Open Access
Add to BibSonomy
Abstract
We report on the growth and characterization of wavelength-thick indium tin oxide (ITO) films deposited using high power impulse magnetron sputtering (HiPIMS) with post deposition processing to achieve an epsilon near zero (ENZ) property at 1550 nm telecom wavelengths. The goal is to fabricate 1550 nm ENZ films for use as claddings for waveguides, resonators, or high-contrast metastructures in photonic devices operated at telecom wavelengths. We developed a HiPIMS growth and post-annealing process to improve on existing ENZ ITO quality and uniformity. By consecutively annealing the ITO film, the plasma frequency gradually shifts, enabling fine tuning of the ENZ wavelength regime from 1800 to 1500 nm. The films were characterized using spectroscopic ellipsometry, transmission electron microscopy, x-ray diffraction, and energy dispersive x-ray spectroscopy. Our micro-analyses shows that the change in the microstructure resulted in the change in the optical properties of the ITO. These findings allow us to control the ENZ property at the desired wavelength and reduce the absorption loss, which is beneficial for device application.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Metamaterials have been extensively studied for their well-known interactions with electromagnetic waves and their ability to be designed with specific ε and µ values. In photonic metamaterials, engineered subwavelength structures with localized resonating centers are either randomly distributed or periodically arranged in the material, enhancing the light-matter interaction. These metamaterials can be made using two types of formation mechanisms to create subwavelength resonators: self-assembled doped dielectric structures, and artificially fabricated, periodically distributed resonator structures. There has been growing interest in thin film epsilon-near-zero (ENZ) metamaterials allowing for unconventional tailoring and manipulation of the light-matter interaction, such as very large electro-refraction, enhanced nonlinearity, and supercoupling effects [1–4]. Thin films of both metal-dielectric multilayer structures and self-assembled doped dielectrics have been studied for their ENZ optical properties [5]. ENZ materials have numerous interesting properties, such as a vanishingly small phase change during wave transit [2], as well as strongly enhanced optical nonlinearity [3–4]. These properties could be exploited in new devices to enable sub-wavelength channels for light guiding and waveform shaping [6]. Our team is developing wavelength-thick ENZ films to provide a platform for new types of optoelectronic devices with enhanced properties and functionalities.
Indium tin oxide (ITO) is widely used in electronic and optoelectronic devices because of its transparent conducting property. Under certain processing conditions, ITO films can reach an ENZ regime (in which the value of both the real and imaginary parts of permittivity, ε1 and ε2, are within -0.5 and 0.5) at C-band telecommunication wavelengths (∼1550nm), which allow its application in many optoelectronic devices operating in telecom bands [1,5]. Wavelength-thick ITO films are desired for devices using ENZ ITO as a propagating medium or cladding, but there are few examples in the literature of ITO films thicker than one micron. Our group has previously worked with wavelength thick ITO with ENZ properties for wavelengths near 1550 nm [7]. These films were grown using DC sputtering, then annealed with high temperatures to obtain the ENZ property. However, there were substantial microstructural nonuniformities throughout the film. These crystallite morphology differences led to a change in the oxygen vacancy concentration, which is one of the primary factors in determining the plasma frequency, and therefore caused a large variation in the permittivity throughout the bulk of the film. The lack of uniformity in optical properties as well as high absorption loss limits the usefulness of these ENZ ITO films in photonic device applications.
ITO films have been made using several different deposition techniques such as electron beam evaporation [8,9], atomic layer deposition [10], RF magnetron sputtering [11], and pulsed laser deposition (PLD) [12,13]. In this work, we demonstrate wavelength thick (>1.5 µm) ITO films with improved film uniformity and a reasonably flat permittivity within the ENZ regime throughout the film by employing a different sputter deposition technique, high-power impulse magnetron sputtering (HiPIMS). HiPIMS utilizes very large voltage pulses for short durations to create a highly dense, ionized plasma. HiPIMS increases film density and smoothness [14] as well as improves the uniformity of conformal deposition on complex surfaces [14,15]. HiPIMS deposition of thin ITO films was previously demonstrated, and it was shown that altering parameters such as deposition power, pulse length, and gas pressure can greatly impact film properties [16].
In this work, we demonstrate how the HiPIMS process allowed us to improve the uniformity of the optical constants in thick ITO films when compared to DC sputtered films. We also demonstrate how the optical constants can be fine-tuned by post-deposition annealing which can change the charge carrier concentration of the films. ITO film conductivity and the plasma frequency are determined by many parameters, including oxygen vacancy concentration [17,18], crystal morphology [19,20], and post deposition annealing conditions [21–23]. Our goal is to show that HiPIMS, together with post-deposition annealing, leads to uniform, wavelength thick ITO films with a tunable ENZ property and lower absorption loss than DC sputtered films, making them suitable for novel optoelectronic and RF-photonic devices.
2. Methods
The 1.6 µm thick ITO films were deposited onto a Si wafer via HiPIMS using a 90/10 In2O3/SnO2 wt % ITO target from Kurt J. Lesker (KJL). The deposition chamber was a DC/RF magnetron sputtering system from KJL with an accompanying HiPIMS unit from Starfire Industries. The deposition was done with Ar gas flow at 3.5 mTorr pressure. The HiPIMS parameters were 15 µs pulse length at 700 Hz with a 250A current and 800W power limit. The dc voltage was held at 600V which led to a target power density of about 140W. This power was chosen to reduce the temperature of the target as well as to minimize arcing events during HiPIMS. These parameters together with the chamber geometry led to a deposition rate of around 5 nm/min. This wafer was then cleaved into different testing samples.
Post-deposition annealing was done in an Allwin21 Rapid Thermal Annealing tool. We conducted multiple annealing processes on individual test samples to study cumulative annealing time dependence effects. Eight rounds of the annealing process were carried out in a nitrogen environment with a flow rate of 5 L/min at 650°C. Annealing times ranged from 10 min to 20 min for a total cumulative time of 2 hrs, as shown in Fig. 1.
Fig. 1. a) Graph showing the dependence of real permittivity ε1 and b) imaginary permittivity ε2 as a function of wavelength. The color red indicates the slice of the film that is at the surface exposed to air and the color blue indicates the slice that is buried next to the Si substrate. The arrow shows the direction of increasing cumulative anneal time. It can be seen that ε1 = 0 trends toward lower wavelengths as the total anneal time increases. The shaded region indicates ENZ regime.
Ellipsometry was done on the annealed samples between each time step using a J.A. Woollam M-2000 variable angle spectroscopic ellipsometer (VASE). The accompanying Woollam analysis software CompleteEASE (v. 6.51) was used to model the VASE results using a Drude + Tauc-Lorentz (TL) model to determine the real and imaginary optical parameters, ε1 and ε2. This model is often used with ITO to describe absorption in the UV-Vis spectral range using the TL model as well as free carrier absorption in the IR range using the Drude model. The permittivity wider wavelength range is therefore:
with,We cannot assume that the permittivity is uniform throughout the thickness of the ITO film because TEM and SEM images show a changing microstructure throughout the bulk of the film as a function of thickness, and it is known that crystal morphology has a role in ITO film properties. Therefore, the ITO was modeled using a graded layer technique which sliced the film into thin, consecutive layers, and Drude + TL model was applied to each individual layer. This method was utilized to account for variations in the permittivity throughout the ITO film depth profile and allow us to characterize the uniformity of the annealed film. Furthermore, to be more confident of the uniqueness of the model fits, we eliminated the film thickness as an unknown parameter by measuring the film cross-section in the SEM and measured each sample at five angles [24].
A ZEISS Auriga scanning electron microscope (SEM), Panalytical MRD- X'pert Pro x-ray diffractometer (XRD), and a JEOL ARM200F TEM equipped with energy dispersive X-ray spectroscopy (EDS) system were used to analyze the film cross-sections.
3. Results
The effect of increased cumulative annealing times is shown in the plot of the ellipsometry model’s results (Fig. 1). The spectra for both the air-exposed top layer and the buried interface layer are shown. The red lines in Fig. 1(a) indicate the real permittivity ε1 in the top layer slice and the blue lines indicate the bottom layer slice, with the black arrow representing the direction of increasing cumulative anneal time. The ITO samples that were annealed for short durations had ε1 > 0 for the entire measurement range (<1600 nm) through the thickness of the films. The ε1 values across the spectrum decreased with increased annealing time. After 2 hours, the samples had ε1 = 0 at 1570 nm and 1500 at the top and bottom slices of the film, respectively. The plasma frequency shifted to lower wavelengths gradually with increasing cumulative anneal time and there were no abrupt changes anywhere within the bulk of the film after any of the consecutive annealing rounds. The ENZ regime are indicated by the shaded areas in both graphs of Fig. 1. Figure 1(b) shows the imaginary permittivity ε2. It can be seen that the top layer (red) increases while the bottom layer (blue) decreases, though both remain in the ENZ regime for all annealing times. The value of ε2 in the ENZ regime for the annealed HiPIMS thick films is 0.36-0.44, which is comparable to other ITO thin films in the literature grown by PLD that have an ENZ ε2 of 0.42 [13].
In our previous work, we used DC sputtering to make the ITO films. We found an abrupt shift in the permittivity partway through the film thickness after it was annealed [7]. The films made by HiPIMS in this work have significantly improved uniformity of ε1 and ε2, as seen in Fig. 2. Figure 2 shows the permittivity at 1550 nm wavelength as a function of depth into the film for each annealing step. There are slight variations throughout the bulk of the film, but the permittivity remains smooth and in the ENZ regime everywhere after 1.33 hrs of cumulative anneal time. In the prior DC sputtered films, the permittivity would drift out of the ENZ regime toward the top and bottom of the film and there was only a portion in the middle that was ENZ [7]. Using HiPIMS improved the uniformity as well as demonstrate the tunability of the permittivity.
Fig. 2. Graph showing ε1 and ε2 at 1550 nm wavelength as a function of distance from the Si-ITO interface. The arrow again indicates the direction of increasing cumulative anneal time. It can be seen that ε1 drops towards zero, while ε2 increases. The uniformity of ε1 remains fairly constant, while ε2 actually increases in uniformity.
Figure 3 compares cross-sectional SEM images of a HiPIMS ITO film and a DC ITO film prepared by cleaving the samples. The microanalyses of the DC ITO film showed some abrupt changes in morphology partway through the film. The microstructure of the DC ITO film changes from nano-grains near the buried ITO-substrate interface into larger equiaxed grains at the surface (Fig. 3(b)). This does not occur in the HiPIMS films (Fig. 3(a)) where the microstructure is composed of columnar grains growing perpendicular to the substrate throughout the entire film. There were no cracks or bulges in the film before or after annealing.
Fig. 3. SEM pictures of a) HiPIMS deposition and b) DC deposition of ITO on a Si wafer. a) The HiPIMS film can be seen to be composed of mainly columnar grains, while the b) DC film is made of equiaxed grains and nano-grains.
To further study the structural and compositional differences in the pre and post annealed films, XRD and STEM-EDS were done and shown in Fig. 4. The structure factor for ITO indicates that for a randomly oriented polycrystalline film the reflections, in order of intensity, correspond to the (222), (440), (400), (622), (211), and (411) planes with the first three being much more intense than the latter three. In both the annealed and unannealed films, the (222) peak is the strongest, as expected, and the (400) and (622) reflections are present with the expected relative intensities. The (440) peak is not present above the noise floor despite the structure factor showing that it should have approximately half the intensity of the (222) peak. Based on the relative intensities of the peaks the as-deposited film has a texture favoring the <111 > and <100 > orientations. This contrasts with other reports of as deposited ITO films made via electron beam evaporation where there is only a broad amorphous peak [8,9]. This could be because the HiPIMS plasma give more kinetic energy to the sputtered particles and therefore some crystallization can take place even on room temperature samples. After annealing for 2 hrs., the (211) intensity increases and a stronger than expected (411) peak appears, indicating the growth of small grains with less dominant orientations at the expense of the larger grains with orientations favored by the growth conditions, leading to a shift in texture. The (400) peak remains absent, however. This change in texture, along with a change in the grain size distribution, may affect the plasma mean free path, contributing to the observed shift in permittivity. This increase in crystallinity is consistent with other ITO films grown with electron beam evaporation and then annealed [8,9,25]. Paine et al. have shown that annealing ITO films can cause structural relaxation in amorphous material which promotes local ordering as well as subsequent crystallization from nucleation and growth mechanisms [8].
Fig. 4. Graphs showing the XRD spectra of the pre- and post-annealed HiPIMS ITO samples. After annealing (2 hrs.), it can be seen that ITO (220) and ITO (411) appear.
The STEM-EDS line profiles as a function of film depth are shown in Fig. 5 along with STEM cross-sectional images. In the as-deposited sample, the amount of oxygen expected is much lower than the theoretical stoichiometric concentration of 60.7%. This was expected from both our previous work on ITO as well as other reports of oxygen deficient ITO films in the literature [7,9,26]. Another contributing factor could be that due to the 1.6 um thickness of the ITO films, it was difficult to prepare a TEM cross-section specimen that left the entire film intact but thin enough for an ideal TEM specimen. As a result, the areas where EDS was performed were very thick, resulting in the relative loss of O characteristic x-ray signal. The O kα1 characteristic x-ray has an energy of 0.52 keV, whereas the In and Sn kα1 lines are at 24.21 keV and 25.27 keV, respectively. The two order of magnitude difference in energies results in a significant difference in the probability of the O kα1 x-rays being scattered before escaping the thick specimen compared to those generated by excitation of In and Sn, resulting in a significant undercounting of O. This introduces a systematic error, resulting in an offset of relative values of O, In, and Sn, but should have little to no effect on the shape of the individual element distributions across the specimens.
Fig. 5. STEM-EDS line profiles showing the elemental composition of the films a) as deposited, b) after annealing. From the EDS, it can be seen that during annealing oxygen is making its way into the sample from pores in the film.
There is an increase in oxygen concentration at the surface of the film after annealing. This could be caused either by trace oxygen impurities present during the anneal which then was picked up by the oxygen deficient film or by a loss in indium which in turn would increase the relative concentration of oxygen.
4. Discussion and conclusion
By using HiPIMS, we have grown and characterized micron-scale-thick ITO films. We tune the film permittivity through post deposition processing and achieve ENZ condition at 1550 nm wavelength. Through multiple microanalysis methods, we show that the thick ITO films developed by the HiPIMS process are more uniform in both microstructure and optical properties compared to our previously developed ITO film grown with a conventional DC sputtering process. By increasing the cumulative anneal time, it was shown that the HiPIMS ITO films changed their microstructure and optical properties.
Optical property (ε1, ε2) changes are caused by shifts in the plasma frequency which is determined by both the carrier concentration and the mean free path. Changes in either of these can be caused by a few possibilities such as changes in crystal size, morphology, and orientation as well as changes in oxygen vacancy concentration [8,9,17–20] or other changes in the bonding states of indium during thermal annealing. Paine et al. showed that the relaxation of amorphous ITO is followed by crystallization of small, ordered domains [8]. This crystallization is 3-dimensional but is forced to be 2-dimensional at the top and bottom boundaries of the film [8]. This change in growth direction may explain why ε1 is an arc with a high point in the middle and low points on either end in Fig. 2.
Our previous work on ENZ ITO demonstrated the growth of wavelength thick ITO films and the potential for them to be used as a cladding material for waveguides, resonator cavities, and other applications. While having a varying permittivity throughout the bulk of a film could be useful in devices that utilize depth dependent permittivity, many applications would require more uniform films. The goal of the present work is to improve the uniformity in these films. The optical properties in the HiPIMS ITO films makes them more suitable as a platform for new optoelectronic devices, high-index-contrast metastructures, and other photonic devices. The ability to fine tune the real permittivity of the ITO film into ENZ regime makes HiPIMS ITO a versatile material choice for novel optoelectronic devices.
Funding
U.S. Army Combat Capabilities Development Command.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. A. Boltasseva and H. A. Atwater, “Low-loss plasmonic metamaterials,” Science 331(6015), 290–291 (2011). [CrossRef]
2. N. Engheta, “Pursuing near-zero response,” Science 340(6130), 286–287 (2013). [CrossRef]
3. M. Z. Alam, I. De Leon, and R. W. Boyd, “Large optical nonlinearity of indium tin oxide in its epsilon-near-zero region,” Science 352(6287), 795–797 (2016). [CrossRef]
4. C. Argyropoulos, P. Chen, G. D’Aguanno, N. Engheta, and A. Alù, “Boosting optical nonlinearities in epsilon-near-zero plasmonic channels,” Phys. Rev. B 85(4), 045129 (2012). [CrossRef]
5. P. R. West, S. Ishii, G. V. Naik, N. K. Emani, V. M. Shalaev, and A. Boltasseva, “Searching for better plasmonic materials,” Laser Photonics Rev. 4(6), 795–808 (2010). [CrossRef]
6. A. Alù, M. G. Silveirinha, A. Salandrino, and N. Engheta, “Epsilon-near-zero metamaterials and electromagnetic sources: tailoring the radiation phase pattern,” Phys. Rev. B 75(15), 155410 (2007). [CrossRef]
7. J. N. Ni, W. L. Sarney, A. C. Leff, J. P. Cahill, and W. Zhou, “Property variation in wavelength-thick epsilon-near-zero ITO metafilm for near IR photonic devices,” Sci. Rep. 10(1), 713 (2020). [CrossRef]
8. D. C. Paine, T. Whitson, D. Janiac, R. Beresford, and C. O. Yang, “A study of low temperature crystallization of amorphous thin film indium-tin-oxide,” J. Appl. Phys. 85(12), 8445–8450 (1999). [CrossRef]
9. T. Neubert, F. Neumann, K. Schiffmann, P. Willich, and A. Hangleiter, “Investigations on oxygen diffusion in annealing processes of non-stoichiometric amorphous indium tin oxide thin films,” Thin Solid Films 513(1-2), 319–324 (2006). [CrossRef]
10. J. W. Elam, D. A. Baker, A. B. F. Martinson, M. J. Pellin, and J. T. Hupp, “Atomic layer deposition of indium tin oxide thin films using nonhalogenated precursors,” J. Phys. Chem. C 112(6), 1938–1945 (2008). [CrossRef]
11. T. C. Gorjanc, D. Leong, C. Py, and D. Roth, “Room temperature deposition of ITO using RF magnetron sputtering,” Thin Solid Films 413(1-2), 181–185 (2002). [CrossRef]
12. J. H. Kim, K. A. Jeon, G. H. Kim, and S. Y. Lee, “Electrical, structural, and optical properties of ITO thin films prepared at room temperature by pulsed laser deposition,” Appl. Surf. Sci. 252(13), 4834–4837 (2006). [CrossRef]
13. H. Ghobadi, Y. Smirnov, H. L. Offerhaus, J. A. Alvarez-Chavez, M. Morales-Masis, and I. De Leon, “Optical properties of highly-crystalline tin-doped indium oxide films in their near-zero permittivity spectral region,” Opt. Mater. Express 12(1), 96–108 (2022). [CrossRef]
14. J. Alami, POÅ Persson, D. Music, J. T. Gudmundsson, J. Böhlmark, and U. Helmersson, “Ion-assisted physical vapor deposition for enhanced film properties on nonflat surfaces,” J. Vac. Sci. Technol., A 23(2), 278–280 (2005). [CrossRef]
15. J. Alami, S. Bolz, and K. Sarakinos, “High power pulsed magnetron sputtering: Fundamentals and applications,” J. Alloys Compd. 483(1-2), 530–534 (2009). [CrossRef]
16. F. C. Carreri, A. Sabelfeld, H. Gerdes, R. Bandorf, M. Vergöhl, and G. Bräuer, “HIPIMS ITO films from a rotating cylindrical cathode,” Surf. Coat. Technol. 290, 65–72 (2016). [CrossRef]
17. M. Chen and S. Chen, “Influence of oxygen deficiency in indium tin oxide on the performance of polymer light-emitting diodes,” Thin Solid Films 517(8), 2708–2711 (2009). [CrossRef]
18. A. Thogersen, M. Rein, E. Monakhov, J. Mayandi, and S. Diplas, “Elemental distribution and oxygen deficiency of magnetron sputtered indium tin oxide films,” J. Appl. Phys. 109(11), 113532 (2011). [CrossRef]
19. G. Giusti, “Deposition and characterization of functional ITO thin films,” Dissertation, University of Birmingham, 2011.
20. A. K. Kulkarni, K. H. Schulz, T. S. Lim, and M. Khan, “Dependence of the sheet resistance of indium-tin-oxide thin films on grain size and grain orientation determined from X-ray diffraction techniques,” Thin Solid Films 345(2), 273–277 (1999). [CrossRef]
21. Z. Xu, P. Chen, Z. Wu, F. Xua, G. Yang, B. Liu, C. Tan, L. Zhang, R. Zhang, and Y. Zheng, “Influence of thermal annealing on electrical and optical properties of indium tin oxide thin films,” Mater. Sci. Semicond. Process. 26, 588–592 (2014). [CrossRef]
22. J. H. Park, C. Buurma, S. Sivananthan, R. Kodama, W. Gao, and T. A. Gessert, “The effect of post-annealing on Indium Tin Oxide thin films by magnetron sputtering method,” Appl. Surf. Sci. 307, 388–392 (2014). [CrossRef]
23. N. M. Ahmed, F. A. Sabah, H. I. Abdulgafour, A. Alsadig, A. Sulieman, and M. Alkhoaryef, “The effect of post annealing temperature on grain size of indium-tin-oxide for optical and electrical properties improvement,” Results Phys. 13, 102159 (2019). [CrossRef]
24. H.G. Tompkins and J.N. Hilfiker, Spectroscopic Ellipsometry: Practical Application to Thin Film Characterization, (Momentum Press, 2015).
25. N. M. Khusayfan and M. M. El-Nahass, “Study of structure and electro-optical characteristics of indium tin oxide thin films,” Adv. Condensed Matter Phys. 2013, 1–8 (2013). [CrossRef]
26. M. Nistor, J. Perrière, C. Hebert, and W. Seiler, “Nanocomposite indium tin oxide films: formation induced by a large oxygen deficiency and properties,” J. Phys.: Condens. Matter 22(4), 045006 (2010). [CrossRef]
