Cobalt-catalyzed carbonylation of benzyl halides using polyethylene glycols as phase-transfer catalysts

Article abstract of DOI:10.1021/om950575i

Hydroxycarbonylation of benzyl and substituted benzyl chloride and bromide derivatives was achieved in good yields (up to 97.6%) and quantitative chemoselectivity, in the presence of Co₂(CO)₈ as the transition metal and polyethylene

Full text of DOI:10.1021/om950575i

3222
Organometallics 1996, 15, 3222-3231
Coba lt-Ca ta lyzed Ca r bon yla tion of Ben zyl Ha lid es Usin g
P olyeth ylen e Glycols a s P h a se-Tr a n sfer Ca ta lysts
Claudia Zucchi1a and Gyula Pa´lyi*,1a
Department of Chemistry, University of Modena, Via Campi 183, I-41100 Modena, Italy
Vilmos Galamb,1b Ella Sa´mpa´r-Szerencse´s, and La´szlo´ Marko´*,1c
Research Group for Petrochemistry of the Hungarian Academy of Sciences, Egyetem-u. 6,
H-8201 Veszpre´m, Hungary
Pei Li and Howard Alper*
Ottawa-Carleton Chemistry Institute, Department of Chemistry, University of Ottawa,
Ottawa, Ontario K1N 9B4, Canada
Received J uly 25, 1995X
Hydroxycarbonylation of benzyl and substituted benzyl chloride and bromide derivatives
was achieved in good yields (up to 97.6%) and quantitative chemoselectivity, in the presence
of Co2(CO)8 as the transition metal and polyethylene glycols as the phase-transfer catalyst
in an organic solvent/20% aqueous NaOH two-phase system under mild conditions (1 bar
CO, room temperature). η1-Benzyl-, η3-benzyl-, and (η1-phenylacetyl)cobalt carbonyls were
investigated as intermediates of this catalytic process.
In tr od u ction
We now report on the cobalt-catalyzed hydroxycar-
bonylation of benzyl halides using biphasic (BP) and
phase-transfer catalytic (PTC) reaction conditions. Poly-
ethylene glycols were chosen as PT agents in the latter
The carbonylation of benzyl halides and their deriva-
tives is one of the most widely investigated transition
metal catalyzed carbonylation reactions.2-4 This reac-
tion is one of the first examples of the combination of
phase-transfer catalysis with transition metal chemistry
(TM-PTC) and is the most promising in terms of
industrial application.5 Surprisingly, while a number
of complexes were tested in TM-PTC studies (Fe-
(CO)5,3w,y [PhC(O)Fe(CO)4],-4f Co2(CO)8,4b,d Na[Co-
(CO)4],3w,4c [Co(CO)4]-/resin,4g Co(CO)3(NO),4e [Co(CN)6]3-/
[Co(CO)4],-4h Pd(PPh3)4, PdCl2(PPh3)24a) as transition
metal catalysts (or precursors), quaternary ammonium
salts were used almost exclusively as PT agents for
these reactions.
The practical application of this reaction is still
limited on economic grounds.5,6 Our main goal was to
achieve some progress in this respect, aiming therefore
to use (i) mild conditions, (ii) an inexpensive transition
metal complex, (iii) a cheap PT agent, which moreover
should be effective in low concentration, and (iv) an
economical solvent, which results in good separation
during workup or no solvent at all, and (v) to avoid
environmental and safety problems.
(3) Transition metal catalysis: (a) Heck, R. F.; Breslow, D. S. J . Am.
Chem. Soc. 1963, 85, 2779. (b) Chandhari, F. M.; Knox, G. R.; Pauson,
P. L. J . Chem. Soc. C 1967, 2255. (c) Rhee, I.; Ryang, M.; Tsutsumi, S.
J . Organomet. Chem. 1967, 9, 361. (d) Ohono, K.; Tsuji, J . J . Am. Chem.
Soc. 1968, 90, 99. (e) Watanabe, Y.; Taniguchi, K.; Suga, M.; Mitsudo,
T.; Takegami, Y. Bull. Chem. Soc. J pn. 1979, 52, 1869. (f) Cassar, L.;
Foa`, M. (Montecatini) Ger. Pat. DOS 1,914,391 (1969). (g) Foa`, M.;
Cassar, L.; Chiusoli, G. P. (Montecatini) Ger. Pat. DOS 2,035,902
(1971). (h) Fukukoa, S.; Ryang, M.; Tsutsumi, S. J . Org. Chem. 1971,
36, 2721. (i) Scheben, J . A.; Ky, E.; Mador, I. L.; Orchin, M. (National
Distillers) U.S. Pat. 3,560,561 (1971). (j) Knowles, R. N. (du Pont de
Nemours) U.S. Pat. 3,636,082 (1972). (k) El Chahawi, M.; Richtzenhain,
H. (Dynamit Nobel, GFR) Ger. Pat. DOS 2,240,398 (1974). (l) El
Chahawi, M.; Richtzenhain, H. (Dynamit Nobel, GFR) Ger. Pat. Dos
2,240,399 (1974). (m) Schoenberg, A.; Bartoletti, I.; Heck, R. F. J . Org.
Chem. 1974, 39, 3318. (n) Stille, J . K.; Wong, P. K. J . Org. Chem. 1975,
40, 532. (o) Hidai, M.; Hikita, T.; Wada, Y.; Fujikara, Y.; Uchida, Y.
Bull. Chem. Soc. J pn. 1975, 48, 2075. (p) Perron, R. C. (Rhone-Poulenc)
Ger. Pat. DOS 2,600,541 (1976). (q) Schneider, K.; Kummer, R.;
Schwirten, K.; Schindler, H. D. (BASF) Ger. Pat DOS 2,521,610 (1976).
(r) El Chahawi, M.; Richtzenhain, H. (Dynamit Nobel, GFR) Swiss Pat.
589,593 (1977). (s) El Chahawi, M.; Prange, U.; Richtzenhain, H.; Vogt,
W. (Dynamit Nobel, GFR) Ger. Pat. DOS 2,553,931 (1977). (t) El
Chahawi, M.; Prange, U.; Richtzenhain, H.; Vogt, W. (Dynamit Nobel,
GFR) Ger. Pat. DOS 2,606,655 (1977). (u) Watanabe, Y.; Tanuguchi,
K.; Suga, M.; Mitsudo, T.; Takegami, Y. Bull. Chem. Soc. J pn. 1979,
52, 1869. (v) Alper, H.; Hashem, K.; Heveling, J . Organometallics 1982,
1, 775. (w) Tanguy, F.; Weinberger, B.; des Abbayes, H. Tetrahedron
Lett. 1983, 24, 4005. (x) Schneider, K.; Kummer, R.; Schwirten, K.;
Schindler, H. D. (BASF) U.S. Pat. 4,424,394 (1984). (y) Tustin, G. C.;
Hembre, R. T. J . Org. Chem. 1984, 49, 1761. (z) El Chahawi, M.
(Dynamit Nobel, GFR) Ger. Pat. DE 2,828,041 (1984). (aa) Foa`, M.;
Moro, A.; Gardano. A.; Cassar, L. (Montedison) U.S. Pat. 4,351,952
(1982). (ab) Cavinato, G.; Toniolo, L. J . Mol. Catal. 1993, 78, 131. (ac)
Ramirez-Vega, F.; Laurent, P.; Cle´ment, J .-C.; des Abbayes, H. J . Mol.
Catal. A 1995, 96, 15.
X Abstract published in Advance ACS Abstracts, April 15, 1996.
(1) (a) Work started at Research Group for Petrochemistry, Hungar-
ian Academy of Sciences, Veszpre´m, Hungary. (b) Present address:
Alkaloida Pharmaceutical Works, H-4440 Tiszavasva´ri, Hungary. (c)
Present address: Institute of Chemical Engineering, Hungarian
Academy of Sciences, Egyetem u. 2, H-8200 Veszpre´m, Hungary.
(2) Reviews: (a) El Chahawi, M.; Prange, U. Chem.-Ztg. 1978, 102,
1. (b) Falge, J . New Syntheses with Carbon Monoxide; Springer: Berlin,
1980; p 294. (c) Cassar, L. Ann. N.Y. Acad. Sci. 1980, 333, 208. (d)
Alper, H. Adv. Organomet. Chem. 1981, 19, 183. (e) Tkatchenko, I. In
Comprehensive Organometallic Chemistry, Wilkinson, G., Stone, F. G.
A., Abel, E. W., Eds.; Pergamon Press: Oxford, U.K., 1982; Vol. 8, pp
188-205. (f) Alper, H. Fundam. Res. Homogen. Catal. 1984, 4, 79. (g)
des Abbayes, H. Isr. J . Chem. 1985, 26, 249; New J . Chem. 1987, 11,
535. (h) Galamb, V. Kem. Kozlemenyek 1985, 63, 227. (i) Petrignani,
J . F. In The Chemistry of the Metal-Carbon Bond; Hartley, F. R., Ed.;
Wiley: New York, 1989; Vol. 5, p 63. (j) Colquhoun, H. M.; Thompson,
D. J .; Twigg, M. V. Carbonylation; Plenum Press: New York, 1991.
(k) Chiusoli, G. P. Transition Met. Chem. 1991, 16, 553.
(4) Transition metal catalysis combined with phase-transfer cataly-
sis: cf. also refs 3v,w,y and the following. (a) Cassar, L.; Foa`, M.;
Gardano, A. J . Organomet. Chem. 1976, 121, C55. (b) Alper, H.; des
Abbayes, H. J . Organomet. Chem. 1977, 134, C11. (c) Cassar, L.; Foa`,
M. J . Organomet. Chem. 1977, 134, C15. (d) des Abbayes, H.; Boloup,
A. J . Chem. Soc., Chem. Commun. 1978, 1090. (e) Gambarotta, S.;
Alper, H. J . Organomet. Chem. 1981, 212, C23. (f) Laurent, P.; Tanguy,
G.; des Abbayes, H. J . Chem. Soc., Chem Commun. 1986, 1754. Solid/
liquid TM-PTC: (g) Cianelli, G.; Foa`, M.; Umani-Ronchi, A.; Gardano,
A. (Montedison) U.K. Pat. Appl. GB 2,079,748 (1982). (h) Amer, I.
Inorg. Chim. Acta 1994, 216, 97.
S0276-7333(95)00575-9 CCC: $12.00 © 1996 American Chemical Society
+
+

Carbonylation of Benzyl Halides
Organometallics, Vol. 15, No. 14, 1996 3223
Ta ble 1. Ca r bon yla tion of Ben zyl Ha lid es a n d
Th eir Su bstitu ted Der iva tives
case. There are surprisingly few reports on the use of
these cheap and efficient7 phase-transfer agents in TM-
PTC.8
The other goal set by this study was to utilize some
recently accumulated knowledge9 on possible organo-
metallic intermediates of the hydroxycarbonylation
reactions10 to gain additional insight into the mecha-
nism.
PEG, M
(mmol)
org
yield,
%
b
mmol
phasea
TONc
PhCH2Cl
15d
15d
12e
12e
15d
15d
25d
50
15
25
30
30
400 (1)
400 (0.5) t-AmOH
400 (1)
400 (1)
t-AmOH
97.6
83
80
37
67
14.6
12.5
9.6
4.4
10
12.8
22.4
33.8
11.7
20.5
25.5
24.0
72
t-AmOH
MP
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
FO
FO
FO
FO
none
600
85
400 (1)
400 (1)
2000 (1)
400 (1)
89.7
67.6
78.3
82
85
80
Exp er im en ta l Section
Starting materials were of commercial origin, their purity
was determined by GLC, and they were distilled, if necessary,
before use. Co2(CO)8 was either made by the method of Szabo´
et al.11 or purchased from Aldrich Chemical Co. Fusel oils
were distillery waste and were distilled before use, using a
1.3 m column filled with glass rings; the first drop generally
appeared at 118 °C, and practically all material came over
below 132 °C headspace temperature. The fractions above 120
°C were used; their composition varied slightly and were
determined by GLC12a (3-Me-butan-1-ol 80-85%, 2-Me-butan-
1-ol 15-20%).
400 (1)
400 (2)
90
30
80
70
20.8
30
none
PhCH2Br
p-MeC6H4CH2Cl
15e
15d
400 (1)
400 (1)
400 (1)
400 (1)
400 (1)
400 (1)
t-AmOH
t-AmOH
t-AmOH
t-AmOH
Mp
65
71
52
22
64
9.6
14.6
18.9
63
66
35
29
30
44
24
39
26
27
34
52
14
63
24.4
10.7
7.8
6.6
19.2
1.4
2.2
2.8
18.9
9.9
5.2
4.4
4.5
6.6
3.6
5.8
3.9
4.0
5.1
7.8
2.1
18.9
m-MeC6H4CH2Cl 15d
o-MeC6H4CH2Br
12d
12d
Me5C6CH2Clf
7.5g
t-AmOH
7.5g
7.5g
400 (0.5) MP
t-AmOH
All operations, until completion of the catalytic reactions,
were performed according to standard Schlenk techniques.13
CO was treated with 10% KOH before use.
p-BrC6H4CH2Br
p-ClC6H4CH2Cl
m-ClC6H4CH2Cl
12e
400 (1)
400 (1)
400 (1)
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
t-AmOH
15d
7.5g
Infrared spectra were recorded with an IR 75 (Carl Zeiss,
J ena, Germany) and Bruker FT-IR IFS 113V or Perkin-Elmer
7.5g
7.5g
7.5g
7.5g
7.5g
7.5g
7.5g
7.5g
1
783 spectrometer, and H- and 13C-NMR spectra were recorded
o-ClC6H4CH2Cl
p-FC6H4CH2Br
400 (1)
400 (1)
with 80-MHz BS-487 (Tesla, Brno, CSFR), Varian EM-360,
Bruker WP-80 (FT, 80 MHz), and Bruker AMX-400 (FT, 400
MHz) instruments; GLC measurements were made with a
Hewlett-Packard HP-5830 A or Varian 6000 gas chromato-
graph with a 1.3 m × 3 mm Porapak P column.
m-FC6H4CH2Br
o-FC6H4CH2Br
400 (1)
400 (1)
1-C10H7CH2Cl
15d
400 (1)
400 (1)
(5) Phenylacetic acid and its derivatives are usually manufactured
using the cyanide route: (a) Ullmans Enczyclopa¨die der technischen
Chemie; Verlag Chemie: Weinheim, Germany, 1976; Bd. 11, p 71. (b)
Pozdnyakovich, Yu. V.; Nazarenko, E. S.; Shein, S. M. Khim. Prom.-
st. 1985, (1), 5.
(6) The relatively high price of “quats” allows their use only for the
synthesis of more precious substituted phenylacetic acid derivatives:
Zaidmann, B.; Sasson, Y.; Neumann, R. Ind. Eng. Chem., Prod. Res.
Dev. 1985, 24, 390.
(7) Suggested originally for organic reactions by: (a) To¨ke, L.; Szabo´,
G. T. Acta Chim. (Budapest) 1977, 93, 421. (b) Szaka´cs, S.; Go¨bo¨lo¨s,
S.; Szammer, J . Magyar. Kem. Folyoirat 1980, 86, 273; Manatsh. Chem.
1981, 112, 833. (c) Szaka´cs, S.; J a´ky, M.; Go¨bo¨lo¨s, S.; Nagy, F. Magyar.
Kem. Folyoirat 1983, 89, 402. (d) Cf. also: Saenger, W.; Suh, I. H.;
Weber, G. Isr. J . Chem. 1979, 18, 253.
15d
12d
2-C10H7CH2Br
a
t-AmOH ) 2-methylbutan-2-ol; MP ) 4-methylpentan-2-one;
FO ) fusel oil; none ) only substrate represented organic phase.
b
Isolated yields; the products were identified by their melting
point and 1H-NMR spectra. c Turnover number: molecule of acid
d
per starting Co atom. 12.5 mL of aqueous NaOH/12.5 mL of
organic solvent; r.t., 1 bar CO; 0.5 mmol of Co2(CO)8. e 10 mL of
15-20% aqueous NaOH/10 mL of organic solvent; r.t., 1 bar CO;
0.2 mmol of Co2(CO)8. f The reaction mixture becames gelatinous
g
as the carbonylation proceeds. 6.25 mL of aqueous NaOH/6.25
mL of organic solvent; r.t., 1 bar CO; 0.25 mmol of Co2(CO)8.
(8) (a) Suzuki, N.; Tsukanaka, T.; Nomot, T.; Ayaguchi, Y.; Izawa,
Y. J . Chem. Soc., Chem Commun. 1983, 515. (b) Alper, H.; J anusz-
kiewicz, K.; Smith, D. J . H. Tetrahedron Lett. 1985, 26, 2263. (c) Wang,
J .-X.; Alper, H. J . Org. Chem. 1986, 51, 273. (d) Li, P.; Alper, H. J .
Org. Chem. 1986, 51, 4354. (e) Calet, S.; Alper, H.; Petrignani, J . F.;
Arzoumanian, H. Organometallics 1987, 6, 1625.
(9) (a) Galamb, V.; Pa´lyi, G. J . Chem. Soc., Chem. Commun. 1982,
487. (b) Galamb, V.; Pa´lyi, G. Coord. Chem. Rev. 1984, 59, 203. (c)
Galamb, V.; Pa´lyi, G.; Ungva´ry, F.; Marko´, L.; Boese, R.; Schmid, G.
J . Am. Chem. Soc. 1986, 108, 3344. (d) Haa´sz, F.; Bartik, T.; Galamb,
V.; Pa´lyi, G. Organometallics 1990, 9, 2773. (e) Ungva´ry, F.; Marko´,
L. Inorg. Chim. Acta, 1994, 227, 211.
(10) Leading mechanistic studies with Co: see also refs 2c,d,g, 3a,
4c, and 9d and the following. (a) Nagy-Magos, Z.; Bor, G.; Marko´, L. J .
Organomet. Chem. 1968, 14, 205. (b) des Abbayes, H.; Buloup, A. J .
Chem. Soc., Chem. Commun. 1978, 1090. (c) des Abbayes, H.; Buloup,
A. J . Organomet. Chem. 1979, 179, C21. (d) Moro, A.; Foa`, M.; Cassar,
L. J . Organomet. Chem. 1980, 185, 79. (e) des Abbayes, H.; Bulloup,
A. Tetrahedron Lett. 1980, 21, 4343. (f) sec-Benzyl derivatives: Fran-
calanci, F.; Foa`, M. J . Organomet. Chem. 1982, 243, 59. (g) des
Abbayes, H.; Buloup, A.; Tanguy, G. Organometallics 1983, 2, 1730.
(h) sec-Benzyl derivatives: Francalanci, F.; Gardano, A.; Abis, L.;
Fiorani, T.; Foa`, M. J . Organomet. Chem. 1983, 243, 87. (i) sec-Benzyl
derivatives: Francalanci, F.; Gardano, A.; Foa`, M. J . Organomet. Chem.
1985, 282, 277.
Ca r bon yla tion Exp er im en ts. All carbonylation reactions
were performed in three-necked reaction vessels equipped with
a gas inlet (through a sintered glass end), outlet, and a GC
septum fixed into a glass tube (to enable injecting the reagents
without opening the reaction vessel). External magnetic
stirring was applied. This reaction vessel was flushed 3 times
with CO before starting the experiments. Since the conditions
used in all the carbonylation experiments were the same, a
representative procedure is described below. Variations in the
experimental parameters are given in Table 1.
Ca r bon yla tion of Ben zyl Ch lor id e in 2-Meth ylbu ta n -
2-ol/Wa ter . 1. Bip h a sic Rea ction Con d ition s. A mixture
of dicobalt octacarbonyl [170 mg (0.5 mmol)], 12.5 mL of
2-methylbutan-2-ol, and 12.5 mL of 20% aqueous NaOH was
(13) (a) Herzog, S.; Dehnert, J . Z. Chem. 1964, 4, 1. (b) Shriver, D.
F.; Drezdzon, M. A. Manipulation of Air-Sensitive Compounds, 2nd
ed.; Wiley: New York, 1986. (c) Melting points of all phenylacetic acid
derivatives prepared in course of this work are reported in: Aldrich
Catalog Handbook of Fine Chemicals 1993-94; Alebrich: Milwaukee,
WI, 1993. (d) 1H-NMR and 13C-NMR spectra of all phenylacetic acid
derivatives are reported: Pouchert, C. J .; Behnke, J . The Aldrich
Library of 13C- and 1H-NMR Spectra; Alebrich: Milwaukee, WI, 1993.
Vol. 2. Reference nos.: Unsubstituted, 981A; 4-Me, 1010C; 3-Me,
1007A; 2-Me, 1001A; 2-F, 1002C; 4-Cl, 1014A; 3-Cl, 1007C; 2-Cl, 1003A;
4-Br, 1014B; 1-naphth, 1041A; 2-naphth, 1041B.
(11) Szabo´, P.; Marko´, L.; Bor, G. Chem. Techn. (Leipzig) 1961, 13,
549.
(12) (a) O¨ tvo¨s, I.; Sze´p, I.; Bikfalvi, J .; Pa´sztor, L.; Pa´lyi, G.
Elelmiszervizsgalati Kozl. (Food Anal. Proc.) 1976, 22, 142. (b) Vogel,
A. I. Practical Organic Chemistry; Longmans: London, 1956; pp 757,
761-762 (used also for preparation of the model sample).
+
+

3224 Organometallics, Vol. 15, No. 14, 1996
Zucchi et al.
stirred at room temperature while 0.2-0.5 mL/s CO gas was
bubbled into the liquid mixture. The dark brown-red mixture
turned light blue after 3-5 min, benzyl chloride (or bromide)
(1.73 mL, 15 mmol) was injected, and after being stirred for 5
min, the solution became deep brown-red again. Stirring,
under carbon monoxide, was continued for an additional 5 h
and the conversion determined by taking aliquots every 0.5
h and analyzing these by GC. After 5 h GC analysis showed
complete conversion of benzyl chloride and at the same time
the color of the reaction mixture changed again, now to lemon-
yellow. After the yellow color developed the reaction mixture
was no longer handled in an inert atmosphere. The reaction
mixture was transferred to a separatory funnel, the aqueous
and organic phases were separated, and the aqueous phase
was subsequently washed with 10 mL of benzene (or toluene)
and then acidified with 36% aqueous HCl (until pH ∼ 1 was
reached). During acidification a precipitate was formed, which
was separated by filtration or decantation. The (wet) precipi-
tate was then extracted with Et2O (3 × 10 mL), and the extract
was dried over anhydrous Na2SO4 or MgSO4. Finally the ether
was removed at reduced pressure and white crystalline phen-
of Et2O. Naphth-2-ylmethyl chloride (670 mg, 3.8 mmol) was
added at 0 °C to the filtered solution under Ar atmosphere.
The reaction mixture was stirred for 6 h at r.t. under Ar
atmosphere, and then the solution was cooled to ∼0 °C and
an Ar stream (∼80 mL/min) was bubbled through it for 5-6 h
until the solvent was completely evaporated. The remaining
dark brown oily product was extracted with 20 mL of n-
pentane, and the solution was filtered and analyzed by IR
spectroscopy. The IR ν(C-O) spectrum was similar to those
of (η3-benzyl)cobalt tricarbonyl compounds. Then the n-
pentane was evaporated with an Ar stream as described above
and a reddish-brown dense oil was obtained (yield 669 mg,
62%). This material could not be obtained in crystalline form
from various solvents and their mixtures at temperatures
ranging between 0 and 80 °C. Anal. Found: C, 59.3; H, 3.4;
Co, 20.5. Calcd for C14H9O3Co: C, 59.18; H, 3.19; Co, 20.74.
1H- and 13C-NMR and IR spectra are reported in Tables 2-6.
(η3-Na p h th -1-ylm eth yl)coba lt tr ica r bon yl was prepared
similarly, and the yield of the brownish-red oily product was
76%. Anal. Found: C, 59.2; H, 3.5; Co, 20.7. Calcd for
C
14H9O3Co: C, 59.18; H, 3.19; Co, 20.74. 1H- and 13C-NMR
1
ylacetic acid was obtained (identified by H-NMR (δ, DMSO-
and IR spectra are reported in Tables 2-6.
d6) 3.56 (s, 2H), 7.24 (s, 5H), and 10.5 (s, br, 1H) and mp 77.0
P r ep a r a tion of η3-(2-Meth ylben zyl)coba lt Tr ica r bon yl.
Na[Co(CO)4] (2 mmol) was prepared in 50 mL of Et2O and
cooled to 0 °C. To this solution, under Ar atmosphere, while
being stirred, 2-methylbenzyl bromide (348 mg, 1.9 mmol) was
added at once. The reaction mixture then left to warm to r.t.
after 30 min, and it was stirred for additional 5 h at r.t. Then
a sample was taken and analyzed by IR spectroscopy, which
showed the disappearance of the characteristic strong band
of [Co(CO)4]- and emergence of a new band system, which
could be assigned to a mixture of approximately 20% RC(O)-
Co(CO)4, 40% RCo(CO)4, and 40% η3-benzylCo(CO)3. Then the
reaction mixture was cooled to 0 °C and an Ar stream (50 mL/
min) was bubbled through it for 10 h, while the solvent was
completely evaporated. The brown residue was extracted by
2 × 10 mL of n-hexane, and the combined extracts were filtered
and chilled to -80 °C overnight. A dark brown microcrystal-
line substance was isolated which turned to be oily when it
was left to warm at ∼-60 °C; yield 0.15 g (32%). Anal.
Found: C, 53.6; H, 3.8; Co, 23.4. Calcd for C11H9O3Co: C,
53.25; H, 3.66; Co, 23.75. NMR and IR spectra are reported
in Tables 2-6.
°C (measured) lit. (76-77 °C12b)). Yield: 1.37 g (67%).
2. P h a se-Tr a n sfer Con d ition s. In these experiments the
phase-transfer agent (PEG) was placed in the reaction vessel
after Co2(CO)8 had been added. All other conditions were as
described above and in Table 1. The substituted arylacetic
1
acid products were identified by melting points13c and H- and
13C-NMR spectra.13d
In fr a r ed Sp ectr oscop ic In vestiga tion of th e Rea ction
Mixtu r es. Rea ction of Dicoba lt Octa ca r bon yl w ith So-
d iu m Hyd r oxid e. In a thermostated (25 °C) reaction vessel,
equipped with a magnetic stirrer, gas buret (paraffin oil),
injection neck and gas inlet/outlet tube, 170 mg (0.5 mmol) of
dicobalt octacarbonyl was dissolved in 10 mL of 2-methylbu-
tan-2-ol (CO atmosphere). The solution was stirred for 5 min,
10 mL of 20% aqueous NaOH was added, and stirring was
continued for ∼1.5 h. At the beginning 4.4 mL gas was
evolved, and then gas absorption was observed with the
volume change at the end being +2.1 mL. Samples were taken
for infrared analysis from the liquid and for GLC from the gas
phase. Infrared analysis showed [Co(CO)4]- (∼1900 cm-1, vs)
as the only cobalt carbonyl species present, distributed in a
70: 30 ratio between the aqueous and organic phase. GC
analysis showed 1.5% H2 and 98.5% CO in the gas phase, the
amount of H2 corresponding to 0.124 H atoms/Co atom.
P r ep a r a tion of (η1-(4-F lu or op h en yl)m eth yl)coba lt Tr i-
ca r bon yl. Na[Co(CO)4] (2 mmol) was prepared in 50 mL of
Et2O, and to this solution, 4-fluorobenzyl bromide (359 mg,
1.9 mmol) was added at once under Ar atmosphere at -20 °C.
This solution was stirred and left to warm to r.t. after 30 min.
Samples were taken after each 1 h, which were analyzed by
IR spectroscopy. A fine white precipitate (NaBr) was formed
and the solution became yellow after 2 h, when IR ν(CO)
spectra indicated the disappearence of the characteristic strong
absorption due to [Co(CO)4]- and a band system corresponding
mainly to RCo(CO)4 (together with some minor bands, due
most probably to an RC(O)Co(CO)4 type complex). Then the
solvent was evaporated at r.t. and reduced pressure. The
brown residue was extracted by 2 × 10 mL of n-hexane, and
the combined extracts were filtered and chilled to -80 °C. A
light yellow microcrystalline substance was separated, yield
0.24 g (46%). Anal. Found: C, 47.3; H, 2.5; Co, 20.9. Calcd
for C11H6O4CoF: C, 47.17; H, 2.16; Co, 21.04. NMR and IR
spectra are reported in Tables 2-6.
P r ep a r a tion of (η1-(2-Ch lor op h en yl)a cetyl)coba lt Tet-
r a ca r bon yl. Na[Co(CO)4] (4 mmol) was prepared in 60 mL
of Et2O, and to this solution 2-chlorobenzyl chloride (612 mg,
3.8 mmol) was added under CO atmosphere at 0 °C. This
solution was stirred and left to warm to r.t. after 30 min. Then
samples were taken after each 1 h, which were analyzed by
IR spectroscopy. IR ν(C-O) spectra indicated after 7 h the
disappearance of the strong band of [Co(CO)4]- and a band
system which corresponds approximately to the mixture of 20%
The aqueous phase was analyzed for Co(II) by adding 70%
aqueous HClO4 until pH ) 1, and the HCo(CO)4 formed from
the [Co(CO)4]- present was removed by bubbling CO through
the solution for 30 min. The Co(II) content was determined
by complexometry photometrically. Found: 4.5 mg of Co(II)
(7.6% with respect to the starting Co2(CO)8).
The above experiment was repeated except that, after the
addition of aqueous NaOH, the mixture was stirred for an
additional 5 min and then 1.73 mL (15 mmol) of benzyl
chloride was added. At the end of the reaction (end of gas
volume change) 1.37 g of PhCH2COOH was isolated (67%) and
31.5% unreacted PhCH2Cl was determined by GLC. The gas
volume changes and GLC analysis showed 10.35 mmol of CO
consumed (69-67% for PhCOOH and 2% for the formation of
carbonate, the latter corresponding to ∼30% of the reactant
Co atoms in Co2(CO)8). The final gas phase contained 7% H2,
which corresponds to 0.54 H atoms/Co atom.
In experiments investigating the fate of the catalyst during
the catalytic reactions (BP or PTC) an identical experimental
technique was used but samples were taken for IR analysis
after each 10-15 min.
Ch a r a cter iza tion of th e In ter m ed ia tes. P r ep a r a tion
of (η3-Na p h t h -2-ylm et h yl)coba lt Tr ica r b on yl. Co2(CO)8
(680 mg, 2 mmol) was transformed into Na[Co(CO)4] in 60 mL
+
+

Carbonylation of Benzyl Halides
Organometallics, Vol. 15, No. 14, 1996 3225
Ta ble 2. 1H-NMR Sp ectr a of Acyl-, Alk yl-, a n d (η3-Ben zyl)coba lt Ca r bon yls (C6D6, δ Va lu e Rela tive to
δTMS 0)
R
RC(O)Co(CO)4
RCo(CO)4
(η3-benzyl)Co(CO)3
PhCH2
7.19-6.97 (m, Ph); 3.92 (s, 2H, CH2) 7.19-6.97 (m, Ph); 3.16 (s, 2H, CH2) 6.65, 6.58, 6.48, 6.15, 6.05, 2.33, 2.12
p-MeC6H4CH2 7.16-6.93 (m, Ph); 3.93 (s, 2H, CH2); 7.1-6.9 (m, Ph); 3.23 (s, 2H, CH2);
6.54, 6.44, 6.19, 6.09 (ab, 27.65, 7.87 CH),
2.38, 1.81, 2.02 (s, CH3)
2.05 (s, 3H, Me)
m-MeC6H4CH2 7.05-7.00, 6.90-6.84 (m, Ph), 3.96
(s, 2H, CH2); 2.08 (s, 3H, Me)
2.13 (s, 3H, Me)
7.05-7.00, 6.90-6.84 (m, Ph); 3.20
(s, 2H, CH2); 2.13 (s, 3H, Me)
6.38, 6.29, 5.97, 5.88, 5.38, 2.57, 1.95
o-MeC6H4CH2 7.02-6.84 (m, Ph); 4.06 (s, 2H, CH2); 7.02-6.84 (m, Ph); 3.20 (s, 2H, CH2); 7.31 (dd, 6.86, 1.77 1H,CH), 6.66 (t, 7.07
2.04 (s, 3H, Me)
2.09 (s, 3H, Me)
1H,CH), 6.50 (d, 7.08 1H,CH), 5.02 (d,
6.64 1H,CH), 2.89, 1.61 (2d, 2.6 CH2)
1.82 (s CH3)
Me5C6CH2
4.43 (s,2H,CH2); 2.05, 1.99 (2s, 12H,
4CH3); 1.69 (s, 3H, CH3)
2.85; 2.02, 1.95 (2s, 12H, 4CH3); 1.81 (s,
3H, CH3)
p-BrC6H4CH2
7.20-7.11, 6.82-6.80, 6.60-6.55
(m, Ph); 3.72 (s, 2H, CH2)
7.06-6.88, 6.79-6.63 (m, Ph); 3.77
(s, 2H, CH2)
7.20-7.11, 6.82-6.80, 6.60-6.55
(m, Ph); 2.89 (s, 2H, CH2)
7.16, 7.98-6.71 (m, Ph); 2.94 (s,
2H, CH2)
a
b
p-ClC6H4CH2
a
a
a
b
m-ClC6H4CH2 7.16, 7.98-6.71 (m, Ph); 3.80 (s, 2H, 7.16, 7.98-6.71 (m, Ph); 2.90 (s,
CH2)
2H, CH2)
7.08-7.04, 6.84-6.70 (m, Ph); 3.21
(s, 2H, CH2)
7.0-6.9 (m, Ph); 3.0 (s, 2H, CH2)
7.08-7.06, 6.81-6.67 (m, Ph); 3.21
(s, 2H, CH2)
b
o-ClC6H4CH2
7.08-7.04, 6.84-6.70 (m, Ph); 4.19
(s, 2H, CH2)
7.08-7.06 (m, Ph); 3.8 (s, 2H, CH2)
7.08-7.06, 6.81-6.67 (m, Ph); 4.19
(s, 2H, CH2)
b
p-FC6H4CH2
o-FC6H4CH2
a
a
b
1-NaphtCH2
7.32-7.06 (m, 7H, CH); 4.41 (s, 2H,
CH2)
a
7.79 (dd, 8.4, 0.88), 7.58 (d, 7.96), 7.54
(d, 8.18), 7.32-7.06 m, 6.69 (d, 5.97),
6.66 (d, 5.98), 4.587 (d, 5.97), 3.68 (d,
2.43) 1.70 (dd, 2.43, 0.66)
2-NaphtCH2
a
7.49-7.16 (m, 7H, CH); 4.08 (s,
2H, CH2)
7.49-7.16, 6.9-6.58 m, 5.07 (d, 1.43),
3.02 (d, 1.67), 1.82 (d, 1.67)
a
Could not be sufficiently enriched in NMR solvent (probably because of insufficient solubility or transformation during the
b
measurement). Some unidentified bands were also observed.
Ta ble 3. 13C-NMR Sp ectr a of Acylcoba lt Ca r bon yls, RC(O)Co(CO)4 (C6D6 δ Va lu e Rela tive to δTMS 0)a
Cquat
CHar
R
COacyl
224.1
COcarb
196.9
CH2
69.0
CH3
PhCH2
ipso
133.4
p-
127.6
o-, m-
o-
m-
129.8, 128.8
129.8
129.6
130.5, 128.7
128.4, 126.9
130.7, 130.5
128.0, 126.3
133.3
p-MeC6H4CH2
m-MeC6H4CH2
o-MeC6H4CH2
Me5C6CH2
224.3
224.0
223.6
222.9
196.9
197.0
197.1
197.5
ipso
4-C
ipso
3-C
ipso
2-C
ipso
130.5
138.1
133.3
138.3
133.5
136.9
129.3
68.8
69.0
68.3
67.7
21.0
21.1
19.5
p-
o-, m-
18.2, 17.1
17.0, 16.8
16.7
132.4, 131.9
p-BrC6H4CH2
223.7
223.9
223.9
222.1
224.2
215.6
196.7
196.7
196.6
196.8
196.7
198.5
ipso
4-C
ipso
4-C
ipso
3-C
ipso
2-C
ipso
4-C
ipso
2-C
121.9
132.4
132.0
133.7
127.5
135.5
133.5
134.5
129.2 (3.0)
162.6 (246.1)
123.0 (16.0)
161.9 (245.8)
131.9
131.4
131.1
128.9
68.1
68.0
69.0
67.2
67.9
65.9
b
p-ClC6H4CH2
b
m-ClC6H4CH2
130.0, 129.9
127.8, 126.7
131.5, 129.6
129.1, 126.9
131.4 (8.0)
115.7 (12.1)
132.3 (4.2)
130.3 (7.6)
124.5 (2.5)
116.2 (21.7)
129.0 128.9
128.7 126.7
126.1 125.5
124.3
b
o-ClC6H4CH2
p-FC6H4CH2
o-
m-
4-C
5-C
6-C
3-C
b
o-FC6H4CH2
1-NaphtCH2
223.7
197.0
Cjunction
1-C
134.1
132.5
130.9
67.3
2-NaphtCH2
224.2
196.9
c
c
69.2
a
b
Tentative assignment. Some unidentified bands were also observed. c Could not be sufficiently enriched in NMR solvent (probably
because of insufficient solubility or transformation during the measurement).
RCo(CO)4 and 80% RC(O)Co(CO)4 compound. Then the sol-
vent was evaporated at r.t. and reduced pressure, and the
brownish residue was extracted by 20 mL of n-hexane. The
extract was filtered and then chilled to -30 °C. A light yellow
microcrystalline substance was obtained, yield 0.75 g (61%).
Anal. Found: C, 44.5; H, 2.1; Co, 17.8; Cl, 10.6. Calcd for
C12H6O5CoCl: C, 44.41; H, 1.86; Co, 18.16; Cl, 10.92. NMR
and IR spectra are reported in Tables 2-6.
Additional (η1-benzyl)cobalt tetracarbonyl, (η3-benzyl)cobalt
tricarbonyl, and (η1-phenylacetyl)cobalt tetracarbonyl deriva-
tives, with different substituents in the phenyl ring were
prepared by essentially similar methods as described above.
+
+

3226 Organometallics, Vol. 15, No. 14, 1996
Zucchi et al.
Ta ble 4. 13C-NMR Sp ectr a of Alk ylcoba lt Ca r bon yls RCo(CO)4 (C6D6 δ Va lu e Rela tive to δTMS 0)2
Cquat
147.7
CHar
R
COcarb
199.2
CH2
19.9
CH3
PhCH2
ipso
p-
126.1
o-, m-
128.9, 128.6
p-MeC6H4CH2
m-MeC6H4CH2
o-MeC6H4CH2
p-BrC6H4CH2
201.2
199.3
199.3
198.8
198.9
198.7
199.0
199.2
204.4
ipso
4-C
ipso
3-C
ipso
2-C
ipso
4-C
ipso
4-C
ipso
3-C
ipso
2-C
ipso
4-C
ipso
2-C
137.4
b
b
20.4
20.1
16.1
18.5
18.5
17.9
14.9
18.8
15.4
20.9
21.3
18.5
138.2
147.5
134.6
146.5
119.7
146.7
134.6
146.3
134.6
139.7
131.0
146.0
b
129.4, 128.8
126.4, 125.8
142.0, 131.1
127.7, 126.9
132.0
130.1
130.0
128.5
130.1, 128.5
126.5, 126.0
130.5, 130.3
129.9, 127.1
130.1 (7.8)
115.8 (21.5)
130.6 (4.0)
127.7 (8.0)
124.6 (3.2)
115.9 (21.7)
128.9, 128.6, 127.6
127.3, 126.9, 126.5
125.2
c
p-ClC6H4CH2
c
m-ClC6H4CH2
c
o-ClC6H4CH2
p-FC6H4CH2
o-
m-
4-C
5-C
6-C
3-C
c
o-FC6H4CH2
115.5 (14.0)
160.3 (246.6)
2-NaphtCH2
201.2
Cjunction
1-C
133.9
131.0
133.1
33.8
a
b
Tentative assignment. Could not be sufficiently enriched in NMR solvent (probably because of insufficient solubility or unfavorable
effect on the equilibria) to obtain (additional) assignable bands. c Some unidentified bands were also observed.
Ta ble 5. 13C-NMR Sp ectr a of (η3-Ben zyl)coba lt Ca r bon yls RCo(CO)3 (C6D6 δ Va lu e Rela tive to δTMS 0)a
Cquat
CH
R
COcarb
203.8
CH2
35.6
CH3
PhCH2
100.2
97.2
98.7
97.7
138.9
126.9
140.7
b
131.1
126.4
129.6
120.1
127.0
128.9
122.3
129.6
p-MeC6H4CH2
o-MeC6H4CH2
Me5C6CH2
203.9
203.9
204.5
35.2
33.9
34.6
21.6
130.8
99.8
16.1
134.3, 131.3
32.0, 25.8
23.1, 17.3
14.3
136.8b
1-NaphtCH2
2-NaphtCH2
203.5
203.8
Cjunction
134.0
132.4
95.2
135.4
132.9
95.9
129.7
128.4
80.8
129.7
126.7
80.8
129.3
127.3
129.0
121.4
35.2
40.2
1-C
Cjunction
128.8
b
127.6
2-C
a
-
b
Tentative assignment. m-MeC6H4CH2 decomposed during NMR measurement. Could not be sufficiently enriched in this solvent
(probably because of insufficient solubility or unfavorable effect of solvent on the equilibria) to obtain (additional) assignable bands.
Experimental conditions in preparations of the model com-
plexes varied as follows.
According to IR analysis of the reaction mixtures (substi-
tuted) benzyl chlorides need 3-10 h for (practically) quantita-
tive conversion of the [Co(CO)4]- anion, while benzyl bromides
need 1-5 h reaction times.
(even using CO atmosphere), while (ii) at the latter under CO
the acyl complex was predominant; under Ar atmosphere the
η3-benzyl derivative was predominant with almost no alkyl
complex in both cases.
Spectroscopic (1H- and 13C-NMR and IR) characterization
of the model complexes is reported in Tables 2-6.
Typical data are as follows: PhCH2Cl, 4-5 h; 4-MeC6H4-
CH2Cl, 3-4 h; 1-C10H7CH2Cl, 4-5 h; 2-ClC6H4CH2Cl, 6-8 h;
PhCH2Br, 2 h; 2-MeC6H4CH2Br, 1-2 h; 2-FC6H4CH2Br, 2-3
h.
Solven t Effect Stu d ies. In a Schlenk vessel under CO or
Ar, Na[Co(CO)4] (2 mmol) was prepared in 30 mL of Et2O and
reacted with 1.9 mmol of (substituted) benzyl chloride (or
bromide) as described earlier in this paper and ref 9c. After
the reaction was completed (monitored by IR) Et2O was
evaporated at r.t. and reduced pressure and 30 mL of a solvent
was added. The solution was stirred under CO at r.t. until
equilibrium was reached (30-120 min) (monitored by IR), and
then the IR spectrum was registered with an expanded
wavenumber scale.
Mod el Exp er im en ts for Testin g th e In ter m ed ia cy of
(η3-Ben zyl)coba lt Com p lexes in th e Ca ta lytic Cycle.30 (η3-
Naphth-1-ylmethyl)cobalt tricarbonyl (1.14 g, 4 mmol) (pre-
pared as described earlier in this paper) was dissolved in 12
mL of toluene. This solution was then divided into four equal
parts (A-D) and treated according to the methods of the PTC
carbonylation experiments as follows. To portion A (3 mL) 3
Yields (ratios) of the cobalt carbonyl complexes depend
strongly on the substituents, the acyl- and the η3-benzyl
complexes being favored with electron donor substituents,
while the acyl and alkyl derivatives are favored with electron
acceptor (F, Cl) substituents. Typical compositions of the
reaction mixtures at the end of the reaction are for the RC-
(O)Co(CO)4/RCo(CO)4/η3-benzylCo(CO)3 ratio as follows: PhCH2
(CO, 80/20/0; Ar, 70/25/5); 2-MeC6H4CH2 (CO: 70/25/5);
3-ClC6H4CH2 (CO, 60/40/0; Ar, 5/70/25); Me5C6CH2 (CO, 80/
20/0; Ar, 40/5/55) (approximate IR data).
The naphthylmethyl and pentamethylbenzyl derivatives
represent particular cases: (i) At the former a characteristic
predominance of the η3-benzyl derivative (>80%) was observed
+
+

Carbonylation of Benzyl Halides
Organometallics, Vol. 15, No. 14, 1996 3227
the use of a corrosive promoter (HI)3t or halogenated
solvent (CH2Cl2).3v,15
We regarded two procedures (ester, ref 3k; acid, ref.
4c) as standards. Our principal aim was to develop
more economical techniques.
We usedsif necessary at allsonly polyethylene glycols
as PT agents for two reasons. First, these are ap-
preciably cheaper than “quats” or crown ethers, and
second, the quaternary ammonium salts may suffer
irreversible degradation under the conditions of cata-
lytic carbonylation, as was demonstrated earlier by one
of us.16
Essentially three procedures were found to be suitable
to prepare phenylacetic acids from the corresponding
benzyl chlorides or bromides.
(a) A PTC system using the 2-methylbutan-2-ol/water
solvent pair provided the best isolated yield for PhCH2-
COOH and acceptable yields for some substituted
derivatives. An alternative to this method proved to be
4-methylpentan-2-one/water in some cases, but the
available data are insufficient to interpret the observed
differences. The BP variant of the tert-amyl alcohol
containing system provides considerably lower yields,
but the reaction takes place. Not even traces of phen-
ylacetic ester and phenylpyruvic acid (or derivatives)4b,d
were detected in these experiments.
Ta ble 6. F T-IR Sp ectr a of Acyl-, Alk yl-, a n d
(η3-Ben zyl)coba lt Ca r bon yls (n -Hexa n e, cm -1
)
R
RC(O)Co(CO)4 RCo(CO)4 (η3-benzyl)Co(CO)3
Me5C6CH2
2104 m
2045 s
2022 vs
2013 s
2065 m
2002 vs
1980 s
1966 s
a
1722 m
2107 m
2020 vs, br
1717 m
2107 m
2042 s
p-ClC6H4CH2
m-ClC6H4CH2
2099 m
2035 s
2013 vs, br
2101 m
2036 s
2054 vs
∼1990 s
1972 s
2063 s
a
2027 vs
1714 m
2107 m
2049 s
2026 vs
2012 vs
1712 w
2108 m
2049 s
2018 vs
o-ClC6H4CH2
o-FC6H4CH2
2100 m
2036 s
2023 vs
2012 vs
2062 s
2001 s
1987 s
2101 m
2036 s
2023 vs
2013 s
2063 s
a
2028 vs
2013 s
1713 w
1-NaphthCH2
2-NaphthCH2
2058 s
1999 s
1988 vs
2058 s
2000 vs
1986 vs
a
a
2106 m
a
2098 m
2033 m
2015 vs
(b) Seeking a more practical variant of these proce-
dures, we found that fusel oil is also a suitable solvent
both under BP or PTC reaction conditions. Fusel oils
of various origins were tested, but no remarkable
differences were observed with these distillates having
only slightly different compositions. Good yields and
turnover numbers could be achieved. There was practi-
cally no difference between the BP and PTC variant in
these solventssi.e. the PT agent had no detectable effect
on the yields. The selectivity was found to be quantita-
tive also in this case.
(c) A third variant, on the other hand, provided a
dramatic demonstration of the effect of the PT agent.
Using only the benzyl chloride/water system (i.e. no
organic “solvent”), we found that in the absence of PEG
no reaction occurs, while if PEG was added, acceptable
yields of phenylacetic acid could be achieved.
a
Could not be sufficiently enriched in this solvent (probably
because of insufficient solubility or unfavorable effect of solvent
on the equilibria) to obtain (additional) evaluable bonds.
mL of 20% aqueous NaOH solution and 60 µL of PEG 400 were
added and stirred (for 20 h) until it became yellow; then it
was treated as the samples in the carbonylation experiments.
To portion B 3 mL of 20% aqueous NaOH solution, 60 µL of
PEG 400, and Co2(CO)8 (17 mg, 0.05 mmol) were added, and
then the sample was treated as portion A.
To portion C was added 1-(choromethyl)naphthalene (5.28
g, 30 mmol), 3 mL of 20% aqueous NaOH solution, and PEG
400 (60 µL). Subsequently the sample was treated as portion
A.
To portion D was added 1-(choromethyl)naphthalene (5.28
g, 30 mmol), 3 mL of 20% aqueous NaOH solution, Co2(CO)8
(17 mg, 0.05 mmol), and PEG 400 (60 µL). Then the solution
was treated as portion A.
Experiments A, C, and D were repeated using (η3-naphth-
2-ylmethyl)cobalt tricarbonyl and 2-(bromomethyl)naphtha-
lene.
Mech a n istic Stu d ies. We report in this section on
the results of experiments (mainly monitored by IR
spectroscopy) concerning (i) the fate of the catalyst
precursor (Co2(CO)8), (ii) the identification of the orga-
nocobalt complexes formed and their distribution in the
liquid phases, (iii) the influence of the nature of the
organic phase on the equilibrium between the organo-
cobalt complexes and the role of the phase-transfer
agent, and (iv) the influence of the substituent on the
aromatic ring on the equilibrium distribution of the
organocobalt compounds. Since the best quality IR
spectra could be obtained in t-AmOH, most of the
Yields of 1-naphthylacetic acid (2-naphthylacetic acid in
parentheses): A, 8.9% (18.6); B, 54%; C, 91.6% (55%); D, 93.4%
(77.4%). (Yields of experiments C and D were calculated with
respect to the sum of the starting η3-complex and naphthyl-
methyl halide.)
Resu lts a n d Discu ssion
P r ep a r a tion of Ar yla cetic Acid s. Excellent results
were achieved in previous investigations in the prepara-
tion of phenylacetic acids3g,r,v,z,aa,4a,c and esters3k,l,n,s,t
from benzyl halides. From the viewpoint of eventual
practical realization, however, most of these methods
have some drawbacks: (i) medium to high pressure (5-
150 bar),3n,s,z,aa,4a (ii) a multicomponent catalyst “system”
which appears to create handling and waste (recycling)
problems,3g,l (iii) insufficient selectivity,3rzaa,14 and (iv)
(15) Low molecular weight halocarbons should be avoided in light
of recent environmental results and regulations. Some leading refer-
ences are as follows: (a) Fabian, P.; Go¨nner, D. Fresenius Z. Anal.
Chem 1983, 319, 890. (b) Grob, K.; Habich, A. J . High Resolut.
Chromatogr.-Chromatogr. Commun. 1983, 6, 11. (c) Grob, K.; Grob,
G. J . High Resolut. Chromatogr.-Chromatogr. Commun. 1983, 6, 133.
(d) Grob, K. J . Chromatogr. 1984, 229, 1. (e) Nunez, A. J .; Gonzales,
L. F.; J ana´k, J . J . Chromatogr. 1984, 300, 127. (f) Ettre, L. S.; Di
Cesare, J . L. Am. Laboratory 1984, 16, 76. (g) Ettre, L. S. Anal. Chem.
1985, 57, 1419A. (h) Croll, B. T.; Sumner, M. E.; Leathard, D. A.
Analyst 1986, 111, 73. (i) Lakszner, K.; Szepesy, L. J . High Resolut.
Chromatogr.-Chromatogr. Commun. 1986, 9, 441.
(14) With regard to preparation of phenylpyruvic acid or its deriva-
tives.
(16) Gambarotta, S.; Alper, H. J . Organomet. Chem. 1980, 194, C19.
+
+

3228 Organometallics, Vol. 15, No. 14, 1996
Zucchi et al.
Sch em e 1
substantial fraction of the latter. The Co2(CO)8 formed
in this manner is then transformed as indicated in
Scheme 1.
It should be pointed out that the essential part of
Scheme 1 together with our finding regarding the H2
evolution can also be formulated as follows:
CO + 2OH- + CO32- + H2
This formulation indicates that in biphasic or PTC
Co2(CO)8/NaOH systems a kind of cobalt-assisted water
gas shift reaction (WGSR)19f-h occurs. This point is now
currently investigated in our laboratories.
Infrared spectroscopic studies of the reaction mixtures
of t-AmOH and H2O and (in a few cases) fusel oil/H2O
systems revealed the presence of dicobalt octacarbonyl
(diminishing) and [Co(CO)4]- (increasing) as the only
cobalt carbonyls in the system in the absence of benzyl
chloride.
The bands of dicobalt octacarbonyl disappeared after
about 5-10 min. This was accompanied by the disap-
pearance of the brown color. At this stage the distri-
bution20a of the tetracarbonylcobaltate anion between
the organic and aqueous phase showed marked differ-
ences depending on the presence or absence of the
phase-transfer agent:
spectroscopic studies were made in the system with this
alcohol as the organic phase. The identity of the
organocobalt intermediates was ensured by 1H- and 13C-
NMR and IR spectra (Tables 2-6) and by analogy of
these spectra to those of isolated (benzyl- and (phenyl-
acetyl)cobalt carbonyls where an X-ray diffraction struc-
ture could be obtained (c.f. Experimental Section and
refs 9b,c, 18c, 21j, and 22).
aqueous phase
organic phase
t-AmOH
t-AmOH/PEG
70%
0%
30%
100%
F a te of th e Ca ta lyst P r ecu r sor . It is generally
believed that in phase-transfer or biphasic systems
dicobalt octacarbonyl undergoes a complicated reaction
depicted in Scheme 1.2dg,17-19 One of the most interest-
ing points of this mechanism is the intermediacy of the
hydroxycarbonylcobalt tetracarbonyl HOC(O)Co(CO)4.
Esters of this compound are known,18 and decarboxy-
lation of one of the esters was recently demonstrated.18f,h
In the present work we controlled the stoichiometry, and
in addition to earlier findings, we detected the evolution
of H2 even at an early stage of the reaction. This can
be regarded18h,19a-e as indirect evidence for the forma-
tion and decomposition (via decarboxylation) of HOC-
(O)Co(CO)4 in the apolar phase according to Scheme 1.
The dramatic increase of the quantity of H2 evolved
during the carbonylation of benzyl chloride may be
explained by assuming a dicobalt octacarbonyl-based
catalytic process. This may result in the reaction
between PhCH2Co(CO)4 and HOC(O)Co(CO)4, initiated
by a radical fission19b-d of the former consuming a
This effect of PEG on the distribution of [Co(CO)4]-
in the t-AmOH/H2O system and its influence on the
yield clearly show that this anion “is needed” in the
organic phase for the hydroxycarbonylation reaction.
One can propose that the “organometallic part of the
catalytic cycle” takes place exclusively in the organic
phase.20b The fact that [Co(CO)4]- is the only organo-
cobalt species present before the benzyl halide is added
supports the earlier supposition2g,10e,g that this anion
is the entry point into the catalytic cycle.
Or ga n ocoba lt Com p ou n d s in th e Rea ction Mix-
tu r es of Hyd r oxyca r bon yla tion . The tetracarbon-
ylcobaltate anion is known to react with benzyl halides
in Et2O when mixtures of (benzyl- and (phenylacetyl)-
cobalt tetracarbonyls and (η3-benzyl)cobalt tricarbonyls
are formed.3a,9a-d,10a As intermediates of this reaction
[PhCH2C(O)Co(CO)3(X)]- (X ) Cl, Br, I) complexes were
identified.18f,21
Several η1-benzyl (1) and η3-benzyl (3) as well as η1-
phenylacetyl (2) cobalt carbonyls were characterized by
IR spectroscopy (unsubstituted, 4-Me, 3-Me, 2-Me,
(17) The stoichiometry was originally established by Hieber at al.:
(a) Hieber, W.; Schulten, H. Z. Anorg. Allg. Chem. 1937, 232, 18, 20.
(b) Hieber, W.; Abeck, W.; Sedlmeier, J .; Abeck, W. Chem. Ber. 1953,
86, 700. (c) Wender, I.; Stenberg, H. W.; Orchin, M. J . Am. Chem. Soc.
1952, 74, 1216.
(18) (a) Heck, R. F. J . Organomet. Chem. 1964, 2, 195. (b) Hieber,
W.; Duchatsch, H. Chem. Ber. 1965, 98, 1744. (c) Milstein, D.; Huckaby,
J . L. J . Am. Chem. Soc. 1982, 104, 6150. (d) Ungva´ry, F.; Marko´, L.
Organometallics 1983, 2, 1608. (e) Tasi, M.; Sisak, A.; Ungva´ry, F.;
Pa´lyi, G. Monatsch. Chem. 1985, 116, 1103. (f) Tasi, M.; Pa´lyi, G.
Organometallics 1985, 4, 1523. (g) Werner, H.; Hofmann, L.; Zolk, R.
Chem. Ber. 1987, 120, 379. (h) Bartik, T.; Kru¨mmling, T.; Marko´, L.;
Pa´lyi, G. Gazz. Chim. Ital. 1989, 119, 307. (i) Bartik, T.; Kru¨mmling,
T.; Kru¨ger, C.; Marko´, L.; Boese, R.; Schmid, G.; Vivarelli, P.; Pa´lyi,
G. J . Organomet. Chem. 1991, 421, 323.
(19) (a) Ungva´ry, F.; Marko´, L. J . Organomet. Chem. 1980, 193, 397.
(b) Orchin, M. Acc. Chem. Res. 1981, 14, 59. (c) Galamb, V.; Pa´lyi, G.
Atti Accad. Sci. Bologna, Cl. Sci. Fis. 1982, 270 (13/9), 157. (d) Pa´lyi,
G.; Ungva´ry, F.; Galamb, V.; Marko´, L. Coord. Chem. Rev. 1984, 53,
37. (e) Murai, S.; Sonoda, N. Angew. Chem. 1979, 91, 896. (f) Roofer-
De Poorter, C. K. Chem. Rev. 1981, 81, 447. (g) Ford, P. C. Acc. Chem.
Res. 1981, 14, 31. (h) Henrici-Olive´, G.; Olive´, S. The Chemistry of the
Catalysed Hydrogenation of Carbon Monoxide; Springer: Berlin, 1984.
(20) (a) Measured on the basis of the area of the ν(C-O) band. This
assumes that the absorption coefficients were the same in different
solvents. For this reason our data are only of qualitative character.
(b) If anionic cobalt carbonyl species other than [Co(CO)4]- would be
formed in the catalytic cycle, these should also be concentrated in the
organic phase.
(21) Analogous complexes were reported. [XYCo(CO)3]- type: (a)
Francalanci, F.; Gargano, A.; Abis, L.; Foa`, M. J . Organomet. Chem.
1983, 251, C5. (b) Anderson, F. R.; Wrighton, M. S. J . Am. Chem. Soc.
1984, 106, 995. (c) Vlaic, G.; Bart, J . C. J .; Foa`, M.; Francalanci, F.;
Clement, R. J . Organomet. Chem. 1985, 287, 369. (d) See also refs 9d
and 18e,f. (e) Ro¨per, M.; Schieren, M.; Heaton, B. T. J . Organomet.
Chem. 1986, 299, 131. (f) Galamb, V.; Pa´lyi, G.; Boese, R.; Schmid, G.
Organometallics 1987, 6, 861. (g) Ro¨per, M.; Kru¨ger, C. J . Organomet.
Chem. 1988, 339, 159. (h) Fachinetti, G.; Funaioli, T.; Masi, D.; Mealli,
C. J . Organomet. Chem. 1991, 417, C32. (η3-benzyl)Co(CO)3 type: (i)
Ullah, S. S.; Molla, M. E.; Karim, M. M.; Begum, R. Indian J . Chem.
1985, 24A, 412. (j) Blecke, J . R.; Burch, R. R.; Caulman, C. R.; Schardt,
B. C. Inorg. Chem. 1981, 20, 1316.
+
+

Carbonylation of Benzyl Halides
Organometallics, Vol. 15, No. 14, 1996 3229
4-MeO, 3,4-(MeO)2, 4-Cl, 2,6-Cl2)9c,10a,21i and the molec-
ular structures of some similar complexes were deter-
mined by X-ray diffraction,9bc,18c,21fj,22 but only a few
1
compounds were studied by H-NMR9,21i spectroscopy.
1
We summarize, in Tables 2-6, most of the H- and
13C-NMR data and the IR ν(C-O) spectra not studied
earlier. These characteristics, considering the similarity
to the spectra of compounds published earlier, and those
also obtained by X-ray diffraction9b,c,18c,21f,j,22 represent
a reliable basis for the identity of 1-3.
An infrared study of the organic and aqueous phase
of the reaction mixtures in the hydroxycarbonylation
experiments revealed the presence of these complexes
but with characteristic variations of the relative con-
centrations. The alcoholic phase, without a PT agent,
shows a high relative concentration of the (phenylace-
tyl)cobalt complex, a much lower concentration of the
alkyl derivative, and very low quantities of the η3-benzyl
complex. (This can also be expected on basis of the
polarity of the solvent, as will be demonstrated later.)
In these biphasic reactions, as hydroxycarbonylation
proceeds, the relative amount of the η3-benzyl complex
increases. The end of the reaction is characterized by
a change in color (see Experimental Section), disap-
pearance of the benzyl chloride (GC), termination of CO
absorption, and disappearance of the bands of all
neutral cobalt carbonyls (including those of (η3-benzyl)-
cobalt tricarbonyl).
The influence of the PT agent is convincingly dem-
onstrated by the IR spectra. In the case of the t-AmOH/
PEG/H2O system, (η3-benzyl)cobalt tricarbonyl is the
only detectable neutral cobalt carbonyl during hydroxy-
carbonylation. If the reaction is run in a biphasic
manner and PEG is added later, this change is observed
immediately after the addition of the PT agent.
The intermediacy of the (η3-benzyl)cobalt tricarbonyl
complexes in the catalytic cycle(s) was demonstrated
also by model experiments23 (cf. Experimental Section).
These experiments led to the following two important
conclusions: (i) the (η3-benzyl)cobalt tricarbonyl com-
plexes show catalytic activity in the carbonylation of the
corresponding benzyl halides; (ii) this catalytic activity
increase in the presence of dicobalt octacarbonyl show-
ing a typical synergetic effect.
These observations allow the following conclusions.
(i) If the final step of hydroxycarbonylation is acceler-
ated, (η3-benzyl)cobalt tricarbonyl type complexes are
accumulated in the reaction mixture (this effect becomes
more important when the η3-benzyl complex is more
stable, as in the case of the naphthyl derivatives). (ii)
In order to explain this phenomenon we have to assume
that the (η3-benzyl)cobalt tricarbonyl complexes are the
true intermediates of the catalytic cycle, not only
representing a “cobalt reservoir” which is present
because of the equilibrium between the η1- and η3-benzyl
species. We suggest therefore that the η3-benzyl com-
plexes may be formed from [PhCH2C(O)Co(CO)3X]- via
decarbonylation to [PhCH2Co(CO)3X]-, followed by loss
of halide. This pathway thus constitutes an additional
loop of the mechanism proposed by des Abbayes.2g (iii)
The effect of the PT agent can be interpreted either in
terms of the des Abbayes hypothesis,2g supposing that
F igu r e 1. Infrared ν(CO) spectra of equilibrium mixtures
of (η1-phenylacetyl)Co(CO)4, (η1-benzyl)Co(CO)4, and (η3-
benzyl)Co(CO)3 in different solvents.
the presence of PEG decreases the interfacial term of
the activation energy for nucleophilic attack of the OH-
ion on the acylcobalt complex at the interface or,
alternatively, that PEG increases the concentration of
NaOH in the organic phase. The first possibility should
be regarded as being more probable since our experi-
ments showed no characteristic change in the partition
of NaOH upon addition of PEG.
Perhaps the most important conclusion of the infrared
investigations of the hydroxycarbonylation reaction
mixtures is that the role of the PT agent is not only to
change the partition of the [Co(CO)4]- anion10e between
the two phases but also to promote a subsequent step
of the catalytic cycle: the cleavage of the acyl complex
by hydroxide ion (most probably by increasing the
reactivity of OH- through coordinating to the Na+
cation20a,d). It remains to be demonstrated whether the
reaction pathway found for t-AmOH/PEG/H2O can be
generalized to other Co2(CO)8-PTC systems.
Solven t a n d Su bstitu en t Effects. The kinetics of
the reaction between benzyl chloride and Na[Co(CO)4]
in various solvents has been investigated by Moro et
al.10d These authors found that the rate of this reaction
decreases in the order anisole > i-Pr2O > n-Pr2O > THF
> AN > DMF. This means that increasing polarity of
the solvent disfavors the activation step of the catalytic
cycle.
Since the polarity of the (η1-benzyl-, (η3-benzyl-, and
(η1-phenylacetyl)cobalt complexes is presumed to be
different, a marked solvent effect is also expected on
the equilibria between these compounds.
In fact it was found that increasing the polarity of
the solvent favors the formation of the acyl complex
(Figure 1). Since this complex is believed to be the
immediate precursor of the end product, more polar
solvents should be preferred. We observed, however,
that solvents of high polarity like DMF or DMSO gave
undesirable side reactions (a kind of “base reaction”
leading to Co(II) and [Co(CO)4]-) with the neutral
(22) Pa´lyi, G.; Zucchi, C.; Bartik, T.; Herbrich, T.; Kriebel, C.; Boese,
R.; Sorkau, A.; Fra´ter, G. Atti Accad. Sci. Bologna, Rend. Cl. Sci. Fis.
1992/93, 281 (14/10), 159.
(23) The concept of these experiments was suggested by one of the
reviewers of this paper. The authors acknowledge gratefully this
suggestion.
+
+

3230 Organometallics, Vol. 15, No. 14, 1996
Zucchi et al.
Ch a r t 1
intermediates. Dioxane and THF were found to favor
pyruvic acid formation in a similar system.3aa Thus a
compromise had to be found and it turned out that
t-AmOH (or components of fusel oil or the branched
ketone used by us) represents just such an advanta-
geous intermediate case of being not too reactive yet
sufficiently polar. A further practical advantage of
these alcoholic systems is that the chemoselectivity is
quantitative: the formation of the corresponding esters
was never observed.
One can speculate whether the (η3-benzyl)cobalt tri-
carbonyls in (even “hard”) donor-type solvents are
present as true η3-benzyl complexes or as (η1-benzyl)-
cobalt tricarbonyl solvent PhCH2Co(CO)3(S) adducts,24-26
since the IR band pattern (A1 + Esthe latter broadened)
would be the same for both, and the observed band
positions are between those of (η3-benzyl)cobalt tricar-
bonyl (in an apolar solvent) and (η1-benzyl)cobalt tri-
carbonyl triphenylphosphine adducts. However at-
tempts to isolate such a solvent adduct failed, but the
smooth and almost quantitative formation of η1-PhRCH2-
Co(CO)3PPh3 complexes from the corresponding η3-PhR-
CH2Co(CO)3 derivatives and PPh3 (cf. ref 9c) supports
this view.
Sch em e 2
It is evident that a considerable substituent effect has
to be expected in the hydroxycarbonylation reaction.
Kinetic studies of the substituent effects in related
systems were performed on the rates of activation of
benzyl chloride10d and bromide10g by the [Co(CO)4]-
anion. Only a very slight substituent effect was de-
tected in the former case, which was interpreted in
terms of interesting mechanistic speculations which will
be discussed later. Our results (Table 1) showed that
the benzyl halides containing electron acceptor substit-
uents (F, Cl) provided relatively low yields of the
hydroxycarboxylated product.
The negative slope of the K vs σp plot of the alkyl/
acyl equilibrium9b provides independent support for the
presently accepted mechanism of CO insertion according
to which “insertion” should be regarded as an intramo-
lecular nucleophilic attack of the alkyl group at one of
the coordinated carbonyl groups.9b,10a,24 On the other
hand, the positive slope of the equilibrium constant of
the η1-alkyl h η3-benzyl equilibrium9b indicates that (η3-
benzyl)cobalt tricarbonyl is an electrophilic center sus-
ceptible to nucleophilic attack by CO.25 One can also
rationalize these findings in terms of the acceptor
substituent on the ring destabilizing η3-coordination and
thus facilitating attack (addition) of the “fourth” CO.
The opposite sense of the two subsituent effects
explains, in part, the preparative problems encountered9
in efforts to enrich reaction mixtures in some individual
complexes of these equilibrium mixtures (for preparative
purposes) and also the compromises which have to be
worked out frequently in carbonylation studies.
A special kind of substituent effect was observed with
1-chloro- and 2-(bromomethyl)naphthalene. The reac-
tions of these compounds with Na[Co(CO)4] provide the
usual acyl/alkyl/η3-benzyl mixture, but here the equi-
librium lies so much on the η3-benzyl side that in an
argon atmosphere the cobalt tetracarbonyl complexes
almost disappear and under 1 bar CO the η3-benzyl-
type complex still predominates.
The high relative stability of the (η3-methylenenaph-
thalene)cobalt tricarbonyls has some interesting conse-
quences: (i) The lower yield in hydroxycarbonylation
of 1-(chloromethyl)naphthalene may principally be due
to the stability of the η3-complex. (ii) The preferential
formation of these compounds is apparently a conse-
quence of the more facile dearomatization of the sub-
stituted ring in naphthalene as compared to the corre-
sponding benzene or thiophene derivatives.27,28 (iii) The
structures of these complexes29 show that the process
(26) In agreement with the supposition of an (η3-benzyl)Co(CO)3 h
(η1-benzyl)Co(CO)3 equilibrium; cf. also ref 23.
(27) (a) King, R. B.; Fronzaglia, A. J . Am. Chem. Soc. 1966, 88, 709.
(b) King, R. B.; Kapoor, R. N. Inorg. Chem. 1969, 8, 2535. (c) Cotton,
F. A.; Marks, T. J . J . Am. Chem. Soc. 1969, 91, 1339. (d) Muetterties,
E. L.; Hirsekorn, F. J . J . Am. Chem. Soc. 1977, 99, 2509. (f) Blecke, J .
R.; Burch, R. R.; Caulman, C. L.; Schardt, B. C. Inorg. Chem. 1981,
20, 1316.
(28) See for example: Coulson, C. A.; O’Leary, B.; Mallison, R. B.
Hu¨ckel Theory for Organic Chemistry; Academic Press: London, 1978;
pp 44, 151-155.
(29) See the Experimental Section of this work and: Sa´mpa´r-
Szerencse´s, E.; Thiele, K. H.; Pa´lyi, G. Unpublished results.
(24) We cite here only some early key papers on this point: (a) Heck,
R. F.; Breslow, D. S. J . Am. Chem. Soc. 1961, 83, 4024. (b) Marko´, L.
Chem. Ind. (London) 1962, 260. (c) Nitschmann, E. R. Chem. Ber. 1964,
97, 2098. Two excellent, recent spectroscopic studies provide indepen-
dent experimental support for the existence of RCo(CO)3 and RC(O)-
Co(CO)3: (d) Roe, D. C. Organometallics 1987, 6, 942. (e) Sweany, R.
L. Organometallics 1989, 8, 175.
(25) (a) Calderazzo, F. Angew. Chem. 1977, 89, 305. (b) Kuhlmann,
E. J .; Alexander, J . J . Coord. Chem. Rev. 1980, 33, 195. (c) Berke, H.;
Hoffmann, R. J . Am. Chem. Soc. 1978, 100, 7224.
+
+

Carbonylation of Benzyl Halides
Organometallics, Vol. 15, No. 14, 1996 3231
is regiospecific (Chart 1B,D); isomers involving the loss
of the aromatic character of both rings (Chart 1A,C) are
obviously less favored.
(which undergo dramatic kinetic enrichment in certain
cases) and their role as intermediates in phase-transfer
reactions was not previously observed.
Con clu d in g Rem a r k s on th e Mech a n ism . This
study accounts for most details of the t-AmOH/PEG/H2O
system. It can reasonably be assumed that most of the
conclusions can be generalized to all biphasic or PTC
carbonylations of benzyl halide type substrates. The
main mechanistic features, summarized in Scheme 2,
lead to the following conclusions:
(i) The role of PEG is not simply to transport
[Co(CO)4]- from the aqueous to the organic phase
(similarly as the “quats” were found to act in earlier
studies2g,4e,g) but also to accelerate the hydrolysis of the
acylcobalt tetracarbonyl complex.
(ii) We demonstrated experimentally the presence of
the acyl-, alkyl- and (η3-benzyl)cobalt complexes in the
hydroxycarbonylation reaction mixtures. The acyl and
alkyl complexes have been known2,10,30 for some time,
but the presence of (η3-benzyl)cobalt-type complexes
Ack n ow led gm en t. The authors acknowledge help-
ful discussions with Drs. T. U. Ka´llay (Budapest), F.
Kova´cs (Va´rpalota), and M. Musulin (Veszpre´m), com-
munication of unpublished experimental data by Prof.
R. L. Sweany (New Orleans, LA), and help in obtaining
the GC and IR measurements by Drs. I. O¨ tvo¨s and K.
Nagy-Perge. The University of Ottawa scientists are
grateful to the Natural Sciences and Engineering Re-
search Council of Canada for support of this work. A
fellowship (to C.Z.) is acknowledged from the Hungarian
Academy of Sciences (Budapest).
Su p p or tin g In for m a tion Ava ila ble: Tables of X-ray
data, positional and thermal parameters, and bond distances
and angles (12 pages). Ordering information is given on any
current masthead page.
(30) Braterman, P. S.; Walker, B. S.; Robertson, T. H. J . Chem. Soc.,
Chem. Commun. 1977, 651.
OM950575I
+
+