Reduction of an hexamethylphosphoramide degradation product: A diazabutadiene

Article abstract of DOI:10.1021/ol901798s

Alkali metal over-reduction of an electron acceptor in the presence of a hydrogen atom donor, in hexamethylphosphoramide (HMPA), results in a radical that is not simply the anion radical of the acceptor. This new species exhibits an enigmatic EPR pattern.

Full text of DOI:10.1021/ol901798s

ORGANIC
LETTERS
2009
Vol. 11, No. 20
4564-4567
Reduction of an
Hexamethylphosphoramide Degradation
Product: A Diazabutadiene
Brad D. Rose, Steven J. Peters,* Richard C. Reiter, and Cheryl D. Stevenson*
Department of Chemistry, Illinois State UniVersity, Normal, Illinois 61790-4160
cdsteVe@ilstu.edu; sjpeter@ilstu.edu
Received August 6, 2009
ABSTRACT
Alkali metal over-reduction of an electron acceptor in the presence of a hydrogen atom donor, in hexamethylphosphoramide (HMPA), results
in a radical that is not simply the anion radical of the acceptor. This new species exhibits an enigmatic EPR pattern. Using 15N and 2H labeling
studies, the “HMPA degradation product” was found to be the anion radical of N,N′-dimethyl-1,4-diazabutadiene. DFT calculations support this
assignment and a mechanism for its formation.
For well over a quarter century, a rather enigmatic and
fleeting EPR signal (Figure 1) has been observed whenever
a variety of substrates are over-reduced with alkali metals
in hexamethylphosphoramide (HMPA) when in the presence
of a source of hydrogen atoms.1 This so-called “HMPA
degradation product” (1•-) has been observed during dehy-
drohalogenations prior to reduction, reduction of alkyne
condensation products, etc.1
absence of ion association,2 we recently decided to identify
the underlying unknown species (1•-) that gives rise to this
troubling spectrum.3 We anticipated a radical containing the
PdO moiety. However, a series of isotopic labeling studies
have revealed 1•- to be a new example of an unligated
diimine anion radical, that of N,N′-dimethyl-1,4-diazabuta-
diene.
It seemed reasonable that the slow over-reduction of a
conjugated organic acid (more electrons added than
Since HMPA is a common solvent for anion radicals and
it normally allows observation of the anion radical in the
(2) Ion association is absent with anion radicals in HMPA. See: (a)
Levin, G.; Jagur-Grodzinski, J.; Szwarc, M. J. Am. Chem. Soc. 1970, 92,
2268–2275. (b) Stevenson, C. D.; Echegoyen, L.; Lizardi, L. R. J. Phys.
Chem. 1972, 76, 1439–1442.
(1) Weak versions of this spectrum were observed during over-reduction
experiments in the following studies: (a) Peters, S. J.; Turk, M. R.;
Kiesewetter, M. K.; Reiter, R. C.; Stevenson, C. D. J. Am. Chem. Soc.
2003, 125, 11212–11268. (b) Gard, M. N.; Kiesewetter, M. K.; Reiter, R. C.;
Stevenson, C. D. J. Am. Chem. Soc. 2003, 125, 16143–16150.
(3) This spectrum causes consternation, as its appearance means that
you have over-reduced and can no longer observe the desired radical.
10.1021/ol901798s CCC: $40.75
Published on Web 09/21/2009
 2009 American Chemical Society

1
Figure 2.
(Right) 31P NMR H-decoupled spectrum of the 15N-
labeled HMPA synthesized as described above recorded at 202.45
MHz. The chemical shift is referenced to 85% phosphoric acid at
0.0 ppm. 1J15N-31P ) 29.9 Hz. (Left) 15N NMR spectrum of the 15N-
labeled HMPA synthesized as described above recorded at 50.70
MHz. The chemical shift is referenced to nitrobenzene at -6.0 ppm.
1J31P-15N ) 29.9 Hz.
Figure 1. X-band EPR spectrum obtained from the potassium over-
reduction of p-nitrophenol in HMPA. The p-nitrophenol simply
serves as a hydrogen donor, and a wide variety of hydrogen atom
donors can be used to generate the same spectrum. Weaker versions
of this spectrum have been observed in many previous studies.
Indeed, three sets of I ) 1/2 spins were found, one of
which results from a group of six, presumably two methyl
groups. Present also are two groups of two spins, due to 15N,
1H, and/or 31P.
organic acid) of high solution electron affinity would yield
desirable concentrations of this HMPA degradation prod-
uct. The careful reduction of p-nitrophenol with a freshly
distilled potassium mirror in HMPA first exhibits the
known EPR pattern for the p-nitrophenol anion radical.4
As the reduction proceeds past one electron transferred
per NO2-C6H4-OH ([K] > [ p-nitrophenol]), H atoms
are released, and this spectrum is replaced by a strong
signal due to 1•- (Figure 1). Even at this signal-to-noise
ratio, an unambiguous interpretation is impossible, due
to the multitude of possible spin combinations.
The conundrum can be resolved with isotopically labeled
HMPAs, but unfortunately none are commercially available.
Only a patent procedure, which is difficult to reproduce and
quite cumbersome, for the synthesis of laboratory amounts
has been reported.5 Furthermore, since HMPA is a common
solvent used in EPR studies, a clean simple synthetic method
for isotopically labeled HMPA was developed.
Specifically, an excess of 15N-dimethylamine or perdeu-
terated dimethylamine was reacted with POCl3 in dry THF
under high vacuum conditions at ambient temperatures. After
sufficient mixing time, the solution was exposed to excess
ammonia. The NH3 helps break up the (CH3)2NH2Cl
precipitate to release more amine for further reaction with
the POCl3. After additional mixing, the excess NH3 was
removed and the remaining solution was dissolved in diethyl
ether and filtered to remove the white salt (NH4Cl). After
removal of the ether, the product was distilled under reduced
pressure to provide a 49% yield of labeled HMPA. The
product and its purity were verified using GCMS and NMR
(Figure 2).
Figure 3. (Upper) Low-field half of the x-band EPR spectrum
obtained from the K over-reduction of p-nitrophenol in per-15N-
HMPA. (Lower) A computer simulation using couplings of 7.870
G for six spins, 4.145 G for two spins, and 8.456 G for two spins.
I ) 1/2 for all of the nuclear spins. The arrow marks the spectral
center and the ∆wpp ) 0.28 G.
To resolve this spin ambiguity, the reduction and EPR
analysis were repeated in HMPA-d18. Surprisingly, in the
deuterated solvent, only I ) 1 spins proved to be present
(see Figure 4); all of the I ) 1/2 spins in 1•- must be due to
protons; and there can be no splittings from 31P.
All that was necessary to obtain the correct simulation was
to determine which and how many of the I ) 1/2 spins, from
Figure 3, must be multiplied by the ratio of gyromagnetic
ratios γ2H/γH or γ14N/γ15N along with changing the spin
numbers from 1/2 to 1. All possibilities were computed, and
the only simulation that agrees with the empirical spectrum
is shown in blue in Figure 4.
The reduction and EPR analysis were repeated in our
newly synthesized per-15N-HMPA. Only nuclei with I ) 1/2
can now be present, and the resulting EPR spectrum (red in
Figure 3) was easy to simulate (blue in Figure 3).
(4) Piette, L. H.; Ludwig, P.; Adams, R. N. J. Am. Chem. Soc. 1962,
84, 4212–4215.
(5) (a) Liu, R. Huaxue Shijie 1991, 32, 154–155. (b) Kenneth, G. L.
Amidation of phosphorus halides with organic amines, US Patent 2,852,550,
1958.
Now, simply converting the couplings from the six
equivalent deuteriums of aD ) 1.21 to aH ) 7.87 G and two
Org. Lett., Vol. 11, No. 20, 2009
4565

this strongly reducing environment), and it has at least a C2
symmetry. These considerations provide an empirical formula
of C2(CH3)2H2N2.
The presence of two different groups of carbons is
supported by the vertical expansion of Figure 5 (see
Supporting Information) revealing splittings from two pairs
of 13C nuclei with splittings of 2.17 and 2.55 G. All of this
suggests that 1•- is a N,N′-dimethyl-1,4-diazabutadiene anion
radical.
As in the alkyl-substituted butadiene anion radicals,6 1•-
can be in the cis or trans configuration (Scheme 1); this issue
can be resolved via DFT calculations.
Scheme 1. cis-trans Equilibrium of the Diimine Anion Radical,
Lying Predominantly in the Direction of the trans Isomer
Figure 4. (Upper) Low-field half of the x-band EPR spectrum
obtained from the K over-reduction of p-nitrophenol in per-
deuterated-HMPA. (Lower) A computer simulation using aD’s )
2
2
1.21 G (6 Hs), 0.637 G (2 Hs), and 6.01 G for two 14N’s. I ) 1
for all of the nuclear spins and the ∆wpp ) 0.20 G. The g-tensor
anisotropy effects due to the nitrogen splitting were included in
the simulation. The arrow marks the spectral center.
equivalent deuteriums of aD ) 0.637 to aH ) 4.15 G and I
) 1 to 1/2 reveals a near perfect fit (see Figure 5) to the
original spectrum shown in Figure 1.
The B3LYP//6-311++G** calculation predicts the trans
isomer to be of lower energy by 4.06 kcal/mol. The B3LYP//
6-311++G**/EPR-III computed electron-nuclear couplings
(blue) compare favorably, except for aN, to the empirical
values (red). Volumes have been written, and not without
controversy,7 concerning calculated aN’s and the reasons for
poor and good agreement with the empirical values. Interest-
ingly, EPR-II predicts an aN of 6.11 G which is very close,
but EPR-II makes poor predictions for the non-nitrogen
coupling constants. EPR-III, however, predicts less than half
of the empirical value of 6.01 G (see Supporting Informa-
tion).
A reasonable mechanism for the initial formation of the
cis configuration of 1•- involves the attack of the phenolic
Figure 5. (Upper) Low-field half of the x-band EPR spectrum
obtained from the potassium over-reduction of p-nitrophenol in
isotopically natural abundance HMPA. (Lower) This spectrum was
computer simulated using coupling constants of 7.870 G for six
H’s, 4.145 G for two H’s, and 6.01 G for two 14N’s. The vertical
arrow marks the spectral center, and ∆wpp ) 0.25 G.
(6) Gerson, F.; Hopf, H.; Merstetter, P.; Mlynek, C.; Fischer, D. J. Am.
Chem. Soc. 1998, 120, 4815–4824.
(7) (a) Hermosilla, L.; Calle, P.; Garcia de la Vega, J. M.; Sieiro, C. J.
Phys. Chem. A 2006, 110, 13600–13608. (b) Engels, B. Calculation of NMR
and EPR Parameters Theory and Applications; Kaup, M., Buhl, M., Malkin,
V. G., Eds.; Wiley-VCH: Weinheim, 2004; pp 483-492. (c) Wetmore,
S. D.; Boyd, R. J. J. Phys. Chem. B 1998, 102, 9332–934. (d) Bartra, R.;
Giese, B.; Spichty, M.; Gescheidt, G.; Houk, K. N. J. Phys. Chem. 1996,
100, 18371–18379.
A somewhat redundant but revealing experiment involves
the use of NO2-C6H4-OD, which expresses no incorpora-
tion of deuterium into 1•-. The HMPA degradation product
is composed only of atoms originating from the HMPA. We
also know that (1•-) is N to N conjugated, having the set of
two protons on carbons (the H-N bond cannot endure under
(8) (a) Wertz, J. E.; Bolton, J. R. Electron Spin Resonance; McGraw-
Hill: New York, 1986; pp 58-60. (b) Atkins, P.; de Paula, J. Physical
Chemistry, 7th ed.; W. H. Freeman: New York, 2002; pp 432-435. (c)
See also: Wiberg, K. B.; Rablen, P. R.; Marquez, M. J. Am. Chem. Soc.
1992, 114, 8654–8668.
(9) Franz, K. D.; Dieck, T. H.; Starzewski, K. A. O.; HohMann, F.
Tetrahedron 1975, 31, 1465–1469.
4566
Org. Lett., Vol. 11, No. 20, 2009

hydrogen atom coming from the p-nitrophenol upon a methyl
group of the HMPA (Scheme 2). After intramolecular
The anion radical of 1 has not been previously reported;
however, over 30 years ago the one exception, mentioned
above, involved the observation of the anion radicals of N,N′-
dialkyldiimenes, where the alkyl groups are sufficiently bulky
to retard polymerization. The system closest to 1•- is shown
below (4•-). Even though the resolution is quite poor (see
Figure 1 in ref 9), the coupling constants compare quite
favorably to ours for 1•-.
Scheme 2.
Proposed Mechanism for the Formation of 1•-
Presumably 1•-, and a variety of its reaction products,
should now be available from a simple one-pot procedure
from one of the most common and useful solvents: HMPA.
When radicals under reductive conditions are involved,
HMPA may be more than just a solvent. Further, the
synthetic procedure described for the isotopically labeled
HMPA should be viable for “HMPAs” where the methyl
groups are replaced with other entities extending the possible
zoo of break down products. We are currently exploring this
avenue.
rearrangement and further reduction, both 1 and 2 could be
formed. The fate of 2 is unknown, but it is clear that it could
not persist under these highly reductive conditions.
1,3-Butadiene (3) is the simplest conjugated hydrocar-
bon, and its corresponding anion radical serves as the
exemplar8a,b and undergraduate text-book target of quantum
mechanics as applied to molecular systems.8b After all, EPR
spin density measurements reveal the quantum mechanical
“location” of the odd electron in this simplest conjugated
“molecular box” (3•-), and the corresponding spin densities
compare quite favorably with the anachronistic Hu¨ckel
theoretical approach.8c The simplest atomic perturbation upon
Acknowledgment. We thank the National Science Foun-
dation for MRI grant (CHE-0722385), and Steve Peters
thanks the Research Corporation for support.
Supporting Information Available: Previously observed
spectra of 1•-, DFT calculation results, and the natural
abundance 13C spectral analysis. This material is available
free of charge via the Internet at http://pubs.acs.org.
3
•- amounts to the replacement of carbons with adjacent (in
the Periodic Table) elements. This is the topic of some
contemporary theoretical work on the neutral analogues.8c
The anionic aza-derivatives are, however, inherently
unstable and have with one exception9 been reported as
bound ligands.10 In this study, the anion radical of N,N′-
dimethyl-1,4-diazabutadiene (1•-) appears as its solvated ion
in HMPA, and its spin densities compare quite favorably
with those from the DFT theoretical approach.
OL901798S
(10) (a) For example, see: Tuononen, H. M.; Armstrong, A. F. Inorg.
Chem. 2005, 44, 8277–8284, and references therein. (b) Schmidt, E. S.;
Jockisch, A.; Schmidbaur, H. J. Am. Chem. Soc. 1999, 121, 9758–9759.
(c) Andappan, M. M. S.; Nilsson, P.; von Schenck, H.; Larhed, M. J. Org.
Chem. 2004, 69, 5212–5218.
Org. Lett., Vol. 11, No. 20, 2009
4567