Vanadium nitride functionalization and denitrogenation by carbon disulfide and dioxide

Article abstract of DOI:10.1039/b111550m

A dramatic difference in behavior is observed for the dithiocarbamate and carbamate complexes [Ar(But^t)N]₃V(NCE₂)Na(THF)₂(E = S or O, respectively), prepared from the corresponding nitride species {[Ar(But^t)N]₃V≡NNa}₂ by way of a nucleophilic addition reaction involving carbon disulfide or dioxide, and is rationalized with the aid of DFT calculations.

Full text of DOI:10.1039/b111550m

Vanadium nitride functionalization and denitrogenation by carbon
disulfide and dioxide†
Justin K. Brask, Víctor Durà-Vilà, Paula L. Diaconescu and Christopher C. Cummins*
Department of Chemistry, Room 2-227, Massachusetts Institute of Technology, Cambridge, MA 02139-4307,
USA. E-mail: ccummins@mit.edu
Received (in Purdue, IN, USA) 12th December 2001, Accepted 6th February 2002
First published as an Advance Article on the web 26th March 2002
A dramatic difference in behavior is observed for the
51V NMR), an analogue of the known sulfide 2-S.4 Removal of
solvent in vacuo and dissolution of the remaining red residue in
diethyl ether followed by filtration through a fine sintered-glass
frit gave, upon drying, dark red microcrystalline 1-S in 94%
yield. Collected on the frit was a white salt identified as sodium
thiocyanate (96% yield) by comparing its 13C NMR spectrum in
D2O with that obtained for a commercially available sample
(133.7 ppm).5 This finding balances the equation of decomposi-
tion of 1-NCS2Na(THF)2 to generate 1-S (Scheme 1). Charac-
terization of 1-S includes the observation of a singlet at 659 ppm
in its 51V NMR spectrum, cf. 676 ppm for 2-S,4 as well as an X-
ray structure.§ Furthermore, following the procedure outlined
by Gambarotta and co-workers for the preparation of 2-S,4 a
dark green solution of 1 in diethyl ether was treated with
elemental sulfur in order to prepare 1-S independently (93%
isolated yield).
dithiocarbamate
and
carbamate
complexes
[Ar-
(But)N]3V(NCE2)Na(THF)2 (E = S or O, respectively),
prepared from the corresponding nitride species {[Ar-
(But)N]3V·NNa}2 by way of a nucleophilic addition reaction
involving carbon disulfide or dioxide, and is rationalized
with the aid of DFT calculations.
The vanadium(III) trisanilide complex [Ar(But)N]3V1 (1, Ar =
3,5-Me2C6H3) provides a robust platform for stabilizing
interesting main-group functionalities.2 For example, the nucle-
ophilicity of the nitrido substituent in (1-NNa)2, generated upon
treatment of 1 with sodium azide, facilitated the preparation of
1-NPCl2 via a metathesis reaction with PCl3. A series of
transformations yielded ultimately the monomeric iminophos-
phinimide 1-NPNBut.2 Herein, we report that addition reactions
involving nucleophilic (1-NNa)2, or the new adamantyl deriva-
tive {[Ar(1Ad)N]3VNNa}n (2-NNa)n, and the electrophiles CS2
or CO2 generate the vanadium trisanilide dithiocarbamate and
carbamate complexes 1-NCS2Na(THF)2, 2-NCS2Na(THF)2
and 1-NCO2Na(THF)2 (Scheme 1). The unanticipated sponta-
neous conversion of 1-NCS2Na(THF)2 to the corresponding
terminal sulfide 1-S (with NaNCS extrusion) is rationalized
with the aid of DFT calculations.
Kinetic data obtained by single-pulse 1H NMR spectroscopy
indicate that the decay of 1-NCS2Na(THF)2 to form 1-S is first-
order in vanadium (kobs at 25 °C = 2.2 0.1 3 1024 s21),† thus
we favor an intramolecular mechanism for this conversion.
Hence, a four-membered V–N–C–S ring is implicated either as
a short-lived (not observed) intermediate complex or in the
transition state leading to thiocyanate extrusion. Species
containing a related M–N(Ph)–C–E ring have been charac-
terized by crystallography (M = Ta, E = S)6 or proposed as
intermediates (M = V, E = O).7
Dark red 1-NCS2Na(THF)2 was isolated in 89% yield by
treatment of a yellow-green THF solution of (1-NNa)2 with 2
equiv. CS2 and expeditious low-temperature (235 °C) work-up.
As a solution in THF-d8, 1-NCS2Na(THF)2 exhibits a partially
resolved 1+1+1 triplet in its 51V NMR spectrum at 254 ppm
Given that the three anilide ligands provide a protected
pocket, which has proven critical for preventing the dimeriza-
tion of derivatives of (1-NNa)2,2 it is surprising that intra-
molecular nucleophilic attack on the V center by one of the S
atoms is possible. It was surmised accordingly that
2-NCS2Na(THF)2, incorporating a more highly constrained
pocket, would exhibit a greater kinetic resistence to thiocynate
extrusion. Forthwith, the dithiocarbamate complexes
2-NCS2Na(THF)2 and 2-N13CS2Na(THF)2 were prepared by a
protocol analogous for, and in yields consistent with, their But
counterparts (vide supra). The new reagent (2-NNa)n employed
in these syntheses was obtained in 64% yield from the reaction
of [Ar(1Ad)N]3V4 (2) with NaN3. The NMR data obtained for
(1J
V 100 Hz),3 due to coupling with one of the adjacent 14
N
14N51
nuclei.‡ This phenomenon is attributed to interaction of the
vanadium center with the nitrido nitrogen, which evidently
experiences a more symmetrical electric field gradient than
those nitrogens contained in the three anilide ligands. Utiliza-
tion of 13C-labeled carbon disulfide, i.e. preparation of
1-N13CS2Na(THF)2, was required to observe the relevant broad
singlet at 237 ppm (Dn1/2 = 340 Hz) in the 13C NMR spectrum.
51V NMR spectroscopy revealed a broad resonance at 256 ppm
(Dn1/2 = 274 Hz) for 1-N13CS2Na(THF)2, likely a consequence
of 13C coupling masking the coupling to 14N.
2-NCS2Na(THF)2 [51V: 217 ppm (t, 1J
14N51
88 Hz)] and
V
Surprisingly, storage of a THF solution of 1-NCS2Na(THF)2
2-N13CS2Na(THF)2 [51V: 220 ppm (br, Dn1/2 252 Hz); 13C:
234 ppm (br, Dn1/2 230 Hz)] in THF-d8 were similar to those
determined for 1-NCS2Na(THF)2 and 1-N13CS2Na(THF)2,
respectively. Indeed, 2-NCS2Na(THF)2 was found to be stable
in solution for extended periods (days) at room temperature.
Only upon refluxing a THF solution of 2-NCS2Na(THF)2 for
several hours was a trace amount of 2-S observed ( < 5% by 1H
and 51V NMR), in addition to traces of the free aniline
HN(1Ad)Ar along with other unidentified products. Note that
the sulfide 2-S, as prepared independently,4 is stable to this
regimen.
at 23 °C for 24 h resulted in complete conversion to 1-S (1H and
In light of the fascinating reactivity of 1-NCS2Na(THF)2, we
were intrigued by the prospect of preparing and probing the
behavior of the carbamate analogue 1-NCO2Na(THF)2. Addi-
tion of excess anhydrous CO2 gas to a cooled (0 °C) THF
solution of (1-NNa)2 and subsequent removal of volatile
material in vacuo generated quantitatively ( > 95%) orange
microcrystalline 1-NCO2Na(THF)2. However, in striking con-
Scheme 1
† Electronic supplementary information (ESI) available: synthetic, spectro-
scopic, analytical, and computational results for all new complexes. Fig. S1:
frontier orbitals of model anions: See http://www.rsc.org/suppdata/cc/b1/
b111550m/
902
CHEM. COMMUN., 2002, 902–903
This journal is © The Royal Society of Chemistry 2002

trast to the behavior of 1-NCS2Na(THF)2, 1-NCO2Na(THF)2
exhibits thermal stability at room temperature or upon refluxing
in THF solution for several hours. The 51V NMR spectrum
obtained for 1-NCO2Na(THF)2 features a very broad resonance
centered at ca. 2230 ppm (Dn1/2 1060 Hz). To verify
independently the accessibility and properties of 1-O, 1 equiv.
pyridine N-oxide was added to a dark green solution of 1 to give
free pyridine and 1-O, which was isolated in 94% yield as a fine
orange powder. A distinctive, relatively sharp singlet at 2171
ppm (Dn1/2 290 Hz) was recorded in the 51V NMR spectrum for
1-O.
To gain an understanding of the difference in stability/
reactivity of 1-NCS2Na(THF)2 vs. 1-NCO2Na(THF)2, DFT
methods were employed.† Model compounds incorporating
dimethylamide groups in lieu of anilide ligands were utilized to
reduce calculation cost. The apparent dichotomy in reactivities
was clarified vastly upon analysis of the frontier MOs for the
complexes under scrutiny (Fig. S1, ESI†). Significantly, the
HOMO–LUMO gap for the model anion (Me2N)3V(NCS2)2 is
less than half of that calculated for (Me2N)3V(NCO2)2,
reducing its relative stability considerably. This difference in
energies is attributed to the combination of two phenomena: (i)
the HOMO of (Me2N)3V(NCO2)2 is stabilized with respect to
that of (Me2N)3V(NCS2)2 by means of a substantial V–N p-
bonding contribution and (ii) the LUMO for (Me2N)3V(NCS2)2
is stabilized by the presence of an N–C bonding interaction,
which is negligible in the LUMO for (Me2N)3V(NCO2)2.
Another factor that may facilitate conversion of
1-NCS2Na(THF)2 to 1-S via intramolecular nucleophilic attack
is the presence of longer C–E bonds (optimized at 1.71 and 1.25
Å for E = S and O, respectively).¶
which we are here concerned. Remarkably good agreement with
the experimental 51V NMR chemical shifts was exhibited,
lending strong support to our structural assignments.¶ Calcula-
tions were performed also on the model terminal chalcogenides
(Me2N)3VE (E = O, S, Se), a series of particular interest since
the chemical shift values differ dramatically in the three real
systems 1-O, 1-S, and 2-Se [ca. 2171 (vide supra), 659 (vide
supra) and 1001 ppm,4 respectively]. As anticipated,8 variations
in the diamagnetic component are minimal ( < 10 ppm),
rendering changes in the shielding of V due mostly to the
paramagnetic contribution. As a consequence of reduced
electronegativity of the terminal substituents, the HOMO–
LUMO gap decreases for the heavier chalcogenide congeners,
i.e. 1-S or 2-Se, facilitating a higher degree of field-induced
mixing between the occupied and vacant frontier orbitals. The
relative paramagnetic contribution to the chemical shielding of
the V center is thus increased,3c,8,9 ultimately accounting for the
observed downfield shifts vis-à-vis 1-O. Isolobal molybdenum
nitride and phosphide derivatives exhibit analogous NMR
phenomena.10
In summary, this work illuminates both experimentally and
by way of DFT calculations the fascinating properties and
behaviors of the new family of dithiocarbamate and carbamate
functionalities constructed atop a vanadium trisanilide plat-
form.
For financial support we gratefully acknowledge the National
Science Foundation (Award CHE-9988806), the National
Science Board (1998 Alan T. Waterman award to C. C. C.), and
the David and Lucile Packard Foundations. J. K. B. thanks
NSERC (Canada) for a post-doctoral fellowship. We thank also
a referee for insightful comments and helpful suggestions.
Fig. 1 depicts calculated relative enthalpy values for the
system comprised of (Me2N)3VN2, CO2 and CS2 as a function
of the important transformations.¶ This scenario permits
comparison of the CO2 and CS2 reaction pathways, keeping
constant the system’s chemical formula along the hypothetical
reaction coordinate (i.e. formally adding CO2 and CS2 as
necessary). Consumption of CS2 as opposed to CO2 is favored
by ca. 23.4 kcal mol21. Moreover, thiocyanate ejection from
(Me2N)3V(NCS2)2 is substantially more exothermic than cor-
responding extrusion of cyanate from (Me2N)3V(NCO2)2. The
observed lack of cyanate extrusion from 1-NCO2Na(THF)2 may
be kinetic in origin, as treatment of 1-O with NaNCO under
forcing conditions similarly resulted in no reaction.
Notes and references
‡ See ESI for the 51V NMR spectrum of 1-NCS2Na(THF)2.
§ Crystal data for 1-S: C36H54N3SV, M = 611.82, monoclinic, space group
C2/c, a = 30.398(5), b = 10.645(2), c = 22.194(4) Å, b = 93.515(3)°, V
= 7168(2) Å3, T = 183(2) K, Z = 8, m(Mo-Ka) = 0.361 mm21, Dc
=
1.134 g cm23, 10 185 reflections measured, 3352 unique (Rint = 0.0436),
3351 observed [I > 2s(I)]. The final R1 and wR2(F2) were 0.0873 [I >
2s(I)] and 0.1606 (all data), respectively. See ESI for an ORTEP drawing of
complex 1-S.† CCDC reference number 176564. See http://www.rsc.org/
suppdata/cc/b1/b111550m/ for crystallographic data in CIF or other
electronic format.
¶ See ESI for details and tables: isotropic shielding, paramagnetic and
diamagnetic contributions, Mulliken and Hirshfeld charges, and relative
enthalpies.
Chemical shielding constants were determined by DFT for
the 51V nuclei in models of all vanadium(
) complexes with
V
1 M. G. Fickes, Ph.D. Thesis, Massachusetts Institute of Technology,
MA, 1998.
2 J. K. Brask, M. G. Fickes, P. Sangtrirutnugul, V. Durà-Vilà, A. L. Odom
and C. C. Cummins, Chem. Commun., 2001, 1676.
3 (a) W. A. Nugent and R. L. Harlow, J. Chem. Soc., Chem. Commun.,
1979, 342; (b) F. Preuss, W. Towae, V. Kruppa and E. Fuchslocher, Z.
Naturforsch., Teil B, 1984, 39, 1510; (c) D. D. Devore, J. D. Lichtenhan,
F. Takusagawa and E. A. Maatta, J. Am. Chem. Soc., 1987, 109, 7408;
(d) P. L. Hill, G. P. A. Yap, A. L. Rheingold and E. A. Maatta, J. Chem.
Soc., Chem. Commun., 1995, 737.
4 K. B. P. Ruppa, N. Desmangles, S. Gambarotta, G. Yap and A. L.
Rheingold, Inorg. Chem., 1997, 36, 1194.
5 A. A. Isab, M. N. Akhtar and A. R. Al-Arfaj, J. Chem. Soc., Dalton
Trans., 1995, 1483.
6 H. Brunner, M. M. Kubicki, J.-C. Leblanc, C. Moise, F. Volpato and J.
Wachter, J. Chem. Soc., Chem. Commun., 1993, 851.
7 K. R. Birdwhistell, T. Boucher, M. Ensminger, S. Harris, M. Johnson
and S. Toporek, Organometallics, 1993, 12, 1023.
8 (a) J. Mason, Polyhedron, 1989, 8, 1657; (b) D. Rehder, Coord. Chem.
Rev., 1991, 110, 161; (c) W. von Philipsborn, Chem. Soc. Rev., 1999,
95.
9 M. Bühl and F. A. Hamprecht, J. Comput. Chem., 1998, 19, 113.
10 G. Wu, D. Rovnyak, M. J. A. Johnson, N. C. Zanetti, D. G. Musaev, K.
Morokuma, R. R. Schrock, R. G. Griffin and C. C. Cummins, J. Am.
Chem. Soc., 1996, 118, 10654.
Fig.
1 Relative enthalpies for states of the system comprised of
(Me2N)3VN2, CO2, and CS2 as a function of relevant transformations.
Values were computed using DFT methods ([V] = V(NMe2)3) consisting
of geometry optimization for each of the depicted local minima. Possible
transition states were not explored.
CHEM. COMMUN., 2002, 902–903
903