The ligand-exchange process of P-Hapical phosphoranes and the thermal formation and pseudorotation of anti-apicophilic spirophosphoranes

FULL PAPER

DOI: 10.1002/ejoc.200500510

The Ligand-Exchange Process of P–H_apicalPhosphoranes and the Thermal

Formation and Pseudorotation of Anti-Apicophilic Spirophosphoranes

Satoshi Kojima,^[a]Kazumasa Kajiyama,^[a]Masaaki Nakamoto,^[a]Shiro Matsukawa,^[a]and

Kin-ya Akiba*^[a][‡]

Keywords: Anti-apicophilic compounds / Hypervalent compounds / Kinetics / Phosphoranes / Pseudorotation

Monocyclic P–H_apicalphosphoranes 10 bearing a ring-opened

Martin ligand are formed upon treatment of the P–H_equatorial

spirophosphorane 7 with more than 2 equiv. of alkyllithium

position. These phosphoranes 6 were found to completely

convert into their energetically more stable C_equatorial,O_apical

stereoisomers 5 upon heating. On the basis of kinetic exami-

nations of the cyclization process of 10, the formation of 6

can be rationalized by the enhanced nucleophilicity of the

hydroxy group in the donor solvents. The isomer 5b was esti-

mated to be more stable than 6b (R = nBu) by 12 kcalmol^–1

from kinetic studies on the pseudorotation of 6b to 5b. This

is in good agreement with theoretical calculations with anal-

ogous 5a and 6a (R = Me), which estimate the difference in

compounds.

A ligand-exchange process of these phos-

phoranes 10 involving the interconversion of the bidentate

Martin ligand with the monodentate Martin ligand was

found to proceed, probably through a hexacoordinate phos-

phorus atom. Heating of 10 in nondonating solvents fur-

nished alkylspirophosphoranes 5, whereas heating of 10 in

donor solvents led to the formation of anti-apicophilic phos-

phoranes 6, a new class of phosphoranes in which the less

apicophilic carbon substituent occupies an apical position

and the more apicophilic oxygen atom occupies an equatorial

energy to be 14.1 kcalmol^–1

.

Germany, 2006)

sensus is that an oxygen atom is much more apicophilic

than a carbon atom. Thus, a phosphorane having a mixture

of oxygen and carbon substituents, and capable of under-

going pseudorotation, would naturally be expected to have

a maximum number of oxygen atoms in the apical positions

in the most stable stereoisomer(s) (I; Scheme 1). As for con-

figurational isomers with an apical oxygen atom and an

equatorial carbon atom reversed (II, IIЈ), they have been

assumed to exist as high-energy intermediates on the pseu-

dorotation coordinate for interconversion of P-epimers of

I. However, such species have never been observed simulta-

neously with I. This is understandable considering the fact

that a rather large difference in energy is expected for them

[empirical estimate: ca. 8.3 kcalmol^–1;^[10]theoretical values

for cyclic phosphorane PH₃(CH₂CH₂O): 7.77^[11]and

2.4 kcalmol^–1;^[9c]for acyclic phosphorane PH₃CH₃OH:^[12]

1.6 kcalmol^–1]. Furthermore, there is a question as to

whether such species actually exist as stationary points on

the pseudorotation reaction coordinate at all.^[12]

One method of obtaining a phosphorane with an apical

carbon atom in the presence of more electronegative hetero-

atoms would be by raising the intrinsic apicophilicity of the

carbon substituent by introducing electron-withdrawing

groups onto it. For example, for carbon and fluorine sub-

stituents the use of fluorinated carbon atoms has led to the

observation of 1 in equilibrium with its equatorial CF₃iso-

mer by IR spectroscopy.^[13]Another method would be to

utilize ring strain (four- or five-membered bidentate li-

Introduction

Stable pentacoordinate 10-P-5^[1]phosphoranes have

served as model compounds for intermediates in reactions

involving the phosphoryl unit in biological systems.^[2]Phos-

phoranes usually assume trigonal-bipyramidal structures in

which there are two distinctive groups of sites − the apical

and the equatorial positions. The positional exchange of

these stereochemically fluxional compounds is usually de-

scribed in terms of Berry pseudorotation (BPR), which has

its basis in bond bending and generally follows a low-energy

process.^[3,4]Because of the electron-rich nature of the hyper-

valent three-center,four-electron apical bond,^[5]elements

(substituents) more capable of stabilizing negative charge

are preferred in these apical positions. Thus, the ability to

stabilize this bond, known as apicophilicity, usually paral-

lels the electronegativity of the elements (or substituents).^[6]

Several experimental^[4,7,8]and theoretical^[9]apicophilicity

scales, mostly based on the relative barriers of pseudorota-

tion, have been proposed in the past. Although apicophilici-

ties are different according to their systems, the general con-

[a] Department of Chemistry, Graduate School of Science, Hiro-

shima University,

1-3-1 Kagamiyama, Higashi-Hiroshima 739-8526, Japan

[‡] Present address: Advanced Research Center for Science and

Engineering, Waseda University,

3-4-1 Ohkubo, Shinjuku-ku, Tokyo 169-8555, Japan

E-mail: akibaky@waseda.jp

Supporting information for this article is available on the

WWW under http://ww w .eurjoc.org or from the author.

218

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

Figure 1.

Scheme 1.

phoranes 4b and 4c, which contain an oxaphosphetane ring,

have also been prepared and characterized by us. Pseudoro-

tation of the O-cis isomer 4b to its O-trans isomer 4c proved

that, unlike 4a and all other reported unusual C_apicalrota-

mers, the intrinsic element apicophilicities had not been al-

tered.^[19b]Furthermore, establishment of methods to obtain

anti-apicophilic phosphoranes has allowed us to examine

the difference in reactivity between isomeric phosphoranes

of types 5 and 6, and thus we have found that the latter

anti-apicophilic phosphoranes are much more reactive

towards nucleophiles than their O-trans isomers.^[19c]In this

paper, we present our kinetic examination of the thermal

cyclization process of P–H_apicalphosphoranes 10 to give O-

cis 6, in conjunction with the pseudorotation process of 6,

and, based upon these results, we present a mechanistic ra-

tionale for the unique formation of 6. We also describe the

stereomutation properties of 10, along with full details of

the preliminarily reported results.

gands), as in compound 2, such that the bidentate ligands

can only span an apical and an equatorial position (Fig-

ure 1).^[14]There are a great many examples of this group of

compounds.^[4]Still another method would be to employ ste-

ric effects and ring strain together. Holmes and co-workers,

for example, have prepared phosphoranes 3 with an apical

carbon atom and equatorial oxygen atoms.^[15]Here, the

eight-membered ring is forced by the bulky tBu groups and

by the bond-angle effect to occupy two equatorial sites,

while the five-membered ring is forced to occupy an apical

and an equatorial site, thus making the monodentate alkyl

group sit at the remaining apical position. As for carbon

and nitrogen substituents, Kawashima and co-workers have

found that 4a is in equilibrium with its C_equatorialstereoiso-

mer by introducing a phenyl group at the nitrogen atom.^[16]

While all of these methods do indeed give unusual C_apical

rotamers, it can be said that the intrinsic apicophilicity of

the elements has been altered by special factors.

By utilizing the Martin ligand,^[17]we have recently

achieved the isolation and characterization of the first anti-

apicophilic phosphorane 6b, which we define as a 10-P-5

phosphorane that is P-epimeric with, although less stable

than, phosphorane 5b, which has a configuration that fulfils

the apicophilicity rules in terms of ordinary substituent

Results and Discussion

Formation of Monocyclic P–H_apicalPhosphoranes 10

During our investigation of the electrophilic substitution

electronegativity (Figure 1). In the case of 6b, the less apico- reactions of phosphoranide 8, generated by deprotonation

philic carbon substituent occupies an apical position and of P–H_equatorialspirophosphorane 7,^[8,20]we noticed that

the more apicophilic oxygen atom occupies an equatorial when alkyllithium reagents were used as base, varying

position (O-cis), while 5b, which has an ordinary configura- amounts of side products in which the alkyl group from the

tion with two apical oxygen atoms (O-trans), exists as the lithium reagent was incorporated, such as 5, were obtained

more stable isomer, as described in a preliminary report.^[18](Scheme 2). In order to form the product 10-P-5 phos-

Evolving from this discovery, we later found that anti-apico- phoranes from the dianionic species 9, a subsequent oxidat-

philic phosphoranes 6 can even be prepared at ambient ive process is required. We speculated that aqueous treat-

temperature or lower by using I₂as the oxidant.^[19a]Phos- ment of 9 had led to diprotonated species such as 10 and

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

219

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

subsequent cyclization with concomitant H₂evolution gave X-ray Structure of the P–H_apicalPhosphoranes 10

the ultimate alkylspirophosphoranes 5. In order to prove

X-ray structural analyses of compounds 10a (R = Me),

this assumption we set out to isolate the supposed interme-

10b (R = nBu), 10c_intra(R = tBu), and 10c_inter(R = tBu)

diate 10. When compound 7 was treated with more than

revealed that the former three have similar structures. In the

2 equiv. of RLi (R = Me, nBu, tBu) in diethyl ether and

case of compound 10c, the different forms could be recrys-

carefully treated with acid, monocyclic P–H_apicalphos-

tallized independently by changing the solvent − hexane/

phoranes 10a [R = Me; ³¹P NMR (CDCl₃): δ = –51.9 ppm

CCl₄for 10c_intraand hexane/MeCN for 10c_inter. The OR-

(¹J_P,H= 276 Hz)], 10b [R = nBu; ³¹P NMR (CDCl₃): δ = –

TEP plots of 10c_intraand 10c_interare shown in Figure 2, with

34.4 ppm (¹J_P,H= 273 Hz)], 10c_intra[R = tBu; ³¹P NMR

selected structural parameters for all four compounds given

(CDCl₃): δ = –14.7 ppm (¹J_P,H= 273 Hz)], and 10c_inter[R

in Table 1. All of them assume a trigonal-bipyramidal struc-

= tBu; ³¹P NMR ([D₆]acetone): δ = –40.1 ppm (¹J_P,H

293 Hz)] were indeed found to form, along with varying

=

ture, as evident from the sum of the three C–P–C angles

in the projected equatorial plane (359.4–360.0°), with the

hydrogen atom positioned apical, as expected. The common

feature among 10a, 10b, and 10c_intrais the presence of intra-

molecular hydrogen bonding between the apical oxygen

atom and the O–H hydrogen of the monodentate Martin

ligand. Possibly in order to accommodate this hydrogen

bonding, the benzene ring of the monodentate Martin li-

gand is distorted, with a P1–C10–C15 angle of about 133°

instead of 120°. On the other hand, 10c_interengages in inter-

molecular hydrogen bonding with an MeCN molecule.

Thus, 10c_intrais unsolvated and 10c_interis solvated. This is

a rare example of a fully characterized pair of unsolvated

and solvated forms due to different modes of hydrogen

bonding and, incidentally, the first involving 10-P-5 phos-

phoranes that we are aware of. As a consequence of the

trans influence, the apical P1–O1 bond (1.91–1.95 Å) is

longer for all four (vide infra) than for other phosphoranes

bearing the Martin ligand, such as spirophosphorane 5, in

which it is about 1.77 Å, when the opposite element is oxy-

gen.^[19c]Furthermore, it seems that intramolecular hydro-

gen bonding in 10a, 10b, and 10c_intraweakens and elongates

the apical P–O1 bond even further, as evident from the val-

amounts of the corresponding 5. Judging from the similar-

1

ity of the H NMR patterns for the two species of 10c, we

assumed them to be isomers. Their actual identities − dif-

fering modes of solvation − were later revealed by X-ray

1

analysis (vide infra). The characteristically small J_P,Hval-

ues (¹J_P,H= 273 Hz) compared with that of 7 (¹J_P,H

=

729 Hz), in which the hydrogen atom is equatorial, indi-

cated that the hydrogen atom is bonded to the phosphorus

atom in the apical position, as already observed for

11.^[17c,21]Isomeric ring-opened phosphanes 12 that could be

imagined to be in equilibrium with 10 were not observed,

even though the presence of an alkyl group was expected to

destabilize the monocyclic 10-P-5 phosphoranes 10.

ues of 1.914, 1.954, and 1.919 Å for 10a, 10b, and 10c_intra

,

respectively, compared with 1.837 Å for 10c_inter. The O2–P1

distance in 10c_inter(2.84 Å) is shorter than that in 10c_intra

(3.10 Å) and the sum of the van der Waals radii of the two

atoms (3.3 Å).^[24]The oxygen atom is also oriented so that

a lone pair can be projected to face and weakly interact

with the phosphorus atom. This coincides with the ³¹P

Scheme 2.

We have previously observed similar amphiphilic proper-

ties in a spirostiboranide that differs from 8 only in the cen-

tral atom.^[22,23]However, no intermediates analogous to 10

or 7 could be observed in the antimony series.

Figure 2. ORTEP drawings of 10c_intraand 10c_intershowing the ther-

mal ellipsoids at the 30% probability level. All the hydrogen atoms

other than P1–H1 and O2–H2 have been omitted for clarity. The

broken lines denote hydrogen bonding.

220

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

chemical shift of 10c_inter[δ_P(CDCl₃) = –43.0 ppm] being intermolecular hydrogen-bonded species 10c_interbeing the

at higher field (more electron-rich) than that of 10c_intra[δ_Ponly observable species in donor solvents. When 10c_inter

(CDCl₃) = –14.7 ppm].

containing a solvent molecule of EtOH was dissolved in

nondonor solvents, a mixture of 10c_intraand 10c_interwas

observed. The presence of 10c_interin nondonor solvents is a

consequence of either interaction with EtOH or adven-

titious H₂O. These results indicate that for 10c, compared

with 10a and 10b, intramolecular hydrogen bonding is less

favorable: the appearance of the intermolecular hydrogen-

bonded species is instantaneous for 10a and 10b upon dis-

solution in donor solvents, whereas equilibration between

the two forms of 10c can be observed by NMR spec-

troscopy.

Table 1. Selected bond lengths [Å] and angles [°] for 10a, 10b,

10c_intra, and 10c_inter

.

10a

1.190(1)

1.914(3)

1.812(5)

1.838(4)

1.808(5)

0.947(3)

1.661(3)

10b

10c_intra

10c_inter

P1–H1

P1–O1

P1–C1

P1–C10

P1–C19

O2–H2

O1–H2

1.450(1)

1.954(2)

1.811(3)

1.847(3)

1.814(3)

0.987(2)

1.653(2)

1.371(2)

1.919(8)

1.82(1)

1.886(10)

1.86(1)

0.586(8)

2.003(6)

1.371(1)

1.837(2)

1.815(4)

1.847(4)

1.869(4)

1.025(2)

N1–H2

P1–O2

1.783(5)

2.827(3)

174.0(1)

84.8(1)

88.4(1)

94.5(1)

131.0(2)

118.0(2)

110.9(2)

127.6(3)

3.098(4)

174.5(1)

83.6(2)

88.0(2)

91.2(2)

116.7(2)

118.0(2)

124.8(2)

133.5(3)

3.071(3)

176.0(1)

83.5(1)

87.6(1)

91.3(1)

116.4(1)

117.0(2)

126.0(2)

132.7(2)

3.142(8)

173.4(2)

83.1(4)

91.6(4)

94.4(5)

111.8(4)

118.4(5)

129.7(5)

134.6(7)

O1–P1–H1

O1–P1–C1

O1–P1–C10

O1–P1–C19

C1–P1–C10

C1–P1–C19

C10–P1–C19

P1–C10–C15

Ligand Exchange of the P–H_apicalPhosphoranes 10

The dynamic behavior of 10b (R = nBu), whose ¹⁹F sig-

nals could be assigned as shown in Figure 3 by signal

broadening due to through-space coupling^[25]between a-

endo and b-endo, and by ¹⁹F NMR irradiation experiments,

was examined by variable-temperature ¹⁹F NMR measure-

ments. No evaluation of 10a (R = Me) could be carried out

because of its labile nature.

Solvent Dependence of the Composition of 10_intraand 10_inter

Since the X-ray structure of 10c revealed the presence of

unsolvated and solvated forms, the solvent influence upon

the composition of 10 was examined (Table 2). In nondonor

solvents, both 10a and 10b show the presence of only the

intramolecular hydrogen-bonded species 10_intra. Dissolution

of 10a in donor solvents, however, resulted in the facile for-

mation of spirophosphorane 5a even at room temperature,

and a small upfield signal due to 10a_intercould be observed

along with that of 10a_intra. Phosphorane 10b also shows the

presence of both forms and is less susceptible towards cycli-

zation. As expected, 10c shows different behavior, with the

Figure 3. ¹⁹F NMR assignment of 10b.

Ordinary pseudorotation involving interconversion be-

tween P-enantiomers should result in coalescence between

the signals corresponding to a-endo and a-exo, and between

those of b-exo and b-endo. However, what was actually ob-

served was the coalescence between signals corresponding

to a-exo and b-exo. Thus, other than the pseudorotation

between P-enantiomers, the presence of a low-energy pro-

cess was revealed, which interconverts the bidentate and the

monodentate Martin ligand by a hydrogen shift. The bar-

Table 2. ³¹P NMR chemical shifts [ppm] of phosphoranes 10a–c in

various solvents.^[a]

Solvent

10a

10b

10c (with EtOH)

intra

–

inter

intra inter intra

inter

Hexane

–35.5

–15.5

–14.5

–14.7

–43.9

53.0^[b]

–

Benzene

–34.2

–42.8

–43.0

–42.2

–40.1

–41.2

52.6^[b]

–

Chloroform

Diethyl ether

Acetone

–34.4

51.9^[b]

–

rier of this dynamic process was estimated to be ΔG^϶

=

17.7 kcalmol^–1(Δν = 106 Hz) at 368 K in [D₈]toluene and

ΔG^϶= 17.4 kcalmol^–1(Δν = 105 Hz) at 363 K in 1,2-

dichlorobenzene, based upon the Gutowsky–Holm approxi-

mation.^[26]A line-shape analysis of spectra taken in [D₈]-

toluene gives an estimate of ΔH^϶= 11.4 1.7 kcalmol^–1

and ΔS^϶= –16.3 4.7 eu.

–35.0 –52.3

–51.4

53.4^[b]

–

68.8^[b]

–

Tetra-

–53.0

hydrofuran

Acetonitrile

Methanol

Pyridine

70.6^[b]

[c]

–51.4

–52.3

–53.2

–39.7

–40.6

–39.3

[c]

Two mechanisms are plausible to account for the ob-

served ligand-exchange reaction, one which involves a ring-

opened 8-P-3 phosphane (12: path A) and one which in-

volves a 12-P-6 transition state (TS-i: path B), where con-

comitant P–O bond formation and bond breaking is ac-

[a] “intra” refers to the intramolecular hydrogen-bonding species,

and “inter” to the intermolecular hydrogen-bonding species.

[b] The signal of the cyclized compound 5a or 5b was also ob-

served. [c] Only the signal of 5a was observed.

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

221

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

Scheme 3.

companied by a hydrogen shift from oxygen atom to oxygen it may well be that the global energy maximum (transition

atom, as shown in Scheme 3. The former path involves two state) of the whole process (path B) is one of the pseudoro-

pseudorotations and a ring opening to form the symmetric tation transition states [Ψ(H) or Ψ(Ar)] and not TS-i.

intermediate 12 and the same three processes to reach the

enantiomer. The other mechanism involves direct transfor-

mation from 10A to 10C, as depicted in path B, followed

by two low-energy pseudorotation processes to give the en-

antiomeric 10A.

Since phosphoranes bearing a P–H bond are usually

found in equilibrium with prototropic ring-opened isomeric

phosphanes such as 12, path A was initially thought more

feasible. However, when 10b was dissolved in MeOD and

the NMR spectrum of the solution measured immediately,

the O–H hydrogen atom was found to be completely ex-

Formation of O-cis Phosphoranes

changed with deuterium, whereas the P–H hydrogen atom

While we were carrying out recrystallization of P–H

was only partially exchanged. Since the ligand-exchange re- phosphorane 10b in MeCN at ambient temperatures, we

action is an extremely rapid process, if path A were opera- obtained a crystalline compound 6b that was neither 10b

tive the P–H hydrogen atom should also have been com- nor 5b.^[18a]This same compound was also found to form

pletely replaced by a deuterium atom due to the symmetric along with 5b when 10b was heated in pyridine at ca. 60 °C

nature of intermediate 12. However, since this is not the and it gradually underwent conversion into 5b. X-ray struc-

case, we believe that path B, which does not involve dissoci- tural analysis (vide infra) of this compound revealed that it

ation of the P–H bond, is more likely. Furthermore, dissol- is an unexpected stereoisomer of 5b bearing a carbon atom

ution of 10a–c in ethereal solutions containing KOH led and an oxygen atom in a trans relationship in the apical

to complete conversion to species which we believe to be positions, i.e., an O-cis relationship. Optimization of the re-

hexacoordinate [from 10a: δ_P(Et₂O) = –153 ppm (¹J_P,H

=

action conditions to furnish O-cis 6b led to a yield of 71%

679 Hz); from 10b: δ_P(Et₂O) = –130 ppm (¹J_P,H= 674 Hz); along with 29% of O-trans 5b, as shown in Scheme 4, upon

from 10c: δ_P(Et₂O) = –117 ppm (¹J_P,H= 335 Hz)] judging heating of 10b in a THF/pyridine (5:1) mixture at 60 °C for

1

from the high-field ³¹P chemical shifts. The J_P,Hvalues for 20 min. Phosphorane O-cis 6a could also be obtained in

10a,b are in good agreement with the reported values for 52% yield, along with 46% of O-trans 5a, upon heating

1

hexacoordinate 13 (R = Me, J_P,H= 620 Hz; R = Ph, J_P,Hof a THF solution of 10a at 40 °C for 1 h. However, the

= 642 Hz). This implies that attaining the hexacoordinate pseudorotation of O-cis 6a to O-trans 5a was found to take

state is facile and therefore our assumption of a transition place gradually even at room temperature. In the case of

state such as TS-i is quite reasonable. Since the overall bar- 10c, O-cis 6c could be obtained in 75% yield along with

rier for the exchange process is essentially the same in tolu- 18% recovery of 10c by refluxing in THF/pyridine (5:1) for

N

ene (E_T= 0.099) and 1,2-dichlorobenzene (E_T= 0.225), 36 h.

222

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

tions,^[12]thus reflecting the larger polarity of the P–O bond

compared with the P–C bond. The trends are practically

the same in O-cis 6c.

Table 3. Selected bond lengths [Å] and angles [°] for 5b, 6b, and 6c.

5b^[a]

6b^[a]

6c^[b]

P1–O1

P1–O2

P1–C1

P1–C10

1.765(2)

1.750(2)

1.822(2)

1.817(2)

1.818(3)

175.75(8)

87.30(9)

90.73(9)

91.3(1)

90.84(9)

87.32(9)

92.9(1)

126.8(1)

116.5(1)

116.6(1)

1.773(3)

1.658(3)

1.806(4)

1.860(4)

1.830(5)

82.9(1)

87.0(2)

170.5(2)

89.0(2)

119.9(2)

87.8(1)

124.2(2)

99.1(2)

114.6(2)

94.9(2)

1.771(3)

1.652(2)

1.830(4)

1.893(4)

1.875(4)

82.9(1)

85.6(1)

169.5(1)

90.4(2)

120.7(2)

87.5(1)

116.3(2)

95.7(2)

121.8(2)

97.7(2)

Scheme 4.

P1–C19

X-ray Structure of the O-cis Phosphoranes

O1–P1–O2

O1–P1–C1

O1–P1–C10

Figure 4 shows an ORTEP drawing of 6b along with 5b

for comparison. Compound 6c was also confirmed to have O1–P1–C19

O2–P1–C1

an O-cis structure. Selected structural parameters are listed

O2–P1–C10

in Table 3. Although the O1–P1–C10 angles for O-cis 6b

O2–P1–C19

and 6c are 170.5 and 169.5°, respectively, and deviate from

C1–P1–C10

C1–P1–C19

C10–P1–C19

the ideal value of 180°, the sums of the angles O2–P1–C1,

C1–P1–C19, and C19–P1–O2 for O-cis 6b and 6c are 358.7,

and 358.8°, respectively, and near the ideal value of 360°,

thus indicating that the two compounds indeed take on trig-

onal-bipyramidal (TBP) structures, although they are some-

what distorted. The two figures clearly show that the two

compounds are in a stereoisomeric relationship. The apical

bond in O-trans 5b was found to be less distorted, with

an angle of 175.75°. A similar trend has been predicted in

theoretical calculations on PH₃F₂, with the F_apical,F_apical

[a] Ref.^[19c][b] The crystal was found to contain two independent

molecules. Values of only one are given.

Solution Structure of O-cis Spirophosphoranes: Rapid

Enantiomer Interconversion

As a consequence of a rapid, single-step pseudorotation

process between enantiomers 6-R and 6-S, with the R group

as the pivot, averaged spectra are observed for 6a–c at am-

bient temperature. A comparison of the NMR spectra of

O-trans 5 and O-cis 6 showed that for all three pairs of

isomeric phosphoranes the ³¹P NMR shift is located at

lower field for 6 by around 15 to 17 ppm [δ_P(CDCl₃) =

–22.6 ppm for 5a and –6.3 ppm for 6a; –18.8 ppm for 5b

and –3.5 ppm for 6b; –9.8 ppm for 5c and 7.7 ppm for 6c].

The ¹J_C(aryl),Pcoupling constants for 5a, 5b, and 5c are 177,

160, and 160 Hz, respectively, whereas those for 6b and 6c

are 88 and 77 Hz; that of 6a is obscured. Since the

¹J_C(alkyl),Pcoupling constants are essentially the same for

each pair (5a: 125 Hz, 6a: 130 Hz; 5b: 116 Hz, 6b: 114 Hz;

5c: 103 Hz, 6c: 114 Hz), it follows that the solution struc-

tures of these compounds are all essentially the same, i.e.,

trigonal bipyramids. Square-pyramid (SP) type structures

would be expected to give coupling constants of similar

magnitude for both aryl groups because both would occupy

basal sites, which is not the case here. Unfortunately, low-

temperature (down to 203 K) ¹³C NMR measurements only

gave broadened spectra for the two inequivalent Martin li-

isomer having an F–P–F angle of 180° and the F_apical

,

F_equatorialisomer (calculated with C_ssymmetry) an apical

bond angle of 171.6°.^[9a]This propensity has also been pre-

dicted with the

F

_apical,F_equatorialisomer of [SiH₃F₂]^–

(175.5°).^[27]On the other hand, the corresponding bonds in

O-cis 6b were found to be different, as expected, with the

apical bonds being typically longer than the corresponding

equatorial bonds (P–O_apical1.773 Ͼ P–O_equatorial1.658 Å

and P–C_apical1.860 Ͼ P–C_equatorial1.806 Å). The percent-

ages of contraction of an equatorial bond compared to that

of an apical bond for 6b are 6% for P–O and 3% for P–

C, and are in good agreement with averaged (among three

conformers) differences between P–O (4%) and P–H (2%)

bonds in PH₃(OH)_apical(OH)_equatorialby theoretical calcula-

1

gand carbon atoms of O-cis 6b and 6c. Since the J_C(alkyl),P

values for O-trans 5b and O-cis 6b coincide the best

1

(115 Hz), from analogy of J_C(aryl),P= 160 Hz for O-trans

1

5b, we can assume that the equatorial J_C(aryl),Pcoupling

constant for O-cis 6b is also 160 Hz, therefore the axial

¹J_C(aryl),Pcoupling constant is calculated to be 16 Hz from

the observed average value of 88 Hz (88 × 2 – 160). This

value relationship (large coupling constant for

Figure 4. ORTEP drawings of 5b and 6b showing the thermal ellip-

soids at the 30% probability level. All the hydrogen atoms have

been omitted for clarity.

P–atom_equatorialand small coupling for P–atom_apical) paral-

1

lels that for J_P,Hin the P–H phosphoranes 10 (H_apical

:

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

223

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

273–293 Hz) and 7 (H_equatorial: 729 Hz) in Scheme 2 (vide tions that species of extra coordination are involved during

supra).

the reaction.^[28]We also have previously observed selective

In the ¹⁹F NMR spectra the two signals observed at hydrocarbon elimination in the intramolecular cyclization

room temperature for O-cis 6b [¹⁹F NMR (CDCl₃, 293 K): reaction of acyclic 8-Si-4 o-silylbenzyl alcohols to 8-Si-4 cy-

δ = –74.9 (q, J_F,F= 8.5 Hz, 6 F), –76.2 ppm (q, J_F,F

4

=

clic alkoxysilanes via 10-Si-5 transition states,^[29a]in the hy-

8.5 Hz, 6 F)] decoalesce into four signals [¹⁹F NMR ([D₈]- drolysis of monocyclic 12-Sb-6 ate complexes to monocyclic

toluene, 193 K): δ = –74.3 (br. s, 3 F), –74.5 (br. s, 3 F), 10-Sb-5 stiboranes,^[29b]and in the transformation of mono-

–75.5 (br. s, 3 F), –76.0 ppm (br. s, 3 F)] at low tempera- cyclic 10-Bi-5 alcohols to 10-Bi-5 spirobismuthanes via 12-

tures. The fact that only one ³¹P signal [³¹P NMR ([D₈]- Bi-6 transition states.^[29c]As for phosphorus, this phenome-

toluene, 219 K): δ = –3.6 ppm] is observed for the com- non is not normally encountered with neutral compounds

pound throughout the temperature range confirms that the of ordinary valency. However, the facile hydrolysis of tetra-

¹⁹F NMR signals are for a single species. Coalescence line- coordinate phosphonium cations is known to proceed via

shape analysis of the better resolved upper-field pair of sig- pentacoordinate intermediates,^[30]and the exchange reac-

nals gave estimated activation parameters for the process, tion of alkoxy and amino groups of phosphites with other

i.e., interconversion by a single BPR, to be ΔH^϶

=

alcohols and amines has been determined to involve penta-

10.0 0.7 kcalmol^–1and ΔS^϶= –3.4 3.0 eu.^[26]The lower- coordinate P–H intermediates.^[4,31]As for transformations

field pair was broadened further at low temperature, pre- involving hypervalent reactants and products, little has been

sumably by the suppression of CF₃-group rotation. Deco- reported as the lack of available thermodynamically stable

alescence of signals for O-cis 6c could not be observed even compounds and their susceptibility to moisture has compli-

at 193 K ([D₈]toluene). This implies that the inversion bar- cated kinetic examinations.^[32]

rier for O-cis 6c is smaller than for O-cis 6b due to the

The rates of cyclization of 10b to yield 5b were measured

N

presence of the bulky tBu group, which could distort the in [D₈]toluene (E_T= 0.099,^[33]DN^N= 0.01^[34]) and o-

N

ground-state structure of O-cis 6c somewhat towards the dichlorobenzene (E_T= 0.225,^[33]DN^N= 0.08^[34]), which

square-pyramidal transition state of pseudorotation can be considered to be nondonor solvents, in the tempera-

(Scheme 5).

ture range 343–373 K by monitoring the change in the ¹⁹F

NMR integrals. All the measurements were found to obey

first-order kinetics (Table 4). The activation parameters ob-

tained are as follows: in toluene, ΔH^϶

=

17.3 0.6 kcalmol^–1and ΔS^϶= –34.7 1.7 eu; in o-dichlo-

robenzene, ΔH^϶14.3 0.2 kcalmol^–1and ΔS^϶

=

–41.0 0.6 eu. Cyclization was found to be slightly faster in

the more polar o-dichlorobenzene, and the negative acti-

vation entropy values were found to be quite large for both

solvents, with that of o-dichlorobenzene being somewhat

larger.

Scheme 5.

Cyclization of 10b to 5b in Nondonor Solvents

Cyclization of 10b to 5b and 6b in Pyridine

In order to gain insight into the mechanism of the cycli-

zation reaction of 10b, we carried out kinetic studies. This

Cyclization in pyridine (308–323 K) was found to yield

cyclization reaction is also interesting from the point of both O-trans 5b and O-cis 6b. Thus, the measurements were

view that it can serve as a model for the heterolytic cleavage analyzed according to Equations (1), (2), and (3), with the

of element–hydrogen or element–carbon bonds to give com- rate constants corresponding to the reactions in Scheme 6,

pounds bearing new element–heteroatom bonds with con- where [5b], [6b], and [10b] denote the concentration of 5b,

comitant elimination of H₂or hydrocarbon. In the case of 6b, and 10b at arbitrary intervals; the subscript “0” denotes

second-row elements and below, there have been sugges- concentration at t = 0.

Table 4. The cyclization reactions of 10b to 5b in nondonor solvents.^[a,b]

Solvent

Temp.

[K]

Rate

[s^–1]

ΔH^϶

ΔS^϶

[eu]

[kcalmol^–1]

Toluene

343

353

363

373

343

353

363

373

(1.76 0.05)×10^–6

(3.95 0.08)×10^–6

(8.15 0.12)×10^–6

(1.46 0.12)×10^–5

(9.87 0.23)×10^–6

(1.84 0.07)×10^–5

(3.32 0.02)×10^–5

(5.55 0.05)×10^–5

17.3 0.6

–34.7 1.7

o-Dichlorobenzene

14.3 0.2

–41.0 0.6

[a] The process was monitored by ¹⁹F NMR spectroscopy. [b] Error is given as standard deviation.

224

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

[10b] = [10b]₀exp{–(k₁+ k₃)t}

FULL PAPER

(1)

(2)

for the formation of both O-trans 5b and O-cis 6b are much

larger, whereas the pseudorotation rates stay essentially the

same (vide infra). The data indicate that the path for direct

O-trans 5b formation (k₃) is entropy-favored and that for

O-cis 6b is enthalpy-driven (k₁). The fact that the pseudoro-

tation of O-cis 6b to O-trans 5b (k₂) is 4–5-fold slower than

the formation of both O-trans 5b and O-cis 6b from 10b is

noteworthy.

[6b] = [10b]₀{k₁/(k₂– k₁– k₃)}[exp{–(k₁+ k₃)t} – exp(–k₂)t]

[5b] = [10b]₀[1 – {(k₂– k₃)/(k₂– k₁– k₃)}exp{–(k₁+ k₃)t} +

k₁/(k₂– k₁– k₃)exp(–k₂)t]

(3)

Scheme 6.

Figure 5, Table 5, and Figure 6 show examples of the

curve fitting of the measurements at 323 K, each depicting

the observed relative concentrations of 10b, 6b, and 5b, the

calculated rates, and the Eyring plot of these rates. Com-

pared with the cyclization in nondonor solvents, the rates

Figure 6. Eyring plot of the rates k₁+ k₃(open circles), k₁(open

squares), k₂(filled squares), and k₃(filled circles) calculated from

curve fitting of the rate profiles of the thermal reaction of 10b in

pyridine (Scheme 6).

Cyclization of 10c to 5c and 7

The cyclization reaction of 10c was carried out after re-

moval of the solvent molecule in vacuo. The rates were

found to be very slow, as a whole, compared with those of

N

10b (Scheme 7, Table 6). p-Xylene (E_T= 0.074, DN^N

=

0.13) and o-dichlorobenzene were used as nondonor sol-

vents in the temperature range 393–423 K. Because of the

complexity of the ¹⁹F NMR spectra of 10c, the ³¹P NMR

spectrum was monitored instead. We found that both O-

trans 5c and 7 were formed, the latter as a consequence of

the loss of the tert-butyl group. In the case of p-xylene as

solvent, whereas an appreciable amount of 7 was observed

at 393 K, only trace amounts were detected at higher tem-

peratures. The rate of formation of 7 was considerably

higher in o-dichlorobenzene than in p-xylene. The absolute

values of the activation entropies for the formation of O-

trans 5c were found to be quite small compared with those

of O-trans 5b formation. In o-dichlorobenzene, for instance,

the value for O-trans 5b is –41 eu whereas it is –10 eu for

O-trans 5c. The activation entropy for the formation of 7

was found to have a larger negative value (–22 eu) than that

of the O-trans 5c forming process (–10 eu).

The effect of using pyridine as solvent was dramatic.

Only the less stable rotamer O-cis 6c was formed at tem-

peratures considerably lower than those required for the

formation of O-trans 5c in nondonor solvents. In this case

it should also be noted that the cyclization of 10b to O-cis

6c is significantly faster than the pseudorotation of O-cis 6c

to O-trans 5c.

Figure 5. Rate profile showing the relative concentrations of 10b

(open circles), 5b (filled circles), and 6b (filled squares) for the ther-

mal reaction of 10b in pyridine at 323 K.

Table 5. The cyclization reactions of 10b in pyridine.^[a,b]

Process

Temp.

[K]

Rate

[s^–1

ΔH^϶

ΔS^϶

[eu]

]

[kcalmol^–1

]

k₁

308

313

318

323

308

313

318

323

308

313

318

323

(1.17 0.07)×10^–4

(1.83 0.05)×10^–4

(2.82 0.26)×10^–4

(4.52 0.20)×10^–4

(2.63 0.08)×10^–5

(4.36 0.11)×10^–5

(7.83 0.38)×10^–5

(1.26 0.04)×10^–4

(1.04 0.14)×10^–4

(1.74 0.09)×10^–4

(3.33 0.41)×10^–4

(6.64 0.43)×10^–4

17.1 0.3

20.3 0.5

23.9 1.3

–21.0 1.1

–13.8 1.7

+0.8 4.2

(10b Ǟ 6b)

k₂

(6b Ǟ 5b)

k₃

(10b Ǟ 5b)

[a] The process was monitored by ¹⁹F NMR spectroscopy. [b] Error

is given as standard deviation.

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

225

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

Table 6. The cyclization reactions of 10c in various solvents.^[a]

Solvent

Product

Temp. [K]

Rate [s^–1]

ΔH^϶[kcalmol^–1]

ΔS^϶[eu]

p-Xylene^[b]

5c

393

403

413

423

393

403

413

423

393

403

413

423

333

343

353

363

(1.14 0.02)×10^–6

(3.66 0.07)×10^–6

(9.58 0.10)×10^–5

(1.46 0.04)×10^–5

(3.15 0.20)×10^–7

(2.06 0.01)×10^–6

(5.73 0.10)×10^–6

(1.45 0.04)×10^–5

(3.22 0.07)×10^–5

(1.81 0.02)×10^–6

(3.84 0.08)×10^–6

(8.06 0.09)×10^–6

(1.92 0.03)×10^–5

(2.75 0.03)×10^–5

(8.43 0.11)×10^–5

(2.04 0.03)×10^–4

(4.93 0.08)×10^–4

32.2 1.1

–4.2 2.6

7

5c

o-Dichlorobenzene^[b]

29.5 0.7

25.0 1.0

22.3 0.6

–9.9 1.8

–21.8 2.6

–12.7 1.8

7

Pyridine^[c]

6c

[a] Error is given as standard deviation. [b] The process was monitored by ³¹P NMR spectroscopy. [c] The process was monitored by ¹⁹F

NMR spectroscopy.

of the O_equatorial,O_equatorialisomer. This rotamer would eas-

ily give the O-trans isomer 5 upon a single pseudorotation.

Last of all, having the H atom already apical is not only

favorable in terms of apicophilicity but also due to the fact

that it is weakly bonded to the phosphorus atom in TBP

structures, thus enabling facile bond dissociation en route

to products via a four-membered cyclic transition state. Al-

though an ionic stepwise mechanism cannot be fully ex-

cluded, we believe it is not likely because a stable anionic

hexacoordinate species would have to be regarded as the

reactive species, and this conflicts with the fact that mono-

deprotonation of 10a (Me) with aqueous KOH in Et₂O re-

sults in the facile formation of a stable hexacoordinate spe-

cies [³¹P NMR: δ = –153 ppm (¹J_P,H= 679 Hz)] that is much

less prone to cyclization than 10a itself.

Scheme 7.

Mechanism of the Cyclization Reaction

If we assume that nucleophilic reaction to a pentacoordi-

nate phosphorus atom occurs with equatorial edge attack,

as is generally accepted,^[35]then rationalization of the

mechanism for formation of the O-cis isomer in a donor

solvent (pyridine) becomes straightforward by considering

configuration 10A as the reacting species, where P–H is api-

cal and P–C(aryl) is equatorial (Scheme 8).^[36]Attack of the

hydroxy oxygen atom from the rear side of the P–C(aryl)

bond to form a pseudohexacoordinate phosphorus species

with concomitant H₂extrusion and pyridine dissociation by

a four-membered concerted reaction, as depicted in TS-cis,

would give O-cis 6 with the newly formed P–O bond be-

coming the new apical bond (P–O1). This mechanistic ra-

tionale is appropriate for the following four reasons. First

of all, since P–H phosphoranes 10 have only one P–O bond,

Scheme 8.

As for the formation of O-trans 5b (nBu) in donor sol-

whereas spirophosphoranes O-trans 5 and O-cis 6 have two, vents, the fact that the rate of formation of O-trans 5b is

it is reasonable to assume that the latter are more stable as similar to that of O-cis 6b (Table 5) makes O-trans 5b for-

hypervalent species. Thus, it is reasonable to assume that mation by unfavorable oxygen attack the σ*-orbital of the

the transition state of the cyclization reaction is reactant- P–C(alkyl) bond in configuration 10A highly unlikely

like (TS-cis) and TBP features of the reactant are retained (Scheme 9). Also, a rather large negative value for the acti-

to a large extent. Secondly, the reactant takes on its lowest vation entropy would have been expected because a more

energy configuration with 10A and thus the formation of progressed and thus tighter transition state would be neces-

O-cis 6 requires only a minimal amount of motion. Thirdly, sary. It follows that the formation of O-trans 5b requires

attack of the oxygen atom to the σ*-orbital of the P–C(aryl) pseudorotation to a different configuration (such as inter-

bond to form the new apical P–O1 bond is expected to be mediate 10F), where P–H is apical and P–O is equatorial,

energetically more favorable than attack to the σ*-orbital and which is inevitably higher in energy than 10A, to pre-

of the P–C(alkyl) bond, which would lead to the formation cede the cyclization (TS-trans). Thus, the enthalpy value of

226

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

23.9 kcalmol^–1(considerably larger than the value of nondonor solvents. However, this was not the case. There-

17.1 kcalmol^–1for the path that gives O-cis 6b: TS-cis) fore, for 10c it is highly likely that the cyclization in nondo-

could be considered the result of a combination of two bar- nor solvents gives O-trans 5c directly via 10F. By analogy,

riers, that of pseudorotation and cyclization. Pseudorota- we expect the cyclization of 10b to be the same. For the

tion would allow 10b to take on configurations in which the cyclization of 10b in nondonor solvents, the fact that ΔS^϶

oxygen atom could attack the electronically favorable σ*- has a large negative value and that ΔH^϶decreases in the

orbital of a P–O bond. Considering the fact that the apico- more polar o-dichlorobenzene (k_{o-dichlorobenzene}/k_toluene

=

philicity order is H Ͼ aryl Ͼ alkyl and that it is the hydro- 3.8–5.6) is consistent with a concerted mechanism, as pro-

gen atom on the phosphorus atom that is removed from the posed for the reaction in pyridine, that differs only in the

molecule, configuration 10F can be considered to be the absence of an interacting solvent molecule. The solvent ef-

reactive intermediate. This species can be generated by a fect for 10c (k_{o-dichlorobenzene}/k_p-xylene= 1.8–2.2) is similar.

three-step pseudorotation process without passing through Concerted reactions such as sigmatropic reactions are

high-energy species, with the bidentate ligand occupying di- known to exhibit only small solvent effects in such a case

equatorial positions. The transition state is probably also as this.^[37]

concerted in this case, with concomitant pyridine dissoci-

Both TS-cis (Scheme 8) and TS-trans (Scheme 9) are ex-

ation (vide supra). The fact that the activation entropy for pected to be lowered in energy in pyridine due to the in-

formation of O-cis 6b is more negative than for O-trans 5b creased nucleophilicity of the hydroxy oxygen atom. How-

formation may be due to a somewhat later transition state ever, since the activation energy for TS-trans is a combina-

and thus more progressive bond formation for the former tion of pseudorotation (not affected by solvent) and a ring-

process. That O-trans 5c (tBu) could not be observed in the

cyclization of 10c in pyridine (Table 6) probably reflects the

fact that the pseudorotation barrier for 10c is much higher

than that for 10b from a steric point of view, since the re-

quired passage through 10E, which bears an apical tBu

group, is thermodynamically quite unfavorable.

Figure 7. Energy profile for the thermal reaction of 10b. The

broken line and bold line correspond to reactions in nondonor sol-

vents and in pyridine, respectively. The curved arrows denote the

Scheme 9.

observed pathways for each group of solvents. The straight arrows

show the degree of stabilization upon changing the solvent.

The mechanism for the formation of the stable O-trans

isomer 5 in nondonor solvents is less obvious, since the tem-

peratures required for the cyclization to O-trans 5b (nBu;

Table 4) exceed those required for pseudorotation of O-cis

6b to O-trans 5b (Table 5). It is possible that the initial prod-

uct is unobserved O-cis 6, which quickly pseudorotates to

O-trans 5 under the reaction conditions, or that 10F

(Scheme 9) is the reactive species as in the donor solvent

case and directly gives rise to O-trans 5c. In the case of 10c

(tBu), the rate of formation of O-trans 5c was found to be

only slightly slower than the pseudorotation of O-cis 6c to

O-trans 5c (Table 6), the latter of which is expected to be

essentially independent of the solvent. Thus, if O-trans 5c

were to form by a sequence of O-cis 6c formation followed

by pseudorotation, then the latter transformation would be

expected to be observed during the cyclization reactions in Scheme 10.

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

227

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

closure processes, the relative degree of stabilization should pected from the pseudorotation between previously exam-

be less for TS-trans, as is observed (Figure 7).

ined diastereomeric O-trans spirophosphoranes.^[8]

Reasoning analogous to that for the direct formation of

5 can be applied to the formation of 7, as depicted in

Scheme 10. X-ray structures show that, compared with 10b

(nBu; 1.812 Å), there is a notable increase in the P–C(alkyl)

bond length in 10c_intra(tBu; 1.854 Å) with still further

elongation in 10c_inter(tBu; 1.878 Å), in which a P–O inter-

action is present (Figure 2), thus implying the potential for

the P–tBu bond to undergo cleavage.

Pseudorotation of the O-cis Isomer to the O-trans Isomer

The rates of pseudorotation of O-cis 6 to O-trans 5, an

apparently irreversible process, were measured by monitor-

N

ing the ¹⁹F NMR spectra in toluene (302–328 K; E_T

=

0.099,^[33]DN^N= 0.003,^[34]β = 0.11^[38]), pyridine (313–

N

328 K; E_T= 0.302, DN^N= 0.85, β = 0.64), and ethanol

Figure 8. Eyring plot of the rates of stereomutation of 6b to 5b in

toluene (open circles), pyridine (open squares), and ethanol (filled

circles).

N

(313–328 K; E_T= 0.654, DN^N= 0.82, β = 0.77) for O-cis

6b, and in p-butyltoluene (expected to have physical proper-

ties similar to p-xylene) for O-cis 6c. All the measurements

conformed to first-order kinetics, and activation parameters

could be estimated from the Eyring plot of the rates

(Table 7, Figure 8). The rates for O-cis 6b in pyridine were

taken from the complex reaction of 10b (Scheme 6). An in-

dependent measurement of isolated O-cis 6b at 313 K (inde-

Relative Stability of the Stereoisomers O-trans 5 and O-cis

6

Here we take it for granted that the stereomutation of O-

pendent: 4.24×10^–5s^–1; curve fitting: 4.36×10^–5s^–1) con- trans 5 [21 (5-S) or 12 (5-R)] to its enantiomer (12 or 21)

firmed the validity of the rates derived from the curve fit- follows Berry pseudorotation. In principle, up to 20 stereo-

ting (vide infra). For O-cis 6b, the rates in pyridine and isomers are possible, as depicted in the Desargues–Levi dia-

ethanol differ from that in toluene by less than 10% at the gram (Figure 9).^[39]For the present group of isomeric spiro-

same temperatures [(1.26, 1.16, and 1.29)×10^–4s^–1, respec- phosphoranes, including O-trans 5 and O-cis 6, a bidentate

tively, at 323 K]. Therefore, only a very small solvent effect ligand cannot span between the two apical positions.

is present and thus mechanisms involving proton-catalyzed Hence, the four stereoisomers 13, 31, 24, and 42 can be

P–O bond dissociation/recombination with intervention of eliminated as probable isomers. In the case where a five-

highly polar species can be disregarded. The magnitude of membered bidentate ligand lies in the equatorial plane, the

the entropy is not so large and the sign is negative, as ex- stereoisomers should be of high energy because of ring

Table 7. Rates and activation parameters of the stereomutation of O-cis 6 to O-trans 5.^[a,b]

Solvent

toluene

Temp. [K]

Rate [s^–1]

ΔH^϶[kcalmol^–1]

ΔS^϶[eu]

6b

302

308

313

318

323

(1.18 0.01)×10^–5

(2.45 0.04)×10^–5

(4.03 0.07)×10^–5

(7.61 0.12)×10^–5

(1.29 0.01)×10^–4

(2.31 0.01)×10^–4

(2.63 0.08)×10^–5

(4.36 0.11)×10^–5

(4.24 0.09)×10^–5

(7.83 0.38)×10^–5

(1.26 0.04)×10^–4

(3.44 0.03)×10^–5

(6.56 0.05)×10^–5

(1.16 0.03)×10^–4

(1.97 0.02)×10^–4

(6.73 0.10)×10^–6

(3.89 0.02)×10^–5

(2.15 0.04)×10^–5

(7.62 0.02)×10^–5

21.8 0.4

–9.0 1.2

328

pyridine

308^[c]

313^[c]

313^[d]

318^[c]

323^[c]

313

20.3 0.5

–13.8 1.7

ethanol

23.1 0.5

27.2 0.8

–5.3 1.5

318

323

328

403

423

443

463

6c

p-butyltoluene

–13.1 1.8

[a] The process was monitored by ¹⁹F NMR spectroscopy. [b] Error is given as standard deviation. [c] Values from the cyclization reaction

of 10b in pyridine. [d] Measured with isolated 6b and not included in the activation parameter calculations.

228

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

strain. Among these isomers, 35, 53, 45, and 54 can be con- energy difference between isomers O-trans 5 and O-cis 6

sidered to be higher in energy than 15, 51, 25, and 52 be- directly. The activation parameters for the transformation

cause isomers of the former group bear two apical carbon of O-cis 6b to O-trans 5b were obtained as ΔH^϶

=

substituents. Thus, since the BPR process requires passage 21.8 kcalmol^–1and ΔS^϶= –9.0 eu (Table 7). Unfortunately,

through at least one of these groups of isomers, it is more however, the interconversion barrier between enantiomers

appropriate to assume the latter. The isomers next highest of O-trans 5 was too high to be measured by methods that

in energy are 34 and 43, which have two oxygen substituents allow the use of symmetric compounds.^[17d]Thus, we re-

in the equatorial plane but do not possess an equatorial sorted to comparing with compounds 14b-endo and 14b-

bidentate ligand. That leaves 23, 41, 32, and 14 (O-cis 6) exo, which differ from O-trans 5 in only having one CH₃

and the most stable isomers 12 and 21 (O-trans 5). Because group in place of one of the four CF₃groups. The averaged

of the symmetry of the molecules, the pairs (15, 52), (25, activation parameters were ΔH^϶= 33.7 kcalmol^–1and ΔS^϶

51), (23, 14), (41, 32), (54, 35), and (53, 45) are degenerate. = –8.0 eu for the interconversion of the pair of 14.^[8]Thus,

Hence, the pathway for the interconversion between O-cis 6b should be less stable than O-trans 5b by about 12

enantiomers of the most stable isomers 12 and 21 is reduced (33.7 – 21.8) kcalmol^–1. Since the oxygen atom in the Mar-

to 12–43–25–41–23–15–34–21 (shown as bold lines) or 12– tin ligand is expected to be slightly more apicophilic than

43–51–32–14–52–34–21 (shown as a combination of bold the oxygen atom attached to the asymmetric carbon atom

and normal lines), which are degenerate. Here the process in 14, the actual difference should be somewhat larger. At

41–23 (or 32–14) corresponds to interconversion of enantio- first glance, this difference of 12 kcalmol^–1would seem

mers of O-cis 6. Therefore, it follows that the interconver- rather large for a compatible pair of stereoisomers, al-

sion between enantiomers of O-trans 5 (12 and 21) and though they are only compatible because of the pseudorota-

transformation of O-cis 6 to O-trans 5 (23 or 41 to 21 or tion barrier that separates them. However, this value is in

12) involve a common pathway and thus the same global good agreement with theoretical calculations on the Me

transition state, whether the actual transition state be of pair of O-trans 5a and O-cis 6a, which have a difference of

TBP structure or of SP structure, or whether the whole con- 14.1 kcalmol^–1.^[19c]The use of the difference in activation

version process involves seven or five steps. Thus, the differ- entropy of the Me pair to account for that of the nBu pair

ence in activation energy between the interconversion and of O-trans 5b and O-cis 6b is justified because a kinetic

the transformation processes would, in principle, give the examination on the pseudorotation of a diastereomeric pair

Figure 9. Simplified Desargues–Levi diagram for the pseudorotation of spirophosphoranes 5 and 6. The two-digit numerals correspond

to the substituents occupying the apical positions. Substituent priority numbers were arbitrarily assigned as shown for configurations 21

and 12. The pathways to consider are shown as bold lines. Paths involving vertices corresponding to 13, 31, 24, and 42, which are not

probable due to the ring effect, and those involving superficially identical compounds due to degeneracy (C₂symmetry of the compounds),

are shown as dotted and normal lines, respectively.

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

229

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

of 14a bearing an equatorial methyl group in the place of

an equatorial nBu group in 14b shows that the primary dif-

ference in the rate of interconversion between the equatorial

Me and nBu series (14a: ΔG^϶= 34.7 kcalmol^–1; 14b: ΔG^϶

= 36.3 kcalmol^–1) is entropy, and that the enthalpy of acti-

vation is essentially the same (14a: ΔH^϶= 33.5 kcalmol^–1;

14b: ΔH^϶= 33.7 kcalmol^–1).^[8]The Berry pseudorotation

pathway of lowest energy for the inversion of O-trans 5

should be the path shown in Figure 9, considering the com-

prehensive analysis carried out by Holmes.^[10,40]

Conclusions

In conclusion, by utilizing the Martin ligand, we have

isolated and characterized the first anti-apicophilic phos-

phoranes O-cis 6, with substituent configurations that viol-

ate the apicophilicity rule, as high-energy stereoisomers of

phosphoranes O-trans 5 with an ordinary O_apical,C_equatorial

array. This was only possible because the phosphoranes O-

cis 6 can be formed under non-equilibration (kinetic) condi-

tions from 10, and because the barrier of pseudorotation

for these pseudorotamers to their more stable stereoisomers

O-trans 5 is sufficiently high enough to freeze the conver-

sion and allow isolation. During the course of these studies

we have also found that monocyclic P–H_apicalphosphoranes

10, which are precursors to O-cis 6, exist as mixtures of

solvated and unsolvated species. A ligand-exchange process

interconverting the bidentate Martin ligand with the mono-

dentate Martin ligand was found to exist. This process

likely proceeds via hexacoordinate transition states. In ad-

dition, a rationale of the cyclization reaction to produce O-

cis 6 has been presented on the basis of a kinetic examina-

tion of the process. The energy difference between O-trans

5 and O-cis 6 (R = nBu) was elucidated to be about

12 kcalmol^–1from kinetic measurements of the transforma-

tion of O-cis 6 to O-trans 5 and between the interconversion

of diastereomeric analogs of O-trans 5. This is in good

agreement with the calculated difference of 14.1 kcalmol^–1

between O-trans 5 and O-cis 6 (R = Me).

There still remains the question of whether the diequato-

rial oxygen isomers (34 and 43), the species attained by a

single pseudorotation of 5b with the nBu group as the pivot,

actually exists as an energy-minimum species. As we can

make a tentative guess from the kinetic parameters deduced

from these studies for our system, intermediate A (A-R or

A-S in Figure 10), were it to exist, is expected to be energeti-

cally less stable than 6b from usual apicophilicity considera-

tions, and thus the energy difference should be more (prob-

ably by at least a few kcalmol^–1) than the value of

12 kcalmol^–1given above for the difference between 5b and

6b. As for the activation energy for the one-step pseudoro-

tation process between 5b and A, it can be estimated to

be roughly the same (10 kcalmol^–1) as that for the process

between enantiomers of 6b because both pseudorotations

involve a single step with the same pivot substituent (nBu).

Thus, the assumed pseudorotation activation barrier from

5b to A is already smaller than the relative energy between

5b and A by about 2 (12 – 10) kcalmol^–1as the lower limit.

Therefore, it is unlikely that A exists as an energy minimum;

it is probably a structure along the reaction pathway, or it

could be that the turnstile mechanism^[41]that is frequently

neglected is the actual operative route.

Experimental Section

General: Melting points were measured with a Yanaco micro melt-

ing point apparatus and are uncorrected. ¹H (400 MHz), ¹³C

(100 MHz), ¹⁹F (376 MHz), and ³¹P NMR (162 MHz) spectra were

recorded with a JEOL EX-400 spectrometer. ¹H NMR chemical

shifts (δ) are given in ppm downfield from residual [D]chloroform

(δ = 7.26 ppm). ¹³C NMR chemical shifts (δ) are given in ppm from

[D]chloroform (δ = 77.0 ppm). ¹⁹F NMR chemical shifts (δ) are

given in ppm downfield from external CFCl₃. ³¹P NMR chemical

shifts (δ) are given in ppm downfield from external 85% H₃PO₄.

Elemental analyses were performed with a Perkin–Elmer 2400

CHN elemental analyzer. All reactions were carried out under N₂.

Figure 10. Energy diagram for Berry pseudorotation involving 5b, 6b, and 14b (exo and endo).

230

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

THF was freshly distilled from Na/benzophenone. Ethanol was dis-

tilled from magnesium. All other solvents were distilled from CaH₂.

Phosphoranes 7^[17c,20b]and 14^[8c]were prepared according to pub-

lished procedures. Preparative thin layer chromatography was car-

yl]phenyl}propan-2-ol (10c): tBuLi (1.70 m in pentane, 3.60 mL,

6.12 mmol) was added to a solution of P–H_equatorialspirophos-

phorane 7 (1.00 mg, 1.94 mmol) in Et₂O (20 mL) at –78 °C. After

removal of the cooling bath, the solution was stirred at room tem-

perature for 30 min. The solution was quenched with 1 m HCl

(20 mL), extracted with Et₂O (3×50 mL), dried with anhydrous

MgSO₄, and concentrated in vacuo. Separation and purification of

the residue was carried out by TLC (hexane/CH₂Cl₂= 5:1) to give

a mixture (1.01 g, 92%) of 10c_intraand 10c_interas a white powder.

ried out on plates of Merck silica gel 60 GF₂₅₄

.

[TBPY-5-15]-1,1,1,3,3,3-Hexafluoro-2-{2-[1-methyl-3,3-bis(tri-

fluoromethyl)-1,3-dihydro-2,1λ⁵-benzoxaphosphol-1-

yl]phenyl}propan-2-ol (10a): MeLi (1.14 m in Et₂O, 5.67 mL,

6.46 mmol) was added to a solution of P–H_equatorialphosphorane

7 (460 mg, 0.891 mmol) in Et₂O (10 mL) at –78 °C. After removal

of the cooling bath, the solution was stirred at room temperature

for 50 min. The solution was quenched with 1 m HCl (3 mL), ex-

tracted with Et₂O (3×30 mL), dried with anhydrous MgSO₄, and

concentrated in vacuo. Separation and purification of the residue

was carried out by TLC (hexane/CH₂Cl₂= 5:1) to give 10a

(322 mg, 68%) and 5a (105 mg, 22%) as white solids. Recrystalli-

zation of 10a from hexane/CH₂Cl₂gave crystals for X-ray analysis.

Recrystallization from hexane/CCl₄gave crystals of 10c_intra. 10c_intra

:

M.p. 108 °C. ¹H NMR (CDCl₃): δ = 8.58 (br. s, 1 H), 8.00–7.89

1

(m, 1 H), 7.80–7.48 (m, 7 H), 5.81 (d, J_H,P= 273 Hz, 1 H), 1.30

3

(d, J_H,P= 23.4 Hz, 9 H) ppm. ¹³C NMR (CDCl₃): δ = 139.3 (d,

1

3

¹J_C,P= 139.7 Hz), 136.7 (d, J_C,P= 132.4 Hz), 136.2 (d, J_C,P

=

3

2

16.6 Hz), 134.6 (d, J_C,P= 12.9 Hz), 134.0, 131.3 (d, J_C,P

2

14.7 Hz), 130.6 (d, J_C,P= 14.7 Hz), 129.5, 129.2 (d, J_C,P

16.6 Hz), 129.0, 128.8 (d, ³J_C,P= 3.9 Hz), 126.5 (d, J_C,P= 16.6 Hz),

1

123.4 (q, J_C,F= 286.8 Hz), 123.1 (q, J_C,F= 290.4 Hz), 122.7 (q,

10a: M.p. 101 °C. H NMR (CDCl₃): δ = 8.86 (br. s, 1 H), 8.10–

1

2

1

¹J_C,F= 288.6 Hz), 122.3 (q, J_C,F= 288.6 Hz), 79.7 (sept, J_C,F

=

7.90 (m, 1 H), 7.79 (br. s, 4 H), 7.55–7.20 (m, 3 H), 6.37 (dq, J_H,P

2

1

= 276, ³J_H,H= 3.9 Hz, 1 H), 2.44 (dd, ²J_H,P= 14.7, ³J_H,H= 3.9 Hz,

3 H) ppm. ¹³C NMR (CDCl₃): δ = 136.7 (d, ²J_C,P= 11.0 Hz), 134.8

31.3 Hz), 78.4 (sept, J_C,F= 31.3 Hz), 43.8 (d, J_C,P= 90.1 Hz),

15.1 ppm. ¹⁹F NMR (CDCl₃): δ = –73.4 (q, J_F,F= 9.5 Hz, 3 F), –

4

3

73.5 (q, J_F,F= 9.5 Hz, 3 F), –74.6 (q, J_F,F= 9.1 Hz, 3 F), –76.1

(br. s), 133.7 (d, J_C,P= 3.7 Hz), 133.5 (br. s), 131.4 (d, J_C,P

=

(q, J_F,F= 10.7 Hz, 3 F) ppm. ³¹P NMR (CDCl₃): δ = –14.7 ppm.

4

14.7 Hz), 131.1 (br. d, ²J_C,P= 12.9 Hz), 129.8 (br. s), 129.6 (d, ¹J_C,P

= 123.2 Hz), 129.1 (br. d, J_C,P= 14.7 Hz), 128.3 (d, J_C,P

141.3 Hz), 128.2 (d, J_C,P= 25.7 Hz), (d, J_C,F= 286.8 Hz), 124.8

3

1

C₂₂H₁₉F₁₂O₂P (574.34): calcd. C 46.01, H 3.33; found C 45.85, H,

3.23. Recrystallization from hexane/EtOH gave crystals of 10c_inter

(containing EtOH). An X-ray sample was recrystallized from hex-

ane/MeCN. 10c_inter: M.p. 83–84 °C (containing EtOH). ¹H NMR

([D₆]acetone, without EtOH): δ = 8.77–8.72 (m, 1 H), 7.87–7.82

(m, 1 H), 7.75–7.63 (m, 4 H), 7.63–7.54 (m, 2 H), 7.48 (s, 1 H),

=

2

1

3

1

(d, J_C,P= 14.7 Hz), 122.6 (q, J_C,F= 284.8 Hz), 122.4 (q, J_C,F

=

290.4 Hz), 122.2 (q, ¹J_C,F= 290.5 Hz), 79.0 (sept, ²J_C,F= 35.0 Hz),

78.9 (sept, J_C,F= 33.1 Hz), 22.4 (d, J_C,P= 99.3 Hz) ppm. ¹⁹F

NMR (CDCl₃): δ = –72.9 (br. s, 3 F), –76.3 (br. s, 3 F), –76.9 (br.

s, 3 F), –77.0 (br. s, 3 F) ppm. ³¹P NMR (CDCl₃): δ = –51.9 ppm.

C₁₉H₁₃F₁₂O₂P (532.26): calcd. C 42.88, H 2.46; found C 42.75, H

2.37.

2

1

3

6.07 (d, J_H,P= 293 Hz, 1 H), 1.17 (d, J_H,P= 19.5 Hz, 9 H) ppm.

¹³C NMR ([D₆]acetone, without EtOH): δ = 141.3 (d, J_C,P

141.3 Hz), 139.0 (d, J_C,P= 5.5 Hz), 137.1 (d, J_C,P= 14.7 Hz),

136.1 (d, ³J_C,P= 14.7 Hz), 135.1 (d, ¹J_C,P= 139.7 Hz), 133.9, 132.7

(d, ²J_C,P= 3.6 Hz), 130.7, 130.1 (d, ²J_C,P= 12.9 Hz), 128.6 (d, ³J_C,P

=

1

4

2

[TBPY-5-15]-1,1,1,3,3,3-Hexafluoro-2-{2-[1-butyl-3,3-bis(trifluoro-

methyl)-1,3-dihydro-2,1λ⁵-benzoxaphosphol-1-yl]phenyl}propan-2-ol

(10b): nBuLi (1.62 m in hexane, 2.00 mL, 3.24 mmol) was added to

a solution of P–H_equatorialspirophosphorane 7 (500 mg,

0.969 mmol) in Et₂O (10 mL) at –78 °C. After removal of the cool-

ing bath, the solution was stirred at room temperature for 60 min.

The solution was quenched with 1 m HCl (10 mL), extracted with

Et₂O (3×30 mL), dried with anhydrous MgSO₄, and concentrated

in vacuo. Separation and purification of the residue was carried

out by TLC (hexane/CH₂Cl₂= 5:1) to give 10b (322 mg, 68%) and

5b (105 mg, 22%) as white solids. Recrystallization of 10b from

hexane/CH₂Cl₂gave crystals suitable for X-ray analysis. 10b: M.p.

3

1

= 7.4 Hz), 125.9 (d, J_C,P= 12.9 Hz), 125.2 (q, J_C,F= 286.8 Hz),

1

124.7 (q, J_C,F= 288.6 Hz), 124.6 (q, J_C,F= 290.5 Hz), 122.3 (q,

¹J_C,F= 288.7 Hz), 81.0 (sept, J_C,F= 29.5 Hz), 80.0 (sept, J_C,F

=

2

29.4 Hz), 38.6 (d, J_C,P= 97.5 Hz), 19.4 ppm. ¹⁹F NMR ([D₆]ace-

1

4

9

tone, without EtOH): δ = –73.3 (qq, J_F,F= 8.9, J_F,F= 3.1 Hz, 3

4

F), –73.6 (q, J_F,F= 9.5 Hz, 3 F), –74.3 (q, J_F,F= 8.6 Hz, 3 F),

–75.5 (qq, J_F,F= 9.2, J_F,F= 2.8 Hz, 3 F) ppm. ³¹P NMR ([D₆]-

acetone, without EtOH): δ = –40.1 ppm. C₂₄H₂₅F₁₂O₃P (620.41):

calcd. C 46.46, H 4.06 (containing EtOH); found C 46.26, H, 3.90.

4

9

[TBPY-5-12]-1-Methyl-3,3,3Ј,3Ј-tetrakis(trifluoromethyl)-1,1Ј-

spirobi[1,3-dihydro-2,1λ⁵-benzoxaphosphole] (6a): A solution of P–

H_apicalphosphorane 10a (50 mg, 9.4×10^–2mmol) in THF (0.5 mL)

was heated at 40 °C for 1 h. The solution was concentrated in

vacuo. Purification of the residue was carried out by TLC (hexane/

CH₂Cl₂= 2:1) to give 5a (R_f= 0.8, 23 mg, 46%) and 6a (R_f= 0.9,

1

119 °C. H NMR (CDCl₃): δ = 9.13 (br. s, 1 H), 8.07–8.02 (m, 1

H), 7.89–7.85 (m, 1 H), 7.80–7.72 (m, 3 H), 7.45–7.41 (m, 1 H),

1

3

7.36–7.32 (m, 1 H), 6.16 (dd, J_H,P= 273, J_H,H= 5.9 Hz, 1 H),

3.21–3.06 (m, 1 H), 2.69–2.52 (m, 1 H), 1.65–1.46 (m, 1 H), 1.45–

3

1.36 (m, 2 H), 1.35–1.22 (m, 1 H), 0.88 (t, J_H,H= 7.3 Hz, 3 H)

2

ppm. ¹³C NMR (CDCl₃): δ = 136.6 (d, J_C,P= 12.8 Hz), 136.1 (d,

1

26 mg, 52%) as white solids. 5a: M.p. 134 °C. H NMR (CDCl₃):

3

¹J_C,P= 139.7 Hz), 134.7, 134.3 (d, J_C,P= 14.7 Hz), 134.1 (d, J_C,P

δ = 8.43–8.38 (m, 2 H), 7.79–7.69 (m, 6 H), 2.11 (d, ²J_H,P= 16.6 Hz,

2

1

3

3 H) ppm. ¹³C NMR (CDCl₃): δ = 136.8 (d, J_C,P= 9.1 Hz), 136.1

= 14.7 Hz), 130.8 (d, J_C,P= 16.5 Hz), 130.2, 129.8 (d, J_C,P

139.7 Hz), 129.7, 129.1 (d, J_C,P= 16.6 Hz), 127.0 (d, J_C,P

=

2

3

2

4

2

(d, J_C,P= 20.2 Hz), 133.7 (d, J_C,P= 3.7 Hz), 131.4 (d, J_C,P

=

1

3

12.9 Hz), 123.4 (d, J_C,F= 290.4 Hz), 122.7 (q, J_C,F= 286.8 Hz),

14.7 Hz), 130.6 (d, J_C,P= 176.5 Hz), 124.8 (d, J_C,P= 14.7 Hz),

122.5 (q, ¹J_C,F= 286.8 Hz), 122.3 (q, J_C,F= 288.6 Hz), 78.9 (sept,

1

122.6 (q, J_C,F= 285.0 Hz), 122.4 (q, J_C,F= 288.6 Hz), 81.3 (sept,

²J_C,F= 31.2 Hz), 78.6 (sept, J_C,F= 31.2 Hz), 34.7 (d, J_C,P

=

2

1

²J_C,F= 31.2 Hz), 25.2 (d, J_C,P= 125.0 Hz) ppm. ¹⁹F NMR

1

90.1 Hz), 27.7, 23.5 (d, J_C,P= 23.9 Hz), 13.6 ppm. ¹⁹F NMR

2

(CDCl₃): δ = –75.1 (q, ⁴J_F,F= 9.5 Hz, 6 F), –75.4 (q, ⁴J_F,F= 9.5 Hz,

6 F) ppm. ³¹P NMR (CDCl₃): δ = –22.6 ppm. C₁₉H₁₁F₁₂O₂P

(530.24): calcd. C 43.04, H 2.09; found C 42.83, H 1.86. 6a: M.p.

(CDCl₃): δ = –72.9 (br. s, 3 F), –76.3 (q, ⁴J_F,F= 9.2 Hz, 3 F), –76.8

4

9

(q, J_F,F= 8.2 Hz, 3 F), –77.0 (q, J_F,F= 9.2, J_F,F= 4.9 Hz, 3 F)

ppm. ³¹P NMR (CDCl₃): δ = –34.4 ppm. C₂₂H₁₉F₁₂O₂P (574.34):

calcd. C 46.01, H 3.33; found C 46.09, H, 3.24.

1

2

127 °C. H NMR (CDCl₃): δ = 7.80–7.31 (m, 8 H), 2.28 (d, J_H,P

= 10.2 Hz, 3 H) ppm. ¹³C NMR (CDCl₃): δ = 136.8 (d, J_C,P

9.1 Hz), 136.1 (d, J_C,P= 20.2 Hz), 133.8 (d, J_C,P= 3.7 Hz), 132.3

(br. s), 132.0 (br. s), 131.4 (d, J_C,P= 14.7 Hz), 130.3 (d, J_C,P

=

[TBPY-5-15]-1,1,1,3,3,3-Hexafluoro-2-{2-[1-(1,1-dimethylethyl)-3,3-

bis(trifluoromethyl)-1,3-dihydro-2,1λ⁵-benzoxaphosphol-1-

=

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

231

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

FULL PAPER

11.1 Hz), 125.6 (d, J_C,P= 9.1 Hz), 124.8 (d, J_C,P= 14.7 Hz), 122.6

(q, J_C,F= 285.0 Hz), 122.4 (q, J_C,F= 288.6 Hz), 122.3 (q, J_C,F

H_apicalphosphorane 10c (300 mg, 0.523 mmol) was treated with

pyridine (4.0 mL, 0.8 mmol) in THF (20 mL) at 60 °C for 36 h. The

solution was quenched with 1 m HCl (50 mL), extracted with Et₂O

(50 mL), and washed with aqueous CuSO₄(50 mL), water (50 mL),

and brine (50 mL). The mixture was then dried with MgSO₄, and

1

2

= 285.9 Hz), 121.9 (q, J_C,F= 285.8 Hz), 79.8 (sept, J_C,F

=

31.3 Hz), 26.2 (d, J_C,P= 129.5 Hz) ppm. ¹⁹F NMR (CDCl₃): δ =

1

4

–75.6 (q, J_F,F= 9.2 Hz, 6 F), –76.4 (q, J_F,F= 8.6 Hz, 6 F) ppm.

³¹P NMR (CDCl₃): δ = –6.3 ppm. C₁₉H₁₁F₁₂O₂P (530.24): calcd. concentrated in vacuo. Purification of the residue was carried out

C 43.04, H 2.09; found C 43.34, H 1.99.

by TLC (hexane/CH₂Cl₂= 3:1) to give 6c (R_f= 0.6, 224 mg, 75%)

and 10c (R_f= 0.4, 53.7 mg, 18 % recovery) as white solids.

Recrystallization of 6c from EtOH gave an X-ray sample. M.p.

[TBPY-5-12]-1-Butyl-3,3,3Ј,3Ј-tetrakis(trifluoromethyl)-1,1Ј-spirobi-

[1,3-dihydro-2,1λ⁵-benzoxaphosphole] (6b): P–H_apicalphosphorane

10b (200 mg, 0.348 mmol) was treated with pyridine (0.07 mL,

0.8 mmol) in THF (5 mL) at 60 °C for 20 min. The solution was

quenched with 1 m HCl (50 mL), extracted with Et₂O (50 mL), and

washed with aqueous CuSO₄(50 mL), water (50 mL), and brine

(50 mL). The mixture was then dried with MgSO₄, and concen-

trated in vacuo. Purification of the residue was carried out by TLC

(hexane/CH₂Cl₂= 2:1) to give 5b (R_f= 0.9, 58.2 mg, 29%) and 6b

(R_f= 0.7, 141 mg, 71%) as white solids. Recrystallization of 5b and

6b from MeCN gave single crystals suitable for X-ray analysis. 5b:

M.p. 109 °C. ¹H NMR (CDCl₃): δ = 8.43–8.38 (m, 2 H), 7.72–7.58

1

138 °C. H NMR (CDCl₃): δ = 7.91–7.88 (m, 2 H), 7.73–7.71 (m,

3

2 H), 7.62–7.52 (m, 4 H), 1.21 (d, J_H,P= 21.0 Hz, 9 H) ppm. ¹³C

1

2

NMR (CDCl₃): δ = 136.3 (d, J_C,P= 77.2 Hz), 134.1 (d, J_C,P

=

2

3

14.7 Hz), 133.8 (d, J_C,P= 12.9 Hz), 131.7, 129.4 (d, J_C,P

3

1

9.2 Hz), 125.4 (d, J_C,PP= 9.2 Hz), 122.64 (q, J_C,F= 286.8 Hz),

122.57 (q, ¹J_C,F= 286.8 Hz), 122.0 (q, ¹J_C,F= 288.6 Hz), 79.6 (sept,

²J_C,F= 31.2 Hz), 44.4 (d, J_C,P= 112.1 Hz), 29.6 ppm. ¹⁹F NMR

1

(CDCl₃): δ = –73.7 (q, ⁴J_F,F= 9.2 Hz, 6 F), –75.9 (q, ⁴J_F,F= 9.5 Hz,

6 F) ppm. ³¹P NMR (CDCl₃): δ = 7.7 ppm. C₂₂H₁₇F₁₂O₂P

(572.32): calcd. C 46.17, H 2.99; found C 46.32, H 3.27.

2

3

(m, 6 H), 2.34 (ddt, J_H,P= 28.3, J_H,H= 12.2, J_H,H= 4.9 Hz, 1 X-ray Crystal Structure Determination of 10a, 10b, 10c_intra, 10c_inter

,

2

3

H), 2.20 (ddt, J_H,P= 12.2, J_H,H= 12.2, J_H,H= 4.4 Hz, 1 H),

and 6c: Crystals suitable for X-ray structure determination were

3

1.86–1.70 (m, 1 H), 1.32–1.20 (m, 3 H), 0.81 (t, J_H,H= 7.3 Hz, 3 mounted on a Mac Science MXC3 diffractometer and irradiated

H) ppm. ¹³C NMR (CDCl₃): δ = 137.1 (d, J_C,P= 9.2 Hz), 136.2

with graphite-monochromated Cu-K_αradiation (λ = 1.54178 Å) for

data collection for 10b, 10c_intra, and 6c, and graphite-monochro-

mated Mo-K_αradiation (λ = 0.71073 Å) for data collection for 10a

and 10c_inter. Lattice parameters were determined by least-squares

fitting of 31 reflections for all compounds with 51° Ͻ 2θ Ͻ 60°.

Data were collected in the 2θ/ω scan mode. All data were corrected

for absorption^[42]and extinction.^[43]The structures were solved by

3

2

4

2

(d, J_C,P= 18.4 Hz), 133.6 (d, J_C,P= 3.6 Hz), 131.3 (d, J_C,P

=

1

3

14.7 Hz), 130.3 (d, J_C,P= 160.0 Hz), 124.8 (d, J_C,P= 14.7 Hz),

122.7 (q, J_C,F= 286.8 Hz), 122.5 (q,¹J_C,F= 290.5 Hz), 81.3 (sept,

1

²J_C,F= 31.2 Hz), 39.2 (d, J_C,P= 115.8 Hz), 25.3 (d, J_C,P

=

1

3

7.4 Hz), 23.6 (d, J_C,P= 23.9 Hz), 13.2 ppm. ¹⁹F NMR (CDCl₃): δ

2

= –75.1 (q, ⁴J_F,F= 9.3 Hz, 6 F), –75.4 (q, ⁴J_F,F= 9.3 Hz, 6 F) ppm.

³¹P NMR (CDCl₃): δ = –18.8 ppm. C₂₂H₁₇F₁₂O₂P (572.32): calcd. direct methods and refined by full-matrix least-squares methods

C 46.17, H 2.99; found C 46.19, H 2.84. 6b: M.p. 115 °C. ¹H NMR

with the teXsan program.^[44]All non-hydrogen atoms were refined

(CDCl₃): δ = 7.75–7.69 (m, 2 H), 7.67–7.65 (m, 2 H), 7.59–7.56 (m, with anisotropic thermal parameters. All hydrogen atoms in 6c and

2

3

4 H), 2.48 (dt, J_H,P= 17.1, J_H,H= 6.8 Hz, 2 H), 1.67–1.54 (m, 2 the hydrogen atoms directly attached to P1 and O2 were located

3

H), 1.39–1.29 (m, 2 H), 0.83 (t, J_H,H= 7.3 Hz, 3 H) ppm. ¹³C from a difference Fourier map calculated at the final or at a late

1

2

NMR (CDCl₃): δ = 135.9 (br. d, J_C,P= 88.3 Hz), 134.2 (d, J_C,Pstage of the structural analysis. All other hydrogen atoms were in-

3

2

= 16.5 Hz), 132.2, 132.0 (d, J_C,P= 11.1 Hz), 130.3 (d, J_C,P

=

cluded in the refinement at calculated positions (C–H = 1.0 Å) ri-

ding on their carrier atoms with isotropic thermal parameters.

CCDC-258229 (6c), -258230 (10a), -258231 (10b), -258232

(10c_intra), and -258233 (10c_inter) contain the supplementary crystal-

11.1 Hz), 125.5 (d, ³J_C,P= 11.0 Hz), 122.5 (d, ³J_C,P= 7.4 Hz), 122.5

1

2

(q, J_C,F= 286.8 Hz), 122.0 (q, J_C,F= 288.8 Hz), 79.9 (sept, J_C,F

= 31.3 Hz), 40.2 (d, ¹J_C,P= 114.0 Hz), 24.7 (d, ³J_C,P= 7.4 Hz), 24.0

(d, J_C,P= 20.2 Hz), 13.3 ppm. ¹⁹F NMR (CDCl₃): δ = –74.9 (q, lographic data for this paper. These data can be obtained free of

2

⁴J_F,F= 8.5 Hz, 6 F), –76.2 (q, J_F,F= 8.5 Hz, 6 F) ppm. ³¹P NMR charge from The Cambridge Crystallographic Data Centre via

4

(CDCl₃): δ = –3.5 ppm. C₂₂H₁₇F₁₂O₂P (572.32): calcd. C 46.17, H

www.ccdc.cam.ac.uk/data_request/cif.

2.99; found C 46.17, H 2.72.

Dynamic NMR Measurements: Samples of 10b and 6b (ca. 15 mg),

dissolved in freshly distilled solvent (0.5–0.6 mL), were sealed sepa-

rately in NMR tubes under N₂. ¹⁹F NMR spectra were measured

in a variable-temperature mode (error within 1 °C). The observed

temperatures were calibrated with the ¹H NMR chemical shift dif-

ference of signals of neat 1,3-propanediol (high-temperature re-

gion) and MeOH (low-temperature region). The rates and acti-

vation free energies at the coalescence point were calculated using

[TBPY-5-11]-1-(1,1-Dimethylethyl)-3,3,3Ј,3Ј-tetrakis(trifluoro-

methyl)-1,1Ј-spirobi[1,3-dihydro-2,1λ⁵-benzoxaphosphole] (5c): A

solution of P–H_apicalphosphorane 10c (150 mg, 0.261 mmol) in o-

dichlorobenzene (10 mL) and CH₃CN (2 mL) was heated at 150 °C

for 4 d. The solution was then concentrated in vacuo. Purification

of the residue was carried out by TLC (hexane) to give 5c (R_f=

0.8, 224 mg, 80%) as a white solid. Recrystallization from hexane

gave an X-ray sample. M.p. 128 °C. H NMR (CDCl₃): δ = 8.42– the Gutowsky–Holm approximation method and the line-shape

1

3

8.38 (m, 2 H), 7.66–7.64 (m, 6 H), 1.13 (d, J_H,P= 20.0 Hz, 9 H) analysis was performed in a narrow temperature range around the

ppm. ¹³C NMR (CDCl₃): δ = 137.2 (d, J_C,P= 9.2 Hz), 136.4 (d,

coalescence point The activation enthalpies and entropies were cal-

culated according to the transition state theory of Eyring by linear

regression.

3

²J_C,P= 16.6 Hz), 132.9 (d, J_C,P= 3.7 Hz), 132.5 (d, J_C,P

=

4

1

2

3

160.0 Hz), 130.9 (d, J_C,P= 12.8 Hz), 124.1 (q, J_C,P= 14.7 Hz),

122.7 (q, ¹J_C,F= 286.7 Hz), 122.4 (q, J_C,F= 288.6 Hz), 81.5 (sept,

1

Kinetic Measurements of the Pseudorotation of 6b and 6c, and Cycli-

zation of 10b and 10c: Samples (ca. 15 mg), dissolved in freshly

distilled solvent (0.5–0.6 mL), were separately sealed in NMR tubes

under N₂. For pseudorotation of 6b and cyclization of 10b and

10c (in pyridine), ¹⁹F NMR spectra were measured in a variable-

temperature mode, and the specified temperatures were maintained

throughout each set of measurements (error within 1 °C). For the

1

²J_C,F= 31.3 Hz), 40.5 (d, J_C,P= 103.0 Hz), 28.2 ppm. ¹⁹F NMR

4

(CDCl₃): δ = –72.4 (q, J_F,F= 10.1 Hz, 6 F), –74.5 (q, J_F,F

=

10. 1 Hz, 6 F) ppm. ^{3 1}P NMR (CDCl₃): δ = –9 . 8 ppm.

C₂₂H₁₇F₁₂O₂P: calcd. C 46.17, H 2.99; found C 46.04, H 2.73.

[TBPY-5-12]-1-(1,1-Dimethylethyl)-3,3,3Ј,3Ј-tetrakis(trifluoro-

methyl)-1,1Ј-spirobi[1,3-dihydro-2,1λ⁵-benzoxaphosphole] (6c): P–

232

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

The Ligand-Exchange Process of P–H_apicalPhosphoranes

FULL PAPER

cyclization of 10c in solvents other than pyridine, the ³¹P NMR

spectra were monitored in a broad-band decoupling mode. Inte-

grals of ³¹P signals were confirmed to be proportional to the

amount of compounds by ¹⁹F NMR spectroscopy. For the observa-

tion of the pseudorotation of 6c, ¹⁹F NMR spectra were measured

at room temp. after heating samples at specified temperatures for

arbitrary periods of time. The temperatures were maintained with

a silicon oil bath (error within 2 °C). The data for those other

than 10b in pyridine, were analyzed by assuming irreversible first-

order kinetics using the equation ln (c₀/c) = kT, in which c₀= con-

centration of reactant at t = 0 and c = concentration of reactant at

arbitrary intervals. For 10b in pyridine, analysis was carried out by

assuming a combination of irreversible first-order competitive and

consecutive reactions (see text).

R. S. McDowell, A. Streitwieser, Jr., J. Am. Chem. Soc. 1985,

107, 5849–5855; c) P. Wang, Y. Zhang, R. Glaser, A. E. Reed,

P. v. R. Schleyer, A. Streitwieser, J. Am. Chem. Soc. 1991, 113,

55–64; d) G. R. J. Thatcher, A. S. Campbell, J. Org. Chem.

1993, 58, 2272–2281.

R. R. Holmes, J. Am. Chem. Soc. 1978, 100, 433–446.

H. Wasada, K. Hirao, J. Am. Chem. Soc. 1992, 114, 16–27.

P. Wang, Y. Zhang, R. Glaser, A. Streitwieser, P. v. R. Schleyer,

J. Comput. Chem. 1993, 14, 522–529.

O. Lösking, H. Willner, H. Oberhammer, J. Grobe, D. L. Van,

Inorg. Chem. 1992, 31, 3423–3427 and references cited therein.

a) R. K. Oram, S. Trippett, J. Chem. Soc., Perkin Trans. 1 1973,

1300–1310; b) S. A. Bone, S. Trippett, M. W. White, P. J. Whit-

tle, Tetrahedron Lett. 1974, 15, 1795–1798.

a) N. V. Timosheva, T. K. Prakasha, A. Chandrasekaran, R. O.

Day, R. R. Holmes, Inorg. Chem. 1995, 34, 4525–4526; b) N. V.

Timosheva, A. Chandrasekaran, T. K. Prakasha, R. O. Day,

R. R. Holmes, Inorg. Chem. 1996, 35, 6552–6560.

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

Supporting Information (see footnote on first page of this article):

Crystallographic data for 10a, 10b, 10c_inter, 10c_intra, and 6c, and

ORTEP drawings of 10a, 10b, and 6c.

a) T. Kawashima, T. Soda, K. Kato, R. Okazaki, Phosphorus

Sulfur Silicon Relat. Elem. 1996, 109–110, 489–492; b) T. Ka-

washima, T. Soda, K. Kato, R. Okazaki, Angew. Chem. Int.

Ed. Engl. 1996, 35, 1096–1098.

Acknowledgments

a) I. Granoth, J. C. Martin, J. Am. Chem. Soc. 1979, 101, 4618–

4622; I. Granoth, J. C. Martin, J. Am. Chem. Soc. 1979, 101,

4623–4626; b) E. F. Perozzi, R. S. Michalak, G. D. Figuly,

W. H. Stevenson III, D. B. Dess, M. R. Ross, J. C. Martin, J.

Org. Chem. 1981, 46, 1049–1053; c) M. R. Ross, J. C. Martin,

J. Am. Chem. Soc. 1981, 103, 1234–1235; d) W. H. Steven-

son III, S. Wilson, J. C. Martin, W. B. Farnham, J. Am. Chem.

Soc. 1985, 107, 6340–6352; e) J. C. Martin, Science 1983, 221,

509–514; f) S. K. Chopra, J. C. Martin, Heteroat. Chem. Het-

eroat. Chem. Heteroatom Chem. 1991, 2, 71–79; g) C. D. Moon,

S. K. Chopra, J. C. Martin, in Phosphorus Chemistry, Develop-

ments in American Science (Eds.: E. N. Walsh, E. J. Griffiths,

R. W. Parry, L. D. Quin), ACS Symposium Series 486, Ameri-

can Chemical Society, Washington, DC, 1992, pp. 128–136.

For preliminary communications, see: a) K. Kajiyama, S. Ko-

jima, K.-y. Akiba, Tetrahedron Lett. 1996, 37, 8409–8412; b) S.

Kojima, K. Kajiyama, M. Nakamoto, K.-y. Akiba, J. Am.

Chem. Soc. 1996, 118, 12866–12867.

a) K. Kajiyama, M. Yoshimune, M. Nakamoto, S. Matsukawa,

S. Kojima, Y. Yamamoto, K.-y. Akiba, Org. Lett. 2001, 3,

1873–1875; b) S. Kojima, M. Sugino, M. Nakamoto, S. Matsu-

kawa, K.-y. Akiba, J. Am. Chem. Soc. 2002, 124, 7674–7675;

c) S. Matsukawa, S. Kojima, K. Kajiyama, M. Nakamoto, Y.

Yamamoto, K.-y. Akiba, S.-Y. Re, S. Nagase, J. Am. Chem.

Soc. 2002, 124, 13154–13170.

a) S. Kojima, K. Kajiyama, K.-y. Akiba, Tetrahedron Lett.

1994, 35, 7037–7040; b) S. Kojima, K. Kajiyama, K.-y. Akiba,

Bull. Chem. Soc. Jpn. 1995, 68, 1785–1797.

The authors are grateful to Central Glass Co. Ltd. for a generous

gift of hexafluorocumyl alcohol. Partial support of this work

through Grants-in-Aid for Scientific Research (nos. 06740488 and

07740501) provided by the Ministry of Education, Science, and

Culture of the Japanese Government is heartily acknowledged.

[1] For the N-X-L designation, see: C. W. Perkins, J. C. Martin,

A. J. Arduengo, W. Lau, A. Alegria, J. K. Kochi, J. Am. Chem.

Soc. 1980, 102, 7753–7759.

[2] a) F. H. Westheimer, Acc. Chem. Res. 1968, 1, 70–78; b) G. R. J.

Thatcher, R. Kluger, Adv. Phys. Org. Chem. 1989, 25, 99–265

and references cited therein.

[18]

[19]

[3] R. S. Berry, J. Chem. Phys. 1960, 32, 933–938.

[4] a) R. R. Holmes, Pentacoordinated Phosphorus − Structure and

Spectroscopy, ACS Monographs 175 and 176, American Chem-

ical Society, Washington, DC, 1980, vols. I and II; b) N. A.

Polezhaeva, R. A. Cherkasoz, Usp. Khim. 1985, 54, 1899–1939;

c) L. N. Markovskii, N. P. Kolesnik, Y. G. Shermolovich, Usp.

Khim. 1987, 56, 1564–1583; d) R. Burgada, R. Setton, in The

Chemistry of Organophosphorus Compounds (Ed.: F. R. Hart-

ley), Wiley-Interscience, Chichester, 1994, pp. 185–272; e) K.-

y. Akiba, Chemistry of Hypervalent Compounds, Wiley-VCH,

New York, 1999.

[20]

[21]

[5] a) G. C. Pimentel, J. Chem. Phys. 1951, 19, 446–448; b) R. J.

Hach, R. E. Rundle, J. Am. Chem. Soc. 1951, 73, 4321–4324;

c) J. I. Musher, Angew. Chem. Int. Ed. Engl. 1969, 8, 54–68.

[6] E. L. Muetteries, W. Mahler, R. Schmutzler, Inorg. Chem. 1963,

2, 613–618.

[7] a) S. Trippett, Phosphorus Sulfur 1976, 1, 89–98; b) M. Eisen-

hut, H. L. Mitchell, D. D. Traficante, R. J. Kaufman, J. M.

Deutsch, G. M. Whitesides, J. Am. Chem. Soc. 1974, 96, 5385–

5397; c) C. G. Moreland, G. O. Doak, L. B. Littlefield, N. S.

Walker, J. W. Gilje, R. W. Braun, A. H. Cowley, J. Am. Chem.

Soc. 1976, 98, 2161–2165; d) G. Buono, J. R. Llinas, J. Am.

Chem. Soc. 1981, 103, 4532–4540; e) L. V. Griend, R. G. Cavell,

Inorg. Chem. 1983, 22, 1817–1820.

[8] a) S. Kojima, M. Nakamoto, K. Kajiyama, K.-y. Akiba, Tetra-

hedron Lett. 1995, 36, 2261–2264; b) S. Kojima, M. Nakamoto,

K. Yamazaki, K.-y. Akiba, Tetrahedron Lett. 1997, 38, 4107–

4110; c) M. Nakamoto, S. Kojima, S. Matsukawa, Y. Yamam-

oto, K.-y. Akiba, J. Organomet. Chem. 2002, 643–644, 441–452;

d) S. Matsukawa, K. Kajiyama, S. Kojima, S.-y. Furuta, Y.

Yamamoto, K.-y. Akiba, Angew. Chem. Int. Ed. 2002, 41,

4718–4722.

For other examples of P–H_apicalphosphoranes see: a) G.-V.

Röschenthaler, R. Bohlen, D. Z. Schomburg, Z. Naturforsch.,

Teil B 1985, 40, 1593–1596; b) R. A. Kemp, Phosphorus Sulfur

Silicon, Relat. Elem. 1994, 87, 83–92.

a) K.-y. Akiba, H. Nakata, Y. Yamamoto, S. Kojima, Chem.

Lett. 1992, 1559–1562; b) S. Kojima, Y. Doi, M. Okuda, K.-y.

Akiba, Organometallics 1995, 14, 1928–1937.

a) G. Trinquier, J.-P. Daudey, G. Caruana, Y. Madaule, J. Am.

Chem. Soc. 1984, 106, 4794–4799; b) J. Moc, K. Morokuma,

Inorg. Chem. 1994, 33, 551–560.

[22]

[23]

[24]

[25]

A. Bondi, J. Phys. Chem. 1964, 68, 441–451.

For reviews on through-space coupling, see: a) J. Hilton, L. H.

Sutcliffe, Prog. Nucl. Magn. Reson. Spectrosc. 1975, 10, 27–

39; b) R. H. Contreras, M. A. Natiello, G. E. Scuseria, Magn.

Reson. Rev. 1985, 9, 239–321; c) R. H. Contreras, J. C. Facelli,

Annu. Rep. NMR Spectrosc. 1993, 27, 255–356.

H. S. Gutowsky, C. H. Holm, J. Chem. Phys. 1956, 25, 1228–

1234.

[26]

[27]

[9] a) H. J. Bestmann, J. Chandrasekhar, W. G. Downey, P. v. R.

Schleyer, J. Chem. Soc., Chem. Commun. 1980, 978–979; b)

T. L. Windus, M. S. Gordon, L. P. Davis, L. W. Burggraf, J.

Am. Chem. Soc. 1994, 116, 3568–3579.

Eur. J. Org. Chem. 2006, 218–234

www.eurjoc.org

233

S. Kojima, K. Kajiyama, M. Nakamoto, S. Matsukawa, K. Akiba

to H results in 12 modes of attack of which five give C_apical

FULL PAPER

[28] For example: M. H. Abraham, P. L. Grellier, in The Chemistry

of the Metal–Carbon Bond, vol. 2 (Eds.:F. Hartley, S. Patai),

John Wiley and Sons, New York, 1985, chapter 2.

,

O_apical, four give C_apical,C_apical, and three give O_apical,O_apical

isomers.

[29] a) Y. Yamamoto, Y. Takeda, K.-y. Akiba, Tetrahedron Lett.

1989, 30, 725–728; b) Y. Yamamoto, H. Fujikawa, H. Fujish-

ima, K.-y. Akiba, J. Am. Chem. Soc. 1989, 111, 2276–2283; c)

Y. Yamamoto, K. Ohdoi, X. Chen, M. Kitano, K.-y. Akiba,

Organometallics 1993, 12, 3297–3303.

[30] H. J. Cristau, F. Plenat, in The Chemistry of Organophosphorus

Compounds (Ed.: F. R. Hartley), Wiley-Interscience, Chiches-

ter, 1994, pp. 45–183.

[37]

[38]

For example, see: C. Reichardt, in Solvents and Solvent Effects

in Organic Chemistry, 2nd ed., VCH Publishers, Weinheim,

1988, pp. 121–283.

β: Solvent hydrogen-bond acceptor basicity. M. J. Kamlet, J.-

L. M. Abboud, M. H. Abraham, R. W. Taft, J. Org. Chem.

1983, 48, 2877–2887.

K. Mislow, Acc. Chem. Res. 1970, 3, 321–331.

A rough estimate would put C considerably higher in energy

than the highest-energy SP.

[39]

[40]

[31] a) B. Tangour, C. Malavaud, M. T. Boisdon, J. Barrans, Phos-

phorus Sulfur Relat. Elem. 1988, 40, 33–39; b) B. Tangour, C.

Malavaud, M. T. Boisdon, J. Barrans, Phosphorus Sulfur Sili-

con Relat. Elem. 1989, 45, 189–195; c) B. Tangour, C. Mala-

vaud, M. T. Boisdon, J. Barrans, Phosphorus Sulfur Silicon, Re-

lat. Elem. 1990, 53, 423–424 and references cited therein.

[32] a) M. Wieber, W. R. Hoos, Monatsh. Chem. 1970, 101, 776–

781; b) M. Wieber, K. Foroughi, Angew. Chem. Int. Ed. Engl.

1973, 12, 419–420; c) C. Malavaud, J. Barrans, Tetrahedron

Lett. 1975, 35, 3077–3080; d) C. Laurenco, R. Burgada, Tetra-

hedron 1976, 32, 2253–2255.

[41] For leading references involving the turnstile mechanism see:

a) I. Ugi, D. Marquarding, H. Klusacek, P. Gillespie, F. Rami-

rez, Acc. Chem. Res. 1971, 4, 288–296; b) P. Gillespie, P. Hoff-

man, H. Klusacek, D. Marquarding, S. Pfohl, F. Ramirez,

E. A. Tsolis, I. Ugi, Angew. Chem. Int. Ed. Engl. 1971, 10, 687–

715; c) F. Ramirez, I. Ugi, Adv. Phys. Org. Chem. 1971, 9, 25–

126; d) R. R. Holmes, Acc. Chem. Res. 1972, 5, 296–303; e) P.

Gillespie, F. Ramirez, I. Ugi, D. Marquarding, Angew. Chem.

Int. Ed. Engl. 1973, 12, 91–119; f) S.-K. Shih, S. D. Peyerim-

hoff, R. J. Buenker, J. Chem. Soc., Faraday Trans. 2 1979, 75,

379–389; g) P. Lemmen, R. Baumgartner, I. Ugi, F. Ramirez,

Chem. Scr. 1988, 28, 451–464.

[42] A. Furusaki, Acta Crystallogr., Sect. A 1979, 35, 220.

[43] C. Katayama, Acta Crystallogr., Sect. A 1986, 42, 19.

[44] teXsan: Single-Crystal Analysis Software, version 1.9, Molecu-

lar Strucuture Corporation, The Woodlands, Texas 77381,

USA, 1998. The program was previously available from Mac

Science Co..

[33] E_T^N: Normalized solvent polarity. C. Reichardt, in Solvents

and Solvent Effects in Organic Chemistry, 2nd ed., VCH Pub-

lishers, Weinheim, 1988, pp 339–405.

[34] DN^N: Normalized donor number. Y. Marcus, J. Solution Chem.

1984, 13, 599–624.

[35] F. Ramirez, G. V. Loewengart, A. T. Elefteria, K. Tasaka, J.

Am. Chem. Soc. 1972, 94, 3531–3536.

[36] Because of the presence of the bidentate ligand (which cannot

assume equatorial,equatorial and apical,apical configurations)

six configurations are possible excluding enantiomers, and the

requirement of oxygen attack from directions other than anti

Received: July 9, 2005

Published Online: November 2, 2005

234

www.eurjoc.org

Eur. J. Org. Chem. 2006, 218–234

Article Doi

The ligand-exchange process of P-Hapical phosphoranes and the thermal formation and pseudorotation of anti-apicophilic spirophosphoranes

DOI: 10.1002/ejoc.200500510

Source and publish data:

Authors:

Article abstract of DOI:10.1002/ejoc.200500510

Full text of DOI:10.1002/ejoc.200500510

Products guided by the article

R&D Labs maybe for 183946-54-5

Relevant to this article

Hot Product