Journal of Colloid and Interface Science 328 (2008) 314–323

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Pore tuned activated carbons as supports for an enantioselective molecularcatalyst

Frederico Maia a, Rui Silva b, Bruno Jarrais b, Ana R. Silva a, Cristina Freire a,∗, Manuel F.R. Pereira b,∗,José L. Figueiredo b

a REQUIMTE, Departamento de Química, Faculdade de Ciências, Universidade do Porto, 4169-007 Porto, Portugalb Laboratório de Catálise e Materiais (LCM), Laboratório Associado LSRE/LCM, Departamento de Engenharia Química, Faculdade de Engenharia, Universidade do Porto, 4200-465 Porto,Portugal

a r t i c l e i n f o a b s t r a c t

Article history:Received 4 June 2008Accepted 10 September 2008Available online 17 September 2008

Keywords:Mesoporous activated carbonsTextural propertiesEnantioselective epoxidation of alkenesMn(III) salen complex

The Jacobsen catalyst was immobilized onto four activated carbons with different average pore sizes,achieved by a gasification process followed by molecular oxygen oxidation. The influence of the texturalproperties of the activated carbon in the immobilization process and in the catalytic performance ofthe Mn(III) heterogeneous catalysts was investigated in detail. Three different catalytic systems werestudied: styrene epoxidation using m-chloroperoxybenzoic acid; 6-CN-2,2-diMeChromene epoxidationusing NaOCl and iodosylbenzene (PhIO) as oxidants. The catalysts tested were active and enantioselectivein the three systems studied. Selectivity towards the desired epoxide only decreases in the case of thematerial with smaller pores, remaining identical to that of the homogeneous phase in all the othermaterials. The enantiomeric excess values (%ee) for alkene epoxidation increase with the pore sizeof the heterogeneous catalysts, and these values are even higher than the homogeneous counterpartsin the styrene epoxidation reaction. Total Mn(III) loadings increase with the pore size, as well astheir distribution within the carbon porous matrix. Characterization of the activated carbons bearingthe immobilized manganese(III) complexes by TPD and XPS point to reaction between carbon surfacephenolate groups and the manganese(III) complexes through axial coordination of the metal centers tothese groups.

© 2008 Elsevier Inc. All rights reserved.

1. Introduction

Chiral compounds are of great importance in the manufactureof fine chemicals, drugs, vitamins, cosmetics and optical materials.For that matter, the use of chiral building blocks has attracted greatattention. Chiral epoxides are an important group of these build-ing blocks, acting as intermediates that can be readily transformedinto several chiral compounds through regioselective ring-openingor functional transfer reactions [1]. Manganese(III) salen complexeshave been proven to be excellent catalysts in the epoxidation ofunfunctionalized olefins, exhibiting high activity and enantioselec-tivity, using a wide range of oxygen sources such as iodosylben-zene, sodium hypochlorite, 3-chloroperoxybenzoic acid, hydrogenperoxide, etc. [2].

Separation and recycling of the homogeneous catalyst is dif-ficult and makes the entire process nonviable economically forindustrial applications, specially in the case of expensive catalysts

* Corresponding authors. Faxes: +351 225081449 (M.F.R. Pereira), +351 220402590(C. Freire).

E-mail addresses: acfreire@fc.up.pt (C. Freire), fpereira@fe.up.pt (M.F.R. Pereira).

0021-9797/$ – see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2008.09.030

with relatively low turnover numbers such as the Jacobsen cata-lyst. This is why the heterogenization of these complexes has beenthe subject of intense study in recent years. The replacement of ahomogeneous catalyst by an active heterogeneous one changes thesynthetic process to a more desirable and clean one providing ad-vantages such as easy handling and product separation, catalystrecovery and lower level of waste [3]. In the case of the het-erogenization of manganese(III) salen complexes, there is even anincrease in the catalyst stability, as the main deactivation process isthe formation of inactive dimeric μ-oxo manganese(IV) species inhomogeneous phase, process which is hindered by local site isola-tion of the complexes in the solid matrix [2,4].

Activated carbons are porous and inexpensive materials whichhave been widely used as solid supports in the heterogenizationof metal complexes [5–13]. These materials possess several oxy-gen superficial groups, which can be selectively tuned by thermaland chemical processes [14], which can then be used to link theactive species covalently. Moreover, their texture can also be al-tered by physico-chemical methods in order to promote selectivechanges in the average pore size, surface area and pore volume.Thus, the final heterogeneous catalyst properties can be selectively

F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323 315

Scheme 1. Anchoring procedure for the Jacobsen catalyst (CAT) onto the modified activated carbons.

changed, introducing different surface environments which can in-crease enantioselectivity and shape selectivity. In this work, a pro-cedure to introduce textural changes in a commercial activatedcarbon, maintaining the surface chemistry identical, is presented.This was achieved by submitting the starting activated carbon toa gasification process during four different times, leading to fourpore tuned materials with different pore sizes, mesopore areasand micropore volumes. The materials were then submitted tothe same oxidation step, in which they were heated in the pres-ence of molecular oxygen to promote the formation of superficialoxygen groups, like phenol groups, which were converted into phe-nolate groups by treatment with NaOH. These groups were thenused to anchor covalently the Jacobsen catalyst, by axial coordina-tion to the metal center, as described previously [11] (Scheme 1).The materials obtained were thoroughly characterized in order tostudy the influence of pore tuning on the metal complex anchor-ing procedure and on its final distribution within the matrix, aswell as in the catalytic performance for styrene and 2,2-dimethyl-6-cyanochromene epoxidation. The new materials, as well as theirprecursors, were characterized by ICP-AES, XPS, TPD and nitrogenadsorption at 77 K.

2. Experimental

2.1. Materials and reagents

The starting carbon material was a NORIT ROX 0.8 activatedcarbon (rodlike pellets with 0.8 mm diameter and 5 mm length).The reagents and solvents used to anchor the metal complexesand in the catalytic experiments were used as-received. (R,R)-(-)-N,N ′-bis(3,5-di-tert-butylsalicyldene)-1,2-cyclohexanediamino-manganese(III) chloride (CAT—Jacobsen catalyst), styrene, 2,2-di-methyl-6-cyanochromene (6-CN-2,2-diMeChromene), m-chloroper-oxybenzoic acid (m-CPBA), 4-methylmorpholine N-oxide (NMO),chlorobenzene and sodium hypochlorite solution were from Ald-rich; all solvents were from Merck (pro analysis), except dichloro-methane and acetonitrile used in the catalytic experiments whichwere from Romil (HPLC grade). Iodosylbenzene (PhIO) was pre-pared as described in literature [15].

2.2. Preparation of the pore tuned activated carbons

The starting material (NORIT ROX 0.8) was added to an aqueoussolution of Co(NO3)2·6H2O (3.5% wt), and the resulting suspensionwas shaken for 7 days. The role of cobalt is to catalyze the gasi-fication of carbon, thereby promoting the formation of mesopores[16]. The activated carbon was then washed with deionized waterand dried in an oven at 373 K, overnight.

Then, the material was gasified in a tubular reactor, using thefollowing experimental conditions: heating from room temperatureto 1173 K at 10 K min−1 under a flow of 100 cm3 min−1 of N2;at 1173 K the gas was changed to CO2, maintaining the flow rateand the material was gasified during four different exposure times,

leading to four different materials: 0 (CA), 60 (CA60), 180 (CA180)and 300 min (CA300). After gasification the cobalt must be re-moved to avoid its possible interference in the reaction. Therefore,the samples were then washed with HCl 0.05 M, in order to re-move cobalt, and with deionized water until pH ∼ 7, and dried inan oven at 373 K, overnight.

Afterwards, all four materials were submitted to the samegas phase oxidation process in a tubular reactor: heating fromroom temperature to 698 K at 10 K min−1 under a flow of100 cm3 min−1 of N2; at 698 K the gas was changed to 5% O2(in N2) maintaining the flow rate for 300 min, and then cooled toroom temperature under a flow of 100 cm3 min−1 of N2 (samplesCAX_O2, where X is the gasification time). These conditions wereselected in order to have a mild oxidation treatment, with the aimof introducing oxygen surface groups but keeping the texture al-most unchanged.

2.3. Anchoring of chiral manganese(III) salen complex onto theactivated carbon

The activated carbons (1.10 g, CAX_O2) were refluxed with anaqueous solution of sodium hydroxide 13.3 mmol (100 cm3) for1 h, in order to convert the CAX_O2 surface phenol into phe-nolate groups (Scheme 1); a decrease in the pH of the aqueoussolutions from 14 to 13 was observed. The materials (CAX_ONa)were washed with deionized water until constant pH (8) and thendried at 393 K in an oven, under vacuum. Then, typically, 0.70 gof CAX_ONa material was refluxed for 8 h with an ethanolic solu-tion of 0.200 mmol of CAT (Jacobsen catalyst) [11]. The anchoringprocess was monitored by UV–vis spectroscopy and a decrease inintensity of the electronic bands of the manganese(III) salen com-plex in the region 200–800 nm was observed [9–12]. In orderto remove any physisorbed complex, the resulting materials werepurified by Soxhlet extraction with ethanol for 8 h. Finally, the ma-terials were vacuum dried in an oven overnight at 393 K (samplesCAT@CAX_ONa).

2.4. Physico-chemical measurements

Nitrogen adsorption measurements were carried out at 77 Kusing a Coulter Omnisorp 100 CX apparatus. The BET specific sur-face areas (SBET) were evaluated using adsorption data in the rel-ative pressure range from 0.05 to 0.12. The micropore volumes(V micro) and mesopore surface areas (Smeso) were determined bythe t-method, using the standard isotherm for carbon materi-als proposed by Reinoso et al. [17]. The adsorption data werealso analyzed using the Dubinin–Radushkevich equation. Since atype IV deviation occurred, two microporous structures were con-sidered, and the corresponding volumes, W01 and W02, were cal-culated [18]. The Stoeckli equation [19] was used to estimate theaverage pore width of the smaller pores (L1), using an affinity coef-ficient for nitrogen (β = 0.34). Their respective surface areas (Sm1)

316 F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323

were calculated assuming slit-shaped pores. The pore size distribu-tion was obtained by using the non-local density theory (NLDFT)applying the kernel file provided by Quantachrome’s data reduc-tion software, where a slit-pore model is assumed.

The TPD profiles were obtained with a custom built set-up,consisting of a U-shaped tubular microreactor, placed inside anelectrical furnace. The mass flow rate of the helium carrier gas(69 μg s−1) and the heating rate of the furnace (5 K min−1) werecontrolled with appropriate units. The amounts of CO and CO2 des-orbed from the carbon samples (0.1 g) were monitored using aSpectramass Dataquad quadrupole mass spectrometer.

Manganese ICP-AES analysis were carried out at “Laboratório deAnálises,” IST, Lisbon (Portugal). X-ray photoelectron spectroscopy(XPS) was performed at “Centro de Materiais da Universidade doPorto” (Portugal), in a VG Scientific ESCALAB 200A spectrometerusing non-monochromatized AlKα radiation (1486.6 eV). Due tothe hardness of the materials it was not possible to use com-pressed pellets to the XPS studies; instead the materials wereglued in a tape and consequently the C and O atomic % cannotbe considered for analysis of the prepared materials. To correctpossible deviations caused by electric charge of the samples, theC 1s band at 285.0 eV was taken as internal standard. The surfaceatomic percentages were calculated from the corresponding peakareas and using the sensitivity factors provided by the XPS manu-facturer.

GC-FID chromatograms were obtained with a Varian CP-3380gas chromatograph using helium as carrier gas and a fused sil-ica Varian Chrompack capillary column CP-Sil 8 CB Low Bleed/MS(30 m×0.25 mm id; 0.25 μm film thickness). The enantiomeric ex-cess values (%ee) of the epoxides were determined using the samechromatograph but using a fused silica Varian Chrompack capil-lary column CP-Chiralsil-Dex CB (25 m × 0.25 mm id; 0.25 μmfilm thickness). Conditions used: 333 K (3 min), 5 K min−1, 443 K(2 min), 20 K min−1, 473 K (10 min); injector temperature, 473 K;detector temperature, 573 K. The reaction parameters %C (sub-strate conversion), %S (alkene epoxide selectivity), TON, TOF andenantiomeric excess values (%ee) were calculated using the fol-lowing formula, where A stands for area of chromatographic peak:%C = {[A(alkene)/A(chlorobenzene)]t=0 h − [A(alkene)/A(chloro-benzene)]t=x h} × 100/[A(alkene)/A(chlorobenzene)]t=0 h; %S =[A(alkene epoxide)/A(chlorobenzene)] × 100/[A(all products)/A(chlorobenzene)]; TON = mmol of converted alkene/mmol Mn;TOF = TON/time of reaction; and %ee = [A(major enantiomer) −A(minor enantiomer)] × 100/[A(major enantiomer) + A(minorenantiomer)].

2.5. Catalytic experiments

The catalytic activity of the heterogeneous catalysts in theasymmetric epoxidation of alkenes was studied under constantstirring conditions using round bottom flasks of 50 cm3 as batchreactors. The experimental conditions were varied according to thesubstrate and oxidant in study, because different substrates and ox-idants show different reactivity. The following systems were used:(i) styrene/m-CPBA-NMO; (ii) 6-CN-2,2-diMeChromene/NaOCl and(iii) 6-CN-2,2-diMeChromene/PhIO. The corresponding experimen-tal conditions were: (i) 0.500 mmol of styrene, 0.500 mmolof chlorobenzene (internal standard), 2.50 mmol of NMO (co-oxidant), 0.100 g of heterogeneous catalyst and 1.00 mol of m-CPBA, in 5.00 cm3 of dichloromethane at 268 K (sodium chlo-ride in an ice bath); (ii) 0.500 mmol of 6-CN-2,2-diMeChromene,0.500 mmol of chlorobenzene (internal standard), 0.100 g of het-erogeneous catalyst and 0.75 mmol of NaOCl, in 5.00 cm3 ofdichloromethane at 273 K (ice bath); and (iii) 0.250 mmol of6-CN-2,2-diMeChromene, 0.500 mmol of chlorobenzene (internal

standard), 0.100 g of heterogeneous catalyst and 0.12 mmol ofPhIO, in 5.00 cm3 of acetonitrile, at room temperature.

During the experiments 0.05 cm3 aliquots were taken from so-lution, with a hypodermic syringe, filtered through 0.2 μm PTFEsyringe filters, and directly analyzed by non-chiral and chiral GC-FID.

The catalysts were then washed sequentially by Soxhlet extrac-tion with 100 cm3 of methanol and 100 cm3 of dichloromethanefor the system styrene/m-CPBA-NMO, with 100 cm3 of dichloro-methane and 100 cm3 of acetonitrile for the system 6-CN-2,2-diMeChromene/NaOCl, or with 100 cm3 of acetonitrile and100 cm3 of dichloromethane for the system 6-CN-2,2-diMe-Chromene/PhIO, for 2 h and dried under vacuum in a horizontaloven at 393 K, overnight. Then, they were reused using the sameexperimental conditions.

To provide a framework for the results obtained using the het-erogeneous catalysts, all three epoxidation systems were also car-ried out in homogeneous medium under experimental conditionscomparable to those described above, using the same amount ofJacobsen catalyst. The contribution of the support itself to the catal-ysis was found to be too small to be significant.

3. Results and discussion

3.1. Characterization of the materials

The results obtained for all samples by the various character-ization techniques (TPD, XPS and elemental analysis) will be dis-cussed in two stages: textural and chemical characterization.

3.1.1. Textural characterizationThe textural properties of all samples were obtained from anal-

ysis of the N2 adsorption isotherms at 77 K (Fig. 1) and the resultsare shown in Table 1. Fig. 2 shows the pore size distributions ob-tained by NLDFT.

As can be seen in Fig. 1, all isotherms are typical of materi-als with micro- and mesopores. Analysis of Table 1 shows that theSBET surface area increases with the extent of the gasification withCO2. There is also an increase in the values of the other texturalparameters with the extent of gasification, namely the Smeso andV micro. The Smeso almost triplicates between CA and CA300, andthe same occurs for the O2 oxidized samples. The volume of mi-cropores increases approximately 50% for the CO2 gasified samples,and the average width of the smaller micropores (L1) duplicates.This increase in the pore size with the extent of gasification par-allels the decrease in the areas of the smaller micropores, Sm1.Hence, the textural properties of the samples increase with theextent of gasification (high burn off values). The NLDFT results cor-roborate the previous observations, since an enlargement of thepore size distribution with the increase of the B.O. is clearly ob-served (Fig. 2a). Comparison of the textural parameters values forthe CO2 gasified samples and O2 oxidized samples in Table 1 showsthat the oxidation step has no significant effect on the texturalproperties of the activated carbon samples.

Within each sample type (CAX) (see for example Figs. 1band 2b), neither the oxidation step nor the treatment with sodiumhydroxide leads to significant changes on the textural properties ofthe materials.

However, with the metal complex anchoring, different ef-fects were observed with the different gasification times. ForCAT@CA_ONa, no significant changes occurred in the textural pa-rameters, suggesting that the complex is mainly located at the ex-ternal surface of particles. This may be due to diffusion constrainsimposed by the pores (small dimensions) to the Jacobsen complexas a consequence of its size: 1.74 by 1.18 nm considering the largest

F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323 317

(a)

(b)

Fig. 1. N2 equilibrium adsorption isotherms at 77 K: (a) samples obtained after oxidation; (b) sample CA300 submitted to different treatments.

Table 1Textural properties for the activated carbon based materials.

Sample B.O. (%) SBET

(m2 g−1)Smeso

a

(m2 g−1)V micro

a

(cm3 g−1)W01

b

(cm3 g−1)W02

b

(cm3 g−1)L1

c

(nm)Sm1

d

(m2 g−1)

CA 5.0 928 127 0.348 0.311 0.037 0.94 666CA60 19.1 1109 231 0.368 0.345 0.031 1.1 621CA180 43.8 1371 291 0.452 0.419 0.044 1.4 583CA300 51.1 1551 329 0.517 0.444 0.084 1.5 582

CA_O2 0.6 905 103 0.346 0.313 0.033 0.85 738CA60_O2 2.3 1057 151 0.387 0.357 0.030 1.0 684CA180_O2 3.0 1394 231 0.499 0.444 0.056 1.4 624CA300_O2 3.2 1578 343 0.537 0.447 0.091 1.6 571

CA_ONae – – – – – – – –CA60_ONae – – – – – – – –CA180_ONa – 1332 223 0.483 0.428 0.055 1.3 663CA300_ONa – 1553 385 0.487 0.421 0.093 1.5 555

CAT@CA_ONa – 903 130 0.335 0.303 0.031 0.95 640CAT@CA60_ONa – 968 187 0.324 0.303 0.033 1.1 549CAT@CA180_ONa – 1101 257 0.352 0.319 0.052 1.6 398CAT@CA300_ONa – 1257 263 0.427 0.356 0.074 1.7 411

a Micropore volume (V micro) and mesopore surface area (Smeso) calculated by the t-method.b W01 and W02 are the micropore volumes associated to small and large micropores, respectively, determined by the Dubinin method.c Average small micropores width.d Surface area of the small micropores.e Not determined due to the lack of material.

318 F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323

(a) (b)

Fig. 2. Pore size distribution obtained by NLDFT: (a) samples obtained after gasification; (b) sample CA300 submitted to different treatments.

(a)

(b)

Fig. 3. TPD spectra obtained for sample CA180 after gasification, oxidation and NaOH treatment: (a) CO and (b) CO2 spectra.

F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323 319

distances of the two farthest atoms in an axis, in perpendicular di-rections in the molecule [13]. In the case of CAT@CA60_ONa andCAT@CA180_ONa, there is a decrease of V micro, suggesting that inthese materials the complex was mainly immobilized within themicropores of large dimensions (W02) or at the entrance of thesmall micropores, which may block their access justifying the de-crease observed in W01. Finally, for CAT@CA300_ONa, there is adecrease of both V micro and Smeso. These variations indicate thatin this material the Jacobsen complex is immobilized in both meso-pores and micropores (the largest ones, W02, and at the entranceof those of lower dimensions, W01). This can be clearly seen inFig. 2b where the NLDFT results show a decrease of the pore vol-ume both in the range of micro and mesopores.

3.1.2. Chemical characterizationFig. 3 shows the TPD profiles of CO and CO2 for the gasified,

oxidized and NaOH treated, activated carbon CA180, which is rep-resentative of all others. Table 2 shows the amounts of CO and CO2released, obtained by integration of the areas under the peaks.

In the case of the CO2 gasified materials, the amount of surfaceoxygen groups are scarce, both for the CO and CO2 spectra, whencompared with the oxidized samples. The lack of a significantamount of surface groups in the CO2 gasified samples can be at-tributed to the high temperatures used in this treatment (1173 K).In the O2 oxidation step, a four-fold increase is observed in theamount of released CO and a three-fold increase for released CO2.After treatment with NaOH, there is a decrease in the amount ofCO released and an increase in the amount of released CO2.

The TPD peaks in the oxidized samples can be tentatively as-signed to the different functional groups by comparison with datataken from literature [14]. The CO spectra show two maxima ataround 950 and 1100 K, which may result from the decompositionof phenols and carbonyls/quinones. In the CO2 profiles, the low

Table 2Amounts of CO and CO2 evolved from the gasified, oxidized and NaOH refluxedactivated carbon samples, obtained by integration of the area under the TPD peaks.

Material CO (μmol g−1) CO2 (μmol g−1)

CA 759 126CA60 784 113CA180 606 81CA300 826 63

CA_O2 2853 318CA60_O2 3178 247CA180_O2 2972 124CA300_O2 3212 218

CA_ONa 2246 547CA60_ONa 2521 434CA180_ONa 2493 428CA300_ONa 2803 442

temperature peak may be assigned to carboxylic acid functions,while the higher temperature peak shows a shoulder at ∼850 Kand a maximum at ∼950 K, which may result from decompositionof carboxylic anhydrides and lactones. Thus, gas phase oxidationwith O2 increases anhydride, lactone, phenol and carbonyl/quinonesurface groups [14].

After reflux with the NaOH solution (CAX_ONa), the CO TPDprofiles show a decrease in the peaks around 950 and 1100 K,which can be explained by reaction of NaOH with phenol andanhydride surface groups. At the same time, there is an increasein the low temperature peak in the CO2 TPD profile, which maybe due to carboxylic acids and carboxylate groups, formed by hy-drolysis of carboxylic anhydrides promoted by the aqueous sodiumhydroxide solution.

The bulk manganese content of the final materials determinedby ICP-AES and XPS is presented in Table 3. As can be seen, thebulk manganese content increases with the gasification time andwith the mesoporous area (Smeso), showing an almost linear trend,as represented in Fig. 4.

The surface chemical composition of the materials was deter-mined by XPS. However, the method used in the preparation ofthe samples (see Section 2) prevents a reliable analysis of C and Oatomic %. In the high-resolution XPS spectra of all samples therewere no peaks due to Co 2p3 indicating that the purification ofthe parent materials was efficient in terms of cobalt removal. Nev-ertheless, there are in almost all the CAX and CAX_ONa materialspeaks due to N 1s and Cl 2p, which are due to some impuritiesderived from chemicals used in the preparation and purificationof the materials that remain within the materials. In this contextthe most relevant surface contents for this work are those of Naand Mn, although the Cl% can also give some qualitative importantinformation, and are summarized in Table 4.

All the materials CAX_ONa show peaks in the Na 1s region in-dicating the presence of sodium within the carbon materials as aresult of the treatment with sodium hydroxide that deprotonatedthe surface oxygen groups with acidic character, specifically thephenol group, which are in high quantity. Furthermore, all the ma-terials CAT@CAX_ONa show the peaks due to Mn 2p centered at642.0 eV, confirming the immobilization of the Jacobsen complex;the observed BE is typical of Mn(III) complexes with salen ligands[20,21], indicating that the oxidation number of Mn and complexstructure has been preserved after the immobilization procedure.

Analysis of the surface Mn contents in Table 4 indicates thatthe materials with the largest pores (CAT@CAX_ONa, X = 180and 300) show lower values than those with the smaller pores(CAT@CAX_ONa, X = 0 and 60). Another interesting feature thatcan be visualized in Table 3 is that the surface Mn content has theopposite behavior as that shown by bulk Mn content in relationto the gasification time and mesoporous area. The decrease of thesurface Mn content, which corresponds to a decrease in the ratio

Table 3Chemical and textural properties of the heterogeneous catalysts.

Heterogeneouscatalyst

tgas

(min)Smeso

(m2 g−1)aPhenolic groups

(μmol g−1)bMn (μmol g−1) Immobilization

yield (%)hICP-AES XPS

CAT@CA_ONa 0 103 804 42 (16c; 27d; 22e) 381 28CAT@CA60_ONa 60 151 816 76 (36c; 67d; 67e) 381 26CAT@CA180_ONa 180 231 880 84 (40c; 71d;g) 178 17CAT@CA300_ONa 300 343 1044 149 (60c; 122d; 71f) 160 22

a Determined by t-method.b Obtained by the deconvolution of the TPD spectra using a method described in [14].c After two successive catalytic cycles for styrene epoxidation with m-CPBA/NMO.d After two successive catalytic cycles for 6-CN-2,2-diMeChromene epoxidation with NaOCl.e After two successive catalytic cycles for 6-CN-2,2-diMeChromene epoxidation with PhIO.f After three successive catalytic cycles for 6-CN-2,2-diMeChromene epoxidation with PhIO.g Not determined due to the lack of enough material.h Obtained as: (amount of Mn by ICP-AES)/((amount of phenolic groups) × (SBET − Sm1)/SBET) × 100.

320 F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323

Fig. 4. Plot of Smeso and Mn content for the different carbon materials as a function of gasification time used in their preparation.

Table 4XPS elemental analysis of carbon based materials (at%).

Material Cl Na Mn

CA0 a a a

CA0_ONa a 0.39 a

CAT_CA0_ONa a 0.51 0.48

CA60 0.13 a a

CA60_ONa 0.14 0.57 a

CAT_CA60_ONa 0.18 0.58 0.49

CA180 0.11 a a

CA180_ONa 0.20 0.52 a

CAT_CA180_ONa 0.18 0.30 0.22

CA300 0.16 a a

CA300_ONa 0.29 0.75 a

CAT_CA300_ONa 0.15 0.36 0.20

a Not detected.

Mn(XPS)/Mn(ICP-AES) from 9.3 to 1.1 indicates that the metal com-plex distribution becomes more homogeneous within the porousmatrix, with the increase in the support porosity, confirming thetextural data referred above.

Some insights into the mechanism of complex immobiliza-tion can be deduced by the analysis of the Na and Cl surfacecontents. Comparing the materials CAX_ONa (X = 180 and 300)to the corresponding materials with the immobilized complex,(CAT@CAX_ONa, X = 180 and 300), there is a decrease in both Naand Cl% (more pronounced in Na%), what is a clear indication ofthe immobilization of the complex through axial coordination ofthe metal center to the phenolate surface groups, as depicted inScheme 1: the phenolate groups exchange the Na+ ions by the Mncomplex, with the concomitant formation of a covalent bond be-tween the metal center and the phenolate groups, which act asa stronger ligand compared with the chloride previously coordi-nated to the Mn center. For the materials with the lower porosity,CAX_ONa and CAT@CAX_ONa with X = 0 and 60, the Na% is almostinvariant, what prevents a reliable conclusion on the immobiliza-tion mechanism for the complex; nevertheless, we think that thesame mechanism must be operative in these materials, since allthe materials behave similarly in the catalytic test.

Assuming that all the complex will be immobilized mostlythrough the phenolic groups, the maximum Mn complex thatcould be anchored would be equal to the amount of those surfacegroups. A deconvolution procedure previously described [14] wasfollowed to estimate the amount of these groups on the CAX_O2samples (see Table 3). It was also considered that only the phe-nolic groups located in the large micropores (W02) and meso-pores would be accessible to the complex (this surface area canbe roughly estimated as SBET − Sm1). Thus, the maximum amount

of Mn(III) complex that could be anchored can be estimated asthe number of phenolic groups times (SBET − Sm1)/SBET (acces-sible phenolic groups). The complex immobilization yield can bethus estimated by dividing the amount of Mn determined by ICP-AES by the amount of accessible phenolic groups (see Table 4); itcan be observed that this value is in the range 17–28%.

3.2. Catalytic experiments

The catalytic activity of the new materials in the epoxidation ofstyrene and 6-CN-2,2-diMeChromene is collected in Table 5.

All new materials acted as enantioselective heterogeneous cat-alysts in the epoxidation of styrene, using m-CPBA as oxidant andNMO as co-oxidant, and 6-CN-2,2-diMeChromene, using NaOCl orPhIO as oxidants.

The heterogeneous catalysts CAT@CA180_ONa (in the epoxi-dation of 6-CN-2,2-diMeChromene with NaOCl and PhIO) andCAT@CA300_ONa (in the epoxidation of 6-CN-2,2-diMeChromenewith PhIO) present higher TON values than those corresponding tothe homogeneous phase, while the remaining prepared catalystspresent lower TON values (Table 5).

In general, in relation to the %ee, the prepared heterogeneouscatalysts show results in the same range as those obtained inhomogeneous phase, or even higher like CAT@CA180_ONa andCAT@CA300_ONa, which present higher values for the epoxidationof styrene with m-CPBA/NMO (Table 5).

In relation to the type of oxidant used in the epoxidation of6-CN-2,2-diMeChromene, it is observed that PhIO is more reactive(higher %C) than NaOCl, although it requires a higher reaction time(72 h) and leads to a smaller enantioselectivity (lower %ee).

It can also be observed that the general trend of the TON of theheterogeneous catalysts, for the epoxidation of the two alkenes,is to increase with the porosity of the activated carbons usedas support: CAT@CA_ONa < CAT@CA60_ONa < CAT@CA300_ONa <

CAT@CA180_ONa (Fig. 5). In relation to the enantioselectivity,the catalysts also show a similar trend for the epoxidation ofstyrene, that is, the increase of porosity also leads to an increasein %ee: CAT@CA_ONa < CAT@CA60_ONa < CAT@CA300_ONa =CAT@CA180_ONa (Fig. 6); for the epoxidation of 6-CN-2,2-diMe-Chromene with NaOCl the trend persists, except for CAT@CA-180_ONa, which is the catalyst with the smaller value of enan-tioselectivity, although they all present values of the same or-der of magnitude (80–85%): CAT@CA180_ONa < CAT@CA_ONa <

CAT@CA60_ONa < CAT@CA300_ONa. For the epoxidation of 6-CN-2,2-diMeChromene with PhIO, the trend is very similar and, onceagain, the catalysts show enantioselectivity values in the same or-der of magnitude (70–73%), with the exception of CAT@CA_ONa

F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323 321

Table 5Asymmetric epoxidation of alkenes catalyzed by the homogeneous and heterogenized Jacobsen catalysta.

Catalyst Alkene/oxidant Run t (h) %Cb %Sb TONc TOF (h−1)d %eee

CAT Styrene/m-CPBA-NMO 1 99 95 114 114 46CAT@CA_ONa 1st 4 19 76 20 5 30

2nd 4 7 49 4 1 13

CAT 6-CN-2,2-diMeChromene/NaOCl 1st 24 10 100 18 1 90CAT@CA_ONa 1st 48 5 89 9 0.2 82

2nd 48 2 100 4 0.1 70

CAT 6-CN-2,2-diMeChromene/PhIO 24 36 100 38 2 75CAT@CA_ONa 1st 72 30 32 9 0.1 64

2nd 72 22 100 22 0.3 45

CAT Styrene/m-CPBA-NMO 1 84 78 46 46 41CAT@CA60_ONa 1st 4 34 86 20 5 36

2nd 4 16 60 7 2 19

CAT 6-CN-2,2-diMeChromene/NaOCl 24 11 100 11 0.5 93CAT@CA60_ONa 1st 48 8 95 9 0.2 84

2nd 48 2 100 2 0.04 79

CAT 6-CN-2,2-diMeChromene/PhIO 24 36 100 20 1 73CAT@CA60_ONa 1st 72 36 88 17 0.2 71

2nd 72 21 89 10 0.1 54

CAT Styrene/m-CPBA-NMO 1 84 95 54 54 40CAT@CA180_ONa 1st 4 67 95 38 10 43

2nd 4 23 80 11 3 28

CAT 6-CN-2,2-diMeChromene/NaOCl 24 17 100 16 1 92CAT@CA180_ONa 1st 48 20 98 18 0.4 80

2nd 48 1 93 1 0.02 72

CAT 6-CN-2,2-diMeChromene/PhIO 24 34 100 16 1 73CAT@CA180_ONa 1st 72 50 96 24 0.3 70

2nd 72 33 100 17 0.2 63

CAT Styrene/m-CPBA-NMO 1st 1 90 84 27 27 40CAT@CA300_ONa 1st 4 74 96 24 6 43

2nd 4 16 79 5 1 21

CAT 6-CN-2,2-diMeChromene/NaOCl 24 33 96 17 0.7 91CAT@CA300_ONa 1st 48 22 84 10 0.2 85

2nd 48 3 100 2 0.03 56

CAT 6-CN-2,2-diMeChromene/PhIO 24 35 100 9 0.4 75CAT@CA300_ONa 1st 72 46 98 16 0.2 73

2nd 72 42 98 12 0.2 653rd 72 22 100 6 0.1 51

a For experimental catalytic conditions used, see experimental Sections 2.3 and 2.4.b Substrate conversion (C%) and alkene epoxide selectivity (S%) were determined by GC as described in Section 2.c Total TON based on the alkene conversion and epoxide selectivity.d TOF = TON/reaction time.e %ee were determined by chiral GC as described in Section 2.

which presents a slightly lower value (64%): CAT@CA_ONa <

CAT@CA180_ONa < CAT@CA60_ONa < CAT@CA300_ONa (Fig. 5).The advantage of supporting homogeneous catalysts is the pos-

sibility of their recycling and reusing, through a simple filtration ofthe heterogeneous catalyst from the reactional medium. Thus, thecatalysts were tested in a 2nd catalytic cycle using the same ex-perimental conditions, after a convenient process of purification inorder to remove all reagents and products that might have beenheld in the porous structure of the activated carbons. The resultsare presented in Table 5.

It is important to highlight that CAT@CA180_ONa, as well asCAT@CA300_ONa, still present a TON value (2nd run) higher thanthat obtained for the homogeneous phase for the epoxidation of6-CN-2,2-diMeChromene using PhIO as oxidant (Table 5). However,it was observed that, upon reutilization, there is in general a de-crease in the activity when compared to the 1st run. This fact canbe attributed to the degradation of the ligand, which can also beoxidized by the oxidant, and to the leaching of the Mn(III) com-plex, since a decrease in immobilized Mn(III) is observed after thecatalysis experiments (Table 3). The leaching of the complex washigher in some cases than others, due to the different resistance

of the complex towards the different oxidants. Therefore, thereis a trend in the percentage of leaching according to the oxidantused: NaOCl < PhIO < m-CPBA. Thus, the Mn(III) complex is moreresistant to NaOCl (lower leaching), while m-CPBA is the most ag-gressive oxidant for the Mn(III) complex and, consequently, leadsto a higher percentage of leaching (Table 3). These results com-bined with the observed trends between TON, %ee with the mate-rial porosity, may be taken as an indication that observed catalyticactivity for the materials has a negligible homogeneous componentdue to the active phase leaching.

Reaction times in heterogeneous phase are higher when com-pared with homogeneous phase reactions due to the slow diffusionof the reagents in the porous structure of the support, which leadsto lower TOF values. Nevertheless, lower diffusion limitations areobserved with the increase of the mesopore area.

4. Conclusions

Four nanostructured activated carbons were prepared with dif-ferent textural properties, namely, different mesoporous and mi-croporous areas and different pore sizes. The Jacobsen catalyst was

322 F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323

(a)

(b)

Fig. 5. Plots of TON and %ee for the different heterogeneous catalysts.

Fig. 6. Plot of %ee vs Smeso of the different catalysts.

immobilized onto these materials, giving rise to four new hetero-geneous catalysts.

TPD data clearly indicate that modification of the activated car-bon by reaction between sodium hydroxide with the surface phe-nol groups took place, originating more reactive phenolate groups.Characterization of the manganese(III) salen heterogeneous cata-lysts by XPS and ICP-AES indicate that the complex anchoring

proceeds through coordination of surface phenolate groups to themanganese(III) center.

All new materials acted as enantioselective heterogeneous cata-lyst in the epoxidation of styrene, using m-CPBA as oxidant, and6-CN-2,2-diMeChromene, using NaOCl or PhIO as oxidants. Re-action times in heterogeneous phase are higher when comparedwith homogeneous phase reactions, however, with the increase of

F. Maia et al. / Journal of Colloid and Interface Science 328 (2008) 314–323 323

the mesopore area, lower diffusion limitations are observed. TheTON of the heterogeneous catalysts, for the epoxidation of the twoalkenes, generally increases with the increase in porosity of theactivated carbons used as support. In relation to the enantioselec-tivity, the catalysts also show a similar trend for the epoxidation ofstyrene, that is, an increase in porosity also leads to an increase in%ee. In the case of CAT@CA180_ONa and of CAT@CA300_ONa, %eefor the epoxidation of styrene with m-CPBA/NMO present highervalues when compared to the homogeneous counterparts. A trendin the percentage of leaching can be established according to theoxidant used: NaOCl < PhIO < m-CPBA.

Acknowledgments

The authors thank NORIT N.V., Amersfoort, The Netherlands, forproviding the activated carbon. This work was funded by Fundaçãopara a Ciência e Tecnologia (Portugal) and FEDER through projectRef. POCI/EQU/57369/2004.

References

[1] H. Zhang, Y. Zhang, C. Li, J. Catal. 238 (2006) 369.[2] T. Katsuki, Coord. Chem. Rev. 140 (1995) 189.[3] C. Vartzouma, E. Evaggellou, Y. Sanakis, N. Hadjiliadis, M. Louloudi, J. Mol. Catal.

A 263 (2006) 77.

[4] K. Srinivasan, P. Michaud, J.K. Kochi, J. Am. Chem. Soc. 108 (1986) 2309.[5] A.R. Silva, C. Freire, B. de Castro, M.M.A. Freitas, J.L. Figueiredo, Microporous

Mesoporous Mater. 46 (2001) 211.[6] A.R. Silva, M. Martins, M.M.A. Freitas, A. Valente, C. Freire, B. de Castro, Micro-

porous Mesoporous Mater. 55 (2002) 275.[7] A.R. Silva, C. Freire, B. de Castro, M.M.A. Freitas, J.L. Figueiredo, Langmuir 18

(2002) 8017.[8] A.R. Silva, M. Martins, M.M.A. Freitas, J.L. Figueiredo, C. Freire, B. de Castro, Eur.

J. Inorg. Chem. (2004) 2027.[9] A.R. Silva, J. Vital, J.L. Figueiredo, C. Freire, B. de Castro, New J. Chem. 27 (2003)

1511.[10] A.R. Silva, J.L. Figueiredo, C. Freire, B. de Castro, Microporous Mesoporous

Mater. 68 (2004) 83.[11] A.R. Silva, B. de Castro, C. Freire, Carbon 42 (2004) 3027.[12] M. Cardoso, A.R. Silva, B. de Castro, C. Freire, Appl. Catal. A 285 (2005) 110.[13] A.R. Silva, V. Budarin, J.H. Clark, B. de Castro, C. Freire, Carbon 43 (2005) 2096.[14] J.L. Figueiredo, M.F.R. Pereira, M.M.A. Freitas, J.J.M. Órfão, Carbon 37 (1999)

1379.[15] P. Piaggio, P. McMorn, D. Murphy, D. Bethell, P.C.B. Page, F.E. Hancock, C. Sly,

O.J. Kerton, G.J. Hutchings, J. Chem. Soc. Perkin Trans. 2 (2000) 2008.[16] J.L. Figueiredo, J.J.M. Órfão, M.C.A. Ferraz, Fuel 63 (1984) 1059.[17] F. Rodríguez-Reinoso, J.M. Martin-Martinez, C. Prado-Burguete, B. McEnaney,

J. Phys. Chem. 91 (1987) 515.[18] A. Linares-Solano, J.L. Figueiredo, J.A. Moulijn (Eds.), Carbon and Coal Gasifica-

tion, Martinus Nijhoff Publishers, Dordrecht, 1986, p. 137.[19] H.F. Stoeckli, L. Ballerini, S. De Bernardini, Carbon 27 (1989) 501.[20] I. Kuzniarska-Biernacka, A.R. Silva, A.P. Carvalho, J. Pires, C. Freire, J. Mol. Catal.

A Chem. 278 (2007) 82.[21] A.R. Silva, V. Budarin, J.H. Clark, C. Freire, B. de Castro, Carbon 45 (2007) 1951.