The TEMPO/copper catalyzed oxidation of primary alcohols to aldehydes using oxygen as stoichiometric oxidant

Article abstract of DOI:10.1007/s00706-005-0349-0

A catalytic system for the selective oxidation of primary alcohols to aldehydes under very mild conditions was developed. The catalytic system is based on TEMPO and Cu(II), which is generated in situ by oxidation of elemental copper and chelated by means of 2,2′-bipyridine. Compared to existing Cu/TEMPO oxidation methods we substantially lowered the amount of copper necessary and discovered that the reaction is dependent on pH. The catalytic system was also tested with polymer-bound TEMPO and new insights into the currently discussed mechanism were derived. Springer-Verlag 2005.

Full text of DOI:10.1007/s00706-005-0349-0

Monatshefte f€ur Chemie 136, 1591–1599 (2005)
DOI 10.1007/s00706-005-0349-0
The TEMPO/Copper Catalyzed Oxidation
of Primary Alcohols to Aldehydes Using
Oxygen as Stoichiometric Oxidant
David Geißlmeir1, Walther G. Jary2, and Heinz Falk1;ꢀ
1
Institute of Organic Chemistry, Johannes Kepler University Linz, A-4040 Linz, Austria
2
DSM Fine Chemicals Austria, Research and Development, A-4021 Linz, Austria
Received May 19, 2005; accepted May 23, 2005
Published online August 4, 2005 # Springer-Verlag 2005
Summary. A catalytic system for the selective oxidation of primary alcohols to aldehydes under very
mild conditions was developed. The catalytic system is based on TEMPO and Cu(II), which is
generated in situ by oxidation of elemental copper and chelated by means of 2,20-bipyridine. Compared
to existing Cu=TEMPO oxidation methods we substantially lowered the amount of copper necessary
and discovered that the reaction is dependent on pH. The catalytic system was also tested with
polymer-bound TEMPO and new insights into the currently discussed mechanism were derived.
Keywords. TEMPO; Nitroxyl radical; Copper; 2,20-Bipyridine; Immobilized catalyst.
Introduction
The selective oxidation of primary alcohols to aldehydes is one of the most impor-
tant functional group transformations in organic synthesis and industrial fine che-
mical production. Classical methods are based on chromium reagents, manganese
dioxide, silver oxide, N-chlorosuccinimide [1], dimethyl sulfoxide activated by
oxalyl chloride [2], and periodanes [3, 4]. These methods are generally only useful
on a laboratory scale because of the toxicity or corrositivity of the reagents, and
because of the amount of chemical waste, which is produced due to the stoichio-
metric nature of these reactions. Modern oxidation methods work on a catalytic
basis, using clean stoichiometric oxidants, like oxygen or hydrogen peroxide. In
addition, nitroxyl radicals like 2,2,6,6-tetramethylpiperidin-1-oxyl (TEMPO) and
its derivatives in combination with copper ions, turned out to be efficient and
selective catalysts in the oxidation of alcohols to the corresponding carbonyl com-
pounds. The first copper=nitroxyl radical catalyzed oxidation has been reported
by Semmelhack et al. in 1984 [5], who managed to oxidize allylic and benzylic
ꢀ
Corresponding author. E-mail: heinz.falk@jku.at

1592
D. Geißlmeir et al.
alcohols at room temperature in DMF. More recently, several copper=TEMPO
catalytic systems have been published: Knochel et al. have used a mixture of
chlorobenzene=perfluorooctane as biphasic solvent system in combination with
perfluoroalkylated bipyridines as copper complexing agents [6, 7]. Ansari and Gree
as well as Jiang and Ragauskas have reported TEMPO catalyzed alcohol oxida-
tions in ionic liquids [8]. Minisci et al. have reported a highly efficient method for
the oxidation of alcohols using TEMPO in combination with Co=Mn or Cu=Mn
nitrates using acetic acid as the solvent. This reaction was carried out under ambi-
ent pressure and temperature [9]. In most recent works Sheldon et al. have inves-
tigated the mechanism of the original Semmelhack procedure [10] and developed a
new catalytic system using CuBr2=TEMPO in combination with 2,20-bipyridine as
complexing agent [11, 12]. Herein, we want to report fundamental improvements
of the catalytic system developed by Sheldon et al. as well as new insights into the
reaction mechanism.
Results and Discussion
Catalytic System
The Sheldon procedure, as shown in Scheme 1, is carried out in an acetonitrile=
water solvent mixture under ambient conditions. Although the reaction is very
environmentally friendly and safe, it has some drawbacks concerning its industrial
use, especially in terms of catalyst concentrations (5 mol% catalysts and co-
catalysts) and high cost of co-catalysts (e.g. potassium tert.-butoxide). We now
improved the catalytic system using benzyl alcohol as a substrate example. It
turned out that under the given reaction conditions a copper salt has not necessarily
to be used. It can be substituted by elemental, fine powdered copper, which is
oxidized in situ to Cu(II) and is then chelated by 2,20-bipyridine (bipy) by stirring
copper powder and bipy for 30 min before adding the other reagents. The reported
differences in reactivity of various copper salts [11], which are due to the counter-
ion, are circumvented in this way. In addition, we discovered that the amount
of TEMPO has a much higher influence on the reaction time than the amount of
copper catalyst. Thus, it was possible to lower the amount of copper to 0.5 mol%
and bipy concentrations to 2.5 mol% maintaining reasonable reaction times (Fig. 1).
By means of ICP analysis of the reaction mixture it turned out that only 0.22 mol%
copper are active. Lowering the amount of copper was one of our main goals,
because it is highly toxic for microorganisms and might cause problems in indus-
trial waste water treatment.
Upon replacing the costly potassium tert.-butoxide with NaOH, it turned out
that the reaction is pH-controlled. The reaction completely stopped at pH ¼ 8 and
the optimum pH was determined to be 13–13.5. By carrying out the reaction in this
Scheme 1

The TEMPO=Copper Catalyzed Oxidation
1593
Fig. 1. Effects of TEMPO concentration on the Cu=TEMPO catalyzed oxidation of benzyl alcohol;
reaction conditions: 2.5 mol% 2,20-bipyridine, 0.5 mol% Cu, pH 13.5, 25ꢁC, O2 atmosphere
(A) 5 mol% TEMPO (B) 2.5 mol% TEMPO
pH range maintained by constantly adding small amounts of NaOH, the influence
of the TEMPO concentration on the catalytic system was shown to be essential.
Thus, by halving the amount of TEMPO, the reaction time approximately doubled
(Fig. 1).
Furthermore, the commercially available immobilized TEMPO derivative
(TEMPO-4-oxymethyl)polystyrene [13, 14] was tested in a scale-up experiment
Fig. 2. Deactivation of polymer bound TEMPO in the oxidation of benzyl alcohol; reaction condi-
tions: 5 mol% TEMPO (2.5 mmol TEMPO=g (TEMPO-4-oxymethyl)polystyrene), 2.5 mol% Cu,
5 mol% 2,20-bipyridine, pH 13, 25ꢁC, O2 (0.3 dm3=min) (A) run 1 (B) run 2

1594
D. Geißlmeir et al.
under the same reaction conditions. After separation of the reaction mixture, the
polymer bound catalyst was washed with a small amount of solvent and reused for
the next experiment. In this way immobilized TEMPO was reused four times and
did not show any significant loss in activity or selectivity. However, during the
course of such a reaction it was observed that conversion rates dropped rather
quickly with reaction time, due to a deactivation of the catalyst. A turnover fre-
quency (TOF) vs. catalyst run time plot according to Ref. [15] showed that after 5 h
and approximately 50% conversion the catalyst activity was already very low
(Fig. 2). Thus, experiments were aborted at 80–85% and the immobilized TEMPO
was reused for the next run.
Substrate Scope
After having obtained promising results with benzyl alcohol, a substrate screening
was carried out using various substituted benzylic alcohols (Scheme 2, Table 1).
Unfortunately, it turned out that substrates, whose aldehydes are likely to form
hydrates under the given reaction conditions [16], are overoxidized to the acid by
about 50%. Substrates with electron donating groups, like methyl- or methoxyben-
zyl alcohols, are selectively oxidized to the aldehyde if substituted in ortho- or
Scheme 2
Table 1. Substrate screening of benzylic alcohols
Run
Substrate
Reaction time=h
Conversion=%
Selectivity=%
1
2-Nitrobenzyl alcohol
3-Nitrobenzyl alcohol
4-Nitrobenzyl alcohol
2-Chlorbenzyl alcohol
3-Chlorbenzyl alcohol
4-Chlorbenzyl alcohol
2-Methylbenzyl alcohol
3-Methylbenzyl alcohol
4-Methylbenzyl alcohol
2-Methoxybenzyl alcohol
3-Methoxybenzyl alcohol
4-Methoxybenzyl alcohol
23
8
92
71
41
42
2a
3a
4
8
82
45
23
4
>99
99
49
5b
56
6
8
>99
91
64
7
8a
30
8
>99
53
67
9
23
23
8
98
97
10
11
12a
>99
90
97
68
8
69
>99
a
Reactions were discontinued, although not finished; b 35 mmol alcohol, 7 mol% TEMPO, 3.5 mol%
2,20-bipyridine, 0.7 mol% Cu

The TEMPO=Copper Catalyzed Oxidation
1595
Scheme 3
Table 2. Substrate screening of aliphatic and allylic alcohols
Run
Substrate
Reaction time=h
Conversion=%
1
2
3
4
5
Citronellol
1-Octanol
2-Octanol
Phenylethanol
Geraniol
29
7
74a
39a
29
29
5
traces
traces
99b
a
b
Selectivity >99%, no further conversion with longer reaction times; selectivity 97%, de 67%
(E)-citral
para-position. For the corresponding aldehydes a resonance structure as shown in
Scheme 3 can be formulated and accordingly these compounds are stabilized
against a hydroxide ion attack and subsequent hydrate formation.
Another substrate screening was carried out on different aliphatic and allylic
alcohols. However, when using standard reaction conditions as described above
only very low conversions were observed. Thus, we decided to raise the amount of
copper to 2.5 mol% and the amount of bipy to 5 mol%. In case of 1-octanol the
reaction stopped at a conversion of 39%, which is remarkably lower than the results
obtained with citronellol and geraniol. The already known selectivity for primary
over secondary alcohols [12] was also proven, because only traces of 2-octanol
were oxidized to the corresponding ketone. It was not possible to completely
oxidize citronellol even when using long reaction times. On the contrary, geraniol
was completely and highly selectively oxidized within 5 h under the same reaction
conditions (Table 2). Accordingly, an allylic or benzylic activation of the alcohol
seems to be a prerequisite for a fast and complete conversion when using a copper=
TEMPO system under these conditions.
Mechanistic Aspects
The first mechanism for the copper=TEMPO catalyzed oxidation of aldehydes
proposed by Semmelhack et al. [5] involves the oxoammonium ion, which is gen-
erated by oxidation or disproportionation of TEMPO [17]. In the Semmelhack
mechanism the hydrogen abstraction of the ꢀ-hydrogen during the oxidation takes
place via an ionic transition state. Later on Sheldon et al. [10, 11] have proposed a
mechanism based on radical mediated ꢀ-hydrogen abstraction in the transition

1596
D. Geißlmeir et al.
Scheme 4
state (Scheme 4). Following this mechanism, Cu(II) forms a complex with a biden-
tate nitrogen ligand like 2,20-bipyridine (I) and then a deprotonated alcohol forms
an alkoxy species with this complex (II). The next step is the coordination of
TEMPO to copper in an ꢁ2 manner (III). Such complexes of copper(II) halides
with ꢁ2 coordinated TEMPO have been reported by Rey et al. and were proven by
X-ray diffraction [18, 19]. Abstraction of the ꢀ-proton by TEMPO then results in a
radical species IV, which breaks up into the corresponding aldehyde, TEMPOH,
and a copper(I) species V after intramolecular one-electron transfer. This is pro-
posed to be the rate determining step of the reaction. The second role of TEMPO is
to reoxidize Cu(I) to Cu(II) and then the catalytic cycle is closed by regeneration of
TEMPOH by oxygen.
The results obtained in our experiments are clearly in favor of the latter radical
mechanism. Stoichiometric oxidations of simple aliphatic alcohols by the sepa-
rately synthetisized oxoammonium salt of TEMPO are known [17]. For example,
1-octanol and 2-octanol are oxidized smoothly to the corresponding aldehyde and
ketone at ambient temperature and in good yields. In our case, however, these two
substrates showed almost no reactivity or did not react at all. On the contrary,
allylic and benzylic alcohols reacted readily using our method, because these
compounds are able to stabilize the radical transition state IVof Scheme 4. Further-
more, according to literature [20–22] oxoammonium ions of various nitroxyl radi-

The TEMPO=Copper Catalyzed Oxidation
1597
cals including TEMPO are rather unstable in aqueous solutions of alkali hydrox-
ides. These compounds are reduced to the nitroxyl radical with simultaneous
formation of hydrogen peroxide. This fact is in conflict with our findings that
the optimum pH for high reaction rates is 13. Another reason to favor the radical
mechanism is the difference in reactivity of substituted benzyl alcohols (Table 1).
Abstraction of the ꢀ-proton via an ionic transition state should be much faster if the
benzylic alcohols are substituted with electron withdrawing groups. Our results
show approximately equal reactivity regardless of the substitution of the benzylic
alcohol. This is due to the fact that radical species can be stabilized by electron
withdrawing as well as electron donating substituents. Neumann et al. [23], who
used TEMPO under acidic conditions, suggested that the reoxidation of TEMPOH
to TEMPO is the rate determining step. This proposal was restricted by Sheldon
et al., who showed, that reoxidation of TEMPOH is much faster at alkaline pH than
under neutral or acidic conditions [10], and thus the reoxidation step may only be
rate determining at low pH values. A radical transition state can also explain the
formation of (Z)-citral in the oxidation of (E)-geraniol (67% de (E)-citral was
observed after the reaction was finished). According to literature [24] rotational
barriers for allyl radicals range between 10 and 67 kJ=mol. Thus, it is likely that a
diastereomerization during the transition state takes place under our conditions.
The presence of water is crucial for the catalytic cycle. In the absence of water
a conversion equal to the amount of TEMPO catalyst added was observed. In our
opinion the most important effect of water and the hydroxide ion is the formation
of a hydroxo-complex together with Cu(II) and bipy, which has been first reported
by Wagner-Jauregg et al. [25]. Usually, Cu(II) forms a 3:1 complex [Cu bipy3]2þ
with excess of bipy, but in this case the copper ion sphere is completely shielded
against a coordination with TEMPO or the alcohol. It has been shown by Ojima
[26] that under aqueous alkaline conditions this complex reacts to the hydroxo-
complex [Cu [OH]2 bipy] and in this case the hydroxo-groups should be readily
replaced by the deprotonated alcohol or TEMPO (compare II and III in Scheme 4).
Conclusions
The initial reaction conditions for the Cu=TEMPO catalyzed oxidation of alcohols
to aldehydes developed by Sheldon et al. were fundamentally revised. The most
important improvement is the replacement of copper halogenide salts by elemental
copper powder, which is oxidized in situ to Cu(II) under the given reaction con-
ditions. In addition, the environmentally toxic copper catalyst was reduced to
0.5 mol%, and it was determined that only 0.22 mol% copper actively catalyzes
the reaction. The discovery that the reaction is controlled by pH gives a further
edge in terms of dosage of co-catalyst like the costly potassium tert.-butoxide,
which can be substituted by NaOH. A substrate screening showed that the catalytic
system is very useful for the oxidation of allyl and benzyl alcohols. However, if the
benzyl alcohols tend to form hydrates under the given reaction conditions, over-
oxidation to the corresponding acids will take place. Immobilized TEMPO was
successfully used along with Cu(0) and bipy in the oxidation of benzyl alcohol on a
larger scale. Our results rather imply a radical reaction mechanism, as proposed by
Sheldon et al., than a mechanism involving the oxoammonium ion. In particular,

1598
D. Geißlmeir et al.
higher reactivity of allyl and benzyl alcohols connected to the stabilization of the
intermediary formed radical species and the observed diastereomerization of allyl
alcohols favor the Sheldon radical mechanism.
Experimental
All chemicals were obtained from commercial suppliers and were used without further purification.
Reaction products were identified and quantified by GC analysis using a Hewlett Packard 5890 series II
gas chromatograph with FID detection equipped with a DB-1 column (30 m ꢂ 0.53 mm) or by HPLC
analysis using a Hewlett Packard 1050 LC system with UV detection (210nm) equipped with a Zorbax
Eclipse-XDB C8 column (4.6ꢂ 150 mm). Compounds were identified by comparison of the retention
times with commercially available references and quantified using nitrobenzene as internal standard, or
by external calibration.
Typical Oxidation Procedure
The oxidation of the alcohols was carried out under O2 in a 100 cm3 double jacketed vessel, thermo-
stated at 25ꢁC and equipped with a magnetic stirrer, thermometer, and pH-electrode. Cu powder
(16 mg, 0.25 mmol) were suspended in 60cm3 CH3CN=H2O 2=1, 195.5 mg 2,20-bipyridine
(1.25 mmol) were added, and the reaction mixture was stirred for 30min whereby it turned deep
red. Then 390 mg TEMPO (2.5mmol) and 200mm3 NaOH 40% were added and the reaction was
started by addition of 50mmol alcohol. During the reaction the pH-value was kept between 12.5 and
13.5 by addition of small amounts of NaOH 40%. Samples to monitor the progress of the reaction were
taken out regularly, acidified with acetic acid, and extracted with CH2Cl2 followed by derivatization
with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) if necessary.
Oxidation Procedure Using Immobilized TEMPO
The oxidation was carried out in a 500 cm3 double jacketed vessel (Schmizo), thermostated at 25ꢁC,
and equipped with reflux condenser, KPG stirrer, gas tube with frit por. 1, and pH-electrode. (TEMPO-
4-oxymethyl)polystyrene (7.5g, 2.5mmol TEMPO=g), 596 mg Cu (9.34 mmol), and 2.93g 2,20-bipyr-
idine (18.76mmol) were suspended in 500 cm3 CH3CN=H2O 2=1 and stirred for 45min. Then
40.6g benzylalcohol (375mmol) were added and O2 was bubbled through the reaction mixture
(0.3dm3=min). During the reaction the pH-value was kept between 12.5 and 13.5 by addition of small
amounts of NaOH 40%. Samples to monitor the progress of the reaction were taken out regularly,
acidified with acetic acid, and extracted with CH2Cl2 followed by derivatization with BSTFA. Reac-
tions were stopped at a conversion between 80% and 90% and the reaction mixture was removed.
(TEMPO-4-oxymethyl)polystyrene was washed with 300 cm3 CH3CN=H2O 2=1 and reused for suc-
ceeding experiments.
References
[1] Hudlicky M (1990) Oxidations in Organic Chemistry. ACS, Washington DC
[2] Mancuso AJ, Huang SL, Swern D (1978) J Org Chem 43: 2480
[3] Dess DB, Martin JC (1983) J Org Chem 48: 4155
[4] Dess DB, Martin JC (1991) J Am Chem Soc 113: 7277
ꢀ
[5] Semmelhack MF, Schmid CR, Cortes DA, Chou CS (1984) J Am Chem Soc 106: 3374
[6] Betzemeier B, Cavazzini M, Quici S, Knochel P (2000) Tetrahedron Lett 41: 4343
[7] Ragagnin G, Betzemeier B, Quici S, Knochel P (2002) Tetrahedron 58: 3985
[8] (a) Ansari IA, Gree R (2002) Organic Letters 4: 1507; (b) Jiang N, Ragauskas AJ (2005)
Tetrahedron Lett 46: 3323
[9] Cecchetto A, Fontana F, Minisci F, Recupero F (2001) Tetrahedron Lett 42: 6651

The TEMPO=Copper Catalyzed Oxidation
1599
[10] Dijksman A, Arends IWCE, Sheldon RA (2003) Org Biomol Chem 1: 3232
[11] Gamez P, Arends IWCE, Reedijk J, Sheldon RA (2004) Adv Synth Catal 346: 805
[12] Gamez P, Arends IWCE, Reedijk J, Sheldon RA (2003) Chem Commun, 2414
[13] Miyazawa T, Endo T, Okawara M (1985) J Polym Sci A: Polym Chem 23: 1527
[14] Miyazawa T, Endo T (1988) J Mol Catal 49: L31
€
[15] Dittmeyer R, Keim W, Kreysa G, Oberholz A (2003) Winnacker-Kuchler: Chemische Technik,
vol 1. Wiley-VCH, Weinheim
[16] McClelland RA, Coe M (1983) J Am Chem Soc 105: 2718
[17] de Nooy AEJ, Besemer AC, van Bekkum H (1996) Synthesis 10: 1153
[18] Caneschi A, Grand A, Laugier J, Rey P, Subra R (1988) J Am Chem Soc 110: 2307
[19] Laugier J, Latour JM, Caneschi A, Rey P (1991) Inorg Chem 30: 4474
[20] Anelli PL, Biffi C, Montanari F, Quici S (1987) J Org Chem 52: 2559
[21] Endo T, Miyazawa T, Shiihashi S, Okawara M (1984) J Am Chem Soc 106: 3877
[22] Osiecki JH, Ullman EF (1968) J Am Chem Soc 90: 1078
[23] Ben-Daniel R, Alsters P, Neumann R (2001) J Org Chem 66: 8650
[24] Korth HG, Trill H, Sustmann R (1981) J Am Chem Soc 103: 4483
[25] Wagner-Jauregg T, Hackley BEJ, Lies TA, Owens OO, Proper R (1955) J Am Chem Soc 77: 922
[26] Ojima H (1963) Nippon Kagaku Zasshi 84: 787; Chem Abstr 61: 45748