To investigate the possible role that a π-olefin–Rh interaction
may have in the reaction of 1 and 2 with rhodium trichloride,
several control experiments were performed. The interaction of
an aqueous suspension of zinc phenylphosphonate with an
aqueous solution of RhCl3 was studied, and no reaction took
place after several days of stirring at room temperature as
evidenced by careful before and after weighing of the metal
phenylphosphonate. Similarly, no reaction took place on
raising the temperature to 80 ЊC for several days. This indicates
that the presence of the vinyl group is important for reactivity.
Changing the surface area of the metal vinylphosphonates to
increase the interaction with rhodium chloride had little effect
on the rate of dissolution and all attempts to further character-
ize the rhodium species in solution, via derivatization with
phosphines or amines, met with little success. However, we were
curious to know whether other transition metal ions would
interact with copper and zinc vinylphosphonates, and the reac-
tion of 1 or 2 with aqueous solutions of cobalt chloride was
studied. After several weeks of stirring either 1 or 2 with CoCl2
in water, no reaction was detected. While olefin complexes of
Co() are known, the stability of these compounds is consider-
ably less than the rhodium() analogs. This is due primarily to
the increased overlap of the filled d-orbitals of the Group 9
transition metal with the π* antibonding orbital of the olefin
ligand in the progression from cobalt to rhodium. Thus we
suggest that the bonding interaction of cobalt ions with 1 and
2 is not sufficiently strong enough to result in the breakdown
of the internal layered structure.
Aqueous solutions of [RhCl3(OH2)3] are known to be moder-
ately acidic, which suggested to us that 1 and 2 were simply
dissolving under acid conditions. Under our experimental con-
ditions the pH of [RhCl3(OH2)3] was measured, and gave values
in the range pH 2.3–4.5 depending on the batch of rhodium
chloride used. A sample of 1 was suspended in a solution of
hydrochloric acid with a known pH of 2.3 and rapidly stirred
for 24 hours. No change in the appearance of the crystals was
observed and the material was recovered without mass loss,
confirming that 1 does not dissolve at this pH. Further work is
in progress to understand the detailed nature of the rhodium-
mediated delamination of metal vinylphosphonates.
Fig. 2 A molecule of 1 showing coordination about the copper atom.
Bond lengths (Å) for the non-hydrogen atoms: Cu–O(2)#1 1.932(3),
Cu–O(1) 1.961(3), Cu–O(3)#2 1.978(3), Cu–O(1W) 2.001(3), Cu–
O(3)#3 2.306(2), P–O(1) 1.527(3), P–O(2) 1.533(2), P–O(3) 1.540(2),
P–C(1) 1.777(4), O(2)–Cu#1 1.932(3), O(3)–Cu#4 1.978(3), O(3)–Cu#5
2.306(2), C(1)–C(2) 1.309(7). Symmetry transformations used to gen-
erate equivalent atoms: #1 Ϫx ϩ 1, Ϫy, Ϫz. #2 x, Ϫy ϩ 1/2, z ϩ 1/2. #3
Ϫx ϩ 1, y ϩ 1/2, Ϫz Ϫ 1/2. #4 x, Ϫy ϩ 1/2, z Ϫ 1/2. #5 Ϫx ϩ 1, y Ϫ 1/2,
Ϫz Ϫ 1/2.
solution of rhodium trichloride was added to a suspension of 1
(molar ratio Rh/Cu = 1.40–1.60/1.00) in water or methanol and
stirred for 1 hour, the blue crystals completely dissolved to give
clear, blood red solutions. The zinc vinylphosphonate 2 reacted
in a similar manner to give homogeneous orange solutions. No
evidence was found for the deposition of copper or zinc metal
that might indicate a redox process taking place. The room
temperature 31P NMR spectrum of the reaction between zinc
vinylphosphonate and Rh() contained two major resonances
at 25.1 and 14.6 ppm with relative intensities 1.0 : 2.9 (although
the line widths were extremely broad, see ESI†). The ratio
of peak intensities did not change on varying the molar ratio
Rh/Zn. The line broadness suggests either dynamic behavior of
the resulting complex, or the presence of oligomeric/polymeric
structures in solution. No significant changes were seen in the
IR spectra of the solutions compared to the solid state
materials, and it is worth noting the absence of a band that
could be attributable to P–OH. In addition, removal of solvent
from the reaction mixtures gave solids that were shown to be
amorphous by X-ray powder diffraction. Redissolving the
solids in water or methanol gave identical IR and NMR spectra
to those recorded during the reaction. All subsequent attempts
to crystallize a single compound from both of the reaction mix-
tures, using a variety of organic solvents (methanol, ethanol,
diethyl ether, THF) failed. Clearly, the highly ordered, layered
structure of copper and zinc vinyl phosphonate has been
destroyed, and it is tempting to suggest that the interaction of
the rhodium metal center with accessible vinyl groups results in
slow delamination (Fig. 3). Delamination of layered organo-
The donors of the Petroleum Research Fund administered by
the American Chemical Society are thanked for support of this
work.
Notes and references
§ Preparation of [Cu(C2H3PO3)]ؒH2O (1). A 1 L round-bottomed flask
was charged with copper sulfate (9.10 g, 57.0 mmol), vinylphosphonic
acid (3.51 g, 57.0 mmol), and de-ionized water (350 mL). Urea (3.38 g,
56.3 mmol) was added to the solution, followed by an aqueous solution
of NaOH (0.10 M), until the pH reached 2.8. The solution was heated
in an oil-bath at 70 ЊC for 72 hours. The resulting crystals were collected
by filtration and dried in air to give 1 as large blue plates (7.82 g, 80%).
Anal. calc. for C2H5PO4Cu: C, 12.81; H, 2.69. Found: C, 12.68; H,
2.77%.
Preparation of [Zn(C2H3PO3)]ؒH2O (2). A 1 L round-bottomed flask
was charged with zinc nitrate (8.37 g, 44.2 mmol), vinylphosphonic
acid (3.42 g, 31.7 mmol), and de-ionized water (350 mL). Urea (3.24 g,
53.9 mmol) was added to the solution, followed by an aqueous solution
of NaOH (0.10 M), until the pH reached 2.8. The solution was heated
in an oil-bath at 70 ЊC for 72 hours. The resulting crystals were collected
by filtration and dried in air to give 2 as white needles (5.44 g, 71%).
Anal. calc. for C2H5PO4Zn: C, 12.68; H, 2.66. Found: C, 12.65; H,
2.63%.
¶ Crystal data for 1: C2H5PO4Cu, M = 187.57, monoclinic, space group
P21/c, a = 9.878(5), b = 7.628(3), c = 7.332(3) Å, U = 549.5(4) Å3, Z = 4,
Dc = 2.267 Mg mϪ3, µ(Mo-Kα) = 4.183 mmϪ1, F(000) = 372, T = 293(2)
K, 1254 independent reflections (Rint = 0.0293) with 2θ < 55Њ. Refine-
ment of 95 parameters converged at final R1 [for selected data with
I > 2σ(I )] = 0.0487, wR2 (all data) = 0.1216.
Fig. 3 Proposed interaction of rhodium complex with pendant vinyl
groups of lamellar phosphonate.
X-Ray powder patterns were recorded on a Scintag XDS-2000 auto-
mated powder diffractometer, fitted with a graphite monochromater,
using Cu Kα radiation. The diffractometer was operated in constant
scan mode of 2.00Њ minϪ1 over the range 5 < 2θ > 60Њ with zero point
determined from an external silicon standard. Step scanning was per-
formed with a step size of 0.03Њ 2θ and a count time of 27.5 min.
phosphonates has previously been reported, for example: the
interaction of alkanethiols with cadmium methylphosphonate
results in destruction of the layered structure, presumably via
formation of stable Cd–S bonds.10
J. Chem. Soc., Dalton Trans., 2002, 824–826
825