Reactivity of palladium(II) methyl complexes towards CO2: Formation of carbonate complexes

Article abstract of DOI:10.1016/j.ica.2003.10.027

Treatment of a benzene solution of (tmeda)PdMe₂ or (dppe)PdMe₂ with carbon dioxide gives the corresponding methyl bicarbonate complex, (L-L)PdMe(O₂COH). These were characterised by NMR spectroscopy and elemental analysis. Under strictly dry conditions no reaction was observed. Recrystallisation of the tmeda bicarbonate complex from acetone yields the corresponding η²-carbonate complex, which was characterised by X-ray crystallography. The reaction probably proceeds through attack by free carbonic acid on the dimethyl complex.

Full text of DOI:10.1016/j.ica.2003.10.027

Inorganica Chimica Acta 357 (2004) 1295–1298
www.elsevier.com/locate/ica
Note
Reactivity of palladium(II) methyl complexes towards CO2:
formation of carbonate complexes
1
Julia Pushkar , Ola F. Wendt
*
Inorganic Chemistry, Department of Chemistry, Lund University, P.O. Box 124, S-221 00 Lund, Sweden
Received 11 June 2003; accepted 7 October 2003
Abstract
Treatment of a benzene solution of (tmeda)PdMe2 or (dppe)PdMe2 with carbon dioxide gives the corresponding methyl bi-
carbonate complex, (L–L)PdMe(O2COH). These were characterised by NMR spectroscopy and elemental analysis. Under strictly
dry conditions no reaction was observed. Recrystallisation of the tmeda bicarbonate complex from acetone yields the corresponding
g2-carbonate complex, which was characterised by X-ray crystallography. The reaction probably proceeds through attack by free
carbonic acid on the dimethyl complex.
Ó 2003 Elsevier B.V. All rights reserved.
Keywords: Crystal structures; Palladium complexes; Carbon dioxide; Carbonate complexes
1. Introduction
2. Experimental
The use of carbon dioxide as a C1 building block has
received a lot of attention recently. The attractiveness of
this approach is due to the fact that CO2 is abundant,
inexpensive, non-toxic and renewable [1]. However,
carbon dioxide is inherently unreactive and thermody-
namically stable and hence only a few processes have
reached industrial scale [1a]. Especially interesting are
reactions that form a new C–C bond such as the classical
carboxylation of phenoxide to give salicylic acid [2,3].
Although p-allyl palladium complexes are known to
insert CO2 there are few reports of such insertions into
late transition metal–carbon r-bonds [4,5]. We are in-
terested in the reactivity of Pd–C bonds towards carbon
dioxide and here we report the reaction of palladium
dimethyl complexes with CO2 to give bicarbonate
complexes. The molecular structure of a palladium g2-
carbonate complex is also reported together with a
possible reaction mechanism.
2.1. General considerations
Benzene solvent and CD3OD were vacuum-distilled
from sodium/benzophenone ketyl and CaH2, respec-
tively. Dry carbon dioxide was obtained by passing it
through a )78 °C trap on the vacuum-line. PdMe2(tmeda)
(1) and PdMe2(dppe) (2) (tmeda ¼ N,N,N0,N0-tetrameth-
ylethylenediamine; dppe ¼ bis(diphenylphosphino) eth-
ane) were prepared according to literature procedures [6].
PdCl2(dppe) (3) was prepared in benzene from the cor-
responding tmeda complex by ligand substitution. All
other reagents are commercially available and were pur-
chased and used as received.
NMR spectra were recorded either on a Varian Unity
300 MHz instrument (1H, 31P) or a Bruker DRX 400
MHz spectrometer (13C). Chemical shifts are given in
ppm downfield from Me4Si using residual solvent peaks
(1H and 13C NMR) or H3PO4 (31P NMR d 0) as internal
and external references, respectively. IR spectra were
recorded as KBr discs on a Bio-Rad FTS 6000 FT-IR
spectrometer. The elemental analyses were performed by
MikroKemi AB, Uppsala, Sweden or by H. Kolbe Mi-
* Corresponding author. Tel.: +46-46-22-28153; fax: +46-46-22-
24439.
E-mail address: ola.wendt@inorg.lu.se (O.F. Wendt).
1
€
kroanalytisches Laboratorium, Mulheim an der Ruhr,
Germany.
€
Present address: Fachbereich Physik, Freie Universitat, Berlin,
Arnimallee 14, D-14195 Berlin, Germany.
0020-1693/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2003.10.027

1296
J. Pushkar, O.F. Wendt / Inorganica Chimica Acta 357 (2004) 1295–1298
SAINT [9]. The structure was solved by direct methods
2.2. Preparations
and refined by full matrix least-square calculations on F2
using SHELXTL 5.1 [10]. Non-H atoms were refined with
anisotropic displacement parameters. All hydrogen at-
oms except those on the water molecule were con-
strained to parent sites, using a riding model. Friedel
pairs were not merged. Based on the systematic extinc-
tions either of the two space groups Pc and P2=c could
be chosen and we initially solved the structure in Pc
based on the value of jE ꢂ E ꢀ 1j (0.605), which indi-
cates a non-centrosymmetric space group. This was
shown to be the correct choice, since there is no twofold
axis present in the refined structure.
cis-Pd(Me)(O2COH)(tmeda) (4). Under an atmo-
sphere of nitrogen a small autoclave was charged with 1
(0.100 g, 0.791 mmol) and dry benzene was added. The
reaction vessel was pressurised with approximately 10
atm of CO2 straight from the gas cylinder and stirred
over-night at room temperature. This gave 0.062 g (52%)
of a white, air-stable solid which was collected on a frit
and washed twice with benzene. NMR (CD3OD, RT):
1H d 0.38 (br, 3H) 2.49 (br, 6H) 2.56 (m, 2H) 2.64 (s,
6H) 2.77 (m, 2H). IR m(C–O) 1621 cmꢀ1. Anal. Calc. for
C8H20N2O3Pd: C, 32.2; H, 6.7; N 9.4; O 16.1. Found: C,
32.2; H, 7.1; N 9.2; O 16.0%.
1
2.4.1. Crystal data and collection and refinement details
C7H18N2O4Pd, M ¼ 300:63, monoclinic, a ¼ 5:8716
cis-Pd(Me)(O2COH)(dppe) ꢁ C6H6 (5). Startingfrom
2
0.129 g of 2 and using a similar procedure as for 4 gave
0.039 g (26%) of 5 as an off-white, air-stable solid. NMR
(CD3OD): 1H d 0.54 (dd, 3JPH ¼ 7.2 and 2.7 Hz, 3H) 2.21
(m, 2H) 2.51 (m, 2H) 7.35–7.85 (m, 20H) 31P {1H} d 36.9
(d, 2JPP ¼ 62 Hz) 62.1 (d) 13C {1H} d 161.4 (s). IR m(C–O)
ꢀ
ꢀ
ꢀ
ð12Þ A, b ¼ 6:9687ð14Þ A, c ¼ 13:600ð3Þ A, b ¼ 95:72ð3Þ°
3
ꢀ
V ¼ 553:69ð19Þ A , space group Pc (no. 7), Z ¼ 2,
l ¼ 1:669 mmꢀ1, 5567 reflections measured, 3301 unique
(Rint ¼ 0:0496) which were used in all calculations. The
final wRðF 2Þ was 0.0910 and the S value 0.999 (all data).
The RðF Þ was 0.0461 (I > 2rðIÞ).
1
1613 cmꢀ1. Anal. Calc. for C28H28O3P2 Pd ꢁ C6H6: C,
2
60.06; H, 5.04. Found: C, 60.16; H, 5.02%.
cis-Pd(O2CO)(dppe) (6). In a round-bottom flask
0.049 g of 3 was dissolved in 20 ml of CH2Cl2. Silver
carbonate (0.029 g) was added and the reaction mixture
was stirred overnight with exclusion of light. The silver
chloride formed was filtered off and the filtrate was
stripped from solvent and recrystallised from CH2Cl2/
pentane giving 0.020 mg of 6. NMR (C6D6): 1H d 2.67 (s,
4H) 6.89–7.01 (m, 12H) 7.64–7.76 (m, 8H) 31P {1H} d
30.2 (s).
3. Results and discussion
3.1. Reactivity towards CO2
All reactions are depicted in Scheme 1. Pressurising a
solution of 1 or 2 in dry C6D6 in a J. Young NMR-tube
with approximately 4 atm of CO2 gave a precipitate
after one day. In solution it was observed that the sig-
nals for the starting materials decreased in intensity but
apart from small signals from unidentified decomposi-
tion products no new product was observed. A peak
always appearing at 0.29 ppm was assigned to methane.
Running the reaction on a preparative scale gave
2.3. NMR experiments
Typically, a J. Young NMR-tube was charged with a
few mg of the complex and evacuated on the high-vac-
uum line. The dry NMR-solvent was then vacuum-
transferred onto the solid at )78 °C. If applicable, CO2
was condensed in at )196 °C to give a pressure of 4–5
atm. The tube was thawed and any reaction was allowed
to evolve.
L
L
O
O
L
L
O
O
Pd
C O
Pd
C O
6 (L-L = dppe)
7 (L-L = tmeda)
2.4. Structure determinations
-
HCO3 /
H2O
Recrystallisation of 4 from acetone under ambient
conditions gave needle-like crystals that were mounted
in a capillary. The X-ray analysis showed that the re-
crystallisation had given cis-Pd(O2CO)(tmeda) (7),
which co-crystallised with one mole of water. Intensity
data were collected at 293 K with a Bruker Smart CCD
system using x scans and a rotating anode with Mo Ka
recryst./
acetone
H2O/
C6D6
CO2/H2O
dry C6H6
L
L
Me
Me
L
L
O2COH
Me
Pd
Pd
1 (L-L = tmeda)
2 (L-L = dppe)
4 (L-L = tmeda)
5 (L-L = dppe)
ꢀ
radiation (k ¼ 0:71073 A) [7]. The intensity was cor-
dry CO2/
dry C6H6
rected for Lorentz, polarisation and absorption effects
using SADABS [8]. The first 50 frames were collected
again at the end to check for decay. No decay was ob-
served. All reflections were merged and integrated using
Scheme 1.

J. Pushkar, O.F. Wendt / Inorganica Chimica Acta 357 (2004) 1295–1298
1297
products 4 and 5 in 25–90% yield. Typical yields are
given in the experimental section; the variation in yield is
probably due to different amounts of water present.
Both products are air-stable white to off-white solids. In
a parallel experiment the carbon dioxide was dried by
passing it through a )78 °C trap on the vacuum-line
before admitting it to the NMR tube. In this case no
precipitate was observed and there was no sign of any
reaction in solution. Compounds 4 and 5 turned out to
be difficult to analyse in solution; they are insoluble in
hydrocarbon and ether solvents and quickly decompose
to the corresponding dichlorides, L2PdCl2, in chlori-
nated solvents.
In dry CD3OD 4 is stable enough to obtain a 1H
NMR spectrum. The complex is asymmetric as seen by
the fact that the two sides of the tmeda ligand are in-
equivalent. There is also a peak at 0.38 ppm assigned to
a Pd–Me group. No other ligands (apart from tmeda
and Me) could be identified in solution, but the ele-
mental analysis speaks in favour of a bicarbonate
complex and this is also supported by the IR-spectrum
which displays a m(C–O) at 1621 cmꢀ1. Furthermore, it
is clear that there is a dynamic process in solution
(broad peaks for the methyl group and the N–Me peaks
on one side) and this is probably an exchange between
coordinated bicarbonate and methanol solvent. Re-
crystallisation of 4 gave the corresponding g2-carbon-
ate, 7 (vide infra).
gation 5 was chosen. To study the possible formation of
an OH-complex, a C6D6 solution of 5 was saturated
with water by addition of 10 ll of water and the reaction
1
was followed by H and 31P NMR spectroscopy. After
10 days there was still more than 90% starting material
in addition to 3 peaks from unidentified decomposition
products, none of which contained the spectral signature
2
(e.g., JPP) of a cis-L2PdMe(OH) complex. Addition of
KHCO3 slowly (several days) yielded methane and one
symmetric metal species that was identified as 6 by
comparison with an authentic sample.
It therefore seems that the presence of water in a
benzene solution of the dimethyl complexes does not
lead to formation of a hydroxide to any measurable
extent. This is in agreement with a recent investigation
of the hydrolysis of similar platinum complexes, which
only undergo reaction under acidic conditions [13]. The
only way the bicarbonate could form via the hydroxide
is thus if there is a fast pre-equilibrium between the
hydroxide and dimethyl complex followed by a slower
CO2 insertion. This seems highly unlikely since that
would involve a fast and reversible C–H activation of
methane. We therefore favour a mechanism in which
molecular carbonic acid is formed from carbon dioxide
and water and this species (or possibly a proton fol-
lowed by free bicarbonate) attacks the dimethyl complex
to give the bicarbonate or carbonate complex along with
methane. It seems that the amount of water present is
crucial in determining whether a bicarbonate or a car-
bonate complex is formed. Although in principle the
methyl bicarbonate complex could rearrange into an
Dissolving 5 in dry CD3OD gives rise to static 1H and
31P NMR spectra. Just as 4, the complex has C1-sym-
metry with inequivalent sides of the dppe ligand as seen
by for example the appearance of two doublets in the
1
31P NMR spectrum. In the H spectrum there is one
peak at 0.54 ppm (dd) corresponding to one methyl
group. The bicarbonate ligand was not observed in the
1H spectrum, presumably due to H/D exchange, but in
the carbon spectrum there is a peak at 161.4 ppm as-
signed to the bicarbonate. Attempts at obtaining X-ray
quality crystals invariably gave a microcrystalline ma-
terial which diffracted poorly. The static behaviour of 5
as compared to 4 is probably due to steric reasons; the
sterically demanding dppe ligand slows down the attack
by the solvent.
The type of reactions reported here are not unprece-
dented. Crutchley et al. [11] report a similar reaction, in
which a trans-bis(phosphine)-bicarbonatomethyl com-
plex is formed from the reaction of CO2 with the cor-
responding dimethyl complex in the presence of water.
Also, it was recently reported that the reaction of a
(tmeda)Pd(OAr)2 complex with CO2 and water gave a
Pd g2-carbonate complex [12].
It has been suggested that the formation of bicar-
bonate from a methyl ligand takes place via water hy-
drolysis to give a hydroxide and methane; the hydroxide
subsequently undergoes insertion of CO2 [11]. We de-
cided to probe the mechanism and for a closer investi-
Fig. 1. Diamond drawing with atomic numbering of 7 ꢁ H2O. The
ꢀ
thermal ellipsoids denote 30% probability. Bond distances (A) and
angles (deg) with estimated standard deviations: Pd–N1 ¼ 2.060(7);
Pd–N2 ¼ 2.041(5); Pd–O2 ¼ 2.015(5); Pd–O3 ¼ 2.008(6); C7–O1 ¼
1.221(8);C7–O2 ¼ 1.335(7); C7–O3 ¼ 1.317(9); N1–Pd–N2 ¼ 84.9(2);
N1–Pd–O2 ¼ 103.4(2); N2–Pd–O3 ¼ 106.2(2); O2–Pd–O3 ¼ 65.4(2);
O1–C7–O3 ¼ 126.3(7); O1–C7–O2 ¼ 123.5(6); O3–C7–O2 ¼ 110.2(6).

1298
J. Pushkar, O.F. Wendt / Inorganica Chimica Acta 357 (2004) 1295–1298
g2-carbonate complex (and split off methane) this does
not seem to happen unless there is additional water
present as for example in the recrystallisation of 4.
Finally, we can conclude that the present dimethyl
complexes are not suitable for CO2 activation in order
to make new C–C bonds. We are currently investigating
more electron rich Pd complexes in this laboratory.
Acknowledgements
Financial support from the Swedish Research
Council, The Crafoord Foundation and the Royal
Physiographic Society is gratefully acknowledged.
References
3.2. Crystal structure
[1] (a) See for example: H. Arakawa, M. Aresta, J.N. Armor, M.A.
Barteau, E.J. Beckman, A.T. Bell, J.E. Bercaw, C. Creutz, E.
Dinjus, D.A. Dixon, K. Domen, D.L. DuBois, J. Eckert, E.
Fujita, D.H. Gibson, W.A. Goddard, D.W. Goodman, J. Keller,
G.J. Kubas, H.H. Kung, J.E. Lyons, L.E. Manzer, T.J. Marks, K.
Morokuma, K.M. Nicholas, R. Periana, L. Que, J. Rostrup-
Nielson, W.M.H. Sachtler, L.D. Schmidt, A. Sen, G.A. Somorjai,
P.C. Stair, B.R. Stults, W. Tumas, Chem. Rev. 101 (2001) 953;
(b) W. Leitner, Coord. Chem. Rev. 153 (1996) 257;
The crystals formed on recrystallisation of 4 from
acetone consist of discrete molecules of 7 packed by
dispersive forces together with one water molecule per
complex. The molecular structure is shown in Fig. 1.
The coordination geometry around Pd is distorted
square-planar with a mean deviation from a least-
ꢀ
squares plane through PdN2O2 of 0.0023 A. The angles
(c) X. Yin, J.R. Moss, Coord. Chem. Rev. 181 (1999) 27.
[2] H. Kolbe, J. Prakt. Chem. 10 (1874) 89.
[3] R. Schmitt, J. Prakt. Chem. 31 (1885) 397.
deviate substantially from 90°, the smallest being the O–
Pd–O angle which is 65.4(2)° due to the g2-coordination
of the carbonate. The carbonate moiety is planar and
has O–C–O angles ranging between 110 and 126°. The
C–O double bond is clearly localised with a C7–O1
[4] (a) R. Santi, M. Marchi, J. Organomet. Chem. 182 (1979) 117;
(b) A. Behr, Angew. Chem. Int. Ed. Engl. 27 (1988) 661;
(c) M. Shi, K.M. Nicholas, J. Am. Chem. Soc. 119 (1997) 5057.
€
[5] D.J. Darensbourg, G. Grotsch, P. Wiegriffe, A.L. Rheingold,
Inorg. Chem. 26 (1987) 3827.
ꢀ
ꢀ
distance of 1.221(8) A which is some 0.1 A shorter than
the other two C–O distances. The molecular structure of
7 was recently reported but in crystals of a different
space group with no water molecules in the unit cell [12].
Angles and distances in [12] are similar to those pres-
ently reported.
[6] W. de Graaf, J. Boersma, W.J.J. Smeets, A.L. Spek, G. van
Koten, Organometallics 8 (1989) 2907.
[7] BrukerAXS, SMART, Area Detector Control Software, Bruker
Analytical X-ray System, Madison, WI, USA, 1995.
[8] G.M. Sheldrick, SADABS, Program for absorption correction,
€
University of Gottingen, Germany, 1996.
[9] BrukerAXS, SAINT, Integration Software, Bruker Analytical X-
ray System, Madison, WI, USA, 1995.
[10] G.M. Sheldrick, SHELXTL 5.1, Program for structure solution and
€
4. Supplementary material
least square refinement, University of Gottingen, Germany, 1998.
[11] R.J. Crutchley, J. Powell, R. Faggiani, C.J.L. Lock, Inorg. Chim.
Acta 24 (1977) L15.
Complete crystallographic data in the CIF format
have been deposited with the CCDC (no. 220158). In
addition, tables of structure factors have been deposited
and are available from the authors upon request.
[12] H. Yasuda, J.-C. Choi, S.-C. Lee, T. Sakakura, Organometallics
21 (2002) 1216.
[13] D.W. Lucey, D.S. Helfer, J.D. Atwood, Organometallics 22
(2003) 826.