Inorg. Phys. Theor.: Ca. 1 Atmosphere of Hydrogen Pressure and

View Article Online / Journal Homepage / Table of Contents for this issue
Inorg. Phys. Theor. 1711
The Preparation and Properties of Tris(triphenylphosphine)halogeno-

rhodium(1) and Some Reactions thereof including Catalytic Homo-
geneous Hydrogenation of Olefins and Acetylenes and their Derivatives
B y J. A. Osborn, F. H. Jardine, J. F. Young, and G. Wilkinson
Published on 01 January 1966. Downloaded by University of South Carolina Libraries on 07/01/2016 19:37:16.
Tris(triphenylphosphine)chlororhodium(l), RhCl(PPh3)3,has been prepared by the interaction of an excess of

triphenyiphosphine with rhodium(ll1) chloride hydrate in ethanol ; the corresponding bromide and iodide are also
described. The dissociation of the complex in various solvents has been investigated, and its reactions with
hydrogen, ethylene, and carbon monoxide and aldehydes studied. Dihydrido- and ethylene complexes have
been isolated and studied by nuclear magnetic resonance (n.m.r.) spectroscopy. Approximate values for the
formation constants of ethylene and propylene complexes have been obtained ; the latter i s lower by a factor of
over lo3. By electron spin resonance spectroscopy, the complex RhCI(PPh,), has been shown to contain trace
amounts of a paramagnetic species, probably a rhodium(l1) complex.
In homogeneous solution the tris(tripheny1phosphine) complexes are exceedingly active catalysts for the rapid
and homogeneous hydrogenation, a t ca. 1 atmosphere of hydrogen pressure and room temperature, of unsaturated
compounds containing isolated olefinic and acetylenic linkages.
The rates of hydrogenation of hept-1 -ene, cyclohexene and hex-1-yne have been studied quantitatively and the
dependence on factors such as substrate and catalyst concentration, temperature, and pressure determined. The
data can be accommodated by a rate expression of the form :
where [S] and [A] are the olefin and catalyst concentrations, respectively, and p is the concentration of hydrogen in
solution.
From the data for cyclohexene the activation energy for the rate determining step is , = 22.9 kcal. mole-l
( A H 1 = 22.3 kcal. mole-l) and the value of ASS = 12.9 e.u.
It is shown that the rate of hydrogen-deuterium exchange under selected conditions is quite slow compared
with the rates of hydrogenation of olefins and, furthermore, that when H,-D, mixtures are used in the reactions,
alkanes and dideuteroalkanes are the major products. Reductions of maleic and fumaric acids with deuterium
shows that cis-addition occurs preferentially. Similarly, in the reduction of hex-2-yne to n-hexane, cis-hex-2-ene
is found to be the major olefin intermediate.
A mechanism for the hydrogenation is proposed in which the metal complex serves as a template to which a
hydrogen molecule and an olefin molecule are briefly co-ordinated before transfer of one to the other takes place.
The low kinetic isotope effect (rate H,/rate D, = 0.9) suggests that synchronous breaking of Rh-H bonds and
making of C-H bonds takes place in the transition state involving two simultaneous three-centre interactions.
THE activation of molecular hydrogen by transition- reduce q ~ i n o n e . ~However, rapid deposition of metal
metal ions or complexes in homogeneous solution is occurs, a complication also encountered in the reduction
well kn0wn.l However, only very few systems have of ferric ion by hydrogen in the presence of hexachloro-
proved to be useful for the catalytic homogeneous rhodate(111) ion.6
hydrogenation of unsaturated compounds. One of the During our studies on the use of reducing agents for
most versatile catalysts, but one which has the disad- the catalytic preparation of rhodium(n1) complexes
vantage of being soluble only in water, is the cobalt(I1)- from RhC1,,3H20, it was observed that molecular
cyanide system; it allows the reduction of conjugated hydrogen could act as a catalyst, e.g., for the preparation
double bonds, as in butadiene, a , P-unsaturated acids of trans-[Rh py4 Cl,]Cl from pyridine solutions of
and aldehydes, but isolated ethylenic linkages are rhodium trichloride (which contain mainly 1,2,6-
unaffected.2 A series of Ziegler-type catalysts soluble Rh py3 C13). Furthermore, it was found that 1,2,6-
in organic solvents allows hydrogenation of olefinic Rh py3 C13 in ethanol can slowly catalyse the reduction
hydrocarbons,lC while a chloroplatinate(I1)-trichloro- of hex-l-ene by hydrogen a t atmospheric pressure (or
stannate(I1) complex is effective for ethylene and acetyl- more rapidly a t higher pressures) without metallic
ene.3 I t has been claimed that higher olefins are not rhodium being deposited.
reduced in ethanolic solution,4aalthough in the presence In order to stabilise the catalyst complex more
of triphenylphosphine, platinum-tin complexes can effectively towards reduction to the metal, and in the
hydrogenate the methyl ester of soy bean hope of obtaining a system catalytically more efficient
I t has long been known that molecular hydrogen in than the ones employing pyridine, we studied rhodium
the presence of soluble rhodium(II1) ammines could
3 R. D. Cramer, E. L. Jenner, R. V. Lindsay, jun., and U. G.
( a ) J. Halpern, A d v . in Catalysis, 1959, 11, 301; (b) J. Stolberg, J . Amer. Chem. SOC., 1963, 85, 1691.
Halpern and R. S. Nyholm, Proc. 3rd Internat. Co?zg. Catal., 4 ( a ) G. C. Bond and I. Hellyer, Chem. and Ind., 1965, 35; (b)
W. hl. H. Sachter, G. C. A. Schuitt, and P. Zwietering, eds., vol. I, J. C. Bailar, jun., and H. Itatani, Inorg. Chem., 1965, 5, 1618.
North Holland Pub. Co., Amsterdam, 1965, p. 25, 146; (c) M. F. M. Iguchi, J . Chern. SOC.Japan, 1939, 60, 1287.
Sloane, A. S. Matlack, and D. S. Breslow, J . Amer. Chem. SOC., 6 R. F. Harrod and J. Halpern, Canad. J . Chem., 1959, 37,
1963, $5, 4014, and references therein. 1933.
J. Kwiatek and J. K. Seyle?, J . Organometallic Chem., 1965, 7 R. D. Gillard, J. A. Osborn, P. B. Stockwell, and G.
3, 421, and references therein. Wilkinson, Proc. Chem. Sac., 1964, 284.
View Article Online
1712 J. Chem. SOC. (A), 1966

complexes of the stronger x-acceptor ligands such as geneous hydrogenation catalysts; most of the detailed
tertiary phosphones and arsines. Our first studies studies have been made on the chloride. These particular
were made using a rhodium(II1) complex stabilised by complexes have proved more convenient than any other
triphenylphosphine. During the preparation of this previously known catalyst systems for allowing the
phosphine complex we found that when an excess of mechanism of the catalytic process to be studied. In
triphenylphosphine was used, a rhodium ( I) complex, succeeding Papers we shall describe hydrof ormylation
KhCl(PPh,),, results. This rhodium(1) complex, and the and hydrogenation studies, using rhodium (111) and other
corresponding bromide and iodide, have proved to be Group VIII metal complex catalysts.
the most effective catalysts yet recognised for the (A) Tris(triphenylphosphine)chlororhodium(I) and its
homogeneous hydrogenation a t normal temperatures Properties.-(a) Preparation and Physical properties.
and pressures of a variety of unsaturated compounds When rhodium(II1) chloride hydrate and an excess
containing both double and triple bondss (ca. 6 mols) of triphenylphosphine are dissolved in boiling
Although patent claims1 have been made that ethanol, reduction occurs and the dark, burgundy-red,
solutions of various metallic salts in phosphites could act crystalline complex, RhCl(PPh,),, is obtained A large
as homogeneous hydrogenation catalysts, stoicheio- excess of triphenylphosphine is essential to prevent the
metric compounds were not characterised, nor were formation of a bridged dimer (see below) and to further
phosphine ligands studied. Since our first public- the reduction of the rhodium(II1) species formed initially.
ations7-9 Vaska has extended his work on hydrogen Since reduction does not occur in ethanol heated under
uptake by square complexes and in preliminary notes l1*l2 reflux unless an excess of triphenylphosphine is present,
reports that ethylene, propylene and acetylene can be and since acetone-water mixtures can be used in place
slowly hydrogenated a t 40-60" by the use of of ethanol as a solvent, the triphenylphosphine is,
IrCl(CO)(PPh,), or at 2 0 4 0 " by IrH(CO)(PPh,), and without doubt, the actual reducing agent. The initial
OsHCl(CO)(PPh,),. As we shall show later, the square product is probably PPh,Cl, which then undergoes
carbonyl complexes are very much less efficient catalysts solvolysis to form OPPh,; the latter is detectable in the
than RhCl(PPh,),. residual solutions. If insufficient ethanol as a solvent
Although hydroformylation reactions of olefins can be is used, an orange crystalline material is at first precipi-
effected using rhodium(111) complexes,8 the rhodium(I) tated, which gives the red complex on continued heating
complex again is a more effective catalyst, and can also under reflux. This orange complex has an analysis
be used for acetylene hydrof0rmy1ation.l~ As we shall identical with that of the red RhCl(PPh,), and may be
show later, RhCl(PPh,), is converted by carbon monoxide a second crystalline form.16 * The corresponding
a t 1 atmosphere pressure to RhCl(CO)(PPh,),, so that bromide, prepared by the addition of an excess of lithium
the active hydroformylation catalyst, is, in fact this bromide to the reaction mixture, can also be isolated
well known complex; this was confirmed by its direct in two crystalline forms. The iodide has been prepared
use. The only previous literature reference to the use similarly, but the conditions of reaction are more critical
of rhodium catalysts is a recent one describing a com- (see Experimental section).
parison of rhodium carbonyl with cobalt carbonyl in The above reaction involving triphenylphosphine
the conventional hydroformylation process.14 Although appears to be unique. By contrast, the interaction of
patent claims for the use of solid-state rhodium catalysts other tertiary aryl alkyl phosphines and arsines with hot
exist, the claims involving soluble catalysts,15i.e., Group ethanolic solutions of rhodium trichloride gives mono-
VIII salts of transition metals in the presence of donor or bi-nuclear rhodium(II1) c0mp1exes.l~ Although
ligands such as R,P, wherein the metal is specifically the interaction of rhodium chloride with triphenyl-
stated to be in an oxidation state of 0 or -1, provide phosphine has been reported,18stoicheiometric quantities
no examples of the use of rhodium complexes. Indeed, were used in order to obtain RhCl,(PPh,),; we have had
not only were authentic stoicheiometric compounds some difficulty in reproducing this preparation and
both unspecified and characterised, but we have shown although orange products are obtained, they appear to
that some of the claims cannot be reproduced. be mixtures which, on repeated crystallisation, give
In the present Paper we discuss the properties of some RhCl(PPh,),.
tris(tripheny1phosphine)halogenorhodium (I) complexes, The only compound of a similar type previously
RhX(PPh,),, (X = C1, Br, or I), and their use as homo- reported, appears to be the diphenylphosphine
RhCl(PHPh,),, obtained by the interaction of PHPh,
* These compounds were also independently prepared by l4 H. Wakamatsu, N i p p o n Kekakzt Zasshi, 1965, $5, 227
Bennett and Longstaff (ref. 16). (Chem. Abs., 1964, 61, 13,173).
* J. A. Osborn, G. Wilkinson, and J . F. Young, Chem. Comm., l5 e.g., B.P. 801,734/1968; 988,941, 988,942, 988,943, 988,9441
1965, 17. 1965.
J. F. Young, J. A. Osborn, F. H. Jardine, and G. Wilkinson, l6 RI. A. Bennett and P. A. Longstaff, Chem. and Iizd., 1965,
Chem. Comm., 1965, 131. 846.
lo e.g., B.P. 946,062/1964. l7 J. Chatt, N. P. Johnson, and B. L. Shaw, J . Chem. SOC.,
l1 L. Vaska and R. E. Rhodes, J . Amer. Chem. SOC.,1965, 87, 1964, 2508.
4970. l8 R. J. Mawby and L. M. Venanzi, Experimentia Sztpplementa,
l2 L. Vaska, Inorg. Nuclear Chem. Letters, 1965, 1, 89. No. 9 ; " Essays in Coordination Chemistry," Birkhauser, Basel,
l3 F. H. Jardine, J. A. Osborn, G. Wilkinson, and J. F. Young, 1964, p. 240.
Chem. and Ind., 1965, 560. l9 R. G. Hayter, Inorg. Chew., 1964, 3, 301.
View Article Online

with [(C,H,),RhCl],. The molecular weight in solution It appears that the dissociation in equation (1) is
was quoted as normal, in contrast t o that of RhCl(PPh,), essentially complete in chloroform or benzene solution,
(see below); no chemical properties were reported. A where the apparent molecular weight remains constant
number of phosphite complexes, e.g., RhCl[P(OPh),], at about half the undissociated value over an ap-
were obtained 2o from [(CO),RhCl],. These were also preciable concentration range.
stated to have normal molecular weights; i t was also On heating benzene solutions of the chloride under
specified that tris-tertiary-phosphine complexes could reflux, the orange-pink complex of stoicheiometry
not be obtained from the carbonyl chloride. RhCl(PPh,), is obtained; over long periods (>24 hr.)
It is of some interest that 5-co-ordinate complexes of precipitation of the dimer occurs even from cold con-
stoicheiometry RhXL, seem to be unstable for alkyl centrated solutions in benzene. The dimer is very
or aryl phosphines or phosphites, despite the existence sparingly soluble in organic solvents but molecular
of 5-co-ordinate species such as Rh(C0)H (PPh,),,,l weights of this substance in solution under hydrogen
Rh(C,Hl,)(SnCl,)(PPh,),,22and RhH(PF3)4.23 It seems (see later) agree with a dimeric formula. A dimeric
reasonable to assume that the difference is due to steric halogen-bridged structure is the most likely, bearing in
factors. I t is pertinent to note that in the triphenyl- mind the well-established behaviour of closely related
phosphine complexes of ruthenium(r1) 24 and pal- rhodium(1) complexes of the type [RhClL,],. The
ladium(I1) 25 an a-hydrogen atom of a phenyl group dimer can be readily converted to RhCl(PPh,), by heat-
occupies an axial position of the metal atom in the ing with a 10 molar excess of triphenylphosphine in
crystalline complexes. The existence of red and orange ethanol. It may be mentioned here that although
forms of RhCl(PPh,), may well be due to differences in RhCl(PPh,), in the solid state, is stable indefinitely in
such secondary interactions. Furthermore, the form- air, the crystalline dimer is not. Oxygen is taken up
ation of the tris-triphenylphosphine derivative, which is and the infrared (i.r.) spectrum of the product shows a
stable only in the crystalline state, may be a result of the strong band at ca. 1150 cm.-l attributable to triphenyl-
stabilisation of a sparingly soluble crystalline lattice due phosphine oxide.
to these interactions. At the present time, we have no information concern-
At 25", the chloride complex is moderately soluble in ing the structure of the dissociated solution species
chloroform or dichloromethane (ca. 20 g./l.), slightly RhCl(PPh,), (A). If we assume that RhCl(PPh,), has
soluble in benzene (ca. 4 g./l.) or toluene, and consider- the square configuration common for Rhl, then on
ably less soluble in acetic acid, acetone and other ketones, consideration of trans-effects one would anticipate that
methanol, ethanol and other alcohols. It is virtually one of the trans-phosphine groups would be labile.
insoluble in light petroleum and cyclohexane. The However, i t seems unlikely that this can be the sole
bromide and iodide are somewhat more soluble in organic reason for this ready lability, and steric repulsion due
solvents. The complexes dissolve and react with donor to overcrowding b y the phenyl groups is likely to be a
solvents, (L), such as pyridine, dimethyl sulphoxide, strong and probably decisive contributory factor. I t
and acetonitrile, to yield complexes of the type seems unlikely that the formally three-co-ordinate
RhCl(PPh,),L. The solutions in dichloromethane or species, ( A ) , would retain the square configuration,
benzene quite rapidly absorb oxygen to give an oxygen with the fourth position occupied by solvent, particularly
complex 26 * (see later) which is more soluble in organic in rather weakly donating solvents such as chloroform or
solvents than is RhCl(PPh,),; the higher solubilities benzene. Furthermore, steric repulsion between the
reported for RhC1(PPh,),16 may be due to inadequate triphenylphosphine groups will tend to widen the
de-gassing of solvents. I t is essential for the determin- P-Rh-P angle, so that the most likely structure for
ation of molecular weights in solution that air be (A) would appear to be one with pyramidally disposed
rigorously excluded ; there are difficulties in ensuring triphenylphosphine, and chloro-groups. The essential
this when the Mechrolab Osmometer is used, but freez- point is that (A) has vacant co-ordination sites which
ing-point depression studies are in agreement with the can be occupied either by weakly bound solvent molecules
osmometric values (see Experimental section). or by other ligand atoms. The existence of vacant
The tristriphenylphosphine halides dissociate in sites is essential to the catalytic activity of (A) as we shall
solution, e.g., see.
soh e nt We note finally that RhCl(PPh,), gives non-conducting
RhCI(PPh,), -
+ RhCI(PPh,), + PPh, (I) solutions in acetone and dimethylformamide.
PPh,
(b) Chemical properties. (i) Reaction with carbon
and the solution species may dimerise: monoxide and aldehydes.-The co-ordinative unsatur-
ation of (A) is demonstrated by its reaction with donor
2RhCI(PPh,), [RhCI(PPh,)212 (2)
23 Th. Kruck and W. Lang, Angew. Chem. Internat. Edn., 1965,

* See behaviour of IrCl(CO)(PPh,), (ref. 26). 4, 870.
2o L. Vallerino, J . Chem. SOC.,1957, 2473. 24 S. J. La Placa and J . A. Ibers, Inorg. Chem., 1965, 4, 778.
21 S. S. Bath and L. Vaska, J . Amer. Chem. SOC., 1963, 85, N. A. Bailey, J . M. Jenkins, R. Mason, and B. L. Shaw,
3500. Chem. Comm., 1965, 237, 296.
22 J . F. Young, R. D. Gillard, and G. Wilkinson, J . Chenz. SOC., 28 L. Vaska, Science, 1963, 140, 809; J. A. Ibers and S. J .
1964, 5176. La Placa, ibid., 1964, 145, 920.
5 T
View Article Online
1714 J. Chem. SOC.(A), 1966

molecules t o form adducts, as noted above, but in RhCl(PPh,),(C,H,), can be readily isolated. Although
particular by its affinity for carbon monoxide, when the dipole moment measurements cannot be made, i t seems
well-known complex trans-bis(triphenylphosphine) car- that this square complex has trans-triphenylphosphine
bony1 chlororhodium(1) is formed. Thus, on saturation groups like the corresponding carbonyl complex ; the
of benzene or chloroform solutions of (A) with carbon 19F n.m.r. spectra of the corresponding tetrafluoro-
monoxide, an immediate and quantitative reaction ethylene complex 31 show that the tetrafluoroethyl-
occurs, the solution changing in colour from red t o group is trans to chlorine and cis t o two mutually trans-
yellow : triphenylphosphine groups.
RhCI(PPh,), +CO trans-RhCI(CO)(PPh,), (3)
Both the chloride and bromide complexes are stable in
the solid state but in solution lose ethylene rapidly.
I t appears to be impossible to reverse this reaction and This dissociation can be accelerated by sweeping the
even the action of molten triphenylphosphine a t 100" solution with nitrogen, but regeneration is quantitative
on the carbonyl complex does not displace carbon on resaturating the solution with ethylene. Thus, un-
monoxide. like the carbon monoxide reaction, this one is
The high affinity of (A) for carbon monoxide is also reversible :
shown by its rapid reaction with aldehydes under mild K2
conditions, when carbon monoxide is abstracted. Thus RhCI(PPh,), + C,H, RhCI(PPh,)2(C,H,) (4)
a solution of (A) in benzene reacts with n-heptaldehyde The n.m.r. spectrum of the ethylene chloro-complex
or valeraldehyde in the cold, or more rapidly on warm- has been measured in deuterochloroform solution in the
ing, to give quantitative yields of trans-RhCl(C0) (PPh,),. presence of varying amounts of ethylene (Table 1 ) .
From heptaldehyde, gas-liquid chromatography (g.1.c.)
shows that the other products are ca. SOYo n-hexane TABLE1
and ca. 20% hexenes. Benzaldehyde gives only the Nuclear magnetic resonance spectra of
complex and benzene. The latter reactions may be RhCI(PPh,),(C,H,) in chloroforin at 26"
regarded as a reverse of the hydroformylation reaction. C,H, (free) Line * Approximate
More detailed studies of the kinetics and mechanism of C,H, (co-ord.) position half-width
the reaction have been made 27 and will be reported 0 262 25
0-5 282 20
separately. During our studies, similar reactions of alde- 1-5 300 5
hydes with the complex have been briefly described; 28 3. 0 300 5
the reaction with carbon monoxide and abstraction a t 300 5
reactions with allyl alcohol and acetate ion have also * c./sec. on low field side of Me,Si. t Free C,H, value T 4.69
in CHC1,.
been noted.16 We have also demonstrated carbon nion-
oxide abstraction reactions with dimethylformamide on The spectra show only a single line due to ethylene,
warming, and with dioxan in the presence of an excess of whose position depends on the et hylene-rhodium
triphenylphosphine on warming. In the latter case, if an concentration ratio. At low ethylene concentrations
excess of triphenylphosphine is not present, heating the peak is broad, half-width ca. 25 c./sec., and even
gives only crystals of the dimer. Methyl ethyl ketone on cooling to -50", no sharpening or splitting was
also gives only the dimer and other ketones appear to observed, although a slight downfield shift (ca. 7 c./sec.)
be unreactive towards carbon monoxide abstraction. occurs due to the greater solubility of ethylene a t lower
I t may be noted that the interaction of rhodium tri- temperatures. With increasing ethylene concentration,
chloride hydrate in the presence of an excess of tri- the line sharpens and moves downfield towards the
phenylphosphine in solvents such as dioxan and di- position of resonance for unco-ordinated ethylene in the
rnethylformamide gives RhCl(C0)(PPh,), 29, and (A) solvent. As in other ethylene complexes, it is clear that
may be an intermediate in such reactions. A similar a rapid exchange between free and co-ordinated ethyl-
reaction is that of RhCl,(PEt,), with allyl alcohol or ene occurs in solution, even at -50". A mean lifetime
ethanolic potassium hydroxide, which again gives the for residence of ethylene on the metal atom can be
carbonyl complex.30 estimated as less than ca. sec., a value comparable
Finally, an adaptation 27 of the aldehyde reaction to those found by Cramer.32
employing aqueous formaldehyde as both a reducing By contrast, 19F n.m.r. spectra show that a t 25"
agent and a source of carbon monoxide, provides the there is no ligand exchange with the corresponding
most rapid and economical preparation of trans- complex of t etrafluoroe t h ylene .31
RhCl(C0)(PPh,), directly from RhC1,,3H20. The ethylene addition reaction (equation 4) can also
(ii) Interaction with ethylene. Ethylene, like carbon be followed spect ropho t omet ricall y . De-gassed solutions
monoxide, reacts rapidly with solutions of (A) in benzene of RhCl(PPh,), in benzene show a band a t 360 mp
or chloroform and the bright yellow, crystalline complex, (E -6000) as a shoulder on a much stronger band
bis(triphenylpliosphine)ethylenechlororhodium( I), present in the ultraviolet (u.v.) spectrum. On passing
27 M. Baird, J . A. Osborn, and G. Wilkinson, unpublished 3O J. Chatt and B. L. Shaw, Chern. and I n d . , 1960, 931.
results. 31 M. J . Mays and G. Wilkinson, J . Chem. Soc., 1965, 6629.
2 8 J . Tsuji and K. Ohno, Tetrahedvon Letters, 1965, 3969. 32 R. Cramer, J , Amev. Chem. SOC.,1964, 86, 217; Inovg.
29 A. Rusina and A . A. VlEek, ATature, 1965, 206, 295. Chern., 1965, 4, 445.
View Article Online

ethylene through the solution, or on dissolving the would appear that the main factor may be an electronic
crystals of RhC1(PPh,),(C,H4) in benzene saturated one. It seems most reasonable to assume that when
with ethylene, a new band at 416 mp (E -3000) appears alkyl groups replace hydrogen in ethylene, hypercon-
a t the expense of the 360 my band. This change is jugation and/or inductive effects lead to a reduction in
reversed on passing oxygen-free nitrogen through the the x-acidity of the carbon anti-bonding orbitals;
solution. The occurrence of isosbestic points at 434 and this in turn will lead to weaker metal- olefin
402 mp shows that only two species are involved in this x-bonding. Furthermore, for substituted ethylenes
reversible process. The relative concentrations of the there must be, of necessity, asymmetry in the metal t o
two species were estimated by subtracting the extra- C=C bond, the axis of which is doubtless a t some angle
polated background resulting from a strongly intensify- to the molecular plane other than the 90" found in
ing band further in the U.V. Thus, for ca. 4 x 1 0 - 4 ~ - ethylene complexes. Such distortion is likely to intro-
solutions saturated with ethylene a t 1 atmosphere, the duce steric factors particularly when, as in the present
ratio of the ethylene complex to the uncomplexed case, there are bulky ligands in positions cis to that
RhCl(PPh,), is ca. 20 : 1. Using the solubility of ethyl- occupied by the olefin. That electronic factors are very
ene in benzene a t 25" and 1 atmosphere (3.403 ml./ml.), important, however, is clear from the high stability
the approximate equilibrium constant for reaction and low lability of the tetrafluoroethylene complex.
(4) is K , -100 mole-l 1. Despite this low complexing ability of mono-olefins
The same procedure was used for propylene except other than ethylene, isolable complexes are easily
that the solution of the ethylene complex in benzene obtained with chelat ing diolefins,l6 hexafluorobut -2-
was swept with propylene at 1 atmosphere for 60 min. yne,,l and diphenyla~etylene.~~
during which time the spectrum was followed. The Finally, an important point concerning the inter-
band a t 416 mp soon decreased in intensity a t the ex- action of olefins with (A) is that, in contrast to a number
pense of the band a t 360 my. When an equilibrium had of well-established cases with other complexes of rho-
been reached, as indicated by constancy of the spectrum, dium, as well as with complexes of palladium and
no measurable intensity above the background could be other metals, there is no evidence for olefin isomerisation.
detected at 416 mp, while the band at 360 my had an For example, there was no cis--trans isomerisation of a
intensity virtually that found for the solution of 1 : 1 mixture of cis- and trans-hex-2-enes or isomerisation
RhC1(PPh3), alone. Hence the ratio of the propylene of hex-l-ene a t 25" over periods up to 24 hr. Slight
complex to that of uncomplexed RhCI(PPh,), must isomerisation occurred when hydrogen was present.
be less than the error in measuring the band intensities, (iii) Reaction with molecular hydrogen ; dihydrido-
say 50,;. ,4 similar result was found for hept-l-ene. complexes. ( a ) When saturated solutions of (A) in
Hence the complex constant for the formation of the solvents such as benzene, chloroform, ethyl acetate,
propylene complex, bearing in mind the greater solu- or glacial acetic acid are shaken with molecular hydro-
bility of propylene in benzene (ca. 17 ml./ml.) is a t least gen, the colour changes from red to pale yellow. Mano-
2000 times smaller than that for ethylene; thus K , metric studies indicate that in all of these solvents 1
<0.05 mole-l 1. Using RhBr(PPh,), the complexity mole of hydrogen is absorbed per mole of rhodium
constant for ethylene is lower by approximately a complex. In chloroform and benzene the dihydrido-
factor of five. We have been unable to isolate an iodo- species is more soluble than RhCI(PPh,),; the di-
complex and the constant is presumably much smaller hydrido-species in glacial acetic acid and ethyl acetate
still. The decrease in ethylene complex forming ability, have about the same low solubility as RhCl(PPh,),.
C1 > Br > I, would be expected, considering the increase Hence, depending on the concentration, pale yellow or
in the trans-effect of the halide if the halide is trans to white crystals of the dihydrido-complex may separate
ethylene. as the hydrogen uptake proceeds. N.m.r. spectra of these
There is relatively little information in the literature solutions, which contain one equivalent of displaced
concerning values for the complexity constants except phosphine show three high-field lines due to hydrogen
for ethylene and substituted styrenes 34b with PdlI atoms bound to rhodium. On sweeping the solutions
33934a
or PtZ1. However, Cramer, in his work on ethylene with nitrogen, the high-field lines disappear and the
dimerisation 35 quotes unpublished results on complexes solution regains its initial red colour. Re-saturation
of the type [RhCl,(olefin) (solvent)]-, stating that ethyl- with hydrogen restores the yellow colour and the n.m.r.
ene is fa\-oured by a factor of about lo3 over but-l-ene. signals. The cycle can be repeated, apparently in-
Similarly, in studies on olefin oxidation by palladium(I1) definitely, showing a reversible equilibrium of the
solutions, it appears qualitatively that, the equilibrium type :
concentrations of the olefin complexes are lower for RhCI(PPh,), + H,
K,
RhCI(PPh,),H, (5)
propylene and but-l-ene than for ethylene.36
Since quite bulky substituents can evidently be Analogous behaviour was observed for solutions of
present on the olefin, a t least, in certain complexes,34bit RhBr(PPh,), and RhI(PPh,),.
33 P. M. Henry, J . Amev. Chem. SOC., 1964, 86, 3246, and 3jR. D. Cramer, J . Amer. Chem. SOC., 1965, 87, 4717.
references therein. 36J . Smidt, Chem. and Ind., 1962, 54.
34 ( a ) S. 1. Shupack and M. Orchin, J . Amer. Chem. SOC.,1963, 37D. N. Lawson, J . A. Osborn, and G. Wilkinson, following
85, 908; ( b ) ibid., 1964, 86, 586. Paper.
View Article Online
1716 J. Chem. SOC.(A), 1966

Absorption of molecular hydrogen has not been lines are attributable to the spin-spin interaction of H
previously reported for a rhodium complex, but the with ,lP. Additional confirmation that this is so, comes
uptake of hydrogen a t normal pressure by the iridium
complex IrCl(CO)(PPh,), is well known.38 The solu-
bility of this hydrido-species was too low for n.m.r.
study, but the corresponding bromide and iodide have
been stated39 to be sufficiently soluble as to yield

n.m.r. spectra which are consistent with an octahedral
species with cis-hydrido-groups, a result in confirmation
of i.r. data. It has been reported40 that the dihydrido-
derivatives of the corresponding rhodium and cobalt
complexes could not be obtained. The uptake of hydro-
gen by the carbonyl species and their ability to act as
hydrogenation catalysts is discussed a t a later stage.
( b ) The nature, spectra, and structure of dihydrido-
species. There is no change in the non-conductance of
acetone solutions of the chloro- and bromo-complexes FIGURE 1 Nuclear magnetic resonance spectra a t 56.4 &Ic./sec.
RhX(PPh,),, (X = C1 or Br), when they are saturated in chloroform solution a t 25" of RhCl(PPh,),H,
with hydrogen; this indicates that the halide ion re-
mains co-ordinated, a result further confirmed by the TABLE2
analysis of the isolated solids. Although crystalline N.m.r. and i.r. data on dihydrido-complexes of RhX(PPh,),,
solids can be obtained fairly readily from hydrogen- X = C1, Br, or I : high-field lines and Rh-H modes
saturated solutions, their characterisation has not been only. N.m.r. relative intensities in parentheses
v cm.-1
easy, due both to the ease of aerial oxidation of, and Solids c
ready loss of hydrogen from, the solutions. Rh-H str. :
The microcrystalline white solid isolated from hydro- Solvent T a Solutionb Rh-H bend
gen saturated chloroform solutions of (A) analyses for Chloride
CDCl, 28.2(2), 21.5(1), 2095m, 2050m 2078m, 2013m;
RhCl(PPh,),H, ; the molecular weight under hydrogen 18-8(1) 785m
is normal. The species may possibly be 5-co-ordinate C6H6 26.5(2), 20.4(1), 2066sh, 2035m 2066sh, 2051m;
in the solid, but in solution 6-co-ordination with solvation 17*8(1) 805m
C,H,N 28.6(1), 27-7(1) 2078sh, 2052m 2143m, 2075m; 8
by chloroform is most likely. The formation of solvates 838n1, 812m
with chlorinated hydrocarbons is confirmed by the isola- CH,CI, - -_ 2099m,J 2065m;a
tion from dichloromethane solutions of a stable yellow 845s
MeC0,Et 28.2(2), 2 1.5(1) - 2075m, 2088m;
crystalline complex [RhC1(PPh,),H,],,CH,C12,which 18*8(1) 840m, 775m
apparently contains quite firmly held and, we propose, MeNCHO 28.2(2), 21.5(1), -. 2137m, 2081m;
18*8(l) -
co-ordinated, bridging dichloromethane. Using ethyl
Bromide
acetate, glacial acetic acid and dimethylformamide as CDC1, 28.2(2), 21.5(1), 2177sh, 2057m - -
solvents for RhCl(PPh,),, white solids may again be 18*8(1)
isolated. The analyses, whilst not entirely satisfactory, 'IjH6 26-5(2),20.4(1), 2066sh, 2035m - -
17*8(1)
show that two triphenylphosphine ligands are present C,H,N 29.2(1), 28-4(1) 2081sh, 2049m 214Pm, 2099m;
for each rhodium atom, and that co-ordinated solvent 845m, 830m
is present. The latter is also clearly shown by the pres- CH,Cl, - - 2103m, 2068m;
847m, 838m
ence of co-ordinated solvent bands in the i.r. spectra of
Iodide
the complexes. Dissolution of these solvated complexes CDCl, 26*2(2),23*1(1),
(which are very sparingly soluble in the parent solvent) 20.5(1)
in deuterochloroform gives an n.m.r. spectrum identical a Measured in CDC1, except where indicated ; values relative
with that given by hydrogenation of RhCl(PPh,), in t o Me,Si as internal reference. Measured in parent solvent.
Spectra taken in Nujol mulls. In pure pyridine: for both
deuterochloroform. Hence these solvates appear to chloride and bromide, T 28.1(1), 27.8(1). e Bands due to
have the same basic stereochemistry as the 5-co-ordinate co-ordinated solvent are also present. f For deuteride, 1507
species in chloroform solution. cm.-l.
The high-field n.m.r. spectra (Figure 1 and Table 2) at Note : The n.m.r. spectra of highly concentrated solutions
may show lines due t o [RhX(PPh,),H,],. For the C1: in
56.4 Mc./sec. consist of three peaks (a weak shoulder CDC1,~27-4; in C6H6 T 28.9. For the Br: in CDCl, T 29.1 ; in
may occasionally be observed on the highest field line- C6H, T 25.6.
this is due to the tetrahydrido-complex of the dimer
discussed later). The peaks A,A' are of intensities which from a study of the corresponding triphenylarsine com-
are equal and half that of peak B. At 40 Mc./sec., the p l e ~ , which ~l has only two lines of equal intensity in
A,A' separation remained constant, showing that these 39 L. Vaska, quoted by A. P. Ginsberg, A d v . Tv ans. Met.
Chem., 1965, 1, 111; cf. also refs. 11, 12.
s* L. Vaska and J. W. Di Luzio, J . 'Amer. Chem. SOC.,1961, 40 See R. S. Nyholm, ref. l ( b ) .
83, 1262; 1962, 84, 679. 41 J. T. Mague and G. Wilkinson, J . Chem. SOC.( A ) , 1966, 1736.
View Article Online

positions similar to those in the phosphine complex. hydrogen bonds. This is to be expected for cis-di-
The absence of fine structure on the lines is discussed hydrido-groups. The dideutero-complex,
below. [R hC1(PPh,) 2D2]2, CH,CI, obtained from dichloromet hane
The spectra are most readily interpreted in t e r n s of solution shows only one Rh-D stretching frequency
an octahedral cis-dihydrido-complex (I) of RhIII formed (vRh-= = 1507 cm.-l; V H / V ~ = 1.39), the other being
from (A) by the addition of hydrogen to give two obscured, presumably, by triphenylphosphine absorp-
Rh-H bonds. t ion.

H H (c) Broad nature of the high-field n.m.r. lines. The
high-field lines of the cis-dihydro-species are rather
broad (ca. 45-70 c./sec.). Fine structure would have
been expected to arise (1) from H-lo3Rh coupling (ca.
~ from HA-HB coupling and (3)
20-30 c . / s ~ c . ) , *(2)
X " ci from additional cis 31Pcoupling. Although such coupl-
(1 1 (m ings will appear under the envelope of the broad high-
field lines, no resolution could be obtained. This
P , PI. and P, = PPh,; X = halide; S = solvent: Py = pyridine
behaviour is in striking contrast to the situation with
For a number of octahedral hydride-phosphine com- the hydrido-species obtained from the dimer.
plexes it has been found that H31P coupling constants The dimer [RhCl(PPh,),], when dissolved in di-
are of the order 80-150 c./sec. for H trans to P, while chloromethane saturated with hydrogen takes u p
the couplings for H cis to P are ca. 15 c./sec. Further- exactly two moles of hydrogen per mole of complex.
more, the T values for H trans to P are generally lower The hydrido-species is reasonably stable, even in air,
than those when H is trans to a halide or a similar and the molecular weight is that for the dimer, showing
ligand.42 The observed spectrum is consistant there- that no cleavage has occurred on hydrogenation. The
fore with (I); a similar spectrum would result if X and S n.m.r. spectrum shows a well-defined fine structure of
in (I) were interchanged and a t present we cannot be overlapping, resolved, double triplets centred a t about
certain of the exact dispositions (see later discussion, 5 26.6. This spectrum results from the interaction of H
however). with lo3Rhto give a doublet (splitting 20 c./sec.), which is

I t is possible that in chloroform or benzene solution, further split into a triplet by two equivalent 31P nuclei
there is an unsolvated 5-co-ordinate species which (splitting 15 c./sec.); it is consistent with a structure
could have a spectrum very similar t o that of solvated involving octahedral rhodium(II1) with trans-hydrogen
species, however, we believe that solvent occupation atoms as in (111). The n.m.r. signal due to this species
of the vacant site always occurs in solution. The solu-
tion species are in any case likely'to be distorted from
true octahedral geometry. I n order to have the ob-
served tmns-coupling it is only necessary that the
hydrogen shall experience the presence of the phos-
phorus atom in the trans-position and this is possible
even if slow ligand exchange is proceeding.
The pyridine solutions of (A) after hydrogenation, the
isolated dihydro-pyridine complex, and the tetrahydrido- is occasionally observed in concentrated hydrogenated
pyridine cleavage product of the dihydrido-dimeric solutions of RhC1(PPh3),, especially when benzene is
species discussed below, all have an identical spectrum used as solvent. On the addition of pyridine or tri-
consisting of two rather broad lines. This is consistent phenylphosphine to solutions of the dimer in the cold
with trans-phosphines and cis-dihydrides as in (11); under hydrogen, cleavage occurs and the process can be
the 7 values are in the region for cis-H-P groups, in followed by changes in the high-field n.m.r. spectrum.
agreement with i.r. data. Although we have not isolated Since trans-dihydrido-groups on the rhodium atom
an acetonitrile complex, the n.m.r. spectra of hydrogen- gave the expected fine structure, i t seemed likely,
ated acetonitrile solutions of (A) are identical with those originally, that some exchange process could be occur-
of the pyridine complex. I t is of interest that the re- ring in the cis-dihydrido-systems which could lead to
organisation of the hydrido-complexes in chloroform line broadening. However, there was no detectable
solution on addition of pyridine or acetonitrile can be change in the appearance of the lines (1) on cooling to
followed via the growth and decay of the respective -70" or (2) on exerting a pressure of ca. 100 atmo-
high-field lines. spheres of hydrogen on the solution in an n.m.r. tube.
The infrared spectra (Table 2) of the various species in I n these solutions under pressure the proton resonance
both the solid state in mulls and in solution generally of dissolved molecular hydrogen 44 a t T 5-57 was readily
show two broad Rh-H stretching frequencies in the observed in addition to all three high-field lines, which
normal region, ca. 2000 cm.-l for transition-metal and
43 J. A. Osborn, R. D. Gillard, and G. Wilkinson, J . Chem.
42 J . Powell and B. L. Shaw, J . Chem. Soc., 1965, 3879, and Soc., 1964, 3168.
references therein. 44 D. F. Evans and J. P. Maher, Chem. and I n d . , 1961, 1960.
View Article Online
1718 J. Chem. SOC. (A), 1966

were unchanged in position. From this fact the mean noted also that, in solution both RhCl(PPh,), and the
lifetime of a hydrogen molecule on the hydrido-complexes corresponding arsine complex take up 1 mole of oxygen
can be estimated as greater than ca. lop3sec. and that solid adducts can be isolated. These and other
On the addition of an excess of PPh,, the line breadth reactions will be described separately.
is apparently reduced suggesting, either some inhibition (B) Catalytic Hydrogenation of Olefins and Acetylenes.
of ligand-dissociation and exchange, or reduction of the -(a) General remarks. In benzene or similar solvents,
concentration of the paramagnetic impurity discussed the tris(tripheny1phosphine)halogenorhodium( I) com-

below. That ligand exchange can occur, however, plexes are extremely efficient catalysts for the homo-
has been shown 41 by the addition of triphenylphosphine geneous hydrogenation of non-conjugated olefins and
to solutions of RhCl(AsPh,),H,, when the changes in the acetylenes a t ambient temperature and pressures of
n.m.r. spectrum can be followed. 1 atmosphere or below. Functional groups such as
The main and unexpected cause of line broadening in keto-, hydroxy-, cyano-, nitro-, chloro-, azo-, ether, ester
the cis-dihydrido-species is due to the presence of trace or carboxylic acid are not reduced under these conditions.
quantities of a paramagnetic species in the starting Although in the present Paper we are concerned with
material RhCl(PPh,), even when this is prepared under quantitative measurements on hept- 1-ene, cyclohexene
nitrogen . and hex-1-yne, a wide variety of unsaturated compounds
(d) Electron spin resonance measurements. Solutions containing olefinic and acetylenic bonds have been
of RhCl(PPh,), in de-gassed dichloromethane at 25" studied. Details will be given separately; however, it
show a weak electron spin resonance (e.s.r.) signal. appears that ; (1) terminal olefins (including non-con-
The two lines at g = 2.19 and g = 2.05 form a clear jugated, non-chelating diolefins such as octa-l,7-diene)
doublet structure centred on g = 2.12, splitting 190 are reduced more rapidly than internal olefins, e.g.,
gauss, with an average line width, AH; -60 gauss. hex-1-ene > hex-2-ene ; (2) cis-olefins are reduced faster
This signal is retained on hydrogenation of the sample. than trans-olefins, eg., cis-hex-2-ene > trans-hex-2-ene ;
Further details of these and related studies will be re- (3) conjugated olefins, such as butadiene or cyclohexa-
ported separately. I t is clear, however, that the broaden- 1,3-diene are not hydrogenated a t 1 atmosphere,
ing of the n.m.r. high-field lines can be attributed to the although they may be reduced a t higher pressures
shortening of the relaxation time of all the nuclei in the (ca. 60 atmospheres) ; (4) chelating non-conjugated
sample by the direct effect of the electron dipole. diolefins, e.g., cyclo-octa-l,5-diene, which may, and in
Further broadening could, of course, result from chemical some cases, do, displace the phosphine ligands from
(including electron) exchange between the diamagnetic the complex, nevertheless, may be hydrogenated very
species and a small concentration of a paramagnetic slowly at 1 atmosphere but again rapidly under pressure ;
rhodium(I1) species. The concentration of the latter is (5) in contrast to some other homogeneous catalytic
estimated as 1 in lo4. systems, ethylene is not hydrogenated (see Discussion
Since the original rhodium source material, RhC1,,3H20 section), although propene, butenes, and higher olefins
shows only a very feeble e.s.r. signal, the paramagnetic are effectively hydrogenated a t 1 atmosphere pressure.
species, perhaps RhC12(PPh3),,must be produced in the The rates of reduction compare favourably with those
reduction process and collected with the crystalline obtained by heterogeneous catalysts such as Raney
RhCl(PPh,),. There is a possibility that the extraneous nickel or Adams catalyst. A comparison of the rates
paramagnetic impurity in RhCl,,SH,O is being selec- of hydrogenation of cyclohexene under normal operat-
tively complexed ; however, this seems unlikely, first ing conditions by RhBr(PPh,), and Adams catalyst
because of the nature of the e.s.r. signal and secondly (equimolar amounts of the metal) is shown in Figure 2.
because RhCl,,SH,O is believed to be free from transition- Using lo-, M-solutions of the catalyst in benzene with
metal salts other than perhaps a trace of iridium. hydrogen pressures of ca. 1 atmosphere, the hydrogen-
Since the dimer hydride has sharp n.m.r. lines, the ation of olefins may be so rapid that the temperature of
dimerisation process in hot benzene must leave the the solution is raised to the boiling point. For example
RhII species in solution. However, on reconversion of moles of RhCl(PPh,), in benzene (80 ml.) containing
the dimer to the monomer by heating the dimer with an cyclohexene (16 g.) absorbed hydrogen at ca. 30 ml./min.
excess of triphenylphosphine in ethanol under reflux, a t 25", which corresponds to the generation of ca. 40
the paramagnetic impurity is regenerated. We have so cal./min. in the solution. Under identical conditions
far failed to obtain a sample of RhCl(PPh,), which is free the bromide and iodide had absorption rates approxi-
from paramagnetic contamination. mately two and three times as fast, respectively.
(iv) 0ther reactions. Tris (triphenylphosphine)chloro- In the kinetic studies, which were made a t constant
rhodium or its solutions interact with a great variety of volume, i t was necessary not only to have thermal
ligands. Thus, they add various molecules 45 in a way control of the reaction vessel, but also to ensure that the
similar to the reactions of square complexes. Methyl reactions were not diffusion controlled due to inadequate
iodide gives the complex RhClI (CH,) (PPh,),ICH, where saturation with hydrogen. Details of the apparatus and
CH,I acts as a donor. Ally1 chloride adds to give allylic the procedure used are given in the Experimental section.
complexes of stoicheiometry, RhCl,(PPh,),(C,H5), while a5 M. C. Baird, D. N. Lawson, J . T. Mague, J . A. Osborn, and
hydrogen chloride gives RhCl,(PPh,),H. It may be G. Wilkinson, Chem. Comm., 1966, 129.
View Article Online

The solid catalyst was added t o the hydrogen- stirring in of hydrogen the colour lightens t o a clear
saturated solution, before the addition of olefin. This yellow-orange. If the stirring is stopped, over a period
procedure was found t o be desirable since the solid (ca. 30 sec. under normal conditions), this ~ 7 e l l 0 ~ -
dissolves rather slowly in benzene whereas the dihydro- orange deepens again to the deep red-brown. The
complex is more soluble, thus assisting dissolution ; reasons for this sequence of changes will be discussed
it also prevents any appreciable prior dimerisation later.
of RhCl(PPh,), in the more concentrated solutions a t Although the trace quantities of hydroperoxides
1 present in cleaned and dried olefins after storage do not
appreciably inhibit catalytic hydrogenation, neverthe-
28 - less, at high olefin concentrations some fall off from
maximum rates is observed if the hydrocarbons are
24 - not freshly purified. I n the quantitative work presented
here, the olefins were distilled under nitrogen or in
20 - vacuo into a burette attached to the reaction flask
$ before use.
-
I
a
I
16 -
12 -
(b) Quantitative measurements. The overall rates of
hydrogenation of hept-1-ene, cyclohexene and hex-1-yne
have been measured and the dependence on temperature,
WQ
pressure, substrate concentration, catalyst concentra-
8 - tion, catalyst nature and co-solvents and/or poisons
investigated.
The rate of hydrogen uptake was obtained from the
tangent to the plot of hydrogen pressure against time
in a closed system. The standard hydrogen pressure
I20 240 360 480 used was 50 cm., and conversion to the rate of consump-
T i m e (sec.) tion of substrate (S) in mole 1.-l sec.-l was made using
FIGURE 2 Qualitative comparison of the rates of homogeneous the stoicheiometric relationship -d[H,]/dt = -d[S]/dt.
and heterogeneous catalysis, (0) moles of RhBr(PPh,), The solubility of hydrogen was considered constant from
in 1 : 1 benzene-ethanol, ( x ) moles of Adams' catalyst in run to run and was assumed to obey Henry's Law;
glacial acetic acid. Solvent volume, 50 ml. with 10 mi. cyclo-
hexene a t 25'. There is no detectable uptake of hydrogen partial pressure of benzene and olefins were taken from
under these concentration conditions and on this time scale standard tables. The amount of substrate consumed a t
using either RhCl(C0)(PPh,), or IrCl(C0)(PPhJ2. The latter the time of measurement (at a pressure of 50 cm.) is
shows very slow uptake a t 40' using hex-1-ene
less than 1% and is considered negligible. The results
temperatures above 25". Furthermore, since the hydrido- shown graphically are generally the average of a t least
species is a very pale yellow, the presence of undissolved two determinations and are reproducible to 4 5 % ; two
catalyst is readily detected, whereas this is sometimes independent sets of apparatus were used, one in which
difficult to ascertain in the deep red solution given by the the olefin was distilled in nitrogen, the other in vaczfo.
catalyst in benzene alone. The addition of olefin There was no evidence of any heterogeneous reaction,
(without stirring) causes immediate darkening of the and identical results were obtained when glass helices
pale yellow solution to deep red-brown; on rapid were added to the reaction flask.
20 -
15 -
.-i
View Article Online
1720 J. Chem. SOC. (A), 1966

(i) Dependence on substrate concentration. A typical (ii) Dependence on hydrogen pressure. The qualita-
plot showing the rate of hydrogen consumption against tive dependence of the rate on hydrogen pressure can be
hept-1-ene concentration is shown in Figure 3. The rate seen from Figure 2. From results under standard
does not increase linearly with increasing olefin concen- conditions, the reciprocal of the rate against the re-
tration but approaches an asymptotic value. The plot ciprocal of the pressure (Figure 6) is a linear plot.
of the reciprocal of this rate (in mole-l 1. sec.) against Values for two catalyst concentrations are given.
'i' 14 -
-
0
X
u
vl
Y
--
'0)
l0-
0
E
I I I I
1.0 2.0 3.0 4.0
1
I
moli? 1.
[0I a f i n]
4 Plot of reciprocal of rate of hydrogenation of hept-1-ene ( x ) and of cyclohexene
FIGURE (0)
against the reciprocal of the olefin
concentration in benzene a t 25" with 1.25 mM-RhCl(PPh,) ,
the reciprocal of the olefin concentration (Figure 4) is Thus, assuming Henry's Law is obeyed, 1/R = a(l/p') +
linear with a positive intercept on the y-axis. Figure 4 p where 9' is the hydrogen pressure and u and p are
also gives. comparable data for cyclohexene. The
reciprocal of this intercept gives the maximum rate
possible under the specified condition, i.e., the asymp-
totic value in Figure 3. Similar results for hex-1-yne
are given in Figure 5.
12 -
lo -
I
I I I I 1
' I 1.0
I
2.0
-I
3.0
2
4.0 5.0
-cm. xIO
0.5 1.0 2.5
1.5 2.0
4%1
'
[HQx-l-yn~]
mole-' 1. FIGURE 6 Dependence of rate of hydrogenation of cyclohexene
with hydrogen pressure in benzene a t 25'; 2-50 mM-concen-
FIGURE 5 Plot of reciprocal of the rate of hydrogen con- trations of RhCl(PPh,), (O),1.25 m M ( x )
sumption by hex- 1-yne against the reciprocal of acetylene
concentration in benzene a t 25" with 1.25 mM-RhCI(PPh,), constants. From the plots we obtain the following data :
1.25 mM catalyst: u = 2.04 x ml. cm. min.-l,
From these results it can be seen that the rate (R) is p = 2.4 ml.-l min. 2-50 mM. catalyst: cx = 1.52 x
related to the substrate concentration [S] by the relation- ml. cm. min.?, p = 1.2 m1,-1 min.
ship 1/R = C(l/[S]) C'. + (iii) Dependence on catalyst concentration. A plot
View Article Online

of the reciprocal of the rate against the reciprocal of the lyst concentration for various concentrations of cyclo-
olefin concentration as in (1) for three catalyst concen- hexene is shown in Figure 8; in the absence of catalyst
trations is shown in Figure 7. The intercepts on the no hydrogen uptake was detectable. The data suggest
pseudo first-order behaviour, but there is evidently a
TABLE3 small intercept on the rate axis, which is discussed later.
Data derived from plot of reciprocal of rate us. reciprocal of (iv) Dependence on temperature. Rates were
cyclohexene concentration for different values of measured at four temperatures ranging from 15-30' for
catalyst concentration; 25", RhCl(PPh,), three cyclohexene concentrations. From the plots of the
Rate
Catalyst Intercept Constant k' reciprocal of the rate against the reciprocal of the olefin
conc. mole-' 1. Gradient Ratio * mole-'
m M / 1 . sec. x 10-3 sec. x lo-, Intercept/gradient set.-' 1.t TABLE4
0.625 4.25 8-50 0.50 0.18 Data derived from the plot of the reciprocal of the rate us.
1-25 3.00 5.25 0.57 0.15 the reciprocal of cyclohexene concentration for varying
1.875 3-25 4.00 0-56 0.13 temperatures
* Assumes concentration of dissolved hydrogen a t 50 cm. Gradient Rate constant k',
pressure, 25" approximately 2.5 x lo-, mole 1.-' (see M. W. T oK sec. x 10-3 mole-, sec.-l 1. x lo2 log,, k'
Cook, University of California, Radiation Laboratory Report
No. 3459). 7 Calculation assumes that the concentration of 288 20.7 3.9 - 1.409
the hydrido species is the same as that of the catalyst. 293 10-6 7.5 - 1.125
298 5.4 14.8 -0.830
303 2.6 30.8 - 0.5 12
y-axis and the gradients are given in Table 3 together
with rate constants derived from them as discussed concentration, which remain linear throughout the
later. The dependence of the rate of reduction on cata- temperature range, the gradients given in Table 4 can
0.5 I. 0 1.5 2-0

I
mole-' I.
[Cyc lohexene]
FIGURE
7 Plot of the reciprocal of the rate of hydrogenation of cyclohexene us. the reciprocal of olefin concentration for catalyst
concentrations of (0) 0.625, (v)
1-25, ( x ) 1.875 m~-Rhcl(PPh,),in benzene a t 25"
t
H
-(Y
: TE 10
I I I I
0.5 I. 0 1.5 2.0
[RhCl (PPh,),] mrnole/l.
FIGURE
8 Rate of hydrogenation of cyclohexene as a function of catalyst concentration in benzene a t 25" with RhCl(PPh,),;
cyclohexene (0) 2 - 4 9 ~ (V)
, 1 . 2 4 5 ~ (, x ) 0 . 8 7 ~
View Article Online
1722 J. Chem. SOC.(A), 1966

be obtained. The calculated rate constant, k' (see later) for the three halides and assumes that the ratio of the
is also given, and a plot of log,,k' against 1/T x lo3 is rates is constant for all olefin concentrations.
shown in Figure 9. A good linear plot is obtained, indicat- (vi) Effect of co-solvents. The addition of polar
ing that the kinetics of the system are not complicated. solvents such as alcohols or ketones to the benzene solu-
tions normally employed gave a substantial increase in
the rate of hydrogenation of the substrates. Figure 11
gives relative rates under otherwise identical conditions

for benzene, 1 : 1 benzene-ethanol and 1 : 1 benzene-
methy lethyl ketone; as for Figure 10, the ratios of the
intercepts to the gradients are constant. The overall
rate is approximately doubled using co-solvent. There
is no correlation of the rate increase with dielectric con-
stant of the co-solvent. The solubility of hydrogen
is not significantly different in the different solvent
mixtures. Other co-solvents that can be used are,
methanol, t-butyl alcohol, isopropanol, phenol, glacial
acetic acid, ethyl acetate, dimethylformamide, dioxan
and dimethyl sulphoxide. Some of these solvents may
be used alone, although the low solubility of the catalyst
in them reduces the effectiveness. Using pure alcohols,
slow abstraction of carbon monoxide may occur over a
9 Plot of log,, k' us ( ~ / T ' K )x
FIGURE lo3 from data given in period of time (several hours at 25") leading to de-
Table 4 activation of the catalyst. More strongly co-ordinating
substances inhibit the rate of hydrogenation and in high
The effect on the kinetics of the small differences in concentrations totally prevent reduction at 25" and
solubility of hydrogen over this temperature range can 1 atmosphere. Such materials are : triphenylphosphine,
be neglected. thiophen, 8-hydroxyquinoline, pyridine and acetonitrile.
0-5 1.0 1.5 2.0

I m o l d ' 1,
[Cyclohexene]
FIGURE
10 Dependence of rate of hydrogenation of cyclohexene on the nature of the catalyst in benzene a t 25". 1.25
mM-Catalyst (0) RhCl(PPh,),, ( x ) RhBr(PPh,),, (v)RhI(PPh,),
From Figure 9, a value for the activation energy, E,, Ethylene, conjugated diolefins and oxygen also poison
of 22.9 kcal. mole-l is obtained; AH$ is 22.3 kcal. the low pressure activity but hydrogenation at higher
mole-l and ASS is +12.9 e.u. pressures is readily effected.
(v) Dependence on the halide. The relative rates for (c) Discussiort of kinetic results. The experiments
RhX(PPh,),, (X = Cl, Br, or I) under identical con- discussed under Section (A) (b) show that in solution,
ditions and using equimolar quantities of the complexes the dissociated catalyst species, RhX(PPh,),, can
is shown in Figure 10. The lines have been drawn so rapidly attain equilibrium with dissolved hydrogen
that the ratio of the intercept to the gradient is constant. and/or olefin to give a complex. For the hydrido-
This is necessary for a valid comparison of the rates complex there is a vacant co-ordination site and for the
View Article Online

olefin complex, two. Thus in the transfer of hydrogen tion and [S] and [A] are the substrate and catalyst
molecules in solution to the olefin in solution via a concentrations, respectively, then the rate, R, is:
catalyst complex species, the rate determining step can
R = - - =d[SI (KK, + KK,)p[SI
be either one or both of two possible paths: (1) attack of
uncomplexed olefin on the dihydrido-complex a t the dt 1 + K19 + K,[SI
vacant site to give a transition state in which both This rate expression assumes the same form if k = 0
hydrogen and olefin are bound to the metal; (2) attack and there is good reason to believe, as outlined later,
of molecular hydrogen on the olefin complex, that this is so, i.e., the attack of hydrogen on the olefin
I I I I
0.5 1.0 1.5 2-0
I mole- 1.
[C y c I oh ex Q n e]
11 Dependence of rate of hydrogenation of cyclohexene with co-solvent (1 : 1 with benzene) a t 25; 1-25mM-RhCl(PPh,),.
FIGURE
(0) Pure benzene, ( x ) methyl ethyl ketone, (0) ethanol
RhX(PPh,),(olefin), to give the same transition state complex does not occur. The rate expression then
as in Scheme 1. reduces to:
From the overall rates of hydrogenation measured, kKlPF1[A1
R=
these paths are indistinguishable and are kinetically
equivalent if it is assumed that the equilibria are set up
1 KlP K,FI + +
quickly. Although, from other experiments described This expression is found to reasonably accommodate
later it appears that (1) is the preferred path, we shall most of the experimental data described earlier.
write the formal scheme [(A) = RhX(PPh,),] : (i) The dependence of the rate on substrate concen-
tration is :
oie:AJ/ \*--,,
--.
ik
olefin
which is identical in form to that obtained in Section
k (B) (b) (i) with
(A)(olefin) __t (A) + paraffin
He
Scheme I
The colours of the reacting solutions noted earlier give Thus for effective catalytic reduction of a substrate,
some indication of the competition between the rapidly k must be large and K, small; however, it is necessary
established equilibria K , and K,. Thus (A)H, is pale to the form of the results that K , should have a finite
yellow, (A) + olefin) fin) deep red-brown (except for value.
ethylene as noted before), while the reacting solutions For ethylene, which of all simple mono-olefins has by
are normally orange, although a t high olefin concentra- far the largest value of K,, catalytic reduction does not
tions or low hydrogen pressures they darken towards occur, even although, as discussed later, transfer of
red-brown. hydrogen from the dihydrido-species to ethylene is
If Kl and K, are the equilibrium constants and k rapid.
and k the rate constants for the reactions as shown in I t is appropriate a t this point to discuss competitive
the Scheme, p is the concentration of hydrogen in solu- experiments involving ethylene and trans-stilbene.
View Article Online
J. Chem. SOC. (A), 1966

If ethylene (ca. 25 cm. pressure) is admitted to the Thus both a and (3 should be proportional to 1/[A] and
benzene solutions of RhCl(PPh,), the bright yellow it can be seen from the values given in (B) (b) (ii) that
complex is formed immediately, as discussed earlier. this is approximately correct.
Admission now of hydrogen (ca. 25 cm. pressure) pro- (iii) From the above section (i) it can be seen that the
duces no change in colour, nor is any gas absorbed. gradient of the plot of 1/R vs. 1/[S] (Figure 7) becomes,
This is in striking contrast to the behaviour of propylene l/k[A] if K,p is >1. The latter is evidently so from the
and but-1-ene under identical conditions, where rapid measurements of the electronic spectrum of the catalys-
hydrogenation takes place. Furthermore, if cyclo- ing solutions, which is that of the pure hydrido-species.
hexene is added to the bright-yellow complex solution Using this relationship, from the data in Table 3, we
under ethylene-hydrogen (1 : 1,total pressure ca. 50/cm.) obtain the rate constant k for the rate-determining step;
no hydrogenation occurs-ethylene is thus an effective the values of k are also listed in Table 3 and the average
poison for the hydrogenation of cther olefins. When value 0.153 was used in calculations. The rate constants
the initial procedure was reversed and the dihydrido- used in the Arrhenius plots are derived similarly from
species were prepared first, admission of ethylene was Figure 7.
followed by a small rapid absorption which rapidly The intercepts and gradients should be proportional
ceased; the solution again assumes the yellow colour to 1/[A] and from Table 3 it appears that this dependence
(and has the absorption spectrum) characteristic for does hold approximately. Moreover, i t can also be seen
RhCl(PPh,),(C,H,). We can thus conclude that this that ratio of the intercept to the gradient for the plot
complex cannot activate molecular hydrogen. The (Table 3) is equal to K,/K,p, which gives K,/K, ap-
possibility that it can activate molecular hydrogen but proximately 1.6 x lo-, mole-, 1. This value, although
that the transfer step to co-ordinated ethylene cannot quantitatively rather smaller than expected, assuming
occur can be dismissed, since a transition state or inter- reasonable values for K , and K,, is, nevertheless, of
mediate complex would have to be identical to that the right order of magnitude.
formed in the reaction of the dihydrido-species and ethyl- The dependence of the rate on the catalyst concentra-
ene. Furthermore, a t high concentrations, n.m.r. tion should be linear:
studies of the ethylene complex under hydrogen show
no high-field lines. The implication is clear that in the R= KK1p[Sl . [A]
kinetic mechanism proposed, we can eliminate the path 1 + KlP + K,PI
involving hydrogen attack on the olefin (or acetylene) However, although a linear dependence was observed
complex. I t might be expected from the trans-effects of in Figure 8, the line does not pass through the origin as
halide on the ethylene complex formation constant required by this equation. This divergence could be due
[Section (A) (b) (ii)] that for the bromide and especially to some dimerisation or reassociation of the catalyst a t
iodide, the higher concentration of uncomplexed species the higher concentrations, thus effectively decreasing the
might allow hydrogenation to proceed. For the iodide rate values and consequently shifting the plots away
only, very slow hydrogenation of ethylene does occur from the origin but still maintaining the approximately
(initial absorption rate 0.5 ml./min.) with a 1 : 1 ratio of linear dependence. For a small association or di-
hydrogen and ethylene a t a total pressure of ca. 50 cm. merisation of the catalyst species the rate expression
The low reactivity of trans-stilbene towards hydrogen- approximates to
ation provides evidence for the requirement that the
R = k[A] - k[AI2
substrate as well as molecular hydrogen, be activated
by complexing, i.e., that the vacant co-ordination site where k > k; it is assumed that the undissociated
on the cis-dihydrido-species is essential to the efficiency monomer and dimer are inactive catalytically. Thus a
of the system. In benzene a t 25 and 1 atmosphere of curve of this form shown as a dotted line in Figure 8
hydrogen, trans-stilbene is only very slowly reduced could result. At higher concentrations of catalyst,
(hours). However, if cyclohexene is added, immediate of course, greater divergence from linearity would occur,
rapid hydrogen absorption occurs a t the rate normal unfortunately, the low solubility of the complex pre-
for cyclohexene alone, while the rate of stilbene re- cludes study in this region. Molecular weight studies
duction continues unaffected. The stilbene is clearly should indicate any appreciable dimerisation but we
unable to occupy the vacant site readily for steric have not been able to determine them under the con-
reasons. ditions used catalytically. However, a t higher concen-
(ii) The dependence on pressure is of similar form, trations in benzene dimerisation plainly occurs. The
ViZ., molecular weight studies would in any case not reliably
+
1 - 1 K,[S] 1 1 detect dimerisation to the extent of 5-10% which is
R - kK,[S][A] (6) -k KSI[A] all that would be required. Any dimer present would
be hydrogenated and as we show in Section (C) the dimer
This agrees with the results in Section (B) (b) (ii) with hydride does not transfer hydrogen to olefins.
the values : (iv) The values of AH3 and A S obtained from these
studies may be compared with those quoted by Halpern
for the homogeneous reduction of iron(II1) by hydrogen
View Article Online

using tetrachloropalladate(II), A H : = 20 kcal. mole-l, the intensities of both lines decreased, while a new line
AS1 = -6.8 e.u. and for the reduction of fumaric acid at T -9.69, due to ethane, appeared. No sign of lines
by hydrogen in the presence of ruthenium(I1) were due to an intermediate containing a Rh-C,H, group
AH1 = 18 kcal. mole-l, AS1 = -8 e.u. produced by ethylene insertion into the Rh-H bond
The overall rate expression can hence be formulated was observed.
as Similar direct transfer experiments were carried out in
k' = 1.1 x 1016 e-22,900/RT 1. mole-1 s w - 1 benzene solutions of (A)H, with ethylene and liquid
olefins, e.g., hex-1-ene. The direct transfers are also
(v) The increase in the catalytic activity of indicated by colour changes in the solutions, the yellow
RhX(PPh,),, X = C1 < Br < I, can be due to two (A)H, colour being replaced by the red-brown of (A)
factors. First, the equilibrium constant, K , is de- (or an (A)-olefin complex) as was observed in the catalytic
creased due to the trans-effect increase C1 < Br < I experiments described earlier. Due to the low concen-
[cf. Section (A) (b) (ii)] and hence according to the rate trations of (A) in the latter, it is impossible to demon-
equation the rate of reduction should increase. Secondly, strate a standing concentration of (A)H,by either n.m.r.
the increasing trans-eff ect of the halide could increase or i.r. methods. However, the electronic absorption
the lability of one of the hydrogen atoms in the cis- spectra of these solutions are identical with that of the
dihydrido-species RhX(PPh,),H,(solvent) ; this would dihydrido-species. Similar experiments were made
be most likely if the halide were trans to the hydride as adding hex-1-ene to solutions of the pyridine hydrido-
in (I). However, as noted earlier, we cannot unequivoc- complex, RhCl(PPh,), py H,, and the dimeric hydride
ally confirm this orientation although the trans-position [RhC1(PPh3),H2l2 and in both cases the high-field lines
for X seems reasonable since it leaves the solvent remained unchanged.
occupied site available for olefin competition mutually (b) Lack of exchange with deuterium. Solutions of
cis to both hydrogen atoms. RhCl(PPh,),, in the absence of olefin, were exposed to
The tram-effect argument has to be used with caution. deuterium gas under catalytic conditions, for periods up
For example, the complex Rh(SnCl,)(PPh,), 22 has to 4 days. Samples for mass spectrometric analysis were
catalytic activity well below that of RhCl(PPh,), de- withdrawn periodically. After making a small correc-
spite the prediction based on the known46 large trans- tion for exchange of deuterium with traces of water,
effect of SnCl,-- that the rate would be very fast. Very etc., in the system in absence of catalyst in the solution,
small changes in electronic effects can significantly as determined by separate experiment, it appeared that
alter the thermodynamic entities involved. An active there was no exchange of deuterium with any component
intermediate must have a rather well-defined thermo- of the catalyst system. Experiments were made using
dynamic stability yet must be sufficiently labile to allow pure benzene, benzene-hexane, 1 : 1 benzene-ethanol,
fast transformations to occur. and similar solutions containing up to S~~concentrations
(vi) The increase in rate achieved by use of polar of triphenylphosphine. No H D other than normal
solvents suggests that the rate determining step involves contamination was evident in any case.
the formation of an activated complex which has a In a similar experiment, a small quantity of catalyst
greater dipole moment than the reacting species. That (ca. lop4moles) in solution was exposed to two separate
is, some charge separation is occurring in the formation amounts of deuterium for periods of 4 days: there was
of the intermediate. This may result from either one no evidence from infrared measurement of any incorpor-
of the hydrogen atoms in RhCl(PPh,),H,(solvent) ation of deuterium in the recovered catalyst.
becoming increasingly positive with respect to the other, From these studies it is clear that the v.-hydrogen atom
possibly due to differences in solvation of the halide ions of the phenyl groups in co-ordinated triphenylphosphine
trans to hydrogen. I t is shown in Section (C) that the are not involved in the catalyst system despite the
co-solvent plays no part in the actual hydrogen transfer possibility in crystals of short distances 24?25 between
mechanism. the metal and the a-hydrogen atoms.
(C) Direct Hydrogen Transfer to OleJins; Deuterium (c) H2-D2 exchange: reduction of hex-1-ene by H,-D,
Studies.-(a) Hydrogen transfer. Since the implications mixtures. The kinetic data show that only one catalyst
are clear that the catalytic reactions involve transfer molecule is involved in the transfer of hydrogen to the
of hydrogen to an olefin via a metal cis-dihydrido-species olefin. Provided the H,-D, exchange process on the
of the type characterised by n.m.r., we have studied the catalyst can be minimised, it should be possible to confirm
direct transfer of hydrogen, necessarily using high this since reduction of hex-1-ene with H,-D, mixtures
concentrations of the dihydrido-species, by n.m.r. should only yield dihydrido- and dideutero-paraffin.
methods. (i) H,-D,exchange. Benzene solutions of RhCl(PPh,),
A hydrogen-saturated deuterochloroform solution of were stirred under 1 : 1 H,-D, mixtures (total pressure
(A) in an n.m.r. tube was frozen and asuitable quantity -60 cm.) and gas samples withdrawn for mass spectro-
of ethylene condensed. On allowing the tube to warm metric analysis after 1+ hr. Using 1.25- and 2 - 5 0 - m ~
up slowly in a low temperature probe, both the high- solutions, the increase of HD in the gaseoqs phase was
field lines of (A)H, and the line due to unco-ordinated 46 See F. A. Cotton and G. Wilkinson, " Advanced Inorganic
ethylene were readily observable at -50". At about 0", Chemistry," 2nd edn., Interscience,Wiley, 1966, p. 1029, 1034.
View Article Online
1726 J. Chem. SOC.(A), 1966

0-9 and 3*7y0,respectively. Thus, the rate of exchange can be used to determine the stereochemistry of the re-
as demonstrated by the appearance of HD in the gaseous action using maleic and fumaric acids since the corre-
phase appears to be slow compared with the rate of sponding dideuterosuccinic acids have been character-
reduction under identical conditions. Moreover, the rate i ~ e d . ~ 'From maleic acid, the succinic acid product
of exchange a t zero time seems to have approximately had a strong band at 1192 crn.-l, but only a very weak
a second-order dependence on the catalyst concentration, band a t 1215 cm.-l, confirming that the major com-
suggesting that the exchange process occurs by a ponent was meso-l,2-dideuterosuccinic acid. This
bimolecular collision process. * indicates that predominantly cis-deuteration of the
(ii) Reduction of hex-1-ene by H,-D, mixtures. double bond has occurred. Confirmation was obtained
It is seen from the above Section that in order to minimise from the fumaric acid which had been similarly deuter-
the effect of the inter-molecular exchange process during ated, and where the product had bands in the region
reduction, the standing concentration of the dihydro- 1220-1 170 cm.? corresponding to DL-sym-1,2-dideutero-
or dideutero-species in solution must be kept to a succinic acid.
convenient but small value. Hence, low concentrations Yet further evidence that the addition of hydrogen or
of RhCl(PPh,), and low total gas pressure, but large deuterium by the complex is cis is obtained from the
concentrations of olefin were used. The gas pressure was reduction of hex-2-yne. Although the reduction of this
maintained ca. 15 cm. by carefully adding equal quanti- acetylene is slow a t 1 atmosphere, it is rapid a t 50
ties of hydrogen and deuterium when required. The atmospheres and 20". If the reaction is stopped before
same experiments were repeated a t total pressure of the hex-2-yne is fully reduced and the products ana-
ca. 60 cm. The results are shown in Table 5. The peak lysed by gas-liquid chromatography (g.1.c.) it is found
that not only is n-hexane present but also hex-2-enes.
TABLE5 X two stage reduction of the acetylene is hence involved
Mass spectrometric analysis of n-hexanes obtained by re- in which free olefin is formed as intermediate. Analysis
duction of hex- 1-ene with hydrogen, deuterium and
-
of the hexene fraction shows that the cis-product is
ca. 1 : 1 mixtures of hydrogen and deuterium in >20 times more abundant than the trans-isomer.
benzene a t 25"
mle * Furthermore, since cis-olefins appear to be more rapidly
Mole 1.-1 3 reduced than their trans-isomers, this must be a mini-
hexene Gas Pa,,cm. 86 87 88
mum value for the cis-trans ratio.
0.728 H, 30 t 20 0.8 -
0.728 D, 30 t - 0.9 23.0 (e) Kinetic isotope effect. The rates of uptake of
0.728 H, + D, 60; 32.7 24 35.5 hydrogen and deuterium by cyclohexene using
1.845 H, + D, 15 + 20.5 4-0 18.0 RhCl(PPh,), under identical conditions are plotted as
* Average uncorrected data for two runs. 7 1.82 m M the reciprocals against the reciprocal of the olefin con-
RhCl(PPh,),. $ 1.54 mbr RhCl(PPh,),.
centration in Figure 12. The lines have been drawn
at m/e = 87 was corrected for: (a) the natural HD such that the gradients are in the same ratio as the
contamination in the hydrogen and deuterium used and intercepts on the y-axis. The ratio of the rate of re-
for the natural abundance of 13C and D in the n-hexane duction using hydrogen to that of deuterium approxi-
which result in C,H13D as contaminant, and ( b ) the mates to 0-9. The significance and implications of this
presence of ions such as C6Hl1D,+ resulting from cracking value are referred to in the Discussion section.
in the ionisation chamber. This correction was obtained
from a sample of hex-1-ene which had been catalytic- DISCUSSION
ally deuterated. The mechanism of catalytic hydrogenation of un-
From the data in Table 5 after these corrections have saturated substrates can conveniently be regarded as
been made, a t the average gas pressure 15 cm. the pro- one involving three steps : (1) hydrogen activation, (2)
ducts of the reduction are C6H14, 50%; C,H13D, G.lyo; substrate activation, (3) hydrogen transfer. We shall
C,H,,D,, 43.976. Thus, there is a 93.9% addition of discuss these steps separately since, if effective reduction
hydrogen and deuterium separately which is in agree- of a substrate is to occur, all three conditions must be
ment with the kinetic data showing a first order depend- met. Although many metal complexes have been shown
ence on catalyst concentration. However, a t an average to activate hydrogen molecules, very few have been
pressure of 60 cm. the products are C6H14, 36-6y0; able to reduce olefins or acetylenes either because the
C,H,,D, 23.7% ; C,H,,D,, 39.7% ; this is a considerable latter have not been activated on the metal complex, or
increase on the yield of H-D product expected to result because transfer of activated hydrogen was kinetically
from the H,-D, exchange process. It is probable that or thermodynamically impossible. The interplay of the
the results in (i) are misleading since the appearance of
HD in the gaseous phase may well fail to reflect the * Rapid exchange could occur if both hydrogen and deuterium
molecules were simultaneously activated by one metal atom.
actual concentration of H D in solution in an equilibpium This would involve the formation of four, mutually cis M-H (or
system, where diffusion is the only driving force to over- D) bonds and, since promotion of the metal t o Rh", as well as
possibly an increase in the co-ordination number, would be re-
come surface barrier effects. quired, this scems highly unlikely [cf. Section (A) of the Dis-
(d) Deuteration and stereochemistry of addition to cussion section!.
olefins. In view of the above results deuterium reduction 47 c. R.Childs and K. Bloch, J. Org. Chem., 1961, 26, 1630.
View Article Online

various factors involved is extremely critical for all three Thus, any ligand which withdraws electron density from
steps to be successfully achieved. the metal increases the promotional energy and reduces
(A) The Hydrogen Activation Process.-The ability the ability of the species to activate molecular hydrogen.
of a metal ion or complex to induce fission of molecular The primary interaction of molecular hydrogen in
hydrogen to form one or, as in our present case, two solution with a metal species must, we believe, involve
metal-hydrogen bonds depends on a number of factors the feeding-in of electron density in filled metal orbitals
which cannot readily be divorced from each other. to an acceptor orbital on one hydrogen atom of the
Some of these factors have been discussed especially hydrogen molecule. The tentative suggestion has been
by Halpern and by Nyho1m.l made by Nyholm that an anti-bonding orbital could be
I
mole-' 1.
[C y c lo h Q X on Q]
FIGURE
12 Kinetic isotope in the hydrogenation of cyclohexene in benzene a t 25" with 1.25 mr\i-RhCl(PPh,),.
(0) Hydrogen, ( x ) deuterium
For a d s complex to be oxidised by hydrogen to give used as an acceptor. On the other hand, Halpernl
a d6 species : has suggested that the bonding electrons of hydrogen
H,+ ML, H2ML.j could attack a vacant metal orbital. The former con-
it would appear that 2EAL-H > EH-E + P , where EJI--]8 cept seems most likely, especially since hydrogen acti-
is the energy of formation of the metal-hydrogen bond, vators are in general electron-rich systems. The pre-
EH---Kis the bond strength for molecular hydrogen and cise mode of operation is, however, speculative although
P is the promotional energy of the metal in the com- the hydrogen p orbitals with their greater spacial ex-
plex. Clearly, if the hydrogen activation reaction is to tension could act as acceptors even though they are of
be reversible then 2E51-H must be approximately equal higher energy. The main point is, however, that the
+
to EJl-n P . The electronic and ligational conditions charge separation induced in the H-H bond by such
for a reversible system to be formed are very critical. electron flow will, in the proximity of the metal atom,
For the present system this can be seen by considering lead to bond weakening, lengthening, and synchronous
the species RhCI(PPh,),L. When L is a ligand of low formation of two metal-hydrogen bonds ; initially a t
x-acidic nature, e.g., chloroform, benzene, acetic acid, least, these should be mutually cis but, depending
ethyl acetate, pyridine, or acetonitrile, activation pro- on the nature of the metal species, rearrangement may
ceeds readily. If, however, L is a strong x-acid, viz., occur. Thus, if the configuration of the complex is such
carbon monoxide, tetrafluorethylene, ethylene or per- that no cis site is immediately available, which is the
,
fluorobut-2-yne no activation occurs a t atmospheric case for IrCl(C0)(PPh,) or (Ph,P),RhCl,Rh (PPh,),
pressure (although activation may occur at elevated then some type of reorganisation, e.g., by edge-displace-
pressures it is not necessarily the same type as at 1 ment, and/or dissociation is required ; trans-dihydrides
atmosphere since ligand displacement may occur). could then result depending on the nature of the metal
Since EEr-* remains constant and Ell-= is not signi- and its ligands. Alternatively, in other cases, solvation
ficantly changed * the principal factor is the promotional by polar solvents could remove the more positiveljr
energy, including energies of re-orientation of ligands. charged atom so that heterolytic fission
* The i.r. M-H stretching frequencies in cis-H,IrCl(CO) (PPh,), M + H, = MH + H+
obtained from the activator IrCl(CO)(PPh,), are 2190 and 2100
cm.-l whilst those in, say, cis-H,RhCl(py)(PPh,), are 2143 and occurs. This is doubtless what happens when octa-
2075 cm.-l. hedral species, e.g., py, RhCl,, activate hydrogen ;
View Article Online
1728 J. Chem. SOC.(A), 1966

hydrogen chloride is in fact eliminated with the form- quite long and thus, together with evidence discussed
ation of py, RhC12H.41 earlier, indicates strongly that the olefin attacks the
The failure of RhCl(C0)(PPh,), to activate hydrogen is hydrido-species. The alternative pathway has a
in contrast to the ability of the iridium analogue to further contra-indication since we should not expect an
give reversible hydrogen activation and a cis-dihydride.* olefin complex to be even able to activate molecular
The difference can be ascribed to the greater special hydrogen.
extent of non-bonding d electron density and the lower It seems clear that there must be an intermediate
pomotion energy in the large third transition series species, or at least an activated complex, in which both
atoms and partly to the somewhat increased strength molecular hydrogen (as cis-MH,) and an olefin are co-
of the Ir-H bonds compared to Rh-H. ordinated to the metal atom. The poisoning and
The failure of the square phosphite complex, competitive experiments indicate that a vacant site
RhCl[P(OPh),], to activate hydrogen at 1 atmosphere o n the hydrido-species, at which the substrate olefin can
may be attributed in part to the higher x-acidity of be activated, is necessary for efficient catalysis. Those
triphenylphosphite compared to triphenylph~sphine,~~ t substrates which cannot approach the activating site
which lowers the non-bonding electron density on the because of steric factors are not readily reduced. In
metal. However, the main difference is that the phos- view of the very rapid and stereospecific cis-reduction,
phite complex does not dissociate in solution thereby it is reasonable also to assume that the vacant site is
providing a vacant site. In view of the higher n-acidity cis to the two M-H bonds as in (I).
of P(OPh), a greater tendency to dissociation would I t is now clear why certain species, even though they
have been expected on the grounds of trans-competition. can activate molecular hydrogen quite rapidly and
I t appears, therefore, that the main factor causing dis- reversibly, e.g., IrCl(C0)(PPh,),, RhCl(PPh,), in pyri-
sociation in the triphenylphosphine complex is the greater dine or acetonitrile, are either unable to catalytically
steric hindrance between the triphenylphosphine lig- hydrogenate olefins at 25" and 1 atmosphere or do so
ands ; the steric hindrance between cis-P(OPh), ligands only very slowly. On the dihydrido-species there is no
appears to be small. vacant site for olefin co-ordination-i.e., they are co-
Finally, the unexpected inability of RhCl(PPh,),(S) to ordinatively saturated. The very slow hydrogenation of
catalytically hydrogenate ethylene has an explanation olefins by the iridium species IrCl(CO)(PPh,),H, a t
entirely analogous to that for the effect of carbon mon- elevated temperatures is best attributed to thermal
oxide in RhCl(CO)(PPh,),. It is known from the tram- dissociation of the phosphine ligands $ which would
effect and other studies that ethylene stands high in the then competitively provide a co-ordination site. A less
x-acid and trans-effect sequence and is indeed com- likely explanation would be direct attack of olefins on
parable to carbon monoxide as a x-acid ligand. Thus the hydrido-species.
RhCl(C,H,)(PPh,), cannot be expected to activate (C) The Transfer Ste#.-All the evidence is in agree-
hydrogen a t 1 atmosphere. There is of course no conflict ment with the initial formation of
here with the experiments demonstrating direct transfer RhC1(PPh,) ,H,(solvent) , displacement of solvent com-
to ethylene of hydrogen by RhCl(PPh,),H,; it is merely petitively by olefin, and stereospecific cis-transfer of
that once the ethylene complex is formed, hydrogen bound hydrogen to the olefin in an actual intermediate
activation ceases. or in an activated complex. Diffusion of the saturated
The problem of why propylene and higher olefins substrate away from the transfer site leaves the complex
can be catalytically hydrogenat ed-or , reformulating, again ready to activate dissolved hydrogen molecules.
why such olefins do not inhibit hydrogen activation- As to the precise manner in which the two hydrogen
is clearly due to the enormously reduced complexity atoms are transferred to the olefin, it is our view, though
constants for these olefins, as discussed in Section this is not unequivocally proven, that essentially
(A) (b) (ii). From the rate equation derived in the simultaneous transfer occurs without the formation of an
kinetic section, if K , is small the rate may be large (in alkyl intermediate. We have no evidence so far from
the absence of steric factors), and vice versa. chemical studies that it is possible to attack one hydro-
(B) The Substrate Activation Step.-Whether there is gen atom in RhCl(PPh,),H, without the involvement of
an actual intermediate species, RhCl(PPh,),H, (olefin), the other-either both hydrogens react as in olefin
cannot be readily ascertained. It may be noted, how- hydrogenation, or both hydrogens are evolved as mole-
ever, that the lifetime of the hydrido-species is evidently cular hydrogen. The principal difference between the
present system and the very few cases in complexes
* [Note added in Proof.--P. B. Chock and J. Halpern ( J . Amer. t See also W. Hieber and E. Linder, Ber., 1961, 94, 1417, for
Chenz. Soc., 1966, 88, 3511) have now described the kinetics of
addition of hydrogen to Irans-IrX(C0)(PPh,),. The rate of difference in dissociation constants of HCo(CO),L [L = PPh, and
addition was markedly dependent on the halide X (increasing in P(OPh)31.
the order C1 < Br < I) and also on the solvent used, increasing $ This would predict that the triphenylphosphine exchange
when more polar solvents replaced benzene. These results are rate should be very slow for IrCl(CO)(PPh,),H, [although it is
very similar t o those found in our studies. The very small exceedingly rapid for square IrCl(C0)(PPh,),] and should
deuterium isotope effect also observed ( k ~ * / k=~ *1.22) could increase with temperature paralleling the increased catalytic
mean that bond breaking in the hydrogen molecule was accom- activity ,
panied by a synchronous metal-hydrogen bond formation 4 8 W. D. Horrocks, jun., and R. C. Taylor, Inorg. Chem., 1963,
possibly through a triangular transition state.] 2, 723.
View Article Online

where an unsaturated hydrocarbon (usually ethylene) vibration (vRh-H -2100 cm.-l) in the transition state a t
is known to attack an M-H bond is that in the latter 25" would lead to a value of (rate H)/(rate D) -4, the
there is only one M-H bond and the attack of olefin is zero-point energy difference of Rh-H and Rh-D being
best regarded as an insertion reaction into the M-H ca. 0.86 kcal./mole. However, it has been proposed
bond. I n view of the appreciable stability of the that in certain hydride transfer reactions in organic
rhodium-ethyl bond (detected by n.m.r.) formed by systems, e.g., those in the Cannizaro reaction,51the alka-
insertion of ethylene into an Rh-H bond in the dimeris- line cleavage of triphenyl~ilane,~, and the hydrolysis of
ation process for ethylene studied in great detail by , ~ ~ low (rate H)/(rate D)
pyridine d i p h e n y l b ~ r a n e where
Cramer 35 and of the high stability of C2H5RhC12(PPh3), values are found, a complete rupture of the X-H bond
a t 25" in solution,45we should have expected our n.m.r. does not occur; the bond is bent so that the hydrogen
direct transfer studies to ethylene to have shown the atom can attach itself to another part of the molecule
presence of an ethyl intermediate. None could be leading to a three-centre transition state in which the
observed. lengthening of the X-H bond occurs synchronouslv
It might have been thought possible to eliminate an with the formation of, say, a C-H bond. This concept
alkyl stage by stereochemical argument, since if an can be applied in the present case for a Rh-H bond and
alkyl intermediate had a lifetime long compared t o the such a mechanism would significantly reduce the (rate
time of free rotation about the M-C bond, both cis- H)/(rate D) value; indeed, the inverse size of this effect
and trans-hydrogen addition should occur equally. indicates that zero-point energy is actually gained in the
However, inspection of the stereochemistry shows that transition state. Thus a " tightening u p " may occur
the insertion of an olefin such as maleic acid into an in the activated complex whereby zero-point energy
M-H bond probably determines itself the overall lost by the bending and lengthening of the Rh-H bond
stereochemistry of the addition process. This is very is more than compensated for by the zero-point energy
obviaus when acetylenes are considered. Thus when gained b y the formation of a C-H bond since the zero-
RC-CR' co-ordinates or occupies the activation site, point energy of a C-H bond is greater than that of the
the groups R,R' will be bent slightly away from the Rh-H bond by -1.43 kcal./mole. Moreover, if the olefin
metal as in (IVa). Transfer of one hydride to give a complex occupies a position cis to both Rh-H bonds
it is possible that by suitable alignment of the double
H bond about this position, both hydrides may be trans-
ferred simultaneously, each by a three-centre transition
state as in Scheme 2. It seems quite likely that all
metal-carbon bond will thus cause R and R to become

I
cis to each other as in (IVb). Although we cannot
determine which hydrogen atom is transferred first, i t is
1 H-C-
likely that it would be the one trans to triphenylphos-

phine ; in the iridium hydrido-complex, IrH,Cl(PEt,Ph),,
the hydrogen trans to phosphorus is reactive towards I
dilute mineral acids whereas the hydrogen trans to chloride Scheme 2
is not.49 The transfer of the second hydride will then
retain the cis-configuration of the product. insertion mechanisms of this type proceed via such
A deeper insight into the mechanism of the transfer triangular transition states. It is pertinent to note in
process is provided by the very low value for the overall this connection that appreciable interaction between a
kinetic isotope effect [Section (C)] where the ratio of the transition metal and C-H bonds of a bound ligand has
rates (H),/(D)is about 0.9. It is well knownm that the been clearly demonstrated in several cases 54 and indeed
kinetic isotope effect results from changes in the activ- in one case,55 that of Ru(Me,PCH,CH,PMe,),, there is
ation energy of the process caused by differences in the evidently an equilibrium between two forms, one with
zero-point energy when the reactants are converted to an co-ordinated phosphine and the other with a Ru-H
activated complex. Complete loss of a Rh-H stretching plus a Ru-CH, bond. Furthermore, there is a close
48 J . Chatt, R. S. Coffey, and B. L. Shaw, J . Chem. SOC.,
1366, 53 M. F. Hawthorne and E. S. Lewis, J . Amer. Chem. SOL,
7393. 1958, 80, 4296.
F. H. Westheimer, Chenz. Rev.,1961, 61, 265. 54 M. L. H. Green, L. Pratt, and G. Wilkinson, J . Chem. SOC.,
5 1 K. B. Wiberg, J . Amer. Chem. SOC.,1954, 76, 5371. 1960, 989, 3753; R. B. King, J . Amer. Chem. Soc., 1963, 85,
52 S. L. Kaplan and K. E. Wilzbach, J . Amer. Chem. SOC., 1922.
1955, 77, 1300. 55 J. Chatt and J. M. Davidson, J . Chena. SOC.,1965, 843.
5u
View Article Online
1730 J. Chem. SOC.(A), 1966

analogy between the transition state postulated here 4.4%). On continued heating under reflux, gradual
and that proposed in hydrogenations using di-imide.S6 conversion of these orange crystals to the deep red form
occurs.
EXPERIMENTAL The infrared spectra of the two forms of the compound
from 4000-600 cm.-1 showed no differences. Their
Microanalyses, molecular weight determinations (Mechro- chemical properties appear to be identical, and, in particular,
lab Osmometer, 37") and microhydrogenations are by the
there is no difference in catalytic behaviour when dissolved.

Microanalytical Laboratory, Imperial College.
N.m.r. spectra were obtained using a Varian Associates TABLE6
model 43100 spectrometer at 56.45 and 40 Mc./sec., and e.s.r.
Molecular weight data for RhCl(PPh,), ( M , required
spectra on a Varian Associates V-4502-15 spectrometer.
924.9). Mechrolab osmometer at 37"
Melting points (uncorrected) were determined on a
Kofler hot-stage microscope. G.1.c. measurements were Concentration Concentration
Solvent (g./la) M Solvent (g./l.) M
made using a Perkin-Elmer model F-11 chromatograph CHC1, ... 5.654 471 C6H6* ... 2.783 480
with flame ionisation detector and 2 m. columns. The 13.856 440 4.025 473
stationary phase was squalene for low boiling hydrocarbons 18.044 462 4.550 492
or silver nitrate-diethyleneglycol where better resolution 23.534 461
was required ; calibrations were made using standard * Above 4-6 g./l. variable results are obtained but these are
samples. Infrared spectra were taken on a Grubb-Parsons generally above 600. Cryoscopic values in benzene : 2-575g./l.,
M , 458; 2.321 g./l.) M , 457.
grating " Spectromaster " and electronic spectra on a
Perkin-Elmer model 350 spectrometer. Mass spectra were The preparation of the complex can also be carried out
obtained using a n A.E.I. MS-9 spectrometer. in aqueous acetone. Thus, RhCl3,3H,O(O.1 g.) in water
Materials.-The rhodium chloride, RhC1,,3H20 was (5 ml.) is added to a hot solution of PPh, (0.6 g.) in acetone
from Johnson Matthey Ltd. ; the only transition-metal (25 ml.) and heated under reflux. Orange crystals of the
impurity is believed to be possibly a trace of iridium.57 complex are deposited after a few minutes (Found: C,
Triphenylphosphine (Albright and Wilson Ltd .) was re- 69.6; H, 4.6%).
crystallised from benzene-ethanol before use. The excess of PPh, used in the preparations can be re-
Hydrogen (British Oxygen Co.) was passed through an covered by the addition of water to the filtrates until
Engelhard " Deoxo catalyst tube and through a silica
"
precipitation begins ; on standing 2-3 days, PPh, crystal-
gel drying tube before admission to the vacuum system. lises out. Recrystallisation from ethanol and ethanol-
Cylinder deuterium gas (Matheson Lecture bottle, 99.5%) benzene gives pure material.
was used without treatment; after equilibration in the (b) Tris(triphenylphosphine)bronzorhodium(I).To a solu-
system the gas was analysed mass spectrometrically for tion of PPh, (12 g., 6M excess) in hot ethanol (250 ml.)
each run. was added a solution of RhC1,,3H20 (2 g.) in warm ethanol
Ethylene, cis- and trans-but-2-ene (Matheson Lecture (70 ml.). After heating under reflux until the solution
bottle), and propylene (I.C.I.) were used without purific- begins to lighten in colour (5 min.), LiBr (8 g.) in hot
ation. Hex-l-ene, hept-l-ene, and cyclohexene (Koch- ethanol (50 ml.) was added and the mixture heated under
Light) were freed from peroxides by acid ferrous sulphate reflux for 1 hr. Orange prisms of the complex were collected
washes, dried over CaCl,, and were stored under nitrogen from the warm solution, washed with de-gassed ether (50
and protected from light. They were freshly distilled ml.) and dried in vacuo. Yield 5.1 g. (64%), m. p., 133-
in nitrogen or in vacuo for each run, directly into a burette 134" (Found: C, 66.8; H, 4.9; Br, 8.2. C,,H,,BrP,Rh
attached to the reaction vessel. Purity was checked by requires C, 66.9; H, 4.7; Br, 8.2%).
g.1.c. Hex-l-yne (Koch-Light) was passed through an A second form of the bromide can be prepared as follows.
activated alumina column under nitrogen and immediately To a solution of RhC1,,3H20 (0.5 g.) in warm ethanol
distilled into the burette. (75 ml.) was added LiBr (2 g.). After heating under reflux
All other solvents were of reagent-grade quality and were (5 min.), hot ethanol (76 ml.) containing PPh, (3 g.) was
dried and de-gassed before use. All experiments were added and the mixture heated under reflux for 30 min.
normally carried out under nitrogen, hydrogen or in vacz~o. Deep brown crystals of the complex were collected from the
(A) Preparations and Reactions.-(a) Tris(tripheny1- warm solution, washed with de-gassed ether (25 ml.) and
~ l ~ o s p h i n e ) c l z l o r o r l o d i u ~ (To
I ) . a solution of freshly re- dried in vacuo. Yield, 1.3 g. (69%), m. p. 129-130"
crystallised PPh, (12.0 g. 6RX excess), in hot ethanol (350 (Found: C, 66.7; H, 4.7; Br, 8.3%).
ml.), was added a solution of rhodium trichloride tri- Again, no differences were found in the i.r. spectra of
hydrate (2 g . ) in hot ethanol (70 ml.) and the solution was the two forms from 4000-600 cm.-1 and their chemical
heated under reflux (30 min.). The hot solution was filtered properties are identical.
and the burgundy-red crystals of the complex were washed (c) Tris(triphenyZ~hosphine)iodorhodium(I).To a solu-
with degassed ether (50 ml.) and dried in vucuo. Yield tion of PPh, (6.0 g., 6M excess freshly recrystallised from
6.25 g. (88% based on Rh), m. p. 157-158" [Found: C, ethanol) in hot ethanol (200 ml.) was added a solution of
70.1; H, 4.9; C1, 3.9; 0, 0.0. C,4H4,C1P,Rh requires C, RhC13,3H,0 (1.0 9.) in warm ethanol (60 ml.) and the mix-
70.1; H, 4.9; C1, 3.5; 0, O-Oo/,]. The molecular weight ture brought to the boil. When lightening of the colour
data is given in Table 6. of the solution occurred and orange crystals began to be
If more concentrated solutions are used, i.e., using a deposited (ca. 5-10 min.), a solution of I i I (4 g.) in hot
total of 200 ml. or less of ethanol, orange crystals of the ethanol (80 nil.) was added quickly dropwise, and the
complex are obtained after 5 min. heating under reflux; 56 S. Hiinig, H. R. Miiller, and W. Thier, Angew. Chem.
these often mixed with a small quantity of the above- Internat. Edn., 1965, 4, 271.
mentioned red complex (Found: C, 69.7; H, 4.9; C1, 67 Dr. F. M. Lever, personal communication.
View Article Online

solution stirred. The mixture heated under reflux for washed with a little ethylene-saturated chloroform and dried
3-4 hr. and brown crystals collected from the warm in air. Yield, ca. 70% (Found: C, 65.5; H, 5.0; C1, 4.9.
solution. These were often contaminated with a material C,8H,4C1P2Rhrequires C, 66.0; H, 4.8; C1, 5.1%).
of higher iodine content, probably the dimeric species The corresponding bromide complex is made in an identical
[RhI(PPh,),],. The product was therefore dissolved in the way from RhBr(PPh,), (Found: C, 62.3; H, 4.7; Br,
minimum quantity of hot benzene containing an excess of 10-5. Ca8H3,BrP,Rh requires C, 62-0; H, 4.6; Br, 10.9yo).
PPh,, the solution filtered and allowed to stand for several The complexes are readily soluble in benzene, chloroform
hours under nitrogen. Deep red prisms of the complex and dichloromethane. The solutions lose ethylene rapidly
were collected, washed with ether and dried in vacuo. but regeneration is complete in the presence of an excess
m. p. 118-120" (decornp.) (Found: C, 64.1; H, 4.8; of ethylene.
I, 12-3. C,4H4,1P,Rh requires C, 63.8; H, 4.4; I, 12.5%). (h) Absorption of hydrogen by RhCl(PPh,), a d
(d) Tetrakis (triphenylphosphine)-t.L.~'-dichlorodirhodizsm(I). [RhCl(PPh,)&. In a standard micro-hydrogenation ap-
Saturated solutions or suspensions of RhCl(PPh,), in benz- paratus, 5 ml. of solvent (benzene, ethyl acetate, glacial
ene, toluene or preferably methyl ethyl ketone when acetic acid) were saturated with hydrogen at 1 atmosphere
heated under reflux under nitrogen, give salmon-pink pressure. Then RhCl(PPh,), (ca. 40-45 mg.) were tipped
crystals of the complex in essentially quantitative yield. into the solvent from a bucket and the additional hydrogen
The collected material was washed with de-gassed ethanol uptake measured. The rate of uptake varied with each
and dried in vacuo (Found: C, 65.4; H, 4.8; C1, 5.3. solvent, but no useful data could be obtained because of
C,,H6,C1,P4Kh2 requires C , 65.1; H, 4.5; C1, 5.4%). the differing solubilities of the complex and the dihydride
The complex slowly absorbs oxygen from the air in the product in each of the solvents. Generally, uptake was
solid state but more rapidly as a suspension in organic complete after 1 hr. I n all cases, the absorption was
solvents in which it is sparingly soluble. Storage should be found to correspond t o 1 mole of hydrogen per mole of
in uacuo. The molecular weight (in chloroform, saturated RhCl(PPh,),. The dimer (37-45 mg.) in dichloromethane
with hydrogen) is that for a dimer: M (found), 1263, absorbed 1-30ml. hydrogen a t s.t.p. ( E 2-05moles hydrogen
(calc.), 1324. If molecular weight measurements are carried per mole).
out under aerobic conditions, low values are obtained, (i) Bis(triphenylphosphine)dihydridocizlororhodiu~(I11). On
presumably because of the formation of the soluble oxy- passing hydrogen through a solution of RhCl(PPh,), (0-5
geno-complex, RhCl(PPh,) ,O,. The dimer in suspension g.) in de-gassed chloroform (10 ml.) for a few minutes, the
undergoes cleavage with carbon monoxide at 25", to give solution turned pale yellow. On concentrating the solution
RhClCO(PPh,),, identified by comparison of its i.r. spectrum under hydrogen to a small volume (-2 ml.) or by the ad-
with an authentic sample. Crystallisation from benzene dition of a little de-gassed ether, pale yellow crystals of the
solution containing an excess of triphenylphosphine gives complex were obtained. These were quickly collected,
R hC1(PPh,) ., washed with a little hydrogen-saturated ether and dried
(e) Intevaction of the complexes with carbon monoxide. i n vaczto (Found: C, 65.2; H, 5.1; C1, 6-2. C3,H,,C1P,Rh
When carbon monoxide is passed through a solution of requires C, 65.0; H, 4-8; C1, 5.3%). The molecular weight
RhCl(PPh,), (0.05 g.) in benzene (10 ml.) the solution in CHCl, under hydrogen is 700 (required 665).
rapidly turns yellow and after removal of solvent, crystals The yellow crystals, even after storing in vucuo for several
of RhClCO (PPh,), are obtained in essentially quantitative days, smell faintly of chloroform and the high chloride
yield. The product was identical (i.r.) with the authentic analysis reflects the retention of some weakly bonded
material prepared from [Rh(CO),Cl], (Found: C, 64.3; chloroform. [RhCl(PPh,),(H,),CHCl, requires C1, 18.1yo].
H, 4.3; C1, 5-3. Calc. for C,,H,,ClOP,Rh, C, 64.4; H, (j) D i[bis(triphenyl~hosphine)dihydrochlor~h odium(~~~)]di-
4.4 ; C1, 5-1"/) . chloromethane. This was prepared in the same manner as
Identical behaviour was found for the corresponding the possibly five co-ordinate species above except using-
bromide and iodide complexes and the products were dichloromethane as solvent. Yellow crystals of the com-
identified by their i.r. spectra. plex were obtained on concentrating the solution under
(f) Interactions of the complexes with aldehydes. A suspen- hydrogen; they were filtered, washed with a little de-
sion of RhCl(PPh,), (0.5 g.) in benzene (10 d.) was warmed gassed ether and dried in vacuo for several days (Found:
(40") and n-heptaldehyde added (0.5 ml.). The solution C, 61.8; H, 5.0; C1, 9-5. CI3H,,Cl4P,Rh2 requires C, 62.0;
rapidly turned pale yellow, and on cooling, crystals of H, 4.7; C1, 10.0%).
RhClCO(PPh,), were deposited from the solution (identified The i.r. spectrum (mull) shows bands due to the presence
by i.r.). Analysis of the residual solution by g.1.c. shows the of dichloromethane. The corresponding dideutero-complex
following yields of products : n-hexane, 80% ; hex-l-ene, was made in an analogous way.
10% ; cis-hex-2-ene, 5% ; trans-hex-2-ene, 5%. (k) Bis (tripheny1phosphine)pyridinedihydridochlororho-
Similar behaviour was found for propionaldehyde, form- dium(m). On passing hydrogen through a solution of
aldehyde, benzaldehyde, cinnamic aldehyde and formic W l ( P P h , ) , (0.25 g.) in pyridine (10 ml.) for 20 min.;
acid. the solution gradually turned pale yellow. On concentra-
The corresponding bromide and iodide were found to tion under hydrogen to ca. 2 ml., crystals of the complex
behave similarly, RhBrCO(PPh,), and RhICO(PPh,), were obtained. These appeared to be yellow but after
being identified as the products. washing with ether (5 ml.) were white. Yield ca. 90%
(g) Bis(tripJzeny1phosphine)etJzylenechlororhodium(1). On (Found: C, 66.6; H, 5.1; C1, 4-8. C,lH,,CINP,Rh requires
bubbling ethylene through a solution of RhCl(PPh,), C, 66.2; H, 5.0; C1, 4.8%).
(0.1 8.) in de-gassed chloroform (10 ml.), the red colour The complex is stable in the solid state in absence of air
rapidly changes to yellow and on concentration of the but in solution, hydrogen is lost slowly; re-saturation with
solution under an ethylene atmosphere, yellow crystals of hydrogen restores the complex.
the complex were obtained. They were collected rapidly, The corresponding bromide can be obtained in an
View Article Online
1732 J. Chem. SOC.(A), 1966

analogous way from RhBr(PPh,), (Found: C, 62.4; H, 4.7; hydrogen starvation or diffusion control. In kinetic runs
Br, 10.3. C,,H,,BrNP,Rh requires C, 62.4; H, 4-7; Br, the initial uptake rates were customarily well below 90 ml./
10.2Y0). min. so that there is little likelihood of hydrogen starvation
As well as Rh-H frequencies i.r. spectra of both com- in solution.
pounds show bands due to both PPh, and pyridine. The catalyst moles, in general) was weighed into a
(1) Other solvent hydrido-complexes. Repetition of the small glass bucket, suspended from a side-arm, which on
preceding experiments using benzene, ethyl acetate, glacial rotation dropped the bucket into the solution.
acetic acid or dimethylformamide gave pale yellow solutions, The standard solvent in kinetic runs was benzene (50
from which almost white crystals were obtained directly ml.) plus petroleum [(30-n) ml.] (hexane fraction, Hopkin
or by precipitation upon concentration of the solution in a and Williams " Spectrosol for U.V. spectroscopy), where n
"
stream of hydrogen. However, the solids were rather is the volume of substrate used. The total volume was
unstable, losing hydrogen. In addition to Rh-H bands hence constant throughout the series of runs; this procedure
(Table 2), i.r. spectra of the ethyl acetate and dimethyl- was adopted to minimise any effect of dielectric constant
formamide products had frequencies assignable to co- changes.
ordinated solvent a t 1734m and 1639m cm.-', respectively. The solvent was de-gassed by careful evacuation and the
(B) N.m.r. Studies.-(a) Ethylene complex. Deutero- system flushed three times with hydrogen. The catalyst
chloroform (0.5 ml.) containing RhC1(PPhJ2(C2H,)(0.05 g.) was added and the solution stirred until the solid had
and one drop of SiMe, in an n.m.r. tube was attached to a dissolved giving a pale yellow solution. The stirring was
vacuum line, frozen in liquid nitrogen, evacuated, and an then stopped and freshly distilled substrate run in from the
appropriate measured amount of ethylene condensed into burette; the colour of the solution became deep brown
it. After sealing, the tube was placed in a Varian low but absorption of hydrogen did not occur until stirring had
temperature probe for the 40 Mc./sec. spectrometer and the begun. The hydrogen was evacuated, the system allowed
spectrum measured over the range -50-25". Repetition to equilibrate (2-3 min.) and the vapour pressure measured
with varying quantities of ethylene led to the results given (the values were reproducible to ca. 0.3 cm.). Hydrogen
in Table 1. was then re-admitted to a total pressure of ca. 60 cm.;
(b) Hydrogen pressure studies. A deuterochloroform this pressure was adjusted so that the critical readings at
(0-5 ml.) solution of RhCl(PPh,), (0.1 g.) containing one the standard 50 cm. partial pressure of hydrogen occurred
drop of SiMe, was saturated with hydrogen and transferred a t least 2-3 min. after initiation of the reaction by starting
to a thick-walled n.m.r. tube. The latter was sealed in to a the stirrer; this period was sufficient to allow the setting
brass adaptor by Ciba Araldite epoxy resin (typeAY/HY 111). up of a steady state. During this reaction time only ca.
The adaptor was connected to stainless steel pressure tubing (1% of the substrate was consumed. IVhere small
(1 mm. outside diam.) terminating in a needle valve and quantities of substrate were used, the total volume of the
connection to pressure gauge and cylinder. The system system was reduced to ca. 0.5 1. by isolation of the storage
was flushed with hydrogen and thereafter the pressure bulb, so that smaller amounts of substrate are consumed
slowly increased to ca. 100 atmospheres. After a few for a larger pressure change.
minutes to allow equilibration (and for safety reasons) the (D) Deuteration Studies.-(a) Reduction of iizaleic atid
tube arrangement was disconnected from the cylinder and fumaric acids. To 1 : 1 benzene-ethanol (50 nil.) containing
the n.m.r. tube inserted in the spectrometer. RhCl(PPh,), (2 x lop4 moles) was added the acid (0-0424
(c) Hydrogen transfer to olefins. A solution of RhCl(PPh,), moles). Deuterium (600 mi.) was admitted and after
(ca. 0.1 g.) in CDCl, (3.0 ml.) in an n.m.r. tube sealed to a complete absorption (ca. 20 hr.) the solution was filtered,
vacuum line, was saturated with hydrogen until yellow ; concentrated in 'uacuo and the crystals of succinic acid so
after addition of a little SiMe,, ethylene was condensed obtained collected and crystallised from acetonitrile :
onto it. After sealing, the n.m.r. spectrum was followed the i.r. spectra were measured in Nujol mulls.
from - 70"-25". (b) Hydro-deuteration of hex-l-ene. After exposure to
In similar experiments, liquid olefins such as hex-l-ene the mixed gases for 60 min. the solution was fractionally
were injected into the hydrido-complex solution with the distilled and the fraction between 64-70' collected. This
tube in the probe. fraction, enriched in n-hexane, was analysed directly
(C) Catalytic Experiments.-The reaction vessel was a mass spectrometrically. The data is collected in Table 5 .
250 ml. flask with an outer jacket which was flattened on
one side to accommodate a magnetic stirrer. It was We acknowledge the award of an 1851 Exhibition Science
attached via a condenser to a vacuum system with storage Scholarship (J. F. Y.), and Inorganic Research Student-
bulbs, manometers, sampling tubes, etc. ship (J. A. 0. and F. H. J.). We thank Albright and
Water from a thermostat bath was rapidly pumped (ca. Wilson Ltd. for gifts of triphenylphosphine. We are also
2 l./min.) through the jacket and the temperature of hydro- indebted to Drs. D. F. Evans, J. F. Gibson, ,4.J. Po&,
genating solutions could be maintained to &0.25', as checked and L. Pratt for valuable discussions and assistance.
by the direct measurement of temperature of these solutions. Finally, we thank Dr. L. B. Hunt, Dr. F. &I. Lever and
The stirrer was a Teflon coated magnet fitted with two Johnson-Matthey Ltd. for their most generous loans of
Teflon flange discs and was driven by the external motor so the rhodium chloride, without which this work could not
that it operated a t the gas-liquid interface. This gives very have been done.
efficient stirring in of hydrogen, and hydrogen absorption INORGANIC
CHEMISTRYRESEARCH LABORATORIES,
rates of up to ca. 250 ml./min. have been observed with IMPERIAL
COLLEGE,
RhI(PPh,), without any apparent limitation due to LONDONS.W.7. [6/335 Received, Mavclz 15th. 19661

Inorg. Phys. Theor.: Ca. 1 Atmosphere of Hydrogen Pressure and

Uploaded by

Copyright:

Available Formats

Inorg. Phys. Theor.: Ca. 1 Atmosphere of Hydrogen Pressure and

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Inorg. Phys. Theor.: Ca. 1 Atmosphere of Hydrogen Pressure and

Uploaded by

Copyright:

Available Formats

View Article Online / Journal Homepage / Table of Contents for this issue

Inorg. Phys. Theor. 1711

The Preparation and Properties of Tris(triphenylphosphine)halogeno-

Tris(triphenylphosphine)chlororhodium(l), RhCl(PPh3)3,has been prepared by the interaction of an excess of

1712 J. Chem. SOC. (A), 1966

Inorg. Phys. Theor. 1713

23 Th. Kruck and W. Lang, Angew. Chem. Internat. Edn., 1965,

1714 J. Chem. SOC.(A), 1966

Inorg. Phys. Theor. 1715

1716 J. Chem. SOC.(A), 1966

been stated39 to be sufficiently soluble as to yield

Inorg. Phys. Theor. 1717

Rh-H bonds. t ion.

however). with lo3Rhto give a doublet (splitting 20 c./sec.), which is

1718 J. Chem. SOC. (A), 1966

concentration of the paramagnetic impurity discussed the tris(tripheny1phosphine)halogenorhodium( I) com-

Inorg. Phys. Theor. 1719

1720 J. Chem. SOC. (A), 1966

Inorg. Phys. Theor. 1721

0.5 I. 0 1.5 2-0

1722 J. Chem. SOC.(A), 1966

gives relative rates under otherwise identical conditions

0-5 1.0 1.5 2.0

Inorg. Phys. Theor. 1723

J. Chem. SOC. (A), 1966

Inorg. Phys. Theor. 1725

1726 J. Chem. SOC.(A), 1966

Inorg. Phys. Theor. 1727

1728 J. Chem. SOC.(A), 1966

Inorg. Phys. Theor. 1729

metal-carbon bond will thus cause R and R to become

likely that it would be the one trans to triphenylphos-

1730 J. Chem. SOC.(A), 1966

there is no difference in catalytic behaviour when dissolved.

Inorg. Phys. Theor. 1731

1732 J. Chem. SOC.(A), 1966

You might also like