THE PREPARATION AND CHARACTERIZATION
OF SOME ALKOXO-NIOBIUM (IV)
COMPLEXES

Thesis for H10 Degree of DH. D.
MICHIGAN STATE UNIVERSITY
Rupert A, D. Wentworth

1963

 

M .p ,U-c—n. M ‘ -A...

 

v—y

THESIS f) LIBRARY I”: }

1" :‘

\v .1..— . . .
.; [m 1ch1gan State I

i Unwcth f ‘
I

3“" i *-—-I ,4

rv—v

MICHIGAN STATE UN'VERSITY

EAST LANSING, MICHIGAN

 

ABSTRACT

THE PREPARATION AND CHARACTERIZATION OF SOME
ALKOXO-NIOBIUM(IV) COMPLEXES

by Rupert A. D. Wentworth

A new class of crystalline niobium(IV) compounds,
the pentachloroalkoxoniobates, have been prepared and character-
ized. The alkoxo group was either methoxo, ethoxo, or
isopropoxo, while the cation was usually a large protonated
organic base (alkylammonium, pyridinium, quinolinium, etc.).
The color of the resulting compound is dependent on the
nature of the cation.

The compounds were prepared by the electrolytic
reduction of NbCl5 dissolved in anhydrous alcohol saturated
with anhydrous HCl, followed by the addition of a hot
alcoholic HCl solution of the organic base.

Magnetic studies indicate Curie-Weiss paramagnetism
and verify the presence of the niobium(IV) ion with 4 d1
configuration. The infrared spectra of the compounds is in

agreement with the formulation [Nb(OR)C15]-2, showing

intense absorption in the region of 1000-1100 cm.-l, which

’

Rupert A. D. Wentworth

is typical of the C—0 stretching vibration in alkoxides, but
not showing absorptions characteristic of the OH group. The
C-O stretching frequency is dependent upon the nature of the
cation to a small degree, while the ring C-H stretching
frequency for the quinolinium compounds is shifted toward
lower frequencies than that observed in (C9H8NIC1 and
(C9H8N)2C0C14.

The dependence of the color upon the nature of the
cation is obvious from the visible reflectance spectra of
the compounds. The alkylammonium compounds show a single
band at about 19,600 cm.-l. The maxima in the spectra of
the pyridinium compounds are always shifted toward higher
frequencies, and are very asymmetric. The spectra of the
quinolinium compounds are much more complex with a broad
absorption band extending from 15,400 to about 22,100 cm.”1
This band seemingly consists of two or three poorly resolved
components.

The limited range of the available spectrophotometer
(14,300 to 25,000 cm.-l) makes unequivocal interpretation of
the spectra of even the simply alkylammonium compounds
impossible. However, the color dependence appears to arise

from an interaction of the aromatic cations with the complex,

possibly through the alkoxo oxygen, resulting in the observed

Rupert A. D. Wentworth

shift in the ring C—H and alkoxo C—O stretching frequencies.

Attempts to prepare the [I\Ib(OR)Br5]—2 and [NbOC15]—3
anions were unsuccessful.

Further investigation of the ethyl alcohol—niobium(IV)
system resulted in the isolation of [NbCl(OC2H5)3(C5H5N)]2.
Forty-three isomers of this dimer are possible. The
inertness of the chloride ions toward solvolysis and pre—
cipitation suggests that chloride bridging exists, reducing
the number of possible isomers to five.

The complexity of the infrared spectrum in the region
between 1000 and 1100 cm._1 is explained in terms of a non—
equivalence of bonding between the oxygens and the metal ion,
resulting from differing trans groups.

The diamagnetism of the dimer is thought to be a
result of a direct overlap of the adjacent metal dXy orbitals.

The reaction of [NbCl(OC2H5)3(C5H5N)]2 with sodium

ethoxide results in the formation of Nb(OC H5)4, which is

2
extremely unstable to hydrolysis and is difficult to purify.
The diamagnetism of an impure sample suggests that it is also
polymeric, but molecular weight studies were not attempted
because of the compound's instability.

Both [NbCl(OC2H5)3(C5H5N)]2 and Nb(OC2H5)4 are
readily converted to (C5H6N)2[Nb(OC2H5)C15], which demonstrates

the equivalence of the oxidation state in the three compounds.

THE PREPARATION AND CHARACTERIZATION OF SOME

ALKOXO-NIOBIUM(IV) COMPLEXES

BY

Rupert A. D. wentworth

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1963

ACKNOWLEDGMENT

The author wishes to express his sincere gratitude

to Professor Carl H. Brubaker, Jr. for his guidance, kindness,

and ready wit; and to the author's wife, Virginia, whose struggles

and spartan existence made this investigation possible.
Appreciation is also extended to R. A. Wentworth

and C. E. Rance for their regular monthly endowments: and

to the Dow Chemical Company, the Eastman Kodak Company, and

the National Science Foundation for financial assistance

during the years of 1961-1963.

ii

TABLE OF CONTENTS

I. INTRODUCTION
II. EXPERIMENTAL . . . . . . . . . . . . . .

Materials

Analytical Methods
Electrolysis Apparatus
Preparation of Compounds

Attempts in Aqueous Solutions

The Pentachloroalkoxoniobates

Properties of the Pentachloroalkoxo—
niobates

Attempts to Prepare the
Pentabromoalkoxoniobate(IV) Anion

Attempts to Prepare the
Oxopentachloroniobate(IV) Anion

Dichlorohexaethoxobis(pyridine)-
diniobium(IV) and its Reactions

Tetraethoxoniobium(IV) and its
Reactions

Spectroscopic Measurements

Magnetic Moment Measurements

Attempts to Obtain X-Ray Diffraction Data
III. DISCUSSION OF RESULTS . . . . . . . . . .

The Pentachloroalkoxoniobates

Dichlorohexaethoxobis(pyridine)diniobium(IV)

Tetraethoxoniobium(IV)

REFERENCES . . . . . . . . . . . .

iii

Page

12
14

14
15

19
20
20
21
25
27
28
32
33
33
56
61

68

LIST OF TABLES

Table Page
I. Magnetic susceptibilities and moments of

some of the alkoxo-niobium(IV)

compounds . . . . . . . . . . . . . . . . 31
II. A comparison of the infrared spectra of

pyridinium ion and ligand pyridine with

that of (C5H6N)2[Nb(OCH3)C15] . . . . . . 34
III. A comparison of the infrared spectra of

quinolinium ion and ligand pyridine

Wlth that of (C9H8N)2[Nb(OCH3)C15] . . . 35
IV. Boiling or sublimation temperatures of

metal ethoxides . . . . . . . . . . . . . 63

iv

Figure

10.

ll.

12.

LIST OF FIGURES

Electrolysis vessel

Infrared spectra of (C9H8N)2[Nb(OC2H5)C15]

and (C9H8N)2[Nb(OC3H7)C15] from 2200 to
3600 cm.‘1 (Perfluorokerosene mull)

Infrared spectra of (C9H8N)Cl and (C9H8N)2CoCl

from 2200 to 3600 cm.-1(Perfluorokerosene
mull) . . . . . . . . . . . . .

4

Infrared spectrum of [(CH3)5N]2[Nb(OC3)C15]
(Nujol mull)

Infrared spectrum of (C H N) [Nb(OCH )Cl ]
. 5 6 2 3 5
(NuJ Ol mull) O O O O O O O O O O O 0

Infrared spectrum of (C H N) [Nb(OCH )Cl ]
. 9 8 2 3 5

(Nu-J Ol mull) . O C O O O C C . 0
Infrared spectrum of [(CH3)2NH

_[Nb(OC H )Cl ]
(Nujol mull) . . . . . .2.5 .5.

212

Infrared spectrum of (C H N) [Nb(OC H )Cl ]
. 5 6 2 2 5 5
(NuJ Ol mull) . C O O O O O O . O .

Infrared spectrum of (C9H8N)

[Nb(OC H )01 ]
(Nujol mull) . . . . . g . . .2.5. .5.

Infrared spectrum of (C9H8N)2[Nb(OC3H7)C15]
(NuJol mull) . . . . . . . . . .

Visible reflectance spectra of some
pentachloromethoxoniobates

Visible reflectance spectra of some
pentachloroethoxoniobates

V

Page

13

37

38

39

40

41

42

43

44

45

49

SO

Figure

13.

14.

15.

16.

17.

Visible reflectance spectra of some
pentachloroisopropoxoniobates

I
Infrared spectrum of [NbC150C2H5)3(C5H5N)]2

(Nujol mull)

Isomers of [NbCl(OC H ) (C H N)] with
. . .2 5 3 5 5 2
chloride bridging . . . . . .

Infrared spectrum of NbIOC H5)4 (Nujol mull)

2

Infrared spectrum of Nujol

vi

Page

51

57

59
65

67

INT RODUC TI ON

Niobium, the congener of vanadium, is commonly
found in the pentavalent oxidation state. However, it has
been known for some years that compounds with lower oxidation
states could be prepared in either solid or aqueous phases.

As early as 1878, Roscoel reported the preparation of niobium
trichloride by the thermal decomposition of the pentachloride,
and in 1912 Ott2 demonstrated that niobium(III) could be
prepared by the electrolytic reduction of mineral acid
solutions of niobium(v). Later work has confirmed the
existence of the tetravalent, trivalent, divalent, and
possibly the zero-valent states, and several examples of
compounds containing mixed oxidation states.

Examples of compOunds containing niobium(IV) have
only recently been prepared and examined. Brubaker3 pre—
pared the first example, NbC14, by the reduction of the
pentachloride with niobium metal. He described this compound
as a crystalline, black material which hydrolyzes and oxidizes

readily. Above 2500 it disproportionates and forms the

trichloride and the pentachloride. Its spectrum in

ethylene glycol monomethyl ether, which undoubtedly com-
plexes the niobium ion in some manner, consists of a single,
broad, asymmetric maximum at about 20,200 cm.“1 (490 mu)
with a shoulder at about 16,000 cm.—1 (625 mu).

Later work led to the preparation of NbBr44 and
NbI4.5 All of the halides were found to be diamagnetic.

Dahl6 has examined the structure of NbI4 and found it to

consist of infinite linear chains of NbI6 octahedra sharing
opposite edges. The niobium atoms are shifted from the
centers of the octahedra toward one another in pairs, re-
sulting in a niobium-niobium distance of only 3.31%. Weak ‘
metal-metal interactions which couple the unpaired electrons
explain these shifts and account for the observed diamagnetism.
The oxidation state of four is confirmed from the observed
identical environments of the niobium atoms.

A qualitative interpretation of the nature of the
bonding about each niobium atom has.been given7 in terms
of simple molecular orbital theory based on octahedral
symmetry. The 4dzz and 4de—y2 metal orbitals plus one 53
and three 5p metal orbitals interact with the orbitals of
corresponding symmetry on the iodide ions to give bonding

o—type molecular orbitals. The unshared electron on each

niobium can be placed in the dxy orbital which overlaps with

an identical orbital on the adjacent niobium atom, giving
rise to the observed diamagnetism. The unoccupied dxz
and dyz orbitals are no doubt utilized in W-bonding with
the filled W—type symmetry orbitals of the iodides.

The existence of the simple hexachloro—complexes
of niobium(IV) has been noted in a phase study of the
tetrachloride with alkali metal chlorides.8 Unfortunately,
the properties of these compounds were not noted.

There is also other evidence for the existence of
this oxidation state. Niobium pentachloride is readily

reduced by pyridine and lithium salts of dialkylamides to

4 and Nb(NR.2)4,lo respectively. The

give (CSHSN)2NbC1 9
insolubility of the former has precluded a molecular weight
determination, but the latter compound has been shown to be
monomeric in benzene, and cannot be an equimolar combination
of niobium(III) and niobium(V). Other examples of
niobium(IV) are found in cyclopentadienyl compounds. 1'12

I The blue-violet (C5H5)4Nb, obtained from the reduction of

the pentachloride with a large excess of CSHSNa is monomeric
in’benzene.
Polarography and spectrophotometry have also

provided information about this oxidation state. Gozzi

and Vivarellil3 polarographically reduced hydrochloric acid

solutions of niobium(V) and observed the visible spectra
of the reduced solutions. In izug HCl, for example, a
red solution was obtained which had a single maximum at
20,900 cm.—l(478 mu) while in 8.§ HCl a blue solution was
obtained with a single maximum at 14,300 cm.-1 (700 mu).
In the more acidic solution they arbitrarily concluded that
the [NbC16]_2 ion is present while in the latter [Nb0C15]-3
or [Nb0C14]-2 are probable. Similar one-electron reductions
have been observed in ethylenediaminetetraacetate, citrate,
lactate, and gluconate solutions.14

The solid compounds discussed thus far bear little
stoichiometric resemblance to those containing other metal
ions with one d electron (d1) e.g. metals from Groups
IV B, V B, and VI B. In these compounds stabilization
of the metal ions by halides, pseudohalides, or oxygen—
containing ligands seems to be necessary. Only rarely is
perfect cubic symmetry attained and most of the complexes
exhibit strong tetragonality.

In Group IV B, only titanium(III) is found in

simple complexes. Two hydrates exist, both having the

stoichiometry TiC13-6H20. Presumably the violet form is

15

[Ti(H20)6]C13 and the green form is [Ti(H20)5Cl]C12'HZO.

It also exists in several anionic halo-complexes, such as

5

(NH4)3[TiF6] and Rb2[TiC15(HZO)] .16 The magnetic moment of
this metal ion is generally very close to the spin—only
value of 1.73 Bohr magnetons.l

Of the d1 metal ions in Group V B, excluding niobium,
only the complexes of vanadium(IV) are known. These are
almost always tetragonal and are vanadyl (VO+2) derivatives,
although this is probably due to their method of preparation.
Many different compounds are known and include aqua, chloro,
sulfato, oxalato, and acetylacetonato derivatives. They are
lall exceptionally stable to oxidation. In general, the
magnetic moment of the metal ion in these compounds is quite
close to the spin-only value.18 One example of supposed
perfect cubic symmetry is KZIVF6].19 However, the magnetic
moment is anomolously high.

Chromium(V), molybdenum(V), and tungsten(V) are
found in a completely analogous series of compounds having
the general stoichiometry Mé[MOX5] (where M' is an alkali
metal ion, M is the Group VI B metal ion, and X is chloride
in the case of Chromium(V) or fluoride, chloride, bromide,
or thiocyanate in the cases of molybdenum(V) and tungsten(V)).
The magnetic moment of Chromium(V) in Rb2[CrOC15] is about
2.1 Bohr magnetons.2O This high value is attributed to the

presence of trivalent chromium. The moments of molybdenum(V)

in a great number of these compounds are very close to the

. 2 . .
spin-only value, 1 while tungsten(V) has not received a
thorough treatment as yet.

Recently the compound C H6N[Mo(OCH3)2C14] was

5

22 . .
prepared during the course of a study of the alcoholySis
of MoClS. Unfortunately, no spectra or magnetic suscepti-
bilities are presented.

These metal ions have come under the close scrutiny

of several theoretical chemists within the last year. In

a series of paper318'23'24 a complete molecular orbital
+ .—

treatment has been applied to VO(H20)52, CrOCl5 2, and

MoOClS—Z. It was shown that simple crystal field theory

could not be adequately utilized with these ions because

of extensive W-bonding between the oxo ligand and the metal

ion, so much so that the linkage is formulated as M E O.

This interaction results in a contraction along this axis

which is, of course, observed in compounds such as VOSO4'5H20.25
The niobium(IV) compounds prepared during the course

of this research, especially those containing the novel

[Nb(OR)C15]-2 ion, are also tetragonal, although probably

not as much as the oxo Species of vanadium, chromium, and

molybdenum, and should be useful to the theoretical chemist

in attaining a more complete knowledge of the electronic

1 .
structure of d metal ions.

EXPERIMENTAL

Materials

 

Niobium Pentahalides.-—Niobium pentachloride was obtained as

 

a gift from the Diamond Alkali Company and was used without
further purification. The manufacturers claim the purity

of this compound was 99.8% with a maximum tantalum con—
centration of 0.02%.

Niobium pentabromide was prepared by elementary
synthesis using niobium sheet purchased from Fairmount
Chemical Company. This was claimed to have 99.8% purity
and contained a maximum of 0.01% tantalum. The reaction
was accomplished in a sealed, L—shaped tube so that the
liquid bromine was at room temperature while the metal was
held at 400°. The bromine was distilled from phosphorus(V)
oxide in a vacuum line into tube containing a weighed amount
of niobium. The tube had been previously dried by flaming
and pumping at 0.01 mm. for four hours. The tube was then
sealed under vacuum and the reaction was allowed to proceed.
When all the metal had reacted, one end of the tube was
cooled in an acetone-dry ice bath until all the bromine had

7

distilled into the cold end, which was then sealed from
the pentabromide with a torch. The criterion for purity was

the absence of any orange niobium oxybromide.

Substituted Ammonium Halides.-—Tetramethylammonium chloride
(Eastman White Label grade) was dried at 800 prior to use.
Alcoholic solutions of dimethylammonium chloride were pre-
pared by allowing a known amount of anhydrous dimethylamine

to be absorbed in the alcohol saturated with HCl.

Pyridine, Quinoline, and Isoquinoline.--These were all Eastman
White Label grade. Pyridine was dried by distillation from
barium oxide and stored in a dry container. Quinoline-and

isoquinoline were used without further purification.

N-Methylpyridinium and N-Methquuinolinium Iodides.-—
N-Methylpyridinium iodide was prepared by allowing stoichio-
metric quantities of pyridine and methyl iodide to react.
Two recrystallizations from ethyl alcohol—benzene mixtures
. o . . o 26

gave yellow, aCicular crystals, m.p. 116.5 (lit. 118 ).

N—Methquuinolinium iodide was prepared and purified
similarly to yield white, acicular crystals, m.p. 145o

(lit. 144.5°).27

Hydrogen Halides and Nitrogen.——Anhydrous hydrogen chloride
and bromide, in cylinders were obtained from the Matheson
Chemical Company and used without further purification.
Nitrogen was purified by passing General Dynamics,
. . . o
Oil-pumped nitrogen over copper turnings at 500 . It was
then bubbled through concentrated sulfuric acid and passed

over anhydrous magnesium perchlorate.

Solvents.--Methy1 alcohol was dried by reaction with
magnesium and subsequent distillation. For the synthesis of
the compounds containing the pentachloroalkoxoniobate(IV)
anions, ethyl and isopropyl alcohols were dried by reaction
with sodium and distilled. However, for the neutral alkoxo-
chloro and alkoxo compounds, completely anhydrous alcohols
were necessary. They were obtained by azeotropic distillation
with benzene under a nitrogen atmOSphere.

Anhydrous benzene was obtained by azeotropic distil—
lation with ethyl alcohol. Chloroform was dried over calcium

chloride and distilled prior to use.

Analytical Methods

Niobium and Halide Analyses.--Samples of the complexes were
oxidized in very dilute nitric acid, made basic with aqueous

ammonia, and digested on a steam bath for several hours.

10
With the compounds [NbCl(OC2H5)3(C5H5N)]2 and Nb(OC2H5)4,

which are insoluble in water, a better procedure consisted
of dissolution in a small amount of ethyl alcohol, addition
of about 10 m1. of 1;M_ammonia, followed by evaporation to
dryness on a steam bath. The residue was then treated in
exactly the same manner as the water-soluble complexes.

The solutions were adjusted to pH = 1 with nitric
acid and filtered to remove the hydrous niobium pentoxide.
Niobium was determined gravimetrically by ignition of the
hydrous oxide. Chloride or bromide was determined by
potentiometric titration of the filtrate with standardized

AgNOB.

Carbon and Hydrogen Analyses.——These were performed by

Spang Microanalytical Laboratory, Ann Arbor, Michigan,

Pyridine Analysis.-—In several instances, a rapid method
for the determination of the pyridine content of a compound

was desirable. Accordingly, the following spectrophotometric

-4
method was developed. Solutions of pyridine up to 1.7 x 10

.M in 0°05.§ H2804 were found to obey Beer's Law at 255.mu

(e = 5.01 x 103 1. mole.1 cm. ). Samples were digested

in aqueous solutions of H2804 of known concentration,

filtered: and the solutions were made up to one liter with

11

sufficient H.2804 and water so that the final concentration

was 0.05 N. The pyridine content was determined spectro-

photometrically at 255 mu.

Alkoxide Determination.--The ethoxide content in Nb(OC2H5)4

was determined by the quantitative oxidation of ethyl alcohol

with K2Cr207.45 To a weighed sample of the tetraethoxide,

a weighed amount of K2Cr207 was added and several ml. of

12.5% H2804 was added. After two hours the unreacted

Cr(VI) was determined iodometrically.

Determination of the Oxidation State of Niobium.--The oxidation
state of a few samples was determined by oxidation of the

compound with standard aqueous KZCrZO7,

followed by back
titration of the excess Cr(VI) with a standard solution of
Fe(NH4)2(SO4)2°6H20. The end point was determined ampero-
metrically using a rotating platinum electrode and a calomel
reference electrode. The current was measured with the

galvonometer of a Sargent Model XXI polarograph.

Molecular Weight.--The molecular weight of [NbCl(0C )

2H5 3(C5H5N)]2
was determined ebullioscopically by Huffman Microanalytical
Laboratories, Wheatridge, Colorado. Due to the sensitivity

of the compound to oxidation, the measurement was performed

12

in a dry nitrogen atmosphere. The stated precision was

i 10%.

Electrolysis Apparatus

 

The electrolysis vessel (Figure 1) consisted of a
125 ml. Erlenmeyer flask with a ground glass top. The flask
was fitted with a two—way stopcock on the bottom, and a bent
side-arm which contained a fine porosity fritted glass disc.
The vertical portion of the side-arm served as the anode
compartment. The level of the mercury pool cathode could
be adjusted by a leveling bulb which was connected to one
arm of the two—way stopcock by a short piece of flexible
tubing. The anode compartment contained the appropriate
alcohol saturated with HCl and a short piece of carbon rod
which served as the anode. The solution was stirred during
the electrolysis by bubbling dry, purified, nitrogen through
the electrolyte and allowing it to escape through an
opening in the top of the flask. The nitrogen also served
to exclude air and moisture.

The current was provided by a Heathkit variable'
voltage power supply (Model PS—3) capable of producing a

maximum current of 0.2 amp.

 

/ Ԥ.\

 

7:

5.

.—. ~-

" :2
m.‘.

~-

--

-

    

.3 ,_ ,_
45;;Enfaflfll ‘" -

 

 

 

 

Figure 1. Electrolysis vessel.

14

Preparation of Compounds

 

Attempts to Prepare Solid Compounds Containiana Reduced

Niobium Species from Aqueous Solutions. Sulfuric Acid

 

Solution.--Sulfuric acid solutions of niobium(V) were pre-
, . . . .7 28 . .
pared by dissolv1ng hydrous niobium pentox1de (containing
between 30-40% H20) in concentrated H2804 at 2100 with
stirring, cooling to 00, and slowly adding water dropwise
to the solution. It was found after repeated athempts that
an acid concentration of 6 M.was necessary to prevent precipi-
tation in solutions that were 0.04 M in niobium.
A solution of these concentrations was electrolyzed
, . 2
at 6.7 volts and a current den51ty of 0.003 amp/cm.
Titration showed the niobium to be trivalent. The solution
was blue and this color was retained on an anion exchange
column, but was readily eluted with water. Attempts were
next made to precipitate the anion using cations of varying
, , , + + + +
Size and charge. The addition of Na , K , NH , (CH3)4N ,

4
[Co(en)3]+3, [Co(NH3)5C1]+2, and [Co(en)2(C204)]+ in 3.M

H2804 failed to produce any precipitate even when present

in large excess.

_Hydrochloric Acid Solutions.—-Solutions that were 0.005M

in niobium(V) were easily prepared by the solution of the

15

pentachloride in concentrated aqueous HCl. Electrolysis over
extended periods of time at a potential of about 10 volts
and a current of 0.15 amp. produced solutions containing
trivalent niobium.
When these solutions were evaporated in a vacuum,
a bright blue cake would form before the solvent was completely
removed. This product was amorphous and contained only small
amounts of chloride ion. It is undoubtedly a hydrous oxide
formed by the hydrolysis of the anionic chloro—niobium species.
Attempts were made to precipitate this anionic
Species by the addition of Na+ or K+ ions and saturation of
the solution with HCl at —200 with no success.
It is unfortunate that the limited supply of the
pentachloride at that time prevented a more detailed study

at higher niobium concentrations.

Compounds Containing the Pentachloroalkoxoniobate(IV) Anion.
Because of the lack of success in preparing simple com-

plexes in aqueous solutions, the electrolysis of alcohol-HC1

solutions of niobium(V) was investigated as a possible route

to such complexes.

Methoxo complexes. Pyr id inium Pentachloromethoxoniobate(IV) . —-

A complete description of the preparation of this compound

“H.

«I.

NJ.

fl.»

5

we.

16

will be given and applies to the preparation of all the
complexes of this type. A solution of niobium(IV) was pre-
pared by the electrolytic reduction of the pentachloride

(4.0 g., 0.015 mole) in 25 m1. of anhydrous methyl alcohol.‘s?II
The potential was adjusted so that a current of 0.075 amp. III!
was maintained throughout the electrolysis. The solution
rapidly became wine—red and then changed to deep brown after
several hours. No hydrogen evolution was noted until about

the theoretical time for a one—electron reduction. Con-
currently, the thin layer of solution between the flask and

the meniscus of the mercury became yellow while the bulk

of the solution remained dark brown. Titration experiments
with portions of the solution showed that the niobium was
completely tetravalent at this point. On one occasion,

the period of electrolysis was extended by 300%, but no

further reduction of the niobium was noted.

The mercury was then removed and the solution drained
into a flask containing slightly more than the stoichiometric
quantity<xfpyridine in hot methyl alcohol saturated with.HC1.
The compound which resulted after several hours of cooling,
was obtained as a red-brown, spiny, crystalline agglomerate.
Filtration was accomplished under a nitrogen atmosphere. The

crystals were washed with three 20 ml. portions of chloroform,

17

and dried by continuous pumping for 24 hours.
Anal. Calcd. for (C5H6N)2[Nb(OCH3)C15]:C, 28.61;
H, 3.27; Cl, 38.45; Nb, 20.13, CSHSN' 34.3. Found: C, 28.67;

H, 3.30; Cl, 38.21; Nb, 20.21, CSHSN, 34.7. Oxidation State:

4.03.

When larger excesses of pyridine were used in the
precipitation, the product was an amorphous, brown cake.
When cold methyl alcohol was substituted for chloroform in
the washing of the crystals, the chloride to niobium ratio
was only 4.88 indicating that some solvolysis had probably

occurred.

Tetramethylammonium pentachloromethoxoniobate(IV).-ePink
crystals after a few minutes. Anal. Calcd. for
[(CH3)4N]2 [Nb(OCH3)C15]: Cl, 39.5; Nb, 20.68. Found: C1, 39.1:

Nb, 20.77.

Quinoliniumypentachloromethoxoniobate(IV).-—Violet crystals

after 20 min. Anal. Calcd. for (C9H8N)2[Nb(OCH3)C15]:
Cl, 31.6: Nb, 16.55. Found: Cl, 31.2: Nb, 16.71.
Isoquinoliniumppentachloromethoxoniobate(IV).-—Red crystals

after 10 min. Anal. Calcd. for (CgHéN)2[Nb(OCH3)C15]:

C1, 31.6; Nb, 16.55. Found: C1, 31.2; Nb, 16.58.

18

Ethoxo Complexes. Dimethylammoniumypentachloroethoxoniobate(IV).--
Peach needles within one minute. Anal. Calcd. for

[(CH3)2NH [Nb(OC2H5)C15]: c1, 43.5; Nb, 22.80. Found:

212
C1, 43.5: Nb, 22.87.
Pyridiniumypentachloroethoxoniobate(IV).—-Orange crystals
within a few minutes. Anal. Calcd. for (CSHSN)2[Nb(OC2H5)C15]:
C, 30.30; H. 3.56; C1, 37.33: Nb, 19.54. Found: C, 30.62;

H, 3.71; Cl, 37.21; Nb, 19.41.

N—MEthylpyridinium pentachloroethoxoniobate(IVL.--Red
crystals after 5 min. Anal. Calcd. for (C6H8N)2[Nb00C2HS)ClS]:

Cl, 35.3. Found: Cl, 35.0.

4—Picoliniumpentachloroethoxoniobate(IV).--Orange crystalline

 

polyhedra after a few minutes. Anal. Calcd. for

(C6H8N)2[Nb(OC2H5)C15]: c1, 35.3; Nb, 18.45. Found:

C1, 35.5; Nb, 18.59.

‘Qpinolinium pentachloroethoxoniobate(IV).-—Crystallization
of this violet compound was initiated by cooling at -770

until one crystal formed, and holding the Solution at

-10° for several hours. Anal. Calcd. for (C9H8N)2[Nb(OC2H5)C151:

C, 42.91; H, 3.79: Cl, 30.80: Nb, 16.12. Found: C, 42.88:

19

H, 3.67; Cl, 30.44; Nb, 15.73.

N-Methylguinolinium pentachloroethoxoniobate(IyL.—-Violet

crystals after 10 min. Anal. Calcd. for (C H N)2[Nb(OC2H5)Cl ]:

10 10 5

Cl, 29.40. Found: Cl, 29.5.

Isoprppoxo complexes. Tetramethylammonium_pentachloro-

 

isopropoxoniobate(IV).--Pink curdy precipitate immediately.

This was not investigated further.

Pyridiniumypentachloroisoprppoxoniobate(IV).--Orange powder

 

formed immediately. Anal. Calcd. for (C5H6N)2[Nb(OC3H7)C15]:

C1, 36.3; Nb, 18.98. Found: Cl, 35.9; Nb, 19.10.

.QuinoliniumypentachloroisoprppoxoniObate(IVL.-—Violet powder

formed immediately. Anal. Calcd. for (C9H8N)2[Nb(OC3H7)C15]:

C1, 30.1; Nb, 15.76. Found: C1, 30.0; Nb, 15.86.

Physical Properties of the Pentachloroalkoxoniobates.--They
are extremely sensitive to air, turning blue and then white.
They are soluble in water to give a brown solution from which
slowly deposits a hydrous brown oxide. In alcohols, solution
is accompanied by decomposition; and the color of the solution
depends on the water content of the alcohol. If the alcohol

is completely anhydrous, the solution is blue; but in the

20

presence of traces of water, the color is brown.

In dimethylformamide, the solution is initially violet,
but quickly changes to blue. These solutions are extremely
sensitive to oxidation, however, and would require the utmost
care in handling.

They are insoluble in dimethylsulfoxide, and common

organic solvents, such as benzene, ethers, and ketones.

Attempts to Prepare Analogous Bromo Complexes.

Several attempts to prepare compounds containing
the [Nb(OC2H5)Br5]_'2 ion, by the electrolytic reduction of
the pentabromide in ethyl alcohol saturated with HBr,
failed. Crystalline substances containing no definite bromide
to niobium ratio were obtained when the pyridinium and
quinolinium cations were employed.l For example, the material
containing the pyridinium ion had a Br/Nb ratio of 5.74.
Significantly, both substances were green and their infra-

red spectra showed no alkoxo groups to be present.

. . -3
Attempts to Prepare Compounds Containing the [NbOClS] Ion.

 

Several attempts to prepare compounds of this class
failed. Two basic methods were tried and will be described.
When 95% and 90% ethyl alcohol solutions (saturated

with HCl) of NbClS(0.6 M3 were electrolyzed and added to

21

stoichiometric quantities of pyridine in the same solvent,
the product in both cases (C5H6N)2[Nb(OC2H5)C15]. This
was verified in both cases by the infrared spectra and by
analysis in former case. .Aggl. Calcd.: C1, 37.3.
Found: Cl, 36.9.

From an 80% ethyl alcohol solution (saturated with
HCl) having the same niobium concentration, a very dark
green solution was obtained after electrolysis and the addition
of pyridine. A dark powder precipitated upon cooling. After
washing with two portions of freshly distilled chloroform and
drying by continuous pumping, a pale olive cake was obtained.
Its infrared spectrum showed no alkoxide to be present, but
the analytical data left considerable doubt concerning the
true nature of the product. 'Agal, Found: C1, 28.0;
Nb, 30.1: Cl/Nb, 2.43.

When the hydrolysis of several of the alkoxo
compounds was attempted with aqueous HCl (oxygen-free)
in the presence of stoichiometric quantities of the organic
base, considerable oxidation always occurred even when extreme

precautions were taken to exclude oxygen.

Preparation and Properties of Dichlorohexaethoxobis(pyridinel—
diniobium(IV).

Preparation.--Niobium pentachloride (6.20 9., 0.0229 mole)

22

in 25 ml. of anhydrous ethyl alcohol saturated with anhydrous
HCl was reduced electrolytically, and the resulting solution
evaporated to dryness. The tacky, dark substance which re—
sulted was dissolved in 20 m1. of anhydrous ethyl alcohol
and 5 m1. of pyridine was added. The color of the solution
changed from a deep purple to an equally deep yellow-red.
Crystallization began within one hour, and was seemingly
complete after another nine hours. The red compound was
collected on a filter under nitrogen atmosphere and washed
with three 10 ml. portions of cold ethyl alcohol: yield,
5.04 g. (64%,based on NbClS). Concentration of the mother
liquor produced only a small amount of the compound.

Anal. Calcd. for [NbC1(OC2H5)3(C5H5N)]2:

CSHSN' 23.1; C1, 10.4: Nb, 27.1. Found: CSHSN' 23.2;

C1, 10.3; Nb, 27.1. Molecular weight calcd. for dimer: 686.
Ebullioscopic molecular weight in CHC13: 674. Molecular
complexity: 1.97.

The dimer is very soluble in chloroform, and
slightly soluble in ethyl and isopropyl alcohols at room
temperature. It is insoluble in acetone, benzene, carbon

tetrachloride, nitrobenzene, dimethylsulfoxide, dimethyl—

formamide, 1,4-dioxane, glacial acetic acid, methylene

23

chloride, l,l'}l"-trichloroethane, and pentachloroethane.
In oxygen-free water, it is insoluble.

When heated in a sealed evacuated tube, this compound
turns brown at about 1000 and liquifies with gas evolution

(pyridine ?) at about 1550.

Reaction of [NbCl(OC2H5)3(C5H5N)]2 with Pyridinium Ion in

 

Alcoholic HCl. A Demonstration of the Oxidationqugte of
Niobium in the Dimer.--To 5 ml. of ethyl alcohol saturated
with HCl, 0.20 g. of the dimer was added. The mixture was
heated gently to accelerate the solution process, and the
color of the solution changed from purple to ochre.

Pyridine (0.2 ml.) was added, and the solution cooled at

-10° for twenty—four hours. The small, bright orange crystals
were collected on a filter under nitrogen, washed with two

10 m1. portions of CHCl and dried by continuous pumping

3!
for twelve hours; yield, 0.12 g. (43%»based on the dimer).
The infrared spectrum of this compound was identical to that

of an authentic sample of (C H6N)2[Nb(OC2H5)C15]. Anal.

5

Calcd. for (C5H6N)2[Nb(OC2H5)C15]: Cl, 37.3. Found: C1, 36.8.

Attempted Solvolysis of [NbC1(OC2H5)3(C5H5N)]2 with Isopropyl

 

Alcohol.-—A small amount of the dimer (ca. 0.1 g.) was heated

24

to 1500 in a sealed, evacuated tube with 25 ml. of isopropyl
alcohol for twenty-four hours. The tube was then opened

in a nitrogen atmosphere, the solvent was distilled away

and a brown, tacky substance remained. Large amounts of
pyridine were found in the distillate. Analysis of the
residue indicated that no solvolytic replacement of chloride
had occurred, and that the nature of the residue was un—

certain. Anal. Calcd. for NbCl(OC2H5)3(C3H50H): Cl, 11.0:
Nb, 29.5. Calcd. for NbC1(OC3H7)3: C1, 11.5; Nb, 30.1.

Found: Cl, 11.7; Nb, 28.7; Cl/Nb, 1.07.

Attempts to Precipitate Chloride Ion from [NbCl(OC2H5)3(C5H5N)]2

 

with Alcoholic Solutions of AgNO3and KI.—-When a concentrated

alcoholic solution of AgNO3.was added t0'a small amount of
the dimer dissolved in anhydrous ethyl alcohol, immediate
decolorization of the solution occurred and was accompanied
by the precipitation of a gray material along with a milky
solid. The gray solid was separated by filtration and washed
successively with alcohol and aqueous ammonia. Qualitative”
analysis indicated it to be silver metal.

A saturated solution of KI was added to a solution

of the dimer without producing any visible effect, even

after heating for several minutes. If a labile chloride

25

ion had been present, then a precipitate of KCl would have

been expected.

Preparation and Reactions of Tetraethoxoniobium(IVL.
Preparation. Attempts by the Ammonia Method.--When an

electrolytically reduced solution of NbCl (2.7 9., 0.0010 mole)

5
in 25 ml. of anhydrous ethyl alcohol saturated with HCl was
evaporated to dryness, a tacky, dark, unknown substance was
obtained in which the Cl/Nb ratio approximated two. The
passage of large quantities of gaseous ammonia, which was
distilled from sodium, into a solution of the substance in
20 ml. of ethyl alcohol and 20 m1. of benzene resulted in
the precipitation of NH4C1, but without any noticeable

change in the color of the solution. The solution was filtered,

and evaporated to give a brown solid containing 8.63% Cl.

Preparation by the Reaction of [NbCl(OC2H5)3(C5H5N)I2

 

with NaOC2H5.--In a typical preparation, 4.97 9. (0.00724 mole)

of [NbCl(OCZH5)3(C5H5N)]2 was added to a hot solution of

0.01448 mole of NaOCZHgin 50 m1. of anhydrous ethyl alcohol.

The bright blue precipitate, which formed initially, slowly
disappeared and a deep red-brown solution formed. This was
evaporated to dryness and the residue dissolved in 40 ml.

of benzene. This solution was filtered to remove NaCl, and

26

evaporated under vacuum at 600 to yield an extremely hydro?
scopic red—brown oil.

Anal. Calcd. for Nb(OC2H5)4: Nb, 34.1; Cl, 0.00.

Found: Nb 35.9; Cl, 0.34. In a second preparation,
found: Nb, 33.0; Cl, 0.47.

Exposing the substance for short periods of time to
the so-called anhydrous nitrogen atmosphere of a dry box
caused the color to change to deep brown and the niobium
content to rise steadily, indicating that hydrolysis was
occurring.

A quantity of the oil was placed in a sublimator
which contained a dry ice-ethylene glycol monomethyl ether
(methyl cellosolve) coolant in the cold-finger. A tacky,
red-brown solid began to collect on the cold finger at 0.002 mm.
and a bath temperature of 1600. A dark brown solid residue
remained in the sublimator.

Anal. Calcd. for Nb(OC2H5)4: Nb, 34.1, OC2H5 , 65.9

Cl, 0.00. Found: Nb, 34.5; OC2H5' 64.9; Cl, 0.00;

OCZHS/Nb, 3.90.

Reaction of Nb(OC2H5)4 with Pyridinium Ion in Alcoholic HCl.

 

A Proof of the Oxidation State of Niobium.-—Unsub1imed

Nb(OC2H5)4 (0.15 g., 0.00055 mole) was dissolved in 3 ml.

27

of anhydrous ethyl alcohol saturated with HCl, and 0.2 m1.

of pyridine were added. The color of the solution changed
from deep brown to ochre. The solution was cooled at -100 for
forty-eight hours. The small, orange crystals were collected
on a filter under nitrogen, washed with two 10 m1. portions

of CHC13, and dried in a vacuum overnight. The infrared
spectrum was identical to that of an authentic sample of

(C5H6N)2[Nb(OC2H5)C15]. Total yield: 0.23 g. (89% based on

Nb(OC2H5)4). Anal. Calcd. for (C5H6N)2[Nb(OC2H5)C15]:
Cl, 37.3. Found: Cl, 37.2.

Spectroscopic Measurements

 

The complete infrared spectra of the compounds were
obtained in NUjol mulls and the use of a Perkin—Elmer Model
21 spectrophotometer. The spectra from 2400-3600 cm.-1 were
obtained with perfluorokerosene mulls and a Beckman IR—7
spectrophotometer. The mulls were prepared in a dry box
under a nitrogen atmosphere.

The visible solution spectra were obtained with a
Beckman Model DK-2 Spectrophotometer, while the reflectance
spectra were obtained with a Bausch and Lomb spectrophoto-
meter with a reflectance attachment. MgCO was used as a

3

standard. In order to prevent oxidation of the solid

28

samples, they were ground with mineral oil and spread between

two circular, microscope cover glasses.

Magnetic Moment Measurements

Magnetic susceptibilities were measured by the
Gouy method using the apparatus and techniques of Vander
Vennen.29 The calculation was facilitated by use of the
equation:

106 x = (Di-fl.— (1)
where X is the gram-susceptibility of the sample; a is a
constant related to the volume of displaced air, and is
simply 0.029 x sample volume; V is the tube constant: F'
is the force experienced by the sample alone, i.e., the
measured force corrected for the force experienced by the
tube alone; and w is the sample weight.

In practice the constant I must be determined for a
particular tube using a material of known susceptibility.
In this research freshly distilled water was used, whose
susceptibility is known to be -0.720 x 106 with dX/dt =
0.0012/degree at 20°.31

The molar susceptibility of the sample is obtained
by multiplying the gram-susceptibility by the molecular

weight. The susceptibility of the metal ion Xi is

29

obtained simply by correction of the molar susceptibility

. . . , 32
for any diamagnetic speCies present. Pascal s constants
were used in this research to estimate the diamagnetism of
the ligands and cations, with the exception of the pyridinium
and quinolinium ions. The values for these were obtained

33
from Holm and Cotton.
In normal paramagnetic substances Xi is related

simply to the absolute temperature as

. s .
XM T (Curie Law) (2)

or C

T-‘PG

 

Xh = (Curie-Weiss Law) (3)

For the latter case a plot of 1/Xh against T allows the
evaluation of C and 6 from the slope and intercept.
A quantity which is most familiar to chemists is

the magnetic moment H of an ion, which is related to Xh as

2
X. = N020 or N06
M 3kT 3k(T+9)

 

(4)

It is easily seen by comparing (2) or (3) with (4) that

-mé
c—3k (5)

where N is Avagadro's number: 3 is the Bohr magneton; and k
‘18 Boltzmann's constant. Simplification leads to

1/2

H = 2.84 C (6)

30

Since the alkoxo compounds are extremely sensitive
to oxidation and hydrolysis, it was necessary that they be
protected from the atmosphere during a measurement. After
W was experimentally determined, the tube was carefully filled
with the ground sample and then sealed. A knowledge of the
weights of the filled tube and glass fragment allowed the
calculation of the sample weight.

The low-temperature studies were performed using a
specially constructed Dewar flask29 and methyl cellosolve-
dry ice and liquid nitrogen coolants.

The results are shown in Table I. The susceptibilitiescmf
the pentachloroalkoxoniobateS{follOW' the Curie-Weiss Law in
all cases, and correspond closely to the spin—only moment for
a d1 metal ion.

One preparation of [NbCl(OC2H5)3(C5H5N)]2 was Slightly
paramagnetic, while another was strongly‘diamagnetic. The
positive value can undoubtedly be attributed to paramagnetic
impurities. Since the magnetic purity of even the second
sample was unknown, it must be regarded as only an indication
of the diamagnetism of the dimer. Similarly, due to the
small supply of relatively pure Nb(OC2H5)4 and its extreme
instability to hydrolysis, the sample was impure and the
susceptibility can only be taken as an indication of complete

spin-pairing.

31

Table I. Magnetic susceptibilities and moments of some of
the alkoxorniobium(IV) compounds.

 

 

 

Temp. Susceptibility 0
Compound QA Xh x 106 6 A H
(CSHSN)2[Nb(OCH3)C15] 299 1020 61 1.72
196 1430
77 2700
(C9H8N)2[Nb(OC2H5)C15] 298 1150 19 1.72
196 1690
77 3660
(C9H8N)2[Nb(OC3H7)C15] 296 1210 20 1.74
196 1790
77 3620
2
[NbCl(0C2H5)3(C5H5N)]2 96 13
299 -1150
Nb(OC2H5)4 299 -100(in C6H6)

 

The magnetic moments are estimated to have a pre-
cision of.i 3%, based on a duplicate determination of

(C5H6N)2[Nb(OCH3)C15] .

32
Attempts to Obtain X-Ray Diffraction Data

Numerous attempts were made to obtain X-ray powder

photographs of (C5H6N)2[Nb(OCH3)C15] and [NbCl(OC2H5)3(C5H5N]2

using the ground powders in sealed 0.3 mm. glass capillary
tubes, and a Debye—Scherrer camera with Cr Ka radiation.

In both cases, the samples were completely white after the
necessary long exposure time. The films did contain some
lines which were almost unreadable. It is not known

whether the decomposition is a result of the X-ray irradiation
or slow oxidation by the diffusion of oxygen through the

thin, glass capillary.

DISCUSSION OF RESULTS

The Pentachloroalkoxoniobates

During the course of this research, twelve crystalline
niobium(IV) compounds, all having the general stoichiometry
of (BH)2[Nb(OR)C15},were prepared and characterized. One

preVious example of this class, (C5H6N)2[Zr(OC2H5)C15], is

known.34 The cationic portion of the niobium compounds always
consisted of a large organic cation (alkylammonium, pyridinium,
quinolinium, etc.), while the alkoxo group was either methoxo,
ethoxo, or isopropoxo. The color of the resulting compound
is dependent on the nature of the cation.

Conclusive evidence that the compounds contain
pyridinium,and quinolinium ions and not ligand pyridine
and quinoline is readily had by comparing the infrared
spectra of the compounds with those of (C5H6N)Cl,, (C H N)

5 6 2

35 36 37

CoCl (C9H8N)Cl, (C9H8N)2CoCl4, and (C9H7N)2CoC12

2!
(Tables 2 and 3). Further support is to be had from the
analogous Nemethylpyridinium and N-methquuinolinium

compounds, which cannot be coordinated to the metal ion

through nitrogen.
33

34

Table 2. A comparison of the infrared spectra of pyridinium
ion and ligand pyridine with that of
(C5H6N)2[Nb(OCH3)C15].

 

 

 

C5H6N ligand CSHSN (C5H6N)2[Nb(OCH3)C15]
1630 1600 s 1630 s
1600. 1570 w 1600 s
1530 1490 s 1525 S
1480 1440 s 1475 s
1325 -—-- 1335 m
1240 1240 vw 1240 m
1200 1215 s 1195 m
1160 1150 m 1160 m

 

3, strong; m, medium;

weak: vw,

very weak.

35

Table 3. A comparison of the infrared spectra of quinolinium
ion and ligand quinoline with that of
(C9H8N)2[Nb(OCH3)C15].

 

 

 

C9H8N ligand C9H7N (C9H8N)2[Nb(OCH3)C15]
1630 1590 1630
1590 1580 1590
1550 1540 1550
1400 1400 1410
1300 1320 1300
1220 1310 1230
1160 1140 1160
1140 1130 1140
1080 1080 1070

 

m, medium;

W:

weak; vw,

very weak.

36

However, the infrared spectra (Figure 2) of the
quinolinium compounds does show some irregularities in the

ring C-H stretching frequency. In both (C9H8N)C1 and

(C9H8N)2CoCl4 this vibration occurs at 3400 cm.-1 (Figure 3),
and is seemingly Sharp and without fine structure. In the
niobium compounds this absorption has shifted to3000cm."l
and exhibits considerable fine structure. This indicates
a weakening of the ring C-H bond. The possible cause will
be outlined after a consideration of other aspects of the
spectra of these compounds.

Alcohols uniformly have intense absorptions in
the region of 1100 cm.-1 which have been assigned to
the C-0 stretching vibration. Such an absorption is
present in all the pentachloroalkoxoniobates; and since
there is no indication of the O-H vibrational mode, the
alkoxide ion must be directly coordinated to the metal
ion. The absorptions associated with the C-0 stretching
vibration are not constant for a given alkoxide, but seem
to depend, at least to a small extent, on the nature of
the cation. Exact frequencies from the spectra presented
in Figure 4 through 10 are: [(CH3)4N]2[Nb(OCH3)C15],

-1 ' -1
1130 cm. , (C5H6N)2[Nb(OCH3)C15], 1095 cm. ,

37

H

.Aaasfi mammonmxouosamuwmv

..80 ooom-oo oomm Scum .maoiemmoocozLNIZSSoos

cam HmaoammmoovnzamAmemov mo mnuommm pmHMHMSH .N musmflm

mumnfiss m>m3 CH wosmsqmum
000m 000m oovm

d u d

 

_maoiemmoocoz_«Azmmmoc----

_maoxmmmooooz_mxzmmoov , om.

uorssrmsuexl quaoxad

 

 

 

 

38

CONN Eoum

w

.AHHSE mammoumxonosamummv

H

I.So ooom 00

Hooomxzmmmoc oom Hoxzmmmoc no 0000000 ooumumcH

mquEss m>m3 SH mucosvmum

 

ooow ooom ooem
oaooUNAmemoSIIII
HUAmeooS I

 

 

 

 

.m ousmflm

39

ma

.AHHsS Honozv

_

m

HoAmmoovnzHNHvammova mo Enhuommm pmumumsH

msouofiz CH sumcmwm>m3

m

m

b

v madmam

 

 

 

 

 

 

d

 

 

 

 

 

d

j

 

 

4

 

 

 

 

40

.AHHSE Honpzv HmaoammoovnzHNAZwmmov mo Esuuummm pmumumsH
msouoflz_sa sumcwam>m3

Ha OH 0 m h w m

.m ousmflm

 

 

- q q 1 u

H
d

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

41

.Aaasfi Hohpzv HmaofimmoovnzHNAmemov mo Esnuummm pmumuwcH .o musmflm
mGOHUHS CH numcwam>mz

ma NH HH OH m m b m m g

g a, q a E q d I

 

 

 

 

 

 

 

 

 

 

 

 

42

.AHHSE Honazv

NH Ha

msouoflz :H sawsmaw>mz

3 m m S o m

.h mnsmam

 

 

 

 

 

 

q . 4 - q . .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

43

ma

.AHHSS Honozv ”maoxmmmoovnzLNAZSSmov no 55000000 omnmumoH

msouoflz SH mnumcmam>m3

NH Ha OH m m h o m

.m musmam

 

 

 

 

 

 

a .I. a q a a . . 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

q

 

 

 

 

 

44

NH

.AHHSS Homozc

HH

OH

0

HmHofimmmoonzamAme UV mo Esuuuwmm pwumnwcH

msouoflz CH numsmHm>mz

m m h o m

.m musmHm

 

 

 

 

 

 

 

q

q q u 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

45

MH

.Aaaoe Honozv

NH HH

HmHUAnmmooVQZHNAZOmOUV mo Esuuommm UmumumsH
mcouon SH sumstm>m3

OH O m b O

1L0

.OH musmHm

 

 

 

 

- - 1 -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

46

(C9H8N)2[Nb(OCH3)C15], 1110 cm.’l; [(CH3)2NH2]2[Nb(OC2H5)C15],

1070 and 1100 cm.'1; (c5

-1
(C9H8N)2[Nb(OC2H5)C15], 1070 and 1090 cm. , and

-1 .
(C9H8N)2[Nb(OC3H7)C15], 1030 and 1100 cm. . A weakening

C-O bond has occurred in the presence of pyridinium and
quinolinium ions.

The ethoxo and isopropoxo compounds exhibit two
absorptions 1000 and 1150 cm.-l. Bradley39 has noted a

similar occurrence in the Spectra of various ethoxides

) )

([Nb(OC2H5 512,[Ti(OC2H5 4]3, etc.) and has concluded the

cause to be the presence of both bridging and terminal

alkoxide.

C)

 

 

 

EDKJ 3)
0 represents the oxygen in ethoxide

The pentachloroalkoxoniobates can be pictured as
monomeric and hexa-coordinated ions. But, if Bradley's

conclusion is true, then polymeric structures containing both

—1
H6N)2[Nb(OC2H5)C15], 1070 and 1090 cm.

I

47

bridged and terminal alkoxide must be imagined. This is a
difficult task. Fortunately, such a process does not seem
necessary when the spectra of ethyl alcohol and sodium
ethoxide are considered. It seems likely that in either

of these compounds all C-O bonds are equivalent: Yet, both
have two absorptions in the 1000-1150 cm.—1 region.

Ethyl alcohol absorbs at 1055 and 1097 cm.-1, while sodium
ethoxide absorbs at 1062 and 1128 cm.-l. Further assurance
is had from the simplicity of the spectra of the pentachloro-
methoxoniobates which exhibit Single maxima between 1090

and 1130 cm.-l. Methyl alcohol and sodium methoxide behave
similarly with absorptions at 1035 and 1074 cm.-1 respectively.
Thus, it seems reasonable to picture the complexes as
monomeric, hexa—coordinated ions.

Some credence is lent to this picture from a
consideration of the magnetic moments and the Curie-Weiss
dependence of the susceptibilities of the complexes (Table I).
This evidence, however, is not an absolute criterion for the
mononuclear nature of the complexes. For example, Earnshaw
and Lewis41 have shown that magnetic interaction between
two metal ions in a binuclear complex is sensitive to the
nature of the bridging group and to the metal-ligand-metal

bond angle. For [(NH3)5Cr-OH—Cr(NH3)5]Br the magnetic

SI

48

moment is very close to that for the free metal ion and

the slight curvature in the plot of l/Xh against T is

only detectible through precise measurements of the suscepti-
bility at a considerable number of temperatures. For the

the

seemingly similar compound, [(NH SCr-O-Cr(NH3)5]Br

3) 4'
moment deviates sharply from the free—ion value at low
temperatures, and a pronounced antiferromagnetic interaction
can be observed in the plot of l/Xh against T. This has
been attributed to an increase in the metal—ligand-metal
bond angle and a concurrent increase in v—bonding between
the oxygen and the metal ions.

The evidence offered to this point is in accord
with a monomeric, hexa-coordinated series of complexes.
However, structurally, they cannot be this simple because
the color of the complexes is dependent on the nature of
the cation. Certainly, in most complexes, their color is
due to d-d transitions within the metal ion, and should be
little effected by the cation.

Reflectance spectra in the visible region were taken
for a number of compounds (Figures 11 through 13). The
simple alkylammonium methoxo and ethoxo salts show a
single (very Slightly asymmetric) band at about 19,600 cm.”1

(510 mu). The maxima in the spectra of the pyridinium

49

.moquoHcoxonumEouoHnomusmm 080m mo muuommm mosmuomHmmH mHQHmH>

mGOHUHEHHHHS GH nvmsoHoRmz

.HH ousoflm

 

 

 

 

324.832 .281 TIE-
22.592 Aide ........

mo 1.183209..on

 

 

 

 

flousqxosqv 31111191398:

50

.moquoHsoxonumouoHnomusom meow mo mnuommm mocmuomHme mHQHmH>

mcopoHSHHHHa SH newsonbwz

0mm

00m

.NH

ouzmHm”

 

£1.33; See.
nihzuooxz éétIa
53.832. «2:7...
as 3603: «ongI

 

1 dII1 Ti

0.
.0

 

 

 

Koueqzosqv aAriSIag

51

.mwquOHSoxomonmomHouoHSomuSmm 0E0» mo muuowmm mUSmuumHmmH mHQHmH>

mSOSoHBHHHHS SH SpmSon>w3

omo cow 0? com

1 q u d a u d d a

4 1 q a a q d 4 q

28.18032 31.. I---
321538 oz 0:1 “.ch

 

.MH muoofim

OO:

 

 

 

fiaueqxosqv SAI'I'BISH

52

compounds are always shifted toward higher frequencies, and
are considerably more asymmetric. The spectra of the
quinolinium compounds are much more complex with a broad
absorption band extending from 15,400 to about 22,100 cm._l.
This band seemingly consists of two or three poorly resolved
components. It is interesting to note that the spectra of
the isoquinolinium pentachloromethoxoniobate(IV) does not
resemble the quinolinium compound, but consists of an
asymmetric band at 20,600 cm.-1. No high intensity bands,
such as might be associated with charge transfer, are pre-
sent in any of the spectra, except at the limiting frequencies
of these measurements for the ethoxo compounds.

The spectra of the alkylammonium compounds should
be most readily interpreted, yet even in these there is
considerable ambiguity in making spectral assignments. It
seems reasonable to assume that these complexes are tetragonal:
and that some w-bonding exists between the alkoxide oxygen
and the metal ion. With these assumptions, the energy level
diagram for the [MoOC15]-2 ion23 should be at least qualita-
2

tively useful in interpreting the spectra of the [Nb(OR)C15]‘

species: both of which possess C4v symmetry.

53

b1 (6x2 — yz)

 

e (d d
xz yz
b (d )
2 xy
The predicted transitions are b2___—_'e' b2___—_’bl (A), and
b§—————.al. In the [MoOC15]-2 ion the first two transitions

occur at 13,800 and 23000 cm._l, respectively. The latter
transition is apparently masked by a charge transfer band.

Due to limitations of the available spectrophotometer,
the only observed transition for the alkylammonium compounds
occurs at 19,600 cm.-l. This might reasonably be assigned

to either the b2—————.e or bz—H———+bl transitions: and,

indeed, arguments for either can be presented.

Since the b2-—————Sb1 transition in the [MoOC15]_2 ion

occurs at 23,000 cm.-l, a value not too different from the
observed transition, a similar assignment might be made
for the [Nb(OR)C15]-2 ion.
42 .
In another approach, Jorgensen argues that this

transition can hardly be that of bi—————ebl, because it

would have about the same energy as A in the cubic complex

‘[NbCl6]-2. This value is unknown at present; however, in

54

the complex [MoCl6]-3, A is 19,200 cm.-l. Since the
increased charge should introduce a 20% larger A, the
value for [NbC16]-2 should be about 23,000 cm.-l. Since
this is larger than that observed, the observed band must

be due to the b e transition.

 

2

The effect of the cation is not easily explained.
The increasing complexity, without a great increase in
intensity, in the visible region seems to argue that the
tetragonal symmetry is even more distorted. The decrease
in the C-0 frequencies may mean that the interaction is
weakening the C-0 bond, while the decrease in the C-H
frequency in the quinolinium cation provides a similar
point of interaction. Perhaps, then, the aromatic
protonated bases are interacting with the complex, possibly
through the alkoxo-group oxygen, and providing even more
distortion of the complex.

This explanation is open to doubt also, since the
vibrational frequencies characteristic of the pyridinium and
quinolinium ring systems (Table 2 and 3) are certainly
not greatly altered, if any, from those of the simple
hydrochlorides.

At this point it does not even seem possible to

say that exact structural determination must await X-ray

55

examination because decomposition of the samples occurs when
they are irradiated with X-rays.

The spectra of the solutions from above the
precipitated solid complexes are not similar to those of
the solids, indicating that the species in solution is not
the same as that in the solid complex. However, the spectra
from all the supernatant solutions are almost identical,with
a single maxima at 21,300 cm.-1 in the region from 14,300 to
25,000 cm.-l. Therefore, for a given alcohol, the species
in solution, whatever its nature, is independent of the
nature of the cation.

Another feature of the pentachloroalkoxoniobates
that is worthy of note is their stability towards hydrolysis.
Since fairly good yields of the pyridinium ethoxo compound
were isolated from 90% ethyl alcohol solutions, it is
obvious that careful drying of the solvent is not necessary
for preparative purposes. Ordinary laboratory absolute
alcohol is sufficient to guarantee good yields of the complexes.

The stability of the alkoxo-niobium bond toward
hydrolysis may seem anomolous, but it would seem to support
the contention that W-bonding exists between the alkoxo-
group oxygen and the metal ion. This would tend to make

oxygen's ordinarily non-bonding electrons unavailable for

56

protonation (a step which could logically precede a dissocia—
tive 8N1 release of the alkoxo group), and strengthen the
niobium-oxygen bond. Jorgensen43 has commented on the

. . . +2 +3
stability of the oxo-group in V0 and M00 toward proto-
nation, even in solutions of high acidity, and has correlated
this property with extensive w—bonding between the oxygen

and the metal ion.

Dichlorohexaethoxobis(pyridine)diniobium(IV)

This interesting dimeric compound possesses so many
non-equivalent ligands that an accurate structural determina—
tion based on chemical and physical properties alone is
impossible. Inspection shows there are forty-three possible
isomers based on the following assumptions.

1. Each niobium is octahedrally surrounded by ligands.

2. Each niobium is maintained in the tetravalent state,
i.e., a niobium(III)-niobium(V) combination is not
allowed.

3. Pyridine cannot act as a bridging ligand, but any
other combination of bridging ligands is possible.
The infrared spectrum (Figure 14) makes it clear
that ligand pyridine is present, since the strong absorption

characteristic of pyridinium ion at 1630 cm._1 is absent.

57

.AHHJE Hohszv NHAmemovmfimmNoovHonzH wo Esuuoomm UmSmumSH .SH mHSOHm

msouoflz SH numSmHm>m3

m.H NH HH OH m w e O m w

 

 

‘

I I..I -. . 4 . . _ .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

58

The region of the Spectrum, in which the characteristic
C-O absorptions occur (ca. 1100 cm.-l), is quite complex.
Exact frequencies are 1010 m, 1040 s, 1000 s, 1120 s
(shoulder),and 1140 cm.-1 m. This is certainly evidence
for non-equivalent alkoxo groups. Moreover, it could be
interpreted as evidence for bridging and terminal alkoxo
groups.

The inability to replace chloride ion by solvolysis
and precipitation as KCl, however, points to the inertness

4.
of this ligand. In the kineticallyvstable [Co(NH3)5Cl] 2

and [Pt(NH Cl]+ ions the chlorides are fairly readily

3)3
replaced. It is only in certain dimeric palladium and platinum
complexes which possess chloride bridging that kinetic inert-
ness of the chloride ion is encountered. It seems logical
then to assume that in the dimeric niobium compound the
chloride ions are bridging groups. With this assumption the
number of isomers is reduced to five (Figure 15).

If this is the case, how can the complexity of
the infrared spectrum between 1000 and 1100 cm...1 be explained?
In the structures shown in Figure 15, all of the oxygens

are not really equivalent, since some are trans.to pyridine,

while others are trans to chloride or oxygen. A non-equivalence

59

 

 

 

 

 

 

._ fQ / f1) fISID/J
N gr <qu N (Ber a)

0 represents oxygen in ethoxide.
N represents nitrogen in pyridine.

Figure 15. Isomers of [NbCl(OC2H5)3(C5H5N)]2 With

chloride bridging.

60

w-bonding between the oxygens and the metal ion, resulting
from differing trans groups, could greatly effect the C-0
absorption spectrum, and cause the observed complexity.

The observed diamagnetism does not afford any
additional evidence concerning the nature of the bridging
groups, but it does provide a stimulus for investigating
the nature of the spin-pairing mechanism. In NbI4 (see
Introduction) the diamagnetism is explained by direct
metal-metal bonding because the metal-metal distance is
only 3.3 3. Using the covalent radius for niobium of
1.3 A and assuming the bridged chloride to have a slightly
larger radius than the covalent radius of 0.99 244 (say
1.1 A), then simple trigonometry allows a calculation of 3.5
A for the niobium-niobium intramolecular distance. This
seems to agree well enough with that observed in NbI4 to
postulate that the spin-pairing occurs by direct overlap of
the adjacent metal dxy orbitals. If the niobium ions are
shifted from the centers of their octahedra toward each other,
as is observed in NbI4, then overlap would be even more
favorable.

The visible and near—infrared spectrum of the dimer
in CHCl and ethyl alcohol is particularly uninformative,

3

consisting only of a shoulder at about 27,400 cm...l adjacent

61

to a large charge transfer band in the ultraviolet region.
The molar extinction coefficient at 27,400 cm.”l using

three separate solutions was 190, 310, and 520 1. mole-1
cm.- . Any of these values are anomolously high for a

d-d transition in an octahedral complex. The lack of agree-

ment and the high values can probably be attributed to

oxidation.

Tetraethoxoniobium(IV)

Prompted by Bradley's numerous successful preparations
of tetravalent metal alkoxides (for example, Ti(IV),

8 9 Sn(IV),50 and

Zr(IV),46 Hf(IV),47 Ce(IV),4 Th(IV),4
V(IV)52) and by the successful preparation of several chloro—
alkoxo niobium(IV) species during this research, the prepar—
ation of tetraethoxoniobium(IV) was attempted.

Bradley has used three general methods for the
preparation of metal alkoxides:

1. The Ammonia Method:

MC14 + 4ROH + 4NH3————, M(OR)4 + 4NH4C1I

This method was applied with success to Ti(IV),
Zr(IV), and Hf(IV). For Ce(IV), which does not form a stable
tetrachloride, the compound (C H N) CeCl was used instead.

5 6 2 6

(C5H6N)2CeC16+ 6NH3 + 4ROH——>Ce(OR)4 + 6NH4C1I + 2 C5H5N

62

2. The Sodium Alkoxide Method:

 

MCl + 4 NaOR

4 M(0R)4 + 4 NaCll

With SnCl4, the ammonia method did not result in
complete precipitation of the chloride as ammonium chloride.
Moreover, attempts to prepare thorium alkoxides by a similar
method were also unsuccessful. >However, the use of sodium
alkoxide resulted in the complete removal of chloride.

3. The Solvolytic Method:
With VC14, neither of the above methods were

successful, but solvolysis of tetrakis(dialkylamido)vanadium(IV)

with alcohol produced the desired alkoxide.

 

M(NR2)4 + 4 ROH MgOR)4 + 4 R NH

2

Thomas, however, was not able to apply this method
successfully to niobium(IV) because of oxidation to
Nb(OR)5.51 He has attributed the lack of success to the
instability of the tetralkoxide in the presence of alcohol,
presumably:

2 Nb(OR)4 + 2 ROH -—————e 2 Nb(OR)5 + H2

In this research, the ammonia method was not
successful, but the sodium alkoxide method, when applied to
[NbC1(OC2H5)3(C5H5N)]2, produced the desired Nb(OC2H5)4

with no apparent oxidation.

63

 

+
[NbC1(OC2H5)3(C5H5N)]2 2 NaOCZH5

I + +
2 Nb(OC2H5)4 2 NaCl 2 CSHSN

The niobium alkoxide is readily purified by
sublimation. Volatility is quite common among the alkoxides
of metals of low atomic number as the following table
illustrates:

Table IV. Boiling or sublimation temperatures of metal
ethoxides (in OC/mm. Hg.).

 

 

 

. 46

Ti(OCZHS)4 103/O.1

* 46
Zr(OCZHS)4 180 /0.1

_ 3+. 47
Hf(OC2H5)4 180 200 /O.l

_ . 48
C€(OC2H5)4 non volatile

_ ‘ 49
Th(OC2H5)4 non volatile
*Sublimed.

Another feature of the metal alkoxides is their
tendency to polymerize. Molecular weight studies on
the niobium compound were not attempted, however, because
of the limited amounts of pure material available and its
extreme sensitivity to oxidation and hydrolysis. The dia-

magnetism of an impure sample suggests that the niobium

64

compound is also polymeric.

The infrared spectrum (Figure 16) in the C-0
vibrational region is only slightly more complex than the
pentachloroethoxoniobates: 1030 m (shoulder), 1040 s
(Shoulder), 1060 s, 1100 s, and 1140 cm."1 s (shoulder).
Since the evidence points to a polymer, then the molecule
must consist of both bridging and terminal ethoxides.

Yet, the spectrum is relatively simple. This simplicity
might support the postulate that the complexity of the
spectrum of [NbCl(OC2H5)3(C5H5N)]2 is due to a trans
effect, rather than a difference due to bridging and
terminal ethoxides.

The visible and near—infrared spectrum of
Nb(OC2H5)4 in ethyl alcohol is almost identical to that of
[NbCl(OC2H5)_3(C5H5N)]2 with only a shoulder at about
27,800 cm.’1 (6 = 20 1. mole"1 cm.—l).

Both Nb(OC2H5)4 and [NbC1(OC2H5)3(C5H5N)]2 are
extremely labile to chloride substitution and depolymerization
in solutions of strong acidity. Lability is to be expected
for d1 metal ions,52 but the ease of depolymerization is

somewhat surprising. This feature has enabled an interesting

cycle of reactions to be completed:

65

 

NH

N H

HH

.,HHSE HOHSZO enmmmoovflz mo Esuuommm poumuwSH

mSOSUHE SH LDOS0H0>03

OH O m b O m

 

 

 

 

 

 

 

.oe meooes

 

 

 

 

66

NaOC H

 

[NbC1(OC (C H N)] :Nb(OC H)

2H5)3 55 2 254

HCl I
C5H5N HCl

(C5H6N)2[Nb(OC2H5) C15]

Since the pentachloroethoxoniobate can be prepared
from either of the two species, they must be niobium(IV)
derivatives and not an equimolar combination of niobium(III)

and niobium(V).

67

 

 

 

 

 

 

.Hohsz mo Esuuommm OmumHOSH .NH musmHm
mSOHUHE SH SumSmHm>m3
MH NH HH JOH o O b O m d
T q 1 u H - . H . -

 

 

 

 

 

10.

11.

12.

13.

H.

F.

C.

H.

V.

REFERENCES

E. Roscoe, Chem. News, 31, 26 (1878).

Ott, Z. Electrochem., 18, 349 (1912).

H. Brubaker, Jr., Ph.D. Thesis, Massachusetts
Institute of Technology, 1952.

Schaefer and K. D. Dohman, Z. anorg. allgem.Chem.,
311, 134 (1961).

Corbett and P. Seabaugh, J. Inorg. Nucl. Chem.,
.6, 207 (1958).

F. Dahl and D. L. wampler, Acta Cryst., 12, 903
(1962).

F. Dahl and D. L. Wampler, J. Am. Chem. Soc , 81,
3150 (1959).

G. Korshunov and V. V. Safanov, Zh. Neorgan. Khim.,
6, 753 (1961).

E. McCarley, B. G. Hughes, J. C. Boatman, and

B. A. Torp, Presented to the Division of Inorganic
Chemistry, American Chemical Society National Meeting
at Washington, D. C., in March, 1962.

C. Bradley and I. M. Thomas, Can. J. Chem., 49,
449 (1962).

O. Fischer and A. Treiber, Ber., 94, 2193 (1961).

C. Brantley, U. S. 2,921,948, Jan. 19, 1960;
Chem. Abstr., 54, 11048 (1960).

D..Gozzi and S. Vivarelli, Z. anorg. allgem. Chem.,
279, 165 (1955).

68

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

69

D. J. Ferrett and G. W. C. Milner, J. Chem. Soc., 1186
(1956).

J. Kleinberg, W. J. Argersinger, Jr., and E. Griswold,
"Inorganic Chemistry," D. C. Heath and Co., Boston,
Mass., 1960, p. 498.

N. V. Sidgwick, "The Chemical Elements and Their
Compounds," Oxford University Press, London,
England, 1952, p. 651.

B. N. Figgis and J. Lewis, "Modern Coordination Chemistry,‘
J. Lewis and R. G. Wilkins, ed., Interscience Publishers,
InCo] New York] NoYo’ 1960’ p0 4060 ‘

See for example, C. J. Ballhausen and H. B. Gray,
Inorg. Chem., 1, 111 (1962).

J. Cavell and R. J. H. Clark, J. Chem. Soc., 2692
(1962).

Calculated from the susceptibilities given in ref. 24.
See for example, ref. 17, p. 444.

H. Funk, F. Schmeil, and H. Scholtz, Z. anorg. allgem.
Chem., 310, 88 (1961).

H. B. Gray and C. R. Hare, Inorg. Chem., 1, 363 (1962).

C. R. Hare, I. Bernal, and H. B. Gray, Inorg. Chem.,
.1, 831 (1962).

G. Andersson, Acta Chem. Scand., 10, 623 (1956).

I. Heilbron, "Dictionary of Organic Chemistry,"
Oxford University Press, New York, N.Y., 1953, p. 275.

Ibid., p. 303.

E. W. Golibersuch and R. C. Young, J. Am. Chem. Soc.,
11, 2404 (1949). '

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

30.
31.
32.

33.
34.

35.

36.

37.

38.
39.

40.
41.

42.
43.

44.
45.

46.

70

Ref.l7, p. 412.
Ibid., p. 415.
Ibido I p0 403.

R. H. Holm and F. A. Cotton, J. Chem. Phys., 31,
790 (1959).

D. C. Bradley, F. M. Abdwel Halim, E. A. Sadek, and
W. Wardlaw, J. Chem. Soc., 2032 (1952).

N. S. Gill, R. H. Nuttall, D. E. Scaife, and D. W. A.
Sharp, J. Inorg. Nucl. Chem., 18, 79 (1961), and
references therein.

E. G. V. Percival and W. Wardlaw, J. Chem. Soc., 1505
(1929).

Prepared analogously to (CSHSN)2CoCl2 in ref. 35.

L. J. Bellamy, "The Infrared Spectra of Complex Molecules,"
Methuen and Co., London, England, p. 95.

C. G. Barraclough, D. C. Bradley, J. Lewis, and I. M.
Thomas, J. Chem. Soc., 2601 (1961).

F. H. Seubold, J. Org. Chem., 21, 156 (1956).

A. Earnshaw and J. Lewis, J. Chem. Soc., 396 (1961).

C. K. Jorgensen, private communication to C. H.
Brubaker, Jr.

C. K. Jorgensen, Acta Chem. Scand., 11, 73 (1957).

T. Moeller, "Inorganic Chemistry," J. Wiley and Sons,
New York, N.Y., 1958, p. 135.

D. C. Bradley, F. M. Abd-el Halim, and W. Wardlaw,
J. Chem. Soc., 3453 (1950).

D. C. Bradley, R. C. Mehrotra, J. D. Swanwick, and
W. Wardlaw, ibid., 2025 (1953).

47.

48.

49.

50.

51.

52.

D.

D.

71

C. Bradley, R. C. Mehrotra, and W. Wardlaw,
ibid., 1634 (1953).

C. Bradley, A. K. Chatterjee, and W. Wardlaw,
ibid., 2260 (1956).

C Bradley, M. A. Saad, and W. Wardlaw, ibid.,
1091 (1954).

C. Bradley, E. V. Caldwell, and W. Wardlaw, ibid.,
4775 (1957).

M. Thomas, Can. J. Chem., 39, 1386 (1961).
Basolo and R. G. Pearson, "Mechanisms of Inorganic

Reactions," J. Wiley and Sons, New York, N.Y.,
1958, p. 110.

CHEMISTRY LIBRARY

 

 

WIIIIIIIIIIII

III

N 3
“0
W13
“-9
m2

1
3

I
|

 

IIIIIIII

I