Open Access
Add to BibSonomy
Abstract
The fabrication of highly reflective aluminum coatings is still an important part of current research due to their high intrinsic reflectivity in a broad spectral range. By using thin seed layers of Cu, CuOx, Cr, CrOx, Au, and Ag, the morphology of sputtered (unprotected) aluminum layers and, consequently, their reflectance can be influenced. In this long-term study, the reflectance behavior was measured continuously using spectrophotometry. Particular seed layer materials enhance the reflectance of aluminum coatings significantly and reduce their long-term degradation. Combining such seed layers with evaporation processes and suitable protective layers could further increase the reflectance of aluminum coatings.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
For the fabrication of highly reflective broadband mirrors, mainly aluminum (Al), silver (Ag), and gold (Au) thin films are used. However, only mirrors with an Al coating provide a high intrinsic reflectivity from the far-ultraviolet (FUV) to far-infrared (FIR) spectral range. Since the technological feasibility of such thin films, Al coatings have been applied for numerous optical components [1–6], particularly as mirrors with suitable protective layers [7–11]. As an example, the FUV mirrors for the proposed NASA space telescopes “Large Ultraviolet Optical Infrared Surveyor” (LUVOIR) [12] and “Habitable Exoplanet Imaging Mission” (HabEx) [13] are based on protected Al layers. Nevertheless, several characteristic features must be considered to fabricate Al coated mirrors with high reflectance.
Highly reflective Al coatings with protective layers require specific deposition conditions. First, in order to fabricate highly reflective Al coatings, a low base pressure within the deposition chamber, low process pressure, high Al deposition rate, and substrate temperatures close to room temperature are necessary [14]. Second, protective layers are deposited in-vacuo immediately after the Al deposition to protect the Al surface from oxidation [15,16]. Otherwise, the self-passivating aluminum(III) oxide (Al2O3) top layer would reduce the reflectance in the FUV spectral range massively [17]. Metal fluoride protective layers are particularly suitable because they are transparent in the FUV. However, fluoride layers deposited at low temperatures show a lower density than the bulk material, indicating a porous structure [18,19]. Compact protective coatings require high mobility of adatoms, which is achieved in evaporation processes by heating the substrates. For sputtering techniques, the higher energy of the sputtered atoms provides this required mobility [20].
Sputtering techniques offer advantages compared to evaporation processes regarding Al coatings. Magnetron sputtering provides a precise thickness control, lateral homogeneity, scalability over large areas, and reliable adhesion strength to substrates [21,22]. Kiyota et al. realized Al sputter rates of about 3.7 nm/s and achieved Al coatings with negligible oxygen inclusions [23]. By using so-called liquid phase targets with a pulsing unit, specific sputter rates up to 24.5 nm/(kW·s) are possible [24]. Furthermore, sputtering enables the deposition of closed protective layers with a high packing density even at room temperature [20]. By using ion beam sputtering (IBS), which gained more and more importance in the last years, the energy of sputtered atoms can be controlled independently of the sputter rate. Fernandez et al. combined evaporated and IBS sputtered layers of both Al and magnesium fluoride (MgF2) [25]. The authors show that the reflectance of these mixed layer systems is almost comparable to all-evaporated layers at short wavelengths. However, the layers exhibit a high surface roughness of 3.4 nm root-mean-square (rms) and sharply dropping reflectance for wavelengths above 140 nm. The high surface roughness of this Al layer might be significantly reduced using thin seed layers.
Seed layers can improve the reflectance of Al coatings. Current research for Al-based FUV mirrors mainly concerns the optimization of deposition parameters and temperature in evaporation processes. Recently, Stempfhuber et al. presented a different approach using so-called seed layers [26,27]. Their thin titanium (Ti) layer underneath the unprotected Al layer reduced the Al surface roughness significantly and enhanced the reflectance in the FUV spectral range. Most recently, Larruquert et al. combined Ti seeded Al coatings with an MgF2 protection layer [28]. The surface roughness reduction at the Al/MgF2 interface significantly increased the reflectance from the FUV to the infrared (IR) spectral range compared to unseeded Al/MgF2 coatings. Especially, the absorption band around 160 nm due to surface-plasmon coupling was almost completely suppressed. Our work concentrates on sputtered Al layers with different thin seed layers to enhance their reflectance as well. In order to evaluate the long-term stability, as essential for space telescope missions, the reflectance of these seeded Al coatings was studied over about 4 years.
This article aims to understand the reflectance behavior of sputtered Al coatings with different seed layer materials. For this purpose, thin seed layers of copper (Cu), copper oxide (CuOx), chromium (Cr), chromium oxide (CrOx), gold (Au), and silver (Ag) were applied to fabricate highly reflective Al coatings. Here, the optimized process parameters for the deposition of highly reflective Al layers were kept constant. These seed layers are promising to modify the nucleation dynamics and morphology of the Al layer so that larger and flat grains with reduced grain boundaries and surface roughness are formed. In order to investigate the influence of the seed layers on long-term stability, the Al coatings were fabricated without a protective layer. Our study shows that particular seed layers significantly and sustainably enhance the reflectance of Al coatings in a broad spectral range.
The results of this article are divided into two main parts. In the first part, the reflectance of bare and seeded Al coatings is investigated. Therefore, spectrophotometric measurements were performed over about 4 years. The second part is dedicated to the layer properties of the Al coatings concerning their reflectance behavior. These properties were studied by scanning electron microscopy (SEM), atomic force microscopy (AFM), X-ray diffraction (XRD), and four-point probe resistivity measurements. Additionally, the adhesive strength was evaluated by tape tests.
2. Materials and methods
As substrate material, single-crystalline silicon (Si) wafers with a native silicon dioxide (SiO2) top layer of 1–2 nm thickness were used. These Si wafers were p-doped, (100)-oriented, double-side polished with an AFM surface roughness (1 × 1 µm2) of about 0.15 nm rms, and 25 × 25 × 0.675 mm3 in size. For cleaning, a multi-stage, ultrasonic-assisted bath cleaning system (Elma Schmidbauer, Germany) was used. The cleaning procedure consisted of alternating surfactants and water (H2O) baths, followed by a final bath of deionized ultrapure H2O.
The depositions were performed with an industrial direct current (DC) magnetron sputtering system MRC-943 (Kenotec SRL, Italy). This system has three sputtering targets with a size of 120 × 380 mm2 and a high-frequency etching station arranged side by side. The magneton cathodes were pulsed with a DC pulse frequency of 100 kHz. Substrates are transferred on a substrate pallet (300 × 300 mm2) via load lock with a lifting mechanism into the deposition chamber. At a distance of about 60 mm, the substrates move horizontally below the targets. A dry-running scroll pump (TriScollTM 600, Varian, USA) and two cryogenic pumps (CTI-Cryogenics On-Board 8 Cryopump, Edwards, UK) ensure a base pressure of (2–4) × 10−5 Pa.
Before the aluminum deposition, seed layers of Cu, Cr, Au, and Ag, as well as CuOx and CrOx, were deposited. The poly-crystalline metal seed layers were sputtered using pure metal targets (purity level better than 3N5). The amorphous metal oxide seed layers of CuOx and CrOx were sputtered reactively using metal targets and molecular oxygen (O2, purity level 5.0). The nominal thickness of all seed layers was varied between 3 nm and 10 nm. In the following, however, the nominal seed layer thickness always refers to 3 nm for Cu, CuOx, Cr, CrOx, and Au and 6 nm for Ag due to the high Ag sputtering yield. Before the seed layer deposition, the substrates were cleaned in-situ using Argon (Ar) plasma to remove H2O and carbon residues from the surface. The plasma was ignited for 60 s with a power of 180 W and an Ar (purity level 6.0 = 99.9999%) flow rate of 35 sccm. In Table 1, the process parameters for the Ar plasma pre-treatment, seed layers, and Al layer deposition are summarized.
Table 1. Process Parameters for the Argon (Ar) Plasma Pre-treatment, Seed Layers, and Aluminum (Al) Layer Deposition Using Direct Current (DC) Magnetron Sputtering With Ar as Sputtering Gas.a
Immediately after the seed layer deposition, the poly-crystalline Al layer was deposited in-situ using a pure Al target (purity level 5N = 99.999%) and Ar as the sputtering gas. In order to achieve highly reflective Al coatings, the lowest possible Ar flow rate of 4 sccm, high cathode power of 4000 W, no bias voltage (0 V), and no substrate heating were applied (see also Table 1). The process pressure within the deposition chamber was about 0.1 Pa. The Al deposition rate corresponds to about 3.7 nm/s, whereas all deposited Al layers exhibit a thickness of (75 ± 2) nm determined by X-ray reflectometry (XRR). For all layer systems, the process parameters for the Al layer were identical. After fabrication of the Al coatings, the samples were stored under ambient conditions, meaning at room temperatures of (23 ± 5) °C and (45 ± 20)% humidity.
For spectrophotometric characterization of the coatings, a commercial double-beam grating spectrometer Lambda 850 (Perkin Elmer, USA) was used. The instrument is equipped with a deuterium and halogen lamp, 60 mm integrating sphere, photomultiplier tube, and an InGaAs-detector. For this study, the reflectance was measured from 200 nm to 1200 nm wavelength without using sophisticated vacuum-ultraviolet (VUV) equipment. The angle of incidence was always 6° so that the light polarization is negligible. The measurement uncertainty varies from about 0.15% for the visible (VIS) and IR to about 0.25% for the ultraviolet (UV) spectral range. Regressions were performed to obtain accurate reflectance values at exemplary selected wavelengths (250 nm, 400 nm, and 1000 nm). By linear or quadratic regressions (λ ± 15 nm) around the selected wavelengths, the noise of the spectrophotometric measurements is negligible. For better visualization, the shown reflectance spectra are slightly smoothed.
The topography of the Al coatings was investigated by SEM and AFM. For SEM analysis, a Schottky field emission SEM Sigma (Carl Zeiss Microscopy, Germany) with an acceleration voltage of 5 kV and a working distance of about 5 mm was used. The in-lens detector provides a resolution of < 2 nm. The AFM measurements were performed with a Dimension Icon (Bruker, USA) equipped with a NanoScope V controller under ambient conditions in an ISO class 7 cleanroom. Using sample areas of 1 × 1 µm2 scanned in tapping mode, the surface roughness was calculated in root-mean-square (rms).
XRR and XRD were performed with the X-ray diffractometer D8 Discover (Bruker AXS, Germany) in Bragg-Brentano geometry. For these measurements, Cu Kα radiation (λ = 0.154 nm) with an acceleration voltage and cathode current of 40 kV and 40 mA, respectively, was used. In order to determine the layer thicknesses, the XRR data were analyzed with the Bruker Leptos 7 software package. The Al crystallite sizes were estimated from the Al(111) peak using Scherrer’s equation with a shape factor of K ≈ 0.94 [29].
The specific electrical resistivity was determined using four-point probe resistivity measurements. Digital multimeters Keithley 2601 (Keithley Instruments, USA) and Keithley 2000 were used as a current source and for voltage measurement, respectively. The contact tips with a distance of 2.54 mm were placed in the sample middle. The specific electrical resistivity was calculated using a correction factor of G = 4.344 [30].
The adhesive strength of the Al layers to the substrate was evaluated by tape tests after about 4 years. Using a 3M 853 TAPE (3M, USA), these tape tests were performed according to ISO 9211–4 [31] and severity levels 02 and 03. If the Al coating was adhesive after a single tape test with severity level 02, the test was repeated at the same area with severity level 03.
3. Experimental results
3.1 Reflectance of bare and seeded Al thin films
In order to understand the influence of different seed layers on the reflectance of aluminum coatings, spectrophotometric measurements were performed. Figure 1(a) shows the specular reflectance of bare Al coatings, meaning without a seed layer, from 200 nm to 1200 nm wavelength. For this bare Al coating, the reflectance as-deposited, after 7 days, and after about 4 years is illustrated. At an exemplary selected wavelength of 400 nm, the as-deposited reflectance R0 and corresponding reflectance losses ΔR after 7 days and after about 4 years are calculated.
Fig. 1. Specular reflectance of Al coatings from 200 nm to 1200 nm wavelength. (a) For a bare Al coating, the reflectance as-deposited (black), after 7 days (red), and after about 4 years (blue) is shown. At a wavelength of 400 nm, the as-deposited reflectance R0 and corresponding reflectance losses ΔR after 7 days and after about 4 years are given. (b) As-deposited reflectance of seeded Al coatings (red, blue, green in solid, and dashed) compared to a bare Al coating (black).
The reflectance of bare Al coatings decreases significantly in the VIS and UV spectral range during storage on air. Figure 1(a) depicts the characteristic spectra of unprotected Al coatings with an oxidized surface. In the near-infrared (NIR) spectral range, the reflectance of the bare Al coating hardly degrades. At a wavelength of 400 nm, the as-deposited reflectance R0 = 91.7% results in reflectance losses of about –0.3% after 7 days and −0.6% after about 4 years. With decreasing wavelength, the reflectance losses increase, while the characteristic profile remains the same. The reflectance losses are primarily caused by oxidation of the Al surface to Al2O3 by contact with air. Due to the self-passivation of this surface, the formation of Al2O3 and corresponding reflectance losses stagnate over time. Overall, the reflectance spectra and degradation of the bare Al coating are comparable to other studies [17,32,33].
Compared to bare Al coatings, particular seed layer materials enhance the reflectance in a broad spectral range. Figure 1(b) shows the as-deposited reflectance of Al coatings with Cu, CuOx, Cr, CrOx, Au, and Ag seed layers compared to a bare Al coating. Seeded Al coatings with Cr, CrOx, and Au exhibit an increased reflectance in the whole spectral range. In contrast, the reflectance of the Cu/Al and Ag/Al layer systems is slightly reduced. For the CuOx seeded Al coating, the reflectance spectrum is almost identical to the bare Al coating without a seed layer. Only in the region of aluminum interband transitions at 830 nm and below 300 nm are minor deviations. The nominal thickness of all seed layers was varied between 3 nm and 10 nm. However, no differences in the as-deposited reflectance spectra of Al coatings with different seed layer thicknesses were observed for the individual seed layer materials. Therefore, only the Al coatings with the thinnest seed layers, namely 3 nm for Cu, CuOx, Cr, CrOx, and Au and 6 nm for Ag, were further investigated in this study. At a wavelength of 400 nm, Cr and Ag seeded Al coatings result in reflectance changes of about +0.3% and -0.3%, respectively, compared to the bare Al coating. These as-deposited reflectance differences are relatively small but enlarge due to atmospheric degradation of the Al coatings.
After about 4 years, Al coatings with seed layers of CuOx, Cr, CrOx, and Au show reflectance losses barely. Figure 2 illustrates the specular reflectance of Al coatings with different seed layers as-deposited and after about 4 years from 200 nm to 1200 nm wavelength. At a wavelength of 400 nm, the as-deposited reflectance R0 and corresponding reflectance losses ΔR after about 4 years are given. For Al coatings with CuOx, Cr, CrOx, and Au seed layers, this as-deposited reflectance is about 91.8% to 92.0% with reflectance losses of ΔR ≈ −0.1%, which are within the measurement uncertainty. As with the bare Al coating, the reflectance losses increase with decreasing wavelength, although much less pronounced. Thus, these seeded Al coatings provide an enhanced reflectance with reduced degradation compared to the bare Al coating without a seed layer.
Fig. 2. Specular reflectance of (a) Cu, (b) CuOx, (c) Cr, (d) CrOx, (e) Au, and (f) Ag seeded Al coatings as-deposited (black) and after about 4 years (red) from 200 nm to 1200 nm wavelength. At a wavelength of 400 nm, the as-deposited reflectance R0 and corresponding reflectance losses ΔR after about 4 years are given.
Aluminum coatings with seed layers of Cu and Ag degrade significantly, as Fig. 2 illustrates. After about 4 years, the reflectance of these seeded Al coatings is substantially reduced in the whole spectral range from 200 nm to 1200 nm wavelength. In contrast to the bare Al coating and seeded Al coatings with CuOx, Cr, CrOx, and Au, the reflectance in the NIR spectral range degraded as well. The reflectance losses increase with decreasing wavelength. At a wavelength of 400 nm, the reflectance losses for the Cu and Ag seeded Al coatings are about −1.2% and −0.8%, respectively. Overall, the Cu/Al and Ag/Al layer systems exhibit a reduced reflectance with elevated degradation compared to the bare Al coating.
The reflectance losses of the bare and seeded Al coatings proceed with exponential decay. Figure 3 shows the reflectance of bare and seeded Al coatings at exemplary selected wavelengths of 250 nm, 400 nm, and 1000 nm, depending on time after deposition. As already described, the CuOx, Cr, CrOx, and Au seeded Al coatings still exhibit a high long-term reflectance after about 4 years. In contrast, the Cu/Al and Ag/Al layer systems degrade significantly compared to bare Al coatings. Their reflectance losses after deposition decay exponentially, as the fitted (dashed) lines illustrate (note the logarithmic time scale). The reflectance approaches a constant value, and the reflectance losses saturate, which results from self-passivation of the Al surface with Al2O3. However, seed layers affect this oxidation process and the corresponding reflectance losses.
Fig. 3. Reflectance of bare and seeded Al coatings at wavelengths of 250 nm (black), 400 nm (red), and 1000 nm (blue) depending on time after deposition. Compared to (a) bare Al coatings, the (b) CuOx, Cr, CrOx, and Au seeded Al coatings still exhibit a high reflectance after about 4 years, whereas the layer systems of (c) Cu/Al and Ag/Al degrade significantly. The dashed lines represent a fitted exponential decay; note the logarithmic time scale.
Seed layers influence the reflectance and reflectance losses of Al coatings sustainably, as Fig. 3 shows. The bare (unprotected) Al coatings degrade at atmospheric conditions over time, with the reflectance losses stagnating after a maximum of about 100 days. The inset (with linear time scale) of Fig. 3(a) demonstrates that the major reflectance changes occur directly after the Al deposition. The CuOx, Cr, CrOx, and Au seeded Al coatings behave similarly to each other. Compared to the bare Al coating, these seeded layer systems exhibit an enhanced reflectance with significantly reduced reflectance losses. After about 100 days, their reflectance also reaches a constant value so that the high reflectance remains. The Cu/Al and Ag/Al systems also behave similarly to each other. Unfortunately, these seeded Al coatings exhibit a reduced reflectance with elevated reflectance losses. However, these reflectance losses become completely apparent only after 2 to 3 years. According to the fitted exponential decays from Fig. 3, the as-deposited reflectance ${\rm R}_{0}^{{\rm fit}}$, corresponding reflectance losses $\Delta {\rm R}^{\rm fit}$, and long-term reflectance ${\rm R}_{\rm t}^{{\rm fit}}$ after about 4 years are summarized in Table 2 quantitatively.
Table 2. Reflectance Behavior of Bare and Seeded Al Coatings at Wavelengths of 250 nm, 400 nm, and 1000 nm.a
To summarize, the influence of different seed layers on the reflectance behavior of Al coatings was investigated by spectrophotometric measurements over about 4 years. For bare Al coatings without seed layer, the reflectance decreases significantly in the VIS and UV spectral range. However, seeded Al coatings with CuOx, Cr, CrOx, and Au provide an enhanced reflectance with reduced degradation compared to bare Al coatings. In contrast, the Cu/Al and Ag/Al layer systems show a slightly reduced reflectance with elevated degradation. The reflectance losses of the bare and seeded Al coatings proceed with exponential decay.
3.2 Relationships between reflectance and layer properties
After a detailed study of the reflectance behavior, the relationships between reflectance and layer properties of bare and seeded aluminum coatings were investigated. Therefore, the topography of the Al coatings was evaluated by SEM and AFM images. The size of the Al crystallites within the layer was estimated using XRD and Scherrer’s equation. By using four-point probe resistivity measurements, the specific electrical resistivity was determined, which could provide information about impurities within the Al layer or the conditions at the grain boundaries. Figure 4 shows SEM images (top view) of Al coatings with different seed layers. Besides, the AFM surface roughness (1 × 1 µm2) is given.
Fig. 4. Scanning electron microscope (SEM) images (top-view) of (a) Cu, (b) CuOx, (c) Cr, (d) CrOx, (e) Au, and (f) Ag seeded Al coatings. For comparison, the SEM image of a (g) bare Al coating is integrated into the image of the CuOx/Al layer system. The scale bar in the bottom right corner is valid for all SEM images, including the image of the bare Al coating. Additionally, the atomic force microscope (AFM) surface roughness (1 × 1 µm2) in root-mean-square (rms) is given.
Particular seed layers modify the topography of Al coatings significantly, as Fig. 4 illustrates. The Al coatings with CuOx, CrOx, and Ag seed layers exhibit a similar surface structure as the bare Al coating. Their surfaces appear uniform overall, and the individual grains with dome-shaped capping are clearly visible. Based on the SEM images, the lateral grain elongation is about 35 nm to 45 nm for the bare and CrOx seeded Al coatings, about 30 nm to 45 nm for the CuOx/Al layer system, and about 22 nm to 45 nm for the Ag/Al layer system. Larger grains (>50 nm) and small pores (<20 nm) occasionally appear on the surface. For the Cr seeded Al coating, the typical dome-shaped capping of the grains is less pronounced. These grains elongate lateral about 33 nm to 50 nm, whereby larger grains (>70 nm) consisting of several smaller grown-together grains occur sporadically. In contrast, the surface of the Cu/Al layer system shows grains with a smooth floe structure. Their lateral grain elongation is about 27 nm to 40 nm with weakly pronounced grain boundaries, whereby larger grains (>50 nm) occur sporadically. The Au seed layer causes the most significant changes in topography compared to the other Al coatings. On the one hand, the grains appear as flat floes with weakly pronounced grain boundaries and lateral elongation of about 30 nm to 70 nm. On the other hand, some grains with a dome-shaped capping occur, which elongate about 25 nm to 50 nm. Sporadically, also larger grains (>80 nm) with flat or dome-shaped cappings occur. Thus, seed layers influence the surface structure of the Al coatings, but also their surface roughness.
By applying seed layers, the surface of Al coatings can be smoothed. In Fig. 4 and Table 3, the surface roughness of the bare and seeded Al coatings is given. Changes in the surface structure of the Al coatings are partly associated with a change in their surface roughness. The CuOx and CrOx seeded Al coatings exhibit a similar surface structure and surface roughness with about 1.7 nm rms as the bare Al coating. Despite the similar surface structure, the surface roughness of the Ag/Al layer system is the highest at 2.2 nm rms. For the Cr and Cu seeded Al coatings, the surface roughness decreases to 1.5 nm rms and 1.2 nm rms, respectively, because their dome-shaped grains are less pronounced. The smooth surface areas of the Au seeded Al coating exhibits a roughness of about 0.8 nm rms. Due to the larger dome-shaped grains, the overall surface roughness is about 1.6 nm rms. However, no significant correlation exists between the surface roughness of these seeded Al coatings and their reflectance or reflectance losses.
Table 3. Layer Properties of Bare and Seeded Al Coatings.a
No significant correlation is observed between the crystallite size within the Al layers and their reflectance behavior. Table 3 lists the estimated Al crystallite sizes of the bare and seeded Al coatings using XRD and Scherrer’s equation. Since XRD provides a lower estimation of the crystallite size, these values are always smaller than the lateral grain sizes obtained by SEM. Nevertheless, the tendencies and determined sizes by SEM and XRD correspond well with each other, indicating that the external appearance and internal structure of the Al layers are similar. The crystallite sizes within the Al layers range from about 27 nm for the Ag/Al layer system to about 39 nm for the Au/Al layer system. However, no significant correlation exists between the crystallite size of the Al coatings and their as-deposited reflectance, reflectance losses, or long-term reflectance after about 4 years.
The specific electrical resistivity of Al coatings correlates with their reflectance behavior. Figure 5(a) shows the as-deposited reflectance of bare and seeded Al coatings depending on their specific electrical resistivity for exemplary selected wavelengths of 250 nm, 400 nm, and 1000 nm. For these wavelengths, the as-deposited reflectance (and long-term reflectance after about 4 years) of the Al coatings increase with decreasing resistivity. Especially for shorter wavelengths, this correlation is more pronounced. The Cr/Al layer system shows the highest as-deposited reflectance and lowest specific electrical resistivity with about 5.1 × 10−8 Ω·m, close to the value for Al bulk material of 2.5 × 10−8 Ω·m [34]. For the Ag/Al layer system, the behavior is reversed with a resistivity of about 6.9 × 10−8 Ω·m. This empirical relationship is not strictly linear, but a moderate to strong correlation is clearly evident. Notably, this correlation was also observed in our previous experiments where the process parameters (e.g., the cathode power and Ar flow rate) of bare Al coatings were varied. This correlation is undoubtedly not theoretically derivable, like the Hagen-Rubens relation, which is not valid in this spectral range from 200 nm to 1200 nm [35]. This empirical relationship probably results from the unique properties of thin Al coatings, which will be discussed in detail later.
Fig. 5. (a) As-deposited reflectance of bare and seeded Al coatings depending on their specific electrical resistivity for wavelengths of 250 nm (black), 400 nm (red), and 1000 nm (blue). With decreasing resistivity, the reflectance increases, especially for shorter wavelengths. (b) Reflectance losses after about 4 years of Al coatings depending on their as-deposited reflectance. The higher the reflectance, the lower the corresponding reflectance losses. Additionally, the seed layer materials are indicated.
The higher the as-deposited reflectance, the lower are corresponding reflectance losses for short wavelengths. Figure 5(b) shows the reflectance losses after about 4 years of Al coatings depending on their as-deposited reflectance for wavelengths of 250 nm, 400 nm, and 1000 nm. The reflectance losses decrease with increasing as-deposited reflectance (and long-term reflectance after about 4 years) of the Al coatings. Especially for shorter wavelengths, this empirical relationship is more pronounced. Based on the as-deposited reflectance of bare and seeded Al coatings, the expected degradation at atmospheric conditions can be estimated. This correlation is also not strictly linear, but a strong correlation is clearly evident. Notably, this empirical relationship was also observed in our previous experiments with the process parameters of bare Al coatings varied.
The bare and seeded Al coatings are adhesive to the substrate. Finally, the adhesion strength of the Al coatings on Si wafers was evaluated by tape tests according to ISO 9211–4 [31] after about 4 years. All bare and seeded Al coatings were adhesive after a single tape test with severity level 02, meaning a quick tape removal. The tape test was repeated a second time at the same sample area with severity level 03, meaning a snap removal of the tape. Again, all Al coatings were entirely adhesive to the substrate indicating a solid bond.
To sum up, the layer properties of Al coatings concerning their reflectance behavior were investigated by SEM, AFM, XRD, four-point probe resistivity measurements, and tape tests. Seeded Al coatings with CuOx, CrOx, and Ag show a similar surface structure as bare Al coatings. Their surface appears uniform, and the individual grains with dome-shaped capping are clearly visible by SEM. For Cr and Cu seeded Al coatings, the surface is smoothed because their dome-shaped grains are less pronounced. For the Au/Al layer system, the grains appear as flat floes with weakly pronounced grain boundaries, whereas some grains with a dome-shaped capping occur. The specific electrical resistivity of the Al coatings correlates with the reflectance behavior. Furthermore, the higher the as-deposited reflectance, the lower are corresponding reflectance losses of the adhesive Al coatings.
4. Discussion
Thin aluminum coatings are essential for numerous optical components due to their high and broadband intrinsic reflectivity. Highly reflective and adhesive Al coatings were deposited by magnetron sputtering using a low process pressure, high deposition rate, no bias voltage, and substrate temperatures near room temperature. With suitable thin seed layers, the morphology of Al layers can be modified to improve their reflectance behavior.
Particular seed layers enhance the reflectance of Al coatings significantly and reduce their long-term degradation. Figure 1(a) shows that the reflectance of bare (unprotected) Al coatings decreases over time due to the native formation of a self-passivating Al2O3 top layer. The reflectance spectra and degradation are comparable to other studies [17,32,33]. Seed layers of Cr, CrOx, and Au enhance the as-deposited reflectance of Al coatings in the spectral range from 200 nm to 1200 nm wavelength, as Fig. 1(b) illustrates. The reflectance of these seeded Al coatings is comparable to (unprotected) evaporated Al coatings [9]. In contrast, Cu and Ag seed layers reduce the as-deposited reflectance slightly compared to the bare Al coating; the CuOx/Al layer system behaves almost identical. The seeded Al coatings also exhibit a different degradation behavior over about 4 years, as shown in Fig. 2. Aluminum coatings with CuOx, Cr, CrOx, and Au seed layers display minor reflectance losses under atmospheric conditions. Compared to the bare Al coating, the degradation of Cu and Ag seeded Al coatings is elevated. The time-dependent reflectance behavior of these Al layer systems is presented both qualitatively and quantitatively in Fig. 3 and Table 2, respectively. For all coatings, the reflectance losses proceed with exponential decay, which is consistent with the logarithmic growth of the native Al2O3 top layer [32]. However, the reflectance losses of the Cu/Al and Ag/Al layer systems reveal only after 2 to 3 years.
The topography of seeded Al coatings does not explain their reflectance behavior. Figure 4 illustrates that seed layers modify the surface structure and surface roughness of Al layers. The CuOx, CrOx, and Ag seeded Al coatings exhibit a similar surface structure compared to the bare Al coating. Only the surface roughness of the Ag/Al layer system is slightly higher at 2.2 nm rms (see also Table 3), which could explain the reduced as-deposited reflectance at shorter wavelengths. In contrast, the Cu/Al layer system is significantly smoothed with 1.2 nm rms but not reproduced in the as-deposited reflectance. For Cu seeded Al coatings, Stempfhuber et al. reported an increased surface roughness and decreased reflectance [26]. However, the total scatter-induced reflectance losses up to 200 nm wavelength are relatively small [36]. For a surface roughness of 1.2 nm rms and 2.2 nm rms, the total scatter-induced reflectance losses are about 0.1% and 0.5% at 400 nm, and about 0.4% and 1.2% at 250 nm wavelength, respectively. Because the deposition parameters for the Al layer were always identical, interactions of Al and the seed layer materials probably cause the reflectance differences.
Seed layers affect the wetting behavior and initial nucleation of aluminum. The morphology and, consequently, the reflectance of the seeded Al coatings depends on both the intrinsic properties of the seed layers and the Al layer. Especially, the surface free energy (SFE) and surface roughness of the thin seed layers combined with the Si substrate play an essential role. The SFE of Cu, Cr, Au, and Ag surfaces is about 1.83 J/m2, 2.30 J/m2, 1.50 J/m2, and 1.25 J/m2, respectively [37]. Compared to the Si wafer with native SiO2 top layer (SFE ≈ 53 mJ/m2 [38]) and Al (SFE ≈ 1.16 J/m2 [37]), this could improve the wetting behavior of aluminum on the seed layers, decrease their diffusion, and lead to more nucleation sites [39]. Notably, the high SFE of Cu and Cr might cause the smoothed dome-shaped Al grains and surface roughness. Especially Cr serves as an adhesion layer for many substrate materials. Unfortunately, the topography of the bare seed layers was not investigated in detail. Future research could investigate these seed layers and their properties more precisely to establish a holistic comprehension of the property inter-relationships between the seed and Al layers.
The bond dissociation energy (BDE) of Al and seed layer materials may influence the Al crystallite size. In addition to the process parameters and thickness of the Al layer, their SFE, BDE, surface- and bulk-diffusion must be considered. The BDE of Cu-Al, Cr-Al, Au-Al, and Ag-Al is about 227 kJ/mol, 223 kJ/mol, 326 kJ/mol, and 184 kJ/mol, respectively [40]. With increasing BDE, the Al crystallite sizes determined by XRD also increase (see also Table 3). Regarding the Al-Al bonds (BDE ≈ 264 kJ/mol), the deposited Al atoms preferentially bond to the seed layer or among each other with higher or lower BDE, respectively. In particular, the Au-Al bond with about 326 kJ/mol is solid and could provide the basis for the large and flat floe-like Al grains. Nevertheless, the topography of the Au seed layer must also be considered. Maniyara et al. obtained a 3-nm-ultra-thin and closed Au layer using a 1 nm thin Cu seed layer [41]. Future research might apply such a Cu-Au seed layer to obtain smooth Al coatings with floe-like Al grains, fewer grain boundaries, and even higher reflectance. Notably, the Au/Al layer system shows no indications of the formation of Au-Al intermetallic compounds (white or purple plague) on Si wafers or transparent borosilicate glass B270 (neither front nor backside). Besides, the Au seeded Al coating is adhesive and hardly exhibits reflectance losses after about 4 years. In contrast, the BDE of Ag and Al is relatively low, which could explain the small Al grains. Unfortunately, no investigations about the hetero-surface diffusion of Al atoms or clusters have been performed so far [42].
The specific electrical resistivity enables an estimation of the expected reflectance behavior of Al coatings. In this study, no significant correlations between the reflectance behavior, surface roughness, and crystallite size of our highly reflective Al coatings were found. However, the specific electrical resistivity of the bare and seeded Al coatings correlates with their reflectance behavior, which in turn correlates with the reflectance losses, as Fig. 5 illustrates. In other words, the lower the specific electrical resistivity (high electrical conductivity), the higher the as-deposited reflectance, and the lower the reflectance losses. Thus, a simple four-point probe resistivity measurement enables an estimation of the expected Al reflectance and reflectance losses, even if these reflectance losses occur only after several years. This empirical relationship is more pronounced at shorter wavelengths and should not be confused with the Hagen-Rubens relation. In contrast, the Hagen-Rubens relation can be theoretically derived for good conductors in general and is valid for metals in the FIR spectral range [35]. Alternatively, we assume that the resistivity of these thin Al coatings is related to their grain boundaries.
An increased resistivity could indicate poorly formed Al grain boundaries within the Al layer and on its surface. Grain boundaries decrease the plasma resonance of electrons at the boundaries, causing a decreased reflectance [43]. Besides, the poorly formed grain boundaries might be more sensitive to oxidation and oxygen penetration, which further reduces the reflectance over time, especially at shorter wavelengths. Furthermore, the crystallographic Al(111) plane is more resistant to oxidation [44,45]. Whether the seeded Al coatings form Al2O3 top layers of different thicknesses is unfortunately unclear. By means of XRR, a determination of the thin Al2O3 layer thickness was unfeasible due to the low density contrast to aluminum. Stempfhuber et al. attributed the Al(111) diffraction peak intensity of their seeded Al coatings to different growth modes and, consequently, the reflectance in the FUV spectral range [26]. However, no significant relationship to the intensity of the Al(111) peak was observed by XRD in this study. Only the Al(111) peak of the Cu and Au seeded Al coatings were one magnitude larger, which could be related to the floe structure of their Al grains.
5. Conclusion
In conclusion, highly reflective and adhesive aluminum thin films with seed layers of Cu, CuOx, Cr, CrOx, Au, and Ag were prepared by DC magnetron sputtering on Si wafers. Aluminum coatings with CuOx, Cr, CrOx, and Au seed layers provide an enhanced reflectance in a broad spectral range with reduced long-term degradation compared to bare Al coatings without a seed layer. This simple approach is directly applicable to existing Al sputtering processes. Combining these seeded Al coatings with evaporation processes and suitable protective layers, such as SiOx, AlF3, or MgF2, could probably further increase their reflectance. In this context, the reflectance behavior in the FUV spectral range of such coatings is promising, as most recently demonstrated with Ti seeded Al/MgF2 mirrors by Larruquert et al. [28]. In contrast, seeded Al coatings with Cu and Ag show a reduced reflectance with elevated degradation. The surface structure of Al coatings with Cr and Cu seed layers is smoothed because the typical grains with dome-shaped capping are less pronounced. Particularly if smooth rather than highly reflective Al coatings are required, Cu seed layers could be applied. By using Au seed layers, the Al grains appear as flat floes with weakly pronounced grain boundaries, whereas some grains exhibit a dome-shaped capping. For the bare and seeded Al coatings, the reflectance behavior correlates with their specific electrical resistivity. With decreasing resistivity, the Al reflectance increases and, furthermore, the reflectance losses after about 4 years decrease. Both empirical relationships enable an estimation of the expected reflectance and long-term degradation at atmospheric conditions for highly reflective Al coatings.
Funding
Fraunhofer Society Attract Project (066-601020); Fraunhofer IOF Center of Excellence in Photonics; Thüringer Aufbaubank; Deutsche Forschungsgemeinschaft; Open Access Publication Fund (433052568); CRC/SFB 1375 “NOA - Nonlinear Optics down to Atomic scales” (398816777).
Acknowledgments
The authors thank Michael Scheler, Thomas Müller, and Steffen Schulze for technical support, Anika Bobzin and David Müller for the preparation of SEM images, Anna Franziska Gottwald and Alexander Bergner for AFM measurements, Michaela Mensing for support with the tape tests, Dr. Kristin Pfeiffer for helpful corrections, as well as Dr. Robert Müller, Dr. Philipp Naujok, and Dieter Gäbler for helpful and enlightening discussions. The authors especially thank Dr. Mark Schürmann and Prof. Dr. Frank Schmidl for their supervision and support of this work. We also thank the reviewers for their constructive comments to improve this paper. Paul Schmitt thanks the Thüringer Aufbaubank (TAB) for promoting his doctoral research studies. Finally, we acknowledge support by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), Open Access Publication Fund of the Thueringer Universitaets- und Landesbibliothek Jena, and the DFG Collaborative Research Center (CRC/SFB) 1375 “NOA - Nonlinear Optics down to Atomic scales”.
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. J. E. Curran, J. S. Page, and U. Pick, “The influence of some evaporation parameters on the structure and properties of thin aluminium films,” Thin Solid Films 97, 259–276 (1982). [CrossRef]
2. V. P. Jaecklin, C. Linder, J. Brugger, N. F. de Rooij, J.-M. Moret, and R. Vuilleumier, “Mechanical and optical properties of surface micromachined torsional mirrors in silicon, polysilicon and aluminum,” Sens. Actuators, A 43(1-3), 269–275 (1994). [CrossRef]
3. A. D. Rakić, A. B. Djurišić, J. M. Elazar, and M. L. Majewski, “Optical properties of metallic films for vertical-cavity optoelectronic devices,” Appl. Opt. 37(22), 5271–5283 (1998). [CrossRef]
4. C. W. Hollars and R. C. Dunn, “Evaluation of thermal evaporation conditions used in coating aluminum on near-field fiber-optic probes,” Rev. Sci. Instrum. 69(4), 1747–1752 (1998). [CrossRef]
5. Z. Ma and G. S. Was, “Aluminum metallization for flat-panel displays using ion-beam-assisted physical vapor deposition,” J. Mater. Res. 14(10), 4051–4061 (1999). [CrossRef]
6. S. van Gils, T. Dimogerontakis, G. Buytaert, E. Stijns, H. Terryn, P. Skeldon, G. E. Thompson, and M. R. Alexander, “Optical properties of magnetron-sputtered and rolled aluminum,” J. Appl. Phys. 98(8), 083505 (2005). [CrossRef]
7. J. Strong, “The Evaporation Process and its Application to the Aluminizing of Large Telescope Mirrors,” Astrophys. J. 83, 401–423 (1936). [CrossRef]
8. J. C. Burridge, H. Kuhn, and A. Pery, “Reflectivity of Thin Aluminium Films and their Use in Interferometry,” Proc. Phys. Soc., London, Sect. B 66(11), 963–968 (1953). [CrossRef]
9. G. Hass, “Reflectance and preparation of front-surface mirrors for use at various angles of incidence from the ultraviolet to the far infrared,” J. Opt. Soc. Am. 72(1), 27–39 (1982). [CrossRef]
10. S. Wilbrandt, O. Stenzel, H. Nakamura, D. Wulff-Molder, A. Duparré, and N. Kaiser, “Protected and enhanced aluminum mirrors for the VUV,” Appl. Opt. 53(4), A125–130 (2014). [CrossRef]
11. L. V. Rodríguez-de Marcos, D. R. Boris, E. Gray, J. G. del Hoyo, A. C. Kozen, J. G. Richardson, S. G. Rosenberg, S. G. Walton, V. Wheeler, E. J. Wollack, J. M. Woodward, and M. A. Quijada, “Room temperature plasma-etching and surface passivation of far-ultraviolet Al mirrors using electron beam generated plasmas,” Opt. Mater. Express 11(3), 740–756 (2021). [CrossRef]
12. LUVOIR Team, The Large UV/Optical/Infrared Surveyor (LUVOIR) Mission Concept Study Final Report. Report, 2019.
13. HabEx Team, The Habitable Exoplanet Observatory (HabEx) Mission Concept Study Final Report. Report, 2020.
14. K. M. McPeak, S. V. Jayanti, S. J. P. Kress, S. Meyer, S. Iotti, A. Rossinelli, and D. J. Norris, “Plasmonic Films Can Easily Be Better: Rules and Recipes,” ACS Photonics 2(3), 326–333 (2015). [CrossRef]
15. M. A. Quijada, S. Rice, and E. Mentzell, “Enhanced MgF2 and LiF over-coated Al mirrors for FUV space astronomy,” Proc. SPIE 8450, 843–852 (2012). [CrossRef]
16. L. V. Rodríguez-de Marcos, J. I. Larruquert, J. A. Méndez, N. Gutiérrez-Luna, L. Espinosa-Yáñez, C. Honrado-Benítez, J. Chavero-Royán, and B. Perea-Abarca, “Optimization of MgF2-deposition temperature for far UV Al mirrors,” Opt. Express 26(7), 9363–9372 (2018). [CrossRef]
17. G. Hass, W. R. Hunter, and R. Tousey, “Reflectance of Evaporated Aluminum in the Vacuum Ultraviolet,” J. Opt. Soc. Am. 46(12), 1009–1012 (1956). [CrossRef]
18. H. K. Pulker and E. Jung, “Correlation between film structure and sorption behaviour of vapour deposited ZnS, cryolite and MgF2 films,” Thin Solid Films 9(1), 57–66 (1972). [CrossRef]
19. O. R. Wood, H. G. Craighead, J. E. Sweeney, and P. J. Maloney, “Vacuum ultraviolet loss in magnesium fluoride films,” Appl. Opt. 23(20), 3644–3649 (1984). [CrossRef]
20. J. I. Larruquert and R. A. M. Keski-Kuha, “Far ultraviolet optical properties of MgF2 films deposited by ion-beam sputtering and their application as protective coatings for Al,” Opt. Commun. 215(1-3), 93–99 (2003). [CrossRef]
21. M. Kolar and P. Mach, “Relationships among properties of sputtered thin films and sputtering process parameters,” in 24th International Spring Seminar on Electronics Technology. Concurrent Engineering in Electronic Packaging (ISSE 2001 Conf. Proc., 2001), pp. 42–46.
22. D. L. Rode, V. R. Gaddam, and J. Haeng Yi, “Subnanometer surface roughness of dc magnetron sputtered Al films,” J. Appl. Phys. 102(2), 024303 (2007). [CrossRef]
23. T. Kiyota, S. Toyoda, K. T. K. Tamagawa, and H. Y. H. Yamakawa, “Aluminum Film Deposition Using an Ultrahigh-Vacuum Sputtering System,” Jpn. J. Appl. Phys. 32(Part 1, No. 2), 930–934 (1993). [CrossRef]
24. D. V. Sidelev, A. V. Yuryeva, V. P. Krivobokov, A. S. Shabunin, M. S. Syrtanov, and Z. Koishybayeva, “Aluminum films deposition by magnetron sputtering systems: Influence of target state and pulsing unit,” J. Phys.: Conf. Ser. 741, 012193 (2016). [CrossRef]
25. M. Fernández-Perea, J. I. Larruquert, J. A. Aznárez, A. Pons, and J. A. Méndez, “Vacuum ultraviolet coatings of Al protected with MgF2 prepared both by ion-beam sputtering and by evaporation,” Appl. Opt. 46(22), 4871–4878 (2007). [CrossRef]
26. S. Stempfhuber, N. Felde, S. Schwinde, M. Trost, P. Schenk, S. Schröder, and A. Tünnermann, “Influence of seed layers on optical properties of aluminum in the UV range,” Opt. Express 28(14), 20324–20333 (2020). [CrossRef]
27. S. Stempfhuber, D. Gäbler, P. Schenk, P. Munzert, and S. Schwinde, Method of Manufacturing an Aluminum Layer and Optical Element, EP3779526A1 (22 July 2020).
28. J. I. Larruquert, C. Honrado-Benítez, N. Gutiérrez-Luna, Á. Ríos-Fernández, and P. López-Reyes, “Far UV-enhanced Al mirrors with a Ti seed film,” Opt. Express 29(5), 7706–7712 (2021). [CrossRef]
29. P. Scherrer, “Bestimmung der Größe und der inneren Struktur von Kolloidteilchen mittels Röntgenstrahlen,” Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-Physikalische Klasse 1918, 98–100 (1918).
30. L. J. Swartzendruber, “Correction factor tables for four-point probe resistivity measurements on thin, circular semiconductor samples,” Technical Note, 1964.
31. DIN Deutsches Institut für Normung e. V., “Optik und Photonik - Optische Schichten - Teil 4: Spezifische Prüfmethoden (ISO 9211-4:2012),” (Beuth Verlag, 2012).
32. R. K. Hart and U. R. Evans, “The oxidation of aluminium in dry and humid oxygen atmospheres,” Proc. R. Soc. A 236, 68–88 (1956). [CrossRef]
33. J. Bartl and M. Baranek, “Emissivity of aluminium and its importance for radiometric measurement,” Meas. Sci. Rev. 4, 31–36 (2004).
34. F. Kohlrausch, “Praktische Physik. Zum Gebrauch für Unterricht, Forschung und Technik,” 24th ed. (B.G. Teubner, 1996).
35. E. Hagen and H. Rubens, “Über Beziehungen des Reflexions- und Emissionsvermögens der Metalle zu ihrem elektrischen Leitvermögen,” Ann. Phys. 316(8), 873–901 (1903). [CrossRef]
36. H. E. Bennett and J. O. Porteus, “Relation Between Surface Roughness and Specular Reflectance at Normal Incidence,” J. Opt. Soc. Am. 51(2), 123–129 (1961). [CrossRef]
37. L. Vitos, A. V. Ruban, H. L. Skriver, and J. Kollár, “The surface energy of metals,” Surf. Sci. 411(1-2), 186–202 (1998). [CrossRef]
38. P. Schmitt, V. Beladiya, N. Felde, P. Paul, F. Otto, T. Fritz, A. Tünnermann, and A. V. Szeghalmi, “Influence of Substrate Materials on Nucleation and Properties of Iridium Thin Films Grown by ALD,” Coatings 11(2), 173 (2021). [CrossRef]
39. J. N. Israelachvili, Intermolecular and Surface Forces, 3rd ed. (Academic Press, 2011).
40. Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies (CRC Press, 2007).
41. R. A. Maniyara, D. Rodrigo, R. Yu, J. Canet-Ferrer, D. S. Ghosh, R. Yongsunthon, D. E. Baker, A. Rezikyan, García de Abajo, F. Javier, and V. Pruneri, “Tunable plasmons in ultrathin metal films,” Nat. Photonics 13(5), 328–333 (2019). [CrossRef]
42. G. Antczak and G. Ehrlich, “Surface Diffusion: Metals, Metal Atoms, and Clusters,” (Cambridge University, 2010).
43. O. Hunderi, “Influence of Grain Boundaries and Lattice Defects on the Optical Properties of Some Metals,” Phys. Rev. B 7(8), 3419–3429 (1973). [CrossRef]
44. P. O. Gartland, “Adsorption of oxygen on clean single crystal faces of aluminium,” Surf. Sci. 62(1), 183–196 (1977). [CrossRef]
45. P. Hofmann, W. Wyrobisch, and A. M. Bradshaw, “The interaction of oxygen with aluminium single crystal surfaces: Mainly ΔΦ aspects,” Surf. Sci. 80, 344–351 (1979). [CrossRef]
