Hindered rotation around a C-?PH bond: A single-crystal EPR study of the diphenyldibenzobarrelenephosphinyl radical

Article abstract of DOI:10.1021/jp9816519

A new phosphine, the diphenyldibenzobarrelenephosphine 2, was designed to study the barrier to rotation of the P-H group around the C-^?P bond. After homolytic scission of a P-H bond by radiolysis, the EPR spectrum of the resulting phosphinyl radical, trapped in a single crystal of 2, was studied at 77 K and at room temperature. The directions of the ³¹P hyperfine eigenvectors were compared with the bond orientations of the undamaged compound as determined from its crystal structure. The temperature dependence of the EPR spectrum was analyzed by using the density matrix formalism; this showed that interaction between the phosphinyl hydrogen and the phenyl ring bound to the ethylenic bond is determinant for explaining the potential energy profile. DFT investigations are consistent with these experimental results.

Full text of DOI:10.1021/jp9816519

J. Phys. Chem. A 1998, 102, 8245-8250
8245
Hindered Rotation around a C-•PH Bond: A Single-Crystal EPR Study of the
Diphenyldibenzobarrelenephosphinyl Radical
Marcin Brynda, The´o Berclaz, Michel Geoffroy,* and Geetha Ramakrishnan
Department of Physical Chemistry, 30 Quai Ernest Ansermet, UniVersity of GeneVa, 1211, GeneVa, Switzerland
Ge´rald Bernardinelli
Laboratory of X-ray Crystallography, 24 Quai Ernest Ansermet, UniVersity of GeneVa,
1211 GeneVa, Switzerland
ReceiVed: March 26, 1998; In Final Form: June 16, 1998
A new phosphine, the diphenyldibenzobarrelenephosphine 2, was designed to study the barrier to rotation of
the P-H group around the C-•P bond. After homolytic scission of a P-H bond by radiolysis, the EPR
spectrum of the resulting phosphinyl radical, trapped in a single crystal of 2, was studied at 77 K and at room
temperature. The directions of the 31P hyperfine eigenvectors were compared with the bond orientations of
the undamaged compound as determined from its crystal structure. The temperature dependence of the EPR
spectrum was analyzed by using the density matrix formalism; this showed that interaction between the
phosphinyl hydrogen and the phenyl ring bound to the ethylenic bond is determinant for explaining the potential
energy profile. DFT investigations are consistent with these experimental results.
Introduction
steric hindrance caused by the presence of such a moderately
cumbersome group as a phenyl ring.
Whereas a large variety of spectroscopic techniques are well-
suited to the study of hindered rotation in diamagnetic mol-
ecules,1 investigation of the motion around a RA•-BR′ bond,
where the atom A bears an unpaired electron, is considerably
more problematic. We have recently reported2 that radical
derivatives of 9-substituted triptycenes potentially offer a good
opportunity for such measurements: after radiolytic scission of
a P-H bond in 1 (Scheme 1) the temperature dependence of
the single-crystal EPR spectrum of R1 is directly dependent
upon the rotation barrier around the C-•PH bond.
SCHEME 1
Experimental Section
The question is to know to what extent this motion, and
therefore the EPR spectrum, is sensitive to structural modifica-
tions of the triptycyl moiety. We have therefore synthesized
the novel compound 2, which, in contrast with 1, has no C3
axis. We show that the corresponding radiogenic phosphinyl
radical R2 gives rise to an EPR spectrum whose temperature
dependence considerably differs from that observed for R1.
To simplify the EPR analysis, spectra were obtained after
deuteration of the phosphinyl group of 2 (D-2), and for the sake
of comparison new measurements were carried out on the
deuterated phosphine 1 (D-1).
Syntheses. As shown in Scheme 2, 2 was synthesized by a
direct Diels-Alder addition3 followed by lithiation of the adduct
and the subsequent reduction of dichlorophosphine. All experi-
ments were performed under dry N2 atmosphere.
9-Bromodiphenyldibenzobarrelene (2a). A glass tube con-
taining 4 g of bromoanthracene and 3.1 g of diphenylacetylene
dissolved in 40 mL of triglyme was sealed and heated at 235
°C for 48 h in an autoclave. After cooling, 300 mL of diethyl
ether was added to the dark brown solution. The residual
From an analysis that combines a density matrix simulation
of the spectra recorded for R2, the crystallographic parameters
measured on 2, and functional density calculations performed
on R2′, we have obtained a detailed description of the rotation
of the P-H group around the C-•P bond in the solid state.
This method appears to be particularly informative about the
10.1021/jp9816519 CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/02/1998

8246 J. Phys. Chem. A, Vol. 102, No. 43, 1998
Brynda et al.
SCHEME 2
solvent was extracted with three times 100 mL of water, and
after the usual workup the resulting brown oily solution was
quickly poured into 200 mL of hexane cooled to 0 °C. The
resulting yellow crystalline powder was dried under vacuum
and recrystallized from an ether/hexane mixture. Yield: 1.37
g [21%]. MS-M+: 436; 355; 256; 178; 151, Mp ) 160 °C,
1H-RMN 5.45 (s, 1H), 6.82 (m, 2H), 6.99 (m, 2H), 7.15 (m,
10H), 7.79 (m, 2H) ppm.
9-Diphenyldibenzobarrelenephosphine (2). n-BuLi (3.15 mL)
(1.6 M in hexane) was slowly added, at -100 °C, to a solution
containing 1.8 g of 2a in 30 mL of dry THF. After 20 min the
temperature was allowed to slowly rise and then maintained at
-65 °C for 30 min. Then the mixture was cooled to -80 °C,
and 0.6 mL of PCl3 was added all at once. The solution was
heated and refluxed for 1 h. After cooling, the mixture was
slowly poured, at 0 °C, in a suspension of 0.3 g of LiAlH4 in
dry hexane. After additional reflux (2 h) a few drops of HCl/
H2O were carefully added to the solution. The solution was
then extracted with 3 × 30 mL of ether, washed, dried, and
evaporated. The final product was separated by silica gel
chromatography using hexane/ether (9:1) as an eluant. Yield:
63% (1 g).
MS-M+: 388, Mp ) 141 °C, H-RMN 2.75; 3.76 (d, 2H,
JPH ) 201 Hz), 5.47 (s, 1H), 6.84 (m, 2H), 6.99-7.25 (m, 12
H), 7.43 (m, 2H), 7.56 (m, 2H) ppm. 31P-RMN: -152 ppm,
JP-H ) 201 Hz.
1
Figure 1. Perspective view of the 9-diphenyldibenzobarrelenephos-
phine 2 with the atom numbering. Ellipsoids are represented with 40%
probability level.
structure was solved by direct methods using MULTAN 87;6
all other calculations used the XTAL7 system and ORTEP8
programs. The chirality/polarity of the structure was refined,
and the absolute structure parameter9 converges to x ) 0.02
(5). The structure was refined by full-matrix least-squares, based
Deuteration of 2. The procedure was similar to that described
for the undeuterated compound. However, 2b was reacted with
LiAlD4, and, after refluxing, two drops of DCl/D2O were added
to the reaction mixture. After extraction and partial evaporation
of the solution, yellow crystals grew on keeping the solution
one night at 0 °C. These crystals were dissolved in THF, and
the deuterated compound D-2 was isolated by chromatography
on deuterated alumina4 with hexane/ether as an eluant. Yield:
39% (0.62 g).
Crystallographic Data and Structure Determination of 2.
C28H21P, Mr ) 388.4; µ ) 1.254 mm-1, F(000) ) 816, dx )
1.26 g‚cm-3, orthorhombic, P212121, Z ) 4, a ) 9.890(1), b )
13.654(2), c ) 15.148(1) Å, V ) 2045.6(4) Å3 from 23
reflections (46° < 2θ < 53°), colorless prism 0.08 × 0.20 ×
0.26 mm mounted on a quartz fiber with RS3000 oil to prevent
degradation. Cell dimensions and intensities were measured
at 200 K on a STOE Stadi4 diffractometer with graphite-
monochromated Cu KR radiation (λ ) 1.5418 Å), ω-2θ scans,
scan width 1.11° + 0.35 tg θ, and scan speed 0.06 deg/s. Two
reference reflections measured every 45 min showed variation
less than 2.5 σ (I). 3° < 2θ < 110°; 0 < h < 10; 0 < k < 14;
0 < l < 15 and all antireflections of these; 3638 measured
reflections, 2505 unique reflections of which 2275 were
observable (|Fo| > 4σ(Fo)); Rint for 1133 equivalent reflections
0.026. Data were corrected for Lorentz and polarization effects
and for absorption 5(A* min, max ) 1.099, 1.343). The
on |F|, using a weight of 1/(σ2(Fo) + 0.0001(Fo )) and gave
2
final values R ) 0.037, ωR ) 0.035, and S ) 2.21(5) for 330
variables and 2275 contributing reflections. The mean and
maximum shift/error on the last cycle was 0.34 × 10-3 and
0.21 × 10-2 respectively. All coordinates of hydrogen atoms
were observed and refined with a fixed value of isotropic
displacement parameters (U ) 0.05 Å2). Three atomic sites
have been observed, for the hydrogen atoms bonded to the
phosphorus. As shown by the difference electron density map,
these hydrogen atoms have been refined with population
parameters of 0.8, 0.8, and 0.4 (Σ of populations ) 2) for H(01),
H(02), and H(03) respectively (Figure 1). The final difference
electron density map showed a maximum of +0.17 and a
minimum of -0.22 e Å-3
.
EPR Measurements. In a typical experiment a large single
crystal of 2 (or D-2 or D-1), obtained by slow evaporation of a
solution in a hexane/diethyl ether (9:1) mixture, was X irradiated
at room temperature under vacuum for 2 h (tungsten anticathode,
30 mA, 30 kV). The irradiated crystal was then glued on a
small brass sample holder designed to allow rotation of the
crystal around the three crystallographic axes (the experimental

Hindered Rotation around a C-•PH Bond
J. Phys. Chem. A, Vol. 102, No. 43, 1998 8247
1
TABLE 1: Selected Bond Lengths (Å), Bond Angles, and
Torsional Angles (deg) for the
9-Diphenyldibenzobarrelenephosphine 2
TABLE 2: g, H, and 31P Coupling Tensors Measured at
300 K for Radical R2
principal axesa
P-C(9)
1.862(5) C(10)-C(10a)
1.401(8) C(10)-C(11)
1.515(8) C(11)-C(12)
1.519(6) P-H(01)
1.397(7) P-H(02)
1.554(7) P-H(03)
1.554(6)
1.533(6)
1.537(6)
1.338(6)
1.37(5)
1.37(5)
1.37(4)
tensor
principal values
λ
µ
ν
C(4a)-C(9a)
C(4a)-C(10)
C(8a)-C(9)
C(8a)-C(10a)
C(9)-C(9a)
C(9)-C(12)
g
2.0050
2.0072
2.0109
0.1846
0.9748
0.1255
-0.7389
0.2219
-0.6362
-0.6480
0.0248
0.7612
31P
1H
527
293
39
-0.4189
-0.9039
0.0859
0.5842
-0.1959
0.7876
0.6951
-0.3801
-0.6102
H(01)‚‚‚H(1)
H(02)‚‚‚H(1)
H(03)‚‚‚H(8)
2.31(7) H(01)‚‚‚H(20)
2.26(6) H(02)‚‚‚H(8)
3.74(6)
2.46(6)
2.5(1)
61
37
33
-0.0899
-0.1259
0.9879
-0.8113
-0.5660
-0.1460
0.5776
-0.8147
-0.0513
2.8(1)
H(03)‚‚‚H(24)
P-C(9)-C(8a)
112.6(3) C(9)-P-H(01)
115.4(3) C(9)-P-H(02)
112.2(3) C(9)-P-H(03)
104.0(4) H(01)-P-H(02)
107.3(4) H(01)-P-H(03)
104.4(4) H(02)-P-H(03)
95(2)
93(2)
114(2)
86(3)
127(3)
132(3)
P-C(9)-C(9a)
a Principal axes for the other sites are obtained from the symmetry
operations: λh, µ, ν; λ, µj, ν; λ, µ, νj.
P-C(9)-C(12)
C(8a)-C(9)-C(9a)
C(8a)-C(9)-C(12)
C(9a)-C(9)-C(12)
two phenyl rings at C(11) and C(12) are conrotatory oriented
(C(9)-C(12)-C(19)-C(20) ) 113.1(5)° and C(10)-C(11)-
C(13)-C(18) ) 145.1(5)°) and show a slide deviation of C(13)
(0.126(8) Å) and C(19) (0.109(8) Å) out of the mean plane of
the C(9)-C(10)-C(11)-C(12) bridge toward the C(1)-C(9a)
phenyl ring.
H(01)-P-C(9)-C(8a) 160(2)
H(01)-P-C(9)-C(9a) 40(2)
H(02)-P-C(9)-C(12) -165(2)
H(03)-P-C(9)-C(8a) -65(3)
H(01)-P-C(9)-C(12) -79(2) H(03)-P-C(9)-C(9a) 175(3)
H(02)-P-C(9)-C(8a) 73(2)
H(02)-P-C(9)-C(9a) -46(2)
H(03)-P-C(9)-C(12) 56(3)
As mentioned in the crystallographic experimental section,
three atomic sites have been observed for the hydrogen atoms
bonded to the phosphorus showing staggered conformations
relative to the barrelene moiety. The coordinates of these
hydrogen atoms have been refined with weak restraints on the
P-H bond lengths. The population parameters observed for
these atomic sites show one major rotamer (60%) and two
equally populated (20%) minor rotamers. For the major
conformation, C-P-H and H-P-H bond angles are in good
agreement with analogous phosphine compounds12 despite short
interatomic interactions with H(1) (Table 1). In the crystal
structure and for the temperature of measurement (200 K), we
can assume that no free rotation occurs about the C-P bond.
This assumption seems to be confirmed by the anisotropy of
the displacement parameters observed on the phosphorus atom
where the main displacement is oriented parallel to the P-H(03)
direction (see Figure 1). Both minor rotamers involving H(03)
show a significant opening of the C-P-H and H-P-H bond
angles.
EPR reference axes X, Y, Z, for 2 and D-2 are respectively
aligned along the a, b, c directions). To prevent any chemical
alteration in the irradiated compound, the crystal was then coated
with an inert adhesive polymer.
The EPR spectra were recorded on a Bruker 200D spectrom-
eter (X-band) equipped with an AST computer recording unit
and a VT1100 temperature controller. The angular variations
of the spectra, in three perpendicular planes, were obtained at
300 K and at 77 K. The EPR tensors were determined by using
an optimization program that calculates the position of the
signals with second-order perturbation theory.
As described in more detail in ref 2, the EPR spectra were
simulated by using the density matrix formalism.10 The intensity
of the EPR signal was calculated as I(ω) ) Re(S+(ω)) ) Tr-
(FS+) ) Tr(F-+). All the calculations were performed in
Liouville space by summing the elements of the vector F-+
)
iIo (U-1ΛU - iω1)-1P where Λ is a diagonal matrix and P is
a vector containing the populations of the various configurations.
In addition to the transition values and the -1/T2 terms, the
elements of the matrix L ) (U-1ΛU - iω1) contain the rate
constants kij. For R2, three configurational sites (J, K, L) are
present; at a given temperature each site is characterized by its
population Pi(i ) j, k, l) and by its exchange parameters with
the two other sites: kih ) Ph‚khi/Pi (with h * i, h ) j, k, l). The
EPR Results. Spectra Obtained with a Single Crystal of 2.
When the magnetic field lies in an EPR reference plane (X, Y,
Z), the EPR spectrum obtained at room temperature with an X
irradiated single crystal of 2 is composed of four lines marked
A and of four lines marked B (see Supporting Information);
correlation time τ is related to k as τ ) (2π × 2.8 × 106k)-1
.
each of these subspectra exhibits hyperfine coupling with a 31
P
The temperature dependence of the spectra was analyzed by
determining, for each temperature, the set of kij values that leads
to a good simulation of the experimental spectrum.
Quantum Mechanical Calculations. DFT calculations were
carried out on a Silicon Graphics computer (IRIS 4D) using
the GAUSSIAN94 package.11 All the results were obtained with
a 3-21G* basis set.
nucleus and a proton. The angular variations of these signals
lead to the g, 1H and 31P hyperfine tensors given in Table 2. In
accord with the crystal structure that indicates the presence of
four molecules per cell, the radicals are trapped along four
orientations; each signal results from the overlap of two lines
since the EPR reference planes are perpendicular to the C2 axes
present in the structure. On all spectra, a line, marked T, is
observed around g ) 2.002 and is probably due to the trapping
of a radical resulting from the homolytic scission of the
C(10)-H bond.
The spectrum is highly temperature sensitive between 100
and 240 K. At 77 K, multiple and partial overlaps of lines occur
and make the resulting pattern so complex that, for some
orientations, the transitions cannot be identified. This was not
unexpected since, in addition to the four orientations of the
phosphinyl radical present in the unit cell, freezing the motion
of the PH group around the C-P bond gives rise, a priori, to
Results
Crystal Structure of 2. Selected geometrical parameters are
reported in Table 1. No significant difference in the geometrical
parameters are observed with the analogous 9-phosphinyltrip-
tycene compound.2 Nevertheless, it should be noted that in 2
the bond length of the C(11)-C(12) bridge is significantly
shorter than in the C(4a)-C(9a) and C(8a)-C(10a) phenyl
bridges. The dihedral angles between the mean planes of the
central barrelene skeleton are 118.8, 119.1, and 122.2°. The

8248 J. Phys. Chem. A, Vol. 102, No. 43, 1998
Brynda et al.
Figure 2. EPR spectrum obtained at 77 K with an X irradiated single
crystal of D-2 (Ho lies in the ab plane, 10° from the a axis).
1
three new sets of g and 31P and H hyperfine tensors. This
difficulty was overcome by studying the deuterated compound.
EPR Spectra of Deuterated 2. An EPR spectrum obtained
at room temperature with an X irradiated single crystal of D-2
exhibits hyperfine coupling with a 31P nucleus only. The
angular dependence of this spectrum leads to g and 31P hyperfine
tensors that are identical with those reported in Table 2 for the
undeuterated radical studied at 300 K. At 77 K, when the
magnetic field is oriented in the EPR reference planes, the
spectrum (Figure 2) exhibits a maximum of eight lines due to
additional nonequivalent orientations of the trapped radical. The
subspectrum A gives rise to “configuration sites” Aj and Ak,
and the subspectrum B generates Bj and Bk. The angular
dependence of these signals lead to the tensors given in Table
3. It is worthwhile mentioning some geometric properties
measured at low temperature: (1) the 31P-T| principal axes of
the two molecular sites make an angle of 110.3°, (2) all the
31P-T| principal axes are oriented perpendicular to the crystal-
lographic C-P direction, and (3) the T|a and T|b directions make
angles of 6.4° and of 3.1° with the normals to the C(8a)C(9)P
and C(12)C(9)P planes, respectively.
The number of observed transitions, their intensity, as well
as their position reversibly vary with temperature between 100
and 240 K. This dependence is illustrated in Figure 3 where
the spectra have been recorded when the magnetic field is
aligned along the crystallographic a direction. For this orienta-
tion the pattern is particularly clear since the two “crystal-
lographic sites” (A and B) are magnetically equivalent and no
overlap occurs between the signals owing to the “configuration
sites”(Aj and Ak, Bj and Bk).
Figure 3. Temperature dependence of the experimental EPR spectrum
obtained with an X irradiated single crystal of D-2. (Ho is aligned along
the a axis). The lines marked k are due to an overlap of the Ak and Bk
lines, j signals correspond to the overlap of the Aj and Bj lines. The
calculated positions of the Al /Bl sites are marked l.
above 180 K, it is impossible to reproduce the change in the
relative intensity of the two exchanging lines (for example, Ak/
Aj, in Figure 3), which is observed between 110 and 170 K.
Careful examination of the spectra recorded in this temperature
range shows that the change in the intensity ratio is accompanied
by a change in the shape of the signal: between 110 and 130
K the broadening of the signal marked j is more pronounced
than that of the signal marked k. This behavior can be simulated
by assuming that a third “site”, marked l (say Al and Bl ), with
a population of 0.04 at 100 K exchanges with the configurational
site j with a rate constant kjl equal to 0.11 and with the site k
with a constant kkl ) 0.30. The resonance positions of the Al
and Bl lines were assumed to occur at the field values calculated
for a phosphinyl radical whose C(9a)C(9)PD dihedral angle is
equal to 180° (Hlow field ) 332 mT, Hhigh field ) 334 mT). The
populations of the J and K structures were assumed to be both
equal 0.48 at 100 K. At higher temperature the population
numbers of J, K, L were calculated accordingly to the Boltzmann
distribution and the kuV (u, V ) j, k, l) values were adjusted by
simulation. In these conditions the lines l are never directly
observed, but they are revealed by the shape modification of
signals Aj, Ak (and Bj, Bk). The resulting simulations are shown
in Figure 4, and the related parameters are given in Table 4.
Simulation. Simulation of the temperature dependence of the
EPR spectrum by using the model of a proton exchanging
between two sites that differ only by their orientation is not
totally satisfactory, whereas the fitting is almost acceptable
EPR Spectra of Deuterated 1. The deuteration of 1 causes a
drastic simplification of the low-temperature spectra and
TABLE 3: g and 31P Coupling Tensors for the Radical D-2 at 77 K
principal axesa
tensor
principal values
λ
µ
ν
site j/k g
2.0014/2.0014
2.0047/2.0059
2.0137/2.0147
865/833
34/74
25/28
-0.5378/0.5961
0.6348/0.4991
-0.5548/0.6289
-0.5095/0.6186
0.6206/0.7800
0.5961/-0.0950
-0.5403/-0.4998
-0.7647/0.8437
-0.3512/-0.1958
-0.5758/-0.4504
0.2689/0.2529
-0.6472/-0.6283
0.1109/-0.1976
0.7542/0.7524
-0.6394/-0.6438
-0.7366/0.5725
0.2203/0.5077
site j/k 31
P
-0.7721/-0.8563
a Principal axes for the other sites are obtained from the symmetry operations: λh, µ, ν; λ, µj, ν; λ, µ, νj.

Hindered Rotation around a C-•PH Bond
J. Phys. Chem. A, Vol. 102, No. 43, 1998 8249
Figure 6. (a) Configurations due to the orientation of the phosphinyl
group in 2 as obtained from the crystal structure. (b) Configurations
•
due to the orientation of the P-H bond in R2 as obtained from the
EPR analysis (φ represents the position of the C(19)-C(24) phenyl
ring).
crystallographic coordinates of the carbon framework and by
using for the CPH fragment the geometrical properties previ-
ously obtained by ab initio calculations on the barrelenophos-
phinyl radical. The resulting curve is shown in Figure 5 (curve
a) together with that obtained, by the DFT method, for the
barrelenephosphinyl radical (curve b); we have also performed
similar calculations for the triptycylphosphinyl radical R1 (curve
c). To detect the influence of the orientation of the C(19)-
C(24) phenyl ring we have performed another set of calculations
by rotating the plane of this phenyl ring of 15° around the
C(12)-C(19) bond. For this arbitrary structure, the potential
energy as a function of the P-H orientation is also shown in
Figure 5 (curve d).
Figure 4. Simulation of the temperature dependence of the EPR spectra
obtained with an X irradiated single crystal of D-2.
TABLE 4: Exchange Parameters and Populations Used for
the Simulation of the EPR Spectra at Various
Temperaturesa
temp
kjl
kjk
kkl
Pj
Pk
Pl
100
120
170
220
300
0.05
2.5
16.0
80.0
400.0
0.11
0.4
4.0
50.0
180.0
0.3
1.3
14.0
250.0
1800.0
0.48
0.47
0.448
0.43
0.41
0.48
0.47
0.448
0.43
0.41
0.04
0.06
0.104
0.14
0.18
a Additional sets of values are given as Supporting Information.
Discussion
The g and 31P hyperfine tensors measured at 77 K for the
radical trapped in a crystal of 2 (see Table 3) are totally
consistent with those expected for the immobile phosphinyl
radical R2: axial symmetry, g| aligned along 31P-T| and close
to 2.002, 31P isotropic coupling constant Aiso ) 310 MHz,
anisotropic coupling constants: τ| ) 538 MHz, τ ) 40 MHz.
Moreover, as for R1, the parallel component of the 31P hyperfine
coupling is oriented perpendicular to the crystallographic C-P
bond direction of the undamaged phosphine. All these proper-
ties, together with the fact that deuteration of 2 suppresses a
1H coupling, show that the expected homolytic scission of the
P-H bond occurs; moreover, the similarity of the tensors
measured for R1 and R2 at 77 K clearly indicates that the EPR
parameters of the motionless phosphinyl radical do not depend
on the nature of the triptycene matrix.
We have already shown,2,14 that the most stable geometry of
a HP-CR3 phosphinyl radical corresponds to the staggered
configuration (HPCR dihedral angle ) 180°). The salient
feature of the spectra obtained with the phosphinyl radical R2
trapped in the crystal of 2 is that each “site” detected at room
temperature seems to generate only two “configurational sites”
at 77 K. This is in contrast with the situation observed with a
crystal of D-1 where the P-D bond of the phosphinyl radical
is blocked, at 77 K, with the same probability along the three
orientations staggering the barrelene C9-C bonds. Information
about the stereochemistry of R2 at low temperature can be
obtained by examining the orientations of the magnetic phos-
Figure 5. DFT calculated variation of the potential energy of some
phosphinyl radicals as a function of the C(12)C(9)PH torsion angle:
(a) R2; whose atomic coordinates are obtained from the crystal structure;
(b) phosphinylbarrelene; (c) phosphinyltriptycene R1; (d) R2′ after 15°
rotation of the C(19)-C(24) ring around the C(12)-C(19) bond.
therefore a considerable improvement of the precision in the
line position measurement.13
Quantum Mechanical Calculations. Radical R2 is too large
to be studied in depth by quantum mechanical calculations. Since
the motion of the P-H bond is probably not affected by the
presence of the C(13)-C(18) ring, we have used the smaller
radical R2′ as a model. To estimate the dependence of the
potential energy on the orientation of the P-H bond we have
therefore carried out DFT calculations on R2′ by using the

8250 J. Phys. Chem. A, Vol. 102, No. 43, 1998
Brynda et al.
the barrier between J and L is drastically dependent upon the
orientation of the phenyl ring C(19)-C(24). The comparison
between the curves a and d of Figure 5 and that of Figure 7
suggests that after the homolytic scission of the P-H bond a
slight reorientation of the phenyl ring increases the barrier
between configuration L and J.
Acknowledgment. We thank the Swiss National Science
Foundation for financial support.
Supporting Information Available: Tables of atomic
coordinates, isotropic and anisotropic displacement parameters,
and bond lengths and angles together with k and P values used
for the EPR spectra simulations; EPR spectra obtained at 300
K with a crystal of 2 and D-2 and at 77 K with a crystal of
D-2; angular variation of the EPR signals (300 K, crystal of 2);
experimental 1/T-dependence of the exchange constants for R2
(11 pages). Ordering and access information is given on any
current masthead page.
Figure 7. Schematic representation of the potential energy variation
of R2 as a function of the C(12)C(9)PH torsion angle. ∆EJL ) 2.46
kcal‚mol-1, ∆EJK ) 2.37 kcal‚mol-1, ∆EKL ) 2.74 kcal‚mol-1, EK
-
EJ = 0 kcal‚mol-1, EL - EK ) 0.5 kcal‚mol-1
.
phorus 3p orbital as given by the two 31P-T| principal axes:
31P-T|a and 31P-T|b are almost perpendicular to the planes
C(8a)C(9)P and C(12)C(9)P, respectively, and imply that R2
adopts the two configurations marked J and K in Figure 6b. In
the radical, therefore, at 77 K the phosphinyl hydrogen atom
never occupies the position associated with that of the atom H3
(Figure 6a), which is the very hydrogen characterized by a small
occupation parameter in the undamaged phosphine at 200 K
(see crystal structure).
The fact that the position corresponding to staggered P-D
and C(9)-C(9a) bonds (configuration L) is practically unoc-
cupied at 77 K is obviously due to the presence of the phenyl
ring bound to the C(12) carbon. The orientation of this ring
increases the repulsion between the phosphinyl hydrogen H(3)
and the hydrogen bound to C(24), and this interaction certainly
destabilizes the configuration L (Figure 6b).
References and Notes
(1) (a) Oki, M. Acc. Chem. Res. 1990, 23, 351, (b) Oki, M.; Matsusue,
M.; Akinaga, T.; Matsumoto, Y.; Toyota, S. Bull. Chem. Soc. Jpn. 1994,
67, 2831.
(2) Ramakrishna, G.; Jouaiti, A.; Geoffroy, M.; Bernardinelli, G. J.
Phys. Chem. 1996, 100, 1086.
(3) Nakadaira, Y.; Sato, R.; Sakurai, H. Chem. Lett. 1985, 643.
(4) Kabalka, G. W.; Pagni, R. M. J. Org. Chem. 1981, 46, 1513.
(5) Blanc, E.; Schwartzenbach, D.; Flack, H. D. J. Appl. Crystallogr.
1991, 24, 1035.
(6) Main, P.; Fiske, S. J.; Hull, S. E.; Lessinger, L.; Germain, G.;
Declercq, J.-P.; Woolfson, M. M. A System of Computer Programs for the
Automatic Solution of Crystal Structures from X-Ray Diffraction Data;
Universities of York, England, and Louvain-la-Neuve, Belgium, 1987.
(7) Hall, S.R.; Flack, H. D.; Stewart, J. M. Eds XTAL3.2 User’s Manual;
Universities of Western Australia and Maryland, 1992.
As shown in Table 4, the configuration L begins to be
populated above 100 K, but it immediately exchanges with
configuration J, whereas its exchange with configuration K is
appreciable only above 130 K. Exchange between configura-
tions J and K starts at 110 K.
(8) Johnson, C. K. ORTEP II; Report ORNL-5138; Oak Ridge National
Laboratory: Oak Ridge, TN, 1976.
(9) Bernardinelli, G; Flack, H. D. Acta Crystallogr. 1985, A41, 500.
(10) (a) Hayes, R. G.; Steible, D.; Tolles, W. M.; Hunt, J. J. Chem.
Phys. 1970, 53, 4466. (b) Benetis, N. P.; Sjo¨quist, L.; Lund, A.; Maruani,
J. J. Magn. Reson. 1991, 95, 523, (c) Heinzer, J. Mol. Phys. 1971, 22, 167.
(d) Bogan, C. M.; Kispert, L. D. J. Chem. Phys. 1972, 57, 3109.
(11) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. V.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Ciolowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Reploges, E. S.; Gomperts R.; Martin R. L.;
Fox D. J.; Binkley, J. S.; Defrees D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. Gaussian 94, Revision B.1; Gaussian
Inc.: Pittsburgh, PA; 1995.
All the parameters measured from the temperature depen-
dence of the spectra (population of the sites, correlation times
-1
ij
k
, rotation barriers) indicate that, in contrast to R1, the three
potential wells of R2 are not identical. The variations of the
rate constants with 1/T show, however, that, in the approxima-
tion of the Arrhenius law, the activation energies remain very
close to each another and are comprised between 2 and 3 kcal
mol-1. A sketch of the principal properties of the potential
profile is shown in Figure 7.
The DFT calculated curves visualize some important effects
of the hydrocarbon framework on the rotation of the P-H bond
around the C(9)-P bond. As expected, passing from the
barrelenyl to the triptycyl increases the barrier (interaction with
the three benzene protons in â position to C(9)) and using the
crystallographic coordinates for R2′ leave the two potential wells
corresponding to configuration J and K very similar whereas
(12) Wesemann, J.; Jones, P. J.; Schomburg, D.; Heuer, L.; Schmutzler,
R. Chem. Ber. 1992, 125, 2187.
(13) The 31P hyperfine couplings obtained at 300 K are equal to 23,
409, 413 MHz and those measured at 77 K are equal to 4, 74, 848 MHz.
Owing to improvement in the precision of our measurements (better
estimation of the temperature gradient, more accurate determination of the
tensor elements), all the experimental points of the curve ln τ-1 ) f(1/T)
are located on a single line and correspond to an energy barrier equal to
2.75 kcal‚mol-1 (previous results: 3.2 and 1.6 kcal‚mol-1).
(14) Bhat, S. N.;. Berclaz, T.; Jouaiti, A.; Geoffroy, M. HelV. Chim.
Acta 1994, 77, 371.