- Email: [email protected]
Density, thickness and composition measurements of TiO2SiO2 thin films by coupling X-ray reflectometry, ellipsometry and electron probe microanalysis-X
Applied Surface Science 253 (2006) 363–366 www.elsevier.com/locate/apsusc
Density, thickness and composition measurements of TiO2–SiO2 thin films by coupling X-ray reflectometry, ellipsometry and electron probe microanalysis-X A. Hodroj a,*, H. Roussel a, A. Crisci b, F. Robaut b, U. Gottlieb a, J.L. Deschanvres a a
Laboratoire des Mate´riaux et du Ge´nie Physique (UMR 5628 CNRS), Ecole Nationale Supe´rieure de Physique de Grenoble, Institut National Polytechnique de Grenoble, Minatec, 3 parvis Louis Ne´el, BP 257, 38016 Grenoble Cedex 1, France b Consortium des Moyens Technologiques Communs, Institut National Polytechnique de Grenoble, BP 75, 38402 St. Martin d’He`res, France Available online 25 July 2006
Abstract Mixed TiO2–SiO2 thin films were deposited by aerosol atmospheric CVD method by using di-acetoxi di-butoxi silane (DADBS) and Ti tetrabutoxide as precursors. By varying the deposition temperatures between 470 and 600 8C and the ratios between the Si and Ti precursors (Si/Ti) from 2 up to 16, films with different compositions and thicknesses were deposited. The coupled analysis of the results of different characterisation methods was used in order to determine the variation of the composition, the thickness and the density of the films. First EPMA measurements were performed at different acceleration voltages with a Cameca SX50 system. By analysing, with specific software, the evolution of the intensity ratio Ix/Istd versus the voltage, the composition and the mass thickness (product of density by the thickness) were determined. In order to measure independently the density, X-ray reflectometry experiments were performed. By analysing the value of the critical angle and the Kiessig fringes, the density and the thickness of the layers were determined. The refractive index and the thickness of the films were also measured by ellipsometry. By assuming a linear interpolation between the index value of the pure SiO2 and TiO2 films, the film composition was deduced from the refractive index value. XPS measurements were also performed in order to obtain an independent value of the composition. A good agreement between the ways to measure the density is obtained. # 2006 Elsevier B.V. All rights reserved. Keywords: TiO2–SiO2; Ellipsometry; XPS measurements; X-ray reflectometry; EPMA
1. Introduction TiO2 and SiO2 materials are commonly used as optical thin films in the visible and near-infrared wavelength ranges. TiO2– SiO2 films have a wide range of achievable refractive index due to the large difference in refractive index between TiO2 and SiO2. This property is useful for the preparation of several optical devices (antireflective coatings, passive or active waveguides). Titanium silicate thin films can also be used as dielectric materials due to a high dielectric constant, a high breakdown field and a low leakage current [1]. We deposited TiO2–SiO2 thin film with an atmospheric aerosol CVD reactor. This process is based on the pyrolysis on a
* Corresponding author. Tel.: +33 4 56 52 93 23; fax: +33 4 56 52 93 01. E-mail address: [email protected] (A. Hodroj). 0169-4332/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2006.06.014
heated substrate of on aerosol produced by ultrasonic spraying of a solution [2,3]. As starting solution, di-acetosi di-butoxi silane (DADBS) and Ti tetra-butoxide precursors are dissolved in organic solvent [4]. Cleaned IR transparent (1 0 0) oriented silicon wafers were used as substrates. The substrate temperature is varied from 470 to 600 8C. Related to the CVD conditions, the deposited films are adherent. As observed by scanning electron microscopy (SEM), the surface morphology is very flat and smooth. This fact is confirmed by atomic force microscopy (AFM) measurement (Fig. 1) which revealed a roughness ranging from 0.25 to 1 nm (RMS). In order to determine the films thickness, their composition and the film density, we proposed a coupled analysis of the results of different characterisation techniques. This coupled analysis is described by approach (Fig. 2). The film density is deduced from X-ray reflectometry (XRR) at a first time. On the other hand, the measured values of electron probe
364
A. Hodroj et al. / Applied Surface Science 253 (2006) 363–366
Fig. 1. AFM topography of TiO2–SiO2 films: the RMS roughness is about 0.28 nm.
Fig. 3. Refractive index of TiO2–SiO2 films as a function of the deposition temperature for different Ti/Si ratios.
3. Results and discussion
Fig. 2. Analysis scheme approach.
microanalysis-X (EPMA) treated by STRATAGEM software can give the mass thickness of the film. Using the thickness obtained by both ellipsometry and XRR techniques, we can deduce the density of the film at a second time. The two values of film density are then compared. 2. Experimental The surface morphology of the deposited sample was observed with a Philips XL30 SEM equipped with a conventional thermoelectronic tungsten gun. Refraction index and thickness were measured on an Ellipsometer at 633 nm. We used a commercial atomic force microscopy from Nanoscope (Veeco 3100). X-ray reflectometry experiments were carried out on a Brucker D500 diffractometer equipped with Cu Ka radiation and a front monochromator. X-ray photoelectron spectroscopy (XPS) was studied with XR3E2 apparatus from vacuum generator in a UHV chamber (10 10 mbar), sample surfaces were irradiated with Mg Ka radiation (1253.6 eV). The ejected electrons were collected by a hemispherical analyser at constant pass energy of 30 eV. Analyses were carried out at the angle between the sample surface and the analyser, namely a = 908. A xenon erosion (at beam energy 3 kV for 15 min) was used to eliminate the surface contamination. Electron probe X microanalysis (EPMA) was performed on a Cameca SX50 system.
The films were systematically characterised by ellipsometry at 633 nm; in order to explore the thickness and the refractive index n. The film thickness ranges from 80 to 120 nm, and the refractive index from 1.57 to 1.97. In Fig. 3, the refractive index of the films is plotted, versus different deposition temperatures for different Ti/Si source ratios. For all solution, when the substrate temperature increases, the refractive index decreases. This result is related to an increase of the SiO2 content in the film. In order to relate the measured values of refractive index of the films to their composition, we performed quantified XPS measurements on various films. As shown in Fig. 4, after erosion, the measured spectra were exempted of any contamination, except xenon line. For as-deposited films, the energy scale was calibrated with the C 1s line at 285.0 eV. Under this condition, the Ti 2p3/2 line is centred at 458.5 eV. After erosion, the energy was calibrated to keep the same value for Ti 2p3/2 line. The Si 2p3/2 and Ti 2p3/2 lines corresponded to single components with position and width values of 102.5/1.7 and 458.5/1.8 eV, which are ascribed to SiO2 and TiO2 contributions, respectively [5,6]. We obtained the film composition within uncertainty of 10% by using sensitivity factors of 0.17 and 1.1 relative to Si 2p3/2 and Ti 2p3/2 lines, respectively. In Fig. 5, we present the refractive index as a function of the percentage of SiO2 relative to TiO2 in the films. We can notice that, for more than 60% of SiO2, the
Fig. 4. XPS spectrum of TiO2–SiO2 films after xenon erosion. The Si 2p3/2 line is plotted in the inset.
A. Hodroj et al. / Applied Surface Science 253 (2006) 363–366
Fig. 5. Refractive index of TiO2–SiO2 films as a function of the silica content, calculated from XPS results.
Fig. 6. X-ray reflectivity data and best fit for a typical sample as function of incident angle.
refractive index depends linearly on the SiO2 content. According to this result, the film composition can be deduced by a measurement of the refractive index. In Fig. 6, we present an example of the results obtained by X-ray reflectometry. An excellent agreement was obtained between X-ray reflectivity data and the model fit. The thickness of the film was determined from the period of interference fringes and the density was deduced from the critical incident angle [7]. For the analysis of thin film by EPMA, it is necessary to perform measurements at different acceleration voltages since the electron beam interacts with the substrate even at low voltages. So, the intensity ratios Ix/Istd are measured at 10, 15, 20 kV. The set of experimental values is analysed by a specific software STRATAGEM [8,9] and by procedure fitting between Table 1 Experimental results Sample
SiP12 SiP20 SiP18 SiP13 SiP19 SiP25
Ellipsometry
XRR
EPMA
XPS
n
e (nm)
e (nm)
d (g/cm3)
d (g/cm3)
%SiO2
1.97 1.81 1.71 1.69 1.62 1.57
80 89 93 101 104 108
80.3 90.5 94 101.7 105 107.5
2.82 2.56 2.55 2.45 2.4 2.2
2.73 2.61 2.56 2.47 2.31 2.13
20 44 60 63 74 82
365
Fig. 7. Film densities obtained by XRR and EPMA as a function of the silica content of the films.
the experimental points with the calculated Ix/Istd curve, the composition and the mass thickness are determined. The Table 1 summarizes the experimental results of ellipsometry, XRR, EPMA and XPS for the samples studied. The refractive index, the thickness, the density and the percentage of SiO2 are presented. Comparing the thickness values obtained by ellipsometry and XRR, we notice the good agreement between the two results. In Fig. 7, we present the obtained density as a function of the silicon content of the films. Film densities are obtained by XRR and EPMA as described in Fig. 2. We note the excellent agreement between the values obtained by the two techniques. Difference is less than 5% inferior to experimental uncertainties. As expected, film densities decrease when the SiO2 content increases. 4. Conclusion Homogeneous TiO2–SiO2 thin films were deposited by aerosol CVD at atmospheric pressure. Their refractive index and their composition are function of the elaboration conditions. Ellipsometry and XRR yield identical results for film thickness. The coupled analysis of ellipsometry, XPS, XRR and EPMA results leads to the variation of the film densities as a function of film composition. A coherent value of density is obtained either by XRR or by ellipsometry and EPMA. Acknowledgements The authors would like to thank Ing. Gregory Berthome (LTPCM, Grenoble, France) for the XPS measurements. This work was supported by the Functional Advanced Materials and Engineering of hybrids and ceramics, European Network of Excellence. References [1] D. Brassard, D.K. Sarkar, M.A. El Khakani, L. Ouellet, J. Vac. Sci. Technol. A 22 (3) (2004) 851. [2] J.L. Deschanvres, F. Cellier, M. Labeau, M. Langlet, J.C. Joubert, J. Phys. C 50 (5) (1989) 695.
366
A. Hodroj et al. / Applied Surface Science 253 (2006) 363–366
[3] M. Labeau, P. Rey, J.L. Deschanvres, J.C. Joubert, G. Delabpoglisse, Thin Solid Films 213 (1992) 94. [4] J.L. Deschanvres, J.C. Joubert, Electrochem. Soc. Proc. 97 (25) (1997) 1071. [5] T.L. Barr, J. Phys. Chem. 82 (16) (1978) 1801.
[6] D. Rats, L. Vandenbulcke, R. Herbin, R. Benoit, R. Erre, V. Serin, J. Sevely, Thin Solid Films 270 (1995) 177. [7] O. Durand, A. De Rossi, A. Bouchier, J. Phys. IV France 118 (2004) 203. [8] Stratagem Version 2.6, SAMx, 4, rue Galile´e, 78280 Guyancourt, France. [9] J.L. Pouchou, F. Pichoir, Scanning Microsc. 7 (1993) 167.
Density, thickness and composition measurements of TiO2–SiO2 thin films by coupling X-ray reflectometry, ellipsometry and electron probe microanalysis-X A. Hodroj a,*, H. Roussel a, A. Crisci b, F. Robaut b, U. Gottlieb a, J.L. Deschanvres a a
Laboratoire des Mate´riaux et du Ge´nie Physique (UMR 5628 CNRS), Ecole Nationale Supe´rieure de Physique de Grenoble, Institut National Polytechnique de Grenoble, Minatec, 3 parvis Louis Ne´el, BP 257, 38016 Grenoble Cedex 1, France b Consortium des Moyens Technologiques Communs, Institut National Polytechnique de Grenoble, BP 75, 38402 St. Martin d’He`res, France Available online 25 July 2006
Abstract Mixed TiO2–SiO2 thin films were deposited by aerosol atmospheric CVD method by using di-acetoxi di-butoxi silane (DADBS) and Ti tetrabutoxide as precursors. By varying the deposition temperatures between 470 and 600 8C and the ratios between the Si and Ti precursors (Si/Ti) from 2 up to 16, films with different compositions and thicknesses were deposited. The coupled analysis of the results of different characterisation methods was used in order to determine the variation of the composition, the thickness and the density of the films. First EPMA measurements were performed at different acceleration voltages with a Cameca SX50 system. By analysing, with specific software, the evolution of the intensity ratio Ix/Istd versus the voltage, the composition and the mass thickness (product of density by the thickness) were determined. In order to measure independently the density, X-ray reflectometry experiments were performed. By analysing the value of the critical angle and the Kiessig fringes, the density and the thickness of the layers were determined. The refractive index and the thickness of the films were also measured by ellipsometry. By assuming a linear interpolation between the index value of the pure SiO2 and TiO2 films, the film composition was deduced from the refractive index value. XPS measurements were also performed in order to obtain an independent value of the composition. A good agreement between the ways to measure the density is obtained. # 2006 Elsevier B.V. All rights reserved. Keywords: TiO2–SiO2; Ellipsometry; XPS measurements; X-ray reflectometry; EPMA
1. Introduction TiO2 and SiO2 materials are commonly used as optical thin films in the visible and near-infrared wavelength ranges. TiO2– SiO2 films have a wide range of achievable refractive index due to the large difference in refractive index between TiO2 and SiO2. This property is useful for the preparation of several optical devices (antireflective coatings, passive or active waveguides). Titanium silicate thin films can also be used as dielectric materials due to a high dielectric constant, a high breakdown field and a low leakage current [1]. We deposited TiO2–SiO2 thin film with an atmospheric aerosol CVD reactor. This process is based on the pyrolysis on a
* Corresponding author. Tel.: +33 4 56 52 93 23; fax: +33 4 56 52 93 01. E-mail address: [email protected] (A. Hodroj). 0169-4332/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2006.06.014
heated substrate of on aerosol produced by ultrasonic spraying of a solution [2,3]. As starting solution, di-acetosi di-butoxi silane (DADBS) and Ti tetra-butoxide precursors are dissolved in organic solvent [4]. Cleaned IR transparent (1 0 0) oriented silicon wafers were used as substrates. The substrate temperature is varied from 470 to 600 8C. Related to the CVD conditions, the deposited films are adherent. As observed by scanning electron microscopy (SEM), the surface morphology is very flat and smooth. This fact is confirmed by atomic force microscopy (AFM) measurement (Fig. 1) which revealed a roughness ranging from 0.25 to 1 nm (RMS). In order to determine the films thickness, their composition and the film density, we proposed a coupled analysis of the results of different characterisation techniques. This coupled analysis is described by approach (Fig. 2). The film density is deduced from X-ray reflectometry (XRR) at a first time. On the other hand, the measured values of electron probe
364
A. Hodroj et al. / Applied Surface Science 253 (2006) 363–366
Fig. 1. AFM topography of TiO2–SiO2 films: the RMS roughness is about 0.28 nm.
Fig. 3. Refractive index of TiO2–SiO2 films as a function of the deposition temperature for different Ti/Si ratios.
3. Results and discussion
Fig. 2. Analysis scheme approach.
microanalysis-X (EPMA) treated by STRATAGEM software can give the mass thickness of the film. Using the thickness obtained by both ellipsometry and XRR techniques, we can deduce the density of the film at a second time. The two values of film density are then compared. 2. Experimental The surface morphology of the deposited sample was observed with a Philips XL30 SEM equipped with a conventional thermoelectronic tungsten gun. Refraction index and thickness were measured on an Ellipsometer at 633 nm. We used a commercial atomic force microscopy from Nanoscope (Veeco 3100). X-ray reflectometry experiments were carried out on a Brucker D500 diffractometer equipped with Cu Ka radiation and a front monochromator. X-ray photoelectron spectroscopy (XPS) was studied with XR3E2 apparatus from vacuum generator in a UHV chamber (10 10 mbar), sample surfaces were irradiated with Mg Ka radiation (1253.6 eV). The ejected electrons were collected by a hemispherical analyser at constant pass energy of 30 eV. Analyses were carried out at the angle between the sample surface and the analyser, namely a = 908. A xenon erosion (at beam energy 3 kV for 15 min) was used to eliminate the surface contamination. Electron probe X microanalysis (EPMA) was performed on a Cameca SX50 system.
The films were systematically characterised by ellipsometry at 633 nm; in order to explore the thickness and the refractive index n. The film thickness ranges from 80 to 120 nm, and the refractive index from 1.57 to 1.97. In Fig. 3, the refractive index of the films is plotted, versus different deposition temperatures for different Ti/Si source ratios. For all solution, when the substrate temperature increases, the refractive index decreases. This result is related to an increase of the SiO2 content in the film. In order to relate the measured values of refractive index of the films to their composition, we performed quantified XPS measurements on various films. As shown in Fig. 4, after erosion, the measured spectra were exempted of any contamination, except xenon line. For as-deposited films, the energy scale was calibrated with the C 1s line at 285.0 eV. Under this condition, the Ti 2p3/2 line is centred at 458.5 eV. After erosion, the energy was calibrated to keep the same value for Ti 2p3/2 line. The Si 2p3/2 and Ti 2p3/2 lines corresponded to single components with position and width values of 102.5/1.7 and 458.5/1.8 eV, which are ascribed to SiO2 and TiO2 contributions, respectively [5,6]. We obtained the film composition within uncertainty of 10% by using sensitivity factors of 0.17 and 1.1 relative to Si 2p3/2 and Ti 2p3/2 lines, respectively. In Fig. 5, we present the refractive index as a function of the percentage of SiO2 relative to TiO2 in the films. We can notice that, for more than 60% of SiO2, the
Fig. 4. XPS spectrum of TiO2–SiO2 films after xenon erosion. The Si 2p3/2 line is plotted in the inset.
A. Hodroj et al. / Applied Surface Science 253 (2006) 363–366
Fig. 5. Refractive index of TiO2–SiO2 films as a function of the silica content, calculated from XPS results.
Fig. 6. X-ray reflectivity data and best fit for a typical sample as function of incident angle.
refractive index depends linearly on the SiO2 content. According to this result, the film composition can be deduced by a measurement of the refractive index. In Fig. 6, we present an example of the results obtained by X-ray reflectometry. An excellent agreement was obtained between X-ray reflectivity data and the model fit. The thickness of the film was determined from the period of interference fringes and the density was deduced from the critical incident angle [7]. For the analysis of thin film by EPMA, it is necessary to perform measurements at different acceleration voltages since the electron beam interacts with the substrate even at low voltages. So, the intensity ratios Ix/Istd are measured at 10, 15, 20 kV. The set of experimental values is analysed by a specific software STRATAGEM [8,9] and by procedure fitting between Table 1 Experimental results Sample
SiP12 SiP20 SiP18 SiP13 SiP19 SiP25
Ellipsometry
XRR
EPMA
XPS
n
e (nm)
e (nm)
d (g/cm3)
d (g/cm3)
%SiO2
1.97 1.81 1.71 1.69 1.62 1.57
80 89 93 101 104 108
80.3 90.5 94 101.7 105 107.5
2.82 2.56 2.55 2.45 2.4 2.2
2.73 2.61 2.56 2.47 2.31 2.13
20 44 60 63 74 82
365
Fig. 7. Film densities obtained by XRR and EPMA as a function of the silica content of the films.
the experimental points with the calculated Ix/Istd curve, the composition and the mass thickness are determined. The Table 1 summarizes the experimental results of ellipsometry, XRR, EPMA and XPS for the samples studied. The refractive index, the thickness, the density and the percentage of SiO2 are presented. Comparing the thickness values obtained by ellipsometry and XRR, we notice the good agreement between the two results. In Fig. 7, we present the obtained density as a function of the silicon content of the films. Film densities are obtained by XRR and EPMA as described in Fig. 2. We note the excellent agreement between the values obtained by the two techniques. Difference is less than 5% inferior to experimental uncertainties. As expected, film densities decrease when the SiO2 content increases. 4. Conclusion Homogeneous TiO2–SiO2 thin films were deposited by aerosol CVD at atmospheric pressure. Their refractive index and their composition are function of the elaboration conditions. Ellipsometry and XRR yield identical results for film thickness. The coupled analysis of ellipsometry, XPS, XRR and EPMA results leads to the variation of the film densities as a function of film composition. A coherent value of density is obtained either by XRR or by ellipsometry and EPMA. Acknowledgements The authors would like to thank Ing. Gregory Berthome (LTPCM, Grenoble, France) for the XPS measurements. This work was supported by the Functional Advanced Materials and Engineering of hybrids and ceramics, European Network of Excellence. References [1] D. Brassard, D.K. Sarkar, M.A. El Khakani, L. Ouellet, J. Vac. Sci. Technol. A 22 (3) (2004) 851. [2] J.L. Deschanvres, F. Cellier, M. Labeau, M. Langlet, J.C. Joubert, J. Phys. C 50 (5) (1989) 695.
366
A. Hodroj et al. / Applied Surface Science 253 (2006) 363–366
[3] M. Labeau, P. Rey, J.L. Deschanvres, J.C. Joubert, G. Delabpoglisse, Thin Solid Films 213 (1992) 94. [4] J.L. Deschanvres, J.C. Joubert, Electrochem. Soc. Proc. 97 (25) (1997) 1071. [5] T.L. Barr, J. Phys. Chem. 82 (16) (1978) 1801.
[6] D. Rats, L. Vandenbulcke, R. Herbin, R. Benoit, R. Erre, V. Serin, J. Sevely, Thin Solid Films 270 (1995) 177. [7] O. Durand, A. De Rossi, A. Bouchier, J. Phys. IV France 118 (2004) 203. [8] Stratagem Version 2.6, SAMx, 4, rue Galile´e, 78280 Guyancourt, France. [9] J.L. Pouchou, F. Pichoir, Scanning Microsc. 7 (1993) 167.