Kinetics and mechanism of the oxidation of some aldoses, amino sugars and methylated sugars by tris(pyridine-2-carboxylato)manganese(III) in weakly acidic medium

Article abstract of DOI:10.1016/S0008-6215(98)00329-2

The kinetics of oxidation of some aldoses, amino sugars and methylated sugars by tris(pyridine-2-carboxylato)manganese(III) have been studied spectrophotometrically in sodium picolinate -picolinic acid buffer medium. The reactions are first order with respect to both manganese(III) and sugar concentrations, but independent with respect to sodium picolinate - picolinic acid buffer medium. The mechanism for the reactions is discussed.

Full text of DOI:10.1016/S0008-6215(98)00329-2

Carbohydrate Research 315 (1999) 70–75
Kinetics and mechanism of the oxidation of some
aldoses, amino sugars and methylated sugars by
tris(pyridine-2-carboxylato)manganese(III) in weakly
acidic medium
Kalyan Kali Sen Gupta *, Bilkis Ara Begum
Department of Chemistry, Jada6pur Uni6ersity, Calcutta 700 032, India
Received 14 July 1998; accepted 1 December 1998
Abstract
The kinetics of oxidation of some aldoses, amino sugars and methylated sugars by tris(pyridine-2-carboxy-
lato)manganese(III) have been studied spectrophotometrically in sodium picolinate–picolinic acid buffer medium.
The reactions are first-order with respect to both manganese(III) and sugar concentrations, but independent with
respect to sodium picolinate–picolinic acid buffer medium. The mechanisms for the reactions are discussed. © 1999
Elsevier Science Ltd. All rights reserved.
Keywords: Kinetics; Sodium picolinate–picolinic acid; Manganese
tion of
D-glucose by manganese(III) has been
1. Introduction
studied [13], and manganese(III) is shown to
behave as a one-electron transfer oxidant. The
present investigation was undertaken in order
to clarify the mechanism of oxidation of some
aldoses, amino sugars, and methylated sugars
by tris(pyridine-2-carboxylato)manganese(III)
in sodium picolinate–picolinic acid buffer
medium. The tris-manganese(III) complex is
very unstable in acid or alkaline medium, but
it is stable in sodium picolinate–picolinic acid
buffer medium in the pH range 4.22–6.45,
which minimizes the possibility of a ligand
substitution reaction.
Manganese can adopt a wide variety of
oxidation states, and this ability is certainly
related to the redox function of the metal ion
in biological systems [1]. The tripositive state
of the metal requires special attention, not
only because of its biochemical relevance in
diverse redox functions [1,2], but also because
it is difficult to stabilize it in aqueous medium
[3], and many of its complexes exhibit unusual
magnetic and structural features [4]. There are
several reports on the kinetics of oxidation of
various substrates by manganese(III) in
perchlorate, sulfate, acetate and pyrophos-
phate media [5]. The important ones involve
reactions of oximes [6], vitamins [7,8] and
amino acids [9–12]. The kinetics of the oxida-
2. Experimental
Reagents.—The aldoses, amino sugars and
methylated sugars were obtained from BDH,
E. Merck or Sigma Chemical Co. Pyridine-2-
* Corresponding author.
0008-6215/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(98)00329-2

K.K. Sen Gupta, A.B. Bilkis / Carbohydrate Research 315 (1999) 70–75
71
carboxylic acid was from Lancaster Synthesis.
All other materials were of certified or reagent
grade.
Instruments.—Spectral measurements were
carried out in the UV–vis region using a
Systronics (India) spectrophotometer. Mag-
netic susceptibility was measured at rt by a
Princeton Applied Research vibrating sample
magnetometer using Hg[Co(SCN)4] as the cal-
ibrant. Cyclic voltammograms were recorded
by a Bio-Analytical System (BAS) CV-27. All
experiments were carried out in a dry argon
atmosphere using a three-electrode configura-
tion. A planar Beckman model platinum elec-
trode was used as the working electrode and
SCE as the reference electrode. C, H and N
analyses were performed by the microanalyti-
cal laboratory using a Perkin–Elmer 240
CHN Analyser. TGA studies were performed
with a Shimadzu Corporation (Japan) TG 50
instrument in a normal atmospheric environ-
ment at a rate of 15 °C/min up to 400 °C. pH
measurements were carried out in an Elico LI
120 (India) pH meter. The EPR spectrum was
recorded with a Varian EPR spectrometer.
Kinetic measurements.—The electronic
spectrum of the manganese(III) complex ex-
hibited [15] two weak shoulders at 21,900
cm−1 (m=257 cm−1 mol−1 dm3) and 24,800
cm−1 (m=407 cm−1 mol−1 dm3). Since the
extinction coefficient of manganese(III) solu-
tion is small in the visible region, the progress
of the reaction was followed in the UV region
where manganese(III) absorbs appreciably at
the concentration region of 0.5–5.0×10−4
mol dm−3. Beer’s law is also obeyed in this
concentration range. The reaction rate was
monitored spectrophotometrically at 350 nm
(m=2710 cm−1 mol−1 dm3) under pseudo-
first-order conditions, i.e., [sugar]ꢀ[tris-
Mn(III)]. The pseudo-first-order rate constant
(kobs) was determined from the plot of log A
(A=absorbance) vs time. The rate constants
(kobs) calculated by the graphical method were
reproducible within 93% error.
Preparation of the manganese(III) com-
plex.—The complex, tris(pyridine-2-carboxyl-
ato)manganese(III) was prepared by
a
modified procedure from that reported in the
literature [14]. MnCl2·6H2O (2.5 g, 10.7 mmol)
was dissolved in water (25 mL). The solution
was warmed, and H2O2 (25 mL, 808.8 mmol)
was added to it. This was followed by the
addition of NaOH (1 g, 25 mmol) dissolved in
water (25 mL). The mixture was stirred vigor-
ously for at least 30 min, and the precipitated
MnO2 was filtered through a sintered glass
crucible. The precipitate was washed thor-
oughly with water and then warmed with an
aq solution of excess pyridine-2-carboxylic
acid with vigorous stirring. The mixture was
then filtered through a sintered glass crucible.
The scarlet–red filtrate deposited red crystals
of tris-manganese(III) on cooling. Anal. Calcd
for [Mn(C5H4NCO2)3]H2O: C, 49.22; H, 3.21;
N, 9.56. Found: C, 49.95; H, 3.30; N, 9.36.
Thermogravimetric analysis of the solid
manganese(III) compound was carried out by
heating a sample (19.39 mg) at the rate of
15 °C/min up to 400 °C. The result indicated
that loss of water took place in one step,
based on the weight-loss calculation for
[Mn(C5H4NCO2)3]H2O, confirming the loss of
only one water molecule. The magnetic mo-
ment of the solid compound at room tempera-
ture (rt, 25 °C) has been found to be 4.92 BM.
This value is not different from that men-
tioned earlier [15]. The structure of the man-
ganese(III) complex has been reported [16] to
be a pseudo-tetragonally distorted octahedron
with the axial bonds slightly elongated. An
attempt has been made to determine the redox
potential in sodium picolinate–picolinic acid
buffer medium. Mn3+ (2×10−4 mol dm−3
)
at pH 6.1 and with NaCl as the supporting
electrolyte is found to undergo an irreversible
one-electron reduction at −0.80 V at 298 K.
Preparation of buffer solution.—The pKa
value of picolinic acid is 5.52 [17]. Buffer
solutions were prepared from standard solu-
tions of sodium picolinate and picolinic acid.
The pH of the buffer solution was checked
against a standard buffer solution with a pH
meter.
Stoichiometry and product analysis.—The
reaction mixtures containing a known excess
of tris[Mn(III)] of 1.3×10−3 mol dm−3 and
[glucose] of 2×10−4 mol dm−3 at fixed pH
6.1 were made. After completion of the reac-
tions, the unreacted [Mn(III)] was determined
by iodometric titration with a standard
sodium thiosulfate solution using starch indi-
cator near the end point [15]. Determination

72
K.K. Sen Gupta, A.B. Bilkis / Carbohydrate Research 315 (1999) 70–75
of unreacted [Mn(III)] showed that 1 mol of
glucose consumes 2 mol of Mn(III). The reac-
tion may be represented as follows:
micro test tube with a drop of a solution of 1
g of FeSO4, (NH4)2SO4·6H2O and 0.5 g of KF
in 100 mL of 0.1 N acetic acid to give a
light-yellow color, which is stable in acetic
acid or dilute mineral acids. The colored reac-
tion is due to the formation of the inner-com-
plex cation between Fe2+ and the picolinate
anion. The color intensity of the reaction mix-
ture with the above reagents was much higher
than that of the buffer mixture, and this was
confirmed spectrophotometrically.
R–CHO+2[Mn(III)(C5H4NCO2)3]+5H2O
“R–COO−
+2[Mn(II)(C5H4NCO2)2(H2O)2]
+2C5H4NCO2− +3H+
(1)
where R= –(CHOH)4CH2OH.
After the kinetic experiments, the reaction
mixtures of the aldoses and amino sugars were
acidified and treated with 2,4-dinitrophenylhy-
drazine hydrochloride. The absence of any
yellow precipitate indicates that neither the
–CH2OH nor the ꢀCHOH group is oxidized.
The iron(III) chloride solution, colored violet
with phenol, when added to the reaction mix-
ture gave a bright yellow coloration [18], indi-
cating that aldonic acids are formed in the
oxidation of the aldoses. The aldoses were
oxidized separately by the oxidant under ki-
netic conditions. After purification and con-
centration, the products were confirmed [18]
by paper chromatography in comparison with
the oxidation products of the respective al-
doses by bromine. Similar experiments when
carried out with the oxidation products of the
amino sugars gave yellow colorations with the
above iron(III) reagent. The amino sugars are
also oxidized [19] by HgO to give aldosaminic
acids, and when iron(III) chloride solution,
colored violet with phenol, is added to an
aldosaminic acid solution, a yellow coloration
is also given. However, the addition of alka-
line hydroxylamine solution to the reaction
mixtures and subsequent addition of 2% FeCl3
and 1% HCl failed to give any mauve col-
oration [20], which indicated that no lactone
formation was taking place during the reac-
tion. This is to be expected since primary
products in the oxidation of cyclic aldoses are
aldonic acids or g- or d-lactones in acidic
solutions, but at pH\2.5 the proportion of
lactones is reduced [21]. Since the pKa value of
gluconic acid is 3.81, the thermodynamically
stable species gluconate ion will be formed at
pH 6.1.
Polymerization test.—Acrylamide (40%)
was added during the course of the reactions.
An immediate haziness appeared during the
oxidation of the substrate by manganese(III).
When an excess of methanol was added to the
reaction mixture, a thick precipitate of poly-
acrylamide was formed, demonstrating that
free radicals intervene in the oxidation pro-
cess. The observation is to be expected since
manganese(III) is a one-electron transfer oxi-
dant that is reduced by the substrates to give
free radicals and manganese(II). Blank experi-
ments in which either the oxidant or the sub-
strate were excluded gave no polymeric
suspension. Moreover, the presence of man-
ganese(II) in the reaction mixture was confi-
rmed by the appearance of the typical six lines
in the EPR spectrum as has been reported in
the literature [22b].
3. Results and discussion
The reactions were carried out at different
[Mn(III)] in the region 1–5×10−4 mol dm−3
but at constant concentrations of substrates,
pH and temperature. The rate constants were
independent of oxidant concentration in each
reaction (Table 1). The pseudo-first-order rate
constants (kobs) were calculated at constant
[Mn(III)], pH and temperature but at different
substrate concentrations. The results plotted
in Fig. 1 indicate that each reaction is first
order with respect to the [substrate]. The de-
crease in the rate with increase in asymmetric
centre (n) is shown in Fig. 2. The plot of
log k2 (k2=kobs/[substrate]) vs ‘n’ is hyper-
bolic in character [23a]. The fact that the
acyclic aldoses react at much faster rates than
the cyclic ones is to be expected since the
That picolinic acid is generated [22a] during
the reaction is confirmed by the following test.
One drop of the test solution is mixed in a

K.K. Sen Gupta, A.B. Bilkis / Carbohydrate Research 315 (1999) 70–75
73
Table 1
Effect of tris[Mn(III)] on the rate of oxidationa
Substrate
k )
obs×104 (s−1
D
D
D
D
-Glucose
-Mannose
-Galactose
-Xylose
0.49990.01
0.70490.01
0.88690.02
1.2590.05
2.2390.03
3.3390.09
4.7190.10
29.490.12
0.26590.01
0.41990.01
0.66590.01
L
-Arabinose
D
-Ribose
-Erythrose
D
DL-Glyceraldehyde
D
D
D
-Glucosamine hydrochloride
-Mannosamine hydrochloride
-Galactosamine hydrochloride
Fig. 2. Dependence of the reaction rate on the number of
asymmetric centres (n). Plot of log k2 vs ‘n’.
a T=313 K, [Mn(III)]=(1–5)×10−4 mol dm−3, pH 6.1,
[substrate]=0.3 mol dm−3
The influence of temperature on the rate
was determined for each reaction. The values
of k2 were calculated at different temperatures
from 308 to 328 K. Least-squares analyses of
log k2/T against 1/T plots (Fig. 3) were used to
obtain the best straight lines from which the
enthalpies of activation (DH") and entropies
of activation (DS") were calculated using the
theory of absolute reaction rates.
The activation parameters are recorded in
Table 2. The slower reactions are character-
ized by negative entropies of activation, unlike
the fast reactions where entropies of activation
are positive [23b]. The much slower rates and
negative entropies of activation obtained in
the present study, as compared to the oxida-
.
concentration of potential –CHO groups is
higher in acyclic aldoses than the cyclic
sugars.
The influence of pH on the reaction rate
was investigated at different pH ranges but at
constant [substrate], [Mn(III)] and tempera-
ture. The values of kobs were found to be
independent of pH (over the range of pH
4.22–6.45). The reaction may be described by
the following rate expression:
d[Mn(III)]
−
=k[S][Mn(III)]
(2)
dt
Fig. 3. Dependence of temperature on second-order rate
constants for the oxidation of substrate by manganese(III).
Fig. 1. Variation of the reaction rate with the [substrate].
Plots of log (k2/T) vs 1/T. (a)
D
-Glucose, (b)
-arabinose, (f)
-erythrose, (h) DL-glyceraldehyde, (i) -glucosamine hy-
-mannosamine hydrochloride (k) -galac-
tosamine hydrochloride.
D
-mannose, (c)
-ribose, (g)
Plots of kobs vs [substrate] at pH 6.1, [Mn(III)]=2.0×10−4
D
-galactose, (d)
D
-xylose, (e)
L
D
mol dm−3 and T=318 K. (“)
D
-Glucose; (ꢀ)
-erythrose; (ꢂ) DL-glyceraldehyde; (ꢃ) -glucosamine hy-
drochloride.
D
-ribose; (ꢁ)
D
D
D
D
drochloride (j)
D
D

74
K.K. Sen Gupta, A.B. Bilkis / Carbohydrate Research 315 (1999) 70–75
Table 2
370.3 K. The value of i, which is higher than
T (318 K) (T=mid-point of the experimen-
tally used range of temperatures), suggests
that enthalpy is a controlling factor [26].
It has been shown [18,27] that cyclic pyra-
noid forms of hexoses and pentoses are in-
volved in their oxidations with the metal ions.
Values of activation parameters of the oxidations of some
aldoses and amino sugars
Substrate
DH"
DS"
(kJ mol−1
)
(J K−1 mol−1
)
D
D
D
D
-Glucose
-Mannose
-Galactose
-Xylose
8394
7494
7194
6694
6294
4892
3592
2692
8794
−52913
−79913
−86913
−99913
−108913
−14997
−18697
−20097
−45913
4
The pyranoid form in the C1 conformation
preponderates [28] in aqueous solution of
hexose and the b anomer where the 1-OH
equatorial is attacked by the oxidants [29].
However, the b anomer of a pentose exists as
L
-Arabinose
D
-Ribose
D
-Erythrose
DL-Glyceraldehyde
4
1
an equilibrium mixture of C1 and C4 forms,
of which the former preponderates. These al-
doses are believed to be preferentially oxidized
D
D
D
-Glucosamine
hydrochloride
-Mannosamine
hydrochloride
-Galactosamine
hydrochloride
7694
7394
−74913
−82913
4
in the C1 b pyranoid form, and hence, initial
product should be the corresponding lactone.
In aqueous solution, the composition of
D
-
erythrose is 25% a-furanose, 63% b-furanose,
ꢁ2% aldehyde, and ꢁ10% aldehydrol [30].
tions of anion of the aldoses by osmium(VIII)
in alkaline medium [23c], justify the experi-
mental observation. The enthalpy of activa-
tion is linearly related [24] to the entropy of
activation (r=0.9988) as plotted in Fig. 4 and
the isokinetic temperature is 369.2 K. The
isokinetic behavior is further supported [25]
by the linear plot of log k% vs log k2 (r=
0.9982), where k% and k2 are 2the second-order
rate constants at2the temperatures 323 and 313
K, respectively (Fig. 4). The isokinetic temper-
ature, i, calculated from an Exner plot, is
Thus
D-erythrose exists predominantly in the
furanoid form as the b anomer. Glyceralde-
hyde, the simplest sugar, consists mainly of
the aldehydrol in dilute solution; in the crys-
talline state (DL) or syrupy state ( ), it is
D
dimeric [31a]. Thus, as the length of the chain
of the aldehydo form increases, the extent of
hydration decreases [31b]. Again the pKa val-
ues of some aldoses are known [32]. The val-
ues are 12.35, 12.29, 12.41, and 12.21 for
D-glucose,
D-xylose,
L
-arabinose and -ribose,
D
respectively. Consequently, the aldoses will re-
main in the undissociated form in the pH
range 4.22–6.45.
The pseudo-first-order rate constants for
the oxidations of both methyl a-
D
-glucopyra-
noside and methyl b-
D
-glucopyranoside have
been determined. The values of kobs obtained
for the oxidations of the two methyl glu-
cosides are 8.5×10−6 and 12.8×10−6 s−1
for methyl a- and methyl b- -glucopyra-
D
nosides, respectively, at [substrate]=3.0×
10−1 mol dm−3, [Mn (III)]=2×10−4 mol
dm−3, pH 6.1 and T=328 K. The oxidations
of methyl a-
D
-glucopyranoside and methyl b-
D
-glucopyranoside are too slow to be studied
Fig. 4. Isokinetic plots for the oxidations of aldoses and
amino sugars by tris[manganese(III)] in buffer medium.
(a, a%), plot of DS" vs DH"; (b, b%), plot of log k% vs log k2.
under the conditions at which other substrates
are oxidized. Consequently, C-1 of the aldoses
appears to be more reactive [33] than C-6. The
reaction proceeds through the intermediate
formation of a free radical between the reac-
(1)
(5)
D
-Glucose, (2)
-arabinose, (6)
-glucosamine hydrochloride (10)
nosamine hydrochloride (11) -galactosamine hydrochloride.
D
-mannose, (3)
-ribose, (7)
D
-galactose, (42)
D-xylose,
L
D
D
-erythrose, (8) DL-Glycer-
-man-
aldehyde, (9)
D
D
D

K.K. Sen Gupta, A.B. Bilkis / Carbohydrate Research 315 (1999) 70–75
75
[2] (a) J.A. Kirby, A.S. Robertson, J.P. Smith, A.C. Thomp-
son, S.R. Cooper, P. Klein, J. Am. Chem. Soc., 103
(1981) 5529–5537. (b) J.C. Depaula, G.W. Brudvig, J.
Am. Chem. Soc., 107 (1985) 2643–2648.
[3] F.A. Cotton, G. Wilkinson, Ad6anced Inorganic Chem-
istry—A Comprehensi6e Text, 4th ed., Wiley, New York,
1980, p. 74.
[4] M.N. Bhattacharjee, M.K. Chaudhuri, H.S. Das Gupta,
D.T. Khathing, J. Chem. Soc., Dalton Trans., (1981)
2587–2588.
[5] G. Davies, Coord. Chem. Re6., 4 (1969) 199–201.
[6] K. Ramakrishnan, K.R. Shankaran, V.S. Srinivasan,
Indian J. Chem., 29A (1990) 843–846.
[7] K.I. Bhat, B.S. Sherigara, Transit. Met. Chem. (London),
19 (1994) 178–181.
[8] K.I. Bhat, B.S. Sherigara, I. Pinto, Indian J. Chem., 33A
(1994) 42–46.
[9] R. Varadarajan, M. Joseph, Indian J. Chem., 19A (1980)
977–979.
[10] B.S. Sherigara, K.I. Bhat, I. Pinto, N.M. Made Gowada,
Int. J. Chem. Kinet., 27 (1995) 675–690.
[11] S. Chandraju, B.S. Sherigara, N.M. Made Gowada, Int.
J. Chem. Kinet., 26 (1994) 1105–1119.
[12] K.S. Rangappa, S. Chandraju, D.S. Mahadevappa,
Transit. Met. Chem. (London), 21 (1996) 519–523.
[13] G.V. Bakore, M.S. Bararia, Z. Phys. Chem., 229 (1965)
245–249.
[14] S.P. Ghosh, P.K. Ray, T.K. Bandyopadhyay, A.K. Deb,
Z. Naturforsch, 36b (1981) 1270–1272.
Scheme 1.
[15] M.M. Ray, J.N. Adhya, D. Biswas, S.N. Poddar, Aust.
J. Chem., 19 (1966) 1737–1740.
[16] B.N. Figgis, C.L. Raston, R.P. Sharma, A.H. White,
Aust. J. Chem., 31 (1978) 2545–2548.
[17] R.C. Weast, CRC Handbook of Chemistry and Physics,
66th ed., CRC Press, Boca Raton, FL, 1986, p. 162.
[18] K.K. Sen Gupta, S. Sen Gupta, S.N. Basu, Carbohydr.
Res., 71 (1979) 75–84.
[19] W. Pigman, The Carbohydrates: Chemistry, Biochemistry,
Physiology, Academic Press, New York, 1957, p. 472.
[20] M. Abdel Akher, F. Smith, J. Am. Chem. Soc., 73 (1951)
5859–5860.
[21] B. Capon, Chem. Re6., 69 (1969) 407–498.
[22] (a) F. Feigl, V. Anger, R. Oesper, Spot Tests in Organic
Analysis, 7th ed., Elsevier, New York, 1966, p 387. (b) S.
Kundu, A.K. Bhattacharya, R. Banerjee, J. Chem. Soc.,
Dalton Trans., (1996) 3951–3957.
tants in the rate-determining step. The free
radical is then rapidly oxidized by another
[Mn(III)] to give products. The different steps
of the reactions involving the hydrated form
of the acyclic aldoses and manganese(III) are
shown in Scheme 1. Irrespective of whether
the substrate is in the cyclic or acyclic form,
the aldonic acids are converted to their anions
at pH 6.1. The observed order with respect to
each reactant and the formation of polymeric
suspension in the presence of acrylamide cor-
roborate the suggested mechanism.
[23] (a) K.K. Sen Gupta, A.B. Bilkis, B. Pal, Carbohydr. Res.,
309 (1998) 303–310. (b) K.J. Laidler, Chemical Kinetics,
2nd ed., Tata McGraw-Hill, New Delhi, 1965, p 235. (c)
K.K. Sen Gupta, A.B. Bilkis, Int. J. Chem. Kinet., in
press.
[24] J.E. Leffler, J. Org. Chem., 20 (1955) 1202–1231.
[25] O. Exner, Nature, 201 (1964) 488–490.
[26] T.A. Iyengar, D.S. Mahadevappa, J. Carbohydr. Chem.,
11 (1992) 37–58.
Acknowledgements
One of the authors (Bilkis Ara Begum) is
grateful to the Government of India for an
ICCR fellowship and to the authorities of the
Bangladesh Atomic Energy Commission for
granting her leave.
[27] K.K. Sen Gupta, S.N. Basu, Carbohydr. Res., 72 (1979)
139–149.
[28] M. Rudram, D.F. Shaw, J. Chem. Soc., (1965) 52–57.
[29] R. Bentley, J. Am. Chem. Soc., 79 (1957) 1720–1725.
[30] R.S. Tipson, D. Horton, Ad6. Carbohydr. Chem.
Biochem., 42 (1984) 36–37.
References
[31] (a) S.J. Angyal, R.G. Wheen, Aust. J. Chem., 33 (1980)
1001–1011. (b) R.S. Tipson, D. Horton, Ad6. Carbohydr.
Chem. Biochem., 42 (1984) 30–31.
[32] C. Degani, Carbohydr. Res., 18 (1971) 329–332.
[33] K.K. Sen Gupta, S. Sen Gupta, S.K. Mandal, A. Ma-
hapatra, J. Chem. Res. (S), (1990) 60–61.
[1] (a) J.L. Sheats, R.S. Czernuszewicz, G.C. Dismukes,
A.L. Rheingold, V. Petrouleas, J. Stubbe, W.H. Arm-
strong, R.H. Beer, S.J. Lippard, J. Am. Chem. Soc., 109
(1987) 1435–1444. (b) M. Koikawa, H. Okawa, J. Chem.
Soc., Dalton Trans., (1988) 641–645.
.