A comparative XPS study of the MgCl2/SiO2/Si interface prepared under UHV conditions and by different organic solvents. A. Siokou1, S. Karakalos1,2, S. Ntais1 and S. Ladas1,2 1
FORTH/ICE-HT, Stadiou Str., Platani Achaias, GR-26504, P.O. Box, 1414, Patras, Greece Department of Chemical Engineering, University of Patras, University Campus, GR26504, Rion, Patras, Greece 2
[email protected] Silicon oxide in combination with magnesium chloride is used as a support for the industrial preparation of some types of Ziegler-Natta catalysts. These catalysts are the best candidates for the low pressure production of linear low-density polyethylene and ethylene co-polymers from the gas phase, with improved optical properties in the film form. By many authors silica is believed to act as an inert support, where MgCl2 simply covers its outer surface while the active species of silica do not affect the catalysts active sites for polymerization [1]. On the other hand some consider that silica behaves as a normal solid state reactant with high surface area. In other words, the active sites into the pores of each particle are accessible for reaction with precursor solutions of the reactants. The usual industrial way of preparation of this type of support is by impregnation of precursor organo-metallic MgCl2 compounds on high surface area (300-700 m2 g-1) silica. The study of this system at a molecular level is a particularly difficult task due to the sensitivity of magnesium chloride to oxygen and humidity. Furthermore, although surface sensitive spectroscopies can provide significant information on the nature of surface species, the surface composition and structure, the poor conductivity of both SiO2 and MgCl2 in combination with the porous nature of the former make surface analysis rather complicated. In the preset work the interaction between MgCl2 and SiO2 is investigated by X-ray Photoelectron Spectroscopy. The porous silica substrate is replaced by a “model” flat support that is electrically conductive providing a surface science compatible “realistic” system. For this purpose a Si(100) wafer is used on which a thin SiO2 layer is grown. The oxide layer is chosen to be thicker than 3 nm to resemble the chemical properties of the real oxide [2] and thin enough to avoid electrostatic charging during XPS measurements. For the purposes of the present study MgCl2 is applied on this support in three different ways: i) in-situ by evaporation under UHV conditions, ii) ex-situ by drop-casting from a sparse (0.1M) tetrahydrofuran (THF) solution and iii) ex-situ by drop-casting from a dense (1 M) ethanol solution. The Si2p, O1s, C1s, Mg2p and Cl2p XPS peaks and Mg(KLL) XAES peak were recorded before and after deposition. All samples where annealed at 723 K for further investigation. As shown in Table I the best defined Mg2p and Cl2p peaks (fwhm=1.9 eV and 2 eV respectively) are recorded after MgCl2 deposition under UHV conditions. The surface atomic ratio is Cl/Mg=2, a value achieved already at sub-monolayer coverage, indicating a molecular deposition of MgCl2 without significant interaction with the substrate. On the other hand, when a 1.3 nm thick overlayer is formed by drop-casting from THF solution the Cl/Mg ratio has a lower value (Cl/Mg=1.2) and the width of the Cl2p and Mg2p peaks is considerably larger (fwhm=3.2 eV and 2.8 eV respectively) indicating an inhomogeneous electronic environment around Mg and Cl atoms due to the presence of trapped solvent molecules. Furthermore, the low concentration of MgCl2 in the THF solution results to the disruption of the MgCl2 structure by THF molecules.
Table I: Cl2p and Mg2p Binding Energies (BE), Mg(KLL) Kinetic Energy (KE), full width at half maximum (fwhm) of all the peaks, modified Auger parameter α΄, Mg/Cl surface atomic ratio, IO1s /ISi2p, IC1s/ISi2p and IMg2p/ISi2p intensity ratios and the initial thickens of the deposit (Ddepos ).
4.2 nm SiO2 MgCl2/ THF Soluti. (O.1M) Dde pos = 1.3 nm 3.4 nm SiO2 MgCl2/ EtOH Solut. (1M) Dde pos = 9.4 nm 12 nm SiO2 MgCl2 (evaporation) Dde pos = 1.8 nm MgCl2 (evaporation) Dde pos = 3.1 nm
Mg2p BE, eV (fwhm, eV)
Cl2p BE, eV (fwhm, eV)
Mg(KLL) KE, eV (fwhm, eV)
IO1s/ ISi2p
IC1s/ ISi2p
200.2 (2.8) 200.2 (3.8)
1178.7 (3.4) 1179.1 (4.3)
1231.4
723Κ
52.7 (3.2) 51.6 (3.6)
3.1
0.24
1.2
3.1
0.92
0.04
1230.7
0.31
3.5 2.2
0.46 0.06
0.06
RT
52.6 (2.1)
200.4 (2.3)
1178.5 (2.5)
723Κ
51.8 (2.8)
1231.1
2.3
5.6
6.9
3.27
200.2 (2.7)
1179.1 (3.3)
1230.9
0.2
3.5 5.4
0.4 0.06
0.23
200 (2) 200.2 (2)
1179 (2.4) 1178.9 (2.2)
1231.2
2
4.4
0.04
0.12
RT
52.2 (1.9) 52.4 (1.9)
1231.3
2.3
3.9
0.03
0.33
723K
52.2 (2.3)
200 (2.3)
1178.4 (2.3)
1230.6
1.4
5.5
0.08
0.03
RT
RT
α΄, eV
Cl/ Mg
I Mg2p /ISi2p
The formation of surface complexes between MgCl2 and THF at least at the top surface layers is not confirmed from the present results but cannot be excluded for layers closer to the interface with silica. When the deposition takes place from the dense ethanol solution a 9.4 nm thick layer is formed consisting mainly of stoichimetric MgCl2. The existence of trapped solvent molecules is inevitable as is evident from the C1s and O1s signals increment that follows deposition. In all the above cases the modified Auger parameter has a value (α΄=1231.3±0.1) similar to the one measured for a 10 nm thick pure MgCl2 layer evaporated on Au foil. This indicates that in all cases that least the top layers of the deposit consist of the stoichiometric compound. At 723 K the O1s and C1s signals decrease in the cases where the deposition has taken place from the solutions, indicating desorption of solvent molecules or their decomposition products. When the first sample is annealed, extended desorption of Cl atoms is observed while Mg remains on the surface in an oxidised state. On the other hand when the thicker deposits are annealed at the same temperature almost 90% of the Mg atoms leave the surface in the form of MgCl2. The amount of Mg atoms that remains on the oxide’s surface is larger in the case of the sample prepared by ethanol solution. The BE of the Mg2p and the value of α΄ indicate extended oxidation of the Mg atoms. Comparison of the state of the surface for the three samples after annealing leads to the conclusion that the existence of solvent molecules or their dissociation products assist Mg oxidation at elevated temperatures. In conclusion the present results show that the existence of the solvent plays an important role on the MgCl2 /SiO2 interfacial properties while the solution concentration controls the amount of MgCl2 deposited. Sparse solutions lead to the formation of substoichiometric and partially oxidized deposit which is not a good candidate for a ZieglerNatta support due to the lack of Cl atoms which are important for the electron transfer between Mg and Ti atoms when the next component of the catalyst (TiCl4) is applied. The experiment that was performed under UHV conditions shows that at RT there is no significant interfacial interaction between MgCl2 and SiO2 proving that silica in this case acts as an inert support used only to control the shape and size of the produced polymer particles. References [1] Union Carbide Corporation Unites States Patend, 4, 302, 566. [2] H.E. Bergna, “The colloid Chemistry of Silica”, ACS, Washinghton, DC, 1990, p.1.
