Influence of Proton Irradiation on Thin Films of AZO and ITO Transparent Conductive Oxides—Simulation of Space Environment

Featured Application

Optoelectronic devices in space missions.

Abstract

Transparent conductive oxides are essential materials for many optoelectronic applications. For new devices for aerospace and space applications, it is crucial to know how they respond to the space environment. The most important issue in commonly used low-Earth orbits is proton radiation. This study examines the effects of high-energy proton irradiation (226.5 MeV) on thin films of aluminium-doped zinc oxide (AZO) and indium tin oxide (ITO). We use X-ray diffraction and electron microscopy observations to see the changes in the structure and microstructure of the films. The optical properties and homogeneity of the materials are determined by spectrophotometry and spectroscopic ellipsometry (SE). Analysis of the chemical states of the elements with X-ray photoelectron spectroscopy (XPS) gives insight into what proton irradiation changes at the surface of the oxides. All measurements show that ITO is less influenced than AZO. The proton energy and fluence used in this study simulate about a hundred years in low Earth orbit. This research demonstrates that both transparent conductive oxide thin films can function under simulated space conditions, with ITO showing superior resilience. The ITO film was more homogenous in terms of the total thickness measured with SE, had fewer defects and adsorbates present on the surface, as XPS analysis proved, and did not show a difference after irradiation regarding its optical properties, transmission, refractive index, or extinction coefficient.

1. Introduction

Transparent electrodes are viable materials for the further development of optoelectronics, such as semitransparent light-emitting diodes, solar cells, or electrochromic devices [1,2,3,4]. Different materials can be used as transparent electrodes, such as polymers, for example, PEDOT: PSS, carbon-based nanomaterials, ultra-thin metal films, or conductive oxides [5,6,7]. Among transparent conductive oxides, two have been well known for years—fluorine-doped tin oxide (FTO) and indium-doped tin oxide (ITO)—while recent research and applications use aluminium-doped zinc oxide (AZO) as an alternative [8,9,10,11].

Optoelectronic devices can be found in the aerospace field, in space missions, satellites, or future establishments for manned missions [12,13]. Photovoltaics in particular are already being used in space applications and will remain a reliable source of electrical energy. Moreover, ITO is used as part of thermal control coatings in space missions [14]. The use of a transparent electrode could broaden the range of possible materials to be used in photovoltaics or include electrochromic systems in space applications [15,16].

Any material used in the space environment will be exposed to cosmic radiation. The space environment is different depending on the location of a satellite. In the widely used low-Earth orbit, the most significant is proton irradiation, where protons are trapped by the Earth’s magnetosphere in the van Allen belts [17]. All charged particles exist in a wide spectrum of energy, from low energy to hundreds of MeV. The protons energies range from 30 keV to 400 MeV at 500 km above the Earth’s surface.

In this study, we investigate how a high-energy proton beam of 226.5 MeV affects the properties of thin films of ITO and AZO transparent conductive oxides. Researchers have previously studied ITO but usually with very low energy. Wei et al. [18] treated ITO thin films with a 100 keV proton beam of different doses, with the highest of 2 × 10¹⁶ cm⁻², which is equal to 7 years in the geomagnetic equator region. The irradiation reduced both the conductivity and the transmittance of ITO, while the sheet resistance increased twice. Khasanshin and Novikov [19] studied how 20 keV proton irradiation influences K208 glass coated with ITO. They found that degradation of the ITO film caused larger changes in the material reflectivity than in the case of a glass without a coating. Oryema et al. [20] studied another transparent conductive oxide, namely FTO. They used a proton beam of 7 MeV and different fluences. With increasing dose of ions, the transmittance of films decreased, and the sheet resistance decreased for lower doses, but at 1 × 10¹⁶ cm⁻², it abruptly rose. They found small changes in the crystal structure of the oxide; crystallinity improved after irradiation, and the peaks shifted to lower angle values.

In this work, we study the influence of high-energy protons on thin films of AZO and ITO. A preliminary simulation using Stopping and Range of Ions in Matter (SRIM-2008) [21] software showed that the influence of this radiation should be low on ITO and AZO thin films. The simulated trajectories of ions showed that the particles would pass through the material in a straight line. SRIM software gives information about energy passed by incident particles into the target material—the results showed that no energy would be transferred into the samples. The structural properties of the films were evaluated with X-ray diffraction (XRD), and the chemical states of the surface were evaluated with X-ray photoelectron spectroscopy (XPS). Optical properties were studied with spectrophotometry, and an optical method of spectroscopic ellipsometry (SE) was used to estimate the homogeneity of thin films. Observations with scanning electron microscopy (SEM) were also made.

2. Materials and Methods

ITO films were purchased from Merck (Darmstadt, Germany)and had a thickness of around 200 nm, deposited on a glass substrate. AZO films were deposited with atomic layer deposition at a thickness of 100 nm using a Beneq P400A reactor (Beneq, Espoo, Finland). The glass substrates were placed in a chamber where they were initially stabilized at temperature of 200 °C and pressure of 1 mbar for 4 h. The low vacuum was achieved using a rotary pump and the pressure was controlled using an MKS Baratron 626D sensor (MKS Instruments, Inc., MA, USA). The precursors of diethylzinc and trimethylaluminum were then alternated with distilled water over a period of 2 h. The opening times of the valves that supply the precursors and the distilled water were 0.3 and 0.5 s, respectively. The irradiation experiment was performed at the Polish Academy of Sciences Institute of Nuclear Physics in Kraków. The energy of proton beams was 226.5 MeV, and the fluence of protons was 10¹¹ cm⁻². Samples were placed on a holder in such a way that the beam was incident directly on the film. A Gantry station was used to perform the irradiation experiment, which was conducted in air ambience and at room temperature. The beam homogeneity over the irradiated area was 9.99%. The samples after irradiation are named with an H letter; therefore, before irradiation they are AZO and ITO, and after irradiation, AZO:H+ and ITO:H+. Characterisation of irradiated samples was carried out after transporting them to laboratories at the AGH University and to NIMP in Romania.

Structural properties were studied with a Panalytical Empyrean diffractometer using a copper anode (λ = 0.154 nm); the parameters for grazing incidence X-ray diffraction measurement were: 2 θ range 20–80°, step 0.04°, time per step 5 s, and omega angle 1°. Homogeneity maps were measured with a J.A. Woolam 2000 M spectroscopic ellipsometer. The measurement angles were 58°, 62°, and 66°. To observe the microstructure with scanning electron microscopy, we used Phenom microscope from Phenom-World. Transmittance of films as measured with an Avantes Sensline Ava-Spec ULS-RS-TEC spectrophotometer and an Avantes AvaLight DH-S-BAL-Hal lamp (Avantes, Apeldoorn, The Netherlands).

The investigation of chemical properties of the surface was conducted with X-ray photoelectron spectroscopy in an analysis chamber equipped with a 150 mm hemispherical electron energy analyser (Phoibos). The analyser operated in fixed transmission mode (FAT) with pass energy of 20 eV. The X-ray source was monochromatised Al Kα with 1486.6 eV energy, at 250 W power (12.5 kV × 20 mA). The charging effects were compensated with a flood gun (1 V × 0.1 mA) and with binding energy correction to the C 1 s contamination core level at 286.4 eV.

3. Results and Discussion

X-ray diffraction studies show no detectable influence of proton irradiation (Figure 1). The diffractograms are identical before and after irradiation. The assignment of peaks was carried out based on the ICDD card #00-001-0929 for ITO and #00-005-0664 for AZO. ITO has the structure of the In₂O₃ cubic Ia3 space group no. 206, and AZO has the structure of the ZnO hexagonal system group P63mc no. 186.

Microscopic observations presented in Figure 2 showed that the films were homogeneous. On the surface of the AZO, there were some precipitates with an increased amount of Zn, as measured using energy dispersive spectroscopy (Figure 2a). The sample after irradiation also exhibited these elements, as well as other impurities (Figure 2b). ITO before and after irradiation had some impurities on the surface whose nature and amount did not change after irradiation. It was possible to observe these samples with higher magnification, which is presented in the insets of Figure 2c,d. ITO films have grains that are hundreds of nanometres in size.

The transmittance of transparent conductive oxides is a crucial property for these materials. For all samples, the transmittance is above 80% for all measured wavelengths. AZO transmittance decreased after irradiation in the 400 to 630 nm range (Figure 3a). ITO transmittance was slightly reduced throughout the measured range (Figure 3b).

The fitting of ellipsometric data to a dielectric function model was performed with CompleteEASE 6.71 software from J.A. Woollam. To create a model for ITO, two Tauc–Lorentz oscillators, one Gaussian oscillator, and the Drude function were used. For AZO, the same functions were used, but with only one Tauc–Lorentz oscillator. Equations (1)–(8) define these functions.

• Tauc–Lorentz oscillator [22]:

$${\epsilon }_{T-L}\left(E\right)={\epsilon }_1-{i\epsilon }_2,$$

(1)

$${\epsilon }_2=\left[{\displaystyle \frac{Amp E_0 Br{\left(E-E_g\right)}^2}{{\left(E^2-E_0^2\right)}^2+{Br}^2{E}^2}}·{\displaystyle \frac{1}{E}}\right], E>E_g,$$

(2)

$${\epsilon }_2=0, E\le E_g,$$

(3)

$${\epsilon }_1={\displaystyle \frac{2}{\pi }}P{\int }_{E_g}^{\infty }{\displaystyle \frac{{\xi }^{\prime }{\epsilon }_2\left(\xi \right)}{{\xi }^2-E^2}}d\xi ,$$

(4)

where ε₁—real part of dielectric function, ε₂—imaginary part of dielectric function, E_g—energy band gap, E—photon energy, E₀—peak central energy, Amp—amplitude, Br—broadening, and P—Cauchy principal value.

• Gaussian oscillator [22]:

$${\epsilon }_{Gaussian}\left(E\right)=Amp(\Gamma \left(Z^{-}\right)+\Gamma \left(Z^{+}\right)+i\left[\exp \left(-{\left(Z^{-}\right)}^2\right)-\exp \left(-{\left(Z^{+}\right)}^2\right)\right]),$$

(5)

where Γ is a convergence series that produces a Kramers–Kronig relation consistent shape of the dielectric function, and Z⁺ and Z⁻ are equivalent to Gaussian “Z-values”:

$$Z^{+}={\displaystyle \frac{E+E_0}{{\sigma }_n}}\hspace{1em}\hspace{1em}\hspace{1em}{Z}^{-}={\displaystyle \frac{E-E_0}{{\sigma }_n}},$$

(6)

σ_n is equivalent to the Gaussian standard deviation.

$${\sigma }_n={\displaystyle \frac{Br}{2\sqrt{ln2}}},$$

(7)

• Drude oscillator [22]:

$${\epsilon }_{Drude}(E)={\displaystyle \frac{-{ħ}^2}{{\epsilon }_0\rho (\tau ·E^2+iħ\mathrm{E})}},$$

(8)

where ρ is resistivity and τ is scattering time.

Figure 4 shows maps of the total thickness calculated for the samples from the ellipsometry measurements. The models are multilayered, with three layers used in the case of both materials. The number of layers was decided on based on knowledge of the samples and by optimizing this number during the data fitting process. The map average of mean square error of the fittings as well as average thickness of model layers and total thickness are presented in Figure 5. The mean square error of the map fitting is very good for ITO samples below 4 and for AZO below 3. For ITO, the thickest layer is the bottom one, whereas for AZO, it is the middle layer. These results are influenced by the different deposition methods used for each material. The ITO was purchased as a pre-manufactured product, suggesting that its production technology is likely more optimized than the single-sample preparation used for AZO. Industrial manufacturing processes tend to yield a better-prepared substrate surface, resulting in no detectable interface in ellipsometry measurements. The two top layers have distinct properties: one forms when the material is exposed to the environment after deposition, while the final layer represents surface roughness and adsorbates that develop during sample handling. In case of AZO, there is an interface layer between substrate and the main bulk of the film. Figure 4 shows maps of calculated total thickness of the films, which is the sum of all layers in the model; in each case, it was three layers. The average thickness of AZO was 103.9 ± 2.2 nm and that for AZO:H+ was 113.2 ± 3.0 nm, which is close to the expected value of 100 nm, while the error and total thickness increased after irradiation. For ITO, the total thickness was 189.2 ± 0.5 nm and that for ITO:H+ was 191.3 ± 1.7 nm; again, the error after irradiation increased, while the total thickness increased slightly.

Optical properties of oxide thin films were determined from the fittings as well, with the thickest layer in each model, i.e., layer #1 for ITO and layer #2 for AZO, as arbitrarily chosen representative points. The thickest layers were used in optical properties analysis as they have the greatest impact on sample optical properties. Figure 6 presents refractive index n and extinction coefficient k of ITO and AZO before and after irradiation. The properties of ITO did not change with the irradiation, while for AZO, they changed slightly. The biggest change was present for the extinction coefficient at wavelengths higher than 1000 nm. This increase is due to free carrier absorption in the material [23]. This result suggests that the number of charge carriers in AZO increases, which could cause an increase in the material’s conductivity.

Figure 7 shows the fittings of the XPS signal for the AZO and ITO samples; the data before irradiation of a sample are presented lower in each figure and those of an irradiated sample higher. For AZO, we measured narrow scans of the photoemission lines of interest: Al 2p, O 1s, Zn 2p, and C 1s. For ITO, we acquired high-resolution scans of In 3d, Sn 3d, and O 1s. All fittings were performed in Igor Pro using the Voigt function, which is a combination of Gaussian and Lorentzian functions [24]. In all spectra, the main contribution to the lattice component is associated, while at higher binding energies (with ~1 eV difference), we have surface components that originate from surface defects caused by under-coordinated bonds.

In the case of AZO, we attributed the components as follows: the lattice components are associated with Al 2p at about 73.6 eV and with Zn 2p at about 1021.4 eV. For O 1s, the binding energies (BE) of Al-O and Zn-O are distinguishable, with Zn-O at about 530 eV and Al-O at about 531.5 eV, while a component with BE between the two is associated with the undercoordinated bonds at the surface [25,26]. In

Table 1. Relative contribution of components in the XPS analysis of AZO and ITO before and after proton irradiation.

Sample	Relative Contribution	Zn [%]	Al [%]	O [%]
AZO	Lattice	76.4	50.3	41.3 Zn-O 23.7 Al-O
	Surface	21.3	30.4
	Defects	2.3	19.2	26.8
	Adsorbates			8.3
AZO: H+	Lattice	64.8	56.5	27.2 Zn-O
	Surface			39.0 Al-O
	Defects		27.8	20.2
	Adsorbates	1.7		13.7
		In [%]	Sn [%]	O [%]
ITO	Lattice	64.1	67.2	45.1
	Surface	30.0	32.8	25.9
	Defects			29.0
	Adsorbates	5.9
ITO: H+	Lattice	64.4	68.1	41.7
	Surface	29.6	31.9	27.0
	Defects			31.3
	Adsorbates	6.1

, we can see the relative contribution of different types of bonds in each spectrum. The surface adsorbates are bound by AlO_x, and since the surface sensitivity of Al is much lower than that of Zn (Al 2p signal comes from the topmost ~7.5 nm, while the Zn 2p signal comes from the topmost ~3 nm), one can assume that the surface is dominated by AlO_x. A second argument for this assumption is that the AlO_x lattice signal increases after irradiation, as well as the Al-O component in the O 1s spectrum, while the Zn contribution associated with the lattice component decreases.

The shift of the signal to higher binding energies is related to the decrease in the conductivity of the surface, as then more energy is needed to extract a core-level electron. The main component of Al 2p shifted from 73.6 eV BE to 73.8 eV after irradiation. The main component of Zn 2p shifted from 1021.3 eV BE to 1021.5 eV. In the case of O 1s, the Zn-O component shifted from 529.8 to 530 eV and the Al-O from 531.8 to 532 eV. All the secondary components of Al 2p, Zn 2p, and O 1s also shifted to higher energies. This indicates that the resistance of AZO should increase after proton irradiation. This behaviour is also consistent with the surface dominated by AlO_x, which is a high band gap insulator [27].

In the case of ITO, the irradiation does not change the surface, as there is no notable shift in the BE of the core levels and also no change in the composition of the surface. Only the intensity component of O 1s that comes from the I₂O₃ lattice is slightly decreased (Figure 7f). In the case of ITO, information about relative concentrations of Sn and In can be extracted from XPS because the kinetic energies of the species of interest are comparable, and thus the surface sensitivities are comparable. Both samples show an In:Sn ratio of 10:1.

The sheet resistance of the films changed only slightly for ITO film (increase of about 1.8%). AZO before irradiation had a resistance of 302.485 ± 3.413 Ω/□, which, after the experiment, decreased to 202.347 ± 0.265 Ω/□ (33% decrease). The results of the four-point probe measurements are summarized in

Table 2. Sheet resistance of studied samples.

Sample	Sheet Resistance [Ω/□]
AZO	302.485 ± 3.413
AZO: H+	202.347 ± 0.265
ITO	9.106 ± 0.001
ITO: H+	9.266 ± 0.010

. The decrease in resistance of AZO after irradiation is in contrast to the XPS results, which indicate increased electrical resistance. As AZO is an oxide semiconductor, its conductivity is influenced by the presence of oxygen vacancies and metallic interstitial defects.

4. Conclusions

In this study, we characterize thin films of conducting oxides, ITO and AZO, before and after proton irradiation. The treatment is meant to simulate the cosmic space environment. XRD characterization proved that no structural changes were induced by the 226.5 MeV proton beam at a fluence of 10¹¹ cm⁻². Microscopic observations with SEM showed that ITO did not change with irradiation, whereas on the surface of AZO, new defects and larger precipitates emerged. Using mapping spectroscopic ellipsometry, we observe that proton irradiation made both oxides less homogenous. XPS analysis clearly shows that irradiation was more impactful for AZO than ITO. Proton irradiation caused more defects and deterioration in AZO. All components in AZO shifted to higher binding energy, which indicates an increase in electrical resistance. While the transmittance and resistance of ITO did not change much, in the case of AZO, the transmittance increased, and the resistance decreased after irradiation. A higher number of defects could be the reason for improved conductivity, as in the case of oxide semiconductors, their conductivity depends on the presence of defects such as oxygen vacancies and metallic interstitials. Higher AZO conductivity corresponds to analysis of optical properties with spectroscopic ellipsometry, where the extinction coefficient behaviour indicates an increase in the free carriers number. The contradiction between XPS and sheet resistance measurements may be due to a different depth of the material being probed. In the case of XPS, only a few nanometres of surface are explored, while electrodes in the four-point probe measurement can penetrate the surface.

This research shows that the high-energy proton environment has a slight negative influence on the properties of transparent conductive oxide thin films, proving they are good candidates as functional layers in space applications.