Organometallics 1983, 2, 157-160 methods. Solvents were dried with NaK alloy. In a typical preparation, 4.9 mmol of Zr(CH2CMe3)4in 20 mL of ether was treated with 20 mmol of Me3PCH2in 30 mL of ether. The solution changed from colorless to red and yielded yellow crystals of I upon cooling to -40 "C. Elemental analysis was satisfactory.21 NMR data were collected on a JEOL FX-SOQ spectrometer with a multinuclear insert. Infrared spectra were taken with a Perkin-Elmer 283B spectrophotometer. Observed bands for I: 1289 (m), 1270 (m), 1057 (m), 1020 (vs), 981 (m), 960 (w), 940 (s), 922 (m), 886 (m), 841 (w), 817 (s), 750 (s), 729 (s), 701 (m), 654 (20) Schmidbaur, H.;Tronich, W. Inorg. Synth. 1978,18, 137-138. (21) Analysis by Schwarzkopf Microanalytical Laboratory, Woodside, NY, 11377, gave an empirical formula of ZrP3.0C12.1H29.0 with four elements totaling 99% of the material. (22) I = S(C - RB) and u(n = [S2(C + RZB)+ K P ] ~ where / ~ , C = total counts, R = twice the ratio of scanning time to total background B, S is the scan rate, and K = 0.05, a factor to reflect instrument stability. (23) All computing was done on a PDP 11/60 using the 1980 EnrafNonius SDP package. (24) Collier, M.R.; Lappert, M. F.; Pearce, R. J. Chem. Soc., Dalton Trans. 1973,445-451. (25) Baldwin, J. C.; Keder, N. L.; Strouse, C. E.; Kaska, W. C. 2. Nuturforsch., E Anorg. Chem., Org. Chem. 1980,35B, 1289-1297.
157
(m), 603 (s), 506 (m), 420 (w) cm-'. X-ray Crystallography. Crystals of air-sensitive I were examined under a layer of degassed Nujol. A selected crystal was mounted in an 0.5-mm glass capillary in an epoxy matrix. Details of data collection are given in Table V. Standard Lorentz and polarization corrections were applied to the data.23 Correction was also made for a total linear decrease in intensity of the standards of about 13% by using program CHORTA.23 Because the precise crystal dimensions were unknown and the absorption coefficient was small, no absorption correction was applied. Since anisotropic refinement produced no unusual thermal parameters, the omission of this correction was deemed acceptable. The structure was solved for the heavy atoms from a Patterson map and completed by standard Fourier and least-squares procedures to final discrepancy indices R = CllFol- lFcll/lFol = 0.059 = 0.082 and an esd of an and R, = (CIIFol- IFc112/CIFo12)1/z observation of unit weight of 2.48. Registry No. I, 83603-93-4; Zr(CH2CMe3)4,38010-72-9; Me3PCH2,14580-91-7. Supplementary Material Available: Listings of anisotropic thermal parameters, additional bond distances and angles, and observed and calculated structure factors (16 pages). Ordering information is given on any current masthead page.
Photolytic Cleavage and Isomerization Reactions of ( 1-a,4-6-q3-0rgano)tricarbonyliron Complexes Involving an (q4-Diene)(q2-olefin)Fe(C0)2Intermediate. Quantum Yields and Mechanistic Studies Takeo Akiyama, Friedrich-Wilhelm Grevels, Johannes G. A. Reuvers, and Peter Ritterskamp Max-Planck-Institut fur Strahlenchemie, 0-4330 Mulheim a.d. Ruhr, Federal Republic of Germany Received April 22, 1982
Photolysis of two isomeric (1-~,4-6-q~-organo)Fe(CO)~ complexes (la/ 1b)-the organic moiety being an adduct of methyl acrylate and 2,3-dimethylbutadiene-results in the interconversion of the two isomers and in the cleavage of the initially formed carbon-carbon bond leading to (q4-2,3-dimethylbutadiene)Fe(C0)3 (2) and methyl acrylate. T h e quantum yields are in the range of 0.03 (1 2), 0.07 (la l b ) , and 0.20 ( l b la). (q4-2,3-Dimethylbutadiene)(q2-methyl acrylate)Fe(CO)2 is the key intermediate in these processes. complex (5) parallel the above The thermal reactions of the analogous (1-~,4-6-q~-organo)Fe(CO)~P(C~H~)~ photoreactions of 1.
-
-
Scheme I
Introduction Prolonged irradiation of pentacarbonyliron i n t h e presence of 1,3-dienes a n d methyl acrylate' or other olefins2 results in carbon-carbon bond formation to give products in which a monoolefin-diene adduct is 1,4-6-q-coordinated to t h e F e ( C 0 ) 3 moiety. T h e reaction proceeds via either )~ (cf. t h e (q2-olefin)Fe(CO), or ( ~ ~ - d i e n e ) F e ( C Ocomplexes ref 3-6), as monitored by infrared spectroscopy a n d con(1) Grevels, F.-W.;Feldhoff, U.; Leitich, J.; Krtiger, C. J. Organomet. Chem. 1976,118,79. In relation to the q3-allylgroup the ester group of la is exo and that of lb endo: Chiang, A.-P.; Krager, C., unpublished results (X-ray structure analysis of lb). (2) Grevels, F.-W.; Salama, I., unpublished results. (3) Whitesides, T.H.; Shelly, J. "Abstractsof Papers", 169th National Meeting of the Americal Chemical Society, Philadelphia, PA, Apr 1975; Americal Chemical Society: Washington, DC, 1975; ORGN-86. (4) Bond, A.; Lewis, B.; Green, M. J. Chem. SOC.,Dalton Trans. 1975, 1109. (5) Green, M.; Lewis, B.; Daly, J. J.; Sanz, F. J. Chem. SOC.,Dalton Trans. 1975,1118. (6) Kerber, R. C.; Koerner von Gustorf, E. A. J . Orgunomet. Chem. 1976,110,345.
-
hv
1
-2CO
hv
I
- CC
1
firmed by separate irradiations starting from the respective reactants.' Irradiation of (q4-diene)Fe(CO)3/fluoroalkyne mixtures yields analogous product^.^,^
0276-7333/83/2302-0157$01.50/00 1983 American Chemical Society
158 Organometallics, Vol. 2, No. I , 1983
Akiyama et al. +
A
~i
4000
[M ' c m
-
I
~
15000 C02CH3 l a
1
I ~10000
I
1
2000
1000
H
2
lb
I
-
3000
0
1
-
I
*.L{
I
'
A
C
5000
O
k
ImMol/il
-2 22
24
26
28
30 32
30
36
38
40
42
44
46
7[~~'cm-']
Figure 1. Electronic spectra of the complexes la, lb, 2, and 3, in n-hexane.
Scheme I1
I
I
2Li
A
I
la
2
Scheme I summarizes the photochemical synthesis of the (2,3-dimethylbutadiene/methyl acrylate)Fe(CO), adduct 1 in n-hexane or diethyl ether a t ambient temperature. Two stereoisomers, la and lb, are formed which differ in
2
L
6
8
10 min
Figure 2. Irradiation of 2.2 X M solutions of the adducts la (A) and lb (B) in n-hexaneat ambient temperature (immersion lamp apparatus, solidex glass, Hg lamp, Philips HPK, 125 W).
Table I. Quantum Yieldsa for the Photolytic Cleavage and Isomerizationof l a and lb expt no. @-la 0 Ib 02 la
la:
3b
R ' = H, R 2 = CO2CH3
4 5 6 7c 8'
l b : R ' = C02CH3, R 2 = H
the orientation of the ester group.' As the reaction proceeds, (v2-methylacrylate)Fe(C0)4 is consumed almost completely whereas, even after extended irradiation, substantial amounts of (~~-2,3-dimethylbutadiene)Fe(CO)~ (2) are remaining present in the reaction mixture. At first glance, this could be attributed to internal light filter effects (see electronic spectra, Figure 1). However, preliminary experiments have shown that the adducts 1, upon irradiation, release methyl acrylate with reformation of 2.l It seemed worthwhile to study this photolytic reaction in some detail since, beyond our interest in this particular case, a more comprehensive knowledge of the mechanism of such a facile metal-assisted reversible C-C coupling of di- and monoolefins may be of general relevance to transition-metal c a t a l y s i ~ . ~ Results and Discussion Irradiation of the two isomeric (cr,q3-allyl)Fe(CO),complexes la or l b in both cases not only brings about the (7) Bottrill, M.; Goddard, R.; Green, M.; Hughes, R. P.; Lloyd, M. K.; Lewis, B.; Woodward, P. J. Chem. SOC.,Chem. Commun. 1975, 253. (8) Bottrill, M.; Davies, R.; Goddard, R.; Green, M.; Hughes, R. P.; Lewis, B . Woodward, P. J. Chem. SOC.,Dalton Trans. 1977, 1252. (9) Jukmont, M.; Teyssig, Ph. Aspects Homogeneous Catal. 1981,4, 99.
1 2
av
expt no. lb
9 lob
av
0.109 0.121 0.106 0.097 0.083 0.109 0.132 0.111 0.109
0.069 0.083 0.076 0.056 0.063 0.056 0.083 0.066 0.069
0.033 0.025 0.023 0.033 0.019 0.024 0.028 0.027 0.028
@-ib
0.260 0.204 0.232
0.213 0.188 0.201
0.033 0.028 0.031
At 340 nm (29 400 cm"), in n-hexane under CO atmosphere; 22 "C; 20% conversion, for details see Experimental Section. Light intensity reduced, as compared to no. 1 + 2 and 9, respectively,by using a 32% a
transmission wire guaze. Light intensity reduced, as compared to no. 4-6, by using a 14% transmission neutral density filter.
formation of (v4-2,3-dimethylbutadiene)Fe(CO),(2) with loss of methyl acrylate but also concomitantly, isomerization, l a s lb, is observed (Scheme 11). No other products are detected if the photolysis is performed under a carbon monoxide atmosphere. Figure 2 shows the course of two experiments which have been carried out on a preparative scale. The extent of photolytic cleavage yielding 2 is nearly the same for each of the two starting materials la and lb. However, as a notable difference, the isomerization l b h la (Figure 2B) is more efficient than
Organometallics, Vol. 2, No. I , 1983 159
Reactions of (1-a,4-6-~-Organo)tricarbonylironComplexes
Scheme IV
Scheme 111 la
lb
lb
la
C02CH3
H
3
4a
11%
1-2
t
CO hv
.F;(CO)3]
the reverse reaction (Figure 2A). These trends are confirmed by the quantum yields (Table I) which indicate an equal, albeit moderate, efficiency for the formation of 2 in each case (& 0.03). On the other hand, the quantum yields for the formation of the respective isomer differ by a factor of about 2.5-3. The sum of the product quantum yields is in excellent accord with the value for the disappearance of the respective starting material. In regard to the mechanism two alternatives may be considered for the initial photolytic step: either homolysis of the metal-carbon u bond (Scheme 111) or dissociation of one carbon monoxide ligand. Spectroscopic studies at low-temperature matrix isolation conditionslO at 10-12 K have shown that detachment of CO is the only detectable primary photoreaction under those circumstances; there is no indication of an isomerization, la (hv)s lb (hv)in either direction. If the photolysis in solution is carried out under an inert-gas atmosphere an additional product (3) can be observed. Enrichment of the solution with this product, such that it can be isolated, is favored by low temperatures and by removal of liberated carbon monoxide. It is identified as (v4-2,3-dimethylbutadiene) (q2-methylacrylate)Fe(C0)2(3),which has previously been prepared" from ($-methyl a~rylate)Fe(CO)~ and 2,3-dimethylbutadiene. At ambient temperature-without irradiation-3 readily takes up one carbon monoxide molecule to yield a mixture of the isomeric adducts la and lb in a -2.5:l ratio.ll No (~~-2,3-dimethylbutadiene)Fe(CO)~ (2) is formed under these conditions. However, irradiation of 3 under CO atmosphere in n-hexane solution at -45 "C -where the thermal reaction yielding 1 is almost negligible-results in complete displacement of methyl acrylate by carbon monoxide yielding 2. From these results we conclude that 3 is the key intermediate in the reactions depicted in Scheme 11, namely, the photochemical interconversion of la and lb and the photolytic formation of 2 as well. As outlined in Scheme IV the photolytic dissociation of one carbonyl ligand from la or Ib, with retention of the carbon skeleton of the organic moiety, yields the 1Belectron intermediates 4a or 4b which have been observed in low-temperature matrices.l0 A t ambient temperature these coordinatively unsaturated species can be expected to undergo spontaneous reductive cleavage to form the Welectron complex 3. The diradical type intermediate shown in Scheme I11 cannot (10) Grevels, F.-W.; Klotzbucher, W. E. Inorg. Chem. 1981,20, 3002. (11) Grevels, F.-W.;Schneider K. Angew. Chem. 1981,93,417;Angew. Chem. Int. Ed. Engl. 1981,20, 410.
I-
4b
AC02CH3
t
2
entirely be excluded. However, a contribution from this alternative pathway ought to be small in view of the fact that the 1a:lb ratio in the course of the photolysis (Figure 2) approaches the same value as that obtained in the thermal reaction of 3 with carbon monoxide.'l Bearing in mind that two consecutive photochemical steps are involved in the conversion 1 2 (Scheme IV) one might expect the efficiency of this process to depend on the intensity of light. At the outset of this study first preparative data seemed in fact to indicate such a light intensity effect,12i.e., that higher intensities would reduce the yield of 1 by favoring the cleavage back to 2. However, the quantum yield determinations at 340 nm (29400 cm-')-where the absorbances of la/lb and the intermediate 3 are not much different (Figure 1)-showed such an effect to be within the experimental error limits (Table I), if at all existing. A solid interpretation of this result requires a more detailed knowledge of the kinetics governing the delicate balance of the reactions involving the species 3 and 4a/4b (Scheme IV). We are anticipating this information from forthcoming kinetic studies using flash photolysis and fast infrared techniques. Starting from 3 not only carbon monoxide but also other ancillary ligands such as phosphines or i~onitriles'~ are able to bring about the oxidative cyclization as long as these ligands are not too bulky. The triphenylphosphine complex 5 (Scheme V) is moderately stable and can be isolated from a concentrated solution containing excess triphenylphosphine. In dilute n-hexane solutions the metal carbonyl bands of 5 (1993 and 1935 cm-') gradually disappear. Complex 3 is formed intermediately and reacts further to yield (v4-2,3-dimethylbutadiene)(triphenylphosphine)Fe(CO)2(6) as the final product. This thermal reaction sequence parallels nicely the photochemical conversion
-
hv
hu
1 ZA 3 - 2
thus confirming our formulation of the reductive cleavage as a thermal process subsequent to the initial loss of an ancillary ligand (Scheme IV). If 5 is allowed to decompose under carbon monoxide atmosphere or in the presence of, e.g., trimethyl phosphite the (a,q3-allyl)ironcomplexes la/lb or 713 are obtained (Scheme V). Further detailed investigations concerning the ligand-induced oxidative cyclization of type 3 complexes are in progress. Although (12) Akiyama, T.; Grevels, F.-W.; Salama, I., Proceedings of the International Conference on Organometallic Chemistry, 9th, Dijon/France, Sept 1979, Abstract D2. (13) Grevels, F.-W., et al., unpublished results.
160 Organometallics, Vol. 2, No. 1, 1983 Scheme V
i i ( C O ) * P(CsH& 6
such a metal-assisted C-C bond formation between a diene and a monoolefin has been classified as a symmetry-forbidden process,'* the reaction proceeds under rather mild conditions i n case of the carbonyliron complex 3. Experimental Section All reactions and manipulations were carried out, unless noted otherwise, under argon and in argon-saturated solvents. Irradiations were performed in an immersion lamp apparatus (solidex glass, X 1 280 nm) by using a mercury lamp Philips HPK 125 W. Spectra were recorded by using the following instruments: IR, Perkin-Elmer 580; UV-visible, Bruins Instruments MIT 20. Melting points were determined on a Reichert Kofler apparatus. Microanalysis was performed by Dornis & Kolbe, Mulheim a.d. Ruhr. Analytical grade solvents (Merck, Darmstadt) and reagents (Merck and Fluka) were used as received. Complexes la/lb,' 2 (by analogy to ref 15), and 311 were prepared according to published procedures. (1) Photolysis of Complexes la/lb. (a) Solutions of 0.120 g (0.39 mmol) of la or l b in 170 mL of n-hexane, purged with carbon monoxide for 5 min, were irradiated. Samples were drawn after 1, 2, 5, and 10 min and analyzed by quantitative infrared spectroscopy using the CO stretching bands at 2006/2062 (la), 1998.5/2057 (lb),and 2049 (2) cm-'; appropriate corrections were made in order to account for tail absorptions of the respective other components. The results are displayed in Figure 2. (b) Quantum Yield Determinations. An electronically integrating actinometer was used (described elsewhere)16which was calibrated by ferrioxalate actin0metry.l' Solutions of l a or lb (-2 X M) in n-hexane were prepared under argon and subsequently saturated with carbon monoxide. The 3-mL samples were irradiated a t 340 nm in quartz cuvettes (d = 1cm) by using a Hanovia 1000-W Hg-Xe lamp in connection with a Schoeffel Instruments GM 250 single-grating monochromator. Light intensities were in the order of (4-30) X einstein min-'. All (14) Pearson, A. G. "Symmetry Rules for Chemical Reactions"; Wiley-Interscience: New York, 1976; p 427. (15) Koerner von Gustorf, E.; Pfajfer, Z.; Grevels, F.-W. 2.Naturforsch., B: Anorg. Chem., Org. Chem., Biochem., Biophys., Biol. 1971, 26, 66. (16) Amrein, W.; Gloor, J.; Schaffner, K. Chimia 1974, 28, 185. London, Ser. A 1956, (17) Hatchard, C. G.; Parker, C. A. Roc. R. SOC. 235,518. Murov, S. L. In "Handbook of Photochemistry";Marcel Dekker: New York, 1976; p 119.
Akiyama et al. experiments were carried out at ambient temperature (22 & 2 "C). Concentrations of la, lb, and 2 were determined, a t 10-20% conversion, by quantitative infrared spectroscopy (see above). (c) Isolation of (q4-2,3-Dimethylbutadiene)(q2-methyl a~rylate)Fe(CO)~ (3). A solution of la (3.70 g, 12 mmol) in 200 mL of diethyl ether was irradiated at -30 "C for 14.5 h. Carbon monoxide was removed by an argon stream. The solvent was evaporated under vacuum a t -30 "C and the residue chromatographed on silica gel (Merck, Darmstadt; Kieselgel 60, 230-400 mesh). The column (5 X 35 cm) was operated at -30 "C. Elution with pentane gave 2 (1.06 g, 40%); 3 was eluted with pentanediethyl ether (7:3). The crude product (1.37 g, 41%) was recrystallized from hexane at -80 "C and identified" by its melting point, infrared and 'H NMR spectroscopy, and mass spectrometry. A third fraction, eluted with diethyl ether, contained la and other, unidentified, products. (2) Photoreaction of 3 with Carbon Monoxide. A solution of 3 (0.24 g, 0.86 mmol) in 180 mL of n-hexane was cooled to -45 "C and saturated with carbon monoxide. Samples were drawn after 10 min and 1 h, degassed, and warmed up to ambient temperature. The infrared spectra showed 3 and minor amounts of l a / l b but no 2. The solution of 3,under carbon monoxide at -45 "C, was then irradiated for 0.25 h. The infrared spectrum showed the almost exclusive formation of (s4-2,3-dimethyl(2), and only trace amounts of la/lb were b~tadiene)Fe(CO)~ detectable. (3)Reaction of 3 with Triphenylphosphine. A solution of 3 (0.21 g, 0.75 "01) in 1.5 mL of n-hexane was added to a solution of triphenylphosphine (0.40 g, 1.5 mmol) in 19 mL of n-hexane a t 0 "C. Yellow crystals started to precipitate after 25 min and were isolated after 1.5 h by inverse filtration, washed twice with 0.5 mL of n-hexane, and dried under vacuum: yield 0.40 g of 5 (98%): mp 75-77 "C; infrared spectrum (KBr) t(C0) 1981 and 1924 cm-l, ;(ester carbonyl) 1681 cm-'. Anal. Calcd for C30H31Fe04P:C, 66.43; H, 5.76; Fe, 10.30; P, 5.71. Found: C, 65.98; H, 6.29; Fe, 10.48; P, 5.78. (4) Reactions of Complex 5. (a) Complex 5 (10.2 mg, 0.019 mmol) was dissolved in 10 mL of n-hexane. As monitored by infrared spectroscopy, ca. 40% of 5 was converted immediately to 3. When the mixture was left standing at ambient temperature, the metal carbonyl bands of (v4-2,3-dimethylbutadiene)(triphenylph~sphine)Fe(CO)~ (6, identified by comparison with an authentic samp1el3) gradually appeared while the concentration of 3 remained nearly constant for about 2 h. After 6.5 h the formation of 6 was almost complete, and only trace amounts of 5, 3, and l a / l b were detectable. (b) With Carbon Monoxide. Complex 5 (11.3 mg, 0.021 mmol) was dissolved in 10 mL of n-hexane. When the mixture was left standing under carbon monoxide, 5 was converted into l a / l b (ca. 5-1O:l ratio). At ambient temperature the reaction was complete after 4-5 h; at 0 "C 1.5 days were required. 3 was not observed as an intermediate, and neither 6 nor 2 was detected. (c) With Trimethyl Phosphite. A solution of 5 (4.8 mg, 0.0089 mmol) and trimethyl phosphite (42.7 mg, 0.34 mmol) in 10 mL of n-hexane was allowed to stand at ambient temperature. The CO stretching bands of 5 gradually disappeared, and those of the analogous trimethyl phosphite complex 7 grew in. After 2.5 h the conversion was complete. 7 was identified by comparison with an authentic sample,13the structure of which was determined by X-ray crystallography.'8
Acknowledgment. We t h a n k the staff of t h e spectroscopic laboratories of this institute for their help and Miss H. Nowak, Mrs. R. Schrader, and M r . K. Schneider for able technical assistance. Registry No. la, 61950-05-8; lb, 61867-50-3; 2, 31741-56-7; 3, 77257-30-8; 5, 83399-55-7; 6, 74753-27-8; 7, 83399-56-8. (18) Chiang, A.-P.; Kruger, C., to be submitted for publication.
