Selective oxidation of para-xylene to terephthalic acid by μ3-oxo-bridged Co/Mn cluster complexes encapsulated in zeolite-Y

Article abstract of DOI:10.1006/jcat.2001.3385

Novel, solid catalysts of μ₃-oxo-bridged Co/Mn cluster complexes were prepared and their catalytic properties (in the "neat" state and when encapsulated in zeolite Y) in the selective oxidation of paraxylene to terephthalic acid with oxygen wer

Full text of DOI:10.1006/jcat.2001.3385

Journal of Catalysis 204, 409–419 (2001)
doi:10.1006/jcat.2001.3385, available online at http://www.idealibrary.com on
Selective Oxidation of para-Xylene to Terephthalic Acid
by -Oxo-Bridged Co/Mn Cluster Complexes
3
Encapsulated in Zeolite–Y
S. A. Chavan, D. Srinivas, and P. Ratnasamy1
National Chemical Laboratory, Pune-411 008, India
Received May 9, 2001; revised August 10, 2001; accepted August 10, 2001
(1–5). The commercial process has been optimized to the
point where typical crude TA purities of 98–99.5% are
achievedinyieldsof96–97mol%based on PXfeed atanox-
idizer contact time of 45–90 min (6–9). However, the follow-
ing improvements are desirable in the commercial process:
(i) replacement of the homogeneous by a heterogeneous
solid catalyst system, and (ii) reduction or elimination of 4-
carboxybenzaldehyde (4-CBA) and other impurities from
TA, which necessitate elaborate hydrogenation and recrys-
tallization procedures in the manufacture of purified TA
required for the polyester industry. One such method of
heterogenization of homogeneous catalysts is encapsula-
tion of active metal complexes inside the pores of zeolites or
zeolitic materials (10–15). Using in situ electronic and elec-
tron paramagnetic resonance (EPR) spectroscopic studies,
we had, earlier, identified the formation of 3-oxo-bridged
heteronuclear Co/Mn cluster complexes (Fig. 1) in the ho-
mogeneous medium during the course of the reaction (16).
These oxo-bridged cluster complexes were postulated as
the active species in the selective PX oxidation reaction.
In the present study, we have, for the first time, synthe-
sized these cluster complexes and studied their spectro-
scopic and catalytic properties, both in the “neat” state as
well as when they are encapsulated in zeolite-Y. The solid
catalysts possess activities comparable to that of the neat
3-oxo-bridged cluster complexes and the conventional
mixture of Co/Mn/Br salts. It is interesting that the key
impurity, 4-CBA, which imparts color to TA, is lower in
abundancewiththesolidcatalysts. Theprobablereasonsfor
the superior performance of these encapsulated catalysts
Novel, solid catalysts of
3-oxo-bridged Co/Mn cluster
complexes, viz., [Co3(O)(CH3COO)6(pyridine)3]+, [Mn3(O)
(CH3COO)6(pyridine)3]+, and CoMn2(O)(CH3COO)6(pyridine)3
(denoted Co3(O), Mn3(O), and CoMn2(O), respectively), encapsu-
lated in zeolite-Y oxidize selectively, para-xylene to terephthalic
acid with dioxygen. The catalysts were prepared by the “flexible
ligand synthesis” method and characterized by X-ray diffraction,
thermal analysis, cyclic voltammetry, Fourier transform in-
frared, diffuse reflectance ultraviolet–visible, X-ray photoelectron
spectroscopy, and electron paramagnetic resonance spectroscopic
techniques. The various physicochemical measurements confirm
the presence and structural integrity of the 3-oxo-bridged cluster
complexes in zeolite cavities. The activity and selectivity of both
the “neat” and encapsulated cluster complexes followed the order
CoMn2(O) > Mn3(O) > Co3(O), revealing the superiority of the
heteronuclear complexes. Under optimal conditions, both neat and
encapsulated cluster catalysts exhibit 100% para-xylene conversion
with >98% selectivity for terephthalic acid. It is important that
the key impurity, 4-carboxybenzaldehyde, is significantly lower
in abundance (than the current commercial catalysts) with one
of the zeolite-encapsulated catalysts, CoMn2(O)–Y. Leaching of
metal ions from the solid catalyst during reaction is minimal and
the catalyst could be recycled without significant loss of activity.
A more facile redox behavior of Co between +2 and +3 oxidation
states in CoMn2(O) (confirmed by cyclic voltammetry) is perhaps
c
2001 Elsevier Science
responsible for high catalytic activity.
Key Words: encapsulated complexes in zeolite–Y; oxo-bridged
Co/Mn cluster catalysts; solid catalyst for para-xylene oxidation;
terephthalic acid; selective oxidation with dioxygen; EPR; spectro-
scopic studies.
are discussed. The advantage of the solid, encapsulated
oxo-bridged heteronuclear Co/Mn cluster complexes is that
they could be recycled without significant loss of activity.
-
3
INTRODUCTION
Terephthalic acid (TA), one of the largest volume com-
modity chemicals, is commercially manufactured, by dioxy-
gen oxidation of para-xylene (PX), in a homogeneous
acetic acid (HOAc) medium, using a catalyst combination
of cobalt and manganese salts and bromide ion promoter
EXPERIMENTAL
Materials
1 To whom correspondence should be addressed. Fax: (91)-020-
5893355. E-mail: prs@ems.ncl.res.in.
Co(CH3COO)2 4H2O, Mn(CH3COO)2 4H2O, glacial
CH3COOH, pyridine, and NaBr were obtained from
409
0021-9517/01 $35.00
2001 Elsevier Science
All rights reserved
c

410
CHAVAN, SRINIVAS, AND RATNASAMY
45 min. The resultant brown solution was stirred for 30 min.
Then, 0.695 g of NaClO4 was added and the stirring was
continued for further 15 min. A brown crystalline prod-
uct (Mn3(O)) was obtained overnight on slow evapora-
tion at 295 K. This was filtered, washed with ethanol, and
dried in vacuum. Elemental analysis was calculated as fol-
lows: C, 37.12%; H, 3.99%; N, 4.56%. Elements found were
C, 37.04%; H, 4.28%; N, 4.48%.
CoMn2(O)(CH3COO)6(pyridine)3(CoMn2(O)). Man-
ganic acetate (2.7 g) was taken in a solution composed
of ethanol (25 ml) and glacial acetic acid (4.2 ml) and
stirred for 10 min until all the manganic acetate dissolved.
To this brown solution Co(CH3COO)2 4H2O (2.5 g)
dissolved in hot pyridine (4 g) was added while being
stirred continuously. The solution was allowed to stand
at 298 K, from which shiny black crystals of 3-oxo-
bridged Co/Mn mixed cluster, CoMn2(O), were obtained.
The crystals were filtered, washed with ethanol, and dried
in vacuum. Elemental analysis was calculated as follows:
C, 43.39%; H, 4.41%; N, 1.87%. Elements found were
C, 43.53%; H, 4.77%; N, 1.87%.
FIG. 1. Structure of 3-oxo-bridged Co/Mn cluster complex.
Mn–Y and Co–Y. Mn(CH3COOH)2 4H2O (4.3 g) dis-
solved in 50 ml of distilled water was added slowly to a sus-
pension of 7 g of zeolite-HY in 50 ml of water. The slurry
was stirred for 4–5 h, at 333 K and then allowed to stand
overnight at 298 K. The solid Mn-exchanged HY was fil-
tered, washed with warm deionized water (about 500 ml),
and dried at 393 K. Co–Y was prepared in a similar man-
ner except that Co(CH3COOH)2 4H2O (4.3 g) was used
in place of Mn(CH3COOH)2 4H2O. The Mn content in
Mn–Y and Co content in Co–Y were estimated to be 3.08
and 10.21 wt%, respectively.
S. D. Fine Chemicals, India. PX, procured from Aldrich
Co., wasusedasreceived. Zeolite-Y(H-form)wasprepared
from NH4Y by calcination.
Catalyst Preparation
[Co3(O)(CH3COO)6(pyridine)3]ClO4(Co3(O)).
This
was prepared by a method analogous to that of Sumner
and Steinmetz (17). Co(CH3COO)2 4H2O (1.25 g) was
taken in a solution of glacial acetic acid (12.5 ml) and pyri-
dine (0.4 ml) and warmed at 323 K while stirring until all the
solid was dissolved. The purple-colored solution was then
cooled to 298 K, and a freshly prepared peracetic acid (ob-
tained from the addition of 0.4 g of glacial acetic acid and
0.7 g of 30% H2O2) was added dropwise over a period of
30 min while stirring. During the addition of peracetic acid,
the color of the solution changed to dark brown. Then, 3 ml
of water was added and refluxed for 1 h at 353 K. A solu-
tion of NaClO4 (0.4 g) dissolved in 20 ml of distilled water
was added to the cooled reaction mixture. Good-quality
microcrystals of Co3(O) were obtained from the solution
kept at 278 K. Elemental analysis was calculated as follows:
C, 35.66%; H, 3.69%; N, 4.99%. Elements found were C,
35.07%; H, 3.74%; N, 4.85%.
CoMn–Y. In the preparation of CoMn–Y, 1.43 g of
Co(CH3COOH)2 4H2O and 4.3 g of Mn(CH3COOH)2
4H2O were used, for 7 g of zeolite HY, for ion exchang-
ing. Zeolite-HY was initially exchanged with Mn ions, as
described above. The Mn–Y thus formed was then used for
further exchange with Co ions to form finally CoMn–Y. The
Co and Mn contents in CoMn–Y were estimated to be 1.02
and 2.10 wt%, respectively.
The encapsulated complexes were prepared by the “flex-
ible ligand synthesis” method (19).
[Co3(O)(CH3COO)6(pyridine)3]–Y(Co3(O)–Y). In
a
typical preparation of encapsulated 3-oxo-bridged Co
cluster complex, 1.5 g of Co–Y was suspended in 15 ml of
HOAc taken in a double-necked, round-bottom flask fitted
[Mn3(O)(CH3COO)6(pyridine)3]ClO4(Mn3(O)). This with an air condenser and an inlet for air. The suspension
complex was prepared as per the method of Vincent et al. was stirred gently for 30 min while passing air. To this was
(18). Mn(CH3COO)2 4H2O (2.5 g) was dissolved in a sol- added 3 ml of pyridine dropwise with stirring for 15 min.
vent mixture composed of ethanol (20 ml), glacial acetic Then 0.5 g of NaBr and 5 ml of distilled water were added
acid (12 ml), and pyridine (3 ml). The resulting solution and the solution was stirred for another 30 min. To this
was stirred while adding N-n-Bu4MnO4 (1.14 g) dissolved 10 ml of aqueous H2O2 (50%) was added dropwise and the
in 10 ml ethanol, in small proportions, over a period of slurry was stirred for another 2–3 h, at 298 K, while passing

SELECTIVE OXIDATION OF p-XYLENE
411
air. The pink solid Co3(O)–Y was filtered, washed with Catalytic Activity Studies and Product Analysis
acetic acid until all the metal complexes adsorbed on the
In a typical oxidation reaction, 2 ml of PX, 38 ml of
HOAc, 5.6 ml of distilled water, 0.0865 g of NaBr, and a
catalyst were taken in a closed titanium-lined Parr 4833 re-
actor which was subsequently pressurized to 200–550 psig
at 298 K with oxidant, air. The reactions were conducted
external surface were removed, and dried at 298 K, under
vacuum. It may be mentioned that all the metal complexes
and clusters of the present study are highly soluble in acetic
acid. Co content in zeolite–Y was estimated to be 0.70 wt%.
[Mn3(O)(CH3COO)6(pyridine)3]–Y(Mn3(O)–Y). En- at 473 K for a period of 15 min to 4 h (the pressure went
capsulated oxo-bridged Mn cluster complex, Mn3(O)–Y up to about 880 psig at reaction temperature, for an ini-
(pale brown), was prepared in a manner similar to that of tial pressure of 550 psig). The products of the oxidation are
Co3(O)–Y, except that Mn–Y was used in place of Co–Y. para-tolyl alcohol (A), para-tolyl aldehyde (B), para-toluic
Manganese content in zeolite–Y was estimated to be acid (C), 4-CBA (D), terephthalic acid (E), and benzoic
1.38 wt%.
acid (F). The liquid products A and B and unreacted PX
were estimated by gas chromatography using a Shimadzu
GC 14 B, SE 30 S.S. packed column. These liquid prod-
ucts and acetic acid were separated from the final reaction
mixture by distillation. The solid products C, D, E, and F
were converted into water-soluble sodium salts by means
of 0.1 N NaOH and acetic acid solutions and estimated
by high-performance liquid chromatography (HPLC)
(Shimadzu LC-9A) using a Shimadzu C-R4A Chro-
matopac, C18 column. The weights of the liquid and solid
[CoMn2(O)(CH3COO)6(pyridine)3]–Y(CoMn2(O)–Y).
Encapsulated oxo-bridged mixed metal complex,
CoMn2(O)–Y (purple), was prepared in a manner
similar to that of Co3(O)–Y except that CoMn–Y was
used in place of Co–Y. The amounts of Co and Mn were
estimated to be 0.38 and 1.31 wt%, respectively.
Procedures
Metal compositions were estimated using inductively products were noted for each run and mass balance was
coupled plasma–atomic emission spectroscopy (Perkin– verified.
Elmer PE-1000) and atomic absorption spectroscopy
Separation and Recycle of the Solid Catalyst
(AAS;VarianSpectrSF-220)techniques. C, H, andNanaly-
ses of neat complexes were done on a Carlo Erba EA 1108
elemental analyzer. Powder X-ray diffractograms of neat
and encapsulated metal cluster complexes were recorded
on a Rigaku Miniflex X-ray diffractometer with a CuK ra-
diation. Thermogravimetric and differential thermal anal-
yses were performed on a Setaram TG-DTA 92 automatic
derivatograph. Fourier transform infrared (FT-IR) spectra
of neat and encapsulated solid complexes were recorded
on a Shimadzu FT-COM 1 spectrophotometer as nujol
mulls. Ultra violet (UV)–visible spectra in normal and dif-
fuse reflectance modes were measured using a UV-101PC
scanning spectrophotometer. X-ray photoemission spectra
(XPS) were recorded on a VG Microtech Multilab ESCA
3000 spectrometer using MgK radiation (h = 1253.6 eV).
EPR spectra were recorded in the temperature range 82–
300 K, using an X-band Bruker EMX spectrometer and
a BVT 3000 temperature controller. The spectra at 77 K
were recorded using a quartz finger dewar (Bruker ER167
FDS-Q). Molecular modeling studies were carried out on a
Silicon Graphics O2 workstation.
The solid, encapsulated cluster catalyst was separated
from the products as follows: The solid product mixture
containing the encapsulated catalyst was treated with 0.1 N
NaOH, until pH of 7, converting all the solid carboxylic
acids into water-soluble salts. The zeolite catalyst was sep-
arated by filtration. No significant leaching was observed
(chemical analysis). The catalyst retained its activity and
could be reused.
RESULTS AND DISCUSSION
Physicochemical Characteristics
The chemical compositions of the metal complexes and
solid catalysts are given in the experimental section.
XRD. No major changes were observed in the X-ray
diffraction (XRD) patterns of zeolite HY on encapsula-
tion of 3-oxo-bridged Co/Mn cluster complexes, indicating
that encapsulation did not alter the framework structure of
zeolite–Y. XPS data (not shown) confirm that Co is present
in the trivalent oxidation state in the neat as well as the
zeolite-encapsulated samples.
To gain deeper insight into the redox properties of Co
and Mn ions, cyclic voltammetry (CV) experiments were
carried out using a BAS CV-50 electrochemical system on
the homogeneous acetic acid solutions (0.004 M) of cobalt
TGA. The thermogravimetric and differential thermo-
and manganese acetate salts, as well as on the neat cluster gravimetric analysis (TG-DTG) profiles of different cata-
complexes, using a Pt-working electrode, a Pt-wire auxil- lysts are shown in Fig. 2. The peak maximum around 480 K
iary electrode, and a saturated calomel reference electrode; in the DTG of Co3(O) (Fig. 2) corresponds to loss of co-
0.1 M tetraethyl ammonium perchlorate was used as a sup- ordinated pyridine, and that at 549 K is due to decomposi-
porting electrolyte.
tion of the acetate ligand. The weight losses in two stages

412
CHAVAN, SRINIVAS, AND RATNASAMY
FT-IR spectroscopy. The FT-IR bands of the zeolite
framework and adsorbed water overlap those of the acetate
and pyridine ligands (Fig. 3). Neat Mn3(O) and Co3(O)
cluster complexes (curves a and b, respectively) show addi-
1
tional peaks around 1092, 1072, 1047, 1032, and 762 cm
due to the counterion (perchlorate). The 3-oxo-bridge in
Mn3(O) shows a characteristic peak at 665 cm 1. The po-
sition of this peak is sensitive to the coordinated metal ion
and occurs at 692 and 685 cm 1 for Co3(O) and CoMn2(O),
respectively (curves b and c, respectively). The spectral
assignments of different bands for neat and encapsulated
complexes are presented in Table 1. The COO vibrational
modes of acetate groups shift to lower energy (i.e., toward
higher wavenumbers) on encapsulation (Fig. 3 and Table 1).
The separation (1 ) between as(COO ) and s(COO ) is
in the range 165–212 cm 1, indicative of bidentate COO
coordination, in syn–syn mode, to two metal ions (20).
UV–visible spectroscopy. Acetic acid solutions of the
neat complexes show an intense UV band, of ligand ori-
gin, around 254 nm and a shoulder at 280 nm (Fig. 4, curves
a–c). In addition to these, a weak, partially resolved band,
characteristic of a 3-oxo-bridge, is also observed at around
320 nm for Mn3(O) and at 355 and 345 nm for Co3(O)
and CoMn2(O), respectively. The position of the latter
band is sensitive to the coordinated metal ion and is at-
tributed to oxygen ⇒ metal charge transfer transitions. This
band is seen even in the diffuse reflectance UV–visible spec-
tra of encapsulated cluster complexes (curves d–f of Fig. 4).
However, due to the dilution in the zeolite matrix, it could
be seen only in Co3(O)–Y (curve d). The observation of
these bands confirms the presence and structural integrity
of these cluster complexes in the cavities of the zeolite.
FIG. 2. TGA (—) and DTG (- - -) profiles of neat and encapsulated
3-oxo-bridged cluster complexes.
in the range 563–663 K correspond to decomposition of
the counterion perchlorate initially into ClO and then into
chlorine. The weight loss in the range 818–936 K corre-
spondstodecompositionofsecondaryproductslikeCoCO3
(formed from the acetate) into cobalt oxide-type materi-
als. In Mn3(O) cluster, the DTG peak at 388 K is due to
the loss of water. Pyridine decomposition is observed at
454–495 K (peak maximum around 484 K). The decom-
position of acetate in Mn3(O) occurs in the range 497–
603 K, with peak maximum at 577 K. The peaks with max-
ima at 616 and 700 K are due to decomposition of the
perchlorate ion. The weight loss with a peak maximum
at around 390 K in CoMn2(O) is due to water loss. The
three DTG peaks, due to loss of pyridine, in the temper-
ature range 423–529 K, arise perhaps from differences in
bond strengths of pyridine coordinated to Co and Mn ions.
The decomposition of the acetate occurs at 563 K. In the
case of the zeolite-Y-encapsulated catalysts, the peaks cor-
responding to adsorbed/bound water dominate those of
acetate and pyridine ligands (Fig. 2). The DTG peaks of
the latter appear as shoulders in the range 493–593 K. The
broad multistep peak in the range 723–923 K is due to the
decomposition of secondary carbonates (formed from
the acetate) during thermal analysis.
FIG. 3. FT-IR spectra (in nujol mull): (a) Mn3(O), (b) Co3(O),
(c) CoMn2(O), (d) Mn3(O)–Y, (e) Co3(O)–Y, and (f) CoMn2(O)–Y.

SELECTIVE OXIDATION OF p-XYLENE
413
TABLE 1
FT-IR Assignments for Neat and Encapsulated Oxo-Bridged Metal Cluster Complexes
Mn3(O)
Co3(O)
CoMn2(O)
1
1
1
1
1
1
Assignment
Neat (cm
2924
)
Encapsulated (cm
2924
)
Neat (cm
2924
)
Encapsulated (cm
2924
)
Neat (cm
2924
)
Encapsulated (cm
)
Asymmetric and symmetric
C–H stretch
2924
COO asymmetric stretch
COO symmetric stretch
CH bending (CH3)
1605
1404
1456
1342
617
1624
1458
1489
1340
1609
1402
1452
1352
623
1635
1463
1463
1616
1404
1448
1350
619
1635
1458
1489
1340
a
a
a
COO deformation
—
—
—
=
C
N stretch (Py)
1558
1545
1558
1550
1553
1545
a Masked by zeolite bands.
EPR spectroscopy. The neat Mn3(O) cluster complex Co3(O) clusters indicate strong antiferromagnetic interac-
shows a broad, isotropic signal at g 2.01 with a tion. The Curie temperature is around 130 K for the Co3(O)
=
peak-to-peak linewidth (1Hpp) of 500 G. The isotropic na- cluster. It is below 77 K for Mn3O. The CoMn2(O), on
ture is due to the small zero-field interaction in the cluster the other hand, is EPR-silent throughout the temperature
complex. The position and 1Hpp of the signal are invariant range.
to temperature, in the range 77–298 K. However, the inten-
The encapsulated Mn3(O) cluster complex (Mn3(O)–Y)
sity of the EPR signal, determined from the double integra- shows a broad EPR signal at g = 2.012 with barely resolved
tion method, increases with decreasing temperature (Fig. 5, manganese hyperfine features superimposed on the broad
curve a), indicating that the antiferromagnetic interaction signal (Fig. 6, curve a). Encapsulated Co3(O) complexes
(among Mn(III) ions, resulting in an s = 2 ground state) is also exhibit a broad signal at g = 2.210 (Fig. 6, curve b),
rather weak in this complex. This is the first report on the the intensity of which decreases at low temperatures and
EPR characteristics of this cluster complex. Its magnetic disappears totally at 77 K. The variation of the signal in-
behavior was reported by Vincent et al. (18), who found the tensity in the neat and encapsulated Co3(O) complexes is
exchange coupling constant (J) to be 10.2 cm 1. Our EPR similar and indicates strong intramolecular antiferromag-
results are in agreement with the small J values reported by netic interactions. Co3(O)–Y samples also revealed trace
them. Polycrystals of the neat Co3(O) cluster complex also amounts of a Mn(II) ion impurity present in the parent
showisotropicsignalsat300Kbutatahigher g value(2.259) zeolite sample (marked by an asterisk in Fig. 6). Mixed
and with a larger linewidth (1075 G). Unlike Mn3(O), metal clusters encapsulated in zeolite–Y showed (curve c) a
the variation with temperature of the signal intensity, g Mn-like spectrum at g = 2.026. A shift in g value from
value, and linewidth (Fig. 5, curves b–d, respectively) of 2.012, typical of a Mn3(O)–Y cluster, is due to the presence
FIG. 4. UV–visible (for neat) and diffuse reflectance UV–visible (for encapsulated) spectra of 3-oxo-bridged cluster complexes:
(a) Co3(O), (b) Mn3(O), (c) CoMn2(O), (d) Co3(O)–Y, (e) Mn3(O)–Y, and (f) CoMn2(O)–Y.

414
CHAVAN, SRINIVAS, AND RATNASAMY
FIG. 6. X-band EPR spectra of encapsulated 3-oxo-bridged Co/Mn
cluster complexes at 298 K: (a) Mn3(O)–Y, (b) Co3(O)–Y. The broad signal
is due to the encapsulated cobalt cluster (the six sharp signals marked by
an asterisk are due to Mn impurity). (c) CoMn2(O)–Y.
FIG. 5. Variation of EPR signal intensity with temperature for
(a) neat Mn3(O) and (b) Co3(O) cluster catalysts. Variation of (c) g
value and (d) linewidth (1Hpp) with temperature for neat Co3(O) cluster
complex.
of Co in the trinuclear unit, confirming the structural in-
tegrity of the CoMn2(O) cluster in zeolite–Y, a conclusion
also reached from the FT-IR and UV–visible spectral data.
Stability of Cluster Complexes under Reaction Conditions
Figure 7 shows the EPR spectra of the Mn3(O) com-
plex and NaBr dissolved in acetic acid and heated in air
(550 psig) to 473 K for 1 h (curves a–c of Fig. 7). The spectral
features reveal the structural stability of these oxo-bridged
cluster complexes under typical reaction conditions. Heat-
ing a mixture of Mn(CH3COO)2 4H2O solution in acetic
acid, NaBr, and air (550 psig), in fact, converts the linear
manganese acetate into a trimeric Mn3(O) cluster complex
(compare curves a and d of Fig. 7). This observation also
confirms that the Mn cluster complexes are formed during
the course of the reaction even with the conventional cata-
lyst system (Co/Mn/Br /acetic acid), as postulated by us in
our earlier publication (16).
FIG. 7. EPR spectra (at 82 K) of Mn3(O): (a) Mn3(O) + HOAc +
NaBr; (b) Mn3(O) + HOAc + NaBr after 1 h at 473 K; (c) Mn3(O)
+ HOAc + NaBr after 1 h at 473 K and 550 psig air; and (d)
Mn(CH3COO)2 4H2O + HOAc + NaBr after 1 h at 473 K and 550
psig air.

SELECTIVE OXIDATION OF p-XYLENE
415
TABLE 2
Effect of Solvent and Halide Ion on the Oxidation of para-Xylene over Co/Mn Catalysts in Homogeneous Medium Reaction Conditionsa
Product distribution (wt%)c
Conversion
Run no.
Catalyst (mmol)
Solvent/halide ion
(wt%)
TONb
A
B
C
D
E
F
1
2
3
4
5
6
7
8
9
Co(OAc)2(0.43) + Mn(OAc)2(0.14)
Co(OAc)2(0.43) + Mn(OAc)2(0.14)
Co(OAc)2(0.43) + Mn(OAc)2(0.14)
Co3(O) (0.041)
Mn3(O)(0.039)
CoMn2(O)(0.041)
1,4-Dioxane, Br
AcOH, Cl
35.1
18.0
100
17.3
46.0
23.5
2.3
4.3
0.5
7.5
11
6
11.2
1.7
0
23.8
14.0
24.0
6.8
0
88.5
74.7
0
76.2
84.6
74.8
52.5
63.7
70.9
75.2
0.3
23.6
0.9
0
1.4
AcOH, Br
1,4-Dioxane, Br
1,4-Dioxane, Br
1,4-Dioxane, Br
AcOHd
32
78
216
106
1
20
2
34
0.1
97.7
1.3
1.2
Co(OAc)2(0.43) + Mn(OAc)2(0.14)
40.7
36.3
29.1
23.3
Co3(O) (0.041)
AcOHd
Mn3(O)(0.039)
CoMn2(O)(0.041)
AcOHd
0
1.6
10
AcOHd
a para-Xylene (2 ml) + NaBr (0.0836 g) + H2O (5.6 ml) + HOAc (38 ml); oxidant, air; reaction time, 45 min; temperature, 473 K; pressure, 550 psig.
b Turnover number (TON), mole of para-xylene converted per mole of metal catalyst.
c Products: A, p-tolyl alcohol; B, p-tolyl aldehyde; C, p-toluic acid; D, 4-carboxybenzaldehyde; E, terephthalic acid; and F, benzoic acid. Liquid
products A and B were estimated using gas chromatography. Products C to F were estimated using HPLC.
d Reaction without any halide ion.
Selective Oxidation of PX
sented in Tables 2–5. The reaction did not proceed to any
significant extent in the absence of catalyst.
In the commercial process for the selective oxidation of
PX to TA, in addition to complete conversion of the PX, it
Effect of solvent and halide ion. When the reactions are
is also essential to get as high a selectivity for TA as possi- carried out with NaCl in place of NaBr (Table 2), the con-
ble. Particularly, it is desirable to suppress or eliminate the versions are markedly lower (compare runs 2 and 3). The
4-CBA impurity. In current practice, 4-CBA occurs to the conversions are also low in the absence of the halide ion
extent of 0.1–0.5% in the crude product. In the following (runs 7–10), or when 1,4-dioxane is used in place of acetic
sections we report the activities of the neat and zeolite- acid (runs 1 and 4–6). Both acetic acid and the bromide ion
encapsulated cluster catalysts. The catalytic data are pre- are essential for high activity and selectivity.
TABLE 3
Oxidation of para-Xylene over Co(CH3COO)3 4H2O and Mn(CH3COO)2 4H2O Salts in Homogeneous Medium Reaction Conditionsa
Product distribution (wt%)d
Run no.
Catalyst (mmol)b
Pressure (psig)
Conv. (wt%)
TONc
A
B
C
D
E
F
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Co(0.43)
200
200
200
200
200
350
350
350
350
350
550
550
550
550
84.7
83.4
89.1
87.4
89.1
100
100
100
100
100
100
100
100
99
36
27
30
28
38
43
32
33
32
42
32
33
32
42
1.6
0
0
0
0
0
0
0
0
0
0
0
0
39.4
40.3
44.4
43.6
44.0
0
0
0
0
0
21.4
23.0
24.6
25.3
24.7
14.1
0.8
1.6
5.0
23.7
0.7
0.36
8.8
6.4
10.4
13.3
12.9
2.2
0.01
0.1
0.2
5.0
28.4
30.1
20.1
17.3
17.2
83.6
98.8
98.2
92.8
70.9
96.9
99.6
99.5
91.7
0.3
0.2
0.6
0.5
1.2
0.1
0.4
0.1
2.0
0.4
0.01
0.01
0.01
0.02
Co(0.43) + Mn(0.14)
Co(0.28) + Mn(0.28)
Co(0.14) + Mn(0.44)
Mn(0.44)
Co(0.43)
Co(0.43) + Mn(0.14)
Co(0.28) + Mn(0.28)
Co(0.14) + Mn(0.44)
Mn(0.44)
Co(0.43) + Mn(0.14)
Co(0.28) + Mn(0.28)
Co(0.14) + Mn(0.44)
Mn(0.44)
0
0
0
0
1.4
0.03
0.05
1.0
0.44
7.3
0
a Reaction conditions are the same as in Table 2. Reaction time, 2 h; temperature, 473 K.
b Sources of Co and Mn are Co(CH3COO)2 4H2O and Mn(CH3COO)2 4H2O, respectively.
c Turnover number (TON), mole of para-xylene converted per mole of metal catalyst.
d Products are the same as in the Table 2.

416
CHAVAN, SRINIVAS, AND RATNASAMY
TABLE 4
Catalytic Activity of Neat Co/Mn Cluster Catalysts in the Selective Oxidation of para-Xylene in Homogeneous Medium: Effect
of Pressure and Reaction Timea
Product distribution (wt%)c
Pressure
(psig)
Reaction
time (min)
Conversion
(wt%)
Run no.
Catalyst (mmol)
Co3(O)(0.041)
TONb
A
B
C
D
E
F
1
2
3
4
5
6
7
8
9
550
15
30
45
15
30
45
120
15
30
45
67.7
75.6
86.7
79.4
64.5
100
100
100
100
100
307
343
393
373
303
470
470
449
449
449
449
449
0.5
0.5
0
0
0
0
0
0
0
26.5
33.6
42.5
37.6
22.5
0
0
0
0
0
20.1
25.3
26.0
27.4
26.1
27.0
1.6
4.0
6.1
4.7
1.8
46.7
33.3
11.7
22.9
10.3
8.0
0.1
0.3
0.9
0.7
5.3
6.4
0.9
0.9
0.3
0.9
0.3
0.1
0.1
0
0
0
0
0.1
19.5
11.2
40.8
64.9
98.2
95.7
93.0
94.6
97.8
99.0
Mn3(O)(0.039)
550
CoMn2(O)(0.041)
550
10
11
12
0
0
0
60
120
100
100
0
0
0.4
0.1
0.8
a Reaction conditions are the same as in Table 2. Reaction temperature, 473 K.
b Turnover number (TON), mole of para-xylene converted per mole of metal catalyst.
c Products are the same as in Table 2.
PX oxidation over cobalt and manganese acetate salts. selectivity is obtained with the cluster catalysts (Mn3(O)
When mixtures of the cobalt and manganese acetates are and CoMn2(O)) (runs 7 and 12 of Table 4). Complete
used, the clusters are formed in about 15 min at higher pres- conversion of PX is achieved within 15 min with CoMn2(O)
sures (350–550 psig) and temperatures (above 473 K) com- (run 8). However, TA selectivity improves when reactions
pared to about 4 h at atmospheric pressure and 363 K. The are continued for a longer time (runs 8–12). Co3(O) ex-
combination of cobalt and manganese acetates in millimole hibits low PX conversions and TA selectivity. Mn3(O), on
ratios of 3 : 1, 1 : 3, and 1 : 1 yield cluster complexes of the the other hand, yields 100% PX conversion and high TA
type Co2Mn(O), CoMn2(O), and a mixture of Co2Mn(O) selectivity at longer reaction times (run 7). It may be noted
and CoMn2(O), respectively (16). Both conversion and se- that the heteronuclear cluster catalyst CoMn2(O) is more
lectivity are higher when Co and Mn are present together active and yields higher TA selectivity than the homonu-
(Table 3). Complete conversion of PX and >95 wt% selec- clear Co3(O) and Mn3(O). The concentration of 4-CBA is
tivity to TA are obtained at 550 psig and 473 K (Table 3, also, in general, less with the former.
runs 7, 8, 11–14).
PX oxidation over solid catalysts. When the complexes
PX oxidation over neat cluster catalysts. At 473 K and a are encapsulated in zeolite–Y, complete conversions are
pressure of 550 psig, 100% PX conversion and >98 wt% TA achieved only at longer reaction times (Table 5), perhaps
TABLE 5
Catalytic Activity of Solid Catalysts: 3-Oxo-Bridged Co/Mn Clusters Encapsulated in Zeolite–Ya
Product distribution (wt%)c
Catalyst
(g mmol of metal)
Conversion
(wt%)
Run no.
Time (min)
TONb
A
B
C
D
E
F
1
2
3
4
5
6
7
8
Mn3(O)–Y (0.075)
60
150
240
60
150
60
74.8
80.4
99.9
65.6
72.0
74.4
100
183
197
245
341
374
151
203
203
0
0
0
13.4
10.5
0
33.1
37.8
0
18.4
14.0
0
47.3
30.8
20.1
13.0
10.7
6.4
8.1
1.7
0.7
37.5
39.8
0.8
<0.01
0.01
10.4
29.7
79.2
15.2
20.7
92.8
98.9
99.4
1.1
0
0
2.5
4.3
0.1
0
Co3(O)–Y (0.036)
CoMn2(O)–Y (0.09)
150
240
0
0
0
0
1.1
0.6
100
0
a Reaction conditions are the same as in Table 2. Catalyst, 0.2995 g; temperature, 473 K; and pressure, 550 psig.
b Turnover number (TON), mole of para-xylene converted per mole of metal catalyst.
c Products are the same as in Table 2.

SELECTIVE OXIDATION OF p-XYLENE
417
TABLE 6
Cyclic Voltametric Data of Neat 3-Oxo-Brideged Co/Mn Cluster Complexes
E1/2
(in V)a
1E
(in V)b
Complex
Redox couple
Co(CH3COO)2 4H2O
Mn(CH3COO)2 4H2O
Co3(O)
A: CoIII/CoII
+0.75
+0.75
+0.69
+0.56
+0.93
0.21
0.30
0.22
0.04
0.05
A: MnIII/ MnII
A: Co3III(O)/CoIICo2III(O)
A: CoIIIMnI2II(O)/CoIIMnI2II(O)
B: CoIIIMnIVMnIII(O)/CoIIIMn2III(O)
CoMn2(O)
a
E1/2, Half-wave potential with reference to saturated calomel electrode.
b 1E, separation between the cathodic and anodic peaks of the redox couple.
due to diffusional limitations in the zeolite. Easy separa- form chlorine atom (and Mn(II)) is not favorable because
bility and reusability (see Experimental section) are ma- the E1/2 oftheMn(III)/Mn(II)couple(1.2V)issmallerthan
jor advantages of the solid catalysts. Another interesting that of the Cl /Cl couple (1.36 V). The redox potentials of
feature with the encapsulated CoMn2(O) cluster catalyst cobalt, manganese, bromide, and chloride ions against an
(run 8) is the very low concentration of the 4-CBA im-
purity (around 0.01%). It may be noted that the activity
and selectivity of both the neat and encapsulated clus-
ter complexes follow the trend CoMn2(O) > Mn3(O) >
Co3(O), confirming the superiority of the heteronuclear
complexes.
Catalyst Leaching Tests
The molecular dimensions of the M3(O) (M = Co, Mn)
˚
cluster is about 10.8 A. It can be accommodated in the su-
˚
percages of zeolite–Y ( 12 A). The cluster is, however,
larger than the openings of the windows of the zeolite’s
supercage, and after formation it is, hence, locked inside
the supercage, making its loss during the reaction difficult.
Leaching of metal ions from the solid catalyst at the end
of the reaction was investigated by AAS and EPR spectro-
scopictechniques. WhileAASdidnotrevealanyleachingof
metal ions, EPR estimated that about 0.5% of the encap-
sulated manganese in the solid catalyst was leached into
solution (about 50 ppm of Mn in solution). This amount
cannot account for the observed catalytic activity. Catalytic
runs with this trace amounts of metal ions in solution exhib-
ited low PX conversions (28 wt%) with para-tolyl alcohol
(A) and para-tolyl aldehyde (B) as products; TA was not
detected.
3-Oxo-Bridged Co/Mn Cluster Complexes
and Mechanism of PX Oxidation
Role of bromide. During the reaction, Co(III) oxidizes
Mn(II) to Mn(III) ions, being itself reduced to Co(II). The
Mn(III) ion, in turn, abstracts an electron from the bro-
mide ion to form a bromine atom. The bromine atom subse-
quently abstracts a H-atom from the –CH3 group (forming
HBr) and initiates the oxidation reaction. A similar abstrac-
FIG. 8. Cyclic voltammograms of (a) Co(CH3COO)2 4H2O,
tion of an electron from the chloride ion by Mn(III) ions to (b) Co3(O), and (c) CoMn2(O). Redox couples A and B are indicated.

418
CHAVAN, SRINIVAS, AND RATNASAMY
FIG. 9. Catalytic oxidation of para-toluic acid over CoMn2(O)–Y.
NH electrode are given as (21).
acid(C). H-atomabstractionfromthemethylgroupinpara-
toluic acid (C) generates a secondary benzyl radical, which
follows the same pathway to yield, eventually, TA (E). The
former reaction (oxidation of PX to para-toluic acid) is eas-
ier, and even mononuclear Co or Mn complexes accomplish
this significantly (16, 23). Further oxidation of para-toluic
acid to TA, however, is difficult and only the catalyst system
Co/Mn/Br , in HOAc, at high pressure and temperature,
achieves 100% conversion of PX with >95% selectivity to
TA. The reduction in ring electron density (in going from
PX to para-toluic acid) makes H-abstraction from methyl
group of para-toluic acid about 4.9 times more difficult.
Monomeric complexes cannot abstract H-atoms from para-
toluic acid; 3-oxo-bridged heteronuclear Co/Mn clusters
(CoMn2(O)) are capable of accelerating the latter reac-
tion step. A tentative reaction mechanism for oxidation by
3-oxo-bridged cluster complexes is shown in Fig. 9. It is
interesting to note that such oxo- and acetato-bridged Fe,
Cu, and Mn complexes are active sites in many oxygenase,
oxidase and catalase enzymes (24–27). Cluster complexes
exhibitlabileintramolecularelectronandspintransfer. This
unique property and the basic nature of the 3-oxo-bridges
account, perhaps, for their superior performance in the se-
lective oxidation of PX.
Co(III) + e ↔ Co(II) (E1/2 = 1.92 V)
Mn(II) ↔ Mn(III) + e (E1/2 = 1.2 V)
Br ↔ Br + e (E1/2 = 1.06 V)
Cl ↔ Cl + e (E1/2 = 1.36 V).
Chlorineatomsarealsolessstablethanbromineandreadily
form chlorine molecules. The unfavorable redox couple as
well as the low concentration of chlorine atoms in solution
are responsible for the low conversions observed with chlo-
ride ions (Table 2). Bromide ions are, hence, the promoters
of choice.
Role of the cluster complexes. The CV data of cobalt
and manganese acetates and Co3(O) and CoMn2(O) are
presented in Table 6. Figure 8 shows the CV of cobalt
acetate, Co3(O) and CoMn2(O) (curves a–c, respectively).
The redox couple A (Fig. 8) is due to changes in the oxida-
tion state of Co (between +2 and +3). CoMn2(O) (curve c)
shows an additional redox couple (B) due to MnIII ↔ MnIV
transitions of one of the two Mn ions in CoMn2(O). The
E1/2 value for couple A is lower for CoMn2(O) (+0.56 V)
compared to Co3(O) (+0.69 V) and cobalt and manganese
acetates (+0.75 V), respectively (Table 6). Lower E1/2
values are indicative of greater ease of electron trans-
fer. The redox couples of CoMn2(O) are reversible, while
those of the others are quasireversible (note the smaller
CONCLUSIONS
3-Oxo-bridged Co/Mn acetate cluster complexes, both
1E values for CoMn2(O) (0.04 V) compared to Co3(O) in the neat state and when encapsulated in the supercages
(0.22 V), cobalt acetate (0.21 V), and manganese acetate of zeolite-Y, are efficient catalysts for the selective oxida-
(0.30 V); Table 6). The low E1/2 and 1E values indicative tion of PX to TA. The more facile redox behavior of Co
of more facile CoII ↔ CoIII transitions in CoMn2(O) are, between +2 and +3 oxidation states in CoMn2(O) com-
perhaps, responsible for its higher catalytic activity.
pared to the homonuclear complexes is responsible for the
The aerial oxidation of PX by a Co/Mn/Br catalyst sys- greater catalytic activity of the former. The zeolite-based
tem follows a free radical chain mechanism (1). The mech- heterogeneous catalysts, in addition to possessing all the
anism of the oxidation reaction is believed (1) to begin superior activity and selectivity characteristics of the homo-
with H-atom abstraction from the methyl group by bromine geneous catalysts, have the advantages of easy separation
atoms (forming HBr). The resultant benzyl radical adds to and reuse. The catalyst is stable at reaction conditions and
O2 and proceeds through the hydroperoxide (22) to para- there is negligible leaching of metal ions into the reaction
tolyl alcohol (A), para-tolyl aldehyde (B), and para-toluic medium. To our knowledge, this is the first solid catalyst

SELECTIVE OXIDATION OF p-XYLENE
419
reported that is both highly active and selective in oxida- 11. Sabater, M. J., Corma, A., Domenech, A., Forne´s, V., and Garcia, H.,
Chem. Commun. 1285 (1997).
12. Ogunwumi, S. B., and Bein, T., Chem. Commun. 901 (1997).
tion of para-xylene to terephthalic acid.
13. Parton, R. F., Bezoukheanova, C. P., Grobet, J., Grobet, P. J., and
Jacobs, P. A., Stud. Surf. Sci. Catal. 83, 371 (1994).
14. Deshpande, S., Srinivas, D., andRatnasamy, P., J. Catal. 188, 261(1999).
ACKNOWLEDGMENTS
15. Chavan, S., Srinivas, D., and Ratnasamy, P., Top. Catal. 11/12, 359
The authors thank Ms. N. E. Jacob for the help in thermal analysis and
(2000).
Prof. Subhash Padhye and Mr. Nitin Gokhale for electrochemical studies.
16. Chavan, S. A., Halligudi, S. B., Srinivas, D., and Ratnasamy, P., J. Mol.
Catal. 161, 49 (2000).
17. Sumner, C. E., Jr., and Steinmetz, G. R., J. Am. Chem. Soc. 107, 6124
REFERENCES
(1985).
18. Vincent, J. B., Chang, H.-R., Folting, K., Huffman, J. C., Christou, G.,
1. Partenheimer, W., Catal. Today 23, 69 (1995).
2. Sheldon, R. A., and Van Santen, R. A. (Eds.), “Catalytic Oxidation:
Principles and Applications.” World Scientific, Singapore, 1995.
3. Sheldon, R. A., in “The Activation of Dioxygen and Homogeneous
CatalyticOxidation”(D. H. R. Barton, A. E. Martell, andD. T. Sawyer,
Eds.), p. 9. Plenum, New York, 1993.
4. Sheldon, R. A., in “Dioxygen Activation and Homogeneous Catalytic
Oxidation” (L. I. Simandi, Ed.), Vol. 66, p. 573. Elsevier, Amsterdam,
1991.
and Hendrickson, D. N., J. Am. Chem. Soc. 109, 5704 (1987).
19. Balkus, K. J., Jr., and Gabrielov, A. G., J. Inclusion Phenom. Mol.
Recognit. Chem. 21, 159 (1995).
20. Mehrotra, R. C., and Bohra, R., “Metal Carboxylates.” Academic
Press, New York, 1983.
21. Douglas, B. E., McDaniel, D. H., and Alexander, J. J., “Concepts and
Models in Inorganic Chemistry.” Wiley, New York, 1994.
22. Metelski, P. D., Adamian, V. A., and Espenson, J. H., Inorg. Chem. 39,
2434 (2000).
23. Jacob, C. R., Varkey, S. P., and Ratnasamy, P., Appl. Catal. A 182, 91
(1999).
5. Parshall, G. W., and Ittel, S. D., “Homogeneous Catalysis.” Wiley-
Interscience, New York, 1992.
6. Saffer, A., and Barker, R. S., U.S. Patent 2,833,816 (1958).
7. PCT Int. Appl. WO 9,931,038 (24 June 1999).
8. Partenheimer, W., and Schammel, W. P., U.S. Patent 4,786,753
(1988).
9. Broeker, J. L., Partenheimer, W., andRosen, B. I., U.S. Patent5,453,538
(1995).
24. Andersson, K. K., and Gra¨slund, A., Adv. Inorg. Chem. 43, 359 (1995).
25. Siegbahn, P. E. M., and Crabtree, R. H., in “Metal-Oxo and Metal-
Peroxo Species in Catalytic Oxidations in Structure and Bonding”
(B. Meunier, Ed.), Vol. 97, p. 125. Springer-Verlag, Berlin, 2000.
26. Beyer, W. F., and Fridovich, I., Biochemistry 24, 6460 (1985).
27. Law, N. A., Caudu, M. T., and Pecoraro, V. L., Adv. Inorg. Chem. 46,
305 (1999).
10. Hutchings, G. J., Chem. Commun. 301 (1999).