Synthesis, X-ray geometry, and anodic behavior of tris[2-(hydroxymethyl)phenyl]phosphane

Article abstract of DOI:10.1021/jp952854p

Tris[(monohydroxymethyl)phenyl]phosphanes 5a and 5b were prepared, and the geometry of the ortho-isomer 5a was determined by X-ray diffraction. 5a was found to be propeller-like in shape, exhibiting a helical conformation with the three hydroxymethyl subs

Full text of DOI:10.1021/jp952854p

J. Phys. Chem. 1996, 100, 4323-4330
4323
Synthesis, X-ray Geometry, and Anodic Behavior of
Tris[2-(hydroxymethyl)phenyl]phosphane
Florence Chalier,† Yves Berchadsky,† Jean-Pierre Finet,† Ge´rard Gronchi,‡
Sylvain Marque,† and Paul Tordo*,†
Laboratoire de Structure et Re´actiVite´ des Espe`ces Paramagne´tiques, Case 521, Unite´ de Recherche Associe´e
n° 1412 au Centre National de la Recherche Scientifique, UniVersite´ de ProVence, AVenue Escadrille
Normandie-Niemen, 13397 Marseille Cedex 20, France, and Laboratoire d’Electrochimie de l’ENSSPICAM,
Unite´ de Recherche Associe´e n° 1410 au Centre National de la Recherche Scientifique, UniVersite´
d’Aix-Marseille III, AVenue Escadrille Normandie-Niemen, 13397 Marseille Cedex 20, France
ReceiVed: September 25, 1995X
Tris[(monohydroxymethyl)phenyl]phosphanes 5a and 5b were prepared, and the geometry of the ortho-isomer
5a was determined by X-ray diffraction. 5a was found to be propeller-like in shape, exhibiting a helical
conformation with the three hydroxymethyl substituents standing on the same side as that with the phosphorus
lone pair. Phosphanes 5a and 5b underwent electrochemical one-electron oxidation, giving rise to nonpersistent
phosphoniumyl cation radicals 5a‚+ and 5b‚+. The low value of the anodic peak potential of 5a was explained
by the fast decay of 5a‚+, which was rapidly rearranged to a phosphoranyl radical through the formation of
a strong P-O bond. The fast oxidation of this intermediate phosphoranyl radical followed by a second
cyclization led to the 1-[2-(hydroxymethyl)phenyl]spirobi[1H,3H-2,1-benzoxaphosphole] 7a, which was isolated
in significant yield.
Introduction
SCHEME 1
The persistence of phosphoniumyl cation radicals, derived
from triarylphosphanes, is largely dependent on the steric
hindrance of the aryl groups.1,2 Electronic effects of the
substituents on the radical cation half-lives have been studied
essentially in the case of through-bond inductive effects1,2 and
electron delocalization into aromatic rings.1,3 X-ray studies and
force field calculations have shown that 2,6-disubstitution on
at least two of the three aryl ligands leads to a significant
flattening of the pyramidal structure of the molecule.2,4 Then a
greater contribution of the phosphorus 3p orbital in the HOMO2,5
results in an increase of its energy level and thus in a greater
oxidizability of triarylphosphanes substituted with methyl sub-
stituents or electron-withdrawing substituents like the chlorine
atom.1,2 On the other hand, in the absence of oxygen, crowding
of the radical center makes the corresponding phosphoniumyl
cation radicals significantly persistent. However, methoxy
ortho- and para-substituted phosphanes present a very low
oxidation peak potential as well as a completely irreversible
oxidation due to the rapid evolution of the radical cations
through intramolecular processes. McEwen et al. pointed out
the faster quaternization reaction rate of triarylphosphanes
substituted in the ortho position by methoxy or dimethylamino
groups compared to their para analogues.6,7 This result was
attributed to an interaction between the 2p electrons of a
heteroatom and the incipient positive PIV center in the early
transition state. Similarly, according to the X-ray crystal
structure of tris[2-(dimethylaminomethyl)phenyl]phosphane,8
intramolecular coordination of the phosphorus PIII atom was
suggested. Furthermore, a through-space heteroatom-phos-
phorus PIII atom interaction in tris[2-(trifluoromethyl)phenyl]-
phosphane resulted in an enhanced electronic density on the
phosphorus atom, leading to an easy electron removal by vertical
ionization.9
In our investigation of the factors influencing the life span
and the structure of triarylphosphoniumyl radicals, we have
studied the anodic behavior of a series of heterosubstituted
triarylphosphanes shown in Scheme 1. In this paper, we will
report on the synthesis, structural studies and electrochemical
investigations of two new phosphanes, the tris[2- or 4-(hy-
droxymethyl)phenyl]phosphanes 5a and 5b.
Experimental Section
Chemicals. All reactions were carried out under a dry,
oxygen-free nitrogen atmosphere. The glassware was dried in
an oven at 120 °C and flushed with nitrogen after assembling.
Solvents were dried by distillation on sodium/benzophenone for
THF and on magnesium for methanol. All solvents for
syntheses, crystallization, and analyses were deoxygenated, and
the samples were stored under nitrogen. Melting points were
taken on a Bu¨chi capillary apparatus and have been left
uncorrected. NMR spectra were obtained on Varian EM 360
A (1H NMR 60 MHz), Bruker AC 100 (31P NMR 40.5 MHz),
and Bruker AM 400 X (1H NMR 400 MHz and 13C NMR 100.6
MHz) spectrometers. Unless otherwise stated, chemical shifts
are reported as δ values relative to internal and external
1
tetramethylsilane for H and 13C NMR, respectively, and to
external 85 wt % phosphoric acid for 31P NMR. Mass spectra
were performed on a Varian MAT 311 spectrometer.
Compounds 2a, 3a [31P NMR (CDCl3) δ -35.9], and 4a [31P
NMR (CDCl3) δ -21.9] were prepared using the Schiemenz
and Kaack procedure.10
* To whom all correspondence should be addressed.
†Universite´ de Provence.
‡Universite´ d’Aix-Marseille III.
Tris[2-(hydroxymethyl)phenyl]phosphane 5a. (a) Reduction
of 4a. A mixture of sodium borohydride (0.58 g, 15.3 mmol)
X Abstract published in AdVance ACS Abstracts, December 15, 1995.
0022-3654/96/20100-4323$12.00/0 © 1996 American Chemical Society

4324 J. Phys. Chem., Vol. 100, No. 10, 1996
Chalier et al.
and tris(2-formylphenyl)phosphane 4a (2 g, 5.8 mmol) in dry
methanol (100 cm3) was stirred at room temperature for 30 min.
The solution was concentrated, and water was added. The white
precipitate was filtered and dried. After recrystallization from
CH2Cl2-ethyl alcohol (95:5), tris[2-(hydroxymethyl)phenyl]-
phosphane 5a was obtained as cream-colored crystals (1.3 g,
65% yield), mp 193-195 °C. 31P NMR: δ (DMSO), -35.2;
δ (CD3OD) -35.5. 1H NMR (DMSO, 400.13 MHz) δ: 7.619
(3H, ddd, JH3-P ) 4.5 Hz, JH3H5 ) 1.0 Hz, JH3H4 ) 7.5 Hz,
H-3), 7.416 (3H, dt, JH4H3 ) JH4H5 ) 7.5 Hz, JH4H6 ) 1.3 Hz,
H-4), 7.184 (3H, dt, JH5H4 ) JH5H6 ) 7.5 Hz, JH5H3 )1.0 Hz,
H-5), 6.626 (3H, ddd, JH6-P ) 4.9 Hz, JH6H5 ) 7.5 Hz, JH6H4
) 1.3 Hz, H-6), 5.274 (3H, t, J ) 4.6 Hz, OH), and 4.548 (6H,
dd, JH-P ) 1.5 Hz, JH-OH ) 4.6 Hz, CH2). 13C NMR (DMSO,
100.6 MHz) δ: 60.87 (CH2, JCP ) 28.17 Hz), 126.27 (C-3,
JCH ) 49.79 Hz, JCP ) 5.03 Hz), 129.10 (C-4, JCH ) 39.74),
131.43 (C-1, JCP ) 12.07 Hz), 132.38 (C-6, JCH ) 49.79 Hz),
and 149.29 (C-2, JCP ) 32.43 Hz). High resolution MS
(HRMS) found, 352.1235; C21H21O3P requires 352.1228; m/e
(70 eV), 352 (6.6%, M+), 349 (3.5), 321 (21.3, M+ - CH2-
OH), 316 (28.1), 303 (64.3), 297 (13), 243 (22.1), 213 (12.5),
196 (16.5), 178 (100), 165 (56.9), 152 (15.9), 137 (19.2), 119
(11.7), 109 (13.9), 107 (23.1, C6H4CH2OH+), 91 (99.3), 89
(16.6), 77 (43.9), 65 (20.4), 63 (11.3), 51 (29.4), 47 (15.5), 39
(19.1), and 31 (23.4, CH2OH+).
organic phases were washed with 5% sodium hydrogen
carbonate aqueous solution and dried over MgSO4. After
distillation of the solvents, the crude semisolid residue was
purified by chromatography (CH2Cl2, neutral alumina) to afford
4b as cream-colored crystals (1.2 g, 55% yield), mp 114 °C,
lit.12 115-115.5 °C. 31P NMR (CDCl3) δ: -4.9 ppm.
Tris[4-(hydroxymethyl)phenyl]phosphane 5b. The reduction
of tris(4-formylphenyl)phosphane 4b (0.941 g, 0.0027 mol) in
dry methyl alcohol (20 cm3) by sodium borohydride (0.315 g,
0.0081 mol) under the conditions used for the reduction of 4a
afforded the triarylphosphane 5b (0.750 g, 79% yield) as white
crystals, mp 185 °C. Found: C, 71.72; H, 6.03. C21H21O3P
requires C, 71.58 and H, 6.01%. 31P NMR (CD3OD) δ: -4.6.
1H NMR (CD3OD, 60 MHz) δ, relative to internal sodium
3-(trimethylsilyl)propionate: 7.6-7 (12H, AA′BB′, ArH) and
4.5 (2H, s, CH2O).
X-ray Crystallographic Determinations. Colorless needles
of 5a suitable for X-ray diffraction were grown from ethyl
alcohol (95%). A crystal with approximate dimensions of 0.6
× 0.4 × 0.3 mm was mounted on a glass fiber. The space
group was assigned as P1, and the crystal system is triclinic.
Intensity data were collected at 298 K on an Enraf Nonius
CAD-4 diffractometer equipped with a graphite crystal incident
beam monochromator using the θ/2θ scan mode in the index
range 0° e 2θ e 48°, hmax ) 10, lmax ) 13.The data were
collected by the CAD-4 Enraf Nonius program.13 The data
reduction and the structure resolution were performed by the
SDP software package13 by using a direct method program
MULTAN.14 The structure was completed by Fourier difference
syntheses (10 hydrogen atoms were located in this way). Eleven
hydrogen atoms were introduced at the ideal position and were
included in the refinement calculations but were not themselves
refined. Atomic scattering and anomalous factors were taken
(b) Synthesis from 2-Bromobenzyl Alcohol. A solution of
n-butyllithium (2.5 M in hexane, 21.4 cm3, 53.4 mmol) was
added dropwise over 30 min into a cooled (-25 °C) solution
of 2-bromobenzyl alcohol (5 g, 0.0267 mol) in dry THF (50
cm3) with the temperature maintained below -20 °C. The
mixture was stirred for 2h at -20 °C. Then phosphorus
trichloride (0.70 cm3, 8.02 mmol) was added slowly at -25 °C
and the mixture was stirred for 2 h at -20 °C. After addition
of a 5% sodium hydrogen carbonate aqueous solution (20 cm3),
THF was distilled under vacuum and the solid was filtered from
the aqueous phase. A white solid (0.96 g) of oxidized products
was removed from the mixture by crystallization in CH2Cl2,
and the tris[2-(hydroxymethyl)phenyl]phosphane 5a (1.035 g,
37% yield) was crystallized from the oily yellow residue by
treatment with CH2Cl2-ethyl alcohol (95:5).
Tris[4-(1,3-dioxacyclopent-2-yl)phenyl]phosphane 3b. The
Grignard reagent prepared from 2-(4-bromophenyl)-1,3-diox-
olane11 (12.82 g, 0.0559 mol) and magnesium fine powder (1.36
g, 0.0559 mol) in dry THF (50 cm3) under sonication was cooled
at 0 °C, and a solution of phosphorus trichloride (1.22 cm3,
0.014 mol) in THF (4 cm3) was then slowly added. The mixture
was stirred for 12 h at room temperature and then was refluxed
for 3 h. After cooling at 0 °C, a saturated aqueous solution of
sodium hydrogen carbonate (10 cm3) was added. The mixture
was extracted with CH2Cl2 (100 cm3). The organic phase was
washed with water, dried on MgSO4, and concentrated under
reduced pressure. Crystallization from THF-ethanol (4:1) at
4 °C gave the title compound 3b as white crystals (3.5 g, 45%
yield), mp 149 °C. Found: C, 67.79; H, 5.72. C27H27O6P
requires C, 67.78 and H, 5.69%. 1H NMR (CDCl3, 60 MHz)
δ: 7.9-7.1 (12H, m, ArH), 5.8 (3H, s, ArCH), and 4.4-3.9
(12H, m, OCH2-CH2O). 31P NMR (CDCl3) δ: -7.1.
from International Tables for X-Ray Crystallography.15
A
sample of 2476 unique reflections was measured, and 2497
reflections with I g 3σ(I) were used in structural determination.
No correction for absorption or secondary extinction was made.
Full matrix least-squares refinement was carried out with w )
1/s2(I) and converged to R ) 0.0436, Rw ) 0.0443 for 2336
observations and 262 refined variables. Final cell constants,
as well as other information pertinent to data collection and
refinement, are listed in Table 1.
Electrochemical Studies. All the electrochemical experi-
ments were performed at room temperature in acetonitrile (SDS)
0.1 M in tetra-n-butylammonium hexafluorophosphate (sup-
porting electrolyte). Acetonitrile was refluxed over a mixture
of KMnO4 (19 g L-1) and K2CO3 (18 g L-1) for 48 h, distilled,
and then redistilled over P2O5 (2 g L-1). The supporting
electrolyte (TBAFP, Fluka, purum) was twice crystallized from
a mixture of ethyl acetate and pentane (20:80) and dried at
60 °C under reduced pressure. Collidine was freshly distilled
before use. Measurements were carried out on dry argon-purged
solutions in three electrode cells fitted with a platinum auxiliary
electrode and a Tacussel XR 110 saturated calomel electrode
as reference. This reference electrode was separated from the
test solution by a compartment containing the supporting
electrolyte solution and closed with a porous ceramic disk. The
working electrode used for both stationary and rotating disk
voltammetric measurements was a Tacussel EDI 101 T platinum
disk electrode (diameter, 2 mm). For coulometric and prepara-
tive electrolyses, the working electrode was a platinum foil (area,
2 or 12.5 cm2) and the used cell had separate anodic and cathodic
compartments. The electrolyses were periodically interrupted
to perform voltammetric monitoring at the rotating disk electrode
(rotation speed, 2500 rpm).
Tris(4-formylphenyl)phosphane 4b. A mixture of tris[4-(1,3-
dioxacyclopent-2-yl)phenyl]phosphane 3b (3 g, 0.0063 mol) and
2 N aqueous HCl (12 cm3, 0.024 mol) in THF (50 cm3) was
refluxed for 15 min and then cooled to room temperature and
extracted with CH2Cl2 (100 cm3). The aqueous phase was
neutralized with a saturated aqueous solution of sodium
hydrogen carbonate and extracted with CH2Cl2. The collected

Tris[2-(hydroxymethyl)phenyl]phosphane
J. Phys. Chem., Vol. 100, No. 10, 1996 4325
TABLE 1: Crystal Data, Data Collection, and Refinement
Parameters
SCHEME 2
molecular formula
formular weight
C21H21O3P
352.37
crystallization
space group
ethanol
P1(triclinic)
crystal size/mm
a/Å
0.6 × 0.4 × 0.3
10.418(4)
b/Å
c/Å
10.840(7)
8.908(6)
R
108.46(4)
112.21(8)
63.97(0)
899
â
γ
V/Å3
SCHEME 3
F(000)
186
2
1.302
formular unit per cell Z
density/g cm-1
data collection
radiation
Enraf-Nonius CAD-4 diffractometer
1MoKR
0.710 69
wavelength λ/Å
absorption coefficient µ/cm-1 1.632
monochromator
collection range 2θ/deg
scan width
h,k,l range
number of reflections
measured
graphite
1 e 2θ e 48
0.8 + 0.35 tan θ
-10/10; -11/11; 0/13
) 7.7 Hz, JHP unknown), H-16 (d 7.27, 1H, m, JH-Hortho ) 7.6
Hz, JH-Hmeta ) 1.5 Hz, JHP unknown)], 7.19 (1H, tdd, JH-Hortho
) 7.6 Hz, JH-Hmeta ) 1.7 Hz, JHP ) 3.3 Hz, H-17), 5.02 (2H,
ABX, JHA-HB ) 14.1 Hz, JHA-P ) 2.9 Hz, HA of ArCH2OP),
4.90 (2H, ABX, JHB-HA ) 14.2 Hz, JHB-P ) 4.7 Hz, HB of
ArCH2OP), 4.49 and 4.15 (2H, AB, JHA-HB ) 11.9 Hz, ArCH2-
OH), and 5.26 (s, 1H, OH). 13C NMR (CDCl3, 100.6 MHz) δ:
2865
unique
2476
2497 (>3σ)
262
used in refinement
number of variables
R and Rw
weight w
goodness of fit
max shift
17
66.16 (d, JCP ) 6.5 Hz, C-21), 66.59 (s, C-19 and C-20),
0.0436; 0.0443
122.90 (d, JCP ) 16 Hz, C-3 and C-9), 127.35 (d, JCP ) 16 Hz,
C-17), 128.19 (d, JCP ) 13.1 Hz, C-5 and C-11), 129.17 (d,
JCP ) 160 Hz, C-1 and C-7), 129.17 (d, JCP ) 3 Hz, C-16),
129.49 (d, JCP ) 12 Hz, C-18), 130.93 (d, JCP ) 14 Hz, C-15),
132.51 (d, JCP ) 3 Hz, C-4 and C-10), 134.91 (d, JCP ) 11 Hz,
1/σ2
0.636
0.01
H atoms
10 hydrogen atoms were localized
on a Fourier map and refined. The
11 other hydrogen positions were
calculated, not refined.
C-2 and C-6), 138.51 (d, JCP ) 10 Hz, C-14), 140.83 (d, JCP
)
177 Hz, C-13), and 148.05 (d, JCP ) 26 Hz, C-2 and C-8). m/e:
350 (11.9%, M+), 349 (22), 332 (10.5), 321 (15.9), 320 (18.5),
319 (100), 303 (11.7), 244 (25.4), 243 (53.8), 228 (30.2), 213
(12.4), 179 (40.9), 178 (32.2), 166 (15.8), 165 (34.1), 137 (12.1),
91 (27.8), 89 (12.3), 78 (10.7), 77 (29.2), 65 (10), 63 (13.4), 57
(13.3), 55 (10.4), 51 (12.7), 47 (11.3), 43 (18.5), 39 (13.4), 31
(13.4), 29 (11.8), and 28 (14.8).
A custom-built potenstiostat and data acquisition system18
was used for cyclic voltammetry. Measurements at the rotating
disk electrode were performed with a Tacussel PRG5 apparatus.
The electrolyses were carried out at constant potential using a
Tacussel PRT 100-1X potentiostat and a Tacussel IG5 electronic
integrator.
Results and Discussion
Preparation of 7a by Electrolysis of 5a. After electrolysis
at 1.1 V/SCE of tris[2-(hydroxymethyl)phenyl]phosphane in the
presence of three equivalents of collidine, 90% of the solvent
was evaporated. The precipitate of tetra-n-butylammonium
hexafluorophosphate was filtered, and the solvent was then
distilled to dryness. From the solid residue, the remaining
electrolyte support was removed by crystallization in pentane-
toluene and then by filtration. The liquid phase was neutralized
with 5% acetic acid aqueous solution and dried on MgSO4.
Solvents were evaporated, and the last traces of acetic acid were
removed by distillation under reduced pressure (1 mmHg). 1-[2-
(Hydroxymethyl)phenyl]spirobi[1H,3H-2,1-benzoxaphos-
phole] was isolated as a cream-colored solid (75 mg from 109
mg of electrolyzed triarylphosphane, 69% yield), mp 128 °C.
Found: C, 71.76; H, 5.38. C21H19O3P requires C, 71.99; H,
5.47%. HRMS found, 350.1058. C21H19O3P requires 350.1072.
IR frequencies (cm-1): 3567 (free O-H), 3450 (O-H), 1081
and 1075 (P-O), 1058 (C-O). 31P NMR (CDCl3) δ: -32.0.
Synthesis of the Triarylphosphanes. The ortho-hydroxy-
methyl substituted triarylphosphane 5a and the para-substituted
isomer 5b were prepared from the respective bromobenzalde-
hyde in a four-step sequence as indicated in Scheme 2. The
tris(2-formylphenyl)phosphane 4a has already been described
by Schiemenz and Kaack.10 Its reduction by sodium borohy-
dride led to the tris[2-(hydroxymethyl)phenyl]phosphane 5a,
which was obtained in 43% overall yield from bromobenzal-
dehyde 1a. By use of the same sequence, the tris[4-(hydroxy-
methyl)phenyl]phosphane 5b was obtained in a 19% overall
yield from bromobenzaldehyde 1b. A more direct access to
phosphane 5a is a modification of the sequence used by Dahl
et al.19 in the synthesis of 3H-2,1-benzoxaphosphole. The
lithium salt of 2-lithiobenzyl alcohol was prepared by reaction
of 2-bromobenzyl alcohol with n-butyllithium. It was then treat-
ed with phosphorus trichloride, which led, after work-up, to 5a
in 37% yield (Scheme 3). The tris(formylphenyl)phosphanes
4a and 4b are relatively acid-sensitive. Indeed, deprotection
of the dioxolanyl derivatives 3a and 3b by acid-catalyzed
hydrolysis required a good control of the reaction conditions.
For example, in the deprotection of 3a catalyzed by p-toluene-
sulfonic acid in acetone, the yield of 4a dropped from 93% to
38% when the reaction time was increased by 1 h. Moreover,
when the deprotection was attempted with 2 N aqueous HCl
1H NMR (CDCl3, 400.13 MHz) δ: 8.31 (2H, dd, JH-Hortho
7.7 Hz, JHP ) 10 Hz, H-6, and H-12), 7.52 (2H, tt, JH-Hortho
)
)
7.7 Hz, JH-Hmeta ) JHP ) 1.5 Hz, H-4, and H-10), 7.45 (2H, br
q, JH-Hortho ) 7.7 Hz, JHP ) 6 Hz, H-5, and H-11), 7.39 (1H,
br dd, JH-Hortho ) 7.6 Hz, JHP ) 17.4 Hz, H-18), 7.33-7.23 16
[m, 4H, H-15 (d 7.30, 1H, m, JH-Hortho ) 7.6 Hz, JH-Hmeta
)
1.4 Hz, JHP unknown), H-3 and H-9 (d 7.28, 2H, m, JH-Hortho

4326 J. Phys. Chem., Vol. 100, No. 10, 1996
Chalier et al.
TABLE 2: Selected Bond Lengths and Angles of Tris[2-(hydroxymethyl)phenyl]phosphane
Bond Lengths (Å)
P-C(1)
1.836(2)
1.846(3)
1.836(3)
1.406(4)
1.392(4)
1.388(3)
1.373(5)
1.384(6)
1.386(3)
O(23)-C(20)
O(24)-C(21)
O(25)-C(22)
C(8)-C(9)
1.370(4)
1.442(3)
1.423(3)
1.398(5)
1.389(3)
1.398(4)
1.377(4)
1.373(6)
1.384(5)
C(2)-C(20)
C(9)-C(21)
C(15)-C(22)
C(14)-C(15)
C(14)-C(19)
C(15)-C(16)
C(16)-C(17)
C(17)-C(18)
C(18)-C(19)
1.512(5)
1.498(3)
1.507(2)
1.399(3)
1.397(3)
1.390(5)
1.379(4)
1.365(5)
1.386(3)
P-C(8)
P-C(14)
C(1)-C(2)
C(1)-C(6)
C(2)-C(3)
C(4)-C(5)
C(3)-C(4)
C(5)-C(6)
C(8)-C(13)
C(9)-C(10)
C(10)-C(11)
C(11)-C(12)
C(12)-C(13)
Bond Angles (deg)
C(1)-P-C(8)
102.4(1)
C(1)-P-C(14)
102.8(1)
C(8)-P-C(14)
101.5(1)
119.2(2)
121.8(2)
119.0(2)
118.9(3)
120.8(3)
121.4(4)
119.9(4)
120.0(3)
122.2(3)
118.9(3)
110.4(3)
P-C(1)-C(2)
119.3(2)
121.5(2)
118.9(2)
119.2(3)
121.1(3)
121.0(3)
120.1(3)
119.8(3)
121.3(2)
119.5(3)
111.4(2)
P-C(8)-C(9)
119.5(2)
121.3(3)
119.1(2)
118.9(2)
121.1(3)
120.9(3)
120.2(3)
119.7(3)
122.1(3)
118.9(3)
111.2(2)
P-C(14)-C(15)
P-C(1)-C(6)
P-C(8)-C(13)
P-C(14)-C(19)
C(2)-C(1)-C(6)
C(1)-C(2)-C(3)
C(1)-C(6)-C(5)
C(2)-C(3)-C(4)
C(3)-C(4)-C(5)
C(4)-C(5)-C(6)
C(1)-C(2)-C(20)
C(3)-C(2)-C(20)
O(23)-C(20)-C(2)
C(9)-C(8)-C(13)
C(8)-C(9)-C(10)
C(8)-C(13)-C(12)
C(9)-C(10)-C(11)
C(10)-C(11)-C(12)
C(11)-C(12)-C(13)
C(8)-C(9)-C(21)
C(10)-C(9)-C(21)
O(24)-C(21)-C(9)
C(15)-C(14)-C(19)
C(14)-C(15)-C(1)
C(14)-C(19)-C(18)
C(15)-C(16)-C(17)
C(16)-C(17)-C(18)
C(17)-C(18)-C(19)
C(14)-C(15)-C(22)
C(16)-C(15)-C(22)
O(25)-C(22)-C(15)
TABLE 3: Comparative Structural Features, from X-ray
Data, of Selected Triarylphosphanes and of the
Tris[o-(hydroxymethyl)phenyl]phosphane
Figure 1. ORTEP diagram of the molecular structure of phosphane
5a.
P-C bonds). The pyramidalization angle R is generally
governed by repulsive nonbonding interactions between the Ar
groups. In the case of the tris(2,4,6-trimethylphenyl)phos-
phane,23 an important flattening was observed (R ) 19.24°)
compared to triphenylphosphane (R ) 25.52°). On the other
hand, tris(2-tolyl)phosphane presented a barely different R angle
(R ) 25.69°). In phosphane 5a, this R angle was even slightly
larger (R ) 26.02°).
For 5a, each of the distances between the phosphorus atom
and the three oxygen atoms (3.555-4.437 Å) is larger than the
sum of the Van der Waals radii of phosphorus and oxygen (3.32
Å).24 By contrast, because of intramolecular coordination, each
of the distances between the phosphorus atom and the three
nitrogen atoms (2.999-3.071 Å) in tris[2-(dimethylamino)-
phenyl]phosphane is smaller than the sum of the Van der Waals
radii of phosphorus and nitrogen (3.4 Å).8
solution in THF for 15 min at 60 °C, a 50% yield of 4a was
obtained. In the case of the phosphane 4b, deprotection was
only performed to completion with HCl catalysis. Only a
mixture of partly deprotected phosphanes was observed in the
reaction with p-toluenesulfonic acid in acetone. This fact
explains the lower overall yield in the para-substituted series.
X-ray Crystallography of Phosphane 5a. Triarylphos-
phanes adopt a propeller-like conformation.20 The X-ray data
obtained for the phosphane 5a are reported in Tables 2 and 3,
and the ORTEP21 view of the molecular structure in Figure 1
shows that 5a adopts a helicoidal geometry, with the three
hydroxymethyl groups standing on the side of the phosphorus
lone pair. The general structural features of 5a are relatively
close to those of the tris(2-tolyl)phosphane22 or the
triphenylphosphane.20 The mean P-C bond length (1.839 Å)
is very close to the value found for the tris(2-tolyl)phosphane
(1.835 Å). However, in the case of 5a, one of the three P-C
bonds is slightly longer (1.846 and 1.836 Å for the two other
The X-ray study of the phosphane 5a thus indicated that there
is no specific difference in the steric effects between a
hydroxymethyl or a methyl group and that the phosphorus atom
does not expand its coordination number by intramolecular
coordination with the three oxygen atoms.
Electrochemical Studies. In cyclic voltammetry at a station-
ary platinum electrode, the anodic oxidation of 5a in acetonitrile

Tris[2-(hydroxymethyl)phenyl]phosphane
J. Phys. Chem., Vol. 100, No. 10, 1996 4327
SCHEME 4
containing 0.1M tetra-n-butylammonium hexafluorophosphate
exhibited an anodic peak potential Ep ) 0.89 V/SCE for a
a
potential scan rate of 0.1 V s-1. Under the same conditions,
the para-substituted phosphane 5b presented a higher anodic
peak potential Ep ) 1.29 V/SCE. For both phosphanes, the
a
absence of a reverse peak, in the 0.1-100 V s-1 range of
potential scan rates, demonstrated the low persistence of the
primary product of the oxidation process. No electron para-
magnetic resonance (EPR) signal could be detected when the
oxidation of 5a or 5b was performed electrochemically within
the cavity of an EPR spectrometer or by γ-irradiation of a freon
matrix25 doped with 5a or 5b. On the other hand, the difference
in the anodic peak potential values for the two phosphanes
clearly demonstrated their different electrochemical behavior
although the inductive electronic effects of their substituents
are similar. The phosphane 5b could be integrated into the
isosteric series of triarylphosphanes bearing one or no methyl
substituent on the ortho-position of each phenyl ring, for which,
in our previous studies,we had found a linear correlation between
Ep and the ∑σ+ Hammett parameter.1,2,26 However, 5a could
not be integrated in this correlation and the ortho-hydroxymethyl
substituent appeared to have a dramatic influence on its anodic
potential value.
Figure 2. Influence of the presence of collidine on the voltammetric
wave of 5a with the following molar ratios of collidine/5a: (a) 0.0;
(b) 0.33; (c) 0.66; (d) 1.0; (e) 1.33; (f) 1.66; (g) 2.0-3.0.
The apparent number of electrons involved in the electron
transfer for the oxidation of 5a can be estimated by comparison
to the rapid monoelectronic oxidation of the tris(2,4,6-tri-
methylphenyl)phosphane (TMP) under the same conditions,
assuming that the diffusion coefficients for TMP and 5a are
not too different.27 For potential scan rates ranging from 0.1
to 10 V s-1, the average ratio of the anodic peak currents was
Figure 3. Voltammetric studies of 5a at the rotating disk electrode in
the presence of perchloric acid with the following molar ratios of acid/
5a: (a) 0.0; (b) 0.35; (c) 0.53; (d) 0.70; (e) 0.88; (f) 1.06.
concentration of collidine up to a constant value reached for a
molar ratio [collidine]/[5a] of 3 (Figure 2). At this point, the
measured current was 2.76 times the current obtained in the
absence of collidine.
Ip (5a)/Ip (TMP) ) 0.64
a
a
Under nitrogen atmosphere, in acetonitrile containing 0.1 M
TBAFP, a preparative electrolysis was conducted, without
collidine, at a potential corresponding to the limiting current of
the 5a anodic wave. The progress of the electrolysis was
monitored by voltammetry at the rotating electrode. A regular
decrease of the oxidation wave of 5a at E1/2 ) + 0.75 V/SCE
was observed together with the appearance and increase of a
reduction wave at E1/2 ) -0.25 V/SCE. Integration of the
electrolysis current value gave a total apparent number of 1.1
electrons exchanged per molecule.
Attribution of the cathodic wave to the reduction of 6a
(Scheme 4), which is generated during the electrolysis, is
supported by the results of voltammetric measurements on 5a
in acidic medium (Figure 3). When a dilute solution of HClO4
in acetonitrile was added to the voltammetric solution of 5a, a
wave appeared at E1/2 ) -0.25 V/SCE, the limiting current of
which varied linearly with the amount of HClO4. A corre-
sponding decrease of the intensity of the oxidation wave of 5a
was also observed.
This value indicates that the primary step of the electrode
reaction is a monoelectronic oxidation. The resultant cation
radical is then quickly consumed in a chemical reaction. The
small magnitude of the ratio of anodic peak currents (0.64) could
imply that the phosphane 5a diffusing to the electrode reacts
partly with the chemical species generated after the primary
oxidation step. In contrast, such a behavior is excluded in the
case of TMP in view of the high persistence of the corresponding
phosphoniumyl radical cation TMP‚+.
If protons are generated during the chemical steps, as
frequently observed in oxidation reactions, protonation of the
phosphane 5a would lead to the hydrophosphonium cation 6a
(Scheme 4), and this could explain the apparent decay in
phosphane concentration and the low value of the ratio
Ip (5a)/Ip (TMP). In order to test this hypothesis, the phosphane
a
a
5a was oxidized in the presence of the slightly nucleophilic
2,4,6-trimethylpyridine (collidine). Under the conditions of our
experiments, oxidation of collidine occurs at a potential peak
of 2.44 V/SCE, thus avoiding any interference with the oxidation
of 5a. In the presence of collidine, the cyclic voltammogram
of 5a showed a greater amplitude of the oxidation peak. In
this case, collidine competes with the phosphane as a proton
acceptor, and the quantity of nonprotonated 5a reaching the
electrode is therefore increased.
When a preparative electrolysis was performed on 5a in the
presence of collidine, with a molar ratio [collidine]/[5a] of 3
(Figure 4), the oxidation wave of 5a decreased during the course
of electrolysis. A plot of the voltametric current I toward the
electric-charge-consumed Q is shown on Figure 5. Extrapola-
tion at I ) 0 for the value of the voltammetric current gave a
total apparent number of 1.8 exchanged electrons per molecule.
In voltammetry at the rotating electrode, the limiting current
corresponding to the oxidation wave of 5a increased with the

4328 J. Phys. Chem., Vol. 100, No. 10, 1996
Chalier et al.
SCHEME 5
Figure 4. Selected voltammetric curves recorded during the progress
of the electrolysis of 5a (0.338 mmol) in the presence of an excess of
collidine (molar ratio 3:1): (a) residual current; (b) residual current
with added collidine; (c) Q ) 0 C; (d) Q ) 8.1 C; (e) Q ) 16.2 C; (f)
Q ) 24.6; (g) Q ) 32.4 C; (h) Q ) 40.5 C; (i) Q ) 48.6 C; (j) Q )
56.7 C.
SCHEME 6
Figure 5. Evolution of the voltammetric currents during the electrolysis
of 5a in the presence of collidine (3:1): (a) limiting current of 5a; (b)
limiting current of collidinium (absolute value); (c) sum.
A cathodic wave appeared at E1/2 ) -0.95 V/SCE, attributed
to the reduction wave of the collidinium cation generated during
the electrolysis by protonation of the collidine. There was no
evidence for the presence of 6a since no wave near -0.25
V/SCE was present.
The electrochemical behavior of 5a in the absence of collidine
or in the presence of a sufficient amount of collidine can be
explained by the mechanisms described in Schemes 5 and 6. In
both cases, the first step is the monoelectronic oxidation of 5a
leading to the cation radical 5a‚+. Intramolecular reaction
between the positively charged phosphorus center and the
oxygen of the ortho-hydroxymethyl substituent of one aromatic
ligand creates a strong P-O bond together with the release of
a proton. The proton is trapped either by the nonoxidized 5a
diffusing to the electrode (Scheme 5) or by collidine (Scheme
6), leading to the corresponding phosphonium or ammonium
cation. The loss of the proton generates a cyclic phosphoranyl
radical in the vicinity of the electrode, and this neutral radical,
which is probably easily oxidized, contributes to the electron
transfer at the actual working potential (ECE mechanism).
Alternatively, the cyclic phosphoranyl radical could react with
the phosphoniumyl cation radical in a homogeneous electron
transfer reaction as presented in Scheme 7 (DISP mechanism).28
The ensuing cationic species undergoes a second intramolecular
attack, leading to the stable 1-[2-(hydroxymethyl)phenyl]spirobi-
[1H,3H-2,1-benzoxaphosphole] 7a with the release of a second
SCHEME 7
proton. In both cases, either an ECE (Scheme 5 and 6) or a
DISP mechanism, the same global reaction balance is obtained.
In the absence of collidine, two electrons are transferred for
the consumption of three molecules of 5a and the formation of

Tris[2-(hydroxymethyl)phenyl]phosphane
J. Phys. Chem., Vol. 100, No. 10, 1996 4329
TABLE 4: Calculation of Ratio between Oxidized Amount
of 5a and the Amount of Collidinium (∆Ccol/∆C5a) Formed
during the Electrolysis of 5a
d
e
C5a
∆Qa (C) Ilab (µA) Ilac (µA) (mol L-1
Ccol
(mol L-1
)
)
∆Ccol/∆C5a
0.00
4.05
8.10
101.0
94.5
88.0
81.0
73.5
66.0
58.5
52.0
45.0
38.0
31.0
24.5
17.5
10.0
0.0 1.47 × 10-3 0.00
7.5 1.37 × 10-3 1.72 × 10-4
16.5 1.28 × 10-3 3.78 × 10-4
25.5 1.18 × 10-3 5.84 × 10-4
34.5 1.07 × 10-3 7.90 × 10-4
44.0 9.60 × 10-4 1.01 × 10-3
53.5 8.51 × 10-4 1.22 × 10-3
61.5 7.57 × 10-4 1.41 × 10-3
71.0 6.55 × 10-4 1.63 × 10-3
80.0 5.53 × 10-4 1.83 × 10-3
88.5 4.51 × 10-4 2.03 × 10-3
97.0 3.56 × 10-4 2.22 × 10-3
106.5 2.55 × 10-4 2.44 × 10-3
116.0 1.46 × 10-4 2.66 × 10-3
1.72
1.99
2.01
1.98
1.98
1.97
1.98
2.00
2.00
1.99
1.99
2.01
2.01
Figure 6. Structure and numbering of the spirobi[benzoxaphosphole]
7a.
12.15
16.20
20.25
24.60
28.35
32.40
36.45
40.50
44.55
48.60
52.65
All these results led us to conclude that the electrochemical
behavior of 5a in acetonitrile and in the presence of an excess
of an organic base is correctly described by Scheme 6. When
no base is added, Scheme 5 may be used to explain the behavior
of 5a, although some mechanistic complications are likely to
occur. Since the phosphane is a less efficient base than collidine
and, moreover, its concentration decreases during the electroly-
sis, then the reversible deprotonation of the phosphonium can
increase the disponibility of 5a. The electrochemical behavior
of 5a may be called a self-deprotonation mechanism, which is
quite similar to a self-protonation mechanism. A very interest-
ing and detailed study on the kinetics of the self-protonation
reaction in organic electrochemical processes has been reported
by Amatore et al.29
a Amount of charge. b Limiting current of oxidation of 5a. c Limiting
current of reduction of collidinium. d Concentration of 5a. e Concen-
tration of collidine.
one molecule of 7a and two molecules of 6a. This is in clear
agreement with the peak current ratio [Ip (5a)/Ip (TMP) ) 0.64]
a
a
and with the appearance, during the electrolysis, of the reduction
wave attributed to 6a. However, this mechanism cannot explain
the coulometric result (n ) 1.1) obtained for the total number
of electrons exchanged. This discrepancy could be imputed to
the importance of the concentration changes during coulometry
and to the reversibility of the phosphane protonation.
It is worth considering that the phosphonium cation 6a
generated during the electrolysis is positively charged at the
phosphorus center and that an intramolecular reaction with the
side chain oxygen atoms could occur. If such a reaction would
be likely to occur at an appreciable rate, the phosphonium cation
prepared by addition of HClO4 in a solution of 5a should not
be stable. But this is not evidenced by the voltammetric
experiments, which showed constant limiting currents under
such conditions.
On the other hand, in the presence of collidine the second
mechanism described in Scheme 6 is in relatively good
agreement with the electrochemical results obtained. For each
molecule of 5a, two electrons must be exchanged, and the
experimental value is n ) 1.80. The presence of the collidinium
cation among the products of electrolysis is supported by the
appearance of the reduction wave at E1/2 ) -0.95 V/SCE.
Voltammetric studies on authentic samples of collidinium
p-toluenesulfonate and collidinium chloride indicated that, in
the concentration range from 5 × 10-4 to 2.5 × 10-3 M, the
limiting current (il) is proportional to the concentration (C) of
collidinium [for collidinium chloride C/il ) 2.29 × 10-5 mol
L-1 mA-1 and for collidinium p-toluenesulfonate C/il ) 2.24
× 10-5 mol L-1 mA,-1 under the same experimental conditions
as those of electrolysis monitoring (2 mm rotating disk electrode
diameter, 2500 rpm)].
Conclusion
According to the X-ray diffraction results, the tris[2-(hy-
droxymethyl)phenyl]phosphane 5a does not exhibit any oxygen-
PIII interaction although the three oxygen atoms are on the same
side as the phosphorus electron lone pair. In the presence of
an efficient proton trap such as collidine, electrochemical
experiments led us to conclude that the low value of the anodic
peak observed for 5a was the consequence of the occurrence
of an ECE (or DISP) global process involving the transfer of
two electrons and the release of two protons. After the first
electron transfer, the geometry of formed 5a‚+ is appropriate to
favor its fast decay through the intramolecular formation of a
P-O bond, leading to an intermediate phosphoranyl radical.
Subsequent oxidation of this phosphoranyl followed by a second
cyclization step finally gave rise to the 1-[2-(hydroxymethyl)-
phenyl]spirobi[1H,3H-2,1-benzoxaphosphole] 7a, which was
isolated in 69% yield by controlled potential preparative
electrolysis. In the absence of collidine, a fraction of 5a was
protonated and became electrochemically inactive at the oxida-
tion potential used in the experiments.
The anodic behavior and the ESR characterization at 77 K
in Freon matrices of the cation radicals of tris[2-(methoxy-
methyl)phenyl]phosphane, tris[2-((methylthio)methyl)phenyl]-
phosphane and tris[2-((dimethylamino)methyl)phenyl]phosphane
will be reported in a forthcoming paper.
References and Notes
(1) Culcasi, M.; Berchadsky, Y.; Gronchi, G.; Tordo, P. J. Org. Chem.
1991, 56, 3536.
The mean value of the proportionality coefficient was used
to determine the quantity of collidinium cations generated during
the electrolysis. For each voltammetric monitoring curve, the
concentration of the remaining 5a (assuming proportionality
between limiting current and concentration) and that of the
collidinium salt were computed (Table 4). As predicted
according to Scheme 6, the ratio between the quantity of
generated collidinium and the quantity of consumed 5a is very
close to 2.
(2) Palau, C.; Berchadsky, Y.; Chalier, F.; Finet, J.-P.; Gronchi, G.;
Tordo, P. J. Chem. Phys. 1995, 99, 158.
(3) Tomaschewski, G. J. Prakt. Chem. 1966, 33, 1339.
(4) Redouane, N.; Vila, F.; Chalier, F.; Gronchi, G.; Tordo, P.
Phosphorus, Sulfur Silicon Relat. Elem. 1993, 77, 101.
(5) Gilheany, D. G. Chem. ReV. 1994, 94, 168.
(6) McEwen, W. E.; Shiau, W.; Yeh, Y. I.; Schulz, D. N.; Pagilagan,
R. U.; Levy, J. B.; Symmes, C., Jr.; Nelson, G. O.; Granoth, I. J. Am. Chem.
Soc. 1975, 97, 1787.
(7) Cairns, S. M.; McEwen, W. E. Heteroat. Chem. 1990, 1, 9.
(8) Chuit, C.; Corriu, R. J. P.; Montforte, P.; Reye´, C.; Declercq,
J.-P.; Dubourg, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1430.
(9) Miller, G. R.; Yankowsky, A. W.; Grim, S. O. J. Chem. Phys. 1969,
51, 3185.
Workup of the electrolysis solution mixture afforded the
spirobi[benzoxaphosphole] 7a (Figure 6), which was isolated
in a 69% yield.
(10) Schiemenz, G. P.; Kaack, H. Liebigs Ann. Chem. 1973, 1480.

4330 J. Phys. Chem., Vol. 100, No. 10, 1996
Chalier et al.
(11) Hoffmann, A. K.; Thomas, W. M. J. Am. Chem. Soc. 1959, 81,
580.
(20) Dunne, B. J.; Orpen, A. G. Acta Crystallogr., Sect. C 1991, 45,
345.
(12) Bartlett, P. A.; Bauer, B.; Singer, S. J. J. Am. Chem. Soc. 1978,
100, 5085.
(21) Johnson, C. K. ORTEP Report ORNL-3794; Oak Ridge National
Laboratary: Tennesse, 1965.
(13) Frenz, B. A. In The Enraf Nonius CAD-4 SDP-A Real Time System
for Concurrent X-ray Data Collection and Crystal Structure Solution in
Computing in Crystallography; Schenk, H., Olthof-Hasekamp, R., van
Koningsveld, H., Bassi, G. C., Eds.; Delft University Press: Delft, 1978;
pp 64.
(14) Main, P.; Fiske, S. J.; Hull, S. E.; Lessinger, L.; Germain, G.;
Declerq, J. P.; Woolfson, M. N. In MULTAN 80, A System of Computer
Programs for the Automatic Solution of Crystal Structure from X-ray
Diffraction Data; Universities of York, England and Louvain: Belgium,
1980.
(22) Cameroun, S. T.; Dahle`n, B. J. Chem. Soc., Perkin Trans. 2 1975,
1737.
(23) Blount, J. F.; Maryanoff, C. A.; Mislow, K. Tetrahedron Lett. 1975,
913.
(24) Bondi, A. J. Phys. Chem. 1964, 68, 441.
(25) Eastland, G. W.; Symons, M. C. R. J. Chem. Soc., Perkin Trans.
2 1977, 833.
(26) Culcasi, M.; Gronchi, G.; Tordo, P. Phosphorus Sulfur Relat. Elem.
1987, 30, 511.
(27) This method was also used by Romakhin, A. S.; Nikitin, E. V.;
Parakin, O. V.; Ygnat’ev, Yu. A.; Mironov, B. S.; Kargin, Yu. M. Zh.
Obshch. Khim. 1986, 56, 259.
(28) Save´ant, J. M. Acc. Chem. Res. 1993, 26, 455. Andrieux, C. P.;
Save´ant, J. M. In Electrochemical Reaction. InVestigation of Rates and
Mechanism of Reactions; Bernasconi, C. F., Ed.; Techniques of Chemistry,
Vol. 6 (2); JohnWiley & Sons: New York, 1986; Chapter 7, p 305. Evans,
D. H. Chem. ReV. 1990, 90, 739, and references therein.
(29) Amatore, C.; Capobianco, G.; Farnia, G.; Sandona`, G.; Save´ant, J.
M.; Severin, G.; Vianello, E. J. Am. Chem. Soc. 1985, 107, 1815.
(15) Ibers, J., Hamilton, W. C., Eds. International Tables for X-Ray
Crystallography; Kynoch Press: Birmingham (present distributor, Kluwer
Academic Publishers: Dordecht), 1974; Vol. IV.
(16) The chemical shift and proton-proton coupling constants were
determined by COSY HH and 1H NMR with phosphorus coupling by using
reverse probe.
(17) The chemical shifts were assigned by HMQC spectrum elucidation.
(18) Berchadsky, Y.; Gronchi, G. Syste`me d’acquisition BYGG-AEL.
Catalogue de l’Agence Nationale du Logiciel (France); Societe´ Franc¸aise
de Chimie: Paris, November 1991.
(19) Dahl, B. M.; Dahl, O.; Trippet, S. J. Chem. Soc., Perkin Trans. 1
1981, 2239.
JP952854P