Polymer International Polym Int 55:854–861 (2006)

Silica-supported bis(imino)pyridyl iron(II)catalyst: nature of the support–catalystinteractionsSaptarshi Ray and Swaminathan Sivaram∗Polymer Science and Engineering Division, National Chemical Laboratory, Pune-411008, India

Abstract: Ethylene polymerizations were performed using silica-supported 2,6-bis[1-(2,6-diisopropylphenylimino)ethyl] pyridine iron(II) dichloride with methylaluminoxane (MAO) as co-catalyst. Silica was calcined at 600, 400and 200 ◦C under vacuum for 8 h. The effect of calcination temperature of silica on the polymerization activity andthe properties of the polymers obtained were examined. Catalyst–support interactions were examined by both achemical method and XPS. It was observed that upon supporting the catalyst on the surface of silica, there is anincrease in the binding energy of the metal center. However, no change in the metal binding energy was observedon supporting the catalyst to silica calcined at different temperatures. Ethylene polymerizations were performedusing MAO as co-catalyst. Catalysts were also prepared by first pretreating silica with MAO, followed by additionof the Fe(II) catalyst and contacting a complex of Fe(II) catalyst–MAO with silica previously calcined at 400 ◦C for8 h. The results indicate that there is no chemical bonding between the support and the catalyst. 2006 Society of Chemical Industry

Keywords: polyethylene; bis(imino) pyridyl iron(II) complex; silica, XPS; supported catalyst

INTRODUCTIONLate transition metal catalysts are now well establishedas a class of high-activity catalysts for olefinpolymerization, capable of providing a whole rangeof polyolefin structures.1–4 These catalysts can beactivated both by alkylaluminum compounds as wellas other non-coordinating Lewis acids.

There has been substantial interest in exploringsupported late transition metal catalysts for olefinpolymerization as drop-in substitutes for supportedmetallocenes in industrially relevant polyolefin pro-cesses such as gas and slurry phase processes.

Although several reports on supported late transi-tion metal catalysts are found in patent literature, veryfew studies on such catalysts have been published.Vaughan and co-workers supported nickel 1,2-diiminecomplexes on silica, pretreated with methylaluminox-ane (MAO).5 The authors showed that the branchingfrequency of the polymer decreases when supportedcatalysts are used. Hong et al.6 supported [1,2-bis(2,6-diisopropylphenylimino)acenaphthene]NiBr2 and (2-py-CH=N-2,6-di-iPrPh)NiBr2 on silica, pretreatedwith MAO or triisobutylaluminum. They observedthat with an increase in polymerization temperaturein the range 0–50 ◦C the branching of the polymersincreased, whereas their molecular weight decreased.

Bis-(imino)pyridine iron(II) was supported on silicacalcined at 800 ◦C and at 450 ◦C, and was pretreatedwith MAO.7 The supported catalysts thus obtainedcould be activated with triisobutylaluminum andshowed high polymerization activity for ethylene at

70 ◦C as well as steady-state kinetics. The authorsproposed that the fixation of the iron complex on thesilica surface occurs by multiple bonding of LFeCl2with the silica surface via interactions of the pyridyland phenyl groups of the ligand L with the OH groupof silica.

Recently Alt and co-workers have reported a seriesof bis[iminopyridyl] iron(II) dichloride catalysts sup-ported on silica containing a small amount of water.8

The support was pretreated with trimethylaluminum(TMA), which partially hydrolyzed on the surface ofsilica. A number of parameters, such as the effect ofAl/Fe ratio, water content on the support, Al/silicamolar weight ratio, influence of hydrogen concentra-tion, reactor residence time and steric and electroniceffects of the catalyst, were examined. It was observedthat the highest polymerization activity was obtainedwhen the water to aluminum ratio was 1.0. This ishigher than that observed in the case of metallocenes,which normally show the highest activity when thewater to aluminum ratio is 0.69. The authors con-cluded that, unlike metallocenes, where methylationis believed to occur by the reaction with TMA presentin MAO, supported iron(II) catalysts do not requirefree TMA for methylation, which can be accomplishedby MAO.

The patent literature reports several approaches tothe preparation of supported late transition metalcatalysts. These are (a) supporting the catalysts onsilica previously pretreated with MAO, in a slurryof toluene,9,10 (b) reaction of catalysts with MAO

∗ Correspondence to: Swaminathan Sivaram, Polymer Science and Engineering Division, National Chemical Laboratory, Pune-411008, IndiaE-mail: s.sivaram@ncl.res.in(Received 4 August 2005; revised version received 2 November 2005; accepted 19 December 2005)Published online 11 May 2006; DOI: 10.1002/pi.2020

2006 Society of Chemical Industry. Polym Int 0959–8103/2006/$30.00

Silica-supported bis(imino)pyridyl iron(II) catalyst

in toluene followed by supporting the resultingcomplex onto silica11–14 and (c) chemically tetheringthe ligands of a late transition metal catalyst onto thesurface of silica.15,16

In this paper we explore the nature of cat-alyst–support interactions involving 2,6-bis[1-(2,6-diisopropylphenylimino)ethyl] pyridine iron(II)dichloride as the catalyst and silica as the support,using X-ray photoelectron spectroscopy (XPS). Sev-eral reaction parameters have been examined, suchas the temperature of calcination of silica, the effectof hydroxyl concentration, and the relative ratio ofpaired and isolated hydroxyls on the activity of thecatalysts.

EXPERIMENTALMaterialsAll operations were performed under argon atmo-sphere using standard Schlenk techniques. Toluenewas dried using standard procedures.. The cata-lysts were stored and transferred in a Labmaster100 m Braun inert atmosphere glove box. SiO2 (grade952), with a surface area of approx. 200 m2 g−1, wasobtained from Davison Chemicals (USA); polymeriza-tion grade ethylene was obtained from the Gas CrackerComplex of Indian Petrochemicals Ltd, Nagothane,India. The purity of ethylene was periodically checkedusing an MBraun moisture and oxygen analyzer andwas found to be better than 2 ppm impurities bothfor moisture and oxygen. MAO was procured fromWitco GmbH (Germany). The aluminum contentof MAO was ∼4.5 mol L−1 and the Al/Me ratiowas ∼1.5. Ligand precursor, 2,6-diacetylpyridine and2,6-diisopropylaniline were procured from Aldrich,USA and used without purification. Ferrous chlo-ride was procured from Sigma USA. 2,6-Bis [1-(2,6-diisopropylphenylimino)ethyl] pyridine iron(II)dichloride (catalyst 1) was prepared according to areported method.17 Anal. (C33H43N3FeCl2) in %: C,65.14; H, 7.12; N, 6.91. Found: C, 64.95; H, 7.25;N, 6.68.

Preparation of silica supported2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron (II) dichloride catalystsSilica-supported 2,6-bis[1-(2,6-diisopropylphenyli-mino)ethyl] pyridine iron(II) dichloride (catalyst 2)Calcinated silica (1 g) was placed in a round-bottomedflask containing a magnetic stirring bar, needle andfitted with a septum adapter. Toluene (10 mL) wasadded to the flask under nitrogen atmosphere. In aseparate round-bottomed flask, 105 mg of catalyst 1(0.17 mmol) was used and 20 mL of toluene addedto it. The catalyst 1 solution was slowly added tothe silica, and the mixture was stirred at 50 ◦Cfor 2 h. The mixture was then cooled, washed andfiltered three times with 10 mL of toluene. Thedeep-blue colored solid obtained was dried undervacuum.

2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl] pyridineiron(II) dichloride on silica, pretreated with MAO(catalyst 3)Calcined silica (1.3 g) was placed in a round-bottomedflask fitted with a septum adapter and containing amagnetic stirring bar. To this was added 15 mL ofdry toluene followed by 0.7 mL of MAO, containing4.2 mmol of aluminum, while stirring. The mixturewas stirred for 2 h at 50 ◦C. The mixture was thencooled, filtered and washed three times with 10 mL oftoluene. The solid was dried under vacuum.

Part (0.8 g) of the above solid was placed in around bottomed flask and 10 mL of toluene wasadded; 80 mg of catalyst 1 (0.13 mmol) was placedin a separate round bottomed flask and dissolved in20 mL of toluene. The catalyst 1 solution was slowlyadded to the silica and the mixture was stirred at 50 ◦Cfor 2 h. The mixture was cooled, washed thrice with10 mL of toluene and filtered. The pale blue coloredsolid obtained was dried under vacuum.

2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl] pyridineiron(II) dichloride, pretreated with MAO and supportedon silica (catalyst 4)Catalyst 1 (100 mg, 0.16 mmol) was placed in a round-bottomed flask fitted with a septum adaptor. Toluene(10 mL) was added to the flask, followed by theslow addition of 0.3 mL of MAO solution, containing1.8 mmol of aluminum. The mixture was stirred for1 h. In a separate flask 1 g of silica was placed andslurried with 10 mL of toluene. A mixture of catalyst1/MAO was added slowly under stirring to the slurryof silica. The mixture was stirred for 3 h at 50 ◦C. Themixture was then cooled, washed three times with10 mL of toluene and filtered. The pale-blue coloredsolid obtained was dried under vacuum.

CharacterizationThe aluminum content of the MAO-pretreated silicawas determined by titration with EDTA (Ethylenediamine tetra-acetic acid) solution

Determination of the metal content inheterogeneous catalystsIron in the supported catalyst was estimated usinga Perkin Elmer-P100 inductively coupled plasmaionization (ICP) technique. The solid supportedcatalyst (100 mg) was boiled in 50 mL H2SO4 2 molL−1 and 0.5 mL of concentrated HNO3 for 30 minuntil dryness. The content was then cooled, followedby addition of water and filtration into a 100 mLstandard flask and the volume of the content made upto the mark using deionized water. The solution wasthen diluted 10 times and subjected to ICP analysis. Ablank solution was also made using the same method.

Estimation of the free hydroxyl content of silicaThe free hydroxyl content of silica, calcined atdifferent temperatures, was estimated by treatingsilica with methyl magnesium iodide and subsequently

Polym Int 55:854–861 (2006) 855DOI: 10.1002/pi

S Ray, S Sivaram

determining the methane gas liberated by the reactionof the hydroxyl groups with methyl magnesium iodide,using a gas burette.

Approximately 100 mg of calcined silica was placedin a 50 mL round-bottomed flask fitted with a septumadapter and a stopcock The flask was connected toa gas burette with a pressure equalizer and an oilreservoir. The gas burette was then flushed withnitrogen at least five times. Then, 5 mL of dry n-butyl ether was added to the silica with a hypodermicsyringe. The flask was cooled to 0 ◦C and the pressureinside the gas burette was adjusted to 1.013 ×105 Pa.Thereafter, the burette was disconnected from theflask. Methyl magnesium iodide (2 mL of 2 mol L−1

solution, in n-butyl ether) was quickly added to theflask using a hypodermic syringe. The mixture wasthen stirred for 5 min. Thereafter, the burette wasconnected to the flask by opening the stopcock, andthe methane generated was allowed to flow into theburette. Once the flow of methane ceased, its volumewas recorded. A blank run was performed withoutsilica. The number of moles of methane evolved pergram of silica was then calculated from the volume ofmethane gas generated.

Extraction of metal from supported catalystsThe supported catalyst (100 mg) was placed in around-bottomed flask and stirred in 10 mL of tolueneat 50 ◦C for 4 h. The solvent was filtered. Thisexperiment was repeated in the presence of MAO. Thetoluene filtrate obtained was evaporated to dryness; theresidue was dissolved in 2 mol L−1 sulfuric acid andused for the measurement of iron content by ICP.

X-ray photoelectron spectroscopyX-ray photoelectron spectra (XPS) were recordedon a VG Scientific (UK) ESCA 3000, using MgKα

(1253.6 eV) as the exciting source. The X-ray sourceoperated at 12 mA and 14.5 kV. The residual gaspressure in the spectrometer chamber during dataacquisition was less than 1.4 × 10−7 Pa. All XPS wererecorded with a pass energy of 50 eV and 4 mm slit.The spectrometer was calibrated by determining thebinding energy of Au4f7/2 (84.0 eV), Ag3d5/2 (368.3 eV)and Cu2p3/2 (932.4 eV) levels using spectroscopicallypure metals (M/s Johnson Matthey). These values arein good agreement with the literature values.18 Theaccuracy of the value was within +0.2 eV. The XPSwere referred to the C1s (285.0 eV).

Pellets (10 mm × 1 mm) were made from solidcatalysts and then introduced into the fast-entryair-lock chamber while retaining the base pressureof the analysis chamber at a base pressure of1.013 × 10−8 Pa. From the air-lock chamber thesamples were transferred to the analysis chamber anddegassed for 15–16 h before analysis. The treatmentof the spectra was performed using standard methodssuch as background subtraction, smoothing andnonlinear (Gaussian) least-square curve fitting. Thepeak position and the intensity were roughly assigned

on the spectra and were used as the crude value forthe curve-fitting program.

Polymerization and characterizationPolymerization of ethylene was performed at 1.013 ×105 Pa pressure in a water-jacketed glass reactor.Ethylene was supplied to the reactor from awater-jacketed gas burette. A circulating thermostatmaintained the temperature of the reactor within±0.2 ◦C. The solid catalysts were introduced intothe reactor in a vial. The desired quantity of solventwas added and saturated with ethylene, by lettingthe ethylene flow from the gas burette into thereactor under vigorous stirring. The polymerizationwas initiated by adding MAO with a hypodermicsyringe. The polymerization was performed for20–30 min. The polymerization was terminated withsome acidified methanol. The polymers obtained wereseparated by filtration and dried under vacuum.

The molecular weights and molecular weightdistribution of the polymers were determined usinga gel permeation chromatograph (GPC) PL-GPC 220at 160 ◦C using trichlorobenzene as the solvent andlow polydispersity polystyrene samples as standards.

RESULTS AND DISCUSSIONFive catalyst samples were examined in this study.Catalysts 2a, 2b and 2c were prepared by supportingcatalyst 1 on silica calcined at 600 ◦C for 8 h, 400 ◦Cfor 8 h and 200 ◦C for 8 h, respectively. Catalyst 3 wasprepared by supporting catalyst 1 on silica previouslycalcined at 400 ◦C for 8 h, and pretreated with MAO.Catalyst 4 was prepared by treating catalyst 1 withMAO, followed by supporting the complex on silicapreviously calcined at 400 ◦C for 8 h.

Nature of metal–support interactionsTable 1 shows the metal content in the five catalysts.No significant change in metal content is observedwhen silica was calcined at different temperatures.The nature of the interactions between the catalystsand silica was examined by two methods, namely, bymeasuring the free hydroxyls present on the surfaceof the silica before and after supporting the catalysts

Table 1. Depletion of free hydroxyls and % Fe in various

silica-supported 2,6- diacetylpyridine-bis-(2,6-diisopropylphenylanil)

iron(II) dichloride catalysts

Free hydroxyls of silicaa

SilicaCatalystNo.

calcination temp.(◦C)/time (h)

Beforesupporting

Aftersupporting % Fe

2a 600/8 h 1.81 1.79 0.562b 400/8 h 2.29 2.29 0.612c 200/8 h 2.47 2.50 0.633 400/8 h 2.21 0.24 0.754 400/8 h 2.10 1.91 0.20

a mmol of OH (g of silica)−1.

856 Polym Int 55:854–861 (2006)DOI: 10.1002/pi

Silica-supported bis(imino)pyridyl iron(II) catalyst

695 700 705 710 715 720 725 730 735

0

1000

2000

3000

C

B

A

Inte

nsi

ty

Binding energy (eV)

Figure 1. Binding energy of Fe (2p3/2) center of catalysts 2a (A), 3(B) and 4 (C): effect of pretreatment of the silica (catalyst 3) andpretreatment of catalyst (catalyst 4).

Table 2. XPS binding energy of silica supported Fe(II) catalyst

Catalystno.

Fe (2p)3/2 bindingenergy (eV)

1 710.82a 711.82b 711.72c 711.83 712.84 712.8

and by the use of X-ray photoelectron spectroscopy.Table 1 shows the free hydroxyl content of catalysts2a–c. It is seen that there is no depletion of freehydroxyls due to the reaction of catalyst 1 with silica.In catalyst 3 there was a depletion of free hydroxyls,presumably on account of the reaction of MAO withthe hydroxyl groups of silica. The free hydroxyl contentof catalyst 4, prior to supporting, is equal to the freehydroxyl found on silica, calcined at 400 ◦C for 8 h.Upon supporting the catalyst, only a small reductionin the free hydroxyl group is observed.

The XPS of the supported catalysts were recordedusing the peak due to Si2p at 103.3 eV (Fig. 1) asinternal reference for purposes of charge correction.XPS of silicon did not provide any desired informationsince the change in the silicon environment was ratherinsignificant. The binding energies of Fe centers aregiven in Table 2. It can be seen that the binding energyof Fe in the homogeneous catalyst is lower than thatof the supported catalyst (Fig. 2). The increase in thebinding energy from the homogeneous catalyst to theheterogeneous catalysts is indicative of the presenceof some weak secondary interactions, between Si–OHand Fe–Cl. However there is no change in the bindingenergy of Fe as a function of temperature of calcinationof silica (catalysts 2a–c, Fig. 3).

It is therefore inferred that the electronic environ-ment of the metal centers is not affected by either thequantity or the nature of the hydroxyl groups present

695 700 705 710 715 720 725 730 735 740

0

200

400

600

800

1000

1

2a

Inte

nsi

ty

Binding energy (eV)

Figure 2. Binding energy of Fe (2p3/2) center in catalysts 1 and 2a.

695 700 705 710 715 720 725 730 735

0

1000

2000

3000

2c

2b

2a

Inte

nsi

ty

Binding energy

Figure 3. Binding energy of Fe (2p3/2) center of catalysts 2a–c: effectof silica calcination temperature.

on silica. This implies that the catalyst has no strongchemical interaction with the surface of silica. Thisobservation is contrary to that of Semikolenova et al.,7

who reported that the support of an iron complexon silica occurs by multiple bonding of the catalystwith silica surface via interaction of pyridyl and phenylgroups of the ligand with the OH group of silica.However, this conclusion was not substantiated byany experimental evidence. XPS, however, providesevidence of any significant bonding between the cata-lyst and the surface hydroxyls. The binding energy ofthe Fe centers of catalysts 3 and 4 is higher than forcatalysts 2a–c (Fig. 1) as well as for the homogeneouscatalyst. This is presumably due to a cationic chargeon iron, which renders electron loss from the metalcenter more difficult. Ma et al.19 reported a similarincrease in the binding energy of the Fe (2p3/2) centerfrom 710.5 eV to 711.8 eV when the iron catalyst wassupported on silica pretreated with MAO.

Polymerization of ethylene using supported2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron(II) dichloride catalystsPolymerization of ethylene was performed with allfive catalysts using MAO as the co-catalyst. The

Polym Int 55:854–861 (2006) 857DOI: 10.1002/pi

S Ray, S Sivaram

Table 3. Ethylene polymerization using catalysts 1, 2a, 2b and 2ca

Activity (kg PE[mol (Fe) h]−1

Polym. Al/FeEntryno.

temp.(◦C)

molarratio

After10 min

After30 min Mp1

b Mp2c

Rpd ×

103

Catalyst 11 30 1000 4670 1510 600 91 700 2.60

Catalyst 2a2 30 1000 4318 1628 2480 98 500 2.093 40 1000 4126 1400 2100 78 500 1.794 50 1000 3355 1380 1850 71 300 1.785 50 500 660 – 1400 22 800 1.466 50 1500 3811 1389 1700 41 200 1.90

Catalyst 2b7 30 1000 4065 1529 2300 92 500 1.868 40 1000 3156 1087 2000 80 000 1.759 50 1000 3108 1150 1600 42 600 1.72

10 50 500 2221 – 1100 23 800 1.6411 50 1500 3463 1241 2100 36 400 1.70

Catalyst 2c12 30 1000 3042 1468 540 91 600 1.7113 40 1000 2982 1431 1150 65 000 1.2814 50 1000 2335 887 3750 46 500 1.2115 50 500 2315 832 1150 22 700 1.0216 50 1500 2395 983 650 32 510 1.23

PE, polyethylene.a Pressure: 1.013 × 105 Pa; time: 30 min; [Fe] = 1–1.5 × 10−6 mol L−1;toluene = 50 mL.b Peak molecular weight of lower molecular weight fraction (g mol−1).c Peak molecular weight of higher molecular weight fraction (g mol−1).d Rate of polymerization (mol s−1).

polymerizations were performed at 1.013 × 105 Paethylene pressure for 30 min. The results are givenin Tables 3 and 4. The polymerization activity isreported after 10 min and 30 min of polymerization.Polymerization activity of the catalysts decreased withan increase in temperature in all cases. The activityof catalysts 3 and 4 was lower than that of catalysts2a–c. It was observed that the catalyst prepared usinga silica support with lower hydroxyl content exhibiteda higher catalytic activity.

Kinetics of polymerizationThe kinetics of polymerization of ethylene was studiedfor all the catalysts. The results are shown in Figs 4–6.The initial rates of polymerization, Rp, were calculatedfrom the linear extrapolation of the initial rate in theascending part of the kinetic curves. Figure 4 showsthe kinetics of polymerization of catalyst 1 and 2a.It is seen that catalyst 1 exhibited an initial surge inits activity, followed by a rapid decay. This is typicalfor homogeneous catalysts. The rate of decay in thecase of a supported catalyst was much slower. Thisdemonstrates the beneficial effect of the silica supporton polymerization kinetics. The initial Rp values for thetwo catalysts were similar, however. Figure 5 showsthe kinetic profiles of catalysts 2a–c. It is observedthat the initial RP increased in the order 2a > 2b > 2c.

Table 4. Ethylene polymerization using catalysts 3 and 4a

Activity, kg PE[mol (Fe) h]−1

Polym. Al/FeEntryno.

temp.(◦C)

molarratio

After10 min

After30 min Mp1

b Mp2c

Rpd ×

103

Catalyst 31 30 1000 1500 830 1400 140 000 0.662 40 1000 1607 857 1400 72 400 0.733 50 1000 911 386 1250 92 500 0.424 50 500 916 382 1450 111 300 0.515 50 1500 828 427 1320 64 300 0.50

Catalyst 46 30 1000 1856 980 3200 191 300 1.127 40 1000 916 498 3100 177 500 0.668 50 1000 1055 520 2800 172 000 0.659 50 500 644 389 2900 273 400 0.47

10 50 1500 997 499 2900 145 200 0.48

a Pressure: 1.013 × 105 Pa; time: 30 min; [Fe] = 2–2.6 × 10−6 mol L−1;toluene = 50 mL.b Peak molecular weight of lower molecular weight fraction (g mol−1).c Peak molecular weight of higher molecular weight fraction (g mol−1).d Rate of polymerization (mol s−1).

0 5 10 15 20 25 300.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

2a

1

mo

l s−1

×10

−3

Time (min)

Figure 4. Kinetics of polymerization with catalysts 1 (entry 1, Table 3)and 2a (entry 2, Table 3): effect of a heterogeneous catalyst.

Thus, the catalyst prepared from a support calcined athigher temperature showed the highest Rp.

The initial rates of polymerization of catalysts 2b, 3and 4 are shown in Fig. 6. In all these cases, the silicaused for catalyst preparation was calcined at 400 ◦Cfor 8 h. The initial rate for 2b was higher than forboth 3 and 4. Thus, pre-reacting the surface hydroxylgroups of silica with MAO had a deleterious effect oncatalyst activity.

Molecular weight and molecular weightdistributionsAll polymers exhibited two distinct peaks in GPC.This observation is similar to what has been observedfor homogeneous catalysts. Figure 7 shows the effectof heterogenization and calcination temperature of

858 Polym Int 55:854–861 (2006)DOI: 10.1002/pi

Silica-supported bis(imino)pyridyl iron(II) catalyst

0 5 10 15 20 25 300.0

0.5

1.0

1.5

2.0

2.5

3.0

2c

2b

2a

Time (min)

mo

l s−1

×10

−3

Figure 5. Kinetics of polymerization of catalysts 2a, 2b and 2c: effectof silica calcination temperature (entry 2, 7 and 12, Table 3).

0 5 10 15 20 250.0

0.5

1.0

1.5

2.0

4

3

2b

M.s

ec-1

x10-3

Time (min)

Figure 6. Kinetics of polymerization of catalysts 2b (entry 7, Table 3),3 and 4 (entry 1 and 6, Table 4): effect of pretreatment of silica (3) andcatalyst (4) with MAO.

silica on the nature of molecular weight distributions.The use of silica calcined at lower temperatures leadsto a marginal decrease in peak molecular weights.Apparently, residual hydroxyl groups have an adverseeffect on molecular weights. This observation is furthersubstantiated by the fact that when silica is pretreatedwith MAO there is a significant increase in peakmolecular weights (Fig. 8). Higher molecular weightpolymers are obtained when a catalyst complex ispre-formed and then supported on silica. Figure 9shows the effect of polymerization temperature on thenature of molecular weight distributions. It is observedthat both Mp1 and Mp2 shift towards lower molecularweight with increasing temperature.

Nature of active sites and mechanism ofpolymerizationA summary of salient results obtained with varioussilica-supported Fe(II) pyridylimine catalysts is given

4 62

0.0

0.1

0.2

0.3

0.4

0.5

log_10 (M_w)

wt.

fra

ctio

n D

C

B

A

Figure 7. GPC of polymers obtained from catalysts 2a (A entry 2,Table 3), 2b (B entry 7, Table 3), 2c (C entry 12, Table 3) and catalyst1 (D entry 1, Table 3).

0.6

0.5

0.4

0.3

0.2

0.0

2 3 4 5 6

0.1

C

B

A

wt.

fra

ctio

n

Log_10 (M_w)

Figure 8. GPC of polymers obtained from catalyst 2b, (A entry 1,Table 3), catalyst 3 (B entry 1, Table 4) and catalyst 4 (C entry 6,Table 4).

in Table 5. Any discussion on the nature of the activesites and the mechanism must take into considerationthe following experimental observations, namely:

(a) Supporting the catalyst on suitably activated silicaincreases the catalyst activity.

(b) The catalyst activity as well as initial Rp decreasewith increasing hydroxyl content on silica (2a >

2b > 2c). However, it is pertinent to recognizethat under the polymerizing conditions usedthese residual hydroxyl contents will react withthe organo-aluminum compounds used as co-catalysts.

(c) Complete removal of all hydroxyl groups bypretreating silica with MAO leads to a significantdrop in catalyst activity as well as initial Rp.

(d) Supporting the catalyst on suitably activated silicaleads to a small increase in molecular weightcompared to soluble catalysts.

Polym Int 55:854–861 (2006) 859DOI: 10.1002/pi

S Ray, S Sivaram

0.0

2 4 6

0.1

0.2

0.3

0.4

0.5 C

B

A

log_10 (M_w)

wt.

fra

ctio

n

Figure 9. GPC of polymers obtained from catalyst 2b, at 30 ◦C (Aentry 7, Table 3), 40 ◦C (B entry 8, Table 3) and 50 ◦C (C entry 9,Table 3).

(e) However, peak molecular weight decreases withdecreasing calcination temperature of the support.

(f) Pretreatment of silica with MAO followed bytreatment with the Fe(II) catalyst leads tosignificant increase in peak molecular weights.

(g) Pre-forming the Fe(II)–MAO complex followedby treatment with silica leads to even higher peakmolecular weights.

Silica support with low hydroxyl content may offerjust enough weak interactions between the catalystand isolated hydroxyl groups on the surface of thesupport as shown in Fig. 10.

Such weak Cl–H hydrogen bonds are ubiquitous inthe literature. These weak interactions can stabilize theactive site and thus alter the nature of polymerizationkinetics (Fig. 10). The small decrease in initial Rp,catalyst activity and peak molecular weight upon usingsilica bearing more hydroxyl groups can be attributedto the partial loss of MAO by the reaction withhydroxyl groups on silica and the consequent decreasein Al/Fe ratio. Complete elimination of hydroxylgroups on silica has an adverse effect on catalystactivity and initial Rp. Under these conditions thecatalyst is most probably held on to the surface ofsilica through weak physical adsorption. Under thesecircumstances the silica support does not exert anystabilizing influence on the active site.

The most significant observation is that the catalystprepared using silica pretreated with MAO produces

OH Cl

Fe

Cl

N

N

N

Ar

Ar

Si

Figure 10. Schematic representation of H–Cl hydrogen bondedsupported Fe(II) pyridylimine catalyst.

polymers with substantially higher peak molecularweights than silica which has not been pretreatedwith MAO. Taken together with the observationthat catalysts 3 and 4 show a reduced rate ofpolymerization, this is indicative of much lower activesite concentration under these conditions. There isconsiderable ambiguity in the literature on the natureof active sites in the case of the Fe(II) iminopyridylcatalyst system.20 Fe(III) species are implicated asactive sites, although their precise nature remains tobe determined. For some reason, either less activesites are formed when a MAO-pretreated silica is used(catalyst 3) or the active sites generated are partiallydestroyed when they are brought into contact withsilica calcined at 400 ◦C for 8 h (catalyst 4).

CONCLUSIONS(1) 2,6-Diacetylpyridine-bis(2,6-diisopropylphenyl-

anil) iron(II) dichloride when supported on sil-ica shows no evidence of any chemical bonding tothe support. The concentration in free hydroxylgroups on silica remains unchanged before andafter supporting the complex.

(2) The electronic environment of iron in the complexalso remains unchanged when silica calcined atdifferent temperatures is used.

(3) The catalytic activity for polymerization ofethylene is higher at higher silica calcinationtemperature. Residual hydroxyl groups of silicaadversely affect catalyst activity.

(4) All polymers exhibit bimodal molecular weightdistributions. This is similar to the behaviorof homogeneous catalysts. The free hydroxylconcentration of silica has an adverse effect onthe molecular weight.

(5) Calcination temperature and hence concentrationof free hydroxyl groups influence the kinetics

Table 5. Polymerization of ethylene using silica-supported Fe(II) pyridylimine catalyst: summary of the resultsa

Catalyst designation

Catalyst 1 Catalyst 2a Catalyst 2b Catalyst 2c Catalyst 3 Catalyst 4

Catalyst activity, kg PE [mol (Fe) h]−1 1510 1628 1529 1468 830 980Initial Rp × 104 mol s−1 26.0 20.9 18.6 17.1 6.6 10.2Mp2 91 700 98 500 92 500 91 600 140 000 191 300

a Conditions: temperature = 30 ◦C; [Al]/[Zr] = 1000; P = 1.013× 105 Pa.

860 Polym Int 55:854–861 (2006)DOI: 10.1002/pi

Silica-supported bis(imino)pyridyl iron(II) catalyst

of polymerization. The initial Rp decreaseswith increasing free hydroxyl content of silica,presumably due to the partial loss of MAO byreaction with the hydroxyl groups on silica andconsequent decrease in Al/Fe ratio.

(6) Activity of the catalyst with MAO-pretreated silicais lower, due to the lower degree of stabilizationof the active sites by the support.

(7) The peak molecular weights of the polymersobtained from the catalyst prepared using sil-ica pretreated with MAO are higher thanwith catalysts in which silica is not pre-treated with MAO. It is even higher whenthe active site is preformed and then treatedwith silica. This is indicative of a much loweractive site concentration. Part of the activesites is probably destroyed during the pro-cess of supporting the catalysts in these twocases.

ACKNOWLEDGEMENTSR acknowledges the Council of Scientific andIndustrial Research, India, for Junior and SeniorResearch Fellowships.

REFERENCES1 Mecking S, Angew Chem Int Ed 40:534 (2001).2 Gibson VC and Spitzmesser SK, Chem Rev 103:283 (2003).3 Ittel SD, Johnson LK and Brookhart M, Chem Rev 100:1169

(2000).

4 Patil AO and Hlatky GG, Beyond Metallocenes; Next-generationPolymerization Catalysts: Proceedings (1st edn). AmericanChemical Society, Columbus, OH (2003).

5 Vaughan GA, Canich JAM, Gindelberger DE and Squire KR,WO 9748736 (1997); Chem Abstr 128:89235 (1998).

6 Hong DS, Seo TS and Woo SI, Stud Surf Sci Catal 130:3867(2000).

7 Semikolenova NV, Zakharov VA, Talsi EP, Babushkin DE,Sobolev AP, Echevskaya LG, et al, J Mol Catal, Part A: Chem182–183:283 (2002).

8 Schmidt R, Welch MB, Palackal SJ and Alt HG, J Mol Catal,Part A: Chem 179:155 (2002).

9 Berardi A and Speakman JG, WO 0024788 (2000); Chem Abstr132:308833 (2000).

10 Gibson VC, Kimberley BS, Maddox PJ and Mastroianni S, WO0015646 (2000); Chem Abstr 132:237522 (2000).

11 Kimberley BS and Pratt D, WO 9946304 (1999); Chem Abstr131:214732 (1999).

12 Kimberley BS and Samson JNR, WO 9946303 (1999); ChemAbstr 131:214731 (1999).

13 Canich JAM, Vaughan GA, Matsunaga PT, Gindelberger DE,Shaffer TD and Squire KR, US 6,492,473 (2002); Chem Abstr128:89234 (2002).

14 Mackenzie PB, Moody LS, Killian CM, Ponasik JA,McDevitt JP and Lavoie GG, US 6303720 (2001).

15 Preishuber-Pflugl P and Brookhart M, Macromolecules 35:6074(2002).

16 Kaul FAR, Puchta GT, Schneider H, Bielert F, Mihalios D andHerrmann WA, Organometallics 21:74 (2002).

17 Britovsek GJ, Bruce M, Gibson VC, Kimberley BS,Maddox PJ, Mastroianni S, et al, J Am Chem Soc 121:8728(1998).

18 Richter K and Peplinsky B, J Electron Spectrosc Relat Phenom13:69 (1978).

19 Ma Z, Sun WH, Zhu N, Li Z, Shao C and Hu Y, Polym Int51:349 (2002).

20 Britvosek GJP, Clentsmith GKB, Gibson VC, Goodgame DML,Mctavish SJ and Pankhurst QA, Catal Commun 3:207 (2002).

Polym Int 55:854–861 (2006) 861DOI: 10.1002/pi