Synthesis, characterization, reactivity and theoretical studies of ruthenium carbonyl complexes containing ortho-substituted triphenyl phosphanes

Article abstract of DOI:10.1016/j.jorganchem.2005.05.016

A series of ruthenium o-phosphane complexes was synthesized and characterized. The reactivity of the prepared complexes was studied by using them as catalysts for the hydroformylation of 1-hexene. The activities depended on the binding mode of the phosphane and on the strength of the ruthenium-phosphane interaction. Strongly coordinated chelating [2-(dimethylamino)phenyl]-(diphenyl) phosphane and [2-(methylthio)phenyl]- (diphenyl) phosphane showed poor activity, while weakly chelated [2-(methoxy)phenyl]-(diphenyl) phosphane and non-chelating phosphanes such as [2-(methyl)phenyl]-(diphenyl) phosphane or [2-(ethyl)phenyl]-(diphenyl) phosphane led to higher activities.

Full text of DOI:10.1016/j.jorganchem.2005.05.016

Journal of Organometallic Chemistry 690 (2005) 3803–3814
www.elsevier.com/locate/jorganchem
Synthesis, characterization, reactivity and theoretical studies
of ruthenium carbonyl complexes containing ortho-substituted
triphenyl phosphanes
a
a,
a
b
M. Andreina Moreno , Matti Haukka *, Sirpa Jaaskelainen , Sauli Vuoti ,
¨ ¨
¨
Jouni Pursiainen , Tapani. A. Pakkanen
b
a
a
Department of Chemistry, University of Joensuu, P.O. Box 111, FI-80101, Joensuu, Finland
b
Department of Chemistry, University of Oulu, P.O. Box 3000, FI-90014,Oulu, Finland
Received 6 March 2005; received in revised form 10 May 2005; accepted 11 May 2005
Available online 20 June 2005
Abstract
A series of ruthenium o-phosphane complexes was synthesized and characterized. The reactivity of the prepared complexes was
studied by using them as catalysts for the hydroformylation of 1-hexene. The activities depended on the binding mode of the
phosphane and on the strength of the ruthenium–phosphane interaction. Strongly coordinated chelating [2-(dimethyla-
mino)phenyl]-(diphenyl) phosphane and [2-(methylthio)phenyl]-(diphenyl) phosphane showed poor activity, while weakly chelated
[2-(methoxy)phenyl]-(diphenyl) phosphane and non-chelating phosphanes such as [2-(methyl)phenyl]-(diphenyl) phosphane or
[2-(ethyl)phenyl]-(diphenyl) phosphane led to higher activities.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Ruthenium; Phosphanes; Hydroformylation
1. Introduction
sidered as hemilabile ligands. Substitution in the ortho
position permits an efficient charge donation from li-
gand to metal and favours chelation, but it also intro-
duces higher steric requirements. Thus, the overall
effect varies according to the nature of the ortho
substituent.
Extensive work has been devoted to the development
of new phosphane ligands and the corresponding metal
complexes in order to tailor the catalytic properties.
Ruthenium compounds with phosphane ligands are
known to have applications in various fields of catalysis,
including hydrogenation [7–9], isomerization [10–14]
and hydroformylation reactions [15–18]. Although
ruthenium–phosphanes are well-known systems, they
are still of interest. Studies on new ruthenium complexes
[19] as well as mechanistic reports [20,21] on catalytic
behaviour have been recently published. Furthermore,
properties such as hemilability have been exploited in
Phosphanes are widely used ligands in complex chem-
istry because of the variety of properties they exhibit.
They can stabilize charge on a metal centre [1,2], gener-
ate vacancies for coordination [3,4], or favour a specific
geometrical configuration in a metal complex providing
high catalytic selectivity towards specific products [5].
They can also act as pro-chiral ligands helping in the
obtaining of chiral products for pharmaceutical applica-
tions [6]. Among different phosphanes, those containing
coordinating subtituents in the ortho-position display
useful properties. If the substituents can switch between
coordinated and uncoordinated mode they can be con-
*
Corresponding author. Tel.: +358 13 251 3346; fax: +358 13 251
3344.
E-mail address: Matti.Haukka@joensuu.fi (M. Haukka).
0022-328X/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.05.016

3804
M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
other applications, for example in the development of
sensors [22] or as a tool in supramolecular chemistry
[23].
(0.39 mmols) diphenylphosphane were dissolved in sepa-
rated flasks in 4 ml of degassed ethanol. The solutions
were combined and stirred overnight. A white solid pre-
cipitate was formed. The solid was filtered, washed with
ethanol, and dried under vacuum. The chelated (op) com-
plex [RuCl2(CO)(g2-op)(op)] 3 was obtained in a similar
way as described by Jeffrey and Rauchfuss [24]. Irradia-
tion of a sample of [RuCl2(CO)2(op)2] 4 dissolved in
CH2Cl2 with UV light for 8 h, yields complex 3 when
the solvent was removed by means of a vacuum, colour-
less crystals of 3 were formed.
In this study, we report on the preparation and char-
acterization of a series of ruthenium carbonyl complexes
containing ortho-substituted phosphane ligands. The
chemical reactivity of the complexes is studied by using
1-hexene hydroformylation as a probe reaction. The for-
mation of the new complexes is further investigated by
means of computational DFT calculations as well as
spectroscopic and crystallographic analysis. The steric
factors of the phosphane ligands are considered from
the point of view of cone angles. The chelation effects
of the bidentate phosphanes over the reactivity of the
coordinated complexes are also discussed. Schematic
pictures of the studied phosphanes are shown in Fig. 1.
2.1.1. Characterization
(OC-6-42)[RuCl2(CO)2(g2-np)] (1); np = [2-(dimeth-
ylamino)phenyl]-(diphenyl) phosphane. m(CO) = 2006,
2069 cmꢀ1 in CH2Cl2, dP (CDCl3) 49.1s. Anal. Calc.
for 1: H, 3.78; C, 49.54; N, 2.63. Found: H, 3.78; C,
49.31; N 2.62%. Yield = 85%.
2. Experimental
(OC-6-43)[RuCl2(CO)2(g2-sp)] (2); sp = [ 2-(methyl-
thio)phenyl]-(diphenyl) phosphane. m(CO) = 2016, 2075
cmꢀ1 in KBr pellets, dP (CDCl3) 55.6s. Anal. Calc. for
2: H, 3.28; C, 47.03. Found: H, 3.28; C, 47.04%.
Yield = 88%.
(OC-6-12)[RuCl2(CO)(g2-op)(op)] (3); op = [2-(meth-
oxy)phenyl]-(diphenyl) phosphane. m(CO) = 1968 cmꢀ1
(in CH2Cl2), dP (CDCl3) 34.0s. Anal. Calc. for 3: H,
3.22; C, 47.32. Found: H, 3.22; C, 47.17%. Yield = 65%.
FT-IR measurements were performed on a Nicolet
Magna 750 spectrometer. 31P NMR of the metal com-
plexes was recorded on a Bruker Avance with a reso-
nance frequency of 250 MHz. Elemental analysis of
the complexes was done on EA1110 CHNS-O equip-
ment (CE instruments). All reactions were performed
under nitrogen, and the solvents were also degassed
prior to use. Crystals were obtained by recrystallization
from a mixture 1:1 of hexane and dichloromethane.
2.2. Synthetic procedure: non-chelated complexes
2.1. Synthetic procedure: chelated complexes
[RuCl2(CO)3]2 (200 mg, 0.39 mmol) and 228 mg (0.83
mmol) of [2-(methyl)phenyl]-(diphenyl) phosphane, (0.79
mmol) of bis[2-(methyl)phenyl]-(phenyl) phosphane,
(0.79 mmol) of [2-(ethyl)phenyl]-(diphenyl) phosphane,
(0.72 mmol) of bis[2-(ethyl)phenyl]-(phenyl) phosphane,
(0.87 mmol) of triphenylphosphane, and (0.78 mmol)
of [2-(methoxy)phenyl]-(diphenyl) phosphane were
[RuCl2(CO)3]2 (200 mg, 0,39 mmol) and 120 mg of
the correspondent ligand op = [2-(methoxy)phenyl]-(di-
phenyl) phosphane (0.78 mmol), sp = [2-(methyl-
thio)phenyl]-(diphenyl) phosphane (0.41 mmol) or
np = [2-(dimethylamino)phenyl]-(diphenyl) phosphane
Me
Et
P
P
P
P
Et
Et
Me
Me
detp
dmep
etp
mep
P
MeS
sp
P
P
MeO
(Me)2N
op
np
Fig. 1. Schematic structures of the ligands under study.

M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
3805
dissolved in separated flasks in 4 ml of degassed ethanol.
The solutions were combined and stirred overnight (18
h). A white-yellowish solid precipitate was formed.
The solid was filtered, washed with ethanol, and dried
under vacuum.
stopped and the autoclave was rapidly cooled to room
temperature and brought to atmospheric pressure, after
which the liquid samples were analysed. The product
distribution is reported as wt%.
The gases CO and H2 used in the hydroformylation
experiments were of 99% and 99.99% purity, respec-
tively. The solvent 1-methyl-2-pyrrolidone (Aldrich
99%) and the internal standard cyclohexane (Merck
99%) were used without further purification and de-
gassed with nitrogen before use. Similarly, 1-hexene
(99%) was degassed prior to use. Gas chromatographic
analyses of the product mixture were recorded on a
Hewlett–Packard 5890 series II chromatograph
2.2.1. Characterization
(OC-6-33)[RuCl2(CO)2(op)2] (4); op = [2-(meth-
oxy)phenyl]-(diphenyl) phosphane. m(CO) = 1994, 2059
cmꢀ1 in KBr pellets, dP (CDCl3) 10.6s. Anal. Calc. for
4: H, 4.48; C, 56.55. Found: H, 4.35; C, 56.29%.
Yield = 95%.
(OC-6-33)[RuCl2(CO)2(pph3)2] (5); pph3 = triphenyl-
phosphane. m(CO) = 1992, 2055 cmꢀ1 for the (OC-6-
33) isomer and 2016.5, 2080.6 cmꢀ1 for the (OC-6-12)
isomer in KBr pellets, dP (CDCl3) 17.5s. Anal. Calc.
for 5: C, 54.31, H, 3.92. Found: C, 53.8; H, 3.54%.
The characterization data is in good agreement with that
reported by Batista et al. [25].
(OC-6-33)[RuCl2(CO)2(mep)2] (6); mep = [2-(me-
thyl)phenyl]-(diphenyl) phosphane. For the (OC-6-33)
isomer: m(CO) = 1978, 2046 cmꢀ1 in KBr pellets, dP
(CDCl3) 16.2s. Anal. Calc. for 6: H, 4.32; C, 59.89.
Found: H, 4.28; C, 59.63%. Yield = 85%. FT-IR
(CH2Cl2) also shows another dicarbonyl isomer located
at 2075 and 2143 cmꢀ1 presumably the (OC-6-12)
isomer.
(OC-6-33)[RuCl2(CO)2(dmep)2] (7); dmep = bis[2-
(methyl)phenyl]-(phenyl) phosphane. m(CO) = 1990,
2052 cmꢀ1 in KBr pellets, dP (CDCl3) 14.5s. Anal. Calc.
for 7: H, 4.74; C, 62.38. Found: H, 4.72; C, 61.56%.
Yield = 87%.
equipped with
a Varian WCOT fused silica 50
M · 0.53 M column and temperature programming.
2.4. X-ray structure determinations
The X-ray diffraction data was collected using a Non-
ius KappaCCD diffractometer using Mo Ka radiation
˚
(k = 0.71073 A). Single crystals of 1–4 and 6–9 were
mounted in inert oil to the cold gas stream of the diffrac-
tometer. The Denzo-Scalepack [26] program package
was used for cell refinements and data reduction. The
structures were solved by direct methods using the
SHELXS-97 or SIR-2002 programs [27,28]. A multiscan
absorption correction based on equivalent reflections
(XPREP in SHELXTL v. 6.14) [29] was applied to all data
(Tmin/Tmax
) values were 0.29924/0.36289, 0.30170/
0.35219, 0.14144/0.19398, 0.24387/0.31003, 0.24317/
0.29114 for 1–4 and 0.14979/0.19128, 0.23453/0.30346,
and 0.24387/0.31003, respectively, for 6–9. All structures
were refined with SHELXL-97 [30] and WinGX graphical
user interface [31]. In 4 the phenyl ring containing
OMe group and one of the plain phenyl rings were dis-
ordered over two sites with occupancies of 0.62:0.48.
Due to the disorder both rings were refined with fixed
(OC-6-33)[RuCl2(CO)2(etp)2] (8); etp = [2-(ethyl)phe-
nyl]-(diphenyl) phosphane. m(CO) = 2012, 2073 for the
(OC-6-33); and 1996, 2058 cmꢀ1 for the (OC-6-12) iso-
mer in CH2Cl2, dP (CDCl3) 15.6s. Anal. Calc. for 8:
H, 4.33; C, 54.01. Found: H, 4.32; C, 53.86%.
Yield = 92%.
˚
C–C distances of 1.390 A. Furthermore, the carbons
[Ru(l-Cl)Cl(CO)2(detp)]2 (9); detp = bis[2-(ethyl)phe-
C5A and C5B were refined with equal anisotropic dis-
placement parameters. In 6 one of the methyl groups
was also disordered between two phenyl rings with occu-
pancies of 0.64:0.36. All of the hydrogens were placed in
an idealized position and constrained to ride on their
parent atom. The crystallographic data is summarized
in Table 1 and the selected bond lengths and angles in
Table 2. Thermal ellipsoid plots of 1, 2, 3, 6 and 9 are
shown in Figs. 2–6. The plots of 4, 7 and 8 are given
as Supplementary material.
ꢀ1
nyl]-(phenyl) phosphane. m(CO) = 2011, 2071 cm
in
CH2Cl2, dP (CDCl3) 36.1s. Anal. Calc. for 9: C, 52.76;
H, 4.24. Found: C, 52.61; H, 4.39%. Solid precipitates
after a week and with stirring at room temperature.
Yield = 65%.
2.3. Catalysis
The hydroformylation reactions were performed in
high-pressure autoclaves (100 ml Berghof) equipped
with a teflon liner. The autoclaves were charged in a
glove box. In a typical experiment, the solvent 1-
methyl-2-pyrrolidinone (5 ml), the standard cyclohexane
(0.2 ml), the olefin 1-hexene (0.5 ml), and the catalyst
were added to the autoclave, which was then pressurized
to 20 bar with synthesis gas CO/H2 1:1. The autoclave
was heated at 120 ꢁC for 17 h. The reaction was then
2.5. Computational details
The geometries of the complexes were optimized
using the B3PW91 hybrid density functional method
and employing 6-31G* as a basis set (for ruthenium:
Huzinagaꢀs extra basis 433321/4331/421) [32]. The geom-
etry optimizations were followed by analytical frequency

3806
M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
calculations to obtain the vibration spectra and station-
ary point of all compounds. The calculations were made
using the GAUSSIAN-03 program package.
3. Results and discussion
Reactions between the ruthenium dimer [RuCl2-
(CO)3]2 and the selected phosphanes yielded two types
of complexes: chelated and non-chelated. The chelated
complexes (np) 1 and (sp) 2 contained only one phos-
phane ligand, while the chelated (op) complex 3 con-
tained two. In contrast, all the non-chelating
complexes contained two phosphane ligands. The reac-
tions are summarized in Scheme 1.
In a typical reaction, FT-IR revealed the formation
of tricarbonyl complexes containing one non-chelating
phosphane ligand as a minor side product in most of
the reactions involving non-chelating phosphanes.
3.1. Catalytic studies
The catalytic activity of all of the complexes was
tested in 1-hexene hydroformylation. The activities are
reported in Table 3. We used catalysis to study the reac-
tivity of the complexes. In the series of chelated phos-
phanes catalytic activity is a good indication of the
hemilabile character of these ligands and chelate effects,
and in the non-chelated series, catalysis allows studying
electronic and steric effects associated with reactivity.
3.1.1. Chelated complexes
The most distinct feature is that the activities of the
chelated compounds are, in general, lower than the
activities of the non-chelated ones. This is due to the
higher stability of the chelated complexes, which are less
eager to participate in the catalytic reaction. When com-
pared with each other, the activities of the chelated
ruthenium phosphane complexes follow the order:
Ru(op) > Ru(sp) > Ru(np). Thus, chelated ruthenium
(np) (1, Fig. 2) is considered the most stable complex be-
cause it is not active under catalytic conditions. The (np)
ligand coordinates strongly through both nitrogen and
phosphorus, resulting in a highly stable structure. The
tendency to produce strongly chelated complexes is typ-
ical of this particular (np) ligand [33,34] and examples of
non-chelated metal complexes of this ligand are, to the
best of our knowledge, not reported. In contrast, other
(np) ligands such as the 2-(diphenylphosphino)-pyridine
(PPh2py) [35–37] and the 2-(diphenylphosphino)-1-
methyl-imidazole (dpim) [38] show slightly better hemil-
abile properties and the non-chelated metal complexes
containing these phosphanes have been identified. In
the latter two cases chelation has to be promoted in con-
trast to our complex in which the chelated product is the
only one observed. The general problem with substi-

M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
3807
Table 2
Selected bond lengths (A) and bond angles (ꢁ) for 1–4 and 6–9
˚
Complex
1
2
3
4 Æ CH2Cl2
6 Æ 2(CH2Cl2)
7 Æ 2(CH2Cl2)
8 Æ 2(CH2Cl2)
9
Ru(1)–Cl(1)
Ru(1)–Cl(1A)
Ru(1)–Cl(2)
Ru(1)–P(1)
Ru(1)–P(2)
Ru(1)–C(1)
Ru(1)–C(2)
Ru(1)–O(2)
Ru(1)–S(1)
Ru(1)–N(1)
2.4065(9)
2.4397(6)
2.4072(7)
2.4525(13)
2.448(2)
2.4646(7)
2.4368(7)
2.4663(7)
2.4648(7)
2.3942(7)
2.3768(8)
2.4597(9)
2.2918(9)
2.4023(7)
2.3003(5)
2.3667(8)
2.3336(8)
2.4232(8)
1.811(3)
2.4362(12)
2.4133(14)
2.4136(14)
1.860(6)
2.441(2)
2.424(2)
2.432(2)
1.951(10)
1.874(8)
2.4335(6)
2.4549(6)
2.4564(6)
1.893(3)
1.870(3)
2.4550(7)
2.4327(7)
2.4220(7)
1.876(3)
1.931(4)
1.875(4)
1.871(4)
1.914(3)
1.876(3)
1.864(3)
1.877(3)
1.874(5)
2.263(2)
2.4007(6)
2.262(3)
Cl(1)–Ru(1)–Cl(2)
C(1)–Ru(1)–C(2)
C(1)–Ru(1)–O(2)
P(1)–Ru(1)–P(2)
P(1)–Ru(1)–S(1)
P(1)–Ru(1)–N(1)
Ru(1)–Cl(1)–Ru(1A)
89.85(12)
90.34(16)
89.28(3)
91.46(11)
93.51(10)
95.76(4)
82.72(15)
90.0(2)
90.7(3)
97.66(2)
93.23(11)
96.63(2)
92.18(11)
87.85(3)
90.11(12)
171.32(11)
177.09(3)
176.57(5)
176.56(6)
176.27(2)
175.93(3)
85.228(18)
81.14(7)
98.89(2)
Fig. 3. Molecular structure of (OC-6-43)[RuCl2(CO)2(g2-sp)] (2).
Thermal ellipsoids are drawn at the 50% probability level.
Fig. 2. Molecular structure of (OC-6-42) [RuCl2(CO)2(g2-np)] (1).
Thermal ellipsoids are drawn at the 50% probability level.
The (sp) complex of ruthenium (2, Fig. 3) shows al-
ready somewhat higher reactivity than the (np) com-
plex, producing alcohols to a small extent. In other
respects, the behaviour of this ligand is quite similar
to the (np). However, the (sp) complex is expected to
be slightly less stable than (np), judging from the catal-
ysis results. In most cases, the (sp) forms chelated com-
plexes where the phosphine is coordinated through the
phosphorus and sulphur atoms. Examples of non-che-
lated complexes are rare, but they are known [39].
Our (np) and (sp) ligands have very similar coordina-
tion behaviour. They both yield exclusively chelated
complexes. However, there are also differences. In the
case of (np) basicity of the amino group is key to the
interaction with the metal while in the case of (sp)
the basic character is much less pronounced. Reports
tuted phosphanes containing nitrogen in the ortho posi-
tion is that when the chelate is formed is hard to break
so the hemilability of these phosphanes is rather limited.
The differences in the behaviour of these phosphanes, as
well as ours, can be attributed to the basicity of the
nitrogen-containing fragment. In the case of pyridine
and imidazol both are basic fragments that can interact
with the ruthenium centre and coordinate. The ruthe-
nium–nitrogen bond can be cleaved with the help of ste-
ric stress introduced by the imidazol or pyridine ring. In
our case, the dimethyl amine fragment is highly basic
and coordinates too strongly because the electron pair
involved in the bond is not delocalized as in the case
of pyridine or imidazol.

3808
M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
Fig. 6. Molecular structure of [Ru(l-Cl)Cl(CO)2(detp)]2 (9). Thermal
ellipsoids are drawn at the 50% probability level. The atom labels ‘‘A’’
indicates atoms at equivalent position (1 ꢀ x, 2 ꢀ y, 1 ꢀ z).
towards the chelated species. This means that once the
ring is formed (fast rate) cleavage of the metal–nitro-
gen or metal–sulphur bond is difficult (slow rate). Fur-
thermore, the study shows that the rate constants are
strongly dependent on the nucleophillicity of the link-
ing atom, in our case the nitrogen in (np) is more
nucleophillic than the sulphur in (sp). These results
provide a ground on which our (np) and (sp) com-
plexes can be compared based on a kinetic parameter.
Among the chelated complexes, the (op) complex 3
displays the best activity. However, complex 3 (Fig. 4)
is very unstable and tends to be converted into 4 [24].
This suggests that the catalytic mechanism is not the
same as in the case of Ru–(sp) and Ru–(np), and that
the structural rearrangement may play an important
role during the formation of the catalytic active species.
Although the (op) ligand has chelating capabilities like
(sp) and (np), it did not chelate under the used experi-
mental conditions. Instead, the obtained main product
was the bis phosphane complex 4. Chelated Ru–(op)
complexes have been previously reported in the litera-
ture [43,44]. However, the preparation method typically
requires the use of higher temperatures, and the synthe-
ses are most commonly carried out under reflux. In the
case of our experiment, room temperature reactions re-
sulted solely in the non-chelated complex.
Fig. 4. Molecular structure of (OC-6-12)[RuCl2(CO)(g2-op)(op)] (3).
Thermal ellipsoids are drawn at the 50% probability level.
Fig. 5. Molecular structure of (OC-6-33)[RuCl2(CO)2(mep)2] (6).
Thermal ellipsoids are drawn at the 50% probability level. The
principle-numbering scheme for 4 and 6, 8 is similar. Figures of these
molecules are supplied as Supplementary material.
Charge donation seems to have an effect on the
strength of the heteroatom-ruthenium bond, in the che-
lated complexes. In the case of (np), both methyl groups
attached to the nitrogen donate charge, making the
nitrogen electron pair more willing to interact with the
metal. In the (sp) and (op) complexes 2 and 3, respec-
tively, there is only one methyl group involved. How-
in the literature regarding the basicity of aryl-substi-
tuted phosphanes suggest that the amino phosphane
is more basic (pKa = 8.65) than for example a methoxy
phosphane (pKa = 4.57) [40,41]. It can be expected that
the methoxy phosphane resembles the thiomethoxy
phosphane and thus assume that aminophosphane is
also more basic than thiomethoxy phosphane. Recent
studies on ring closure kinetics of bidentate hemilabile
(np) and (sp) ligands on platinum [42] showed that the
ring closure rate constant is faster than the rate con-
stant for the opening, suggesting that the equilibrium
between chelated and non-chelated complexes is shifted
˚
ever, the covalent radii of sulphur is larger (1.02 A)
˚
than the radii of oxygen (0.73 A) [45]. Thus, an effective
overlap of the sulphurꢀs p-orbitals with the rutheniumꢀs
d-orbitals is possible, making the ruthenium–sulphur
interaction stronger than in the case of ruthenium-
oxygen.

M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
3809
CO
Cl
X
Cl
P
Ru
Cl
X
P
OC
Ru
Cl
P
X
OC
np (1)
sp (2)
opa(3)
X
P
Cl
CO
Ru
Cl
OC
OC
CO
CO
Cl
Cl
Ru
CO
+
hν
+CO
Y
P
pph3 (5)
mep (6)
dmep (7)
etp (8)
Y
CO
Cl
Cl
CO
Y
Ph2P
PhP
OC
Y
Cl
Cl
CO
Ru
Ru
Ru
Cl
Y
OC
Y
PPh
PPh2
CO
Cl
opb(4)
detp (9)
Scheme 1. General reaction products observed during the reaction of [RuCl2(CO)3]2 and X = chelating phosphanes; Y = non-chelating phosphanes.
(op)a = chelated complex, (op)b = non-chelated complex. Reversibility was observed only with the (op) ligand.
For the catalytic process to take place is essential that
the precursor is activated. In the case of the chelated
complexes we calculated the energy required for the
breakdown of the chelate ring and coordination of car-
bon monoxide (see Scheme 2, for example). Our results
show that, in the case of (np) complex 1, the energy
requirements (3 kJ/mol) are low, and in principle the che-
late breakdown is possible. However, the lack of reactiv-
ity in this complex suggests that the process is not
favourable. For the (sp) complex 2, the energy needed
for the chelate to break down (ꢀ41 kJ/mol) is more
favourable than in the case of (np). This result is consis-
tent with the experimental results, since the complex
shows some activity. In the case of (op) complex 3, cleav-
age of the chelate is highly favoured and was observed
experimentally when the complex rearranged to form 4.
This rearrangement proceeds in solution. When the com-
plex 4 irradiated with UV light in a closed vessel, the
product obtained was 3 [24]. However, once the irradia-
tion was stopped, the complex started to re-coordinate
CO to form complex 4 (see Scheme 2). The product ob-
served after re-coordination of CO was the cis isomer
(OC-6-33)[RuCl2(CO)2(op)2] suggesting that the five-
coordinate intermediate isomerises from trans to cis be-
fore the CO coordinates. This fluxional behaviour of
the (op) ligand has been the subject of intense study and
several reports in the literature mention the facile cleav-
age of the ruthenium–oxygen bond by CO [46–49].

3810
M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
The fluxional behaviour has direct implications over
the catalytic mechanism. The chelates can break down
by carbon monoxide coordination, generating mono-
mers that can then follow a dissociative reaction mech-
anism, such as the detachment of one phosphane to
create a vacancy for coordination. However, fluxional
behaviour is restricted to the (op) complex. Thus, che-
late breakdown results in the generation of a more reac-
tive species. The (sp) complex 2 shows no obvious
fluxional behaviour and as discussed above ring opening
in this compound is very slow resulting in a low concen-
tration of active complex in solution. Finally, the inert
behaviour of the (np) complex 1 is indicative of its
inability to react. As a consequence, cleavage of the
ruthenium–heteroatom bond by CO follows the order:
(op) > (sp) > (np). This is a good indication of the differ-
ent hemilabile character of these phosphanes.
3.1.2. Non-chelated complexes
The reaction between the ruthenium dimer [RuCl2-
(CO)3]2 and the non-chelating phosphanes (etp),
(dmep), and (mep) yielded monomers containing two
carbonyls, two chlorides and two phosphanes (com-
plexes 5 and 6–8; see, for example, Fig. 5). We also
prepared the non-chelated (op) complex 4. This partic-
ular complex has been reported previously involving
the carbonylation of the bis-chelated complex [Ru-
Cl2(op)2] [24]. Herein, we report a simple one step syn-
thesis for this complex at room temperature, affording
high yields and purity.
When the reaction was performed using the bis[2-
(ethyl)phenyl]-(phenyl) phosphane (detp) the unex-
pected dimer [RuCl2(CO)2(detp)]2 9 was obtained
(see Fig. 6). Since this was the only example of such a
product among the non-chelated complexes, we exam-
ined its mechanism of formation in more detail using
DFT methods. The most probable reaction routes to
complex 9 from the reaction of [RuCl2(CO)3]2 and the
(detp) ligand are illustrated in Eqs. ((1)–(4) according to
the computational results. Complexes A and B are the
proposed intermediates, which have not been isolated.
½RuCl2ðCOÞ ꢁ þ 2ðdetpÞ
3 2
!
ꢀ14.3 kJ=mol 2½RuCl2ðCOÞ ðdetpÞꢁ ðAÞ
ð1Þ
ð2Þ
3
½RuCl2ðCOÞ ðdetpÞꢁ ðAÞ þ ðdeptÞ
3
34.1 kJ=mol
!
½RuCl2ðCOÞ ðdetpÞ ꢁ ðBÞ þ CO
2
2
2½RuCl2ðCOÞ ðdetpÞ ꢁ ðBÞ
2
2
ꢀ29.2 kJ=mol
!
½Ruðl-ClÞðClÞðCOÞ ðdetpÞꢁ ð9Þ þ 2ðdetpÞ
2
2
ð3Þ
ð4Þ
2½RuCl2ðCOÞ ðdetpÞ ꢁ ðAÞ
3
2
38.2 kJ=mol
!
½Ruðl-ClÞðClÞðCOÞ ðdetpÞꢁ ð9Þ þ 2ðCOÞ
2
2

M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
3811
-99.7 kJ/mol
Ph2
P
CH3
O
OMe
Ph2
P
Cl
Ru
CO
Cl
Cl
Ru
CO
P
hν
OC
P
Cl
OC
Ph2
Ph2
OMe
Scheme 2. Rupture of the chelate and rearrangement of (OC-6-12)[RuCl2(CO)(g2-op)(op)].
In order to obtain experimental support for the se-
quence proposed, we performed the synthesis step by
step. In step 1 (Eq. (1)), stoichiometric amounts of
ruthenium and phosphane were dissolved separately in
ethanol, the solutions were mixed and allowed to react,
and the reaction was monitored by IR. After 4 h reac-
tion the tricarbonyl complex (A in Eq. (1)) was identified
by FT-IR (m(CO) = 1999s, 2062s (broad), 2135s cmꢀ1 in
EtOH). In the second step, an excess of phosphane was
added and the solution was heated under reflux for 2 h.
A gas sample taken from the reaction vessel revealed
clearly the release of carbon monoxide (m(CO) = 2143
cmꢀ1, with typical CO fine structure). Furthermore,
the FT-IR of the solution (signals at 2009s, 2071s
cmꢀ1 in EtOH) confirmed the loss of a carbonyl from
the tricarbonyl complex A (Eq. (1)). The new dicarbonyl
pattern can be related to C (Eqs. (3) and (4)).
The energies required to form the dicarbonyl mono-
mer (Eq. (2)) and dicarbonyl dimer (Eq. (4)) are very
similar, and they both involve release of carbon monox-
ide. The formation of the dimer via the dicarbonyl
monomer B (Eq. (3)) was considered because most of
the non-chelating phosphanes form monomers of this
type. However, direct formation of the dimer from the
tricarbonyl monomer (Eq. (4)) is also possible and re-
quires approximately the same amount of energy as
the reaction in Eq. (2). When the solution was kept un-
der reflux for another 4 h, a clear orange-yellow precip-
itate of C was obtained. FT-IR of the solid revealed the
same carbonyl signals at 2009s and 2071s cmꢀ1 in
CH2Cl2. Elemental analysis of the solid and 31P NMR
confirmed that the product was [Ru(l-Cl)Cl(CO)2-
(detp)]2. Although halide-bridged dimers containing
small phosphines have been previously reported
[38,50–53] examples of ruthenium carbonyl halide-
bridged dimers containing bulky phosphanes are less
common [54–56].
In 1-hexene hydroformylation, the activities of the
complexes containing alkyl-substituted phosphanes
were, in general, higher than the activity of unsusbsti-
tuted triphenylphosphane. The alkyl-substituted phos-
phanes are selective towards the production of
alcohols. In all of the cases the terminal alcohol was
the main product. Alcohols are the result of further
hydrogenation of the aldehydes produced, showing that
the catalyst operates by means of two different mecha-
nisms. All of the ruthenium complexes with alkyl-substi-
tuted phosphanes were monomers (see Fig. 5, for
example) except the one with bis, detp = bis[2-(ethyl)-
phenyl]-(phenyl) phosphane which is a dimer 9. In order
to study the catalytic activation process in more detail,
we modelled the classical generation of a vacancy on
the ruthenium centre by assuming a dissociative mecha-
nism for both the monomers and the dimer. The results
indicate that the detachment of a phosphane ligand
from the monomer is, as expected the most probable ini-
tiation step (see Scheme 3). The reaction energy for this
step is 53 kJ/mol, while the release of carbonyl would re-
quire 180 kJ/mol. This is due to the steric requirement
introduced by substitution in the ortho position in the
phosphane ligand. Experimental evidence of such
detachment was observed when the 31P NMR experi-
ments showed free phosphane in solution after catalysis
except in the case of the [Ru(l-Cl)Cl(CO)2(detp)]2
dimer.
With the dimer the process goes along through a dif-
ferent route. We have modelled the generation of a va-
cancy via cleavage of the chlorine bridge using DFT
methods. This route was suggested because the chlorine
bridge is known to break readily in the presence of
Cl
Cl
Cl
Ru
Cl
53.3 kJ/mol
PhP
PPh
PhP Ru
+
PhP
CO
CO
OC
OC
Scheme 3. Generation of a vacancy in a monomeric phosphane-containing complex.

3812
M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
CO
Ru
Cl
Cl
Cl
Cl
Cl
CO
P
P
CO
Cl
Ru
CO
135.2 kJ/mol
P
Ru
2
CO
CO
Scheme 4. Cleavage of the dimer and generation of a vacancy.
coordinating solvents [57–59]. The energy requirements
for such a cleavage are higher in comparison with the
detachment of the phosphane ligand in a monomer
(see Scheme 4). However, the dimer is highly active,
and cleavage of the chlorine bridge generates monomers
with a vacancy for coordination. Furthermore, no evi-
dence of release of phosphane was observed when 31P
NMR experiments were performed on the catalytic solu-
tion suggesting that detachment of the phosphane in the
dimer is unlikely to be the way in which the dimer
activates.
4. Conclusions
The most active catalysts can be obtained by using
weakly bound non-chelating phosphanes, since removal
of the phosphane is favourable. Similarly, the lowest
activities are related to the complexes, which contain
strongly bonded chelated phosphanes. This is primarily
due to the stabilization effect associated with the forma-
tion of the chelate ring. More stable chelated [2-
(dimethylamino)phenyl]-(diphenyl) phosphane and
[2-(methylthio)phenyl]-(diphenyl) phosphane complexes
1–2 are less willing to participate in the catalytic reac-
tions. Even though [2-(methoxy)phenyl]-(diphenyl)
phosphane, is able to chelate the oxygen–ruthenium
interaction is weaker than the corresponding N–Ru or
S–Ru in the cases of [2-(dimethylamino)phenyl]-(diphe-
nyl) phosphane and [2-(methylthio)phenyl]-(diphenyl)
phosphane complexes. Thus, [2-(methoxy)phenyl]-(di-
phenyl) phosphane behaves more like any of the non-
chelated phosphines.
The degree of substitution in the ligand has no
marked effect on the catalytic performance. Similar cat-
alytic activity was observed with both mono-substituted
and di-substituted phosphanes. Activities in the non-
chelated series are related to the steric properties of
the phosphanes. Electronic effects are irrelevant in the
ruthenium complexes containing alkyl-substituted phos-
phanes when compared to each other. However,
activities of the complexes containing substituted phos-
phanes are higher compared to unsusbstituted phos-
phane. In this case the electronic effects seems to have
influence, since the presence of substituents may stabi-
lize the active intermediates. The steric stress induced
by the substituents in the ortho position may also facil-
itate detachment. In other words, substitution in the
phosphane increases activities but once substituted the
effect is not enhanced by increasing the number of sub-
stituents in the ligand.
3.2. Steric effects
The possible steric effect introduced by the phos-
phane ligands on the coordinated complexes was esti-
mated by calculating the Tolman cone angles (Table
4). The cone angles were calculated from both experi-
mentally determined crystal structures and from DFT
optimized structures. The differences between the cone
angles calculated from the X-ray structures and from
the optimized structures can be attributed to the packing
effects in the crystals. According to the Tolman defini-
tion, a wider cone introduces more steric limitations.
Most of the chelated phosphanes display lower values
for the cone angle. This means less steric stress, and
hence more stable complexes. Again, this is in accor-
dance with the observed catalytic activity (Table 3).
Table 4
Calculated cone angles for the coordinated complexes
Catalyst
Calculated cone angle, h (ꢁ)
X-ray
structure
Computationally
optimized structures
[RuCl2(CO)2(g2-np)]
[RuCl2(CO)(g2-op)(op)]
[RuCl2(CO)2(g2-sp)]
[RuCl2(CO)2(op)2]
177
182a–188b
220
189a–197b
204a–206b
194a–196b
184a–190b
197a–197b
192a–192b
176
182a–188b
174
210a–230b
202a–205b
201a–188b
202a–212b
188a–193b
187a–190b
[RuCl2(CO)2(pph3)2]
[RuCl2(CO)2(mep)2]
[RuCl2(CO)2(etp)2]
In the chelated series, catalytic activity seems to be re-
lated mainly to the difficulties encountered by the che-
lated complexes to generate
stabilization via chelate effect reduces the concentration
of active intermediates in solution. The poor hemilabile
character of the (np) and (sp) complexes results in imme-
a
vacancy. Thus,
[Ru(l-Cl)Cl(CO)2(detp)]2
[RuCl2(CO)2(dmep)2]
a
Phosphane 1.
Phosphane 2 in the complex.
b

M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
3813
[16] V.K. Srivastava, R.S. Shukla, H.C. Bajaj, R.V. Jasra, J. Mol.
Catal. A 202 (2003) 65–72.
diate ring closure and the generation of an inactive com-
plex. On the other hand, the fluxionality exhibited by the
hemilabile (op) complex proves to be a disadvantage
when the catalytic precursor in non-chelated since high
temperature promotes chelation.
[17] E. Lozano Diz, A. Neels, H. Stoeckli-Evans, G. Suss-Fink,
Polyhedron 20 (2001) 2771–2780.
[18] M. Beller, B. Zimmermann, H. Geissler, Chem. Eur. J. 5 (4)
(1999) 1301–1305.
[19] C. Schluenken, M.A. Esteruelas, F. Lahoz, L.A. Oro, H. Werner,
Eur. J. Inorg. Chem. 12 (2004) 2477–2487.
[20] C.A. Sandoval, T. Ohkuma, K. Muniz, R. Noyori, J. Am. Chem.
Soc. 125 (44) (2003) 13490–13503.
Acknowledgement
[21] C. Beddie, M.B. Hall, J. Am. Chem. Soc. 126 (42) (2004) 13564–
13565.
Financial support provided by the Academy of Fin-
land is gratefully acknowledged (M.H.).
[22] C.W. Rogers, M.O. Wolf, Chem. Commun. (1999) 2297–2298.
[23] M.S. Khoshbin, M.V. Ovchinnikov, C.A. Mirkin, L.N. Zakharov,
A.L. Rheingold, Inorg. Chem. 44 (2005) 496–501.
[24] J.C. Jeffrey, T.B. Rauchfuss, Inorg. Chem. 18 (10) (1979) 2658–
2666.
Appendix A. Supplementary data
[25] A.A. Batista, J. Zuckerman-Schpector, O.M. Porcu, S.L. Queiroz,
M.P. Araujo, G. Oliva, D.H.F. Souza, Polyhedron 13 (1994) 689–
691.
Crystallographic data for the structural analysis has
been deposited with Cambridge Crystallographic Data
Centre. CCDC-263091 to CCDC-263098 for 1–4 and
6–9, respectively, contains the supplementary crystallo-
graphic data for this paper. Copies of this information
may be obtained free of charge from The Director,
CCDC, 12 Union Road, Cambridge CB2 1EZ, UK
(fax: +44 1223 336033; e-mail: deposit@ccdc.cam.ac.uk,
or http://www.ccdc.cam.ac.uk). Supplementary data
associated with this article can be found, in the online
version at doi:10.1016/j.jorganchem.2005.05.016.
[26] Z. Otwinowski, W. Minor, in: C.W. CarterJr., R.M. Sweet
(Eds.), Processing of X-ray Diffraction Data Collected in Oscil-
lation Mode, Methods I Enzimology, Macromolecular Crystal-
lography, Part A, vol. 276, Academic Press, New York, 1997, pp.
307–326.
[27] G.M. Sheldrick, SHELXS-97: Program for Crystal Structure Deter-
mination, University of Go¨ttingen, Go¨ttingen, Germany, 1997.
[28] M.C. Burla, M. Camalli, B. Carrozzini, G.L. Cascarano, C.
Giacovazzo, G. Polidori, R. Spagna, SHELXTL 6.14, J. Appl.
Crystallogr. 36 (2003) 1103.
[29] G.M. Sheldrick, SHELXTL V. 6.14, Bruker Analytical X-ray
Systems, Bruker AXS, Inc., Madison, WI, USA, 2003.
[30] G.M. Sheldrick, SHELXL-97: Program for Crystal Structure
Refinement, University of Go¨ttingen, Go¨ttingen, Germany, 1997.
[31] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837.
[32] S. Huzinaga (Ed.), Gaussian Basis Sets for Molecular Calcula-
tions Physical Sciences Data 16, Elsevier, Amsterdam, 1984.
[33] E.S. Chandrika, S. Ma, S.J. Rettig, B.R. James, W.R. Cullen,
Inorg. Chem. 36 (1997) 5426–5427.
References
[1] F. Van Gastel, J.F. Corrigan, S. Doherty, N.J. Taylor, A.J.
Carty, Inorg. Chem. 31 (1992) 4492–4498.
[2] S.P. Nolan, T.R. Belderrain, R.H. Grubbs, Organometallics 16
(1997) 5569–5571.
[34] C.B. Pamplin, E.S. Ma, N. Safari, S.J. Rettig, B.R. James, J. Am.
Chem. Soc. 123 (2001) 8596–8597.
[3] A. Beguin, H.C. Bo¨ttcher, G. Suss-Fink, B. Walther, J. Chem.
Soc., Dalton Trans. (1992) 2133–2134.
[35] I. Moldes, E. Encarnacion, J. Ros, A. Alvarez-Larena, J.F.
Piniella, J. Organomet. Chem. 566 (1998) 165–174.
[36] A. Dervisi, P.G. Edwards, P.D. Newman, R.P. Tooze, S.I. Coles,
M.B. Hursthouse, J. Chem. Soc., Dalton Trans. (1998) 3771–3776.
[37] A. Dervisi, P.G. Edwards, P.D. Newman, R.P. Tooze, J. Chem.
Soc., Dalton Trans. (2000) 523–528.
[4] S.Y. Kang, T. Yamabe, A. Naka, M. Ishikawa, K. Yoshizawa,
Organometallics 21 (2002) 150–160.
[5] M. Kranenburg, P.C. Kamer, P.W. Van Leeuwen, W.N. Piet,
Eur. J. Inorg. Chem. 2 (1998) 155–157.
[6] C. Darcel, E.B. Kaloun, R. Merdes, D. Moulin, N. Riegel, S.
Thorimbert, J.P. Genet, S. Juge, J. Organomet. Chem. 624 (2001)
333–343.
[38] A. Caballero, F.A. Jalon, B.R. Manzano, G. Espino, M. Perez-
Manrique, A. Mucientes, F.J. Poblete, M. Maestro, Organomet-
allics 23 (2004) 5694–5706.
[7] J. Elek, L. Nadasdi, G. Papp, G. Laurenczy, F. Joo, Appl. Catal.
A 255 (2003) 59–67.
[39] L. Hirsivaara, M. Haukka, S. Jaaskelainen, R. Laitinen, E.
Niskanen, T.A. Pakkanen, J. Pursiainen, J. Organomet. Chem.
579 (1–2) (1999) 45–52.
[8] X. Dong, C. Erkey, J. Mol. Catal. A 211 (2004) 73–81.
[9] H. Horvath, G. Laurenczy, A. Katho, J. Organomet. Chem. 689
(2004) 1036–1045.
[40] C.A. Streuli, Anal. Chem. 32 (1960) 985–987.
[41] T. Allman, R.G. Goel, Can. J. Chem. 60 (6) (1982) 716–722.
[42] R. Romeo, L. Monsu-Scolaro, M.R. Plutino, A. Romeo, F.
Nicolo, A. Del Zotto, Eur. J. Inorg. Chem. (2002) 629–638.
[43] C.W. Roggers, B.O. Patrick, S.J. Rettig, M.O. Wolf, J. Chem.
Soc., Dalton Trans. (2001) 1278–1283.
[10] H. Sertchook, D. Avnir, J. Blum, F. Joo, A. Katho, H.
Schumann, R. Weimann, S. Wernik, J. Mol. Catal. A 108
(1996) 153–160.
[11] C. Slugovc, E. Ruba, R. Schmid, K. Kirchner, Organometallics
18 (1999) 4230–4233.
[44] Y. Yamamoto, K.-I. Sugawara, T. Aiko, J.-F. Ma, J. Chem. Soc.,
Dalton Trans. (1999) 4003–4008.
[12] R.C. Van der Drift, M. Vailati, E. Bowman, E. Drent, J. Mol.
Catal. A 159 (2000) 163–177.
[45] J.E. Huheey, E.A. Keiter, R.L. Keiter, in: Jane Piro (Ed.),
Inorganic Chemistry, fourth ed., HarperCollins College Publish-
ers, New York, 1993, p. 253 (Chapter 8).
[13] A. Salvini, P. Frediani, F. Piacenti, J. Mol. Catal. A 159 (2000)
185–195.
[14] A. Salvini, F. Piacenti, P. Frediani, A. Devescovi, M. Caporali, J.
Organomet. Chem. 625 (2001) 255–267.
[46] E. Benedetti, G. Braca, G. Sbrana, F. Salvetti, B. Grassi, J.
Organomet. Chem. 37 (1972) 361–373.
[47] M.I. Bruce, F.G.A. Stone, J. Chem. Soc. A (1967) 1238.
[15] P. Smolen˜ski, F.P. Pruchnik, Z. Ciunik, T. Lis, Inorg. Chem. 42
(2003) 3318–3322.

3814
M.A. Moreno et al. / Journal of Organometallic Chemistry 690 (2005) 3803–3814
[48] P. Braunstein, D. Matt, Y. Dusausoyz, Inorg. Chem. 22 (1983)
2043–2047.
[54] A.V. Marchenko, J.C. Huffman, P. Valerga, M. Jimenez-Tenorio,
M.C. Puerta, K.G. Caulton, Inorg. Chem. 40 (2001) 6444–6450.
[55] M-K. Chung, G. Ferguson, V. Robertson, M. Shalf, Can. J.
Chem. 79 (2001) 949.
[49] P. Braunstein, D. Matt, D. Nobel, S.-E. Bouaoud, B. Carluer, D.
Grandejan-Lemoine, J. Chem. Soc., Dalton Trans. (1986) 415.
[50] B. deKlerk-Engels, J.H. Groen, K. Vrieze, A. Mockel, E. Lindner,
Inorg. Chim. Acta 195 (1992) 237.
[56] T. Gottschalk-Gaudig, K. Folting, K.G. Caulton, Inorg. Chem.
38 (1999) 5241–5245.
[51] E. Lindaer, A. Mikkel, M. Steimann, H. Kiihbauch, R. Fawzi,
H.A. Mayer, Inorg. Chem. 32 (1993) 1266–1271.
[57] S.D. Drouin, D. Amoroso, G.P.A. Yap, D.E. Fogg, Organomet-
allics 21 (2002) 1042–1049.
[52] T.G. Southern, P.H. Dixneuf, J.-Y. Le Marquille, D. Grandjean,
Inorg. Chem. 18 (11) (1979) 298.
[58] V. Cadierno, P. Crochet, J. Dez, S.E. Garca-Garrido, J. Gimeno,
Organometallics 22 (2003) 5226–5234.
[53] R. Sanchez-Delgado, M. Medina, F. Lopez-Linares, A. Fuentes,
J. Mol. Catal. A 116 (1997) 167–177.
[59] B.F.G. Johnson, R.D. Johnston, J. Lewis, J. Chem. Soc. A (1969)
792–796.