Silica-immobilized chromium colloids for cyclohexane autoxidation

Article abstract of DOI:10.1002/anie.200602736

(Figure Presented) In a stable condition: Oxyhydroxide colloids of chromium deposited on silica gel are selective, stable catalysts for the autoxidation of cyclohexane (CyH). The Cr^III colloids, generated in situ, are immobilized by a column precipitation chromatographic technique (see picture). An increased yield of cyclohexanone (QO) is obtained.

Full text of DOI:10.1002/anie.200602736

Communications
Chromium Colloids
DOI: 10.1002/anie.200602736
Silica-Immobilized Chromium Colloids for
Cyclohexane Autoxidation**
Eric Breynaert, Ive Hermans,* Bert Lambie,
Guido Maes, Jozef Peeters, AndrØ Maes, and
Pierre Jacobs
The selective oxidation of cyclohexane (CyH) with molecular
oxygen still constitutes a major challenge for chemists.[1] The
synthesis of cyclohexanone (QO) is of special interest,[2] as it
is a key chemical for the production of polyamides such as
nylon 6 and nylon 6,6. It is generally recognized that the
thermal reaction of CyH with O2 proceeds via a complex
radical chain mechanism, with a limit of 5% on the
conversion to prevent overoxidation of the desired prod-
ucts.[3–5] Currently, the industrial autoxidation of CyH is often
carried out in the presence of small amounts of dissolved
cobalt salts.[3,4] Cobalt ions are able to assist in the generation
of new radicals from the primary chain-propagation product,
cyclohexyl hydroperoxide (CyOOH), via a Haber–Weiss
catalytic cycle [reactions (1) and (2)].[1,3–8] This catalytic
decomposition is indeed much faster than thermal initiation
through a bimolecular reaction of CyOOH with CyH or
[9]
QO.
CyOOH þ Co2þ ! CyOC þ Co3þOHÀ
ð1Þ
ð2Þ
CyOOH þ Co3þOHÀ ! CyOOC þ H2O þ Co2þ
Chromium, however, is known to catalyze additionally the
dehydration of CyOOH to QO, which results in a much more
favorable product distribution.[3,4] Our recently developed
insights into the mechanism of CyH autoxidation[10–12]
prompted us to look at the opportunities for using Cr as a
catalyst. The earlier mechanistic view,[3–5] which assumes that
QOis responsible for the majority of by-products, is at odds
with our results that stem from a combined experimental and
theoretical approach to the subject.[10,11]
[*] E. Breynaert, Dr. I. Hermans, Prof. Dr. A. Maes, Prof. Dr. P. Jacobs
Centrum voor Oppervlaktechemie en Katalyse
K. U. Leuven
Kasteelpark Arenberg 23, 3001 Heverlee (Belgium)
Fax: (+32)16-321-998
E-mail: ive.hermans@biw.kuleuven.be
B. Lambie, Prof. Dr. G. Maes, Prof. Dr. J. Peeters
Departement Chemie
K. U. Leuven
Celestijnenlaan 200F, 3001 Heverlee (Belgium)
[**] This work was performed in the frame of IAP, GOA, IDECAT, and
CECAT projects. E.B., I.H., and B.L. acknowledge fellowships from
K. U. Leuven, FWO-Vlaanderen, and IWT, respectively.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
7584
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7584 –7588

Angewandte
Chemie
On the one hand, QO was found to be barely more
reactive than CyH, which indicates that its overoxidation can
only account for a small fraction of the by-products. On the
other hand, CyOOH was identified as being very reactive
towards the chain-carrying peroxyl radicals, on account of its
weakly bonded a-H atom.[13] This hitherto overlooked,
though very fast, a-H abstraction results in the formation of
known that polyvalent ions can induce flocculation and
facilitated deposition in a colloidal system,[20] and that sulfate
ions can influence the CrIII hydrolysis,[21] CaSO4 was premixed
with the dry silica support. The synthesis was carried out in a
CPC system characterized by a high collector-particle radius
and superficial velocity, in combination with a column flow
reversal cycle, to achieve a homogeneous distribution of
QO, cyclohexanol (CyOH), and the majority of the by- chromium. After chromium deposition, the column packing
products (Scheme 1).[10,11] It can therefore be rationalized that
was vacuum-dried to remove water and residual hydrazine.
Dynamic light scattering (DLS) analysis of a liquid sample
from the recirculation reactor reveals the presence of 6-nm
particles that are reasonably stable in time (Figure 1). These
nanosized Cr colloids are fixed on the silica gel by the CPC
technique, and they remain invisible by SEM. It should be
emphasized that our synthesis procedure yields nanoparticles
that are much smaller than those commonly reported.[21,22]
Scheme 1. Formation of QO, CyOH, and by-products via fast propaga-
tion reactions in the CyH autoxidation.[10,11]
the catalytic dehydration of CyOOH to QO can boost the
selectivity of the reaction. Equally important is that the use of
Cr could eliminate the need for a separate decomposition step
of CyOOH. However, given its noxious nature, immobiliza-
tion of Cr is an important prerequisite to its industrial
breakthrough.
Retention of (isomorphically) substituted Cr in Cr/
AlPO4-n under autoxidation conditions was difficult.[1,14]
Despite several attempts, the heterogeneous nature of the
active site in Cr/AlPO4-n or Cr/silicalite-like materials could
not always be established unambiguously. Indeed, rigorous
testing for the heterogeneity of so-called heterogeneous Cr
catalysts is crucial, as many of these materials slowly release
Cr under oxidizing conditions.[14] For the conventional Cr/
AlPO4-5, experimental evidence was found that the Cr is not
incorporated in the framework, but rather is present as
octahedral CrIII ions at the surface of the zeolite crystal.[15]
However, it cannot be excluded that changing the synthesis
conditions[16,17] might result in stable, supported, Cr-based
autoxidation catalysts. The performance of recently discov-
ered three-dimensional mesoporous chromium oxide in
autoxidation reactions remains to be examined.[18] Alterna-
tively, Cr immobilization is possible by covalent anchoring of
its complexes to a solid support.[19] Although these materials
Figure 1. Particle size distribution as determined by DLS. Liquid
sample withdrawn from the recirculation reactor without passing over
the CPC column, measured after a) 4, b) 8, and c) 12 days (initially
filtered over a 0.22-mm filter to remove dust particles).
The UV/Vis spectra of the green aqueous phase (Figure 2)
are characteristic of octahedral Cr3+ ions.[23] The Cr speciation
in solution does not change significantly over the range of free
chromium concentrations used (10–40 mm).[24] The average
extinction coefficients of (23 ꢀ 5) and (17 ꢀ 5)mÀ1 cmÀ1 at 422
À
are more resistant to leaching, labile C H bonds in the
organic spacer can be slowly oxidized under the severe
autoxidation conditions.
Herein, a new approach to the immobilization of Cr is
presented, aimed at the design of stable autoxidation
catalysts. Use is made of a novel column precipitation
chromatographic (CPC) technique, which was recently devel-
oped to quantify oxidic eigencolloids of TcIV.[20] CrIII eigen-
colloids, generated in situ by reduction of dosed K2Cr2O7 with
hydrazine, were precipitated onto a silica support in a
chromatography column. The deposition of colloids flowing
over a support column is interpreted mechanistically as a
combination of coagulation of the colloids and heterocoagu-
lation and heteroadagulation onto the packed bed.[20] As it is
Figure 2. UV/Vis spectra of the solution phase depending on the
amount of K2Cr2O7 added to the CPC system (0.25, 0.50, 0.75,
1.5 mmol).
Angew. Chem. Int. Ed. 2006, 45, 7584 –7588
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7585

Communications
and 582 nm, respectively,[24] are in good agreement with
Table 1: Product yield at 3% conversion during the autoxidation of CyH
at 403 K.
reported values for (oligomeric) chromium species.[25]
Column precipitation[20] is the dominant immobilization
mechanism, as the achieved loadings of up to 13 mggÀ1 are
higher than the maximal loading one could possibly obtain
with ion exchange.[24,26] The diffuse reflectance spectra (DRS)
of the dry chromium-loaded solids[24] are very similar to the
liquid-phase UV/Vis spectra and also reveal the presence of
octahedral Cr3+ species.[23]
Experiment
QO
CyOH
CyOOH
By-products
pure[a]
23.5
60.0
37.5
31.0
24.3
50.5
40.0
1.2
1.5
5.5
14.5
10.5
new Cr catalyst[b]
Co catalyst[c]
[a] Pure CyH autoxidation, reaction time=1150 min. [b] Cr-catalyzed
autoxidation (30 ppm Cr as 13.0 mg Cr per g silica gel), reaction time=
150 min. [c] Co-catalyzed autoxidation (homogeneous reaction with
5 ppm CoII octoate (2-ethylhexanoate), ICN Pharmaceuticals), reaction
time=150 min.
The X-band EPR spectra of Cr-loaded silica gel and a
liquid aliquot from the recirculation reactor are analogous
(Figure 3 a and b). The intense and broad b signal around g =
2 is characteristic for Cr3+ ions surrounded by six oxygen
CyOOH,[30] also explaining the rather remarkable linear O2
consumption (Figure 5a). Normally the CyOOH concentra-
tion initially increases almost exponentially,[9,10] thus causing a
large increase in the rate of initiation, which results in an
exponential increase in the O2 consumption. The high QO
yield relative to that of by-products compensates the slightly
higher by-product yield that results from overoxidation of
QO.
Immediately after a catalytic run, an EPR analysis of the
catalyst with 1.60 mgCrgÀ1 reveals the presence of an addi-
tional sharp g signal (Figure 3c), often assigned to a single CrV
species.[23,28] However, other studies attribute this signal to a
mixed-valence trinuclear cluster of the type CrVIOCrIIIOCrVI
(pseudo Cr5+).[31] Such in situ generated CrVI species could be
the actual active sites of the catalyst. Indeed, a study of the
CyOOH dehydration activity of chromium demonstrated the
formation of an intermediate cyclohexyl peroxychromate
ester between CyOOH and CrVI, which is able to decompose
to QO.[30] When the used catalyst is stored for a few days in a
desiccator, the sharp g signal disappears, which suggests that it
corresponds to a very reactive Cr species.
Figure 3. EPR spectra of a) a liquid sample withdrawn from the CPC
reactor, b) Cr-loaded silica gel (1.60 mg Cr per g gel), and c) sample (b)
after an autoxidation reaction at 403 K. All spectra were recorded at
110 K (9.588 GHz).
To evaluate the stability of the catalyst, the solid was
separated at 353 K from the liquid phase after approximately
5% conversion (reaction at 403 K), to avoid re-adsorption of
potentially leached Cr species. The liquid phase was checked
for Cr by electrothermal vaporization (ETV) atomic absorp-
tion spectroscopy (AAS) and the concentration was below
5 ppb for all tested samples. Important is the fact that the
same amount of Cr was measured when the catalysts were
stirred for the same time at the same temperature in CyH,
without O2. Meanwhile, DLS analysis of this CyH liquid
revealed particles of several hundred nanometers, probably
caused by abrasion and attrition of the silica powder in the
batch reactor.
This finding presents strong evidence that the new Cr
catalyst is chemically stable under the current process
conditions, and that the very small amount of Cr detected in
the liquid phase is a result of physical impact of the stirrer on
the material. It is expected that these effects will be absent if
the reaction is carried out in a tubular reactor. Anyway, the
(remaining) activity of the supernatant at 403 K can be safely
attributed to the oxidation-enhancing nature of the products.
Indeed, Figure 4 illustrates that an artificial autoxidation
mixture, devoid of any chromium and mimicking the super-
natant composition, exhibits the same activity, thus pointing
to the heterogeneous nature of the new colloid-based system.
ligands in an octahedral coordination.[23,28] This finding, in
combination with the UV/Vis and DRS information, indicates
that the chromium nanoparticles identified in solution are
immobilized on the silica support. No change in the surface
area upon Cr loading was observed with N2 adsorption.[29] X-
ray diffraction showed no sign of a crystalline chromium oxide
phase, and small-angle X-ray scattering (SAXS) was unable
to characterize nanoparticles because of the structure of the
silica support (broad peak below 2q = 88).[24]
The catalytic performance of the new Cr-on-silica catalyst
was evaluated in the autoxidation of CyH with 30 ppm of Cr,
at a temperature of 403 K where thermal autoxidation is very
slow. It was found that the new material (13.0 mgCrgÀ1) is
slightly more active than homogeneous [Cr(acac)3] (acac =
acetylacetonate; Dp(O2) after 300 min equals 1.6 and
1.4 MPa, respectively), which should be ascribed to an ideal
dispersion of Cr colloids with a high surface area. The
turnover frequency (TOF) of the immobilized Cr species was
fairly independent of their loading. The CyOOH concentra-
tion remained very low throughout the reaction,[24] compared
to that in the pure autoxidation reaction at the same
temperature (Table 1).[11,24] This low concentration is the
result of the catalytic dissociation and catalytic dehydration of
7586
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7584 –7588

Angewandte
Chemie
Experimental Section
In the CPC system, the content of a 100-mL stirred reactor was
recirculated over a 25-mm-diameter column packed with silica
powder (20 g) containing CaSO4·2H2O(10 wt%, dry premixed), at
a flow rate of 20 mLminÀ1. Silica 60 from Fluka with a particle size of
40–63 mm and a BET specific surface of 440 m2 gÀ1 was used as inert
support material. A flow reversal was automatically initiated every
10 min to target a homogeneous Cr distribution over the column.
After an initial equilibration of the column with the reactor volume
(pH set at 4.0 ꢀ 0.1 with HCl), the formation of chromium(III) colloid
was achieved in situ by dosing the desired amount of K2Cr2O7 at a rate
of 3 mmolminÀ1 to the reducing stirred tank, which contained excess
hydrazine (ratio N2H2/CrVI ꢁ 1.3:1). A synthesis pH of 4 was chosen,
as PHREEQC modeling of the solution speciation indicates a mainly
oligomeric nature of oversaturated CrIII solutions, which favors the
formation of Cr colloids.[32] The Cr content was measured by flame
AAS after microwave digestion of the solid material.[33]
Figure 4. The activity at 403 K of a) fresh Cr catalyst (5.5 mgCrgÀ1),
b) the hot-filtered supernatants (separated after 210 min), c) an artifi-
cial mixture containing the same amount of CyOOH (ca. 1 mm),
CyOH (34 mm), QO (99 mm), and adipic acid (ca. 1 mm), and d) pure
CyH autoxidation.
The autoxidation of CyH (50 mL, HPLC grade) was studied in a
100-mL stainless-steel Parr reactor, stirred at 500 rpm. The headspace
was filled with O2 at an initial room-temperature pressure of
2.75 MPa. The pressure was monitored continuously during the
experiment. Products and by-products were collected in excess
acetone and analyzed by GC (CP-Sil5 column, Chrompack), after
Probably, the stability can be attributed to the amorphous
nature of the oxyhydroxide Cr colloids, which allows the
accommodation of different coordination modes of Cr during
a redox cycle.
As can be seen in Figure 5, the activity of the recycled
catalyst is slightly lower than that of the fresh catalyst,
although the oxidation remains much faster than a pure CyH
autoxidation. After this initial decrease in performance, the
activity is maintained during several recycles. The initial loss
in activity before reaching a steady conversion points to the
existence of a short transient period, the details of which are
not well understood at this moment.
silylation
with
N-methyl-N-(trimethylsilyl)trifluoroacetamide
(MSTFA).[24] 1-Heptanol was used as external standard to correct
for sensitivity differences in the flame ionization detector.
UV/Vis, diffuse reflectance, and X-band EPR spectra were
recorded on a Perkin–Elmer Lambda 12 spectrometer, a Cary 5 UV/
Vis/NIR spectrophotometer, and a Bruker ESP 300E instrument with
a rectangular TE104 cavity, respectively.
Received: July 10, 2006
Published online: October 16, 2006
Keywords: autoxidation · chromium · colloids · radicals ·
.
supported catalysts
[1] U. Schuchardt, D. Cardoso, R. Sercheli, R. Pereira, R. S.
da Cruz, M. C. Guerreiro, D. Mandelli, E. V. SpinacØ, E. L.
Pires, Appl. Catal. A 2001, 211, 1.
[2] a) T. Maschmeyer, R. D. Oldroyd, G. Sankar, J. M. Thomas, I. J.
Shannon, J. A. Klepetko, A. F. Masters, J. K. Beattie, C. R.
Catlow, Angew. Chem. 1997, 109, 1713; Angew. Chem. Int. Ed.
Engl. 1997, 36, 1639; b) V. Kesavan, P. S. Sivanand, S. Chan-
drasekaran, Y. Koltypin, A. Gedanken, Angew. Chem. 1999, 111,
3729; Angew. Chem. Int. Ed. 1999, 38, 3521.
[3] R. A. Sheldon, J. K. Kochi, Metal-Catalyzed Oxidations of
Organic Compounds, Academic Press, New York, 1981.
[4] G. Franz, R. A. Sheldon in Ullmannꢀs Encyclopedia of Industrial
Chemistry, Wiley-VCH, Weinheim, 2000.
[5] S. Bhaduri, D. Mukesh, Homogeneous Catalysis: Mechanisms
and Industrial Applications, Wiley, New York, 2000.
[6] N. M. Emanuel, Z. K. Maizus, I. P. Skibida, Angew. Chem. 1969,
81, 91; Angew. Chem. Int. Ed. Engl. 1969, 8, 97.
[7] D. L. Vanoppen, D. E. De Vos, M. J. Genet, P. G. Rouxhet, P. A.
Jacobs, Angew. Chem. 1995, 107, 637; Angew. Chem. Int. Ed.
Engl. 1995, 34, 560.
[8] M. Nowotny, L. N. Pedersen, U. Hanefeld, T. Maschmeyer,
Chem. Eur. J. 2002, 8, 3724.
Figure 5. Comparison of the activity at 403 K of a) fresh CPC-synthe-
sized catalyst (2.5 mgCrgÀ1), b) after one catalytic run, c) after four
recycles, and d) the pure CyH autoxidation.
In conclusion, we have presented a new approach to the
synthesis of heterogeneous chromium catalysts, based on the
immobilization of Cr colloids generated in situ. The active
materials are demonstrated to increase the yield of cyclo-
hexanone in the autoxidation of cyclohexane. Strong evidence
for the heterogeneous nature of the catalysis is supplied.
More work is in progress to assess the nature of the support
for heterogeneous chromium colloids in autoxidation catal-
ysis, as well as that of colloids in other (liquid phase)
reactions.
[9] I. Hermans, P. A. Jacobs, J. Peeters, Chem. Eur. J. 2006, 12, 4229.
[10] I. Hermans, T. L. Nguyen, P. A. Jacobs, J. Peeters, ChemPhys-
Chem 2005, 6, 637.
[11] I. Hermans, P. A. Jacobs, J. Peeters, J. Mol. Catal. A 2005, 251,
221.
Angew. Chem. Int. Ed. 2006, 45, 7584 –7588
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7587

Communications
[12] I. Hermans, P. A. Jacobs, J. Peeters, ChemPhysChem 2006, 7,
1142.
[13] L. Vereecken, T. L. Nguyen, I. Hermans, J. Peeters, Chem. Phys.
Lett. 2004, 393, 432.
[14] R. A. Sheldon, M. Wallau, I. W. C. E. Arends, U. Schuchardt,
Acc. Chem. Res. 1998, 31, 485.
[15] B. M. Weckhuysen, R. A. Schoonheydt, Zeolites 1994, 14, 360.
[16] J. Kornatowski, G. Zadrozna, M. Rozwadowski, B. Zibrowius, F.
Marlow, J. A. Lercher, Chem. Mater. 2001, 13, 4447.
[17] A. M. Baele, D. Grandjean, J. Kornatowski, P. Glatzel, F. M. F.
de Groot, B. M. Weckhuysen, J. Phys. Chem. B 2006, 110, 716.
[18] A. K. Sinha, K. Suzuki, Angew. Chem. 2005, 117, 275; Angew.
Chem. Int. Ed. 2005, 44, 271.
[19] See, for example: I. C. Chisem, J. Rafelt, M. T. Shieh, J. Chisem,
J. H. Clark, R. Jachuck, D. Macquarrie, C. Ramshaw, K. Scott,
Chem. Commun. 1998, 1949.
[20] E. Breynaert, A. Maes, Anal. Chem. 2005, 77, 5048.
[21] S. Music, M. Maljkovic, S. Popovic, R. Trojko, Croat. Chem. Acta
1999, 72, 789, and references therein.
[22] See, for example: a) T. Tsuzuki, P. G. McCormick, Acta Mater.
2000, 48, 2795; b) M. Ocaꢀa, J. Eur. Ceram. Soc. 2001, 21, 931;
c) M. Baꢀobre-Lꢁpez, C. Vꢂzquez, J. Rivas, M. A. Lꢁpez-
Quintela, Nanotechnology 2003, 14, 318; d) M. V. Borysenko,
V. M. Bogatyrov, E. N. Poddenezhny, A. A. Boiko, A. A.
Chuiko, J. Sol–Gel Sci. Technol. 2004, 32, 327; e) D.-W. Kim,
S.-I. Shin, J.-D. Lee, S.-G. Oh, Mater. Lett. 2004, 58, 1894; f) W. S.
Zhang, E. Brꢃck, Z. D. Zhang, O. Tegus, W. F. Li, P. Z. Si, D. Y.
Geng, K. H. J. Buschow, Phys. B 2005, 358, 332.
[23] B. M. Weckhuysen, I. E. Wachs, R. A. Schoonheydt, Chem. Rev.
1996, 96, 3327.
[24] See Supporting Information.
[25] H. Stꢃnzi, W. Marty, Inorg. Chem. 1983, 22, 2145.
[26] For example, at pH 4.68 the ion-exchange capacity (determined
with CrCl3) was found to be 0.024 meqgÀ1, in good agreement
with those determined for PbII, CuII, and NiII at the same pH
value.[27] The pH value of 4.68 is above that of our synthesis (4.0),
and more than twice that of the point of zero charge (PZC),
determined to be 2.12 by potentiometric titration.
[27] H. H. Tran, F. A. Roddick, J. A. OꢄDonnell, Water Res. 1999, 33,
2992.
[28] B. M. Weckhuysen, R. A. Schoonheydt, F. E. Mabbs, D. Colli-
son, J. Chem. Soc. Faraday Trans. 1996, 92, 2431.
[29] Even if the Cr colloids have a specific surface area as high as
1000 m2 gÀ1, a loading of 5 mg Cr per g support would induce
only a negligible increase in surface area.
[30] W. Buijs, R. Raja, J. M. Thomas, H. Wolters, Catal. Lett. 2003, 91,
253.
[31] P. G. Harrison, N. C. Lloyd, W. Daniell, J. Phys. Chem. B 1998,
102, 10672, and references therein.
[32] The Cr speciation in solution was modeled using PHREEQC, in
combination with the Lawrence Livermore National Library
(LLNL 85 2005-02-02) database; precipitation was disabled
during the modeling because of the slow kinetics. http://
www.geo.vu.nl/users/posv/phreeqc/index.html (PHREEQC for
Windows, version 2.12.04 built using PREEQC 2.12.1-669).
[33] Powder (ca. 200 mg) was dissolved in concentrated perchloric
acid (1 mL), HNO3 (2 mL), and HCl (1 mL).
7588
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7584 –7588