Anion effects in the formation of the active catalyst in the Ruhrchemie - Rhone-Poulenc aqueous biphasic hydroformylation process. Are there any?

Article abstract of DOI:10.1139/v05-088

The water soluble [Rh(OAc)(CO)(mtppms)₂] containing monosulfonated triphenylphosphine ligands was prepared for the first time and its hydrogenation was studied in aqueous solutions. In the presence of additional mtppms, the reaction yielded [RhH(CO)(mtppms)₃], a close analogue of [RhH(CO)(mtppts)₃], the immediate catalyst precursor in the Ruhrchemie - Rhone-Poulenc aqueous biphasic hydroformylation process. The extent of the [Rh(OAc)(CO)(mtppms)₂] → [RhH(CO)(mtppms) ₃] transformation strongly depended on the solution pH, similar to the case of the hydrogenation of [RhCl(CO)(mtppms)₂] studied earlier. In this respect, RhCl₃·aq and Rh(OAc)₃·aq can be used equally well for the in situ preformation of [RhH(CO)(mtppts) ₃], although the latter is the preferred choice in the industrial process.

Full text of DOI:10.1139/v05-088

1033
RAPID COMMUNICATION / COMMUNICATION RAPIDE
Anion effects in the formation of the active
catalyst in the Ruhrchemie – Rhône-Poulenc
aqueous biphasic hydroformylation process. Are
there any?1
József Kovács, Ferenc Joó, and Carl D. Frohning
Abstract: The water soluble [Rh(OAc)(CO)(mtppms)2] containing monosulfonated triphenylphosphine ligands was pre-
pared for the first time and its hydrogenation was studied in aqueous solutions. In the presence of additional mtppms,
the reaction yielded [RhH(CO)(mtppms)3], a close analogue of [RhH(CO)(mtppts)3], the immediate catalyst precursor in the
Ruhrchemie – Rhône-Poulenc aqueous biphasic hydroformylation process. The extent of the [Rh(OAc)(CO)(mtppms)2] →
[RhH(CO)(mtppms)3] transformation strongly depended on the solution pH, similar to the case of the hydrogenation of
[RhCl(CO)(mtppms)2] studied earlier. In this respect, RhCl3·aq and Rh(OAc)3·aq can be used equally well for the in
situ preformation of [RhH(CO)(mtppts)3], although the latter is the preferred choice in the industrial process.
Key words: rhodium, water-soluble, hydrides, sulfonated phosphines, biphasic.
Résumé : On a préparé pour la première fois les complexes [Rh(OAc)(CO)(mtppms)2] qui sont solubles dans l’eau et
qui contiennent des ligands triphénylphosphines monosulfonées et on a étudié leur hydrogénation dans des conditions
aqueuses. En présence de mtppms additionnel, la réaction a conduit à la formation du [RhH(CO)(mtppms)3], un ana-
logue apparenté au [RhH(CO)(mtppts)3], le précurseur immédiat du catalyseur utilisé dans le procédé d’hydroformylation
biphasique aqueuse de la société Ruhrchemie – Rhône-Poulenc. Le degré de transformation de [Rh(OAc)(CO(mtppms)2] →
[RhH(CO)(mtppms)3] dépend fortement du pH de la solution, ce qui est semblable à ce qui a été observé antérieure-
ment dans le cas de l’hydrogénation du [RhCl(CO)(mtppms)2]. À cet effet, le RhCl3·aq et le Rh(OAc)3·aq peuvent être
utilisés aussi bien l’un que l’autre pour la préformation in situ du [RhH(CO)(mtppts)3], même si le dernier est préféré
dans le procédé industriel.
Mots clés : rhodium, soluble dans l’eau, hydrures, phosphines sulfonées, biphasique.
[Traduit par la Rédaction] Kovács et al. 1036
Introduction
[RhH(CO)(mtppts)3] (mtppts = 3,3′,3′′-phosphinetriylbenz-
enesulfonic acid, meta-trisulfonated triphenylphosphine) (4).
In practice, the catalyst is preformed from suitable precur-
sors, most commonly from hydrated rhodium(III) acetate
and mtppts under synthesis gas pressure, and used without
isolation (5). Under actual hydroformylation conditions,
[RhH(CO)(mtppts)3] is involved in various chemical equilib-
The Ruhrchemie – Rhône-Poulenc process for the aque-
ous/organic biphasic hydroformylation of propene (1–3) was
introduced into the industrial scene more than two decades
ago. The key to the success of this process was the invention
of an utterly hydrophilic, active, and stable catalyst
Received 11 January 2005. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 12 May 2005.
Dedicated to Professor Howard Alper in recognition of his outstanding achievements in organic and organometallic chemistry and
catalysis.
J. Kovács. Homogeneous Catalysis Research Group, Hungarian Academy of Sciences, University of Debrecen, P.O. Box 7, 4010
Debrecen, Hungary.
F. Joó.2 Homogeneous Catalysis Research Group, Hungarian Academy of Sciences and Institute of Physical Chemistry, University
of Debrecen, P.O. Box 7, 4010 Debrecen, Hungary.
C.D. Frohning. Regnitstr. 50, 46485 Wesel, Germany.
1This article is part of a Special Issue dedicated to Professor Howard Alper.
2Corresponding author (e-mail: fjoo@delfin.unideb.hu).
Can. J. Chem. 83: 1033–1036 (2005)
doi: 10.1139/V05-088
© 2005 NRC Canada

1034
Can. J. Chem. Vol. 83, 2005
ria, leading to the formation of [RhH(CO)2(mtppts)2],
[RhH(CO)2(mtppts)], and [RhH(CO)(mtppts)2] (2–5). Of
these compounds, [RhH(CO)(mtppts)2] is the active catalyst
for the formation of linear products, and these equilibria
have been studied in detail (6) to obtain optimum rate and
selectivity performance of the in situ formed catalyst.
Despite the continuous research and development, a few
questions with regard to the reaction mechanism still deserve
scrutiny. One of these is the effect of pH on the rate of
hydroformylation. It is known that an increase in pH from
5.5 to 7.5 brings about a 35% increase in the reaction rate of
the hydroformylation of propene (1). A similar rate increase
was observed in the reaction of octene-1 (7). Since the
chemical reaction does not involve H+ or OH–, the answer
may involve the pH-sensitivity of the formation and further
equilibria of the catalytically active rhodium complexes.
Some time ago, it was shown that the hydrogenation of
the related, water-soluble [RhCl(CO)(mtppms)2] complex
and 85% H3PO4 (31P), respectively. Conditions of the pH-
potentiometric measurements are described in detail in
refs. 9 and 10.
RhCl3·aq was purchased from the Pressure Chemical Co.
(Pittsburgh, Pennsylvania), D2O and DMSO-d6 from Cam-
bridge Isotope Laboratories, and H2, N2, Ar, and CO from
Linde (Hungary). Other reagents were obtained from
Aldrich and used without further purification. mtppms (Na+-
salt) (11) and [Rh(OAc)2]2·2MeOH (12) were prepared as
described in the literature.
[Rh(OAc)(CO)(mtppms)2] was prepared by a modification
of the literature method (13) for the synthesis of
[Rh(OAc)(CO)(PPh3)2]. To a suspension of 0.1 g (0.4 mmol
Rh) [Rh(OAc)2]2·2MeOH in 10 mL ethanol was added
0.25 mL of 50% HBF4 and the solution was stirred at 60 °C
for 20 h. To the resulting green solution, a solution of
640 mg (1.6 mmol) mtppms and 545 mg (4 mmol)
NaOAc·3H2O in 10 mL ethanol was added at room tempera-
ture and the mixture was refluxed for 3 h. An orange solid
separated that was filtered and washed with ethanol. This
raw product ([Rh(OAc)(mtppms)3] together with inorganic
impurities, NaOAc, NaBF4) can be directly used for further
synthesis. It was suspended in 5 mL of fresh ethanol, and
bubbled with CO at room temperature, upon which the color
changed gradually to orange then to yellow with dissolution
of the solid. The reaction with CO overnight gave a cream-
coloured precipitate. The complex was filtered, washed with
ethanol, and dried. Weight: 300 mg (67% overall yield). IR
(cm–1): 1978 ν(CO), 1399 νs(COO), 1579 νas(COO), 1038,
(mtppms
=
(3-diphenylphosphino)benzenesulfonic acid,
meta-monosulfonated triphenylphosphine) in aqueous solu-
tion was, indeed, governed by the pH (8). In the presence of
excess mtppms and chloride, the hydrogenation of [RhCl(CO)-
(mtppms)2] to yield [RhH(CO)(mtppms)3] did not take place
below pH 4.5 but reached 100% conversion at pH 10. Impor-
tantly, at basicities between pH 5 and 8, the conversion in-
creased gradually with the gradual increase of pH, and that
referred to a pH-dependent mobile equilibrium (eq. [1]).
[1]
[RhCl(CO)(mtppms)2] + mtppms + H2
= [RhH(CO)(mtppms)3] + H+ + Cl–
1
1195 ν(SO3). H NMR (360 MHz, DMSO-d6, r.t., ppm) δ:
0.67 (s, CH3COO), 6 to 7 (m, aromatic protons). 31P NMR
(360 MHz, DMSO-d6, r.t., ppm) δ: 35.3 (d, 1J(Rh-P) =
138 Hz). Anal. calcd. (%): C 47.28, H 3.97, S 6.47; found: C
46.30, H 4.07, S 6.49.
Although the conditions of this model study and those of the
industrial hydroformylation strongly differ, one may expect
similar equilibria be involved in the pH dependence of the
hydroformylation reaction. However, albeit hydrated RhCl3
is
a suitable starting material for the synthesis of
Results and discussion
[RhH(CO)(mtppts)3], industry prefers Rh(OAc)3·aq (OAc– =
CH3COO–) for the preformation of the catalyst to avoid the
presence of the strongly nucleophilic chloride (4, 5). On the
contrary, the pH-static hydrogenations [8] were performed at
constant ionic strength maintained by KCl. The question
arose, therefore, whether it is justified to draw conclusions
regarding the pH effects on the formation of [RhH(CO)-
(mtppts)3] based on studies of chloride-containing systems,
or to put it another way, whether there is an effect of the
anion of the precursor on the formation of this hydro-
formylation catalyst. In the following, we report the results
of a pH-static hydrogenation study of [Rh(OAc)(CO)-
(mtppms)2] under chloride-free conditions.
[Rh(OAc)(CO)(mtppms)2] was synthesized for the first
time by a modified procedure of Mitchell et al. (13) for the
synthesis of [Rh(OAc)(CO)(PPh3)2], based on the reaction of
4+
the green dirhodium(II) cation (Rh2 with mtppms) fol-
lowed by carbonylation with CO. Replacement of PPh3 with
mtppms in the procedure described by Spencer (14) for the
preparation of [Rh(OAc)(PPh3)3] from RhCl3·aq did not
yield the expected acetatorhodium(I) complex, but afforded
[RhCl(mtppms)3]. The cream-coloured [Rh(OAc)(CO)-
(mtppms)2] dissolves well in water and in DMSO, but is
only slightly soluble in ethanol and insoluble in apolar or-
ganic solvents. According to the IR and NMR spectra, it has
a square-planar geometry with a monodentate acetato ligand
(νas(COO) – νs(COO) = 180 cm–1) (15) and with the phos-
phine ligands in the trans position.
Experimental
All manipulations were done under an argon or nitrogen
atmosphere. All solvents were purified by distillation and
carefully deaerated before use.
Aqueous solutions of [Rh(OAc)(CO)(mtppms)2] react
with H2 and in the presence of excess mtppms yield the
known [RhH(CO)(mtppms)3] (8, 16). This reaction was stud-
ied in a pH-static hydrogenation apparatus as described in
detail for [RhH(CO)(mtppms)3] (8). Briefly, the pH of the
deoxygenated solvent (0.1 mol/L NaClO4 to maintain con-
stant ionic strength) was set to a given value in the range 4 <
pH < 10 using HClO4 or KOH, and a known amount of the
complex was dissolved in it under an argon atmosphere. De-
IR spectra were recorded on a PerkinElmer Paragon 1000
1
PC FT-IR spectrometer in KBr discs. H and 31P NMR spec-
tra were recorded on Bruker AV 360 equipment in D2O or
(CD3)2SO (DMSO-d6). Chemical shifts are referenced to re-
sidual solvent peaks further referenced to external 2,2-
dimethyl-2-silapentane-5-sulfonate (DSS) sodium salt (1H)
© 2005 NRC Canada

Kovács et al.
1035
Fig. 2. Proton production upon dissolution (᭹, ᭿) and subse-
quent hydrogenation (᭡, ᭜) of [Rh(OAc)(CO)(mtppms)2] (᭹, ᭡)
and [RhCl(CO)(mtppms)2] (᭿, ᭜) at various pH. Conditions (᭹,
᭡): [Rh] = 1.8 mmol/L, [mtppms] = 5.4 mmol/L, [NaClO4] =
0.1 mol/L, p(H2) = 1 bar (1 bar = 100 kPa), T = room tempera-
ture; (᭿, ᭜): [Rh] = 2.4 mmol/L, [mtppms] = 7.2 mmol/L,
[NaClO4] = 0.1 mol/L, p(H2) = 1 bar (1 bar = 100 kPa), T =
35 °C.
Fig. 1. Proton production upon dissolution (᭹) and subsequent
hydrogenation (᭡) of [Rh(OAc)(CO)(mtppms)2]. [Rh] =
1.8 mmol/L, [mtppms] = 5.4 mmol/L, [NaClO4] = 0.1 mol/L,
pH = 7.0, p(H2) = 1 bar (1 bar = 100 kPa), T = room tempera-
ture.
0.6
0.5
0.4
0.3
0.2
0.1
0
1
0,8
0,6
0,4
0,2
0
0
5
10
15
20
25
30
t (min)
Scheme 1. Equilibria in solutions of [RhX(CO)P2] (X = OAc– or
Cl–, P = mtppms) under H2 in the presence of excess mtppms.
A
+ P
+ H2
[RhH(CO)P3] + H+ + X-
[RhX(CO)P2]
+H2
+ OH-
+P
- H2O
D
B
C
- X-
+ P -P
[Rh(OH)(CO)P2]
3
5
7
9
11
+ H2
pH
- H2O
E
[RhH(CO)P2]
increasing pH showing that the reaction goes more and more
extensively towards the formation of [RhH(CO)(mtppms)3],
reaching a practically complete conversion at pH 10. It is
important to note that all reactions in Scheme 1 are revers-
ible, and the pH-static hydrogenations reveal the extent to
which [Rh(OAc)(CO)(mtppms)2] (together with [Rh(OH)-
(CO)(mtppms)2]) are converted to [RhH(CO)(mtppms)3].
This extent is strongly pH dependent. Taking the closely
similar chemical properties of mtppms and mtppts, we may
assume that during the preformation of the catalyst in the
Ruhrchemie – Rhône-Poulenc process [RhH(CO)(mtppts)3]
takes part in similar equilibria. Since [RhH(CO)(mtppts)3] is
the immediate precursor of the true hydroformylation cata-
lysts (under syngas [RhH(CO)2(mtppts)] and [RhH(CO)-
(mtppts)2]), the actual proportion of rhodium that may enter
the catalytic cycle(s) is strongly influenced by the pH, and
this explains (at least in part) the higher catalytic activities
observed at pH 7.5 vs. 5.5.
pending on the actual pH, hydrolysis of the complex to
[Rh(OH)(CO)(mtppms)2], the mtppts analog of which is de-
scribed in ref. 17, occurred to a certain extent, resulting in
proton release to the solution. However, the acid formed in
this reaction was neutralized by KOH supplied by the
autoburette and the pH remained constant; furthermore, the
extent of hydrolysis could be calculated from the volume
(amount) of KOH consumed. Figure 1 shows such a “titra-
tion curve” and the first “step” on the graph refers to the hy-
drolysis of the complex under an inert atmosphere (at the
particular pH, i.e., 7.0 used in the experiment). Bubbling H2
through the equilibrated solution resulted in further KOH
consumption, indicating a heterolytic split of H2 in its reac-
tion with [Rh(OAc)(CO)(mtppms)2] (and with [Rh(OH)-
(CO)(mtppms)2] produced by the partial hydrolysis). This is
the second step on Fig. 1. The reactions involved are shown
on Scheme 1. Of particular interest is the fact that at this pH,
the [H+]/[Rh] ratio levels off at about 0.55 showing that only
55% (i.e., by far less than 100%) of all Rh reacted to yield
[RhH(CO)(mtppms)3].
Data of the pH-static hydrogenation of [RhCl(CO)-
(mtppms)2] are also shown in Fig. 2 for comparison. It can
be concluded that there is no substantial effect of the anionic
ligand (Cl– or OAc–) on the pH dependence of the formation of
[RhH(CO)(mtppms)3]. In aqueous solutions, the heterolytic split
of H2 by transition metal complexes (such as shown by rxn. [1]
and Scheme 1) is largely facilitated (18, 19) by the strong
Similar measurements were made at several pH values
and the results are shown on Fig. 2. It is seen that
[Rh(OAc)(CO)(mtppms)2] does not react with H2 below
pH 4. However, the [H+]/[Rh] ratio increases steadily with
© 2005 NRC Canada

1036
Can. J. Chem. Vol. 83, 2005
+
hydration of the resulting H , ∆Hhyd = –1091 kJ mol–1 (20),
and the particular anion. It seems that the relatively small
difference in the hydration ethalpies of Cl– and OAc–,
–381 kJ mol–1 (20) and –334 kJ mol–1 (21), respectively,
FJ) and by the European Commission (MRTN-CT-2003–
503864; AQUACHEM).
o
References
does not modify substantially the overall energy balance.
1. B. Cornils and W.A. Herrmann (Editors). Aqueous phase
organometallic catalysis — Concepts and applications. 2nd ed.
Wiley-VCH, Weinheim, Germany. 2004.
2. J. Herwig and R. Fischer. In Rhodium catalyzed hydro-
formylation. Edited by P.W.N.M van Leeuwen and C. Claver.
Kluwer, Dordrecht, The Netherlands. 2000. p. 189.
3. E. Kuntz. CHEMTECH, 17, 570 (1987).
4. B. Cornils. Top. Curr. Chem. 206, 133 (1999).
5. C.D. Frohning and C.W. Kohlpaintner. In Aqueous phase
organometallic catalysis — Concepts and applications. 2nd ed.
Edited by B. Cornils and W.A. Herrmann. Wiley-VCH,
Weinheim, Germany. 2004. p. 377.
6. I.T. Horváth, R.V. Kastrup, A.A. Oswald, and E.J. Mozeleski.
Catal. Lett. 2, 85 (1989).
The hydration enthalpy of OH , ∆Hhyd = –460 kJ mol–1 (20),
–
o
is higher than that of Cl– and OAc–; furthermore, formation
of H2O adds an additional driving force to reaction E
(Scheme 1), so [Rh(OH)(CO)(mtppms)2] is expected to react
preferentially over [Rh(OAc)(CO)(mtppms)2] and [RhCl(CO)-
(mtppms)2]. Whether [RhH(CO)(mtppms)3] (the resting state
of the catalyst at high phosphine concentrations) is formed
on path A, on path (B + C), or on path (A + E + D) is not
known from these investigations, but its equilibrium concen-
tration is pH-governed, independent of the way of forma-
tion. [Rh(CO)(OAc)(mtppms)2] shows a somewhat higher
tendency towards hydrolysis than [RhCl(CO)(mtppms)2], al-
though at room temperature, the extent of this hydrolysis
does not exceed 20%.
7. R.M. Deshpande, Purwanto, H. Delmas, and R.V. Chaudhari.
J. Mol. Catal. A: Chem. 126, 133 (1997).
Based on the above findings, it can be concluded that the
pH of the aqueous phase has the same effect on the catalytic
activity of Rh(I)-carbonyl-hydrides independent of the anion
of the precursor being chloride or acetate. Nevertheless, the
industrial preference for Rh(III)-acetate may be justified on
other grounds, e.g., less corrosion or more favourable kinet-
ics of the preformation of the catalyst. It is understood that
the chemistry going on in the industrial hydroformylation re-
actor is far more complex than that described in this commu-
nication, and that the operational conditions of the
Ruhrchemie – Rhône-Poulenc process for the aqueous/
organic biphasic hydroformylation of propene (120 °C,
50 bar syngas (1 bar = 100 kPa)) are far from those of our
measurements. This may influence the positions of the equi-
libria on Scheme 1, however, no other species were detected
under pressure in case of the related mtppts complexes (6).
Therefore, we think that the pH effects observed under more
realistic hydroformylation conditions (1, 7) can also be re-
lated to the base-promoted formation of the actual hydro-
formylation catalysts, [RhH(CO)2(mtppts)] and [RhH(CO)-
(mtppts)2], as described previously.
8. F. Joó, J. Kovács, A.Cs. Bényei, L. Nádasdi, and G.
Laurenczy. Chem. Eur. J. 7, 193 (2001).
9. F. Joó, J. Kovács, A.Cs. Bényei, and Á. Kathó. Angew. Chem.
Int. Ed. 37, 969 (1998).
10. F. Joó, J. Kovács, A.Cs. Bényei, and Á. Kathó. Catal. Today
42, 441 (1998).
11. F. Joó, J. Kovács, Á. Kathó, A.Cs. Bényei, T. Decuir, and D.J.
Darensbourg. Inorg. Synth. 32, 1 (1998).
12. G.A. Rempel, P. Legdzins, H. Smith, and G. Wilkinson. Inorg.
Synth. 13, 90 (1972).
13. R.W. Mitchell, J.D. Ruddick, and G. Wilkinson. J. Chem. Soc.
A, 3224 (1971).
14. A. Spencer. J. Organometal. Chem. 93, 389 (1975).
15. T.A. Stephenson, S.M. Morehouse, A.R. Powell, J.P. Heffner,
and G. Wilkinson. J. Chem. Soc. A, 3632 (1965).
16. A.F. Borowski, D.J. Cole-Hamilton, and G. Wilkinson. Nouv.
J. Chim. 2, 137 (1978).
17. W.A. Herrmann, J.A. Kulpe, W. Konkol, and H. Bahrmann. J.
Organometal. Chem. 389, 85 (1990).
18. B.R. James. Homogeneous hydrogenation. Wiley, New York.
1973.
19. F. Joó. Aqueous organometallic catalysis. Kluwer, Dordrecht,
The Netherlands. 2001.
20. D.W. Smith. J. Chem. Ed. 54, 541 (1977).
21. E. Gallicchio, L.Y. Zhang, and R.M. Levy. J. Comput. Chem.
23, 517 (2002).
Acknowledgements
This work was supported by the Hungarian National Re-
search Foundation (OTKA F037358 to JK and T043365 to
© 2005 NRC Canada