Dichroic laser mirrors with mixture layers and sandwich-like-structure interfaces

1. INTRODUCTION

Dichroic laser mirrors are usually used as harmonic separators [1,2], beam combiners [3], or beam splitters [4] and play an important role in many laser applications, including inertial confinement fusion (ICF) lasers [5], petawatt femtosecond lasers [6], high power fiber lasers [3,7], compact $Q$-switched or mode-locked lasers [8–10], and other emerging lasers [11]. The ideal dichroic laser mirror for high-power lasers requires a significantly different reflection or transmission property and a high laser-induced damage threshold (LIDT) at two different wavelengths of interest. Unfortunately, a traditional dichroic laser mirror (TDLM) composed of alternating high- and low-refractive-index ($n$) pure materials often has difficulty achieving excellent spectral performance and high LIDTs at two wavelengths simultaneously [4]. Generally, TDLM for UV-NIR laser applications is achieved by alternately e-beam deposited ${\mathrm{HfO}}_2$ layers and ${\mathrm{SiO}}_2$ layers [12]. Sometimes, ${\mathrm{Al}}_2{\mathrm{O}}_3$ is chosen instead of ${\mathrm{HfO}}_2$ as the high-$n$ material, which shows improved LIDT but requires a relatively large total number of coating layers [13,14]. There is a trade-off between the required optical performance and LIDT because suitable candidate coating materials are limited. In recent years, the library of available coating materials is expanded by co-evaporated or co-sputtered oxide mixtures [15–17]. The mixture materials provide us with adjustable optical gap values and optical constants, show superior properties over pure materials [16,18], and are attractive for many applications [19–21]. In addition to the coating material itself, it is also necessary to consider the interface-related issues of the traditional discrete interface, which is one of the key factors affecting LIDT. The co-evaporated interface with a graded-refractive index shows a significant increase in LIDT compared with the traditional discrete interface [22,23]. Therefore, by appropriately designing mixture materials and optimizing the interface, it is expected to realize an ideal dichroic laser mirror suitable for high-power lasers.

Here, we propose to use mixture materials combined with a novel sandwich-like interface to meet the challenging requirements of dichroic laser mirrors. First, the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture coating is compared with pure ${\mathrm{HfO}}_2$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ monolayer coatings as well as ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ nanolaminate coatings in terms of microstructure and optical and mechanical properties. Then, a mixture-based dichroic laser mirror (MDLM) is designed and prepared, which uses ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture material as a high-$n$ layer with adjustable $n$ and optical bandgap and pure ${\mathrm{SiO}}_2$ as a low-$n$ material. The schematic diagram of the MDLM design is shown in Fig. 1. The interface between the ${\mathrm{SiO}}_2$ layer and the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture layer is a sandwich-like-structure interface, which can be represented by “${\mathrm{SiO}}_2-{\mathrm{HfO}}_2$ gradient material$|{\mathrm{HfO}}_2|{\mathrm{HfO}}_2\text{−}{\mathrm{Al}}_2{\mathrm{O}}_3$ gradient material.” As we will show, this arrangement allows the MDLM design to have excellent spectral performance and high LIDTs at two wavelengths. Our MDLM design strategy, prepared using e-beam evaporation [18], shows improved performance over TDLM and opens new avenues for a variety of laser coatings.

 

Fig. 1. Schematic diagram of the proposed MDLM design.

Download Full Size | PPT Slide | PDF

2. MATERIALS AND METHODS

A. Preparation of the Coatings

All coatings are deposited on fused silica and BK7 substrates using e-beam evaporation, in which the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture layer and the sandwich-like-structure interface are obtained by dual e-beam co-evaporation [18]. BK7 and fused silica substrates are used for stress measurement and other measurements, respectively. Before deposition, the coating chamber is heated to 473 K and evacuated to a base pressure of $9\times {10}^{-4}\mathrm{Pa}$; then, the substrate is cleaned with plasma ions at a bias voltage of 80 V for 120 s. The deposition rates for ${\mathrm{HfO}}_2$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ in monolayer and nanolaminate coatings and the deposition rates for ${\mathrm{SiO}}_2$ in multilayer coatings are 0.1, 0.1, and 0.2 nm/s, respectively. The deposition rates for ${\mathrm{HfO}}_2$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ in the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture coating are 0.028 and 0.072 nm/s, respectively. The deposition rates for ${\mathrm{HfO}}_2$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ in the mixture layer of the MDLM coating are 0.05 and 0.05 nm/s, respectively. The oxygen pressures of the ${\mathrm{SiO}}_2$ layer and other layers are $5.0\times {10}^{-3}$ and $1.3\times {10}^{-2}\mathrm{Pa}$, respectively.

The schematic diagram of the deposition process of the MDLM coating is shown in Fig. 2. The deposition rate is monitored by a quartz crystal monitor installed at the corresponding side of the coating material. To illustrate the deposition process of the sandwich-like-structure interface, the interface of the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture layer on the ${\mathrm{SiO}}_2$ layer is taken as an example. First, the deposition rate of the ${\mathrm{SiO}}_2$ layer is reduced from the set value to 0, and the deposition rate of the ${\mathrm{HfO}}_2$ layer is increased from 0 to the set value at the same time, thereby forming the ${\mathrm{SiO}}_2-{\mathrm{HfO}}_2$ gradient material. Then, the deposition rate of the ${\mathrm{HfO}}_2$ layer is kept constant within the set time. At last, the deposition rate of the ${\mathrm{Al}}_2{\mathrm{O}}_3$ layer is increased from 0 to the set value, while the deposition rate of the ${\mathrm{HfO}}_2$ layer remains constant, thereby forming the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ gradient material.

 

Fig. 2. Schematic diagram of the deposition process of MDLM coating.

Download Full Size | PPT Slide | PDF

B. Characterization of the Coatings

X-ray diffraction (XRD) (PANalytical Empyrean) is used to characterize the structure information of the coating. A VUV spectrometer (LZH ML 6500) and a UV-VIS-NIR spectrometer (Perkin Elmer Lambda 1050) are employed to measure the transmittance spectra in the ranges of 150–200 nm and 200–1200 nm, respectively. The reflectance spectra in the VIS region are calculated from the transmission data neglecting absorption. The refractive indices and optical bandgaps are estimated using the commercial thin film software (Essential Macleod) and the Tauc equation [24], respectively. The laser-induced temperature rise in the multilayer coating is obtained from the finite-element method (FEM) simulation.

The elemental composition profiles are determined by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha) using a monochromatic Al $\mathrm{K}\mathrm{\alpha }$ (1486.6 eV) X-ray source. The spectra are recorded after every 20 s of etching with 1 keV Ar⁺ ions.

A 632.8 nm wavelength interferometer (ZYGO Mark III-GPI) is employed to inspect the sample surfaces before (substrates) and 60 days after the deposition (coatings) in a controlled environment with a temperature of $23\pm 1.5°\mathrm{C}$ and relative humidity of $45\%\pm 5\%$. The coating stress is obtained from Stoney’s formula.

The absorption of the coating at 1064 nm is measured by a home-made system based on the surface thermal lensing technique. The interfacial adhesion is characterized by a scratch test using a nano indenter (KLA Tencor). A load is gradually increased from 20 μN to 50 mN for the scratch test.

The 1-on-1 LIDT is tested according to ISO 21254. An $s$-polarized $2\omega$ Nd:YAG laser with a 7.7 ns pulse width and a $p$-polarized $1\omega$ Nd:YAG laser with a 12 ns pulse width are used for 532 and 1064 nm LIDT measurements, respectively. The test is performed at an angle of incidence of 45°. The effective beam sizes on the sample surface for 532 and 1064 nm LIDT measurements are approximately $0.32{\mathrm{mm}}^2$ and $0.072{\mathrm{mm}}^2$, respectively. Fifteen sites are tested for each energy fluence. The damage morphology is characterized by a focused ion beam scanning electron microscope (FIB-SEM, Carl Zeiss AURIGA CrossBeam). The chemical composition of the damaged area is analyzed by energy dispersive spectroscopy (EDS, Oxford X-Max, $50{\mathrm{mm}}^2$).

3. EXPERIMENTAL RESULTS AND DISCUSSION

A. Properties of the Pure Monolayer, Nanolaminate, and Mixture Coatings

A ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture coating has been prepared and compared with ${\mathrm{HfO}}_2$ and ${\mathrm{Al}}_2{\mathrm{O}}_3$ monolayer coatings and ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ nanolaminate coatings. The design structures of the nanolaminate coatings are listed in Table 1. The microstructure and optical properties of the coatings have been studied. Multiple sharp diffraction peaks indicating crystallinity are obtained in the XRD pattern of the ${\mathrm{HfO}}_2$ monolayer coating [Fig. 3(a)]. Two broad peaks at 21.8° and 31.4°, attributed to the amorphous fused silica substrate and ${\mathrm{HfO}}_2$ (111), are observed in the nanolaminate and mixture coatings. The thinner the ${\mathrm{HfO}}_2$ sublayer, the less obvious the peak at $2\theta =31.4°$.

Table 1. Design Information and Extracted Thickness of the Nanolaminate Coatings^a

View Table | View all tables in this article

 

Fig. 3. Microstructure and optical property of the pure monolayer, nanolaminate, and mixture coatings. (a) XRD spectra, (b) transmittance, and (c) optical bandgap versus $n$ of ${\mathrm{HfO}}_2$ monolayer, ${\mathrm{Al}}_2{\mathrm{O}}_3$ monolayer, ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ nanolaminate, and mixture coatings. (d) Surface figure change ($\mathrm{\Delta }\text{Power}$) caused by the ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ nanolaminate and mixture coatings.

Download Full Size | PPT Slide | PDF

To determine $n$ and the optical bandgap, the transmittance spectra of the substrate and coating are measured [Fig. 3(b)]. The determined $n$ and surface figure change of the mixture and nanolaminate coatings are close [Figs. 3(c) and 3(d)]. The optical bandgap of the nanolaminate coating increases as the thickness of the sublayer decreases, and the mixture layer shows the largest optical bandgap. The dispersion curves of the refractive indices of ${\mathrm{HfO}}_2$ monolayer, ${\mathrm{Al}}_2{\mathrm{O}}_3$ monolayer, and ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture coatings are derived using the envelope method (Fig. 4). The thicknesses of the ${\mathrm{HfO}}_2$ monolayer, ${\mathrm{Al}}_2{\mathrm{O}}_3$ monolayer, and ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture coatings are fitted to be 393.3, 513.8, and 433.5 nm, respectively.

 

Fig. 4. Dispersion curves of refractive indices of ${\mathrm{HfO}}_2$ monolayer, ${\mathrm{Al}}_2{\mathrm{O}}_3$ monolayer, and ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture coatings.

Download Full Size | PPT Slide | PDF

The mixture layer has two advantages over the nanolaminate layers: the deposition process is simpler, and the bandgap is larger when $n$ is close. The mixture is therefore chosen as the high-$n$ material for MDLM coating in this work. This allows one to develop MDLM coatings with excellent optical and LIDT properties.

B. Mixture-Based Dichroic Laser Mirrors with Sandwich-like-Structure Interfaces

An MDLM coating is designed to achieve an $s$-polarized reflectivity (Rs) higher than 99.5% at $527\pm 15\mathrm{nm}$ and a $p$-polarized transmittance (Tp) higher than 98% in the range of 720 to 1064 nm for an angle of incidence of 45°. This MDLM coating can act as a $1\omega$ and $2\omega$ harmonic separator in fusion-class lasers or a pump/signal beam separator in petawatt-class Ti-sapphire laser systems. The coating structure is as follows: $\text{substrate}|2\mathrm{L}{(0.5\mathrm{M}\mathrm{L}0.5\mathrm{M})}^{14}4.45\mathrm{L}\text{|air}$. Here, $\mathrm{M}$ and $\mathrm{L}$ represent ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture and ${\mathrm{SiO}}_2$ layers with a quarter-wavelength optical thickness (QWOT) at 595 nm (${\lambda }_0$), respectively. The numbers before $\mathrm{M}$ and $\mathrm{L}$ represent the optical thickness in units of the QWOT of the corresponding material. The design thickness of the sandwich-like-structure interface is about 12 nm, including ${\mathrm{SiO}}_2-{\mathrm{HfO}}_2$ gradient material ($\sim 4\mathrm{nm}$), ${\mathrm{HfO}}_2$ material ($\sim 4\mathrm{nm}$), and ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ gradient material ($\sim 4\mathrm{nm}$). The $n$ of $\mathrm{L}$ and $\mathrm{M}$ at 595 nm are 1.446 and 1.80, respectively. According to $n_M{d}_M=n_L{d}_L={\lambda }_0/4$, the physical thicknesses are $d_{2L}=205.74\mathrm{nm}$, $d_{0.5M}=41.32\mathrm{nm}$, $d_L=102.87\mathrm{nm}$, and $d_{4.45L}=457.77\mathrm{nm}$. A thick ${\mathrm{SiO}}_2$ overcoat layer ($4.45\mathrm{L}$) is used as a protective layer, and the resulting electric (E)-field intensity (incident angle at 45°, $s$-polarized light) of the coating–air interface is close to zero at 532 nm. For comparison, the MDLM coating and a TDLM coating with a $\text{substrate}|2\mathrm{L}{(0.5\mathrm{H}\mathrm{L}0.5\mathrm{H})}^{10}4.45\mathrm{L}\text{|air}$ structure are prepared. $\mathrm{H}$ represents ${\mathrm{HfO}}_2$ layers with a QWOT at 595 nm, and the physical thickness $d_{0.5H}=38.66\mathrm{nm}$. The elemental percentage profiles from the high-$n$ layer to the low-$n$ layer in the TDLM and MDLM coatings are compared in Fig. 5.

 

Fig. 5. Elemental percentage profiles from the high-$n$ layer to the low-$n$ layer.

Download Full Size | PPT Slide | PDF

Obvious sharp diffraction peaks are not observed in the XRD pattern of MDLM but in that of the TDLM coating [Fig. 6(a)]. The MDLM coating shows a lower optical absorption in the short wavelength region. Both coatings show Rs higher than 99.5% at $532\pm 15\mathrm{nm}$ and Tp higher than 98% in the range of 720 to 1064 nm as designed [Fig. 6(b)]. The surface profiles of the samples before and after coating are compared in Fig. 6(c); the stresses of the TDLM and MDLM coatings are calculated to be $-20.1\mathrm{MPa}$ and $-81.2\mathrm{MPa}$, respectively. The larger compressive stress of MDLM coating indicates that it can withstand a lower humidity environment than the TDLM coating. The surface and cross-section morphologies after the scratch test [Fig. 6(d)] show that the delamination damage of the MDLM coating is more moderate than that of the TDLM coating, which indicates that the MDLM coating has better mechanical properties. The measured absorption (at 1064 nm) of the MDLM coating is 6 ppm (parts per million), which is half of 12 ppm of the TDLM coating. The laser damage probability as a function of fluence for the two coatings is shown in Fig. 6(e). The damage probability curve under 532 nm laser irradiation is used to extract the defect parameters by using the model developed by Krol et al. [25], assuming the existence of two classes of defects with different values of the LIDT $T_i$ and area density $D_i$ (integrated over the thickness) for both coatings. The obtained defect parameters are listed in Table 2, in which the $\mathrm{\Delta }{T}_i$ is the threshold standard deviation. More than two typical damage morphologies in one damage site are observed after $p$-polarized 1064 nm laser irradiation, indicating a more complex damage mechanism, which makes the defect-parameter-extract model difficult to apply. Overall, the LIDT (the maximum laser fluence corresponding to the damage probability of 0%) of the MDLM coating is almost twice that of TDLM coating at both wavelengths.

 

Fig. 6. Microstructure and optical property of the TDLM and MDLM coatings. (a) XRD spectra. (b) Transmittance (left: incident angle at 0°; middle: incident angle at 45°, $p$-polarized light) and reflectance spectra (right: incident angle at 45°, $s$-polarized light). (c) Surface figures of the samples before and after coating. (d) Surface and cross-section morphologies after the scratch test. (e) Single-pulse damage probability as a function of the input fluence.

Download Full Size | PPT Slide | PDF

Table 2. Extracted Defect Parameters

View Table | View all tables in this article

Typical damage morphologies of the TDLM and MDLM coatings irradiated by an $s$-polarized 532 nm laser [Figs. 7(a)–7(d)] suggest that both coatings have two different morphological features. The two morphological features in the TDLM coating are shell-type pits [Fig. 7(e)] and flat bottom pits [Fig. 7(f)] observed at laser fluence just above the LIDT and the higher laser fluences, respectively. The two morphological features in the MDLM coating are flat bottom pits [Fig. 7(g)] and nodular-ejected pits [Fig. 7(h)] observed at laser fluence just above the LIDT and the higher laser fluences, respectively. Plasma scalds surrounding the flat bottom pits and nodular-ejected pits are observed at a relatively high laser fluence [Figs. 7(i) and 7(j)]. In general, the shell-type pits are smaller and shallower than the flat bottom pits, and the size of the shell-type and flat bottom pits does not strongly depend on the laser fluence. A nanoscale pinpoint is observed inside the most damaged sites, which indicates that the damage morphology is related to nanoscale defects.

 

Fig. 7. Damage morphology imaged by SEM and the depth profile of the marked area measured by FIB. (a)–(d) Damaged sites and (e)–(h) schematic diagram of the damage morphologies after irradiation of an $s$-polarized 532 nm laser. (i) and (j) Plasma scald induced by $s$-polarized 532 nm laser in MDLM coating.

Download Full Size | PPT Slide | PDF

More morphological features are observed after $p$-polarized 1064 nm laser irradiation [Figs. 8(a)–8(i)] because the defect distribution depends on the polarization and wavelength of the study [26]. As shown in Figs. 8(a) and 8(b), multiple pits are observed in the damaged area induced by a single laser shot. At higher laser fluence, plasma scalds are also observed. The typical morphologies are further classified according to their features. In addition to flat bottom pits [Fig. 8(j)] and nodular-ejected pits [Fig. 8(k)] similar to those observed after $s$-polarized 532 nm laser irradiation, two other morphological features observed may be related to substrate defects [27–29], namely, craters [Fig. 8(l)] and pits with microcracks [Fig. 8(m)]. The craters are similar to those observed in fused silica [30]. The microcrack damage may be related to the extensive fractures in the subsurface layer of fused silica. Fracture-induced absorption [27] may cause local temperature rise under laser irradiation and induce damage. The energy dispersive spectroscopy characterization confirms the presence of coating material in the microcracks (Fig. 9), which indicates that local high temperatures are generated under laser irradiation, causing the coating material to melt and enter the cracks. Based on the observed typical damage morphologies, damage may be initiated at three locations: the seeds of the nodules [Figs. 7(d), 8(d), and 8(h)], nano-precursors at the interface of the ${\mathrm{SiO}}_2$ layer on the high-$n$ layer [Figs. 7(a)–7(c), 8(c), and 8(g)], and substrate defects especially for the $p$-polarized 1064 nm laser [Figs. 8(e), 8(f), and 8(i)].

 

Fig. 8. Damage morphology imaged by SEM and the depth profile of the marked area measured by FIB. (a) and (b) Full field-of-view of the damaged area. (c)–(i) Typical damaged sites marked with asterisks. (j)–(m) Schematic diagram of the damage morphologies after irradiation of a $p$-polarized 1064 nm laser.

Download Full Size | PPT Slide | PDF

 

Fig. 9. EDS characterized chemical composition of the damaged site of the TDLM coating and MDLM coating induced by 1064 nm $p$-polarized laser.

Download Full Size | PPT Slide | PDF

An FEM simulation is used to investigate the 1064 nm $p$-polarized laser-induced temperature rise in the two coatings. The extinction coefficients ($k$) of the ${\mathrm{HfO}}_2$ layer ($1.36\times {10}^{-6}$) and ${\mathrm{HfO}}_2-{\mathrm{Al}}_2{\mathrm{O}}_3$ mixture layer ($4.52\times {10}^{-7}$) are calculated by using Eq. (1) [31], based on the measured absorption, neglecting the absorption of the ${\mathrm{SiO}}_2$ layers:

For comparison, the input laser parameters for the TDLM coating and MDLM coating are consistent. Figure 10 shows the simulated temperature distributions in the event that a maximum temperature of 1600°C is generated in the TDLM coating, and the simulated temperature rise distributions in the MDLM coating under the same laser input parameters. The maximum temperature in the TDLM coating is almost twice of that of the MDLM coating. Compared with the TDLM coating, the MDLM coating shows a lower laser-induced temperature rise due to its smaller absorption, which is consistent with the higher LIDT observed.

 

Fig. 10. Simulated laser-induced temperature rises in TDLM and MDLM coatings.

Download Full Size | PPT Slide | PDF

4. SUMMARY

In summary, we have proposed and experimentally demonstrated a new MDLM coating with mixture layers and sandwich-like-structure interfaces. The proposed MDLM coating shows excellent spectral performance, and an LIDT that is almost twice of that of the TDLM coating at the two wavelengths of interest for the following reasons. First, the mixture layer in the MDLM coating has a larger optical bandgap and a lower absorption, resulting in a smaller temperature rise under the same fluence laser irradiation; second, the sandwich-like-structure interface allows the MDLM coating to exhibit enhanced mechanical properties. We believe that the described concept opens new avenues for improved dichroic mirror coatings and other laser coatings and can benefit many areas of laser technology that rely on high-quality laser coatings.