Inorganica Chimica Acta, 18 (1976) 209-213 209

0 Elsevier Sequoia S.A., Lausanne - Printed in Switzerland

Kinetics of Decomposition of a Molybdenum(I) Dinitrogen Complex

C. M. ELSON*
A.R.C. Unit of Nitrogen Fixation, University of Sussex, Brighton NBl 9QJ, Sussex, U.K.
Received July 17, 1975

Thecomplex, trans-[Mo(Nz)z(Ph,PCH,CH,PPH2),]+; Experimental

decomposes with the loss of both Nz ligands in tetra-
hydrofuran-methanol (26% V/V)-0.1 N LiX electrolyte The molybdenum(O) dinitrogen complex was pre-
(X = Ct and CEO,-) via a first order dissociative mecha- pared by Dr. R.L. Richards. All solvents and lithium
nism. The rate constant of decomposition at 25” C in salts were dried before use. The electrolyte employed
the presence of lithium chloride is 2.4 ?I 0.2 x l&’ s-’ in this study was tetrahydrofuran_methanol(26% V/V)
and the activation parameters are: E, = 24 + 2 Kcal containing 0.1 N lithium chloride or perchlorate. All
mot’ and AS* = +I 9 3~ 5 cal deg-’ mot’. The kinetic solutions were approximately 1mM in complex and
parameters for the perchlorate electrolyte are similar. dissolved air was removed by purging with high purity
The molybdenum-containing product of decomposition argon.
disproportionates to produce MO(O) and MO(H) species The working electrode was either a vitreous carbon
which further react, in some cases, with the unions of electrode (0.46 cm’) for cyclic voltammetry, or a large
the electrolyte. All reactions are initiated and monitor- vitreous carbon plate (14.1 cm”) for controlled
ed by electrochemical techniques. potential electrolyses. The counter electrode was a
piece of platinum foil separated from the sample com-
partment by a fritted glass disc. The reference electrode
Introduction was a saturated potassium chloride-calomel electrode
(S.C.E.) separated from the sample by a glass frit and
Recently interest has been focused on the reactions of a salt bridge which contained 0.4M tetrabutylammo-
trans - bisdinitrogenbis (1,2 - diphenylphosphinoethane) nium iodide in methanol-water (1: 1).
molybdenum(O), trans-[Mo(N&dpez] (A), with or- The potential of the working electrode was controlled
ganic and hydrogen halidesl, sulfuric acid’ and on its relative to the S.C.E. by a Chemical Electronics TR
interaction with iron-sulfur cluster compounds3. In our 70/2A Potentiostat. Potential sweeps and steps were
studies of the redox behaviour and X-ray photoelectron generated by a Chemical Electronics RBI Waveform
spectra of the dinitrogen complexes of tungsten and Generator. Sweep rates ranged between 30 and 500 mV
molybdenum and analogues4, we confirmed that A s-i and the chronoamperograms used pluses of 1,3 and
could be reversibly oxidized to trans-[Mo(N&dpe$, 10 s. The current-potential curves were displayed on a
(A+)‘. Furthermore, we established6 that A did not Bryans X-Y recorder (model 26000 A3).
interact with the ferredoxin-model compound, [NBu”& The elevated pressure studies were performed in a
[Fe&(SEt),] (where Et = C,H5 and Bu” = nC,H,), one-litre rocking autoclave (Baskerville and Lindsay,
nor did acidification of a mixture of the two compounds Manchester) which was connected to cylinders of high
produce ammonia. This last reaction did demonstrate purity gases. Electrical contact to the interior of the
that iron(II1) chloride, derived from the reaction of autoclave was achieved by passing wires through a
HCl with the iron-sulfur cluster, oxidised A and under PTFE gland. The electrolytic cell used in conjunction
heterogeneous conditions yielded the salt, trans-[Mo with the autoclave consisted of a three-necked round
(N&dpeZ][FeCb]. However, the tetrachloroferrate mo- bottom flask which was wrapped in cotton wool to
lybdenum(1) compound was never prepared pure and provide thermal insulation. The three electrodes, work-
was always contaminated with a non-nitrogen contain- ing, counter and reference, were mounted inside the
ing species7 due to chemical decomposition. This report vessel. The sample was flushed three times at pressures
describes the kinetics of decomposition of trans-[Mo of 10 atm prior to the experiments.
(N&dpe$ in a tetrahydrofuran (thf)-methanol (Me
OH) solvent and the nature of the decomposition pro-
ducts. Results

* Present address: Department of Chemistry, Saint Mary’s The electrochemical techniques of cyclic voltammetry
University, Halifax B3H 3C3, Nova Scotia, Canada. and chronoamperometry were found to be the most
 -t
210 C. M. Elson

convenient methods of producing known quantities of

t-]Mo(N&dpe$, A+, and monitoring its disappear-
ance. The reaction scheme can be written:

r
truns-[Mo(Nz)zdpez]-e- P truns-[Mo(N&
4df (1)
truns-[Mo(N&dpe$ % bans-[MO(S),
dpe$ + 2N (2)
ia
where S is a solvent molecule derived from the electro-
lyte.
From the cyclic voltammograms, it was established
I
zspa

that A underwent a reversible one-electron diffusion-

controlled oxidation step’: the separation between the
anodic peak and half-peak potentials was 6.5 mV and
the separation between the anodic and cathodic peaks
was 67 mV (theoretical values = 59/n mV where n Figure 2. A double potential-step chronoamperogram of a 1
represents the number of electrons transferred); also mM solution of rrans-[Mo(N,)2dpez] between Ei = -0.45 V
the Malachesky R factor’ equalled 4.86 (theoretical and Er = -0.05 V fort = 3s at 22.5” C.
value = 4.92n”‘). The one exception to the predicted
behaviour was that the cathodic peak current was less ommetry (Figure 2). Initially a potential, Ei = -0.45V,
than the theoretical value which is characteristic of a was chosen at which there was no electrode reaction but
chemical reaction succeeding the electron transfer step’. which was in the diffusion controlled region for the
The presence of the following chemical reaction was reduction of the oxidation product, A’. The potential
demonstrated by recording cyclic voltammograms at was then stepped from Er to a value, Er = --O.OW, in
various temperatures (Figure 1). The effects of increas- the diffusion-controlled region of the oxidation of A
ing temperature were: 1) the oxidation peak current, and the oxidation current was recorded against time.
ip,, increased since the diffusion coefficient increased After a pre-selected interval, T, the potential was stepped
with temperature; 2) the anodic peak potential, Ep,, back to Ei and the oxidised material, A+, which had not
shifted due to differences in reference and junction undergone decomposition was reduced. The rate con-
potentials and 3) the cathodic peak height, ip,, decreas- stants were determined from the ratio of the corrected
ed due to the increased rate of transformation of At. cathodic to anodic currents at various times, t, and from
The rate constant of decomposition, k, was determin- a series of standard curves for various values of r. The
ed by the method of Schwarz and Shain” assuming results of the studies in both lithium chloride and
that the reaction was first or pseudo-first order. The perchlorate electrolytes are presented in Table I. Each
method employed double-potential step chronoamper- value was the result of approximately six determina-
tions and the standard deviations were of the same
14.2T
magnitude as found in other studies”. The values
were reproducible to within the standard deviations.
225 “C The activation energies, determined by applying the
method of least squares to standard Arrhenius plots

TABLE I. The Rate Constants for Decomposition of trans-

[Mo(N&dpe$ at Various Temperatures.

Electrolyte Temperature k x 10 (8’)

(“C)

thf-MeOH-LiCI 14.2 f 0.1 0.45 f 0.05

22.5 _+ 0.1 2.01 f 0.2
29.5 f 0.1 5.1, f 0.6
38.6 f 0.1 11.9 + 1.0

thf-MeOH-LiClO, 14.0 + 0.1 0.54 f 0.05

20.6 f 0.1 1.9, + 0.2
Figure 1. Cyclic voltammograms between -0.45 and -0.05 V 28.5 f 0.1 6.09 f 0.6
of a 1 mM solution of trurrs-[Mo(N2)2dpe2] at various tempera- 36.4 + 0.1 14., f 1.5
tures.
Molybdenum(Z) Dinitrogen Complex 211

TABLE II. The Rate Constants for Decomposition of trans-

[Mo(N&dpe$ at Various Pressures in thf-MeOH-LiC1.

Pressurising Pressure k x 10’ (s-r)

Gas (at@

N2 1+0 3.6 + 0.4

(T = 12°C) 40 f 2 6.3 f 0.6
80 + 2 15., -t 1.6
120 f 2 20.3 + 2.0

1+0 3.2 + 0.3

I (“T’= 11°C) 39 * 2 10.3 ?z 1.0
3.2 3.3 3A 3.5 81 f 2 17.a + 1.8
120* 2 22.3 + 2.1
103/T (“K-I’) 151 f2 27.2 f 2.5

Figure 3. An Arrhenius plot of In k versus l/T(‘K-‘) for de-

composition of trans-[Mo(N,)rdpe$ in thf-MeOH-LiCIO,.
1.0

(Figure 3) were 24 + 2 and 26 f 2 kcal moT’ for the log !!I

( h )
chloride and perchlorate electrolytes respectively. The
interpolated values of k at 25’ C were 2.4 + 0.2 x 10-l
s-I and 3.3 ? 0.3 x 10-l s-’ for the same electrolytes
0.5
which correspond to half-lives of reaction of 2.9 and
2.1 s respectively. The entropy of activation was cal-
culated from the rate constant at 25’ C using the theory
of Absolute Reaction Rates12. The AS* values obtained
were + 19 + 5 and + 26 +_ 7 cal deg-’ mar’ in the
chloride and perchlorate electrolyte respectively.
The cyclic voltammograms and chronoamperograms 60 160
at the various temperatures were always diffusion con- P (atm)
trolled. This implied that the chemical reaction follow-
ing the electrode reaction did not regenerate A. Further- Figure 4. A plot of log(k,/kr) against pressure (atm) for de-
more, it was found that the rate constant at a given composition of tram- [Mo(NZ)zdpez]+ in thf-MeOH-LiC1 under
temperature was independent of the concentration of A. Ar.
The role of solvent was less clearly defined since
restrictions on the composition of the electrolyte were non-linear (Figure 4) which precluded further cal-
imposed by the solubility and stability of A and the culations.
electrochemical method. In solvent mixtures which The product of the decomposition reaction was cha-
were greater than 50% V/V MeOH, the dinitrogen com- racterised by bulk-scale electrolyses of A. The oxida-
plex, A, slowly decomposed and in electrolytes where tions of A were performed at O.OV in both chloride
the MeOH concentration was less than 10% V/V the and perchlorate media. In the chloride electrolyte, the
cyclic voltammograms became distorted and poorly oxidation was not well behaved and very prolonged.
defined. However, using electrolytes which were 26% The resultant solution was reddish pink in colour dis-
and 52% MeOH, the rate of decomposition of A+ was playing adsorption bands at 356 and 534 nm and
found to be approximately 25% slower in the MeOH- exhibiting a reversible redox couple at Ep, = -0.16V
rich case. and Ep, = -0.lOV with both oxidised and reduced
The decomposition of A+ was also studied at elevated forms present. Both these features were characteristic
pressures by employing an autoclave. The results of of the couple, truns-[MoClzdpe2]/truns-[MoClzdpez]t,
the determination of the rate constant at various after allowance was made for the effect of different sol-
pressures of Nz and Ar are given in Table II. The tem- vents4v6. The product solution when A was oxidised in
peratures at which the two sets of data were collected perchlorate electrolyte was pale yellow and displayed a
were somewhat different since initial flusing of the reversible reduction wave at Ep, = -0.78V and Ep, =
autoclave vessel cooled the solutions to different extents. -0.72V and a further irreversible reduction wave at
Plots of the logarithm (base 10) of the ratio of the rate -1.29V. Recovery of the molybdenum product showed
constants at elevated pressure to the rate constant at that both N, ligands had been lost but that OH and
one atmosphere and passing through the origin were ClO, groups were present [IR bands (Nujol) at 1683
212 C. M. Elson

and 3500 cm-’ and at 1080 cm-‘]. When A was dis- The reaction was envisioned as proceeding in two
solved in thf-MeOH-LiC10, and allowed to stand for steps: the first step was rate determining and involved
several hours, a reversible oxidation wave appeared at the dissociation of one of the coordinated Nz groups;
Ep, = -0.71V and Ep, = -0.79V. Hence, the main subsequently, the second Nz ligand was very rapidly
products of the gross-scale oxidation of A in perchlorate replaced. Such a mechanism was supported by cyclic
media were truns-[Mo(S)zdpez]+ (where S is a mole- voltammetry which detected only truns-[Mo(NJz
cule of the electrolyte) and a higher oxidation state dpe$ and truns-[Mo(S)zdpez]+ between + 1.0 and
molybdenum species. -1 .OV. Moreover, a dissociative pathway was predicted
by the large positive entropies of activation and by the
decreased rate in the 52 % methanol electrolyte. Transi-
Discussion tion states which are less polar than the reactants are
implicated when the rate constant decreases as the
The decomposition reaction involved the loss of the dielectric constant of the medium increases13. Since
ligating N2 groups from trans-[Mo(N&dpe$, A+. coordinated dinitrogen has been found to be negatively
The lability of the dinitrogen groups was demonstrated charged when bound to rheniumI and molybdenum4,
by both the electrolytic oxidation of A in perchlorate an activated complex of reduced polarity was expected
media which produced truns-[Mo(S)2dpez]f and by when a molybdenum-dinitrogen bond was cleaved. By
heterogeneous ferric chloride oxidations of A in etha- comparison, attack of a solvent molecule (MeOH or
nol which yielded N, (0.2-1.0 mol/Mo) and truns- thf) in an associative mechanism would have preserved
[MoClzdpez][FeC14]6. George and Siebold have also or enhanced the polarity of the transition state.
reported that tr~ns-[Mo(N~)~dpe~]I~ decomposed in The controlled potential oxidations of A at O.OV
solution with loss of Nz. produced truns-[MoC12dpez] and [runs-[MoClzdpeJc
The electrochemical behaviour eliminated mecha- in chloride media and fruns-[Mo(S)zdpez]+ and a
nisms of decomposition involving disproportionation of higher oxidation state molybdenum complex in the
A+ or reduction of A+ by solvent since both these pro- perchlorate electrolyte. Therefore, it appeared that the
cesses would produce A near the electrode surface and initial product of the decomposition reaction, rrurans-
it was found that diffusion alone brought A to the [Mo(S),dpe2]+, disproportionated relatively slowly:
electrode. Furthermore, since the reaction obeyed
pseudo-first order kinetics and the rate constant was 2 truns-[Mo(S)2dpez]f + truns-[Mo(S)zdpe2]
independent of concentration of A, a dimerisation reac- + truns-[Mo(S)zdpez]2+ (3)
tion was ruled out.
Volumes of activation could not be calculated from In the lithium chloride electrolyte, the molybdenum(I1)
the results of the elevated pressure studies since the dication reacted with chloride ions to generate the
applied pressures were insufficient to produce a linear known complex, truns-[MoClzdpez]. It was unlikely
plot of log(k,/kl) versus pressure. However, since the that chloride reacted directly with truns-[Mo(S)zdpez]+
same results were obtained with nitrogen and argon since chloride was not involved in the decomposition
gases between O-150 atm, it was apparent that A+ step and, since reduction of truns-[MoCl,dpe,] did not
was not stabilised by increased pressure of Nz and that yield a stable MO(~) chloro complex but, instead,
the back reaction (reverse of 2) could be neglected in truns-[Mo(S)2dpez] was produced4. The dichloride
the kinetic treatment. complex was oxidisable at -0.13V and was, therefore,
The similarities in the activation energies and entro- converted under the electrolysis conditions to truns-
pies determined in chloride and perchlorate electrolytes [MoC12dpe2]+. Furthermore, the other disproportiona-
indicated that the same mechanism was operative in tion product, truns-[Mo(S)2dpe,] (Ep, = -0.72V), was
both media. Also, since the rate constants at 25°C in oxidised to regenerate truns-[Mo(S)zdpe2]+. Such a
the two electrolytes differed by only 30%, it was con- reaction sequence accounted for the prolonged electro-
cluded that the rate determining step did not involve lyses and the production of some truns-[MoCl,dpe,]
the attack of anions on A+. Therefore, the N2 ligands [FeC&] when trurzs-[Mo(N&dpeJ was oxidised under
were replaced by solvent molecules. It was not possible homogeneous condition8 by one equivalent of ferric
to distinguish whether thf or MeOH was the entering chloride. In the perchlorate media, truns-[Mo(S)2dpez]+
group since the rate constants determined in the 26% could have been formed either directly from the decom-
and 52% MeOH electrolytes differed by only 25 %. position of A+ or by oxidation of the disproportionation
However, the infrared spectrum of the products of the product, truns-[Mo(S)zdpe2]. However, since a mo-
bulk-scale oxidation of A in perchlorate media was lybdenum complex in an oxidation state greater than
consistent with the formulation, truns-[Mo(MeOH)l one was observed, disproportionation also occurred in
dpez][C104] which suggested that methanol was the the perchlorate electrolyte. The failure to produce
entering group but satisfactory elemental analysis could stable MO(I) species by reduction of truns-[MoC&dpe,]
not be obtained. further supported the disproportionation scheme.
Molybdenum(Z) Dinitrogen Complex 213

References 7 T.A. George and C.D. Siebold, Znorg. Chem., 12, 2548
(1973).
1 a) J. Chatt, G. A. Heath and R.L. Richards, Chem. Comm., 8 R.S. Nicholson and I. Shain, Analyt. Chem., 36, 706
444 (1972). b) ibid., 1010. (1964).
2 J. Chatt, A.J. Pearman and R.L. Richards, Nature, 253 9 P.A. Malachesky, Analyt. Chem., 41, 1493 (1969).
(5486), 39 (1975). 10 W. M. Schwarz and I. Shain, J. Phys. Chem., 69,30 (1965).
3 a) Chem. Eng. News, ~1.5 (Sept. 24, 1973). b) E.E. van 11 S. P. Perone and W. J. Kretlow, Analyt. Chem., 38, 1760
Tamelen, J.A. Gladys2 and C.R. Brulet, J. Am. Chem. (1966).
SOL, 96, 3020 (1974). 12 K. J. Laidler, “Reaction Kinetics”, Vol I, Pergamon Press,
4 J. Chatt, C. M. Elson and G. J. Leigh, unpublished results. 69 (1963).
5 T.A. George and CD. Siebold, J. Am. Chem. Sot., 94, 13 F. Basolo and R.G. Pearson, “Mechanisms of Inorganic
6859 (1972). Reactions”, 2nd ed., Wiley, New York, 1967, ~291.
6 J. Chatt, C.M. Elson and R.L. Richards, Chem. Comm., 14 J. Chatt, C.M. Elson, N.E. Hooper and G.J. Leigh, J.
189 (1974) plus unpublished results. Chem. Sot. Dalton Trans., 2392 (1975).