THE PREPARATION AND CHARACTERIZATION OF
SOME MIXED-UGAND COMPLEXES 0F TlTANIUM (HI)

Thasis for the Degree of Ph. D.
MECHEGAN STATE UNIVERSITY
GLEN RIC-HARD HOFF
2970

 

 

         
      

LIBRAP. 1’11
Michigan State
- . University

This is to certify that the
thesis entitled

The Preparation and Characterization of Some

Mixed-Ligand Complexes of T1tan1um(III)

presented by

Glen Richard Hoff

has been accepted towards fulfillment
of the requirements for

& degree in 21516—11533!

Major professor ifiv
DateWJir F1 70

0-169

ABSTRACT
THE PREPARATION AND CHARACTERIZATION OF
SOME MIXED-LIGAND COMPLEXES OF TITANIUM(III)
By

Glen Richard Hoff

The complexes of titanium(III) chloride with the
coordinating solvents acetonitrile, tetrahydrofuran, dioxane,
and isopropanol were investigated. The solvent adducts
TiCl3-nL were prepared by the direct reaction of the solvent
with titanium trichloride or with many of the titanium(III)
complexes. Addition of tetraethylammonium chloride to the
solvent adducts suspended in the solvent yielded
(C2H5)uN[TiClu°2L]. Compounds of the type TiCl3°L and the
chloroalkoxo complex Ti2Cl3(C3H7O) were obtained by heating
the solvent adducts under vacuum. A new class of compounds of
the type TiC13L2L', containing two different non-halide ligands,
were prepared by the addition of ligand L' to the adduct
TiClB-nL suspended in the solvent L and by the dissolution
of TiCl3-L' in the solvent L. The reactions and inter—

conversions of all these compounds were studied, especially

with regard to non-halide ligand substitution.

Glen Richard Hoff

The infrared spectra of the complexes generally showed
the shifts of absorption bands associated with coordination
of the solvents. Far infrared spectra were too broad to
provide useful information about the symmetries or bond
strengths of the compounds. Two absorption bands were present
in the visible spectra which indicated that all the compounds
contained some elements of tetragonal or lower symmetry.
Jorgensen's rule of average environment was obeyed in the
complexes of the type TiCl3L2L'. The frequencies of the
absorption bands increased with increasing chloride content
of the complexes in contrast to Jorgensen's rule.

Magnetic studies over a range of temperatures indicated
normal paramagnetism for most of the complexes and several
ligand field parameters were obtained. As was expected,
the distortion and t2g electron delocalization decreased
and the magnetic moment became more dependent upon tempera-
ture as the chlorideznon-halide ligand ratio was increased.
No correlations between several ligand field parameters were
found in complexes of the type T1013L2L'. The compounds
TiCl 'CH CN and TiCl 'CHHBO were antiferromagnetic and

3 3 3

TiZCl3(C3H7O)3 was diamagnetic.

The electron spin resonance Spectra were measured at
297 and 77°K. The spectra indicated that compounds of the

l . , 1
types TiCl3L2L and (02H5)uN[TiClu 2L] as well as T1013

and TiCl3-HC3H8O had cis configurations. In some complexes

the separation of the t

°2CuH802

2% sublevels was either too small or

Glen Richard Hoff

too great to permit the observation of more than a single
absorption peak even at 77°K. Since most of the complexes
had an average g value close to 1.89 it was difficult to
confirm the existence of any definite correlations of the g
value with ligand field parameters. It appeared that dis-
tortion of the complexes and the approach of the g values
toward 2.00 increased in the same order as the ligand field
strengths of the complexes.

Several of the substances prepared changed color as
they were cooled to 77°K. These thermochromic materials had
charge-transfer bands which narrowed and.shifted to higher
frequencies to produce the color change. In addition, when
some of the thermochromic substances were cooled to 77°K
new peaks appeared in their esr spectra. It is felt that
the thermochromism and unusual esr spectra of some of the
substances is somehow related to the presence of both

titanium(III) and titanium(IV).

THE PREPARATION AND CHARACTERIZATION OF

SOME MIXED-LIGAND COMPLEXES OF TITANIUM<III>

By

Glen Richard Hoff

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1970

t/‘I 1‘ ' i" "

. '4... ‘3 2 '3
(HI-w.” {~1,.‘..,'_",‘w -,.,.~'

1"“ za‘“ " ‘f’/

To my wife, Janet, and our family

ii

ACKNOWLEDGMENT

The author wishes to extend his thanks and appreciation
to Professor Carl R. Brubaker, Jr. for his interest and
guidance during this investigation.

Financial support from the National Science Foundation

was deeply appreciated.

iii

TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS. . . . . . . . . . . . . . iii
LIST OF TABLES . . . . . . . . . . . . . . V
LIST OF FIGURES. . . . . . . . . . . . . . Vi
INTRODUCTION. . . . . . . . . . . . . . . 1

EXPERIMENTAL. O O O O O O O O O Q 0 O O O 19

Materials. . . . . . . . . . . . . 19
Analytical Methods. . . . . . . . . . . 22
Experimental Apparatus and Technique. . . . . 23
Preparation of the Compounds . . . . . 2A
Spectroscopic Measurements . . . . . . . 31
Magnetic Moment Measurements . . . . . . 32
RESULTS AND DISCUSSION . . . . . . . . . . . 35
Preparation and Reactions of the Complexes. . . 35
Infrared Spectra . . . . . . . . . . . 39
Electronic Absorption Spectra . . . . . . . “8
Magnetic Moments . . . . . . . . . . . 50
Electron Spin Resonance Spectra . . . . . . 68
Thermochromism . . . . . . . . . . . . 72

BIBLIOGRAPHY. O O O O O 0 O O O O O O O O 75

iv

Table

II.
III.

IV.

VI.

LIST OF TABLES
Page

Low Temperature Baths for Magnetic Suscepti-
bilities . . . . . . . . . . . . . . 33

Electronic Absorption Spectra of Compounds. . . A9
Magnetic Properties of Titanium(III) Complexes . 51

Ligand Field Parameters Calculated from

Magnetic Data 0 o 0 Q 0 Q 0 0 o o o o 67
The g Values from esr Spectra of the Complexes . 7O
Thermochromic Titanium Complexes . . . . . . 72

Figure

20.

LIST OF FIGURES

Page
Relative energy of the titanium(III) d orbitals
in octahedral, tetragonal, and trigonal
environments. . . . . . . . . . . . . 6
Magnetic behavior of typical (1) paramagnetic
and (2) antiferromagnetic substances as a
function of temperature . . . . . . . 1A

Infrared spectrum of TiCl3‘3CH3CN. . . A0
Infrared spectrum of TiCl3-3CQH80. . . A0
Infrared spectrum of TiCl3s2CuH802 . . . Al
Infrared spectrum of TiCl3oAC3H80. . . Al
Infrared spectrum of TiCl3(CuH80)2(CHBCN) . A2
Infrared spectrum of TiCl3(CuH802)2(CH3CN). A2
Infrared spectrum of TiCl3(CuH802)2(CuH8O). A3
Infrared spectrum of TiCl3(CAH802)2(C3H80)' A3
Infrared spectrum of TiCl3-CH3CN . . . . AA
Infrared spectrum of T1C13°CAH8O . . . . AA
Infrared spectrum of TiCl3«CuH802. . A5
Infrared spectrum of TiZCl3(C3H7O)3 . . . A5
Infrared spectrum of (C2H5)uN[TiClu°2CH3CN] A6
Infrared spectrum of (02H5)uN[TiClus2CuH80] A6
Infrared spectrum of (C2H5)4N[T1Clu-2CNH802] A7
Graph of “err vs. £3 for TiCl3~3CH3CN . . 5A
Graph of “eff Vs. §$ for TiCl3~BCuH80 . . 55
Graph of “err vs. §$ for Tici3~2CuH802 . . 56

vi

Figure Page

21. Graph of “eff vs. €$~for TiCl3‘AC3H8O . . . . 57
22. Graph of ”eff vs. §$-ror TiCl3(CuH80)2(CH3CN). . 58
kT

23. Graph of vs. -for TiCl3(CuH8O2)(CH3CN). . 59

ueff

2A. Graph of ”eff vs. -for TiCl3(CAH8O2)2(CAH8O) , 60

25. Graph of —— for TiCl3(CuH8 (C3H80) . 61

TI
kT
T"
kT
“eff VS' A! O2)2
kT
26. Graph of “eff vs. IT for TiCl3.CuH8O2 . . . . 62
kT
At
kT
T'-
kT
IT

27. Graph of —— for (C2H5)uN[T1Clu.2CH3CN] . 63

ueff VS'

28. Graph of for (C2H5luNETiClus2CuHBOJ . 6A

ueff vs.

29. Graph of for (C2H5)4N[T1Clu'2CuH8O2]. 65

ueff VS'

30. Representative electron spin resonance spectra . 69

vii

INTRODUCTION

Ever since the beginning of coordination chemistry
more than 170 years ago, chemists have been concerned with
theories about the bonding in transition metal complexes.
As the theories are deveIOped and refined they are more
able to account for many of the properties of these coordi-
nation complexes. Compounds of transition metals which
have only one d electron provide the simplest of systems
with which to test bonding theories as measurements such
as the optical and electron spin resonance spectra and
magnetic behavior can be correlated with the behavior of
the lone d electron. The symmetries of the complexes
generally are easy to predict and are only slightly
distorted from ideal point group symmetry. Comparisons
between molecular orbital parameters calculated from
theoretical considerations and those obtained experi-
mentally are especially useful for compounds in which the
symmetry is lower than octahedral. The d1 transition
metal complexes studied in this laboratory include those
of titanium(III),l vanadium(IV),2 niobium(IV),3
molybdenum(V),u and tungsten(V).5

There are several advantages in the use of
titanium(III) instead of other metals in a study of d1
transition metal complexes. The properties of titanium(III)

1

compounds are generally sufficiently different from those
of the other oxidation states of titanium so that the com-
plexes can usually be isolated and characterized. In
addition, titanium(III) does not form the oxo compounds
which make up a large part of the chemistry of vanadium(IV),
molybdenum(V), and tungsten(V). In most complexes with
monodentate ligands the titanium atom is octahedrally
coordinated. In those cases where the ligand:metal ratio
is less than 6:1 one or more of the ligands often bridge
two titanium atoms to complete the coordination sphere. ’
Thus characterizations can be based upon distortion or
reduction of octahedral point group symmetry. However,
titanium(III) compounds are very susceptible to attack by
oxygen and water and previous work in their characteriza-
tion has been limited.

Titanium(III) chloride reacts with a number of
coordinating solvents to form complexes of the type TiCl3°nL
(n = l-A). Few of the compounds are more than sparingly
soluble in the solvent from which they were made and none
will dissolve in any other solvent without reaction. The
titanium atom appears to be six-coordinate in all of the
complexes. In alkyl cyandides titanium trichloride reacts

6-8

to form TiCl ~3RCN (R = methyl, ethyl, n-propyl).

3
Titanium(III) chloride coordinates with three acetone

molecules but only two diethylketones or benZOphenones.8’9

The complex HTiClu-CquOO is made from diethylether saturated

with HC1.l Titanium trichloride adds three molecules of

tetrahydrofuran(THF) but only two of dioxane and one of
those is reported to be eliminated with prolonged heating

and stirring in dioxane.8’9 TiCl3°l.5L is formed in

1,2—dimethoxyethane or ethylene glycol dimethyl ether.9’10

While titanium(III) chloride reacts with ammonia to form

the white hexammine complex,11 it adds only two

trimethylamines and undergoes solvolysis with methylamine

12,13

or ethylamine. Under appropriate conditions titanium

.2,5dimethylformamide,lu
16

-l.5bipyridine. All three

trichloride forms TiCl3

TiC13-2bipyridine,15 and T1013

chlorides are ionic in the amine complexes TiCl -3L

3
(L = ethylenediamine, propylenediamine) and TiCl3-

2diethylenetriamine.7’l7 TiCl3-Aethylenediamine has been

7,18

reported but its existence is questioned. In alCohols

titanium forms complexes of the type TiCl -AL (L =

3

methanol, ethanol, is0propanol, sec—butanol, cyclohexanol)

19,20

in which one of the chlorides is ionic. Titanium(III)

bromide and iodide also react with coordinating solvents to

form compounds analogous to those of the chloride.6’18’20’22

There are several alternative methods to prepare
complexes of titanium(III) halides. Addition of the ligand
to titanium trichloride suspended in another solvent is

used to prepare compounds containing alcohols, alkoxide

1,16,23

ions, nitriles, and amines. In many cases an inter-

mediate complex such as TiCl -2(CH3)3N or HTiClu-Et20 is

3

used to prepare compounds not obtainable by direct

6,2A

synthesis. Trimethylamine is displaced from

TiCl3-2(CH3)3N in dioxane and the very soluble
TiCl3'3dioxane is formed.9 Tetraethylammonium halide
added to TiX3-3CH3CN (X = Cl, Br) in acetonitrile yields
the polymeric (C2H5)uN[TiXu°2CH3CN].25 The intermediate
TiBr3-2(CH3)3N, used to prepare a variety of titanium(III)

bromide complexes,is prepared by the reduction of

22

titanium(IV) bromide with trimethylamine. TiCl -nCH OH

3 3
(n = A,5) is made by the electrolytic reduction of

titanium(IV) chloride in methanol.2u Several complexes

eliminate non—halide ligands when heated in a vacuum and

compounds of the type TiCl3-2L (L = NH 11:1“,24

25,26

3, THF, DMF),

TiCl °L (L = THF, CH3CN), and (C2HS)AN[T1XAJ

3
(X = Cl, Br) are formed.25

Complexes in whichthe titanium atom is coordinated
to six identical, non-bridging ligands generally exist
only in solution. Systems of the type [TiL6]n containing
water, halides, alcohols, urea, acetonitrile, and amines

27-30

have been characterized. Solid compounds which have

been isolated include the hexammine complex,5
- 25
(CSHSN)3[T1X6] (X Cl, Br), and Ti(CH30N)6(I3)3 which
has been prepared by the oxidation of powdered titanium
metal with iodine in acetonitrile.31
Titanium(III), having only one d electron, has a 2D

ground state. In an octahedral ligand field this ground

state is split into two levels (Figure l). The upper level,

2

Eg, is separated from the ground 1evel,2T2 by an energy A0.

g!
The orbital degeneracy in each of these two levels is

lifted by components of lower symmetry. Generally the
ligand field symmetry about the titanium atom is reduced
from cubic (Oh) to tetragonal (DAh) or trigonal (D3)
although other symmetries are also possible depending upon
the nature of the ligands.

Trigonal distortion, caused by elongation or com—

pression along the C3 axis of the octahedron, leaves the

2BE state of the cubic field unchanged but splits the

2T state into a 2E and a 2A1 state, with an energy

28

separation 5 Tetragonal deformation, caused by com—

1'
pression or elongation along the CA axis, splits the 2T283.

ground state into an upper 2E8 and a lower 282g level
separated by an energy 61. The 2Eg state of the octahedral

field is split into a 2A level, and a 2B level,

lg lg
separated by an energy 62. The ordering of these two

levels, shown in Figure l, are those which are most
8,28,32

widely used. They are the reverse of those used

by McDonald gt al. which must be considered when

spectrochemical data are compared.7

The 2T + 2Eg transition in octahedral titanium(III)

2%
complexes falls between 13,000 and 20,000 cm-l. The inten-
sity of the peak (e’b3 to 12) is consistent for transitions

which are Laporte forbidden but which appear because of

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2E
2 S
E c > ------- <
10 Dq A0 E 10 Dq
2E ( )2
‘ ’- - \\\‘
a ,/’<
1 ,’
1 2g
D3 Oh
trigonally regular
distorted octahedron
octahedron
Figure 1.

Relative energy of the titanium(III)

 

 

 

 

 

 

2B
,I' f I 1g
//’
6
\ 2
\\ l
\\_ 2A
1%
10 Dq
,,( )2E
,’ f S
\ 6
\. 1
\\ A L 23
2S
Duh
tetragonally
distorted
octahedron

d orbitals in octahedral, tetragonal, and

trigonal environments.

33

vibronic coupling. However, the expected sharp
symmetrical ligand-field peak is not seen. The actual
peak is broad and asymmetric, showing a distinct doublet
structure which indicates that the ligand field must
contain at least a tetragonal component. As previously

indicated, a tetragonal deformation will split both the

ground term (2T2g) and the upper term (2Eg) so that two
2 2 2 2

transitions, 82% + A18 and 82g + Blg’ would be
expected. The tetragonal distortion is considered to be
3A,35

the result of the Jahn-Teller effect, which removes

the degeneracy of the original cubic field orbitals, and

1 for titanium(III)

is on the order of 1,000 to 3,000 cm-
complexes.

The difference in the energies of the two transitions
corresponds to the separation energy 62 of the upper
orbitals. Because data concerning the concommitant splitting
of the ground level are not obtainable from optical spectro—
scopy, the usual practice for a d1 system is to make no
correction in either the upper or ground states for lower-
symmetry components in the octahedral ligand field.36 The

energy A0 of the 2T + 2Eg transition is taken to be that

28
of the more intense upper band of the asymmetric ligand
field band. The energy A0 is also equal to 10 Dq, where
Dq is the cubic field splitting parameter.
To facilitate comparisons of ligand field strengths

when the ligands are varied, the practice of ignoring lower

ligand-field symmetries also extends to tetragonal and
trigonal complexes. Although the perfect octahedral field
no longer exists, the energy of the most intense d-d
transition is set equal to 10 Dq. In tetragonal complexes
this corresponds to the 282% + 281g transition. It can be
seen from figure one that trigonal complexes have only one
upper state. No amount of trigonal distortion, however
great, removes the degeneracy of this Eg state. Thus only
a single ligand field peak would be expected to appear in
the optical spectrum. However, the spectrum of trigonal
complexes shows the same double structure observed in
octahedral and tetragonal complexes so a tetragonal com—
ponent must also be present. For trigonal complexes the
energy of the Alg + Blg transition is taken to be equal to
10 Dq.

Visible and ultraviolet spectra have been obtained
for many of the titanium(III) complexes reported. All
show the asymmetric or double structure caused by tetra—
gonal distortion. Several ligands, along with their Dq

values, fit into the spectrochemical series in the follow-

ing order:

CH3CN (2390) > H20 (2010) > (NH2)200 (175A) > (CH3)2CO (1710)

> CH3OH (1680) > CAH802 (1590) > CAHBO (1570) > C1 (1370).

it is assumed that Jorgensen's rule of average environment

37

is obeyed. The absorption spectrum of a solution of the
trisacetonitrile complex of titanium(III) chloride is the
same as the reflectance spectrum of the solid. However, the
two spectra are clearly different for the tristetrahydrofuran
or bisdioxane complexes which would indicate that the nature
of the absorbing species in solution is different from that
of the solid.8

Charge transfer bands, generally more intense than
ligand field bands, appear in the ultraviolet region of the
spectrum. They arise from transfer of the d electron of
titanium to antibonding n molecular orbitals on the ligands
or transfer of n-electrons from the ligands to vacant d
orbitals of titanium. The exact nature of the transitions
is unknown but tentative assignments have been made in
some cases.6’8’9

The infrared spectra of titanium(III) complexes
generally Show band shifts and splittings associated with
coordination. For example, the 2253 cm..1 band of CH3CN is
CN.8 The

shifted 35 cm_1 higher in the compound TiCl °3CH

3 3
bands assigned to symmetric and asymmetric C—O-C stretching
vibrations of the cycljx:ethers tetrahydrofuran and dioxane
are shifted to lower wave numbers and split upon coordina-

8,9

tion through the oxygen. Information about the
symmetry of the complex may be obtained from the titanium—

halogen bands in the far infrared spectrum. A cis (C3V)

lO

complex should have two infrared active metal-halogen
vibrations while a trans (C2v) complex should have three.
On this basis Clark has shown that TiCl °3CuH80 is trans

3
and TiCl '3CH3CN is cis.38 However, the far infrared

3
bands are often too broad to provide any useful informa-
tion.

Data concerning the variation of the magnetic moment
with temperature can provide useful information about axial
distortion, spin-orbit coupling, and electron delocaliza-
tion in titanium(III) complexes. Extensive reviews of

this area are available33’39’uO

and only a brief summary
will be given here.

Paramagnetic susceptibilities, from which the magnetic
moments are calculated, vary inversely with temperature

according to the Curie law

or the Curie-Weiss law

_ C
xm - T+0
In these equations Xm is the molar susceptibility,
corrected for the diamagnetic contribution of titanium and
the ligands, T is the absolute temperature, and C is the
Curie constant, characteristic for each substance. 0, the

Weiss constant which takes intermolecular interactions

11

into account, is the temperature at which a plot of l/xm
against T intercepts the T axis.
The magnetic moment is calculated from the molar

susceptibility by

u = 2.8A (me)%

for compounds which obey the Curie law.uo’ul

The quantity
(T+0) must be used in place of T if the Curie-Weiss law is
obeyed. However, in many cases the magnetic susceptibility
data do not fit either law. The usual practice in these
cases is to use the Curie law to determine the magnetic
moment at a specific temperature and label this moment the
effective magnetic moment “eff at the specified tempera-
ture.“2

In the simplest case where there is no orbital
angular momentum the magnetic behavior of the complex
arises only from electron spin angular momentum. As a
result, the spin-only magnetic moment is 1.73 Bohr magnetons
for titanium(III). In molecules with a high degree of
ligand field symmetry orbital angular momentum is present
and will couple with the spin angular momentum. Spin-
orbit coupling has a considerable effect upon the variation
of the magnetic moment with temperature. Kotani has shown
that the magnetic moment of a purely octahedral (11 complex
A3

should approach zero as the temperature approaches zero.

However, there are no titanium(III) complexes which behave

12

in this manner. Both axial distortion of the ligand
field symmetry and orbital delocalization of the t2g electron
onto the ligands cause the magnetic moment to tend toward
the spin-only value.
Figgis has presented a method for the treatment of
the magnetic susceptibility data of titanium(III) systems
from which estimates can be made of the following factors
which quench the orbital angular momentum and cause the
magnetic moment to approach the spin—only value:
1. orbital delocalization (l-k) which is the
fraction of time the t2g electron spends
upon the ligand atoms
2. spin orbit coupling A' which is reduced from
the free ion value A by complexationu5
3. energy separation 61 between the orbitals
derived from the splitting of the 2T28
state by axial distortion.
Figgis has calculated theoretical, effective magnetic
moments at several temperatures for a range of values for

k (the amount of time the electron spends on the titanium

atom) and v, where

A positive value for v, and thus for 6 indicates that

l,

the orbital singlet from the splitting of the 2T28 state

l3

lies lowest and is the ground state. A plot is made of the
experimental magnetic moments against kT/X where k is

the Boltzmann constant, T is the absolute temperature, and
A is the free ion spin orbit coupling constant. The
resultant curve is compared to the theoretical curves of
Figgis and the best values of k and v are determined by
the best fit of the curves. The value of A' is determined
by substituting A' for A in the abscissa of the plot of
the experimental data. The value of A', a fraction of A,
is then adjusted until the best match of the curves is
made. Although ligand fields of low symmetry cause the
magnetic moment to tend toward the spin-only value, the
effect is not great even for compounds whiCh have axial
ligand field components ten times the spin-orbit coupling
constant. The effect of orbital delocalization is much
greater.

In the preceeding discussion it has been assumed
that the individual titanium atoms act independently of
each other. Generally this is true but in at least two
cases exchange interactions cause the compound to exhibit
antiferromagnetism. Figure 2 shows the variation of xm
with temperature for paramagnetic and antiferromagnetic
substances. Above the Neel point the behavior of both
types of compounds is similar, but below the Neel tempera-
ture xm of the antiferromagnetic compound rapidly

decreases with temperature. Even above the Neel

1A

 

>5

Nee1-———)
point

 

 

 

T———-—>

Figure 2. Magnetic behavior of typical (1) paramagnetic
and (2) antiferromagnetic substances as a
function of temperature.

15

temperature the magnetic moment of an antiferromagnetic
titanium(III) compound is markedly reduced from the spin-
only value.

Antiferromagnetism may come from intermolecular
interactions, extending throughout the Crystal, or intra-
molecular interactions, limited to within the molecule.
With intermolecular interactions exchange between para-
magnetic centers takes place through intervening atoms,
known as "super-exchange." Super-exchange is destroyed
when the interacting ions are separated by dissolution
or dilution in a solid isomorphous diamagnetic host.
Intramolecular interactions are unchanged by dissolution
or dilution. Earnshaw gt 31.46 have presented a mathemati-
cal treatment for the determination of the number of
interacting spins S and the value of the exchange integral

J for the equation AB = 2JSiS The treatment of the

k.
magnetic data to find n, the number of spins and conse-
quently the number of members of a polymeric chain, and J

is similar to the treatment of magnetic data of paramagnetic

complexes in the Figgis method. Titanium trichloride

A7

exhibits antiferromagnetism but titanium tribromide does

A8

not The only complex reported to show antiferromagnetic

behavior is TiCl3-CH3CN.25
Electron spin resonance (esr) measurements can

provide valuable information concerning the structure,

magnetic behavior, and molecular orbital parameters of

16

transition metal complexes. There are many books and

reviews which cover the fundamentals and applications of

“9

electron spin resonance. Several references which deal

specifically with titanium(III) in different crystal field
50 51

symmetries are Kuska, Carrington and Longuet-Higgens,

and Gladney and Swalen.52

It would be expected that titanium(III), as a d1
electron system presenting the simplest of crystal field
problems, would have been extensively studied. However, a
review of the literature has shown that this is not the
case. In octahedral lattice sites, such as with cesium
titanium alum, the spin—lattice relaxation time is very
short because of the presence of a low-lying excited state
and resonance is only detected at liquid helium tempera-
tures.53-56
Compounds of lower symmetry generally exhibit a

resonance at room temperature, although the resonance may

be anisotropic and only an average value of the gyromagnetic
ratio is seen. Among the titanium(III) complexes of lower
symmetry for which a resonance has been observed are those

57,58 59 1,60

containing acetylacetone, halides, or alkoxides.

Giggenbach and Brubaker presented the esr spectrum of

several mixed complexes containing chloride ion, methanol,

1

methoxide ion, or acetonitrile. The resonance spectrum of

trichlortrisacetonitriletitanium(III) has been observed

by McDonald gt al.1u and Giggenbach.61

1?

T1A8’ which has a natural abundance of 73.9%, has no

nuclear spin and consequently does not give rise to hyperfine
splitting of the epr spectrum. Tiu7, 7.3% natural abundance,
and T1A9, 5.5% natural abundance, have nuclear spins of 5/2
and 3/2 respectively. Hyperfine splittings caused by these
two nuclei have been observed by Waters and Makisg and

McGarvey.57

The hyperfine patterns are best observed at
-A5° and at other temperatures the resonance broadens and
only a broad asymmetric signal is seen.1

Thermochromism is the property of a compound to change

color when the temperature is changed.u9’62

Generally this
occurs in solids when a lowering of the temperature causes
the charge transfer bands to sharpen and tail less into the
visible region of the spectrum. In this case the color
change should be gradual and no other changes in properties
should take place. When thermochromism is associated with
a phase change in the solid the color change should be abrupt
(except for hysteresis) and other changes in the properties
of the solid may occur. There are no previous reports of
thermochromism being observed in titanium(III) complexes.
There are very few reports of titanium(III) complexes
of the type TiCl3L2L' where L and L' are different non-halide
ligands.. One of the objectives of this research was to pre-
pare and characterize compounds of this type which contain

acetonitrile, tetrahydrofuran, dioxane, and isopropanol as

the non-halide ligands. Another objective was to prepare

18

and study the series of compounds T1013.3L, [TiClu-2L]-,

and TiCl3aL. In TiCl3-L chloride ions bridge titanium atoms
so that each titanium is coordinated to five chloride ions

and one non-halide ligand. The experimental data were examined
to find correlations between changes in the properties of

the complexes and the nature of the ligands. In some cases

the properties of the T101 L' were expected to be inter-

3L2

mediate to those of the parent complexes TiCl onL and TiCl onL'.

3 3
The location of the d-d absorption bands could be predicted
from Jorgensen's rule of average environment and the Dq values
of the ligands. The symmetries of the TiCl3L2L' complexes
were lower than those of the parent complexes. The changes
in symmetry were expected to be apparent in the esr spectra
and magnetic behavior of the compounds. The effect of an
increased halide content was studied in the series TiCl3-3L,
(TiClu°2L]-, TiCl3-L. However, TiCI3-L may show additional
changes because of the presence of bridging chloride ions.
As very few esr spectra of titanium(III) chloride adducts
have been reported, the spectra were obtained for all of

the compounds in this research to observe the effect of

the symmetry and the nature of the ligands upon the g value.

EXPERIMENTAL

A. Materials

 

Titanium Trichloride.-—Anhydrous titanium trichloride

 

was obtained from K & K Laboratories, Inc. and Research
Organic/Inorganic Chemical Company. Titanium(III) and
chloride determinations gave an analysis of 99.91% titanium
(III) chloride and the compound was used without further
purification. It was stored in a Schlenk vessel to

facilitate subsequent transfer.63

Solvents.--6uAll solvents were refluxed with an
appropriate drying agent under a nitrogen atmosphere except
where indicated. Purity of the solvents was checked by
examination of their proton magnetic resonance spectrum for
evidence of impurities.

Acetonitrile was stirred with 200 mesh silica gel
several hours and then stirred overnight with calcium hydride.
After evolution of hydrogen had ceased, two grams of phos-
phorus pentoxide per liter of solvent was added and the
acetonitrile was refluxed continuously. It was distilled
at a high reflux ratio immediately before use.

Chloroform was washed with water and dried with calcium

chloride. After the chloroform had refluxed 2A hours with

19

2O

phosphorus pentoxide it was distilled and stored in the
absence of light before being used.

Dioxane was stored over sodium hydroxide. It was
refluxed with sodium and distilled after the surface of the
sodium had become shiny and the purple color of the disodium
benzophenone complex was formed.65

Trimethylamine (anhydrous liquid from Eastman Organic
Chemicals) was transferred to a Schlenk vessel and stored
at —5°. It was distilled at a reduced pressure into the
reaction vessel which was at -78°.

Isopropyl alcohol was continuously refluxed with
sodium isopropoxide and distilled immediately before use.

Tetrahydrofuran, diethyl ether, and benzene were
refluxed continuously with lithium aluminum hydride and
distilled immediately before use.

Tetraethylammonium Chloride.--Tetraethylammonium
chloride was obtained from Eastman Organic Chemicals.

Twice it was ground, boiled with dry benzene to remove the
water, and dried under vacuum. This method gave a product

of higher purity than by merely drying the tetraethylammonium
chloride in an oven at 100°.

Ggggg,--Oil-pumped prepurified nitrogen from Liquid
Carbonics was passed through a heated activated copper
catalyst (BTS catalyst - BASF R 3-11 from Badische Anilin- &
Soda-Fabrik AG) at 150° to remove oxygen and through Aquasorb

(indicating moisture absorbant, containing phosphorus pentoxide,

21

from Mallinckrodt Chemical Works) to remove water. The
nitrogen in the controlled atmosphere box was continuously
recirculated through the Aquasorb and heated BTS columns.
The condition of the atmosphere in the controlled atmosphere
box was checked by observing the burning time of an exposed
light-bulb filament.

Hydrogen chloride from Matheson Gas Products was
anhydrous and was used without further purification.

Analytical Reagents.--Silver nitrate was ground and
dried at 180°, reground and stored in a vacuum desiccator.

A 16.989 g sample was dissolved in one 1 of distilled water
to make a 0.1000 N. solution.

A 0.100 N cerium(IV) sulfate solution was prepared
from a standard solution from Fisher Scientific and the con-
centration checked by titration with ferrous ammonium sulfate.
Ferroin (1,10-phenanthroline ferrous sulfate complex) was
used as the indicator.

Dilute sulfuric acid was prepared by boiling demineral-
ized distilled water, adding enough sulfuric acid to make
a A M solution, boiling the mixture for another five minutes,
and then cooling the solution in a special Schlenk vessel.
Nitrogen was rapidly, and continuously bubbled through the

solution during its preparation and cooling.

B. Analytical Methods

 

Carbon, Hydrogen, and Nitrogen Analysis.--These analyses

 

were performed by the microanalytical laboratory of the
Institute of Water Research, Michigan State University, East
Lansing, Michigan. Although special handling precautions

were taken to avoid decomposition of the samples, they reported
the samples very difficult to analyze. Identical samples

did not always give similar results.

Chloride Analysis.--Generally the solid sample was

 

quickly dissolved in water but in cases where hydrogen
chloride may be lost the sample was first cooled to 77°K

and dilute sulfuric acid added. After the sample was slowly
warmed to room temperature it was treated in the usual
manner.

It was necessary to oxidize the titanium(III) cataly—
tically with air prior to the chloride determination. This
was done by the addition of one milliliter of a saturated
solution of copper sulfate in 9M sulfuric acid to the dissolved
sample and stirring the solution until the color changed
from violet to yellow.

Chloride was determined by a differential potentiometric
titration with a standard silver nitrate solution.66 Silver-
silver chloride electrodes and a Corning Model 10 pH meter
were used.

Titanium(III) Analygis.--Titanium(III) was determined
by a cerimetric technique. As the sample was sensitive to

oxidation, it was cooled to 77° K in a Schlenk vessel and

23

deoxygenated dilute sulfuric acid was added. After the
solution had warmed to room temperature, it was titrated
with cerium(IV) sulfate. Ferroin was used as the indicator.
A nitrogen atmosphere was maintained over the solution during
the dissolution and titration of the sample.

The total titanium content was determined spectrophoto-
metrically as the titanium(IV) hydrogen peroxide complex

in sulfuric acid.67

Beer's law is obeyed at A10 nm for
solutions containing up to 75 mg titanium(IV) per liter.

The analysis was performed on a portion of the solution
prepared for chloride analysis, in which the titanium(III)

had been catalytically oxidized. Some samples could not

be analyzed in this manner because the solution was turbid

and the absorbance reading was thus abnormally high. Although
the total titanium analysis also measured any titanium(IV)

impurity present in the original sample, the two methods

gave very close agreement in most cases.
C. Experimental Apparatus and Technique

All of the titanium(III) samples prepared were sensi-
tive to water and oxygen. All reactions and manipulations
were carried out under a nitrogen atmosphere. Schlenk tube
techniques, similar to those recently presented by Herzog
gt gl.,63 were used throughout. One difference is that
inner ground glass joints were generally used on Schlenk
vessels to facilitate transfer and so that compounds would

not come into contact with the silicone lubricant on the joint.

2A

An important factor in the successful preparation of
the titanium(III) complexes was the use of a vacuum mani—
fold in connection with the nitrogen line. Whenever a
glassware connection was made or broken, nitrogen was forced
into the vessels through the side—arm stopcocks. This
positive flow of nitrogen, about four liters per minute,
prevented air from entering the vessel. When the vessel
was closed, it was immediately evacuated and the atmosphere
replaced with fresh nitrogen. Clean glassware, dried at
180° for several hours, was evacuated and filled with nitrogen
several times while it cooled.

As the nitrogen of the controlled atmosphere box was
not nearly as pure as that in the Schlenk vessels, which
could be changed quickly, the box was little used except
for the storage of compounds in Schlenk vessels and for the
manipulations not amenable to Schlenk tube techniques.

These included preparation of mulls for optical spectroscopy

and the filling of esr and magnetic susceptibility tubes.

D. Preparation of the Compounds

 

Trichlorotrisacetonitriletitanium(III).--Eight grams

 

of titanium trichloride was added to 300 ml of acetonitrile.
The mixture was heated and the solvent allowed to reflux
several hours. The color of the mixture changed from violet
to blue during this time. Half of the solvent was then

removed by vacuum evaporation which also served to cool the

25

remaining mixture. The cold solution was filtered and the
dark blue crystals were washed with portions of acetonitrile
and dried under vacuum.

Ti, 17.27; Cl, 38.3A;

Analysis. Calc. for TiCl 06H

3 9N3‘
C, 25.98; H, 3.27; N, 15.15. Found: Ti, 17.29; Cl, 38.68;

C, 26.3A; H, 3.51; N, 1A.78.

Trichlorotristetrahydrofurantitanium(III).—-This com—

 

pound was prepared in a manner similar to that of the above
acetonitrile complex. Finely divided turquoise blue crystals
were obtained.

Analysis. Calc. for T1013012H2u03: Ti, 12.93;
Cl, 28.70; C, 38.86; H, 6.53. Found: Ti, 12.98; Cl, 28.70;

c, 38.8A; H, 6.5A.

Trichlorobisdioxanetitanium(III).--This compound was

 

prepared in a manner similar to that of the above acetonitrile
complex. The reaction time was about 16 hours. A pale
green powder was collected.

Analysis. Calc. for TiCl3C8H16OA: Ti, 1A.50; Cl, 32.19;
C, 29.06; H, A.89; Found: Ti, 1A.5A; Cl, 32.22; C, 28.78;
H, 5.03.

Longer reaction times (60, 100, and 2A0 hours) gave

a product similar in appearance, analysis, and behavior.

Trichlorotetrakis(2-propanol)t1taniumg I l.-uThis

compound was also prepared in a manner similar to that of

26

the above acetonitrile complex. The reaction was complete
in two hours and yeilded a blue powder.

Analysis. Calc. for TiCl Ti, 12.15;

3012H320A‘
C1, 2Mi96; C, 36.50; H, 8.18. Found: Ti, 12.13; C1, 26.88;

C, 36.00; H, 7.95.

Trichloroacetonitriletitanium(III).—-A portion of

 

TiCl3-3CH3CN was heated at 100° under a vacuum for several

hours until the evolution of a gas ceased. The residue

was a brown powder.

Analysis. Calc. for TiCl3C2H3N: Ti, 2A.A2; Cl, 5A.A5;
C, 12.28; H, 1.55; N, 7.17. Found: Ti, 2A.5A; C1, 5A.l8;
C, 11.98; H, 1.65; N, 6.86.

The reaction will proceed at 70° but at temperatures

above 100° or with prolonged heating a dark violet-brown

product was formed. This may contain TiCl3 as titanium and

chloride analyses were higher than those above.
In some cases a yellow sublimate was collected which

appeared to be TiClu-2CH CN. It was thermochromic, turning

3
white as it was cooled to 77°K.

Analysis. Calc. for TiCluCuH6N2: Ti, 17.62; C1, 52.17.

Found, Ti, 17.72; C1, 53.12.

Trichlorotetrahydrofurantitanium(III).--A sample of

TiCl3i3CuH80 was heated under vacuum at 75°. A gas was

evolved and the sample turned light green and then grey-

green. The light green substance, which may be TiCl3-2CuH80,

27

could not be isolated. The gray-green solid turns violet
as it is cooled to 77°K.

Analysis. Calc. for TiCl3cuH8O: Ti, 21.16, Cl, A6.98;
C, 21.20; H, 3.56. Found: Ti, 21.20; 01, A6.80; O, 20.63;
H, 3.62.

A thermochromic yellow sublimate, apparently
TiClA'ZCAH8O’ was also obtained when the samples were briefly
exposed to the atmosphere.

Analysis. Calc. for TiClAC8H16O: Ti, 1A.35; Cl, A2.A7.

Found: Ti, 1A.58; Cl, A2.6A.

Trichlorodioxanetitanium(III).—-TiCl 2CAH802 heated

3

to 1A0° under vacuum gave a finely divided cream powder.

 

Analysis. Calc. for TiCl3CAH8O2: Ti, 19.76; Cl, A3.89;
C, 19.81; H, 3.33. Found: Ti, 19.98; C1, A3.7A; C, 19.36;

H, 3.31.

Trichlorotris(2-prgpoxy)dititanium(III).--A sample of

 

TiCl3-AC3H80 was heated under vacuum at 75° for ten hours.

Hydrogen chloride and 2-propanol were evolved and the
remaining solid turned red-brown.

Analysis. Calc. for Ti2Cl3C9H2103: Ti, 25.25;
C1, 28.03. Found: Ti, 26.70; 01, 29.33.

Trichloroacetonitrilebistetrahydrofurantitanium(III).--

 

Four grams of TiCl °3CAH8O was suspended in tetrahydrofuran

3
and twenty milliliters of acetonitrile was added. After the

solution was stirred and heated nine hours the volume was

28

reduced by vacuum evaporation to forty milliliters. An
equal quantity of hexane was added and the solution was
stirred vigorously. The blue solid which precipitated was
collected on a filter and dried under vacuum.

Analysis. Calc. for TiC13ClOH1902N: Ti, 1A.11;
C1, 31.33; C, 35.35; H, 5.65; N, A.l3. Found: Ti, lA.25;
C1, 31.30; C, 3A.A8; H, 6.02; N, A.01.

A different preparative method was tried in which
three milliliters of tetrahydrofuran was added to 5.85 g
TiCl3-CH3CN suspended in benzene. After the mixture was
stirred and heated six hours it was filtered. The gray-

green solid, collected by filtration and dried under vacuum,

was thermochromic and appeared to be TiCl3-CuH80.

Trichloroacetonitrilebisdioxanetitanium(III).--Several

 

grams of TiCl ~3CH CN were suspended in dioxane and the

3 3
mixture was heated and stirred five hours. The very thick
blue-green mixture was filtered and the solid was washed
with dioxane and dried under vacuum. The blue-green powder
is difficult to handle as it becomes charged with static
electricity when it is ground.

Analysis. Calc. for TiCl OAN: Ti, 12.90;

3ClOH19
01, 28.63; C, 32.31; H, 5.16; N, 3.77. Found: Ti, 12.98;
Cl, 28.66; C, 33.Ao; H, 5.15; N, A.78.

The preparation worked equally well when TiCl3°CH3CN

was used as the starting material or when acetonitrile was

added to T1C13~2CuH802 in dioxane.

29

Trichlorotetrahydrofuranbisdioxanetitanium(III).--A

 

quantity of TiCl3-CuH80 was suspended in dioxane and the
mixture was heated and stirred. After two hours the solid
had dissolved and the solution was brown. The volume of
the solution was reduced to forty milliliters under a
vacuum and an equal quantity of diethyl ether was added.
Immediate precipitation resulted and the solid was washed
with dioxane and dried under a vacuum.

Analysis. Calc. for TiCl Ti, 11.90;

3012H2A05‘
Cl, 26.A2; C, 35.79; H, 6.01. Found: Ti, 12.00; C1, 26.38;
C: 35-13; H, 5.31-

The use of TiCl3-3CuH80 as the starting material often

did not give a homogeneous precipitate.

Trichloro(2-propanol)bisdioxanetitanium(III).--Ten

 

milliliters of isopropyl alcohol was added to six grams of
TiCl3o2CuH802 in dioxane and the mixture was stirred and

heated overnight. The volume of the solution was reduced

under vacuum to A0 m1 and an equal quantity of diethyl ether
was added while the mixture was vigorously stirred. The
solution was filtered and the pale blue precipitate was

dried under vacuum.

Analysis. Calc. for TiCl3C11H2A05: Ti, 12.26, Cl, 27.23;
C, 33.8A; H, 6.19. Found: Ti, 12.16; Cl, 27.28; C, 33.15;
H, 6.08.

3O

Tetraethylammonium Tetrachlorobisacetonitriletitanate(III).--

 

A known quantity of TiCl ~3CH3CN was dissolved in a solution

3
of 10% acetonitrile in chloroform. Upon addition of 2.3
equivalents of tetraethylammonium chloride to the blue solution,
the mixture immediately turned yellow. After one-half hour

the mixture was brown and after it had been stirred overnight
the mixture was green. The solution was filtered and a

pale blue-green product was collected and dried under vacuum.

Analysis. Calc. for TiCluC Ti, 11.91;

12H26N3‘
C1, 35.27; C, 35.85; H, 6.52; N, 10.A5. Found: Ti, 12.01;

01, 35.15; C, 3A.91; H, 6.1A; N, 10.38.

Tetraethylammonium Tetrachlorobistetrahydrofuran-

titanate(III).—-A known quantity of TiCl

 

3.3CAH8O was
dissolved in tetrahydrofuran and 2.5 equivalents of
tetraethylammonium chloride was added. The mixture was
stirred overnight. The solution was filtered and a pale
green powder was collected and dried under vacuum.

Analysis. Calc. for TiClACl6H 602N: Ti, 10.32;

3
Cl, 30.55; C, A1.A0; H, 7.82; N, 3.02. Found: Ti, 10.36;
Cl, 30.52; C, A1.31; H, 8.12; N, 3.25.

When the green powder was washed with chloroform it
immediately turned orange-brown. After it had been dried
under vacuum, the powder was salmon and was thermochromic.
This material was also obtained from a solution of 10%

tetrahydrofuran in chloroform. The material was non-

stoichiometric with a titaniumzchloride ion ratio of 1:A.5.

31
Analysis. Found: Ti, 11.58, Cl, 37.58.

Tetraethylammonium Tetrachlorobisdioxanetitanate(III).--

A weighed quantity of TiCl '2CuH8O2 was suspended in dioxane

3
and 2.5 equivalents of tetraethylammonium chloride was added.
After the mixture had been stirred overnight it was filtered
and the pale green powder which was collected was dried under
vacuum.

Analysis. Calc. for TiClACl6H36OAN: Ti, 9.67;
C1, 28.6A; C, 38.81; H, 7.31; N, 2.83. Found: Ti, 9.61;
C1, 28.70; C, 38.59; H, 7.52; N, 2.95.

The same material was also prepared from a solution

of 10% dioxane in chloroform.

E. Spectroscopic Measurements

 

Optical Spectra.——The infrared spectra of the compounds

 

were determined by means of Nujol mulls on sodium chloride
plates on a Unicam SP-200 spectrophotometer. Far infrared
spectra were recorded on a Perkin-Elmer A57 grating spectro-
photometer with the sample enclosed in cesium iodide or
polyethylene plates. The visible and ultraviolet spectra
were obtained by use of a Unicam SP-800 spectrophotometer
and a Cary 1A spectrophotometer. Solid samples were mulled
with Nujol and enclosed in quartz plates. As most of the
Compounds dissolve only with reaction no solution spectra
Of the complexes were obtained. A Beckman DU spectrophoto-

meter was used for the spectrophotometric analysis of titanium.

32

Electron Spin Resonance Spectra.-—The X—band esr

 

spectra of the solid complexes were recorded at 297°K and
77°K. Compounds which were thermochromic were also studied
at a range of intermediate temperatures. The X-band spectro-
photometers used were a Varian E-A esr spectrometer system
and the Varian A501-0A spectrometer previously described.“9
First derivative curves were recorded on an X—Y recorder
withthe X axis proportional to the magnetic field strength.
A sample of pitch in potassium chloride (g = 2.0028) was
used as a calibration standard. An E-257 variable tempera-
ture unit was used in the variable temperature studies.

At lower temperatures the microwave power had to be greatly
reduced to avoid saturation of the samples. The g values

were calculated from the klystron frequency and the measured

magnetic field.

F. Magnetic Moment Measurements

 

Magnetic susceptibilities were measured by the Guoy
method by use of equipment similar to that previously

described.69

A specially designed Dewar vessel was used
for temperatures lower than room temperature. A constant
flow of helium was used to provide heat transfer from the
sample and to prevent water from condensing on the sample

tube.

33

TABLE I.——Low Temperature Baths for Magnetic Susceptibilities.

 

 

Temperature Bath
77°K Liquid Nitrogen
1A8°K Petroleum Ether Slush
175.5°K Methanol Slush
195°K Dry Ice—isopropanol
209.7°K Chloroform Slush
250.3°K Carbon Tetrachloride Slush

 

The magnetic susceptibility was calculated from the

experimental data according to the equation

lo6x z d+B&F-5)

s
where x is the gram susceptibility of the sample, a is the
correction for the displacement of air by the sample, B is
the tube constant, F is the measured force on the sample
and the tube, 6 is the force on the empty tube, and WS is
the weight of the sample. The correction for displaced air,
a, was eliminated by weighing the empty tube filled with
nitrogen. The constants B and 5 must be measured for every
field strength and temperature used. Although borosilicate
glass is diamagnetic, it was found that impurities caused
0 to be positive and on the order of 10-15 milligrams.

To evaluate B the tube was standardized with Hg[Co(SCN)u]
Which has a susceptibility of 16.AAx106 cgs units of 20°.

This standard obeys the Curie-Weiss law with xmm(T+10)_l.7O

3A

The magnetic susceptibilities of all compounds were
measured at 77, 195, and 297°K. If the sample exhibited
normal paramagnetism the magnetic susceptibilities were also
measured at several intermediate temperatures so that an
accurate plot of the effective moment against gg-could be
obtained. As is usual for titanium(III) solvent adducts,
no corrections were made for the small temperature inde-
pendent paramagnetism of the samples. The molar susceptibility
was obtained by multiplying the gram-susceptibility by the
molecular weight. The metal ion susceptibility was then
found by applying corrections for the diamagnetism of the

ligands as computed from Pascal's constants.71

RESULTS AND DISCUSSION
A. Preparation and Reactions of the Complexes

During the course of this research the complexes of
titanium(III) chloride with several coordinating solvents
were studied. A new class of compounds, containing two
different non-halide ligands, as well as several new compounds
of classes previously reported were prepared. In addition,
several known compounds were prepared to complete the
characterizations. The reactions of all these compounds
were studied, especially with regard to non-halide ligand
substitution.

Titanium trichloride reacted with the coordinating
solvents acetonitrile, tetrahydrofuran, and isopropanol to
form the adducts TiCl °3CH

CN, TiCl -3CuH80, and TiCl ~Ac H80

3 3 3 3 3
reSpectively. When the adduct of one solvent was dissolved

in another solvent, all of the non-halide ligands were replaced
by the latter solvent. However, TiC13(CuH80)2(CH30N) was

formed when TiCl °3CuH80 was dissolved in a solution of 20%

3
acetonitrile in tetrahydrofuran.

The solid adduct produced by the reaction of titanium(III)
chloride with dioxane was TiCl3°2CuH802 in which only two
dioxanes were coordinated with the titanium. Clark gt gt.8

had concluded from the uv spectra that a different species

35

36

exists in solution though the species was not identified.

Fowles gt a1.9 reported that a dioxane solution of the very

soluble complex TiCl '3CAH8O2 gave a similar spectrum.

3

Fowles gt 21- also reported that with prolonged reaction

times (up to sixty hours) TiCl 'CAH802 was formed. However,

3
in this investigation reaction times as long as 2A0 hrs.

produced only TiCl ‘ZCAH8O2° As each titanium was coordinated

3

to only two dioxane molecules, compounds of the type
TiCl3(CuH802)2L were prepared by the addition of the solvent

ligand L to TiCl '2CAH802 suspended in dioxane. This type

3
of complex was also produced when a titanium(III) solvent

adduct was dissolved in dioxane.

When the solvent adducts TiCl °nL (L = acetonitrile,

3

tetrahydrofuran, dioxane) were heated under vacuum, compounds

-L were formed. Fowles and Russ25 had

3
reported that a green product was obtained when TiC13°3CH3CN

of the type TiCl

was heated one-half hour at 100°. In this research it was
found that longer heating times were necessary and that the
substance which remained after the evolution of a gas had
ceased was brown. However, elemental analysis of this

-CH CN.

3 3
'3CAH8O was heated under vacuum at 75° it

substance showed it was TiCl

When TiCl3

first turned light green and then gray-green. The light
green substance could not be isolated as the color of the
material was never uniformly light green. Analysis of the
gray-green compound showed it was TiCl

3
and Thurn2u reported that a green compound, TiCl

'CuHBO. Seifert

3‘2CuH80,

37

was produced when TiCl '3CAH80 was heated under vacuum at

3
70°. If the temperature was then increased to 90° another

mole of tetrahydrofuran was lost and gray TiCl 'CAH8O remained.

3
During the course of the present research it was also found
that if the samples had been previously exposed to air a
yellow sublimate was also formed. The sublimates were
primarily TiClu°2L'butcontained some titanium(III). The
TiClu-2L complexes and TiCl3-CuH80 were thermochromic.

When compounds of the type TiC13(CuH802)2L were heated under
vacuum at 70° they eliminated the ligand L to form

TiCl3-20uH802. When the temperature was raised to 100°

a mole of dioxane was then lost to produce the final product

TiCl3-CuH8O2. TiCl3(CuH8O)2(CH3CN) yielded TiCl

when heated under vacuum to 80°.

3'CuH80

Generally the compounds of the type TiCl3-L reacted
with coordinating solvents in the same manner as titanium
trichloride. However, in dioxane the ligand L was not
eliminated when the titanium coordinated with two dioxane
molecules. This is a convenient method of preparation of
2)2L (L = acetonitrile,
tetrahydrofuran). Attempts were made to produce other

the complexes TiC13(CuH80

compounds by similar methods. After the addition of

acetonitrile to TiC13~CuH80 or TiC13-CuH8O2 in stoichlometrlc

amounts to form TiCl3(CH3CN)2(CuH80) and TiC13(CH3CN)2(CuH802)

respectively, the starting materials were recovered unchanged.

38'

When a stoichiometric amount of tetrahydrofuran was added
to TiCl3-L (L = acetonitrile, dioxane) to produce compounds
of the type TiCl3(CuH80)2L, the only product was
TiCl3-CuH80.
Both hydrogen chloride and isopropyl alcohol were
liberated when TiCl3-A03H80 was heated under vacuum at 75°.
The red material which remained was (probably) a polymeric E1
chloroalkoxo complex with the simplest formula T12C13(C3H70)3.
This material did not appear to react with coordinating
solvents which indicated that the alkoxide ligands were not A
easily replaced. r
The reaction of tetraethylammonium chloride with

titanium(III) adducts TiCl onL in tetrahydrofuran or dioxane

3
produced compounds of the type (02H5)uN[TiClu-2L]. However,
in acetonitrile the reaction yielded only non-stoichiometric
materials with a variable chloride content. In a chloroform-
acetonitrileor chloroform-dioxane solvent system the compounds
(C2H5)N[TiClu-2L] (L = acetonitrile, dioxane) were produced.
When the reaction took place in a chloroform-tetrahydrofuran
system, or when (C2H5)N[TiC1u~2CuH80] was washed with
chloroform, the tetrahydrofuran was eliminated from the
complex and the composition of the solid product approached
that of (02H5)uN[TiC1u]. If complexes of the type LiC13L2L'
were used in place of TiCl3-nL in the reactions, the ligand

L' was eliminated and (02H5)uN[T101u-2L] was formed.

39

B. Infrared Spectra

 

The infrared spectra of the complexes were determined
in Nujol mulls and the results are presented in figures 3
through 17. In general the spectra showed the shifts in
absorption bands associated with coordination of the solvents.
The CN stretching vibration of acetonitrile (2253.5 cm-l)
was shifted upward 35 cm.1 upon coordination through the ]
nitrogen. The combination band at 2288 cm-1 was shifted
upward 25 cm-l. In tetrahydrofuran coordination through
the oxygen was indicated by the lowering and splitting of E
the symmetric (909 cm-1) and asymmetric (1071 cm-1) C—O-C
stretching vibrations of the free ligand by ’60 cm-l.
Coordination through oxygen in dioxane lowered and split
the free ligand bands occuring at 1125 and 883 cm-1. The

spectrum of TiCl '2CAH8O2 showed the presence of both free

3
and coordinated oxygen while the spectrum of TiCl3'CuH8O2
showed only coordinated oxygen. It was apparent from the
spectrum of Ti2CI3(C3H7O)3 that the compound was a
chloroalkoxo complex and no isopropyl alcohol remained.

The spectra of complexes of the type TiC13L2L' showed some
shifts of absorbtion bands from their positions in the
parent complexes TiCl3onL and TiCl3-nL'. Considerable
changes in the spectra were observed when the chloride:non-

halide ligand ratio of the complexes was increased. These

Changes may have arisen from changes in both the symmetries

A0

H. f

L

 

 

 

 

 

 

 

 

 

 

l A

4000 2000 1600 1200 800

CH!

 

Figure 3. Infrared spectrum of TiCl3s3CH3CN.

’1 r1
A W l 7
l

l h.-- l A I l A _L

4000 2000 1600 1200 800
cm"1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure A. Infrared spectrum of TiCl3s3CuH80.

Al

/\ «

 

 

 

 

 

 

 

 

 

 

 

L l

4000 2000 1600
cm—1

 

 

 

1200 800

 

Figure 5. Infrared spectrum of TiCl3-204H802.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. A U L 1-1.“.

4000 2000 1600 1200 800
cm"1

 

Figure 6. Infrared spectrum of TiCl3~AC3H8O.

A2

wfl ,.

j A L 1 1 L UL

4000 2000 1600 1200 800
cm-1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 7. Infrared spectrum of TiCl3(CuH80)2(CH3CN).

 

“N 0

 

 

 

 

 

 

 

 

 

 

 

‘ L A A l 1 A 4 A L

4000 2000 1600 1200 800
cm‘ 1

 

 

 

Figure 8. Infrared spectrum of TiCl3(CuH802)2(CH3CN)-

A3

 

 

A A 7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A 1 l j j l l 1

4000 2000 1600 1200 800
cm’1

 

Figure 9. Infrared spectrum of TiCl3(CAH802)2(CAH80)'

 

fl M (HM

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.01.”..-

4000 2000 1600 1200 800
cm"1

 

Figure 10. Infrared spectrum of TiC13(CuH802)2(C3H8O).

 

AA

 

/\

 

 

 

 

 

 

 

I l l L

4000 2000 1600

cm

 

1200 800

Figure 11. Infrared spectrum of T1013~CH30N.

”n

 

/\l

 

 

 

 

 

 

 

 

 

 

- A A 4 A L _L A U.

4000 2000 1600 1200 800

cm"1

 

Figure 12. Infrared spectrum of TiC13~CuH80.

 

A5

 

A

 

 

 

 

 

 

 

 

 

 

L n L a l l L

4000 2000 1600 1200 800

cm

 

 

Figure 13. Infrared spectrunlof TiCl3~CuH802.

 

”\W
A A A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A n A 1 1 L 4 L l 4

4000 2000 1600 1200 800
cm"1

 

Figure 1A. Infrared spectrum of T12013(C3H70)3.

A6

 

 

 

 

 

 

 

 

 

 

 

 

J 1 l L

 

 

lie L n L

4000 2000 1600 1200 800

cm

 

Figure 15. Infrared spectrum of (C2H5)4N[T101u-2CH CN]

3

 

 

D N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b

.1 A A A I A A _L A

4000 2000 1600 1200 800
cm'1

 

 

Figure 16. Infrared spectrum of (C2H5)uN[TiClu~2CuH8O]

A7

 

/\ n

 

 

 

 

 

 

 

 

 

 

 

 

l A

1200 800

 

 

P l l I l

4000 2000 1600
cm—1

 

Figure 17. Infrared spectrum of (C2H5)4N[T101u-2CAH802]

A8

and the ligand field strengths of the complexes. The far
infrared spectra absorption bands were too broad to provide
useful information about the symmetries or the ligand bond

strengths.

C. Electronic Absorption Spectra

 

The electronic absorption spectra of the complexes were .
determined in Nujol mulls and the results are presented in .1
Table 11. Titanium was six-coordinate in all of the complexes
which were studied. Although the symmetries of the complexes
varied widely, all the spectra except that of (02H5)uN[TiCln] j
had two d-d absorption bands. Since no amount of trigonal
distortion of the basic octahedral point group symmetry would
have removed the degeneracy of the upper 2E d orbital level,
the complexes contained some elements of tetragonal or
lower symmetry.

The spectrochemical series for the ligands studied in
this research is CH3CN>C3H80> CAH802>CAH8O' The locations

of the absorption bands of the complexes of the type

TiCl L L' were about 100 cm"1

3 2

from Jorgensen's rule of average environment and the Dq values

higher than the values predicted

of the ligands. In the [TiClu-2LJ- complexes the locations
of the absorption bands were almost identical, although

' ~
Jorgensen 3 rule would have predicted VCH3CN>vCAH8O2~VCAH8O.

A9

 

 

 

TABLE II.-—Electronic Absorption Spectra of Compounds.a’b

-1 c

Compound 2 Absogption Bands (gm ) 2
B2 " B1 32” Al

TiCl3°3CH30N 17,100 1A,700
T1013'3CMH80 1A,700 13,500
TiCl3-2CuH802 1A,800 12,800
TiCl3-AC3H80 16,700 1A,300
TiC13(CuH80)2(CN3CN) 15,600(15,550) 14,000(13,900)
TiCl3(CuH802)2(CHZCN) 15,680(15,570) 13,325(13,A33)
TiC13(CuH802)2(CuH80) 1A,780(1A,770) 13,100(13,03A)
TiCl3(CuH8)2)2(C3H80) 15,A00(15,1A0) l3,220(l3,110)
TiCl3-CH3CN 16,000(1A,832) 11,900(12,900)
TiC13-CuH80 17,300(1A,o32) 13,000(12,500)
TiCl3-CuH802 17,900(1A,065) 1A,600(12,270)
T12C13(C3H7O)3 18,950 17,A50
(C2H5)uN[TiC1u-2CH3CN] 15,150(15,800) 1A,150(13,800)
(C2H5)uN[T101u-2CAH8O] 1A,900(1A,200) 1A,150(l3,000)
(C2H5)uN[TiCIu°2CuH802] 15,000(1u,270) 1A,100(12,730)

[(C2H5)uN][TiCln]

12,Aoo

 

aall spectra were determined in Nujol mulls.

bvalues in parentheses are those expected from Jorgensen's rule
of average environment and the average Dq values of the

ligands.

cassignments assume tetragonal or lower symmetry.

"fl

50

Only one low-lying absorption band was seen for the material
with non—integral stoichiometry (C2H5)uN[Ti01n]. The absorption
bands in compounds of the type TiCl3tL were considerably higher
than would have been expected from Jorgensen's rule. In
addition,the position of acetonitrile in the spectrochemical
series was shifted so that it became the weakest of the ligands
studied. It was expected from Jorgensen's rule that the
location of the absorption bands would have shifted to lower
frequencies as the ch1oride;non-halide ligand ratio was
increased. However, in this investigation it was found that
the frequencies of the absorption bands increased with increas-

ing chloride content of the titanium(III) complexes.

D. Magnetic Moments

 

The magnetic susceptibilities and effective magnetic
moments of the complexes at several temperatures are presented
in Table 111. Most of the compounds exhibited normal Curie-
Weiss paramagnetism. The complexes TiCl -CH CN and

3 3

TiC13-CuH80, in which chloride ions probably bridge two
titanium atoms, were anti-ferromagnetic. A similar compound,
TiCl3-CuH802, exhibited normal paramagnetism but with a much
greater temperature dependence than the other paramagnetic
complexes studied. The chloroalkoxo complex Ti2Cl3(C3H7O)3
was diamagnetic.

Several ligand field parameters were obtained from

graphs of ”eff against.kT/A' (figures 18-29). The experimental

51

TABLE III.--Magnetic Properties of Titanium(III) Complexes.

 

 

' 6
Compound Tgmp. Xm x 10 “eff
K cgs units B.M.
TiC13-3CH3CN 297 12A1 1.72
250.3 1A20 1.69
209.7 16A5 1.67
175.5 1902 1.6A
1A8 2222 1.63
77 A019 1.58
TiCl3°3CuH80 297 12A6 1.72
250.3 1A38 1.70
209.7 16A6 1.67
196 1738 1.66
175.5 1931 1.65
1A8 2220 1.63
77 A021 1.58
TiCl -20 H O 297 1232 1.72
3 u 8 2 250.3 1391 1.68
209.7 161A 1.65
196 1705 1.6A
175.5 1873 1.63
1A8 2122 1.59
77 3670 1.51
TiC13-AC3H8O 297 1270 1.7A
250.3 1A16 1.69
209.7 1616 1.65
196 1692 1.6A
175.5 1859 1.62
1A8 2120 1.59
77 3738 1.52
TiCl (C H 0) (CH CN) 297 1381 1.77
3 u 8 2 3 250.3 1A91 1.73
209.7 1708 1.70
196 1796 1.69
175.5 1988 1.68
1A8 2311 1.66
77 A122 1.60
TiCl (C H 0 ) (CH CN) 297 1230 1.72
3 u 8 2 2 3 250.3 1370 1.66
209.7 1575 1.63
196 1621 1.60
175.5 17A5 1.57
1A8 1957 1.53
77 3302 1.A3

TABLE III.--Continued.

52

 

 

Compound Temp. Xm x “eff
K cgs units B.M.
TiCl (C H 0 ) (C H 0) 297 12A3 1.72
3 u 8 2 2 8 8 250.3 1A01 1.68
209.7 1587 1.6A
196 1686 1.63
175.5 1815 1.60
1A8 2060 1.57
77 3533 1.A8
TiCl (C H 0 ) (C H 0) 297 1255 1.73
3 8 8 2 2 2 8 250.3 1A28 1.70
209.7 1670 1.68
196 17A8 1.66
175.5 1926 1.65
1A8 2231 1.63
77 3928 1.56
T1013.CH3CN 297 789 1.37
196 971 1.2A
77 1928 1.09
196 356 0.75
77 305 O.A3
TiCl -C H O 297 1357 1.80
3 u 8 2 250.3 1A77 1.73
209.7 1625 1.66
196 1624 1.60
175.5 1776 1.59
1A8 1990 1.5A
77 2269 1.19
Ti2CI3(C3H7O)3 diamagnetic
(C H ) N[TiCl °2CH CN] 297 1305 1.77
2 5 u u 3 250.3 1A78 1.73
209.7 1661 1.68
196 1739 1.66
175.5 1918 1.65
1A8 2192 1.62
77 3761 1.53

Table III.--Continued.

53

 

 

Compound Tgmp. x 10 “eff

K cgs units B.M.

(C H ) N(TiCl °2C H O] 297 1318 1.78
2 5 u u u 8 250. 1A61 1.72
209. 166A 1.68

196 1758 1.67

175. 1870 1.63

1A8 2165 1.61

77 3668 1.51

(C H ) NETiCl °2C H O J 297 1265 1.7A
2 5 u u u 8 2 250. 139A 1.68
209. 1600 1.6A

196 1638 1.61

175. 1765 1.58

1A8 2008 1.55

77 32A2 1.A2

 

“eff

 

5A

 

Figure 18.

 

*4:

Graph of “eff vs. §$-ror T1013~3CH3CN.

Circles are experimental points.
Line calculated from reference AA with
k = 0.7, v = A.5, A' = 0.95 A.

 

“eff

 

55

 

 

 

 

 

0.2 0.A 0.6 0.8 1.0 1.2 1.A

kT
IT

Figure 19. Graph of Nerf vs. %$ for TiCl3s3CuH80.

Circles are experimental points.
Line calculated from reference AA with
k = 0.7, v = A.5, A' = 0.95 A.

ueff

56

 

 

 

Figure 20.

 

vs. §$-for TiC13~2CuH802.

Graph of “eff

Circles are eXperimental points.
Line calculated from reference AA with
k = 0.7, v = 3, A' = 0.95 A.

 

57

 

 

 

 

 

g
A!
Circles are experimental points.

Line calculated from reference AA with
k = 0.75, v = 3, A' = 0.95 A.

Figure 21. Graph of u vs. for TiCl ~AC H80.

eff 3 3

1.8

1.7

ueff

1.6

58

 

 

Figure 22.

 

0.4 0.6 0.8 1.0 1.2 1.A

kT
A!

kT
Graph of “eff vs. T7 for TiCl3(CuH80)2(CN3CN).

Circles are experimental points.
Line calculated from reference AA with
k=0.9,v=6,)\'=)..

1.7

1,6

ueff

1.5

1.A

59

 

 

 

 

 

 

’32.
A:
Circles are experimental points.

Line calculated from reference AA with
k=0.7, v=2, 1' =0.95 A.

Figure 23. Graph of “eff vs. for TiCl3(CuH802)(CH30N).

60

 

1.7

1.6

“eff

1.5

 

 

l.“

 

 

E.
it

Figure 2”. Graph of ”eff vs. §$ for TiCl3(CuH802)2(CuH8O).

l Circles are experimental points.
Line calculated from reference nu with
k = 0.7, v = 2.5, A' = 0.95 A.

61

 

ueff

 

 

 

 

Figure 25. Graph of “eff vs. g; for TiCl3(CuH802)2(C3H80).

Circles are experimental points.
Line calculated from reference HQ with
k = 0.7, v = u, A' = 0.90 A.

 

62

 

 

 

 

1.8 .
CD
L-
Ci
1.6 r . 0
ID
1:...
CH
391.“ r
1.2 _ .
1.0
l L *2— l
0.2 0.1: 0.6 0.8 1.0 1.2 1.14
kT
3‘...

Figure 26. Graph of “eff vs. §$ for TiCl3~CuH8O2.

Circles are experimental points.
Line calculated from reference uu with
k = 0.8, v = 0.1, A' = 0.90 A.

63

 

1.7

eff

3'1.6

1.5

 

 

 

 

0.2 0.“ 0.6 0.8 1.0 1.2 l.“

. kT O
Flgure 27. Graph of “eff vs. IT for (C2H5)uN[TiClu.2CH3CN].

Circles are eXperimental points.
Line calculated from reference 44 with
k = 0.8, v = 3.5, A' = 0.90 A.

6H

 

 

 

C
1.7 - ‘. l
O
O
O
O
21.6 -
(l)
1.5 , .
0.2 0.4 0.6 0.8 1.0 1.2 1.4
fl
Ah
Figure 28. Graph of "eff vs. §$ for (C2H5)HN[TiClu-2CMH80].

Circles are experimental points.
Line calculated from reference nu with
k = 0.8, v = 3.0, M = 0.90 A.

 

“eff

65

 

1.7 .
1.6 .
1.5 '
1,H .

 

 

 

 

 

 

Figure 29.

0.“ 0.6 0.8 1.0 1.2 l.“

kT
A“

kT .
Graph of “eff vs. YT for (C2H5)uN[TiClu‘2CuH8O2].

Circles are eXperimental points.
Line calculated from reference uu with
k = 0.75, v = 2.0, A‘ = 0.95 A.

66

curves were fitted to the proper theoretical curves calculated
from the values given by Figgis.uu Values of k, the orbital
delocalization factor, A', the spin-orbit coupling constant,
and v, the distortion factor, were obtained by this procedure.

The energy of separation, 6 of the orbital levels derived

1’
from the 2T2g ground state term was calculated from the equa-
tion v=6l/A. The value of 61 was positive in all of the
complexes which indicated that the orbital singlet derived

from the 2T state was the lowest energy level. The ligand

2%
field parameters which were calculated from the magnetic data
are given in Table IV.

The results were examined to find if any correlations
existed between the values of the ligand field parameters
3 and 61 and other ligand field parameters of the complexes.
Theoretically a complex which has an undistorted octahedral
symmetry should have a magnetic moment which falls to zero
as the temperature approaches zero. In complexes having lower
symmetry and t2g electron delocalization the moment approaches
the spin—only value. Thus it would be expected that as the
symmetries of the complexes were increased the magnetic
moments should be more temperature dependent. This was the
case in the series of compounds with different chloride:non-
halide ligand ratio in which both the amount of orbital
delocalization (l-k) and the energy of separation 61
decreased with increased chloride content. However, with
the complexes of the type TiCl L L' no correlations were

3 2
found between the values of the ligand field parameters k

 

 

67

TABLE IVr—Ligand Field Parameters Calculated from Magnetic Data.

 

 

Compound k v A'(Cm-l) 51(Cm-l)
TiCl3°3CH3CN 0.7 u.5 1M6 660
T1013-3cuH80 0.7 u.5 146 660
TiCl3°2CuH802 0.7 3.0 1u6 uuo
T1013-uc3H80 0.75 3.0 1M6 uuo
TiCl3(CuH80)2(CH3CN) 0.9 6.0 15M 925
TiC13(CuH802)2(CH3CN) 0.7 2.0 1M6 290
TiCl3(CuH802)2(CuH8O) 0.7 2.5 1M6 365
TiCl3(CuH802)2(C3H80) 0.7 u.0 139 550
TiCl3-CuH802 0.8 0.1 139 in
(C2H5)uN[TiClu-2CH3CN] 0.8 3.5 139 u85
(C2HS)HN[T101u-2cuH80] 0.8 3.0 139 M15
(C2H5)uN[TiClu-2CMH802] 0.75 2.0 1H6 290

 

68

and 6 and either the ligand field strength (lODq) or 6

l 2,
the energy of separation of the upper levels derived from

the 2E state.
2s

E. Electron Spin Resonance Spectra

 

The esr spectra of the powdered solid complexes were
determined at 297 and 77°K. The results are presented in
Table V. Figure 30 shows the three types of spectra generally
found. Spectra of the type A occur when g, the Landé spectro-
scopic splitting factor, is isotropic or, more commonly, the
temperature is too high to allow resolution of anisotropic
g values. Compounds with type B spectra generally have an
axis of symmetry and gx=gy¢gz. In the titanium(III) com-
pounds investigated during the course of this research
gi(gx and gy) was greater than g||(gz). Type C spectra are
expected when the complex has no axial symmetry and gx#gy#gz.
-3CH CN and TiCl

The esr spectra of TiCl ~3CuH8O are

38

3

symmetries.

3

3v 2v
the titanium atom is six—coordinate in TiCl

3

expected for the respective C and C Although

3-2CuH802 it is
not certain which functional group bridges the titanium
atoms. If the complex had a dioxane molecule in the sixth
position or the two dioxane molecules present were Erggg,

a type B spectrum would have been observed. Instead a
three-line type C spectrum was observed which suggested that
a bridging chloride ion occupied the sixth position and the
complex had a gig configuration. TiCl3-UC3H80, which had

only two chloride ions coordinated to the titanium atom,

also had a three-line type C spectrum and a cis configuration.

69

 

 

 

 

 

 

 

TYPE A
.-—
4
TYPE B
TYPE C

 

Figure 30. Representative electron spin resonance spectra.

70

TABLE V.-—The g Values from esr Spectra of the Complexes.a

 

 

Compound Temp. Type of
OK spectrum g1 82 83
TiCl3~3CH3CN 297 A 1.907
77 B 1.921 1.883
TiCl3s3CuH80 297 A 1.88u
77 B 1.89u 1.8u9
TiCl3-2CuH802 297 B 1.972 1.880
77 C 1.982 1.880 1.826
T1013.403H80 297 A 1.901
77 C 1.977 1.899 1.828
T101 (0 H 0) (CH CN) 297 A 1.909
3 u 8 2 3 77 A 1.909
TiCl3(CuH8O2)2(CH3CN) 297 C 1.963 1.890 1.836
77 C 1.987 1.903 1.837
TiCl (c H 0 ) (C H 0) 297 A 1.885
3 u 8 2 2 u 8 77 c 1.939 1.898 1.827
T101 (0 H 0 ) (c H 0) 297 A 1.887
3 u 8 2 2 3 8 77 B 1.911 1.855
T1013~CH30N 297 A 1.911
77 A 1.907
TiCl3.CuH80 297 A 1.898
77 A 1.898
TiCl .c H 0 297 A 1.879
3 u 8 2 77 A 1.881
(C2H5)uN[TiClu~2CH3CN] 297 c 1.9u1 1.869 1.807
77 c 1.9u0 1.869 1.806
(C2H5)uN[TiClu‘2CuH80] 297 c 1.977 1.885 1.796
77 c 1.978 1.886 1.796
(c H ) N[TiCl .20 H 0 ] 297 0 1.983 1.886 1.815
2 5 u u u 8 2 77 c 1.983 1.885 1.813

 

a.
for S GCtI’a Of ’6 e B P - and = o

71

Compounds of the type TiCl3L2L' have no axis of symmetry

and were expected to give a three-line type C spectrum.
However, this type of spectrum was observed only for

T1C13(CuH802)2(CH3CN) and TiCl3(CUH802)2(CHH80)' The

unsymmetrical spectrum of TiCl3(CuH80 (C3H80) may have

2)2
been an unresolved three—line spectrum. The spectrum of
TiCl3(CuH80)2(CH3CN) remained a single absorption peak even
at 77°K. This was caused by the large splitting of the
ground state 2T2g (61=990 cm-1 from the magnetic data) which
quenched the orbital angular momentum and caused g to become
approximately isotropic. The spectra of compounds of the

type TiCl -L were expected to show the axial symmetry of these

3
complexes. However, the spectra had only a single absorption

peak. The spectrum of TiCl ~CuH802 was very broad due to

3
the small separation 61 of the t2g sublevels. As TiCl3-CHBCN
and TiCl3-CuH8O were antiferromagnetic, the separation energy

could not be evaluated from the magnetic data. Powdered
antiferromagnetic substances generally give broad, inhomogeneous
resonance spectra.72 Compounds of the type (C2H5)uN[TiC1u-2L]
gave three-line type C spectra even at room temperature. The
spectra indicated that the compounds have gig configurations.

As the symmetries of the complexes became more distorted
from octahedral symmetry, the orbital angular momenta were
quenched and the average g values were expected to approach
2.00. Since most of the complexes had average g values close

to 1.89 it was difficult to establish the existence of any

72

definite correlations. However, the distortion of the complexes
and the approach of the average g values to 2.00 appeared to
be in the same order as the ligand field strengths of the

complexes with the ligands in the order CH CN>C

3 3H80>

CMH802~CHH80>C1'

F. Thermochromism

 

Several of the complexes prepared during the course of
this investigation were thermochromic. The compounds and the
color changes that occur when they are cooled to 77°K are
given in Table VI. Thermochromism may be caused by a change
in phase or ligand geometry, in which case the color change
is abrupt, or a narrowing and shifting of charge transfer
bands into the ultraviolet region of the spectrum which causes
a gradual change in color. There have been no previous

reports of thermochromic titanium complexes.

TABLE VI.--Thermochromic Titanium Complexes.

 

 

Compound Color at

297°K 77°K
TiCl3'CuH80 gray-green violet
TiClu-ZCH3CN yellow white
TiClu-2CuH8O yellow white
(C2H5)uN[TiClu-2CH3CN] green blue
I a
'(C2H5)uN[TiCln]" salmon yellow

aa material with non-integral stoichiometry.

73

Several other physical properties of the thermo-
chromic complexes were qualitatively examined to observe if
the color changes could be correlated with other changes.
The optical spectra showed a shift of the lowest-energy charge
transfer band toward higher energies when the compounds were
cooled. The shift was especially noticeable in the material
with non-integral stoichiometry "(02H5)uNETiClx(CAH80)yJ"’
obtained from (02H5)u[TiClA'2CuH8O]’ in which the charge-
transfer band at 21,000 cm-1 at room temperature merged into
the higher-energy bands. The charge-transfer band of
TiCl3vCuH8O was located such that the compound was thermo-
chromic. In TiCl °CAH802 and TiCl

°CH ON the energies of the

3 3 3
first charge-transfer bands were respectively too high and
too low for the compounds to exhibit thermochromism. No abrupt
changes were found in the magnetic behavior of the thermo—
chromic complexes as they were cooled.

Changes in the esr spectra and the average g value of
the complexes occured when several of the compounds were
cooled. The spectra of TiClu'2CH3CN and TiClu°2CuH80 had
a weak absorption peak at g=1.937 due to an unknown
titanium(III) impurity. As the compound was cooled the
spectrum of TiCl3-3CH3CN appeared and at 77°K the intensity
of this signal was very high. A similar change took place
with TiClu-2CuH80. The spectrum of "(C2H5)uN[TiClx(CuH80)yJ",

which was prepared from (C2H5)uN[TiClu-2CHH80], and

74

(C2H5)uN[TiClu-2CH3CN], which appeared to have been partially
oxidized, showed only a very sharp, single absorption peak
g=2.00 at room temperature. When the compounds were cooled
this peak gradually disappeared and the spectra of the
respective (C2H5)uN[TiClu-2L] type compounds appeared. The
changes were rapid and reversible as the temperatures were
changed. In contrast, pure samples of the respective tetra-
ethylammonium titanium complexes did not exhibit thermo-
chromism or unusual esr spectra. It is felt that the
thermochromism and the changes in the esr spectra of these
compounds is somehow related to the presence of both
titanium(III) and titanium(IV) in the samples though the
connection is not apparent. The esr spectrum of TiCl3-CuH80
did not change except for the customary narrowing when the

compound was cooled.

BIBLIOGRAPHY

75

10.

ll.

12.

13.

BIBLIOGRAPHY

Giggenbach, W., and Brubaker, C. H., Jr., Inorg. Chem.

Bereman, R. D., and Brubaker, C. H., Jr., Inorg. Chem.
8, 2480 (1969).

Wentworth, R. A. D., and Brubaker, C. H., Jr., Inorg.
Chem., 1, A7 (1964).

Rillema D. P., and Brubaker, C. H., Jr., Inorg. Chem.,
9. 1655 (1969).

Reagan, W. J., and Brubaker, C. H., Jr., Inorg. Chem.,

Duckworth, M. W., Fowles, G. W. A., and Hoodless, R. A.,
J. Chem. Soc., 5665 (1963).

McDonald, G. D., Thompson, M., and Larsen, E. M., Inorg.
Chem., 7, 648 (1968).

Clark, R. J. H., Lewis, J., Machin, D. J., and Nyholm,
R. S., J. Chem. Soc., 379 (1963).

Fowles, G. W. A., Hoodless, R. A., and Walton, R. A.,
J. Chem. Soc., 5873 (1963).

Fowles, G. W. A., Greene, P. T., and Lester, T. E.,
J. Inorg. Nucl. Chem., 29, 2365 (1967).

Schumb, W. C., and Sundster, R. F., J. Amer. Chem.
300-. 22. 596 (1933).

Fowles, G. W. A., "Progress in Inorganic Chemistry,"
Vol. VI, Cotton, F. A., ed., Interscience Publishers,
New York, New York, 1964.

Antler, M. and Laubengayer, A. W., J. Amer. Chem. Soc.,
11. 5250 (1955).

76

1A.

15.

l6.

l7.

l8.

19.

20.

21.

22.

23.
2A.

25.

26.
27.

28.

29.

30.

77
Ehrlich, P., and Siebert, W., Z. Anorg. Allg. Chem.,
303, 96 (1960).

Fowles, G. W. A., Hoodless, R. A., and Walton, R. A.,
J. Inorg. Nucl. Chem., 21, 391 (1965).

Siddiqui, M. T., Ahmad, N., and Rahman, S. M. F., Z.
Anorg. Allg. Chem., 336, 110 (1965).

Clark, R. J. H., and Greenfield, M. L., J. Chem. Soc.,
A. “09 (1967).

Edwards, D. A., Fowles, G. W.A., and Walton, R. A., J.
Inorg. Nucl. Chem., 21, 1999 (1965).

Giggenbach, W. and Brubaker, C. H., Jr., Inorg. Chem.,
1. 129 (1968).

Schlafer, H. L., and Gotz, R., z. Anorg. Allg. Chem.,
iéé, l (1964).

Fowles, G. W. A., Lester, T. E., and Russ, B. J., J.
Chem. Soc., A, 805 (1968).

Fowles, G. W. A., and Walton, R. A., J. Chem. Soc.,
4953 (1964).

Winter, G., Inorg. Nucl. Chem. Lett., 2, 198 (1966).

Seifert, J. J., and Thurn, F. M., Z. Anorg. Allg.
Chem., 372, 79 (1970).

Fowles, G. W. A., and Russ, B. J., J. Chem. Soc.,
A. 517 (1967).

Kern, R. I., J. Inorg, Nucl. Chem., 23, 1105 (1962).

Schlafer, H. L., and Gotz, R., Z. Phys. Chem. (Frankfurt
am Main), 52, 97 (196A).

Hartmann, H., Schlafer, H. L., and Hansen, K. H., Z.
Anorg. Allg. Chem., 289, MO (1957).

Schlafer, H. L., Lenz, W., and Staab J., Z. Phys. Chem.
(Frankfurt am Main), 62, 290 (1968).

Gutmann, V., Leitmann, 0., Scherhauger, A., and Czuba,
H., Monatsh. Chem., 28, 188 (1967).

31.

32.

33.

34.
35.

36.
37.

38.

39.

40.

41.

42.

43.
44.

45.
46.

47.

48.

78
Hathaway, B. J., and Holah, D. G., J. Chem. Soc., 537
(1965).

Ballhausen, C. J., and Gray, H. B., Inorg. Chem., l,
111 (1962).

Figgis, B. N., "Introduction to Ligand Field Theory,"
Interscience Publishers, New York, New York, 1966.

Jorgensen, C. K., Acta. Chem. Scand., 22, 73 (1957).

Leihr,IL, and Ballhausen, C. J., Ann. Phys., 1; 304
(1958).

Clark, R. J. H., J. Chem. Soc., 417 (1964).

Jorgensen, C. K., "Absorption Spectra and Chemical
Bonding in Complexes," Pergamon Press, New York,
New York, 1962, p. 109.

Clark, R. J. H., Spectrochim. Acta, 2;, 955 (1965).
Figgis, B. N., and Lewis, J., "Progress in Inorganic
Chemistry," Vol. VI, Cotton, F. A., Interscience

Publishers, New York, New York, 1964.

Earnshaw, A., "Introduction to Magnetochemistry,"
Academic Press, New York, 1968.

Van Vleck, J. H., "Electric and Magnetic Susceptibilities,"
Oxford University Press, London, 1932.

Cotton, F. A., and Wilkinson, G., "Advanced Inorganic
Chemistry," Interscience Publishers, New York, 1966
p. 641.

Kotani, M., J. Phys. Soc. Jap., 3, 293 (1949).

Figgis, B. N., Trans. Faraday Soc., 51, 198 (1961).

Dunn, T. M., J. Chem. Soc., 623 (1959).

Earnshaw, A., Figgis, B. N., and Lewis, J., J. Chem.
Soc., A, 1656 (1966).

Lewis, J., Machin, D. J., Newham, I. E., and Nyholm,
R. S., J. Chem. Soc., 2036 (1962).

Klemm, W., and Krose, E., Z. Anorg. Allg. Chem., 253,
209 (1947).

 

49.

50.

51.

52.

53.
54.
55.

56.
57.
58.

59.

60.

61.
62.

63.

64.

65.

79

Bereman, R. D., Ph. D. Thesis, Michigan State University,
1969.

Kuska, H. A., Ph. D. Thesis, Michigan State University,
1965.

Carrington, A., and Longuet-Higgens, H. C., Quart. Rev.
(London), 15, 427 (1960).

Gladney, H. M. and Swalen, J. D., J. Chem. Phys., 32,
1999 (1965).

Bijl, D., Proc. Phys. Soc. (London), A63, 405 (1950).
Bleany, B., Proc. Phys. Soc. (London), A63, 407 (1950).
Bleany, B., Bogle, G. S., Cooke, A. H., Duffus, R. J.,
O'Brien, M. C. M., and Stevens, K. W. H., Proc.
Phys. Soc. (London), A68, 57 (1955).
Jarrett, H. S., J. Chem. Phys. 21, 1298 (1957).
McGarvey, B. R., J. Chem. Phys., 28, 388 (1963).

Waters, E. L., and Maki, A. H., Phys. Rev., 125, 233
(1962).

Gair'yanov, N. S., Danilova, A. V., and Shagidullin, R. R.,
Opt. Spektrosk., 11, 212 (1962).

Giggenbach, W. and Brubaker, C. H., Jr., Inorg. Chem.,
1. 129 (1968).

Giggenbach, W., unpublished results.

Bereman, R. D., and Brubaker, C. H., Jr., J. Inorg.
Nucl. Chem., in press.

Herzog, S., Dehnert, J., and Luhder, K., "Technique
of Inorganic Chemistry," Vol. VII, Jonassen, H. B.,
and Weissberger, A., eds., Interscience Publishers,
New York, New York, 1968.

Perrin, D. D., Armarego, W. L. F., and Perrin, D. R.,
"Purification of Laboratory Chemicals," Pergamon Press,
Oxford, 1963.

Carter, H. V., McClelland, B. J. and Warhurst, E, Trans.
Faraday Soc., fig, 343 (1960).

66.

67.

68.

69.

70.

71.

72.

80

Walton, H. F., "Principles and Methods of Chemical
Analysis," Prentice Hall, Inc., Englewood Cliffs, New
Jersey, 1964, p. 236.

International Union of Pure Applied Chemistry, Commission
on Spectroscopy and Other Optical Procedures for
Analysis, "Tables of Spectrophotometric Absorption
Data of Compounds Used for the Colorimetric Deter-
mination of the Elements," London: Butterworths, 1963.

Figgis, B. N., and Lewis, J., "Modern Coordination
Chemistry," J. Lewis and R. G. Wilkins, Ed. New York:
Interscience Publishers, Inc.

Vander Vennen, R. E., Ph. D. Thesis, Michigan State
University, 1954.

Figgis, B. N. and Nyholm, R. S., J. Chem. Soc., 4190
(1958).

Foex, G., "Constantes Selectionnes: Diamagnetisme et
Paramagnetisme. Relaxation Paramanetiquex," Masson
& Cie, Paris, 1957.

Bersohn, M. and Baird, J. C., "An Introduction to
Electron Paramagnetic Resonance," New York: W. A.
Benjamin, Inc., 1966.