1. Introduction

Thin film coatings for deep UV wavelengths are predominately produced by evaporation methods (thermal and E-beam). Factors limiting the performance of evaporated films include surface roughness, porosity, absorption, and defect density. To enhance the films, they are typically deposited at high substrate temperatures (above 300°C). The thermal stress created upon cooling down from the deposition temperature often restricts these films to be deposited on fluoride substrates.

Efforts to improve the quality and performance of deposited fluoride films includes the use of ion beam sputtering [1,2]. While the energy of ion beam sputtering removes some of the fluorine atoms during the sputtering process, this deficiency was compensated for by introducing a fluorine-based gas in the deposition environment.

A method of ion beam sputtered fluoride films suitable for a production environment using a commercially available dual ion beam sputtering system is presented in this work. The resulting optical and material properties as well as suitable substrates are discussed for ${\mathrm{AlF}}_3$, ${\mathrm{LaF}}_3$, and ${\mathrm{GdF}}_3$ films. Results for both single layer and multilayer coating designs are presented. Results are focused on the 193 nm wavelength of the ArF excimer laser.

2. Deposition system and parameters

A dual ion beam sputtering system (Veeco SPECTOR) was used to deposit films of ${\mathrm{AlF}}_3$, ${\mathrm{LaF}}_3$, and ${\mathrm{GdF}}_3$. Only the main deposition ion beam was used in this work. The assisting ion source was not used. Xenon sputtering gas was used with beam voltages and beam currents in the 700–900 V and 175–225 mA ranges, respectively. Sputtering with ion beams results in a fluorine-deficient film, and this contributes to absorption losses in the coating. To compensate for the fluorine depletion, ${\mathrm{NF}}_3$ gas was added as a chamber background gas. Previously reported studies used fluorine [1,2] or ${\mathrm{CF}}_4$ [3] gases. ${\mathrm{NF}}_3$ was chosen here because it is relatively less hazardous and has a lower cost than ${\mathrm{F}}_2$ or ${\mathrm{CF}}_4$.

Two deposition systems were used interchangeably in this work. One was configured with a load-locked single-axis rotation with a 300 mm diameter substrate fixture. The other was not load-locked and had a $4\times 200\mathrm{mm}$ diameter planetary fixture. No noticeable difference was observed between the films coated with these two systems.

No heat was added during the deposition process. Chamber temperatures during deposition remained low, at below 30°C for a short antireflective coating runs and below 40°C for longer high-reflective (HR) coating runs.

3. Results

Single layers of ${\mathrm{AlF}}_3$, ${\mathrm{LaF}}_3$, and ${\mathrm{GdF}}_3$ were coated on double-side polished UV-grade fused silica and ${\mathrm{CaF}}_2$ substrates, and on single-sided polished Si wafers. The deposition rates, refractive indices, extinction coefficients, and film stress are shown in Table 1. The refractive index and extinction coefficient were determined by measuring single layers of deposited fused silica and ${\mathrm{CaF}}_2$ substrates using an ellipsometer, and also by Cauchy model, fit to the transmittance data. Film stress of ${\mathrm{GdF}}_3$ and ${\mathrm{AlF}}_3$ was measured from single layers deposited on Si wafers.

Table 1. Single Layer Results

View Table | View all tables in this article

The transmittance of each single layer coating, compared with uncoated substrate, is shown in Fig. 1. In the wavelength range studied, loss in the ${\mathrm{GdF}}_3$ and ${\mathrm{AlF}}_3$ films are low, whereas ${\mathrm{LaF}}_3$ exhibit some loss.

 

Fig. 1. Transmittance data of single layer ${\mathrm{GdF}}_3$ (left, solid line), ${\mathrm{AlF}}_3$ (center, solid line), ${\mathrm{LaF}}_3$ (right, solid line), and uncoated substrates (dashed line).

Download Full Size | PDF

Three-layer antireflective (AR) coatings, optimized for 193 nm at normal incidence, were designed with ${\mathrm{LaF}}_3/{\mathrm{AlF}}_3$ and ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$ combinations. The AR coatings were deposited on both sides of double-side polished fused silica and ${\mathrm{CaF}}_2$ substrates. The coatings were run on time–power, meaning that no in situ layer endpoint monitoring was used. Figure 2 shows the measured transmittance data for the AR coated substrates. There was no measurable difference in the spectral performance of the AR coatings between the fused silica and ${\mathrm{CaF}}_2$ substrates. At 193 nm, the transmittance of the ${\mathrm{LaF}}_3/{\mathrm{AlF}}_3$ AR-coated substrates is 99.0%, and 99.1% for the ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$ AR-coated substrates. The corresponding measured reflectance was 0.2% for ${\mathrm{LaF}}_3/{\mathrm{AlF}}_3$ and 0.06% for ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$. The three curves in each graph in Fig. 2 show three sets of double-sided coated substrates, for a total of six consecutive coating runs. This demonstrates stable and repeatable deposition without in situ layer endpoint monitoring.

 

Fig. 2. Transmittance data of AR coating on both side of fused silica, ${\mathrm{LaF}}_3/{\mathrm{AlF}}_3$ (left) and ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$ (right).

Download Full Size | PDF

A high reflective (HR) coating, centered at 193 nm for normal incidence, with total coating thickness of 1.25–1.4 μm, was coated on fused silica and ${\mathrm{CaF}}_2$ substrates. Figure 3 shows the reflectance and transmittance data of the ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$ HR coatings. Again, there was no measurable difference between the two different substrates. At 193 nm, the measured reflectance was in the range of 97.9%–98.5% and transmittance was 0.44%–0.51%. These values show that, although there is some loss in the film at 193 nm, they are comparable with commercially available E-beam-deposited 193 nm HR coatings.

 

Fig. 3. Transmittance and reflectance data of ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$ HR coating. The two solid lines are reflectance on fused silica and ${\mathrm{CaF}}_2$ substrates. The dashed line is transmittance on fused silica, and the dotted line is transmittance on ${\mathrm{CaF}}_2$.

Download Full Size | PDF

Atomic force microscopy (AFM) analyses of the substrate surfaces were obtained before and after coating to measure the increase of surface roughness due to the coating [Fig. 4]. One $2\mathrm{\mu m}\times 2\mathrm{\mu m}$ area was imaged near the center of each sample. Table 2 summarizes the AFM analysis results of the single layer, antireflective (AR), and HR coatings on super polished fused silica substrates. The single layer of ${\mathrm{GdF}}_3$ coating and the AR coatings increase the surface roughness by approximately 0.1 nm or less, which is typical of ion beam sputtering of oxide films. Single layers of ${\mathrm{AlF}}_3$ shows some increase in roughness, of about 0.6 nm. The combination of ${\mathrm{GdF}}_3/{\mathrm{AlF}}_3$ gives a significant increase in percentage roughness compared with ${\mathrm{LaF}}_3/{\mathrm{AlF}}_3$ coatings. Further studies are needed to understand the cause of this increase.

 

Fig. 4. AFM images of coated surfaces on fused silica substrates.

Download Full Size | PDF

Table 2. Surface Roughness (before and after Coating)

View Table | View all tables in this article

The residual optical losses still present in the films is thought to be caused by surface roughness, which causes scattering on the surface, especially of ${\mathrm{AlF}}_3$, and due also to the fluorine deficiency, as it was unable to fully compensate for the fluorine depletion in the films.

To assess coating durability, the single layers, AR, and HR coatings of ${\mathrm{AlF}}_3$, ${\mathrm{LaF}}_3$, and ${\mathrm{GdF}}_3$ were tested for adhesion, analyzed before and after cleaning, and subjected to thermal cycling. The adhesion of all the single layers, AR, and HR coatings passed a tape test on fused silica, ${\mathrm{CaF}}_2$, and Si substrates. Cleaning the coated surfaces with methanol-soaked cleanroom-grade wipes did not cause any observable physical degradation or changes to the optical performance. Single layers were subjected to a temperature ramp, from room temperature to 75°C, followed by a dwell at 75°C for 2 hours, and then cooled down to room temperature. The films did not show any observable physical degradation when inspected under high intensity fiber light and no change was measured in their optical performances.

Additional coatings of more than 150 layers and total thickness of greater than 6 μm have been deposited on both ${\mathrm{CaF}}_2$ and fused silica substrates with no optical or physical performance differences between the coatings on the two substrates.

4. Conclusion

Ion beam-sputtered fluoride films can be made with the expected low roughness and repeatability inherent to ion beam sputtering. The low temperature of the process allows for the use of standard fused silica substrates with optical and physical performances equivalent to ${\mathrm{CaF}}_2$ substrates.