Interface and material engineering for zigzag slab lasers

Laser damage of zigzag slab lasers occurs at interface between laser crystal and SiO2 film. Although an additional HfO2 layer could be used to manipulate electric-field on the crystal-film interface, their high absorption and polycrystalline structure were unacceptable. SiO2 was then doped in HfO2 to suppress its crystallization and to achieve low absorption by annealing. HfxSi1−xO2 nanocomposite layers were then inserted between laser crystal and SiO2 film to minimize electric-field at crystal-film interface. Laser damage resistance of this new architecture is two times higher than that of traditional zigzag slab lasers.

The doping concentration of Nd:YAG crystal is about 0.05%. A 3 μm SiO 2 coating is deposited on the Nd:YAG crystal to achieve TIR and it is also connected to heat sink. The working laser passes a zigzag path between two TIR surfaces in the crystal and the pumping laser enters from the end of the crystal. It is difficult to directly study the laser damage characteristics of the zigzag slab lasers, therefore, an analog using Nd:YAG prisms coated with SiO 2 film was studied instead, as shown in Fig. 1(b). The Nd:YAG prisms with 0.05% doping and a size of 30 millimeters to a side were used for this study. The incident angle at crystal-film interface is set as 55° to achieve TIR.
The EFI distribution within the Nd:YAG crystal and SiO 2 thin film is given in Fig. 1(c). The standing-wave EFI is created in the subsurface of the crystal and the evanescent-wave EFI evolves in the SiO 2 coating. Because S-polarized EFI rather than P-polarized EFI could be reduced to zero at abrupt interfaces, only S-polarized laser damage is discussed here. The difference between S-polarized and P-polarized laser damage can be found in our previous work 16 .
The LIDT testing was performed using 1064 nm, 10 ns pulses from a Nd:YAG laser having a TEM 00 mode. The raster scan method was used to obtain the LIDTs 17,18 . The LIDT of the traditional TIR configuration is about 10 ± 2 J/cm 2 ( Table 1). The cross-sectional view of one representative damage morphology was revealed using a focus ion beam (FIB) technique, as shown in Fig. 1(d). It can be seen that the depth of the damage site is equal to the thickness of the SiO 2 coating, meaning that the laser damage initiated from the crystal-film interface. Both standing-wave EFI and evanescent-wave EFI are close to their maximum value at the crystal-film interface where nano-sized absorbers are concentrated, laser damage occurred preferentially at this interface, which reflects that evanescent-wave EFI at least plays a similar role in laser damage like standing-wave EFI. The formation of the damaged crater can be explained phenomenologically, as given in Fig. 1(e). The nano-sized absorbers at crystal-film interface is heated by the laser pulse to a very high temperature at which the defect-surrounding matrix is converted to the absorbing medium through photoionization by ultra violet (UV) radiation 19 , thermionic emission of electrons 20 and the heat-transfer-induced band-gap collapse 21 . When the materials are heated to the melting temperature, thermomechanical damage occurs and results in blistering of the coating from Nd:YAG crystal.
Minimizing EFI at crystal-film interface using HfO 2 high index material. EFI in the traditional TIR configuration is fixed, additional approach must be found to engineer EFI at the crystal-film interface to improve LIDT. The HfO 2 film was explored to engineer the EFI at the crystal-film interface due to its optimal properties in the region from near ultraviolet to near infrared 22 . It was inserted between the Nd:YAG crystal and SiO 2 coating to minimize EFI at crystal-film interface, as shown in Fig. 2(a). By choosing thickness of the HfO 2 layer to be 230 nm, a destructive interference was created at the crystal-film interface and the EFI at this interface was decreased to zero, as given in Fig. 2(b). The side effects of adding a HfO 2 layer is that evanescent-wave EFI at the new HfO 2 -SiO 2 interface is quite strong.
SiO 2 and HfO 2 coating on large optics is usually prepared using electron beam evaporation (EBE) process. However, the packing density of coatings prepared by EBE process is also low, and a lot of voids exist in the coatings, making their environmental stability and thermal conductivity poor. Here, the SiO 2 and HfO 2 coating was prepared using Ion assisted deposition (IAD) to achieve dense microstructure with good environmental stability and thermal conductivity. The absorption of SiO 2 coatings prepared by IAD process could be very low, however,  HfO 2 coatings prepared by IAD process are always highly absorbing 23 . Its absorptivity was 2 orders higher than that of the SiO 2 coating, as given in Table 2.
The damage testing indicated that LIDT of this TIR configuration was only 6 ± 2 J/cm 2 (Table 1) and it is even worse than that of the traditional TIR configuration. The cross-sectional view of the representative damage morphology was examined to explain the LIDT result. Figure 2(c) shows that the laser damage initiated from the HfO 2 -SiO 2 interface. The reason is that the highly absorbing HfO 2 coating leads to a concentration of intensive nano-sized absorbers at HfO 2 -SiO 2 interface. Laser damage was created at this interface due to strong standing-wave EFI and evanescent-wave EFI here. The LIDT result reflects that the HfO 2 -SiO 2 interface is even more vulnerable to the laser damage than the crystal-film interface.
Our previous study showed that, when the absorption of both coatings is very low, the HfO 2 -SiO 2 interface is much more resistant to laser damage than the crystal-film interface 11 . So, annealing was attempted to reduce the absorption of the HfO 2 coating. The annealing temperature increases from 400 to 600 degrees centigrade with an increment step of 100 degrees centigrade. Table 2 shows that annealing at 600 degrees centigrade reduced the absorptivity of HfO 2 coatings by two orders and its absorption reaches to the level that is close to that of SiO 2 film. However, the cracking of the annealed HfO 2 film was identified using a Normarski microscope, as shown in Supplementary Figure S1. Previous studies reported that the crystalline states of HfO 2 films were changed during annealing, leading to volume change of the films, inducing strong associate stress and resulting in film cracking 22,24 . X-ray diffraction spectrometry (XRD) measurement revealed the change of the crystalline states during annealing, as given in Fig. 3(a). The XRD spectra shows that the HfO 2 films are polycrystalline and these films most contain the monoclinic phase of HfO 2 with preferred (111) orientation. The intensities of all the peaks become significantly higher when the annealing temperature increases. The large increase in peak intensity after high temperature annealing indicates that crystallite size has increased significantly. This leads to volume change of the films, induces strong associate stress and results in film cracking. As a result, annealing is not suitable to reduce the absorption of HfO 2 films.
Reducing EFI at crystal-film interface using Hf x Si 1−x O 2 nanocomposite. Two approaches have been proposed to suppress the crystallization of HfO 2 coatings during annealing. Nanolaminate is one approach that adds thin amorphous layers periodically into HfO 2 coatings to interrupt the crystallization 25 . However, this   approach introduces many interfaces that will reduce the thermal conductivity and also reduce LIDT. It is not suitable to the current application. Nanocomposite is another approach that dopes amorphous material such as SiO 2 in HfO 2 to suppress crystallization [26][27][28] . It is possible to anneal Hf x Si 1−x O 2 nanocomposite to reduce its absorption without changing its microstructure.
Here, Hf x Si 1−x O 2 nanocomposite was prepared using a co-evaporation process with IAD. It was explored to engineer EFI at the crystal-film interface. The schematic of the configuration using Hf x Si 1−x O 2 nanocomposite is given in Fig. 4(a). The refractive indices of the Hf x Si 1−x O 2 nanocomposites can be varied in a wide range. Figure 4(b) presents the EFI distributions using Hf x Si 1−x O 2 nanocomposites with three different refractive indices. Although zero EFI at crystal-film interface could be achieved by selecting proper thickness of each Hf x Si 1−x O 2 nanocomposite layer, higher refractive index of the Hf x Si 1−x O 2 nanocomposite results in lower EFI in the coatings. It is beneficial to use the highest refractive index of the Hf x Si 1−x O 2 nanocomposite layer that does not crystalize at high annealing temperature. But the limitation is that the higher refractive index, the lower concentration of SiO 2 in Hf x Si 1−x O 2 nanocomposite, and the worse resistant to crystallization at high annealing temperature. In our experiments, the refractive index of 1.87 at the wavelength of 1064 nm is the highest one that was still amorphous when the Hf x Si 1−x O 2 nanocomposite layer was annealed at 600 degrees centigrade. When its layer thickness is 280 nm, the EFI at the crystal-film interface is zero.
The unannealed Hf x Si 1−x O 2 nanocomposite layer was still highly absorbing. High temperature annealing effectively reduced its absorptivity to the level that is comparable to the SiO 2 layer, as given in Table 3. The XRD measurement of the Hf x Si 1−x O 2 nanocomposite film annealed at different temperatures is given in Fig. 3(b). There are no sharp diffraction peaks in the XRD spectra of unannealed and annealed Hf x Si 1−x O 2 films, which shows that the Hf x Si 1−x O 2 films maintained amorphous microstructure during annealing.
The LIDT of this novel TIR configuration is 18 ± 2 J/cm 2 (Table 1), which is almost 2 times higher than that of the traditional TIR configuration (10 ± 2 J/cm 2 ). A cross-sectional view of the representative damage morphology of this TIR configuration is given in Fig. 4(c). The damage site has a larger and more irregular shape, and its depth is much deeper than the thickness of coatings. It can be seen that the laser damage initiated from a deep location within the subsurface of Nd:YAG crystal. EFI at the crystal-film interface is zero, laser damage would not start from here. Moreover, the absorption of the annealed Hf x Si 1−x O 2 nanocomposite is very low, so the Hf x Si 1−x O 2 -SiO 2 interface has less nano-sized absorbers and is more resistant to laser damage compared to the crystal-film interface and the subsurface of the Nd:YAG crystal. It also reflects that the evanescent-wave EFI does not play a more detrimental role in laser damage than standing-wave EFI. As a result, only the nano-sized absorbers in the subsurface of the Nd:YAG crystal can trigger the laser-induced damage at the depth where peaks of EFI are located 11 . The absorptivity and density of nano-sized absorbers in the subsurface is much less than that at crystal-film interface, so the LIDT of this novel TIR configuration is almost 2 times higher than that of traditional one.

Conclusion
Interface and material engineering were performed to improve LIDT of zigzag slab lasers. A layer of high index material of HfO 2 was inserted between the Nd:YAG crystal and SiO 2 film to reduce the EFI at the crystal-film interface to be zero. However, the absorption of HfO 2 layer is too big and it was not practicable to reduce its absorption by annealing. Hf x Si 1−x O 2 nanocomposite was then synthesized using co-evaporation process to solve this issue. Hf x Si 1−x O 2 nanocomposite layer maintained its amorphous microstructure during annealing and achieved extremely low absorption. A novel configuration of zigzag slab lasers was proposed using Hf x Si 1−x O 2 nanocomposite layer, whose LIDT was about two times higher than that of the traditional configuration of zigzag slab lasers.

Methods
Sample preparation. The Nd:YAG prisms were first carefully cleaned by ultrasonic cleaning process. The SiO 2 films and HfO 2 films were prepared using ion assisted deposition (IAD) process, and the Hf x Si 1−x O 2 nanocomposite films were prepared using co-evaporation process with IAD techniques. A schematic of co-evaporation process with IAD techniques is shown in Supplementary Figure S2. The Hf x Si 1−x O 2 nanocomposite film was deposited on the substrates by co-evaporating HfO 2 and SiO 2 with IAD techniques. The two quartz crystal monitors were used to monitor the evaporation rates of two materials, and the two programmable masks were used to control the evaporation rates of two materials. The refractive indices of the Hf x Si 1−x O 2 nanocomposites can be varied in a wide range with different relative evaporation rates of two materials. The details of the cleaning and deposition processes can be found in our previous papers 18,29,30 . Annealing processes of the films were performed in air at atmospheric pressure at 400, 500 and 600 °C for 3 hours.
Characterization methods. The LIDT testing was performed using 1.064 μm, 10 ns pulses from a Nd:YAG laser having a TEM 00 mode, a beam diameter of 1 mm and a repetition rate of 10 Hz. The raster scan method 17 was used to obtain the LIDTs. The laser fluence increased to the "damage fluence" with a 2 J/cm 2 increment. The damage morphologies may change during the raster scan test, so a single shot method was used to obtain representative damage morphologies. The damage morphologies were observed and characterized by an optical microscope, a scanning electron microscope (SEM) and a focus ion beam (FIB) equipment. X-ray diffraction spectrometry (XRD) was used to characterize the crystallization of films with Cu Kα as the X-ray source. Data were collected in the θ-2θ (locked couple) mode from 2θ of 20-70° with a step size of 0.02°. Optical microscope was used to identify the cracking of the annealed HfO 2 coatings.
Photo-thermal technique was used to measure the absorptions of the films. The details of the absorption measurement can be found in our previous paper 23 . The absorptivity of coatings can be calculated by knowing the electric-field intensity (EFI) within them using the formula where A is the absorption, α is the absorption coefficient, n 0 and n 1 are the refractive index of incident medium and film, E is the EFI and l is the thickness of the layer. The absorptivity of SiO 2 film and HfO 2 film prepared using electron beam evaporation (EBE) process and IAD process respectively is given in Supplementary Table S1.