Reversal of regioselectivity in Wacker-Type oxidation of simple terminal alkenes and its paired interacting orbitals (PIO) analysis

Article abstract of DOI:10.1246/bcsj.78.1555

The trend of aldehyde selectivity observed in the oxidation of 1-decene with PdCl₂(CH₃CN)₂/CuCl₂ in alcohol (ROH) under O₂ can be evaluated by the overlap populations for Pd-C and C-OR bonds in the oxypalladation step with paired interacting orbitals (PIO) analysis. The use of t-BuOH affords decanal in 84% regioselectivity.

Full text of DOI:10.1246/bcsj.78.1555

Short Articles
Bull. Chem. Soc. Jpn., 78, 1555–1557 (2005) 1555
PdCl (MeCN) (0.1 mmol)
2
2
CuCl2 (0.3 mmol)
Reversal of Regioselectivity in
+
ROH
O2 (1 atm), 50°C, 0.5h
Wacker-Type Oxidation of Simple
Terminal Alkenes and Its Paired
Interacting Orbitals (PIO) Analysis
CHO
+
O
1
2
R
Yield / %
1 : 2
Me
Et
26
22
11
7
97 : 3
84 : 16
58 : 42
16 : 84
i-Pr
t-Bu
Toshihiko Ogura, Ryuichiro Kamimura,
1
ꢀ
Akinobu Shiga, and Takahiro Hosokawa
Scheme 1.
tridecane as the internal standard. The results given in
Scheme 1 show that increasing the bulkiness of alcohol in-
creases the regioselectivity for aldehyde formation. In the case
of MeOH, only a little decanal was formed, and the C=C bond
isomerization of 1-decene occurred. The use of t-BuOH gives
decanal in 84% regioselectivity (0.5 h). In this case, virtually
no C=C bond isomerization took place.
Department of Environmental Systems Engineering,
Kochi University of Technology, 185 Miyanokuchi,
Tosayamada, Kochi 782-8502
1Sumitomo Chemical Co., Ltd., 6 Kitahara,
Tsukuba 300-3294
5
Received February 2, 2005; E-mail: hosokawa.takahiro@kochi-
tech.ac.jp
With these observations in hand, our attention was directed
toward verifying the efficacy of PIO analysis developed by
4
Fujimoto. The PIO analysis is a method for unequivocally
determining the orbitals which should play dominant roles in
chemical interactions between two fragments A and B. This
method provides not only visualization of bond-forming step
as the orbital phase, but also the suitability of the bond forma-
tion is evaluated by the value of overlap population (OP). The
PIO analysis based on the extended H u¨ ckel method can be
readily carried out by software called LUMMOXÔ recently
developed by Sumitomo Chemical Co., Ltd in Japan.4
Among the studies of PIO analysis made by Fujimoto, fruitful
insights have been given on the stereochemistry of the Wacker
reaction, where the structural model was constructed by
The trend of aldehyde selectivity observed in the oxida-
tion of 1-decene with PdCl2(CH3CN)2/CuCl2 in alcohol
(
ROH) under O2 can be evaluated by the overlap populations
for Pd–C and C–OR bonds in the oxypalladation step with
paired interacting orbitals (PIO) analysis. The use of t-BuOH
affords decanal in 84% regioselectivity.
c,6
Simple terminal alkenes such as 1-decene, when treated
1
with PdCl2 in water under O2, give methyl ketones. No alde-
hydes are formed via the nucleophilic attack of H2O at termi-
nal olefinic carbon. To get aldehydes from non-branched 1-al-
kenes is a challenging subject in the chemistry of the Wacker
ꢂ
ꢂ
4a
ethene, PdCl3 , and OH . For our present analysis, this
Fujimoto model was slightly modified, as shown in Fig. 1.
The modifications are as follows. (i) Propene was chosen
2
reaction. In 1993, Wenzel reported that 1-octene was oxidized
to octanal in t-BuOH by a catalyst system of PdCl2(CH3CN)2,
CuCl2, and LiCl, where the regioselectivity was 57% in 1%
instead of 1-decene for simplifying calculation. (ii) The bond
7
ꢀ
length of newly formed C–OR bond was selected as 2.20 A.
3
(iii) The bond lengths for propene C –CH and O–C bond of
2
chemical yield. This is the only document so far reported.
3
ꢀ
alkoxy group (OR) are selected to be 1.54 and 1.42 A, respec-
tively, as the usual values. Other bond lengths and angles are
taken from the values of Fujimoto model.4 Shown in Fig. 1
are the models for methyl ketone (model 1) and aldehyde
Our present study shows that, under the conditions similar to
those of Wenzel, the bulkiness of alcohols as the nucleophile
is evidently responsible for the formation of aldehyde. Howev-
er, a notion of alcohol bulkiness is one factor governing the
regioselectivity. The other important one is the ease of Pd–C
bond formation. We report here that this insight is given by
the paired interacting orbitals (PIO) analysis, a simple device
a
(model 2), each of which corresponds to the reaction model
1
.42Å
R
R
O
O
of molecular orbital calculation, which clarifies local charac-
teristics of the bond-forming step.4
105°
2
.20Å
165°
CH3
H
H
H
1
.40Å
100°
C1 C
2
At first, we found that a higher aldehyde selectivity could
be attained in t-BuOH by using a simple combination of
PdCl2(CH3CN)2 and CuCl2 (1:3). The use of LiCl and/or
CuCl reduced the regioselectivity. The experimental procedure
is as follows. A suspension of PdCl2(CH3CN)2 (0.1 mmol),
CuCl2 (0.3 mmol), and alcohol (1 mL) was stirred at room
temperature for 0.5 h under O2 (balloon), and then a solution
of 1-decene (1 mmol) in alcohol (1 mL) was added. After
C1
C2
H
H
1
.10Å
2.20Å
Pd
90°
1
.54Å
CH3
H
Cl
Cl
116°
Pd
90°
Pd-Cl = 2.30Å
Cl
Cl
Cl
Cl
model 1 (ketone)
model 2 (aldehyde)
ꢁ
the resulting solution was stirred at 50 C for 0.5 h under O2
(balloon), the reaction was stopped by adding hexane (1 mL)
and ether (1 mL). The products were analyzed by GLC using
Fig. 1. Reaction models for methyl ketone (model 1) and
aldehyde (model 2).
Published on the web August 4, 2005; DOI 10.1246/bcsj.78.1555

1
556 Bull. Chem. Soc. Jpn., 78, No. 8 (2005)
Ó 2005 The Chemical Society of Japan
Table 1. Total Overlap Population (ꢁOP) of All Paired
Interacting Orbitals in Models 1 and 2
In summary, the present study shows the following results.
(1) The formation of Pd–C bond in the oxypalladation step is
facile. (2) The relative ease for making Pd–C and C–OR bond
at either C1 or C2 is responsible for the regioselectivity. (3)
The PIO analysis clarifies the local characteristics of the
bond-forming step in the present reaction. We believe that
these facts must contribute to the organic chemistry of palla-
dium as parts of fundamental information.
Model 1 (ketone)
OP
Model 2 (aldehyde)
ꢁOP
ꢁ
R
Pd–C
C–OR ꢁꢁOP Pd–C
C–OR ꢁꢁOP
Me þ0:112 ꢂ0:087 þ0:025 þ0:003 ꢂ0:028 ꢂ0:025
Et þ0:112 ꢂ0:088 þ0:024 þ0:003 ꢂ0:033 ꢂ0:030
i-Pr þ0:112 ꢂ0:105 þ0:007 þ0:003 ꢂ0:052 ꢂ0:049
t-Bu þ0:104 ꢂ0:669 ꢂ0:564 þ0:003 ꢂ0:090 ꢂ0:087
References
1
a) J. Tsuji, Synthesis, 1990, 739. b) J. Tsuji, ‘‘Palladium
for oxypalladation step. Subsequent Pd–H elimination leads to
8
Reagents and Catalysts, Innovation in Organic Synthesis,’’ John
Wiley & Sons, New York (1995), pp. 19–124. c) P. M. Henry,
‘‘Handbook of Organopalladium Chemistry for Organic Synthe-
sis,’’ ed by E.-i. Negishi, John Wiley & Son, New York (2002),
Vol. II, pp. 2119–2139. d) T. Hosokawa, PETROTECH, 25, 801
(2002).
vinyl ether, which reacts with ROH to afford acetal. The ace-
tal or vinyl ether thus formed could be attacked by water gen-
erated in situ to yield methyl ketone or aldehyde.3
,8b,9
In any
case, the regioselectivity is determined in the stage of oxypal-
ladation.
2
For related references, see: a) T. Hosokawa, S. Aoki, M.
Since the reaction occurs via the formation of two bonds
of Pd–C and C–OR as depicted in Fig. 1, calculations were
performed for two sets of A and B in each model, i.e., A/
B = PdCl3/remaining and OR/remaining for Pd–C and C–
OR bond, respectively. The total overlap populations (ꢁOP)
of all PIOs deduced in each model are summarized in
Table 1. At first glance, one can see that ꢁOP values for
Pd–C are all positive, whereas all ꢁOP values for C–OR are
negative. Namely, the overlap interaction between Pd and C
is favorable, whereas the C–OR bond is repulsive. In other
words, the formation of Pd–C takes place effectively, but not
for C–OR. The PIO analysis also shows that the negative value
Takano, T. Nakahira, Y. Yoshida, and S.-I. Murahashi, J. Chem.
Soc., Chem. Commun., 1992, 1559. b) T. Hosokawa and S.-I.
Murahashi, Acc. Chem. Res., 23, 49 (1990). c) T. Hosokawa and
S.-I. Murahashi, ‘‘Handbook of Organopalladium Chemistry for
Organic Synthesis,’’ ed by E.-i. Negishi, John Wiley & Son,
New York (2002), Vol. II, pp. 2141–2159.
1
0
3
4
T. Wenzel, J. Chem. Soc., Chem. Commun., 1993, 862.
a) H. Fujimoto and T. Yamasaki, J. Am. Chem. Soc., 108,
5
78 (1986). b) H. Fujimoto, Acc. Chem. Res., 20, 448 (1987). c)
http://www.rsi.co.jp/kagaku/cs/pio/
5
In the case of MeOH, ꢃ24% of 1-decene was isomerized
to 2- and 3-decene (5:1) under the conditions described in the text,
and a mixture of 3- and 4-decanone (ꢃ1:1) was formed in ꢃ22%
yield for 0.5 h. Similar trends were observed with EtOH. How-
ever, in the cases of i-PrOH and t-BuOH, virtually no isomeriza-
tion of 1-decene occurred in 0.5 h.
(overlap repulsion) for C–OR increases with increasing the
steric hindrance of OR group toward propene moiety.
The overlap population of each PIOs reflects the extent of
electron delocalization between two fragments of A and B.
The larger the overlap population, the more smoothly the bond
formation occurs. In the case of MeOH, the overlap population
of C–OR in model 1 (ketone) is smaller than that in model 2
6
7
T. Motoki and A. Shiga, J. Comput. Chem., 25, 106 (2003).
The distance of C–OR bond was changed from 2.40 to 1.80
Aꢀ by the interval of 0.20 Aꢀ , and calculations for the extended
(
aldehyde), indicating that the model 1 leading to methyl ke-
H u¨ ckel energies of the interacting system C, which was construct-
ꢂ ꢂ
tone is less likely. Thus, the preferential formation of methyl
ketone observed is not justified by the value of C–OR. Com-
pensation by the positive overlap population of Pd–C is
required for it. The same trend is also seen with EtOH and
i-PrOH. In the case of t-BuOH, inversely, the C–OR bond
formation is important in determining the regioselectivity, be-
cause the very large negative value of C–OR is not compensat-
ed by the positive value of Pd–C. Obviously, model 1 is not
preferable in this case. In short, the PIO analysis explicitly
shows that the present reaction is controlled by the competitive
nature of positive and negative overlap populations with
revealing their characteristics.
ed by fragment A (propene + PdCl3 ) and fragment B ( OMe or
ꢂ
ꢀ
O-t-Bu), were performed. As the result, the distance of 2.20 A
was selected as appropriate by judging the value of ꢂE ¼
EC ꢂ ðEA þ EBÞ, where the symbols of EC, EA, and EB represent
the extended H u¨ ckel energies of C, A, and B noted above. The
bond lengths and angles in alkoxy groups (OR) were taken from
ꢁ
ꢀ
the usual values, such as C{C ¼ 1:54 A and H{C{H ¼ 109:5 .
a) T. Hosokawa, T. Ohta, S. Kanayama, and S.-I.
8
Murahashi, J. Org. Chem., 52, 1758 (1987). b) T. Hosokawa,
Y. Ataka, and S.-I. Murahashi, Bull. Chem. Soc. Jpn., 63, 166
(
1990). c) T. Hosokawa, T. Yamanaka, M. Itotani, and S.-I.
Murahashi, J. Org. Chem., 60, 6159 (1995).
As the reaction proceeds, water or hydroperoxypalla-
9
The above considerations allow us to propose that the sum
ꢁꢁOP) of ꢁOP value for Pd–C and C–OR can formally serve
dium(II) formed in situ must attack the alkene to induce the ke-
tonization (see Ref. 8b). In fact, the relative formation of decanal
in t-BuOH decreased with reaction time; e.g., 1=2 ¼ 29=71 in
(
as the indicator of the reaction pathway. Since ꢁꢁOP value of
þ0:025 with MeOH in model 1 (ketone) is larger than that of
ꢂ0:025 in model 2 (aldehyde), methyl ketone is produced. In
the case of t-BuOH, ꢁꢁOP of ꢂ0:087 in model 2 is larger
than that of ꢂ0:564 in model 1. Thus, aldehyde becomes the
major product. The same trend is again confirmed with EtOH
and i-PrOH. The reactivity of ROH appears to correlate with
the ꢁꢁOP value, albeit not exactly.
1
2% yield for 1 h.
1
0
For the analysis with MeOH, one H atom of methyl group
was placed on the plane perpendicular to the plane containing
C=C bond, and the C–OMe bond of this conformer was then ro-
ꢁ
tated by 180 . Between these two conformers thus formed, the ex-
tended H u¨ ckel energies (E ) of the interacting system (see Ref. 7)
C
were compared, and the more stable conformer was selected. The

T. Ogura et al.
OP value listed in Table 1 was taken from that of the more stable
Bull. Chem. Soc. Jpn., 78, No. 8 (2005) 1557
ꢁ
ꢁ
of 60 . When the extended H u¨ ckel energies (EC) of 6 conformers
thus formed were compared, it was found that 2 or 3 conformers
among them were highly stable and that their stabilities were not
much different. Thus, the ꢁOP values for EtOH and i-PrOH
listed were deduced by averaging the ꢁOP values of these 2 or
3 conformers.
conformer. For the analysis with t-BuOH, the same treatment as
above was performed by replacing H atom with Me group. For
the analysis with EtOH and i-PrOH, one H atom of Et or i-Pr
was placed on the same position as that in the case of MeOH.
The C–OR bond (R ¼ Et or i-Pr) was then rotated by the interval