Homogeneous modeling of ammoxidation chemistry: Nitrile formation using a soluble analogue of MoO3

Article abstract of DOI:10.1039/a703790b

The reaction of the soluble MoO₃ analogue [Mo₆O₁₉]^2- with Ph₃P=NCH₂Ph replicates key features of heterogeneous ammoxidation chemistry by producing moderate yields of PhC=N through the propo

Full text of DOI:10.1039/a703790b

View Article Online / Journal Homepage / Table of Contents for this issue
Homogeneous modeling of ammoxidation chemistry: nitrile formation using a
soluble analogue of MoO3
Thomas R. Mohs, Yuhua Du, Bruce Plashko and Eric A. Maatta*
Department of Chemistry, Kansas State University, Manhattan, Kansas 66506, USA
The reaction of the soluble MoO3 analogue [Mo6O19]22 with
Ph3PNNCH2Ph replicates key features of heterogeneous
ammoxidation chemistry by producing moderate yields of
PhC·N through the proposed intermediacy of the benzyl-
allylimido- and benzylimido-hexamolybdates would provide
the closest approximation yet available of purported ammoxida-
tion surface species. We now report that the reaction of
[Mo6O19]22 with the benzylimido delivery reagent
Ph3PNNCH2Ph affords benzonitrile in moderate yield (37%)
thus replicating essential features of heterogeneous ammoxi-
dation chemistry through the proposed intermediacy of an
unstable benzylimido hexamolybdate system [NBu4]2-
[Mo6O18(NCH2Ph)], 1.
imido hexamolybdate complex [Mo6O18(NCH2Ph)]22
.
The SOHIO/BP process for the heterogeneous oxidation of
propene by O2 in the presence of ammonia to yield acrylonitrile
(i.e., the ‘ammoxidation’ of propene) constitutes the largest
volume example of an allylic oxidation in commercial practice.1
A minimalistic representation of this remarkable transformation
is provided in Scheme 1. It is now well established that the
bismuth oxide component of the catalyst serves to abstract a
hydrogen atom from propene, affording an allyl radical;2
subsequent C–N bond formation and redox events occur at sites
within the MoO3 component.3 Drawing on the extensive studies
of Grasselli and Burrington,4 key nitrogenous surface species
proposed to be involved in this chemistry include imido
{Mo·NH}, allylamido {MoNNHCH2CHNCH2}, allylimido
The addition of 1.47 mmol of Ph3PNNCH2Ph to an equimolar
amount of [NBu4]2[Mo6O19] in 30 ml MeCN at room temp.
produces a brown coloration within 1 min of mixing. After this
solution was heated at 82 °C under N2 for 72 h, analysis by GC–
MS techniques revealed the presence of benzonitrile (0.54
mmol, 37%) and N-benzylidenebenzylamine (0.25 mmol, 34%)
along with quantitative production of triphenylphosphine oxide.
Using an identical protocol, the thermolysis of Ph3PNNCH2Ph
in the absence of hexamolybdate produced only a trace of
benzonitrile (0.009 mmol, 0.6%) and a modest amount of
triphenylphosphine oxide (0.17 mmol, 11.7%); importantly, the
latter product signals the presence of residual water in the
acetonitrile solvent. 1H NMR monitoring throughout the
reaction between Ph3PNNCH2Ph and [Mo6O19]22 provided
only broad and uninformative spectra, inferring the presence of
paramagnetic species. In confirmation, aliquots withdrawn
from the reaction solution were analyzed by EPR spectroscopy
as frozen glasses and gave spectra identical to that published for
{Mo·NCH2CHNCH2},
and
allylideneamido
{MoNNNCHCHNCH2} fragments. Several groups have re-
produced aspects of the proposed transformations in studies of
various mononuclear systems.5–7
H
C
H
C
450 °C
+ NH3 + 3/2 O2
+
3 H2O
H2C
CH3
Bi2O3•n MoO3
H2C
C
N
the brown trianion [Mo6O19]32 11
.
Scheme 1
In order to test whether PhCHNNCH2Ph is an intermediate in
the formation of benzonitrile, a MeCN solution containing
equimolar quantities of the imine and the hexamolybdate was
treated and analyzed as above. No reaction occurred, suggesting
that the imine is formed on a reaction pathway different from
that leading to benzonitrile. Benzyl amine was suspected as the
imine precursor since various polyoxometalates are known to
effect this transformation.12 Accordingly, the reaction of 1.47
mmol amounts of PhCH2NH2 and [Mo6O19]22 in MeCN was
performed as described above and produced a brown solution
immediately; subsequent GC–MS analysis revealed the produc-
tion of PhCHNNCH2Ph (0.37 mmol, 46%) and only minor
amounts of PhCN (0.029 mmol, 1.8%). It is important to note
that the 0.25 mmol of PhCHNNCH2Ph produced in the reaction
between Ph3PNNCH2Ph and [Mo6O19]22 requires far more
benzyl amine (0.50 mmol) than can be accounted for by
hydrolytic decomposition of Ph3PNNCH2Ph alone (0.17 mmol);
this discrepancy can be accommodated by the reasonable
assumption that the reaction between Ph3PNNCH2Ph and
[Mo6O19]22 affords an intermediate which is more sensitive
than Ph3PNNCH2Ph toward hydrolytic release of PhCH2NH2.
Since an initial redox reaction between Ph3PNCH2Ph and the
hexamolybdate is unlikely,† and given the demonstrations of
forming complexes [Mo6O18(NR)]22 in reactions between the
hexamolybdate and Ph3PNNR reagents,10a,d we propose that our
observations are best accommodated by formation of an
In order to further refine these homogenous analogues of
proposed ammoxidation events we sought to transfer this
chemistry into a coordination environment which more closely
resembles that provided by bulk MoO3. As shown in Fig. 1, the
soluble hexamolybdate cluster [Mo6O19]22 possesses an MoO6
coordination sphere8 which is conspicuously similar to that
within MoO3.9 Realizing that a large number of organoimido-
substituted hexamolybdates recently have been accessed
through reactions of [Mo6O19]22 with various imido delivery
reagents,10 this structural correspondence suggested to us that
O
O
1.671 Å
1.677 Å
1.735 Å
1.948 Å
1.873 Å
O
1.904 Å
O
O
Mo
Mo
O
O
O
1.948 Å
O
O
2.251 Å
1.951 Å
1.982 Å
2.319 Å
2.332 Å
O
O
unstable benzylimido hexamolybdate [Mo6O18(NCH2Ph)]22
,
2-
avg. Mo environment in [Mo6O19
]
Mo environment in MoO3
1, which undergoes two modes of decomposition. In accord
with observations on related d0 benzylimido13,14 (and allyl-
imido6) complexes, 1 should be extremely moisture sensitive
22 8
Fig. 1 Comparison of the Mo environments in [Mo6O19
MoO3
]
and in
9
Chem. Commun., 1997
1707

View Article Online
2–
Footnotes and References
[Mo6O19
]
+
Ph3P NCH2Ph
* E-mail: eam@ksu.edu
Ph3P
O
–
† p-CH3C6H4NNPPh3 (Epa = 0.74 V vs. SCE) is easier to oxidize than is
PhCH2NNPPh3 (Epa 1.06 V vs. SCE), yet the former reacts with
[Mo6O19
22 to give only metathesis without electron transfer.10a
=
[Mo5O18(Mo
N
CH2Ph)]2–
]
1
1 R. K Grasselli, in Heterogenous Catalysis, ed. B. L. Shapiro, Texas
A&M University Press, College Station, TX, 1984, p. 182.
2 J. D. Burrington, C. T. Kartisek and R. K. Grasselli, J. Org. Chem.,
1981, 46, 1877; W. Martin and J. H. Lunsford, J. Am. Chem. Soc., 1981,
103, 3728.
(H+ / 2 e–) transfer
hydrolysis
)]2–
CHPh
2–
[Mo5O17(OH)( Mo
N
[Mo6O19
]
+ PhCH2NH2
2
3 J. Haber and B. Grzybowska, J. Catal., 1973, 28, 489; B. Grzybowska,
J. Haber and J. Janas, J. Catal., 1977, 49, 150; J. D. Burrington and R.
K. Grasselli, J. Catal., 1979, 59, 79; J. D. Burrington, C. T. Kartisek and
R. K. Grasselli, J. Catal. 1983, 81, 489.
PhCH NCH2Ph
PhC
N
4 R. K. Grasselli and J. D. Burrington, Ind. Eng. Chem. Prod. Res. Dev.,
1984, 23, 394 and references therein.
Scheme 2 Proposed decomposition pathways for 1
5 D. M.-T. Chan and W. A. Nugent, Inorg. Chem., 1985, 24, 1422.
6 E. A. Maatta and Y. Du, J. Am. Chem. Soc., 1988, 110, 8249.
7 J. Belgacem, J. Kress and J. A. Osborn, J. Mol. Catal., 1994, 86, 267.
8 H. R. Allcock, E. C. Bissell and E. T. Shawl, Inorg. Chem., 1973, 12,
2963.
and its hydrolysis (to produce benzyl amine and hexamo-
lybdate) will initiate the reaction sequence leading to
PhCHNNCH2Ph (Scheme 1).
9 L. Kihlborg, Ark. Kemi, 1963, 21, 357.
10 (a) Y. Du, A. L. Rheingold and E. A. Maatta, J. Am. Chem. Soc., 1992,
114, 345; (b) J. B. Strong, R. Ostrander, A. L. Rheingold and E. A.
Maatta, J. Am. Chem. Soc., 1994, 116, 3601; (c) R. J. Errington, C. Lax,
D. G. Richards, W. Clegg and K. A. Fraser, in Polyoxometalates: From
Platonic Solids to Anti-Retroviral Activity, ed. M. T. Pope and A.
Mu¨ller, Kluwer, Dordrecht, 1994, p. 105; (d) A. Proust, R. Thouvenot,
M. Chaussade, F. Robert and P. Gouzerh, Inorg. Chim. Acta, 1994, 224,
81; (e) W. Clegg, R. J. Errington, K. A. Fraser, S. A. Holmes and A.
Scha¨fer, J. Chem. Soc., Chem. Commun., 1995, 455; (f) J. L. Stark, A.
L. Rheingold and E. A. Maatta, J. Chem. Soc., Chem. Commun., 1995,
1165; (g) J. L. Stark, V. G. Young, Jr. and E. A. Maatta, Angew. Chem.,
Int. Ed. Engl., 1995, 34, 2547.
11 M. Che, M. Fournier and J. P. Launay, J. Chem. Phys., 1979, 71,
1954.
12 R. Neumann and M. Lissel, J. Org. Chem., 1991, 56, 5707; K.
Nakayama, M. Hamamoto, Y. Nishiyama, and Y. Ishii, Chem. Lett.,
1993, 1699.
The conversion {[Ph–CH2–N]22 ? Ph–C·N} requires the
formal export of 2 H+ and 4 e2. Precedent suggests that the
a-CH2 hydrogens within 1 should be acidic6,13,14 while the oxo
sites within 1 will display enhanced basicity10b,g as a result of
the effective electron donation provided by the benzylimido
ligand. These combined attributes suggest that the decomposi-
tion of 1 to produce benzonitrile is triggered by initial H+
migration, probably in a pairwise10g fashion; an accompanying
2 e2 reduction of its hexamolybdate cage will produce a
reduced benzylideneamido dianion 2 (Scheme 2).6,13 While
further reduction of 2 is improbable, several species in the
reaction solution should be capable of oxidizing 2 {the most
potent of which is [Mo6O19]22}. Such an oxidation would
increase the acidity of the remaining methine hydrogen atom,
facilitating a second H+–2 e2 transfer process, allowing the
release of benzonitrile. Efforts are underway to isolate com-
plexes analogous to 1 and 2.
13 Y. Du, A. L. Rheingold and E. A. Maatta, J. Chem. Soc., Chem.
Commun., 1994, 2163.
14 Y. Du, Ph.D. Thesis, Kansas State University 1992.
We thank the Office of Basic Energy Sciences, Department
of Energy (USA) for support of this work.
Received in Bloomington, IN, USA, 30th May 1997; 7/03790B
1708
Chem. Commun., 1997