JEOS RP ISSN03 | Page 348

J. Eur. Opt. Society-Rapid Publ. 22, 33( 2026) 341

Table 2. Compared HfO2 thins films prepared by PEALD with different ALD tools.

Fig. 1. Wavelength dependence of the refractive index of the fabricated hafnia thin films.

sample. The surface roughness of the top interface was determined to be approximately 1.51 nm( root mean square), with an autocorrelation length of about 5.3 nm. These values correspond to dense, smooth oxide layers produced with PEALD and are comparable to HfO 2 thin films on silicon reference samples deposited under similar conditions with 0.8 nm rms( XRR) [ 29 ]. The low surface roughness is critical for optical applications, as elevated topography would introduce scatter losses and degrade coating performance in the ultraviolet spectral range.

To evaluate film quality and optical functionality, the extracted refractive index was compared with a previous PEALD-HfO 2 study. The thin film examined in this work is designated Type A, while the comparison sample is designated Type B( compare Fig. 1 and Table 2). To minimize the influence of impurity contributions on the optical properties, we compare only processes performed with the same precursor( TDMAH) same oxidizing agent( O 2-plasma) and at the same deposition temperature( 100 ° C), since these are the main parameters affecting carbon and nitrogen levels [ 30 ]. The low deposition temperature was chosen to reduce crystallization in the thin films, even though lower temperatures often lead to increased carbon contamination because of incomplete precursor oxidation.

The refractive index of both HfO 2 thin films exhibits a pronounced wavelength dependence, with a normal dispersion i. e. high values in the visible and near-ultraviolet spectral range and a gradual decrease toward longer wavelengths. In the ultraviolet region, the refractive index increases significantly as the photon energy approaches the optical band edge, while absorption remains negligible over a wide spectral range down to 230 nm( compare band edge of 5.6 eV [ 33 ]). This behaviour is highly advantageous for optical coatings and metasurfaces requiring strong phase modulation with minimal optical losses.

While both HfO 2 thin films demonstrate excellent optical properties and comparable transparent regions, Type A and Type B show slight deviations in refractive index. These differences can be attributed to two distinct sources: differences in metrology and deposition conditions.

For Type A thin films, the universal dispersion model( UDM) was employed to extract optical constants [ 34 ]. In contrast, optical constants of Type B thin films were

extracted from VUV transmittance and reflectance spectra using the Lorentz Calculator( LCalc) [ 41 ]. These different analytical methods can introduce systematic deviation, particularly in the VUV region where extinction is dominated by complex inter-band transitions. Nevertheless, both methods are sufficiently mature to provide reliable data and inherently capture the material response, including any influence of i. e. residual impurities.

Typically, different deposition parameters are the main source of such differences. The dominant factor separating the two film types are the different PEALD reactor configuration and associated plasma parameters, which are known to strongly influence HfO 2 film density and refractive index [ 31 ]. Type A films were deposited using an OpAL PEALD reactor with a remote Inductively Coupled Plasma( ICP) source, while Type B samples were produced on a SILAYO-ICP330 system( Sentech Instruments GmbH), equipped with a direct ICP configuration featuring a Planar Triple Spiral Antenna( PTSA) source. These distinct plasma architectures lead to markedly different energy input and ion bombardment characteristics. The SILAYO PTSA-ICP source delivers higher plasma power density and enhanced ion flux, enabling more effective energetic assistance during film growth. This results in denser films: Type B achieves 8.25 g / cm 3 compared to 7.90 g / cm 3 for Type A( Table 2). Film density is directly reflected in the refractive index; the denser Type B films exhibit higher n- values across the transparent spectral range. This densitydependent optical property improvement is a well-established phenomenon in oxide thin films and confirms that plasma parameter optimization is a critical control lever for tailoring optical performance in PEALD-HfO 2 systems.

X-ray diffraction measurements conducted at the deposition temperature( 100 ° C) indicate that both Type A and Type B films are X-ray amorphous( not shown here), consistent with low-temperature PEALD processes. However, X-ray amorphous coatings can contain nanocrystalline domains, short-range crystalline order, or point defects that influence optical properties. This may also explain the moderate deviations, especially observed below the absorption edge between the two film sets. The higher density of type B can encourage slight structural rearrangement, such as localized crystallization or structural densification, and affect the absorption profile below the band gap. Despite these structural subtleties, both films maintain low