Preparation, crystal structure determination and properties of adducts of indium methylene compounds with Group 15 donors

Article abstract of DOI:10.1016/S0022-328X(00)00178-9

The reaction of InBr with CH₂Br₂ in 1,4-dioxane or acetonitrile yields the corresponding solvate of Br₂InCH₂Br, which on reaction with E(C₆H₅)₃ (E=P, As, Sb) gives the indium(III)-Group 15 dimetallo-methane derivatives, Br₃InCH₂E(C₆H₅)₃. The crystal structures of the two related compounds (E=As, Sb) have been determined by X-ray crystallography. For Br₃InCH₂As(C₆H₅)₃, cell constants a=15.553(7), b=21.646(8), c=12.920 (10) ?; space group Pbca, Z=8, R=0.064, R_w=0.054, and for Br₃InCH₂Sb(C₆H₅)₃, a=15.439(6) ?, b=22.016(4) c=13.138(5) ?; space group Pbca, Z=8, R=0.062, R_w=0.049. Results are also reported for I₃InCH₂As(C₆H₅)₃ (cell constants a=12.0122(2), b=15.8526(3), c=13.4180(3) ?, β=109.933(1)°; space group P2_1/n, Z=4, R=0.0481, R_w=0.0431) and Br₃InCH₂N(C₂H₅)₃ (a=7.362(2), b=14.700(2), c=13.049(1) ?, β=98.90(1)°; space group P2_1/n, Z=4, R=0.0418, R_w=0.0368). The structural results are compared with those for other organoindium ylids with the same general structure Br₃InCH₂L (L=P(C₆H₅)₃; 1,1,3,3,-tetramethyl-2-thiourea; N,N,N′,N′-tetramethylethanediamine). Semi-empirical quantum mechanical calculations, using the PM3 method, were carried out on the series Br₃InCH₂E(C₆H₅)₃ (E=P, As, Sb), and the calculated structural parameters are compared with the values determined by X-ray crystallography. The presence of an ylid ligand H₂C^δ--^δ+E(C₆H ₅)₃ in the organoindium compounds is confirmed. Mass spectra and thermogravimetric analysis strongly suggest that the thermal decomposition of the compounds occurs via fission of the indium-carbon bond, leading to the corresponding ylid, which can be trapped by reacting the X₃InCH₂E(C₆H₅)₃ compounds with HBr to produce the onium derivative [CH₃E(C₆H₅)₃]⁺ [InX₄]^- (E=P, As; X=Br, I).

Full text of DOI:10.1016/S0022-328X(00)00178-9

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 603 (2000) 203–212
Preparation, crystal structure determination and properties of
adducts of indium methylene compounds with Group 15 donors
L. de Arau´jo Felix a, C.A.F. de Oliveira a, R.K. Kross a, C. Peppe a, M.A. Brown b,
D.G. Tuck b,*, M.Z. Hernandes c, E. Longo c, F.R. Sensato c
a Departamento de Quimica, CCEN, Uni6ersidade Federal da Paraiba, Campus I, 58.059-900 Joa˜o Pessoa-PB, Brazil
b Department of Chemistry and Biochemistry, Uni6ersity of Windsor, Windsor, Ont., Canada, N9B 3P4
c Departamento de Quimica, Uni6ersidade Federal de Sa˜o Carlos, Sa˜o Carlos-SP, Brazil
Received 17 September 1999; received in revised form 2 February 2000
Abstract
The reaction of InBr with CH2Br2 in 1,4-dioxane or acetonitrile yields the corresponding solvate of Br2InCH2Br, which on
reaction with E(C6H5)3 (E=P, As, Sb) gives the indium(III)-Group 15 dimetallo-methane derivatives, Br3InCH2E(C6H5)3. The
crystal structures of the two related compounds (E=As, Sb) have been determined by X-ray crystallography. For
,
Br3InCH2As(C6H5)3, cell constants a=15.553(7), b=21.646(8), c=12.920 (10) A; space group Pbca, Z=8, R=0.064, Rw=
,
,
0.054, and for Br3InCH2Sb(C6H5)3, a=15.439(6) A, b=22.016(4) c=13.138(5) A; space group Pbca, Z=8, R=0.062,
,
Rw=0.049. Results are also reported for I3InCH2As(C6H5)3 (cell constants a=12.0122(2), b=15.8526(3), c=13.4180(3) A,
i=109.933(1)°; space group P21/n, Z=4, R=0.0481, Rw=0.0431) and Br3InCH2N(C2H5)3 (a=7.362(2), b=14.700(2), c=
,
13.049(1) A, i=98.90(1)°; space group P21/n, Z=4, R=0.0418, Rw=0.0368). The structural results are compared with those for
other organoindium ylids with the same general structure Br3InCH2L (L=P(C6H5)3; 1,1,3,3,-tetramethyl-2-thiourea; N,N,N%,N%-
tetramethylethanediamine). Semi-empirical quantum mechanical calculations, using the PM3 method, were carried out on the
series Br3InCH2E(C6H5)3 (E=P, As, Sb), and the calculated structural parameters are compared with the values determined by
X-ray crystallography. The presence of an ylid ligand H2Cl−ꢀl+E(C6H5)3 in the organoindium compounds is confirmed. Mass
spectra and thermogravimetric analysis strongly suggest that the thermal decomposition of the compounds occurs via fission of
the indiumꢀcarbon bond, leading to the corresponding ylid, which can be trapped by reacting the X3InCH2E(C6H5)3 compounds
with HBr to produce the onium derivative [CH3E(C6H5)3]+ [InX4]− (E=P, As; X=Br, I). © 2000 Elsevier Science S.A. All
rights reserved.
Keywords: Crystal structure determination; Indium methylene compounds; Group 15 donors
1. Introduction
alkyl) has allowed the study of the properties of these
complexes, which may function as carbene transfer
For some years, interest in the organometallic com-
reagents. It has been argued elsewhere that the chem-
pounds of indium was related to their use in semicon-
istry of such molecules is associated with the tautomeric
ductors, but the recent literature [1,2] has shown that
structure IB arising from intramolecular halide transfer
interesting organic transformations, such as the cyclo-
from the carbon atom to the metal center (Scheme 1) to
propanation of electron-deficient alkenes and the epoxi-
dation of carbonyl compounds, can be achieved by
reaction with indium metal and dialkyldihalides. The
give a species in which the carbon atom is highly
electron deficient. This description predicts that a-
halogeno-alkylindium(III) compounds will have two
preparation of indium(III) compounds [3–6] of general
sites available for electron pair donation, namely the
structure X2InCR1R2X (R1=H; R2=H, halogen,
indium atom and the halogenomethylic carbon. In
agreement with this, we have shown that the addition
of halide anion to X2InCHRX (R=H), leads to the
* Corresponding author. Fax: +1-519-9737098.
E-mail address: dgtuck@uwindsor.ca (D.G. Tuck)
corresponding salts of [X3InCHRX]− (X=Cl, Br, I],
0022-328X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(00)00178-9

204
L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
Hewlett Packard 5890-II spectrometer operating in the
E.I. mode.
All preparative work was conducted under dry nitro-
gen, using standard vacuum line techniques, up to the
isolation of the Br3InCH2ER3 compounds, which are
air stable.
2.2. Preparati6e
Scheme 1.
2.2.1. Br3InCH2N(C2H5)3
InBr (0.58 g, 2.95 mmol) and CH2Br2 (0.63 ml, 1.56
g, 9 mmol) were stirred together in acetonitrile (20 ml)
to dissolution of InBr; at this point all the volatiles
were removed under vacuum to ensure absence of
excess CH2Br2. The oil obtained was redissolved in
acetonitrile (20 ml) and to this solution was added
N(C2H5)3 (0.41 ml, 0.30 g, 2.95 mmol). The reaction
mixture was stirred for 12 h, at which point the powder
that precipitated was collected by filtration and dis-
carded. The solution was concentrated under vacuo to
1/3 of its initial volume and ethanol (95%; 8 ml) added;
slow evaporation of the solution (open to the air) lead
to precipitation of Br3InCH2N(C2H5)3 (0.51 g, 1.10
mmol, 37%) as colorless single crystals. Anal. Found:
C, 18.1; H, 3.74; N, 2.98; In, 24.0; Br, 50.2; Calc. for
C7H17NInBr3 C, 17.9; H, 3.65; N, 2.98; In, 24.4; Br,
and the crystal structures of such anions (X=Cl, Br)
have been established recently by X-ray methods. We
have also synthesized adducts derived directly from
structure IB, with the ligand attached directly to the
carbon atom, including the compounds Br3InCH2P-
(C6H5)3, Br3InCH2N(Me)2CH2CH2NMe2 and Br3In-
CH2SC(NMe2)2, and these have also been characterised
fully [3–9].
We have now extended this latter part of the work by
preparing and structurally characterising adducts of
Br2InCH2Br with triethylamine, triphenylarsenic and
triphenylantimony, and of triphenylarsenic and
I2InCH2I. In each case the product is the appropriate
X3InCH2ER3 compound (E=N, As, Sb). The struc-
tural and thermal properties of these species are dis-
cussed in the light of model calculations on the bromo
compounds.
1
51.0. H-NMR (acetone-d6); 1.45, t, 9H (NCH2CH3);
3.20, s, 2H (InCH2); 3.55, q, 6H (NCH2CH3) (all in
ppm, Me4Si=0). 13C-NMR; 9.23 (NCH2CH3); 59.03
(NCH2CH3) ppm (Me4Si=0). InCH2 was not detected.
An analogous experiment with equimolar quantities
of Br2InCH2Br and N(C6H5)3 in acetonitrile involved
stirring for 12 h at room temperature (r.t.), followed by
refluxing for 1 h. Evaporation of volatile components
gave a white solid, from which 67% of the initial
N(C6H5)3 was recovered. There was no evidence of
adduct formation.
2. Experimental
2.1. General
Indium monobromide was prepared by heating metal
and trihalide together in a sealed tube at 450°C.
Methylene bromide (Aldrich) was dried over Linde 4A
molecular sieves. Triphenylarsenic and antimony
(Aldrich) were used as supplied. 1,4-Dioxane and te-
trahydrofuran were dried over sodium and benzophe-
none, and distilled just before use from the blue ketyl
form. Acetonitrile was dried over CaH2, and distilled
before use. Methylene chloride, diethyl ether, acetone
and ethanol (95%) were ACS grade (Merck) and were
used as supplied.
Indium analysis was by back-titration of excess
EDTA with thorium nitrate, and bromide analysis by
the Volhard method, as described previously [8]. Micro-
analysis was by Canadian Microanalytical Services Ltd.
Proton and carbon-NMR spectra were recorded on a
Bruker AC300L spectrometer. Thermogravimetric anal-
ysis was carried out on a Shimadzu TGA-50 instru-
ment, under a nitrogen atmosphere (flow rate 20 ml
min−1); the sample (ca. 10 mg) contained in an alumina
cell, was heated over the range 30–800°C at a constant
rate of 10°C min−1. Mass spectra were recorded on a
2.2.2. Br3InCH2E(C6H5)3
InBr (0.50 g, 2.56 mmol) was suspended in 1,4-diox-
ane (20 ml), and CH2Br2 (1.0 ml, 2.48 g, 14.2 mmol)
added via a syringe. The reaction mixture was stirred
until the monobromide dissolved (ca. 3–4 h). All the
volatiles were removed under vacuum, and the mixture
kept under vacuum to ensure complete removal of
excess CH2Br2. This led to the formation of the bis-
dioxane adduct of BrInCH2Br, which was redissolved
in dioxane (20 ml). Solid E(C6H5)3 (E=As, 0.78 g, 2.56
mmol; E=Sb, 0.90 g, 2.56 mmol) was added to the
solution, and the reaction mixture stirred for 6 h to
yield a colorless solution from which the individual
compounds were isolated according to the following
procedures.
2.2.2.1. Br3InCH2As(C6H5)3. The solution from the re-
action was filtered to remove small amounts of solid

L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
205
residue, and solvent removed partially under vacuum to
the onset of crystallisation; on standing, complete pre-
cipitation occurred, giving Br3InCH2As(C6H5)3 (1.40 g,
2.07 mmol, 81%) as cubic crystals, which were dried in
vacuo. Under vacuum, the crystallinity was lost to give
an amorphous white solid, and the analytical results,
and the NMR spectra, are for this latter material.
Suitable single crystals for X-ray analysis were grown
by dissolving the compound in dichloromethane and
reprecipitating by the slow addition of diethyl ether in
a closed vessel. Anal. Found: In, 17.5; Br, 35.5; Calc.
Br, 44.5; Calc. for C19H18Br4InP; In, 16.1; Br, 44.9.
1H-NMR (CDCl3): 7.4–7.9, m, 5H; 2.86, d, JH, JPꢀH
=
13.2 Hz. 18C-NMR (CDC13); 136.3, d, JPꢀH 3.0 Hz,
aromatic para; 133.7, d, JPꢀH=10.8 Hz, aromatic or-
tho; 131.5, d, JPꢀH=13.1 Hz, aromatic meta; 119.0, d,
J
PꢀH=89.0 Hz, aromatic ipso; 11.4, d, JPꢀH=58.7 Hz,
methyl.
2.2.4.2. E=As, X=Br. Br3InCH2As(C6H5)3 (0.41 g,
0.61 mmol) was dissolved in acetonitrile (10 ml) and
HBr (0.144 g, 1.77 mmol, 0.30 ml of 48% aqueous
solution) added. The reaction mixture was stirred for 16
h, and then refluxed for 1.5 h. Removal of all volatiles
under vacuum gave an oil, which was dissolved in hot
ethanol (10 ml); upon cooling, [(C6H5)3AsCH3][InBr4]
(0.33 g, 0.44 mmol, 72%) precipitated as yellow needles,
which were collected under filtration and dried under
vacuo. Anal. Found: In, 14.2; Br, 41.0; Calc. for
1
for C19H17AsBr3In; In, 17.0; Br, 35.5. H-NMR (ace-
tone-d6) 7.7–7.1, m, 15H (C6H5) 2.21, s, 2H(InꢀCH2).
13C 134.51, 133.80, 132.18, 131.15, 130.62, (C6H5) 9.58
(InCH2) ppm.
2.2.2.2. Br3InCH2Sb(C6H5)3. All the volatiles from the
reaction solution were removed under vacuum to leave
a colorless oil; addition of toluene (10 ml) caused the
precipitation of a white solid, which was filtered in
vacuo, and recrystallised in 1:1 v/v acetone–ethanol
(95%, 20 ml) to yield crystals of Br3InCH2Sb(C6H5)3
suitable for X-ray analysis (1.20 g, 1.66 mmol, 65.0%).
Anal. Found: C, 31.8; H, 2.44; In, 15.4; Br, 33.5; Calc.
for C19H17SbBr3In; C, 31.6; H, 2.38; In, 15.9; Br, 33.2.
1
C19H18AsBr4In; In, 15.2; Br, 42.3 H-NMR (CDCl3):
7.6–7.9, m, 5H; 2.81, s, 1H. 13C-NMR (CDCl3): 134.7,
aromatic para; 132.0, aromatic ortho; 131.3 aromatic
meta; 121.0 aromatic ipso; 9.2 methyl.
E=As, X=I. I3InCH2AsPh3 (1.00 g, 1.22 mmol)
and HI (57% in water, 0.16 ml, 0.16 g, 1.22 mmol) were
stirred in acetonitrile (20 ml) for 3 h. The resultant
yellow precipitate was collected, and the mother liquor
concentrated under vacuo to 50% of its initial volume,
when crystals of [(C6H5)3AsCH3][InI4] (0.61 g, 0.64
mmol, 53%) precipitated. Anal. Found: In, 11.8; I, 52.3;
Calc. for C19H18AsInI4; In, 12.2; I, 53.8. 1H-NMR
(CDCl3) 7.92–7.77 m, 5 H (C6H5), 3.20 s, 1H, CH3,
ppm.
2.2.3. I3InCH2As(C6H5)3
InI (0.30 g, 1.23 mmol) and CH2I2 (0.10 ml, 0.33 g,
1.23 mmol) were stirred together in THF (20 ml) at 0°C
to complete dissolution of InI; AsPh3 (0.3 g, 1.23
mmol) was added and the mixture stirred for 2 h.
Volatiles were removed under high vacuo to give an oil
which was recrystalized in acetone–ethanol (25 ml, 1:1,
v/v) yielding I3InCH2AsPh3 (0.75 g, 0.92 mmol, 75%) as
colorless single crystals. Anal. Found: In, 13.6; I, 46.9;
Calc. for C19H17 AsInI3; In, 14.1; I, 46.7. 1H-NMR
(acetone-d6): 2.45 s, 2H (InCH2As), 7.61–7.71m, 15h
(AsPh3) 13C-NMR: 7.87 (InCH2As), 124.18, 129.67,
131.40, 133.14, 134.61 (AsPh3) ppm.
2.2.5. [(C6H5)3 SbCH2Br][InBr4]
Br3InCH2Sb(C6H5)3 (0.30 g, 0.42 mmol) and a 0.5 M
solution of Br2 in CCl4 (0.83 ml, 0.42 mmol) were
stirred together in CH2Cl2 (10 ml) at r.t. for 12 h, by
which time the red color of the bromine had dis-
charged. Volatiles were removed under vacuum to give
a brown solid, which was recrystallised in acetone–
ethanol (95%; 1:1, v/v) to yield [(C6H5)3-
SbCH2Br][InBr4] (0.27 g, 0.31 mmol, 74%). Anal.
Found: In, 13.5; Br, 45.6; Calc. for C19H17Br5InSb; In,
13.0; Br, 45.3. 1H-NMR (acetone-d6) 7.75–7.97, m,
15H; 5.14, s, 2H. 13C-NMR: 136.2; 134.1; 131.2; 123.9;
13.3 ppm.
2.2.4. [(C6H5)3ECH3][InX4]
2.2.4.1. E=P, X=Br. InBr (0.41 g, 2.10 mmol) and
CH2Br2 (0.5 ml, 1.24 g, 7.1 mmol) were stirred in
acetonitrile (10 ml) until the InBr dissolved. Volatiles
were removed under vacuo, and the product kept under
vacuum to ensure removal of excess CH2Br2; the resul-
tant oil was redissolved in CH3CN (10 ml) and
P(C6H5)3 (0.55 g, 2.10 mmol) added to this solution,
which was stirred for 1 h, after which HBr (170 mg,
2.10 mmol, 0.35 ml of 48% aqueous solution) was
added and the reaction stirred for 1 h. Volatiles were
again removed under vacuum, and the resultant white
solid recrystallised in 20 ml of acetone–ethanol (1:1,
v/v) to yield [(C6H5)3PCH3][InBr4] (0.58 g, 0.81 mmol,
39%) as colorless microcrystals. Anal. Found: In, 15.8;
2.2.6. Attempted reactions with electrophiles
In addition to the reaction with HBr and Br2 de-
scribed above, we investigated the possible reactions of
the Group 15 adducts of Br2InCH2Br with a series of
electrophiles (benzyl chloride and bromide, allyl bro-
mide and iodide, trityl bromide, p-chlorobenzaldehyde).
No reaction was detected in any of these systems after
refluxing for some hours in THF or dioxane, and the
X3InCH2L
compounds
(X=Br,
L=N(C2H5)3,

206
L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
Table 1
Summary of crystal data, intensity collection and structural refinement for compounds 1–4
Compound
1
2
3
4
Formula
fw Mr
C7H17NInBr3
469.77
C
674.80
19H17AsInBr3
C19H17SbInBr3
721.63
C19H17AsInI3
815.77
Crystal system
Space group
Monoclinic
P21/n
Orthorhombic
Pbca (No.61)
15.553(7)
21.646(8)
12.920(10)
Orthorhombic
Pbca (No.61)
15.439 (6)
22.016(4)
13.138(5)
Monoclinic
P21/n
,
a (A)
7.362(2)
14.700(2)
13.049 (1)
98.90(1)
1395.1(4)
2.237
12.0122(2)
15.8526(3)
13.4180(3)
109.933(1)
2402.04(8)
2.256
,
b, (A)
,
c, (A)
i (°)
3
,
V (A )
4349.4(3.8)
2.06
4465.7(2.3)
2.41
dcalc (g cm−3
)
Z
4
8
8
4
Crystal dimensions (mm)
Absorbtion coefficient (mm−1
Temperature (°C)
0.2×0.2×0.2
10.25
25
0.1×0.1×0.1
62.35
23
0.1×0.1×0.1
50.45
23
0.1×0.1×0.1
6.21
25
)
Data collection range (2q) (°)
Total number of rflections measured
2.10–28.33
4052
−50
4120
−45
4059
1.97–25.0
12 103
Number of unique data variables
R
2535:110
0.0418
717:77
0.064
911:123
0.062
4189:122
0.0481
Rw
0.0368
0.054
0.049
0.0431
max D/l in final cycle
0.03
1.36, −0.89
0.01
1.61, −0.93
−3
,
zmax in final difference map (e A
)
0.98, −0.56
1.80,−0.82
As(C6H5)3, Sb(C6H5)3; X=I, L=As(C6H5)3) were re-
covered at the end of this treatment.
crystals to a quartz fiber mounted on a Siemens SMART
three-circle diffractometer equipped with graphite-
monochromatized Mo–Ka radiation and a CCD area
detector, and controlled by a Pentium-based PC run-
ning the SMART software package [16]. Omega-scans
were used in such a way that an initial 180° scan range
consisting of 0.3° intervals was followed by three fur-
ther 120, 180, and 120° scans with ƒ offsets of 88, 180,
and 268°, respectively. This strategy samples the sphere
of reciprocal space up to 2q=56.62°. Cell parameters
were refined using the centroid values of 300 reflections
with 2q angles up to 56.62°. Raw frame data were
integrated using the SAINT program [17]. The structure
was solved by direct methods [18]. An empirical absorp-
tion correction was applied to the data using the pro-
gram SADABS [19]. For compound 4, only indium,
iodine and arsenic atoms were refined anisotropically.
Figs. 1 and 2 show the structures of 1 and 4; 2 and 3
are essentially isostructural with 4, as is the
2.3. Crystallographic studies
The procedures for Br3InCH2As(C6H5)3 (2) and
Br3InCH2Sb(C6H5)3 (3) were essentially identical. Suit-
able crystals were mounted on glass fiber and X-ray
diffraction data were collected on a Rigaku AFC6S
instrument, with graphite monochromated Mo–Ka ra-
,
diation (u=0.71069 A). Cell constants and orientation
matrices for data collection were obtained using 25
strong reflections in the range of 22.54B2qB32.35°.
The space groups were identified from the systematic
absences, and confirmed subsequently by the successful
structure refinements. Decay and absorption correc-
tions (PSI scan) were applied, and the data were also
corrected for Lorentz and polarization effects. The
structures were solved by direct methods. For com-
pound 3, the indium, bromine and antimony atoms
were refined anisotropically, and all others isotropi-
cally; hydrogen atoms were introduced initially in ideal
positions and subsequently refined isotropically. Crys-
tals of compound 2 diffracted poorly, and in order to
solve the structure with the available data it was neces-
sary to treat the phenyl rings as units, using the pro-
,
grammed
parameters
R(CꢀC)=1.40
A
and
CꢀCꢀC=120°. The programs and other data used are
listed in [10–15], and details of the data collection are
given in Table 1.
For Br3InCH2N(C2H5) (1) and I3InCH2As(C6H5)3
(4), diffraction data were collected by attaching single
Fig. 1. Molecular structure of Br3InCH2N(C2H5)3, 1 (ORTEP dia-
gram), with atoms shown as 30% probability ellipsoids.

L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
207
Table 3
Calculated total energies (in eV)
Energy
a,b
Br3InCH2N(C6H5)3
Br3InCH2P(C6H5)3
Br3InCH2As(C6H5)3
3792.53701, 3792.22379
3752.31714, 3751.70850
3755.17799, 3754.65526
3779.81651, 3779.11470
1266.80743
a,b
a,b
a,b
Br3InCH2Sb(C6H5)3
Br2InCH2Br c
c
Br3InCH2
InBr3
1265.90056
1118.18624
c
N(C6H5)3
2524.36436
P(C6H5)3
2482.84111
As(C6H5)3
2486.62269
Sb(C6H5)3
2512.91410
H2CN(C6H5)3
H2CP(C6H5)3
H2CAs(C6H5)3
H2CSb(C6H5)3
2671.09632
2631.02178
2633.94666
2658.06945
Fig. 2. Molecular structure of I3InCH2As(C6H5)3, 4, (ORTEP dia-
gram), with atoms shown as 30% probability ellipsoids. Hydrogen
atoms have been omitted for clarity.
a See Section 2.
b See Section 2.
c Values from [9].
triphenylphosphine derivative. The structural data rele-
vant to the present paper are given in Table 2. Com-
plete tables of bond distances and angles, atomic
coordinates and thermal parameters are available as
supporting information.
were those determined by X-ray structural studies.
There are two entries in Table 3 for the calculated total
energy for each adduct. For (a), the structure was
allowed to refine freely, but this procedure resulted in
models with InꢀCꢀE angles of ca. 90°, much smaller
than the experimental values (ca. 115–118°); no other
significant differences from the experimental values
were observed. Procedure (b) gives the total energies
obtained by constraining the InꢀCꢀE bond angles close
to the experimental values, and we consider only those
results obtained by this latter procedure in the
discussion.
2.4. Calculation methods and models
The PM3 semiempirical method [20,21] implemented
in the MOPAC (version 5.0) program package [22] was
used with standard indium parameters for this proce-
dure. The geometry optimization of the molecules via
gradients and minimization of energy were carried out
using the BFGS algorithm, which makes use of first
and second derivatives. The initial structural parame-
ters for the Br3InCH2E(C6H5)3 (E=P, As, Sb) models
For all the calculated structures, the InꢀBr bond
distances are considerably shorter than the experimen-
tal values. This difference is inherent in the PM3
method, and has been discussed earlier [21,22]. The
Table 2
,
Important bond distances (A) and angles (°)
a
Br3InCH2NEt3
Br3InCH2PPh3
Br3InCH2AsPh3
Br3InCH2SbPh3
I3InCH2AsPh3
Bond length
InꢀX b (av)
InꢀCH2
2.529(1)
2.185(5)
1.507(6)
1.519(1)
2.512(2)
2.18(1)
1.79(1)
1.80(1)
2.508(6)
2.17(5)
1.97(5)
1.89(2)
2.509(5)
2.16(3)
2.03(3)
2.14(3)
2.724(1)
2.184(7)
1.924(6)
1.929(7)
c
EꢀCH2
EꢀC(R) c(av)
Bond angles
XꢀInꢀX b,d
105.3(1)[5.1]
113.4(2)
119.2(4)
109.9(4)[1.7]
109.0(4)[4.8]
106.8(1)[5.2]
112.1(3)[12.0]
118.9(6)
110.1(5)[1.9]
108.8(6)[2.0]
106.7(2)[5.5]
112(1) [12]
115(2)
111(1) [3]
108(1) [3]
106.4[4.3]
112.4[5.0]
118(2)
110(1)[1]
108(1)[3]
108.7(1)[1.5]
110.0(2)[14.7]
119.8(3)
110.2(3)[2.2]
108.0(3)[2.9]
b,d
XꢀInꢀCH2
InꢀCH2ꢀE c
CH2ꢀEꢀC(R) c,d
C(R)ꢀEꢀC(R) c,d
a Results from [3].
b X=Br, I.
c E=N, P, As, Sb.
d Parentheses indicate typical experimental uncertainty; value in square brackets is the difference between highest and lowest values.

208
L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
reasons for the effect are not clear, but two relevant
factors are (i) the exchange and coulombic integrals at
the zero differential overlap approximation are ex-
cluded, and (ii) the method considers only indium s-
and p-orbitals, and so neglects any contributions from
metal d-orbitals, which may be important in the stabili-
sation of these molecules.
for adducts for Br2InCH2Br [3,9], but lower than in
Inꢀ(CH2)nꢀIn compounds [23,24]. The EꢀCH2 bond
distances show a dependence on E, which is discussed
below.
The r(InꢀBr) distances in the three triphenyl deriva-
tives are constant within the experimental uncertainty,
,
and close to the value of 2.518 A (av) found for the
anion Br3InCH2Br−, but significantly lower than in the
neutral
adduct
Br2(THF)2In(CH2)In(THF)2Br2
,
3. Results and discussion
(r(InꢀBr)=2.517(1), 2.544(1) A) [23] and the corre-
sponding anion [Br3In(CH2)4InBr3]2− (r(InꢀBr)=
,
3.1. Structural results
2.543(3) A (av)) [24]. There is little difference between
the structures of Br3InCH2As(C6H5)3 and I3InCH2As-
The four molecules whose structures are reported
here have a number of important features in common
with each other, and with the previously reported
Br3InCH2P(C6H5)3 [3]. The summary of bond angles in
Table 2 shows that the stereochemistry at indium and
at the Group 15 atom is distorted tetrahedral in both
cases. The InꢀCH2ꢀE angles are also reasonably con-
stant in all five molecules, at 11892°, and the bond
distances are in keeping with these consistencies. In
(C6H5)3, apart from the indiumꢀhalogen values.
3.2. Quantum mechanical calculations
Table 3 gives the calculated total energies of the four
molecules Br3InCH2E(C6H5), including the hypothetical
E=N species, and for various related molecules. The
two sets of values for the adducts refer to the different
methods of calculation described in Section 2.4. Table 4
presents the calculated energy levels for the HOMO
and LUMO orbitals in the adducts. The results in
Table 3 lead to calculated DEr values for adduct forma-
tion, by the process
,
particular, the InꢀCH2 length of 2.1890.01 A is essen-
tially constant, and in agreement with previous results
Table 4
Calculated HOMO and LUMO energy levels (in eV)
Br2InCH2Br+E(C6H5)“Br3InCH2E(C6H5)3
(1)
Structure
HOMO
LUMO
Br3InCH2N(C6H5)3
Br3InCH2P(C6H5)3
Br3InCH2As(C6H5)3
Br3InCH2Sb(C6H5)3
−11.590
−13.556
−10.318
−9.906
−9.009
−10.520
−3.630
−5.155
−1.486
−1.270
−1.811
−3.825
of N= −1.052, P= −2.060, As= −1.225 and Sb=
+0.607 eV, or −101, −198, −118 and +59 kJ
mol−1. The value for the nitrogen system predicts that
a stable ylid Br3InCH2N(C6H5)3 should be accessible; in
fact this compound could not be obtained, which may
reflect the steric crowding at N in triphenylamine, since
the analogous triethylamine compounds has been pre-
pared. Equally, the formation of the Sb(C6H5)3 adduct
is predicted to be endothermic, whereas the compound
is in fact stable. The calculated energies for Eq. (1) are
probably at the limit of the reliability of the calcula-
tions, so that conclusions based on small differences
between large values are not reliable. On the other
hand, the calculated bond lengths and angles, calcu-
lated net atomic charges, bond orders and atomic or-
bital populations are given in Tables 5 and 6, and the
agreement between the values in Table 5 and the exper-
imental values in Table 2 is satisfying. The differences
in bond lengths and angles are below 5%, except for the
InꢀBr distances, which seems to be an inherent
difficulty in this method (see above). The good fit
speaks to the reliability of the calculated parameters.
One very significant feature of the Br3InCH2E(C6H5)3
structures is the calculated charge distribution (Table
6), which reveals a considerable negative charge at the
carbon atom bonded to the metallic center (as high as
−0.6 for the phosphorus and arsenic compounds), and
a
Br2InCH2Br
a
Br3InCH2
a Values from [9].
Table 5
,
Calculated bond distances (A), bond angles (°), and bond orders, for
Br3InCH2E(C6H5)3 (E=N, P, As, Sb)
N
P
As
Sb
Bond length
InꢀC
EꢀCH2
EꢀCar
2.18
1.50
1.52
2.18
2.16
1.72
1.76
2.20
2.16
1.83
1.87
2.20
2.14
2.18
2.12
2.19
InꢀBra
Bond angles
InCH2
BrꢀInꢀCH2
H2CEꢀCar
112.7
111.5
110.0
120.1
112.6
111.1
112.1
112.3
110.7
116.2
110.9
112.0
Bond orders
InꢀE
0.89
0.89
0.90
0.90
InꢀCH2
EꢀCH2
0.91
0.98
0.95
0.91
0.91
0.83
0.89
0.95
0.86
0.91
0.69
0.77
a
EꢀCar
a Mean value.

L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
209
Table 6
The calculated energies in Table 3 allow DE to be
Atomic orbital electron populations and net atomic charges in
Br3InCH2E(C6H5)3 (E=N, P, As, Sb)
derived for the processes
Br3InCH2E(C6H5)3“InBr3+H2Cl−ꢀl+E(C6H5)3 (2)
Populations
N
P
As
Sb
For E=P, DE=2.500, As=2.522, and Sb=2.859
(in eV), or 241, 243 and 276 kJ mol−1, which suggests
that decomposition will only occur at high tempera-
tures, as indeed was observed in the TGA experiments
(see below), in which decomposition was detected above
170°C for all three species. The data further imply that
free ylids are unlikely to be involved as readily accessi-
ble intermediates in reactions starting with the
Br3InCH2E(C6H5)3 compounds.
Atom (orbital)
In(s)
In(px)
In(py)
In(pz)
C(s)
C(px)
C(py)
C(pz)
E(s)
E(px)
E(py)
E(pz)
0.913
0.750
0.680
0.638
1.221
1.106
0.924
1.118
1.176
0.942
0.998
0.939
0.919
0.741
0.669
0.621
1.222
1.136
1.107
1.125
1.222
0.621
0.658
0.599
0.913
0.924
0.728
0.673
0.617
1.221
1.112
1.119
1.118
1.183
0.700
0.721
0.665
0.743
0.677
0.635
1.231
0.980
0.805
1.062
1.836
0.771
0.864
0.757
3.3. Spectroscopic results
The NMR results for the CH2 protons in the
Br3InCH2E(C6H5)3 compounds are collected together in
Table 7, and compared with the calculated atomic
charge at the carbon atom in question; results are also
shown for the ether, 1,4-dioxane and bromide deriva-
tives of Br2InCH2Br, in which the ligand is known to be
bonded to the indium atom [9]. There is a clear rela-
Net Charges
Br a
−0.30
0.02
−0.37
0.94
0.14
−0.10
0.13
−0.31
0.05
−0.59
1.90
0.15
−0.14
0.11
−0.30
0.07
−0.57
1.73
0.15
−0.13
0.12
−0.30
0.02
−0.08
0.77
0.11
−0.10
0.12
In
C
E
H(CH2) a
a,b
Car
Har
a,b
1
tionship between the observed H chemical shift and
a Mean value.
the net atomic charge at the methylene carbon atom in
the E(C6H5)3 derivatives, so that for E=As and Sb, the
resonances at 2.21 and 2.83 ppm reflect the higher
electronic density at the carbon of the arsenic deriva-
tive. Equally, the antimony compound, in which the
atomic charge on the methylene carbon is similar to
those in the ether, dioxane and bromide adducts, shows
a proton resonance close to that in to these compounds,
and Br3InCH2SC(NMe2)2 and the triphenylarsenic ad-
duct have similar chemical shifts (2.44 ppm [3,8]) and
structures. We were not able to detect the carbon
resonance of the antimony derivative. This is a com-
mon problem in organoindium compounds, in which
there is believed to be strong coupling with the indium
atom (I=9/2), and in addition antimony itself has two
high-spin isotopes (121Sb, I=5/2 and 123Sb, I=7/2).
The 13C resonances of the methylene group of the
phosphorus and arsenic derivatives, at 10.20 and 9.58
ppm, are at considerably higher field than that in the
parent molecule Br2InCH2Br. 1/2 Et2O (17.6 ppm) [3],
again agreeing with the increased electron density at the
methylene carbon of the Br3InCH2E(C6H5)3 com-
pounds, as predicted by the calculations.
b ar=aromatic.
Table 7
1H- and 13C-NMR chemical shifts (ppm; Me4Si=0) and net atomic
charges of the methylene group in Br3InCH2E(C6H5)3 and
Br2In(L)nCH2Br compounds
Compound
Net atomic charge
1H
13C
a
Br3InCH2P(C6H5)3
−0.59
−0.57
−0.08
nc e
−0.07
−0.01
2.30
2.21
2.83
3.07
2.77
2.75
10.20
9.58
b
b
Br3InCH2As(C6H5)3
Br3InCH2Sb(C6H5)3
nd f
Br2In(Et2O)1/2CH2Br c,d
Br2In(diox)2CH2Br a,d
[Br3InCH2Br]− a,d
17.6
18.96
18.30
a In CD3CN.
b In (CD3)2CO.
c NMR spectra in CD2Cl2 solution.
d Results from [9].
e nc, not calculated.
f nd, not detected.
the increased positive electron density at the het-
eroatom E, all of which establishes unequivocally the
presence of the ylid H2Cl−ꢀl+X(C6H5)3 in the
Br3InCH2E(C6H5)3 structures. This is a particularly in-
teresting conclusion when one considers that the
organoindium ylids are formed from a starting mate-
rial, Br2InCH2Br, in which a great deal of positive
3.4. Mass spectra and thermal decomposition
The E.I. mass spectra of the Br3InCH2E(C6H5)3 com-
pounds are given in Table 8. No molecular ions were
detected, but the fragment [Br2InCH2E(C6H5)3]+ result-
ing from the rupture of one indiumꢀbromine bond was
observed in each case, as was the ylid [H2CꢁE(C6H5)3]+,
which results from the fission of the indiumꢀcarbon
bond
charge is situated on the carbon atom, Br3Inl−ꢀ
l+
CH2, in one tautomeric form.
The formation of the free ylids, H2Cl−ꢀl+E(C6H5)3,
from the organoindium complex is not an easy process.

210
L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
BrInCH2E(C6H5)3“Br3In+[CH2ꢁE(C6H5)3]+
(3)
group (calc. 63.5%); the loss of 11.8% corresponds to
removal of the third C6H5 group (calc 11.4%), and the
final residue of 10.0% is assigned to elemental arsenic
(calc. 10.0%).
The thermal decomposition pathway, together with
the fragmentation pattern in the mass spectrum of
Br3InCH2As(C6H5)3 suggests the thermal decomposi-
tion mechanism:
Additional observed fragments are all derived from
the [CH2=E(C6H5)3]+ylid, and include [E(C6H5)3]+,
[E(C6H5)2]+, [E(C6H5)]+, [H5C6ꢀC6H5]+, and [C6H5]+
.The TGA curves, showing weight loss as a function of
temperature, were obtained for all the Br3InCH2E-
(C6H5)3 compounds. Fig. 3 depicts the result for the
arsenic compound, which is the best behaved example.
There are three separate stages of weight loss, over the
temperature ranges 178–326°C (14.5% loss), 326–
453°C (63.7% loss), and finally an 11.8% loss at 453–
552°C. The residual mass is equal to 10.0% of the initial
quantity of adduct. The first stage corresponds to loss
of CH2 and C6H5 groups (calc 13.5%), and that of
63.7% to the removal of InBr3 and a second C6H5
Br3InCH2As(C6H5)3(s,1)
“Br3In(s,1)+H2CꢁAs(C6H5)3(s,1)
H2CꢁAs(C6H5)3(s,1)
(4)
“As(C6H5)2(s,1)+1/n(CH2)n(g)+1/2(C6H5)2(g)
As(C6H5)2(s,1)“As(C6H5)(s,1)+1/2(C6H5)2(g)
(5)
(6)
Table 8
Mass spectrometric results: m/e, with relative abundances (%) in parentheses, for Br3InCH3E(C6H5)3(E=P, As, Sb)
Fragment
P
As
Sb
639(0.8)–641(1.9)–643(1.9)–645(0.8)
366(8.3)–368(6.00)
365(13.3)–367(11.9)
[Br2InCH2E(C6H5)3]+
[H2CꢁE(C6H5)3]+
[HCꢁE(C6H5)+3 or
H2CꢁE(C6H5)2C6H4]+
[E(C6H5)3]+
549(4.6)–551(8.0)–553(3.8)
276(44.6)
593(7.1)–595(12.8)–597(4.4)
320(35.2)
275(100)
319(80.9)
262(9.5)
185(11.9)
183(42.6)
nd a
306(29.1)
229(15.6)
227(56.8)
152(100)
154(10.0)
77(1.4)
352(8.3)–354(6.5)
275(19.9)–277(8.0)
273(21.1)
198(100)–200(66.3)
154(50.7)
[(EC6H5)2]+
[E(C6H5)]+
[E(C6H5)]+
[(C6H5)2]+
nd a
77(6.4)
[C6H5]+
77(17.5)
a nd=not detected.
Fig. 3. TGA curve for Br3InCH2As(C6H5)3.

L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
211
As(C6H5)(s,1)“As+1/2(C6H5)2(g)
(7)
3.5. Preparati6e and mechanistic aspects
The Group 15 ylids CH2E(C6H5)3, are strong nucle-
ophiles, and are known to react with suitable electron
pair acceptors, and we therefore studied the reaction of
the organoindium ylids with the simplest electrophile,
namely the proton, in order to verify the possibility of
using these complexes as ylid sources. We were success-
ful in preparing the [CH3E(C5H6)3][InBr4] salts for E=
P, As, by the process
Scheme 2.
at this atom. With HBr, the cleavage of the InꢀCH2E
bond is in agreement with the known chemistry of
organoindium compounds, and in this the small elec-
trophile (H+) in the highly polar HBr molecule inter-
acts favorably with the InꢀC bond. The absence of a
reaction between HBr and the antimony ylid can be
correlated with the near lack of polarity in the InꢀC
bond in this latter molecule (see Table 6). These exam-
ples apart, attack on Br3InCH2E(C6H5)3 by bulky elec-
tophiles will be hindered sterochemically, as well as
being energetically unfavorable.
Br3InCH2E(C6H5)+HBr“[CH3E(C6H5)3] [InBr4]
(8)
The onium salts were characterised by elemental
analysis and NMR spectroscopy (see Section 2); the
spectrum of the phosphorus derivative compares well
with an authentic sample of the triphenylmethylphos-
phonium cation, including the PꢀH coupling constants
reported as 88.6 (ipso), 10.7 (ortho), 12.9 (meta), 3.0
(para) and 57.1 Hz (methyl) [23]. The compound
I3InCH2As(C6H5)3 undergoes a reaction similar to Eq.
(8).
4. Conclusions
The a-halogeno-indium carbenoids, of general struc-
ture X2InCR1R2X are of considerable interest in or-
ganic synthesis. The results described here and
elsewhere clearly characterise two electron deficient
sites in these compounds, namely the metal center and
the carbon atom. The lack of electron density at the
carbon is a result of a strong interaction between the
halogen and the metallic center (Scheme 3) and the
ultimate result of halide transfer is the tautomeric struc-
ture IB, which has an electrophilic carbene group.
Reaction with a series of selected ligands, L gives
organoindium ylids of the general structure,
X3InCR1R2L, particularly when L is ligand capable of
accommodating a positive charge, such as amine, phos-
phine, arsine, stibine, etc. and one result of this process
is therefore an unpollung of the carbene from the initial
electrophilic X2InCR1R2X compound to the final nucle-
ophilic organoindium ylid X3InCR1R2L. Calculations
in this and earlier work confirm these effects, and
equally provide a rationale for the spectroscopic prop-
erties of these ylids, for their thermal decomposition
reactions, and for the lack of reactivity with bulky
electrophiles.
The antimony derivative does not react with HBr
under the conditions used succesfully for the phospho-
rus and arsenic analogues, but the reaction of
Br3InCH2Sb(C6H5)3 with bromine gave the addition
product formulated as [(C6H5)3SbCH2Br][InBr4]. The
NMR spectrum of this product (Section 2) shows the
SbꢀCH2ꢀBr resonances at 5.14 ppm (1H) and 13.3 ppm
(13C); there is a considerate shift downfield for the
methylene protons, compared to the parent molecule
(2.83 ppm), consistent with the formation of the stibo-
nium cation[BrCH2Sb(C6H5)3]+ by the reaction
Br3InCH2(C6H5)3+Br2“[BrCH2Sb(C6H5)3][InBr4]
(9)
These reactions with HBr and Br2 were the only
successful results in the attempts at reacting
Br3InCH2E(C6H5)3 adducts with electrophiles. There
are two ways of understanding these results. The first is
to consider the high energies associated with Eq. (2);
for all three adducts, the formation of the appropriate
ylid is highly endoergic, in agreement with the ther-
mogravimetric results, so that the energetics of a reac-
tion between the adduct and an electrophile are
thermodynamically unfavourable. The second approach
is to consider the charge distribution in the InCh2E core
of the adducts, using the results in Table 6. For the
phosphorus compound, for example, one has Scheme 2
with a considerable degree of negative charge on the
carbon atom, and a large positive charge on phospho-
rus, so that the carbon is surrounded by four positive
centers which will inhibit attack by large electrophiles
Scheme 3.

212
L. de Arau´jo Felix et al. / Journal of Organometallic Chemistry 603 (2000) 203–212
[7] J.A. Nobrega, C. Peppe, Z. Tian, M.A. Brown, D.G. Tuck,
Acknowledgements
unpublished results.
[8] A.C. Souza, C. Peppe, Z. Tian, M.A. Brown, D.G. Tuck,
Organometallics 12 (1993) 3354.
[9] A.B. de Carvalho, M.A.M.A. de Maurera, I.A. Nobrega, C.
Peppe, M.A. Brown, D.G. Tuck, M.Z. Hernandes, E. Longo,
F.R. Sensato, Organometallics 18 (1999) 99.
[10] C.J. Gilmore, J. Appl. Crystallogr. 17 (1984) 42 MITHRIL.
[11] P.T. Beurskens, DIRDIF. Technical Report, 1, Crystallography
Laboratory, Nijmegen, Netherlands, 1984, p. 1984.
[12] International Tables for X-ray Crystallography, Kynock Press,
Birmingham, UK, 1974. Tables 2.2A, 2.3.1.
[13] J.A. Ibers, W.C. Hamilton, Acta Crystallogr. 17 (1964) 781.
[14] TEXAN-TEXRAY Structure Analytical Package; Molecular Struc-
ture Corp. Houston, TX, 1985.
[15] C.K. Johnson, ORTEP II. Report ORNL-5138, Oak Ridge Na-
tional Laboratory, Oak Ridge, TN, 1976.
This work was supported in part by Operating
Grants (to C.P.) form the Conselho Nacional de Desen-
volvimento Cientifico e Tecnologico — CNPq CAPES
and PAPESQ — PB Borazil). Two of us (L.A.F. and
R.K.K) thank the CNPq for the award of scholarships
(PIBIC-UFPb program). Support from the Natural Sci-
ences and Engineering Research Council of Canada (to
D.G.T.) is also acknowledged. We thank Professor
D.W. Stephan and Dr Glenn Yap (University of Wind-
sor) for the X-ray crystallographic studies of com-
pounds 1 and 4.
[16] SMART: Data Collection Software, Version 4.050; Siemens Ana-
lytical X-ray Instruments Inc.: Madison, WI, 1996.
[17] SAINT: Data Reduction Software, Version 4.050; Siemens Ana-
lytical X-ray Instruments Inc. Madison, WI, 1996.
References
[18] SHELTXL 5.04/VMS: An Integrated System for Solving, Refining
and Displaying Crystal Structures from Diffraction Data;
Siemens Analytical X-ray Instruments Inc. Madison WI, 1995.
[19] G.M. Sheldrick, SADABS: Empirical Absorption Correction Pro-
gram, University of Gottingen, Gottingen, Germany, 1996.
[20] J.P.P. Stewart, J. Compt. Chem. 10 (1989) 209.
[21] J.P.P. Stewart, Method J. Compt. Chem. 10 (1989) 221.
[22] J.P.P. Stewart, Quantum Chem. Prog. Exch. Bull 3 (1983) 43.
[23] M. Tschinkl, A. Schier, J. Riede, E. Schmidt, G.P. Gabbai,
Organometallics 16 (1997) 4759.
[1] S. Araki, Y. Butsugan, J. Chem. Soc., Chem. Commun. (1989)
1286
[2] S. Araki, T. Hirashita, K. Shimizu, K.T. Ikeda, Y. Butsugan,
Tetrahedron 52 (1996) 2803.
[3] T.A. Annan, D.G. Tuck, M.A. Khan, C. Peppe, Organometallics
10 (1991) 2159.
[4] J.E. Santos, C. Peppe, M.A. Brown, D.G. Tuck, Organometal-
lics 15 (1996) 2201.
[5] L.N.H. Arakaki, M.Sc. thesis, Universidade Federal da Paraiba,
Brazil.
[6] C. Peppe, D.G. Tuck, unpublished results.
[24] M.A. Khan, D.G. Tuck, Acta Crystallogr. Sect. B. Struct. Sci.
B38 (1982) 803.
.