In Situ Monitoring of Pulsed Laser Annealing of Eu-Doped Oxide Thin Films

Abstract

Eu³⁺-doped oxide thin films possess a great potential for several emerging applications in optics, optoelectronics, and sensors. The applications demand maximizing Eu³⁺ photoluminescence response. Eu-doped ZnO, TiO_2, and Lu₂O₃ thin films were deposited by Pulsed Laser Deposition (PLD). Pulsed UV Laser Annealing (PLA) was utilized to modify the properties of the films. In situ monitoring of the evolution of optical properties (photoluminescence and transmittance) at PLA was realized to optimize efficiently PLA conditions. The changes in optical properties were related to structural, microstructural, and surface properties characterized by X-ray diffraction (XRD) and atomic force microscopy (AFM). The substantial increase of Eu³⁺ emission was observed for all annealed materials. PLA induces crystallization of TiO₂ and Lu₂O₃ amorphous matrix, while in the case of already nanocrystalline ZnO, rather surface smoothening0related grains’ coalescence was observed.

1. Introduction

Eu³⁺-doped oxide thin films possess great potential for several emerging applications in optics, optoelectronics, and sensors, i.e., waveguides, display luminophores, imaging detectors, solar cells, and scintillators [1,2,3,4,5,6,7,8,9,10,11]. Eu³⁺ was the dopant of choice in relation to the previously mentioned applications due to its strong emission in the visible part of spectra centered at around 612 nm [12]. As an example of well-known Eu³ dopant host matrices we can mention semiconducting ZnO, TiO_2, and dielectric Lu₂O₃ oxides. The Eu³⁺-doped thin films are fabricated by a variety of methods [1,2], e.g., ion implantation [13], plasma-enhanced chemical vapor deposition [3,4,14], electrochemical deposition [5], hydrothermal deposition [15,16], chemical bath deposition [6], spraying [17,18], sputtering [4,7,8,19,20,21], evaporation [22,23], pulsed laser deposition (PLD) [24,25,26,27], matrix-assisted pulsed laser evaporation technique (MAPLE) [28], and sol-gel [9,10,29,30,31,32,33].

Post-deposition thermal treatment is usually required to activate the Eu ions and optimize the photoluminescence (PL) response [4,8,9,13,14,16,17,19,20,21,26,28,29,30,32,33]. However, the methods used for conventional thermal annealing require heating the samples to high temperature (>700 °C) for sufficiently long times in order for thermal diffusion of the defects to occur. The high-temperature processing limits the flexibility and practical applicability of these methods toward development of various optical and electro-optical devices. To address just the main disadvantages, only heat-resistant substrates can be used, undesirable crystallization of amorphous host might also occur, and, finally, the treatment of the entire sample may be undesirable in the device production.

The utilization of Pulsed Laser Annealing (PLA) to obtain and tune optically active Eu³⁺-doped films may profit from numerous advantages [34,35], i.e., (if necessary) horizontally structured local-rapid laser rating, ability to achieve ultra-high temperature in local scale, better controlled vertical thermal profile, controlled thermal profile, rapid processing, possibility of remote sample processing, and easy process automation. PLA plays, therefore, an important and challenging role in the development of high-tech electronic, optoelectronic, photonics, and sensing devices.

PLA technique offers a great flexibility in the number of control parameters available, i.e., laser wavelength, fluence (F_L), repetition rate (f_rep), number of shots, and ambient atmospheres [36,37,38]. Finding the optimal PLA conditions strongly depends on sample properties, i.e., film thickness, optical properties, and microstructure. Therefore, in order to find the optimal PLA conditions for maximizing Eu³⁺ PL response, a number of testing samples will be necessary. In situ monitoring of PLA can be an effective way to reach the optimal conditions accurately and fast.

In this paper, we report on in situ monitoring optical properties of Eu-doped ZnO, TiO₂, and Lu₂O₃ thin films prepared by PLD within their processing by PLA using ArF excimer laser. Photoluminescence and optical transmittance measurement techniques were implemented. The effect of PLA on morphology, microstructure, and composition was evaluated.

2. Materials and Methods

ZnO:Eu, Lu₂O₃:Eu, and TiO₂:Eu thin films were fabricated by PLD using Nd:YAG laser (λ = 266 nm, τ = 6 ns) and KrF laser (λ = 248 nm, τ = 20 ns). The targets were Eu₂O₃:ZnO (1% at. Eu), Eu₂O₃:Lu₂O₃ (3% at. Eu), and Eu₂O₃:TiO₂ (1% at. Eu). Depositions were performed in an oxygen ambient. The vacuum chamber was pumped down to an ultimate pressure of 5.10⁻⁴ Pa. The films were grown on fused silica substrates maintained at room temperature. The distance between substrate and target was 50 mm. The deposition conditions are listed in

Table 1. The deposition and laser annealing conditions.

Material	Fluence PLD (J⋅cm−2)	O2 Pressure (Pa)	Deposition Time (min)	Thickness (nm)	Fluence PLA (J⋅cm−2)
ZnO:Eu (1% at. Eu)	5	15	8.3	840–1000	175
TiO2:Eu (1% at. Eu)	9	1	25	800	175
Lu2O3:Eu (3% at. Eu)	5	2.5	20	110-150	150

.

PLA was performed using an ArF laser (λ = 193 nm, τ = 20 ns) Compex PRO 205 F (Coherent, Inc., Santa Clara, CA, USA). To perform PLA the laser beam was focused using converging lens on the sample at the angle of 70°. The optical properties were analyzed sequentially after each PLA pulse or burst of pulses that were fired at f_rep = 1 Hz. The signal for in situ PL was taken from the sample edge using an optical fiber connected to a spectrometer Horiba JY Triax iHR550 (Kyoto, Japan) equipped with LN cooled CCD Symphony. The same ArF laser was used for spectra excitation at lowered F_L = 10 mJ⋅cm⁻² with f_rep = 5 Hz. Optical transmittance was in situ probed perpendicularly using light generated by a DH 2000 source (Ocean Insight, Orlando, FL, USA) and analyzed by a spectrometer. The PLA conditions are listed in Table 1 and the experimental setup is shown in Figure 1.

The structure of the films was characterized by X-ray diffraction (XRD) using an X-Ray diffractometer (EMPYREAN, Malvern Panalytical, Malvern, UK) with Cu anode (λ = 1.5406 Å) by grazing incidence (omega = 0.85°) and Bragg–Brentano measurements at room temperature. Morphology was examined by atomic force microscopy (AFM) (Dimension Icon, Bruker, Billerica, MA, USA). Chemical composition was determined by Scanning electron microscopy-Energy-dispersive X-ray spectroscopy (SEM-EDX) (Fera 3, TESCAN, Brno, Czech Republic) and the Orbis^® PC Micro-XRF spectrometer (EDAX, Ametek, Berwyn, PA, USA), Rh X-ray tube, 40 kV, polycapillary optics, a diameter of 30 µm of incident beam, and a silicon-drift detector with resolution FWHM~130 eV/Mn. The thickness measurement took place on an Alphastep IQ device (KLA-Tencor Corporation, Milpitas, CA, USA) with a scan length of 5 mm, stylus force of 5.5 mg, and scan speed of 50 µm·s⁻¹. Each line was measured two times. The stylus diamond tip had a radius of 5 µm and an angle of 60°.

3. Results and Discussion

The deposited films of all investigated materials exhibited characteristic Eu³⁺ PL ⁵D₀→⁷F_j (j = 0, 1, 2, 3, and 4) emission related to f-f transition. The observation of the PL emission confirmed that the energy transfer processes in Eu³⁺ excitation must take place considering excitation wavelength λ = 193 nm (6.42 eV). It can be mainly related to host absorption and marginally in case of Lu₂O₃:Eu also to the O²⁻–Eu³⁺ charge transfer band (CTB) [39,40]. Following hyper-sensitive electric dipole transition ⁵D₀→⁷F₂, its Stark splitting and ⁵D₀→⁷F₂/⁵D₀→⁷F₁ (magnetic dipole transition) emission ratio (asymmetry ratio), we could correlate the effect of PLA on the structural properties’ modification. It provided information related to the surrounding defects and disorder around the Eu³⁺ ions. The ⁵D₀→⁷F₁ transition was not considerably influenced by the surrounding of Eu³⁺ in the particular host. The transition ⁵D₀→⁷F₂ dominated all emission spectra from Eu^3+,, suggesting the local symmetry around Eu³⁺ was low and deviated from an inversion center. The ⁵D₀→⁷F₀ transition was only allowed, taking into account the electric dipole selection rule, in the following 10 site symmetries: C_s, C₁, C_n, and C_nv (n = 2, 3, 4, 6) [12].

The values of E_g listed in

Table 2. Properties, modifications after PLA.

Material	Eu3+ (5D0→7F2) PL Emission Enhancement	Band Gap Eg (eV)		Roughness S a S q (nm)
	PLA	As Deposited	PLA	As Deposited	PLA
ZnO:Eu	4.1×	3.28	3.25	14.6 18.1	3.5 4.6
TiO2:Eu	5×	3.24	3.23	0.5 0.6	1.4 1.8
Lu2O3:Eu	2.8×	5.47	5.51	2.0 3.8	1.5 3.7

were derived using Tauc plot—plotting (αhν)^1/m against photon energy (hν), where m is a parameter related to the type of transition, m = ½, and 2 corresponds to direct allowed transition and indirect allowed transition, respectively. TiO₂ in anatase phase is considered as an indirect semiconductor, where m = 2 [41] while ZnO is a direct semiconductor (m = ½) [42] and Lu₂O₃ is considered as a direct ultrawide band gap semiconductor (m = ½) [43]. The E_g values were obtained with error of ±0.02 eV.

As for the chemical composition of the films, Eu dopant concentrations obtained by EDX and XRF analyses corresponded to those of the targets and any substantial variation after PLA was not detected.

In the following, let us discuss the properties’ modifications of each material, Eu-doped ZnO, TiO_2, and Lu₂O₃ by PLA, in more details.

3.1. ZnO:Eu

The evolution of optical properties (PL emission and transmittance spectra) of ZnO:Eu films with increasing number of PLA shots is shown in Figure 2 and Figure 3, respectively. Characteristic Eu³⁺ red emission was detected at 587.3 nm and 593.8 nm (⁵D₀→⁷F₁), 609.9 nm and 618.8 nm (⁵D₀→⁷F₂), 654.8 nm (⁵D₀→⁷F₃), and 691.5 nm, 692.8 nm and 704.4 nm (⁵D₀→⁷F₄). The spectra in Figure 2 revealed that different emitting centers related to Eu may coexist in ZnO:Eu films, which is commonplace for wide band gap materials of wurtzite structure [44]. Eu³⁺ ions in ZnO:Eu films generally occupy the substitutional sites of Zn²⁺ ions. This fact turns ⁵D₀→⁷F₁ and ⁵D₀→⁷F₂ transitions to be allowed. The observation of ⁵D₀→⁷F₂ transition splitting suggests that Eu³⁺ can be at a substitutional Zn site (C_3v symmetry). We could not observe clear evidence of ⁵D₀→⁷F₀ transition whose appearance in PL spectra of ZnO:Eu was related with Eu³⁺ presence in other sites (e.g., the interstitial sites or surface/grain boundary of ZnO nanocrystals) [33,45,46]. A potential drawback of our interpretation can be the relatively low intensity of the PL signal. However, if we expect only a small fraction of Eu³⁺ in the sample located at ZnO nanocrystals’ surface, Eu³⁺ emission lines’ intensities and asymmetric ratio would increase with the increasing number of PLA shots, while the asymmetric ratio saturated at N_# = 400, as it is depicted in Figure 2. The enhancement of Eu³⁺ PL intensity at 609.9 nm by a factor of 4.1× was obtained. The PL signal intensity and asymmetric ratio variations might be related to microstructure modification, as displayed in Figure 4, where grains coalescing took place. It was reported that the energy relaxation process of excited Eu³⁺ ions should play a more important role toward PL efficiency enhancement than the energy transfer process from ZnO matrix to Eu³⁺ ions [47].

The transmittance spectra shown in Figure 3 confirm good quality of the film because transmittance of ~80% was reached in the visible light spectrum (VIS) band. There was not noticeable any variation in transmittance spectra, which confirmed that defects like oxygen vacancies were not introduced during the PLA processing. Band gap value decreased after PLA from E_g~3.28 eV to 3.25 eV. Recently, we observed similar E_g values and trends at ZnO:Eu film PLD deposited at a substrate temperature of 300 °C [24].

AFM images of the film surface presented in Figure 4 revealed that PLA caused interconnecting of the grains accompanied by substantial lowering of surface roughness, as depicted in Table 2. We could attribute this effect to reaching the melting threshold of ZnO:Eu thin film. Similar results of UV PLA on the surface morphology of ZnO thin film of comparable thickness deposited by PLD on quartz substrate were reported at higher F_L ≥ 140 mJ/cm^2, where coalesced nanoclusters kept the wurtzite crystalline structure [36]. The PLA-induced coalescence of nanoparticles was also reported at ZnO [48] and ZnO:Eu [37] films of lower thickness < 60 nm on metal and Si substrates, respectively.

XRD patterns of as-deposited and PLA-treated ZnO:Eu film, shown in Figure 5, did not reveal any substantial changes in structural properties. This fact is in good correlation with our previous results dedicated to PLA of ZnO:Eu film by KrF laser [37]. The film exhibited a hexagonal wurtzite structure of fibrous texture along the c-axes with random lateral orientations of crystallites. The FWHM of the strongest (002) peak was used to estimate the average crystallite size by using the Scherrer’s formula and were ≈19 nm. This correlated with the AFM measurement data. The presence of additional phases, such as Eu₂O_3, was not detected. Considering the larger ionic radius of Eu²⁺ (1.13 Å) than Eu³⁺ (0.95 Å) in comparison with Zn²⁺ (0.74 Å), substitution of Eu³⁺ ions at Zn²⁺ sites seemed to be more favorable than that of Eu²⁺ ions but charge compensation must take place [24].

3.2. TiO₂:Eu

The dependence of optical properties of TiO₂:Eu films, i.e., PL emission and transmittance spectra, on the number of PLA shots is presented in Figure 6 and Figure 7, respectively. The transmittance spectra underline the high quality of the film because any evident absorption in VIS that can be related to defects such as oxygen vacancies and a presence of Ti suboxides did not appear [28]. We could observe a weak broadened emission line of Eu³⁺ for the as-grown, which is characteristic for the amorphous phase [49]. The increase of Eu³⁺ PL intensity accompanied by sharpening lines of split ⁵D₀→⁷F₂ transition, as depicted in Figure 6, as well as a slight increase of transmittance (1.3%) are corelated to surface modification and gradual crystallization of TiO₂ matrix. The significant increase of Eu³⁺ PL intensity (~5×) was detected at N_# = 50, where transmittance decreased and further remained constant. TiO₂ nanocrystallites formed by PLA may act as sensitizers absorbing the excitation energy more efficiently [31,49]. We can recognize splitting ⁵D₀→⁷F₂ transition to peaks located at 611.1 nm, 614 nm, 616.9 nm, 624 nm, and 626.1 nm. The splitting may indicate Eu³⁺ located in D₂ and S₄ symmetry sites of TiO₂ [49]. The slight decrease of Eu³⁺ PL intensity of ⁵D₀→⁷F₂ transitions at 616.9 nm and 624 nm after N_# = 200 can be attributed to the creation of well-crystalized anatase TiO₂ matrix [28,50]. Because the intensity of ⁵D₀→⁷F₁ transition was weak, it made it impossible to conclude on its splitting. The ⁵D₀→⁷F₀ transition could not be followed at all, which may exclude Eu³⁺ occupying C_2v symmetry sites. Concerning transmittance, we must notice that its variation within PLA processing was generally rather weak. Its relative variation was less than 2%, which is slightly above the measurement error.

We did not obtain any shift in E_g value after PLA, as listed in Table 2. Similar values of E_g~3.22 eV for TiO₂ material [51] and anatase-like thin TiO₂ films [52] were reported. There are reports of higher Eg~3.3 eV for TiO₂:Eu films deposited by the spray pyrolysis technique. Our E_g values suggested we did not introduce defects in the films since lowering E_g from 3.03 eV to 2.52 eV was observed after aerodynamic levitated laser annealing of defect-rich (oxygen vacancies) TiO₂ nanoparticles [53]. Lower E_g~2.7 eV was also reported for TiO₂:Eu films prepared by MAPLE, where E_g value shift was related to interstitial Eu doping [28].

An AFM image of the film surface, shown in Figure 8, revealed clear evidence of crystallization of the film after PLA, as observed by XRD. The growth of grains is reflected by the increasing roughness, as depicted in Table 2.

XRD analyses, shown in Figure 9, revealed the amorphous character of as-grown film, while the PLA-treated film exhibited polycrystalline tetragonal anatase phase TiO₂, as referred in the JCPDS Card NO: 96-500-0224 file. Peak positions and reflections’ planes derived from XRD patterns were 25.38° (011), 36.78° (013), 37.68° (004), 48.03° (020), 53.68° (015), and 55.08° (121). Crystallites’ size was derived using Sherrer’s formula and was ≈ 46.1 nm. Anatase TiO₂ phase was also obtained by KrF PLA of nanoparticles at F_L = 34 mJ/cm^2, while at a higher F_L a rutile phase appeared [54]. The crystallization of divalent- or trivalent europium-containing oxide compounds or europium titanate [55] was not detected by XRD. The large mismatch in the ion radii of Eu³⁺ and Ti⁴⁺, 0.95 A and 0.68 A, respectively, may cause a crystal structural distortion and also make difficult the substitution of Ti⁴⁺ by Eu³⁺ within the anatase lattice [31].

3.3. Lu₂O₃:Eu

The PL emission spectra of Lu₂O₃:Eu film, shown in Figure 10, exhibited a strong luminescence signal of Eu³⁺ attributed to transition ⁵D₀→⁷F_j (j = 0–4) as well as a lower intensity signal from transition ⁵D₁→⁷F_j (j = 1–2). The excitation efficiency by ArF laser at λ_exc = 193 nm profited from Lu₂O₃ host lattice broad absorption band centered around λ = 210 nm [40]. Let us notice that the appearance of ⁵D₁→⁷F_j (j = 1–2) transitions is much less common in Eu³⁺-doped materials. These transitions could be observed in compounds with low-lattice phonon energy, e.g., Lu−O that leads to multiphonon relaxation [12,56]. They were also reported at films grown by chemical vapor deposition (CVD) [3]. Concerning ⁵D₀→⁷F_j (j = 0–2) transitions, the PL spectrum revealed similar features as the spectrum (excited at λ_exc = 254 nm) of PLD-prepared Lu₂O₃:Eu film with 5% Eu content [27]. Transitions were composed of peaks at 580.2 nm for ⁵D₀→⁷F₀ (Eu³⁺ in the noncentrosymmetric C₂ site [57]); 582.2 nm, 586.7 nm, 593.1 nm, and 600.3 nm for ⁵D₀→⁷F₁ (Eu³⁺ in the C₂ and S₆ (centrosymmetric) sites [40,58]); 611.0 nm, 612.6 nm, 624 nm, and 632 nm for ⁵D₀→⁷F₂ (Eu³⁺ exclusively in the C₂ site [40,58]); 650.2 nm and 664 nm for ⁵D₀→⁷F₃; 708 nm, 710.1 nm, and 713.5 nm for ⁵D₀→⁷F₄; 533.4 nm for ⁵D₁→⁷F₁; and 553.6 nm for ⁵D₁→⁷F₂. The most intense peak observed at 611 nm is characteristic of the cubic phase of rare earth sesquioxides [25,32,40].

We noticed the most substantial change in the film optical properties, shown in Figure 10 and Figure 11, after the first PLA shot. The most substantial increase of Eu³⁺ intensity emission and optical transmittance was obtained. The E_g value shifted from 5.47 eV to 5.51 eV, where the latter values corresponded to those reported for evaporated Lu₂O₃ films [43,59]. This may be attributed to both surface and structure modifications (as observed by AFM and XRD in Figure 12 and Figure 13, respectively). The asymmetric ratio increased from 7.9 to 9.5 reached at N_# = 2, which is reflecting the site symmetry decreasing related with film structure modification. The asymmetric ratio value was then decreasing to 9 and not changing for N_# = 4–7. PLA led to the maximum enhancement of 2.8 times for Eu³⁺ intensity emission at 611.0 nm. The energy transfer from Eu³⁺ residing in S₆ to Eu³⁺ in C₂ may take place because the levels of Eu³⁺ in the site S₆ lie above those for site C₂. It was reported, based on following the ⁵D₀→⁷F₁/⁵D₀→⁷F₀ ratio, that the rate of the energy transfer may change with the crystallites’ sizes (it is lower for larger crystallites) or because the location of Eu³⁺ in C₂ site is more common for small crystallites [57,60]. We did not observe any considerable variation of this ratio within the PLA process, although the crystallites’ size varied, as confirmed by XRD.

AFM analyses, shown in Figure 12, revealed similar morphology for as-deposited and PLA-treated film with typical S_q = 3.7 nm. Slight surface smoothing was observed after PLA and S_a decreased from 2.04 nm to 1.48 nm.

A substantial improvement of the crystalline structure after PLA treatment was evidenced by XRD diffraction patterns, depicted in Figure 13. While as-deposited film was rather amorphous, the film revealed a polycrystalline structure after seven PLA shots. The estimated crystallites’ size was 10.9 nm. The similar XRD pattern was observed for Lu₂O₃:Eu films deposited by PLD at a substrate temperature higher than 400 °C [25] and by magnetron sputtering with thermal postdeposition treatment at 900 °C for 2 h [20]. The reflections in the XRD pattern can be assigned to a body-centered cubic (I 21 3) structure of Lu₂O₃, according to JCPDS Card NO: 96-101-0596 file. The Eu₂O₃ phase was not detected in the patterns, confirming the Eu dopant was substituting Lu in Lu₂O₃ host, which was supported by the closed values of the ionic radii of 0.95 Å and 0.85 Å for Eu³⁺ and Lu³⁺, respectively [25]. The broad hump around 2θ = 21° is characteristic for amorphous fused silica substrate.

4. Conclusions

The in situ monitoring system of optical properties (photoluminescence and transmission) at pulsed laser annealing processing of Eu-doped oxide thin films was demonstrated. Eu-doped ZnO, TiO₂, and Lu₂O₃ films were fabricated by PLD at room temperature. The system allowed us to find quickly the optimal PLA conditions (ArF laser at λ = 193 nm, fluence, and number of PLA shots) for maximizing Eu³⁺ photoluminescence. The Eu³⁺ emission enhancement of 4.1×, 5× and 2.8×, after 800, 400, and 7 PLA shots, for ZnO, TiO₂, and Lu₂O₃ matrix was obtained, respectively. Transmission was markedly modified at Lu₂O₃:Eu film, where also E_g increased to 5.51 eV. On the other hand, ZnO:Eu and TiO₂:Eu films exhibited lowering E_g to 3.25 eV and 3.23 eV, respectively. TiO₂:Eu and Lu₂O₃:Eu as-deposited films exhibited an amorphous character that was modified by PLA to anatase and cubic nanocrystalline phase, respectively. In the case of ZnO:Eu, the application of PLA induced rather surface smoothening-related grains’ coalescence.