(Diphosphane monosulfide)platinum(II) complexes for hydroformylation reactions: Their catalytic activity and a high-pressure NMR mechanistic study

Article abstract of DOI:10.1002/ejic.200501016

The neutral complexes of formula [κ²P,S-(dppmS)Pt(CH ₃)-(Cl)] (1), [κ²P,S-(dppeS)Pt(CH₃)(Cl)] (2) and [κ²P,S-(dpppS)-Pt(CH₃)(Cl)] (3) [dppmS = Ph₂PCH₂P(S)Ph₂

Full text of DOI:10.1002/ejic.200501016

FULL PAPER
DOI: 10.1002/ejic.200501016
(Diphosphane monosulfide)platinum(II) Complexes for Hydroformylation
Reactions: Their Catalytic Activity and a High-Pressure NMR Mechanistic
Study
Antonio Lofù,[a] Piero Mastrorilli,*[a] Cosimo F. Nobile,[a] Gian P. Suranna,[a]
Piero Frediani,[b] and Jonathan Iggo[c]
Keywords: Heteroditopic ligands / Hydroformylation / NMR spectroscopy / P,S ligands / Platinum
The neutral complexes of formula [κ2P,S-(dppmS)Pt(CH3)-
(Cl)] (1), [κ2P,S-(dppeS)Pt(CH3)(Cl)] (2) and [κ2P,S-(dpppS)-
carbonyl complex [κ2P,S-(dppeS)Pt(CH3)(CO)]+ (9), which is
formed from 4a and CO, does not undergo insertion to give
the acetyl complex, even under 50 bar of syngas. Thus, the
role of SnCl2 is not only to create a vacant site for CO coordi-
nation and to lower the energy barrier for the hydro-
genolysis, but also to assist migration of the alkyl group in
the CO insertion step. High-pressure NMR studies of the
working reaction solution, under steady-state conditions,
found no evidence for intermediates in which the phosphane
sulfide group shows hemilabile behaviour during catalysis.
Pt(CH3)(Cl)] (3) [dppmS
= Ph2PCH2P(S)Ph2; dppeS =
Ph2P(CH2)2P(S)Ph2; dpppS = Ph2P(CH2)3P(S)Ph2] are active
catalyst precursors for the hydroformylation of 1-octene in
methyl isobutyl ketone. The order of reactivity found is
3 Ͼ 2 Ͼ 1. Surprisingly, the cationic complexes [κ2P,S-
(dppeS)Pt(CH3)(CH3CN)]BF4 (4a) and [{κP,µ-κS-(dppeS)-
Pt(CH3)}2][BF4]2 (5) are less active than the analogous neutral
complex 2. High-pressure NMR studies revealed that, at
20 °C under 1 to 50 bar of syngas, and in the presence of
SnCl2, both 2 and 4a react immediately to form the same ace-
tyl complex (8). However, in the absence of SnCl2, the methyl
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Introduction
acetyl intermediate. The use of chelating diphosphanes
brings about considerable activity enhancement, especially
in the case of dppb,[22] and Scrivanti et al. have performed
a mechanistic study intended to explain why the seven-
membered metallacycle is the most efficient.[23]
Hydroformylation of alkenes is one of the most impor-
tant catalytic reactions in industry, which runs plant pro-
cesses with Co- or Rh-based catalysts.[1] On the other hand,
there is an increasing interest in platinum-catalysed olefin
hydroformylation, mainly devoted to the optimisation of
chemo- and regioselectivity of the reaction and the synthe-
sis of linear aldehydes.[2]
Since Parshall’s pioneering work using PtCl2 in the ionic
liquid [(NEt4)SnCl3],[3] many catalytic systems,[4–7] includ-
ing some specifically suited to enantioselective cataly-
sis,[8–15] have been developed. A common feature of all these
systems is the use of SnCl2 in combination with the PtII
salt or complex. Several mechanistic studies on the catalytic
cycle[16–19] and on the role of the SnCl2[20,21] have appeared,
and these have concluded that the SnCl2 generates trichlo-
rostannate species by reaction with chloroplatinum() cata-
lyst precursors and promotes the hydrogenolysis of the Pt–
Heteroditopic ligands are chelating compounds endowed
with different chemical functionalities capable of coordinat-
ing to a metal centre. The very different trans effects pro-
duced by the dissimilar donor atoms in these ligands offer
the potential to introduce widely different activity and se-
lectivity at selected coordination sites in metal complexes
and has resulted in such ligands receiving increasing atten-
tion in the field of organometallic chemistry and cataly-
[25–31]
sis.[24] Interest in P S heteroditopic ligands
is driven
∧
by the ability of sulfur to coordinate strongly to soft metal
centres.[32,33] In the framework of our studies into the coor-
dination chemistry of heteroditopic ligands,[34,35] we re-
cently became interested in the monosulfides of diphos-
phane ligands (P PS)[36] and have studied the synthesis and
∧
reactivity of neutral and cationic methyl complexes of plati-
[a] Department of Water Engineering and of Chemistry (DIAC)
Polytechnic of Bari, and C.N.R. ICCOM-Sezione di Bari,
4 Via Orabona, 70125 Bari, Italy
∧
num() with the P PS heteroditopic ligands dppmS and
dppeS.[37]
Fax: +39-080-596-3611
The heteroditopic behaviour of the diphosphane monos-
E-mail: p.mastrorilli@poliba.it
∧
[b] Department of Organic Chemistry, University of Florence,
13 Via della Lastruccia, 50019 Sesto Fiorentino (FI), Italy
[c] Department of Chemistry, University of Liverpool,
P. O. Box 147, L69 7ZD Liverpool, UK
ulfide (P PS) ligands has recently been exploited in Rh-
and Ir-catalysed methanol carbonylation.[38,39] Consider-
ably less attention has been paid to the use of Pt complexes
2268
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 2268–2276

(Diphosphane monosulfide)platinum() Complexes for Hydroformylation Reactions
FULL PAPER
of heteroditopic ligands as catalysts for olefin hydrofor- equivalents of dpppS affords [trans-(κP-dpppS)2Pt(CH3)-
mylation.[7,40,41] Here we report our results using both neu- (Cl)] (6) in high yield, in which the two dpppS ligands are
tral and cationic methylplatinum complexes of diphosphane coordinated through the phosphorus atoms only.
monosulfide ligands in combination with SnCl2 as catalysts
for the hydroformylation of 1-octene. In particular, we have
Hydroformylations
focused our interest on the roles of the heteroditopic ligand
and the cocatalyst, and on the effects of chelating ring size
in/on the catalytic reaction. The study has been comple-
mented by high-pressure NMR experiments aimed at gain-
ing insights into the reaction mechanism.
Preformed complexes 1–6 were used as catalyst precur-
sors for the hydroformylation of 1-octene (Scheme 1) in the
presence of five equivalents of SnCl2·2H2O as cocatalyst, in
methyl isobutyl ketone as solvent (Table 1). Alternatively,
catalysts were formed in situ from [(cod)Pt(CH3)Cl] or
[(cod)PtCl2] and the appropriate ligand (this experimental
procedure has been validated; compare entries 4 and 7 with
entries 3 and 6 of Table 1). The progress of the reaction was
monitored by gas consumption. The reaction times of all
catalytic tests were kept below 4 h in order to minimise the
formation of side products deriving from aldol condensa-
tion between formed aldehydes and between aldehyde and
solvent. In all tests, the amount of condensation products
was less than 10%, while the regioselectivity towards nor-
mal aldehyde ranged from 79 to 91%.
Results and Discussion
Synthesis
The neutral and cationic methylplatinum() complexes of
diphosphane monosulfides used in this study are depicted
in Figure 1. Neutral complexes of general formula [κ2P,S-
∧
∧
(P PS)Pt(CH3)Cl] [P PS: Ph2PCH2P(S)Ph2, dppmS (1);
Ph2P(CH2)2P(S)Ph2, dppeS (2); Ph2P(CH2)3P(S)Ph2,
dpppS, (3)] were synthesised by reaction of [(cod)Pt-
(CH3)(Cl)] (cod = 1,5-cyclooctadiene) with one equivalent
of the relevant ligand.[37] Reaction of 2 with silver tetra-
fluoroborate in MeCN affords the cationic complex [κ2P,S-
(dppeS)Pt(CH3)(CH3CN)][BF4] (4a), in which the chloride
is replaced with acetonitrile. The cationic dimer [{κP,µ-κS-
(dppeS)Pt(CH3)}2][BF4]2 (5), in which the sulfur atoms
bridge the two platinum atoms, is obtained quantitatively
on heating solid 4a at 363 K under vacuum, as described
previously.[37] Reaction of [(cod)Pt(CH3)(Cl)] with two Scheme 1. Hydroformylation of 1-octene.
Figure 1. Cationic and neutral complexes used in catalysis.
Eur. J. Inorg. Chem. 2006, 2268–2276
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2269

A. Lofù, P. Mastrorilli, C. F. Nobile, G. P. Suranna, P. Frediani, J. Iggo
FULL PAPER
Table 1. Platinum-catalysed hydroformylation of 1-octene.[a]
Entry
Catalyst
t
[h]
Conversion
[%]
Yield of aldehydes
[%]
Regioselectivity[b]
[%]
TOF[c]
[h–1]
1
2
3
4
5
6
7
8
9
1
4
4
3.5
3
1
1.25
1
1
4
4
63
25
39
56
49
50
60
54
51
70
46
81
77
85
84
66
79
82
76
91
81
6
dppe + [(cod)Pt(CH3)Cl]
100
100
100
100
100
100
100
76
10
16
16
50
48
54
51
18
12
2
dppeS + [(cod)Pt(CH3)Cl]
dppp + [(cod)Pt(CH3)Cl]
3
dpppS + [(cod)Pt(CH3)Cl]
dppb + [(cod)Pt(CH3)Cl]
6
10
dppb + [(cod)Pt(CH3)Cl]
100
(2:1 mol/mol)
dppeS + [(cod)PtCl2]
dppeS + [(cod)PtCl2]
dppp + [(cod)PtCl2]
11
2
2.5
1
3.5
3.5
4
100
90
100
52
52
3
33
60
38
27
32
0
75
85
50
87
89
–
17
24
38
8
9
–
12[d]
13
14
4a
15
5
16[e]
4a
[a] Reaction conditions: olefin (1-octene: 5.8 mmol), olefin/Pt/Sn = 100:1:5; T = 353 K, P(CO) = P(H2) = 25 bar; solvent: methyl isobutyl
ketone (5.8 mL). The reaction times are those required to reach the stated conversions. [b] Selectivity: n-nonanal/total aldehydes. [c] TOF:
Turn over frequency as (mol aldehydes)(mol Pt)–1(time)–1. [d] Pt/SnCl2 = 1:1. [e] Reaction performed in the absence of SnCl2.
Effect of Ring Size
the n:iso ratio. The diminished electron density on Pt might
also be expected to retard hydrogenolysis thereby lowering
the activity of the catalyst, and this is indeed observed.
In agreement with previous reports on the effect of che-
late ring size in diphosphaneplatinum()-based catalytic
systems,[42] we found a pronounced increase in both activity
and selectivity as the ring size is increased. Thus, 1 (five-
membered platinacycle) affords incomplete substrate con-
version after 4 h, and 25% yield in nonanals (Table 1, en-
try 1), whereas with 2 (six-membered platinacycle) we ob-
served complete conversion after 3.5 h with a 56% yield in
nonanals (entry 3). Complete conversion after 1.25 h and
60% yield in aldehydes (Table 1, entry 6) was observed with
3, thereby clearly demonstrating the beneficial effects of a
larger metallacycle. This effect has previously been reported
and ascribed to a lowering in the activation energy for the
hydrogenolysis step.[22,23]
Effect of Ligand/Metal Ratio
We have also investigated the effect of the ligand/metal
∧
∧
ratio using both P PS and P P ligands. Thus, with [trans-
(κP-dpppS)2Pt(CH3)Cl] (6) as the catalyst precursor the re-
action gave the highest chemoselectivity towards hydrofor-
mylation products (70% aldehydes), the remainder being
octane and internal octenes, with the highest regioselectivity
towards n-nonanal (91%). However, the catalytic activity
was lower, with conversion of 76% in 4 h (Table 1, entry 9)
compared to that obtained using a ligand/metal ratio of 1
(Table 1, entry 6). A similar behaviour was observed for the
∧
P P system using a mixture of [(cod)Pt(CH3)Cl] and 1,4-
Effect of the Heteroditopic Ligand
bis(diphenylphosphanyl)butane (dppb) as catalyst precur-
sor (Table 1, entries 8 and 10). Once again, activity is lower
and selectivity to linear product higher. This is not surpris-
ing as the 2:1 ligand to metal ratio may lead to the blocking
of coordination sites, thereby reducing the activity,[5,42] and
increase the steric hindrance at the metal, thus improving
the regioselectivity.[5,44]
The performance of catalyst systems incorporating
∧
heteroditopic P PS ligands is comparable with that of anal-
∧
ogous systems containing P P ligands that give equally
sized platinacycles, although the P P complexes were found
∧
to be more stable under hydroformylation conditions, no Pt
black being observed in the products at the end of the reac-
[43]
∧
tion, in contrast to the P PS-containing systems.
Thus,
using catalysts formed in situ from [(cod)Pt(CH3)Cl] and
the appropriate ligand the P PS ligands give slightly more
Role of SnCl2
∧
regio- and chemoselective catalyst systems (compare en-
Given that cationic PtII species are normally invoked in
tries 1 and 2, 4 and 5, and 7 and 8 of Table 1), although the the catalytic cycle, and the role of the SnCl2 is widely be-
∧
P P ligands show comparable or greater activity. Replace- lieved to be the generation of such cationic species by ab-
–
ment of a diphosphane by a diphosphane monosulfide li- straction of Cl– to form SnCl3 , we tested the mono- and
gand on the catalyst precursor will reduce the electronic dinuclear cationic complexes 4a and 5 in the catalysis and
density at the metal centre, which should favour Markovni- were surprised to find that these complexes gave much
koff-type hydride attack on the olefin to give the linear Pt– lower activity and productivity than 2 (27% and 32% yield
alkyl intermediate and account for the observed increase in of nonanals after 3.5 h for 4a and 5, respectively, see en-
2270
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 2268–2276

(Diphosphane monosulfide)platinum() Complexes for Hydroformylation Reactions
FULL PAPER
tries 14 and 15 of Table 1) even though SnCl2 was present the catalytic reaction is known to proceed smoothly in ke-
in the reactions. The similar catalytic activities exhibited by tonic solvents.
4a and 5 suggest the formation of the same active species
from these cationic precursors, i.e. MeCN is not retained at
HP-NMR Behaviour of the Catalyst Precursors 2 and 4a
Pt and does not interfere in the catalysis. In the absence of
SnCl2 cocatalyst (Table 1, entry 16) 4a is almost inactive.
Thus, independent of the charge on the PtII precursor, the
addition of SnCl2 is necessary to trigger the hydrofor-
Compound 2 is poorly soluble in acetone. However,
upon addition of SnCl2 (Sn/Pt = 5:1), rapid dissolution of
the solids occurred, accompanied by a colour change from
pale straw yellow to dark orange, indicating the formation
of a new species (7a; Scheme 2), which we propose to be
–
mylation,[5,15] and the formation of SnCl3 (which is pos-
sible starting from 2 but not from 4a or 5 since chloride is
not available in the latter systems) further increases the re-
action rate.
1
the SnCl2 adduct of 2. Thus, the H NMR spectrum of 7a
(Table 2, entry 2) shows a resonance at δH = 0.43 ppm,
which can be assigned to a methyl group directly bound to
Pt. The 31P{1H} NMR spectrum of 7a shows resonances at
similar chemical shifts to those of 2 in CD2Cl2. However,
Interestingly, using [(cod)PtCl2] as the platinum source
(and 5 equiv. of SnCl2) affords catalyst systems that are less
chemo- and regioselective than those using [(cod)Pt-
(CH3)Cl], from which only one equivalent of SnCl3– can be
obtained by halide abstraction, as the Pt source (compare
entries 4 and 11 and 5 and 13 of Table 1). However, using
equimolar amounts of Sn and Pt (Table 1, entry 12) and
[(cod)PtCl2] as the Pt source affords higher chemo-and re-
gioselectivities, similar to those obtained using 2 with a Sn/
Pt molar ratio of 5:1. This seems to indicate a detrimental
effect of a Sn/Pt molar ratio higher than 1 when using
[(cod)PtCl2] as the Pt source.
3
1
the JP,P,
2JP,Pt and JP,Pt coupling constants are signifi-
cantly different from those of 2, clearly indicating the for-
mation of a new methylplatinum complex. The chemical
shifts and coupling constants (Table 2, entry 2) indicate that
the methyl group remains trans to the phosphane sulfide
1
donor. Thus, Tóth et al. have reported a JP,Pt value of
1628–1836 Hz for CH2PPh2 trans to methyl in the structur-
ally similar diphosphane complexes [(BDPP)Pt(CH3)(L)] (L
1
= Cl, SnCl3),[18] to be compared with JP,Pt = 3869 Hz in
7a. We can suggest at least three possibilities for the group
occupying the fourth coordination site at Pt in 7a: i) ace-
Summary of the Catalytic Results
–
tone, in which case SnCl3 is present as the counterion; ii)
∧
SnCl3– directly bonded to platinum; or iii) Pt-bound SnCl2.
We can discount the latter hypothesis since, in an otherwise
identical experiment but using one equivalent of SnCl2,
complex 7a was again obtained in high yield.
The use of heteroditopic P PS ligands in Pt/Sn olefin
hydroformylation systems affords catalysts that compare
favourably with analogous systems that use P P ligands.
∧
No significant variation in regioselectivity was observed
using cationic vs. neutral complexes as catalyst precursors,
thus indicating that regioselectivity is established in a PtII
intermediate that is identical for both cationic and neutral
∧
∧
catalyst precursors. For both P P and P PS systems, ac-
tivity is related to ring size, with seven-membered rings giv-
ing the most active systems. The correlation of activity with
ring size is consistent with a chelating structure being main-
tained in the rate-determining steps of the catalysis, al-
though it has been suggested that the dppb may, alterna-
tively, form oligomeric/polymeric metal species.[23] The
presence of SnCl2 is essential to the generation of effective
catalyst systems from both neutral and cationic precursors,
thus indicating that the role of SnCl2 is more than simply
the generation of an active site at Pt by removal of Cl–.
However, the presence of excess SnCl3– is detrimental to the
catalysis under our conditions.
Scheme 2.
In an attempt to differentiate the remaining possibilities
we measured the 31P{1H}NMR spectrum of 4a in [D6]ace-
tone since we believed that acetone would displace the coor-
dinated MeCN to give [Pt(dppeS)(CH3)(acetone)]+, the cat-
ion of the first hypothesis, as a result of mass action. In this
experiment, in addition to the resonances of 4a and its di-
mer 5,[37] two new resonances were observed in the
31P{1H}NMR spectrum at chemical shifts, and with coup-
High-Pressure NMR Study
A high-pressure NMR study of the reaction was per- ling constants, different to those of 2 + SnCl2 reported
formed using both a HP-NMR bubble column reactor de- above (compare entries 2 and 3 in Table 2). We assign these
veloped in Liverpool[45] and conventional sapphire tubes resonances to the new complex [Pt(dppeS)(CH3)(acetone)]+
1
with the aim of gaining additional insight into the mechan- (4b). The larger value of JP,Pt (5003 Hz in 4b vs. 4606 Hz
istic differences responsible for the reactivities of the neutral in 2) confirms the displacement of the acetonitrile by ace-
(2) and cationic (4a) complexes. Except where otherwise tone, a ligand with a weaker trans influence. These observa-
stated, [D6]acetone was used as solvent for this study since tions allow us to exclude 4b or 5 as the complex 7a and
Eur. J. Inorg. Chem. 2006, 2268–2276
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2271

A. Lofù, P. Mastrorilli, C. F. Nobile, G. P. Suranna, P. Frediani, J. Iggo
Table 2. Main 31P{1H} NMR spectroscopic data for the complexes (80 MHz, [D6]acetone if not otherwise specified, 290 K).
FULL PAPER
Entry
Complex formed
δP
δPS
3JP,P
2JP,Pt
1JP,Pt
Other data
1[37]
2
2 in CH2Cl2
2 + SnCl2
10.1
10.4
36.6
36.7
8
18
58
74
4606
3869
poorly soluble in acetone
[(dppeS)Pt(CH3)(SnCl3)] (7a)
1H NMR: δCH = 0.43 ppm
3
3
2
(d, JH,P = 6, JH,Pt = 76 Hz)
3
4
4a in acetone
[(dppeS)Pt(CH3)(acetone)]+ (4b)
(plus 4a and 5)
6.2
39.4
38.8
5
56
83
5003
4001
2 +SnCl2
+
13CO
[(dppeS)Pt{C(O)CH3}(L)]n+ (8)
–1.4
17
1H NMR: δCH = 1.77 ppm
3
(4.5 bar)
(s). 13C{1H} NMR: δCO
224 ppm (s, JC,Pt = 815 Hz)
=
1
5
2 + SnCl2 + syngas
(50 bar)
4a in CH2Cl2
8
–1.4
38.8
16
83
4005
6[37]
5
16.1
7.6
9.6
38
4331
4695
4036
3995
(plus 4a)
39.7
35.6
38.8
7
17
16
55
72
83
7
8
4a + SnCl2
[(dppeS)Pt(CH3)(SnCl2)]+ (7b)
4a + SnCl2 + syngas
(50 bar)
8
–1.4
9
4a + 13CO (6.7 bar)
[(dppeS)Pt(CH3)(CO)]+ (9)
10.8
40.5
12
59
3367
1H NMR: δCH = 0.57 ppm
3
3
(d, JH,P = 6.4 Hz. 13C{1H}
NMR: δCO = 177 ppm (d,
2JC,P = 147 Hz)
10
4a + syngas (50 bar)
2 (or 4a) + SnCl2 + 1-oc- 10
9
10.8
0.0
40.5
38.7
n.d.
n.d.
59
n.d.
3367
4208
11[a]
tene + syngas (50 bar)
plus 11
9.0
40.6
n.d.
n.d.
4001
[a] At 353 K; n.d. = not determinable.
–
support the hypothesis of SnCl3 directly bonded to plati- 20 min, when 31P{1H} NMR spectroscopy confirmed the
1
1
num. Moreover, the reduction of JP,Pt, from 4606 Hz in 2 formation of 8. The H NMR spectrum (Table 2, entry 4)
to 3869 Hz in 7a is consistent with displacement of chloride shows the disappearance of the Pt–CH3 signal of 7a to be
–
by SnCl3 , a ligand with a stronger trans influence. How- replaced by a sharp doublet at δ = 1.77 ppm attributable to
ever, the 119Sn{1H} spectrum of 7a shows no resonance in the methyl group of the acyl moiety (2JC,H = 5 Hz).[48] In
the region δ = –300 to +300 ppm over the temperature agreement with this assignment, the 13C{1H} NMR spec-
–
range 298–213 K, which suggests that the SnCl3 ligand is trum of this sample shows a sharp singlet at δC = 224 ppm
involved in a fluxional process.[46] Support for this hypothe- flanked by platinum satellites, as expected for a Pt–13C(O)
sis is the marked broadness of the 31P{1H} signal of the P CH3 group. The magnitude of the JP,Pt coupling constant
1
trans to this site (∆ν1/2 = 110 Hz, T = 290 K).
(4001 Hz) is not consistent with the acetyl group occupying
Finally, reaction of an acetone solution of 4a (comprising the site trans to P, for which Tóth has reported a value of
4a, 4b and 5) with SnCl2 gives a bright yellow solution of about 1400 Hz for the analogous diphosphane complexes
a new complex (7b) quantitatively (by 31P{1H} NMR spec- [(BDPP)Pt(COMe)L]n+ (L = Cl, n = 0;L = CO, n = 1).
troscopy; Table 2, entry 7). The NMR spectroscopic data Therefore, we propose that the acetyl group is trans to the
for 7b are similar to those of 7a with the exception of 1JP,Pt
,
sulfur donor,[18] a disposition also observed in related
which increases from 3869 Hz in 7a to 4036 Hz in 7b, thus dppeS–Rh complexes.[38]
indicating that 7a and 7b have a similar overall structure
The identity of the ligand occupying the fourth coordina-
but a different ligand, of similar trans influence, occupying tion site on Pt is more difficult to establish. The value of
the fourth coordination site at Pt. Since Cl– is not present 1JP,Pt in 8 is significantly smaller than that observed in 4b
in acetone solutions of 4a, thereby precluding the formation (5003 Hz), which seems to exclude acetone. No signal at-
–
of SnCl3 , we suggest this ligand is SnCl2,[47] i.e. 7a is the tributable to a Pt–CO group is observed in the 13C{1H}
adduct formed between [Pt(dppeS)(CH3)Cl] (2) and SnCl2, NMR spectrum, coupling between phosphorus and 13CO is
and 7b its cationic analogue [Pt(dppeS)(CH3)(SnCl2)]+. not observed, and the 31P NMR chemical shift and coup-
Again, no clear 119Sn NMR resonance could be observed ling constants are unchanged on purging the solution with
in the case of 7b, presumably due to fluxionality, as con- nitrogen,[49] observations that exclude both tightly bound
firmed by the broadness of the 31P{1H} NMR signals.
and labile CO as the ligand at the fourth site on Pt. Starting
Both 7a and 7b react under 50 bar of syngas at 290 K to from 4a, SnCl3– cannot be present, which therefore excludes
–
give 8 (quantitatively by 31P{1H} NMR spectroscopy), SnCl3 . However, an identical 31P{1H} NMR spectrum is
which is an acyl complex derived from CO insertion into obtained starting from 2 and one equivalent of SnCl2, when
the Pt–CH3 bond (Table 2, entries 5 and 8). In a separate SnCl2 is presumably reacted (Figure 2). Thus we cannot as-
experiment, 7a was treated with 13CO (4.5 bar) at 290 K in sign with certainty the ligand occupying the fourth site on
a sapphire tube. After pressurizing with 13CO, the dark- Pt in 8. This contrasts with the analogous P P acylplatinu-
∧
orange colour of 7a progressively lightened as 13CO dif- m() system[18,19] in which the fourth coordination site is
fused into solution to become straw yellow after about occupied by CO even at low pressure.
2272
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 2268–2276

(Diphosphane monosulfide)platinum() Complexes for Hydroformylation Reactions
FULL PAPER
NMR study of the working catalyst. Figure 3 shows the
31P{1H} HP-NMR spectra, recorded in our bubble column
reactor at 353 K and 50 bar of syngas, of solutions prepared
from either 2 or 4a (0.010 ), SnCl2 (5 equiv.) and 1-octene
(100 equiv.). The spectra of both solutions show resonances
of the same two species (10 and 11). The similarity of the
31P{1H} NMR spectroscopic data of 10 and 8 (Table 2, en-
tries 4 and 11) allow us to formulate 10 as the acyl complex
[Pt(dppeS){C(O)(C8H17)}(L)]n+, in which an octyl chain de-
rived from octene is present, as required by the catalysis.
1
The close similarity of JP,Pt in 8 and 10 is consistent with
the same L in both complexes. The identity of 11 (δP
=
1
9.0, δPS = 40.6 ppm; JP,Pt = 4001 Hz) is less certain. The
Figure 2. 31P{1H} HP-NMR spectra of the acyl complex 8 ob- significant difference in the 31P{1H} NMR spectra of 10
tained from 2 (a) or 4a (b) (entries 5 and 8 of Table 2 respectively).
and 11 is not consistent with 11 also being an acyl complex,
with 10 and 11 being the acyls derived from the n- and
The ready formation of the acetyl complex 8 from the
iso-alkyl intermediates. The similarity in the 31P{1H} NMR
reaction of either 2 or 4a with CO in the presence of SnCl2
spectroscopic data of 11 with those of the methyl complexes
(Scheme 3) contrasts with our previous report of the forma-
7a and 7b suggests 11 as the octyl intermediate
tion of [Pt(dppeS)(CH3)(CO)][BF4] (9), which is inert to
[Pt(dppeS)(C8H17)(L)]n+ (Scheme 5). The resonances of 10
methyl migration in CH2Cl2, in the absence of SnCl2.[37] We
and 11 were not observed in an analogous experiment using
have confirmed this result for 4a in acetone (Scheme 4),
4a as the catalyst precursor in the absence of SnCl2, consis-
which gives 9 even under 6.7 bar of 13CO or 50 bar of
tent with 10 and 11 being intimately associated with the
syngas (Table 2, entries 9 and 10). The dependence of the
working catalyst (the lack of reactivity in the absence of
SnCl2 is discussed above). Furthermore we changed the re-
acting gas from CO/H2 to H2 (50 bar), in order to convert
the working catalyst into a hydride species, but we observed
only the slow disappearance of the resonances of both 10
and 11. Furthermore, no high-field signals, ascribable to
hydrides, were observed in the 1H HP-NMR spectrum, con-
sistent with our attribution of these complexes to the non-
anoyl- and octylplatinum() resting states.
∧
CO insertion reaction on SnCl2 for P PS ligand complexes
contrasts with the situation for PtII complexes of diphos-
phane[18,19,50] or P N ligands,[24] for which no such require-
∧
ment is reported. The reluctance of CO to insert into the
Pt–alkyl bond in the absence of SnCl2 provides a ready ex-
planation for the initially surprising inactivity of the cat-
ionic complex 4a in olefin hydroformylation in the absence
of a co-catalyst (entry 6 of Table 1).
Scheme 3.
Figure 3. In situ 31P{1H} HP-NMR spectra showing the steady
state of catalytic reactions using 2 (a) or 4a (b).
Scheme 4.
The observation of both resting states is consistent with
both CO insertion and hydrogenolysis being slow in this
catalytic system, since the use of a bubble column reactor
HP-NMR Behaviour of the Working Catalyst
The understanding of the chemistry of the catalyst pre- allows us to eliminate artifacts resulting from poor gas de-
cursors outlined above provides a firm basis for an HP- livery to the reaction.
Eur. J. Inorg. Chem. 2006, 2268–2276
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2273

A. Lofù, P. Mastrorilli, C. F. Nobile, G. P. Suranna, P. Frediani, J. Iggo
FULL PAPER
Scheme 5.
temperature: 553 K; carrier: nitrogen). GCMS data (EI = 70 eV)
were acquired with an HP 6890 instrument using an HP
Conclusions
Complexes 1, 2 and 3 in the presence of SnCl2 are active
19091S-433
HP-5MS
5% phenylmethylsiloxane
column
catalyst precursors for the hydroformylation of 1-octene (30.0 m×250 µm×0.25 µm) coupled with an HP 5973 mass spec-
1
trometer (injector temperature: 553 K; carrier: helium; 70 eV). H,
13C{1H}, 31P{1H} and 119Sn{1H} NMR spectra were recorded with
a Bruker Avance 400 MHz. HPNMR spectra were recorded with a
Bruker AM200SWB spectrometer using a home built HP-NMR
bubble column reactor.
with yields ranging from 25 to 60% and regioselectivity in
linear aldehyde higher than 79%. The reactivity is depend-
ent on the ring size of the platinacycle, with the seven-mem-
bered ring being the most active. Tin halides have been
shown to fulfil three essential roles in the Pt/diphosphane
monosulfide/SnCl2 catalysed hydroformylation of higher al-
kenes: (i) removal of halide from the platinum centre as
Synthesis of [κ2P,S-(dpppS)Pt(CH3)Cl] (3): A solution containing
an equimolar amount of dpppS (129 mg) in CH2Cl2 (10 mL) was
added dropwise over about 2 h to a solution of [(cod)Pt(CH3)Cl]
(102.6 mg, 0.290 mmol) in CH2Cl2 (5 mL) and the mixture stirred
vigorously at room temperature overnight. The solvent was then
evaporated to about one fifth of the volume and addition of diethyl
ether caused the precipitation of a pale-yellow powder. Filtration
followed by washing with diethyl ether (3×5 mL) and drying under
vacuo afforded 3 in high yield (180 mg, 90%). M.p. 493 K (dec.).
C28H29ClP2PtS (690.07): calcd. C 48.73, H 4.24, S 4.65; found C
–
SnCl3 , which can then (ii) function as a labile ligand, and
(iii) for diphosphane-monosulfide-based catalyst systems,
tin() chloride plays an essential additional role in activa-
ting the alkyl complex for the CO insertion step, presum-
ably by coordination to the CO oxygen. The observation of
both the alkyl and acyl intermediates in solutions of the
working catalyst, and in the absence of the gas delivery
problems that are often encountered in sapphire NMR
tubes, indicates that, at least for the diphosphane mono-
sulfide studied here, there are two slow steps in the reaction.
It is interesting to note that tin halides appear to play a role
in accelerating both slow steps.
48.43, H 4.38, S 4.32. IR (KBr): ν = 3051 cm–1 (m), 2882 (m), 1481
(m), 1435 (vs), 1136 (vs), 927 (s), 834 (s), 743 (vs), 690 (vs), 582 (s,
P=S, str.), 518 (s), 275 (m, Pt–Cl). LC-MS: exact mass calcd. for
C28H29ClP2PtS: 689.08 amu; APCI; found 724.7 [M + Cl]–. 1H
NMR (400 MHz, CDCl3, 298 K): δ = 0.79 [dd, JH,P = 3.2, JH,PS
= 1.2, JPt,H = 74 Hz, 3 H, CH3], 1.76–1.97 (m, 2 H, CH2), 2.91–
˜
3
4
3
3.00 (m, 2 H, CH2), 3.04–3.14 (m, CH2), 7.35–7.50 (m, 6 H, Harom),
7.50–8.09 (m, 14 H, Harom.) ppm. 13C{1H} NMR (101 MHz,
2
1
CDCl3, 298 K): δ = –3.4 (d, JP,C = 6, JPt,C = 640 Hz, CH3), 18.3
(s, CH2) 24.6 (d, JP,C = 40 Hz, CH2P), 28.9 (m, CH2PS), 128.1–
133.3 (Carom.) ppm. 31P{1H} NMR (162 MHz, CDCl3, 298 K): δ =
4.3 (s, 1JP,Pt = 4565 Hz, PPh2), 37.3 (s, 2JPt,PS = 49 Hz, Ph2PS) ppm.
195Pt{1H}NMR (86 MHz, CDCl3, 298 K): δ = –4337.5 ppm (dd,
Experimental Section
All reactions (at least two replicates) were carried out under nitro-
gen using standard Schlenk techniques. Methyl isobutyl ketone was
refluxed with a little KMnO4, washed with aq. NaHCO3, dried
over CaCl2 and purified by passing through a small column of
activated alumina and freshly distilled prior to use. 1-Octene was
purchased from Acros and purified by percolation through a short
plug of neutral alumina prior to use.
2
1JPt,P = 4565, JPt,PS = 49 Hz) ppm.
Synthesis of [trans-(κP-dpppS)2Pt(CH3)Cl] (6): A solution of two
equivalents of dpppS (139 mg, 0.312 mmol) in CH2Cl2 (10 mL) was
added dropwise over about 2 h to a solution of [(cod)Pt(CH3)Cl]
(55.43 mg, 0.156 mmol) in CH2Cl2 (10 mL) and the mixture stirred
vigorously at room temperature overnight. The solvent was then
evaporated to about one fifth of the volume and addition of diethyl
ether caused the precipitation of a white powder. Filtration fol-
lowed by washing with diethyl ether (3×5 mL) and drying under
vacuo afforded 6 in high yield (159 mg, 90%). M.p. 508 K (dec.).
C55H55ClP4PtS2 (1134.6): calcd. C 58.22, H 4.89, S 5.65; found C
SnCl2·2H2O was purchased from Carlo–Erba and used as received.
The ligand dpppS was prepared by adapting the procedure pub-
lished for dppmS or dppeS.[36] Complexes 1, 2, 3, 4a and 5 were
synthesised as described.[37] Anhydrous SnCl2 for the mechanistic
study was purchased from Aldrich. CDCl3 was purchased from Al-
drich and [D6]acetone from Cambridge.
C,H,S elemental analyses were carried out with a Eurovector
CHNS-O Elemental Analyser. IR spectra were recorded with a
Bruker Vector 22 instrument. Chromatographic analyses were car-
ried out on Hewlett–Packard 6890 instruments using a 19091Z-236
HP-1 methylsiloxane capillary column (60.0 m×250 µm×1.00 µm)
58.37, H 4.92, S 5.61. IR (KBr): ν = 3052 cm–1 (m), 2935 (m), 1480
˜
(m), 1435 (vs), 1102 (vs), 953 (s), 803 (m), 749 (s), 691 (vs), 610 (s,
free P=S), 493 (s), 276 (m, Pt–Cl). LC-MS: exact mass calcd. for
1
C55H55ClP4PtS2: 1133.20 amu; APCI; found 1097.2 [M – Cl]+. H
3
2
NMR (400 MHz, CDCl3, 298 K): δ = –0.10 (t, JH,P = 4, JH,Pt
=
or
a
HP 19091J-413 HP-5 phenylmethyl siloxane column
73 Hz, 3 H, CH3), 2.03–2.17 (m, 4 H, CH2), 2.68–2.84 (m, 8 H,
(30.0 m×320 µm×0.25 µm; injector temperature: 553 K; FID
2CH2), δ = 7.29–7.85 (m, 40 H, Harom) ppm. 13C{1H} NMR
2274
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 2268–2276

(Diphosphane monosulfide)platinum() Complexes for Hydroformylation Reactions
FULL PAPER
2
(101 MHz, CDCl3, 298 K): δ = –12.8 (t, JC,P = 6, 1JC,Pt = 659 Hz,
CH3), 18.2 (s, CH2), 26.7 (m CH2–P), 33.5 (m, CH2PS), 128.2–
133.7 (Carom) ppm. 31P{1H} NMR (162 MHz, CDCl3, 298 K): δ =
[14] J. F. Cámer, A. Aaliti, N. Ruiz, A. M. M. Bultó, C. Claver, C. J.
Cardin, J. Organomet. Chem. 1997, 530, 199–209.
[15] L. Dahlenburg, S. Mertel, J. Organomet. Chem. 2001, 630, 221–
243.
[16] M. Gómez, G. Muller, D. Sainz, J. Sales, X. Solans, Organome-
tallics 1991, 10, 4036–4045.
[17] G. Cavinato, G. De Munno, M. Lami, M. Marchionna, L.
Toniolo, D. Viterbo, J. Organomet. Chem. 1994, 466, 277–282.
[18] I. Tóth, T. Kégl, C. J. Elsevier, L. Kollár, Inorg. Chem. 1994,
33, 5708–5712.
41.9 (s, Ph2PS), 21.5 (s, JP,Pt = 3050 Hz, Ph2P) ppm. 195Pt{1H}
NMR (86 MHz, CDCl3, 298 K): δ = –4578.4 ppm (t, JPt,P
1
1
=
3059 Hz) ppm.
Hydroformylation Experiments: In a typical experiment using pre-
formed catalysts, a solution of the platinum complex (0.058 mmol),
1-octene (5.8 mmol) and SnCl2·2H2O (0.29 mmol) in 5.8 mL of
methyl isobutyl ketone was transferred under nitrogen into a 50-
mL stainless steel autoclave. The reaction vessel was pressurised to
50 bar total pressure (CO/H2 = 1:1) and the magnetically stirred
mixture was heated to 353 K in a thermostated apparatus. The re-
action was monitored by following the drop in pressure.
[19] T. Kégl, L. Kollár, L. Radics, Inorg. Chim. Acta 1997, 265, 249–
254.
[20] G. K. Anderson, H. C. Clark, J. A. Davies, Inorg. Chem. 1983,
22, 427–433.
[21] H. J. Ruegg, P. S. Pregosin, A. Scrivanti, L. Toniolo, C. Bot-
teghi, J. Organomet. Chem. 1986, 316, 233–241.
[22] Y. Kawabata, T. Hayashi, I. Ogata, J. Chem. Soc., Chem. Com-
mun. 1979, 462–463.
For the “in situ” procedure, the ligand, dissolved when possible in
about 1.5 mL of methyl isobutyl ketone, was added to a solution
of the platinum precursor in about 1.5 mL of methyl isobutyl
ketone and the mixture was kept under vigorous stirring for 30 min.
After this time SnCl2·2H2O, 1-octene and methyl isobutyl ketone
(up to 5.8 mL of solvent) were added to the solution. The pressure
was monitored throughout the reaction. After cooling and venting
of the gas, the pale-yellow solution was immediately analysed by
GLC. Conversion of 1-octene and yield of aldehydes were calcu-
lated using dodecane as internal standard.
[23] A. Scrivanti, C. Botteghi, L. Toniolo, A. Berton, J. Organomet.
Chem. 1988, 344, 261–275.
[24] G. P. C. M. Dekker, A. Buijs, C. J. Elsevier, K. Vrieze,
P. W. N. M. van Leeuwen, W. J. J. Smeets, A. L. Spek, Y. F.
Wang, C. H. Stam, Organometallics 1992, 11, 1937–1948.
[25] S. O. Grim, J. D. Mitchell, Inorg. Chem. 1977, 16, 1762–1770.
[26] R. Colton, J. Ebner, B. F. Hoskins, Inorg. Chem. 1988, 27,
1993–1999.
[27] J. Browning, G. W. Bushnell, K. R. Dixon, R. W. Hilts, J. Or-
ganomet. Chem. 1993, 452, 205–218.
[28] T. C. Blagborough, R. Davis, P. Ivison, J. Organomet. Chem.
1994, 467, 85–94.
HP-NMR Experiments: In a typical experiment, a solution of
0.058 mmol of the desired Pt complex, together with the specified
amounts of SnCl2 and 1-octene (Table 2) in [D6]acetone (6 mL)
were injected into the HP-NMR bubble column reactor against a
counter stream of N2, CO or syngas, as appropriate. The reactor
was sealed, pressurised, and heated to the desired temperature
(Table 2), upon which the 31P{1H} NMR spectrum of the sample
was recorded.
[29] S. M. Aucott, A. M. Z. Slawin, J. D. Woollins, Eur. J. Inorg.
Chem. 2002, 2408–2418.
[30] S. M. Aucott, A. M. Z. Slawin, J. D. Woollins, Polyhedron 2003,
22, 361–368.
[31] D. E. Berry, J. Browning, K. R. Dixon, R. W. Hilts, Can. J.
Chem. 1988, 66, 1272–1282.
[32] J. R. Dilworth, N. Wheatley, Coord. Chem. Rev. 2000, 199, 89–
158.
[33] R. Romeo, L. Monsù Scolaro, M. R. A. Plutino, F. N. Romeo,
A. Del Zotto, Eur. J. Inorg. Chem. 2002, 629–638.
[34] I. Brassat, U. Englert, W. Keim, D. Keitel, S. Killat, G. P. Sur-
anna, R. Wang, Inorg. Chim. Acta 1998, 280, 150–162.
[35] I. Brassat, W. Keim, S. Killat, M. Moethrath, P. Mastrorilli,
C. F. Nobile, G. P. Suranna, J. Mol. Catal. A 2000, 157, 41–58.
[36] G. P. Suranna, P. Mastrorilli, C. F. Nobile, W. Keim, Inorg.
Chim. Acta 2000, 305, 151–156.
Acknowledgments
We thank Dr. G. Ciccarella for LC/MS measurements. The Italian
MURST (PRIN 2004 project, prot. 2004030719) and COSTD30
(fellowship for STSM to A. L.) are gratefully acknowledged for
financial support.
[37] P. Mastrorilli, C. F. Nobile, G. P. Suranna, F. P. Fanizzi, G. Cic-
carella, U. Englert, Q. Li, Eur. J. Inorg. Chem. 2004, 1234–1242.
[38] M. J. Baker, M. F. Giles, A. Guy, M. J. Taylor, R. J. Watt, J.
Chem. Soc., Chem. Commun. 1995, 197–198.
[39] L. Gonsalvi, H. Adams, G. J. Sunley, E. Ditzel, A. Haynes, J.
Am. Chem. Soc. 2002, 124, 13597–13612.
[1] G. W. Parshall, S. D. Ittel, Homogeneous Catalysis, 2nd edition,
Wiley Interscience, 1992, p. 106.
[2] G. Petöcz, Z. Berente, T. Kégl, L. Kollár, J. Organomet. Chem.
2004, 689, 1188–1193.
[3] G. W. Parshall, J. Am. Chem. Soc. 1972, 94, 8716–8719.
[4] C. Y. Hsu, M. Orchin, J. Am. Chem. Soc. 1975, 97, 3553.
[5] I. Schwager, J. F. Knifton, J. Catal. 1976, 45, 256–267.
[6] P. Meessen, D. Vogt, W. Keim, J. Organomet. Chem. 1998, 551,
165–170.
[7] L. A. Van der Veen, P. K. Keeven, P. C. J. Kamer, P. W. N. M.
van Leeuwen, J. Chem. Soc., Dalton Trans. 2000, 2105–2112.
[8] P. Haelg, G. Consiglio, P. Pino, J. Organomet. Chem. 1985, 296,
281–290.
[40] S. Gladiali, E. Alberico, S. Pulacchini, L. Kollár, J. Mol. Catal.
A 1999, 143, 155–162.
[41]
D. D. Ellis, G. Harrison, A. G. Orpen, P. Hirihattaya, P. G.
Pringle, J. G. de Vries, H. Oevering, J. Chem. Soc., Dalton
Trans. 2000, 671–675.
[42]
[43]
[44]
[45]
[46]
T. Hayashi, Y. Kawabata, T. Isoyama, I. Ogata, Bull. Chem.
Soc. Jpn. 1981, 54, 3438–3446.
As expected, no Pt black formed in the reaction carried out
with complex 6, in which the Pt atom is bound to two P atoms.
F. Ancillotti, M. Lami, M. Marchionna, J. Mol. Catal. 1990,
58, 331–344.
J. A. Iggo, D. Shirley, N. C. Tong, New J. Chem. 1998, 22,
1043–1045.
Over this temperature range the combination of broadening of
the 119Sn signal and its inherent weakness presumably pre-
cludes its observation.
[9] G. Parrinello, J. K. Stille, J. Am. Chem. Soc. 1987, 109, 7122–
7127.
[10] G. Consiglio, S. C. A. Nefkens, A. Borer, Organometallics 1991,
10, 2046–2051.
[11] S. Gladiali, D. Fabbri, L. Kollár, J. Organomet. Chem. 1995,
491, 91–96.
[12] A. Scrivanti, S. Zeggio, V. Beghetto, U. Matteoli, J. Mol. Catal.
A 1995, 101, 217–220.
[13] A. Scrivanti, V. Beghetto, A. Bastianini, U. Matteoli, G. Men-
chi, Organometallics 1996, 15, 4687–4694.
[47]
It is known that solvated SnCl2 can act as a donor ligand to-
ward transition metals. See, for instance: F. A. Cotton, G. Wil-
Eur. J. Inorg. Chem. 2006, 2268–2276
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2275

A. Lofù, P. Mastrorilli, C. F. Nobile, G. P. Suranna, P. Frediani, J. Iggo
FULL PAPER
kinson, Advanced Inorganic Chemistry, 4th edition, Wiley, New
[49] Complex 8 is stable with respect to CO deinsertion on purging
the solution with N2.
[50] A. Scrivanti, A. Berton, L. Toniolo, C. Botteghi, J. Organomet.
Chem. 1986, 314, 369–383.
York, p. 380.
1
[48] There is a noticeable difference in the H NMR chemical shift
of the methyl group of 8 when it is prepared using one (δ =
1.68 ppm) or five (δ = 1.77 ppm) equivalents of SnCl2. This
difference may be due to an interaction of excess tin chloride
with the acyl oxygen (see ref.[20]) resulting in a downfield shift
of the CH3 signal.
Received: November 15, 2005
Published Online: April 4, 2006
2276
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 2268–2276