The hexacyanotitanate ion: Synthesis and crystal structure of [NEt4]3[Ti(III)(CN)6]·4MeCN

Article abstract of DOI:10.1021/ja962773m

The hexacyanotitanate salt, [Et₄N]₃[Ti(CN)₆]·4MeCN, has been prepared by addition of tetraethylammonium cyanide to the titanium(III) triflate salt Ti(O₃SCF₃)₃(MeCN)₃. The orange crystalline product has been characterized by X-ray diffraction, and the d¹ anion is only slightly distorted from ideal O(h) symmetry. The anion resides on a center of symmetry and is characterized by the following parameters: Ti-C = 2.195(2), 2.197(3), and 2.213(3) A?; C-N (av) = 1.141(4) A?; C-Ti-C (cis) = 88.01(9), 88.02(9), 89.02(9), and 89.78(9)°; C-Ti-C (trans) = 180°. In addition to the crystallographic study, details of the IR (ν(CN) = 2071 cm^-1), EPR, and UV-vis spectra (Δ(o) = 22800 cm^-1) are given. Crystal data for [Et₄N]₃[Ti(CN)₆]·4MeCN are as follows: monoclinic, space group I2/a, a = 18.171(6) A?, b = 12.200(4) A?, c = 20.989(5) A?, β = 91.17(2)°, V = 4652(2) A?³, Z = 4, wR² = 0.2054 for 3831 data, 27 restraints, and 318 parameters.

Full text of DOI:10.1021/ja962773m

J. Am. Chem. Soc. 1997, 119, 6251-6258
6251
The Hexacyanotitanate Ion: Synthesis and Crystal Structure of
III
[
NEt ] [Ti (CN) ]‚4MeCN
4
3
6
William R. Entley, Christopher R. Treadway, Scott R. Wilson, and
Gregory S. Girolami*
Contribution from the School of Chemical Sciences, The UniVersity of Illinois at
UrbanasChampaign, 601 South Goodwin AVenue, Urbana, Illinois 61801
X
ReceiVed August 8, 1996. ReVised Manuscript ReceiVed April 28, 1997
Abstract: The hexacyanotitanate salt, [Et4N]3[Ti(CN)6]‚4MeCN, has been prepared by addition of tetraethylammonium
cyanide to the titanium(III) triflate salt Ti(O3SCF3)3(MeCN)3. The orange crystalline product has been characterized
1
by X-ray diffraction, and the d anion is only slightly distorted from ideal Oh symmetry. The anion resides on a
center of symmetry and is characterized by the following parameters: Ti-C ) 2.195(2), 2.197(3), and 2.213(3) Å;
C-N (av) ) 1.141(4) Å; C-Ti-C (cis) ) 88.01(9), 88.02(9), 89.02(9), and 89.78(9)°; C-Ti-C (trans) ) 180°. In
-
1
addition to the crystallographic study, details of the IR (νCN ) 2071 cm ), EPR, and UV-vis spectra (∆o ) 22800
-
1
cm ) are given. Crystal data for [Et4N]3[Ti(CN)6]‚4MeCN are as follows: monoclinic, space group I2/a, a )
3
2
1
8.171(6) Å, b ) 12.200(4) Å, c ) 20.989(5) Å, â ) 91.17(2)°, V ) 4652(2) Å , Z ) 4, wR ) 0.2054 for 3831
data, 27 restraints, and 318 parameters.
Introduction
(CN)6] (which should be soluble in water as judged from the
properties of other hexacyanometalates); instead he proposed
that the product was Ti(OH)3. In 1963, Heintz reported that
the reaction of stoichiometric amounts of titanium trichloride
and potassium cyanide in water, under rigorously anaerobic
The preparation of new molecular-based magnetic materials
has become one of the fastest growing areas of inorganic and
materials chemistry today.1
-3
Recent work has convincingly
shown that molecular magnets with magnetic ordering temper-
atures approaching (and sometimes exceeding) 300 K can be
prepared by using cyanometalates of stoichiometry [M(CN)6
III
conditions, gave a deep blue product of stoichiometry K3[Ti -
7
(
CN)6] as judged from microanalytical data. When the reaction
n-
]
was carried out in the presence of excess cyanide, a deep blue
as molecular building blocks.4 Such anions can be treated with
a variety of paramagnetic transition metal cations to prepare
magnetic materials that are analogues of the well-known solid
Prussian blue. Generally, solids with higher magnetic ordering
temperatures are obtained if early transition metal cyanometa-
,5
III
product of stoichiometry K3[Ti (CN)6]‚KCN was obtained.
Successive washings with distilled water gave the original
III
reaction product K3[Ti (CN)6]. Heintz did not comment on
the unexpected finding that his product was insoluble in water.
In 1961 Schl a¨ fer and G o¨ tz reported that titanium tribromide
and potassium cyanide reacted in liquid ammonia to give a dark
4-
3-
lates such as [V(CN)6 ] and [Cr(CN)6 ] are employed as
building blocks.4,5 In part, this correlation reflects the higher
III
8
green complex which analyzed well for K5[Ti (CN)8]. The
authors argued that the dark green product was an octahedral
energies of the d-orbitals in early transition metals and the
consequent greater degree of delocalization of spin density onto
the cyanide ligands.
In view of this correlation, it would be of interest to prepare
titanium-substituted analogues of Prussian blue. Before this can
be accomplished, however, it is necessary to synthesize a
hexacyanotitanate building block. Hexacyanometalate com-
III
cyanide complex best represented by the formula K3[Ti -
(CN)6]‚2KCN. In 1962 Golding and Carrrington suggested that
this dark green complex was actually eight-coordinate based
on the similarity of its EPR spectrum to those of the known
V
V
eight-coordinate cyano compounds K3[W (CN)8] and K3[Mo -
9
III
n-
(CN)8]. In 1969 Bose and Basu concluded that K5[Ti (CN)8]
was eight-coordinate based on magnetic susceptibility and UV-
vis data, but that the EPR spectrum was more consistent with
plexes of stoichiometry [M(CN)6 ] have long been known for
every first-row transition metal between vanadium and cobalt;
in contrast, although the reactions of titanium(III) with cyanide
salts have been the subject of much discussion, the identities
of the products obtained are far from firmly established.
In 1906 Grossman reported that titanium trichloride and
1
0
a six-coordinate structure.
More recent investigations of the reaction products of
titanium(III) and cyanide have been carried out by Nicholls and
11
co-workers. In 1974, these authors repeated Heintz’s aqueous
preparation but found that the only titanium product obtained
was the dark blue titanium(III) hydroxide. Nicholls et al. also
reinvestigated the work of Schl a¨ fer and G o¨ tz; they suggested
excess potassium cyanide react in aqueous solution to give an
immediate black precipitate.6 No analytical data were presented,
III
but Grossman stated that the precipitate could not be K3[Ti -
X
Abstract published in AdVance ACS Abstracts, June 15, 1997.
III
that the dark green reaction product of stoichiometry K5[Ti -
(1) Miller, J. S.; Epstein, A. J. Angew. Chem., Int. Ed. Engl. 1994, 33,
(
CN)8] was actually a seven-coordinate monocapped trigonal
3
1
85-415.
III
(
2) Kahn, O. Molecular Magnetism; VCH Publishers, Inc.: New York,
prismatic complex of formula K4[Ti (CN)7]‚KCN. In 1980,
993.
(
3) Magnetic Molecular Materials; Gatteschi, D., Kahn, O., Miller, J.
S., Palacio, F., Eds.; NATO ASI Series E, Vol. 198; Plenum: New York,
1
(7) Heintz, E. A. Nature 1963, 197, 690.
(8) Schl a¨ fer, H. L.; G o¨ tz, R. Z. Anorg. Allg. Chem. 1961, 309, 104-
109.
991.
(
4) Ferlay, S.; Mallah, T.; Ouah e´ s, R.; Veillet, P.; Verdaguer, M. Nature
(9) Golding, R. M.; Carrington, A. Mol. Phys. 1962, 5, 377-385.
(10) Bose, S. K.; Basu, S. J. Chim. Phys. (Paris) 1969, 66, 1985-1986.
(11) Nicholls, D.; Ryan, T. A.; Seddon, K. R. J. Chem. Soc., Chem.
Commun. 1974, 635-636.
1
995, 378, 701-703.
(
(
5) Entley, W. R.; Girolami, G. S. Science 1995, 268, 397-400.
6) Grossman, H. Chem.-Ztg. 1906, 907.
S0002-7863(96)02773-4 CCC: $14.00 © 1997 American Chemical Society

6252 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Entley et al.
III
Nicholls and co-workers also prepared the gray-green insoluble
Table 1. Crystal Data for [NEt
at -75 °C
4
]
3
[Ti (CN)
6
]‚4MeCN (1‚4MeCN)
III
III
rubidium and cesium compounds Rb5[Ti (CN)8] and Cs4[Ti -
CN)7] by treating TiBr3 with excess RbCN or CsCN in liquid
(
space group: I2/a
a ) 18.171 (6) Å
b ) 12.200 (4) Å
c ) 20.989 (5) Å
â ) 91.17 (2)°
R ) γ ) 90°
V ) 4652(2) Å3
Z ) 4
12
ammonia, respectively. The authors reiterated their conclusion
mol wt ) 758.99
that each of these complexes contains the C2V monocapped
-
3
F
µ
calcd ) 1.084 g cm
III
4-
trigonal prismatic [Ti (CN)7 ] anion.
-1
calcd ) 2.23 cm
Nicholls and co-workers also claimed that heating K5[TiIII-
size ) 0.35 × 0.30 × 0.20 mm
(
CN)8] to 280 °C in vacuo caused the gray-green solid to become
diffractometer: Enraf-Nonius CAD4
radiation: Mo K Rj , λh ) 0.71073 Å
monochromator: graphite crystal, 2θ ) 12°
scan range, type: 3.86 < 2θ < 48.94°, ω/θ
scan speed, width: 3-16° min , ∆ω ) 1.00[1.00 + 0.35 tan θ]°
rflctns: 3948 total, 3834 unique, 2893 with I > 2σ(I)
12
dark gray.
Extraction with liquid ammonia gave a black
III
residue that analyzed as K3[Ti (CN)6]. The powder X-ray
diffraction pattern of this material suggested that it was not
isomorphous with the known hexacyanometalates K3[M (CN)6],
-
1
III
where M is Cr, Mn, or Fe.
internal consistency: R ) 0.0239
i
Finally, Alexander and Gray have discussed the electronic
R
1
[I > 2σ(I)] ) 0.0521a
2
variables ) 318
constraints ) 27
spectrum of [Ti(CN)63-], but apparently based their discussion
wR
[all data] ) 0.2054b
13
on the data reported by Schl a¨ fer and G o¨ tz.
a
b
) [Σw(F 2
- F 2
)2/Σw(F 2)2]1/2.
R
1
) Σ (||F | - |F ||)/Σ|F
o
c
o
|. wR
2
o
c
o
Thus, the current view of homoleptic titanium(III) cyanide
anions is that they are either six-, seven-, or eight-coordinate,
and are either dark blue, dark green, or black in color. We
now report the synthesis and crystal structure of the first well-
documented titanium(III) hexacyanometalate. In contrast to
previous reports, [Et4N]3[Ti(CN)6] is pale orange when crystal-
1
line and bright yellow when powdered. Octahedral d ions are
of considerable theoretical interest but remain rather rare; the
1
present salt is the first in which a d ion is in a strong ligand
field. The present paper focuses on the physical and spectro-
scopic properties of this species; descriptions of our efforts to
prepare Prussian blue analogues from this building block will
be described elsewhere.
Results and Discussion
Synthesis of [NEt4]3[Ti(CN)6]. As pointed out by Nicholls
and co-workers, the high acidity of the titanium(III) cation in
aqueous solution prevents the formation of a titanium cyanide
Figure 1. ORTEP diagram of the [Ti (CN)
6
3-] anion. The 35%
probability density surfaces are shown.
III
complex in water, titanium(III) hydroxide and hydrogen cyanide
being formed instead.1
1,12
Thus, cyanotitanates must be prepared
III
Treatment of Ti (O3SCF3)3(MeCN)3 with excess (NEt4)CN
in acetonitrile at 45 °C for 10 min followed by cooling to -20
°C gives pale orange prisms of the hexacyanotitanate(III) salt
NEt4]3[Ti (CN)6]‚4MeCN (1‚4MeCN). This reaction must be
carried out at 45 °C or below: refluxing the reaction mixture
causes decomposition of the product. The crystals readily
desolvate to give [NEt4]3[Ti (CN)6] (1). The unambiguous
in non-aqueous media if they are to be prepared at all. Some
workers have investigated the preparation of cyanotitanates in
8
,11,12
liquid ammonia,
but this solvent is less than ideal owing
III
[
to its ability to coordinate strongly to transition metals. We
have taken a different approach and have investigated the
preparation of cyanotitanates in organic media such as aceto-
III
1
4
nitrile.
characterization of this hexacyanotitanate(III) anion is described
below.
Our first need is a source of titanium(III), preferably halide
free, which is soluble in acetonitrile; a trifluoromethanesulfonate
1
5
derivative would be ideal but is unknown for titanium(III).
III
Ti (O SCF ) (MeCN) + 6(NEt )CN + MeCN f
3
3 3
3
4
We find that treatment of titanium(III) chloride with neat
trifluoromethanesulfonic acid affords a light blue-white solid.
This reaction is rather slow and is complete only after about 72
h at 25 °C. Heating the solution above 50 °C must be avoided:
at higher temperatures, the resulting mixture turns black.
Recrystallization of the light-blue solid from acetonitrile yields
the dark-purple “triflate” salt Ti(O3SCF3)3(MeCN)3.
III
[NEt ] [Ti (CN) ]‚4MeCN + 3(NEt )(O SCF )
4
3
6
4
3
3
1‚4MeCN
Crystal Structure of [NEt4]3[Ti(CN)6]‚4MeCN. Single
crystals of 1‚4MeCN were obtained by cooling saturated
acetonitrile solutions to -20 °C. The orange prisms lose solvent
readily, and quickly become amorphous unless they are handled
while still wet with the supernatant. Crystal data are collected
in Table 1, and selected bond distances and angles are presented
in Table 2.
TiCl + 3HO SCF + 3MeCN f
3
3
3
Ti(O SCF ) (MeCN) + 3HCl
3
3 3
3
The overall stoichiometry of the crystals is [NEt4]3[Ti(CN)6]‚
(
(
12) Nicholls, D.; Ryan, T. A. Inorg. Chim. Acta 1980, 41, 233-237.
13) Alexander, J. J.; Gray, H. B. J. Am. Chem. Soc. 1968, 90, 4260-
4
MeCN. Compound 1‚4MeCN crystallizes in the monoclinic
4
271.
space group I2/a with four formula units per unit cell; the
(14) The synthesis of cyanotitanates in acetonitrile has previously been
asymmetric unit contains 1.5 cations, 0.5 anions, and 2
suggested but not implemented: Chadwick, B. M.; Sharpe, A. G. AdV. Inorg.
3-
acetonitrile molecules. The unique [Ti(CN)6 ] anions reside
Chem. Radiochem. 1966, 8, 83-176.
on crystallographic inversion centers (Figure 1) and are ordered.
(15) Dixon, N. E.; Lawrance, G. A.; Lay, P. A.; Sargeson, A. M.; Taube,
+
H. Inorg. Synth. 1986, 24, 243-250.
Of the three [NEt4 ] cations in each formula unit, two are related

III
Synthesis of [NEt4]3[Ti (CN)6]‚4MeCN
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6253
Table 2. Selected Distances (Å) and Angles (deg) for
The metal-carbon bond lengths decrease from 2.20 Å in
III
[NEt
4
]
3
[Ti (CN)
6
]‚4MeCN (1‚4MeCN)
16-31
hexacyanotitanate(III) to 1.89 Å in hexacyanocobalte(III).
Distances
By taking the metal-carbon bond distance and subtracting the
metal ionic radius, one obtains a parameter ∆ that is corrected
for the different sizes of the metals in the series; ∆ corresponds
to the effective radius of the carbon atom of the cyanide group.
Ti-C(1)
2.197(3)
2.195(2)
2.213(3)
1.143(4)
1.139(3)
1.142(3)
1.517(6)
1.513(8)
1.512(6)
1.514(8)
N(7)-C(71)
1.521(4)
1.492(4)
1.499(3)
1.530(3)
1.533(9)
1.518(7)
1.518(9)
1.520(6)
1.516(5)
1.534(5)
1.538(5)
1.496(4)
Ti-C(2)
N(7)-C(73)
N(7)-C(75)
N(7)-C(77)
C(61)-C(62)
C(63)-C(64)
C(65)-C(66)
C(67)-C(68)
C(71)-C(72)
C(73)-C(74)
C(75)-C(76)
C(77)-C(78)
Ti-C(3)
C(1)-N(1)
C(2)-N(2)
C(3)-N(3)
N(6)-C(61)
N(6)-C(63)
N(6)-C(65)
N(6)-C(67)
3
-
Table 3 shows that ∆ is largest for [Ti(CN)6 ] and smallest
3-
for [Co(CN)6 ]. The trend suggests that, upon proceeding from
Ti to Co, either there is an increase in metal-to-ligand π-back-
bonding, an increase in ligand-to-metal σ-bonding, or both. As
usual, the synergism of σ- and π-bonding effects makes it
difficult to determine which of these phenomena is most
important.
In hexacyanometalates, π-back-bonding is known to play a
role as shown by changes in metal-carbon bond lengths upon
Angles
C(1)-Ti-C(2)
C(1)-Ti-C(3)
C(2)-Ti-C(3)
Ti-C(1)-N(1)
Ti-C(2)-N(2)
Ti-C(3)-N(3)
89.02(9) C(71)-N(7)-C(73) 111.8(2)
90.22(9) C(71)-N(7)-C(75) 106.2(2)
88.02(9) C(71)-N(7)-C(77) 109.3(2)
1
6,32,33
oxidation of the metal center.
For example, the metal-
II
4-
III 34-36
carbon bond length increases upon oxidation of [Fe (CN)6]
178.9(2)
177.4(2)
178.9(2)
C(73)-N(7)-C(75) 112.1(3)
C(73)-N(7)-C(77) 106.0(2)
C(75)-N(7)-C(77) 111.5(2)
N(6)-C(61)-C(62) 113.8(7)
N(6)-C(63)-C(64) 115.2(5)
N(6)-C(65)-C(66) 114.1(7)
N(6)-C(67)-C(68) 114.9(5)
N(7)-C(71)-C(72) 114.9(3)
N(7)-C(73)-C(74) 115.3(3)
N(7)-C(75)-C(76) 114.1(3)
N(7)-C(77)-C(78) 114.5(2)
III
3-
to [Fe (CN)6] despite the smaller ionic radius of Fe .
An analogous situation has been observed in the manganese
II
4-
III
3- 22,36,37
derivatives [Mn (CN)6] and [Mn (CN)6] .
The de-
C(61)-N(6)-C(63) 111.0(4)
C(61)-N(6)-C(65) 106.4(6)
C(61)-N(6)-C(67) 111.1(4)
C(63)-N(6)-C(65) 111.6(4)
C(63)-N(6)-C(67) 105.8(6)
C(65)-N(6)-C(67) 111.0(4)
crease in ∆ from Ti to Co is consistent with an increase in the
total amount of metal-cyanide π-back-bonding as the number
of t2g electrons on the metal centers increases from 1 (Ti) to 6
(
Co). If this view is correct, then the total amount of π-back-
3-
bonding in [Ti(CN)6 ] is rather small. Nevertheless, it is likely
that the single t2g electron in this complex is delocalized into
the cyanide π* orbitals to a greater degree than are the t2g
electrons of hexacyano derivatives of later transition metals,
owing to the high energy of the d-orbitals of Ti relative to
those of later M centers.
by symmetry: these reside on general positions and are ordered.
The third [NEt4 ] cation lies almost exactly on a 2-fold axis
III
+
III
and is disordered. There are no unusually short contacts
between the cations and anions. The acetonitrile molecules are
solvates of cocrystallization and are well separated from the
titanium centers. A stereoview down the b axis of the unit cell
is shown in Figure 2.
The [Ti(CN)63-] anions adopt nearly ideal octahedral geom-
etries. The cis C-Ti-C angles all lie between 88.0(1) and 92.0-
1)°; the trans C-Ti-C angles are exactly 180° by symmetry.
The Ti-C-N linkages are almost precisely linear: the three
independent Ti-C-N angles are 178.9(2), 177.4(2), and 178.9-
2)°. One of the three independent Ti-C distances is very
slightly longer than the other two: Ti-C(1) ) 2.195(2) Å, Ti-
C(2) ) 2.197(3) Å, and Ti-C(3) ) 2.213(3) Å. The crystal-
lographic inversion center ensures that the two longest Ti-C
distances are mutually trans. Thus, the structure of the
Ti(CN)6 ] anion may be described as a nearly ideal octahedron
that shows (at best) an almost vanishingly small tetragonal
elongation of 0.017 Å.
The decrease in ∆ from Ti to Co, however, could also reflect
changes in the ligand-to-metal σ bonding; as the energies of
the metal orbitals drop from Ti to Co, donation from the cyanide
5
σ orbital should increase and the metal-carbon bond should
become relatively shorter. This latter explanation is more
consistent with the IR spectroscopic results (see below).
The trivalent chromium and cobalt hexacyanometalates both
adopt nearly ideal octahedral structures as expected since these
ions are not susceptible to Jahn-Teller distortions. The trivalent
manganese and iron hexacyanometalates (which have electroni-
(
(
4
5
cally degenerate t2g and t2g ground states, respectively) are
susceptible to Jahn-Teller distortions, but in fact the structures
are only weakly distorted and the degree of distortion is cation
dependent. The [Ti(CN)6 ] anion, which has a t2g1 ground
state, should also show a Jahn-Teller distortion but again its
structure has nearly ideal Oh symmetry. These observations
illustrate that the t2g orbitals are at best weakly bonding even
3-
3-
[
The carbon-nitrogen distances within the cyanide ligands
are nearly identical and average 1.141(3) Å. This distance is
essentially identical with those of other transition metal cyanide
(
25) Figgis, B. N.; Skelton, B. W.; White, A. J. Aust. J. Chem. 1978,
31, 1195-1198.
(26) Figgis, B. N.; Kucharski, E. S.; Raynes, J. M.; Reynolds, P. A. J.
Chem. Soc., Dalton Trans. 1990, 3597-3604.
III
3-
16-21
complexes in the series [M (CN)6 ], where M is chromium,
(
27) Capparelli, M. V.; Murgich, J. Acta Crystallogr., Sect. C 1995, 51,
manganese,1
6,22-23
iron,
16,24-27
or cobalt
16,28-31
(Table 3).
3
56-358.
(
58.
28) Barkhatov, V.; Zhdanov, H. Acta Physicochim. URSS 1942, 16, 43-
(
16) Sharpe, A. G. The Chemistry of Cyano Complexes of the Transition
Metals; Academic: New York, 1976.
17) Jagner, S.; Ljungstr o¨ m, E.; Vannerberg, N. G. Acta Chem. Scand.,
(29) Kohn, J. A.; Townes, W. T. Acta Crystallogr. 1961, 14, 617-621.
(
(30) Chadwick, B. M.; Sharpe, A. G. J. Chem. Soc. A 1966, 1390-
Sect. A 1974, 28, 623-630.
1391.
(
18) Figgis, B. N.; Reynolds, P. A. Inorg. Chem. 1985, 24, 1864-1873.
(31) Curry, N. A.; Runciman, W. A. Acta Crystallogr. 1959, 12, 674-
678.
(32) Sano, M.; Kashiwagi, H.; Yamatera, H. Inorg. Chem. 1982, 21,
(19) Figgis, B. N.; Reynolds, P. A. J. Chem. Soc., Dalton Trans. 1987,
1
747-1753.
20) Figgis, B. N.; Forsyth, J. B.; Reynolds, P. A. Inorg. Chem. 1987,
(
3837-3841.
2
6, 101-105.
(33) Jones, L. H. Inorg. Chem. 1963, 2, 777-780.
(34) Swanson, B. I.; Ryan, R. R. Inorg. Chem. 1973, 12, 283-286.
(35) Swanson, B. I.; Hamburg, S. I.; Ryan, R. R. Inorg. Chem. 1974,
13, 1685-1687.
(
21) Iwata, M. Acta Crystallogr., Sect. B 1977, 33, 59-64.
(22) Gupta, M. P.; Milledge, H. J.; McCarthy, A. E. Acta Crystallogr.,
Sect. B 1974, 30, 656-661.
23) Ziegler, B.; Haegele, R.; Babel, D. Z. Naturforsch. B 1989, 44, 896-
02.
24) Vannerberg, N. G. Acta Chem. Scand. 1972, 26, 2863-2876.
(
(36) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751-767.
(37) Tullberg, A.; Vannerberg, N. G. Acta Chem. Scand., Sect. A 1974,
28, 551-562.
9
(

6254 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Entley et al.
Figure 2. Stereoview of the unit cell contents down the b axis, showing the packing of the cations and anions and the acetonitrile solvate molecules.
III
]3- Complexes
6
Table 3. Structural and Spectroscopic Data for [M (CN)
a
∆b
(Å)
c
c
c
d
d
(Å)
CN
d
(Å)
MC
r
M
ν
CtN
ν
M-C
δ
M-CN
-1
10Dq
(cm-1)
(cm-1)
(cm
)
(cm )
-1
compd
(Å)
III
[
NEt
4
]
3
[Ti (CN)
6
] (1)
1.141(3)
2.202
2.077
0.81
0.755
0.72
0.69
0.685
1.392
1.322
1.260
1.260
1.205
2071
2128
2112
2118
2129
300
339
361
389
416
390
22800
26600
34000
34950
34500
III
e
e
K
K
K
K
3
3
3
3
[Cr (CN)
[Mn (CN)
[Fe (CN)
[Co (CN)
6
]
6
1.14
1.15
1.14
458
483
506
564
III
f
f
]
1.98
1.95
III
g
g
6
]
III
h
h
6
]
1.15
1.89
a
Reference 36. b ∆ ) dMC - r
. c Reference 45. d Reference 13. e References 17-21. References 22-23. g References 24-27. h References
f
M
2
8-31.
in these cyanometalate complexes (as opposed to the strongly
metal-ligand antibonding character of the eg orbitals).
redissolve the solvated crystals (1‚4MeCN) in acetonitrile gave
a similar gray precipitate, but the solution did develop a slight
orange color.
The single crystal structures of other titanium(III) complexes
with six equivalent ligands also have essentially identical metal-
ligand bond lengths. For example, the titanium(III) centers in
the â-alum CsTi(SO4)2‚12H2O are octahedrally coordinated to
Addition of deoxygenated water to crystalline samples of 1
instantly results in the formation of dark-blue titanium(III)
hydroxide and hydrogen cyanide. Over time, the dark-blue
material turns into a white solid (presumably titanium dioxide).
This behavior agrees with the observations of Nicholls and co-
workers, who noted that deep-blue titanium(III) hydroxide was
the only product obtained by treatment of titanium trichloride
3
8
six water molecules in sites of nearly perfect Oh symmetry.
Hexakis(urea)titanium(III) iodide and perchlorate also adopt
structures in which all of the titanium-oxygen bonds are equal
in length.3
9,40
The titanium(III) centers in the latter two
1
1,12
compounds, however, do not adopt perfect octahedral coordina-
tion geometries. Instead, the coordination polyhedra of the
titanium(III) centers are trigonally distorted: one face of the
octahedron is rotated by ca. 5° relative to the other about a
pseudo-C3 axis. Although these complexes adopt nearly ideal
octahedral geometries, their electronic spectra clearly show that
with potassium cyanide in water.
Spectroscopic Characterization of [NEt4]3[Ti(CN)6]. The
infrared spectrum of 1 taken as a Nujol mull displays a very
-1
strong CtN stretch at 2071 cm ; weak features are also noted
-1
at 2200, 2103, and 2093 cm (Figure 3). The feature at 2200
-1
-1
cm is assigned to acetonitrile, while the strong 2071-cm
41-43
they are Jahn-Teller distorted (see below).
band must be assigned to the t1u stretching mode of the
III
Reactions of [NEt4]3[Ti(CN)6]. Once [NEt4]3[Ti (CN)6]
crystallizes out of solution it is difficult to redissolve. Attempts
to dissolve desolvated crystals of 1 in acetonitrile or acetonitrile
solutions of tetraethylammonium cyanide resulted in the forma-
tion of a gray precipitate; at no point was the clear orange color
characteristic of the mother liquor observed. Attempts to
3-
[
Ti(CN)6 ] anion. (For comparison, the black solid of empiri-
III
cal formula K3[Ti (CN)6] prepared by Nicholls and Ryan
-1 12
displayed a single C≡N stretch at 2088 cm . ) The appear-
ance of additional weak C≡N stretches in the solid-state IR
spectrum of 1 is a common feature of crystalline hexacyano-
metalate complexes, and reflects the low site symmetry of the
1
6,44-46
anion in the solid state lattice.
(
(
(
38) Sygusch, J. Acta Crystallogr., Sect. B 1974, 30, 662-665.
39) Davis, P. H.; Wood, J. S. Inorg. Chem. 1970, 9, 1111-1116.
40) Figgis, B. N.; Wadley, L. G. B. Aust. J. Chem. 1972, 25, 2233-
The CtN stretching frequency for [NEt4]3[Ti(CN)6] (1) is
III
lower than those of other K3[M (CN)6] salts, where M is Cr,
2
234.
4
5
Mn, Fe, or Co (Table 3). This result supports the proposal
(
41) Clark, R. J. H. In ComprehensiVe Inorganic Chemistry; Bailar, J.
C., Jr., Emeleus, H. J., Nyholm, R. S., Trotman-Dickenson, A. F., Eds.;
that ligand-to-metal donation from the 5σ orbital of the cyanide
Pergamon: Oxford, 1973; pp 355-417.
III
III
(
42) The elpasolite-type compounds Rb2Na[Ti Cl6] and Cs2K[Ti Cl6]
(44) The appearance of multiple bands in the CtN stretching region is
even more common for the divalent hexacyanometalates.15
(45) Griffith, W. P.; Turner, G. T. J. Chem. Soc. A 1970, 858-862.
(46) Trageser, G.; Eysel, H. H. Inorg. Nucl. Chem. Lett. 1978, 14, 65-
70.
behave similarly: they show electronic but not structural consequences of
4
3
a Jahn-Teller effect.
43) Ameis, R.; Kremer, S.; Reinen, D. Inorg. Chem. 1985, 24, 2751-
754.
(
2

III
Synthesis of [NEt4]3[Ti (CN)6]‚4MeCN
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6255
III
Figure 4. The reflectance electronic spectrum of [NEt
1) as a Nujol mull.
4
]
3
[Ti (CN)
6
]
(
The reflectance electronic spectrum of 1‚4MeCN as a Nujol
mull is shown in Figure 4. The intense feature at 260 nm (38500
cm ) is assigned to the M f L charge transfer band, as judged
III
4 3 6
Figure 3. The infrared spectrum of [NEt ] [Ti (CN) ]‚xMeCN
-
1
(
1‚xMeCN) as a Nujol mull. The band at 2200 cm is due to residual
-1
acetonitrile.
from the studies of other hexacyanometalate anions by Alex-
ligands is weakest for titanium and strongest for cobalt: donation
from this orbital (which is weakly carbon-nitrogen antibonding)
13
ander and Gray. The reflectance spectrum also shows a broad
d-d absorption centered at 439 nm, which is assigned to the
4
7
is known to increase the C-N stretching frequencies. The
trivalent titanium center has the highest-energy orbitals in this
series owing to its smaller effective nuclear charge compared
to other trivalent hexacyanometalates, so that ligand-to-metal
donation is disfavored. While a low νCN frequency might also
be rationalized in terms of strong metal-to-ligand π-back-
bonding, such an explanation would not explain the long Ti-C
bond (see above) or the low frequency of the Ti-C stretch (see
below).
2
2
2
2
T2g f Eg transition. The T2g f Eg band should be split if
3-
the [Ti(CN)6 ] anion is Jahn-Teller distorted, but we do not
see such a splitting in the solid state spectrum.
2
2
The energy of the T2g f Eg transition is a direct measure
-1
of the d-orbital splitting: 10Dq is calculated to be 22800 cm
As expected, this value is larger than those seen for other Ti L6
complexes in which the ligands are lower in the spectrochemical
.
III
III
50
III
51
series: the 10 Dq values for [Ti (H2O)6]Cl3, (NH4)3[Ti F6],
III
52
-1
and (C5H6N)3[Ti Cl6] are 18 600, 17 100, and 11 800 cm ,
respectively. In the trivalent hexacyanometalates, Alexander
and Gray have demonstrated that 10 Dq increases across the
-
1
In the region between 250 and 450 cm , the IR spectra of
hexacyanometalates typically show two bands: a lower-
frequency metal-carbon stretch and a higher-frequency metal-
-1
first row from chromium (10Dq ) 26600 cm ) to cobalt (10Dq
cyanide deformation mode.3
3,45
In the trivalent hexacyanomet-
-1 13
-1
)
34500 cm ). The 22 800-cm value we observe for 1 is
alates of chromium, manganese, iron, and cobalt the metal-
completely consistent with this trend (Table 3).
carbon stretching and metal-cyanide bending frequencies
A solution UV-vis spectrum was obtained from a sample
generated by adding tetraethylammonium cyanide to Ti(O3-
SCF3)3(MeCN)3 in acetonitrile. This spectrum displays two
correlate well with the metal-carbon bond lengths:4
4,48,49
as the
metal-carbon bond length decreases, the frequencies increase
Table 3). The IR spectrum of 1 features a strong band at 390
(
-1
-1
features at 384 (26050 cm ) and 423 nm (23650 cm ). This
spectrum differs from that seen for solid samples of [NEt4]3-
-
1
-1
cm and a medium intensity feature at 300 cm . As judged
from the trends discussed above, we assign these bands to the
titanium-cyanide bending mode and the titanium-carbon
stretching vibration, respectively. The low frequencies of these
modes are consistent with the rather long metal-carbon bond
length observed for 1. In contrast, the black residue of empirical
[
Ti(CN)6], and we cannot rule out the possibility that in solution
2-
a species such as [Ti(CN)5(MeCN) ] is present. Alternatively,
if the spectrum is actually that of the [Ti(CN)6 ] ion, then the
3-
splitting of the d-d band into two components separated by
-1
2
400 cm reflects the presence of vibronic coupling between
III
formula K3[Ti (CN)6] prepared by Nicholls and Ryan showed
bands at 434 and 311 cm , which these investigators assigned
to the cyanide bending and Ti-C stretching modes, respec-
tively. These results are not consistent with our observations.
2
the excited electronic Eg state and the ꢀg normal vibration;
-
1
equivalently, the single electron in the eg* orbital will induce a
12
(50) Hartmann, H.; Schl a¨ fer, H. L.; Hansen, K. H. Z. Anorg. Allg. Chem.
956, 284, 151-161.
(51) Nassiff, P. J.; Couch, T. W.; Hatfield, W. E.; Villa, J. F. Inorg.
1
(
47) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 4th Ed.; Wiley: New York, 1986; pp 272-280.
Chem. 1971, 10, 368-373.
(52) Hartmann, H.; Hansen, K. H.; Schl a¨ fer, H. L. Z. Anorg. Allg. Chem.
1957, 289, 40-69.
(
48) Jones, L. H. J. Chem. Phys. 1964, 41, 856-863.
49) Jones, L. H. J. Chem. Phys. 1962, 36, 1209-1215.
(

6256 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Entley et al.
Figure 6. Temperature dependence of the reciprocal susceptibility of
Figure 5. Temperature dependence of µeff per formula unit of [NEt
4 3
] -
III
4 3 6
[NEt ] [Ti (CN) ] (1). The Curie-Weiss fit is shown by the straight
III
[Ti (CN)
6
] (1) in an applied magnetic field of 500 G.
line through the data points.
Jahn-Teller distortion, and the splitting represents the conse-
temperature; this may be a consequence of second-order Zeeman
53,54
quent lifting of the degeneracy.
For comparison, other well-
2,64,65
perturbations.
It is difficult to distinguish this effect,
III
characterized Ti L6 complexes exhibit electonic spectra that
however, from the presence of weak antiferromagnetic inter-
molecular interactions, which have been observed in other
-
1 41
contain two bands split by 1500 to 3100 cm .
Electronic spectra have been reported for some of the other
compounds claimed over the years to be cyanotitanates: Schl a¨ fer
43
octahedral titanium(III) complexes in the solid state. Analyses
1
of magnetic data for d complexes are commonly carried out
III
and G o¨ tz reported that “K5[Ti (CN)8]” in liquid ammonia
according to Figgis’s method, which involves three fitting
-
1 8
displayed two bands at 22 300 and 18 900 cm , and Nicholls
parameters: k (the orbital angular momentum reduction factor),
III
and Ryan reported that the black “K3[Ti (CN)6]” residue they
2
δ (the splitting within the T2g level), and λ (the reduced spin-
-
1
obtained displayed a band at 22 100 cm and a shoulder at
66
orbit coupling pertinent to the lowest term for the free ion).
-
1 12
1
8 300 cm .
The general appearance of these spectra
III
The magnetic data for [NEt4]3[Ti (CN)6] correspond ap-
suggests that these workers may have obtained solids that
contained the [Ti(CN)63 ] anion, but the dark colors of their
bulk samples certainly do not agree with the yellow to orange
color we observe for 1 and its acetonitrile solvate. Most likely,
previous samples thought to contain the [Ti(CN)63 ] anion were
probably contaminated with other titanium(III) species or with
titanium(IV) centers, in which case charge transfer processes
could explain the dark colors noted.
proximately to the following values for these parameters: k ≈
-
-1
-1
0
.75, δ ≈ 450 cm , and λ ≈ 150 cm . The positive value of
2 2
δ implies that the orbital singlet B2g derived from T2g is the
3-
ground state of the [Ti(CN)6 ] ion, as it is for essentially all
-
titanium(III) species. It is important to point out, however, that
the magnetic data do not necessarily provide unique values for
the fitting parameters: Carlin has shown that several quite
different sets of values can give equally good fits to the data
Magnetic Susceptibility. The temperature dependence of
62
for titanium(III) complexes.
III
the magnetic moment per formula unit of [NEt4]3[Ti (CN)6]
III
The magnetic moment of [NEt4]3[Ti (CN)6] (1) above 100
(1) in an applied magnetic field of 500 G is shown in Figure 5;
K is slightly field dependent: in an applied magnetic field of
the curve is qualitatively identical with that found for many
other octahedral titanium(III) complexes.51,55-62 At room
temperature, the magnetic moment of 1.64(4) µB falls in the
typical range of 1.50-1.70 µB observed for other magnetically-
dilute pseudooctahedral titanium(III) complexes.6 As the
temperature is lowered, µeff slowly decreases and ultimately
reaches a value of 1.30(1) µB at 5 K. Plots of ø- vs T are
linear between 290 and 170 K; the Weiss constant θ, determined
from the equation ø ) C/(T - θ), is -45 K (Figure 6).
The magnetic data are completely in accord with the presence
of a single unpaired electron in [NEt4]3[Ti(CN)6]. The magnetic
moments of virtually all titanium(III) salts decrease with
1
000 G, the room temperature magnetic moment is 1.55(5) µB.
Thus for this compound the susceptibility decreases with
increasing field. While the origin of this field dependence is
unclear, similar magnetic behavior has been observed for some
other titanium(III) complexes: the susceptibilites of several
adducts of TiCl3 with nitrogen donors decrease with increasing
field, with the field dependence becoming less pronounced at
3
1
56
lower temperatures.
Electron Paramagnetic Resonance Spectra. The EPR
spectrum of the orange-colored acetonitrile solution prepared
by adding (Et4N)CN to Ti(MeCN)3(O3SCF3)3 consists of a
single isotropic signal at g ) 1.984 (line width ) 9 G) at 4 K.
We cannot rule out the possibility that this signal is due to a
(
53) Jorgensen, Acta Chem. Scand. 1957, 11, 73-85
(54) Liehr, A. D.; Ballhausen, C. J. Ann. Phys. (New York) 1958, 3, 304-
2-
species of lower symmetry such as [Ti(CN)5(MeCN) ] (see
above). The EPR spectrum of a powdered sample of 1 at 25
3
19
(
55) Clark, R. J. H.; Lewis, J.; Machin, D. J.; Nyholm, R. S. J. Chem.
°
C shows an isotropic resonance at g ) 1.971 (peak-to-peak
Soc. 1963, 379-387.
56) McDonald, G. D.; Thompson, M.; Larsen, E. M. Inorg. Chem. 1968,
, 648-655.
(
line width ) 30 G). This g value is consistent with those of
other titanium(III) species, although usually relaxation effects
make it difficult to observe EPR spectra for such species at room
7
(57) Baker, J.; Figgis, B. N. Aust. J. Chem. 1980, 33, 2377-2385.
(58) Davis, P. H.; Wood, J. S. Chem. Phys. Lett. 1969, 4, 466-468.
(59) Baker, J.; Figgis, B. N. Aust. J. Chem. 1975, 28, 439-442.
(60) Schl a¨ fer, H. L.; Lenz, W.; Staab, J. Z. Phys. Chem. (Munich) 1968,
4
3,61,65,67,68
temperature.
When the powdered sample is cooled,
6
1
4
2, 290-311.
61) Giggenbach, W.; Brubaker, C. H., Jr. Inorg. Chem. 1969, 8, 1131-
137.
(64) Mabbs, F. E.; Machin, D. J. Magnetism and Transition Metal
Complexes; Chapman and Hall: London, 1973.
(65) Carlin, R. Magnetochemistry; Springer-Verlag: New York, 1986.
(66) Figgis, B. N. Trans. Faraday Soc. 1961, 57, 198-203.
(67) Gladney, H. M.; Swalen, J. D. J. Chem. Phys. 1964, 42, 1999-
2010.
(
(
62) Carlin, R. L.; Terezakis, E. G. J. Chem. Phys. 1967, 47, 4901-
906.
(63) Brant, P.; Tornqvist, E. G. M. Inorg. Chem. 1986, 25, 3776-3779.

III
Synthesis of [NEt4]3[Ti (CN)6]‚4MeCN
J. Am. Chem. Soc., Vol. 119, No. 27, 1997 6257
the signal at g ) 1.971 broadens slightly (peak-to-peak line
width ) 55 G) but otherwise remains unchanged. In addition,
however, two smaller resonances grow in: one near g ) 1.897
becomes apparent at temperatures below 250 K, and a second
near g ) 1.779 becomes apparent at temperatures below 100
K. At 3.2 K, these two resonances appear at g ) 1.901 (line
width ) 70 G) and 1.778 (line width ) 140 G). These
resonances are tentatively ascribed to an axially-distorted
standard procedure of washing the solid with diethyl ether dissolves
the crude product.14 The infrared spectrum of the recrystallized triflate
salt contains a noticeable OH stretch; this is most likely due to the
presence of traces of water or triflic acid in the recrystallized product.
III
4 3 6
Tetraethylammonium Hexacyanotitanate(III) [NEt ] [Ti (CN) ]
III
(
1). To a purple solution of Ti (O
in acetonitrile (22 mL) was added a clear solution of NEt
.1 mmol) in acetonitrile (40 mL). A dark brown-black precipitate
3
SCF
3
)
3
(MeCN)
3
(0.70 g, 1.1 mmol)
4
CN (1.42 g,
9
formed immediately. The reaction mixture was warmed to 45 °C for
ca. 10 min (higher temperatures result in decomposition of the product);
at this point the solution was cloudy and orange. The solution was
filtered, concentrated to ca. 50 mL, layered with toluene (2 mL), and
stored at -20 °C. After 24 h, a large crop of clear orange crystals
was isolated. If crystallization does not occur within 24 h, additional
toluene should be layered on top of the clear orange solution and cooling
should be continued. The clear orange crystals were dried in vacuum,
3-
[Ti(CN)6 ] ion. For comparison, Rb3[TiCl6] also gives an axial
EPR spectrum, and the effects of the Jahn-Teller distortion on
4
3
the electronic structure of this ion have been analyzed.
Concluding Remarks. Numerous investigators have sought
to prepare cyanotitanate(III) complexes, but the results until now
have been doubtful, inconclusive, or contradictory. By treating
titanium(III) triflate with tetraethylammonium cyanide in ac-
at which point they desolvated and turned into a yellow powder.
etonitrile, we have obtained the first authentic hexacyanotitanate-
III
Powdered samples of [NEt
4
]
3
[Ti (CN)
6
] (1) are extremely air sensitive
III
(
III) complex [NEt4]3[Ti (CN)6]. The magnetic and spectro-
and rigorous precautions must be taken to avoid exposing the sample
to the atmosphere. Yield: 0.50 g (74%). Mp: 250 °C dec. Anal.
Calcd for C H N Ti: C, 60.6; H, 10.17; N, 21.2; Ti, 8.05. Found:
scopic data clearly are consistent with the presence of a
hexacyanotitanate(III) anion in this salt. Single-crystal X-ray
diffraction studies unequivocally establish the presence of an
30
60
9
-
1
C, 60.2; H, 10.32; N, 20.4; Ti, 7.88. IR (cm ): 2200 (m), 2103 (w),
2093 (w), 2071 (vs), 1400 (vs), 1306 (s), 1272 (m), 1186 (vs), 1175
III
3-
almost undistorted octahedral structure for the [Ti (CN)6
]
(vs), 1080 (m), 1033 (s), 1006 (vs), 792 (vs), 669 (w), 639 (w), 606
anion in the solid state. Our attempts to use this hexacyanoti-
tanate(III) salt as a building block for the construction of
molecular magnets will be reported in due course.
(
m).
Crystallographic Studies.70 Single crystals of [NEt
]
3
[Ti (CN)
III
]‚
4
6
4
MeCN grown from acetonitrile were mounted while still wet with
the mother liquors on glass fibers with Paratone-N oil (Exxon) and
immediately cooled to -75 °C in a cold nitrogen gas stream on the
diffractometer. Standard peak search and indexing procedures gave
rough cell dimensions, and inspection of the diffraction pattern
confirmed the crystal symmetry. Least-squares refinement of 25
reflections yielded the cell dimensions given in Table 1.
Data were collected in one quadrant of reciprocal space ((h, -k,
-l) by using the measurement parameters listed in Table 1. Systematic
absences for hkl (h + k + l * 2n) and h0l (h, l * 2n) were consistent
with space groups Ia and I2/a. The average values of the normalized
structure factors suggested the centric choice I2/a, which was confirmed
by successful refinement of the proposed model. The measured
intensities were reduced to structure factor amplitudes and their esd’s
by correction for background, scan speed, and Lorentz and polarization
effects. While corrections for crystal decay were unnecessary, absorp-
tion corrections were applied, the maximum and minimum transmission
factors being 0.945 and 0.907. Systematically absent reflections were
deleted and symmetry equivalent reflections were averaged to yield
the set of unique data.
Experimental Section
All operations were carried out in vacuum or under argon. Aceto-
nitrile and toluene were was distilled under N
respectively, before use. Distilled water was sparged with argon prior
to use. HO SCF (Lancaster) was distilled in vacuum (bp 20 °C at 1
Torr) to remove colored impurities before use. (NEt )CN was prepared
2 2
from CaH and Na,
3
3
4
69
according to the method of Andreades and Zahnow. Anhydrous TiCl
Cerac) was used as received.
Microanalyses were performed by the University of Illinois Mi-
3
(
croanalytical Laboratory. The IR spectra were recorded as Nujol mulls
between KBr plates on a Perkin-Elmer 1750 FTIR instrument. Far-IR
spectra were recorded as Nujol mulls between CsI plates on a Nicolet
7
50 Magna-IR Spectrometer employing a Solid Substrate beamsplitter
and a DGTS polyethylene detector. Reflectance spectra were recorded
on a Hitachi U-3300 spectrophotometer equipped with an integrating
sphere. Magnetic measurements were carried out with a low-field (1
T) Quantum Design MPMS SQUID magnetometer. The diamagnetic
4 3 6
correction for [NEt ] [Ti(CN) ] was estimated by using Pascal’s
2
6
4,65
-6
3
-1
constants:
were recorded on a Bruker ESP 300 spectrometer.
Tris(acetonitrile)titanium(III) Trifluoromethanesulfonate Ti -
SCF (MeCN) . To TiCl (1.0 g, 6.5 mmol) at -78 °C was added
ø
dia ) -410 × 10 cm mol . The X-band EPR spectra
The structure was solved by using direct methods (SHELXS-86)
and weighted difference Fourier methods. The correct positions for
the Ti atom and the non-hydrogen atoms of one of the tetraethylam-
monium cations were deduced from an E-map. Subsequent least-
squares refinement and difference Fourier calculations revealed the
positions of the remaining non-hydrogen atoms including two inde-
III
(O
3
3
)
3
3
3
trifluoromethanesulfonic acid (2.2 mL, 25 mmol). The dark-purple
reaction mixture was slowly warmed to room temperature and then
heated to 30 °C for 48 h (heating to higher temperatures causes
decomposition). The resulting light blue-white precipitate was
contaminated with a significant amount of unreacted dark-purple TiCl
An additional aliquot of trifluoromethanesulfonic acid (2.0 mL, 23
3
-
pendent solvate molecules. The [Ti(CN)
6
] anion resides on an
+
inversion center, and one of the [NEt
4
] cations (N6A-C68A) is
3
.
disordered about a 2-fold axis. The N-C and C-C distances of the
+
disordered [NEt
4
] cation were restrained to 1.52 and 1.54 Å,
mmol) was added, and the mixture was heated to 30 °C for an additional
respectively, and chemically equivalent 1,3 C‚‚‚C distances were
restrained to be equal with an effective standard deviation of 0.01 Å.
Methyl hydrogen atom positions were optimized by rotation about the
C-C bond with idealized C-H, C‚‚‚H, and H‚‚‚H distances. The
remaining hydrogen atoms were included as fixed idealized contributors
with C-H ) 0.99 Å. Non-hydrogen atoms were refined anisotropi-
cally, while the displacement parameters for hydrogen atoms were set
to 1.2 times Ueq for the adjacent non-hydrogen atom. The relative site
occupancy factor for the disordered acetonitrile positions (C81A-N8A)
converged at 0.819(8). The quantity minimized by the least-squares
2
4 h. The solvent was removed by vacuum distillation at 25 °C, and
the resulting turquoise solid was extracted into acetonitrile (40 mL).
The purple extract was filtered, and the filtrate was concentrated to ca.
7
mL and stored at -20 °C for 12 h to give a sticky dark-purple solid.
This solid was recrystallized from acetonitrile to give analytically pure
material. Yield: 2.43 g (60%). Anal. Calcd for C
Ti: C,
7.48; H, 1.47; N, 6.80; Ti, 7.75. Found: C, 17.42; H, 1.45; N, 7.16;
9
9 3 9 9
H N F O
1
-
1
Ti, 7.68. IR (cm ): 3235 (s), 2322 (vs), 2293 (vs), 1653 (m), 1356
(
(
vs), 1239 (vs), 1200 (vs), 1156 (vs), 972 (vs), 813 (w), 769 (m), 634
vs), 605 (vs), 570 (w), 534 (m), 506 (s). The excess triflic acid must
2
2
2
2
2
2
program was Σw(F
o
- F
c
) , where w ) 1/[σ (F
o
) + (0.1429P) +
2
2
be removed from the crude pale-blue product in vacuum since the
0.8085P] with P ) (F
o
+ 2F
c
)/3. The analytical approximations to
the scattering factors were used, and all structure factors were corrected
(68) McGarvey, B. R. In Transition Metal Chemistry; Carlin, R. L., Ed.;
Marcel Decker: New York, 1966; Vol. 3, pp 89-201.
69) Andreades, S.; Zahnow, E. W. J. Am. Chem. Soc. 1969, 91, 4181-
190.
(70) For a description of the crystallographic procedures and programs
employed, see: Jensen, J. A.; Wilson, S. R.; Girolami, G. S. J. Am. Chem.
Soc. 1988, 110, 4977-4982.
(
4

6258 J. Am. Chem. Soc., Vol. 119, No. 27, 1997
Entley et al.
for both real and imaginary components of anomalous dispersion.
Successful convergence was indicated by the maximum shift/error of
Quantum Chemicals, E. I. DuPont de Nemours, and Amoco
for fellowships to W.R.E.
0
1
.009 for the last cycle. Final refinement parameters are given in Table
. The largest peak in the final difference Fourier difference map (0.441
-
3
Supporting Information Available: Tables of positional
e Å ) was located in the vicinity of N7; the map had no other
significant features. A final analysis of variance between observed and
calculated structure factors showed no dependence on amplitude or
resolution.
parameters, anisotropic displacement parameters, and full listing
III
of bond distances and angles for [NEt4]3[Ti (CN)6]‚4MeCN
(7 pages). See any current masthead page for ordering and
Internet access instructions.
Acknowledgment. We thank the Department of Energy
under Grant DEFG02-91ER45439 for support of this work and
JA962773M