beta.-Hydride Elimination and Regioselectivity in the Rhodium

Kin Wah Mak and Kin Shing Chan. Journal of the American Chemical Society ... James L. Leighton and Deirdre N. O'Neil. Journal of the American Chemical...
0 downloads 0 Views 589KB Size
Organometallics 1995, 14, 4644-4650

4644

/%HydrideElimination and Regioselectivity in the Rhodium-CatalyzedHydroformylation of Open Chain Unsaturated Ethers Raffaello Lazzaroni,*$tRoberta Settambolo,*and Gloria Uccello-Barretta' Centro di Studio del C N R per le Macromolecole Stereordinate ed Otticamente Attive, Dipartimento di Chimica e Chimica Industriale, via Risorgimento 35, Universita di Pisa, 56126 Pisa, Italy, and Istituto di Chimica Quantistica ed Energetica Molecolare del CNR, via Risorgimento 35, Universita di Pisa, 56126 Pisa, Italy Received February 2, 1995@ The influence of the temperature on the regioselectivity i n the rhodium-catalyzed hydroformylation of ethyl vinyl ether and ethyl allyl ether has been investigated in the range 20-120 "C. For both substrates, the amount of branched aldehyde decreases as the temperature increases, this effect being more pronounced for ethyl allyl ether than for ethyl vinyl ether. The 2H NMR investigation of the deuterioformylation products obtained a t partial conversion, with particular attention to the unconverted substrates, points out the different behavior of the metal alkyl intermediates: at room temperature their formation is irreversible, at high temperature it becomes reversible for both isomers in the case of ethyl vinyl ether, but only for the branched one in the case of ethyl allyl ether. These results fully account for the observed influence of the temperature as well as the substrate structure on the regioselectivity, Scheme 1

Introduction The influence of the reaction parameters and substrate structure on the regioselectivity of the hydroformylation of open chain oxygenated unsaturated compounds in the presence of rhodium-based catalytic precursors has been previously investigated.lI2 Nevertheless, no complete and satisfactory explanation of the observed phenomena has yet been provided. We are interested in the mechanistic aspects of this process and, in particular, the explanation of the observed' influence of the temperature on the distribution of the isomeric aldehydes (regioselectivity) in the Rhr(CO)~z-catalyzed hydroformylation of two simple oxygenated olefins, i.e., ethyl vinyl ether (1)and ethyl allyl ether (21, containing the heteroatom in a or P position with respect to the vinyl group (Scheme 1, Figure 1). In accord with the generally accepted mechanism of the rhodium-catalyzed hydroformylation, linear and branched alkyl-metal c o m p l e x e ~ , arising ~ - ~ from the insertion of olefin in the Rh-H bond, are assumed to be important intermediates (Scheme 2). The reverse process, i.e., the dissociation of the metal-alkyls via P-hydride elimination, can play an important role in the regioselectivity of the reaction, especially if it occurs with different rates for the two isomeric rhodium alkyls. As already shown in our * To whom correspondence should be addressed. Dipartimento di Chimica e Chimica Industriale. Istituto di Chimica Quantistica ed Energetica Molecolare del CNR. Abstract published in Advance ACS Abstracts, August 1, 1995. (1)Lazzaroni, R.; Bertozzi, S.; Pocai, P.; Troiani, F.; Salvadori, P. J . Organomet. Chem. 1986,295,371. (2) (a) Amer, I.; Alper, H. J. A m . Chem. SOC. 1990,112, 3674. (b) Doyle, M.P.; Shanklin, M. S.; Zlokazov, M. V. Synlett. 1994,615. (3) Evans, J.; Schwartz, J.;Urquhart, P. W. J. Organomet. Chem. 1976,81, C37. (4) (a) Bennet, M. A,; Crisp, G . Organometallics 1986,5,1792. (b) Ibid. 1986,5, 1800. (5) (a) Ungvary, F.; Markb, L. Organometallics 1982,1, 1125. (b) Ibid. 1986,5,2341. (6) Pino, P. J. Organomet. Chem. 1980,200,223. +

COM2

/=:-

Et0

1

Rhq(C0),2

r_/CHO

Et0

+

Et0 5

4

L

EtOMCH0

+

E

1

0

7

CHO 6

7

previous studies on simple olefins, such as 1-hexene7 and deuterioformylation experiments can be very useful in order to gain a deeper insight into this aspect.6J0J1 For this reason, we have studied the deuterioformylation of both 1 and 2 (Scheme 3) at two different temperatures (20 and 100 O,C) and partial substrate conversion: the crude reaction mixtures obtained were directly investigated by 2HNMR spectroscopy, as previously done in the case of styrene,12in

4

@

(7) Lazzaroni, R.;Uccello-Barretta,G . ;Benetti, M. Organometallics 1989,8,2323. (8)Lazzaroni, R.; Settambolo, R.; Raffaelli, A,; Pucci, S.; Vitulli, G. J . Organomet. Chem. 1988,339,357. (9) Raffaelli, A,; Pucci, S.; Settambolo, R.; Uccello-Barretta, G.; Lazzaroni, R. Organometallics 1991,10,3892. (10) Considio. G.:Morandini, F.: Haele. - R.: Pino, P. J. Mol. Catal. 1990,60,365. ( l l ) ( a ) Bianchi, M.; Piacenti, F.; Frediani, P.; Matteoli, U. J. Organomet. Chem. 1977,137,361. (b) von Bezard, D. A.; Consiglio, G.; Morandini, F.; Pino, P. J.Mol. Catal. 1980,7, 431.

0276-733319512314-4644$09.00/00 1995 American Chemical Society

Organometallics, Vol. 14, No. 10,1995 4645

Rh-Catalyzed Hydroformylation of Unsaturated Ethers

Scheme 2

6

L CHO 7

CHO

the presence of the oxygen atom directly bonded t o the vinyl group eliminates the possibility of isomerization t o internal olefin during the B-hydrideelimination, and only two isomeric aldehydes (4 and 5) are obtained as products of the reaction (Scheme 2). By contrast, the allyl ether can isomerize to internal olefin l-ethoxypropene (3)(Scheme 2). This process complicates the investigation of the influence of the temperature on the regioselectivity in the case of 2 because the aldehyde 8 (Scheme 2) could form in the reaction mixture. To avoid this problem we carried out our experiments at partial conversion of the substrate (30%)) when an excess of terminal olefin was still present. Under these conditions no formation of aldehyde 8 was observed.

NORMAL

ALOEHYDE 0

0

0

Results and Discussion 20

40

60

80

100

120

Tp.,

Figure 1. Hydroformylation (total pressure 100 atm, CO: Hz 1:l) of the ethyl vinyl ether (0)and ethyl allyl ether (0)at partial conversion. Influence of the reaction temperature on the regioselectivity.

order to carefully observe the position of the deuterium atoms in the unconverted olefins. It is noteworthy that, in the case of the vinyl ether,

Hydroformylation Experiments. The hydroformylation of 2 was carried out with Rh4(C0)12 as catalytic precursor in the temperature range 20-120 "C and at a total pressure of 100 atm (CO:Hz,l:l). Reactions were stopped after approximately the same drop of gas pressure, corresponding to a 30% substrate conversion (12)Uccello-Barretta, G.; Lazzaroni, R.; Settambolo, R.; Salvadori,

P.J. Organomet. Chem. 1991,417,111.

4646 Organometallics, Vol. 14, No. 10, 1995

Lazzaroni et al.

Scheme 3

F

Et0

1

COD2

-

[cat]

D

CDO

1,3+6

l.?(l)d2-7

into aldehydes. In this way correct values for the regioselectivity of the terminal double bond alone were obtained. The composition of reaction mixture was determined by GC analysis. Conversion of the oxygenated substrate to aldehydes was evaluated by GC analysis using toluene as internal standard. At 20 "C no isomerization of the substrate was observed: only the aldehydes 6 and 7 (Scheme 2)were formed; hence, in these conditions, the reaction is completely chemoselective. By contrast, a t higher temperature the substrate does isomerize t o some extent. In particular, a t 100 "C, the chemoselectivity ((6+7)/3)- is 57/43. As shown in Figure 1, the normal aldehyde content (and thus the regioselectivity) is particularly affected by the increase of the temperature. This value changes from 30% at 20 "C to 59% a t 100 "C, the regioselectivity being in favor of the branched isomer (7/6 = 70/30) at room temperature and in favor of the linear one (7/6 = 41/ 59)at 100 "C. Polymeric compounds and hydrogenation products of olefins or aldehydes were not observed. Hydroformylation of 1was carried out under the same experimental conditions employed for 2. As already reported,l the ratio of aldehydes 4 and 5 (Scheme 2)does not depend to any appreciable extent on the degree of conversion because 1 cannot isomerize. As shown in Figure 1, the linear isomer amount increases slightly with increase of the temperature, from 18% a t 20 "C t o 24% a t 100 "C. Deuteriofomylation of Ethyl AUyl Ether (2). In order to elucidate the influence of the reaction temperature on the regioselectivity, deuterioformylation of 2 a t partial substrate conversion (30%)has been investigated at 20 and 100 "C under CO and D2 (1:l)constant pressure of 100 atm. Deuterioformylation and hydroformylation experiments give approximately the same regioselectivity when the reaction is carried out under identical conditions of temperature and pressure. The crude reaction mixtures were directly investigated by 2HNMR spectroscopy, allowing a rapid identification of the deuterated species, thanks to the similarity of the chemical shift ranges of lH and 2Hnuclei. In Figure 2a the spectrum of the crude reaction mixture obtained by deuterioformylation at 20 "C shows only the signals of the two expected (Table 1) aldehydes, 1,3-dideuterio4-ethoxybutanal (1,3-d2-6) and l-deuterio-2-(deuteriomethyl)-3-ethoxypropanal(1,2(1)-d2-7), with no olefinic deuterium resonances being observed.

I

Figure 2. 2H NMR spectrum (46 MHz, 25 "C, C6D6 as external standard) of the crude reaction mixture in benzene, obtained by deuterioformylation of the ethyl allyl ether (2) at (a) 20 "C and at (b) 100 "C. Table 1. 2HNMR Chemical Shifts"of Mono- and Dideuterated SubstratesbArising from the Deuterioformylation of Ethyl Allyl Ether (2) in the Presence of Rlu(C0)12 4 PPm (Z)-3dl-3 (E)-3-dl-3 1,3d2-6 1,2(1)d2-7 1,3-d2-6 1 3 Ud2-7

(Z)-EtO-CH-CH-CH2D @)-E tO-CH-CH-CH2D EtO-CH2-CHD-CH2-CDO EtO-CH2-CH(CH2D)-CDO EtO-CH2-CHD-CH2-CDO EtO-CH2-CH(CH2D)-CDO

1.38 1.60 9.35 9.44

1.53 0.80

Referred to C6D6 as external standard; 46 MHz, C6&,25 "c. 6 = 4-ethoxybutanal, 7 = 2-methyl-3ethoxypropanal. a

*3

= 1-ethoxypropene,

In the 2H NMR spectrum of the crude reaction mixture obtained a t 100 "C (Figure 2b) two resonances at 1.60 and 1.38 ppm are present in addition to the aldehyde signals. These signals can be ascribed to the two geometric isomers (E)-3-deuterio-l-ethoxypropene ((E)-3-dl-3) and (2)-3-deuterio-l-ethoxypropene ((27-3dl-3) arising from the presence of deuterium in the position 3 (EtO-CH-CH-CH2D species). The forma-

Rh-Catalyzed Hydroformylation of Unsaturated Ethers

Organometallics, Vol. 14,No.10,1995 4647

Scheme 4

2

I

Rh b

IC

tion of 2 and E monodeuterated 3 can be rationalized on the basis of the reaction sequence depicted in Scheme 4. The n-complex I, formed by coordination of 2 to a rhodium deuteride species, gives rise to the isomeric alkylrhodium intermediates n and b. If this step is irreversible, the alkyl isomers will be converted into normal 1,3-d2-6 and branched 1,2(1)-d2-7 aldehydes, respectively. By contrast, if the formation of the rhodium alkyls is reversible, via a ,&hydride elimination process, then the n-complexes rhodium hydride/monodeuterated 2 (Ia and Ib) and rhodium hydride/monodeuterated 3 (IC)will form. Exchange of labeled olefins with unlabeled 2, present in excess a t low conversion, could give the n-complex I1 (Rh-W2 unlabeled), which originates the monodeuterated isomeric aldehydes, and the free deuterated olefins 2-deuterio-3-ethoxypropene (2-dl-2) (from the linear alkyl), (E)-and (2)-1-deuterio3-ethoxypropene (l-d1-2), and (E)-and (2)-3-dl-3 (from the branched one). The 2HNMR analysis (Figure 2) indicates that at room temperature the metal alkyls are formed irreversibly and are completely transformed into aldehydes (no resonances arising from olefinic deuterium atoms were present). In contrast, at high temperature only the formation of the branched alkyl is reversible, the linear alkyl intermediate giving the expected linear aldehyde. In particular only 3-deuterio1-ethoxypropene was formed, the 2HNMR signals at 1.60 and 1.38 ppm corresponding to the CH2D groups of the two geometrical isomers (E)and (21,respectively. Deuterioformylation of Ethyl Vinyl Ether (1). Deuterioformylation of 1 was carried out under the same experimental conditions adopted for 2. Also in this case, deuterioformylation and hydroformylation experiments give approximately the same regioselectivity. In Figure 3a is reported the 2HNMR spectrum of the crude reaction mixture obtained by deuterioformylation a t 20 "C. Only the signals corresponding to the two expected aldehydes 1,3-d2-3-ethoxypropanal (1,3-d2-4)and 1,3-d2-2-ethoxypropanal(1,3-d2-5) (Table 2) are present. In the 2H NMR spectrum of the crude reaction mixture formed in the reaction at 100 "C (Figure 3b),

in addition to the signals of the isomeric aldehydes, four new resonances were observed. In particular the three resonances a t 6.44, 4.16, and 3.97 ppm are due to deuterium nuclei on the vinyl moiety of 1. The signal at 6.44 ppm comes from the isotopic species EtOCD=CHz (1-d1-1)having the deuterium a t position 1of the double bond. The two absorptions at 4.16 and 3.97 ppm come from the two geometrical isomers (E)-2deuterio-1-ethoxyethene ((E)-2-d1-1)and (2)-2-deuterio1-ethoxyethene ((2)-2-d1-1)arising from the presence of deuterium in position 2 (EtO-CH=CHD species). The signal at 2.15 ppm can be ascribed to CHD a to the carbonyl group of the normal aldehyde 1,2-dideuterio3-ethoxypropanal (1,2-d2-4).13 Analysis of the 2HNMR data leads to important conclusions that can be summarized as follows: At low temperature deuterated 1 is not present in the reaction mixture. The two rhodium alkyl isomers n* and b1 (Scheme 5) must, therefore, be formed irreversibly and be converted completely into the expected 1,3-dideuterated aldehydes. At high temperature the formation of both branched and normal alkyls is reversible; they undergo ,&hydride elimination to give the two rhodium-hydride n-complexes IIIa and IIIb, in which 1-or 2-dl-1, respectively, is coordinated. When these complexes undergo intermolecular exchange with unlabeled 1, in addition to the complex IV (Rh-W1 unlabeled), which gives the monodeuterated isomeric aldehydes, free (E)-2-d1-1, (Zb2-dl-1, and 1-d1-1 are formed in solution. The three resonances in the range of vinylic protons found in the 2HNMR spectrum at high temperature are in agreement with this interpretation. In addition, complex IIIb gives rise to the linear rhodium-alkyl intermediate ,I1, from which the isomeric 1,2-dideuteratednormal aldehyde can be formed, thus accounting for the signal a t 2.15 ppm. A similar sequence for the complex IIIa would yield the aldehyde 1,2-dideuterio-2-ethoxypropanal(1,2-dz-S), via the branched alkyl-rhodium intermediate b",but no direct evidence for the presence of this species has been found (13)GC-MS analysis of the crude reaction mixture recovered at 100 "C and 30% of conversion showed that the mono- and dideuterated species were present in both isomeric aldehydes, the trideuterated ones ) in the branched isomer. being present in very low amount ( ~ 4 %only

Lazzaroni et al.

4648 Organometallics, Vol. 14,No. 10,1995

-L

ra

b ~ , . l . . , . . , . ( . . . . , . . . . r . . . . , . . . . ( . . . . , . . . - .I I

9

I

7

1

5

1

2

.

.

L .

,

r

4

. ..c.,

Figure 3. 2HNMR spectrum (46 MHz, 25 "C, C6D6 as external standard) of the crude reaction mixture in benzene, obtained by deuterioformylation of the ethyl vinyl ether (1) at (a) 20 "C and at (b) 100 "C. Table 2. 2H NMR Chemical Shifts" of Mono- and Dideuterated Substrates*Arising from the Deuterioformylation of Ethyl Vinyl Ether (1) in the Presence of Rh(C0)12 4 PPm 1-d1-1 (E)-2-d1-1 (2)-2-d1-1

1,3-d2-5 1,2-d2-5 1,3-d2-4 1,2-d2-4

EtO-CDeCHz (E)-EtO-CH=CHD (2)-EtO-CH=CHD CDO EtO-CH(CHzD)-CDO EtO-CD(CDO)CH3 EtO-CHD-CH2-CDO EtO-CH2-CHD-CDO

6.44 4.16 3.97 9.46 0.99 3.27 3.34 2.15

a Referred to C6D6 as external standard; 46 MHz, C&, 5 = 2-ethoxypropanal, 4 = 3-ethoxypropanal.

25 "C.

in the 2HNMR spectrum of the crude reaction mixture (Figure 3b). However, the 2H NMR spectra of an isomerically omogeneous sample of the branched aldehyde (obtained by preparative GLC of the reaction mixture) shows a resonance a t 3.27 ppm of very low intensity with respect to those of CHzD and CDO, which must be assigned to the deuterium atom bound to the tertiary atom of 1,2-d2-5. Therefore, when the reaction is carried out at 100 "C and low conversion, the 1,2-

dz-5 species is obtained in very low amount and its characteristic resonance a t 3.27 ppm is masked by the resonance at 3.34 ppm due to the linear isomer 1,3-d24.14

Conclusions The results of this investigation clearly show that the rearrangement of the isomeric metal-alkyl intermediates under hydroformylation conditions strongly depends on the reaction temperature and substrate structure and can be deduced by direct 2HNMR observation of the crude deuterioformylation reaction mixtures. In the case of ethyl allyl ether 2, at room temperature no isomerization to internal olefin 3 was observed in the unconverted substrate and no deuterium is present in (14)It is noteworthy that the 2H NMR spectrum of the crude reaction mixture arising from a deuterioformylation run of ethyl vinyl ether a t 100 "C and 60% of conversion showed two partially superimposed resonances at 3.27 and 3.34 ppm due to the branched and linear aldehydes, respectively. In particular the signal a t 3.27 ppm corresponds to the isotopic species EtO-CD(CDO)C& (X= H or D) arising from the branched alkyl-rhodium intermediate b (Scheme 5 ) or from deuterioformylation of 1-deuterio-1-ethoxyethene( l-dl-l)formed, under hydroformylation condition, as described above.

Rh-Catalyzed Hydroformylation of Unsaturated Ethers

Organometallics, Vol. 14, No. 10, 1995 4649

Scheme 5

111.

ad' 1

IV

ldl-I

+ W-D =Bra

.Rh-D

ni

=P a

If-4.5

the residual 2. Hence, the olefin addition to the metal hydride is not reversible under these conditions. At higher temperatures only 3-deuterio-1-ethoxypropeneis formed as a consequence of the branched metal-alkyl intermediate rearrangement (Scheme 4). A similar behavior of the branched alkyls has been proposed in the rhodium-catalyzed hydroformylation of cyclic vinyl and allyl ethers.15 Thus, (i) at room temperature the regioselectivity of formation of the isomeric aldehydes reflects the regioselectivity of formation of the metalalkyl intermediates, and (ii)the different behavior of n and b metal-alkyls toward the P-hydride elimination accounts for the decrease of the branched in favor of the normal one observed a t high temperature. The linear alkyl is converted to linear aldehyde, while the branched one partially converts into branched aldehyde and partially regenerates a n-metal complex, the latter process becoming more significant as the temperature increases. In the case of ethyl vinyl ether 1, at room temperature no deuterium is present in the residual substrate. Hence, the olefin addition to the metal hydride is not reversible under these conditions. At higher temperatures (E)- and (2)-1-deuterio-2-ethoxyethene, as well as 1-deuterio-1-ethoxyetheneare formed (Scheme 5) as a consequence of rearrangement of both metal-alkyl intermediates. The above findings indicate that (i) a t room temperature the regioselectivity of (15) Polo, A.; Claver, C.; Castillbn, S.; Ruiz, A. Organometallics 1992, 11, 3525.

formation of the isomeric aldehydes reflects the regioselectivity of formation of the metal-alkyl intermediates; (ii) the very similar behavior of branched b and linear n metal alkyls toward P-hydride elimination accounts for the slight influence of the temperature on the regioselectivity. The apparent significant increase of the linear aldehyde with the increase of the temperature found in the case of the hydroformylation of ethyl allyl ether has been previously observed for the hydroformylation of styrene.16 For both substrates, a t high temperatures the branched alkyl undergoes P-hydride elimination to a very large extent unlike the linear one in which this process occurs in very low amount. Different behavior has been observed for ethyl vinyl ether. In this case both rhodium-alkyl isomers give rearrangement a t high temperature, therefore the regioselectivity is slightly influenced. In simple 1-alkenes such as l-hexene17 the /?-hydride elimination occurs for both rhodium-alkyl isomers at high temperature, but the rearrangement of the branched alkyl predominates over that of the linear one. As a consequence the regioselectivity changes from 52/48 at room temperature to 71/29 at 100 "C. The very different behavior of ethyl vinyl ether with respect to styrene is unexpected: the linear rhodium(16)Lazzaroni, R.;Raffaelli, A.; Settambolo, R.; Bertozzi, S.; Vitulli, G . J.Mol. Catal. 1989,50,1. (17)Lazzaroni, R.; Pertici, P.; Bertozzi, S.; Fabrizi, G. J.Mol. Cutal. 1990,58,75.

Lazzaroni et al.

4650 Organometallics, Vol. 14, No. 10,1995 Scheme 6

The above findings clearly point out the usefulness of the deuterioformylation reaction at partial substrate conversion in the investigation on the regioselectivity. In addition, these results also show that 2H NMR spectroscopy directly on the crude reaction mixtures is a valuable tool for mechanistic investigations in catalytic processes.

Z = OEt: CH20Et

alkyl intermediate undergoes /3-hydride elimination in the case of the ether, but not for the styrene, although both liberated olefins are stabilized, having the double bond conjugated with an ethereal oxygen or an aromatic ring. It is noteworthy to point out the high regioselectivity in favor of the branched aldehyde observed for both ethers 1 and 2 at room temperature when there are no side reactions, such as /3-hydride elimination. Under these conditions the regioselectivity reflects the relative rates of formation of the isomeric alkylrhodium intermediates. As already reported for styrene,16the prevalence of the branched isomer must be related to the polarization required for the formation of a rhodiumcarbon bond, having a partial negative charge on the carbon atom and a positive one on the rhodium atom. As shown in Scheme 6, the delocalization of the negative charge on the Z group due t o the electron-withdrawing inductive effect of the oxygen atom is stronger in the intermediate leading to the branched isomer than the normal isomer. As a consequence the branched chain intermediate is stabilized and is preferred, resulting in regioselective formation of the branched aldehyde. This effect is especially significant in the case of the ethyl vinyl ether because the heteroatom is adjacent to the negative charge (regioselectivityin favor of the branched isomer, 82/18) rather than for ethyl allyl ether (regioselectivity, 70/30).

Experimental Section Benzene was dried over molecular sieves and distilled under nitrogen. The starting vinyl and allyl ethers were commercially available. Rh4(CO)12 was prepared as reported in the literature.ls GC analyses of the reaction mixtures were performed on a Perkin-Elmer 8500 chromatograph equipped with a 12 m x 0.22 mm BP1 capillary column, using helium as carrier gas. 2H NMR spectra of the crude products in benzene were recorded on a Varian VXR 300 spectrometer operating at 46 MHz for 2H. Chemical shifts were referred to CsDs as external standard. Hydroformylation or Deuterioformylation of Vinyl and Allyl Ethers: General Procedure. A solution of ether mmol) in benzene (10 mmol) and Rh4(C0)12 (5 mg; 6.7 x (10 g) was introduced by suction into an evacuated stainless steel autoclave. Carbon monoxide was introduced, the autoclave was then rocked and heated to the reaction temperature, and hydrogen or deuterium was rapidly introduced up to the desired total pressure. When the gas absorption reached the value corresponding to the fixed conversion, the reaction mixture was siphoned out. The degree of conversion was measured by GLC, using toluene as an internal standard.

Acknowledgment. This work was supported by the research program Progetto Finalizzato per la Chimica Fine 11. OM950087A (18)McCleverty, J.;Wilkinson, G. Inorg. Synth. 1966,8,211