Synthesis and application of chiral phosphorus ligands derived from TADDOL for the asymmetric conjugate addition of diethyl zinc to enones

FULL PAPER

Synthesis and Application of Chiral Phosphorus Ligands Derived from

ADDOL for the Asymmetric Conjugate Addition of Diethyl Zinc to Enones

T

Alexandre Alexakis,*^[a]Jonathan Burton,^[a]Johann Vastra,^[a]Cyril Benhaim,^[a]

[a]

ˆ

´ ´

´

Xavier Fournioux, Alexandra van den Heuvel, Jean-Marc Leveque, Frederique Maze,

and Stephane Rosset^[a]

Keywords: Zinc / Copper catalysis / Asymmetric conjugate addition / T / Phosphorus

Asymmetric conjugate addition of diethylzinc to cyclohexen-

2-one, chalcone, and benzalacetone has been found to occur

with 0.5% copper(II) triflate and 1% chiral phosphite. Cyclic

phosphites derived from TADDOL gave excellent to moderate

enantiomeric excesses. The nature of the exocyclic substitu-

ent of the dioxaphospholane ring is important, but the chiral

induction is imposed by the TADDOL framework. Syntheses

of all the TADDOL ligands are described.

Introduction

nard or a dialkylzinc reagent. This latter possibility, intro-

duced by us in 1993,^[5b]has found increasing popularity and

several chiral phosphorus ligands have since been developed

both by ourselves^[6]and by other authors.^[7]Some of the

most efficient are shown in Scheme 1.

CarbonϪcarbon bond formation through conjugate ad-

dition is best achieved by the use of organocopper re-

agents.^[1]The asymmetric version of this reaction is an ac-

tively studied topic.^[2]Although there have been several suc-

The copper catalyst CuX is extremely important in these

cessful examples involving the use of stoichiometric chiral reactions. Although in our first report^[5b]we used CuI as

auxiliaries on the enone partner, much effort is currently the copper source, we later found Cu(OTf)₂to be a superior

being directed towards the development of catalytic pro- catalyst.^[8]Furthermore, we showed that a trivalent phos-

cesses based on the modification of the organometallic phorus ligand has a strong accelerating effect. Hence, the

partner,^[3]particularly with chiral external ligands.^[4]In this possibility of a ligand-accelerated asymmetric catalysis

context, chiral trivalent phosphorus ligands have emerged arises when a chiral trivalent phosphorus ligand is em-

as being among the most efficient reagents.^[5]In the cata- ployed. As little as 0.5% of copper salt and 1% of chiral

lytic reactions, the primary organometallic is either a Grig- ligand are sufficient to furnish a high yield of the conjugate

Scheme 1. Ligands used for asymmetric conjugate additions

adduct.^[8]The reaction is fast in poorly or non-coordinating

solvents such as toluene, dichloromethane, or even diethyl

ether.

As far as the chiral trivalent phosphorus ligand is con-

cerned, we first used a ligand that had proved very success-

[a]

Department of Organic Chemistry, University of Geneva,

`

30 quai Ernest Ansermet, 1211 Geneve 4, Switzerland

Fax: (internat.) ϩ 41-22/328-7396

E-mail: alexandre.alexakis@chiorg.unige.ch

Eur. J. Org. Chem. 2000, 4011Ϫ4027

WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000

1434Ϫ193X/00/1224Ϫ4011 $ 17.50ϩ.50/0

4011

A. Alexakis et al.

FULL PAPER

ful with stoichiometric lithium diorganocuprates.^[5a,5b]Dur-

ing our work on extending the class of chiral phosphorus

ligands to chiral phosphanes^[6a]and to C₂-symmetrical

phosphites derived from tartrates,^[6b]several other reports

have described excellent results obtained with amino-

phosphinites

and

phosphites

derived

from

binaphthol.^{[7aϪ7c,7e,7h,7l,7m]}For our part, we became interes-

ted in another class of C₂-symmetrical diols, namely the

Scheme 3. Enones used in the present study

Ts

(2,3-O-isopropylidene-1,1,4,4-tetraphenylthrei-

tol).^[5c,5d,7f]These chiral auxiliaries were first used by See-

bach’s group with several impressive applications in asym-

metric synthesis.^[9]We report herein our full results in this

field of investigation.^[10]T is easily prepared from

tartaric acid esters, is ten times cheaper than chiral binaph-

thol,^[11]and its trivalent phosphorus derivatives are rather

stable towards moisture and oxidation. The preparation of

these trivalent phosphorus derivatives is described in the

last part of the article.

The results with simple achiral alcohol derivatives in the

exocyclic position are shown in Table 1. With these ligands,

the chirality is imposed solely by the T framework.

As regards the cyclic enone, cyclohexenone, a comparison

of Entries 1 and 2 or 3 shows that a bulky alcohol slightly

increases the enantioselectivity. A phenyloxy group gave no

significant result (Entry 4). Ligands 1 and 4 were already

known to give very low asymmetric inductions in Rh-cata-

lyzed hydroformylation reactions.^[12]In all cases, the abso-

lute stereochemistry of the conjugate adduct proved to be

the same. With benzalacetone, an acyclic enone, all these

ligands led only to poor ee values. The last Entry (5) refers

to an attempt to use the diphosphite, which is in tautomeric

equilibrium with the P^Vform. The results are similarly dis-

appointing. Thus, it is clear that the effect of the T

moiety as the sole source of chirality is insufficient to pro-

duce high levels of asymmetric induction with both cyclic

and acyclic enones.

Results and Discussion

All the ligands used in this study have the same basic

skeleton (Scheme 2). We have varied the exocyclic substitu-

ent from alcohols, to amines and carbon derivatives. Several

structural variations have been tested in order to gain in-

sight into the relative importance of the T part and

of the steric influence or coordinative ability of the exo-

cyclic part of the ligand.

Table 1. T phosphites with an achiral exocyclic alcohol de-

rivative

Scheme 2. T ligands

All the experiments were performed under the same con-

ditions. First, the complex between Cu(OTf)₂(0.025 mmol)

and the chiral ligand (0.05 mmol) was formed (in a 1:2 ra-

tio) by stirring the mixture at room temperature for 30 min,

Et₂Zn (7 mmol) was then added at Ϫ20 °C, followed finally

by the enone (5 mmol). Usually, the reactions were per-

a

b

Yield of isolated adduct. Ϫ ee determined by G.C. with chiral

capillary column (Lipodex E-0.2 µm, 50 m, 0.25 mm).

A more complex ligand may be prepared by combining

formed at Ϫ20 °C and were allowed to proceed for the times (Ϫ)-T with a chiral exocyclic alcohol (Table 2),

indicated in the tables. Most often, the reactions were run thereby introducing a second source of chiral information.

in CH₂Cl₂solution, in which slightly higher enantiomeric Hence, the possibility of matched and mismatched com-

excesses are generally achieved than in toluene. However, binations arises. Both have been tested, with the two enanti-

with certain classes of ligands, the best ee values were ob- omeric alcohols, or, where the chiral alcohol was only avail-

tained only in toluene. Typically, three enones were tested: able in a single enantiomeric form, we replaced (Ϫ)-T

cyclohexenone as a representative of cyclic s-trans enones, by its enantiomer (ϩ)-T. Thus, for cyclohexenone,

benzalacetone as a representative of acyclic s-cis enones, the ligands prepared with (ϩ)-fenchol (Entries 6 and 7) and

and chalcone, the most extensively tested substrate in the (Ϫ)-borneol (Entries 8 and 9) gave similar or slightly better

literature, albeit a very atypical one (Scheme 3).

results than the simpler ligand with just cyclohexanol (2)

4012

Eur. J. Org. Chem. 2000, 4011Ϫ4027

Chiral Phosphorus Ligands

FULL PAPER

Table 2. T phosphites with a chiral non-functionalized exo- and (ϩ)-menthol (5% and 50% ee, respectively). In both

cyclic alcohol derivative

cases, the same absolute configuration of the adduct is ob-

tained. Thus, although the sense of the asymmetric induc-

tion is imposed by the T part, the degree of enantio-

selection is strongly affected by the exocyclic moiety.

Some experimental modifications were then attempted

using the aforementioned matched ligand 11, which led to

the highest ee yet achieved. For example, the use of toluene

as solvent showed, once again, that CH₂Cl₂is somewhat

better (cf. Entries 11a and 11b). The addition of a small

amount of MeOH (10%) led to a small but not insignificant

improvement (Entry 11c). The aim of this experiment was

to check whether aged solutions of Et₂Zn (containing alco-

holates) may affect the enantioselectivity. Conversely, we

˚

also tested the addition of small or large amounts of 4-A

molecular sieves, as such an additive has been reported to

improve the enantioselection.^[7f]In the present case, using

ligand 13 (Entry 13d) or ligand 33 (Entry 33c, Table 5), no

such effect was observed.

Arylcyclohexanols are generally excellent chiral auxiliar-

ies,^[13]much better than menthol, and sometimes even

better than phenylmenthol. We recently described a facile

preparation of this class of compounds in highly enantioen-

riched form through enantioselective nucleophilic ring-

opening of cyclohexene oxide.^[14]Indeed, ligand 13, pre-

pared from phenylcyclohexanol, gave a better ee (60%,

Entry 13a) than ligand 11 (with menthol, Entry 11a) in

CH₂Cl₂solution. However, to our surprise, this ligand be-

haved much better in toluene solution (Entry 13b), giving

ethylcyclohexanone with up to 96% ee. Et₂O also proved to

be a good solvent in this case (88% ee, Entry 13c). This

unusual behavior in toluene was likewise observed with all

the ligands derived from arylcyclohexanols. The results ob-

tained with the 1-naphthyl (ligands 14 and 15) and 2-naph-

thyl series (ligands 16 and 17) are comparable. Again,

marked differences were observed between the matched and

mismatched combinations of ligands. In all these cases,

toluene appeared to be a better solvent, perhaps due to the

better solubility of these ligands. Their lower solubility in

CH₂Cl₂may also account for the low chemical yields ob-

tained with chalcone and benzalacetone.

Generally speaking, acyclic enones gave poor enantiose-

lectivities. The best results were obtained with ligand 13

(chalcone: 50% ee, Entry 13d; benzalacetone: 35% ee, Entry

13f). Thus, another general trend emerges: A ligand that

improves the enantioselection with cyclohexenone also im-

proves it with chalcone. In contrast, benzalacetone seems

to be less sensitive to the matched or mismatched combina-

a

b

Yield of isolated adduct. Ϫ ee determined by G.C. with chiral

tions of the ligand.

capillary column (Lipodex E-0.2 µm, 50 m, 0.25 mm). Ϫ ^cReaction

is performed in toluene. Ϫ ^d10 mol-% MeOH is added to the mix-

In order to ascertain whether interactions other than

steric ones may be useful for improved enantioselection, we

introduced several functionalized alcohols in the exocyclic

position (see Table 3). The homopropargylic system (ligand

18) was selected for π interactions with the Cu atom (Entry

ture. Ϫ ^eReaction is performed in diethyl ether. Ϫ 4-A molecular

f

˚

sieves were added.

(Table 1, Entry 2, 18% ee). A small but significant difference 18), but the result achieved was comparable to those ob-

in ee is apparent between ligands 8 and 9, corresponding to tained with the non-functionalized alcohols (see Table 1),

matched and mismatched combinations. This difference is indicating a rather low influence of the triple bond. The β-

even more striking with ligands 10 and 11 derived from (Ϫ)- dimethylamino group (ligand 19) was chosen for its ability

Eur. J. Org. Chem. 2000, 4011Ϫ4027

4013

A. Alexakis et al.

FULL PAPER

to coordinate to the organometallic reagent through the ni- that during the course of our work, ligands 26 and 27 were

trogen atom (Entry 19). Although the ee was of the usual also described by Feringa et al. and tested with cyclohex-

low magnitude, a noteworthy reversal of the absolute con- enone and cyclopentenone.^[7f]A bulky diisopropylamino

figuration was observed. Therefore, strong coordination of group (ligand 26, Entry 26) gives a moderate degree of en-

the nitrogen atom to the Cu or Zn atom seems to be in- antioselectivity with the usual absolute configuration, i.e.

volved.

that imposed by the chirality of (Ϫ)-T. Interestingly,

a small dimethylamino group (ligand 27, Entry 27) not only

leads to a higher ee (35%), but also inverts the enantioselec-

tion! The reason for the switch in absolute induction is un-

clear at the present time. The divergence from the reported

value of 54% ee for the same experiment^[7f]could be due to

slight differences in the experimental procedure. Two other

examples (Entries 29 and 30) correspond to a matched and

mismatched combination with a chiral exocyclic amine. Al-

though this exocyclic amino group markedly improved the

results with binaphthol instead of T,^[7b]ligands 29

and 30 were completely inefficient in the T case.

Thus, all the above ligands gave poor to moderate results

with cyclohexenone. In contrast to the aforementioned re-

sults, ligand 28, with an even smaller methylamino group

(Entry 28), gave the highest ee (49%) of this study for

benzalacetone. In this case, coordination of the metal center

by a nitrogen atom cannot be invoked and other effects

must be operative.

Table 3. T phosphites with a functionalized exocyclic alco-

hol derivative

Table 4. T phosphonamidite with an exocyclic amino deriv-

ative

a

b

Yield of isolated adduct. Ϫ ee determined by G.C. with chiral

capillary column (Lipodex E-0.2 µm, 50 m, 0.25 mm). Ϫ ^cReaction

d

is performed in toluene. Ϫ 1 mol-% MeOH is added to the mix-

e

ture. Ϫ Using CuI instead of Cu(OTf)₂.

In the other examples in Table 3, a heteroatom and an

additional source of chiral information are combined.

Again, matched and mismatched combinations may be

formed, both of which had to be checked. When a chiral

center was introduced in the substituents at the nitrogen

atom, a matched combination boosted the ee to 38% (li-

gand 21, Entry 21a), again with a reversal of the absolute

configuration, whereas the ee with the mismatched com-

bination amounted to just 5% (Entry 20a). The situation

was not so clear-cut with derivatives of (ϩ)- or (Ϫ)-ephed-

rines (ligands 22 and 23), nor with the chiral oxazolines

(ligands 24 and 25), both combinations giving low to very

low ee values. It is interesting to note that the latter ligands

(24 and 25) gave excellent results in the Rh-catalysed hy-

drosilylation of ketones.^[15]

a

b

Yield of isolated adduct. Ϫ ee determined by G.C. with chiral

capillary column (Lipodex E-0.2 µm, 50 m, 0.25 mm). Ϫ ^cReaction

d

is performed in toluene. Ϫ 1 mol-% MeOH is added to the mix-

e

ture. Ϫ Using CuI instead of Cu(OTf)₂.

Finally, Table 5 shows the results of some tests with alkyl

It has been shown that the chiral phosphorus amidites and aryl phosphinites. Phosphinite 33 has already been suc-

derived from binaphthol (see Scheme 1) are capable of in- cessfully used in Rh-catalyzed asymmetric hydroformyl-

ducing very high levels of enantiocontrol in conjugate addi- ation and hydrosilylation reactions.^[12]On the other hand,

tion reactions.^[7a,7b]Therefore, a number of T-derived alkyl phosphinamidites gave a moderate degree of enanti-

phosphorus amidites were tested under the standard condi- oselectivity in the copper-catalyzed conjugate addition to

tions; the results are shown in Table 4. It should be noted cyclohexenone.^[7g]As seen in Entries 31 and 33, a simple

4014

Eur. J. Org. Chem. 2000, 4011Ϫ4027

Chiral Phosphorus Ligands

FULL PAPER

Table 5. T phosphonites with an exocyclic carbon framework enantioselection. Similarly disappointing results were ob-

tained using 1-naphtho-T (ligand 36) or 2-naphtho-

T (ligand 37), although the latter has been reported

to be more efficient in other kinds of reactions.^[12]Neverthe-

less, this represents a quite promising means of introducing

new variations in this class of T ligands.

Except in the case of the T-arylcyclohexanol li-

gands 13, 15, and 17, the above results are not among the

best ones in the asymmetric conjugate addition of di-

ethylzinc to enones. However, they allow a better under-

standing of the various factors that affect the efficiency of

these T-derived chiral phosphites. Due to the modu-

lar nature of these ligands, it should be a facile task to fur-

ther vary the nature of the aromatic group on the T

moiety as well as the nature of the acetal protecting group.

Therefore, ligand tuning should be possible, allowing access

to highly selective catalysts.

A final series of experiments was carried out with the aim

of checking whether any radical-type mechanism might be

involved in the above reactions. Indeed, a recent report

showed that in the presence of oxygen diethylzinc may un-

dergo conjugate addition to cyclohexenone.^[16]If isopropyl

iodide (up to 6 equivalents) is admixed, only the iPr moiety

is transferred, and not the Et group. In the experiment

shown in Scheme 4, we did not observe any trace of the iPr

adduct. This clearly demonstrates that no radical mechan-

ism is involved in our copper-catalyzed reactions.

a

b

Yield of isolated adduct. Ϫ ee determined by G.C. with chiral

capillary column (Lipodex E-0.2 µm, 50 m, 0.25 mm). Ϫ ^cReaction

d

is performed in toluene. Ϫ 1 mol-% MeOH is added to the mix-

ture. Ϫ ^eUsing CuI instead of Cu(OTf)₂. Ϫ 4-A molecular sieves

f

˚

were added.

Scheme 4. Proof of non-radical mechanism

Synthesis of the TADDOL Ligands

exocyclic n-butyl (ligand 31) or phenyl group (ligand 33)

leads to some of the highest ee values in this study, with

The synthesis of all of the above ligands is not as trivial

values of 48Ϫ54% with cyclohexenone. This is quite a sur- as it might appear at first glance. Due to the pronounced

prising result, since the only stereochemical information steric congestion between the phenyl groups of T and

comes from T itself! The absolute configuration is the exocyclic substituent, different methods were applied

that usually observed with (Ϫ)-T as sole source of according to the specific ligand needed. The most obvious

stereochemical information. Attempts to improve the ee and classical method (Method A) is summarized in

through the use of a bulkier alkyl substituent (ligand 32 Scheme 5 and the ligands prepared in this way are shown.

with tBu) or other aryl substituents such as o-anisyl (ligand It involves initial preparation of the T-PϪCl interme-

34) or ferrocenyl groups (ligand 35) did not improve the diate in the presence of NEt₃as an HCl scavenger, followed

Scheme 5. Synthesis of T ligands according to Method A

Eur. J. Org. Chem. 2000, 4011Ϫ4027

4015

A. Alexakis et al.

FULL PAPER

Scheme 6. Synthesis of T ligands according to Method AЈ

Scheme 7. Synthesis of T ligands according to Method B

Scheme 8. Synthesis of T ligands according to Method C

by addition of the required alcohol.^[17]The course of both force.^[18]The reaction of the secondary or tertiary alcohol

steps of the reaction could easily be monitored by ³¹P with PCl₃to give the ROϪPCl₂species has to be carried

NMR spectroscopy.

out in the absence of NEt₃to avoid the formation of the

An alternative to the above method is to use the alco- corresponding chloride (RϪCl) and/or di- or triphosphites.

holates instead of the free alcohol. Thus, the bis(lithium) The formation of the ROϪPCl₂intermediate could easily

alcoholate of T is prepared and treated with PCl₃. In be monitored by ³¹P NMR. NEt₃was then added, followed

a separate flask, the lithium alcoholate, lithium amide, or immediately by solid T. This method is particularly

organolithium reagent is prepared and then added to the well suited for ligands bearing an aryl cyclohexanol moiety.

T-PϪCl intermediate. This method (Method AЈ) The ligands prepared according to this method are shown

avoids the formation of triethylammonium chloride (which in Scheme 7.

is sometimes difficult to filter off) since it generates only

In a few particular cases, we took advantage of the com-

LiCl salts. It is also the only method possible when an alkyl mercial availability of some intermediates (MeOϪPCl₂,

or aryl group is required in an exocyclic position. Scheme 6 PhOϪPCl₂, tBuϪPCl₂, PhϪPCl₂). These ligands have al-

summarizes the ligands prepared in this way.

ready been prepared by Seebach,^[19]and we followed the

A reverse method (Method B) was used when a bulky described procedure exactly (Method C), as shown in

alcohol was needed in the exocyclic position. It involves the Scheme 8.

construction of the seven-membered ring in the last step,

Finally, two ligands were prepared in a different way. Li-

thereby taking advantage of cyclization as a driving gand 27 was obtained by reaction of T with

4016

Eur. J. Org. Chem. 2000, 4011Ϫ4027

Chiral Phosphorus Ligands

FULL PAPER

value for the optically pure material, [α]²_D³ϭ ϩ10.5 (c ϭ 2.5 in

EtOH).^[21]Ϫ Chalcone was recrystallized from ethanol and benzal-

acetone from pentane; 2-cyclohexenone was used as received. Dry

THF was distilled from sodium in a recycling still using benzo-

phenone ketyl as indicator. Other solvents were purified by stand-

ard techniques.^[22]Reactions in non-aqueous media were carried

out under dry nitrogen or dry argon unless indicated otherwise.

HMPT,^[7f,19]while ligand 5 was isolated as a by-product

from the various ligand preparations (see Scheme 9).

Typical Procedure for the Conjugate Addition: A 25-mL three-

necked flask was charged with Cu(OTf)₂(9 mg, 0.025 mmol),

CH₂Cl₂(3 mL), and the appropriate ligand (0.05 mmol) either neat

or as a 1  solution in toluene. The resulting mixture was stirred for

30 min and then cooled to Ϫ40 °C, whereupon diethylzinc (7 mL of

a 1  solution in hexane, 7 mmol) was added. The solution was

allowed to warm to Ϫ20 °C, whereupon a solution of the appropri-

ate enone (5 mmol) in CH₂Cl₂(3 mL) was added dropwise. The

reaction mixture was stirred for 1Ϫ3 h at Ϫ20 °C (and for a further

4 h at 0 °C in cases where no evolution was observed) and then

quenched by the careful addition of 2  hydrochloric acid (ca.

5 mL). Diethyl ether (ca. 10 mL) was added at 0 °C and the re-

sulting mixture was stirred until clear. The organic phases were

combined, dried (MgSO₄), and the solvent was removed in vacuo.

Where necessary, chromatography provided the pure adduct.

Scheme 9. Synthesis of T ligands 5 and 27

Conclusion

The T class of chiral trivalent phosphorus ligands

would seem to be very promising for copper-catalyzed en-

antioselective conjugate addition. Although the chirality of

the T backbone is insufficient for achieving high

levels of enantiocontrol, the external chirality introduced by

a chiral alcohol moiety attached to the phosphorus atom

allows an increase of the enantiomeric excess of the conju-

gate adduct up to 96% in the best case. Steric factors play

an essential role and a computational model for the design

of new ligands may be expected on this basis. Some of these

ligands have already been used in the literature for other

transition metal catalyzed reactions. All the new ligands de-

scribed in this study may also be tested in these reactions,

thus widening the scope of their potential applications.

3-Ethylcyclohexanone: Purification by flash chromatography: R_fϭ

0.48 (silica, hexane/AcOEt, 80:20). The enantiomeric excess was

measured by chiral GC: Lipodex E, 25 m, 50 cm³·s^Ϫ1, T ϭ 60 °C,

t_Rϭ 24.72 min (R) enantiomer, t_Rϭ 28.15 min (S) enantiomer;

yellow oil. Ϫ ¹H NMR: δ ϭ 2.50Ϫ0.70 (many m). Ϫ ¹³C NMR:

δ ϭ 25.0 (C-7), 28.6 (C-8), 31.0 (C-4), 36.0 (C-5), 38.8 (C-3), 41.1

(C-6), 47.9 (C-2), 210.9 (C-1).

1,3-Diphenylpentanone: Purification by flash chromatography: R_fϭ

0.44 (silica, pentane/Et₂O, 95:5) gave an ivory-colored solid. The

enantiomeric excess was measured by polarimetry {ref.^[21][α]²_D³

ϭ

ϩ10.5 (c ϭ 2.50 in EtOH) pure (S)-(ϩ) adduct} or by HPLC

[Chiralpak AD, 250 mm, 2% iPrOH in hexane, 1 mL·min^Ϫ1, t_R

ϭ

Experimental Section

7.71 min (ϩ)-(S) enantiomer, t_Rϭ 10.04 min (Ϫ)-(R) enantiomer].

1

Ϫ H NMR: δ ϭ 0.80 (t, J ϭ 7.3 Hz, 3 H, 5-H₃), 1.50Ϫ1.90 (m, 2

General Remarks: ¹H NMR spectra were recorded with Bruker AC-

200 (200 MHz) and AC-400 (400 MHz) spectrometers. Chemical

shifts are quoted in ppm relative to tetramethylsilane (δ ϭ 0) and

were referenced to the residual protons in the solvent. Proton-de-

coupled ¹³C-NMR spectra were recorded with Bruker AC-200

(50 MHz) and AC-400 (100 MHz) spectrometers in the solvent in-

dicated. Chemical shifts are quoted relative to tetramethylsilane

H, 4-H), 3.15Ϫ3.35 (m, 3 H, 2-H₂and 3-H), 7.10Ϫ7.60 (m, 8 H,

aryl H), 7.85Ϫ7.98 (m, 2 H, 2Ј-H, 6Ј-H). Ϫ ¹³C NMR: δ ϭ 12.2

(CH₃CH₂), 29.3 (CH₃CH₂), 43.1 (PhCH), 45.6 (CH₂CO), 126.3 (C

arom.), 127.7, 128.1, 128.5, 128.6, 132.9, 137.4, 144.8, 199.1 (CO).

4-Phenylhexanone: Purification by flash chromatography: R_fϭ 0.4

(silica, cyclohexane/EtOAc, 80:20) gave a yellow oil. Ϫ C₁₂H₁₆O,

M_rϭ 176. The enantiomeric excess was measured by polarimetry

{ref.^[21][α]²_D²ϭ ϩ30 (c ϭ 2.30 in EtOH) pure (S)-(ϩ) adduct} or

´

(δ ϭ 0). Ϫ Mass spectra were recorded at the Universite Pierre et

Marie Curie, Paris, under chemical ionization (CI) conditions using

NH₃as the carrier gas. Ϫ Optical rotations were measured with a

PerkinϪElmer 241 polarimeter, using a cell of 1 dm path length.

The concentration (c) is expressed in g/100 mL. Ϫ Microanalyses

by chiral GC [enantiomer separation: Lipodex E, 25 m, 75 cm³s^Ϫ1

,

T ϭ 75 °C, t_Rϭ 35.05 min (S) enantiomer, t_Rϭ 37.66 min (R)

enantiomer. Ϫ ¹H NMR: δ ϭ 0.78 (t, J ϭ 7.3 Hz, 3 H,

CH₃CH₂CH), 1.62 (m, 2 H, CH₃CH₂CH), 2.02 (s, 3 H, COCH₃),

2.72 (d, J ϭ 7.2 Hz, 2 H, CH₂COPh), 3.03 (m, 1 H, PhCH),

7.33Ϫ7.15 (m, 10 H, ArH). Ϫ ¹³C NMR: δ ϭ 11.9 (CH₃CH₂), 29.3

(CH₃CH₂), 30.6 (COCH₃), 42.9 (PhCH), 50.5 (CH₂COPh), 126.3

(C arom.), 127.5, 128.4, 144.2 (C_q), 208.0 (CO).

´

were carried out by the staff of the Universite Pierre et Marie Curie

Microanalytical Department. Ϫ Melting points (m.p.) were deter-

mined using a Büchi SMP-20 melting point apparatus and are un-

corrected. Ϫ Analytical thin-layer chromatography (TLC) was car-

ried out on Merck pre-coated 0.25 mm thick plates of Kieselgel 60

F₂₅₄; TLC plates for use with phosphorus ligands were first neutral-

ized by exposing them to the vapour from an open bottle of ammo-

nia. Gravity chromatography was carried out using Merck Kiesel-

Synthesis of the Ligands

(1R,7R)-4-Methoxy-9,9-dimethyl-2,2,6,6-tetraphenyl-3,5,8,10-tetra-

gel and P SIL G/UV₂₅₄. Ϫ Enantiomeric excesses for oxa-4-phosphabicyclo[5.3.0]decane (1); Method C: Preparation ac-

the conjugate addition products derived from 2-cyclohexenone and

benzalacetone were measured by chiral GC (chiral capillary column

Ϫ Lipodex E, 0.2 µm, 50 m, 0.25 mm). The enantiomeric excess

for the conjugate addition product derived from chalcone was

measured by comparison of the optical rotation with the literature

cording to the literature.^[18]To a stirred solution of (Ϫ)-T

(1.16 g, 2.5 mmol) in THF (10 mL) at Ϫ70 °C, butyllithium

(3.4 mL of a 1.5  solution in hexanes, 5.13 mmol) was added at

such a rate as to keep the internal temperature below Ϫ50 °C. The

reaction mixture was then stirred at Ϫ70 °C for 10 min, allowed to

Eur. J. Org. Chem. 2000, 4011Ϫ4027

4017

A. Alexakis et al.

FULL PAPER

warm to room temperature, and maintained under these conditions

0.54 (s, 3 H, CH₃-T), 0.89 [s, 9 H, (CH₃)₃C], 1.01 (s, 3 H,

for 3 h. It was then cooled to Ϫ70 °C once more, whereupon neat CH₃-T), 5.33 (d, J ϭ 8.5 Hz, 1 H, 7-H), 6.12 (d, J ϭ 8.5 Hz,

MeOPCl₂was added dropwise at such a rate as to keep the internal

temperature below Ϫ70 °C. The resulting mixture was allowed to

warm to room temperature, stirred for 1 h, and then the solvent

was removed in vacuo. The resulting semi-solid residue was stirred

with toluene (10 mL) for 1 h and then filtered through Celite. The

solvent was subsequently removed in vacuo; purification of the res-

idue by flash chromatography (cyclohexane/diethyl ether/triethyla-

mine, 95:5:1) followed by crystallization from pentane furnished the

title compound as a white crystalline solid (1.13 g, 2.1 mmol, 86%);

1 H, 1-H), 7.15Ϫ6.99 (m, 10 H, ArH), 7.89Ϫ7.68 (m, 10 H, ArH).

Ϫ

¹³C NMR: δ ϭ 26.3, 27.1, 30.1, 30.2, 77.6, 80.2, 80.3, 82.6, 82.9,

83.5, 83.6, 84.8, 85.0, 112.3 (C-9), 127.1, 127.3, 127.4, 127.7, 127.8,

127.9, 129.2, 129.5, 142.5, 142.8, 147.1. Ϫ ³¹P NMR: δ ϭ 132.33.

Ϫ MS (CI, NH₃): m/z (%) ϭ 585 [M ϩ 16 ϩ H]^ϩ(5). Ϫ C₃₅H₃₇O₅P

(568.6): calcd. C 73.93, H 6.56; found C 73.77, H 6.96.

(1R,7R)-9,9-Dimethyl-2,2,6,6-tetraphenyl-4-phenyloxy-3,5,8,10-

tetraoxa-4-phosphabicyclo[5.3.0]decane (4); Method C: Preparation

according to the literature.^[18]To a stirred solution of (Ϫ)-T

(1.16 g, 2.5 mmol) in THF (10 mL) at Ϫ70 °C was added butylli-

thium (3.4 mL of a 1.5  solution in hexanes, 5.13 mmol) at such

a rate as to keep the internal temperature below Ϫ60 °C. The reac-

tion mixture was stirred at Ϫ70 °C for 10 min, then allowed to

warm to room temperature, and maintained under these conditions

for 3 h. It was then cooled to Ϫ70 °C once more, whereupon neat

PhOPCl₂was added dropwise so as to keep the internal temper-

ature below Ϫ60 °C. The resulting mixture was allowed to warm

to room temperature, stirred for 18 h (for convenience), and then

the solvent was removed in vacuo. The resulting semi-solid residue

was stirred with toluene (10 mL) for 1 h and then filtered through

Celite. The solvent was removed in vacuo and purification of the

residue by flash chromatography (cyclohexane/diethyl ether/tri-

ethylamine, 95:5:1) followed by crystallization from pentane fur-

nished the title compound as a white solid (1.17 g, 1.99 mmol,

m.p. 189Ϫ191 °C (from pentane) (ref.^[19]190Ϫ192 °C). Ϫ [α]¹_D⁸

ϭ

Ϫ223 (c ϭ 1.05 in CH₂Cl₂) {ref.^[18][α]^r_D^.t.ϭ Ϫ236.7 (c ϭ 1.17 in

1

CHCl₃)}. Ϫ H NMR: δ ϭ 0.61 (s, 3 H, CH₃-T), 0.88 (s, 3

H, CH₃-T), 3.28 (d, 3 H, CH₃O), 5.49 (dd, J ϭ 8.2 and

1.5 Hz, 1 H, 7-H), 5.62 (d, J ϭ 8.2 Hz, 1 H, 1-H), 7.12Ϫ6.95 (m,

12 H, ArH), 7.82Ϫ7.66 (m, 8 H, ArH). Ϫ ¹³C NMR: δ ϭ 26.6,

27.2, 49.4, 81.5, 82.8, 82.9, 83.4, 85.9, 86.0, 113.3 (C-9), 128.3,

129.3, 142.2, 146.8, 147.0. Ϫ ³¹P NMR: δ ϭ 133.80.

(1R,7R)-4-Cyclohexyloxy-9,9-dimethyl-2,2,6,6-tetraphenyl-3,5,8,10-

tetraoxa-4-phosphabicyclo[5.3.0]decane (2); Method A: To a stirred

solution of (Ϫ)-T (1.16 g, 2.5 mmol) and triethylamine

(1.24 mL, 8.5 mmol) in THF (10 mL) at 0 °C was slowly added

PCl₃(236 µL, 2.6 mmol). A white precipitate was deposited and

the resulting mixture was stirred at room temperature for 10 min

and then cooled to 0 °C once more. A solution of cyclohexanol

(270 mg, 2.7 mmol) in THF (10 mL) was then added and stirring

was continued for 1 h at room temperature. The reaction mixture

was diluted with diethyl ether and filtered through a pad of Celite.

The solvent was removed in vacuo, and the residue was taken-up

in diethyl ether and filtered through Celite; this procedure was re-

peated until addition of ether produced a clear solution. Purifica-

tion by flash chromatography (cyclohexane/diethyl ether/triethyl-

amine, 95:5:1) followed by crystallization from a mixture of diethyl

ether and pentane furnished the title compound as a white solid

(693 mg, 1.16 mmol, 47%); m.p. 189Ϫ191 °C (pentane/diethyl

80%); m.p. 178Ϫ179 °C (pentane) (ref.^[18]177Ϫ177.5 °C). Ϫ [α]¹_D⁸

ϭ

Ϫ167.2 (c ϭ 1.36 in CHCl₃) {ref.^[19][α]¹_D⁸ϭ Ϫ172.8 (c ϭ 1.09 in

1

CHCl₃)}. Ϫ H NMR: δ ϭ 0.70 (s, 3 H, CH₃-T), 0.79 (s, 3

H, CH₃-T), 5.13 (d, J ϭ 8.3 Hz, 1 H, 7-H), 5.59 (d, J ϭ

8.3 Hz, 1 H, 1-H), 6.58 (d, J ϭ 8.3 Hz, 2 H, ArH), 7.63Ϫ7.01 (m,

23 H, ArH). Ϫ ¹³C NMR: δ ϭ 26.9, 27.0, 80.6, 82.5, 82.6, 85.4,

86.6, 113.5 (C-9), 120.3, 120.4, 127.6, 128.3, 129.1, 129.4, 129.5,

141.7, 146.3, 152.5. Ϫ ³¹P NMR: δ ϭ 128.12.

(1R,7R)-9,9-Dimethyl-4-hydrido-4-oxo-2,2,6,6-tetraphenyl-3,5,8,10-

tetraoxa-4-phosphabicyclo[5.3.0]decane (5): This compound was

isolated from various ligand preparations as a white crystalline

solid; m.p. 226Ϫ227 °C (decomp.) (pentane/diethyl ether). Ϫ

1

ether). Ϫ [α]¹_D⁸ϭ Ϫ154.8 (c ϭ 0.75 in CHCl₃). Ϫ H NMR: δ ϭ

0.65 (s, 3 H, CH₃-T), 0.93 (s, 3 H, CH₃-T), 1.34Ϫ1.12

(m, 6 H, cyclohexyl), 1.70Ϫ1.58 (m, 3 H, cyclohexyl), 1.81Ϫ1.75

(br. m, 1 H, cyclohexyl), 4.28Ϫ4.22 (m, 1 H, CHO-cyclohexyl), 5.07

(dd, J ϭ 8.3 and 1.4 Hz, 1 H, 7-H), 5.29 (d, 1 H, J ϭ 8.3 Hz, 1-

H), 7.37Ϫ7.28 (m, 12 H, ArH), 7.63Ϫ7.50 (m, 8 H, ArH). Ϫ ¹³C

NMR: δ ϭ 23.9, 25.3, 26.2, 26.8, 34.1, 73.3, 73.4, 81.1, 82.2, 82.5,

84.4, 84.5, 112.3 (C-9), 127.0, 127.2, 127.4, 127.7, 128.0, 128.7,

129.0, 141.7, 146.4. Ϫ ³¹P NMR: δ ϭ 134.10. Ϫ MS (CI, NH₃):

m/z (%) ϭ 595 [M ϩ H]^ϩ(Ͻ 1), 373 (100). Ϫ C₃₇H₃₉O₅P (594.7)

calcd. C 74.73, H 6.61; found C 74.88, H 6.76.

1

[α]¹_D⁸ϭ Ϫ289.9 (c ϭ 1.56 in CHCl₃). Ϫ H NMR: δ ϭ 0.59 (s, 3

H, CH₃-T), 0.79 (s, 3 H, CH₃-T), 5.24 (d, J ϭ 8.0 Hz,

1 H, 7-H), 5.38 (d, J ϭ 8.0 Hz, 1 H, 1-H), 7.12 (d, J ϭ 7.3 Hz, 1

H, 4-H), 7.45Ϫ7.30 (m, 16 H, ArH), 7.65Ϫ7.61 (m, 4 H, ArH). Ϫ

¹³C NMR: δ ϭ 28.2, 28.7, 81.7, 72.0, 90.6, 90.7, 116.3 (C-9), 128.7,

128.8, 130.3, 130.6, 130.7, 145.1, 145.6. Ϫ ³¹P NMR: δ ϭ Ϫ3.13.

Ϫ MS (CI, NH₃): m/z (%) ϭ 530 [M ϩ NH₄]^ϩ(1), 431 (100). Ϫ

C₃₁H₂₉O₅P (512.5): calcd. C 72.65, H 5.70; found C 72.51, H 5.68.

(1R,7R)-9,9-Dimethyl-2,2,6,6-tetraphenyl-4-(1,3,3-trimethylbicyclo-

[2.2.1]hept-2-yloxy)-3,5,8,10-tetraoxa-4-phosphabicyclo[5.3.0]decane

(6); Method B: To a stirred solution of (ϩ)-fenchol (0.77 g,

5.0 mmol) in THF (5 mL) was slowly added PCl₃(436 µL,

(1R,7R)-4-(tert-Butyloxy)-9,9-dimethyl-2,2,6,6-tetraphenyl-3,5,8,10-

tetraoxa-4-phosphabicyclo[5.3.0]decane (3); Method B: To a stirred

solution of PCl₃(218 µL, 2.5 mmol) and triethylamine (1.08 mL,

7.8 mmol) in THF (10 mL) at 0 °C, a solution of tert-butyl alcohol 5.0 mmol) as a solution in THF (4 mL, 1 mL rinse) and the re-

(185 mg, 2.5 mmol) in THF (10 mL, 1 mL rinse) was added over a

period of 0.5 h. The reaction mixture was then allowed to warm to

ambient temperature and maintained under these conditions for

1.5 h. Solid (Ϫ)-T (1.16 g, 2.5 mmol) was then added and

stirring was continued for 1.5 h. Diethyl ether was added, the re-

sulting precipitate was removed by filtration through Celite, and

the solvent removed in vacuo. Purification of the residue by flash

chromatography (cyclohexane/EtOAc/triethylamine, 95:5:1) fol-

lowed by crystallization from pentane furnished the title compound

as a white solid (1.22 g, 2.15 mmol, 86%); m.p. 174Ϫ176 °C (pen-

tane). Ϫ [α]¹_D⁸ϭ Ϫ133.4 (c ϭ 1.06 in CHCl₃). Ϫ ¹H NMR: δ ϭ

sulting mixture was stirred for 1 h. Thereafter, ³¹P NMR indicated

the presence of fencholϪPCl₂(δ ϭ 174.60). The reaction mixture

was then cooled to Ϫ10 °C and triethylamine (2.08 mL, 15 mmol)

was slowly added. A slightly exothermic process was noted, accom-

panied by the formation of a white precipitate. The reaction mix-

ture was allowed to warm to room temperature, maintained under

these conditions for 0.25 h, and then cooled to 0 °C. Solid (Ϫ)-

T was added, the reaction mixture allowed to warm to room

temperature, and stirring was continued for 2 h prior to dilution

with diethyl ether. The solids were removed by filtration through a

pad of Celite, the solvent was removed in vacuo, and the residue

4018

Eur. J. Org. Chem. 2000, 4011Ϫ4027

Chiral Phosphorus Ligands

FULL PAPER

was taken up in diethyl ether. The resulting solution was filtered

as before; this procedure was repeated until the addition of ether

produced a clear solution. Purification by flash chromatography

(cyclohexane/EtOAc/triethylamine, 98:2:1) furnished the title com-

pound as a white foam (0.97 g, 1.5 mmol, 30%). Ϫ [α]¹_D⁸ϭ Ϫ163

0.50 (s, 3 H, CH₃-T), 0.85, 0.84 (2 s, 6 H, CH₃-bornyl), 1.02

(s, 3 H, CH₃-T), 2.20Ϫ1.20 (5 m, CH and CH₂bornyl), 4.75

(m, 1 H, CHO-bornyl), 5.08 (dd, J ϭ 8.3 Hz, J_HPϭ 1.8 Hz, 1 H,

CHO-T), 5.19 (d, J ϭ 8.3 Hz, 1 H, CHO-T), 7.6Ϫ7.2

(m, 20 H, ArH). Ϫ ³¹P NMR: δ ϭ 135.76. Ϫ C₄₁H₄₅O₅P (648.7).

1

(c ϭ 0.50 in CHCl₃). Ϫ H NMR: δ ϭ 0.37 (s, 3 H), 0.76 (s, 3 H),

(1S,7S)-9,9-Dimethyl-2,2,6,6-tetraphenyl-4-(1,7,7-trimethylbicyclo-

[2.2.1]hept-2-yloxy)-3,5,8,10-tetraoxa-4-phosphabicyclo[5.3.0]decane

(9); Method B: To a stirred solution of (Ϫ)-borneol (0.77 g,

5.0 mmol) in THF (5 mL) was slowly added PCl₃(436 µL,

5.0 mmol) as a solution in THF (4 mL, 1 mL rinse) and the re-

sulting mixture was stirred for 1 h. Thereafter, ³¹P NMR indicated

the presence of borneolϪPCl₂(δ ϭ 177.80). The reaction mixture

was then cooled to Ϫ10 °C and triethylamine (2.08 mL, 15 mmol)

was slowly added. A slightly exothermic process was noted, accom-

panied by the formation of a white precipitate. The reaction mix-

0.83 (s, 3 H), 1.15 (s, 6 H), 4.10 (d, J_HPϭ 14 Hz, 1 H, CHO-

fenchyl), 5.12 (m, 2 H, CHO-T), 7.8Ϫ7.2 (m, 20 H, ArH).

Ϫ

87.5, 87.6, 87.7, 126.0, 128.4, 129.0, 143.0. Ϫ ³¹P NMR: δ ϭ

138.41. Ϫ C₄₁H₄₅O₅P (648.7): calcd. C 75.90, H 6.99; found C

76.32, H 7.21.

¹³C NMR: δ ϭ 18.8, 22.6, 28.6, 32.4, 36.4, 38.3, 40.4, 71.8, 71.9,

(1S,7S)-9,9-Dimethyl-2,2,6,6-tetraphenyl-4-(1,3,3-trimethyl-

bicyclo[2.2.1]hept-2-yloxy)-3,5,8,10-tetraoxa-4-phospha-

bicyclo[5.3.0]decane (7); Method B: To a stirred solution of (ϩ)-

fenchol (0.77 g, 5.0 mmol) in THF (5 mL) was slowly added PCl₃ture was allowed to warm to room temperature, maintained under

(436 µL, 5.0 mmol) as a solution in THF (4 mL, 1 mL rinse) and

the resulting mixture was stirred for 1 h, whereupon ³¹P NMR

indicted the presence of fencholϪPCl₂(δ ϭ 174.6). The reaction

mixture was cooled to Ϫ10 °C and triethylamine (2.08 mL,

these conditions for 0.25 h, and then cooled to 0 °C. Solid (ϩ)-

T was added, the reaction mixture was allowed to warm to

room temperature, and stirring was continued for 2 h prior to dilu-

tion with diethyl ether. The solids were removed by filtration

15 mmol) was slowly added. A slightly exothermic process was through a pad of Celite, the solvent was removed in vacuo, and the

noted, accompanied by the formation of a white precipitate. The

reaction mixture was allowed to warm to room temperature, main-

tained under these conditions for 0.25 h, and then cooled to 0 °C.

Solid (ϩ)-T was added, the reaction mixture allowed to

warm to room temperature, and stirring was continued for 2 h prior

to dilution with diethyl ether. The solids were removed by filtration

through a pad of Celite, the solvent was removed in vacuo, and the

residue was taken-up in diethyl ether. The resulting solution was

filtered as before; this procedure was repeated until the addition of

ether produced a clear solution. Purification by flash chromato-

graphy (cyclohexane/EtOAc/triethylamine, 98:2:1) furnished the

residue was taken up in diethyl ether and filtered as before; this

procedure was repeated until the addition of ether produced a clear

solution. Purification by flash chromatography (cyclohexane/

EtOAc/triethylamine, 98:2:1) furnished the title compound as a

white foam (0.49 g, 0.75 mmol, 15%). Ϫ [α]²_D⁰ϭ Ϫ106 (c ϭ 0.50 in

CHCl₃). Ϫ ¹H NMR: δ ϭ 0.50 (s, 3 H), 0.85 (s, 3 H), 0.90 (s, 3 H),

0.92 (s, 3 H), 1.05 (s, 3 H), 2.20Ϫ1.01 (3 m, CH and CH₂bornyl),

4.76 (m, 1 H, CHO-bornyl), 5.06 (dd, J ϭ 8.3 Hz, J_HPϭ 1.8 Hz,

1 H, CHO-T), 5.14 (d, 1 H, J ϭ 8.3 Hz, CHO-T),

7.8Ϫ7.2 (m, ArH). Ϫ ¹³C NMR: δ ϭ 13.2, 19.4, 22.6, 26.8, 28.6,

29.1, 40.4, 44.2, 55.2, 69.4, 71.8, 87.5, 102.8, 126.0, 128.4, 129.0,

143.0. Ϫ ³¹P NMR: δ ϭ 135.34. Ϫ C₄₁H₄₅O₅P (648.7): calcd. C

75.90, H 6.99; found C 76.07, H 7.10.

title compound as a white foam (1.07 g, 1.65 mmol, 33%). Ϫ [α]²_D⁰

ϭ

1

ϩ146 (c ϭ 0.50 in CHCl₃). Ϫ H NMR: δ ϭ 0.31 (s, 3 H), 0.90 (s,

3 H), 1.02 (s, 3 H), 1.10 (s, 3 H), 1.23 (s, 3 H), 1.80Ϫ1.00 (m, 9 H,

CH and CH₂-fenchyl), 4.17 (d, J_HPϭ 13.5 Hz, 1 H, CHO-fenchyl),

5.00 (d, J ϭ 8.44 Hz, 1 H, CHO-T), 5.20 (dd, J ϭ 8.44 Hz,

J_HPϭ 2.8 Hz, 1 H, CHO-T), 7.8Ϫ7.2 (m, 20 H, ArH). Ϫ

¹³C NMR: δ ϭ 18.8, 22.6, 28.6, 32.4, 36.4, 38.3, 40.4, 71.8, 71.9,

87.5, 87.6, 87.7, 126.0, 128.4, 129.0, 143.0. Ϫ ³¹P NMR: δ ϭ

138.11. Ϫ C₄₁H₄₅O₅P (648.7).

(1R,7R)-4-[(1R,2S,5R)-2-Isopropyl-5-methylcyclohexyloxy]-9,9-

dimethyl-2,2,6,6-tetraphenyl-3,5,8,10-tetraoxa-4-phosphabicyclo-

[5.3.0]decane (10); Method B: To a stirred solution of (Ϫ)-menthol

(391 mg, 2.5 mmol) in THF (5 mL) was slowly added PCl₃(218

µL, 2.5 mmol) as a solution in THF (4 mL, 1 mL rinse) and the

resulting mixture was stirred for 1 h. Thereafter, ³¹P NMR indic-

ated the presence of mentholϪPCl₂(δ ϭ 177.00).^[18]The reaction

mixture was then cooled to Ϫ10 °C and triethylamine (1.07 mL,

7.75 mmol) was slowly added. A slightly exothermic process was

noted, accompanied by the formation of a white precipitate. The

(1R,7R)-9,9-Dimethyl-2,2,6,6-tetraphenyl-4-(1,7,7-trimethylbicyclo-

[2.2.1]hept-2-yloxy)-3,5,8,10-tetraoxa-4-phosphabicyclo[5.3.0]decane

(8); Method B: To a stirred solution of (Ϫ)-borneol (0.77 g,

5.0 mmol) in THF (5 mL) was slowly added PCl₃(436 µL, reaction mixture was allowed to warm to room temperature, main-

5.0 mmol) as a solution in THF (4 mL, 1 mL rinse) and the re-

sulting mixture was stirred for 1 h. Thereafter, ³¹P NMR indicated

the presence of borneolϪPCl₂(δ ϭ 177.80). The reaction mixture

was then cooled to Ϫ10 °C and triethylamine (2.08 mL, 15 mmol)

was slowly added. A slightly exothermic process was noted, accom-

panied by the formation of a white precipitate. The reaction mix-

ture was allowed to warm to room temperature, maintained under

tained under these conditions for 0.5 h, and then cooled to 0 °C.

Solid (Ϫ)-T was added, the resulting mixture was allowed to

warm to room temperature, and stirring was continued for 1 h prior

to dilution with diethyl ether. The solids were removed by filtration

through a pad of Celite, the solvent was removed in vacuo, and the

residue was taken up in diethyl ether and filtered as before; this

procedure was repeated until the addition of ether produced a clear

these conditions for 0.25 h, and then cooled to 0 °C. Solid (Ϫ)- solution. Addition of pentane induced crystallization to provide the

T was added, the resulting mixture was allowed to warm to

room temperature, and stirring was continued for 2 h prior to dilu-

tion with diethyl ether. The solids were removed by filtration

through a pad of Celite, the solvent was removed in vacuo, the

residue was taken up in diethyl ether and filtered as before; this

procedure was repeated until the addition of ether produced a clear

solution. Purification by flash chromatography (cyclohexane/

EtOAc/triethylamine, 98:2:1) furnished the title compound as a

white foam (1.07 g, 1.65 mmol, 33%). Ϫ ¹H NMR (400 MHz): δ ϭ

title compound as a white solid (800 mg, 49%); m.p. 155Ϫ156 °C

(pentane). Ϫ [α]¹_D⁸ϭ Ϫ176.1 (c ϭ 1.14 in CH₂Cl₂). Ϫ ¹H NMR:

δ ϭ 0.54 (s, 3 H, CH₃-T), 0.76 (d, J ϭ 7.2 Hz, 3 H, CH₃-

menthyl), 0.77 (d, J ϭ 6.6 Hz, 3 H, CH₃-menthyl), 0.88 (d, J ϭ

7.0 Hz, 3 H, CH₃-menthyl), 0.97 (s, 3 H, CH₃-T), 1.26Ϫ1.20

(m, 1 H, menthyl), 1.12Ϫ0.46 (m, 4 H, menthyl), 1.45Ϫ1.38 (m, 2

H, menthyl), 2.06Ϫ2.03 (m, 1 H, menthyl), 2.43Ϫ2.37 (m, 1 H,

menthyl), 4.03 (dq, J ϭ 10.0 Hz and 4.4 Hz, 1 H, CHO-menthyl),

5.58 (dd, J ϭ 8.3 Hz and 1.6 Hz, 1 H, 7-H), 5.67 (d, J ϭ 8.3 Hz, 1

Eur. J. Org. Chem. 2000, 4011Ϫ4027

4019

A. Alexakis et al.

H, 1-H), 7.16Ϫ6.97 (m, 12 H, ArH), 7.89Ϫ7.72 (m, 8 H, ArH). Ϫ compound as a white foam (1.17 g, 1.75 mmol, 35%). Ϫ [α]²_D⁰

FULL PAPER

ϭ

1

¹³C NMR: δ ϭ 15.9, 21.6, 22.2, 23.1, 81.4, 25.3, 25.8, 27.3, 31.6,

Ϫ110 (c ϭ 0.55 in CHCl₃). Ϫ H NMR: δ ϭ 0.33 (s, 3 H, CH₃-

34.2, 44.0, 48.6, 48.7, 76.0, 76.3, 82.1, 82.5, 84.2, 112.5 (C-9), 127.1, T), 1.10 (s, 3 H, CH₃-T), 2.60Ϫ0.81 (6 m, CH and

127.4, 127.5, 127.8, 129.2, 129.0, 141.4, 145.9, 146.5. Ϫ ³¹P NMR: CH₂phenylcyclohexyl), 4.47 (m, 1 H, CHO-phenylcyclohexyl),

δ ϭ 143.11. Ϫ MS (CI, NH₃): m/z (%) ϭ 651 [M ϩ H]^ϩ(5), 431 4.95 (d, J ϭ 8.1, 1 H, CHO-T), 5.01 (dd, J ϭ 8.1 Hz, J_HP

(100). Ϫ C₄₁H₄₇O₅P (650.8): calcd. C 75.67, H 7.28; found C 75.58,

ϭ

2 Hz, 1 H, CHO-T), 7.81Ϫ7.23 (m, 25 H, ArH). Ϫ ¹³C

H 7.36.

NMR: δ ϭ 25.0, 25.4, 25.6, 27.2, 33.3, 35.1, 51.3, 80.8, 81.4, 82.0,

82.1, 109.4, 112.1, 126Ϫ128 (Ph), 142.6, 145.8. Ϫ ³¹P NMR: δ ϭ

139.90. Ϫ C₄₃H₄₃O₅P (670.8).

(1R,7R)-4-[(1R,2S,5R)-2-Isopropyl-5-methylcyclohexyloxy]-9,9-

dimethyl-2,2,6,6-tetraphenyl-3,5,8,10-tetraoxa-4-phosphabicyclo-

[5.3.0]decane (11); Method B: To a stirred solution of (Ϫ)-menthol

(1R,7R)-9,9-Dimethyl-2,2,6,6-tetraphenyl-4-[(1S,2R)-phenylcyclo-

(391 mg, 2.5 mmol) in THF (5 mL) was slowly added PCl₃(218 hexanyloxy]-3,5,8,10-tetraoxa-4-phosphabicyclo[5.3.0]decane (13);