edmundson Ed. , Dictionary of Organophosphorus Com-
Ž .
b
5899;
A.A. Smaardijk, S. Noorda, F. van Bolhuis, H.
pounds, Chapman and Hall, London, 1988.
31 T. Yokamatsu, T. Yamagishi, S. Shibuya, Tetrahedron: Asym-
Ž
.
Ž .
x
Wynberg, Tetrahedron Lett. 26 1985 493; c Z. Glowacki, M.
Ž
.
Ž .
d Z.
Ž
.
Hoffmann, Phosphor. Sulfur Silicon 55 1991 169;
Glowacki, M. Hoffmann, J. Rachon, Phosphor. Sulfur Silicon
Ž .
metry 4 1993 1779; 1783; N.P. Rath, C.D. Spilling, Tetrahe-
Ž
.
dron Lett. 35 1994 227; I.C. Baldwin, J.M.J. Williams, R.P.
Beckett, Tetrahedron: Asymmetry 6 1995 ; V.J. Blazis, K.J.
Koeller, C.D. Spilling, J. Org. Chem. 60 1995 931; T. Arai,
M. Bougauchi, H. Sasai, M. Shibasaki, J. Org. Chem. 61 1996
Ž
.
Ž
.
82 1993 39;
e T. Yokamatsu, T. Yamagishi, S. Shibuya,
Ž
.
Ž .
Ž
.
Tetrahedron: Asymmetry 4 1993 1779; 1783;
Hammerschmidt, Tetrahedron: Asymmetry 4 1993 109;
J.K. Kozlowski, N.P. Rath, C.D. Spilling, Tetrahedron 51 1995
f Y.F. Li, F.
Ž
.
Ž .
Ž
.
g
.
Ž
2926; T. Arai, K. Aoe, K. Okamura, T. Date, M. Shibasaki,
6385.
Angew. Chem. Int. Ed. Engl. 1996, 35, 104.
32 See for example S. Shambayati, W.E. Crowe, S.L. Schreiber,
w
w
x
w
w
x
17 R. Hulst, N.K. de Vries, B.L. Feringa, Tetrahedron: Asymmetry
Ž
.
Ž
.
5 1994 699.
Angew. Chem. Int. Ed. Engl. 29 1990 256 and references
therein.
x
18 The H-phosphonate ester 4 is an unwanted byproduct in this
particular system but it has very interesting properties as a
phosphonylation agent in its own right in a stoichiometric
variant of the Pudovik reaction. P.G. Devitt, T.P. Kee, Tetrahe-
dron 51 1995 10987.
19 T. Glonek, J.R. Van Wazer, J. Phys. Chem. 80 1976 639; D.
x
Ž
.
33 C.A. Hunter, J. Chem. Soc. Chem. Commun. 1991 749; S.C.
Hirst, P. Tecilla, S.J. Geib, E. Fan, A.D. Hamiliton, Isr. J.
Ž
.
Chem. 32 1992 105; J. Smith, K. Ariga, E.V. Anslyn, J. Am.
Ž
.
Ž
.
Chem. Soc. 115 1993 362; C. Raposo, M. Almaraz, M. Crego,
w
x
Ž
.
M.L. Mussons, N. Perez, M.C. Caballero, J.R. Moran, Tetrahe-
Ž
.
Ž
.
Stanislawski, J.R. Van Wazer, Anal. Chem. 52 1980 96; A.B.
Shortt, L.J. Durham, H.S. Mosher, J. Org. Chem. 48 1983
3125.
dron Lett. 35 1994 7065; C. Raposo, M. Martin, M.L. Mus-
sons, M. Crego, J. Anaya, M.C. Caballero, J.R. Moran, J.
Ž
.
Ž
.
Chem. Soc. Perkin Trans. I 1994 2113; Y. Hamuro, S.J. Geib,
w
w
w
w
x
Ž
.
20 J.K.M. Sanders, B.K. Hunter, Modern NMR Spectroscopy: A
Guide for Chemists, OUP, Oxford, 1987, p. 167.
21 F. Hammerschmidt, Y-F. Li, Tetrahedron 50 1994 10253 and
references therein.
A.D. Hamilton, Angew. Chem. Int. Ed. Engl. 33 1994 446.
w
w
x
34 See for example, M.T.P. Viet, V. Sharma, J.D. Wuest, Inorg.
x
Ž
.
Ž
.
Chem. 30 1991 3026; F.B-Gariepy, K. Hoogsteen, V. Sharma,
Ž
.
J.D. Wuest, Inorg. Chem. 30 1991 4140; T. Bach, D.N.A.
x
Ž
.
22 D.G. Gorenstein in, Phosphorus-31 NMR: Principles and Appli-
Fox, M.T. Reetz, J. Chem. Soc. Chem. Commun. 1992 1634;
V. Sharma, M. Simard, J.D. Wuest, J. Am. Chem. Soc. 114
cations, Academic Press, New York, 1984 ch. 2.
x Ž .
Ž
.
23
a We are currently probing the origins of diastereoisomeric
1992 7931; M. Simard, J. Vaugeois, J.D. Wuest, J. Am.
31 P NMR chemical shift differences Dd of this and related
Ž
.
Ž
.
Chem. Soc. 115 1993 370.
x
Ž
.
chiral derivatising systems with secondary alcohol substrates
comprising the structural motif L-C6 H4CH OH R M.C.
Mitchell, T.P. Kee, work in progress . Two trends are worthy of
brief comment. Firstly, Dd is sensitive to the position of
substitution on the phenyl ring; decreasing in the order ortho
35 K. Ishihara, H. Yamamoto, J. Am, Chem. Soc. 116 1994
w
.
Ž
. x Ž
1561; K. Ishihara, M. Miyata, M. Hattori, T. Tada, H. Ya-
Ž
.
mamoto, J. Am. Chem. Soc. 116 1994 10529; V.K. Aggarwal,
E. Anderson, R. Giles, A. Zaparucha, Tetrahedron: Asymmetry
Ž
.
6 1995 1301.
Ž
.
Ž
.
Ž
.
Ž
.
5.40 )meta 4.76 ) para 4.28 -bromo substituents Table 2 .

(
)
S.R. DaÕies et al.rJournal of Organometallic Chemistry 550 1998 29–57
57
w
x
Ž
.
Ž .
36 See for exmple M. Mentz, T.A. Modro, J. Chem. Soc. Perkin
Chem. Soc. 1934 137; l P.C. Cherry, W.R.T. Cottrell, G.D.
Ž
.
Ž
.
Trans. 2 1995 2227.
Meakins, E.E. Richards, J. Chem. Soc. C, 1968 459; R.M.
Przeslawski, S. Newman, E.R. Thornton, M.M. Joullie, Synth.
Ž .
Commun. 25 1995 2975.
w
w
x
x
Ž
.
37 C. Agami, T. Rizk, Tetrahedron 41 1985 537.
38 M.P. Bonner, E.R. Thornton, J. Am. Chem. Soc. 113 1991
Ž
.
w
x
w
x
1299.
44 In their original report 16a Wynberg et al. did not include a
rationale for how the chiral aminoalcohol interacts with either
organophosphorus or carbonyl reagent but given the rather
w
w
x
x
Ž
.
39 Y.B. Xiang, K. Snow, M. Belley, J. Org. Chem. 58 1993 993.
40 C. Kouklovsky, A. Pouilhes, Y. Langlois, J. Am. Chem. Soc.
Ž
.
w
x
112 1990 6672.
atypical nature of the carbonyl substrate used o-NO2C6 H4CHO
w
w
x
x
Ž
.
41 R. Gamboni, C. Tamm, Helv. Chim. Acta. 69 1986 615.
42 K. Soai, S. Niwa, Chem. Rev. 92 1992 833 and references
and the ability for strong hydrogen bonds to the nitro sub-
stituent, several possible modes of interaction are possible. We
describe here only the one which we believed to be the more
likely at the time.
Ž
.
therein.
w
x Ž .
43 a A.H. Beckett, N.T. Lan, G.R. McDonough, Tetrahedron 25
Ž
.
Ž .
w
w
w
x
1969 5689; 5693; b R.A. Chittenden, G.H. Cooper, J. Chem.
45 D.F. Shriver, P.W. Atkins, C.H. Langford, Inorganic Chemistry,
2nd ed., Oxford University Press, Oxford, 1994 p. 407.
46 F.J. Carver, C.A. Hunter, R.J. Shannon, J. Chem. Soc. Chem.
Ž
.
Ž .
Soc. C, 1970 49; c K. Saigo, M. Yamamoto, K. Morimura,
H. Nohira, Chem. Lett. 1979 545; d G. Helmchen, A. Selim,
D. Dorsch, I. Taufer, Tetrahedron Lett. 24 1983 3213; e R.
Noyori, S. Suga, K. Kawai, S. Okada, M. Kitamura, N. Oguni,
M. Hayashi, T. Kaneko, Y. Matsuda, J. Organomet. Chem. 19
Ž
.
Ž .
x
Ž
.
Ž .
Ž .
Commun. 1994 1277.
x
47 K.A. Connors, Binding Constants: The Measurement of Molecu-
lar Complex Stability, Wiley-Interscience, New York, 1987.
Ž
.
Ž .
w
x
Ž
.
1990 382; f K. Tanaka, H. Ushio, Y. Kawabata, H. Suzuki,
48 A.D. Attie, R.T. Raines, J. Chem. Ed. 72 1995 119.
Ž
.
Ž .
w
x
J. Chem. Soc. Perkin Trans 1, 1991 1445;
J.F.G.A. Jansen, A.H.M. de Vries, Tetrahedron 50 1994 4479;
Ž . Ž .
g
B.L. Feringa,
49 P.W. Atkins, Physical Chemistry, 2nd ed. Oxford University
Ž
.
Press, Oxford, 1982, ch. 27.
Ž
.
w
x
h S.E. Denmark, I. Rivera, J. Org. Chem. 59 1994 6887; i
50 V. Sum, C.A. Baird, M. Thornton-Pett, T.P. Kee, J. Chem. Soc.
Ž
.
Ž .
Perkin Trans. I 1994 3183 and references therein.
H. Rupe, V.A. di Tommasi, Helv. Chim. Acta. 20 1937 1078;
Ž .
w x
51 A.J. Schultz, R.K. Brown, R.R. Schrock, J. Am. Chem. Soc.
j A. Marquet, M. Dvolaitzky, D. Arigoni, Bull. Chim. Soc. Fr.
Ž
.
Ž .
Ž .
103 1981 170.
9 1966 2956; k W.C. Evans, J.M. Ridgion, J.L. Simonsen, J.