A mechanistic study on alcohol oxidations with oxygen catalysed by TPAP-doped ormosils in supercritical carbon dioxide

Article abstract of DOI:10.1002/adsc.200404365

The heterogeneous oxidation of various alcohols with oxygen catalysed by TPAP-doped ormosils in scCO₂ at 75°C and 22.0 MPa has been studied in detail. Sol-gel segregation of TPAP into the inner porosity of an organically modified silica (ormosi

Full text of DOI:10.1002/adsc.200404365

FULL PAPERS
A Mechanistic Study on Alcohol Oxidations with Oxygen
Catalysed by TPAP-Doped Ormosils in Supercritical Carbon
Dioxide
Sandro Campestrini,a,* Massimo Carraro,a Rosaria Ciriminna,b Mario Pagliaro,b
Umberto Tonellatoa
a
`
Universita di Padova, Dipartimento di Scienze Chimiche; CNR, Istituto per la Tecnologia delle Membrane, via Marzolo
1, 35131 Padova, Italy
Fax: (þ39)-49-827-5239, e-mail: sandro.campestrini@unipd.it
b
Istituto per lo Studio dei Materiali Nanostrutturati, CNR, via Ugo La Malfa 153, 90146 Palermo, Italy
Received: December 1, 2004; Accepted: March 4, 2005
Abstract: The heterogeneous oxidation of various al- sole reaction products. Investigation of the oxidation
cohols with oxygen catalysed by TPAP-doped ormo- mechanism shows that the catalytic process exhibits
sils in scCO2 at 758C and 22.0 MPa has been studied a first-order dependence on the amount of catalyst,
in detail. Sol-gel segregation of TPAP into the inner a fractional order on the alcohol concentration and
porosity of an organically modified silica (ormosil) a negative order for oxygen pressures higher than
xerogel along with the use of a reaction medium 0.2 bar. Evidence is presented for an associative oxi-
which does not dissolve the catalyst, prevents aggre- dation mechanism simultaneously involving TPAP, or-
gation of oxidation-inactive ruthenium derivatives ganic substrate and oxygen.
without the need of chemical tethering. Thus, at least
140 TONs may be obtained in the oxidation of pri- Keywords: carbon dioxide; ormosil; oxidation mecha-
mary alcohols with the formation of aldehydes as nism; oxygen; ruthenium; sol-gel
Introduction
Among these aerobic metal catalysts, ruthenium is
probably the most promising thanks to its impressive
Alcohol oxidation yielding ketones and aldehydes is a selectivity in the presence of several oxidisable
chemical transformation of primary importance in the groups;[4] in particular, the species tetra-n-propylam-
chemical industry, carbonyl compounds being precur- monium perruthenate (TPAP) is a very efficient and se-
sors of a variety of valuable fine chemicals including lective catalyst for the aerial oxidation of primary, sec-
drugs, vitamins and fragrances.[1] Traditional alcohol ox- ondary, benzyl and alkyl alcohols either dissolved in
idations, on the other hand, are still carried out by indus- conventional organic solvent[10–12] or even in an ionic
try with stoichiometric amounts of toxic and hazardous liquid.[13]
metal oxidants such as dichromate and permanganate in
Similarly to what happens with many catalytic transi-
volatile organic solvents producing enormous amounts tion metal ion pairs,[14] sol-gel encapsulation of TPAP
of waste.[2] Societal demand for fundamental chemical into ceramic matrices affords heterogeneous aerobic
transformations that are not harmful to the environment catalysts of different efficiency either for conversions
is pressing, and an economically and environmentally in conventional solvent[15,16] or in carbon dioxide in the
viable replacement of current alcohol oxidation proc- supercritical state,[17] an intrinsically non-toxic and envi-
esses with heterogeneous catalytic conversions is a sci- ronmentally benign reaction medium ideally suited for
entific challenge for chemists both in industry and acad- heterogeneous catalytic applications in the fine chemi-
emy.[3]
Obviously, the utilization of oxygen (and even better
cals industry.
In all cases, most functional groups, i.e., carbon-car-
of air) as stoichiometric oxidant for the selective oxida- bon double bond, epoxy, indole, acetal functions, etc.,
tion of organics yielding water as sole by-product is the remain unaffected by TPAP and high yields of carbonyl
final, ideal goal of prolonged research efforts that in the compounds are isolated.[10,11,18] However, in spite of the
last few years have resulted in a number of highly selec- well established synthetic scope of TPAP as a stoichio-
tive catalytic conversions based on Ru,[4] Pd,[5] Cu,[6] metric oxidant and aerobic catalyst, little is known about
Co,[7] V,[8] Os[9] catalytic species.
the intimate mechanism of alcohol oxidations.
Adv. Synth. Catal. 2005, 347, 825–832
DOI: 10.1002/adsc.200404365
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
825

Sandro Campestrini et al.
FULL PAPERS
In a mechanistic study[19] on the stoichiometric alcohol fied silica. These catalysts are prepared as previously de-
oxidations by perruthenate anion in water, Lee and scribed[17] by sol-gel entrapment of TPAP in silica glasses
Congson suggested that the oxidation proceeds through obtained using methyltrimethoxysilane (MTMS) and
the formation of an intermediate involving a bond be- tetramethyl orthosilicate (TMOS) as precursors of the
tween Ru(VII) and the a-alcoholic carbon, followed final xerogel. Catalysts with various surface hydropho-
by fragmentation into a radical species which is further bicities are easily obtained by varying the relative
oxidised to the final carbonyl product.
amount of methyl-modified precursor utilised in the
This radical mechanism is also supported by Griffithꢁs synthesis of the xerogels (i.e., 0%, 25%, 75%, 100% of
study on diol oxidations catalysed by perruthenate in the MTMS), whereas catalytic doped glasses with different
presence of sodium bromate (BrO3ꢀ) and dipersulphate pore sizes and entrapment modalities of TPAP are ob-
(S2O82ꢀ) as terminal oxidants in water at alkaline pH.[20] tained by varying the process employed for the xerogels
However, on going from aqueous to organic solvents, preparation, i.e., the amount of water and the presence
the oxidative process seems to be forced to follow a dif- of NaF.[17] The highersuitability of scCO2 as reaction me-
ferent pathway. Indeed, contrary to what happens in dium for aerobic oxidation of alcohols with respect to
aqueous solution,[19] TPAP oxidizes cyclobutanol to cy- conventional liquid solvents of comparable polarity is
clobutanone in organic media in good yields[18] thus sug- revealed by the fact that TPAP encapsulated in un-
gesting that homolytic processes do not play a major modified SiO2 is practically inactive in toluene[16] where-
role, at least in stoichiometric oxidations. Moreover, as it exhibits some activity in scCO2.[17a] Results obtained
Lee has further showed that the stoichiometric oxida- in scCO2 indicate that the best performing catalysts re-
tion of propan-2-ol by TPAP in CH2Cl2 is strongly auto- quire both a high degree of surface hydrophobicity as
catalytic and ascribed this outcome to the coordination well as high amounts of water and co-solvent in the
of unreacted TPAP with colloidal particles of the prod- sol-gel polycondensation process and therefore we uti-
uct (RuO2).[21]
lised in this study a 75% methylated silica prepared ac-
More recently Kozhevnikov has demonstrated that in cording to protocol A (see Experimental Section)
the catalytic oxidation of primary alcohols with oxygen named TP-Me3A.
catalysed by TPAP in toluene, in addition to the expect-
The catalytic oxidations of alcohols were performed in
ed products – namely aldehydes and minor amounts of a stainless steel, 10-mL reaction vessel at a final total
carboxylic acids – the reaction mixtures contain by- pressure of 22.0 MPa and at 758C, in the presence of
þ
products arising from solvent oxidation and NPr4 deg- O2 (0.2–7 bar at 258C) and TP-Me3A as catalyst. The
radation.[22] Furthermore, the catalytic process exhibits molar ratios of the components of the reaction mixtures
an initial faster stage followed by slower kinetics which were approximately TPAP/alcohol/oxygen¼1/10–200/
suggests the involvement of different active ruthenium 50–350; in which the oxygen excess over substrate and
species. In the same study, the authors confirmed the im- the final pressure of 22 MPa were used to ensure a com-
possibility of catalyst recycling for homogeneous aero- plete conversion of benzyl alcohol and complete solubil-
bic oxidation after a maximum TON of ca. 20.
ity of the substrate, respectively.
Extending our mechanistic kinetic studies on the oxi-
The corresponding rate constants (kcat.) were obtained
dation of various alcohols by oxygen in the presence of a from integrated pseudo-first-order plots, i.e.,
TPAP-doped ormosil in scCO2,[17] we reasoned that ln([aldehyde]t¼ 1 ꢀ[aldehyde]t¼t) vs. time, which are
physical segregation of [NPr4þ][RuO4ꢀ] ion pairs into linear up to 80–90% reaction as shown in Figure 1.
the glass framework would prevent catalyst and/or its
derivatives self-aggregation, thus simplifying the kinetic slopes of such a plot. The linearity of the plots confirms
analysis of the system. that catalyst deactivation does not occur, at least in the
In particular, the kcat. values reported represent the
In addition, the utilization of scCO2 as reaction me- time intervals examined. When the highest alcohol
dium allows us to work in a single phase as far as it con- over TPAP ratio was used (ca. 200), a deviation from lin-
cerns the substrate and the terminal oxidant, namely earity is observed, thus suggesting a possible catalyst de-
gaseous oxygen. Based on the new results reported in activation or even oxygen starving. In this case the kcat.
the following, we propose a mechanistic heterolytic was calculated taking into account only the linear part
pathway in which O2 coordinates to the ruthenium cen- of the plot and it can be considered as an initial rate. Ta-
tre and then dehydrogenates the alcoholic substrate.
ble 1 lists the results of some experiments aimed at eval-
uating the reproducibility of the catalytic reactions and
the kinetic order in TPAP.
Results and Discussion
A first, relevant feature of these catalyzed reactions is
the extremely high selectivity observed for the aldehyde
Carbonyl compounds can be efficiently synthesised in formation. A rationale for this outcome is that the highly
scCO2 by alcohol oxidation with molecular oxygen cata- hydrophobic catalyst employed spills out the water
lysed by TPAP-doped ormosils, whose activity is largely formed during alcohol oxidation, thus preventing alde-
superior to perruthenate encapsulated within un-modi- hyde hydration and consequently over-oxidation to car-
826
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 825–832

Alcohol Oxidations with O2 Catalysed by TPAP-Doped Ormosils in scCO2
FULL PAPERS
Table 1. Oxidation of benzyl alcohol (4.83ꢁ10ꢀ2 mmol) with oxygen (1 bar at 258C) catalysed by ormosil TP- Me3A (content
of TPAP: 0.5ꢁ10ꢀ2 mmol per 100 mg), in scCO2 at 758C and 22 MPa.
Entry
TP-Me3A [mg]
kcat. ꢁ102 minꢀ1
kcat. ꢁ102 minꢀ1
Alcohol
conversion [%]
Benzaldehyde
yield [%]
(average)
1
2
3
4
5
6
7
8
100
100
100
20
20
20
4.66
4.13
3.45
0.99
0.64
0.77
0.29
0.056
4.08
0.8
100
100
98.2
97.0
71.0
98.1
96.7
70.5
10
5
Figure 1. Plot of log([aldehyde]t¼ 1 ꢀ([aldehyde]t¼t) vs. time
in the oxidation of benzyl alcohol (4.83ꢁ10ꢀ2 mmol) with
oxygen (1 bar at 258C) catalysed by 20 mg of TP-Me3A
(TPAP content of 1.0ꢁ10ꢀ3 mmol) in scCO2 at 22 MPa and
758C.
Figure 2. Kinetic order in catalyst relative to benzyl alcohol
oxidation with O2 (1 bar at 258C) catalysed by TP-Me3A
(content of TPAP: 0.5ꢁ10ꢀ2 mmol per 100 mg ormosil), in
scCO2 at 758C, and 22 MPa.
pared to traditional solid catalysts obtained by heteroge-
boxylic acid. Moreover, scCO2 may help in water remov- nisation of a materialꢁs external surface, which in gener-
al through its well known ability of drying solid matrices. al do not show a first order law dependence for catalyst.
Entries 1–3 and 4–6 in Table 1 indicate a fairly good
Kinetic data for the oxidation of benzyl and several
agreement in the measurement of pseudo-first-order other alcohols under different experimental conditions
constants in repeated experiments. In fact, it should be are shown in Table 2.
noted that the oxidative process is entirely heterogene-
Entries 9–11 indicate a linear dependence of initial
ous, with the leach-proof catalyst being immiscible in the velocities (Ro) for benzyl alcohol oxidation versus sub-
supercritical CO2 phase; therefore, some deviations strate concentrations (Figure 3), at least for the lowest
could be expected due, for instance, to the non-homoge- values, thus suggesting first order kinetics in substrate.
neity in the catalyst particles used in different experi-
ments.
At the highest alcohol concentration investigated, a
negative deviation from linearity is observed, which
A clear activity enhancement is observed upon sol-gel might point to substrate coordination to the catalyst.
encapsulation since a 1 mol % catalytic amount of en- Unfortunately, the limited alcohol solubility in the reac-
trapped TPAP can now be used to afford high yields of tion medium limits the exploration of a wider concentra-
benzaldehyde (entry 7), while with homogeneous per- tion range.
ruthenate in solution a minimum of 10 mol % is re-
Entries 12–14, referring to experiments carried out at
quired to observe optimal yields.[10,11] Indeed, entry 8 varying partial oxygen pressures, show a negative de-
has a turnover number of 140 which is one of the highest pendence of rate constants on oxygen concentration (In-
thus far reported for an Ru-based aerobic catalyst.
sert in Figure 3). Comparison between entries 14, 16 and
Finally, the first order kinetics in catalyst (Figure 2) 19, 20 shows the higher reactivity ofprimary with respect
shows another unique feature of sol-gel entrapped com- to secondary hydroxyl functions in the oxidation of both
Adv. Synth. Catal. 2005, 347, 825–832
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
827

Sandro Campestrini et al.
FULL PAPERS
Table 2. Oxidation of various alcohols with oxygen catalysed by TP-Me3A (content of TPAP: 0.5ꢁ10ꢀ2 mmol each 100 mg of
ormosil) in scCO2 at 758C and 22 MPa.
Entry
Substrate [mmolꢁ102]
TP-Me3A [mg]
PO2 [bar]
kcat. ꢁ104 [minꢀ1
]
Ro ꢁ106 [M minꢀ1
]
9
10
11
12
13
14
15
16
17
18
19
20
21
Benzyl alcohol [9.66]
Benzyl alcohol [4.83]
Benzyl alcohol [1.21]
Benzyl alcohol [4.83]
Benzyl alcohol [4.83]
Benzyl alcohol [4.83]
1-D-Benzyl alcohol [4.83]
1-Phenylethanol [4.83]
3-Chlorobenzyl alcohol [4.83]
4-Methylbenzyl alcohol [4.83]
1-Octanol [4.83]
5
5
5
5
5
10
10
10
10
10
10
10
10
1.0
1.0
1.0
0.21
7.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
4.74
5.63
4.43
8.56
4.43
29.40
17.60
10.2
31.4
19.3
7.0
5.72
3.40
0.67
5.16
2.68
17.75
10.65
6.16
18.96
11.65
4.23
1.75
n.r.
2-Octanol [4.83]
Benzyl methyl ether [4.80]
2.9
n.r.
ful substrates, i.e., p-nitro- and p-methoxybenzyl alco-
hols.
The rate constants determined by individual experi-
ments may be correlated in a Hammett plot with the ap-
propriate sigma values. The linear correlation among
the three points is rather poor (R¼0.82) and therefore
we will not speculate on the absolute value of the result-
ing Hammett rho value (þ0.34), nor on the significance
of curved Hammett plot. However, the positive sign of
rho is worthy of note which is, in effect, opposite to
that found in benzyl alcohol oxidations with oxygen cat-
alyzed by metallic Ru,[4] thus corroborating the hypoth-
ꢀ
esis that in our system RuO4 is the active species
throughout the catalytic process.
Moreover, the partial negative charge developed on
the benzyl carbon in the transition state nicely matches
the peculiar reactivity of TPAP towards primary alco-
hols, indicating that the superior reactivity of primary al-
cohols is due to electronic rather than steric factors.
Entries 14 and 15 in Table 2 show that benzyl alcohol
and its isotopically marked derivatives react at different
rates, thus determining a kinetic isotopic effect (KIE) of
Figure 3. Kinetic order in substrate and oxygen (insert) for
benzyl alcohol oxidation with O2 catalysed by TP-Me3A
(content of TPAP: 0.5ꢁ10ꢀ2 mmol per 100 mg of ormosil),
in scCO2 at 22 MPa and 758C.
benzyl and alkyl alcohols, which is a well established fea- 1.67. This value is rather small for a primary KIE, but it
ture of TPAP in stoichiometric oxidations.[11,15] A corol- should be considered that the isotopically marked sub-
lary of this outcome is that the active cꢀatalytic species strate used, namely PhCHDOH, still bears a hydrogen
should be the perruthenate anion RuO4 and not other atom on the alkyl carbon and therefore it can lose either
ruthenium species produced in the course of the process. the deuterium atom giving PhCHO, or the hydrogen, af-
Interestingly, under the experimental conditions fording PhCDO. As a consequence, the deuterium sub-
where smooth alcohol oxidation is observed, the oxida- stitution is reflected only partially in the kinetic meas-
tion of benzyl methyl ether (run 21) does not occur, sug- urement leading to a KIE value lower than expected.
gesting that substrate coordination to TPAP takes place
In order therefore to get a more realistic KIE, we de-
ꢀ
through the formation of an Ru O bond rather than an termined the extent of hydrogen removal relative to
ꢀ
Ru C bond.
deuterium by the ratio PhCDO/PhCHO formed in the
In order therefore to gather further information on course of PhCHDOH oxidation assessed by mass spec-
the intimate mechanism of alcohol dehydrogenation, trometry. This would be easily done by considering the
the effect of a few substituents on the oxidation rate of mass peak generated only by PhCHO and a second
different benzyl alcohols has been studied (entries 14, one generated by PhCDO only. Unfortunately, this is
17, 18). The number of derivatives examined was limited the case only for the 106 mass peak (MPhCHO) generated
by the scarce solubility in scCO2 of other potentially use- by PhCHO, while all other peaks are resulting from the
828
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 825–832

Alcohol Oxidations with O2 Catalysed by TPAP-Doped Ormosils in scCO2
FULL PAPERS
Table 3. Values of PhCDO/PhCHO determined by means of
two different mass peaks ratios [105/106¼(Dꢀ2)/M and 107/
106¼D/M] in various samples taken at appropriate times.
60’ 169’ 236’ 300’ 1303’ 1412’ 1502’ Average
(Dꢀ2)/M 1.89 1.92 1.91 1.86 1.91 1.86 1.83 1.88
D/M
2.13 2.14 2.14 2.13 2.12 2.08 2.07 2.12
contributions of both species. However, it is still possible
to separate the two contributions by knowing the contri-
bution of pure PhCHO. By solving these calculations for
the 107 (MPhCDO) and 105 (MPhCDOꢀ2) mass peaks and for
seven samples of the reaction mixture taken at various
times, two sets of KIE values were obtained (Table 3).
In particular, from the 105/106 ratio an average KIE of
1.88 is obtained while from the 107/106 ratio the average
value calculated is 2.12, giving an overall average of 2.0.
The value thus determined is, as predicted, higherthan
that measured through the ratio of the kinetic constants,
and should be compared with the maximum theoretical
Figure 4. Dependence of normalised rate constants for benzyl
alcohol (4.83ꢁ10ꢀ2 mmol) oxidation with O2 (1 bar) cata-
lysed by TP-Me3A (5.0ꢁ10ꢀ3 mmol) in scCO2 at 22 MPa
and 758C, as a function of catalyst recycling and washing sol-
&
!
*
vent ( : ethyl ether; : n-hexane; : dichloromethane).
value for a primary KIE of 4.6 at 758C. A KIE value of ported TPAP does not dissolve at all, the efficiency de-
2.0 clearly indicates that hydrogen abstraction from the cay is noticeably slower.
a carbon occurs in the rate-determining step of alcohol
oxidation.
These observations seem to corroborate the hypoth-
esis of a TPAP inactivation caused by self-aggregation
The results of the present work with sol-gel entrapped of Ru derivatives induced by the solvent, even though
TPAP further demonstrate the validity of sol-gel encap- a chemical interaction with the solvent cannot be ruled
sulation as a means for preparing highly efficient cata- out. In fact, when the catalyst is recovered and washed
lysts and of scCO2 as advantageous alternatives both with CH2Cl2, a fast development of gas bubbles is in-
to chemical tethering[23] for the heterogenisation of per- duced and indeed a pungent smell of chlorine can be de-
ruthenate as well as to volatile organic solvents which, as tected during desiccation. This is not the case with car-
typical VOCs, upon reaction must be disposed of ac- bon dioxide, which is a superior reaction medium also
cording to strict regulations. In fact, while the reactant due to its chemical inertness and is easily vented off by
alcohol and O2 molecules dissolved in scCO2 easily enter simple pressure release in the reaction system, leaving
the vast xerogel inner porosity (0.24 cm3/g) where they the solid catalyst physically and chemically unmodified,
react at the catalytic centre, TPAP itself remains insolu- ready for reuse.
ble in scCO2 and is therefore scarcely prone to any of the
self-aggregation phenomena which are typical of Ru- tives should be taken into consideration in assuming a
mediated oxidations. plausible mechanism for alcohol oxidations with O2 cat-
From a general mechanistic point of view, two alterna-
In the attempt to elucidate the pathway of TPAP inac- alysed by perruthenate. In particular, the oxidation may
tivation, we thus carried out catalyst recycle experi- proceed either through a radical or a bielectronic, heter-
ments recovering the catalyst from scCO2 medium and olytic mechanism. In addition, the coordination of the
washing it with various organic solvents.
substrate to the catalyst may or may not be a prerequi-
Depending on the solvent utilised, the results in Fig- site in both pathways.
ure 4 show that after only 10 TONs performed in the
The lack of any by-products arising from a hypothetic
course of the first experiment, the catalyst retains only radical, together with a significant absolute Hammettꢁs
10–40% of the initial efficiency in the second experi- rho value, seem to rule out a free radical mechanism.
ment (first recycle). Such an efficiency drop continues As far as the involvement of an intermediate is con-
with further recycling and after a total of five runs cerned, the negative deviations from linearity in Ro ver-
(four recycles) the catalyst becomes inactive.
sus substrate concentration might be indicative of a Mi-
Interestingly, the rate of catalyst degradation depends chaelis–Menten type behaviour as well as of catalyst de-
on the nature of the solvent used for recovering and activation or even of the reached limit of substrate solu-
washing the active material. In dichloromethane, which bility. Hence, the oxidative addition of an alcohol to an
is a good solvent for unsupported TPAP, the inactivation oxo-metal double bond to give an intermediate product
ꢀ
ꢀ
takes place very quickly and already after the first recy- may involve either the O HorC Hbond(seeScheme 1).
cle the catalyst is almost inactive. On the other hand, in It has been reported that alcohol oxidations by ruth-
diethyl ether and even more in hexane, where unsup- enate and perruthenate in basic aqueous solutions pro-
Adv. Synth. Catal. 2005, 347, 825–832
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
829

Sandro Campestrini et al.
FULL PAPERS
same homogeneous catalysts used in organic solvent,[16]
in terms of selectivity, stability and activity.
The origin of the higher activity is probably related to
the limited mobility of TPAP molecules within the cata-
lyst cage filled by scCO2, which prevents self-aggrega-
tion of ruthenium derivatives formed in the oxidation
process. The drying power of scCO2, in association
with the very hydrophobicnature ofthe catalyst, induces
fast water removal from the inner cavities of the catalyst
thus affording an extremely high selectivity for aldehyde
Scheme 1.
ceeds through via path (b) of Scheme 1.[21] In our system, (or ketone) production.
however, the remarkable difference in reactivity be-
The kinetic analysis of the system reveals that an ester-
tween benzyl alcohol and the corresponding methyl type intermediate (RCH2-O-Ru(OH)O3ꢀ), similar to
ether clearly demonstrates that an organometallic deriv- that proposed for Cr(VI) alcohol oxidation, is involved
ative is not involved in the catalytic process. In fact, such in the catalytic process. Furthermore, the peculiar effect
a difference cannot be attributed to a steric effect since of oxygen pressure suggests that the role of the terminal
the secondary alcohol 1-phenylethanol (run 16), bearing oxidant is not to simply reoxidise the reduced form of
steric hindrance similar to that of benzyl methyl ether, is the Ru catalyst but, on the contrary, rather to dehydro-
smoothly oxidised under the same experimental condi- genate the alcohol coordinated to the metal. The pri-
tions.
As a consequence, the difference in reactivity must the rate-limiting step of the overall process.
originate from the different ability of the two substrates The recognition of the main parameters governing the
mary KIE eventually indicates that dehydrogenation is
in generating a reactive intermediate via (a) of oxidative process is of paramount importance for fur-
Scheme 1. Interestingly, the kcat. dependence on oxygen ther improvements which will be reached by appropri-
concentration shows a negative trend, similarly to what ate modifications of the chemical-physical factors which
happens in aerobic alcohol oxidations catalysed by met- have been elucidated in this study. Among these param-
allic Pd,[24] which is generally ascribed to the so-called eters, the possibility of further increasing TPAP segrega-
“catalyst over-oxidation”, a phenomenon probably in- tion must be duly considered in the design of new, more
volving the formation of palladium oxide on the metal durable, Ru-based sol-gel catalysts[25] for industrial
surface which inactivates the catalyst.
aerobic oxidations in scCO2.
On the other hand, in our system such a rationale can-
not be invoked: in fact, in RuO4ꢀ the Ru metal centre is
already present in one of its highest oxidation states. An
alternative rationale, accounting for all the experimen-
tal findings, is that the alcohol oxidation is actually car-
ried out by the oxygen in the coordination sphere of
the metal centre. Such a ternary complex involving per-
ruthenate, alcohol and oxygen reasonably accommo-
dates in a single general mechanistic picture the first or-
der kinetics in catalyst and the less-than-first order ki-
netics in both, alcohol and oxygen.
In fact, the inhibitory effect of high concentrations of
both oxygen and alcohol might be the result of a de-
crease in the concentration of the effective 1:1:1 ternary
complex, which leads to carbonyl products by an intra-
molecular hydride abstraction. Moreover, the occur-
rence of a primary KIE indicates that hydrogen abstrac-
tion from the a-carbon within the hypothesised inter-
mediate is the rate-determining step of the overall proc-
ess.
Experimental Section
Materials
1-D-Benzyl alcohol was prepared from benzaldehyde by re-
duction with NaBD4 in refluxing ether and purified by chroma-
tography (silica gel, CH2Cl2). Methyl benzyl ether was pre-
pared from benzyl alcohol by methylation with methyl iodide
in the presence of KOH in dimethyl sulphoxide and purified
by chromatography (silica gel, dichloromethane:hexane,
1:1). Methyltrimethoxysilane (MTMS), tetramethyl orthosili-
cate (TMOS), tetra-n-propylammonium perruthenate
(TPAP), benzyl alcohol, 3-chlorobenzyl alcohol, 4-methylben-
zyl alcohol, 1-methylbenzyl alcohol, 1-octanol, 2-octanol,
methanol, n-dodecane, were purchased from Sigma-Aldrich
and used without further purification. Ultra-pure water (Milli-
pore Type 1 quality) was used in all the preparations.
Catalyst Preparation
A typical catalyst, TP-Me3A of batch A, obtained in the ab-
sence of NaF as polymerisation catalyst, was prepared by add-
ing MTMS (4.90 mL) and TMOS (1.95 mL) to a solution of
TPAP (55.5 mg) in MeOH (1.80 mL) cooled in an ice bath fol-
lowed by the addition of H2O (5.70 mL) under fast stirring. The
Conclusions
In conclusion, we have demonstrated that in aerobic ox-
idation of alcohols in scCO2 TPAP-doped ormosils ex-
hibit a superior catalytic performance relative to the sol gelled rapidly and the resulting alcogel was sealed and left
830
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 825–832

Alcohol Oxidations with O2 Catalysed by TPAP-Doped Ormosils in scCO2
FULL PAPERS
to age at room temperature for 2 days prior to drying in an oven
at 508C until reaching constant weight (5 days). The catalytic
xerogel thus obtained was powdered, washed under reflux
(CH2Cl2 ꢁ2, 408C) and dried at 508C before use.
Hall, London, 1995; b) P. T. Anastas, M. M. Kirchhoff,
Acc. Chem. Res. 2002, 35, 686.
[3] R. A. Sheldon, I. C. W. E. Arends, G. J. ten Brink, A.
Dijksman, Acc. Chem. Res. 2002, 35, 774.
[4] K. Yamaguchi, N. Mizuno, Chem. Eur. J. 2003, 9, 4353.
[5] a) B. A. Steinhoff, S. S. Stahl, Org. Lett. 2002, 4, 4179;
b) G. J. ten Brink, I. W. C. E. Arends, R. A. Sheldon,
Adv. Synth. Catal. 2002, 344, 355.
Oxidation Procedure in scCO2
The apparatus for conducting reactions in scCO2 was described
in detail elsewhere.[26] Catalytic oxidations of alcohols in scCO2
were carried out in a 10-mL cylindrical high-pressure vessel. At
the entrance and exit of the vessel were placed two 5-mM pore
frits. The vessel was fitted with a magnetic bar for stirring, cat-
alyst (5–100 mg), substrate (5 mL), and internal standard (5 mL
n-dodecane) were charged, after which it was purged with air
(for reactions performed at 0.20 bar O2) or pure gas chromato-
graphic-grade O2 (for reactions performed at 1.0 bar O2); loss
of reactant and standard is negligible in the feeble stream of
gases utilised, thanks to their high boiling points (2168C for
n-dodecane and 2058C for benzyl alcohol). The reactor was
thus sealed and placed into the SFC oven thermostatted at
758C. Pressurisation is achieved in about three minutes by
pumping liquid carbon dioxide to the desired final pressure
(22.0 MPa) while the reaction mixture was kept under stirring
at 400 rpm (by means of an alternating magnetic field stirrer).
Control experiments showed that a complete dissolution of al-
cohols and n-dodecane in scCO2 occurs in about 10–15 min. At
appropriate time intervals, reaction samples were withdrawn
through the stainless steel restrictor maintained at 908C by
means of a 6-way HPLC valve, and thus trapped in dichlorome-
thane. The samples were analysed by GLC (Hewlett Packard
6890 system) on a 30 m EC-1000 capillary column (I.D.:
0.25 mm, thickness of film: 0.25 mM). The product concentra-
tion was measured using a previously determined response fac-
tor.
[6] B. Betzemeier, M. Cavazzini, S. Quici, P. Knochel, Tetra-
hedron Lett. 2000, 41, 4343.
[7] V. B. Sharma, S. L. Jain, B. Sain, Tetrahedron Lett. 2003,
44, 383.
[8] Y. Maeda, N. Kakiuchi, S. Matsumura, T. Nishimura, T.
Kawamura, S. Uemura, J. Org. Chem. 2002, 67, 6718.
[9] C. Dçbler, G. M. Mehltretter, U. Sundermeier, M. Eck-
ert, H. C. Militzer, M. Beller, Tetrahedron Lett. 2001,
42, 8447.
´
[10] I. E. Marko, P. R. Giles, M. Tsukasaki, I. Chelle-Reg-
naut, C. J. Urch, S. M. Brown, J. Am. Chem. Soc. 1997,
119, 12661.
[11] R. Lenz, S. V. Ley, J. Chem. Soc. Perkin Trans. 1 1997,
3291.
[12] B. Hinzen, R. Lenz, S. V. Ley, Synthesis 1998, 977.
[13] S. V. Ley, C. Ramarao, M. D. Smith, Chem. Commun.
2001, 2278.
[14] For several examples, see, for instance: a) J. Blum, D. Av-
nir, H. Schumann, Chemtech 1999, 29, 32; b) D. Avnir,
L. C. Klein, D. Levy, U. Schubert, A. B. Wojcik, “Orga-
no-silica sol-gel materials” in: The Chemistry of Organic
Silicon Compounds, Vol. 2, Part 3, (Eds.: Z. Rappoport,
Y. Apeloig), John Wiley & Sons: New York, 1998, pp.
2317–2362.
[15] A. Bleloch, B. F. G. Johnson, S. V. Ley, A. J. Price, D. S.
Shepard, A. W. Thomas, Chem. Commun. 1999, 1907.
[16] R. Ciriminna, M. Pagliaro, Chem. Eur. J. 2003, 9, 5067.
[17] a) R. Ciriminna, S. Campestrini, M. Pagliaro, Adv. Synth.
Catal. 2003, 345, 1261; b) R. Ciriminna, S. Campestrini,
M. Pagliaro, Adv. Synth. Catal. 2004, 346, 231; since
2002, the British chemicals manufacturer Thomas Swan
& Co. operates in UK a dedicated plant for highly effi-
cient, heterogeneous catalytic production in scCO2 of a
variety of valuable fine chemicals obtained with unprece-
dented purity grades. See also at the URL: http://
www.thomas-swan.co.uk.
Determination of PhCDO/PhCHO Ratios
PhCDO/PhCHO ratios were determined by GC-MS analyses
of various reaction mixture samples relative to the oxidation
of PhCHDOH, carried out with a Hewlett-Packard 5890 gas
chromatograph connected with a Hewlett-Packard 5970 mass
selective detector using a 15 m SE-30 capillary column,
0.25 mm i.d. The 106 mass peak was taken as being representa-
tive of the PhCHO concentration. The 105 and 107 mass peaks,
having taken into account the PhCHO contribution, are repre-
sentative of the PhCDO concentration. Both 105/106 and 107/
106 ratios represent a measure of the PhCDO/PhCHO ratio
and therefore a measure of the hydrogen removal grade rela-
tive to deuterium.
[18] W. P. Griffith, Chem. Soc. Rev. 1992, 179.
[19] D. G. Lee, L. N. Congson, Can. J. Chem. 1990, 68, 1774.
[20] A. J. Bailey, W. P. Griffith, S. I. Mostafa, P. A. Sherwood,
Inorg. Chem. 1993, 32, 268.
[21] D. G. Lee, Z. Wang, W. D. Chandler, J. Org. Chem. 1992,
57, 3276.
[22] M. Hasan, M. Musawir, P. N. Davey, I. V. Kozhevnikov, J.
Mol. Catal. A 2002, 180, 77.
References and Notes
[23] Perruthenate tethering to the inner poreꢁs surface of a
modified mesoporous silicate MCM-41 as described in
ref.[15] has been shown to be an effective strategy to ach-
ieve a re-usable heterogeneous catalyst. Such a compo-
site material provides superior catalytic performance
probably thanks to the reduced possibility of forming ox-
idation-inactive colloidal RuO2.
[1] T. Fey, H. Fischer, S. Bachmann, K. Albert, C. Bolm, J.
Org. Chem. 2001, 66, 8154.
[2] Normally, 15 to 20 kg of waste are produced along with
1 kg of useful product. For example, more than 8ꢁ105
tons of C4 ketones and 2ꢁ106 tons of C1 – C3 aldehydes
are produced annually world-wide, see also: a) J. H.
Clark, Chemistry of Waste Minimisation, Chapman &
Adv. Synth. Catal. 2005, 347, 825–832
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
831

Sandro Campestrini et al.
FULL PAPERS
[24] a) G. Jenzer, D. Sueur, T. Mallat, A. Baiker, Chem. Com-
mun. 2000, 2247; b) G. Jenzer, T. Mallat, A. Baiker, Cat-
al. Lett. 2001, 73, 5; c) M. Besson, P. Gallezot, Catal. To-
day 2000, 57, 127; d) J. D. Grunwaldt, M. Caravati, M.
Ramin, A. Baiker, Catal. Lett. 2003, 90, 221.
[25] Sol-gel silica-entrapped metal catalysts are now commer-
cially available. Recently, the British fine chemicals man-
ufacturer Avecia licensed Johnson Matthey sol-gel tech-
nology for preparing organic-inorganic silica hybrid gels
doped with chiral ligands named CTIS CACHyTM (CTIS:
Chiral Technologies Interface System, CACHy: Catalytic
Asymmetric CyanoHydrin) which allow efficient large-
scale enantioselective conversion of aldehydes and ke-
tones into chiral cyanohydrins for the pharmaceutical in-
dustry. For a recent account on the applications of the
sol-gel catalyst technology in organic chemistry, see
also: R. Ciriminna, M. Pagliaro, Curr. Org. Chem. 2004,
18, 1851.
[26] S. Campestrini, U. Tonellato, Adv. Synth. Catal. 2001,
343, 819.
832
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 825–832