Correlation between Photoluminescence Properties of Surface Defects and Laser-Induced Damage Threshold of Fused Silica

Laser-induced damage threshold is the main limitation for fused silica optics in high-power laser applications. +e existence of various defects near the surface is the key factor for the degradation of the threshold. In this work, the photoluminescence spectra at different regions of the damaged and recovered fused silica samples are recorded to analyze the correlation between photoluminescence of surface defects and laser-induced damage threshold.+e experimental data concluded the inverse proportional correlation between fluorescence and laser-induced damage threshold value. +e weak photoluminescence is the guarantee of the high laser-induced damage threshold, and then the higher local Si nanocluster concentration corresponds with the higher laserinduced damage threshold value for the fused silica optics after CO2 laser treatment. +e investigation reveals that photoluminescence measurement can be employed to check the quality of pristine fused silica and evaluate the tendency of the laserinduced damage threshold value. +e current results are helpful for understanding the evolution of interaction from CO2 laser treatment and fused silica optics and can provide the guide of process technology for the high quality of fused silica optics.


Introduction
Fused silica optics is widely used in high-power laser facilities such as inertial confinement fusion (ICF) due to its high transparency and excellent thermal and chemical stabilities. However, there exist some intrinsic defects and many kinds of flaws on the surface associated with the fabrication process of fused silica optics. ese defects could absorb ultraviolet laser and cause the initial laser damage [1] and then further degrade the laser-induced damage threshold (LIDT) of fused silica optics. Some approaches, such as ion beam etching [2,3], chemical etching [1,4], and CO 2 laser treatment [5,6], have been employed to mitigate the initial damage sites. Among these methods, CO 2 laser treatment is one important method to repair the initial damage site by heating effect. Some investigations demonstrated that the LIDT may return to the normal level or higher after CO 2 laser treatment [7,8]. Although there have been some reports on the optical and mechanical effects of CO 2 laser treatment for scratches and laser-induced damage sites on the surface of fused silica, as well as the physical model of the secondary treatment or annealing process to eliminate residual stress [9,10], there is no report about the evolution of defects on the fused silica surface during this process.
Some works revealed that the intrinsic defects and laserinduced structural defects in fused silica optics include the E′-center [11], which comprised an unpaired electron of single silicon bonded to just three oxygens, oxygen-deficiency center (ODC) [12], nonbridging oxygen hole center (NBOHC) [13], peroxy radical (POR) [14], and Si nanocluster [15]. e presence of defects in the silica matrix can dramatically change its structural, electrical, and optical properties. So, photoluminescence (PL) spectra are one significant tool to understand the evolution of the defects in the fused silica optics during high-power laser irradiation.
According to PL analysis, the electron structural information of various defect species and the distribution of defects on the surface should be obtained for fused silica optics.
Some groups investigated the luminescence properties of the defects in fused silica in the past few decades. Demos et al. assigned the luminescent bands around 400 nm and 650 nm to ODC and NBOHC defects [16], separately. e other emission bands in the range from 450 nm to 600 nm are attributed to the combination of the self-trapped exciton (STE).Švrček et al. concluded that the emission band of the Si nanocluster can be in a wide range around 700 nm. e concentration of various defects in the laser-induced initial damage site would increase dramatically compared with the pristine area. e initial damage site will probably expand exponentially due to higher defect density as reported [17][18][19].
In this paper, PL properties of defects near the surface for the laser damage site and CO 2 laser treatment site are systematically investigated. e correlation between the distribution of different defects and LIDT in the same site is analyzed.

Experimental
e experiments were conducted on UV-grade Corning 7980 fused silica samples with 40 mm square and a thickness of 4 mm. ey were firstly etched in the buffered hydrofluoric (HF) acid solutions and then cleaned by deionized water.
Laser-induced damage sites were generated on the output surface of pristine samples by using a triple-frequency Nd:YAG laser (Saga) which delivers a pulse width of 6.3 ns at 355 nm with a diameter of 800 μm at 1/e 2 . e diameter of the damage area is about 250 μm. e CO 2 laser of the repairing treatment is a commercial radio frequency power excited CO 2 laser (Coherent GEM-100L), with a maximum output power of 100 watt and power stability of ±3%.
e photoluminescence spectra were recorded by using a spectrometer (LabRAM HR800, Jobin Yvon) with a confocal microscope and a multichannel air-cooled CCD detector. e excitation sources were an Ar ion laser (λ 0 � 514.5 nm; source power: 20 mW) and He-Cd laser (λ 0 � 325 nm; source power: 40 mW). All spectra were recorded at room temperature.
Laser-induced damage threshold is obtained by the R-on-1 testing mode. e schematic diagram of the measurement setup is shown in Figure 1. On each selected site, the laser fluence was progressively increasing, until the first damage occurred. For each set of parameters, a series of sites were selected and analyzed, and the results given in this work are based on a statistical analysis on the fused silica samples. and 2(c), respectively. e shoulder peak around 400 nm is from ODC defects. e broad emission bands centered at 490 nm and 580 nm can be definitely ascribed to the transition from the laser-induced STE [12,13]. e peaks around 650 nm and 750 nm are assigned to the luminescence bands of NBOHC and Si nanocluster defects, respectively. e weak emission indicates the low concentration for these defects in the clear pristine sample. ese defects in pristine fused silica optics can have a high absorption for the UV laser and become an initial laser damage site according to existing reports [1,11,12]. In other words, the low defects' density can guarantee the high LIDT value.

Results and Discussion
e contribution from the NBOHC and Si nanocluster is hardly observable for the pristine region, but the obvious peaks occur for the two defects in the damaged region as seen in Figure 2. Figure 3 compares the photoluminescence spectra of the pristine region, ion beam etching-treated region, and the edge region of the laser-induced damage site with the excitation of a 514.5 nm laser.
e sharp peaks at 526 nm, 536 nm, and 543 nm are the response to Raman modes for Si-O-Si, fourfold rings, and threefold rings in fused silica, respectively [11]. e broad weak fluorescence of curve #2 in Figure 3 for the pristine region is similar to that in Figure 2 with 325 nm excitation. Ion beam etching process can smooth the surface of the optics and mitigate a part of the existing surface defects.
e microscopic image of the etching region is shown in Figure 3(c). It is smoother and has higher uniformity compared with the pristine region.
Except for Raman peaks, no emission bands can be detected for the ion beam etching-treated sites as seen in curve #3 of Figure 3(a).
is verified that the ion beam etching process can remove the flaws localized on the surface and enhance the LIDT value. e strong luminescent band centered at 580 nm is quite prominent as shown in curve #1 of Figure 3(a) for the edge region of the laser-induced damage area. It indicates that the laser-induced damage process generates a higher concentration of defects. In general, the type and relative concentration of defects depend on the peak position and intensity of PL spectra, respectively. As the description given above, the laser-induced self-trapped excitons dominate the emission around 580 nm. ese defects could result in the damage of fused silica under ultrahigh power running of the UV laser due to the strong absorption. Investigating the influence on the LIDT from these defects, the values of LIDT at different sites are measured by the R-on-1 testing mode [20]. e LIDT, the type, and the relative concentration of defects for different regions are given in Table 1.
e LIDT values for the pristine surface and ion beamtreated region are 13.9 J/cm 2 and 17.2 J/cm 2 , separately. e low LIDT value of 4 J/cm 2 is the response to the edge region of the laser-induced damage site. e results from Table 1 reveal that the PL intensity is inversely proportional to the LIDT value of the fused silica optics. Further comparing the contribution from different defects for the pristine region and damaged region, the laser-induced STEs dominate the damaged threshold value. ese results indicate that the PL can be employed to check the quality of pristine fused silica and evaluate the possible LIDT value based on the confirmed reference standard sample. CO 2 laser treatment is the most effective method for fused silica materials to mitigate the defects on the surface. Figure 4(a) shows the optical microscopy image of the CO 2 laser treatment site. e PL spectra of three different regions are shown in Figure 4(b). is laser-induced damage site has been overtreated by the CO 2 laser, and the ablation trace can be observed in its edge region. e PL spectrum for the edge region has relative high intensity around 500 nm from the laser-induced STE and the shoulder at 400 nm from the ODC according to the previous reports [13,14]. e measured LIDT value in this region is 4.5 J/cm 2 , and it supports the same inverse proportional correlation between PL band intensity centered at 500 nm and the LIDT.
In spite of the weak fluorescence for center regions 1 and 2, a new luminescence band centered at 750 nm obviously occurs.
is band is ascribed to the Si nanoclusters. e fabrication process and Si-implanted a-SiO 2 after 1100°C annealing can generate the nanoclusters [21]. ese inducing paths are all related to the CO 2 laser heating process. e intensity at center region 1 is higher due to near the center and bathing in higher temperature during the laser irradiation.
ese results are quite helpful to understand the mechanism of defects' evolution of the CO 2 laser treatment process. It should also be noticed that the LIDT value in center region 1 of the CO 2 laser-treated site is 19.2 J/cm 2 higher than 17 J/cm 2 for center region 2 as shown in Figure 5. is indicates that CO 2 laser-treated fused silica optics can run under higher fluence compared to the pristine fused silica optics.
Combining Table 1 with Figures 4 and 5, we can conclude that the weak photoluminescence can be taken as the criterion of the high LIDT value; furthermore, the higher nanocluster contribution protects the damage of fused silica under higher laser power running. In fused silica materials,  e PL intensity is normalized. e intensity of the pristine region possesses 40% of the intensity for the damaged site. e value in the bracket for the pristine region is the relative ratio to the PL intensity of the damaged region. the self-trapped exciton is known to be correlated with the absorption precursor according to the existing reports; this is consistent with the results shown in Figure 5. e relationship between the LIDT and Si nanocluster needs to be studied further.

Conclusions
PL technique combined with the laser-induced damage threshold measurement is employed to investigate the distribution of the defects on the surface of the fused silica sample. e inverse proportional correlation between fluorescence and LIDT value is built according to the experimental results. e weak photoluminescence and the higher Si nanocluster concentration can enhance the LIDT value for the fused silica optics after CO 2 laser treatment. e investigation reveals that the fluorescence can be employed to check the quality of pristine fused silica and evaluate the possible laser-induced damage threshold value. e wide emission from the ODC and self-trapped excitons dominates the low LIDT value. e current results are helpful for understanding the evolution of action from CO 2 laser treatment and fused silica optics.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Disclosure
Qiao Chen and He Wang are the coauthors.

Conflicts of Interest
e authors declare that they have no conflicts of interest.  Figure 4. Each photoluminescence spectrum area is normalized to that at the edge region.
Laser and Particle Beams 5