Download Optical properties of Al2O3 thin films grown by

Optical properties of Al2O3 thin films
grown by atomic layer deposition
Pradeep Kumar,1,* Monika K. Wiedmann,2 Charles H. Winter,2 and Ivan Avrutsky1
1
Electrical and Computer Engineering, 5050 Anthony Wayne Drive, Wayne State University,
Detroit, Michigan 48202, USA
2
Department of Chemistry, 5101 Cass Avenue, Wayne State University,
Detroit, Michigan 48202, USA
*Corresponding author: [email protected]
Received 18 May 2009; revised 14 August 2009; accepted 11 September 2009;
posted 15 September 2009 (Doc. ID 109791); published 25 September 2009
We employed the atomic layer deposition technique to grow Al2 O3 films with nominal thicknesses of 400,
300, and 200 nm on silicon and soda lime glass substrates. The optical properties of the films were investigated by measuring reflection spectra in the 400–1800 nm wavelength range, followed by numerical
fitting assuming the Sellmeier formula for the refractive index of Al2 O3 . The films grown on glass substrates possess higher refractive indices as compared to the films on silicon. Optical waveguiding is
demonstrated, confirming the feasibility of high-index contrast planar waveguides fabricated by atomic
layer deposition. © 2009 Optical Society of America
OCIS codes:
310.6860, 310.2790, 160.4760, 160.4670.
1. Introduction
Finding a material platform for low-loss, high-index
contrast optical waveguides has been an important
task in integrated optics since its inception. Integrated optical devices and systems are applied in optical communication and optical sensors, as well as
emerging applications, such as on-chip optical clock
distribution and optical chip-to-chip interconnects
[1]. These applications require strong light confinement to a waveguide core, which can be achieved
using a large index step between the core and the
claddings. For telecommunication devices, strong
confinement is needed to achieve a high degree of integration and allow for sharp bends of channel waveguides. For sensors, it is important to have a strong
field at the waveguide surface to attain high sensitivity. Many materials and thin-film manufacturing
technologies have been tried, but the quest is far
from over. Among all-dielectric waveguide systems,
the strongest confinement is found in silicon-on0003-6935/09/285407-06$15.00/0
© 2009 Optical Society of America
insulator (SOI) waveguides [2], which have a silicon
core (refractive index nSi ¼ 3:48) and silica
(nSiO2 ¼ 1:45), or even air, claddings. The SOI waveguides so far seem to be an ideal technical solution
for planar photonic integrated circuits. Besides perfect optical characteristics, this material system is
absolutely compatible with the complementary metal-oxide-semiconductor (CMOS) [3] fabrication process in microelectronics. However, because of the
method by which the SOI wafers are manufactured
(wafer bonding followed by cleaving), it appears to be
difficult to implement a more sophisticated photonic
circuitry [4] in SOI systems in which, for example,
multiple optical layers are required. The architecture of some integrated optical devices, such as large
optical switches, may require waveguide crossovers
with negligible cross talk, which, again, is hardly realizable in a single-layer photonic circuit. Also, SOI
has limited applicability in optical sensors, since silicon is not transparent in the visible range. Another
CMOS-compatible material system for optical waveguides is silicon oxynitride [5], which is a mixture of
silicon oxide and silicon nitride typically manufactured by plasma-enhanced chemical vapor
1 October 2009 / Vol. 48, No. 28 / APPLIED OPTICS
5407
deposition. With a high nitrogen content in the film,
the refractive index may reach nSi3N4 ¼ 2:2. High nitrogen content films, however, tend to show high optical loss. With low nitrogen content, the index step is
reduced and strong light confinement is no longer
possible. Magnetron deposition [6] and various sputtering techniques [7,8] can also be used for thin-film
manufacturing. They are successfully used in multilayer dielectric mirrors, but as far as high-index optical waveguides are concerned, the roughness of the
film surface is often too large, which leads to strong
light scattering. Among non-CMOS technologies, solgel [9] and ion exchange [10] have been employed.
Ion exchange is probably one of the oldest and
best-optimized waveguide manufacturing technologies. The index step in ion-exchange waveguides is
small. Sol-gel waveguides generally show good performance with large index steps in systems such
as TiO2 =SiO2, but incompatibility with CMOS limits
applications of this technology.
In the present work, we study the optical properties of alumina films prepared by atomic layer deposition (ALD) [11–15]. Because of the self-limiting
nature of ALD growth, the films typically show perfect thickness uniformity across the entire wafer.
With carefully prepared substrates, the top surface
of the film is expected to be practically atomically
flat. This eliminates, or at least greatly reduces, light
scattering at surface imperfections—the major
source of optical losses in thin-film waveguides.
ALD, compared to other thin-film fabrication techniques, has low growth rates, since film growth occurs
through the sequential pulsing of two or more different precursor gases, separated by inert gas purges.
The ALD mechanism ensures that growth is self-limited, resulting in smooth, uniform films. However,
the exceptional quality of the ALD films makes this
technology quite promising for fabrication of integrated optical devices and systems. Here we present
data on the refractive index of ALD alumina films in
the spectral range from 400 to 1800 nm deduced from
the reflectance spectra, and demonstrate planar optical waveguides.
25 mm and the optical properties of the Al2 O3 films
were investigated by measuring the reflection spectra at five different positions on each sample 5 mm
apart from each other. MATLAB curve fitting tools
were used for analysis of the reflection spectrum
data. The experimentally measured reflection spectra were numerically fit to the expected calculated
spectra [16], assuming the Sellmeier formula for alumina. The fitting results provided the thickness and
refractive index of the film in the entire spectral
range of interest. Using the Lavenberg–Marquardt
algorithm, the experimental data were analyzed by
minimizing the mean-square differences between
the measured reflection spectra and the calculated
spectra, which are generated from the fitting model
at the corresponding wavelength, angle of incidence,
and substrate and cover refractive indices at corresponding wavelengths. The refractive indices for
silicon and soda lime glass were taken from
Refs. [17,18], respectively.
The fitting algorithm also included adjustable factors to account for slow drift of the intensity of the
light source used in the measurements. Reflection
spectra were taken from the range of 400 to
1800 nm, so, for this wide range of wavelength, the
light source has some drift in intensity, which introduces error in the reflection spectra since the effect of
drift of light source depends upon the wavelength.
So, including the effect of drift of light source, we
used a first-order correction in the calculated reflection spectrum:
2. Experimental and Theoretical Details
3.
Optical Properties
A.
Refractive Index
Al2 O3 films were grown on silicon (100) and soda
lime substrates in a Picosun R-75 ALD reactor using
trimethylaluminum (TMA) and water [15]. The deposition temperature was 300 °C. Films with nominal thicknesses of 400, 300, and 200 nm were
grown with 4400, 3300, and 2200 deposition cycles,
respectively. Each cycle consisted of a 0:1 s TMA
pulse, followed by a 3:0 s nitrogen purge, a 0:1 s water
pulse, and, finally, another 3:0 s nitrogen purge.
The reflection spectra have been measured using a
DigiKrom 240 monochromator (Spectral Product)
and a halogen lamp as a light source. Fused quartz
was used for normalizing intensity, since the refractive index of fused quartz is well known. The incident
angle was 5:6° and the beam size was 0:5 mm. The
substrate dimensions were approximately 39 mm ×
5408
APPLIED OPTICS / Vol. 48, No. 28 / 1 October 2009
Rexp ¼ RCalu ðC0 þ C1 ðλ − C2 ÞÞ;
ð1Þ
where C0 and C1 are adjustable parameters, λ is the
wavelength, the value of C2 is 100 nm, which includes the effect of short wavelength, and RCalu is
the calculated reflection spectrum.
The fitting parameters were Sellmeier coefficients,
film thickness, C0 , and C1 . The fits were carried out
for each position for each sample and, then, for the
refractive index of alumina, the results from the fitting were averaged for the same substrate.
The films on silicon substrates possess a very highindex contrast and, thus, show a very strong interference in the reflection spectra. As long as the refractive index of the alumina film is much lower than
that of the silicon substrate, the reflection spectrum
minima contain information about the refractive index and film thickness [16], while the spectrum maxima are defined by the optical properties of the
silicon substrate. The reflection minima are more
clearly pronounced in the logarithmic scale. Therefore, to obtain more accurate information about
the refractive index and thickness of the Al2 O3 film
on the silicon substrate, the logarithm of the reflection spectrum was fitted to the logarithm of the
Fig. 1. (Color online) Experimental and calculated reflection
spectra for Al2 O3 films on (a) silicon and (b) soda lime glass.
calculated reflection spectrum. The error analysis
of fitting justifies the choice of the logarithmic scale
for structures of this type—low-index film on a
high-index substrate. Figures 1(a) and 1(b) show
the measured and calculated reflection spectra of
Al2 O3 films on silicon and soda lime glass substrates,
respectively.
The quality of the fit was investigated in detail by
the mean-squared error (MSE) [14], defined as
N i
Rexp − Ricalc 2
1X
MSE ¼
;
N i¼0
σi
ð2Þ
where N is the total number of measurements and σ i
is the standard deviation of experimental data
points. The MSE values for Al2 O3 films on silicon
and soda lime glass are 0.0006 and 0.0070, respectively. The positions of maxima and minima in the
reflection spectrum depend on the product of film refractive index and film thickness, and there is strong
anticorrelation between them. By introducing the
scaling factor (Sn) in refractive index and film thickness, we construct the contour plot for MSE. The
scaling factor (Sn) introduced a controllable error
Fig. 2. (Color online) MSE variation with respect to the scaling
factor for refractive index and thickness of Al2 O3 films on (a) silicon
and (b) soda lime glass.
in refractive index and film thickness to calculate
and contour plot the MSE. Figure 2 shows the contour plot for MSE with respect to the scaling factors
for the refractive index and film thickness, and the
film refractive index is equal to the product of scaling
factor (Sn) and film refractive index (nf ) given by
Table 1. The error plot is constructed in such a
way that the minimum error is achieved when both
scaling factors are equal to unity for the film thickness and its refractive index. The contour closest to
the minimum corresponds to doubling the fitting error. An order of magnitude with larger error in the
case of a film on a soda lime glass substrate is due
to the substantially smaller index contrast and, consequently, much weaker interference in the reflection
Table 1.
Silicon
Soda lime glass
Sellmeier Coefficients for Al2 O3 Films
A0
A1
λ1 (nm)
1.31838
1.33247
1.32052
1.40684
121.594
119.992
1 October 2009 / Vol. 48, No. 28 / APPLIED OPTICS
5409
Fig. 3. (Color online) Refractive index of Al2 O3 films on silicon
and soda lime glass.
spectra. The fitting results are excellent in comparison to those of Kim et al. [14] and Toyoda et al. [19].
In the numerical fit of the reflection spectra, the
film refractive index was defined according to the
Sellmeier-like dispersion formula:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
A λ2
nf ðλÞ ¼ A0 þ 2 1 2 ;
λ − λ1
ð3Þ
where the A0 term accounts for the contribution of
short wavelength absorption to the refractive index
at a longer wavelength, while A1 and λ1 are the relative oscillator strength and the resonant wavelength, respectively. Table 1 gives the values of A0,
A1 , and λ1 for Al2 O3 films on silicon and soda lime
glass deduced from analysis of the reflection spectra
shown in Fig. 1.
Figure 3 shows the dispersion curves of the refractive index of Al2 O3 films on silicon and soda lime
glass in the wavelength range from 400 to 1800 nm.
The refractive indices at 633 nm were 1.64 and 1.67
for Al2 O3 films on silicon and soda lime glass, respectively. The refractive indices are in good agreement
with those of Räisänen et al. [20] and Ott et al. [21]
and higher than the 1.62–1.64 reported by Kim et al.
[14]. The silicon and soda lime glass substrates are
commonly studied materials, and the refractive in-
Table 2.
Substrate
Silicon
Soda lime glass
5410
dex is known by a factor of 0:001. So, to study
the effect of this error on film refractive index, we introduced 0:001 error into the refractive indices of
silicon and soda lime glass substrates. The shaded
area in Fig. 3 shows the effect of discrepancy of substrate refractive index on film refractive index.
The error introduced by the substrate refractive index in film refractive index was 0:001 for silicon
substrate and 0:003 for soda lime glass substrate.
The refractive index of the Al2 O3 films on silicon compared to that of films on soda lime glass is apparently
lower, by about 0.03, which is much higher than the
error introduced by the discrepancy of substrate refractive index. Since Si wafers typically have a
2–3 nm layer of native SiO2 [22], this may contribute
to the reduced overall index of the film as it is extracted from the reflection spectra. The observed difference, however, is about an order of magnitude
larger than what may be caused by the existence
of the native oxide, and it is about the same for
the film thickness from 200 to 400 nm. The different
optical properties, already mentioned in Ref. [14], apparently indicate somewhat different film densities
on different substrates.
B.
Film Uniformity
The thickness of the Al2 O3 films was measured at 10
different points, which were 2:5 mm apart from each
other on each sample. The results show that the
film thickness is uniform over the entire substrate.
Table 2 provides the range of film thickness, mean
film thickness, and the standard deviation obtained
from fitting. Figure 4 shows the thicknesses of Al2 O3
films on silicon and soda lime glass, which are close
to the nominal values of 400, 300, and 200 nm.
4.
Thin-Film Waveguide
The waveguide properties were investigated for the
Al2 O3 films on soda lime glass. For the theoretical
calculation, the effective refractive index [23] for different modes was calculated for all three samples
and the refractive index and film thickness were taken from the fitting results. For experimental verification, the prism coupling method [24–26] was used
for measuring the effective refractive index. The effective refractive index for each mode is directly related to the angle of incidence at which incidence
light was guided through the waveguide. The effective refractive index is defined as
Range of Film Thickness, Mean Thickness, and Standard Deviation Obtained from Fit
Experimental Film
Thickness (nm)
Range of Film Thickness (nm)
Obtained from Fit
Mean Thickness
(nm)
Standard
Deviation
400
300
200
400
300
200
415–403
312–304
204–199
402–377
302–279
192–178
410.56
307.58
201.82
392.63
296.54
181.87
6.41
6.29
4.31
3.65
2.03
1.43
APPLIED OPTICS / Vol. 48, No. 28 / 1 October 2009
Table 3.
Experimental and Calculated Effective Refractive Index for
Al2 O3 Films on Soda Lime Glass at 632:8 nm
Effective Refractive Index (neff )
Experimental
Fig. 4. Thicknesses of Al2 O3 films on silicon and soda lime glass.
sinðθm Þ
neff ¼ nP · sin φ þ arcsin
;
nP
ð4Þ
where nP is the prism refractive index, φ is the base
prism angle, and θm is the angle at which incidence
light is guided through the waveguide. Figure 5(a)
shows the schematic diagram of our experiment.
Calculated
Film Thickness (nm)
TE-pol.
TM-pol.
TE-pol.
TM-pol.
400
300
200
1.5765
1.5474
1.5159
1.5614
1.5289
-
1.5909
1.5700
1.5270
1.5745
1.5477
-
A He–Ne laser emitting at a wavelength of
632:8 nm and a prism with a vertex angle of
60:19° and a refractive index of 1.7520 were used.
Figure 5(b) shows the guided mode for the 400 and
300 nm thick Al2 O3 films on soda lime glass. Table 3
gives the experimental and calculated effective refractive indexes for the Al2 O3 films on soda lime
glass. At 632:8 nm, the 400 and 300 nm thick films
support fundamental TE- and TM-polarized modes
and 200 nm thick films support only the fundamental
TE-polarized mode.
For characterization of the waveguide losses, the
intensity of the guided mode tracks was measured as
a function of the distance. The losses in waveguide
are defined as
I ¼ I 0 expð−αxÞ;
ð5Þ
where α is defined as losses in waveguide. To calculate losses in db=cm,
αdb ¼ ð10=LÞlog10 ðαÞ;
ð6Þ
where L is length in cm travel by the guided light in
waveguide. Figure 6 shows the natural log of intensity as a function of distance (millimeters). The
slopes of the intensity curves correspond to the losses
in the range from 7 to 11 dB=cm.
Fig. 5. (a) Schematic diagram of experiment and (b) mode excitation for Al2 O3 films on soda lime glass at 632:8 nm.
Fig. 6. (Color online) Loss in Al2 O3 film waveguides at 632:8 nm.
1 October 2009 / Vol. 48, No. 28 / APPLIED OPTICS
5411
5. Conclusion
In summary, the ALD technique was employed to
fabricate Al2 O3 films on silicon and soda lime glass
substrates. The refractive index of the films was
found by fitting the reflection spectra measured in
the 400–1800 nm spectral range and assuming the
Sellmeier formula with MSE values of 0.0006 and
0.0070 for Al2 O3 films on silicon and soda lime glass,
respectively. The refractive index of Al2 O3 films on
soda lime glass was found to be higher than that
in the case of the silicon substrate. High-index contrast optical waveguides are demonstrated.
P. K. and I. A. acknowledge support from the National Science Foundation (NSF) through a subcontract in grant IIP-0741171. M. K. W. and C. H. W.
acknowledge support from the Army Research Office,
grants W911NF-07-0489 and W911NF-06-1-0102.
References
1. M. Haurylau, G. Chen, H. Chen, J. Zhang, N. A. Nelson,
D. H. Albonesi, E. G. Friedman, and P. M. Fauchet, “On-chip
optical interconnect roadmap: challenges and critical directions,” IEEE J. Sel. Top. Quantum Electron. 12, 1699–1705
(2006).
2. M. Galli, D. Gerace, A. Politi, M. Liscidini, M. Patrini,
L. C. Andreani, A. Canino, M. Miritello, R. L. Savio, A. Irrera,
and F. Priolo, “Direct evidence of light confinement and emission enhancement in active silicon-on-insulator slot waveguides,” Appl. Phys. Lett. 89, 241114 (2006).
3. Y. Vlasov and S. McNab, “Losses in single-mode silicon-oninsulator strip waveguides and bends,” Opt. Express 12,
1622–1631 (2004).
4. D. J. Lockwood and L. Pavesi, “Silicon fundamentals for photonics applications,” in Silicon Photonics, Vol. 94 of Top. Appl.
Phys. (Springer-Verlag, 2004), pp. 1–50.
5. K. Worhoff, P. V. Lambeck, and A. Driessen, “Design, tolerance
analysis, and fabrication of silicon oxynitride based planar optical waveguides for communication devices,” J. Lightwave
Technol. 17, 1401–1407 (1999).
6. P. J. Kelly and R. D. Arnell, “Magnetron sputtering: a review of
recent developments and applications,” Vacuum 56, 159–172
(2000).
7. A. T. Hunt, W. B. Carter, and J. K. Cochran, Jr., “Combustion
chemical vapor deposition: a novel thin-film deposition technique, “Appl. Phys. Lett. 63, 266–268 (1993).
8. R. K. Singh and J. Narayan, “Pulsed-laser evaporation technique for deposition of thin films: Physics and theoretical model,” Phys. Rev. B 41, 8843–8859 (1990).
9. S. I. Najafi, T. Touam, R. Sara, M. P. Andrews, and
M. A. Fardad, “Sol-gel glass waveguide and grating on silicon,”
J. Lightwave Technol. 16, 1640–1646 (1998).
5412
APPLIED OPTICS / Vol. 48, No. 28 / 1 October 2009
10. G. L. Yip and J. Albert, “Characterization of planar optical
waveguides by Kþ-ion exchange in glass,” Opt. Lett. 10,
151–153 (1985).
11. E. Langereis, J. Keijmel, M. C. M. van de Sanden, and
W. M. M. Kessels, “Surface chemistry of plasma-assisted
atomic layer deposition of AlO studied by infrared spectroscopy,” Appl. Phys. Lett. 92, 231904 (2008).
12. S. B. S. Heil, P. Kudlacek, E. Langereis, R. Engeln,
M. C. M. van de Sanden, and W. M. M. Kessels, “In situ
reaction mechanism studies of plasma-assisted atomic
layer deposition of AlO,” Appl. Phys. Lett. 89, 131505
(2006).
13. M. Thitsa, Y. Song, and S. Albin, “Effect of oxidation, etching,
and thin-film deposition on silicon photonic crystals,” J. Electrochem. Soc. 155, H351–H356 (2008).
14. Y. Kim, S. M. Lee, C. S. Park, S. I. Lee, and M. Y. Lee, “Substrate dependence on the optical properties of Al2 O3 films
grown by atomic layer deposition,” Appl. Phys. Lett. 71,
3604–3606 (1997).
15. R. L. Puurunen, “Surface chemistry of atomic layer deposition:
a case study for the trimethylaluminum/water process,” J.
Appl. Phys. 97, 121301 (2005).
16. O. S. Heavens, Optical Properties of Thin Solid Films (Dover, 1991).
17. E. Palik, Handbook of Optical Constants of Solids (Academic, 1985).
18. M. Rubin, “Optical constants and bulk optical properties of
soda lime silica glasses for windows,” Sol. Energy Mater.
12, 275–288 (1985).
19. K. Toyoda, H. Kumagai, M. Matsumoto, and M. Obara,
“Comparison study of Al2 O3 optical crystalline thin films
grown by vapor combination of AlðCH3 Þ3 =N2 O and
AlðCH3 Þ3 =H2 O2 ,” Jpn. J. Appl. Phys. 32, 6137–6140
(1993).
20. P. I. Räisänen, M. Ritala, and M. Leskelä, “Atomic layer
deposition of Al2 O3 films using AlCl3 and AlðOiPrÞ3 as precursors,” J. Mater. Chem. 12, 1415–1418 (2002).
21. A. W. Ott, J. W. Klaus, J. M. Johnson, and S. M. George, “Al3 O3
thin film growth on Si (100) using binary reaction sequence
chemistry,” Thin Solid Films 292, 135–144 (1997).
22. M. Ritala and M. Leskelä, in Handbook of Thin Film
Materials, H. S. Nalwa, ed. (Academic, 2002), Vol. 1,
pp. 103–159.
23. H. P. Urbach and G. Rikken, “Spontaneous emission from a
dielectric slab,” Phys. Rev. A 57, 3913–3930 (1998).
24. J. Seligson, “Prism couplers in guided-wave optics—design
considerations,” Appl. Opt. 26, 2609–2611 (1987).
25. P. K. Tien and R. Ulrich, “Theory of prism-film coupler and
thin-film light guides,” J. Opt. Soc. Am. 60, 1325–1337
(1970).
26. O. Hidalgo, Y. Berencen, and J. Rodriguez, “Prism coupling
characterization of planar optical waveguides made by silver
ion exchange in glass,” Phys. Status Solidi C 2, 3746
(2005).