Multilayer Alumina and Titania Optical Coatings ...

Multilayer Alumina and Titania Optical Coatings Prepared by Atomic Layer Deposition Nemo Biluš Abaffy*a, Peter Evansb, Gerry Trianib and Dougal McCullocha a Applied Physics, Royal Melbourne Institute of Technology 124 LaTrobe Street, Melbourne, Australia b Institute of Materials Engineering, ANSTO Menai, NSW 2234, Australia ABSTRACT The microstructure and optical properties of alumina and titania multilayer coatings prepared using atomic layer deposition (ALD) has been investigated. The titania layers were prepared using TiCl4+H2O as the precursors while two different precursors, Al(CH3)3+H2O and AlCl3+H2O, were used to deposit the alumina layers. The results show that ALD can be used to produce amorphous, stoichiometric alumina and titania thin films with uniform thicknesses at low temperatures (120 ºC). An antireflective coating design based on 4 alternating layers of titania and alumina was prepared and the resulting reflectance compared to theoretical calculations. The results demonstrate that ALD is a suitable technique for the deposition of optical thin films at temperatures compatible with thermally sensitive substrates. Keywords: ALD, atomic layer deposition, thin film, dielectric, alumina, titania, optical coating, antireflection

1. INTRODUCTION Atomic Layer Deposition (ALD) shows considerable promise for the synthesis of advanced broadband optical coatings. The non-line-of-sight capability of this technique produces film coverage on all surfaces exposed to the precursor gas flows. In addition, ALD is a surface mediated process that permits precise control over the thickness of each coating layer, while maintaining uniformity and producing sharp interfaces between the layers. Another benefit of ALD is that depending on the reaction kinetics, low temperature depositions are possible enabling the coating of temperature sensitive substrates. Al2O3 and TiO2 are widely used optical materials in multilayer optical thin films due to their high transparency, contrasting refractive indices and the ease with which they can be deposited by various methods. There have been several reports of the use of ALD to prepare Al2O3 and TiO2 films [1-5]. Several possible precursors can be used in the ALD process to produce Al2O3 and TiO2. For example, TiO2 can be produced by the vapour phase reaction of TiCl4+H2O [3], while Al2O3 can be formed by the reaction between precursor couples Al(CH3)3+H2O [1] or AlCl3+H2O [2,6]. There is debate as to the best precursor to employ for optical applications, particularity when low deposition temperatures are used. In this paper, we investigate low temperature depositions of Al2O3 and TiO2 using these different precursors. The optical properties, microstructure and composition of the coatings are explored using spectroscopic ellipsometry, electron microscopy and Auger electron spectroscopy respectively. A multilayer antireflective coating based on alumina and titania was then synthesized at 120°C and its optical response compared to theoretical models.

2. EXPERIMENTAL METHODS Alumina and titania thin films were deposited on Si substrates using a flow type reactor (Microchemistry ALCVDTMF120) at a reaction chamber temperature of 120ºC with nitrogen as the carrier gas. Figure 1 shows a schematic of this ALD system. The appropriate carrier gas flows were chosen for the depositions with the primary flow set to 350sccm and the secondary 200sccm. TiCl4, H2O and Al(CH3)3 precursors were used in the form of liquids at room temperature while the AlCl3 precursor is in solid form at low temperatures. The vapour pressure of the solid AlCl3 was increased by raising its temperature to 90ºC and maintaining it throughout the deposition. Large enough purge times after the precursor pulses were used in order to remove the physisorbed species on the sample during the deposition process and the appropriate primary and secondary N2 flows were chosen.

Nanostructured Thin Films, edited by Geoffrey B. Smith, Akhlesh Lakhtakia, Proc. of SPIE Vol. 7041, 704109, (2008) · 0277-786X/08/$18 · doi: 10.1117/12.794618

Proc. of SPIE Vol. 7041 704109-1 2008 SPIE Digital Library -- Subscriber Archive Copy

Samples were prepared for cross sectional Transmission Electron Microscopy (TEM) by mechanical polishing using a tripod polisher, followed by Ar ion beam thinning to electron transparency. Microstructural observations were carried out in a JEOL 2010 TEM operating at 200kV. Electron Energy Loss Spectroscopy (EELS) was performed using a Gatan Imaging Filter (GIF-2000). The compositions of the coatings were investigated using X-ray Photoelectron Spectroscopy (XPS) and Auger Electron Spectroscopy (AES) depth profiling. Both these techniques were performed on a VG Microlab 310F. An Al unmonochromated X-ray source operating at a power of 300W and a 15kV excitation voltage was used for the XPS analysis. The sample was tilted such that the electron analyser normal to the sample surface collected the escaping electrons. The analysed area for XPS was approximately rectangular and 5 x 1 mm2. Al2O3 and TiO2 powder standards were used to calibrate the sensitivity factors prior to the measurements. In the case of the AES analysis, elemental depth profiles were obtained using 3kV Ar sputtering. The experimental reflectance of the deposited coating was measured using Ocean optics software and equipment. Optical characterization of the films was performed using variable angle spectroscopic ellipsometry (VASE) acquired by a Sopra-GESP5 (UV-VIS) ellipsometer.

ED EDEL

Primary flow

Secondary flow

___N___ Substrate holder

EEEEE Heating elements Fig.1. A schematic of the Microchemistry ALCVDTM-F120 ALD system is shown. The N2 primary flow carries the precursor chemical to the reaction zone where the species reacts with the surface of the substrates before the excess and the physisorbed species are taken away from the reaction zone during the purging cycle. The temperature was maintained throughout the deposition process.

3. RESULTS AND DISCUSSION 3.1 Composition and optical properties of single layer coatings Single layer films were prepared in order to evaluate deposition rate and film composition and optical properties. X-ray Photoelectron Spectroscopy (XPS) was used to determine the stoichiometries of the materials produced from the different precursors (see Table 1). The TiCl4 and AlCl3 precursors were found to produce stoichometric TiO2 and Al2O3 at 120ºC. XPS was also used to determine whether there was any chlorine present in the films. The amount of chlorine was found to be less than 1% for both chlorine based precursors. Alumina films prepared using a Al(CH3)3 precursor were also found to be stoichometric. The growth rate per cycle for the alumina and titania coatings are shown in Table 1 and are consistent with those published previously. Table 1. The composition of single layer titania and alumina films prepared by ALD from the precursors shown and their reported growth rates as well as the growth rates found experimentally at 120ºC. Layer

Precursors

Alumina

Al(CH3)3+H2O

Atomic Percentage (%)

Growth per cycle (Å/cycle)

Al

Ti

O

Reported elsewhere

Experimental

40

0

60

1.1-1.3 [1] (180ºC)

0.8 (120ºC)

Proc. of SPIE Vol. 7041 704109-2

Layer

Titania

Atomic Percentage (%)

Precursors

Growth per cycle (Å/cycle)

Al

Ti

O

Reported elsewhere

Experimental

AlCl3+H2O

40

0

60

0.8 [2] (150ºC)

1.2 (120ºC)

TiCl4+H2O

0

34

66

0.5 [3] (150ºC)

0.5 (120ºC)

VASE was used to analyse the optical characteristics of the single layers of Al2O3 and TiO2 that were deposited using Al(CH3)3, AlCl3 and TiCl4 precursors. A single Tauc-Lorentz dispersion function [7] was used to model the acquired ellipsometric angle Ψ and the phase difference ∆. An example of the experimental data along with the Tauc-Lorentz fit is shown in Figure 2(a). Figures 2(b), (c) and (d) show the refractive index and absorption coefficient functions determined using this fitting procedure. The materials all fall within the reported refractive index ranges of Al2O3 and TiO2 deposited by various methods [8-10]. The absorption coefficient at low wavelengths is substantial in the case of TiO2, which is due to the fact that TiO2 is a lower band-gap material than Al2O3 and so a much larger proportion of the higher energy light is absorbed by the material. There was no substantial difference found between the refractive indices of Al2O3 deposited by using the two different precursors.

>21IIIIIIIIIIIIIIIIIITJ LdIIY)

2.8 2.6

::

300

400

500

600

700

800

Wavelength (nm)

(a)

(b)

1.83 IE-5

1.82 \ —.—n

1.81

lEG

IE-7

-.--......

1.77

lEG

1.76 1.75 300

400

500

600

700

800

300

400

500

600

Wavelength (nm)

wsgth{!}

(c)

(d)

700

000

Fig. 2. (a) Example Tauc-Lorentz function fitting to a single layer of a TiO2 film. The resulting refractive indices n and absorption coefficients k of (b) TiO2, (c) Al2O3 (AlCl3 precursor) and (d) Al2O3 (Al(CH3)3 precursor) are also shown


3.2 Characterisation of a bi-layer TiO2/Al2O3 coating A bi-layer sample consisting of 32nm of titania and then 132nm of alumina on a silicon substrate was prepared using TiCl4 and Al(CH3)3 precursors. The AES depth profile from this sample is shown in Fig 3. The upper most alumina layer was found to be approximately uniform in composition with depth. Due to the nature of sputtering which results in the creation of an uneven crater, particularly at long sputtering times, the titania layer is not well resolved. In addition, the depth resolution of this technique is not sufficient to evaluate the nature of the interfaces between layers, which appear diffuse.

100

4.'.

Si }.' "v'

90 80

0

70 60

.

50 40

E

30 20 10 0

0

2000

4000

6000

8000

10000

Sputtering time (s)

Fig. 3. AES depth profile of a bi-layer sample consisting of 132 nm of alumina and 32 nm of titania on a silicon substrate.

A cross-sectional TEM image of this bi-layer sample is shown in Fig. 4. Sharp interfaces between the layers are observed. This is a feature of the ALD process which has been noted previously [5]. The selected area diffraction patterns of the TiO2 and Al2O3 are shown in Figures 4(b) and (c) respectively. The diffraction patterns from both layers consisted of rings which demonstrate that the layers are disordered. This is consistent with previous ALD studies that have shown that Al2O3 and TiO2 are amorphous for this temperature range [5,11-12]. Note that due to the limited spatial resolution available, the diffraction pattern from the TiO2 layer contains spots which originate from the crystalline Si substrate. The lack of large crystallites in the layers makes them highly suitable for optical applications since the presence of crystal facets can degrade optical performance. EELS was also performed to examine the Electron Loss Near Edge Structure (ELNES) of the Ti, Al and O edges. ELNES provides information on the local bonding arrangements and therefore can be used to distinguish between different materials. Figures 5(a) and (b) show the Al L edge and the O K edge from the alumina layer shown in Figure 4. These spectra are consistent with published standards [13] for crystalline alumina confirming that even though the ALD deposited material is amorphous, the local bonding arrangements resemble those of Al2O3. Figure 5(c) shows the Ti L and O K edges for the titania layer, which also agree with those obtained from crystalline standards.


(b) Silicon

(a)

(c)

Fig. 4. (a) Cross-sectional TEM image of a bi-layer sample consisting of 132 nm of alumina and 32 nm of titania on a silicon substrate. The selected area diffraction patterns of the (b) Al2O3 and (c) TiO2 layers. Note that the diffraction pattern of the TiO2 layer also includes diffraction spots from the crystalline Si substrate.

A1203A1 Ledge

Experimental EELS Atlas

(a)

60

80

100

Electron loss (eV)

120

140

(b)

500

550

600

650

700

Electron loss (eV)

Experimental EELS Atlas

0 K edge

I (c) Fig. 5. Electron loss near edge structure of the (a) Al L edge and (b) O K edge in the alumina layer and (c) the Ti L edge at 460 eV and the O K edge at 530 eV in the titania layer


3.3 Multilayer Optical Coating A TiO2/Al2O3/TiO2/Al2O3 antireflective coating on a Si substrate was designed using the FilmStar software package [14]. Figure 6(a) shows the design with layer thicknesses optimised to give the minimum average reflectance on Si, from 450 to 750nm at perpendicular incidence. Figure 6(b) shows the corresponding optical response as predicted by FilmStar.

60

50

Layer I)

11 mm

ALl)3

40

La\er C

l9nm

Ti(I)2

_____________

a) C-)

30 C-)

Layer B

Layer A

ll9nm

ALL)3

53nrn

Ti(I)2

Si substrate

a) a)

20

10

0

400

500

600

700

800

Wavelength (nm)

(a) (b) Fig. 6. (a) Diagram of the 4 layer antireflection design and (b) the simulated optical response of the 4 layer coating at perpendicular incidence.

An attempt was then made to prepare this coating using ALD and AlCl3 and TiCl4 precursors for the alumina and titania layers. The AES depth profile of this sample is shown in Figure 7. The composition of the alumina layers appears uniform and at the correct stoichiometry. As was the case in the bi-layer sample, the TiO2 layers are not well resolved due to sputter broadening. This broadening effect is most pronounced in the case of the initial TiO2 layer which was deposited onto the silicon substrate which appears to show mixing into both the silicon substrate and neighbouring alumina layer. The 4 layer coating is much thicker than the bi-layer coating shown in Figure 3 resulting in an even greater washing out of the layers.


.

**1$w4w44$øh 40

E

30 20 10 0 0

20000

40000

60000

Sputtering time (s)

Fig. 7. AES depth profile of the 4 layer coating on a silicon substrate

Figure 8(a) shows a cross-sectional TEM image of the 4 layer coating. The four layers (A to D) expected from the design shown in Figure 6(a) are clearly visible. A thin (approximately 15nm) surface contamination layer was also observed in the sample, most likely introduced during specimen preparation. No evidence of inter mixing between layers is evident. The image and an EELS line scan, which was used to measure the presence of Ti as a function of thickness, show relatively sharp interfaces between layers. The alumina layers were measured to be 125nm for layer B and 112nm for layer D, in excellent agreement with those expected (119nm and 111nm – see Figure 6(a)). The titania layers were measured to be 94nm for layer A and 28nm for layer C, much thicker than was intended. One possible reason for the discrepancy in the expected and the resultant titania layer thicknesses may be due to the anatase inclusions into the predominantly amorphous layer which in turn caused the differences in the growth rates of the titania. This suspicion is supported by a study previously reported by Sammelselg et al. [15] which also showed nonhomogeneous growth of titania at 150ºC due to the formation of the anatase phase. However, no evidence of anatase inclusions was observed in the TEM or EELS analysis. The more likely reason for the discrepancy in the titania thicknesses may be due to the procedure used for depositing the multilayer coating shown in Figure 6(a). The F-120 ALD reactor is primarily designed for operation with two precursors such that each is introduced into the deposition chamber through separate vapour transfer lines. The use of three precursors for the present multilayer coating required the TiCl4 and AlCl3 to be introduced through the same set of interconnected gas flow lines. We suspect that this approach has perturbed the pressure balance within the system, causing the observed increase in the deposition rate of TiO2. Further studies, aimed at clarifying and overcoming these limitations, are presently underway and will form the subject of a future report. While analyzing the sample using electron microscopy, it was found that electron beam irradiation encourages crystallization within both the titania and alumina layers. Figures 8(c) and (d) show bright and dark field images of the sample following irradiation in the electron beam for 30 minutes. Crystallites between 25-35nm in size were observed in the alumina layers while smaller (6-12nm) crystallites were observed in the titania layers. This result indicates that the material does not have a high degree of thermal stability under electron bombardment. Further work will employ a heating specimen holder to explore the crystallization process as a function of substrate temperature. Figure 9 shows the reflectance of the 4 layer coating compared with that calculated using the FilmStar software package [14] and layer thicknesses measured from the TEM image in Figure 8(a). There is reasonable agreement between experiment and theory. The increase in reflectance at approximately 450 and 650nm compared to that which was calculated in Fig. 6(b) for the optimised design is due to the deposition of thicker titania layers.


lOOnm

Ti Intensity (arbitary scale)

A BCD

0

50

100

150

200

250

300

350

Distance from surface (nm)

(a) (b)

lOOnm (c)

(d)

Fig. 8 (a) Cross-sectional TEM image of the 4 layer antireflection coatings prepared using ALD at 120oC. Layers A and C are titania, while B and D are alumina. (b) EELS line scan showing the level of Ti as a function of distance from the sample surface. (c) and (d) show bright field and dark field images of a section of the sample which has been irradiated by the electron beam. Nano-crystals of TiO2 and Al2O3 have formed in the layers.


400

60— 55— 50—

F — - Experi'iie'itaI I

Si'iit'Iatio'i

40—

35U

t•

3025—

20 _ I

-

N

15:] 400

500

600

700

600

Wavelength (nm) Fig. 9. The reflectance measured at 20 degrees for the 4 layer multilayer coatings prepared using ALD. For comparison, also shown is the simulated reflectance at 20 degrees for a 4 layer coating with layer thicknesses measured from TEM.

4. CONCLUSION In this work, a study of the microstructure and the optical properties of aluminium and titanium oxides deposited using ALD at 120ºC has been made. The depositions were carried out using TiCl4, AlCl3, Al(CH3)3 and H2O as precursors. XPS was used to verify the stoichiometry and AES depth profiling was performed to check the compositional uniformity of the deposited materials. TEM and EELS were used to investigate the microstructure of the coatings, which were found to be amorphous and contained local bonding configurations consistent with titania and alumina. The optical properties of the Al2O3 and TiO2 were characterized using VASE and a comparison was made between the alumina films deposited using the different precursors. The optical properties of the ALD deposited material was found to match well with published data for Al2O3 and TiO2. A multilayer antireflective silicon coating design was made and attempted to be deposited. An in-depth TEM study of the multilayer coating revealed that while the alumina layers were the desired thicknesses, the titania layers were thicker than expected possibly due to the perturbation of the pressure balances within the system. The multilayer coating was also found to be thermally unstable under electron irradiation as was evident from nano-crystal formation in both the alumina and titiania layers during observation in the TEM. The calculated reflectance of the multilayer coating using the thicknesses found from the TEM analysis compared well to that measured experimentally.

ACKNOWLEDGMENTS

The authors would like to thank Ju L. Peng for assistance in the TEM work and the Defence Science and Technology Organisation (DSTO) for making this research possible through their funding as well as the Australian Institute of Nuclear Science and Engineering (AINSE) and the Australian Nuclear Science and Technology Organisation (ANSTO) for providing financial assistance (Award No. 4516) and the ALD facilities.


REFERENCES [1] [2] [3] [4] [5]

[6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

Puurunen, R. L., “Surface chemistry of atomic layer deposition: A case study for the trimethylaluminum/water process”, J. App. Phys. 97, 121301 (2005) Räisänen, P. I., Ritala, M. and Leskelä, M., “Atomic layer deposition of Al2O3 films using AlCl3 and Al(OiPr)3 as precursors”, J. Mater. Chem. 12, 1415-1418 (2002) Aarik, J., Aidla, A., Mändar, H. and Uustare, T., “Atomic layer deposition of titanium dioxide from TiCl4 and H2O: investigation of growth mechanism”, App. Surf. Sci. 172, 148-158 (2001) Triani, G., Evans, P. J., Mitchell, D. R. G., Attard, D. J., “Atomic layer deposition of TiO2 / Al2O3 films for optical applications”, Proc. SPIE 5870, 587009 (2005) Mitchell, D. R. G., Attard, D. J., Finnie, K. S., Triani, G., Barbé, C. J., Depagne, C. and Bartlett, J. R., “TEM and ellipsometry studies of nanolaminate oxide films prepared using atomic layer deposition”, App. Surf. Sci. 243, 265277 (2005) Heyman, A. and Musgrave, C. B., “A quantum chemical study of the atomic layer deposition of Al2O3 using AlCl3 and H2O as precursors”, J. Phys. Chem. B 108, 5718-5725 (2004) Jellison Jr., G. E., Modine, F. A., “Parameterization of the optical functions of amorphous materials in the interband region”, App. Phys. Lett. 69, 371 (1996); 69, 2137 (1996) Mardare, D. and Hones, P., “Optical dispersion analysis of TiO2 thin films based on variable-angle spectroscopic ellipsometry measurements”, Mat. Sci. and Eng. B68, 42-47 (1999) Lichtenstein, T., Handbook of thin film materials, Rochester, New York (1979) Franta, D., Ohlidal, I., Petrýdes, D., “Optical characterization of TiO2 thin films by the combined method of spectroscopic ellipsometry and spectroscopic photometry”, Vacuum 80, 159-162 (2005) Ritala, M., Saloniemi, H., Leskelä, M., Prohaska, T., Friedbacher, G., Grasserbauer, M., “Studies on the morphology of Al2O3 thin films grown by atomic layer epitaxy”, Thin Solid Films 286, 54-58 (1996) Aarik, J., Aidla, A., Kiisler, A., Uustare, T. and Sammelselg, V., “Effect of crystal structure on optical properties of TiO2 films grown by atomic layer deposition”, Thin Solid Films 305, 270-273 (1997) Ahn, C. C. and Krivanek, O.L., EELS Atlas, Gatan Inc., (1983) FTG Software Associates, FilmStar, © 1991-2008 Sammelselg, V., Rosental, A., Tarre, A., Niinistö, L., Heiskanen, K., Ilmonen, K., Johansson, L.-S., Uustare, T., “TiO2 thin films by atomic layer deposition: a case of uneven growth at low temperature”, App. Surf. Sci. 134, 7886 (1998)

Proc. of SPIE Vol. 7041 704109-10