Electrocatalytic reactions of a manganese-pyrylium complex

Article abstract of DOI:10.1021/om020405n

A salt of a metal-pyrylium σ-complex, pentacarbonyl-3-(2-methoxy-5,6-diphenylpyrylium)manganese(I) tetrafluoroborate (1BF₄), has been prepared by the action of Mn(CO)₅BF₄ on methyl propiolate and diphenylacetylene. The pyrylium complex shows unusual resistance to electrophilic cleavage of the Mn - C bond by H⁺ and a low aptitude for migratory insertion reactions. The complex was studied by spectroscopic, voltammetric, and controlled potential electrolysis methods. 1BF₄ displays two ligand-based one-electron reductions (E₁° = -1.33 V, E_pc,2 = -2.41 V vs Fc), the first of which is Nernstian. Its reactivity is enhanced by one-electron reduction, and it undergoes subsequent dimerization in THF (k_273K = 285 M^-1 s^-1) and CH₂Cl₂ (k_273K = 260 M^-1 s^-1) and a CO dissociation equilibrium (K ≈ 1.4 × 10⁵, k = 0.7 s^-1). In the presence of triphenylphosphine, a slow electrocatalytic CO-substitution reaction occurs for which a mechanism and rate constants have been determined by digital simulation. When 1BF₄ is reduced in the presence of HSnPh₃ and a proton donor, another electrocatalytic process results, presumably forming H₂ at the electrode surface.

Full text of DOI:10.1021/om020405n

3434
Organometallics 2002, 21, 3434-3442
Electr oca ta lytic Rea ction s of a Ma n ga n ese-P yr yliu m
Com p lex
Michael J . Shaw* and J aime Mertz
Department of Chemistry, Box 1652, Southern Illinois University Edwardsville,
Edwardsville, Illinois 62026
Received May 21, 2002
A salt of a metal-pyrylium σ-complex, pentacarbonyl-3-(2-methoxy-5,6-diphenylpyrylium)-
manganese(I) tetrafluoroborate (1BF 4), has been prepared by the action of Mn(CO)5BF4 on
methyl propiolate and diphenylacetylene. The pyrylium complex shows unusual resistance
to electrophilic cleavage of the Mn-C bond by H+ and a low aptitude for migratory insertion
reactions. The complex was studied by spectroscopic, voltammetric, and controlled potential
electrolysis methods. 1BF 4 displays two ligand-based one-electron reductions (E1° ) -1.33
V, Epc,2 ) -2.41 V vs Fc), the first of which is Nernstian. Its reactivity is enhanced by one-
electron reduction, and it undergoes subsequent dimerization in THF (k273K ) 285 M-1 s-1
)
and CH2Cl2 (k273K ) 260 M-1 s-1) and a CO dissociation equilibrium (K ≈ 1.4 × 105, k ) 0.7
s-1). In the presence of triphenylphosphine, a slow electrocatalytic CO-substitution reaction
occurs for which a mechanism and rate constants have been determined by digital simulation.
When 1BF4 is reduced in the presence of HSnPh3 and a proton donor, another electrocatalytic
process results, presumably forming H2 at the electrode surface.
In tr od u ction
cently, reports that metal ions such as Mg2+ serve as
cofactors to bind a substrate molecule in place during
H-transfer from NADH analogues5 have captured our
attention.
Transition-metal complexes that promote the homo-
geneous electrocatalytic generation of hydrogen from
protic substrates have garnered increasing interest in
recent years.1,2 These systems are of interest because
of the need for technologies that produce energy sources
such as H2 on demand. The mechanism of H2 production
often involves reduction and subsequent protonation of
a metal complex to generate a metal-hydride.1 This
hydride ligand reacts with H+ to regenerate the precur-
sor complex and H2. Such reactivity has many parallels
to the models of the NAD/NADH system, which have
received considerable attention in recent years3 and
which have proven very amenable to electrochemical
study.4 Both metal-hydrides and NADH serve as H-
donors to electrophilic substrates other than H+. Re-
A metal complex designed so that a transition-metal
atom is in close proximity to the active site of an NADH
analogue might display cooperative reactivity between
the two sites.6 We have recently prepared several
metal-pyrylium complexes that fit this description.7
Like pyridinium ions, many pyrylium ions can accept
H- equivalents at the para-position.8 The resulting 4H-
pyrans are known to be hydride donors.9 Pyrylium ions
are also easier to reduce than the corresponding pyri-
dinium ions, owing to the greater electronegativity of
the oxygen atom. This property should make catalytic
activity available at more benign potentials.10 However,
cationic hydrocarbyl groups such as pyridinium or
pyrylium bound to the metal via a M-C σ-bond are very
* Corresponding author. E-mail: michsha@siue.edu. Fax: 618 650-
3556.
(1) Divisek, J .; Schmitz, H.; Steffen, B. Electrochim. Acta 1994, 39,
1723. (b) Bhugun, I.; Lexa, D.; Save´ant, J . M. J . Am. Chem. Soc. 1996,
118, 3982. (c) Grass, V. Lexa, D.; Save´ant, J .-M. J . Am. Chem. Soc.
1997, 119, 7526, and references therein.
(2) (a) Zassinovich, G.; Mestroni, G.; Gladiali, S. Chem. Rev. 1992,
92, 1051. (b) Fache, F.; Schulz, E.; Tommasino, M. L.; Lemaire, M.
Chem. Rev. 2000, 100, 2159. (c) Bullock, R. M.; Voges, M. H. J . Am.
Chem. Soc. 2000, 122, 12594. (d) Luan, L. Song, J .-S.; Bullock, R. M.
J . Org. Chem. 1995, 60, 7170. (e) Troutman, M. V.; Appella, D. H.;
Buchwald, S. L. J . Am. Chem. Soc. 1999, 121, 4916.
(5) (a) Ohno, A.; Oda, S.; Ishikawa, Y.; Yamazaki, N. J . Org. Chem.
2000, 65, 6381. (b) Cheng, J .-P.; Lu, Y.; Zhu, X.; Mu, L. J . Org. Chem.
1998, 63, 6108.
(6) Some intermolecular reactivity of this type is known; see: (a)
Lo, H. C.; Buriez, O.; Kerr, J . B.; Fish, R. H. Angew. Chem., Int. Ed.
1999, 38, 1429. (b) Rudler, H.; Audouin, M.; Parlier, A.; Martin-Vaca,
B.; Goumont, R.; Durand-Re´ville, R.; Vaissermann, J . J . Am. Chem.
Soc. 1996, 118, 12045-12058
(7) Shaw, M. J .; Light, S.; Mertz, J .; Ripperda, S. Manuscript in
preparation.
(3) (a) Eisner, U.; Kuthan, J . Chem. Rev. 1972, 72, 1. (b) Stout, D.
M.; Meyer, A. I. Chem. Rev. 1982, 82, 223. (c) Fukuzumi, S. Advances
in Electron-Transfer Chemistry; Mariano, P. S., Ed.; J AI Press:
Greenwich, CT, 1992; pp 67-175. (d) Pestovsky, O.; Bakac, A.;
Espenson, J . H. J . Am. Chem. Soc. 1998, 120, 13422. (e) Pestovsky,
O.; Bakac, A.; Espenson, J . H. Inorg. Chem. 1998, 37, 1616. (f)
Marcinek, A.; Adamus, J .; Huben, K.; Gebicki, J .; Bartczack, T. J .;
Bednarek, P.; Bally, T. J . Am. Chem. Soc. 2000, 122, 437.
(4) (a) Anne, A.; Fraoua, S.; Moiroux, J .; Save´ant, J .-M. J . Am. Chem.
Soc. 1996, 118, 3938. (b) Anne, A.; Fraoua, S.; Grass, V.; Moiroux, J .;
Save´ant, J .-M. J . Am. Chem. Soc. 1998, 120, 2951. (c) Mairanovskii,
S. G. Catalytic and Kinetic Waves in Polarography; Plenum: New York,
1968; pp 245-261.
(8) (a) Kuthan, J .; Sebek, P.; Bo¨hn, S. Adv. Heterocycl. Chem. 1995,
62, 19. (b) Ohkata, K.; Akiba, K.-Y. Adv. Heterocycl. Chem. 1996, 65,
283.
(9) Pozharskii, A. F. Heterocycles in Life & Society; An Introduction
to Heterocyclic Chemistry & Biochemistry & the Role of Heterocycles
in Science, Technology, Medicine & Agriculture: Wiley: New York,
1997; and references therein
(10) (a) Balaban, A. T.; Dinculescu, A.; Dorofeenko, G. N.; Fischer,
G.; Koblik, A. V.; Mezheritskii, V. V.; Schroth, W. Pyrylium Salts:
Syntheses Reactions and Physical Properties: Advances in Heterocyclic
Chemistry Supplement 2; Academic Press: New York, 1982. (b)
Farcasiu, D.; Balaban, A. T.; Bologa, U. L. Heterocycles 1994, 37, 1165.
(c) Ohkata, K.; Akiba, K.-Y. Adv. Heterocycl. Chem. 1996, 65, 283.
10.1021/om020405n CCC: $22.00 © 2002 American Chemical Society
Publication on Web 06/22/2002

Manganese-Pyrylium Complex
Organometallics, Vol. 21, No. 16, 2002 3435
rare.11,12 Despite the large number of pyrylium com-
pounds known, few σ-bonded organometallic pyrylium
complexes have been prepared and investigated.12,13
This oversight is surprising in light of the utility of
pyrylium salts in organic synthesis and their applica-
tions as dyes, laser dyes, photosensitizers, and corrosion-
inhibiting additives in paints.9,10 They also have a rich
redox chemistry, undergoing reduction to form stable
radicals, dimers, or pyrans, depending on steric factors
and on the reducing agent. The coupling of this redox
chemistry to the recognized properties of transition-
metal complexes has the potential to yield systems with
useful physical and chemical properties. For example,
such ligands would be expected to be less electron-rich
than neutral ligands and are expected to have properties
consistent with this relative lack of electron density.
In this paper, we outline the preparation of a new
organometallic pyrylium complex and its electrochem-
istry under conditions where migratory insertion and
electrophilic M-C bond cleavage is possible. We present
herein evidence that such cationic ligands are attached
to the metal via robust M-C σ-bonds which resist
electrophilic cleavage and migratory insertion reactions.
Furthermore, this compound displays the ability to
undergo two types of electrocatalytic reaction depending
on the reaction conditions, and we will demonstrate the
utility of square wave voltammetry to supplement the
information that is available by cyclic voltammetry in
this investigation.
first step of the reaction involves η2-coordination of the
methyl propiolate to the metal center, as originally
proposed by Rosenblum for the preparation of Fp-
lactone complexes.15 There have been recent reports of
other electrophile-mediated cyclization reactions to
prepare pyrylium salts.18
The physical properties of 1+ are consistent with the
positive charge being distributed around the pyrylium
ring. For example, 1+ has bands in its IR spectrum at
2135, 2084, 2040, and 2017 cm-1, which may be com-
pared to those of Mn(CO)5CH3 (2110, 2030 (shoulder),
2010, and 1986 cm-1) and Mn(CO)5I (2125, 2043, 2012,
and 1982 cm-1). The overall intensity pattern and the
shift to higher wavenumbers for 1+ indicate a similar
geometry at a slightly less electron-rich Mn center, as
expected for a species with a cationic hydrocarbyl ligand.
1
The H NMR spectra reveal signals due to the H atom
on the 4-position of the pyrylium ring at 8.45 ppm in
1+. This highly deshielded signal confirms that the
positive charge primarily resides on the pyrylium ring,
leaving a relatively uncharged metal center. The pyry-
lium ligand in 1+ also seems to be the primary redox
site, as expected from a comparison of pyrylium and Mn-
(CO)5R redox potentials (vide infra). This compound is
air stable in the solid state and in CDCl3 solution for
several days.
Resu lts a n d Discu ssion
(1) Syn th esis of P en ta ca r bon yl-3-(2-m eth oxy-5,6-
d ip h en ylp yr yliu m )m a n ga n ese(I) (1+BF 4). The abil-
ity of metal cations to condense organic compounds into
cyclic structures is well-documented.14 In particular, the
ability of Fp+ (Fp ) η5-C5H5(CO)2Fe) to condense
acetylenic esters with alkenes to form cyclic products
including six-membered lactone rings15 prompted Leg-
zdins to determine that [CpM(NO)2]+ salts (Cp′ ) Cp,
Cp*; M ) Cr, Mo, W) form the corresponding lactones
as the sole products.12 Using an alkyne in the place of
an alkene resulted in the formation of the pyrylium
complexes instead of lactones.
We have found that [Mn(CO)5]BF416,17 and [Mn(CO)5]-
PF6 also mediate the condensation of methyl propiolate
and diphenylacetylene to form an organometallic pyry-
lium complex (eq 1).
Further studies into which metal cations mediate this
process and the overall mechanism of this reaction are
underway, but it seems reasonable to suggest that the
The positive charge on the pyrylium ring has a strong
influence on the reactivity of 1+. Most Mn(CO)5R (R )
alkyl, aryl) complexes undergo CO-insertion reactions
when treated with external ligands and undergo im-
mediate Mn-C bond cleavage in the presence of H+.19
1H NMR experiments reveal that complex 1+, on the
other hand, is unaffected by equimolar HBF4 for at least
1
2 days in CDCl3. Similarly, the H NMR and 31P NMR
spectra of solutions that contain 1+ and 2 equiv of PPh3
undergo no change over a period of a week, conditions
where a neutral Mn(CO)5R complex would undergo
migratory insertion to form a phosphine-substituted acyl
complex. The robust Mn-pyrylium bond in 1+ can be
explained by the fact that the positively charged ligand
is unlikely to be a good nucleophile. Thus, its bond to
the metal center is not subject to electrophilic cleavage
by H+, nor does it display a migratory aptitude, since it
appears to be not nucleophilic enough to attack a
coordinated CO ligand.
Electr och em istr y of Com p lex 1+. In CH2Cl2 at 0
°C, complex 1+ displays two cathodic features by cyclic
voltammetry as shown in Figure 1. Similar electrochem-
istry is observed in THF. The first reduction occurs at
E°′ ) -1.33 V vs Fc, and its anodic return wave is
visible at scan rates >50 mV/s. This feature appears to
be due to a one-electron process with fast charge
transfer due to the 1+/0 couple as indicated by compari-
son of the ∆Ep values to those of the ferrocene standard
(e.g., both are 70 mV at 200 mV/s) and bulk coulometry
(11) Garnovskii, A. D.; Sadimenko, A. P. Adv. Heterocycl. Chem.
1998, 72, 1.
(12) Legzdins, P.; McNeil, W. S.; Vessey, E. G. Organometallics 1992,
11, 2718.
(13) Ohe, K.; Mike, K.; Yokoi, T.; Nishino, F.; Uemura, S. Organo-
metallics 2000, 19, 5525, and references therein.
(14) (a) Lautens, M. Klute, W.; Tam, W. Chem. Rev. 1996, 96, 49.
(b) Fruhauf, H.-W. Chem. Rev. 1997, 97, 523. (c) Cotton, F. A.;
Wilkinson, G.; Murillo, C. A.; Bochman. Advanced Inorganic Chemistry,
6th ed.; Wiley: New York, 1999, and references therein. (d) Collman,
J . P.; Hegedus, L. S.; Norton, J . R.; Finke, R. G. Principles and
Applications of Organotransition Metal Chemistry; University Science
Books: Mill Valley, 1987; and references therein.
(15) Rosenbloom M.; Scheck, D. Organometallics 1982, 1, 397.
(16) Reger, D. L.; Coleman J . Organomet. Chem. 1977, 131, 153.
(17) Chaffee, S. C.; Sutton, J . C.; Babbitt, C. S.; Mayer, J .; Guy, K.
A.; Pike, R. D.; Carpenter, G. B. Organometallics, 1998, 17, 5586.
(18) Tovar, J . D.; Swager, T. M. J . Org. Chem. 1999, 64, 6499.
(19) J ohnson, M. D. Acc. Chem. Res. 1978, 11, 57.

3436 Organometallics, Vol. 21, No. 16, 2002
Shaw and Mertz
Ta ble 2. Ra te Con sta n ts for Dim er iza tion of 10 in
THF
T (K)
rate constant (M-1 s-1
)
295
273
251
233
950
285
142
80
pyrylium group. It can also be argued that the one-
electron nature of the second redox process in 1+ is
inconsistent with the usually observed two-electron
behavior of Mn(CO)5X complexes, which produce the
well-known [Mn(CO)5]- anion upon reduction. The
presence of this anionic species is clearly evident in
scans of Mn(CO)5I as an anodic peak at -0.62 V vs Fc,
which appears following reduction of the starting iodide
complex. This wave is conspicuously absent from scans
of 1+, which means that little or no [Mn(CO)5]- is
produced under these conditions.
F igu r e 1. CV of 1.0 mM 1BF4 at 200 mV/s, in 0.15 M
NBu4PF6 in CH2Cl2 at 0 °C at a 1 mm Pt disk electrode.
Ta ble 1. Electr och em ica l Da ta for 1+ in CH2Cl2/
0.15 M NBu 4P F 6 a t 0 °C
The reduction of 1+ was studied in some detail in CH2-
Cl2 at 0 °C. When the method of Nicholson and Shain21
is used to extract first-order rate constants from CVs
recorded between 50 and 1600 mV/s, the k values
increase systematically with increasing scan rate for
both THF and CH2Cl2 by a total of 300%. Since dimer-
ization of pyranyl radicals is a well-known consequence
of the reduction of pyrylium ions, simulations were
performed using DIGISIM22 taking into account a
second-order follow-up reaction. This approach led to a
good match between the experimental and simulated
voltammograms when k1 ) 260 M-1 s-1. Under these
conditions, simulated reversibilities were within the
experimental reversibilities for each of the 5 scan rates
considered (50-1600 mV/s). At high concentrations, a
scan rate (mV/s) ∆Ep(1+/0) (mV) Ipa/Ipc (1+/0
)
Epc(10/-) (V)
50
100
200
400
800
82
84
82
91
102
122
0.46
0.53
0.62
0.72
0.81
0.87
-2.36
-2.38
-2.41
-2.45
-2.48
-2.53
1600
experiments (vide infra). The chemical reversibility (Ipa
/
Ipc) for this feature is over 0.9 in cyclic voltammograms
(CVs) recorded at scan rates in excess of 3200 mV at 0
°C, but at lower scan rates, the reduction is less
chemically reversible (Table 1). The reversibility im-
proves as the temperature is lowered, so that at -45
°C the first reduction appears completely Nernstian at
scan rates faster than 400 mV/s. At temperatures over
0 °C, the follow-up reaction for this reduction seems to
be primarily a relatively slow dimerization process, most
likely involving coupling of the neutral pyranyl radicals.
A second reduction due to the 10/- couple occurs at Epc
) -2.41 V vs Fc at 200 mV/s in CH2Cl2 (Table 1). It is
chemically irreversible at 3200 mV/s and lower scan
rates above 0 °C. This couple is the same height as the
1+/0 couple and is presumably due to another one-
electron process.
These two redox processes are most likely ligand
based. Pyrylium ions are known to undergo one-electron
reversible reductions in the range -0.6 to -1.29 V vs
Fc.10b This range is consistent with 1+ possessing a
rather electron-rich pyrylium ligand as the primary
redox site. It is also known that Mn(CO)5X (X ) alkyl,
aryl, trialkylstanyl, halide) complexes undergo two-
electron reductions to form the pentacarbonylmanga-
nese(-I) anion and that the reduction potentials show
little dependence on the identity of ligand X.20 In our
hands, the most easily reduced complex of this group,
namely, Mn(CO)5I, undergoes two closely spaced ir-
reversible reductions at -1.93 and -2.12 V vs Fc at 0°
in CH2Cl2. The IR data indicate a similar amount of
electron density on the metal center in 1+ and Mn-
(CO)5I, so it is unlikely that the metal’s redox potential
would change by 600 mV to more positive values
because of substitution by an iodide ligand by the
wave due to the dimerized product appears at Epc
)
-1.8 V vs Fc. This peak is barely visible in 1.0 mM
solutions (Figure 1) but can be clearly seen in 5.0 mM
solutions. This wave decreases in intensity with in-
creasing scan rate in a manner consistent with a second-
order reaction with the same rate constant as for the
process that diminishes the amount of 10 at the elec-
trode surface. It also diminishes when the temperature
is lowered and disappears upon addition of an H atom
source such as HSnPh3. The latter reagent indicates
that the radical dimerization process is slower than
atom-abstraction processes for 10.
The reduction of 1+ was also examined in THF over
a 67° temperature range. The reversibility of the 1+/0
couple at 20 and 0 °C was analyzed by the method of
Nicholson and again resulted in first-order rate con-
stants that were dependent on scan rate. Simulation
(with DIGISIM) of this EC process as a second-order
reaction results in good agreement (<6% difference)
between the observed and the simulated reversibilities
for scan rates where Ipa/Ipc < 0.85 (i.e., 50-800 mV/s at
22 °C). At 0° there is good agreement between the rate
constants found in THF (285 M-1 s-1) and CH2Cl2 (260
M-1 s-1). Furthermore, this wave is less chemically
reversible at higher concentrations, as expected for a
second-order process. Rate constants were found for
other temperatures (Table 2), but there also seems to
(21) Nicholson, R. S.; Shain, I. Anal. Chem. 1964, 36, 706. (b) Shain,
I. Anal. Chem. 1966, 38, 1406.
(22) Rudolph, M.; Reddy. D. P.; Feldberg, S. W. Anal. Chem. 1994,
66, 589A.
(20) (a) Dessy, R. E.; Stary, F. E.; King, R. B.; Waldrop, M. J . Am.
Chem. Soc. 1988, 88, 471. (b) Morris, M. D. Adv. Electroanal. Chem.
1974, 7, 80, and references therein.

Manganese-Pyrylium Complex
Organometallics, Vol. 21, No. 16, 2002 3437
be another process that diminishes the amount of 10
present at the electrode surface, which becomes rela-
tively important at low temperatures. The presence of
another process can be inferred from the curvature of
the van’t Hoff plot of these data. At high temperature,
the associative dimerization reaction is important, but
at lower temperatures, a CO dissociation may predomi-
nate (vide infra).
In both THF and CH2Cl2 a chemically irreversible
anodic peak is present after the first reduction wave is
scanned at slow rates (Figure 2). This feature appears
at Epa ) -0.39 V vs Fc (200 mV/s), and it diminishes in
intensity as the scan rate is raised, with the concomitant
increase in chemical reversibility of the 1+/0 couple. This
peak is ascribed to the oxidation of the dimer to re-form
1+. Electrolysis of a 0.60 mM solution of 1+BF4 with a
carbon electrode at -1.45 V vs Fc at -45 °C results in
the passage of 1 Faraday-equivalent of charge, which
confirms the one-electron nature of this process. After
electrolysis, the solution displays an irreversible ca-
thodic peak at -2.31 V vs Fc and the previously
observed irreversible anodic peak at -0.39 V vs Fc.
Scanning of this anodic peak results in enhancement
of the signal due to residual 1+ present, thereby indicat-
ing that the electrolysis product can be converted back
into the starting material on the CV time scale. Regen-
eration of the starting material following oxidation of
the product dimer is expected, given the known behavior
of other bulky dimeric pyrans and dihydropyridines.10
F igu r e 2. CV of 5.0 mM 1BF4 at 200 mV/s, in 0.10 M
NBu4PF6 in THF at 22 °C at a 1 mm Pt disk electrode.
NMR peaks at 2.8, 3.9, 7.1-8.2, 8.4, and 8.8 ppm, and
electrospray MS indicates a parent ion at 903 m/z, which
corresponds to C45H29O14Mn2. This formula can be
rationalized by loss of CO with the gain of an OH group.
Although attempts to crystallize this substance have
thus far been unsuccessful, it is clearly a product of a
dimerization reaction and may be due to an OH--
induced ring-opening reaction in which the acyclic
ligand displaces a carbonyl ligand from the metal.
On the basis of the second-order kinetics arising from
the CV investigation, the electrochemical and spectro-
scopic data of the product species generated by elec-
trolysis, the synthetic efforts, and the known reduction
chemistry of pyrylium ions, it is reasonable to conclude
that 1+ dimerizes upon reduction, through the sterically
accessible 2- and 4-position on the ring, as shown in eq
2.
Further characterization of the electrolytically gener-
ated product revealed it to consist of a mixture of species
resulting from 2,2′-, 2,4′-, and 4,4′-coupling of the
pyranyl radicals. The electrolysis product was extracted
with hexane from the supporting electrolyte after
removal of the CH2Cl2 under reduced pressure. This
solution displayed ν(CO) IR bands at 2125, 2069, 2045,
2033, 2008, 1964, and 1746 cm-1. Significantly, the IR
data do not match Mn2(CO)10 (ν(CO)’s at 2045, 2011,
1978 cm-1), which is the product of reduction of Mn-
(CO)5Br and Mn(CO)5I, a fact that lends further support
to the organic ligand as the site of electron transfer.
However, when the hexane was removed and this
1
sample was dissolved in CDCl3, the H NMR spectrum
revealed five significant peaks between 3.0 and 3.8 ppm.
1
From a consideration of the H signals expected for the
uncomplexed pyrans, four methoxy resonances are
expected for a mixture of the 2,2′-, 2,4′-, and 4,4′-coupled
products, and the final peak in this region is appropriate
for the 4-4′ CH group. Smaller peaks in this region may
indicate that some coupling also occurs through the
6-position. Additionally, a peak at 4.65 ppm and a peak
at 9.65 ppm are suggestive of the 2,4- and the 2,2-
coupled products, respectively, along with phenyl reso-
nances. The NMR results suggest that dimerization
occurs best through the sterically available 2- and
4-positions.
Electr och em istr y in th e P r esen ce of P P h 3. Be-
cause of the known migratory insertion chemistry of
RMn(CO)5 complexes, the redox behavior of 1+ in the
presence of external ligands was investigated. CO
substitution in this system fits a dissociative electro-
catalytic mechanism where (OC)4(PPh3)Mn(methoxy-
diphenylpyrylium), 3+, forms at the electrode surface.
No evidence for migratory insertion was found. This CO-
dissociation equilibrium is likely the other process that
complicates the kinetics of the EC dimerization of 1+
at low temperature.
In Figures 3-5 are displayed CVs and square wave
voltammograms (SWVs) recorded in the presence of
triphenylphosphine and simulations of these data.
These data were analyzed using ESP 2.423 because this
program can simulate SWV data in addition to CV data.
Synthetic attempts to prepare and separate these
isomers have been unsuccessful. These products react
with chromatographic supports such as silica gel. While
reduction of 1+PF6 with Zn in DME yields a similar
mixture of isomers as judged by IR and 1H NMR,
chromatography on silica gel affords a small amount of
an impure product. This substance had IR bands at
2122, 2068, 2027, 2000, 1951, 1750, and 1691 cm-1 and
(23) ESP 2.4 by Dr. Carlo Nervi is available at http://lem.ch.unito.it/
without charge for academic use.

3438 Organometallics, Vol. 21, No. 16, 2002
Shaw and Mertz
F igu r e 3. CVs of 1.3 mM 1BF4 at 50, 100, and 200 mV/s: (A) 1.3 mM PPh3; (B) 13 mM PPh3; (C) simulation, 1.3 mM
PPh3; (D) simulation, 13 mM PPh3.
The CV simulations thus obtained were in good agree-
ment with DIGISIM simulations using the same mech-
anism and parameters. A number of mechanisms were
considered, but only the following explained the gross
and subtle features of each CV and SWV recorded
regardless of the phosphine concentration. While the
theory of square wave voltammetry is well-developed24
and there are a number of tools to simulate SWV data,
its use in determining electrode mechanisms in orga-
nometallic chemistry has been sparse, even though it
offers complementary information to CV.
product’s potential would be more positive than that of
1+. Nucleophilic attack on the pyrylium ring could occur
at the 2-, 4-, or 6-position, but would result in the
formation of a pyran complex, which would be reduced
at potentials typical of Mn(CO)5R species, i.e., <-2 V
vs Fc as shown for the dimeric complex 2 discussed
earlier. If nucleophilic attack could occur on 10, then it
should be more facile upon 1+, a possibility ruled out
by earlier NMR experiments, thereby ruling out the
notion that nucleophilic attack on the ring is rate-
limiting.
A comparison of Figures 3a (1.3 mM PPh3) and 3b
(13 mM PPh3) shows that a 10-fold increase in the PPh3
concentration has very little effect on the CV data. A
similar effect can be seen in Figures 4a and 4b for the
SWV data. Coordination of PPh3 must therefore occur
after the rate-limiting step since the concentration of
PPh3 has only a minor effect on the amount of 3+/0
generated at the electrode surface. Since CO loss must
occur at some point in the substitution reaction, we
postulate that a slow CO loss from 10 must form a
coordinatively unsaturated intermediate, (OC)4MnR, 4,
which quickly reacts with PPh3 to form 30. The CV data
are otherwise quite insensitive to mechanism, and
several other schemes were found to account for the CV
data, but these failed to fit the four sets of SWV data.
1
+ + e- h 10
1° + 1° f 2
10 h 4 + CO
E° ) -1.33 V vs Fc
(3)
(4)
k1 ) 2.6 × 102 M-1 s-1
k2 ) 0.7 s-1, k-2 ) 1.0 × 104 M-1 s-1 (5)
4 + PPh3 f 30
k3 ) 1.0 × 103 M-1 s-1 (6)
3
+ + e- h 30
E° ) -1.41 V vs Fc
(7)
Reaction 4 is the dimerization reaction 2 that occurs
in the absence of added ligand. Since the rate of this
reaction is comparable to the rate at which 3+ is
generated, it must be included in the simulations. If it
is excluded, the simulated voltammograms display
higher reversibility than the experimental data at scan
rates fast enough so that no 3+/0 couple is visible (>400
mV/s). Other possible scenarios were also considered.
The rate-limiting step represented by eq 5 cannot be
CO insertion, because the resulting acyl group would
be electron-withdrawing from the pyrylium ring and the
To simulate the data in Figure 5, eq 5 has to be
reversible, with k-2 . k2. The value of k-2 has little
effect on the simulations of CVs and SWVs shown in
Figures 3 and 4. The data in Figure 5 allow for a more
precise estimation of this parameter because of their
longer time scale. These voltammograms were obtained
by holding the electrode at a potential sufficiently
negative to reduce 1+ for time τ and then rapidly
sweeping to positive potentials to analyze the composi-
tion of the “cloud” of products generated at the electrode
(24) (a) Osteryoung, J . Acc. Chem. Res. 1993, 26, 77. (b) O’Dea, J .
J .; Wikiel, K.; Osteryoung, J . J . Phys. Chem. 1990, 94, 3628.

Manganese-Pyrylium Complex
Organometallics, Vol. 21, No. 16, 2002 3439
F igu r e 4. SWVs (pulse width ) 5 mV, pulse height ) 50 mV) of 1.3 mM 1BF4 at 1, 5, and 10 Hz: (A) 1.3 mM PPh3; (B)
13 mM PPh3; (C) simulation, 1.3 mM PPh3; (D) simulation, 13 mM PPh3.
F igu r e 5. SWVs (pulse width ) 2 mV, pulse height ) 50 mV, 100 Hz) of 1.3 mM 1BF4 where electrode was held at -1.7
V vs Fc for time τ ) 1, 2, 4, and 8 s before scan. Arrows indicate trend in peak height with increasing τ: (A) 1.3 mM PPh3;
(B) 13 mM PPh3; (C) simulation, 1.3 mM PPh3; (D) simulation, 13 mM PPh3.
surface. A value of k-2 ) 1.0 × 104 M-1 s-1 ( 10% was
estimated from these data. This value is quite reason-
able for a process where a ligand attaches to a 16-
electron coordinatively unsaturated species.25 The equi-
librium constant K ) 7 × 10-5 for eq 5 can be estimated
from k2 and k-2. This reaction occurs independently of
whether PPh3 is present and must contribute to the
lower chemical reversibility of the 1+/0 couple observed
at slow scan rates at low temperatures and the curva-
ture of the van’t Hoff plot for the reduction of 1+
discussed earlier.
Another key observation from the SWV data is that
the sum of the cathodic currents due to 1+/0 and 3+/0
equal the current of 1+/0 in the absence of PPh3. This
effect can be seen most clearly in Figure 5. At longer
times, more 3 is generated, at the expense of 1 present

3440 Organometallics, Vol. 21, No. 16, 2002
Shaw and Mertz
at the electrode surface. This observation means that
3
+ must be generated by an electrocatalytic mechanism;
that is, 1+ is reduced to 10, which is quickly converted
to 3. At E°(1+/0), 30 is reoxidized at the electrode surface
to 3+, resulting in a net decrease in the cathodic current
observed for 1+/0. When the scan reaches E°(3+/0), the
3
+ at the electrode surface is rereduced. The electroca-
talysis for this reaction is fairly slow, being limited by
the rate of CO dissociation from 10. It should be noted
that this CO dissociation is 3700 times faster than the
CO dissociation observed for Mn(CO)5Br.26 The SWV
data allow for the estimation of the value of k3 as well.
The value of k-2 must be larger than k3 by about a factor
of 10; otherwise the simulated signal for 3+/0 in Figures
4c and 4d at 10 Hz is smaller than that at 5 Hz.
Similarly, the value of k3 has to be slightly larger than
k2; otherwise the 3+/0 signal is too small in all the CVs
and SWVs.
F igu r e 6. (A) CV of 5.0 mM 1BF4 at 800 mV/s; (B) in the
presence of 16.8 mM HSnPh3.
cal, and electrochemical properties confirm the assign-
ment made from the electrochemistry.
Electr och em istr y of 1+ in th e P r esen ce of HSn -
P h 3 a n d P r oton Don or s. Hydrogen atom abstraction
reactions are typical of radicals.28 It appears that 10 may
abstract H atoms from HSnPh3 (a known H atom
donor)29 presumably to become one or more pyran
complexes designated in eq 9 as “5”.
Overall, the mechanism described by eqs 3-7 fits the
data very well, as shown by the close match between
the data and the simulations shown in Figures 3-5.
While several other mechanisms were considered, they
produced simulations that only fit the data collected by
one technique (e.g., CV) at one PPh3 concentration, often
at only one scan rate. For example, a mechanism where
PPh3 substitution occurred via the dimer 2 was ruled
out because the reversibilities for the simulated CVs and
SWVs at low sweep rates were too high when the data
fit high sweep rates. The diffusion coefficients used for
these simulations were arbitrarily set at 1.0 × 10-5 cm2
s-1 since, as previously observed by Kochi for the
electrocatalytic mechanism, the diffusion coefficient has
only a minor effect on the shape of the voltammo-
grams.27 Indeed, when this parameter was varied by a
factor of 10 for any individual species, little change was
observed in the simulation thus obtained. Similarly, fast
electron-transfer reactions and transfer coefficients of
0.5 were assumed.
The identity of 3+ was probed during a bulk coulom-
etry experiment. A 1.0 mM solution of 1+BF4 was
electrolyzed at -1.2 V vs Fc in the presence of 1.3 mM
PPh3. A total of 0.07 Faraday-equivalents of charge was
passed through the solution, which changed from yellow
to dark orange during the exhaustive electrolysis. The
low charge count confirms the catalytic nature of the
ligand substitution process. After electrolysis, a single
reversible wave at -1.41 V vs Fc, due to 3+/0, was
observed and the solution displayed IR bands at 2070,
2024, 2000, and 1971 cm-1. The pattern is consistent
with a tetracarbonylmanganese complex which is more
electron rich than the starting material 1+.
Since pyrans are known to act as H- donors, mild
sources of H+ were added to the cell to see if catalytic
H2 production would occur. The loss of H- from a pyran
complex 5 would regenerate 1+, thereby closing a
catalytic cycle as described below.
1+ + e- h 10
E° ) -1.33 V vs Fc
10 + H• f 5
5 + H+ f 1+ + H2
(8)
(9)
(10)
Addition of excess HSnPh3 to solutions of 1+ drasti-
cally alters the appearance of the cyclic voltammograms
(Figure 6) in both CH2Cl2 and THF. The wave is no
longer reversible at low scan rates, and the peak
potential moves to slightly more positive values. This
behavior is consistent with an H atom abstraction
process, as is the fact that the wave due to the dimer 2
disappears when the tin hydride is added. If the scan
rates are increased or the temperature decreased, the
reduction feature becomes more reversible, which indi-
cates that this reaction is not much faster than the
dimerization reaction described in eq 4. At high scan
rates (>1600 mV/s) and low temperatures (<-20 °C),
the voltammograms are indistinguishable from those
recorded in the absence of HSnPh3.
When a scan is performed prior to and immediately
after the addition of water or a carboxylic acid to a THF
solution of 3+ and HSnPh3, a 20% increase in current
is observed. The wave also becomes more plateau-
shaped. These observations are consistent with mild
catalysis at the electrode surface. Unfortunately, sub-
sequent scans show that the complex decomposes under
these conditions, with the peak at Epc ) -1.30 V vs Fc
decreasing in intensity over time. Interestingly, the 1+/0
couple itself is unaffected by water, mesitol, 2-propanol,
or methanol in the absence of HSnPh3 on the CV time
3+PF6 was synthesized by the reaction of 1 equiv of
PPh3 with 1PF6 in DME in the presence of a catalytic
amount of Zn dust, and its spectroscopic, microanalyti-
(25) (a) Pearson, J .; Cooke, J . Takats, J . J ordan, R. B. J . Am. Chem.
Soc. 1998, 120, 1434, and references therein. (b) Boese, W. T.; Ford,
P. C. In Electron-Transfer Reactions; Isied, S. S., Ed.; American
Chemical Society: Washington, 1997; p 221. (c) Poli, R., Owens, B. E.;
Linck, R. G. J . Am. Chem. Soc. 1992, 114, 1302. (d) Peaver, K. A.;
Holl, M. M. B.; Carpenter, G. B.; Rieger, A. L.; Rieger, P. H.; Sweigart,
D. A. Organometallics 1995, 14, 512.
(26) Angelici, R. J .; Basolo, F. Inorg. Chem. 1963, 2, 728.
(27) (a) Zizelman, P. M.; Amotore, C. Kochi, J . K. J . Am. Chem. Soc.
1984, 106, 3771. (b) Astruc, D. Electron Transfer and Radical Processes
in Transition Metal Chemistry; VCH Publishers: New York, 1995;
Chapter 6.
(28) (a) Trogler W. C. Organometallic Radical Processes; Elsevier:
New York, 1990. (b) Iqbal, J .; Bhatia, B.; Nayyar, N. K. Chem. Rev.
1994, 94, 519.
(29) (a) Amatore, C. J . Org. Chem. 1996, 61, 9402, and references
therein. (b) Curran, D. P.; Hadida, S.; Kim, S.-Y.; Luo Z.; J . Am. Chem.
Soc. 1999, 121, 6607, and references therein.

Manganese-Pyrylium Complex
Organometallics, Vol. 21, No. 16, 2002 3441
distilled from sodium-benzophenone ketyl into a 500 mL
storage flask charged with fresh benzophenone ketyl. All
solvents were vacuum transferred directly into reaction ves-
sels. Methyl propiolate (Acros), diphenylacetylene (Acros),
tetrafluoroboric acid (Acros), cyclopentadienyldicarbonyliron
dimer (Aldrich), and dimanganese decacarbonyl (Alfa-Aesar)
were obtained commercially and were used as received. Mn-
scale. The catalytic increase in current also occurs when
chloroacetic acid or phenylacetic acid is added to the
solution, thereby indicating that alcohols are not acidic
enough to be H+ donors to the pyran complex.
Preparative attempts to characterize the electrode
products have been unsuccessful. Bulk electrolysis of a
sample of 1+ at a Pt electrode in the presence of excess
HSnPh3 results in the passage of 1-Faraday equivalent
of charge, and the neutral products can be extracted
from the supporting electrolyte and analyzed by 1H
NMR. This approach indicates that a plethora of prod-
ucts are formed as judged by the 20 or so resonances in
the region appropriate for the methoxy group. It is not
clear whether these final products result from decom-
position of a small number of species initially generated
at the electrode. Reaction of 1+ with HSnPh3 or HSnBu
with Zn in DME yields results similar to electrolytic
preparation. Bulk electrolysis of 1+ in the presence of
proton donors at a Hg-pool electrode also leads to
complete consumption of the complex with the passage
of ca. 1-Faraday equivalent of charge. No gas evolution
was observed. While the color changed from yellow to
dark orange during electrolysis, the product solution
was devoid of IR bands due to metal-carbonyl-containing
species.
30
(CO)5CH3 and NBu4PF6 were prepared as detailed in the
literature. NMR spectra were obtained with a 300 MHz Varian
300 Unity Plus spectrometer. IR spectra were obtained using
a Mattson Research Series IR spectrophotometer. NMR spec-
tra were obtained on samples prepared in a drybox under an
anhydrous atmosphere of argon.
General procedures used for electrochemistry have been
described elsewhere. Electrochemical measurements were
performed using an EG&G PAR 263 potentiostat interfaced
to a computer. Electrodes and cells were obtained from BAS,
and a vacuum cell was custom-built by ChemGlass. Bulk
electrolysis was performed using a BAS bulk electrolysis cell
contained within the vacuum cell. The working electrode was
made of reticulated vitreous carbon. Simulations were per-
formed using DIGISIM software available from BAS or ESP
2.4.24
Syn th esis of 1BF 4. Mn(CO)5CH3 (0.285 g, 1.35 mmol) was
dissolved in CH2Cl2 (20 mL), and HBF4‚Et2O (54%, 1.86 mL,
1.35 mmol) was added by syringe. The mixture was covered
and stirred in the dark for 3 h at 20 °C. Methyl propiolate
(0.121 mL, 1.35 mmol) was added by syringe and the mixture
stirred in the dark for 10 min. Diphenylacetylene (0.242 g, 1.1
mmol) was added, and the mixture was stirred in the dark
for 48 h. The solution was filtered through Celite supported
on a porous medium glass frit and concentrated under reduced
pressure until the solution volume was ca. 5 mL. About 1 mL
of Et2O was added and the solution cooled to -30 °C overnight.
The yellow crystals that formed were isolated by cannulation.
Two further crops were obtained for a total yield of 0.15 g
(15%). Anal. Calcd: C, 50.77; H, 2.59. Found: C, 51.01; H, 2.61.
1H NMR (CDCl3) δ: 8.45 (s, 1H, C5H(OCH3)Ph2); 7.65-7.28
(m, 10H, Ph2) 4.62 (s, 3H, OCH3). 13C NMR (CDCl3) δ: 60.05
ppm (OCH3). IR: ν (CO) 2135, 2084, 2040, 2017 cm-1, ν-
(pyrylium CO) 1630 cm-1. E°′(CH2Cl2): -1.33 V vs Fc. FAB-
MS: 457 m/z (P+ - BF4).
Syn th esis of 1P F 6. Mn(CO)5Br (1.05 g, 3.82 mmol) was
dissolved in a CH2Cl2 (25 mL) solution of AgPF6 (0.963 g, 3.81
mmol). A yellow precipitate immediately formed, and the
mixture was stirred for 30 min at 20 °C. Methyl propiolate
(0.56 g mL, 6.66 mmol) was added by syringe and the mixture
stirred in the dark for 10 min, whereupon the color of the
mixture lightened slightly. Diphenylacetylene (0.68 g, 3.81
mmol) was added, and the mixture was stirred in the dark
for 48 h. The solution was filtered through Celite supported
on a porous medium glass frit, Et2O (10 mL) was added, and
the solution was concentrated under reduced pressure until
its volume was ca. 10 mL. The yellow crystals that formed
upon cooling to -30 °C overnight were isolated by cannulation.
Two further crops were obtained for a total yield of 0.69 g
(30%). Anal. Calcd: C, 45.86; H, 2.34. Found: C, 45.92; H, 2.29.
1H NMR(CDCl3) δ: 8.45 (s, 1H, C5H(OCH3)Ph2); 7.65-7.28 (m,
10H, Ph2) 4.62 (s, 3H, OCH3). 13C NMR (CDCl3) δ: 60.05 ppm
(OCH3). IR: ν (CO) 2135, 2084, 2040, 2017 cm-1, ν(pyrylium
CO) 1631 cm-1. E°′(CH2Cl2): -1.33 V vs Fc.
Con clu sion s
Complex 1+BF4 is a surprisingly robust compound,
which resists electrophilic Mn-C bond cleavage by H+
and CO-insertion reactions, but has a rich electrochem-
istry, which results from the union of a metal complex
with a pyrylium ligand. In the absence of coordinating
ligands, the complex slowly dimerizes after reduction
through the coupling of pyranyl radicals. This coupling
produces several isomeric products. After reduction, the
complex also appears to undergo a dissociative equilib-
rium involving loss of CO 3700 times faster than that
observed for Mn(CO)5Br. This equilibrium heavily fa-
vors the starting material. The coordinatively unsatur-
ated intermediate 4 is rapidly trapped by both CO and
PPh3 to regenerate the starting material or complex 3,
respectively. The CO-dissociation equilibrium constant
and the rate constants for the coordination of these two
ligands to 4 were estimated through simulations of
cyclic voltammetric and square wave voltammetric data,
both data sets being necessary to completely discern the
kinetics of this reaction. The CO-dissociation process
ultimately controls the rate of an electrocatalytic ligand
substitution reaction.
On the CV time scale, complex 1+ appears to be
involved in a catalysis which involves HSnPh3 and weak
acids, presumably to generate H2. Unfortunately, the
electrode products decompose during bulk electrolysis
in the presence of protic species, precluding attempts
to quantify this electrocatalytic reaction on a prepara-
tive scale.
Syn th esis of 3P F 6. 1PF6 (0.088 g, 0.15 mmol) was reacted
in DME (10 mL) with PPh3 (0.038, 0.015 mmol) in the presence
of a trace of Zn dust. The yellow solution changed to orange
upon stirring overnight. The solvent was removed under
reduced pressure and the residue recrystallized from CHCl3/
Et2O. A total of 0.091 g of yellow needles was recovered (74%
yield). Anal. Calcd: C, 57.43; H, 3.49. Found: C, 57.33; H, 3.39.
1H NMR(CDCl3) δ: 7.72 (s, 1H, C5H(OCH3)Ph2); 7.00-7.55 (m,
Exp er im en ta l Section
All reactions and subsequent manipulations involving or-
ganometallic reagents were performed under anaerobic and
anhydrous conditions in an atmosphere of prepurified dini-
trogen or argon. Diethyl ether, hexanes, dichloromethane, and
chloroform-d were purified by distillation from CaH2 into 500
mL storage vessels charged with additional CaH2. THF was
(30) McKinney, R. J .; Crawford, S. S. Inorg. Synth. 1989, 26, 155.

3442 Organometallics, Vol. 21, No. 16, 2002
Shaw and Mertz
25H, Ph,) 4.46 (s, 3H, OCH3). 13C NMR (CDCl3) δ: 60.02 ppm
(OCH3). IR: ν (CO) 2084, 2084, 2017, 1993 cm-1, ν(pyrylium
CO) 1686 cm-1. E°′(CH2Cl2): -1.43 V vs Fc. Electrospray MS:
ment of Chemistry, the College of Arts and Sciences,
and the Graduate School of Southern Illinois University
Edwardsville, for their continuing financial support.
Mass spectrometry was provided by the Washington
University Mass Spectrometry Resource, an NIH Re-
search Resource (Grant No. P41RR00954).
691 (P+ - PF6), 663 (P+ - PF6 - CO), 635 m/z (P+ - PF6
-
2CO).
Ack n ow led gm en t. This research was supported by
a Cottrell College Science Award from Research Cor-
poration (CC5264). We also acknowledge the Depart-
OM020405N