Reactivity of mesoporous palladium yttria-stablilized zirconia for solution phase reactions

Carl Poulin, Matthew A. Brown, Yamile A. Wasslen, Catherine M. Grgicak,Keith Fagnou, and Javier B. Giorgi

Abstract: A reduced mesoporous 5% Pd-YSZ precatalyst was observed to be very active for the Heck reaction provid-ing up to 770,000 TON. Three-phase tests have experimentally confirmed that a leached homogeneous species is re-sponsible for the activity. Residual Pd concentration in solution is minimal (<0.10 ppm) and therefore this material haspotential applications because it can be easily removed by filtration. This paper describes the synthesis, structural char-acterization, and catalytic activity of the mesoporous Pd-YSZ material. The material was prepared using anamphiphillic surfactant templating method resulting in a homogeneous dispersion of Pd in an ordered mesoporousstructure. Material properties were fully characterized using X-ray diffraction, surface area analysis, and electron mi-croscopy.

Key words: mesoporous catalysts, palladium, yttria-stabilized zirconia, Heck reaction, Pd-YSZ.

Résumé : On a observé qu’un précatalyseur 5% Pd-YSZ mésoporeux réduit est très actif pour la réaction de Heck quipeuvent fournir jusqu’à 770,000 TON. Des essais en trois phases ont permis de confirmer expérimentalement qu’uneespèce homogène lessivée est responsable pour cette activité. La concentration résiduelle de palladiium dans la solutionest minimale (<0,10 ppm); ce matériau a donc des applications potentielles compte tenu du fait qu’il peut être récupérépar filtration. Dans ce travail, on décrit la synthèse, la caractérisation de la structure et l’activité catalytique du matérielmésoporeux Pd-YSZ. Ce produit a été préparé en faisant appel à la méthode de matrice à agent de surface amphiphilequi fournit une dispersion homogène du palladium dans une structure mésoporeuse ordonnée. Les propriétés du maté-riau ont été complètement caractérisées à l’aide de la diffraction des rayons X, l’analyse de la superficie de la surfaceet la microscopie électronique.

Mots clés : catalyseurs mésoporeux, palladium, yttrium stabilisé par du zirconium, réaction de Heck, Pd-YSZ.

[Traduit par la Rédaction] Poulin et al. 1528

Introduction

Palladium-catalyzed transformations have become wide-spread in the construction of new carbon–carbon andcarbon–heteroatom bonds (1–5). Recent advances in the de-velopment of new homogeneous palladium catalysts haveenabled these transformations to be carried out even withvery deactivated substrates such as aryl chlorides under mildconditions (6). In the context of product purity, however, theuse of homogeneous catalysts has detractions including thefact that the palladium must be removed from the organicproduct at the end of the reaction (7, 8). This can sometimesbe very challenging and costly. With this challenge in mind,chemists have been studying the use of heterogeneous palla-dium catalyst systems as viable alternatives to the more in-tensely studied homogeneous catalysts (9). In recent years, avariety of heterogeneous supports have appeared including

carbon, silicates, zeolites, and block copolymers (10–16).Alternatively, homogeneous catalysts have been “heterogenized”via immobilization of active homogeneous catalysts on vari-ous solid supports (17–26). With all of these heterogeneouscatalysts, however, there is debate with respect to the truenature of the active catalyst (27). Some catalysts do appearto function as true heterogeneous catalysts (28, 29), whileothers have been shown to be catalyst reservoirs(precatalysts) from which an active catalyst leaches into so-lution (28, 30–38). Analysis of these reactions is furthercomplicated by a third option in which a release and capturemechanism involving catalyst leaching and subsequentreadsorption onto the matrix support occurs upon reactioncompletion. Since the question of catalyst form is of para-mount importance, more studies are warranted in this area,particularly when new catalysts and (or) supports are beingdesigned and studied. Simple coupling reactions such asHeck or Suzuki can be used as the standard to test catalystperformance of heterogeneous catalysts.

In the design of heterogeneous catalysts, maximization ofthe active surface area is the key to performance togetherwith the nature of the catalytic species. Metal catalysts suchas palladium have been shown to be in their most activestate when speciated as nanoparticles (39). In this state, amaximum of palladium sites are exposed while the catalyticproperties are controlled by the electronic structure of the

Can. J. Chem. 84: 1520–1528 (2006) doi:10.1139/V06-154 © 2006 NRC Canada

1520

Received 15 July 2006. Accepted 9 September 2006.Published on the NRC Research Press Web site athttp://canjchem.nrc.ca on 8 November 2006.

C. Poulin, M.A. Brown, Y.A. Wasslen, C.M. Grgicak,K. Fagnou, J.B. Giorgi.1 Center for Catalysis Research andInnovation, Department of Chemistry, University of Ottawa,10 Marie Curie, Ottawa, ON K1N 6N5, Canada.

1Corresponding author (e-mail: jgiorgi@uottawa.ca).

nanoparticle and influenced by its support. The optimizationof the active surface area was revolutionized with the dis-covery of mesoporous structures, which provide a maximumtotal surface area with channels of the appropriate size forreactant transport (40–42). Examples of the improved per-formance of mesoporous silica-based catalysts date back totheir petrochemical origins (43) and enhanced propertiescontinue to surface, as recently demonstrated for ethanehydrogenolysis and ethylene hydrogenation (44), hydrogena-tion of phenyl acetylene (45), and anti-stokes luminescenceof encapsulated dyes (46). A particularly appropriate exam-ple has recently used mercaptopropyl modified silica to an-chor Pd atoms to the SBA-15 zeolitic framework (16). Themodified zeolite is an excellent Pd scavenger and has beenused successfully as a Pd catalyst for low-turnover high-yield coupling reactions with minimal Pd leaching.

A considerable effort has been made over the past15 years to obtain mesoporous metal oxide materials thatpossess enhanced properties in catalysis, charge transport,and host–guest affinities. Since the family of silicate meso-porous molecular sieves was first synthesized (40–42), ef-forts have been made toward the synthesis of mesoporoustransition metal oxide materials (47–50) and, in particular,mesoporous yttria and zirconia materials have been recentlydeveloped (51–53). The general synthetic methodology toprepare mesoporous oxides uses an amphiphillic surfactantas a template. In solution, the surfactant agglomerates intomicelles and the inorganic oxide framework grows around it.Upon removal of the organic species, a mesoporous frame-work remains. In many cases, the mesoporous framework isimpregnated with the metal catalyst, however we used thismethodology to create not only the support framework butalso the homogeneous distribution of the nanoparticle-catalyst within the framework. Coproduction of the catalystwith the support produces a homogeneous distribution of themetal and is particularly suitable for high metal loadingswhere metallic connectivity and (or) conductivity is desired.For low metal loadings, coproduction ensures a homoge-neous distribution of the catalyst in the porous framework.The obstacle of relatively low solubility of the metal precursorscan be overcome by the synthesis of yttrium and zirconiumglycolates as precursors for the preparation of yttria-

stabilized zirconia (YSZ) with mesoporous structure (54–56).

In this paper, we report the synthesis and characterizationof a versatile palladium catalyst supported on a mesoporousyttria-stabilized zirconia (YSZ) framework. These materials,which are characterized by a high surface area, have the po-tential to be used in a liquid–solid regime for the catalysis oforganic reactions (57, 58), which has been validated by dem-onstrating high turn-over numbers in the Heck reaction. Cat-alyst structure along with reaction kinetics and mechanismsare discussed.

Experimental

Pd-YSZ preparationThe general synthesis involves the formation of yttrium and

zirconium glycolates separately by reaction of ZrCl4/(OEt)3with ethylene glycol. The solutions are combined and aPd+2 source is added in appropriate proportion according tothe desired loading. Cetyltrimethylammonium bromide(CTAB) is added as a template to form the mesoporousstructure. The solution is allowed to age and finally the re-sulting precipitate is filtered, washed, and dried. This “as-synthesized” (dried) material consists of the micellar CTABarray around which the Pd-YSZ matrix has been formed.Upon calcination, the organic component is eliminated giv-ing rise to the porous structure (Fig. 1).

Details of the procedure are as follows. Zirconiumethoxide (5.0 g, Strem Chemicals Inc., 99.9+%) and2.4 equiv. of NaOH (1.8 g) were added to a flask containingexcess ethylene glycol (51 mL, spectroscopic grade, Sigma-Aldrich Chemical Ltd.). The resulting mixture was thenrefluxed at 180 °C under flowing nitrogen for 24 h. Excessethylene glycol was distilled off leaving a thick yellowishgel (zirconium glycolate). Separately, yttrium glycolate wasprepared by adding yttrium acetate tetrahydrate (1.2 g, AlfaAesar, 99.9%) to ethylene glycol (31 mL) and stirring. Thezirconium and yttrium glycolates were then mixed in adropwise fashion while the zirconium glycolate was still hot(~50 °C). The resulting solution was added to a polypropy-lene bottle containing a Pd+2 source (Pd(NO3)2, Sigma-Aldrich Chemical Ltd., 99.999%, 2.0 mL solution 10 wt% in

© 2006 NRC Canada

Poulin et al. 1521

Fig. 1. Mechanism of formation of mesoporous Pd-YSZ.

10 wt% nitric acid), cetyltrimethylammonium bromide(CTAB, Sigma-Aldrich Chemical Ltd., 4.1 g), NaOH(1.7 g), and water (120 mL). The mixtures were stirred atRT for 25 min and then put into an oven for 5 days at 80 °C.The solid products obtained were collected by vacuum filtra-tion and washed two times with water to remove residualchlorine and CTAB. Samples were then dried at 120 °C for6 h and denoted as-synthesized. This procedure produces aPd loading of 5%; for different loading the Pd concentrationis simply adjusted.

Calcination to 600 °C or 950 °C was performed in air for2 h to ensure the crystallization of the YSZ. The calcinationprocedure also produces PdO. When reduced Pd is required,reduction was performed at 500 °C under H2 flow for 2 h forsamples previously calcined to 600 °C, and at 800 °C underN2 flow for 2 h for samples previously calcined to 950 °C. Areduced material containing 5 mol% Pd and calcined to950 °C prior to reduction shall be denoted as 5%Pd-YSZ950and one precalcined to 600 °C as 5%Pd-YSZ600.

Pd-YSZ characterizationPowder composition and crystallite sizes were determined

by X-ray diffraction (XRD) (Phillips PW1830) using Cu Kαradiation with a wavelength of 1.54 Å. Crystalline phaseswere assigned using the Powder Diffraction File database(ICCD, 2001). The crystallite size analysis of each phasewas performed using the Williamson–Hall methodology,which takes into account crystallite strain (59). Vibrationalcharacterization of the powders prior to calcination was de-termined using Attenuated Total Reflection Fourier trans-form IR spectroscopy (ATR-FTIR) (Bruker Equinox 55).The BET surface area of the calcined powders were deter-mined after reduction in flowing H2 (5% in N2) at 500 °Cfor 2 h, using a Quantachrome Autosorb 1-C instrument. ABJH analysis provides pore size information and the averagepore size was calculated using the Kelvin equation.

General procedure for the Heck reaction ofbutylacrylate with iodobenzene

To a round-bottomed flask fitted with a reflux condenserwas added iodobenzene (4.0 mmol, 0.45 mL), butylacrylate(4.8 mmol, 0.69 mL), N,N-diisopropylethylamine (DIPEA)(8 mmol, 1.4 mL), and N,N-dimethylformamide (DMF)(1.4 mL). To this solution was added the precatalyst 5% Pd-YSZ (0.004 mmol, 9.7 mg). The reaction was then heated to140 °C. Reaction progress was monitored by GC-MS analy-sis of the crude reaction mixture against a decane internalstandard. For all reactions, the mol ratio iodobenzene–butylacrylate–DIPEA was kept at 1:1.2:2.

Triple phase test

Preparation of resinTo a mixture of Wang resin (0.3 equiv.) and iodobenzene

(1 equiv.) in DMF (0.1 mol/L) was added dicyclohexyl-carbodiimide (1 equiv.) and a few grains of 4-dimethyl-aminopyridine as catalyst. The reaction mixture was heatedat 60 °C for 4 days. The solution was filtered and washedtwice with each of the following solvents: water, methanol,dichloromethane (DCM), and diethyl ether.

Capping of the unreacted hydroxyls on the resinTo a suspension of the resin (1 equiv.) in dichloromethane

(0.1 mol/L) was added acetic anhydride (5 equiv.) andpyridine (5 equiv.). The reaction was stirred for 24 h at RT.The reaction mixture was filtered and washed twice witheach of the following solvents: water, methanol, DCM, andether to afford the acetate protected resin.

Reaction and cleavage of compound from the resinTo a mixture of acetate protected resin (1 equiv.) was

added butyl acrylate (3 equiv.), DIPEA (3 equiv.), Pd-YSZ(0.05 equiv.) in DMF (0.16 mol/L). The reaction mixturewas heated at 130 °C for 3 h. The resin (1 equiv.) was fil-tered and washed thoroughly before being stirred in a 1:1mixture of trifluoroacetic acid and dichloromethane(0.5 mol/L) for 2 h. The resulting mixture was filtered andwashed twice with each of the following solvents: methanol,DCM, and ether. The filtrate was put under reduced pressureat 70 °C to remove the volatiles as well as the trifluoroaceticacid. This afforded compound 5 (eq. [1]).

Thiol scavenger testTo a suspension of thiol resin (0.5 equiv.) was added

iodobenzene (1 equiv.), butyl acrylate (1.2 equiv.), DIPEA(2 equiv.), and Pd-YSZ (0.01 equiv.) in DMF (2.7 mol/L).The reaction mixture was heated to 130 °C and monitoredby GC. After prolonged reaction time, no greater than 2%conversion was detected.

Results and discussion

Pd-YSZ material propertiesThe as-synthesized material containing the residual hydro-

carbon micelles possesses long-range order with d-spacingof 40 Å, as seen by the low-angle peak in the XRD pattern(Fig. 2). Upon calcination in air to 600 °C, the hydrocarbonis eliminated leaving a porous mesostructure with d-spacingof 41 Å. The low-angle peak is decreased in intensity, con-sistent with a partial disruption of the long-range order. Theobserved change in d-spacing is consistent with a thickeningof the pore walls as the crystals of YSZ become larger (56).

© 2006 NRC Canada

1522 Can. J. Chem. Vol. 84, 2006

Fig. 2. Small angle XRD pattern of 20%Pd-YSZ for (a) as-synthesized, (b) calcined to 600 °C, and (c) calcined to 950 °C.

Finally, upon calcination to 950 °C the material is describedas having lost its long-range order.

The shape of the low-angle peak for the as-synthesizedmaterial is observed as very broad and asymmetrical. Such ashape has been observed by Ozin and coworkers (54–56) insimilar materials and discussed in terms of a poor long-range order of the mesoporous structure. An alternativeexplanation is the superposition of two peaks for differentorientations of the long-range order structure (53). Kim et al.(53) assigned their peak at 2θ ≈ 3.9° to the (110) reflectionof a hexagonal arrangement of meso-zirconyl chloride. Inthis case, the metastable structure present when thesurfactant is still in the pores is disturbed when the organiccomponent is removed, leaving a single peak in the XRDpattern of the calcined samples.

The mesoporous structure of the material calcined to600 °C is also reflected by the intermediate surface area,typically in the range of 150 m2 g–1, and a slightly asymmet-ric pore size distribution that peaks at 19 Å in diameter withan average pore size diameter of 26 Å, as calculated by theKelvin equation. This value is smaller than those obtainedfor perfectly structured mesoporous materials such aszeolites, but it is an order of magnitude larger than mostcommercial YSZ catalyst powders (Maketech Materials 70%NiO/8YSZ, 9.8 m2 g–1 and 34% NiO/8YSZ, 6.63 m2 g–1).The surface area decreases to �10 m2 g–1 upon calcination to950 °C indicating the predominant collapse of the porousstructure.

Removal of the pore-forming molecule CTAB from thematerial is achieved during the calcination process and iscomplete by 600 °C. This is illustrated in Fig. 3, whichshows ATR-FTIR spectra of a sample as-synthesized and af-ter calcination. The two peaks present at 2850 and 2916 cm–1

in the as-synthesized sample are characteristic of the C–Hstretching modes in the CTAB long chain. These peaks areremoved upon calcination. The regions around 1500 and3400 cm–1 are characteristic of water and can be seen to de-crease substantially upon calcination. Peaks at �1000 cm–1

are associated with yttrium and (or) zirconium species, buttheir identification is difficult due to the large broadening(60).

It is interesting to note that the CTAB species are not re-moved at a single temperature and in fact they appear to bepartially decomposed and trapped within the porous struc-ture. Thermogravimetric analysis in helium shows weightloss at several distinct temperatures (Fig. 4). Correlation ofthe weight loss with mass analysis shows that in the temper-ature range below 60 °C, water is the main componentdesorbed. The hydrocarbon moieties represented by the ethylgroup (m/e = 29) are desorbed in the subsequent two peaksat 280 and 440 °C. Trapping and decomposition of theCTAB is suggested by the simultaneous observation of zir-conium (m/e = 91) and the tertiary amine group (m/e = 57),which come out from the matrix at 440 °C. These observa-tions have direct implications on the stability of themesoporous structure. The binding and (or) trapping ofCTAB to the YSZ network is sufficiently strong that the re-moval of the surfactant disrupts the pores. A more inertsurfactant may improve the thermal stability of the YSZ po-rous structure.

Palladium and YSZ crystallites are present in the material

as individual phases as seen from their characteristic powderdiffraction pattern. The large width of their peaks suggestsextremely small crystallite sizes, particularly for YSZ. In-creasing the calcination temperature results in an increase insize of the YSZ crystallites, as observed by a decrease inpeak widths in the XRD patterns (Fig. 5).

During the calcination process, the crystallites of palla-dium originally present (Fig. 5) also increase in size. How-ever, as the temperature rises Pd is oxidized to PdO. At atemperature of 840 °C, PdO is reduced to Pd once again.The reduction of PdO to Pd can be readily observed in theTGA experiment performed in air as a sharp decrease inmass (Fig. 6). It is interesting to note that the reoxidation ofPd to PdO during cooling occurs more gradually around650 °C, a hysteresis of 190 °C. The observed temperature ofPdO reduction in air is high for zirconia-based catalysts, butit is consistent with literature (61–63) for other Pd-oxide cat-alysts. The ratio of oxidized vs. metallic palladium observedin the XRD (at RT, after calcination to 950 °C, Fig. 5) de-pends on the sample cooling rate in air. Calcination to

© 2006 NRC Canada

Poulin et al. 1523

Fig. 3. ATR-FTIR of 20%Pd-YSZ for (a) as-synthesized and(b) calcined to 600 °C. The CTAB spectrum is shown for com-parison.

Fig. 4. Thermogravimetric analysis of 20%Pd-YSZ in Helium.The evolution of m/e = 29 (�), m/e = 57 (�), and m/e = 91 (�)as a function of temperature are also shown.

600 °C in air produces only the PdO phase. To obtain metal-lic Pd exclusively, the samples can be calcined and cooled inan inert atmosphere.

Palladium sintering effects in heterogeneous catalysis area general concern for temperature treatments above 800 °Cbecause coarsening leads to a decrease in catalytic activity(61–63). In these mesoporous Pd-YSZ materials, palladiumcrystallites remain constant in size up to �800 °C during theoxidation and subsequent reduction process. Calcination tohigher temperatures does increase the Pd crystallite size, butonly to a limited extent (Table 1). The ultimate average crys-tallite size upon calcination depends on the palladium con-tent with a maximum coarsening of three-fold observed for20% Pd loading. The observed increase in crystallite sizecompares very well to increases of over 10-fold for Pd-Al2O3 (64, 65) and nanoengineered 3%Pd-YSZ (66) cata-lysts at similar temperatures. For all samples, the YSZ aver-age crystallite size after calcination to 950 °C is around20 nm, regardless of Pd loading. The relatively stablemesostructure of the YSZ, which partially survives the calci-nation process, seems to be responsible for the low degree ofsintering of palladium. Figure 7 shows an SEM image of

5%Pd-YSZ950R (reduced palladium) where the homoge-neous nature and the small particle size of the material canbe assessed.

Application to the Heck ReactionA 5 mol% material with the highest possible surface area

was chosen for catalytic experiments. The mesoporous5 mol% Pd-YSZ framework possesses long-range order witha d-spacing of 40 Å (by XRD). Crystallites sizes of 10 nmand 5 nm for Pd and YSZ respectively were measured forsamples calcined to 600 °C. This highly reactive 5%Pd-YSZ600R has a total surface area of 150 m2/g with an aver-age pore size diameter of 26 Å. For comparison, a 5% Pdloading material calcinated to 950 °C in the reduced formwas chosen (5%Pd-YSZ950R). Higher temperatures duringthe material preparation process disrupt the long-range orderof the mesoporous structure and lead to the overall reductionof surface area to 10 m2/g. However, up to 1000 °C the in-crease in the crystallite size of the two components is rela-tively small, �2×. Generally, catalytic materials with smallertotal surface areas are not expected to perform as well underreaction conditions. However, for selected reactions the

© 2006 NRC Canada

1524 Can. J. Chem. Vol. 84, 2006

Fig. 5. XRD pattern of 20%Pd-YSZ for (a) as-synthesized,(b) calcined to 600 °C, and (c) calcined to 950 °C, with Pd (�),PdO (#), YSZ (�) peaks.

Fig. 6. Thermogravimetric analysis of 20%Pd-YSZ in air.Heating and cooling traces are shown. PdO→Pd (�) upon heat-ing, T = 860 °C, Pd→PdO upon cooling (�), T = 680 °C.

Samplecomposition Conditions

Pd crystallites(nm±esd)a

PdO crystallites(nm±esd)a

YSZ crystallites(nm±esd)a

5% Pd-YSZ Calc. 600 °C Not present 9.3±2.7 5.0±0.5Calc. 950 °C 29.8±9.4b 21.6±1.4b 14.1±0.2

20% Pd-YSZ As-synthesized 41.8±6.9 Not present AmorphousCalc. 600 °C Not present 39.6±6.5 4.2±0.1b,c

Calc. 950 °C 133±32 14.8±1.1 21.8±1.150% Pd-YSZ As-synthesized 34.8±4.0 Not present Amorphous

Calc. 600 °C Not present 19.9±2.9 4.7±0.1b,c

Calc. 950 °C 90±16 22.3±2.0 21.1±1.2

Note: Crystallite sizes were measured by XRD using the Williamson–Hall analysis. Multiple peak analysisconsisted of, 4 peaks for Pd, 6 peaks for YSZ, and 4 peaks for PdO.

aEstimated standard deviation is based on the peak fitting/refinement routine (JADE diffraction analysis).bSingle peak analysis (111) performed by Scherrer method.cOnly a few broad peaks are available for analysis. YSZ is mostly amorphous in these samples.

Table 1. Crystallite sizes for selected materials.

high-temperature treatment may be preferred because it addsmechanical, chemical, and thermal stability to the catalysts,which have a direct impact on the reusability of the catalysts.

Preliminary work showed that both of these precatalystswere active for the Heck reaction, even prior to the reductionstep. The subsequent sections compare the performance ofthese two precatalysts and provide a discussion of their ef-fectiveness and the reaction mechanism involved for a stan-dard Heck coupling reaction.

Reaction optimizationWith an active catalyst in hand, the reaction conditions

were optimized with respect to solvent, base, temperature,and concentration. In early screens, dimethylformamide(DMF) emerged as the optimal solvent. A base scan also re-vealed that the best outcomes were achieved with organicbases such as diisopropylethylamine (DIPEA). Inorganicbases such as potassium bicarbonate lead to lower turn-overnumbers (TONs) and slower reaction. All subsequent opti-mization studies were performed in DMF with DIPEA asbase.

Initial optimization reactions were performed with 5%Pd-YSZ950R (reduced Pd). From three experiments performedat 120, 130, and 140 °C, it was determined that faster reac-tions and improved TONs were achieved at 140 °C. For ex-ample, after 60 min of reaction time less than 10%conversion was observed at 120 °C compared with 47% and65% conversion at 130 and 140 °C, respectively (Table 2,entries 1–3).

These reactions are very sensitive to concentration as il-lustrated in Table 2, entries 4–7. Four different iodobenzeneconcentrations were assayed (0.2, 0.5, 1.0, and 1.2 mol/L)

and each of the reactions was analyzed after 30 min. Atthese intervals, neither reaction at 0.2 and 0.5 mol/L showedany conversion to product. On the other hand, reactions per-formed at 1.0 and 1.2 mol/L had progressed to 35% and75% conversion after 30 min, respectively, showing that afaster initial reaction occurs at elevated concentration. It isinteresting to note that in Table 2, entries 4–7, the substrateto catalyst ratio (S/C) is kept constant and therefore the ef-fect of concentration is different from the homeopathic ef-fect described by de Vries (27) where a large S/C ratioprevents Pd nucleation and subsequent deactivation of thecatalytic species.

A comparison of the two precatalysts 5%Pd-YSZ950Rand 5%Pd-YSZ600R was carried out. From these experi-ments (Table 2, entries 8–11), the 5%Pd-YSZ600R emergesas a superior catalyst as evidenced by the percentage conver-sion at 30 min (Table 2, entry 10 vs. 8) and 60 min (Table 2,entry 11 vs. 9). With these conditions in hand, reactionswere performed to determine the maximal TON that couldbe achieved with this new catalyst. We were delighted to seethat under optimal conditions, up to 770,000 TON could beachieved after approximately 7 days (Table 2, entry 12).

Elucidation of the active speciesWith an active catalyst in hand, we were interested in de-

termining unequivocally if this was a heterogeneous catalystor if palladium was leaching off of the support to produce anactive homogeneous species. Two different approaches weretaken to answer this question based on the three-phase testconcept (67–70). In the first case, the aryl halide was an-chored to a Wang resin (71). If the true catalyst in these pro-cesses was purely heterogeneous then interaction betweenthe supported substrate and the matrix bound catalyst shouldnot be possible, thus preventing reaction. In contrast, if pal-ladium leaching occurs and if this soluble palladium speciesis catalytically active, then Heck reaction at the solid-supported 4 should be observed. When supported 4 was re-acted under standard conditions followed by separation ofthe solid phase, washing of the Wang resin, and cleavage ofthe substrate–product mixture from the solid support viatreatment with trifluoroacetic acid in dichloromethane (72,73), approximately 50% conversion to 5 was observed(eq. [1]). This lends strong support to the notion that a ho-mogeneous active catalyst is generated under the reactionconditions via leaching from the YSZ support.

© 2006 NRC Canada

Poulin et al. 1525

Fig. 7. Scanning electron micrograph of 5%Pd-YSZ950. Image istaken with an r-filter for improved contrast. High palladium con-centration areas appear brighter, although Pd is homogeneouslydistributed in this sample as confirmed by energy dispersionmeasurements.

O

Wangresin

1) Pd-YSZ, DIPEA,

DMF, 140 ºC

2) TFA, CH2Cl2(cleavage from

solid support)

O

I

HO

O

I

HO

O

OBu

O

OBu

O

50% conversion

4

5

[1]

To determine if an active heterogeneous catalyst specieswas also contributing to the direct arylation reaction, anotherthree-phase test was conducted. In this experiment 1 was re-acted under standard conditions in the presence of a silicasupported thiol-based scavenger resin. Since thiols areknown to have a very high affinity for palladium, the pres-ence of this heterogeneous metal scavenger should sequesterthe homogeneous catalyst as it is produced. On the otherhand, a heterogeneous palladium species that is catalyticallyactive should not be poisoned by the presence of the hetero-geneous metal scavenger. After 16 h, less than 5% of the de-sired product had been formed while an identical reactionrun in parallel without the scavenger resin reached 100%conversion (eq. [2]). This indicates that the heterogeneouspalladium is not catalytically active on the timescale exam-ined and that the observed reactivity arises from the in situgeneration of a homogeneous catalyst.

Having shown that the mechanism of reaction is homoge-neous, we investigated how much Pd is lost during the pro-cess. As it was not possible to identify Pd in solution, wedetermined the amount of palladium before and after reac-tion to obtain an upper limit to the amount of palladium re-maining in solution after reaction. Figure 8 shows the XRDpattern of catalysts 5%Pd-YSZ950R before and after under-going a reaction. XRD was chosen to search for PdO, whichcould have been formed during reaction, but was not ob-served. Although XRD is not a traditional quantitative tech-nique, we can clearly see that the ratio of Pd to YSZ has

decreased substantially. Since YSZ is not affected by the re-action, it can be considered as an internal standard for theXRD measurement allowing us to use the intensity ratio ofthe variable (Pd) to the standard (YSZ) for quantization pur-poses. A quantitative value can be obtained by measuringthe intensity ratio of the most intense peaks for Pd and YSZ(at 40 and 30 °, respectively) and comparing this to a cali-bration curve obtained for variable amounts of Pd in YSZ.The technique cannot measure atomic Pd in the sample, butafter a sintering treatment to allow palladium aggregation,no change was observed. If we consider that all the Pd notpresent in crystalline form in our sample has remained in so-lution, we obtain an upper limit to the amount of Pd leached.Estimating that about 80% of the Pd was lost during reac-

© 2006 NRC Canada

1526 Can. J. Chem. Vol. 84, 2006

Entry CatalystSubstrate–CatalystRatio Temp (°C) Conc. [B]a

ReactionTime (min) % Conc.b

1 A 100 140 0.2 60 652 A 100 130 0.2 60 473 A 100 120 0.2 60 94 A 1 000 140 0.2 30 05 A 1 000 140 0.5 30 06 A 1 000 140 1.0 30 357 A 1 000 140 1.2 30 758 A 1 000 130 0.5 30 139 A 1 000 130 0.5 60 4910 B 1 000 130 0.5 30 4111 B 1 000 130 0.5 60 6112 B 1 000 000 140 1.0 7 days 77

Note: Catalyst A: 5%Pd-YSZ950R and B: 5%Pd-YSZ600R. Conditions: Iodobenzene, butylacrylate, N,N-diisopropylethylamine (DIPEA), and N,N-dimethylformamide (DMF) added to a round-bottomed flask followed by the addition of the catalyst and heating to the indicated temperature.

aConcentration of iodobenzene.bDetermined by GC-MS analysis of the crude reaction mixture against a decane internal standard.

Table 2. Heck reaction optimization.

IO

OO

O

Pd(YSZ)

NEtiPr2

DMA

temperature1 2 3

SilicabasedResin

SH

No ReactionI

OBu

O

Pd-YSZ, DIPEA,DMF, 140 ºC

[2]

Fig. 8. Powder XRD pattern of 5%Pd-YSZ950. Diffractogramshave been normalized to the peak at 2θ = 30° (the most intenseYSZ peak) for (a) before undergoing reaction and (b) after un-dergoing reaction.

tion, this would correspond to �0.10 ppm of Pd in the finalsolution (e.g., Table 2, entry 12). This value is consistentwith best effort for Pd on activated carbon (0.05 ppm) (74)and only slightly worse than the mercaptopropyl-modifiedsilica (0.003 ppm), which is a Pd scavenger (16).

Structure–reactivity analysisThe activity of different precatalysts can be analyzed in

terms of total number of turn-overs and how fast that TONcan be achieved, namely kinetics. In fact, the optimization ofreaction parameters will be most sensitive to the kinetics ofreaction. For these reasons, we chose the lower surface areacatalyst, which we expected would be slower, to optimizethe other reaction parameters. Table 2 (entries 1–9) shows asubset of the reaction optimization data for the Heck reac-tion using 5%Pd-YSZ950R.

Our previous work demonstrated that if reactions wereallowed sufficient time (24 h), the conversion for 100:1(substrate–catalyst) reactions was >95% for a wide range ofconditions. Therefore, the difference in reactivity for the twocatalysts was assessed by comparison of the percentage con-version vs. time for different catalysts and conditions. Forexample, entries 8–11 in Table 2 for a substrate–catalystratio of 1000:1 show that after 0.5 h of reaction 5%Pd-YSZ600R has a TON 3× higher than 5%Pd-YSZ950R. Thisresult is consistent with the heterogeneous catalysis litera-ture since more active catalyst is usually exposed in highersurface area materials. In this case, because of the homoge-neous nature of the reaction mechanism the higher surfacearea and smaller Pd particle size of 5%Pd-YSZ600R facili-tate the extraction of Pd to solution, thereby making the re-action faster.

However, the support seems to be playing a more impor-tant role than simply controlling the surface area of availablepalladium. It is interesting to note that the activity of the cat-alyst is also reflected in the total number of turn-overs.5%Pd-YSZ600R attains 770,000 turn-overs after 7 days, ap-proximately twice as much as 5%Pd-YSZ950R. This effectmay be accounted for by the mesoporous structure of 5%Pd-YSZ600R, which may be preventing the active homoge-neous Pd species from diffusing into the bulk of the solution(encapsulation) where it can be easily de-activated (27).

Despite the homogeneous nature of the catalysts due toleaching of Pd into solution, 5%Pd-YSZ600R compares wellwith other heterogeneous precatalysts discussed in the litera-ture. While TONs for homogeneously catalyzed Heck reac-tions have been documented in excess of 8.0 × 106 (75),TONs with heterogenous precatalysts are typically muchlower. In rare cases, TONs in slight excess of 106 have beenobtained demonstrating that the Pd-YSZ catalyst examinedin these studies compares favourably with other catalyst sys-tems. The highest value we have found in the literature isslightly higher (TON 1,150,000 (29, 76)) for a complexpolymer-based ligand-coordinated palladium catalyst. Incontrast, palladium supported on activated carbon, with sim-ilar cost per gram of catalyst as our precatalyst, has beenshown to achieve TONs of 36,000 (74, 77).

Conclusions

Pd-YSZ has been prepared using an amphiphillic

surfactant templating method resulting in a homogeneousdispersion of Pd on an ordered mesoporous structure. ThePd particle size and composition properties are easily con-trollable. The material is resistant to high temperatures,although some coarsening is observed and the long-rangeorder is maintained only up to intermediate temperatures.

Reduced Pd-YSZ showed high activity for the Heck reac-tion. Up to 770,000 TON were achieved with a 5% Pd-loaded catalyst having a surface area of 150 m2/g. Thisresult compares favourably with other heterogeneousprecatalysts in this reaction. Three-phase tests were con-ducted to ascertain the nature of the active catalyst species.These tests revealed that a homogeneous catalyst is leachingfrom the solid support and that the palladium that remainson the matrix is not responsible for the observed reactivity.Despite the homogeneous nature of the active species, theheterogeneous precatalyst remains interesting because it canbe easily filtered out leaving behind only residual amountsof Pd (<0.10 ppm).

Acknowledgments

The authors thankfully acknowledge funding from theNatural Sciences and Engineering Research Council of Can-ada (NSERC) and the contribution of the Center for Cataly-sis Research and Innovation at the University of Ottawa.

References

1. J. Tsuji. Palladium reagents and catalysts: innovations in or-ganic synthesis. John Wiley & Sons, New York, 1995.

2. G. Poli, G. Giambastiani, and A. Heumann. Tetrahedron, 56,5959 (2000).

3. J. Tsuji. Transistion metal reagents and catalysts: innovationsin organic synthesis. John Wiley & Sons, New York, 2000.

4. F. Diederich and P.J. Stang. Metal catalyzed cross-couplingreactions. Wiley-VCH, New York, 1998.

5. E.-I. Negishi and A. De Meijere. Handbook of organo-palladium chemistry for organic synthesis. Wiley-VCH, NewYork, 2002.

6. A.F. Littke and G.C. Fu. Angew. Chem. Int. Ed. 41, 4176(2002).

7. C.E. Garrett and K. Prasad. Adv. Synth. Catal. 346, 889 (2004).8. K. Konigsberger, G.-P. Chen, R.R. Wu, M.J. Girgis, K. Prasad,

O. Repic, and T.J. Blacklock. Org. Process Res. Dev. 7, 733(2003).

9. M.L. Toebes, J.A. van Dillen, and K.P. de Jong. J. Mol. Catal.A: Chem. 173, 75 (2001).

10. H.-U. Blaser, A. Indolese, A. Schnyder, H. Steiner, and M.Studer. J. Mol. Catal. A: Chem. 173, 3 (2001).

11. A.L. Dantas Ramos, P. da Silva Alves, D.A.G. Aranda, and M.Schmal. Appl. Catal. A: Gen. 277, 71 (2004).

12. C.P. Mehnert and J.Y. Ying. Chem. Commun. (Cambridge),2215 (1997).

13. C.P. Mehnert, D.W. Weaver, and J.Y. Ying. J. Am. Chem. Soc.120, 12289 (1998).

14. M.K. Richmond, S.L. Scott, and H. Alper. J. Am. Chem. Soc.123, 10521 (2005).

15. K. Okumura, K. Nota, K. Yoshida, and M. Niwa. J. Catal. 231,245 (2005).

16. C.M. Crudden, M. Sateesh, and R. Lewis. J. Am. Chem. Soc.127, 10045 (2005).

17. N.E. Leadbeater and M. Marco. Chem. Rev. 102, 3217 (2002).

© 2006 NRC Canada

Poulin et al. 1527

18. C.A. McNamara, M.J. Dixon, and M. Bradley. Chem. Rev.102, 3275 (2002).

19. B. Cetinkaya, N. Gurbuz, T. Seckin, and I. Ozdemir. J. Mol.Catal. A: Chem. 184, 31 (2002).

20. S.C. Schurer, S. Gessler, N. Buschmann, and S. Blechert.Angew. Chem. Int. Ed. 39, 3898 (2000).

21. J. Schwarz, V.P.W. Bohm, M.G. Gardiner, M. Grosche, W.A.Herrmann, W. Hieringer, and G. Raudaschl-Sieber. Chem. Eur.J. 6, 1773 (2000).

22. M. Mayr, M.R. Buchmeiser, and K. Wurst. Adv. Synth. Catal.344, 712 (2002).

23. W.A. Herrmann, C. Brossmer, K. Oefele, C.-P. Reisinger, T.Priermerier, M. Beller, and H. Fischer. Angew. Chem. Int. Ed.Engl. 34, 1844 (1995).

24. M. Beller, H. Fischer, W.A. Herrmann, K. Oefele, and C.Brossmer. Angew. Chem. Int. Ed. Engl. 34, 1848 (1995).

25. C. Baleizao, A. Corma, H. Garcia, and A. Leyva. Chem.Commun. (Cambridge), 606 (2003).

26. C. Baleizao, A. Corma, H. Garcia, and A. Leyva. J. Org.Chem. 69, 439 (2004).

27. J. de Vries. J. Chem. Soc. Dalton Trans. 421 (2006).28. B.M. Choudary, S. Madhi, N. Chowdari, M.L. Kantam, and B.

Sreedhar. J. Am. Chem. Soc. 124, 14127 (2002).29. Y.M.A. Yamada, K. Takeda, H. Takahashi, and S. Ikegami.

Tetrahedron, 60, 4097 (2004).30. A. Cwik, Z. Hell, and F. Figueras. Adv. Synth. Catal. 348, 523

(2006).31. A. Corma, D. Das, H. Garcia, and A. Leyva. J. Catal. 229, 322

(2005).32. C. Rocaboy and J.A. Gladysz. Org. Lett. 4, 1993 (2002).33. D.E. Bergbreiter, P.L. Osburn, and J.D. Frels. Adv. Synth.

Catal. 347, 172 (2005).34. K. Yu, W. Sommer, M. Weck, and C.W. Jones. Adv. Synth.

Catal. 347, 161 (2005).35. C.C. Cassol, A.P. Umpierre, G. Machado, S.I. Wolke, and J.

Dupont. J. Am. Chem. Soc. 127, 3298 (2005).36. M. Nowotny, U. Hanefeld, H. van Koningsveld, and T.

Maschmeyer. Chem. Commun. (Cambridge), 1877 (2000).37. R.B. Bedford, C.S.J. Cazin, M.B. Hursthouse, M.E. Light, K.J.

Pike, and S. Wimperis. J. Organomet. Chem. 633, 173 (2001).38. Y. Ji, S. Jain, and R.J. Davis. J. Phys. Chem. B, 109, 17232

(2005).39. M. Che and C.O. Bennett. Adv. Catal. 36, 55 (1989).40. J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T.

Kresge, K.D. Schmitt, C.T.-W. Chu, D.H. Olson, E.W.Sheppard, S.B. McCullen, J.B. Higgins, and J.L. Schlenker. J.Am. Chem. Soc. 114, 10834 (1992).

41. J.S. Beck, J.C. Vartuli, G.J. Kennedy, C.T. Kresge, W.J. Roth,and S.E. Schramm. Chem. Mater. 6, 1816 (1994).

42. C.T. Kresge, M.E. Leonowitcz, W.J. Roth, J.C. Vartuli, andJ.S. Beck. Nature (London), 359, 710 (1992).

43. B.P. Pelrine, K.D. Schmitt, and J.C. Vartuli. Mobil Oil Corp.,USA. US 5105051 A 19920414. 1992. p. 17.

44. R.M. Rioux, G. Song, J.D. Hoefelmeyer, P. Yang, and G.A.Somorjai. J. Phys. Chem. B, 109, 2193 (2005).

45. N. Marin-Astorga, G. Pecchi, T.J. Pinnavaia, G. Alvez-Manoli,and P. Reyes. J. Mol. Catal. A: Chem. 247, 145 (2006).

46. M.H. Bartl, B.J. Scott, G. Wirnsberger, A. Popitsch, and G.D.Stucky. Chem. Phys. Chem. 4, 392 (2003).

47. D.M. Antonelli and J.Y. Ying. Chem. Mater. 8, 874 (1996).48. D. Grosso, G.J.d.A.A. Soler Illia, E.L. Crepaldi, B. Charleux,

and C. Sanchez. Adv. Func. Mater. 13, 37 (2003).49. B. Lee, D. Lu, J.N. Kondo, and K. Domen. J. Am. Chem. Soc.

124, 11256 (2002).50. P. Yang, D. Zhao, D.I. Margolese, B.F. Chmelka, and G.D.

Stucky. Chem. Mater. 11, 2813 (1999).51. M. Yada, H. Kitamura, M. Machida, and T. Kijima. Inorg.

Chem. 37, 6470 (1998).52. M.S. Wong and J.Y. Ying. Chem. Mater. 10, 2067 (1998).53. A. Kim, P. Bruinsma, Y. Chen, L.-Q. Wang, and J. Liu. Chem.

Comm. (Cambridge), 161 (1997).54. M. Mamak, N. Coombs, and G. Ozin. Adv. Mater. 12, 198

(2000).55. M. Mamak, N. Coombs, and G. Ozin. J. AM. Chem. Soc., 122,

8932 (2000).56. M. Mamak, N. Coombs, and G.A. Ozin. Chem. Mater. 13,

3564 (2001).57. V. Calo, A. Nacci, A. Monopoli, A. Detomaso, and P. Iliade.

Organometallics, 22, 4193 (2003).58. T.H. Galow, U. Drechsler, J.A. Hanson, and V.M. Rotello.

Chem. Commun. (Cambridge), 1076 (2002).59. G. Williamson and W. Hall. Acta Metall. 1, 22 (1953).60. C.M. Grgicak, R.G. Green, W.-F. Du, and J.B. Giorgi. J. Am.

Ceram. Soc. 88, 3081 (2005).61. R.J. Farrauto, J.K. Lampert, M.C. Hobson, and E.M. Water-

man. Appl. Catal. B, 6, 263 (1995).62. K.-I. Fujimoto, F.H. Ribeiro, M. Avalos-Borja, and E. Iglesia.

J. Catal. 179, 431 (1998).63. R.J. Farrauto, M.C. Hobson, T. Kennelly, and E.M. Waterman.

Appl. Catal. A, 81, 227 (1992).64. M. Lyubovsky and L. Pfefferle. Catal. Today, 47, 29 (1999).65. O. Leonte, M. Birjega, D. Macovei, P. Pausescu, N. Popescu-

Pogrion, and M. Georgescu. 28, 347 (1986).66. O. Vasylkiv, Y. Sakka, Y. Maeda, and V.V. Skorokhod. J. Eur.

Ceram. Soc. 24, 469 (2004).67. I.W. Davies, L. Matty, D.L. Hughes, and P.J. Reider. J. Am.

Chem. Soc. 123, 10139 (2001).68. J. Rebek and F. Gavina. J. Am. Chem. Soc. 96, 7112 (1974).69. J. Rebek, D. Brown, and S. Zimmerman. J. Am. Chem. Soc.

97, 454 (1975).70. J.J. Rebek. Tetrahedron, 35, 723 (1979).71. A. Bianco, J. Furrer, D. Limal, G. Guichard, K. Elbayed, J.

Raya, M. Piotto, and J.-P. Briand. J. Comb. Chem. 2, 681(2000).

72. Y. Wang and S.R. Wilson. Tetrahedron Lett. 38, 4021 (1997).73. A.S. Kiselyov and R.W. Armstrong. Tetrahedron Lett. 38,

6163 (1997).74. R.G. Heindenreich, J.G. E. Krauter, J. Pietsch, and K. Kohler.

J. Mol. Catal. A: Chem. 182–183, 499 (2002).75. F. Miyazaki, K. Yamaguchi, and M. Shibasaki. Tetrahedron

Lett. 40, 7379 (1999).76. Y.M.A. Yamada, K. Takeda, H. Takahashi, and S. Ikegami.

Tetrahedron Lett. 44, 2379 (2003).77. R.G. Heindenreich, K. Kohler, J.G.E. Krauter, and J. Pietsch.

Synlett, 7, 1118 (2002).

© 2006 NRC Canada

1528 Can. J. Chem. Vol. 84, 2006