Polarization control in modern laser systems is essential in several current applications: interference lithography [1], refractive index modulations in transparent materials [2], absorption enhancement in laser–matter interactions [3], etc. For solid state lasers, birefringent crystals (such as ${{\rm YVO}_4}$) can be used as gain media and generate linear polarization [4], but, for microchip lasers, isotropic materials are more preferable due to their exclusively better properties [5]. Additional consideration must also be included when higher averaged powers and ultrashort pulses are used, or compact solutions with shorter resonators are needed. Optical components and especially coatings have to withstand extreme intensities without losing their optical performance. Therefore, compact, preferably coating-based solutions for high power polarized laser emission are highly desirable.

Several different methods, mostly implemented using glancing angle deposition (GLAD) technology, have been realized for low band-gap thin-film-based polarizing components. For circular polarization selectivity, chiral nanostructures can be formed with helical non-homogeneity along a fixed axis [6]. For transformation to linear polarization, nematic structures of sculptured thin films can be employed, namely, zig-zag nanostructures, which are similar to biaxial crystals and have three characteristic refractive indices [7]. When such a film is irradiated at a zero angle of incidence (AOI), two principal axes are present, the fast and slow axes, which are parallel and perpendicular to the shadowing plane, respectively [8]. The shadowing plane is the plane parallel to the $\alpha$ rotation axis and vapor flux direction [see Fig. 1(a)]. Birefringence of anisotropic thin film depends on deposition angle $\chi$. Specifically selected combinations of anisotropic thin films in a multilayer structure can lead to novel components. Changing the orientation of an individual film’s optical axis results in a polarizer with highly reflected (R) linear polarization and highly transmitted (T) perpendicular polarization. A few such designs have been realized using ${{\rm TiO}_2}$, ${{\rm Ta}_2}{{\rm O}_5}$, and ${{\rm WO}_3}$ materials [9,10]. The combination of both nanostructures also was implemented for more complex optical coating, capable of inverting polarization in the transmission mode [11]. All of aforementioned investigations describe polarizing coatings, but all of them used materials with comparatively low optical resistivity, limiting their applications in high power

 

Fig. 1. (a) Principle scheme of the GLAD system, (b) effective refractive index dispersions of silica films deposited by the GLAD system at 0°, 66°, and 70° angles, and (c) principal scheme of the cross section of the high-contrast polarizer multilayer coating consisting of anisotropic layers with perpendicularly aligned optical axes.

Download Full Size | PDF

During the last 5 years, a new approach has been presented for coatings with extreme resistivity to laser radiation. All-silica-based high reflectivity mirrors showed the potential to reach more than ${60}\;{{\rm J/cm}^2}$ LIDT at a 355 nm wavelength in the pulsed nanosecond regime [13]. Afterwards, anisotropic all-silica multilayer coatings have been implemented for waveplate formation, which showed impressive performance under high laser power density (laser induced damage threshold, ${\rm LIDT} = {24.4}\;{{\rm J/cm}^2}$) [8].

In this Letter, we present a new all-silica coating polarizer, which is also capable of withstanding high density radiation. In order to demonstrate the versatility of the presented approach, several coating designs have been modeled, and two of them are fabricated together with the full-scale measurements and analysis necessary for the implementation of polarizres into high power microlaser systems. Two polarizing coatings at the wavelength of 355 nm have been formed using two stepper motors based on the GLAD system [Fig. 1(a)]. Afterwards, optical and structural analyses have been performed including spectrophotometric, atomic-force microscopy (AFM), scanning electron microscopy (SEM), and optical resistivity measurements.

Thin films, formed by the GLAD method, can virtually exhibit any refractive index ranging from bulk material to air [14]. Additionally, birefringence of the layers can also be tuned by changing the deposition angle and the thickness of zig-zag subdeposits [15]. In present research, for the formation of anisotropic thin films, depositions at glancing angles (in this case 66° and 70°) were performed using the so-called serial bideposition technique [15]. Substrates were rotated every 6 s in half-turns around its axis for the formation of columnar thin films with orientation perpendicular to the substrate plane with elliptical shape cross sections. The effective refractive indices of silica thin films, deposited at angles of 0°, 66°, and 70°, are presented in Fig. 1(b). A relatively dense silica layer, deposited at a 0° angle, exhibits the isotropic refractive index of 1.43. Effective refractive indices of anisotropic layers for the plane between vapor flux and substrate normal have lower values compared to the perpendicular plane due to the self-shadowing effect during thin-film growth. Therefore, the aforementioned plane performs a function of the “fast axis” and the plane in the perpendicular direction is the “slow axis”. As shown in Fig. 1(b), refractive indices match in the fast and slow directions for thin films deposited at 66° ($n_{\text{fast}}^{66} = 1.27$) and 70° ($n_{\rm{slow}}^{70} = 1.27$) angles, respectively. For opposite directions of the films, the same layers have the difference in refractive indices of 0.054. Therefore, deposition of both single layers sequentially and misaligning their optical axis by 90° with respect to each other would result in a difference of spectral performance. The principal scheme of sequentially misaligned layers with elliptically shaped cross sections is presented in Fig. 1(c). Fixing the optical thicknesses of all layers in the presented multilayer coating will result in two distinct spectral performances for two perpendicular linear polarizations. For T polarization, the refractive indices match, and minimal Fresnel reflections occur between the interfaces of the coating. For perpendicular polarization, namely, R polarization, indices differ, and spectral performance depends on the thicknesses of individual layers.

 

Fig. 2. (a) Modeled transmittances for R and T polarizations of the high-contrast design depending on the number of the stacks (including the back side of the component), (b) modeled transmittances of the middle-contrast design depending of the number of modifications, and (c) measured transmittance for R and T polarizations at zero AOI of fabricated polarizers.

Download Full Size | PDF

The obtained dispersions of effective refractive indices [in Fig. 1(b)] of silica thin films were used in modeling multilayer coatings. The multilayer coating consists of layers sequentially depositing at angles of 66° and 70° and changing the orientation of the films’ optical axis by 90°. Such a sequence results in matched indices for T polarization. Therefore, optical thicknesses of individual layers have low impact on transmittance/reflectance for this linear polarization. For R polarization, indices have the largest difference, and, therefore, optical thicknesses have a high impact on optical performance. In order to gain the maximal reflectance in R polarization at the wavelength of 355 nm, each layer was set at one quarter wavelength optical thickness (QWOT). Therefore, the physical thickness of films, deposited at angles of 66° and 70°, were 68.1 nm and 70.7 nm, respectively. Choosing 12 layers (six pairs of films, deposited at angles of 66° and 70°) as one stack, several designs were modeled for demonstration of the high-contrast polarizer’s principal idea [see Fig. 2(a)]. Modeled transmission spectra indicate high transmission for T polarization, which is independent of the number of stacks in the coating. As for R polarization, the reflectivity increases by increasing the number of stacks within the coating to virtually any value. It is worth mentioning that dense layers (deposited at a 0° angle) were inserted between the stacks in order to strengthen the coating structure and increase the mechanical resistivity. Since the optical thickness of each dense thin film was two QWOTs, it had minimal influence on spectral performance. As can be seen from the figure, the presented approach allows us to gradually control the transmittance of R polarization and stabilize it for T polarization at the same time. In order to gradually vary the transmittance of T polarization, additional modifications were added to the design. The modification means that for the multilayer design three pairs of anisotropic layers are removed, and one high reflective pair (angle of 70° and dense silica layer of QWOT) is added. Figure 2(b) indicates how transmittance of the R polarization remains stable, and it can gradually be changed for T polarization. The presented approach indicates how transmittance for each polarization can be tuned independently to perpendicular polarization by choosing one of the two paths. Two modeled designs of high- and middle- contrast polarizers were experimentally verified and characterized in great detail.

Both designs were deposited using a SIDRABE electron-beam evaporation plant and measured with spectrophotometer PhotonRT with polarized light at 0° (AOI). Measured transmittance graphs are presented in Fig. 2(c). The high-contrast sample consisted of four stacks and exhibited the transmission of 40.4% and 85.3% for R and T polarizations, respectively. The obtained difference in transmission values differs by almost 10% from the theoretical due to experimental errors in the physical thicknesses and refractive indices of the coatings. Also, nano-sculptured layers can exhibit inhomogeneities in structure and optical properties as a consequence [8]. The middle-contrast experimental sample of three stacks and three modifications exhibited the transmission of 35.1% and 62.5% for R ant T polarizations, respectively. The obtained values again differ by almost 10% from the theoretical due to experimental errors in the coatings’ physical thicknesses and refractive indices. However, such discrepancies between the modeled and experiment results do not preclude the demonstration of the concept.

Structural analysis of the experimental samples has been performed in order to investigate the possible causes of optical losses and structural inhomogeneity. The surface morphology and inner structure of the high-contrast polarizer is shown in Fig. 3. Two SEM images are combined in Fig. 3(a): on the left side of the multicolor bar, the cross section is parallel to the R polarization direction, and, on the right side, the cross section is perpendicular to the R polarization direction (parallel to T polarization).

 

Fig. 3. Structural investigation: (a) SEM images of high-contrast polarizer (the colors of the marked central bar indicate substrate (black); layer deposited at 70° (red); layer deposited at 66° (blue); dense layer (gray)) and AFM images of (b) high-contrast and (c) middle-contrast polarizers morphologies.

Download Full Size | PDF

The multilayer structure consists of nanostructured thin films, deposited at angles of 66° and 70° by the serial bideposition method and dense layers inserts. In the beginning of the film growth, the average width of the individual columns is around 25 nm. The tangible shape of the columns cannot be distinguished due to the spread and the coalescence of the columns at interfaces. Previous research indicates that cross sections of the nano-columns, formed by deposition at either 66° or 70° angles, have elliptical cross sections [16]. The cross section of the multilayer parallel to R polarization consists of relatively densely packed columns in thin films, coated at a 66° angle, and more porous layers, coated at a 70° angle. Alternatively, anisotropic single layers, deposited at a 70° angle, with the thickness of more than 400 nm, have thicker columns, which may cause increased optical scattering [8]. Such a phenomenon was not observed when individual anisotropic films were deposited in an orthogonal orientation. Scattering from the relatively large features (25 nm and more) can still cause some optical losses in the UV spectral range. Additionally, the dense interlayers were embedded after each stack for better mechanical robustness.

The surface roughness of the fused silica (FS) substrates used in experiments did not exceed 0.5 nm. GLAD is known to increase the surface roughness of the optical coatings [8,14]. Analysis of AFM images indicates that the surface roughness of the high-contrast and middle-contrast polarizers is 2.9 nm and 2.7 nm, respectively. Comparatively large surface roughness can also be considered as the essential reason for optical losses.

The high-contrast polarizer was tested by high power laser radiation in order to determine the LIDT value. An EKSPLA NL220 pulsed nanosecond Nd:YAG laser was used for the evaluation of damage probability at the wavelength of 355 nm using the one-on-one test method [17]. Pulse duration of ${\sim}{3}\;{\rm ns}$ and microfocus approach with focused beam diameter of 60 µm were chosen in order to separate damages on local defects from intrinsic resistance of investigated surfaces. Damage probability versus radiation fluence for the R and T polarizations is presented in Fig. 4(a). The errors of LIDT measurements did not exceed $\pm {5}\%$ of the measured values. The lowest fluence level, at which damage initiation occurred, was ${29.5}\;{{\rm J/cm}^2}$, when the experimental sample was irradiated in R polarization. For T polarization, the lowest fluence, at which the damaged site was observed, was at the level of ${35.5}\;{{\rm J/cm}^2}$. In order to evaluate the intrinsic resistivity of the coating more accurately, damage thresholds were also evaluated at 50% probability level, therefore neglecting the influence of low resistivity defects [18]. Re-evaluated values were ${39.0}\;{{\rm J/cm}^2}$ and ${48.5}\;{{\rm J/cm}^2}$ for the R and T polarizations, respectively. All of the damaged sites were inspected with optical microscopy and measured by a profilometer in order to determine the possible causes for induced damage. Two typical damaged sites are presented in Fig. 4(b). Several spots are visible in both cases, indicating the defect-based damage morphology. Both images were taken of damaged spots at the fluency level with probability of 100%. Afterwards, damaged sites were measured by a profilometer for depth determination. For R polarization, the depth of the damaged sites increased from 234 nm to 523 nm when the energy fluence was changed from ${42.7}\;{{\rm J/cm}^2}$ to ${65.1}\;{{\rm J/cm}^2}$. For T polarization, no tendency for the depths of the damaged sites was found. Measured depths varied from around 500 nm to 3000 nm at similar levels of fluence.

 

Fig. 4. Resistance to laser radiation of the high-contrast polarizer: (a) damage probability for T and R polarizations and (b) optical microscopy (black bars represent 100 µm) images representing typical damage morphologies.

Download Full Size | PDF

The presented research demonstrates the polarizer, which operates at a zero AOI and is highly resistive to laser radiation. Instead of forming a Brewster type design, GLAD technology enabled us to form anisotropic layers. Accordingly, instead of isotropic films with high and low refractive indices, only silica material has been used by varying the inner structure of individual layers and, therefore, obtaining necessary birefringence for polarizing functions at a zero AOI. However, few issues still have to be addressed for technological implementation in microlaser systems.

The main concern regarding the presented component is its porous structure. The coating is sensitive to environmental changes due to open voids of the silica structure. Several investigations indicate the adsorption of water on the structures of silica films [19]. One of the main applications of the presented polarizers is to deposit them on the active medium. Typically, such components are sealed or otherwise protected from the environment. Therefore, only handling throughout the transportation must be addressed.

Other concern is regarding the mechanical stability of the coating. Single layer columnar structures are extremely fragile and mechanically unstable. The presented coatings are covered with a dense silica layer as the last film of the coating. It provides the mechanical endurance during the mechanical cleaning of residues after the deposition process. For very thick multilayer coatings of polarizers, the top dense coating can be insufficient. Therefore, dense silica layers can be incorporated within the overall coating design without drastically affecting the spectral properties or optical resistivity.

All of aforementioned issues have to be tackled in the future works.

Current research presents an all-silica polarizing coating, which can be applied on virtually any substrate. Spectral performance can be controlled individually for each perpendicular polarization by changing the number and thickness of isotropic and anisotropic layers. Since all films consist of the same material, namely ${{\rm SiO}_2}$, overall coating can withstand extreme laser radiation power densities without any damage, hence, opening new possibilities for high power compact solid state laser systems.