Enantioselective palladium-catalysed allylation of 1,5-dimethylbarbituric acid

Article abstract of DOI:10.1002/(SICI)1099-0682(199801)1998:1<43::AID-EJIC43>3.0.CO;2-S

Allylation of 1,5-dimethylbarbituric acid (BS) with allyl acetate using in situ catalysts of palladium(II) acetylacetonate and chiral phosphane imine ligands results in the enantioselective formation of 5-allyl-1,5-dimethylbarbituric acid (ABS) as the main product with up to 34% ee and 3,5-diallyl-1,5-dimethylbarbituric acid (AABS) as a possible by-product, also with up to 34% ee. This reaction is a type of allylic alkylation, the stereoselectivity of which is difficult to control because the new stereocenter is formed in the nucleophile attacking from the side opposite to the metal atom. Classical optically active ligands do not give any enantioselectivity in this palladium-catalysed reaction. Chiral phosphane imine ligands, however, are a successful class of compound, synthesized by Schiff base condensation of 2-formylphenyl(diphenyl)phosphane with optically active primary amines. An optimisation of this ligand type showed that the substituents at the stereogenic center in the imine part should be a hydroxymethyl group and a bulky alkyl group, with the best ligand being the L-tert-leucinol derivative. A screening of other types of chiral ligand, e.g. phosphane amines and phosphane trisimines, has also been performed. NMR experiments and a molecular modelling study of the cation [(η³-allyl)Pd(2a)₂]⁺ were carried out (tripos force field). The enantioselectivity of the phosphane imine ligands is explained by an interaction of the chiral side arm of one of the ligands, which extends to about 3 A above the allyl plane, with the incoming nucleophile.

Full text of DOI:10.1002/(SICI)1099-0682(199801)1998:1<43::AID-EJIC43>3.0.CO;2-S

FULL PAPER
Asymmetric Catalysis, 115[᭛]
Enantioselective Palladium-Catalysed Allylation of 1,5-Dimethylbarbituric Acid
Henri Brunner*a, Irmgard Demla, Wolfgang Dirnbergera, Bernd Nuberb, and Walter Reißera
Institut für Anorganische Chemie der Universitäta,
D-93040 Regensburg, Germany
Fax: (internat.) ϩ49(0)941/9434439
E-mail: henri.brunner@chemie.uni-regensburg.de
Institut für Anorganische Chemie der Universitätb,
D-69120 Heidelberg, Germany
Received August 4, 1997
Keywords: Asymmetric catalysis / Allylic alkylation / 1,5-Dimethylbarbituric acid / Palladium catalysts /
Chiral phosphane imine ligands
Allylation of 1,5-dimethylbarbituric acid (BS) with allyl ace- yl)phosphane with optically active primary amines. An opti-
tate using in situ catalysts of palladium(II) acetylacetonate
misation of this ligand type showed that the substituents at
and chiral phosphane imine ligands results in the enantiose- the stereogenic center in the imine part should be a hydroxy-
lective formation of 5-allyl-1,5-dimethylbarbituric acid (ABS)
as the main product with up to 34% ee and 3,5-diallyl-1,5-
dimethylbarbituric acid (AABS) as a possible by-product,
methyl group and a bulky alkyl group, with the best ligand
being the -tert-leucinol derivative. A screening of other ty-
pes of chiral ligand, e.g. phosphane amines and phosphane
L
also with up to 34% ee. This reaction is a type of allylic alky- trisimines, has also been performed. NMR experiments and
lation, the stereoselectivity of which is difficult to control be-
a molecular modelling study of the cation [(η3-allyl)Pd(2a)2]+
cause the new stereocenter is formed in the nucleophile at- were carried out (tripos force field). The enantioselectivity of
tacking from the side opposite to the metal atom. Classical
optically active ligands do not give any enantioselectivity in
this palladium-catalysed reaction. Chiral phosphane imine li-
the phosphane imine ligands is explained by an interaction
of the chiral side arm of one of the ligands, which extends to
about 3 A above the allyl plane, with the incoming nucleo-
˚
gands, however, are a successful class of compound, synthe- phile.
sized by Schiff base condensation of 2-formylphenyl(diphen-
tylcyclohexanone[3], enantioselectivities of 62 and 81% ee,
Introduction
respectively, were obtained in the alkylation of allyl acetate.
The enantioselective palladium-catalysed allylic alky-
lation is one of the most studied subjects in stereoselective
synthesis[2][3][4][5]. Usually the new stereogenic center is
formed at one of the carbon atoms of the allylic system.
The enantioselectivity of this type of allylic alkylation can
be controlled by optically active phosphane ligands bound
to the Pd catalyst, which forms an intermediate η3-allyl
complex in the catalytic cycle. High optical inductions
have been obtained for systems, in which the new
asymmetric center is generated in the allylic
moiety[2][3][4][5][6][7][8][9][10][11][12][13][14][15]. However, there is
an unsolved problem in the palladium-catalysed allylic alky-
lation involving the creation of an asymmetric center in the
nucleophile, usually a carbanion, which attacks the η3-allyl
system in the CϪC bond forming step. Control of the en-
antioselectivity of this type of allylic alkylation is difficult
because the nucleophile approaches the (η3-allyl)Pd com-
plex from the side opposite to the metal atom (and its op-
tically active ligand). By using special nucleophiles, such as
the benzophenoneimine of methyl glycinate[18][19] or 2-ace-
The enantiomers of chiral barbituric acids show different
effects inside the organism[20][21][22][23][24]. These optically
active barbituric acids are synthesized by the resolution of
racemates[20][25][26] or by regioselective condensation of N-
methylurea with optically active cyanoacetates[21][22][24]. Al-
though interest in barbituric acids acting as sedative/hyp-
notic and antiepileptic agents is decreasing, this is not true
for barbituric acids which function as short time anesthet-
ics. An important example is Methohexital (Brevimy-
tal )[20][27][28][29][30][31][32][33][34], an allylated barbituric acid,
the synthesis of which in principle could be attempted by
stereoselective palladium-catalysed allylic alkylation. An
achiral allylation of the unsubstituted barbituric acid has
been reported[35]
.
As a model system for the synthesis of Methohexital by
allylic alkylation, we chose the reaction of 1,5-dimethylbar-
bituric acid with allyl acetate as the allylic component,
which we introduced in 1994 together with the pertinent
enantiomer analysis[36]. In these previous studies we showed
that classical ligands, such as (Ϫ)-Diop, (Ϫ)-Prophos and
(Ϫ)-Norphos, only gave the allylated product 5-allyl-1,5-di-
Part 114: Ref.[1]
.
[᭛]
Eur. J. Inorg. Chem. 1998, 43Ϫ54
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1434Ϫ1948/98/0101Ϫ0043 $ 17.50ϩ.50/0
43

H. Brunner, I. Deml, W. Dirnberger, B. Nuber, W. Reißer
FULL PAPER
methylbarbituric acid in racemic form[36], whereas with the Scheme 2. The phosphane imines 2-13, the phosphane amines 14a,
14b, the phosphane trisimines 15a, 15b, the osphane oc-
phosphane imine ligands 2a, 2b, and 11, an enantiomeric
taimine 16, the deltacyclanephosphane 17 and the
excess of 10 to 13% was obtained[36]
.
osphane oxazolines 18-20
In this paper we describe our investigations into the en-
antioselective palladium-catalysed allylation of 1,5-dimeth-
ylbarbituric acid with allyl acetate in which we varied the
base, solvent, temperature, concentration, Pd/ligand ratio
and ligand type[37]. In all, 134 different chiral ligands have
been tested[37][38]. We present an explanation for the enanti-
oselectivity of the phosphane imine ligands in the (η3-al-
lyl)Pd complex[37]. The preparation of new optically active
phosphane imine, phosphane amine, and phosphane tris-
imine ligands is reported[37]
.
Synthesis of New Optically Active PN Ligands
In the previous studies the phosphane imine ligands 2a,
2b, and 11 had given the best results in the Pd-catalysed
allylation of 1,5-dimethylbarbituric acid with allyl acet-
ate[36]. The synthesis of the new PN ligands described here
was carried out bearing in mind information gained from
the enantioselectivities obtained in the catalytic model reac-
tion between 1,5-dimethylbarbituric acid/allyl acetate.
Schiff base condensation (Scheme 1) of 2-formylphenyl(di-
phenyl)phosphane 1[39][40][41] with the optically active pri-
mary amines (R)-(Ϫ)-2-amino-1-butanol, (S)-(ϩ)-2-amino-
1-butanol, -alaninol, -valinol, -leucinol, (2S,3S)-(ϩ)-2-
amino-3-methyl-1-pentanol
(-isoleucinol),
(S)-(ϩ)-2-
amino-3,3-dimethyl-1-butanol (-tert-leucinol), -α-phenyl-
glycinol, -phenylalaninol, -norephedrine, (1S,2S)-(ϩ)-2-
amino-1-phenyl-1,3-propanediol, -tert-butyl tert-leucinate
and (ϩ)-dehydroabietylamine in refluxing methanol and di-
chloromethane, respectively, gave the corresponding phos-
phane imines 2Ϫ13 (Scheme 2). Recrystallisation of the dry
residues from petroleum ether (boiling range 40Ϫ60°C)
provided colourless or white-yellow crystals and powders,
except in the case of 6, which is a colourless oil.
Scheme 1. Synthesis of phosphane imines and phosphane amines
yl)phosphane[40][41][42] with an excess of (R)-(Ϫ)-2-amino-
Reduction of the phosphane imines 2a and 2b with so- 1-butanol and (S)-(ϩ)-2-amino-1-butanol, respectively, in
dium borohydride at Ϫ10°C in methanol (Scheme 1) gave refluxing dichloromethane. 15a and 15b were isolated as
the phosphane amines 14a and 14b (Scheme 2).
yellow crystalline solids after recrystallisation from petro-
The phosphane trisimines 15a and 15b (Scheme 2) were leum ether (boiling range 40Ϫ60°C). The crystals were suit-
prepared by Schiff base condensation of tris(2-formylphen- able for X-ray analysis. Figure 1 shows the molecular struc-
44
Eur. J. Inorg. Chem. 1998, 43Ϫ54

Asymmetric Catalysis, 115
FULL PAPER
Standard Reaction: The Allylation of 1,5-Dimethyl-
barbituric Acid
In the catalytic model reaction, 1,5-dimethylbarbituric
acid (BS) is allylated with allyl acetate to give 5-allyl-1,5-
dimethylbarbituric acid (ABS) as the main product and 3,5-
diallyl-1,5-dimethylbarbituric acid (AABS) as a possible by-
product (Scheme 3).
Scheme 3. Enantioselective palladium-catalysed allylation of 1,5-
dimethylbarbituric acid (BS) with allyl acetate
In dichloromethane BS is deprotonated with a small ex-
cess of a base, such as DBU or NEt3. To the clear solu-
tionwas added 1 mol% of palladium(II) acetylacetonate,
[Pd(acac)2], 4 mol% of a monodentate optically active phos-
phane imine ligand and a mixture of dichloromethane and
toluene. The addition of allyl acetate starts the reaction.
Typically, the clear solution is stirred at 38°C for 24, 48 or
72 hours. A qualitative TLC test (silica 60, Merck; dichloro-
methane/acetonitrile 25:1) shows how much ABS, AABS
and allyl acetate are present at a specific point of conversion
(yellow spots after dipping into a dilute KMnO4 solution).
The reaction is stopped with 0.2  hydrochloric acid. In
ordert to remove the excess base, base/acetic acid adduct,
and the unreacted BS, the organic layer is treated with 0.2
 hydrochloric acid and washed three times with water.
Evaporation to dryness removes the excess allyl acetate.
Due to the slight water-solubility of ABS, about 3% of the
yield of ABS is lost in the work-up procedure described.
This loss, however, cannot be avoided because repeated ex-
traction of the organic layer with hydrochloric acid and
water is necessary to remove BS quantitatively, as this com-
pound would be detected as a broad peak below the (Ϫ)-
ABS peak in the GC analysis with a Chirasil-Val-L column.
The enantiomeric excess of monoallylated barbituric acid
ABS and diallylated barbituric acid AABS was determined
by GC on a Chirasil-Val-L column and on a Lipodex E
column, respectively. The first peak detected on the Chira-
sil-Val-L column was assigned to (Ϫ)-ABS (configuration
unknown) by correlation with an enantiomerically enriched
sample (see below). For AABS, the symbols (1.) and (2.),
ture of one of the two independent phosphane trisimine
molecules 15b in the unit cell. For details and the CSD
number see Experimental Section.
Figure 1. Molecular structure of one of the two independent phos-
phane trisimine molecules 15b as ORTEP plot (thermal ellipsoids
are depicted with 50% probability)
Eur. J. Inorg. Chem. 1998, 43Ϫ54
45

H. Brunner, I. Deml, W. Dirnberger, B. Nuber, W. Reißer
FULL PAPER
respectively, indicate whether the first or the second de- phenylphosphane in dichloromethane at 25°C (base DBU)
tected peak of the two AABS enantiomers on the Lipodex gave 64% ABS and traces of AABS (Table 1, entry 1). Thus,
E column was in excess. Chemical yields of ABS and AABS the palladium-catalysed allylation of BS is an ideal example
were determined on the same Chirasil-Val-L column using of a ligand accelerated catalysis[44]. Under the same con-
the internal standard benzil. In some cases, e.g. if the chemi- ditions, but at Ϫ17°C, a yield of only 4% ABS was obtained
cal yield was below 3%, the by-product 5-chloromethyl-1,5- after 72 h[37]. The use of [(η3-allyl)PdCl]2 in place of [Pd-
dimethylbarbituric acid was detected. This compound re- (acac)2] afforded 60% ABS along with 11% AABS, the dou-
sults from the reaction of the carbanion of BS with the bly allylated product (entry 2). Nickel dichlorophosphane
solvent dichloromethane[37]
.
complexes (1 mol% of catalyst) gave yields below 2% for
Samples of optically active ABS dissolved in dichloro- ABS. With 10 mol%, yields of up to 29% for ABS were
methane did not change their enantiomeric excess after half obtained[38]. Therefore, only Pd-catalysed reactions with 1
a year. In a racemisation test under standard catalytic con- mol% of [Pd(acac)2] were investigated further.
ditions at 25°C, (ϩ)-ABS (12.4% ee) was dissolved in di-
Variation of the base, i.e. replacement of DBU by tri-
chloromethane and then [Pd(acac)2] and PPh3 were added. ethylamine, at 38°C in dichloromethane yielded 98% ABS,
After 72 h a sample was examined by GC [12.2% ee (ϩ)]. and these conditions are close to those chosen as the stand-
The base DBU was then added. After 48 h, work-up using ard conditions described later (entry 3). The use of BSA
the usual procedure gave 11.8% ee (ϩ) (GC reproducibility [N,O-bis(trimethylsilyl)acetamide] as the base in equimolar
about ±0.5% ee). Thus, ABS is configurationally stable dur- quantities with respect to BS, in dichloromethane at 25°C,
ing catalysis.
gave 63% of ABS (entry 4), and this is comparable with
An enantiomerically enriched sample of (Ϫ)-ABS was the results obtained using DBU (entry 1). When a 1.5 
obtained by recrystallisation from water/ethanol (175:1). LDA·THF solution in cyclohexane was added to BS dis-
After this recrystallisation the ee in crystals had diminished, solved at Ϫ72°C in THF (colourless precipitate), low yields
whereas it increased in the mother liquor. A crystal of of ABS and AABS were obtained[37]. Stirring a mixture of
17.4% ee (Ϫ)-ABS had the optical rotations [α]2D5 ϭ Ϫ0.36 BS, NBu4OH, the catalyst, and allyl acetate in a two-phase
25
and [α]
ϭ Ϫ2.4 (c ϭ 5, CH2Cl2). An extrapolation to system for 72 h at 25°C gave 59% ABS and 6% AABS (en-
436
optical purity gives [α]2D5 ϭ ±2.1 and [α] ϭ ±13.8 for the try 5). Presynthesis of the tetrabutylammonium salt of BS
enantiomers of ABS.
25
436
(NBu4BS) makes the addition of a base unnecessary[37]
.
Thus, NBu4BS in a dichloromethane/toluene mixture (5:1)
at 25°C (homogeneous solution) afforded a yield of 44%
ABS and 1% AABS after 24 h (entry 6).
The Diallylated 1,5-Dimethylbarbituric Acid AABS
The formation of AABS depends on the reaction con-
ditions. When the catalysis is carried out with a 3-fold ex-
cess of the base DBU and a 21-fold excess of allyl acetate,
AABS is produced in 98% yield in refluxing THF after 24
h. After 6 h a yield of 34% ABS and 32% AABS is obtained
in this experiment. The formation of AABS depends on the
ligand used in the catalyst, although a correlation between
the ligand structure and the amount of AABS cannot be
found. Under standard conditions (25°C or 38°C) the 134
ligands tested give the by-product AABS in yields ranging
from 0 to 19%[37][38]. Ligands which form 7- to 9-membered
The optically active bases (Ϫ)-quinine, (ϩ)-quinidine,
(Ϫ)-cinchonidine and (ϩ)-cinchonine, in combination with
the achiral ligand triphenylphosphane, induced an enantio-
meric excess of up to 6% ee in ABS, with the yields varying
from 71 to 90% (entry 7)[37]. 2-Formylphenyldiphenylphos-
phane (1) is the parent compound of the imines 2Ϫ13. As
this compound could be formed on hydrolysis of these imi-
nes, it was tested as a ligand in the catalytic allylation. In-
deed, 1 in combination with (Ϫ)-quinine as a base induced
6% ee (ϩ) for ABS (81% yield) and 4% AABS (entry 8).
chelate rings containing
a dioxolane structural ele-
Catalyses with the Phosphane-Imine Ligands 2a and 2b
ment[37][38] (e.g. Diop) give, in all cases, AABS, due to the
high catalytic activity of their complexes. For example, with
the ligand (1S,2S)-1,2-bis[2Ј-(diphenylphosphanyl)phenyl]-
1,2-dimethoxyethane[43], the reaction was almost complete
after 1 h, yielding 78% ABS and 8% AABS. However, in
the case of the the phosphane imine ligand 2a, formation
of ABS and AABS had not even begun after 1 h[37]. Inter-
estingly, a mixture of the starting material BS and the dial-
lylated product AABS, in the absence of allyl acetate, under
catalytic conditions gave the monoallylated product ABS,
i.e. the formation of the diallylated product AABS is revers-
With a 1:2 ratio of [Pd(acac)2]/2a in 10 ml of dichloro-
methane at 25°C, a yield of 14% and an enantiomeric excess
of 8% of (ϩ)-ABS was obtained (Table 2, entry 1). Increas-
ing the Pd/ligand ratio to 1:4 and the amount of dichloro-
methane to 15 ml increased the enantiomeric excess to 10%
ee and 13% ee for (ϩ)-ABS, respectively (entries 2, 3). At a
reaction temperature of 38°C the yield improved signifi-
cantly to 63% ABS and 9% AABS without loss of enanti-
oselectivity (entry 4). With 15 ml of a mixture of dichloro-
methane/toluene (2:1) as the solvent, the enantiomeric ex-
cess rose to 20% ee for (ϩ)-ABS and 20% ee (2.) for AABS
(entry 5), with the yield being 66% ABS and 12% AABS
(typical standard conditions). The use of triethylamine and
(Ϫ)-quinine as bases did not improve the enantiomeric ex-
ible[37]
.
Catalyses with the Ligand PPh3
When the catalysis was attempted with [Pd(acac)2] with- cess[37]. Little or no enantioselectivity with ligand 2a was
out a ligand, there was no turnover. Addition of tri- observed in solvents such as acetonitrile, methanol, tetra-
46
Eur. J. Inorg. Chem. 1998, 43Ϫ54

Asymmetric Catalysis, 115
Table 1. Palladium-catalysed allylation of 1,5-dimethylbarbituric acid (BS) with allyl acetate using PPh3 and 1 as achiral ligands
FULL PAPER
Entry[a] Ligand[b]
Base
[ml] Solvent
Temp.
[°C]
Time
[h]
Yield[c]
ee [%][d]
Yield[c]
ABS [%] ABS config. AABS [%]
1
PPh3
PPh3
PPh3
PPh3
PPh3
PPh3
PPh3
1
DBU
DBU
10 CH2Cl2
10 CH2Cl2
25
25
38
25
25
25
38
38
24
24
24
24
72
24
24
24
64
60
98
63
59
44
73
81
0
0
11
0
2[e]
3
0
NEt3
15 CH2Cl2
0
0
4
BSA[f]
6 CH2Cl2
0
5
NBu4OH solution in MeOH
none; substrate NBu4BS
(Ϫ)-quinine
5 H2O ϩ 20 toluene
25 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
0
6
6
0
1
7
6 (ϩ)
6 (ϩ)
1
8
(Ϫ)-quinine
4
[a]
[b]
[c]
1 Mol% [Pd(acac)2] was used as procatalyst. Ϫ
Pd:ligand ratio ϭ 1:4. Ϫ
Yields were determined by GC on a Chirasil-Val-L
[e]
[d]
column, yields < 0.5% are given as 0%. Ϫ
allyl)PdCl]2 was used as procatalyst. Ϫ BS:BSA ratio ϭ 1:1 ϩ ഠ 1 mg KOAc.
The ee of ABS was measured by GC on a Chirasil-Val-L column. Ϫ 0.5 Mol% [(η3-
[f]
Optimisation of the Lead Structure of 2a
hydrofuran, chloroform and hexamethylphosphoric tri-
amide (base DBU)[37]. Thus, the best solvent for the al-
lylation of BS is a dichloromethane/toluene mixture. As BS
is only sparingly soluble in toluene, the toluene content is
limited in order to keep the reaction mixture homogeneous.
The ligand 2b, the enantiomer of 2a, verified with 19%
ee for (Ϫ)-ABS and 20% ee (1.) for AABS the reproduc-
ibility of the catalysis and the product analysis under the
standard reaction conditions (entry 6). Double stereoselec-
tion came into play for the optically active base (Ϫ)-quinine
in combination with the enantiomeric phosphane imines 2a
and 2b. Thus, 2a provided 16% ee (ϩ)-ABS (78% yield) and
2b 12% ee (Ϫ)-ABS (entries 7, 8). With (ϩ)-quinidine in
combination with 2a, 3% ee was obtained for (ϩ)-ABS un-
der the standard conditions[37]. The bases (Ϫ)-cinchonidine
As already mentioned, the phosphane imine ligand 2a
gave an enantioselectivity of 20% ee (ϩ)-ABS and 20% ee
(2.) AABS using the base DBU in the solvent mixture di-
chloromethane/toluene (Table 2, entry 5; Table 3, entry 1),
and this is superior to all the classical ligands. Therefore,
the structure of 2a, which carries a hydroxymethyl group
and an ethyl group at the stereogenic center in its imine
substituent in the ortho position of one of its phenyl rings,
seems to be a lead structure for catalytic allylation in which
the new stereocenter is formed at the carbanionic center of
the deprotonated nucleophile BS. Therefore, we set out to
optimize this lead structure.
Ligand 3 is similar to ligand 2b, having a methyl group
and (ϩ)-cinchonine, together with ligand 2a, induced 9% ee instead of an ethyl group and the same hydroxymethyl
and 10% ee (ϩ)-ABS and yields of 87 and 82%, respectively, group in the ligand side arm, and this gives rise to an en-
in a reaction which started as a suspension in a 6:1 mixture antiomeric excess of 9% for (Ϫ)-ABS with DBU (entry 2)
of dichloromethane and toluene and which became clear and of 8% for (Ϫ)-ABS with triethylamine . Thus, a de-
after 41 h under standard conditions. The bases BSA and crease in the size of the alkyl group in the ligand side arm
LDA·THF with ligand 2a afforded low ABS yields[37]. With reduces the enantioselectivity. Therefore, we increased the
NBu4BS and the ligands 2a and 2b, 12% ee (ϩ)- and (Ϫ)- size of the alkyl group in the ligands 4Ϫ7, a strategy which
ABS was obtained, but with a yield of only 20 and 17%, turned out to be successful. The phosphane imine 4, carry-
respectively, in a dichloromethane/toluene mixture of 5:1 ing an isopropyl group instead of the ethyl group in 2b, gave
under the standard reaction conditions[37]
.
21% ee for (Ϫ)-ABS and 21% ee (1.) for AABS with DBU
Table 2. Palladium-catalysed allylation of 1,5-dimethylbarbituric acid (BS) with allyl acetate using the phosphane imine ligands 2a and 2b
Entry[a] Ligand Pd/ligand
ratio[a]
Base
[ml] Solvent
Temp. Time
Yield[b]
ee [%][c]
Yield[b]
ee [%][c]
[°C]
[h]
ABS [%] ABS config. AABS [%] AABS config.
1
2
3
4
5
6
7
8
2a
2a
2a
2a
2a
2b
2a
2b
1:2
1:4
1:4
1:4
1:4
1:4
1:4
1:4
DBU
DBU
DBU
DBU
DBU
DBU
10 CH2Cl2
10 CH2Cl2
15 CH2Cl2
15 CH2Cl2
25
25
25
38
72
72
72
72
72
72
24
48
14
26
22
63
66
66
78
80
8 (ϩ)
10 (ϩ)
13 (ϩ)
13 (ϩ)
20 (ϩ)
19 (Ϫ)
16 (ϩ)
12 (Ϫ)
0
0
Ϫ
Ϫ
Ϫ
12 (2.)
20 (2.)
20 (1.)
Ϫ
0
9
10 CH2Cl2 ϩ 5 toluene 38
10 CH2Cl2 ϩ 5 toluene 38
12
7
(Ϫ)-quinine 10 CH2Cl2 ϩ 5 toluene 38
(Ϫ)-quinine 10 CH2Cl2 ϩ 5 toluene 38
1
3
Ϫ
[a]
[b]
1 Mol% [Pd(acac)2] was used as procatalyst. Ϫ Yields were determined by GC on a Chirasil-Val-L column, yields < 0.5% are given
as 0%. Ϫ [c] The ee of ABS was measured by GC on a Chirasil-Val-L column, ee of AABS was measured by GC on a Lipodex E column.
Eur. J. Inorg. Chem. 1998, 43Ϫ54 47

H. Brunner, I. Deml, W. Dirnberger, B. Nuber, W. Reißer
FULL PAPER
(entry 3), which is slightly better than the enantioselectivity nantly (Ϫ)-ABS, whereas the ligands with the (R) configu-
of 2b (Table 2, entry 6). Triethylamine and (Ϫ)-quinine, ration (2a, 8) give (ϩ)-ABS (bases DBU and triethylamine).
respectively, in combination with 4 afforded 18% ee for (Ϫ)-
ABS with 89 and 75% yield, respectively[37]. Further en-
largement of the alkyl group, i.e. to an isobutyl group in
Screening of the Ligands 12؊20
ligand 5, provided 28% ee for (Ϫ)-ABS and 28% ee (1.) for
Under standard conditions, nitrogen, nitrogen/oxygen
AABS with DBU (entry 4). Triethylamine and (Ϫ)-quinine, and nitrogen/sulfur ligands [e.g. 2,2Ј-bipyridine, (Ϫ)-spar-
respectively, together with 5 gave 24% ee and 23% ee for teine, pyridine-oxazolines, etc.] only give yields below 1%
(Ϫ)-ABS[37]. The ligand 6, bearing a sec-butyl group, gave for ABS in the Pd-catalysed allylation of 1,5-dimethylbarbi-
a similar result for (Ϫ)-ABS with 26% ee with DBU (entry turic acid[38]. Therefore, the ligand screening concentrated
5) and 20% ee with triethylamine (yield 97%)[37]. The best on phosphane ligands[37][38]. Here, we only report represen-
enantiomeric excess in the palladium-catalysed system BS/ tative results.
allyl acetate was obtained with the phosphane imine 7,
The phosphane imines 12 and 13, the phosphane amines
which contains a tert-butyl substituent. 7 induced 34% ee 14a and 14b, the phosphane trisimines 15a and 15b, the
for (Ϫ)-ABS with 77% yield and 1% yield for AABS [(34% phosphane octaimine 16, the deltacyclanephosphane 17 and
(1.)] with DBU (entry 6). Ligand 7 with triethylamine as a the phosphane oxazolines 18Ϫ20 (Scheme 2) were tested in
base afforded 29% ee and 89% yield of (Ϫ)-ABS and 1% the standard palladium-catalysed allyllation of 1,5-dimeth-
yield of AABS with 30% ee (1.) (entry 7).
ylbarbituric acid (BS) with allyl acetate.
Table 3. Palladium-catalysed allylation of 1,5-dimethylbarbituric acid (BS) with allyl acetate using the phosphane imine ligands 2aϪ11
Entry[a] Ligand[a]
Base
Solvent
Time
[h]
Yield[b]
ABS [%]
ee [%][c]
Yield[b]
ee [%][c]
ABS config. AABS [%] AABS config.
1
2
2a
3
DBU
DBU
10 CH2Cl2 ϩ 5 toluene
15 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
15 CH2Cl2 ϩ 5 toluene
15 CH2Cl2 ϩ 5 toluene
15 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
15 CH2Cl2 ϩ 5 toluene
10 CH2Cl2 ϩ 5 toluene
72
72
48
72
72
72
72
72
48
24
24
66
74
75
71
82
77
89
40
41
67
71
20 (ϩ)
9 (Ϫ)
12
7
2
6
4
1
1
0
0
4
1
20 (2.)
Ϫ
3
4
DBU
21 (Ϫ)
28 (Ϫ)
26 (Ϫ)
34 (Ϫ)
29 (Ϫ)
12 (ϩ)
11 (Ϫ)
2 (Ϫ)
21 (1.)
28 (1.)
26 (1.)
34 (1.)
30 (1.)
Ϫ
Ϫ
Ϫ
Ϫ
4
5
DBU
5
6
DBU
6
7
DBU
7
7
NEt3
8
8
DBU
9
9
DBU
10
11
10
11
DBU
(Ϫ)-quinine
6 (ϩ)
[a]
[b]
1 Mol% [Pd(acac)2] was used as procatalyst, Pd : ligand ratio ϭ 1:4 at 38°C. Ϫ Yields were determined by GC on a Chirasil-Val-L
[c]
column, yields < 0.5% are given as 0%; The ee of ABS was measured by GC on a Chirasil-Val-L column, ee of AABS was measured
by GC on a Lipodex E column.
We subsequently replaced the alkyl group at the stereo-
Though resembling the phosphane imine ligands 2Ϫ11,
genic center with a phenyl and a benzyl group while retain- ligand 12, a tert-butyl tert-leucinate, gave only 1% ee for
ing the hydroxymethyl substituent (ligands 8 and 9). Phos- (Ϫ)-ABS with DBU and triethylamine, respectively. Ligand
phane-imine 8 induced 12% ee and 13% ee for (ϩ)-ABS 13 bears the alkylaromatic system of (ϩ)-dehydroabietyl-
with DBU (entry 8) and triethylamine, respectively. Similar amine with no hydroxy group in its side arm, and this af-
results were obtained with the ligand 9 (entry 9). The results forded racemic ABS with DBU and 4% ee (ϩ)-ABS with
with the ligands 8 and 9 show that phenyl and benzyl sub- (Ϫ)-quinine. With 2Ϫ3% ee for (ϩ)- and (Ϫ)-ABS, the
stituents in the ligand side arm are not as efficient as alkyl phosphane amine ligands 14a and 14b were much less en-
groups. Ligand 10 differs slightly from ligand 3 in that it antioselective than the corresponding phosphane imine li-
has a hydroxybenzyl group instead of a hydroxymethyl gands 2a and 2b. The change from one-arm phosphane im-
group, and this gave only 2% ee for (Ϫ)-ABS with DBU ines 2a and 2b to the three-arm phosphane trisimines 15a
(entry 10) and 2% ee for (ϩ)-ABS with (Ϫ)-quinine. Thus, and 15b did not lead to an improvement in the optical in-
the hydroxymethyl substituent, essential for the optical in- duction of ABS. With 15a an enantiomeric excess of 6% ee
duction, cannot be replaced by a hydroxybenzyl substitu- and a yield of only 2% for (Ϫ)-ABS was found with the
ent. With ligand 11, which compared to 2b contains a sec- base DBU. Interestingly, with 15a ABS was formed with
ond hydroxy group and an additional phenyl substituent in the inverse configuration in comparison to 2a. Ligand 15b
its chiral side arm, no enantioselectivity was found for ABS in combination with (Ϫ)-quinine provided (Ϫ)-ABS with
using DBU[37]. With (Ϫ)-quinine and ligand 11, 6% ee (ϩ)- 3% ee and in only 1% yield.
ABS was induced (entry 11).
With the phosphane octaimine 16[45], which bears -alan-
Uniformly, within the phosphane imine ligand family inol in its eight side arms, an enantiomeric excess of 11%
2Ϫ11, the ligands with the (S)-configuration of their N- ee for (ϩ)-ABS was obtained. The incorporation of (R)-
bound stereogenic center (2bϪ7, 9Ϫ11) induce predomi- (Ϫ)-2-amino-1-butanol, -valinol, -leucinol, -phenylalan-
48
Eur. J. Inorg. Chem. 1998, 43Ϫ54

Asymmetric Catalysis, 115
FULL PAPER
inol in the ligand arms and changing the ligand backbone Figure 2. Conformation of the (η3-allyl)Pd complex containing two
phosphane imine ligands 2a after energy minimisation
to the 1,2-disubstituted phenylene bridge decreased the en-
antioselectivity for ABS[37]. The deltacyclanephosphane
17[46] gave 8% ee for (ϩ)-ABS with DBU. The combination
of 17 with (Ϫ)-quinine led to an improvement to 11% ee
for (ϩ)-ABS (double stereoselection). The phosphane ox-
azoline ligands 18Ϫ20[47][48] induced up to 99% ee in al-
lylations in which the stereogenic center is formed in the
allylic system and up to 13% ee in allylations in which the
stereogenic center is formed in the nucleophile[49][50]. In the
catalytic allylation of BS with allyl acetate, an enantiomeric
excess of only about 1% ee for (ϩ)-ABS was obtained.
Phosphane-oxazoline 20 in combination with (Ϫ)-quinine
provided 4% ee for (ϩ)-ABS, probably an effect of (Ϫ)-
quinine.
A Model for the Catalytically Active Phosphane-Imine Pd
Complex
The in situ system consisting of the dimeric (η3-allyl)pal-
ladium chloride and the phosphane imine 2a was investi-
gated by 31P{1H}-NMR spectroscopy. In dichloromethane
using a Pd/2a ratio of 1:1, signals at δ ϭ 36.34 (broad) (0.6
P), 32.81 (0.1 P), 32.63 (0.1 P) and 21.82/21.74 (0.2 P) were
observed, the most intense of which could be due to a Pd
complex containing 2a as a PϪN chelating ligand similar
above the allyl plane, can influence the incoming 1,5-di-
methylbarbiturate anion (Figure 2), e.g. by forming hydro-
gen bonds, thus differentiating the NϪH and the NϪMe
side of the BS anion (Figure 3), which explains the enantio-
to the Pd complexes of the phosphane oxazoline li-
gands[8][9][11][13]. On changing from a Pd/2a ratio of 1:1 to
a ratio of 1:4, all the signals present in the 1:1 experiment
disappeared. Two new doublets of equal intensity arose at
selectivity of long-arm ligands of the type 2a.
δ ϭ 45.86 and 18.67 (coupling constant 341.8 Hz), belong-
ing to the complex [(η3-allyl)Pd(2a)2]Cl. Excess 2a exhibited
Figure 3. Differentiation between the NϪH and the NϪMe side of
the incoming 1,5-dimethylbarbiturate anion in the formation of the
a singlet at δ ϭ Ϫ9.92 with double the intensity compared
to the signals of the coordinated 2a. In addition there were
small signals at δ ϭ 33.28 (oxide of 2a), 23.27 and 15.20
(broad) (CH2Cl2/CD2Cl2, external 85% H3PO4, 162 MHz,
21°C). After seven days the signal at δ ϭ 15.20 had disap-
peared and that at δ ϭ 33.28 had increased (increasing oxi-
dation). Thus, using a Pd to ligand ratio of 1:4, a (η3-al-
lyl)Pd complex formed containing two inequivalent phos-
phane imine ligands 2a.
new stereogenic center
Based on X-ray studies[51][52][53][54][55] and calculations[56]
of (η3-allyl)Pd complexes a molecular modelling study of
an (η3-allyl)Pd complex with two phosphane imine ligands
2a was carried out using the tripos force field. Figure 2
shows a conformation of this complex obtained on energy
minimisation.
We thank the Fonds der Chemischen Industrie and the BASF,
Ludwigshafen, for support of this work, and PD Dr. S. Dove for
carrying out the force field calculations.
The two phosphane imine ligands 2a turn out to be dif-
ferent. The hydroxy group of one phosphane imine extends
˚
to about 3 A above the allyl plane, whereas the hydroxy
Experimental Section
group of the other phosphane imine remains below the allyl
plane in accord with their NMR inequivalence (Figure 2).
Furthermore, a substituted and an unsubstituted phenyl
ring, one from each of the two phosphane imines, form a
π-stacking pair which participates in the conformational or-
ganisation of the two monodentate ligands. Taking into ac-
count this phenyl/phenyl interaction, the two monodentate
NMR spectra were recorded on a Bruker WM 250 or on a
Bruker ARX 400 instrument. Chemical shifts (δ) are reported in
parts per million (ppm) vs. internal tetramethylsilane (TMS) (1H)
and external 85% H3PO4 (31P{1H}). Ϫ FD mass spectra were ob-
tained on a Finnigan MAT 95, and for EI mass spectra a Varian
MAT 311A was used. Ϫ Gas chromatography measurements were
carried out on a Fisons Instruments GC 8000 series 8130. For inte-
gration of gas chromatograms a Spectra-Physics SP 4270 was used.
ligands in the (η3-allyl)Pd complex can be considered a
˚
ЉchelatingЉ ligand. The hydroxy group, which is about 3 A Ϫ Infra red spectra were recorded on a Beckman Gitterspektro-
Eur. J. Inorg. Chem. 1998, 43Ϫ54
49

H. Brunner, I. Deml, W. Dirnberger, B. Nuber, W. Reißer
meter IR 4240. Ϫ Optical rotations were measured on a Perkin- (CDCl3, 162 MHz): δ ϭ Ϫ9.69 (s). Ϫ C23H24NOP (361.42): calcd.
FULL PAPER
Elmer polarimeter 241 (1-dm cell). Ϫ Melting points were obtained
on a Büchi SMP 20 (uncorrected). The X-ray structure analysis
was carried out on a AED II diffractometer.
C 76.44, H 6.69, N 3.87; found C 76.15, H 6.76, N 3.95.
(Ϫ)-2-[N-(S)-1Ј-Hydroxybut-2Ј-ylcarbaldimino]phenyl-
(diphenyl)phosphane (2b): Variant A with 3.45 mmol (307 mg, 325
µl) of (S)-(ϩ)-2-amino-1-butanol. Recrystallisation analogous to 2a
(946 mg, 76%), m.p. 84.5Ϫ85°C. Ϫ [α]2D5 ϭ Ϫ73 (c ϭ 1, EtOH). Ϫ
IR, 1H NMR, 31P{1H} NMR analogous to 2a. Ϫ C23H24NOP
(361.42): calcd. C 76.44, H 6.69, N 3.87; found C 76.19, H 6.84,
N 3.91.
(Ϫ)-Quinine, (Ϫ)-cinchonidine, and (ϩ)-cinchonine were pur-
chased from Merck, (ϩ)-quinidine from Fluka, (R)-(Ϫ)-2-amino-
1-butanol, (S)-(ϩ)-2-amino-1-butanol, (1S,2S)-(ϩ)-2-amino-1-
phenyl-1,3-propanediol, (ϩ)-dehydroabietylamine, -tert-butyl tert-
leucinate from Merck, -alaninol, -valinol, -leucinol, -α-phenyl-
glycinol, -phenylalaninol, -norephedrine from Fluka, (2S,3S)-
(ϩ)-2-amino-3-methyl-1-pentanol (-isoleucinol) from Aldrich, (S)-
(ϩ)-2-amino-3,3-dimethyl-1-butanol (-tert-leucinol) from Sigma
or Fluka and used without further purification. 2-Formylphenyl(di-
phenyl)phosphane (1)[39][40][41] and tris(2-formylphenyl)phos-
phane[40][41][42] were prepared as described in the literature. New
(Ϫ)-2-[N-(S)-1Ј-Hydroxyprop-2Ј-ylcarbaldimino]phenyl-
(diphenyl)phosphane (3): Variant A with 3.45 mmol (259 mg) of
(S)-(ϩ)-2-amino-1-propanol (-alaninol). Recrystallisation from 50
ml of petroleum ether (boiling range 40Ϫ60°C): colourless crystals
(995 mg, 83%), m.p. 62Ϫ63°C. Ϫ [α]2D5 ϭ Ϫ19 (c ϭ 1, CH2Cl2). Ϫ
IR (KBr): ν˜ ϭ 1646 cmϪ1 (CϭN). Ϫ 1H NMR (CDCl3, 400 MHz):
analytical data are reported for 1. The phosphane octaimine 16[45]
,
4
4
the deltacyclanephosphane 17[46] and the osphane oxazolines
18Ϫ20[47][48] were prepared using procedures described in the litera-
ture. All the syntheses and catalytic reactions were carried out with
δ ϭ 8.72 (d, JP ϭ 4.0 Hz, 1 H, azomethine), 7.80 (ddd, J ϭ 1.5
Hz, 4JP ϭ 3.8 Hz, 3J ϭ 7.6 Hz, 1 H, ArϪH3), 7.40 (pseudo dt,
4J ϭ 1.3 Hz, 3J ϭ 7.5 Hz, 1 H, ArϪH), 7.35Ϫ7.21 (m, 11 H,
4
3
3
ArϪH), 6.89 (ddd, J ϭ 1.2 Hz, JP ϭ 4.5 Hz, J ϭ 7.7 Hz, 1 H,
ArϪH6), 3.51Ϫ3.34 (m, 3 H, NCHϪCHHЈϪOH), 1.89 (s br, 1 H,
OH), 0.96 (d, 3J ϭ 6.3 Hz, 3 H, NCHϪCH3). Ϫ 31P{1H} NMR
(CDCl3, 162 MHz): δ ϭ Ϫ9.86 (s). Ϫ MS (FD, CH2Cl2); m/z (%):
347.0 (100) [Mϩ]. Ϫ C22H22NOP (347.40): calcd. C 76.06, H 6.38,
N 4.03; found C 75.74, H 6.28, N 3.98.
dried solvents under
Schlenk techniques.
a nitrogen atmosphere using standard
2-Formylphenyl(diphenyl)phosphane (1): Yellow crystals, m.p.
115Ϫ116°C. Ϫ IR (KBr): ν ϭ 1701, 1679 cmϪ1 (CϭO). Ϫ 1H
˜
4
NMR (CDCl3, 400 MHz): δ ϭ 10.50 (d, JP ϭ 5.4 Hz, 1 H, alde-
hyde), 7.97 (ddd, 4J ϭ 1.8 Hz, 4JP ϭ 3.7 Hz, 3J ϭ 7.2 Hz, 1 H,
ArϪH3), 7.51Ϫ7.43 (m, 2 H, ArϪH4 and ArϪH5), 7.37Ϫ7.25 (m,
10 H, ArϪH), 6.99Ϫ6.95 (m, 1 H, ArϪH6). Ϫ 31P{1H} NMR
(CDCl3, 162 MHz): δ ϭ Ϫ10.99 (s). Ϫ C19H15OP (290.30): calcd.
C 78.61, H 5.21; found C 78.66, H 5.07.
(Ϫ)-2-[N-(S)-1Ј-Hydroxy-3Ј-methylbut-2Ј-ylcarbaldimino]-
phenyl(diphenyl)phosphane (4): Variant A with 3.45 mmol (356 mg)
of (S)-(ϩ)-2-amino-3-methyl-1-butanol (-valinol). Recrystallisa-
tion from 50 ml of petroleum ether (boiling range 40Ϫ60°C):
colourless crystals (795 mg, 61%), m.p. 51Ϫ53°C. Ϫ [α]2D5 ϭ Ϫ68
(c ϭ 1, CH2Cl2). Ϫ IR (KBr): ν˜ ϭ 1643 cmϪ1 (CϭN). Ϫ 1H NMR
General Procedure for the Preparation of the Phosphane Imines
2Ϫ13. Ϫ Variant A: 3.45 mmol of the optically active primary
amine was dissolved in 10 ml of methanol. 2-Formylphenyl(diphen-
yl)phosphane (1) (3.45 mmol; 1.00 g) and 10 ml of methanol were
added to the above solution with stirring and the mixture refluxed
for 3 h. After cooling to room temp. the methanol was removed
and the resulting residue was dried and recrystallized as described
for the individual phosphane imines.
4
(CDCl3, 400 MHz): δ ϭ 8.61 (d, JP ϭ 3.8 Hz, 1 H, azomethine),
4
4
3
7.78 (ddd, J ϭ 1.5 Hz, JP ϭ 3.8 Hz, J ϭ 7.6 Hz, 1 H, ArϪH3),
7.42 (pseudo dt, 4J ϭ 1.2 Hz, 3J ϭ 7.5 Hz, 1 H, ArϪH), 7.35Ϫ7.20
(m, 11 H, ArϪH), 6.89 (ddd, J ϭ 1.2 Hz, JP ϭ 4.4 Hz, J ϭ 7.7
4
3
3
Hz, 1 H, ArϪH6), 3.60 (ddd, J ϭ 5.7 Hz, J ϭ 7.4 Hz, J ϭ 11.1
Hz, 1 H, NCHϪCHHЈϪOH), 3.54 (ddd, 3J ϭ 3.4 Hz, 3J ϭ 8.2
Hz, 2J ϭ 11.1 Hz, 1 H, NCHϪCHHЈϪOH), 2.87 (pseudo dt, 3J ϭ
3.4 Hz, 3J ϭ 7.2 Hz, 1 H, CHHЈϪNCHϪCH), 1.80 (ddd, JP ϭ 3.3
Hz, 3J ϭ 5.7 Hz, 3J ϭ 8.2 Hz, 1 H, CHHЈϪOH), 1.68 (pseudo
3
3
2
Variant B: 3.45 mmol (or 4.14 mmol) of the optically active pri-
mary amine was dissolved in 30 ml of dichloromethane. 2-Formyl-
phenyl(diphenyl)phosphane (1) (3.45 mmol; 1.00 g), 1 g of Na2SO4,
and 10 ml of dichloromethane were added to the above solution
with stirring and the mixture refluxed for 3 h. After cooling to
room temp. the Na2SO4 was filtered off and washed with dichloro-
methane. The dichloromethane was removed and the resulting resi-
due was treated as in variant A.
3
3
oct., J ϭ 6.8 Hz, 1 H, CH3ϪCHϪCH3Ј), 0.81 (d, J ϭ 6.8 Hz, 3
H, CHϪCH3), 0.58 (d, J ϭ 6.8 Hz, 3 H, CHϪCH3Ј). Ϫ 31P{1H}
3
NMR (CDCl3, 162 MHz): δ ϭ Ϫ9.33 (s). Ϫ MS (FD, CH2Cl2);
m/z (%): 375.1 (100) [Mϩ]. Ϫ C24H26NOP (375.45): calcd. C 76.78,
H 6.98, N 3.73; found C 76.47, H 6.99, N 3.77.
(Ϫ)-2-[N-(S)-1Ј-Hydroxy-4Ј-methylpent-2Ј-ylcarbaldimino]-
(ϩ)-2-[N-(R)-1Ј-Hydroxybut-2Ј-ylcarbaldimino]phenyl- phenyl(diphenyl)phosphane (5): Variant A with 3.45 mmol (404 mg)
(diphenyl)phosphane (2a): Variant A with 3.45 mmol (307 mg, 325 of (S)-(ϩ)-2-amino-4-methyl-1-pentanol (-leucinol). Recrystalli-
µl) of (R)-(Ϫ)-2-amino-1-butanol. Recrystallisation from 50 ml of sation analogous to 4: colourless needles (962 mg, 72%), m.p.
petroleum ether (boiling range 40Ϫ60°C): colourless needles (1.02
68Ϫ69°C. Ϫ [α]2D5 ϭ Ϫ81 (c ϭ 1, CH2Cl2). Ϫ IR (KBr): ν˜ ϭ 1637
1
4
g, 82%), m.p. 84.5Ϫ85°C. Ϫ [α]2D5 ϭ ϩ73 (c ϭ 1, EtOH). Ϫ IR cmϪ1 (CϭN). Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 8.71 (d, JP
ϭ
(KBr): ν˜ ϭ 1639 cmϪ1 (CϭN). Ϫ 1H NMR (CDCl3, 400 MHz): 4.3 Hz, 1 H, azomethine), 7.81 (ddd, J ϭ 1.3 Hz, JP ϭ 3.8 Hz,
4
4
4
4
4
δ ϭ 8.68 (d, JP ϭ 4.0 Hz, 1 H, azomethine), 7.80 (ddd, J ϭ 1.5
3J ϭ 7.6 Hz, 1 H, ArϪH3), 7.41 (pseudo dt, J ϭ 1.2 Hz, 3J ϭ 7.5
Hz, 4JP ϭ 3.8 Hz, 3J ϭ 7.7 Hz, 1 H, ArϪH3), 7.41 (pseudo dt, Hz, 1 H, ArϪH), 7.35Ϫ7.20 (m, 11 H, ArϪH), 6.87 (ddd, 4J ϭ
3
3
4J ϭ 1.2 Hz, J ϭ 7.6 Hz, 1 H, ArϪH4 or ArϪH5), 7.35Ϫ7.20 (m,
1.1 Hz, 3JP ϭ 4.5 Hz, J ϭ 7.6 Hz, 1 H, ArϪH6), 3.50 (m, 2 H,
4
3
3
3
3
3
11 H, ArϪH), 6.88 (ddd, J ϭ 1.2 Hz, JP ϭ 4.5 Hz, J ϭ 7.6 Hz, NCHϪCH2ϪOH), 3.31 (ddt, J ϭ 3.9 Hz, J ϭ 9.4 Hz, J ϭ 5.4
1 H, ArϪH6), 3.52 (d, J ϭ 5.4 Hz, 2 H, NCHϪCH2ϪOH), 3.09 Hz, 1 H, CH2ϪNCHϪCH2), 1.85 (s br, 1 H, OH), 1.30 (ddd, 3J ϭ
3
(ddt, 3J
ϭ
4.8 Hz, 3J
ϭ
8.5 Hz, 3J
ϭ
5.4 Hz,
1
H,
4.1 Hz, 3J ϭ 9.4 Hz, 2J ϭ 12.8 Hz, 1 H, NCHϪCHHЈϪCH),
CHHЈϪNCHϪCH2), 1.82 (s br, 1 H, OH), 1.42 (dqd, 3J ϭ 4.8 Hz, 1.21Ϫ1.09 [m, 1 H, CHHЈϪCH(CH3)2], 1.08 (ddd, 3J ϭ 3.9 Hz,
2
3J ϭ 7.4 Hz, J ϭ 13.7 Hz, 1 H, NCHϪCHHЈϪCH3), 1.30 (qdd, 3J ϭ 9.1 Hz, 2J ϭ 12.8 Hz, 1 H, NCHϪCHHЈϪCH), 0.72 (pseudo
3J ϭ 7.4 Hz, 3J ϭ 8.5 Hz, 2J ϭ 13.7 Hz, 1 H, NCHϪCHHЈϪCH3), t, J ϭ 6.6 Hz, 6 H, H3CϪCHϪCH3Ј). Ϫ 31P{1H} NMR (CDCl3,
3
0.62 (pseudo t, 3J ϭ 7.4 Hz, 3 H, CHHЈϪCH3). Ϫ 31P{1H} NMR 162 MHz): δ ϭ Ϫ10.05 (s). Ϫ MS (FD, CH2Cl2); m/z (%): 389.1
50
Eur. J. Inorg. Chem. 1998, 43Ϫ54

Asymmetric Catalysis, 115
(100) [Mϩ]. Ϫ C25H28NOP (389.48): calcd. C 77.10, H 7.25, N 3.60; MHz): δ ϭ 8.29 (d, 4JP ϭ 3.8 Hz, 1 H, azomethine), 7.59 (ddd,
FULL PAPER
found C 77.02, H 7.27, N 3.55.
4J ϭ 1.4 Hz, 4JP ϭ 3.8 Hz, 3J ϭ 7.6 Hz, 1 H, ArϪH3), 7.38 (pseudo
4
3
dt, J ϭ 1.3 Hz, J ϭ 7.5 Hz, 1 H, ArϪH), 7.35Ϫ7.10 (m, 14 H,
ArϪH), 6.95Ϫ6.92 (m, 2 H, ArϪH), 6.89 (ddd, 4J ϭ 1.2 Hz, 3JP ϭ
4.3 Hz, 3J ϭ 7.7 Hz, 1 H, ArϪH6), 3.61Ϫ3.51 (m, 2 H,
NCHϪCH2ϪOH), 3.45Ϫ3.37 (m, 1 H, CH2ϪNCHϪCHHЈ), 2.71
(dd, 3J ϭ 5.3 Hz, 2J ϭ 13.4 Hz, 1 H, NCHϪCHHЈϪPh), 2.48 (dd,
(Ϫ)-2-[N-(S)-1Ј-Hydroxy-(3ЈS)-methylpent-2Ј-ylcarbald-
imino]phenyl(diphenyl)phosphane (6): Variant A with 3.45 mmol
(404 mg) of (2S,3S)-(ϩ)-2-amino-3-methyl-1-pentanol (-isoleuci-
nol). Recrystallisation analogous to 4: colourless crystals at Ϫ20°C,
which form a highly viscous oil at room temp. (798 mg, 59%). Ϫ
[α]2D5 ϭ Ϫ72 (c ϭ 1, CH2Cl2). Ϫ IR (film): ν˜ ϭ 1648 cmϪ1 (CϭN).
2
3J ϭ 8.3 Hz, J ϭ 13.4 Hz, 1 H, NCHϪCHHЈϪPh), 2.12 (s br, 1
H, OH). Ϫ 31P{1H} NMR (CDCl3, 162 MHz): δ ϭ Ϫ9.14 (s). Ϫ
MS (FD, CH2Cl2); m/z (%): 423.2 (100) [Mϩ]. Ϫ C28H26NOP
(423.49): calcd. C 79.41, H 6.19, N 3.31; found C 79.12, H 6.37,
N 3.32.
1
4
Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 8.61 (d, JP ϭ 4.0 Hz, 1 H,
4
4
3
azomethine), 7.78 (ddd, J ϭ 1.5 Hz, JP ϭ 3.8 Hz, J ϭ 7.6 Hz,
1 H, ArϪH3), 7.42 (pseudo dt, 4J ϭ 1.2 Hz, 3J ϭ 7.5 Hz, 1 H,
ArϪH), 7.35Ϫ7.20 (m, 11 H, ArϪH), 6.89 (ddd, 4J ϭ 1.2 Hz, 3JP ϭ
4.4 Hz, 3J ϭ 7.7 Hz, 1 H, ArϪH6), 3.63 (dd, 3J ϭ 7.3 Hz, 2J ϭ
11.2 Hz, 1 H, NCHϪCHHЈϪOH), 3.54 (dd, 3J ϭ 3.4 Hz, 2J ϭ
(ϩ)-2-[N-(1ЈR,2ЈS)-1Ј-Hydroxy-1Ј-phenylprop-2Ј-ylcarbal-
dimino]phenyl(diphenyl)phosphane (10): Variant B with 3.45 mmol
(522 mg) of (1R,2S)-(Ϫ)-2-amino-1-phenyl-1-propanol (-nore-
phedrine). Recrystallisation from 18 ml of petroleum ether (boiling
range 40Ϫ60°C)/THF (5:1): colourless crystals (918 mg, 63%), m.p.
74.5Ϫ75°C. Ϫ [α]D25 ϭ ϩ8, [α]D25 ϭ ϩ50 (c ϭ 1, CH2Cl2). Ϫ IR
(KBr): ν˜ ϭ 1649 cmϪ1 (CϭN). Ϫ 1H NMR (CDCl3, 400 MHz):
3
11.2 Hz, 1 H, NCHϪCHHЈϪOH), 2.97 (pseudo dt, J ϭ 3.4 Hz,
3J ϭ 7.3 Hz, 1 H, CHHЈϪNCHϪCH), 1.90 (s br, 1 H, OH),
1.50Ϫ1.39 (m, 1 H, CHϪCHHЈϪCH3), 1.10 (dqd, 3J ϭ 3.8 Hz,
3J ϭ 7.3 Hz, 2J ϭ 13.2 Hz, 1 H, CHϪCHHЈϪCH3), 0.83Ϫ0.71
(m, 1 H, H3CϪCHϪCHHЈ), 0.77 (d, 3J ϭ 6.9 Hz, 3 H, H3CϪCH),
0.67 (pseudo t, 3J ϭ 7.3 Hz, 3 H, CHHЈϪCH3). Ϫ 31P{1H} NMR
(CDCl3, 162 MHz): δ ϭ Ϫ9.36 (s). Ϫ MS (FD, CH2Cl2); m/z (%):
389.2 (100) [Mϩ]. Ϫ C25H28NOP (389.48): calcd. C 77.10, H 7.25,
N 3.60; found C 76.88, H 7.48, N 3.75.
4
4
δ ϭ 8.72 (d, JP ϭ 4.0 Hz, 1 H, azomethine), 7.82 (ddd, J ϭ 1.3
Hz, 4JP ϭ 3.8 Hz, 3J ϭ 7.6 Hz, 1 H, ArϪH3), 7.41 (pseudo dt,
4J ϭ 1.2 Hz, 3J ϭ 7.5 Hz, 1 H, ArϪH), 7.37Ϫ7.19 (m, 16 H,
4
3
3
ArϪH), 6.89 (ddd, J ϭ 1.1 Hz, JP ϭ 4.6 Hz, J ϭ 7.8 Hz, 1 H,
ArϪH6), 4.63 (d, J ϭ 3.6 Hz, 1 H, NCHϪHCOHϪPh), 3.49 (dq,
3
(Ϫ)-2-[N-(S)-1Ј-Hydroxy-3Ј,3Ј-dimethylbut-2Ј-ylcarbaldimino]-
phenyl(diphenyl)phosphane (7): Variant A with 3.45 mmol (404 mg)
3
3J ϭ 3.6 Hz, J ϭ 6.5 Hz, 1 H, H3CϪNCHϪHCOH), 3.15 (s br,
1 H, OH), 0.77 (d, 3J ϭ 6.5 Hz, 3 H, H3CϪNCH). Ϫ 31P{1H}
NMR (CDCl3, 162 MHz): δ ϭ Ϫ9.80 (s). Ϫ MS (FD, CH2Cl2); m/z
(%): 423.1 (100) [Mϩ]. Ϫ C28H26NOP (423.49) и 0.5 THF: calcd. C
78.41, H 6.58, N 3.05; found C 78.16, H 6.73, N 3.30.
of
(S)-(ϩ)-2-amino-3,3-dimethyl-1-butanol
(-tert-leucinol).
Recrystallisation analogous to 4: colourless needles (969 mg, 72%),
m.p. 85Ϫ86°C. Ϫ [α]D25 ϭ Ϫ66 (c ϭ 1, CH2Cl2). Ϫ IR (KBr): ν˜ ϭ
1
1648 cmϪ1 (CϭN). Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 8.59 (d,
(ϩ)-2-[N-(1ЈS,2ЈS)-1Ј,3Ј-Dihydroxy-1Ј-phenylprop-2Ј-ylcarbal-
dimino]phenyl(diphenyl)phosphane (11): Variant A with 3.45 mmol
(577 mg) of (1S,2S)-(ϩ)-2-amino-1-phenyl-1,3-propanediol.
Recrystallisation from 150 ml of petroleum ether (boiling range
40Ϫ60°C): microcrystalline colourless solid (1.24 g, 82%), m.p.
64Ϫ66°C. Ϫ [α]2D5 ϭ ϩ88 (c ϭ 1, EtOH). Ϫ IR (KBr): ν˜ ϭ 1650
4JP ϭ 3.8 Hz, 1 H, azomethine), 7.78 (ddd, 4J ϭ 1.4 Hz, JP ϭ 3.8
4
Hz, 3J ϭ 7.7 Hz, 1 H, ArϪH3), 7.42 (pseudo dt, 4J ϭ 1.2 Hz, 3J ϭ
7.5 Hz, 1 H, ArϪH), 7.35Ϫ7.20 (m, 11 H, ArϪH), 6.89 (ddd, 4J ϭ
3
3
3
1.1 Hz, JP ϭ 4.3 Hz, J ϭ 7.7 Hz, 1 H, ArϪH6), 3.64 (ddd, J ϭ
3.2 Hz, 3J ϭ 10.4 Hz, 2J ϭ 11.0 Hz, 1 H, NCHϪCHHЈϪOH),
3.54 (ddd, 3J ϭ 3.9 Hz, 3J ϭ 9.2 Hz, 2J ϭ 11.0 Hz, 1 H,
1
4
cmϪ1 (CϭN). Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 8.30 (d, JP
ϭ
6
3
3
NCHϪCHHЈϪOH), 2.86 (ddd, JP ϭ 0.8 Hz, J ϭ 3.2 Hz, J ϭ
9.2 Hz, 1 H, NCHϪCHHЈ), 1.46 (pseudo td, JP ϭ 3.9 Hz, 3J ϭ
3.9 Hz, 3J ϭ 10.4 Hz, 1 H, CHHЈϪOH), 0.74 [s, 9 H, C(CH3)3].
4
4
3.2 Hz, 1 H, azomethine), 7.60 (ddd, J ϭ 1.4 Hz, JP ϭ 3.8 Hz,
3J ϭ 7.6 Hz, 1 H, ArϪH3), 7.43 (pseudo dt, J ϭ 1.2 Hz, 3J ϭ 7.5
4
Hz, 1 H, ArϪH), 7.38Ϫ7.19 (m, 16 H, ArϪH), 6.93 (ddd, 4J ϭ 1.2
Ϫ
31P{1H} NMR (CDCl3, 162 MHz): δ ϭ Ϫ9.59 (s). Ϫ MS (FD,
Hz, 3JP ϭ 4.5 Hz, 3J ϭ 7.7 Hz, 1 H, ArϪH6), 4.66 (d, 3J ϭ 5.3
CH2Cl2); m/z (%): 389.1 (100) [Mϩ]. Ϫ C25H28NOP (389.48): calcd.
Hz,
1
H, NCHϪHCOHϪPh), 3.61Ϫ3.34 (m,
3
H,
C 77.10, H 7.25, N 3.60; found C 77.12, H 7.25, N 3.52.
NCHϪCH2ϪOH), 2.76 (s br, 1 H, OH), 1.70 (s br, 1 H, OH). Ϫ
31P{1H} NMR (CDCl3, 162 MHz): δ ϭ Ϫ5.69 (s). Ϫ C28H26NO2P
(439.49): calcd. C 76.52, H 5.96, N 3.19; found C 75.70, H 5.96,
N 3.38.
(ϩ)-2-[N-(R)-1Ј-Hydroxy-2Ј-phenyleth-2Ј-ylcarbaldimino]-
phenyl(diphenyl)phosphane (8): Variant A with 3.45 mmol (473 mg)
of
(R)-(Ϫ)-2-amino-2-phenyl-1-ethanol
(-α-phenylglycinol).
Recrystallisation analogous to 4: yellowish solid (884 mg, 63%),
(Ϫ)-2-[N-(S)-3Ј,3Ј-Dimethyl-tert-butyloxy-2Ј-butyrylcarbald-
imino]phenyl(diphenyl)phosphane (12): Variant B with 4.14 mmol
(775 mg) of (S)-(ϩ)-2-amino-3,3-dimethyl-tert-butyloxybutyric
acid (-tert-butyl tert-leucinate). Recrystallisation analogous to 4:
yellowish crystals (1.37 g, 86%), m.p. 94Ϫ95°C. Ϫ [α]2D5 ϭ Ϫ99 (c ϭ
1, CH2Cl2). Ϫ IR (KBr): ν˜ ϭ 1742 cmϪ1 (CϭO), 1643 (CϭN). Ϫ
1H NMR (CDCl3, 400 MHz): δ ϭ 8.88 (pseudo qd, J ϭ 0.6 Hz,
m.p. 47Ϫ49°C. Ϫ [α]2D5 ϭ ϩ141 (c ϭ 1, CH2Cl2). Ϫ IR (KBr): ν˜ ϭ
1
1649 cmϪ1 (CϭN). Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 8.65 (d,
4
4JP ϭ 3.5 Hz, 1 H, azomethine), 7.71 (ddd, 4J ϭ 1.3 Hz, JP ϭ 3.8
Hz, 3J ϭ 7.6 Hz, 1 H, ArϪH3), 7.41 (pseudo dt, 4J ϭ 1.2 Hz, 3J ϭ
7.5 Hz, 1 H, ArϪH), 7.38Ϫ7.23 (m, 11 H, ArϪH), 7.19Ϫ7.15 (m,
3
3 H, ArϪH), 6.96Ϫ6.92 (m, 3 H, ArϪH), 4.41 (dd, J ϭ 5.2 Hz,
3J
ϭ 7.4 Hz, 1 H, NCHϪCHHЈ), 3.70Ϫ3.62 (m, 2 H,
4
4JP ϭ 5.4 Hz, 1 H, azomethine), 8.18 (ddd, 4J ϭ 1.4 Hz, JP ϭ 4.0
NCHϪCHHЈϪOH), 2.46 (s br, 1 H, OH). Ϫ 31P{1H} NMR
(CDCl3, 162 MHz): δ ϭ Ϫ8.37 (s). Ϫ MS (FD, CH2Cl2); m/z (%):
409.2 (100) [Mϩ]. Ϫ C27H24NOP (409.47): calcd. C 79.20, H 5.91,
N 3.42; found C 78.98, H 5.96, N 3.19.
Hz, 3J ϭ 7.8 Hz, 1 H, ArϪH3), 7.38 (pseudo ddt, J ϭ 0.6 Hz, 4J ϭ
3
1.3 Hz, J ϭ 7.5 Hz, 1 H, ArϪH), 7.36Ϫ7.23 (m, 11 H, ArϪH),
4
3
3
6.85 (dddd, J ϭ 0.6 Hz, J ϭ 1.3 Hz, JP ϭ 4.7 Hz, J ϭ 7.7 Hz,
1 H, ArϪH6), 3.43 [d, J ϭ 0.6 Hz, 1 H, OOCϪNCHϪC(CH3)3],
1.40 [s, 9 H, COOC(CH3)3], 0.87 [s, 9 H, NCHϪC(CH3)3]. Ϫ
31P{1H} NMR (CDCl3, 162 MHz): δ ϭ Ϫ14.11 (s). Ϫ MS (FD,
CH2Cl2); m/z (%): 459.2 (100) [Mϩ]. Ϫ C29H34NO2P (459.60):
calcd. C 75.79, H 7.46, N 3.05; found C 75.79, H 7.32, N 3.03.
(Ϫ)-2-[N-(S)-1Ј-Hydroxy-3Ј-phenylprop-2Ј-ylcarbaldimino]-
phenyl(diphenyl)phosphane (9): Variant B with 3.45 mmol (522 mg)
of
(S)-(Ϫ)-2-amino-3-phenyl-1-propanol
(-phenylalaninol).
Recrystallisation from 90 ml of pentane: yellowish solid (1.01 g,
69%), m.p. 42Ϫ43°C. Ϫ [α]2D5 ϭ Ϫ167 (c ϭ 1, CH2Cl2). Ϫ IR
(KBr): ν˜ ϭ 1652, 1641 cmϪ1 (CϭN). Ϫ 1H NMR (CDCl3, 400
(ϩ)-2-[N-(1ЈR,11ЈR,12ЈS)-Dehydroabietylcarbaldimino]-
phenyl(diphenyl)phosphane (13): Variant B with 4.14 mmol (1.18 g)
Eur. J. Inorg. Chem. 1998, 43Ϫ54
51

H. Brunner, I. Deml, W. Dirnberger, B. Nuber, W. Reißer
FULL PAPER
of (ϩ)-dehydroabietylamine. Recrystallisation from 50 ml of etha-
nol/methanol (3:2): colourless crystals (1.36 g, 71%), m.p.
74Ϫ75°C. Ϫ [α]2D5 ϭ ϩ13 (c ϭ 1, CH2Cl2). Ϫ IR (KBr): ν˜ ϭ 1638
evaporated to dryness. Recrystallisation from 90 ml of petroleum
ether (boiling range 40Ϫ60°C) afforded yellow crystals of 15a (or
15b). Yield 0.90 g (80%).
1
4
cmϪ1 (CϭN). Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 8.87 (d, JP
ϭ
(ϩ)-Tris[2-(N-(R)-1Ј-hydroxybut-2Ј-ylcarbaldimino)-
phenyl]phosphane (15a): M.p. 106.5Ϫ107°C. Ϫ [α]2D5 ϭ ϩ105 (c ϭ
4
4
5.1 Hz, 1 H, azomethine), 7.99 (ddd, J ϭ 1.1 Hz, JP ϭ 3.9 Hz,
3J ϭ 7.7 Hz, 1 H, ArϪH3), 7.37Ϫ7.20 (m, 12 H, ArϪH), 7.15 (d,
1, EtOH). Ϫ IR (KBr): ν ϭ 1647 cmϪ1 (CϭN). Ϫ 1H NMR
˜
3J ϭ 8.2 Hz, 1 H, ArϪH), 6.97 (dd, J ϭ 1.7 Hz, J ϭ 8.2 Hz, 1
4
3
4
(CDCl3, 400 MHz): δ ϭ 8.45 (d, JP ϭ 2.0 Hz, 3 H, azomethine),
2
H, ArϪH), 6.85Ϫ6.82 (m, 2 H, ArϪH), 3.38 (d, J ϭ 11.9 Hz, 1
4
3
7.65Ϫ7.59 (m, 3 H, ArϪH3), 7.38 (pseudo dt, J ϭ 1.2 Hz, J ϭ
H, NϪCHHЈ), 3.26 (dd, J ϭ 1.0 Hz, 2J ϭ 11.9 Hz, 1 H,
7.5 Hz, 3 H, ArϪH), 7.16 (pseudo t, 3J ϭ 7.4 Hz, 3 H, ArϪH),
NϪCHHЈ), 2.81 (sept., 3J ϭ 6.9 Hz, 1 H, H3CϪCHϪCH3),
6.92 (ddd, J ϭ 1.2 Hz, JP ϭ 3.8 Hz, J ϭ 7.7 Hz, 3 H, ArϪH6),
3.76 (s br, 3 H, OH), 3.51Ϫ3.37 (m, 6 H, NCHϪCHHЈϪOH),
3.02Ϫ2.94 (m, 3 H, CHHЈϪNCHϪCHHЈ), 1.25Ϫ1.10 (m, 3 H,
NCHϪCHHЈϪCH3), 0.92Ϫ0.78 (m, 3 H, NCHϪCHHЈϪCH3),
0.41 (pseudo t, 3J ϭ 7.3 Hz, 9 H, CHHЈϪCH3). Ϫ 31P{1H} NMR
(CDCl3, 162 MHz): δ ϭ Ϫ10.12 (s). Ϫ MS (EI, 70 eV); m/z (%):
559.2 (100) [Mϩ], 486.3 (95) [Mϩ Ϫ C4H8OH], 470.3 (39), 459.3
(15) [Mϩ Ϫ CHNC4H8OH]. Ϫ C33H42N3O3P (559.69): calcd. C
70.82, H 7.56, N 7.51; found C 70.38, H 7.64, N 7.45.
4
3
3
3
2.77Ϫ2.62 (m, 2 H, ArϪCH2ϪCH2), 2.18 (d, Jaxial ϭ 12.8 Hz, 1
H, C3CϪH), 1.83Ϫ1.15 (m, 8 H, 4 CH2), 1.22 (d, J ϭ 6.9 Hz, 6
3
H, H3CϪCHϪCH3), 1.18 (s, 3 H, C3CϪCH3), 0.92 (s, 3 H,
C3CϪCH3). Ϫ 31P{1H} NMR (CDCl3, 162 MHz): δ ϭ Ϫ13.82 (s).
Ϫ MS (EI, 70 eV); m/z (%): 557.3 (29) [Mϩ], 542.2 (24) [Mϩ
Ϫ
CH3], 303.2 (100) [Ph2P(C6H4)CHNHCHϩ2 ], 302.1 (75)
[Ph2P(C6H4)CHNCH2ϩ], 288.1 (64) [Ph2P(C6H4)CHNϩ].
Ϫ
C39H44NP (557.76): calcd. C 83.98, H 7.95, N 2.51; found C 83.62,
H 7.99, N 2.58.
(Ϫ)-Tris[2-(N-(S)-1Ј-hydroxybut-2Ј-ylcarbaldimino)-
phenyl]phosphane (15b): M.p. 106.5Ϫ107°C. Ϫ [α]2D5 ϭ Ϫ105 (c ϭ
Synthesis of the Phosphane amines 14a and 14b: 750 mg (2.08
mmol) of 2a (or 2b) was dissolved in 30 ml of methanol at Ϫ10°C
followed by the addition of 100 mg (2.64 mmol) of NaBH4. The
solution was allowed to come to room temp. within 4 h and heated
for 2 h at 50°C. After removal of the methanol, 25 ml of water
and 120 ml of dichloromethane were added to the residue. The
dichloromethane layer was extracted twice with 25 ml of water and
dried with Na2SO4. After filtration the dichloromethane was re-
moved and the resulting oil was recrystallized with 20 ml of petro-
leum ether (boiling range 40Ϫ60°C) to give colourless crystals at
Ϫ20°C. At room temp. the crystals of 14a (or 14b) became a highly
viscous oil. Yield: 665 mg (88%).
1
1, EtOH). Ϫ IR, H NMR, 31P{1H} NMR and MS analogous to
15a. Ϫ C33H42N3O3P (559.69): calcd. C 70.82, H 7.56, N 7.51;
found C 70.68, H 7.55, N 7.53.
X-ray Structure Analysis of 15b: C33H42N3O3P (559.69); crystal
dimensions 0.30 ϫ 0.45 ϫ 0.50 mm; two independent molecules in
the unit cell; crystal system triclinic; space group C1/1, P1, (1); unit
˚
cell dimensions a ϭ 10.068(6), b ϭ 11.611(7), c ϭ 15.125(9) A, α ϭ
3
˚
88.08(5), β ϭ 73.55(5), γ ϭ 70.35(5)°, V ϭ 1593.3 A ; Z ϭ 2; den-
sity dcalcd. ϭ 1.17g/cm3; µ(Mo-Kα) ϭ 0.12 mmϪ1; 3.0° < 2θ < 50.5°;
total no. of reflections 8224, unique reflections 7653, unique reflec-
tions with I > 2.5σ(I) 6037; F(000) ϭ 600; diffractometer AED II;
temp. Ϫ70°C. The structure was solved by direct methods using
the SHELXTL PLUS version 4.2/800 program system. Hydrogen
(ϩ)-N-[2-(Diphenylphosphanyl)benzyl]-N-[(R)-1Ј-hydroxybut-
25
2Ј-yl]amine (14a): [α]2D5 ϭ ϩ11, [α]
ϭ ϩ75 (c ϭ 1, CH2Cl2). Ϫ
365
1H NMR (CDCl3, 400 MHz): δ ϭ 7.41 (ddd, J ϭ 1.2 Hz, JP
ϭ
4
4
atoms were calculated by the HFIX program; R ϭ 0.044; Rw
ϭ
4.5 Hz, J ϭ 7.6 Hz, 1 H, ArϪH3), 7.36Ϫ7.23 (m, 11 H, ArϪH),
7.18 (pseudo dt, 4J ϭ 1.2 Hz, 3J ϭ 7.6 Hz, 1 H, ArϪH), 6.92 (ddd,
4J ϭ 1.2 Hz, 3JP ϭ 4.5 Hz, 3J ϭ 7.6 Hz, 1 H, ArϪH6), 4.10 (s very
br, 2 H, OH and NH), 3.98 (dd, 4JP ϭ 1.6 Hz, 2J ϭ 12.6 Hz, 1
3
3
3
˚
˚
0.033; residual electron density max. 0.36 e/A , min. Ϫ0.39 e/A .
Further details of the crystal structure investigation may be ob-
tained from the Fachinformationszentrum Karlsruhe, D-76344 Eg-
genstein-Leopoldshafen (Germany), on quoting the depository
number CSD-407358.
4
2
H, ArϪCHHЈϪNH), 3.95 (dd, JP ϭ 1.2 Hz, J ϭ 12.6 Hz, 1 H,
ArϪCHHЈϪNH), 3.60 (dd, 3J ϭ 3.8 Hz, 2J ϭ 10.9 Hz, 1 H,
NCHϪCHHЈϪOH), 3.24 (dd, 3J ϭ 5.9 Hz, 2J ϭ 10.9 Hz, 1 H,
1,5-Dimethylbarbituric Acid (BS): Handling of BS has been de-
scribed in the literature[57][58][59][60], but a synthesis and characteri-
sation was not given. 16.6 g (0.722 mol) of sodium was dissolved
under reflux in 230 ml of absolute ethanol under a nitrogen atmos-
phere. After cooling to about 60°C, 25.9 g (0.350 mol) of N-meth-
ylurea and 20 ml of ethanol and, after 15 min, 59.8 ml (0.350 mol,
61.0 g) of methylmalonic acid diethyl ester were added (colourless
precipitate). After 2 h reflux the solvent was removed. The dry
residue was dissolved in 400 ml of hot water and, with vigorous
stirring, 100 ml of conc. hydrochloric acid was added. After cooling
to room temp. the precipitated product was filtered off and washed
with 400 ml of cooled diethyl ether. Recrystallisation from 600 ml
of absolute ethanol/methanol (3:1) provided colourless crystals,
which were washed with petroleum ether (boiling range 40Ϫ60°C)
and dried under high vacuum at 60°C for several days. Yield 34.7
g (63%), m.p. 170Ϫ171°C. Ϫ IR (KBr): ν˜ ϭ 1703 cmϪ1 (CϭO). Ϫ
1H NMR ([D6]DMSO, 250 MHz): δ ϭ 11.27 (s, 1 H, NH), 3.72
(q, 3J ϭ 6.9 Hz, 1 H, HCCH3), 3.06 (s, 3 H, NϪCH3), 1.34 (d,
3J ϭ 6.9 Hz, 3 H, HCCH3). Ϫ C6H8N2O3 (156.14): calcd. C 46.15,
H 5.16, N 17.94; found C 46.21, H 5.18, N 17.90.
3
3
NCHϪCHHЈϪOH), 2.51 (pseudo dq, J ϭ 3.8 Hz, J ϭ 6.3 Hz,
1 H, CHHЈϪNCHϪCH2), 1.36Ϫ1.27 (m, 2 H, NCHϪCH2ϪCH3),
0.82 (t, J ϭ 7.4 Hz, 3 H, CH2ϪCH3). Ϫ 31P{1H} NMR (CDCl3,
3
162 MHz): δ ϭ Ϫ15.44 (s). Ϫ MS (FD, CH2Cl2); m/z (%): 363.1
(100) [Mϩ]. Ϫ C23H26NOP (363.44): calcd. C 76.01, H 7.21, N 3.85;
found C 74.70, H 7.10, N 3.82.
(Ϫ)-N-[2-(Diphenylphosphanyl)benzyl]-N-[(S)-1Ј-hydroxybut-
25
2Ј-yl]amine (14b): [α]2D5 ϭ Ϫ11, [α] ϭ Ϫ75 (c ϭ 1, CH2Cl2). Ϫ
365
1H NMR, 31P{1H} NMR and MS analogous to 14a.
Ϫ
C23H26NOP (363.44): calcd. C 76.01, H 7.21, N 3.85; found C
75.72, H 7.28, N 3.82.
Synthesis of the Phosphane trisimines 15a and 15b: 3.79 ml (40.4
mmol, 3.60 g) of (R)-(Ϫ)-2-amino-1-butanol [or (S)-(ϩ)-2-amino-
1-butanol] was dissolved in 10 ml of dichloromethane and 1 g (7
mmol) of Na2SO4 was added. A solution of 0.70 g (2.0 mmol) of
tris(2-formylphenyl)phosphane[40][41][42] in 10 ml of dichlorometh-
ane was added dropwise during 15 min. After 16 h reflux the yellow
solution was filtered and the dichloromethane was evaporated. The
excess amine was removed by heating under high vacuum at 100°C
Enantioselective Palladium-Catalysed Allylation of 1,5-Dimeth-
for 6 h. The resulting yellow oil was dissolved in 50 ml of dichloro- ylbarbituric Acid (BS) with Allyl Acetate:150 mg (0.961 mmol) of
methane, stirred for 15 min and filtered. The clear solution was
BS and 1.01 mmol of base were dissolved in 10 ml of dichlorometh-
52
Eur. J. Inorg. Chem. 1998, 43Ϫ54

Asymmetric Catalysis, 115
FULL PAPER
ane at 38°C under a nitrogen atmosphere. The following bases were (ligand PPh3). The THF was removed and 30 ml of dichlorometh-
used: DBU (150 µl, 153 mg), triethylamine (140 µl, 102 mg), BSA ane was added followed by standard work-up. The resulting liquid
(247 µl, 205 mg), 1.5  LDAиTHF solution in cyclohexane (673
µl), NBu4OH solution in methanol (12.5%) (2.59 ml, 2.10 g), (Ϫ)-
was dissolved in dichloromethane and filtered through aluminia
to remove traces of ABS. For further purification a bulb to bulb
quinine (328 mg), (ϩ)-quinidine (328 mg), (Ϫ)-cinchonidine (297 distillation was carried out to give a colourless liquid at 130°C and
mg) and (ϩ)-cinchonine (297 mg). BS and base can be replaced by 3 Torr. Yield 222 mg (98%), b.p. ഠ 80°C at 3 Torr. Ϫ IR (film):
1
382 mg (0.961 mmol) of NBu4BS. After 5 min stirring 2.93 mg ν˜ ϭ 1689 cmϪ1 (CϭO). Ϫ H NMR (CDCl3, 400 MHz): δ ϭ 5.82
3
3
3
(0.0096 mmol) of palladium(II) acetylacetonate, 0.0384 mmol of a
(pseudo tdd, J ϭ 6.0 Hz, Jcis ϭ 10.2 Hz, Jtrans ϭ 17.1 Hz, 1 H,
monodentate ligand or 0.0192 mmol of a bidentate ligand and 10 NϪCHHЈϪCHϭCHH), 5.54 (tdd, 3J ϭ 7.5 Hz, Jcis ϭ 10.1 Hz,
3
ml of a 1:1 mixture of methylene choride/toluene were added to the 3Jtrans ϭ 17.0 Hz, 1 H, CϪCH2ϪCHϭCHH), 5.28 (pseudo qd,
3
clear solution. After 2.0 min the reaction was started by the ad-
2/4J ϭ 1.3 Hz, Jtrans ϭ 17.1 Hz, 1 H, NϪCHHЈϪCHϭCHHtrans),
dition of 115 µl (1.07 mmol, 107 mg) of allyl acetate. The solution 5.21 (pseudo qd, 2/4J
ϭ
1.3 Hz, Jcis
ϭ
10.2 Hz,
1
H,
3
3
was stirred for 24, 48 or 72 h at 38°C. For work-up the solution
was diluted with 10 ml of dichloromethane and extracted succes-
NϪCHHЈϪCHϭCHcisH), 5.10 (md, Jtrans ϭ 17.0 Hz, 1 H,
CϪCH2ϪCHϭCHHtrans), 5.08 (md, Jcis ϭ 10.1 Hz, 1 H,
3
4
3
sively with 10 ml and 5 ml of 0.2  hydrochloric acid and three CϪCH2ϪCHϭCHcisH), 4.50 (pseudo tdd, J ϭ 1.3 Hz, J ϭ 6.0
times with 5 ml of water. The organic layer was dried with Na2SO4.
The solid was filtered off and washed with 10 ml of dichlorometh-
Hz, 2J ϭ 14.6 Hz, 1 H, NϪCHHЈϪCHϭCHH), 4.44 (pseudo tdd,
3
2
4J ϭ 1.3 Hz, J ϭ 6.0 Hz, J ϭ 14.6 Hz, 1 H, NϪCHHЈϪCHϭ
ane. Removal of the solvent from the filtrate left the dry product CHH), 3.30 (s, 3 H, NϪCH3), 2.69 (d, 3J ϭ 7.5 Hz, 2 H,
ready for GC analysis. To the product was added about 25 mg of
benzil (GC standard) and 6 ml of dichloromethane. This solution
was used to determine the enantiomeric excess of ABS and the
yield of ABS and AABS by GC on a Chirasil-Val-L column. Fur-
CϪCH2ϪCHϭCHH), 1.55 (s, 3 H, CϪCH3). Ϫ GC-MS (EI, 70
eV); m/z (%): 236.0 (100) [Mϩ], 221.0 (99) [Mϩ Ϫ CH3], 194.9 (17)
[Mϩ Ϫ C3H5], 138.0 (90), 41.0 (59) [C3Hϩ5 ]. Ϫ C12H16N2O3
(236.27): calcd. C 61.00, H 6.83, N 11.86; found C 60.95, H 7.04,
thermore, a chromatographic analysis of 3 ml of the solution on N 11.63.
alumina was carried out in a Pasteur pipette, in which only AABS
and benzil eluted. The eluate was used for the determination of the
enantiomeric excess of AABS on a Lipodex E column.
5-Chloromethyl-1,5-dimethylbarbituric Acid: Found in catalyses
with yields of ABS <3%. Ϫ C7H9ClN2O3 (204.61). Ϫ GC-MS (EI,
70 eV); m/z (%): 204.0 (14) [Mϩ], 189.0 (30) [Mϩ Ϫ CH3], 169.0
(100) [Mϩ Ϫ Cl], 155.1 (6) [Mϩ Ϫ CH2Cl], 69.0 (62), 40.9 (27).
Conditions: 25 m Chirasil-Val-L fused silica capillary column
(0.25 mm inner diameter, 0.12 µm film thickness, from Chrom-
pack), column temp. 130°C, He pressure 1.2 bar, injector temp.
250°C, detector temp. 230°C [flame ionisation], retention times:
(±)-AABS 4.24 min, standard benzil 17.10 min, (Ϫ)-ABS 19.09
min, (ϩ)-ABS 20.33 min, separation factor α[(ϩ)-ABS/(Ϫ)-ABS] ϭ 1.07,
resolution R[(ϩ)-ABS/(Ϫ)-ABS] ϭ 1.80, mass correlation factors:
fbenzil/ABS ϭ 1.77 and fbenzil/AABS ϭ 1.50; retention times for the
enantiomers of 5-chloromethyl-1,5-dimethylbarbituric acid were
34.74 and 35.32 min, respectively. Ϫ 50 m Lipodex E fused silica
capillary column (0.25 mm inner diameter, coated with octakis(2,6-
Molecular Modelling: The calculations were carried out with the
program Sybyl 6.1 using the standard tripos force field on a Silicon
Graphics Indigo workstation (tripos metal file with standard geo-
metries and van der Waals parameters of transition metal atoms).
After establishment of the geometry of the (η3-allyl)Pd system and
the two P atoms, the P substituents were attached in their standard
geometries, the ortho-substituent being oriented towards the allyl
plane. A conformational analysis with energy minimisation was
carried out.
di-O-pentyl-3-O-butyryl)-γ-cyclodextrine,
from
Macherey-
[1]
H. Brunner, A. Winter, J. Organomet. Chem., in press.
[2]
Nagel), column temp. 110°C (after 45 min the column temp. was
raised by 12°C/min to 170°C), H2 pressure 2.07 bar, injector temp.
250°C, detector temp. 240°C (flame ionisation), retention times:
(1.)-AABS 42.45 min, (2.)-AABS 43.84 min, standard benzil 65.0
min, separation factor α[(2.)-AABS/(1.)-AABS]
R[(2.)-AABS/(1.)-AABS] ϭ 1.74.
J.-C. Fiaud in Metal Promoted Selectivity in Organic Synthesis
(Eds.: A. F. Noels, M. Graziani, A. J. Hubert), Kluwer, Dord-
recht, 1991, p.107.
[3]
T. Hayashi in Catalytic Asymmetric Synthesis (Ed.: I. Ojima),
VCH, New York, 1993, p. 325.
[4]
ϭ 1.03, resolution
B. M. Trost, D. L. Van Vranken, Chem. Rev. 1996, 96, 395.
[5]
B. M. Trost, Acc. Chem. Res. 1996, 29, 355.
[6]
U. Leutenegger, G. Umbricht, C. Fahrni, P. v. Matt, A. Pfaltz,
5-Allyl-1,5-dimethylbarbituric Acid (ABS): Main product of cat-
alysis under standard reaction conditions. Maximum yield 184 mg
(98%) (Table 1, entry 3). Recrystallisation from water/ethanol 175:1
provided colourless crystals, m.p. 131Ϫ131.5°C. Ϫ IR (KBr): ν˜ ϭ
1758, 1716, 1680 cmϪ1 (CϭO). Ϫ 1H NMR (CDCl3, 400 MHz):
Tetrahedron 1992, 48, 2143.
[7]
B. M. Trost, D. L. Van Vranken, C. Bingel, J. Am. Chem. Soc.
1992, 114, 9327.
[8]
P. v. Matt, A. Pfaltz, Angew. Chem. 1993, 105, 614; Angew.
Chem., Int. Ed. Engl. 1993, 32, 566.
[9]
J. Sprinz, G. Helmchen, Tetrahedron Lett. 1993, 34, 1769.
[10]
δ ϭ 9.03 (s, 1 H, NH), 5.60 (tdd, 3J ϭ 7.5 Hz, Jcis ϭ 10.1 Hz,
3
G. J. Dawson, C. G. Frost, J. M. J. Williams, S. J. Coote, Tetra-
hedron Lett. 1993, 34, 3149.
3Jtrans ϭ 17.0 Hz, 1 H, CH2ϪCHϭCHH), 5.14 (md, Jtrans ϭ 17.0
3
[11]
O. Reiser, Angew. Chem. 1993, 105, 576; Angew. Chem., Int. Ed.
3
Hz, 1 H, CH2ϪCHϭCHHtrans), 5.11 (md, Jcis ϭ 10.1 Hz, 1 H,
CH2ϪCHϭCHcisH), 3.28 (s, 3 H, NϪCH3), 2.71 (d, J ϭ 7.5 Hz,
2 H, CH2ϪCHϭCHH), 1.57 (s, 3 H, CϪCH3). Ϫ GC-MS (EI, 70
eV); m/z (%): 195.9 (6) [Mϩ], 180.9 (100) [Mϩ Ϫ CH3], 41.0 (32)
[C3Hϩ5 ]. Ϫ C9H12N2O3 (196.21): calcd. C 55.09, H 6.16, N 14.28;
found C 55.09, H 6.23, N 14.09.
Engl. 1993, 105, 547.
3
[12]
P. v. Matt, O. Loiseleur, G. Koch, A. Pfaltz, C. Lefeber, T.
Feucht, G. Helmchen, Tetrahedron: Asymmetry 1994, 5, 573.
[13]
J. V. Allen, S. J. Coote, G. J. Dawson, C. G. Frost, C. J. Martin,
J. M. J. Williams, J. Chem. Soc., Perkin Trans. 1 1994, 1, 2065.
[14]
J. M. Brown, D. I. Hulmes, P. J. Guiry, Tetrahedron 1994, 50,
4493.
[15]
B. M. Trost, B. Breit, S. Peukert, J. Zambrano, J. W. Ziller, An-
3,5-Diallyl-1,5-dimethylbarbituric Acid (AABS): By-product of
catalysis under standard reaction conditions (yield 0Ϫ19%). To
produce AABS, the standard catalysis was carried out with a 3-fold
excess of the base DBU (450 µl) and a 21-fold excess of allyl acetate
(2.20 ml) in refluxing THF (10 ml) with a reaction time of 24 h
gew. Chem. 1995, 107, 2577; Angew. Chem., Int. Ed. Engl. 1995,
34, 2386.
[16]
P. Wimmer, M. Widhalm, Chem. Monthly 1996, 127, 669.
[17]
A. Pfaltz, Acta Chem. Scand. 1996, 50, 189.
[18]
ˆ
´
`
J.-P. Genet, D. Ferroud, S. Juge, J. R. Montes, Tetrahedron Lett.
1986, 27, 4573.
Eur. J. Inorg. Chem. 1998, 43Ϫ54
53

H. Brunner, I. Deml, W. Dirnberger, B. Nuber, W. Reißer
FULL PAPER
[19]
[41]
[42]
ˆ
´
`
J. -P. Genet, S. Juge, J. R. Montes, J. -M. Gaudin, J. Chem.
H. Brunner, A. F. M. M. Rahman, Chem. Ber. 1984, 117, 710.
M. R. Whitnall, K. K. (M.) Hii, M. Thornton-Pett, T. P. Kee,
J. Organomet. Chem. 1997, 529, 35.
A. Terfort, H. Brunner, J. Chem. Soc., Perkin Trans. 1, 1996,
1467.
Soc., Chem. Commun. 1988, 718.
[20]
[21]
W. J. Doran, J. Org. Chem. 1960, 25, 1737.
J. Knabe, W. Rummel, H. P. Büch, N. Franz, Arzneim.-Forsch./
Drug Res. 1978, 28 (II), 7, 1048.
[43]
[44]
[22]
[23]
J. Knabe, Arzneim.-Forsch./Drug Res. 1989, 39 (II), 11, 1379.
J. Knabe, H.-D. Höltje, Lehrbuch der Pharmazeutischen Chemie,
Wissenschaftliche Verlagsgesellschaft, Stuttgart, 1991.
J. Knabe, Pharmazie 1993, 12, 883.
D. J. Berrisford, C. Bolm, K. B. Sharpless, Angew. Chem. 1995,
107, 1159; Angew. Chem., Int. Ed. Engl. 1995, 34, 1059.
M. Janura, Dissertation, Universität Regensburg, 1997.
H. Brunner, A. Reimer, Chem. Ber. 1997, 120, 1495.
J. V. Allen, G. J. Dawson, C. G. Frost, J. M. J. Williams, S. J.
Coote, Tetrahedron 1994, 50, 799.
[45]
[46]
[47]
[24]
[25]
[26]
[27]
J. Knabe, R. Kräuter, Arch. Pharm. 1965, 298, 1.
J. Knabe, R. Kräuter, K. Philipson, Tetrahedron Lett. 1965, 571.
United States Patent, 2872448 (Eli Lilly; ert. 3. 2. 1959; Prior.
4. 4. 1956).
[48]
M. Peer, J. C. de Jong, M. Kiefer, T. Langer, H. Rieck, H.
Schell, P. Sennhenn, J. Sprinz, H. Steinhagen, B. Wiese, G.
Helmchen, Tetrahedron 1996, 52, 7547.
[28]
[29]
[30]
[31]
C. Lehmann, Das Ultrakurznarkotikum Methohexital, Springer-
Verlag, Berlin, 1972.
[49]
[50]
I. C. Baldwin, J. M. J. Williams, R. P. Beckett, Tetrahedron:
Asymmetry 1995, 6, 1515.
S. Ebel, Synthetische Arzneimittel, VCH, Weinheim, 1979, p.
101.
I. C. Baldwin, J. M. J. Williams, R. P. Beckett, Tetrahedron:
Asymmetry 1995, 6, 679.
A. Kleemann, J. Engel, Pharmazeutische Wirkstoffe, Thieme
Verlag, Stuttgart, 1982, p. 578.
[51]
[52]
A. E. Smith, Acta Cryst. 1965, 18, 331.
W. Forth, D. Henschler, W. Rummel, Pharmakologie und Toxi-
T. Ohta, T. Hosokawa, S.-I. Murahashi, K. Miki, N. Kasai,
Organometallics 1985, 4, 2080.
kologie, Bibliographisches Institut
Mannheim, 1987, p. 484.
& F. A. Brockhaus,
[53]
[54]
[55]
J. W. Faller, C. Blankenship, B. Whitmore, S. Sena, Inorg. Chem.
1985, 24, 4483.
[32]
[33]
S. Ebel, H. J. Roth, Lexikon der Pharmazie, Thieme Verlag,
Stuttgart, 1987, p. 430.
H. Yang, M. A. Khan, K. M. Nicholas, Organometallics 1993,
12, 3485.
H.-P. Kuemmerle, G. Hitzenberger, K. H. Spitzy, Klinische
Pharmakologie, Band 6, ecomed Verlagsgesellschaft, München,
1988.
P. Barbaro, P. S. Pregosin, R. Salzmann, A. Albinati, R. W.
Kunz, Organometallics 1995, 14, 5160.
[34]
[35]
[56]
[57]
[58]
´
Bundesverband der Pharmazeutischen Industrie, Rote Liste
1996, Editio Cantor Verlag, Aulendorf, 1996, Nr. 65010.
M. Prat, M. Moreno-Man˜as, J. Ribas, Tetrahedron 1988, 44,
7205.
K. J. Szabo, Organometallics 1996, 15, 1128.
E. Stuckey, Q. J. Pharm. Pharmacol. 1942, 15, 370.
R. J. Kuffner, M. T. Bush, L. J. Bircher, J. Am. Chem. Soc. 1957,
79, 1587.
[36]
[37]
[38]
[39]
[40]
[59]
[60]
H. Brunner, J. Fürst, Inorg. Chim. Acta 1994, 220, 63.
W. Reißer, Dissertation, Universität Regensburg, 1997.
W. Reißer, Diplomarbeit, Universität Regensburg, 1994.
A. E. Senar, W. Valient, J. Wirth, J. Org. Chem. 1960, 25, 2001.
G. P. Schiemenz, H. Kaack, Liebigs Ann. Chem. 1973, 1480.
B. H. Bakker, N. P. E. Vermeulen, D. D. Breimer, A. van der
Gen, Recl. Trav. Chim. Pays-Bas 1979, 98, 341.
´
G. Toth, S. Makleit, Acta Chim. Acad. Sci. Hung. 1981, 107,
139.
[97180]
54
Eur. J. Inorg. Chem. 1998, 43Ϫ54