Se,Se′-diphenyl carbonodiselenoate: Preparation and characterization of its molecular structure and thermal reactivity

Article abstract of DOI:10.1039/b103499p

Reaction of dipotassium nitroacetate with benzeneselenenyl bromide produces the title compound as a thermally unstable product, which has been characterized by X-ray crystallography.

Full text of DOI:10.1039/b103499p

Se,SeA-Diphenyl carbonodiselenoate: preparation and characterization
of its molecular structure and thermal reactivity
Jason T. Lowe,a A. Chandrasekaran,b Roberta O. Day*b and William Rosen*a
a Department of Chemistry, University of Rhode Island, Kingston Rhode Island 02881, USA.
E-mail: wrosen@chm.uri.com
b Department of Chemistry, University of Massachusetts, Amherst, MA 01003-9336, USA.
E-mail: rday@chemserv.chem.umass.edu
Received (in Corvallis, OR, USA) 10th April 2001, Accepted 6th June 2001
First published as an Advance Article on the web 3rd July 2001
Reaction of dipotassium nitroacetate with benzeneselenenyl
bromide produces the title compound as a thermally
unstable product, which has been characterized by X-ray
crystallography.
CI/MS spectrum of 3 exhibited a molecular ion centered at 343
amu with appropriate isotopic satellite peaks for a compound
having two selenium atoms in its molecular structure. Major
fragment peaks were observed at 312, 237 and 159 amu as well
as the base peak at 78 amu. The EI/MS spectrum was identical
to that of 2. (Carbonodithioates fragment in a similar manner in
the EI/MS.7) (6) The UV-VIS spectrum of 3 exhibited no
absorbance above 300 nm in contrast to 2 which has a
characteristic thermochromic8 band at 337 nm (log e = 3.1).
While this spectral information was indicative that 3 had a
molecular structure that contained two phenyl selenium moie-
ties, in addition to a carbonyl group, we were uncertain whether
our compound was similar to Se,SeA-dibenzyl carbonodisele-
noate (5) prepared by Devillanova et al.9 (Reported spectral
information available for 5 consisted only of an IR band at 1680
cm21 which was attributed to a carbonyl band.) Consequently
we determined the X-ray crystal structure for 3 which
confirmed our suspicions that we had isolated Se,SeA-diphenyl
carbonodiselenoate as a rather unstable molecular entity. We
hoped that the reasons for the instability of 3 might then become
apparent.
A perspective view of the molecular structure obtained for 3
is shown in Fig. 1. (We had to collect the data at 173 K because
of the instability of 3 in the X-ray beam.¶) The space group was
found to be C2/c with a twofold axis passing through the
carbonyl function.∑ The unit cell contained four molecules of 3.
The bond lengths for the phenyl selenium moieties were found
to be almost identical10 to those of 2 including the Se–Cipso bond
length = 191.5(3) pm. The bond angles around the carbonyl
function of 3 are somewhat different from those of diphenyl
carbonate (6);11 for 3 Se–C(O)–Se is 110.2(2)°, Se–C = O is
124.9(1)°, and C(O)–Se–Cipso is 97.1(1)° while for 6 O–C(O)–
O is 104.6°, O–C = O is ~ 127.5° and C(O)–O–Cipso is
~ 118.5°. The bond lengths around the carbonyl function of 3
appeared to be normal, and similar to those of 6 when
comparable: for 3 Se–C(O) is 193.5(2) pm, and C = O is
118.5(5) pm while for 6 C = O is 119.1 pm.
Dipotassium nitroacetate, 1, was first synthesized by Steinkopf1
in 1909. Since that time the yield and simplicity of preparation
of 1 has been improved a number of times.2 The reported utility
of employing this easily prepared dianionic organic salt in
reaction with electrophilic reagents has been somewhat limited
however. Reaction with electrophilic reagents such as the
halogens,3 the strong mineral acids,4 and several transition
metal salts5 have been noted but our preliminary attempts to
directly react 1 with organo-electrophilic reagents proved
frustrating.† One exception, we discovered, was the direct
reaction of benzeneselenenyl bromide with 1 in methanol,‡
which produced CO2 and diphenyl diselenide, 2, along with two
other organic products, 3 and 4. This communication thus
reports our characterization of one of these three organic
products, 3, and additionally elaborates upon its easy thermo-
lytic conversion to produce 2. (The chemical characterization of
4 is still under investigation.§)
Addition of 1 to an ice-cold solution of benzeneselenenyl
bromide in methanol resulted in three TLC separable organic
compounds after effervescence of CO2 had subsided. Separa-
tion of the three organic compounds using silica gel chromato-
graphy could be achieved by elution with hexane–methylene
chloride mixtures. The initial band off the column was easily
identified as 2. The compounds 3 and 4 were then eluted in
sequence, with both compounds decomposing to 2 if strict
cooling conditions were not maintained (see Scheme 1).
Compound 3 could be crystallized by slow evaporation of the
elution solvent at 0 °C to deliver X-ray quality crystals. Initial
observations on crystals of 3 indicated that yellow crystals of 2
began to grow from the colorless crystals if 3 was allowed to sit
at room temperature or above for any extended period of time.
(Storage in the refrigerator was mandatory.)
Spectral analysis of compound 3 yielded the following
important information: (1) the solution IR spectrum (CCl4)
exhibited a carbonyl stretch at 1711 cm21 in addition to bands
similar to those exhibited by 2. Attempts to obtain the IR
spectrum of 3 as a KBr pellet resulted in a spectrum identical to
that of 2. (2) The 1H-NMR spectrum of a solution (CDCl3) of 3
exhibited phenyl resonance lines only. (3) The 13C-NMR
spectrum (CDCl3) of 3 exhibited a carbonyl resonance at d =
182.7 in addition to four resonance lines in the aromatic region.
(4) The 77Se-NMR spectrum (CDCl3) of 3 exhibited a single
resonance line at d = 749 relative to dimethyl selenide.6 (5) The
There are several interesting interactions within 3. That
exhibited between the two intramolecular selenium atoms is one
of them. These two selenium atoms are separated by a distance
of 319 pm which is well within the sum of two Van der Waals
radii for selenium ( ~ 400 pm). Whether there is a partial
Fig. 1 ORTEP representation of 3 with the b axis being the 2-fold axis and
roughly in the plane of the page. Atoms with the same label, but
superscripted 2, are related by a crystallographic 2-fold axis.
Scheme 1
1390
Chem. Commun., 2001, 1390–1391
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b103499p

‡ Preparation of 3: 1.0 g of benzeneselenenyl bromide was dissolved in 15
ml of cold (0 °C) methanol and placed in a round bottom flask containing
a stir bar. Solid, freshly prepared 1, 0.77 g, was added in a controlled manner
so that effervescence did not become too vigorous. Monitoring of progress
was done by TLC over the duration of the ~ 1 h reaction. During the course
of the reaction the color changed from dark red to a light yellow. Rotary
evaporation of the resultant mixture was done under vacuum at approx-
imately 0 °C after adding some silica gel. The product, adsorbed onto silica
gel, was then mixed with 70:30 hexane–CH2Cl2 and added to the top of a
chromatography column. Elution was commenced and the fractions were
monitored using TLC. The organic products were eluted from the column in
sequence (numerically). The fractions containing 3 were combined and
allowed to evaporate in the cold until clear colorless crystals had formed. %
Yield of 3 = 19%; mp = 48–50 °C.
intramolecular bond between the two selenium atoms of 3,
implying some hypervalency,12 is not immediately obvious.
Secondly, while the overall molecular structure for 3 is similar
to that exhibited by 6, the comparative bond angles around the
carbonyl function are different implying that any interaction
between selenium atoms in 3 is very different from those
manifested by the oxygen atoms in 6. It was therefore of interest
to determine the thermodynamic activation parameters asso-
ciated with the thermal conversion of 3 to yield 2.
Dilute solutions of 3 could be quantitatively converted to 2
while monitoring changes in the visible spectrum at 350 nm. (A
clean isosbestic point was observed at ~ 300 nm.) Compound 3,
which is colorless, smoothly and completely converted to
§ The molecular characterization of 4 has proven difficult since it more
readily decomposes to 2 than even 3. We believe 4 may be either a hydrate
or hemihydrate of 3, since 4 exhibits CI/MS data with ion peaks centered
around major masses, having two and four selenium atoms respectively, at
359 and 699 amu.
1
yellow 2 in ethylene glycol solvent within 2 h at 60 °C and ⁄2 h
at 90 °C. The first order (or pseudo-first order) kinetics were
linear over the entire temperature range studied, after taking into
account the thermochromic properties of 2.8 The thermody-
namic activation values of Ea = +65.8 kJ mol21 and DSa
=
¶ The compound 3 decomposed in the X-ray beam within 5 min at rt.
2227 J deg21 could be derived using the method of Arrhenius.
The low activation energy for the thermolysis reaction thus
indicated that the process of forming the 229 pm bond in 2 may
be partially on the way to completion within the molecular
structure of 3 itself. Consequently the preliminary conclusion
that there may be a partial Se–Se bond in 3 seems plausible. (See
footnotes ** and †† which were added in proof.)
∑ X-Ray study of 3: Data was collected on
a Nonius KappaCCD
diffractometer. Satisfactory crystal stability was obtained by mounting the
crystal inside of a sealed, thin walled glass capillary and by collecting the
data at 173(2) K. Crystal data for 3: C13H10OSe2, FW
= 340.13,
monoclinic space group C2/c, a 2140.38(5), b 523.60(2), c =
=
=
1291.76(4) pm, b = 119.960(2)°, V = 1.25423(7) 3 109 pm3, Z = 4, Dcalc
= 1.801 g cm23, m = 5.871 mm21. Intensities were obtained for 2055
reflections (2qmax = 50.1°) of which 1104 were unique. Refinement on F2
was based on all of the data. The final agreement factors for the 941 unique
The negative entropy of activation generated during the
thermolysis reaction of 3, where one molecule fragments to two
molecular products, may imply that the associated disorder in
the overall reaction can be offset, to some extent, by some
factors of orderliness in the activated complex. The first factor
implied is that the solvent (ethylene glycol) could be involved in
creating more order during the thermolysis process. This
possibility is buttressed by our observation that thermolysis of 3
in hydrocarbon solvent requires higher temperatures and is thus
slower than in HOCH2CH2OH.** A second factor implicated is
that a shorter than usual movement of the phenyl selenium
substituents, during the extrusion of CO, may be inherent in 3.
As the crystal structure∑ of 3 shows the angle between the
phenyl selenium groups is compressed in comparison to a
normal sp2 hybridized situation around a carbonyl group. (It has
to be noted here that the compression of the analogous angle in
6 is even greater.11) And a third possible factor is the proposed
hypervalent interaction between the selenium atoms of 3. This
proposal bolsters a hypothetical mechanistic process where the
thermolytic transition state structure is a loosely organized
diselenacyclopropanone ring structure. Consequently contrast-
ing the thermal lability of 3 with the thermal stability of 6
suggests that these factors may not be available during
analogous bond formation processes associated with the parent
chalcogen.†† Comparing the molecular structures of Se,SeA-
diphenyl carbonodiselenoate, 3, and S,SA-diphenyl carbonodi-
thioate (7) to that of diphenyl carbonate,11 6, should prove
interesting.‡‡
reflections with I > 2sI are R
163698.
= 0.0262 and Rw = 0.0595. CCDC
** As pointed out by referee K this observation ‘…strongly argues for a
polar transition state…’ that is ‘…hydrogen bonded resulting in increasing
order in the transition state.’ We agree with this comment and note that a
diselenacyclopropanone molecular entity would most likely be polar and H-
bonded in ethylene glycol.
†† This same referee K has stated: ‘…the factors that the authors ascribe to
the thermal lability of 3 are said to be not available for thermally stable 6.
A more obvious difference for this differing reactivity is that the
thermodynamic driving force is substantially different i.e. a C(O)–O bond is
stronger than a C(O)–Se bond and an O–O bond is much weaker than an Se–
Se bond.’ We again agree and we wish to thank this referee for these cogent
ideas lending further support to our diselenacyclopropanone transition state
proposal.
‡‡ Thermolytic studies of carbonodithioate esters have been referenced13 to
7 but the molecular structure of 7 has not been determined as far as we are
aware.
1 W. Steinkopf, Chem. Ber., 1909, 42, 3925.
2 S. Koto, M. Koyama and S. Zen, Org. Synth., 1988, Coll. Vol. V,
797.
3 W. Steinkopf and M. Kuhnel, Chem. Ber., 1942, 75B, 1323.
4 H. Feuer, H. B. Hass and K. S. Warren, J. Am. Chem. Soc., 1949, 71,
3078.
5 D. J. Sutor, F. J. Llewellyn and H. S. Maslen, Acta Crystallogr., 1954,
7, 145.
6 H. C. E. MacFarland and W. MacFarland, Multinuclear NMR, ed. J.
Mason, Plenum, New York, 1987, p. 417.
7 C. Djerassi, P. Brown and J. B. Thompson, J. Am. Chem. Soc., 1966, 88,
We (ROD and AC) would like to thank NSF (CHE-9974648)
and the University of Massachusetts for funds to purchase the
Nonius KappaCCD diffractometer. JTL and WR would like to
thank the URI Foundation for partial support to fund this
research and to Michael Callahan for recording the CI/MS for
us.
4049.
8 H. Kunkely, A. Vogler and J. K. Nagle, Inorg. Chem., 1995, 34,
4511.
9 F. A. Devillanova, G. Verani and L. Henriksen, Tetrahedron, 1980, 36,
1451.
10 R. Marsh, Acta Crystallogr., 1952, 5, 458.
11 J. A. King, Jr. and G. L. Bryant, Jr., Acta Crystallogr., Sect. C., 1993, 49,
550.
12 N. Furukawa and S. Sato, Chemistry of Hypervalent Compounds, ed. K.
Akiba, Wiley-VCH, New York, 1999, p. 241.
13 E. Marianucci, F. Pilati and C. Berti, Polym. Degrad. Stab., 1986, 16,
297.
Notes and references
† For example: reaction of 1 with simple alkyl halides in methanol (or other
solvents) produced no identifiable materials other than the starting
compounds.
Chem. Commun., 2001, 1390–1391
1391