Preparation of d,l-phenylalanine by amidocarbonylation of benzyl chloride

Article abstract of DOI:10.1021/jo951802t

The preparation of d,l-phenylalanine via amidocarbonylation of benzyl chloride with acetamide and CO/H₂ is described. The rate of the reaction is dependent upon the CO pressure below 250 bar, but independent of the hydrogen pressure. A reaction temperature of 100°C gives optimum yields. A relatively large amount of the catalyst, Co₂(CO)₈, is needed for complete conversion because of inhibition caused by hydrogen chloride which is formed during the reaction. Addition of NaHCO₃ removes HCl as insoluble NaCl, resulting in improved conversion and selectivity of the reaction. It also allows the use of a stoichiometric amount of acetamide, whereas a 2-to 3-fold excess of acetamide is needed for complete conversion of benzyl chloride without NaHCO₃. Amidocarbonylation of benzyl alcohol gave d,l-phenylalanine in only 8% yield.

Full text of DOI:10.1021/jo951802t

1842
J . Org. Chem. 1996, 61, 1842-1846
P r ep a r a tion of d ,l-P h en yla la n in e by Am id oca r bon yla tion of
Ben zyl Ch lor id e
J ohannes G. de Vries,* Rene´ P. de Boer, Marjo Hogeweg, and Esther E. C. G. Gielens
DSM-Research, Dept. Fine Chemicals/ Intermediates, P.O. Box 18, 6160 MD Geleen, The Netherlands
Received October 3, 1995X
The preparation of d,l-phenylalanine via amidocarbonylation of benzyl chloride with acetamide
and CO/H2 is described. The rate of the reaction is dependent upon the CO pressure below 250
bar, but independent of the hydrogen pressure. A reaction temperature of 100 °C gives optimum
yields. A relatively large amount of the catalyst, Co2(CO)8, is needed for complete conversion because
of inhibition caused by hydrogen chloride which is formed during the reaction. Addition of NaHCO3
removes HCl as insoluble NaCl, resulting in improved conversion and selectivity of the reaction.
It also allows the use of a stoichiometric amount of acetamide, whereas a 2- to 3-fold excess of
acetamide is needed for complete conversion of benzyl chloride without NaHCO3. Amidocarbon-
ylation of benzyl alcohol gave d,l-phenylalanine in only 8% yield.
Sch em e 1. Th e Holla n d Sw eeten er Com p a n y’s
Asp a r ta m e P r ocess
In tr od u ction
The Holland Sweetener Company’s Aspartame process
(Scheme 1)1 hinges on the use of an enzyme for the
crucial coupling between the two amino acid components
N-(benzyloxycarbonyl)aspartic acid (Z-Asp-OH)2 and phen-
ylalanine methyl ester (H-Phe-OMe). This has a number
of advantages, one of which is the ability to use racemic
phenylalanine as feedstock as the enzyme will only
convert the S-enantiomer. As a consequence much effort
has been devoted to the development of an economic
process for the production of d,l-phenylalanine.
A promising procedure for the preparation of amino
acids is Wakamatsu's amidocarbonylation reaction3,4
which allows the preparation of amino acids from alde-
hydes containing at least one R-proton, CO, and a limited
class of amides, in particular acetamide. The necessity
of the R-proton is best explained by assuming the
intermediacy of the enamide (Scheme 2). This is cor-
roborated by the fact that an enantiopure aldehyde with
a chiral carbon atom in the R-position gave rise to a
racemic amino acid derivative when submitted to stan-
dard amidocarbonylation conditions.5
Sch em e 2. Mech a n ism of th e Am id oca r bon yla tion
Rea ction
Precursors that will form aldehydes under the condi-
tions of the amidocarbonylation reaction such as olefins,6,7
epoxides,8 and allylic alcohols9 can also be used as
substrates. Ojima was able to prepare N-acetylphenyl-
alanine (1) in excellent yield by amidocarbonylation of
styrene oxide in the presence of a titanium cocatalyst to
accelerate the rearrangement of the epoxide to the
aldehyde.8 However, styrene oxide is not an attractive
feedstock for reasons of cost and toxicity.
More interesting is the patent disclosure of Wakamatsu
et al. claiming the amidocarbonylation of benzyl halides
and alcohols.10 Presumably, this proceeds via phenyl-
acetaldehyde which is formed from benzyl chloride or
alcohol by hydroformylation. The hydroformylation of
benzyl chloride with Co2(CO)8 in the presence of a base
has been described by others and gives phenylacetalde-
hyde in good yield.11 The hydroformylation of benzyl
X Abstract published in Advance ACS Abstracts, February 15, 1996.
(1) Oyama, K. In Chirality in Industry; Collins, A. N., Sheldrake,
G. N., Crosby, J ., Eds.; J ohn Wiley & Sons: New York, 1992, p 237.
(2) The following abbreviations were used: Z-Asp-OH ) N-(Benzyl-
oxycarbonyl)aspartic acid, H-Phe-OH ) phenylalanine, H-Phe-OMe )
Phenylalanine methyl ester, Ac-Phe-OH ) N-acetylphenylalanine,
DMF ) dimethylformamide, MIBK ) methyl isobutyl ketone, EtOAc
) ethyl acetate, BzCl ) benzyl chloride, HOAc ) acetic acid.
(3) Wakamatsu, H.; Uda, J . and Yamakami, N. J . Chem. Soc., Chem.
Commun. 1971, 1540.
(4) Lin, J .-J .; Knifton, J . F. In Homogeneous Transition Metal
Catalyzed Reactions, Advances In Chemistry Series 230, Moser, W. R.,
Slocum, D. W., Eds.; American Chemical Society: Washington, D.C.,
1992, p 235.
(5) Parnaud, J .-J .; Campari, G.; Pino, P. J . Mol. Catal. 1979, 6, 341.
(6) Stern, R.; Hirschauer, A.; Commereuc, D.; Chauvin, Y.; US
Patent 4,264,515 (to Institute Franc¸ais du Petrol), 1981.
(7) Ojima, I.; Okabe, M.; Kato, K.; Boong Kwon, H.; Horva´th, I. T.
J . Am. Chem. Soc. 1988, 110, 150.
(10) Yukata, T.; Yamakani, N.; Honma, M.; Komachiya, Y.; Waka-
matsu, H. US Patent 3,996,228 (to Ajinomoto Co. Inc.), 1976.
(11) (a) Takano, T.; Suzukamo, G.; Ishino, M.; Ikimi, K. European
Patent, 34430 (to Sumitomo Chemical Co. Ltd.), 1981. (b) Ayabe, M.;
Hirano, H.; Kibayashi, I.; Shimizui, T.; Kosai, Y. European Patent
109679 (to Denki Kagaku Kogyo K. K.), 1984.
(8) Ojima, I.; Hirai, K.; Fujita, M.; Fuchikami, T. J . Organomet.
Chem. 1985, 279, 203.
(9) Hirai, K.; Takahashi, Y.; dand Ojima, I. Tetrahedron Lett. 1982,
23, 2491.
0022-3263/96/1961-1842$12.00/0 © 1996 American Chemical Society

Preparation of d,l-Phenylalanine
J . Org. Chem., Vol. 61, No. 5, 1996 1843
Sch em e 3. Am id oca r bon yla tion of Ben zyl
Ch lor id e
alcohol was not described before; the expected product
is phenylacetic acid. It seems that in both cases the
acetamide catalyzes the hydroformylation reaction in
some way. The effect seems general for amides. It was
found that in the presence of a stoichiometric amount of
DMF rather than acetamide, aldehydes were indeed
formed, both from benzyl halides and alcohols under
conditions of the amidocarbonylation reaction.10 The
origins of this amide effect are unknown, but the effect
could be caused by the amide functioning as ligand to
the cobalt catalyst.
F igu r e 1. Time dependence of conversion and selectivity in
the standard experiment. (b), Conversion benzyl chloride; (O)
selectivity benzyl chloride; (9) conversion acetamide; (0)
selectivity acetamide; (f) yield of 1.
The amidocarbonylation of benzyl chloride to give d,l-
N-acetylphenylalanine (1) (Scheme 3) is also described
in this patent.10 In view of the fact that benzyl chloride
is a cheap and readily available commodity we decided
to further develop this reaction to make it amenable to
large scale production. The process as described in the
patent has several drawbacks that need to be overcome
before scale-up can be attempted. Very high pressures,
i.e. between 250 and 300 bar, are needed for good yields
of acylated amino acid. Relatively large amounts of the
catalyst, Co2(CO)8, are used which would require an
efficient recycle process to be viable. Good yields of
product (up to 79%) are obtained only if a two- to three-
fold excess of acetamide is used, but this might hamper
the purification procedure. Also, the selectivity with
respect to acetamide is rather poor. This is even more
concerning as acetamide is the most expensive feedstock
in the process.
F igu r e 2. Influence of temperature on conversion and selec-
tivity. (b) Conversion benzyl chloride; (O) selectivity benzyl
chloride; (9) conversion acetamide; (0) selectivity acetamide;
(f) yield of 1.
Longer reaction times gave somewhat higher conversions
but this was offset by a decrease in selectivity. The
reaction is initially very rapid, but after about 10 min a
much slower phase sets in. This can be explained by
inhibition of the catalyst, presumably caused by one of
the products of the reaction. The lower selectivities after
shorter reaction times point to incomplete conversion of
intermediates such as phenylacetaldehyde.
The temperature profile (Figure 2) also indicates that
the initially chosen standard temperature of 100 °C is
indeed an optimum. At 47 °C condensation between
phenylacetaldehyde and acetamide hardly occurs, judging
by the low conversion of acetamide. Surprisingly, only
a small amount of phenylacetaldehyde was found. This
is partially explained by the poorer selectivity of the
hydroformylation of benzyl chloride. The hydrogenation
product toluene is formed with 12% selectivity against
4% at 110 °C, but most of the products are unknown. At
140 °C selectivities are clearly deteriorating.
The results of our attempts to address these problems
are described in this report.
Resu lts a n d Discu ssion
The amidocarbonylation of benzyl chloride with aceta-
mide (Scheme 3) was explored. A standard experiment
was devised based on the best results of Wakamatsu et
al10 (See footnote a in Table 1). All reaction parameters
were varied, one at the time, and the effects on rate,
selectivity, and catalyst stability were recorded. From
a screening of six different industrially relevant solvents12
methyl isobutyl ketone (MIBK) emerged as the solvent
of choice, giving the highest yield of 1.
Rea ction Tim e a n d Tem p er a tu r e. The results of
the standard experiment (Table 1, entry 1) raise the
question whether the yield of 1 could further be increased
by extending the reaction time, or by raising the tem-
perature, as all starting materials had not yet been
converted. This was verified and the results are compiled
in Figures 1 and 2. From Figure 1 it is clear that after
85 min the optimum yield of 1 has indeed been reached.
P r essu r e a n d CO/H2 Ra tio. Figure 3 shows the
effect of the total reaction pressure on the yield of 1.
N-Ac-Phe-OH is only formed at pressures in excess of 100
bar; above 280 bar the maximum yield seems to have
been achieved. The results in Table 1 show that the
hydrogen pressure is not the rate-limiting factor, which
means that hydrogenation of the cobalt acyl intermediate,
(12) MIBK, acetone, EtOAc, dioxane, DMF, and toluene where
tested.

1844 J . Org. Chem., Vol. 61, No. 5, 1996
de Vries et al.
F igu r e 4. Effect of NaHCO3 on conversion and selectivity of
amidocarbonylation. (b) Conversion benzyl chloride; (O) se-
lectivity benzyl chloride; (9) conversion acetamide; (0) selectiv-
ity acetamide; (f) yield of 1.
F igu r e 3. Relation between yield of 1 and total pressure. (f)
Yield of 1.
Ta ble 1. CO/H2 Ra tioa
pressure CO/H2 convb
conv
selc
sel
% yield
Ta ble 2. Am ou n t of Ca ta lysta
exp
(bar)
ratio BzCl CH3CONH2 BzCl CH3CONH2 of 1
mol% conv
conv
sel
sel
% yield
of 1
l
2
3
288
265
275
1:1
3:1
23:1
74
81
81
43
37
43
82
79
78
71
85
73
61
63
63
exp
catb
BzCl CH3CONH2 BzCl CH3CONH2
1
2
3
4
5
6
7
3.25
3.25c
3.25d
7.5
41
32
49
43
22
35
43
41
51
58
41
17
42
35
82
83
80
20
12
30
23
74
81
69
17
3
a
Conditions: Benzyl chloride (40 mmol), acetamide (80 mmol),
21
20
61
83
80
Co2(CO)8 (5 mmol), solvent (40 mL), CO/H2 (1:1, 275 bar), 100 °C,
85 min. All conversions, selectivities, and yields are percentages.
58
12.75
25.75
38.5
74
100
100
b
Conversion is defined as Moles of converted starting material/
Moles of starting material initially present × 100. c Selectivity is
defined as: Yield of 1 in moles/converted starting material in moles
× 100.
a
b
For conditions see footnote a in Table 1. 1 mol % catalyst )
2 mol% Co. c T ) 140 °C. t ) 160 min.
d
the presumed final step in the hydroformylation of benzyl
chloride, is not the rate-determining step.
Prolongation of the experiment with 3.25 mol % catalyst
to 160 min (exp 3) showed that both benzyl chloride
conversion and N-Ac-Phe-OH yield had only marginally
increased. Using the same amount of catalyst at 140 °C
(exp 2) led to breakdown of conversion and selectivity as
observed before.
The need for a large amount of catalyst can only be
satisfactorily explained by catalyst deactivation.
Ca ta lyst Dea ctiva tion . An obvious reason for the
occurrence of catalyst deactivation is the formation of
stoichiometric amounts of HCl in the reaction; chloride
is a well known catalyst poison for cobalt catalysts in CO
chemistry.14 In addition, HCl catalyzes the hydrolysis
of acetamide. Addition of NaHCO3 should circumvent
all these problems by neutralizing the reaction mixture
and removing chloride as insoluble NaCl.
Figure 4 shows the effect of the addition of varying
amounts of NaHCO3 on conversion and selectivity of the
reaction. There clearly is a beneficial effect on the
conversion of both benzyl chloride and acetamide. This
effect is optimal with slightly less than stoichiometrical
amounts of NaHCO3. Apparently, some measure of
acidity is necessary to catalyze the condensation of
acetamide with phenylacetaldehyde.
The experiment in which only 3.25 mol % of catalyst
was used was also repeated with addition of 1 equiv of
NaHCO3. The outcome was rather disappointing; even
after 160 min less than 1% of 1 had been formed. Most
revealing was the fact that in this case there was an
excellent conversion (86%) of benzyl chloride, mostly to
phenylacetaldehyde, whereas acetamide remained largely
unconverted (3%). Lack of acid catalyst for the conden-
sation step is one possible explanation for the inhibition
of the amidocarbonylation of phenylacetaldehyde. In-
deed, no reaction occurred when amidocarbonylation of
phenylacetaldehyde was attempted in the presence of 1
Pino et al. showed that the amidocarbonylation of
butyraldehyde is dependent upon the CO pressure at 70
°C between 10 and 60 bar; above 110 bar the rate shows
a slightly negative dependence upon the CO pressure.5
The amidocarbonylation of phenylacetaldehyde proceeds
rapidly at 140 bar 1:1 CO/H2.13 This seems to indicate
that the CO pressure dependence below 250 bar is related
to the first step, the hydroformylation of benzyl chloride
(See Scheme 3).
Am ou n t of Ca ta lyst. Separation and recycle of the
catalyst is often one of the most expensive steps of a
process that uses homogeneous catalysis. The cobalt
catalyst can be recycled by oxidation with air in the
presence of HOAc14,15 or oxalic acid.16 This results in the
formation of Co(II) salts which can be extracted with
water. Treatment of the aqueous Co(OAc)2 solution with
CO/H2 at high pressure will regenerate HCo(CO)4, the
presumed active species. Obviously, the size of this
recycle stream will have a large impact on investment
and operating costs. In the standard experiment as much
as 12.5 mol % of catalyst is used; reduction of this amount
is highly desirable.
Not surprisingly, there was a clear relationship be-
tween catalyst concentration and benzyl chloride conver-
sion (Table 2). At the conditions of the standard experi-
ment 25 mol % of catalyst was needed for complete
conversion. The question remains whether a large
amount of catalyst is really needed when the reaction is
allowed to proceed longer or at higher temperatures.
(13) Lin, J .-J .; Knifton, J . J . Organomet. Chem. 1991, 147, 99.
(14) Falbe, J . Synthesen mit Kohlenmonoxid; Springer Verlag, New
York, 1967, p 19.
(15) Falbe, J . New Syntheses with Carbon Monoxide; Springer
Verlag, New York, 1980; p 80.
(16) El-Chahawi, M.; Meyer, G. German Patent 3345411 (to Hu¨ls
Troisdorf AG), 1985.

Preparation of d,l-Phenylalanine
J . Org. Chem., Vol. 61, No. 5, 1996 1845
a
Ta ble 3. Aceta m id e/Ben zyl Ch lor id e Ra tio a n d Effect of Na HCO3
CH3CONH2/
BzCl
additives
and changes
conv
BzCl
conv
CH3CONH2
sel
BzCl
sel
% yield
of 1
exp
CH3CONH2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1
1
1
1.5
2
3
1
1
1
1
2
1
1
1
2
2
-
60
63
62
67
74
92
90
96
98
99
86
99
100
87
99
98
59
59
64
79
56
75
81
82
65
85
87
77
79
63
81
41
84
83
80
80
60
35
86
74
-
90
91
85
89
57
90
ND
90
73
79
47
35
46
54
61
60
76
82
76
79
54
80
41
71
82
78
140 °C
160 min
-
42
-
41
-
NDb
85
NaHCO3 (20)
NaHCO3 (30)
NaHCO3 (40)
NaHCO3 (40)c
d
90
89
89
47
89
ND
79
56
50
NaHCO3(40)d
NaHCO3(40)e
NaHCO3(50)
NaHCO3(30)
NaHCO3(40)
a
b
d
For conditions see footnote a in Table 1. Not determined. c t ) 160 min. 20 mL of MIBK. e No solvent.
Sch em e 4. Am id oca r bon yla tion of Ben zyl Alcoh ol
equiv of NaHCO3. Fine tuning the amount of NaHCO3
seems an obvious way to make more progress in this area.
Addition of 1 equiv amount of pyridine resulted in full
conversion of benzyl chloride, but neither 1 nor phenyl-
acetaldehyde was formed. Quaternization of pyridine
with benzyl chloride seems a likely explanation for this
result.
A more radical way to solve the chloride problem would
be to use benzyl alcohol as starting material. Amidocar-
bonylation of benzyl alcohol (Scheme 4) does indeed occur
under the standard conditions but conversion and selec-
tivity were much lower than with benzyl chloride.17
Aceta m id e/Ben zyl Ch lor id e Ra tio. All of the above
experiments were carried out with an acetamide/benzyl
chloride ratio of two. Figure 5 clearly shows the “linear”
relationship between acetamide/benzyl chloride ratio and
conversion of benzyl chloride. It is not clear at this point
what the origins are of this effect. It might mean that
the condensation of acetamide with phenylacetaldehyde
is rate determining, but it is also possible that acetamide
is a catalyst for the hydroformylation step, depending on
which step is rate determining. Another explanation
could be that acetamide plays a role in the prevention of
catalyst inhibition.
If the reaction with 1 equiv of acetamide was allowed
to proceed during 160 min, conversion and yield did not
increase. This clearly points to an earlier onset of
catalyst inhibition. Increasing the temperature to 140
°C gave lower selectivities leading to a lower yield of 1.
The effect of the addition of NaHCO3 in these experi-
ments was even more dramatic (Table 3). Addition of
0.75 equiv of NaHCO3 resulted in an 82% yield of 1 (exp
8). A result which could not be improved by the use of 2
equiv of acetamide (exp 15). The selectivities were even
better for the lower amount of acetamide.
F igu r e 5. Influence of acetamide benzyl chloride ratio. (b)
Conversion benzyl chloride; (O) selectivity benzyl chloride; (9)
conversion acetamide; (0) selectivity acetamide; (f) yield of
1.
Lowering the amount of solvent gave rise to lower
selectivities and yields (Table 3, exp 11), but addition of
bicarbonate allows the amount of solvent to be halved
(Table 3, exp 12). Total omission of solvent was not
possible and led to sharply reduced yields.
Sid e P r od u cts. The main side product is toluene (2)
which is formed in amounts of 4-6 mol % by hydrogena-
tion of benzyl chloride. Smaller amounts of phenyl-
acetaldehyde (3) and its hydrogenation product phenyl-
ethanol (4) are also formed. But even in reactions where
the selectivity in benzyl chloride is low the amount of
these products never exceeds 2 mol %. In a number of
reactions minor amounts of 2-acetamidostyrene (5) were
formed (both E and Z). In the GC of all reaction mixtures
azalactone (6) was observed. Formation of this product
is readily explained by internal attack of the amide
carbonyl group on the cobalt acyl intermediate (Scheme
5).18 However, even when pure 1 was injected this peak
was observed, suggesting that thermal dehydration of 1
to 6 is possible. This throws doubt on the presence of
this compound in the amidocarbonylation mixtures. The
formation of azalactone can be induced by addition of
molecular sieves to the amidocarbonylation reaction.19
(17) Results of the amidocarbonylation of benzyl alcohol under
conditions of the “Standard Experiment”: PhCH2OH: conv 22%, sel
38%; CH3CONH2: conv 17%, sel 24%, yield N-Ac-Phe-OH: 8%.
(18) Magnus, P.; Slater, M. Tetrahedron Lett. 1987, 28, 2829.
(19) Izawa, K.; Nishi, S.; Asada, S. J . Mol. Catal. 1987, 41, 135.
Con clu sion
The amidocarbonylation reaction is a very promising
way to prepare d,l-N-acetylphenylalanine. We were able

1846 J . Org. Chem., Vol. 61, No. 5, 1996
de Vries et al.
ylethanol. Amidocarbonylation experiments up to 180 bar
pressure were carried out in a 50 mL Parr autoclave (Hastelloy
C) equipped with a rupture disk, a manometer and a propeller
type stirrer which was operated at 700 rpm. Higher pressure
experiments were performed in Carius tubes (Hastelloy C)
with an empty volume of 220 mL, equipped with a manometer
and a rupture disc. The tubes were heated in a stainless steel
heating block mounted on a rocker. Toluene, benzyl chloride,
2-phenylethanol, phenylacetaldehyde, and benzyl alcohol were
analyzed by GC on a 25 m CP-Sil 5CB column (0.32 mm i.d, 2
µm film thickness). Initial temperature 60 °C, 60-250 °C (10
°C/min). Injection temperature 280 °C, detection temperature
280 °C. Acetamide was analyzed by GC on a 25 m CP-wax 51
column (0.25 mm i.d., 0.2 µm film thickness). Initial temper-
ature 60 °C, 60-250 °C (10 °C/min). Injection temperature
250 °C, detection temperature 250 °C. N-Ac-Phe-OH, N-For-
Phe-OH, and H-Phe-OH were analyzed by HPLC (HP 1090)
on a 25 × 0.4 cm Nucleosil 120-5-C18 column using gradient
elution (1 mL/min). Solvent A: 50 mM phosphate buffer pH
2.5, solvent B: 20% A/80% acetonitrile. Column temperature
40 °C, injection volume 25 µL. Waters 480 UV detection (257
nm). TLC analyses were performed on Merck silica plates.
Sch em e 5. Sid e P r od u cts in th e
Am id oca r bon yla tion of Ben zyl Ch lor id e
Sta n d a r d Exp er im en t. A Carius tube was flushed with
nitrogen for 10 min, filled with benzyl chloride (5.06 g, 40
mmol), acetamide (4.73 g, 80 mmol), ethyl benzene (500 µL, 4
mmol, internal standard), MIBK (40 mL), and Co2(CO)8 (1.76
g, 5 mmol), and closed. The tube was brought to 275 bar with
CO/H2 (1:1) and placed in a preheated heating block at 100
°C. The tube was shaken at this temperature for 85 min. After
this period the tube was removed from the heating block and
forcibly cooled to room temperature by immersion in cold
water. After 30-60 min the tube was opened and the contents
transferred to a volumetric flask (100 mL). The tube was
carefully rinsed with DMF to ensure complete removal of the
contents. The washings were also added to the volumetric
flask, and more DMF was added up to 100 mL. The solution
was analyzed by GC (two different columns, see Materials and
Methods) and HPLC. In a number of cases GC-MS was
performed to characterize unknown components. All ami-
docarbonylation experiments described in the preceding sec-
tions were variations of this standard experiment. The
experiments in the Parr autoclave were carried out in a similar
manner.
to greatly improve the reaction by addition of NaHCO3.
This addition obviates the need to use a two-fold excess
of acetamide and also allows us to reduce the amount of
solvent. Fine tuning the amount of NaHCO3 might
permit further reduction of the substrate catalyst ratio
which is still unacceptably high.
Exp er im en ta l Section
Ma ter ia ls a n d Meth od s. All chemicals were commercial
products that were used as received. Solvents were degassed
with nitrogen before use. DMF was freed from dimethylamine
and formic acid by storing over molecular sieves (4 Å) for at
least 48 h. Acetamide was dried under high vacuum for 24 h.
The benzyl chloride used was stabilized with 0.25% propylene
oxide. Phenylacetaldehyde was stabilized with 15% 2-phen-
J O951802T