PHOTOLYSIS 0F CRYSTALUNE
TRISOXALATOCOBALTATE (III)
IN A DILUENT CRYSTAL

Thesis for the Degree of M. S.
MICHIGAN STATE UNIVERSITY
ANNE FENERTY
1969

-..;; I: .Ikt...:. . _ . . ’3

«has»:

 

LIBRARY ‘

 

Michigan State i
University ‘
i
r“ ‘ W
1:: "amaze

 
 
    
   

  

‘ 800K BINDERY INC.

LIBRARY BINDE Rs
SPRINGPOUT. IICII‘II
t A A a... '

ABSTRACT

PHOTOLYSIS OF CRYSTALLINE TRISOXALATOCOBALTATE(III)
IN A DILUENT CRYSTAL '

BY

Anne Fenerty

Trisoxalatocobaltate(III) was incorporated into the
NaMg[Al(C204)3]-9H20 lattice by isomorphous1 replacement of
Al(C204)g— to permit spectrophotometric analysis of the
dark green compound.2

Trisoxalatocobaltate(III) undergoes the following
photolytic reactions:

1. Intramolecular oxidation and reduction.

2. Substitution.

3. Racemization.

It was generally believed3 that substitution takes
place when ligand field bands are irradiated and redox re-
actions take place in the charge transfer region. Some evi—
dence of redox reaction at higher wave lengths has been
found, however, and it was suggested4 that the tail end of
the charge transfer band extends into the ligand field region,

To determine the type and mechanism of the reaction at
differing wavelengths we set out to measure quantum yields
on irradiation, by following the decrease in absorbance of

the ligand field bands. This proved to be impossible as

Anne Fenerty
the decomposition products of the reaction absorb at these
wavelengths.

The crystal study method has been adopted for the fol-

lowing reasons:

(1) Dilution.

(2) Identification of the reaction products trapped
in the crystal which slows down radical recombina—
tion.

(3) Observation of polarizations due to the D3 sym—
metry of the crystal, and of the lower symmetries
of the products.

(4) Elimination of solvent effects.

The following results were obtained:

(1) We found the reaction to be identical in both of
the ligand field bands.

(2) On irradiation a band of high molar extinction

develops at 320-350 mg. It is polarized, and at-

 

tributed to5 the 4a1 > 2b1 transition in C027}
which has C2v symmetry.

(3) Co(C204)34— has been identified as the main re-
duction product, through the polarization of the

ligand field transitions, which show D3 symmetry.

The suggested mechanism is:

Anne Fenerty

O O
/
C// hv I \C/

 

 

 

4— 7- 1.

0x2 Co(H20 )2 + €02 + co2 ————> co2 +

0x2 Co II I > OX3CoII + C02.
0/C §O

(4) An approximate first order rate constant was
calculated. The half life on irradiation is be-
tween 2.3—7.4 minutes. This is using the lowest
_optical density value as Da)’ which is quite
possibly incorrect as any further decrease is ob-

scured by the very intense band growing in at

320 mu.

PHOTOLYSIS OF CRYSTALLINE TRISOXALATOCOBALTATE(III)

IN A DILUENT CRYSTAL

By
A e“ 5
Anne Fenerty
A THESIS

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

MASTER OF SCIENCE

Department of Chemistry

1969

(3.5é/73
L’Zéifi?

ACKNOWLEDGMENT

I wish to thank Dr. S. T. Spees for his guidance,
encouragement and constant availability for discussions
and suggestions involving the problems in this project.
In addition I acknowledge the help Dr. Carl Brubaker gave
me in understanding the theoretical aspects underlying
this work. Finally, I thank my husband, Michael for his

patience and support.

ii

TABLE OF CONTENTS

Page
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 1

HISTORICAL . . . . . . . . . . . . . . . . . . . . . 4
EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . 7
Preparation of Compounds . . . . . . . . . . . 7
Actinometry . . . . . . . . . . . . . . . . . . 11
Preparation of NaMg[Co(C204)3] Crystals . . . . 13
Irradiation . . . . . . . . . . . . . . . . . . 14

RESULTS AND DISCUSSION . . . . . . . . . . . . . . . 21

Investigation of the Near U.V. Band . . . . . . 24
Changes in the Two Visible Region Ligand

Field Bands on Irradiation . . . . . . . . . 29
Results Obtained by Irradiation at Different

Wavelengths . . . . . . . . . . . . . . . . . 34
Polarization of the Ligand Field Band Spectra

of the Decomposition Products . . . . . . . . 36

SUMMARY . . . . . . . . . . . . . . . . . . . . . . 58
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 60

APPENDIX . . . . . . . . . . . . . . . . . . . . . . 62

iii

TABLE

II.

III.

IV.

LIST OF TABLES

Page

Energy distribution of a 2500w Xe—Hg arc lamp 15

Monochromator filters . . . . . . . . . . . . 17
Rate constants for the photodecomposition . . 49
Polarized spectra of the irradiated crystals. 51

iv

FIGURE

10.

11.

12.
13.

14.

15.
16.

17.

Visible
aqueous

Visible
aqueous

Visible

spectrum
solution

solution
solution

spectrum

LIST OF FIGURES

Of CO(C204) 2(H20)2

spectrum of Co(C204)3

of an irradiated

K3[CO(C204)3] mull.. o . . . . .

Development of the 320 mu band .

Development of the 320 mu band

Polarization of the 320 uv peak in an ortho—

axial dilute crystal . . .

Thermal reaction

Thermal

Irradiation with white light.

610 band . . . . . .
Spectrum for Figure 9 . . . . . .
Irradiation with 360- 390 mu light. Axial
section, 610 band. . . . . . . . . .
Log (OD - ODoo) y§_time graph of Figure 11..
Spectrum for Figure 11 . . . . . .

reaction

2-

Irradiation #2 with 360-390 mu light.

section, 610 mu band . . . . . . . . .
Log (OD - ODGD) y§_time graph for Figure 14.
Spectrum for Figure 14 . . . . . .

Irradiation of 417 mu, monitored at 610 mu and

427 mu.

Axial s

ection .

in

3..

Axial

Axial section,

Page

10

22

23
25

26

28
32
33

37
38

39
4O

41

42
43

44

45

LIST OF FIGURES (cont.)

FIGURE

18.

19.
20.
21.
22.
23.
24.

25.
26.
27.
28.
.29.

30.

Ultra violet spectrum after 1 hr irradia-

tion with 417 mu light . . . . . . .
Entire spectrum of Figure 17 . . . .
Irradiation at 572 mu . . . . . . .

Axial section for Figure 20, 610 mu

Orthoaxial section for Figure 20, v Spectrum

Orthoaxial section for Figure 20,

Irradiation of an axial section with
light . . . . . . . . . . . . . . .

Irradiation of an axial section with
A > 578 mu light . . . . . . . . . .

Irradiation of an orthoaxial section
A > 450 mu light . . . . . . . . . .

Irradiation of an orthoaxial section

A = 427 mu, 7 spectrum . . . . . . .
Irradiation of an orthoaxial section
A = 427 mu, 0 Spectrum . . . . . . .
Irradiation of an orthoaxial section
A = 360-390 mu, 0 spectrum . . . .

Irradiation of an orthoaxial section
A = 360-390 mu, F spectrum . . . . .

vi

0 spectrum

553 mu

Page

46
47
48
62
63

64

65

66

67

68

69

7O

71

INTRODUCT ION

Trisoxalatocobaltate(III) is a hard acid—hard base
complex. With 24Dq crystal field stabilization energy
and three bidentate chelate rings, the complex is kinetical-
ly inert. It has been the subject of much work including
both thermal, and conventional photochemical activations,
flash photolysis7, and of e.s.r. studies, on the radicals
produced on irradiating a frozen aqueous solutions.

Work involving a diluent crystal has been originated
by Piper and Carlin2 to study the visible polarized spectra
of trisoxalatometallates.

Trisoxalatocobaltate(III) has pseudo octahedral sym-
metry if only the oxygen atoms directly coordinated are
considered. The symmetry is lowered to D3 due to the
chelating rings. As a result of this the Laporte forbidden
d-d transitions become allowed electronically. Carlin
and Piper measured the splitting of the octahedral excited

states and determined the trigonal field strength parameter.

The low spin d6 CoIII ion has tgg configuration.
Ground state is 1A19. In the visible region the two spin
allowed excited states tzgeg in an octahedral
field are: 1T1g and 1ng. These are split as follows:

1

2

Free ion

 

 

 

 

 

 

 

configuration —Qh— (strong fleld) D3
1
E
1 b
T
29, fi//
1
\ A1
1
1 E
T a
lg /
1
5D ‘\\ A2
1 1
19 A1

 

 

Due to the low symmetry; transitions are only allowed
for certain orientations of the electric vector of the
incident light. The crystal is definitely dichroic and
looks green when viewed in light parallel to the C3 axis.
It is blue in light incident normal to the axis.

By taking the triple direct product for each transi—
tion in each of the orientations and finding out if it con-
tains the totally symmetric, irreducible representation the

polarization of the spectrum can be calculated. It is as

 

 

 

follows;
1 -J- "'1Ea
T
<1) 1. -------..-/
1
\ H 1A2
1T 0-1— 1Eb
2
(2) 1A1 """"" >

1A (not allowed
1b I I
any d1rection)

 

3

The 1A1 —-> 1T1 transition is found at 608 mu in the
solution and is split into 610 mu for 1A1 ——¢ 1A2 and
602 mu for 1A1---—> 1Ea. The 1A1 -—$ 1T2 transition is
at 427 mu.

The method used by Piper and Carlin was to cut sections
from the crystal both perpendicular to and parallel to the
C3 axis. The former they called axial and the latter ortho-
axial sections. When the spectrum of an orthoaxial section
was examined in polarized light the selection rules were
confirmed.

The spectrum obtained with light polarized parallel
and perpendicular to the C3 axis was called the w and o
spectrum, respectively. The 0 Spectrum was identical to
the axial spectrum.

We adapted this method to study the effects of irradi-

ation on a single crystal.

HISTORICAL

Photolysis of trisoxalatocobaltate(III) has centered
on the compound in solution. Only a limited number of crys-
tal studies were made.9r10

Due to the fact that it belongs to the D3 group, the
compound is dissymmetric. Both its photoracemization11
and partial photoresolutionlz, using circularly polarized
light, were studied.

Due to the thermal instability of Co(C204)3_, isotope
studies of ligand exchange reactions are difficult to per-
form. The C2042— exchange was followed, however, by pre—
cipitation and radioassay of the free oxalate as CaC204°H2013.
It was found that oxalate exchange has a half life of
> 130 hr. at 50°, compared to a thermal decomposition half
life of 6.5 hrs at 50°. This proves that the thermal de-
composition proceeds by a mechanism not involving reversible
equilibration of C2040 or C204-.

We believe that the mechanism of thermal decomposition
is one of intramolecular oxidation-reduction, identical to
the photochemical reaction. The work done on thermal re—
actions in the solid during this study, is not enough,how—

ever, to draw definite conclusions.

4

5
. . . '3- ,
The most characteristic reaction of Co(C204)3 is the
. . 2-. . ..

redox reaction. The ligand C204 is ea31ly ox1dizable,
and the process involves homolytic bond breaking by transfer
of a o bonding electron primarily located on the ligand,
into an empty eg orbital, located on the metal.

Trisoxalatocobaltate(III) has a charge transfer band

in the near uv at 242 mu with a molar absorbance of

 

20,0002. Irradiation in this region produces reduction of
III II . . .

Co > Co With the primary quantum yield half the

reduced CoIII.

 

In an aqueous solution Spectrum the 1A1 > 1T1

 

transition is at 608 mu and the 1A1 > 1T2 at 427 mu.

There is no Splitting of the excited states in solution

 

and the molar extinction is 152 for the 1A1 > 1T1,

 

and 200 for the 1A1 > 1T2 band14.

While redox reactions are usual in the charge
transfer region their existence in the long wavelength
ligand field bands has been questioned4.

Several theories have been advanced to explain the
phenomenon:

(1) Promotion of a d electron from a tzg orbital
to an e orbital creates electron-ligand repulsion, fol-
lowed by possible bond breakageG.

(2) Due to the wide band widths the pure electronic

energy differes from the band maxima. If the charge trans-

fer band would have the lower electronic energy it could

6
extend below the ligand field transition band. In this case
the oxidation-reduction reaction could be caused by energy
absorbed in the ligand field band, through radiationless

conversion to the charge transfer band.

EXPERIMENTAL

Preparation of Comppunds

K3[Co(C204)3]. Prepared according to Palmer15 from oxalic

 

acid, potassium oxalate and cobalt(II) acetate. The latter
was substituted for cobalt(II) carbonate. The complex
cobalt(II) oxalate was oxidized by Pboz in the dark, to
prevent photochemical reduction. After recrystallization
the salt was dried over CaClz for a week.

While the usual formula is K3[Co(C204)3]-3 1/2 H20,
C2042- analysis and measurement of e at 427 and 608
yielded the anhydrous formula. The value of €427 deter—

mined from an average of three samples yielded:

202 155
6427 = :33 - 200 and €608 = :23 = 152.

This agrees well with that of KaripideslG: €419 = 208,
Furthermore, there was no loss in weight on drying the

salt over P205 for a week.

C2042_ Analysis: Folhmdng Palmer's15 method for the di—u-

 

hydroxotetraoxalatodicobaltate(III) ion, the salt was de-

III

composed and Co was precipitated as cobalt(III) oxide, by

8
using excess sodium hydroxide. After filtration the decantate
was acidified, heated to boiling, and titrated with a 0.10mg

KMnO4 solution.

C204 = in K3[Co(C204)3] Calculated: 60.0%
Found: (1) 59.0
(2) 59.2.

From these results we conclude that water is absent.

Coz+.AnalySis. The amount of Co2+ in K3[Co(C204)] was
determined by simultaneous spectrophotometric analysis.
Analytical wavelengths are: 602, 625 mu.

To an aqueous solution of the salt, acetone and a thio-
cyanate solution was added in 4:5:1 proportion17.

The blue Co(SCN):- has an absorption band at 625 mu
for which we determined 6625 = 2180 i 5% and 6602 = 1725 i
5%.

For Co(C204)33— 6602 = 152 and €625 : 130.

The amount of Co2 in K3[Co(C204)3] is:

(1) 0.95%

=1

(2) 1.27%

Calculations were doublechecked by the known concentrations

3-
of Co(C204)3 .

NaMg[Al(CaO§)a]'9H20. Prepared according to Palmer15 and

 

checked against the known molar extinction: €258 = 141.

9

Preparation of Cobalt(II)ngalate. As complex oxalates of
Co2+ are products of Co(C204)33_ photolysis, an attempt
was made to determine their spectrum and composition.

Job's method of continuous variation was used in an
attempt to distinguish between Co(C204)2(H20)22- and
Co(C204)34_ . By using a 3:1 ligand—metal ratio a purple
complex salt was prepared along with the insoluble pink

simple salt. A 4:1 ratio gave a similar but more concen-

trated spectrum. See Figure 1.

2+ 2- ___. ___.
Co(H20)6 + c204 <——> COC204 <-—

 

K1 _
K2 ,C2042
K
4.. . 3 2—
C0(C204)3 .¢___> C0(C204)2(H20)2

The concentration of Co(C204)34— is expected to be low as
K1 = 3.72, and 3&2 = 2.31, as determined by ion exchange.18

The visible Spectrum of the Co(H20)62+ ion has been
determined by us and the results agree well with those of
Roberts and Field29. The visible absorption maximum was
found at 510 mu, while their value was 512 mu. Our value
of 6510 £55, agrees with their e512 = 5.

A 3:1 oxalate-ligand ratio results in the Shift of the
absorption maximum to 535 mu, with a shoulder at 510 mu and
at 490 mu. See Figure 1. This agrees with their value of
535 mu.

No uv peak was found, till 185 mu. A wide, weak ir

peak was found at 1215 mu. Cotton and Wilkinson19 Show

10 535mu

A_ 3:1 metal-ligand
ratio

§_ 4:1 metal-ligand

ratio
490 B
535 mu
505
A

490

Figure 1. Visible spectrum of
Co(C204)2(H20)2 in aqueous solution.

11

this peak at 1210 mu for Co(H20)62+

The 3:1 ligand—metal Spectrum is attributed to
C0(¢204)2(H20)22-
It is expected that the cis-Co(C204)2(H20)22- isomer is
present, as it is the likely product, formed on removal of
an oxalate ligand. Rapid isomerization is, however, expected
as Co2+ complexes are labile.
Co(C204)34_ has D3 symmetry. This serves as an im—

portant distinction between these two compounds in the solid

phase, as shown by the polarized spectrum.

Actinometry

 

The original purpose of this investigation was to measure
quantum yields for the photodecomposition of trisoxalato-
cobaltate(III), at dflfierent wavelengths.

Due to the relative inaccuracy of radiometers, an at—
tempt was made to utilize a chemical actinometer,
K[Cr(NH3)2(SCN)4] (Reinecke salt), according to Adamson's20

2

method. A 3 x 10- M_solution of the salt was prepared.

The reaction involves aquation of one ligand.

Cr(NH3)2(SCN)4' + H20 —hls Cr(NH3)2(SCN)3(H20) + SCN-

The quantum yields for the photoaquation at different wave-
lengths are known, and from the measured concentration and

volume of the solution, the light intensity can be calculated:

(lg-CPI
dt "T
Ia = intensity of absorbed light
c = concentration of the solution
V = volume of the solution
¢ = quantum yield.
Ideally I = Io which required the concentration of

a

solution to be high enough to absorb all the light. Ac-
cording to Adamson, reflected light is negligible; we found,
however, that it was 9%, from a glass cell, at normal inci-
dence, and corrected for it.

The solution was photolyzed using a neon laser, A =
6328 2, and a 30 ml volume, 10 cm long glass cell. After
adding the solution the cell was stoppered and stirred
magnetically.

The concentration of the photoreleased thiocyanate was

determined as the iron thiocyanate complex:

6 = 4.3 x 103, at 450 mu.
max

In order to prevent aquation of more than one ligand
the percent photolysis must be kept low: 0.61% at 600 mu,
and generally between 0.6—10%.

The intensity of the neon laser was about the same
as that of our Hg-Xe lamp, 14g; 10-6 einsteins/sec. This
is several magnitudes higher than Adamson's (10-9 einsteins/

sec).

13

Thus, to avoid excessive photoreaction, one can either
dilute the solution or shorten the irradiation time. Neither
of these alternatives is satisfactory; in the case of a di—
lute solution correction has to be made for light transmit-
ted through the solution, and for light reflected from the
back window. Short irradiation does not allow for varia-
tions of light intensity and may also introduce timing errors.

We found also that the prepared Reinecke salt12 con-
tained free thiocyanate, and so did its ammonium salt, from
which it is prepared.

Further work on actinometry was abandoned when it be-
came clear that the optical density of our crystal on irradi—
ation is a function of more than one Species, and quantum

yields cannot be measured exactly.

Preparation of NaMg[Co(C204)3] Crystalsv

 

Solutions containing ten, five, and one percent by
weight, of K3[Co(C204)3] in saturated NaMg(Al(C204)3]-9H20
were prepared. The solutions were evaporated slowly while
standing in a dark place. A precipitate, believed to be
CoC204 from the thermal decomposition, formed, and had to
be filtered off.

A1(C204)33— is isomorphously1 replaced by Co(C204)33-,
and when the solution became saturated the dilute crystals
precipitated. A seed crystal was taken and hung from a

thread into the solution, and as it grew the solution was

frequently filtered to remove the decomposition product.

14

The crystals belong to the hexagonal system and grow
quite readily from solution. The 1% trisoxalatocobaltate(III)
crystals were not transparent throughout, showing some A
'cobalt(II) oxalate inclusion. ‘This was quite likely present
in the opaque five and ten percent crystals too. Growing
crystals at a lower temperature might be necessary to pre—
vent decomposition.

The crystals were cleaved with a razor blade. Axial
sections were those split normal to the C3 axis of the D3
ion.l . This is also the C6 axis of the hexagonal sys—
tem. These sections were relatively easier to cut, while
the orthoaxial sections, cut parallel to the main axis
broke and had to be sandpapered, and then rinsed with dis-

tilled water to prevent reflection from the surface.

Irradiation

 

The crystal sections were mounted in a metal Spool
which was placed into a holder to which a Nicol prism polar—
izer was attached. This equipment has been described pre—
viouslylo.

The light source was a high pressure, 1000 volt Hanovia
xenon-mercury arc lamp, supplied with a Hanovia compact arc
lamp power system, which contained the current regulator and
rectifier. The lamp operates at 15 amps, 60 volts fkf., and
is air cooled by a fan in a Schoeffel lamp housing. The
lamp is vented through the top of the housing into the fume—

hood to exhaust any ozone formed by the arc.

15
Output is mainly in the visible and in, with lesser
amounts of uv light. The energy distribution (Table I)

for a similar system is taken from a Schoeffel manual.

Table I. Energy distribution of a 2500w Xe—Hg arc lamp

 

 

 

Wavelength-Range(8) Energy (w) Percent of Total
2000-4000 2.95 18.9
4000-7000 6.34 40.5
7000-14,000 . 6.35 40.6

 

Monochromator filters could be inserted into the lamp
sleeve. ADue to the high power output they had to be water-
cooled. We found that the focal length of the converging
lens in the lamp housing was too short, and a convex lens
had to be added to focus the light on the crystal. The mount
was placed directly into the focal point of the lens, with
the polarizer in back (attached photo). Only two runs were
made with polarized light.

Various wavelength regions were selected by employing
appropriate interference and glass blocking filters. Their
spectral characteristics were determined spectrophotometri—
cally. All but one of the filters used had the secondary ir
windows blocked out. The results from this one filter,

T.p. = 365 mu, were ambiguous, and are not shown in this

work. Filter spectra were established in the 200 mu—700mu

region.

16

 

 

17

Table II. Monochromator filters.

 

 

 

 

Filter Spectral Slit Peak Transmittance
Designa— Width (mu) Wavelength Optical Remarks
tion (mu) Density
5850 300—490 360-390 .07
365 316-458 365 .29 IR window at 675
417 397-447 417 .54
427 416-458 427 .34
553 532—576 553 .42
572 552-602 577 .43
460-90—2 450-IR 470-IR .10 Short wavelength
cutoff
2424 600-IR 578-IR .07 Short wavelength
cutoff

 

After irradiation the mount was immediately placed
into the sample compartment of a Cary 14 Spectrophotometer.
Spectra were obtained using polarized light.

Before analysis, the position of the light beam from
the slit was centered on the crystal. After this adjustment
at the beginning of each run, care was taken not to move the
spool in the mount as this introduces a large error due to
the high optical density and local nonhomogeneity of the
crystal.

After analysis the crystal was immediately returned
for further photolysis. During this time the instrument
was zeroed again and the indicator was returned to starting
position. Thus, there was very little delay between photol—
ysis and the Scanning of the spectrum. This was done to

minimize the chances of radical decay.

18

Light intensity was measured using a YSI Kettering
Model 65 Radiometer. The sensing probe was a dual thermistor
bolometer. (Accuracy was i 5% full scale in use. For the
Hg-Xe light source the readings were off Scale and a 0.9
Bausch and Lomb neutral density filter was used to get them
on to the highest (2.5 x 106 ergs/cm2 sec) scale.

Tests were made to ascertain that all the light inten-
sity incident on the crystal was absorbed. Light was cen—
tered on the crystal and the intensity of the incident and
transmitted light measured. Light measurements behind the
crystal were identical to the background light in the room.

No light could be lost by reflection from the mount,
as the light was focused down to the size of the crystal
section.

The optical density of both sides of an irradiated and
visibly partially decomposed crystal are the same. We
conclude from these tests that Ia = Io, and there is no
scattering due to the decomposition product.

Bausch and Lomb 0.6 and 0.9 neutral density filters
were also used for Spectral measurements to compensate for
the high optical density of the crystal. -The filters were
calibrated in the 400—700 mu region. A box to hold the
filters was made up for the reference compartment of the
Cary 14. Thin sections of dilute (1%) crystals could be
run without filters.

‘ The mount and the spool holding the crystal section,

as well as the polarizer, were calibrated in this region.

19

To study the uv spectrum, the crystal was inserted
into a ring, cut from a doublefaced foamed adhesive. The
ring was then stuck on a uv silica cuvette. The polarizer
was an Abbe prism, which allowed more of the light to pass
through the crystal than the Nicol prism.

Quantum yield calculations were made for an aqueous
solution of trisoxalatocobaltate(III). This solution was

irradiated with the 6328 A light from a helium-neon laser.

___....T...

The energy output of this laser is at the same order as

that of our Hg-Xe arc lamp (I = 106 ergs/sec), but the plane
polarization and monochromaticity could have increased
effects on the crystal.

Quantum yields were determined by laser irradiation of
the solution in a liquid cell, stirred by a magnetic stirrer.
A reference solution was kept at the same temperature in
the room. The effect of irradiation was determined by dif-
ferential spectrophotometry to compensate for any decomposi-
tion due to thermal effects and to uv irradiation from
the room light.

It was found that transmitted light was negligible,
when compared to the background radiation. Therefore Io - Ia
where Io = incident light and Ia = absorbed light. This

neglects the small amount of light, which is reflected from

the front window at a normal angle of incidence.

20

The effect of thermal decomposition on the optical dens-
ity of a crystal was determined by the following method: A
crystal section was fitted into the metal spool of the mount.
The mount was then placed into a test tube which was wrapped
with black electrical tape to prevent any photochemical re—
action. It was then placed into a larger test tube, which
loosely sealed on top with cotton wool, and immersed into

a stirred water bath at 70°.

RESULTS AND DISCUSSION

The visible spectrum of potassium trisoxalatocobalt-
ate(III), Figure 2, shows two well defined peaks in this
region. We could therefore chose to irradiate either band.
Polarizations of ligand field transitions of the reaction
products were investigated using c and w spectra of
orthoaxial sections. The 0 section spectrum and that of
an axial section are identical.

We found that the optical density of the crystal on
irradiation was not a direct function of time. At the
start, the optical density decreased, and on prolonged
irradiation it increased. This was not consistent, however,
but varied from crystal to crystal. The optical density
of the host crystal itself also increased upon irradiation,
which might be due to thermal effects.

The discernible change in the spectrum on irradiation
is found in the Shape of the absorption bands. Both bands
broaden and then flatten out completely. This is also
seen with K3[Co(C204)3] alone, Figure 3. The higher fre-
quency ligand field band disappears first. The optical
density increases greatly in the area of the band, a phenom-
enon also found in the v spectrum, where this 427 mu

band is forbidden.
21

 

        

. , o o
“incomiommi “Iowmwxroomjr 1 .. "omvimviiiixiiimm
own
.coflusmom msomsvm
CH I .AvOmUvoo mo II
Enuummmm coausaom N.o
wHQHmH> .N musmflm
II omv
m.o
wulo can
We
OH .1
m o 0
Rule
afiw
one,
bmvmi Arbfiv

ONV

 

23

.HHSE HnAvONOVOUHAM emanacmnnfl

no mo Ednuommm mHQHmH>

.m museum

one Lv

 

com TV

no

as .3

24
These results suggest a strong absorption band develop-

ing in the near uv. It was located, as described before, at

320-350 mu
Investigation of the Near Ultraviolet Absorption Band

The crystal Spectrum of NaMg[Al(C204)3]'9H20 was deter—
mined to find out if it could be a source of the band. It
was found that the e = 141 band at 258 mu was shifted
toward the low energy region, 260-275 mu in the crystal,
but no band was found in the 320-350 region on irradiation.

The charge transfer band, at 242 mu, of trisoxalato-
cobaltate(III) is too far to cause this absorption.

The solution spectrum of trisoxalatocobaltate(II)
exhibits no uv band till 180 mu. ‘

vThe near ultraviolet absorption band, which is vari-
ously found between 320-350 mu; has been attributed by us,
to the carbon dioxide radical ion.

One of the reasons for this is that the optical density
of the band, after short irradiation, seems independent of
the trisoxalatocobaltate(III) concentration. This would
be the case if the primary product of the photolysis is a
radical with a high extinction coefficient. That this is
the case is proven by the ease with which the band grows,
on irradiation, to obscure the ligand field bands, Figures
4, 5.

The other reason supporting this theory is a report

by Carrington and McLachlan5, showing the carbon dioxide

320 25

 

(A) Before irradiation

05427 .575
05320 1.83

(B) After 10 min irradi-
ation with 360—390 mu
light

 

Figure 4. Develop-
ment of the 320 mu
band.

26

A_ Before irradiation
B_ After 10 min irra-
diation with 360-390 mu
330 light. Different crys-

tal from Figure 4.

lw

In»

420
400

-OD i 0.3

Figure 5. Development of 320 mu band.

27

radical ion of sz symmetry, with the extra electron lo-
calized in a carbon Sp2 orbital. The transition giving

rise to a reported band at 29,000 cm—1 (345 mu) is from an
antibonding a1 to an antibonding b1 orbital, i;e;

431 ‘—"> Zbl o

 

Z =A1

allowed only for the _
4a1 > Zbl x direction - Bl
Y = B2

We measured the polarized spectrum of an orthoaxial
section, which was irradiated for ten minutes, with light
from 532-576 mu. Absorption of light here is by the second,
long wavelength ligand field band.

The spectrum before irradiation, exhibited the usual
260 mu band, due to the host crystal. After fifteen min-
utes irradiation, a very intense band developed around
290 mu, still showing a Shoulder at 260 mu. Since the
absolute orientation of the crystal is not known,the direc-
tion is not known, but in one orientation of the polarizer
the absorption was less intense and the original 260 mu band
was better resolved, Figure 6.

The 290—350 mu band was found to be present in both
dilute and more concentrated crystals. Its increase was
found when irradiating with light of the following wave—

lengths: 360—390, 417 and 533 mu.

   

.Hmummuo prHHU amaxmonuuo :m CH Mmom >5 own map mo GOHDMNHHmHom .0 ouswflm

 

A3

¢.o.|iv . a

+ snows :2 mom
com QpHB Goaumflvmuna GHE.OH kumm Amy

coaumscmwnw mnommm Adv

    

29

Changes in the Two Visible Region Ligand Field Bands

Upon Irradiation

As these bands represent different transitions,

 

1A1

 

> 1T1 and 1A1 > 1T2 respectively, the reactiv—
ities might be expected to differ. Sarma found the decay
on irradiation, faster in the 610 mu band. It is obvious
that the effect of the high intensity uv band is to make
exact calculations impossible. The reason for the higher
quantum yield at the 610 mu band might well be the fact that
its decrease, due to decomposition, is less masked by the
uv band than that of the 427 band.

Quantum yields for the trisoxalatocobaltate(III) ion
in solution were made, using a neon laser (A = 6328 2)..
Assuming photodecomposition to be of first order the follow-
ing formula was used:

dc ¢Ia ¢ = quantum yield in
(1) ——- = --——— moles per einstein

V = volume of solution
in liters

c = concentration in
moles/liter.
(2) f—dc V = f¢ Io dt

(3) Integration between limits: (c1 - c2)V = ¢I0(t2 - t1)

(C1 - C2)V
(t2 ' t1)Io

 

6
ll

(4)

I0
from Beer's law log I_'

30
where: 2 = cell path
6 = molar extinction coefficient
c = concentration in moles/liter
I = transmitted light

OD = optical density.

 

 

Substituting:
(ODl " OD2)V (0131 " 0132) = differ-
(5) ¢ - e £(t2 — t1)Io 'ence in optical density
‘ of irradiated and non-
irradiated solutions.
OD1-OD2 t2 - t1(sec) V(£) Io(einsteinS/sec) e610 £(cm)
.008 4.5 x 103 .026 2.32 x 10'6 152 10

 

¢ = 1.09 x 10’5 moles/einstein, for irradiation with 6328 R
light, monitored at the 610 band.

.This result agrees with that of Knetenlz, who could not
measure photodecomposition accurately enough by chemical
analysis of Co(III), but estimated the limiting quantum
yield from experimental accuracy to be not greater than
1.5 x 10’5, at 610 mu.

When crystal sections were irradiated we could not com—
pensate for thermal effects. It was known that the thermal
instability of the solution due to decomposition was con—
siderably slowed down in the crystal. Some simple experi-
ments were made to determine the change in optical density

in the crystal on heating.

31
The thermal reaction was fitted to the first order rate

equation:

dc _
' at " kc
C:2 t2 (3
Integrating - f §§_ = k dt .'. k = E2f%—-log-l
C1 t1 2 1 C2
From Beer's law
_ OD . 2.3 ODl
C - -— O O k - — O —
e2 t2-t1 g 002

The change in optical density with prolonged heating
time was not consistent. As was observed in the photochemi-
cal reactions the ligand field bands flattened but the
overall optical density increased. At the start of the
heating, both ligand field bands were raised by the same
amount. The raising of the whole Spectrum might be due to
light scattering by the heated crystal. See Figures 7, 8.

Some irreproducibility was introduced by the removal
of the spool from thenount, for heating, followed by re-
placement in the mount. The position could not be exactly
reproduced.

The structure of the unit cell of the diluent crystal
with chromium substituted for aluminum has been established
by Frossard21 by single crystal X-ray diffraction; there are
two formulae units per unit cell. The compound is a nona—
hydrate. There are two hexaaquomagnesium ions and six non-

coordinated water molecules per unit cell.

 

 

can own

 

m o m o we
own
as
E
2 fie
3
com
Am me ma
5 5
mm ms.m mv.m «\H va
o5 vs.m wv.m «\H on Amv
cs ew.m 5H.“ m Amy
mm vfi.m Hw.H H: o AHV
05 name suvoo cameo u A

.20fluommn Hmfinmne .b musmflm

E.

33

em

Hm

hm

\.\../._\.\.
[_u(|I\\\lo lo/ \\\
;/ \\ fin
3+

Hm

hm

A.uqoov Gofluomwu HmEHmna .w musmam

34

.Assuming the cobalt compound to have the same structure,
the formation of the hydrated electron on heating is likely.
Its spectrum was determined by Hartzz: €72°°(m$x) = 15,800.
At 600 mu the extinction coefficient is around 10,000. The
spectrum is suppressed by known electron scavengers such as
1130+, 02. 1120.23

-To test this, air was flushed out with nitrogen and the
test tube sealed, and heated to see if there were a change
in the spectrum or an increase in the reaction rate. No
noticeable difference was found when heating a sample for
two hours in the presence, respectively absence of oxygen.

We conclude that the hydrated electron is absent, and
the carbon dioxide radical, possibly because of its proximity
to it, attacks another molecule of trisoxalatocobaltate(III)
preferentially to oxygen. -

This is the mechanism postulated for the photochemical
decomposition too. While this work is definitely preliminary
the results indicate that the thermal and photochemical re-

actions are identical.
.Results Obtained by Irradiating at Different Wavelengths

(1) Change in the Shape of the spectral bands.
(2) Polarization effects in orthoaxial spectra.

(3) Rate constants.

-Irradiations were carried out at the following wave-

lengths (mu):

35

white light — no filter > 470
360—390 553
365 577
417 > 578
427 632.8

At the high energy side of the second ligand field band
360 mu is the absorption minimum (Fig. 3). The low energy
minimum is at 500 mu. Energy between these limits is ab-
sorbed by the Amax = 427 mu band, while higher wavelengths
are absorbed at the Amax = 608 mu band. The charge trans-
fer band is at 242 mu and is not directly effected.

Assuming that the initial optical density is due to
trisoxalatocobaltate(III) alone, and that the first minimum

in the OD yg time plot represents ODoo' rate constants were

calculated. AS photochemical reactions are usually first

order:
ln[trisoxalatocobaltate(III)] = —kt + C
. ln (out-0000) =-kt+C
. . . _ 2.3 ODl'OD
Plotting ln(ODt ODGD) y§_time yields. k — t2-t1 log 653:6522 .

Unfortunately, in crystals of low trisoxalatocobaltate(III)
concentration ODOD was reached before a sufficient number of
measurements could be made. For these runs the rate constant
was calculated from the difference between two concentrations.

For more concentrated crystals log (OD - ODoo) y§_time
was plotted, and constants were calculated both ways. Calcu-
lation of constants from two concentration measurements were

labeled A, and those calculated from the graph were

36
designated B. For Spectra and graphs see Figures 9-20.
For rate constants see Table III.

The crystal section irradiated with 572 mu light was
orthoaxial, the others were axial. Half lives for decomposi-
tion ranged between 2.32-7.35 minutes. While the results
are of the same order of magnitude they are not consistent
enough to Show the expected increase in magnitude with in-
creasing light intensity.

.Reaction rates are essentially identical at both ligand

field bands.

Polarization of the Ligand Field Band Spectra of

the Decomposition Products

Spectra of crystals containing a low concentration,
about 1% by weight, of trisoxalatocobaltate(III) Show the
presence of two visible absorption bands abSent in the more
concentrated crystal. These bands are centered at 535-

540 mu and at 567-570 mu. Their intensities are low rela—
tive to that of the two ligand field bands of trisoxalato—
cobaltate(III) and upon irradiation their intensities change.
The change is dependent on the orientation of the crystal

for an orthoaxial section. See Figures 21—30 in Appendix.

We attribute these bands to a low symmetry decomposi-
tion product. This compound could be formed in one or both
of two ways.

(1) By a slow thermal reaction during crystal growth.

(2) By a photochemical decomposition during irradiation.

37

 

 

I
o
2.0"
CD
1 .5 db
0
o
1'3 : : : e : : If a 5 T
0 . 4 8 12 16 20
t (min)

Figure 9. Irradiation with white light. Axial section,
610 band.

38
fl

Figure 10. Spectrum for Figure 9.
Original

1
2 5 min irradiation
3 10 min irradiation
4

20 min irradiation

 

 

6 66

39

2 .7 Au.
Figure 11. .Irradiation with 360-390 mu
light. Axial section, 610 band.
8 v
G
O
I
Q
8
U‘ “T
0
F1
" o
O
2.0__
.r o
1 .8 A I : : 5 J
10 2D 30 40 50 1

 

 

t (min)

40

Figure 12. Log (OD - ODa>)-!§ time graph
for Figure 11.

 

 

 

t(min)

‘-

35

41

 

OD610
(1) Before irradiation 2.69 (off scale)
(2) 5 min. irradiation 2.20
(3) 20 min. irradiation 1.87
(4) 50 min. irradiation 1.67
(5) 200 min. irradiation 1.74
Not all points were plotted 610
to make spectrum clearer. ¢ (2)

 

Figure 13. Spectrum for Figure 11.

 

 

 

42

 

I
1.5.
L
1
1.4.1
O
OD’
O
O
1.3..
’ O
O
0 G (D
1.2._
0
1.1g 1 ‘ I I L # I L I , I I l
10 50 t (min) 100
Figure 14. Irradiation #2 with 360—390 mu. light. Axial

section at 610 mu band.

 

43

I
H
O

i

-1.5-

 

 

 

8
D
O
I
Q
3
m
I-l
-2.0 1
-2.5 -
—2.6 J l 1 4 :
0 10 20 _ 30 4o 50
t(min)

Figure 15. Log (OD — ODOO) y§_time graph for Figure 14.

 

44

 
    
      

owe

.vH musmfim How Eduuommm

 

 

 

 

 

1' .
,
'6’," 5"!

\ CHE. mH

K. 5.2 S

\I"
ll!“

    

own

.®H wusmflm

 

45

.COHuomm Hmaxé
.18 haw Ucm SE OHw um UmHOUHQOE .18 5H6 um GOHDMHUMHHH .bH mnsmflm

ACHEVmEflB

 

 

 

 

o o
om. ow om ow om mm m H. 1”
IS
no
So IV \
L
a no
/V
Io. a
Re Iv

46

 

{Changed to
Ivisible source

 

470
Figure 18. Ultra violet Spectrum after 1 hr irradi-

ation with 417 mu light
Note: Different crystal from Figure 17.

\ own own

 

 

47

 

 

coaumatmhua .cae so va
goaumatmhha .gaa am Amv
coahmaemhaa .cae ma Amy

Eduuommm Hmcfimfluo AHV

 

NH musmflm mo Eduuummm muflucm .mH musmwm

48

Crystal 1 orthoaxial section B 427
0 spectrum
Crystal 1 orthoaxial section O 610

 

. Crystal 1 orthoaxial section A 602 v spectrum
Crystal 2 axial section X 610 Spectrum
3 .0"- X/’x X
“\x-_/
OD -

2.5

 

 

 

2.0--

 

1.8 § ‘ L ‘ 1
10 50 t(min) Ibo

-h

Figure 20. Irradiation at 572 mu.

49

 

 

 

 

 

 

Table III. Rate constants for the photodecomposition.
Time OD log k(sec-1)
(sec) (OD - ODoo) (min) (einsteins/sec)
Figure 9. Irradiation with white light, monitored at the
‘ 610 mu band
0 1.69 - .43 2.17x10'3(A) 5.3 Not deter-
mined
5 1.35 ' -1.52
7 1.32 ODOO
Figure 11. Irradiation with 360-390 mu, monitored at 610 mu
0 2.69 -0.10 1.36x10'3(A) 6.15 3::10‘6
5 2.20 -0.51
20 2.09 -0.70 1.13x10‘3(B)
35 1.89 ODaD
Figure 14. ,Same as above, different crystal
0 1.59 -0.46 6.4x10"3(A) 3.96 2.341110”6
5 1.39 -0.70 2.9x10-3(B)
10 1.35 -0.80
15 1.31 -0.92.
20 1.27 -1.10
45 1 .19 ODCD
Figure 17. .Irradiation at 417 mu, monitored at 610 mu
0 1.29 —1.22 1.56x10'3(A) 7.35 5.65x10-8
15 1 .24 -2 .00
20 1.23 OD

CD

 

50

Table III. (Continued)

 

 

Time OD log k(sec-1)
(sec) (OD - ODGD) (min) (einsteins/sec)

f

 

Figure 17. .Irradiation at 417 mu, monitored at 427 mu

0 1.56 -1.3 1.67x10'3(A) 6.9 5.65x10'8
15 1.52 -2.0
18 1.51 0000

 

Figure 20. Irradiation at 572 mu, monitored at 610 mu

0 2.12 -0.96 4.8x10‘3(A) 2.38 2.32x10‘6
5 2.04 —1.52
7 2.01 0000

 

Figure 20. Irradiation at 572 mu, monitored at 427 mu
0 2.51 —o.89 4.95x1o'3(A) 2.32 2.32x10‘6
,5 2.42 —1.38

8 2.38 ODOD

 

51
The compound could not be seen in the more concentrated
crystal as it would be obscured by the intense 610 mu band.
The polarization for an orthoaxial section of the ligand
field transitions for the decomposition products were as

follows:

Table IV. Polarized spectra of the irradiated crystals

 

 

Irradiation(mu) 570 band - W 580 band - o 540 band - o

 

Run 1 I’OD incr. Marked incr. {Marked incr.
360—390
Run 2 OD incr. .Marked incr. Marked incr.
427 OD incr. Slight incr. -Slight incr.
> 450 OD incr. -Marked incr. Marked incr.
572 Slight incr. Slight incr. Slight incr.
577 Unchanged Unchanged Unchanged

 

The 427 mu transition is allowed for the o (i) spec-
trum only. The 540 mu transition which is polarized is al—
lowed in the same direction, and is absent in the v spec-
trum.

Light absorbed at the long wavelength 610 mu band does
not effect the spectrum noticeably, probably because the
energy absorbed is too small.

The decomposition product absorbing in this region could
be either bisoxalatodiaquocobaltate(II) or trisoxalatoco—
baltate(II). The former could be either the gig or the trans

isomer or both, as cobalt(II) complexes are labile. The

E——+>

52
gig isomer is the prevalent one formed from trisoxalato-
cobaltate(III) by loss of a ligand. This isomer has C2 sym-
metry. It is a likely product as the unit cell structure
of the complex in the diluent crystal contains six moles of
noncoordinated water.

However, calculations and spectra support the D3 sym-
metry of trisoxalatocobaltate(II). The reaction mechanism
postulated by us presupposes the presence of both compounds.
The aquo complex could presumably undergo anation by oxalate
and form the trisoxalato complex, which would then be the

final product of the reduction.

Calculations for is— bisoxalatodiaquocobaltate(II ) . C2 .
(1) For a d7 ion the ground state Russell-Saunders term
is 4F.

(2) Crystal field splits this state.24

 

 

 

9.2 32
4 ___ _____--
/A29 A
i_4'r
4P lg
4T ______________ A+ZB
4F 29
4 —— —-— ___-—-——- +
T19 A 2B
Act—'7‘

(3) in C2 the representations are as follows:25
z = A Taking the triple products the following

(x,y) = B selection rules are found:

 

53

 

 

A > A
l.

A > B
H

B > B

 

Since the ground state Tlg is Split and this split-
ting is probably too small to notice the Spectrum would not

appear to be polarized:

 

 

b

4T

291
A
//// a ‘\ J. Bb
4T ___—
19 \\\\ J.

4 ‘r-r' Aa
a \\ ll B
b

41:: ._______/A “\A
lg —_—\ /%
Ba
In the case of D3 symmetry the selection rules are of

course, the same as for the cobalt(III) ion but the states

are split differently for the d7 ion, i.e.:

54

 

 

 

 

 

 

 

 

Oh D3
4 _ H
'119 — A2 + E A2 > A1 E > A1
.11
A2 J— > E E --'1--——-> E
x _L_
‘5ng = A1 + E A2 > A2 E > A2
4A = A
29 2
x
—— A
_/ 2 4
A2 X A2 A T19(P)
/ _-l—_

 

 

4
The Tlg

its polarization was not investigated. A wide band of low

> 4ng transition is in the infrared region and

molar extinction was found at 1210 mu in solution. and was

attributed to this transition.

 

The 4T

1g > 4Azg transition is polarized in the same

 

direction as the 1A

1 o a 0
1g > .ng tranSition (427 mu) 1n

the reactant. Its polarization was therefore easy to find.

The spectrum was interpreted as follows:

 

 

 

. 4 4
v1 . Tlg > T29 at 1210 mu
92 : 4119(6) > 4119(9) at 570-580 mu
v3 : 4T1g > 4A29 at 535-540 mu
Values for Dq’ B, and v2 were calculated from v1
and v3.26
1

B - (v3 - 2V1 + 10 Dq)/15

55

v1 = 1210 mu or 8260 cm'1
v3 = 535 mu or 18,700 cm"1
Dq = 935 cm.1 B = 770 cm-1

v2 = v1 + 10 Dq = 17,610 cm.1 or 567 mu.

The v2 value is in excellent agreement with our experi-
mental value. B is lower than the value found for it in
the Tanabe and Sugano diagrams, where B = 970 cm.1 for
Co(II). This is due to the fact that the separation be-
tween the Russel-Saunders terms is less in the complex than
in the free ion, but the diagrams do not consider the nephel—
auxetic effect. The Dq value is of the magnitude usually

found for oxalate in the literature.

 

 

Calculating v1, v2 and v3 for the above values
of Dq and B, we get from the Tanabe and Sugano diagrams:
A _ 9350 =
Bi _ ——770 12.2 -1
4T > 4T2 = 9250 cm
1
4T1 > 4A2 = 16900 cm“1
4T1(F)--> 411(9) = 18,500 cm"1

The assignment for the latter two transitions are in-
terchanged from our assignments but the two states differ

only slightly in energy, in this region. The primary re-
6 e

29 9’

and evidence for the existence of hexacoordinated low Spin

duction product is expected to be low spin Co(II), t

Co(II) products has recently been reported in the litera-

ture.27 Due to the high pairing energy of 22,500 cm’1 19

the final oxalate compound is expected to be high spin

56

5 2
e

29 9
ion with excess oxalate Should be identical to the product

i.e. t . Thus the spectrum of the hexaquocobalt(II)

 

of the photodecomposition. There are, however, two visible
bands in the spectrum of the solid and one in the solution.

-The hexaquocobalt(II) ion spectrum Shows the absence

 

of the 4Tlg > 4Azg transition which is expected to be
weak. This may be the case with the oxalate in solutions,
too.

Conceivable too, in the solid the trisoxalatocobalt-
ate(II) ion responsible for the absorption is only present
in trace amounts in solution because the low K value.
This also accounts for the fact that the spectrum of a 4:1
ligand-metal solution is identical to that of a 3:1 solu-
tion.

Additional support for the existence of the reduced
oxalate complex is Wendland's9 work. Reflectance Spectro-
photometry established the disappearance of the high energy
ligand field band and a subsequent decrease in reflectance
at 540 mu.

We were unsuccessful in incorporating bisoxalatodiaquo—
cobaltate(II) into the diluent crystal. Its volume and that
of trisoxalatoaluminate are sufficiently different to make
substitution difficult.

An attempt was made to investigate the other decomposi-
tion products C02, and C021 . A mull made of potassium
trisoxalatocobaltate(III) was irradiated and a gas was pro-

duced. The spectrum could not be determined as the mull

57
was rearranged. This could be due either to C03, or to a
decomposition product of the mineral oil (Nujol).
Another mull, made up between Nacl plates, this time of
the dilute crystal, was irradiated and its IR spectrum was
analyzed to identify the gas. The IR spectra were run from

1

650 cm" to 4000 cm-1. The co2 fundamentals are:

1

v1 = 1286 cm'1 , v2 = 667 cm" , v3 = 2449 cm‘1.25

1388 cm"1

All three peaks were present before irradiation. 360-
390 mu light was centered on the mull for 25 minutes.

After the irradiation, when the spectrum was retraced,
no new peaks were found. There was a five percent decrease
in transmittance at v1 and v3: none at v2. We hoped
that as C02T is isoelectronic with N02 some of the IR peaks
could be identified in the new peaks after irradiating.

More work would have to be done to assure that the mull
is tight. The sample in the mull was rearranged on irradia—
tion. If this is due to thermal decomposition of the Nujol
possibly cooling, during photolysis, is necessary.

The photodecomposition is undoubtedly reversible as
cobalt(II) is readily oxidized by molecular oxygen in the
presence of complexing ligands, but the increased optical
density cannot be accounted for by this reaction. .Upon ir—
radiating for twenty minutes with uv light the crystal
surface is covered with the pink decomposition product but

the optical density is still increasing at the ligand field
bands of trisoxalatocobaltate(III). As discussed previously,

this is attributed to COZTL

SUMMARY

The postulated mechanism for the photodecomposition was
supported by tentative spectrophotometric identification of
some of the decomposition products i;g., C021 and Co(C204);-.

Cause of the high intensity band as a C02? transition
was fairly well established. The fact that a long lived
radical intermediate is present was suggested by others.
Parker and Hatchard's7 flash photolysis experiment with tris-
oxalatocobaltate(III) and trisoxalatoferrate(III) assumed it
to be HCOZ' or HC204' .-

E.s.r. studies on trisoxalatocobaltate(III) in the
diluent crystal Show a number of resonances which disappear
on irradiation.26 In place of them a large, narrow signal,
typical of an organic radical, and not present initially,
grows in.26 As Co3+ is diamagnetic, and octahedral Co2+
gives no detectable esr Spectrum at room temperature, the
system is particularly amenable to esr study. This will
be the next step in the identification of the C027. radical.

Work with solutions containing the hydrated electron
show C02 to react with e;q at k = 7.7 x 109 M71, sec-1.28
When C02 is added to de-aerated water the intensity of ab-

sorption, due to the hydrated electron decreases.23 This

58

59
suggests that in our system of oxalate cages C02? ion, rather
than egq is the stable radical product of the photodecompo—
sition.

Ultra violet spectra of solids have not been studied in
detail, and our method has been rather makeshift. A clear,
plastic mount which would hold the crystal in a definite
orientation could help to pin the C027' spectrum down exactly.

Further work is suggested by Eaton and Stuarts, who did
esr work on frozen, acidified solution of trisoxalato-
cobaltate(III). On irradiation their primary products were
H' and HC204' . They tentatively identified c025" as a
secondary product, resulting from further attack by H' .

If aqueous, non-acidified solutions were studied H' would
not be expected to be a primary product, also, the system
would be free of interference from sodium resonance from the

diluent crystal.

BIBLIOGRAPHY

10.
11.

12.

13.

14.

15.

BIBLIOGRAPHY

Powder X-ray data from this laboratory.

Piper, T. S. and R. L. Carlin. "Polarized visible
spectra of crystalline trisoxalatometallates", g;
Chem. Phys., 35, 1809 (1961).

Balzani, V., et al., Inorganica Chimica Acta Reviews,
1, 7 (1967).

 

Porter, G. B., J. G. W. Doerin , and S. Karanka, g,
Am. Chem. Soc., 84“ 4027 (1962?.

 

Carrington, A. and A. D. McLachlan, Introduction to
Magnetic Resonance, New York, Harper and Row, 1967.

 

Adamson, A. W. and A. H. Sporer, J. Am. Chem. Soc.,
QR, 3865 (1958).

Parker, C. A. and C. G. Hatchard, J. Phys. Chem., 63,
622 (1959).

Eaton, D. R. and S. R. Stuart, J. Phys. Chem., 72/
400 (1968).

Wendland, v. W. and E. L. Simmons, J. Inorg. and
Nuclear Chem., 22/ 2420 (1966).

Sarma, A. C., M.S. Thesis, University of Minnesota (1968).

 

Spees, S. T. and A. W. Adamson, Inorg. Chem., 1/ 531
(1962).

Kneten, N. C., M.S. Thesis, University of Minnesota
(1967).

Graziano, F. D., and G. M. Harris, J. Phys. Chem., fig,
330 (1959).

This work.

Palmer, W. G., Experimental Inorganic Chemistry, London
Cambridge University Press (1959).

60

TT-i _u7-_

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

61

Raripides, G. J., Ph.D. Thesis, University of Illinois
1964 .

Sandell, E. G., Colorimetric Determination of Traces of
Metals, New York, Intersceince (1950)

 

Schubert, J; L., E. L. Lind, W. M. Westfall, R. Pleger and
N.C. Li, J. Am. Chem. Soc., 82, 4799 (1958).

Cotton, F. A. and G. Wilkinson, Advanced Inorganic
Chemist£Y. New York, Interscience (1967).

Wegner, E. E. and A. W. Adamson, J. Am. Chem. Soc., 88,
394 (1966). ‘

Frossard. J., Schweiz. mineral petrog. Mitt., 88, 1
(1956).

Hart, E. J., Ann. Rev. of Nucl. Science, 18' (1965).

Hart, E. J., J. W. Boag, J. Am. Chem. Soc., 22x 4090
(1962).

Wilson, E. B., J. C. Decius and P. C. Cross, Molecular
Vibrations, New York, McGraw Hill Book Co., 1955.

 

Cotton, F. A., Chemical Applications of Group Theory,
New York, Interscience (1867).

Spees, S. T., Private communication.

.Kataoka, N., and H. Kon, J. Am. Chem. Soc., 88, 2978

(1968).

Hart, G. S., et al., Disc. Faraday Soc., 88, 193 (1963).

 

Roberts, G. L. and F. H. Field, J. Am. Chem. Soc., 18,
4232 (1950).

 

APPENDIX

 

1
590 2
, 3
’ 4
/ 5
,I 6
I
I
l 580
I
I
I
I
I
I
I
I
I
I
I
/
I
U
I
I
I
I
.7 570
f”, Figure 2 1 .
O”

 

0 2.96 ‘.\
7 2.94 \ \1
27 3.00 \ ‘
53 3.02 \ ‘
73 3.01 ..
111 3.04 \
640
\
\
\
)
I
‘.
('
\-\
\
650

Axial section for Figure 20,
610 mu.

 

 

cl)-

.Ednuommm 6 .ON musmflm How scanomm HMflxmonuHO .NN musmflm

 

63

 

 

own P b n p - - I\L\\ell.“lu|0|~lll-l| I”dlMfl’ - m p . n k
a I a a q a J .\I «I a a _ _ . Illa J a I .
A, \3/ \\\\.\A IIIIII won 1.77}
\\ /I\ \“e Ill IY/
\\ \\t II //
\ \ II I‘ll-III
\ s\ II
\\A.\ I/
\\ .l
\ S .
I
\ a men ,
I
x /
\ 8......
I
//
/ Ar
,, Se
/
z
I
z
x
x
a
I
a
z
I
/
66nd 63625 :9 ommnoomuuew.H me e x
no. a n» m ,
no. .6. me a .,
3. m o C x
no 255 it /
mafia ,

-D

.Esuuowmm 0

.ON mudmflm How :ofluomm HMHXmOQuHO .mm musmfim

 

05¢

 

GO

A

me.n new A

m 1 I mm: II . 1
en nn 8 can can 8 ad H e oe.m e .
mo.m m teen sue me.n m 1

teen «on wo.m m ne.m m .
mn.m H nn.n n x

no A in

65

 

.uzmfla
:8 «mm £UH3 :ofluomm amaxw am

we coaumflvmunH .VN muzmflm

own

a. H: «\H mcflvc

mm»;
AVV va

mH.H

.1E whm A K mo unmfla QUHB cofluomm amaxm cm mo coauMHUmnuH .mm musmflm

 

 

mum

m.o

add

6 mqw

6

mad
coflumflomuafl .cfla ooH Amv

coflumflumuufl .cfls mm va .I:

coflumflomuufl .cfia mv Amy o.o
GOHuMflUmnufl .GHE m ANV
coflpmflcmuufl mnemon “av

* .I.

v.0

 

 

 

 

GO
own
Esuuummwxo
Fowm....oo.m._o.$..P..P
— _ . u - q 1 . . 1 1 J 4 d 4 q
m
all.“ I,
ll \
\I'o’r.
o/o / \s“I.\I.IO\Io\\
/./. I \s~\
/. I Av \w\
, x\ _ .
/\)A\ ow wv H ¢
,):.4\ Ohm mv mv.H m
AHV Amy Amv mH wv.H m
_ o mv.H H
oom iii no %
u

 

EDHuUmmm h

cm mo cofluMHUmuuH

 

 

 

 

Esuuummm b

.unmfla 1e omw A K
nuflB :oHuumm Hafixmonuno

.mm musmflm

.Esuuommm b
.18 bmv n K nuHB cofiuumm Hmwxwonuuo

cm mo COHuMHUMHHH .hN «Human

 

 

 

 

 

 

H.\-l-1-”.I/M/
<
oom
mHH mo.m w //
om mo.m m /./
oH HH.m m :
V o moé. H . x
. x /
.2 H g g. ;
u .
o8 com own \7 . /
m8
omH.

 

69

 

.Eduuommm o .18 haw n A nuaB
coagomm Hmfixmosuuo am no GOHuMHUMHHH

 

mm mudmflm

 

.EsHuommm o .18 ommlomm n K QUHB aofiuomm HMfixmoauno am no GOHuMHUmHHH .mm mndmflm

     

H 41 _ + \A A
m o on own one own

 

 

mm.H .meHH .cHe A
mm.H .meHH .cHa om A
mm.H .cmHHH .cHe mw A
vm.H .omHHH .cHa m Amv
A
o

Hm.H .meHH muonn

71

1.85
1.82
1.81
1.69
1.66

 

 

Figure 30.

Irradiation

k

= 360-390

of an orthoaxial section with

mu,

F spectrum.

 

 

”'Tfiifiuifilmjijnifiu[giLflflgiflfiiiflMifijWV