SURFACE AND INTERFACE ANALYSISSurf. Interface Anal. 27, 728–734 (1999)

XPS Investigation of the Electronic Environmentin Selected Heterogenized Zirconocene Catalysts

M. Atiqullah, 1* M. Faiz, 2 M. N. Akhtar, 1 M. A. Salim,2 S. Ahmed1 and J. H. Khan1

1 Metallocene Catalysts Research Laboratory, Center for Refining and Petrochemicals, The Research Institute, King FahdUniversity of Petroleum and Minerals, Dhahran 31261, Saudi Arabia2 Surface Science Laboratory, Department of Physics, King Fahd University of Petroleum & Minerals, Dhahran 31261, SaudiArabia

Ethylbisindenyl zirconium dichloride (Et(Ind) 2ZrCl 2) and the MAO methylalumoxane (MAO) co-catalystwere heterogenized on Davision silica 955 partially dehydroxylated at 275°C, following the concept of equi-librium adsorption. The influence of MAO on the electronic environment resulting from the heterogenizationwas investigated using XPS. Heterogenization of Et(Ind)2ZrCl 2 and MAO on the above silica generatedtwo types of zirconocenium cations (Cation 1 and Cation 2), independent of the heterogenization meth-ods. Based on the postulated surface chemistry, Cation 1 is presumed to be in the form of an ion-pair[SiO]−[Et(Ind) 2ZrCl] Y, whereas Cation 2 is presumed to be a trapped multi-coordinated crown complex ofMAO. In the absence of MAO, only Cation 1 is formed. The present study provides some support for thepostulated surface chemistry regarding heterogenization of Et(Ind)2ZrCl 2 and MAO on silica. Copyright 1999 John Wiley & Sons, Ltd.

KEYWORDS: XPS; catalysis; Zr

INTRODUCTION

Metallocene catalysts have started to shape the future ofolefin polymerization technology because of their versa-tile superiority over the current Ziegler–Natta catalysts.However, the present and future challenge in this fieldconcerns how to develop an industrially cost- and energy-effective olefin polymerization process. In this regard,heterogenization of the metallocene catalyst precursorsplays a key role to attain the desired particulate morphol-ogy for the resin and to overcome the problem of solventremoval in the polymerization process. A solvent-free pro-cess is economical but the resin morphology influences thepacking density, particulate flow behaviour and polymerprocessability.1

The literature reveals2,3 several outstanding problemsin heterogeneous metallocene catalysis and the associatedslurry olefin polymerization. Heterogenized metallocenesshow a drastic decrease in catalytic activity and the cat-alyst precursors leach off the support. The molecularweight distribution broadens and the average molecu-lar weight and melting point of the resulting polyolefinsincrease. In slurry polymerization, the monomer encoun-ters heat and mass transport resistance due to the hydro-dynamic boundary layer at the surface of the catalystparticle and intra-particle pore diffusion. The problem ofdesigning a suitable particulate morphology for the cata-lyst and conducting the polymerization at operating con-ditions that minimize these transport resistances has yet tobe solved. The industrial success of metallocene catalysts

* Correspondence to: M. Atiqullah, Metallocene Catalysts ResearchLaboratory, Center for Refining and Petrochemicals, The ResearchInstitute, KFUPM, Dhahran 31261, Saudi Arabia.E-mail: matiq@kfupm.edu.sa

depends on overcoming the above inter-related problems.The purpose of this paper is to address those problemsthat concerns the electronic environment resulting fromheterogenization.

Heterogenized metallocene catalysts must be character-ized adequately in order to understand why the catalyticactivity decreases, how the polymer chain propagates,transfers and terminates and how the stereoregulationoccurs, particularly in the presence of the solid support.High-resolution solid-state [13C]NMR spectroscopy withcross-polarization and magic angle spinning (CPMAS)and x-ray photoelectron spectroscopy (XPS), which com-plement each other, are appropriate techniques for char-acterizing the heterogenized catalysts.

The technique of [13C]NMR-CPMAS was applied tocharacterize CpŁ2Th.13CH3/2 (CpŁ D pentamethyl cyclo-pentadienyl) supported on -alumina and magnesiumchloride.4 – 6 From the difference in the chemical shiftvalues between the pure and supported compoundthorocenium cation was concluded to be generated in thesolid state. Such a finding also was observed with the solidreaction product of Cp2Zr.13CH3/2 and methylalumoxane(MAO), using solid-state NMR.7

Gassman and Callstrom8 characterized the interactionbetween selected C2v symmetric, achiral zirconocenes[Cp2ZrCl2, Cp2ZrCH3Cl and Cp2Zr.CH3/2] and MAO dis-solved in toluene using XPS (Table 1). They evaluatedthe resulting interaction by measuring the Zr 3d5/2 bind-ing energies using a specially designed probe that hasbeen described elsewhere.9 The reaction of MAO with theabove zirconocenes showed a binding energy of 182.4 eV,which exceeds the value of the corresponding pure zir-conocenes, indicating that the catalytic species in the mix-ture was electron-deficient compared to the correspondingpure components.

CCC 0142–2421/99/080728–07 $17.50 Received 2 June 1998Copyright 1999 John Wiley & Sons, Ltd. Revised 25 January 1999; Accepted 4 February 1999

XPS OF Zr CATALYSTS 729

Table 1. Effect of substitution of chloride s-ligands inCp2ZrCl 2 by methyl groups on the bindingenergy in a solution environment8

Catalyst sample Peak of Binding energydescription interest (š0.1 eV)

Cp2ZrCl2 Zr 3d5/2 181.7Cp2ZrClCH3 Zr 3d5/2 181.2Cp2Zr.CH3/2 Zr 3d5/2 180.7MAO-pretreatedCp2ZrCl2, Cp2ZrClCH3 and Zr 3d5/2 182.4Cp2Zr.CH3/2

When the above solution was saturated with ethylene,the binding energy decreased slightly because of theformation of polyethylene. A new catalytic species havinga binding energy of 182.1 eV was formed. This means thatthe starting catalytic species changed with the progressionof polymerization. Substitution of the Cl�-ligand withthe CH3 group did not affect the binding energy of themixture under either situation, with or without ethylene.This confirms that all the zirconocenes remained in thesame electronic environment.8

However, the literature does not show any work con-cerning the characterization of heterogenized metallocenesusing XPS. Consequently, the electronic environmentresulting from the mass transfer process occurring dur-ing heterogenization and the interaction between the met-allocenes, MAO and the support remains unexplored.This motivated us to undertake the current study, whichcharacterizes the C2 symmetric, chiral Et(Ind)2ZrCl2, het-erogenized in the presence of MAO on silica, usingXPS. Unlike the work of Gassman and Callstrom,8

this work considers the indenyl (Ind) ligand, which isa benzoannelated version of the Cp ligand. We havechosen Et(Ind)2ZrCl2 because it homopolymerizes ethy-lene and propylene as well as copolymerizes them with˛-olefins.2

EXPERIMENTAL

Catalysts and materials

Silica 955 (Davison Grace), having a surface area of300 m2 g�1, an average pore volume of 1.65 cm3 g�1

and a pore size of 220̊A, was used. Both Et(Ind)2ZrCl2and MAO (10% w/w in toluene), with an average molec-ular weight of 800 and degree of oligomerization of 14,were procured from Witco, Germany. Toluene was pur-chased from Aldrich, UK and purified by refluxing in asodium benzophenone mixture for¾2 days until a dark-blue colour was obtained. All manipulations were carriedout in an inert atmosphere of argon using the Schlenktechnique.10

Support dehydroxylation

The silica support was partially dehydroxylatedin vacuoby heating at 275°C for 24 h in a B̈uchi furnaceand then was cooled to room temperature and storedunder argon. Note that metallocene supported on partially

dehydroxylated silica has been reported to show increasedolefin polymerization activity.11 – 13

Catalyst preparation

The Et(Ind)2ZrCl2 was heterogenized on the above par-tially dehydroxylated silica by equilibrium adsorption.The following three heterogenization methods were used:

(1) Method 1: adsorption of the zirconocene on silicafirst and then treatment of the supported zirconocenewith MAO.

(2) Method 2: adsorption of MAO on silica first and thenaddition of the zirconocene in the final step.

(3) Method 3: addition of MAO-pretreated zirconoceneto silica.

Note that equilibrium adsorption avoids undesirable side-reactions causing precipitation of the catalyst precursorsoff the support.

In Method 1, the dehydroxylated silica was first slur-ried in the toluene. Then Et(Ind)2ZrCl2 was dissolvedin toluene at 60°C and added to the silica–tolueneslurry. The excess liquid was decanted using a pipette.The solid part was washed with toluene to obtain theSiO2/Et(Ind)2ZrCl2 precursor. Then MAO was added tothis precursor slurry in toluene, which was stirred at 60°Cfor 4 h (to promote the diffusive mass transport process).The excess liquid was decanted using a pipette and thesolid part was washed three times with toluene. It wasfinally dried in vacuo at 50°C to a free-flowing powderand stored under argon.

For Method 2, the above procedure was repeated,reversing the order of addition of Et(Ind)2ZrCl2 and MAO.However, the concentration gradient and the mixing con-dition with respect to the diffusing components were keptabout the same.

In Method 3, Et(Ind)2ZrCl2 was first preactivated withMAO using a molar ratio of 1 : 23 to generate the zir-conocenium cation before heterogenization. This valueclosely approximates to the limiting catalyst/co-catalystratio used in the slurry olefin polymerization.

Details of the three heterogenization methods appear inTable 2.

X-ray photoelectron spectroscopy experiments

The XPS spectra of the experimental catalysts wereobtained using a VG Escalab MkII photoelectron spec-trometer. Details of this spectrometer are described else-where.14 The spectrometer energy was calibrated by fixingthe Cu 2p3/2 and Au 4f7/2 peaks at binding energies of932.67 and 83.98 eV, respectively.15

The powdered catalyst samples were first embeddedin an indium foil placed in a metal sample-holder. Thenthey were introduced into the fast entry air-lock chamberwithout affecting the vacuum in the analysis chamber,which maintained a base pressure of 10�10 mbar.

From the air-lock chamber, the samples weretransferred to the analysis chamber. Unmonochromated,1486.6 eV Al K̨ photons were used at 130 W as theincident beam. The kinetic energy of the photoelectronswas analysed using a multi-channeltron hemisphericalenergy analyser with a pass energy set at 20 eV in fixedanalyser transmission (FAT) mode.

Copyright 1999 John Wiley & Sons, Ltd. Surf. Interface Anal. 27, 728–734 (1999)

730 M. ATIQULLAH ET AL.

Table 2. Heterogenization experimental conditions versus catalyst sample description

Heterogenization experimental conditionsSample Catalyst Final componentno. sample description Precursor preparation addition

1 Silica/Et(Ind)2ZrCl2/MAO Silica/Et(Ind)2ZrCl2: 6.55 g silica in30 ml toluene; 0.165 g zirconocenein 50 ml toluene; mixing time ¾2 hat 60 °C and 16 h at roomtemperature

MAO; 28 ml 10 wt.%MAO; mixing time ¾4 h;temperature ¾60 °C

2 Silica/MAO/Et(Ind)2ZrCl2 Silica/MAO: 6.55 g silica in 30 mltoluene; 28 ml 10 wt.% MAO;mixing time ¾4 h; temperature¾60 °C

Et(Ind)2ZrCl2: 0.165 gzirconocene in 50 mltoluene; mixing time¾2 h at 60 °C and 16 h atroom temperature

3 Silica/MAO-pretreatedEt(Ind)2ZrCl2

MAO-pretreated Et(Ind)2ZrCl2:0.165 g zirconocene in 50 mltoluene; 70 ml 10 wt.% MAO;pretreatment time ¾30 min at 60 °C

Silica: 6.55 g silica in30 ml toluene; mixingtime ¾2 h at 60 °C and16 h at room temperature

Spectra were acquired using the VG-Eclipse softwarepackage, which incorporates several data analysis fea-tures such as background subtraction, smoothing andnon-linear least-squares curve-fitting.16 A Shirley-typebackground17 was subtracted from all of the spectra. Datawere smoothed using a cubic function. Silica is an inertsupport, so its Si 2p level binding energy.103.4 eV/15

was used to correct for the energy shift due to surfacecharging.

The spectral peaks were first located roughly throughvisual inspection of an XPS spectrum. Then these approx-imate peak positions were used as initial values by theabove curve-fitting program to accomplish the exact fit-ting and obtain the fitted parameters, such as exact peakposition, full width at half-maximum (FWHM), peakheight, peak area and Gaussian/Lorentzian (G/L) ratio.

The spectral broadening effects due to the electron energyanalyser (Gaussian component) and the natural line width(Lorentzian component) of the x-ray source were removedusing the deconvolution procedure. The Zr 3d regions ofthe catalyst samples were best fitted with five peaks hav-ing a G/L ratio of¾0.3. The widths of the peaks werenot kept fixed while fitting the curves. Nevertheless, theFWHM of the identical peaks of the catalyst samples doesnot exceed theš0.2 eV uncertainty limit.

RESULTS

Table 3 shows the Zr 3d5/2 binding energies of the ref-erence samples Et(Ind)2ZrCl2, and silica/Et(Ind)2ZrCl2,

Table 3. Zirconium 3d binding energies (eV) and spin-orbit split (3d5=2–3d3=2) in Set 1 and Set 2in the XPS spectra of heterogenized Et(Ind)2ZrCl 2

Sample Catalyst sample Peak of Binding energyno. description interest (š0.2 eV)

1 Silica/Et(Ind)2ZrCl2/MAO Zr 3d3/2 Set 1 184.7Set 2 185.7

Zr 3d5/2 Set 1 182.6Set 2 183.6

Spin-orbit (3d5/2 3d3/2) Split Set 1 2.1Spin-orbit (3d5/2 3d3/2) Split Set 2 2.1

2 Silica/MAO/Et(Ind)2ZrCl2 Zr 3d3/2 Set 1 185.1Set 2 186.1

Zr 3d5/2 Set 1 182.8Set 2 183.9

Spin-orbit (3d5/2 3d3/2) Split Set 1 2.3Spin-orbit (3d5/2 3d3/2) Split Set 2 2.3

3 Silica/MAO-pretreated Et(Ind)2ZrCl2 Zr 3d3/2 Set 1 184.9Set 2 185.9

Zr 3d5/2 Set 1 182.9Set 2 183.9

Spin-orbit (3d5/2 3d3/2) Split Set 1 2.0Spin-orbit (3d5/2 3d3/2) Split Set 2 2.0

Reference samplesEt(Ind)2ZrCl2 Zr 3d5/2 182.0Silica/Et(Ind)2ZrCl2 Zr 3d5/2 182.7

Surf. Interface Anal. 27, 728–734 (1999) Copyright 1999 John Wiley & Sons, Ltd.

XPS OF Zr CATALYSTS 731

which were measured to be 182.0 and 182.7, respec-tively. It also summarizes the binding energy of the Zr3d spectral lines and the spin-orbit splits.3d5/2–3d3/2/in the XPS spectra of the catalyst samples as a func-tion of the catalyst preparatory methods. All the bindingenergy values were obtained using the VG-Eclipse peak-fitting program. The average FWHM value of these peakswas 1.2 eV.

Figure 1 (the fitted XPS spectra of the catalyst samples)shows that both Zr 3d5/2 and 3d3/2 spectral lines splitinto Set 1 and Set 2 components. Peak E, appearingas a shoulder at¾181.7 eV, is close to the Zr 3d5/2

binding energy of the Et(Ind)2ZrCl2 reference sample (seeTable 3), therefore the above low-energy shoulder may beattributed to unreacted Et(Ind)2ZrCl2 whose Zr 3d3/2 line(at ¾183.7 eV) is masked by the Set 2 3d5/2 line. Thisfinding, combined with the slightly varying FWHM of thepeaks, could have contributed to altering the ratio of the

corresponding. 3d5/2/3d3/2 peak intensities from 3 : 2. Thebinding energy of Al 2p remains constant at¾75.0 eV inall the samples.

The Set 1 3d5/2 binding energy values for Samples 1,2 and 3 are 182.6, 182.8 and 182.9 eV, respectively. TheSet 2 3d5/2 binding energy values, on the other hand, are183.6, 183.9 and 183.9 eV, respectively. In each set, thesebinding energy values are close to one another. However,the average Set 1 and Set 2 values are higher than the Zr3d5/2 binding energy of Et(Ind)2ZrCl2 by 0.8 and 1.8 eV,respectively. The average Set 1 and Set 2 Zr 3d5/2 bindingenergy values differ from each other by 1.0 eV. The spin-orbit .3d5/2–3d3/2/ split of Set 1 equals that of Set 2 fora given catalyst sample.

Figure 2 is an XPS spectrum of the silica/Et(Ind)2ZrCl2reference sample, showing that the Zr 3d5/2 binding energyline really occurs at¾182.7 eV. Note that this value isclose to the average Set 1 value.

Figure 1. Fitted XPS spectra of the silica/Et(Ind)2ZrCl2/MAO catalyst (Sample 1), the silica/MAO/Et(Ind)2ZrCl2 catalyst (Sample 2) andthe silica/MAO-pretreated Et(Ind)2ZrCl2 catalyst (Sample 3) in the Zr 3d region.

Copyright 1999 John Wiley & Sons, Ltd. Surf. Interface Anal. 27, 728–734 (1999)

732 M. ATIQULLAH ET AL.

Figure 1. (continued).

Figure 2. The XPS spectrum of the silica/Et(Ind)2ZrCl2 reference sample in the Zr 3d region.

DISCUSSION

TheSet1 Zr 3d5/2 bindingenergy of Sample1 is explainedasfollows. It approximatesto theZr 3d5/2 bindingenergyof the silica/Et(Ind)2ZrCl2 referencesampleand exceedsthe binding energy of Et(Ind)2ZrCl2 by 0.6 eV. Thisincrease in binding energy shows that interaction ofEt(Ind)2ZrCl2 with silica makesZr electron-deficientandproducesa zirconoceniumcation(Cation1). The genera-tion of a single cation in the absenceof MAO hasbeenobserved,however for thorocenesupportedon MgCl2and partially dehydroxylatedalumina,using [13C]NMR-CPMAS.4–6

The Set 2 Zr 3d5/2 binding energy of Sample 1 isgreaterthanthatof thesilica/Et(Ind)2ZrCl2 referencesam-ple by 0.9 eV. This meansthat treatmentwith the MAOco-catalystmakesCation 1 more electron-deficient,pro-ducinga differentzirconoceniumcation(Cation2).

A comparisonof Set1 andSet2 Zr 3d5/2 bindingener-gies of Samples2 and 3 with thoseof the Et(Ind)2ZrCl2and silica/Et(Ind)2ZrCl2 referencesamplesalso supportsthe abovefindings. This meansthat two zirconoceniumcationsarealsoproducedin Samples2 and3.

Within experimentaluncertainty, the Set 1 Zr 3d5/2

binding energies for all the catalyst samplesare closeto one another.Therefore,Cation 1 resides in nearlythe samesolid-stateelectronic environmentdespite the

Surf. InterfaceAnal. 27, 728–734 (1999) Copyright 1999JohnWiley & Sons,Ltd.

XPS OF Zr CATALYSTS 733

variation in the heterogenization procedures. The sameobservation holds for Cation 2 because of the closenessof the Set 2 Zr 3d5/2 binding energies. However, theelectronic environment of Cation 1 differs from that ofCation 2.

Generation of the above zirconocenium cations (onereferring to Set 1 and the other to Set 2) appears tosupport the following postulated surface chemistry (seeScheme 1).

In Method 1, Et(Ind)2ZrCl2 reacts with the silica surfaceOH groups first to form Cation 1, presumably in the formof a zirconocenium ion-pair [SiO]�[Et(Ind)2ZrCl]C.18,19

When it is treated subsequently with MAO, the fol-lowing occurs: MAO, being a Lewis acid, methylatesEt(Ind)2ZrCl2; consequently, a bridged zirconoceniummethyl species (Cation 2), presumed to be trapped and sta-bilized in multi-coordinate crown alumoxane complex(es),is produced:20,21 these zirconocenium cations provide thepolymerization catalytic sites and are postulated to differin the electronic and steric effects with respect to the silicasurface.3,18,19

In Method 2, MAO first reacts with the silica sur-face OH groups to form stable Si–O–Al bonds. The

surface-anchored MAO plays the following roles: methy-lation, Cl� ion abstraction and crown alumoxane com-plex(es) formation (Cation 2).20 – 22 The abstracted Cl�

is also postulated to be trapped and stabilized in multi-coordinate alumoxane complex(es).20,21

In Method 3, the preactivation of Et(Ind)2ZrCl2 byMAO in solution first forms, through the same mech-anism referred to in Method 2, the crown alumoxanecomplex(es)23 (Cation 2), which then anchor(s) with thesilica surface OH groups.

Chenet al.18 postulated that in Method 1 (Sample 1)treatment with MAO generates an ion-pair in the form of[SiO]�[Et(Ind)2ZrCH3]C. The present XPS results refutethis postulation by indicating that the resulting zirconoce-nium cation appears as crown alumoxane complex(es)(Cation 2).

The origin of Cation 1 in Methods 2 and 3 canbe explained as follows: MAO and MAO-pretreatedEt(Ind)2ZrCl2 selectively react in Methods 2 and 3, respec-tively, with the silica surface OH groups; consequently, aportion of the total surface OH groups remains unreacted.In Method 2, a part of the finally added Et(Ind)2ZrCl2reacts with the remaining OH groups to produce Cation 1.

Scheme 1. Surface chemistry related to the formation of the zirconocenium cations (Cations 1 and 2) in the heterogenized catalystsamples3,18 23.

Copyright 1999JohnWiley & Sons,Ltd. Surf. InterfaceAnal. 27, 728–734 (1999)

734 M. ATIQULLAH ET AL.

However, in Method 3, the Et(Ind)2ZrCl2 in equilibriumwith the MAO-pretreated Et(Ind)2ZrCl2 solution providesthe source for reaction with the OH groups to generateCation 1.

CONCLUSION

Characterization by XPS showed that heterogenizationof zirconocene and MAO, independent of the heteroge-nization methods, generated two types of zirconoceniumcations: Cation 1 and Cation 2. Based on the postulatedsurface chemistry, Cation 1 is presumed to be in the formof an ion-pair [SiO]�[Et(Ind)2ZrCl]C whereas Cation 2is trapped in the multi-coordinated crown of MAO (seeScheme 1). In the absence of MAO, only Cation 1 isformed.

In a slurry olefin polymerization system, Cations 1and 2 will provide different active sites that will influ-ence the rates of intiation, propagation, chain transfer andtermination in different ways. Consequently, the result-ing molecular weight distribution will be broad, whichhas been reported in the literature2 for heterogeneousmetallocene-based polyolefins.

We recommend that the present findings be investigatedfurther using [13C]NMR-CPMAS. This is what we plan todo in the future.

Acknowledgements

The authors acknowledge the support provided by the Research Instituteand the Physics Department of the King Fahd University of Petroleum& Minerals at Dhahran, Saudi Arabia for the present study. The authorsespecially thank Khurshid Alam and Munawwar Khan for their technicalassistance.

REFERENCES

1. F. Mahmud, M. Atiqullah, N. Jarrah and I. Hussain, Eur.Polym. J. 28, 1039 (1992).

2. O. Olabisi, M. Atiqullah and W. Kaminsky, J. Macromol. Sci.& Rev. Macromol. Chem. Phys. C37, 519 (1997).

3. M. R. Ribeiro, A. Deffieux and M. F. Portela, Ind. Eng. Chem.Res. 36, 1224 (1997).

4. P. J. Toscano and T. J. Marks, J. Am. Chem. Soc. 107, 653(1985).

5. P. J. Toscano and T. J. Marks, Langmuir 2, 820 (1986).6. D. Heddon and T. J. Marks, J. Am. Chem. Soc. 110, 1647

(1988).7. C. Sishta, R. M. Hathorn and T. J. Marks, J. Am. Chem. Soc.

114, 1112 (1992).8. P. G. Gassman and M. R. Callstrom, J. Am. Chem. Soc. 109,

7875 (1987).9. P. G. Gassman, D. W. Macomber and S. M. Willging, J. Am.

Chem. Soc. 107, 2380 (1985).10. R. J. Errington, Advanced Practical Inorganic and Metalor-

ganic Chemistry, 1st Edn. Blackie Academic and Profes-sional, London (1997).

11. S. Collins, W. M. Kelly and D. A. Holden, Macromolecules25, 1780 (1992).

12. T. E. Nowlin, F. Y. K. Lo, R. S. Shinomoto and P. P. Shirodkar,US Patent No. 5,332,706 (July 1994).

13. F. Y. K. Lo and A. L. Pruden, International Patent No. WO95/11,263 (April 1995).

14. E. E. Khawaja, Z. Hussain, M. S. Jazzar and O. B. Dabbousi,J. Non. Cryst. Solids 93, 45 (1987).

15. D. Briggs and M. P. Seah, Practical Surface Analysis, 2ndEdn, Vol. 1. Wiley, New York (1990).

16. A. Proctor and R. M. A. Sherwood, Anal. Chem. 52, 2315(1980).

17. D. A. Shirley, Phys. Rev. B5, 4709 (1972).18. Y-X. Chen, M. D. Rausch and J. C. Chien, J. Polym. Sci. A 33,

2093 (1995).19. W. Kaminsky, F. Renner and H. Winkelbach, Worldwide

Metallocene Conference, Metcon 94, Houston, Texas, USA,25 27 May 1994.

20. K. Soga and M. Kaminaka, Macromol. Chem. Phys. 195, 1369(1994).

21. J. C. W. Chien and D. He, J. Polym. Sci. A 29, 1603 (1991).22. F. Ciardelli, A. Altomare and M. Michelotti, Catal. Today 41,

149 (1998).23. C. J. Harlan, S. G. Bott and A. R. Barron, J. Am. Chem. Soc.

117, 6465 (1995).

Surf. Interface Anal. 27, 728–734 (1999) Copyright 1999 John Wiley & Sons, Ltd.