View Article Online / Journal Homepage / Table of Contents for this issue

J. CHEM. SOC. DALTON TRANS. 1994 1271

Electron-gain and -loss Centres in Oxyanions containing S-S

Bonds: An Electron Spin Resonance Studyt
Shuddhodan P. Mishraa and Martyn C. R . Symons*-b
aNuclear and Radiation Chemistry Laboratory, Department of Chemistry, Baneras Hindu University,
Varanasi - 22 1005, India
Department of Chemistry and Biological Chemistry, University of Essex, Wivenhoe Park,
Colchester C 0 4 3S0, UK
Published on 01 January 1994. Downloaded by Bahir Dar University on 6/14/2021 2:54:57 PM.

The oxyions S,O,Z-, S,O,2- and S,O,2- have been exposed t o ionizing radiation at 77 K in the hope of
trapping the primary electron-gain and -loss centres for study by ESR spectroscopy. In order t o distinguish
features for the electron-capture centres uniquely, dilute glasses containing these anions were also
irradiated at 77 K and their spectra studied. In all cases, in addition to central features for the normal 32S
centres, various species containing one have been detected. Possible structures for these species are
proposed. Changes that occurred on warming t o room temperature are discussed in the light of these
identifications.

In our early studies of the action of ionizing radiation on solvent.' Unfortunately, this system is only suitable for neutral
inorganic salts, using ESR spectroscopy in attempts to identify solutes.
and study 'primary' radiolysis products, the pure compounds One of the most important results of these studies has been
were generally used.'*2 A wide range of radicals were detected, the discovery that electron addition into CJ bonds, giving CT*
and in many cases identified, but there were often ambiguities radicals often occurs.1o Although the bonding is weak, such
because of the dual nature of radiation damage. Thus electron- radicals are now being recognized as important intermediates
loss and -gain centres are expected to be the primary products, in many CI* Radicals can be formed either by
but one or both of these often break down unimolecularly to electron capture, equation (l), or by electron loss followed by
give other products, or they may react with neighbouring ions. bond formation [equations (2) and (3)]. The most stable species
Another complication is that electron return may result either in
very low yields or in bond homolysis, to give pairs of radicals
trapped in close proximity. Even more confusing is the fact that,
during the process of ionization and electron trapping, both the
loss centres (holes) and the gain centres may migrate by electron
transfer, and if there are suitable impurities present in trace
amounts these may act as hole or electron traps, giving radicals
from unknown precursors.
We therefore tried to make use of this migration propensity seen to be those in which A = B (i.e.ALA-, say), and when the
by deliberate doping. These studies showed that doping with CTbonding is between atoms of the second or higher rows in the
anions having one less charge than the host usually resulted in Periodic Table.
specific electron capture at the dopant, electron loss being Our aim herein was to probe the three sulfur-sulfur bonded
confined to the host. The opposite occurred if the dopant had ions S 2 0 4 2 - , S2OS2- and S20,2-, the main interest being in
one extra charge. For example, Po4,- in a sio44-lattice gave the mixed species, S 2 0 5 2-, because few mixed CJ and o* radicals
anions whilst B0,3- ions gave B0,02 - radical anions are known. Here, if bond breaking occurs after electron gain or
when doped into calcium ~ a r b o n a t eand
, ~ ClO, - gave C10,'2 - loss, we would expect to detect SO,'- or SO3'- from these
in this host.5 These results usually gave reliable values for the dimer anions. We have studied the pure compounds in the hope
specific gain or loss centres, enabling us to return to the pure of detecting primary electron-gain and -loss centres and, on
irradiated compounds and identify both centres4y5 annealing, the expected breakdown products, SO,'- and SO,'-.
A further development was the use of ionizing solvents such As a check on identification, we have used dilute solutions in
as methanol, which can be frozen to give glassy solids. Since CD,OD to induce electron capture selectively.
such glasses are well depicted as being infinitely viscous liquids
the solute remains dilutely dispersed, and on exposure to
ionizing radiation many solutes capture electrons exclusively.6 Experimental
We have recently helped to develop a technique used Sodium dithionate (BDH), dithionite (Lancaster) and p-0x0-
originally by Hamill,7 and by Shida and Takemura,' forming tetraoxodisulfate (Fisons) were the highest grades available and
exclusively electron-loss centres from solutes. The most widely were used as received. Salt solutions in CD,OD and C D 3 0 D
used solvent is CFC1,. In these solvents, solvent molecules + D 2 0 were ca. 0.1-1.0 mol drn-,. The pure salts were cooled,
rapidly scavenge ejected electrons, but the hole centres are as very fine powders, to 77 K, and irradiated with 6oCoy-rays
mobile uia electron transfer and hence can reach dilute solute or with 150 MeV (2.4 x J) X-rays at this temperature, to
molecules which will transfer an electron to the solvent doses of up to 10 kGy. The solutions were deuteriated and
reactions, if their ionization potentials are less than that of the frozen as very small glassy beads. The salts were also irradiated
at room temperature at higher doses in order to study low yields
of secondary products. Since the natural abundance of 33S is
t Non-SI unit employed: G = lo4 T. only 0.74%,and since it has I = g, the major central features are
 View Article Online

1272 J. CHEM. SOC. DALTON TRANS. 1994

due to 32Sradicals and it is necessary to study the other regions I 3265G
at high gain in order to detect centres containing one 33S.Those
containing two 33Sare too dilute to detect.

Results and Discussion

Pertinent results are summarized in the Tables and compared
with data from other sources therein. Some typical spectra are
shown in the Figures. The room-temperature samples gave well
defined spectra of SO3'-, or parallel 33Sfeatures for SO,'-, or
both (Fig. 1). These results give us bench-mark parameters for
the present matrixes. (One of us has previously shown that ESR
data for SO,'- and related species vary slightly from one matrix
environment to another.' 3,
Published on 01 January 1994. Downloaded by Bahir Dar University on 6/14/2021 2:54:57 PM.

The SO,'- radical anion, which is isoelectronic with the

remarkably stable radical, ClO,', is itself quite stable, and is
formed reversibly when S2042- ions are dissolved in protic
solvents such as water.14 It is a .n radical, the 33S hyperfine
coupling showing large A l l values and small A l values
(Table l ) . 1 5 Fig. 1 First-derivative ESR spectrum for Na2S20, after exposure to
The SO,'- radical has a pyramidal structure, the singly ionizing radiation at 77 K and warming to room temperature, showing
occupied molecular orbital (SOMO) on sulfur being hybridized, features assigned to radicals containing 33S, identified as 33S03'-
with considerable 3s character.16.17It is far more reactive than radicals
SO,'-, and there is no tendency for its dimer, S 2 0 6 , - , to
dissociate at ambient temperatures.
Sulfur-centred o* radicals have been studied extensively, and Table 1 The ESR parameters for 3 3 S 0 2 * pand 3 3 S 0 3 * -in various
the electron adducts of organic disulfides, RSzSR-, are thought matrices
to be important intermediates in the interconversion of sulf- 33Shyperfine coupling/G
'
hydryl compounds into disulfides. Also, the corresponding
o*-radical cations, R,S-SR, +,are well characterised by ESR Radical Ref. A A, Aiso gll g,
and optical spectroscopy and are remarkably stable.' 8 0 , 1 9 , 2 0 33s02'-a 58 (-)7 15 2.011 2.003
The mixed disulfide, R,S-SR is also thought to be important,,' b 59 (-)8 14 2.009 2.002
but there is some controversy regarding its identification. 2 2 c - - 15 - -

This species is of interest in relation to our present study of 33S03'- d 153 115.5 128 2.0035 2.0035
o,s-so23 centres. e 150 120 130 2.0034 2.0034
a 155 117 130 2.0034 2.0034
Sodium Dithionate.-At 77 K there are two well defined sets "This work. Ref. 15. 'Ref. 14. Ref. 16. Ref. 17.
of 33S features, as previously observed.,, We originally
identified both these centres as the o* adduct, [0,S-S0,]3-,
which was thought to be asymmetrically located in the crystal Table 2 The ESR parameters for various o1 and o* radicals formed
such that the SOMO was unequally distributed, giving a 33S from irradiated S 2 0 6 2 - ,S 2 0 5 2 -and S 2 0 h 2 - ions.
coupling of A II = 165 G, A , = 135 G for one and A II = 158 G,
A , = 130G for the other. This seems to be quite reasonable, but 3S hyperfine
coupling/G
it leaves the corresponding electron-loss centre unidentified.
This could be a centre, similar to SO,- ions,''2 having a SOMO Radical Matrix A, Aim gll,g,
confined to one or more of the oxygen ligands on one side of the
0233%S033- Na,S205 122 90 101 2.002,2.004
anion, as in 'OSO,-SO,-. This would only give a small 33S 100 2.002, 2.004
O233sIS0,3- Na,S20, 121 89
coupling, but large positive g shifts are expected. In fact, there is 0333sIS033- Na2S206 158 135 143 2.0025,2.0025
only one intense central feature for 32S radicals, centred at g = 0, s*so3 - Na2S206 165 130 142 2.0025,2.0025
2.0025, showing little or no g anisotropy. Hence oxygen-centred
species seem to be unlikely. The alternative, not previously
considered, is electron loss from the S-S o bond giving the Table 3 Approximate orbital populations for the 3s and 3p sulfur
one-electron radical, [O,S-SO,] -. Such one-electron-bonded orbitals deduced from the isotropic and anisotropic 33Sdata given in
centres are much less common than their three-electron Tables 1 and 2
analogues, but they should have comparable stability. Known
o1centres include AgAg', Me,C=CMe,+ and H,BoBH,-.,~- 26
Radical p: s ratio (p + s) (%)
If only one of the centres is the o* species and the other is a o1 3 3 ~ ~ 2 ' - 32 70
centre, then we propose that the data be reassigned such that the 33S03*- 4.2 54.5
O3 3%so3 - 2.3 37.6 (75.2")
larger A l l is taken with the smaller A , and vice versa. This 0333s.so3- 3.7 51 (102")
combination is given in Table 2. The reasoning behind this 0233sIS033-(a) 4.6 45
reassignment is that the o*anion should dissociate to give SO,' -
+ SO3,- whilst the o1centre would give SO,'- and SO,. In the " Total spin density on the two sulfur atoms. bThe alternative,
02S-33S033-,was not clearly defined.
former, the second product, SO,2-, is markedly pyramidal,
whilst SO, is planar. Hence we infer that the SO, units in the o*
centre should be more pyramidal than those in the o1unit. This
is the case using these alternative combinations of data. Also, the the usual way,27 using the (3s)' values of 1237 G given by
SOMO for the species assigned the a' structure is strongly Morton and Preston 2 8 but the (3p)' value of 57 G recommended
confined to the two sulfur atoms, whilst there is greater in ref. 27. We have previously found that this choice gives the
delocalization for the o* SOMO (Table 3). This contrast is again most reasonable results for a wide range of radicals.
reasonable when comparing the lower and upper (J orbitals. The We conclude that the two sulfur centres, detected in equal
values for 3s and 3p character given in Table 3 were computed in yields for irradiated Na,S,O,, may well be the primary (J' and
 View Article Online

J. CHEM. SOC. DALTON TRANS. 1994 1273

o* centres. To check the o* assignment, we studied S,O,'- ions

as dilute solutions in a methanol glass. The features were poorly
defined, but were in reasonable agreement with the results for 50 G
the cr* species given (Table 2).
*
It is perhaps surprising that the data for 33S03'- radicals
(Table 1) are quite similar to those assigned to S20603- and
S,O,'- since in the latter only 50% of the SOMO is associated
with each ,,SO3 unit. However, the same behaviour is found for
phosphorus centres, where the spectra are far better defined and
hence more reliably interpreted. For example, the radical cation
P(OMe),'+ has A l l z 900 G and A , x 713 G, whilst the o*
dimer (MeO),P-P(OMe),+ has A , , z 727 G and A , z 619
G." Again, the trends are the same: the P(OR), units become
Published on 01 January 1994. Downloaded by Bahir Dar University on 6/14/2021 2:54:57 PM.

more pyramidal in the dimer, the p :s ratio falling from 3.9 to

2.6 : 1, and the SOMO is far more confined to phosphorus in the
dimer. No clear ,,S-containing species were obtained for frozen
methanolic solutions.

Sodium CL-Oxo-tetraoxodiSuIfate.-The asymmetric anion Fig. 2 First-derivative ESR spectrum for Na,S,O, after exposure to
[0,S-S0,]2 - is considered before [02S-S02]2 - because our ionizing radiation at 77 K, showing 33S features also assigned to
interpretation of results for the latter hinge on those for the [O2SISOJ3 - radicals (a)
former. In this case, considering the o1 and o* species as
probable primary centres, there will be two distinct sets of 33S 9, 2-
satellite lines, one for 0233s-so,species and one for S-9
o,s-33so,species. If some dissociation also occurs, then both I
S-0
' I
SO3'- and SO,'- are expected to form. Since the electron
affinity of the SO, unit is greater than that for the SO, unit, we
expect the SOMO for the o1centre to be centred more on the Fig. 3 Structure for the S2042-anion showing possible weak bonding
SO, unit, whilst that for the o* species will be centred more on between the oxygen ligands
the SO, half of the radical anion. Hence there should be four
sets of features. One is identified as belonging to the o* anion, from SO,'- radicals, but it was less than 1% of that given on
and because of the form of the parameters is tentatively assigned irradiation. The structure of the S204' - ion is as shown in Fig.
to [ 0 , 3 3 S S 0 , ] 3 - . This is because the p:s ratio is relatively 3. The oxygen ligands eclipse, implying weak bonding via
large (4.6: 1) (Table 3), as expected for the SO, unit, since this oxygen, whilst the S-S bond is unusually long. Hence we had
gives SO,'- on dissociation. The spin density is ca. 45% on expected primary electron loss and gain to occur at this weak
sulfur, which is relatively high, since SO2'- itself has a spin bond giving S,O,'- + S204'3-. However, the S-S bonds in
density of only ca. 65% on sulfur,l5 using the same atomic these radicals will be even weaker, and we had therefore
parameters. expected to observe dissociation to give SO,'- + SO, and
This centre ( a ) is of importance, since it is also clearly formed SO,'- + SO,'-. Certainly SO,'- is a product. If the other 33S
in irradiated dithionite (Fig. 2). We suggest that 0'- radicals features are due to the primary loss and gain (0' and o*)
are formed in this system and that these add to one side of the species, then there is a surprising degree of orbital hybridization
S 2 0 4 2 - units to give [0,S-S03]3-. It seems that the species in the SOMO of both. Addition of 0'- to S z 0 4 2 - would give
most readily detected is the 0 2 3 3 Sderivative. If this is correct, the o* radical anion [02SS0,]3-. This is the species, a, that is
both sets of spectra must also contain features for [ O z S common to both S2042- and S 2 0 s 2- systems, and we consider
33S03]3- . There are indeed other common features, but these the identification to be firm. After allowing for the two sets of
are poorly defined, and overlap with the ++ features of the quartets that this o* radical anion generates, there are no other
other centres. Thus, we are satisfied that the required features well defined 33Sfeatures remaining. Hence we suggest that a is,
are present, but hesitate to attempt to produce 33S-hyperfine effectively, the electron-loss centre. The electron-gain centre is
data therefrom. then presumed to be SO,'- formed together with SO,'- by
It seems certain that the o* species is formed, as expected, but dissociative electron capture. Studies using methanol were
the identities of the electron-loss centres are less sure. There is a conducted by cooling to just above the glass temperature to
weak feature in the g 2.01 region which is probably the g,,, remove SO,'-, followed by cooling to 77 K and irradiating.
feature for SO,'- radicals formed in relatively low abundance. Only features for SOz'- were detected. Thus, our search for the
The expected 40 G doublet for 'OH radicals is definitely absent. o1and o* centres for this ion have failed.
We therefore suggest that, as with S 2 0 6 2 - ,electron loss is from We conclude that, for S,O,'- and S206'-, both o1and o*
the S-S bond, giving an asymmetric o1 structure which species are formed on electron-loss and -gain, but that
contributes to the central broad line. This centre, for reasons dissociation to give SOz'- + SO, or SO2,- occurs even at
given above, should have a SOMO biased towards the SO, unit 77 K for S204'- ions.
and hence should show a large 33S hyperfine splitting. Apart
from the features already assigned, there are no resolved
features of the type expected. The o' centre would dissociate to Acknowledgements
give SO, and SO,'-. Features for 33S0,'- clearly grow on We thank the Commission for the European Communities for a
annealing to room temperature, suggesting that the o1radical is contract with India supporting this study.
present, but are not evident at 77 K.
The results for methanolic glasses containing S 2 0 s 2- ions
confirm the identification of a as the electron-capture centre.
References
1 P. W. Atkins and M. C. R. Symons, The Structure of Inorganic
Sodium Dithionite.-The ESR spectrum for this salt, Radicals, Elsevier, Amsterdam, 1967.
irradiated at 77 K, shows a central asymmetric singlet, a g,,, 2 M. C. R. Symons, Ado. Chern., 1968,82, 1.
feature at 2.02 1 , and several sets of 33Scomponents in the wings 3 M. C. R. Symons, J. Chem. Phys., 1970,53,857.
(Fig. 2). Prior to irradiation this salt gave a very weak signal 4 R. S. Eachus and M. C. R. Symons, J. Chem. SOC.A , 1968,2438.
 View Article Online

1274 J. CHEM. SOC. DALTON TRANS. 1994

5 R. S. Eachus and M. C. R. Symons, J. Chem. SOC.A , 1968,2433. 19 B. C. Gilbert, D. K. C. Hodgemann and R. 0.C. Norman, J. Chem.
6 M. C. R. Symons, in Radicaux Libres Organiques, Centre National SOC.,Perkin Trans. 2, 1973, 1748.
de la Recherche Scientifique, Paris, 1977, vol. 287, p. 105. 20 D. M . R. Rao, M. C. R. Symons and B. W. Wren, J. Chem. SOC.,
7 W. H. Hamill, in Radical Ions, eds. E. T. Kaiser and L. Kevan, Perkin Trans. 2, 1982, 1681.
Wiley, New York, 1968, p. 239. 21 M. C. R. Symons, J. Chem. SOC.,Perkin Trans. 2, 1976,908.
8 T. Shida and Y. Takemura, Radiat. Phys. Chem., 1983,21, 157. 22 J. H. Hadley and W. Gordy, Proc. Natl. Acad. Sci. USA, 1975,72,
9 M. C. R. Symons, Chem. SOC.Rev., 1984,13,393. 3486.
10 A. Abu-Raqabah and M. C. R. Symons, J. Chem. SOC.,Faraday 23 M. C. R. Symons, J. Chem. SOC.,Dalton Trans., 1979, 1468.
Trans., 1990, 3293. 24 I. G. Smith and M. C. R. Symons, J. Chem. Rex, 1979, ( S )382.
1 1 R. Breslow, M. Brand1 and A. D. Adams, J. Am. Chem. Soc., 1987, 25 P. H. Kasai and D. McLeod, J. Chem. Phys., 1969,51, 1250.
109,3799. 26 T. A. Claxton, R. E. Overill and M. C. R. Symons, Mol. Phys., 1974,
12 A. Abu-Raqabah and M. C. R. Symons,J. Am. Chem.SOC.,1990,112, 27, 701.
8614. 27 M. C. R. Symons, Chemical and Biochemical Aspects of Electron Spin
13 M. C. R. Symons, J. Chem. SOC.A , 1970, 1998. Resonance Spectroscopy, Van Nostrand Reinhold, London, 1978.
14 P. W. Atkins, A. Horsfieldand M. C. R. Symons, J. Chem. SOC.,1964, 28 J. R. Morton and K. F. Preston, J. Magn. Reson., 1978,50, 577.
Published on 01 January 1994. Downloaded by Bahir Dar University on 6/14/2021 2:54:57 PM.

5220. 29 A. Hasegawa, G. D. G. McConnochie and M. C. R. Symons,J. Chem.

15 D. Schoemaker and C. T. Shirkey, Phys. Rev., 1972,136, 1562. Soc., Faraday Trans. I, 1984, 1005.
16 G. Lind and R. Kewley, Can. J. Chem., 1972,50,43.
17 M . J. Lin and J. H. Lunsford, J. Phys. Chem., 1975,79,892.
18 D. J. Nelson, R. L. Petersen and M. C. R. Symons, J. Chem. Soc.,
Perkin Trans. 2, (a) 1977,2005; (b) 1978, 1978. Received 9th December 1993; Paper 3/07294K