We investigate defects forming in Ce3+-doped fused silica samples following exposure to nanosecond ultraviolet laser pulses and their relaxation as a function of time and exposure to low intensity light at different wavelengths. A subset of these defects are responsible for inducing absorption in the visible and near infrared spectral range, which is of critical importance for the use of this material as ultraviolet light absorbing filter in high power laser systems. The dependence of the induced absorption as a function of laser fluence and methods to most efficiently mitigate this effect are presented. Experiments simulating the operation of the material as a UV protection filter for high power laser systems were performed in order to determine limitations and practical operational conditions.
© 2014 Optical Society of America
Cerium doped crystalline and glass materials have been extensively studied for their importance in a range of applications. Specifically, trivalent cerium doped materials have been investigated and/or utilized as a) detectors of elementary particles and ionizing radiation, b) phosphors, c) ultraviolet radiation absorbing filters, emitters and activators for energy transfer and d) tunable solid state laser materials suitable for operation in the blue and near ultraviolet spectral region. The cerium ions are found in the trivalent (Ce3+) or tetravalent (Ce4+) state [1–5 ] and their relative concentration within the host material largely depends on the oxidizing-reducing conditions during the material fabrication. The energy separation between the lowest 5d levels and the 4f levels in Ce3+ is generally large and variations between different host materials arise from the strength of the interaction between the Ce3+ electrons and the electrons of the surrounding host material ions .
Typically Ce4+ ions give rise to a broad absorption spectrum in the ultraviolet region due to a charge transfer band but exhibit no luminescence [1–5 ]. On the other hand, Ce3+ is one of the few rare-earth ions that present electric-dipole allowed transitions between 4f and 5d electronic states resulting in large absorption coefficients and short luminescence lifetimes. The luminescence spectrum is typically broad, covering significant parts of the near ultraviolet to blue spectral region and has lifetimes on the order of tens of nanoseconds. Phosphorescence and solarization are characteristic traits that are commonly observed in Ce3+ doped materials arising from different types of defects formed in the material.
Solarization refers to the loss of transmission in the visible and near infrared spectral range following exposure to high intensity laser radiation at resonant (ultraviolet) excitation frequencies [1,7,8 ]. Solarization (along with excited state absorption) is a major limiting factor for the use of Ce3+ as the active ion for lasing in many host materials, causing suppression of optical gain. In implementations as ultraviolet (UV) absorbing filter, such as in laser flash lamps, solarization introduces transmission losses in the nominally transparent spectral range. This effect is attributed to the formation of color centers via excited state absorption which are stable at room temperature having lifetimes on the order of a few days or longer. It has been shown that exposure of the material to lower intensity laser pulses can reduce the density of the color centers (solarization-induced absorption) [1,7 ].
Phosphorescence refers to a prolonged emission following exposure to UV light, high-energy particles, or ionizing radiation [9–11 ]. It represents a major limiting factor for the use of Ce3+ as the active ion in scintillating materials and has been assigned to the recombination of trapped electrons and holes [12,13 ]. The lifetime of the phosphorescence at room temperature is typically on the order of tens of minutes and is strongly dependent on the temperature.
A number of crystalline structures have been investigated as hosts of Ce3+ ions, mainly as scintillator  or laser materials [7,15,16 ]. In addition, various glass compositions have been explored as alternative hosts due to their potential to lower the cost of mass production, to enable higher doping levels and for their flexibility to encompass various shapes and sizes including the fabrication of fibers [8,10, 17–22 ]. One of the key characteristics of the optical properties of some Ce3+ doped glasses is the change in the absorption spectrum in the ultraviolet region under high intensity resonant excitation. This change has been assigned to the loss of one electron by the Ce3+ ions leading to a new center (referred to as Ce3++) that is not identical to Ce4+, exhibits no luminescence and has an absorption spectrum in the UV similar to that of Ce4+ ions [2,5 ]. The differences in the absorption spectra between Ce3++ and Ce4+, arises from the difference in the interaction with the nearest neighbors, which have different configurations (sites of Ce3+ incorporation are different from those of Ce4+).
In this work we investigate the formation and the relaxation kinetics of defects in Ce3+ doped fused silica samples following exposure to nanosecond (ns) UV laser pulses. This work is motivated by the potential use of this material as an ultraviolet light absorbing filter in high power, large aperture laser systems that generate near ultraviolet light via frequency conversion of near infrared pulses (such as in neodymium doped glass lasers). Such filters may be used before the frequency conversion components to protect the optics from unwanted UV light back-reflections. The host matrix was selected due to extensive experience in producing large aperture optical components and its high resistance to laser damage. The results indicate that the solarization generated by exposure to laser pulses can be most efficiently eliminated via exposure to 400 nm light and that this effect is dose dependent. It is demonstrated that the removal of the color centers is accompanied by a visible emission that has the same spectral characteristics as that of the Ce3+, 5d→4f luminescence. Experiments simulating the operation of the material as a UV protection filter for high power laser systems are also discussed.
2. Experimental design
Various experimental systems were utilized to perform this work. Measurement of the transmission spectra were performed using a Shimadzu model UV-1600 PC spectrophotometer while light emission measurements were performed using a Jobin Yvon TRIAX 320 spectrometer equipped with a liquid nitrogen cooled detector (Princeton Instruments) under excitation obtained from an argon ion laser, compact diode lasers or narrow band LED sources. Samples were also exposed to a large aperture (≈3 cm diameter) laser beam operating at 351 nm that has been described in detail elsewhere .
Experiments designed to investigate the performance of the material as a UV absorbing filter were performed using the experimental system shown in Fig. 1(a) . It involves a Nd:YAG laser system in which the first, second and third harmonics are separated (after properly tuning the harmonic generation crystals) and passed through polarization elements to control the beam power. The pulse duration was about 3.3 ns at half width at half maximum (HWHM). The power of each beam and the pulse repetition rate (adjustable between 0.1 and 100 Hz) were computer controlled and were adjusted as needed for the execution of the experiments. In these experiments, only the first and third harmonics of the laser output were utilized (1064 nm and 355 nm, respectively) with each beam focused using a nominally 1-meter focal length lens. A fused silica plate was used to divert part of the beam to the reference arm of our diagnostic instrumentation. The samples were positioned at the focal point of the beam (indicated as position P1) or at about half the distance from the focusing lens (position P2). The beam profiles at the output of the sample at P1 were recorded using two CCD cameras (one for each wavelength) via image relaying using a 6.5 X magnification lens system. In this arrangement, 1 pixel on the CCD camera captures the image of about 1 µm2 on the output surface of the sample. The reference arm is configured using identicalinstrumentation to record the beam profile at the same location along the beam propagation path without the presence of the sample.
The two pairs of CCD cameras recording the input and output beam profiles at each wavelength were calibrated to enable the correlation of the counts recorded by the reference CCD to the energy reaching the sample. This also enabled the estimation of the fluence at each location within the beam profile recorded by the reference camera. This configuration enabled us to monitor a) the energy and spatial profile of the input beams, b) the relative intensity of the output beam compared to the input (transmittance) and, c) the beam profile and its modification due to propagation inside the sample. The spatial intensity profiles of the 355 nm and 1064 nm beams in location P1 have a nearly Gaussian shape with a beam diameter at 1/e of peak intensity of about 95 µm and 248 µm, respectively. The beam-to-beam stability exhibited modulation in both, the spatial dimensions of the beam and at the location of the center of the beam. The Raleigh range of the focused beam is estimated to be about 5.9 cm, thus providing a nearly collimated propagation of the beam through the samples (about 1 cm thick) used in this work.
Experiments to investigate the decay dynamics of color centers (associated with solarization of the material) under narrow band illumination at different wavelengths were performed using the system depicted in Fig. 1(b). A set of bright LED sources centered at about 400 nm, 460 nm, 530 nm and 660 nm were used to directly illuminate the sample while images of the sample were continuously acquired using a CCD camera at a rate of 1 image per 6 seconds. A white screen illuminated by an LED source operating at a different wavelength was used as the background “bright field” for image acquisition by the CCD camera after passing through a narrow band filter to eliminate the LED light directly illuminating the sample used to reduce the color centers. The inset show the image of a solarized site (about 2.3 mm in diameter) after exposure of the sample to the 355 nm laser beam (in location P2) in the experimental system shown in Fig. 1(a). The reduction of solarization-induced absorption was analyzed providing information the relaxation dynamics of the color centers.
Cerium doped fused silica samples were obtained from two manufacturers (Heraeus, Germany and Asahi Glass Co., Japan). The total Ce concentration was determined using sample dissolution and inductively coupled plasma mass spectroscopy (Thermo iCAPQ quadrapole ICPMS). Concentrations of Ce3+ and Ce4+ ions were quantified spectroscopically using reference samples prepared under different redox conditions as outlined in . The first sample was 1.3 cm thick and had an estimated concentration of Ce3+ and Ce4+ ions of 847 ppm and 38 ppm, respectively. The second sample was 1 cm thick and had an estimated concentration of Ce3+ and Ce4+ ions of 447 ppm and 26 ppm, respectively. The results obtained from each sample exhibited the same trends and behaviors. For this reason, only the experimental results obtained from the higher concentration sample are shown in Figs. 2-7 .
3. Experimental results
Figure 2(a) shows the transmission spectrum of the Ce-doped fused silica sample before (profile #2) and after (profile #3) exposure to 351 nm, 5 ns laser pulses at a fluence of about 5 J/cm2. The corresponding spectrum of un-doped fused silica (profile #1) with the same thickness has been included for comparative purposes. The absorption edge at about 400 nm is due to the Ce3+ ions. The increased absorption in the visible and near infrared spectral region after exposure of the sample to the UV laser pulses is assigned to the formation of color centers (solarization) and is parallel to previous observations of “coloring” following exposure to ultraviolet or ionizing radiation and similar to previous observations in other Ce3+ doped glass and crystalline materials [1,7,8 ]. In addition, the difference observed in the absorption spectrum of the doped material in the 200 to 250 nm region before and after solarization is attributed to the formation of Ce3++ species [2,5 ]. Subsequent excitation of the sample with low intensity ultraviolet light (obtained from a 365 nm LED light source) reveals that the exposed (to the UV laser pulses) site also exhibits a slightly reduced photoluminescence, an observation that is also suggesting that there is photo-oxidation of some of the Ce3+ ions associated with the formation of Ce3++ species. The photoluminescence spectrum during 355 nm pulsed excitation is shown in Fig. 2(b). This spectrum is similar to that reported previously using Ce3+ doped silica samples synthesized using the sol-gel method but the lifetime of the emission (shown as inset in Fig. 2(b)) was measured to be 96 ± 0.5 ns, slightly longer to about 50 ns reported in sol-gel synthesized silica samples .
Following exposure to high fluence 355 nm pulses, the material exhibits a characteristic long-lived emission (phosphorescence), typical in Ce3+ doped materials [9–11 ]. The spectral profile of the phosphorescence shown in Fig. 3(a) is practically identical to that of the photoluminescence (shown Fig. 2(b)). The decay of the phosphorescence at room temperature was found to exhibit power-law dependence (≈t−1) with a lifetime on the order of 1 min. as shown in the insert of Fig. 3(a). This emission has been assigned to the recombination of trapped electrons and holes involving the 5d→4f transition during the relaxation process. This relaxation pathway commonly leads to power-law relaxation dynamics [12,13 ].
A salient finding during the execution of the experiments is the observation that the solarized material exhibits a characteristic emission, visible even via naked eye, when exposed to CW light in the blue or green spectral region. This was observed using blue and green LED light sources as well as the laser output from an argon ion laser and a 530 nm diode laser. The observation of this emission using 400 nm or shorter wavelengths is hindered by the much stronger produced photoluminescence and was not measured in this set of experiments. It was also found that this emission terminates with the removal of the solarization (color centers). It is therefore assumed to be part of the relaxation process of the electrons trapped in the color centers after they absorb a photon in the visible range. The emission spectrum from a solarized site immediately after turning-on a CW diode laser operating at 530 nm is shown in Fig. 3(b) (profile #1). The peak #2 arises from the Rayleigh scattered 530 nm laser light as it transverses the sample. The spectrum obtained after subtraction of the recorded spectra at the onset of excitation with the 530 nm CW laser beam and 20 seconds later is denoted as profile #3. This spectrum represents the transient emission generated during the removal of the solarization of the material and is assigned to the relaxation of electrons trapped at the color centers. This spectrum is identical under any excitation wavelength but also is identical to that of the photoluminescence and phosphorescence of the material (Figs. 2(a) and 3(a) , respectively). This suggests the involvement of 5d→4f radiative transition in the process of removal of the solarization, suggestive of the relaxation pathway of the trapped electrons.
A schematic depiction of the excitation and relaxation pathways, leading the emission spectra shown in Figs. 2(b), 3(a) and 3(b) , is shown in Fig. 3(c). Absorption of UV photons facilitates the population of the excited state followed in part via the radiative relaxation leading to the emission signal shown in Fig. 2(b). In the presence of a high intensity pulses, excited state absorption (ESA) leads to excitation of electrons in the conduction band which subsequently return to the excited state via non-radiative relaxation (NRR). However, a number of electrons are trapped by shallow traps or color centers. In the first case, thermal excitation leads to recombination of the trapped electrons with holes resulting in the phospo--rencence shown in Fig. 3(a) [9–11 ]. Absorption of light by electrons trapped in the color centers leads to the absorption in the visible spectrum (solarization effect), which represents transitions between states within the color center electronic system. The experimental results indicate that absorption of light can lead to return on the electron to the Ce3+ system (with effectiveness increasing with decreasing wavelength) leading to the generation of the emission shown in Fig. 3(b). In all cases, the origin of the emission is the same and as a result, the spectral profiles of the observed emission spectra are the same (see Figs. 2(b), 3(a) and 3(b) ).
To more accurately quantify the solarization effect, experiments were performed using a large aperture laser system . A sample with dimensions of 5 x 5 x 1.3 cm that had all sides polished was exposed to a single 351 nm laser pulse having nearly flattop temporal shape with temporal duration of about 5 ns and nearly flattop spatial intensity distribution with an average fluence of about 5.3 J/cm2 and a diameter of about 3.3 cm. Following exposure, the sample was protected from any exposure to ambient light while transferred to the experimental system depicted by Fig. 1(b). The recorded side-view (obtained through the polished side of the sample) and front-view images of the sample are shown in Figs. 4(a) and 4(b) respectively. The induced solarization (color centers) results in the area exposed to the laser beam to appear as a dark area in the image shown in Fig. 4(b). On the other hand, the side view image of the sample shown in Fig. 4(a) demonstrates that the density of the color centers (transmission loss) inside the sample is decreased with distance from the input surface (left side of the sample in the image shown) and nearly diminishes at the exit surface. This is not surprising as the intensity of the laser beam decreases with propagation inside the sample due to the strong absorption by the Ce3+ ions. Since the fluence of the laser beam inside the sample can be estimated and the induced solarization (transmission loss) can be experimentally measured, the dependence of the solarization induced absorption as a function of depth, and thus laser fluence, can be obtained. In addition, as the dimension of the solarized region is known, an estimation of the solarization induced absorption coefficient as a function of laser fluence can be obtained.
The result of this estimation is shown in Fig. 4(c), providing the solarization induced absorption coefficient at 530 nm (used for imaging) as a function of the laser fluence. The absorption coefficient at any other wavelength can be estimated using the spectral profile of the solarization-induced absorption shown in Fig. 2(a). The results show that the absorption coefficient increases with laser fluence nearly linearly between about 0.2 and 2 J/cm2 and thereafter saturates and even decreases for higher fluences. This salient behavior will be discussed in more detail later.
As mentioned earlier, the intended application of the material under investigation is to be used as a UV absorbing filter to protect optical components of the laser that are designed to handle only the fundamental frequency. However, the induced solarization extends absorption in the spectral range of the fundamental laser frequency causing undesired loss of laser power.
It has been previously demonstrated [1, 7 ] that exposure of the solarized region to laser pulses in the green spectral region or lower intensity UV light reduces the strength of solarization-induced absorption. This suggests that the process of solarization removal may be associated with linear excitation. We hypothesized that it may be possible to remove the solarization using low intensity CW light sources. To explore this hypothesis, we utilized a series of LED light sources centered at 660 nm, 530 nm, 460 nm and 400 nm having a FWHM bandwidth between 15 and 25 nm. The experiments were performed using the experimental system depicted in Fig. 1(b). These experiments were performed following exposure of the sample to 355 nm laser pulses at about 4 J/cm2 using the experimental system shown in Fig. 1(a) (sample position P2).
Figure 5 shows a typical example results representing the recovery of the solarization induced transmission loss through the sample at 530 nm as a function of time of exposure of the solarized site to 460 nm CW light at three different radiative flux densities of about 3.5, 8.8 and 21 mW/cm2. The results indicate that such illumination conditions can effectively remove the solarization-induced absorption. The results shown in Fig. 5 were obtained using CW light, supporting our hypothesis that the relaxation of the electrons from the color center sites involves a linear absorption process. Thus, the overall removal of solarization-induced absorption may be only dependent on the accumulated exposure to the CW light and not on the instantaneous excitation flux density. This was confirmed by plotting the transmission as a function of accumulative exposure under different excitation flux densities. The inset of Fig. 5 represents the response at three different excitation flux densities at 460 nm shown in Fig. 5 when expressed as a function of accumulative exposure. The results indicate that the behaviors measured merge to a single profile when expressed as accumulative exposure. The same behavior was observed using other excitation wavelengths. This behavior enables the predictive estimate of the exposure time needed at a specific excitation flux density in order to achieve a specific level of the solarization induced absorption at any wavelength as shown in Fig. 6.
Figure 6(a) demonstrates the recovery of the solarization induced transmission loss through the sample at 530 nm (quantified by the measured transmittance) as a function of the exposure time and accumulative exposure for four different excitation wavelengths (660 nm, 530 nm, 460 nm and 400 nm) using in all cases the same radiative flux density of about 19.3 mW/cm2. The results indicate the strong dependence of the removal of solarization as a function of wavelength with the shorter wavelength being far more effective. In particular,CW illumination at about 400 nm leads to the most efficient removal of the solarization induced absorption. As example, the measured transmittance loss at 530 nm after accumulative exposure to about 10 J/cm2 of light at 400 nm, 460 nm, 530 nm and 660 nm are 0.012, 0.033, 0.074 and 0.15, respectively. Experiments with even shorter wavelengths were not performed because the ultraviolet absorption edge of the material (see Fig. 2(a)) extends up to about 400 nm, thus limiting constant propagation through the sample. The experimental data under 400 nm illumination are fit to an exponential decay in order to provide an empirical description of the transmittance of the solarized material after exposure to 400 nm light for practical implementation of the results of this work. Specifically, the empirical function used is expressed as:Fig. 2, the loss of transmittance at the fundamental laser frequency (1051 nm) is about 0.2 to that measured at 530 nm.
To better quantify the strength of removal of solarization for the different CW exposure wavelengths used in this work, the experimental data were used to calculate the change of the transmittance per unit time under the same radiative flux density (19.3 mW/cm2) normalized by the instantaneous transmittance loss. This provides a measure of the rate by which electrons are released from the color centers. The results are shown in Fig. 6(b) in a range over which a power low adequately fits the experimental data. The fits are of the form:
An significant concern when using this material as UV protective filter is that the repeated build-up of solarization and subsequent exposure to 400 nm light to reverse the effect could give rise to new type of defects that cannot be removed. To assess this concern, we exposed the material to 355 nm laser pulse using the system depicted in Fig. 1(a) at an average fluence of about 4 J/cm2 at a repetition rate of 0.1 Hz. The sample was simultaneously exposed to 3 LED sources operating at 400 nm that were able, during the time between exposure to 355 nm pulses (10 sec.), to reduce the transmission loss at 530 nm from about 33% to about 3%. After exposure under these conditions to 300 pulses, the sample was removed and exposed to 400 nm light until complete removal of the solarization. The sample was then reexamined and found to bare no measurable loss of transmission or other observable optical modification. The experiments were repeated multiple times with same result. This suggests that the process of repeated solarization removal does not cause any measurable deterioration of the optical quality of the material.
The experimental system shown in Fig. 1(a) was utilized to simulate operational conditions of the material as ultraviolet absorbing filter under high power pulsed laser excitation. With the sample placed in position P1, the 355 nm and 1064 nm beams were aligned to overlap in the sample. The proper alignment of the beams was confirmed by viewing their position on the output beam diagnostic camera after proper selection of the filters to enable simultaneous visualization of both beams. As the diameter of the 1064 nm beam is about 2.5 times larger to that of the 355 nm beam, the later is covering only the peak portion of the nominally Gaussian profile of the 1064 nm beam.
Typical examples of the experimental results are shown in Fig. 7 where the transmittance of the 1064 nm beam after propagation through the sample is measured before, during and after exposure to 355 nm pulses. The average fluence of the 1064 nm and 355 nm pulses was about 6 and 5 J/cm2, respectively. The first set of experiments is represented on Fig. 7 by the square data points starting with the initial pulses (denoted as “before exposure”) where only the 1064 nm beam propagated through the sample. Because the material is transparent at this wavelength, we normalize all transmission measurements using these initial data points as reference. The 355 nm pulses were thereafter turned on and were simultaneously propagated with the 1064 nm pulses without any temporal delay. The results show that there is an immediate decline in the transmission at 1064 nm beam as shown in the data group denoted as “during exposure”. The inset of Fig. 7 shows the recorded images by the diagnostic cameras of the 1064 nm pulse capturing its spatial profile during the first simultaneous exposure. Specifically, image (a) shows the 1064 nm beam profile recorded by the input beam diagnostics (equivalent to the beam entering the sample) while image (b) shows the 1064 nm beam profile recorded by the output beam diagnostics (equivalent to the beam exiting the sample). The difference in the intensity at the center of the beam arises from reduction of the transmission of the 1064 nm beam at the location irradiated by the 355 nm beam. To better quantify this effect, we normalized the intensity of the output profile to that of the input profile as shown in image (c). Analysis of this image indicates that as much as 40% of the 1064 nm photons are absorbed at the location of peak intensity of the 355 nm beam.
The transmission of the 1064 nm pulses remained reduced after termination of exposure to 355 nm pulses. This effect may be assigned to the solarization of the material by the 355 nm pulses and can be predicted to be on the order of a few percent based on the transmission spectra of the solarized material shown in Fig. 2(a). The sample was subsequently exposed to more that 150 additional 1064 nm pulses during which a slow reduction of the transmittance loss is observed (see data set on Fig. 7 denoted as “after exposure to 355 nm pulses”). A second set of measurements was performed using the same protocol with the additional feature that after exposure to 355 nm pulses, the material was left unexposed to any type of light for 10 minutes before continuing with exposure with the last set of 1064 nm pulses. This was to allow the defects species responsible for phosphorescence to completely decay (see decay profile shown in inset of Fig. 3(a)) in order to separate any possible residual absorption included by the phosphorescence related defects from the residual absorption of the solarization-inducing color centers. This second set of data is shown as solid circles in Fig. 7 and they practically overlap with the results of the first set of experiments. Consequently, we assign the residual absorption at 1064 nm to solarization induced by the exposure to 355 nm pulses and the subsequent recovery to the partial removal of the solarization arising from the exposure of the material to the 1064 nm pulses.
The absorption and emission characteristics as well as the presence of phosphorescence and solarization (see Figs. 2 and 3(a) ) are analogous to previous observations in various Ce3+ doped glass and crystalline materials. The origin and decay dynamics of the phosphorescence has been discussed in other systems and will not be further discussed in this work. However, the observation that the removal of solarization under exposure to CW light is accompanied by an emission, which is revealing a radiative relaxation involving the 5d→4f transition (see Fig. 3(b)) has not been reported before, to the best of our knowledge, in any Ce3+ doped material. It was previously postulated that the relaxation of the color center involves only nonradiative pathways  through direct relaxation to the 4f level. A related effect has been reported in the case of Ce3+ doped silicate glasses  that was referred to as “photo-stimulated luminescence”. Specifically, it was reported that when the material was exposed to He-Ne laser at 633 nm, after irradiation by X-ray, a photo-stimulated emission at about 400 nm was observed which decayed when a He-Ne laser continuously irradiated the sample. It is therefore likely that the X-rays induced solarization of the material and the subsequent exposure to CW 633 nm light reproduced the same processes reported in this work.
Visible, by naked eye, solarization is observed for laser fluences as low as 2.5 mJ/cm2 at 355 nm. Additional exposure to about 3 more pulses produces only a very small additional increase in absorption with no additional increase after that. This indicates that a balance between the new generated color center and those removed after exposure to each 355 nm laser pulse is attained . The results shown in Fig. 4 quantify the solarization induced in this material as a function of the laser fluence. The results shown in Fig. 4(c) demonstrate that the increase of the absorption due to solarization as a function of laser fluence discontinues at a fluence of about 2.6 J/cm2. In fact, the results show that there is a reversal of this behavior at higher fluences. This can be explained by considering two of the salient traits of the material, the phosphorescence and the removal of solarization via exposure to low intensity light. Specifically, we propose that the phosphorescence light causes reduction of the solarization by exposing the color centers to light centered in the blue spectral region. As the solarization due to the 355 nm pulses at higher fluences (above about 2.6 J/cm2) saturates, the subsequent exposure to light from phosphorescence, which increases with fluence, causes removal of more solarization as a function of the 355 nm laser fluence. This explains the observed decrease in the residual solarization as a function of fluence above about 2.6 J/cm2.
The results shown in Figs. 5 and 6 demonstrate and quantify the removal of solarization as a function of the CW light exposure parameters. This enables a practical implementation of this method as per the example described by Eq. (1). Specifically, the results shown in Fig. 6(a) indicate that about 99.8% transmission at the fundamental frequency (99% transmission at 530 nm) can be achieved in about 8 minutes of exposure to 400 nm light at about 20 mW/cm2, which is easily attainable using a currently available bright LED light source. Employment of LED arrays (or other CW light sources) can correspondingly reduce the required exposure time. However, the exposure time required to remove the solarization (which is also dependent on the acceptable level of solarization-induced absorption) sets a limit to the repetition rate for the operation of the laser.
The results shown in Fig. 7 provide an understanding of the transient absorption at 1064 nm arising from excited state absorption during simultaneous exposure to 355 nm light pulses. This effect is well studied in the context of Ce3+ doped materials as tunable laser [1,7,17 ]. This transient absorption decays with the relaxation of the excited state electron population, quantified by the lifetime to the photoluminescence of the material shown in Fig. 2(b). This indicates that, for the use of this material as a UV filter in high power laser application, the third harmonic pulse should follow and be temporally separated from the fundamental frequency pulse in order to avoid loss of transmission at the fundamental frequency.
The further decrease in the transmission during exposure to multiple 355 nm pulses observed in the group of data denoted as “during exposure” in Fig. 7 is assigned to two effects. First, experiments have shown that exposure to multiple pulses further increases the solarization of the material by a small amount, although most of the solarization will occur during the first exposure pulse. The second and more dominant effect in our experiments relates to the beam pointing stability. Specifically, as the center of the beam moves around covering an area that has a diameter about twice that of the each individual laser pulse, a larger part of the 1064 nm beam is exposed to a solarized material. This leads to a continuous decrease of the transmission observed in the experimental results.
The results shown in Fig. 6(b) and quantified by the fitting parameters provided in Eq. (3) indicate the increased efficiency with decreasing wavelength of low intensity light to remove the solarization induced absorption. This efficiency declines with increasing wavelength nearly exponentially and is notably different to the observed solarization induced absorption spectrum (see Fig. 2(a)) where the maximum transmittance loss is observed at 530 nm while the transmittance at 400 nm is about the same as at 660 nm. This indicates that transition to higher energy states compared to those dominating the solarization-induced absorption spectrum (within the electronic level system of the color centers involved) is required for the release of the electrons. A more detailed understanding of this process is outside the scope of this work and will be addressed in future work.
Ce3+-doped silica is a suitable material for use as ultraviolet light protective filter in high power laser systems with two main limitations that must be addressed a priori: a) The strong excited state absorption when the fundamental and third harmonic (UV) pulses temporally overlap; b) The induced solarization of the material subsequent to exposure to UV pulses. The first issue is manageable by ensuring that the temporal duration of the pulse in combination with the optical path of the back-reflected UV pulse cannot support their temporal overlap inside the material. The second issue can be addressed by exposing the material to CW light, such as the 400 nm LED sources used in our experiments. Higher intensity sources can further proportionally reduce this “desolarization” time, thus increasing the operating repetition rate of the laser system.
This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-JRNL-439171. We thank Raluca A. Negres and Mary A. Norton for helping with execution of experiments. The authors also wish to acknowledge Asahi Glass Company (AGC) and Heraeus Quartz America for contributing the Ce:Silica glass samples that were studied in this work.
References and links
1. K.-S. Lim and D. S. Hamilton, “Optical gain and loss studies in Ce3+:YLiF4,” J. Opt. Soc. Am. B 6(7), 1401–1406 (1989). [CrossRef]
2. J. S. Stroud, “Photoionization of Ce3+ in Glass,” J. Chem. Phys. 35(3), 844–850 (1961). [CrossRef]
3. M. Brandily-Anne, J. Lumeau, L. Glebova, and L. B. Glebov, “Specific absorption spectra of cerium in multicomponent silicate glasses,” J. Non-Cryst. Solids 356(44-49), 2337–2343 (2010). [CrossRef]
4. A. M. Efimov, A. I. Ignatev, N. V. Nikonorov, and E. S. Postnikov, “Spectral Components that Form UV Absorption Spectrum of Ce3+ and Ce(IV) Valence States in Matrix of Photothermorefractive Glasses,” Opt. Spectrosc. 111(3), 426–433 (2011). [CrossRef]
5. A. Paul, M. Mulholland, and M. S. Zaman, “Ultraviolet absorption of cerium (III) and cerium(IV) in simple glasses,” J. Mater. Sci. 11(11), 2082–2086 (1976). [CrossRef]
6. G. Liu and B. Jacquier, Spectroscopic Properties of Rare Earths in Optical Materials Springer Series in Materials Science (Springer, 2005), Vol. 83.
7. A. J. Bayramian, C. D. Marshall, J. H. Wu, J. A. Speth, S. A. Payne, G. J. Quarles, and V. K. Castillo, “Ce: LiSrAlF6 laser performance with antisolarant pump beam,” J. Lumin. 69(2), 85–94 (1996). [CrossRef]
8. G. B. Blinkova, Sh. A. Vakhidov, A. Kh. Islamov, I. Nuritrinov, and Kh. A. Khaidarova, “On the Nature of Yellow Coloring in Cerium-Containing Silica Glasses,” Glass Phys. Chem. 20, 283–287 (1994).
9. M. Yamaga, Y. Ohsumi, T. Nakayama, and T. P. J. Han, “Persistent phosphorescence in Ce-doped Lu2SiO5,” Opt. Mater. Express 2(4), 413–419 (2012). [CrossRef]
10. N. Chiodini, M. Fasoli, M. Martini, E. Rosetta, G. Spinolo, A. Vedda, M. Nikl, N. Solovieva, A. Baraldi, and R. Capelletti, “High-efficiency SiO2:Ce3+ glass scintillators,” Appl. Phys. Lett. 81(23), 4374–4377 (2002). [CrossRef]
11. M. Nikl, K. Nitsch, E. Mihokova, N. Solovieva, J. A. Mares, P. Fabeni, G. P. Pazzi, M. Martini, A. Vedda, and S. Baccaro, “Efficient radioluminescence of the Ce3+ -doped Na–Gd phosphate glasses,” Appl. Phys. Lett. 77(14), 2159–2161 (2000). [CrossRef]
12. D. Jia and W. M. Yen, “Trapping Mechanism Associated with Electron Delocalization and Tunneling of CaAl2O4:Ce3+, A Persistent Phosphor,” J. Electrochem. Soc. 150(3), H61–H65 (2003). [CrossRef]
13. M. Yamaga, Y. Tanii, N. Kodama, T. Takahashi, and M. Honda, “Mechanism of long-lasting phosphorescence process of Ce3+-doped Ca2(a)l2SiO7 melilite crystals,” Phys. Rev. B 65(23), 235108 (2002). [CrossRef]
14. K. W. Kramer, P. Dorenbos, H. U. Gudel, and C. W. E. van Eijk, “Development and characterization of highly efficient new cerium doped rare earth halide scintillator materials,” J. Mater. Chem. 16, 2773–2780 (2006). [CrossRef]
15. K. H. Yang and J. A. DeLuca, “UV fluorescence of cerium doped lutetium and lanthanum trifluorides, potential tunable coherent sources from 2760 to 3220 Å,” Appl. Phys. Lett. 31(9), 594–597 (1977). [CrossRef]
17. G. Q. Xu, Z. X. Zheng, W. M. Tang, and Y. C. Wu, “Spectroscopic properties of Ce3+ doped silica annealed at different temperatures,” J. Lumin. 124(1), 151–156 (2007). [CrossRef]
18. G. Q. Xu, Z. X. Zheng, W. M. Tang, and Y. C. Wu, “Spectroscopic properties of Ce3+ doped silica annealed at different temperatures,” J. Lumin. 124(1), 151–156 (2007). [CrossRef]
19. A. Kucuk and A. G. Clare, “Optical properties of cerium and europium doped fluoroaluminate glasses,” Opt. Mater. 13(3), 279–287 (1999). [CrossRef]
20. T. Murata, M. Sato, H. Yoshida, and K. Morinaga, “Compositional dependence of ultraviolet fluorescence intensity of Ce3+ in silicate, borate, and phosphate glasses,” J. Non-Cryst. Solids 351(4), 312–316 (2005). [CrossRef]
21. G. K. DasMohapatra, “A spectroscopic study of cerium in lithium–alumino–borate glass,” Mater. Lett. 35(1-2), 120–125 (1998). [CrossRef]
22. T. I. Prokhorova and O. M. Ostrogina, “The spectral-luminescence properties of vitrous silicas containing cerium,” Fiz. Khim. Stekla 7, 678–685 (1981).
23. M. C. Nostrand, T. L. Weiland, R. L. Luthi, J. L. Vickers, W. D. Sell, J. A. Stanley, J. Honig, J. Auerbach, R. P. Hackel, and P. J. Wegner, “A large aperture, high energy laser system for optics and optical component testing,” Proc. SPIE 5273, 325–333 (2004). [CrossRef]
24. R. Reisfeld, A. Patra, G. Panczer, and M. Gaft, “Spectroscopic properties of cerium in sol-gel glasses,” Opt. Mater. 13(1), 81–88 (1999). [CrossRef]
25. J. Qiu, N. Sugimoto, Y. Iwabuchi, and K. Hirao, “Photostimulated luminescence in Ce 3+-doped silicate glasses,” J. Non-Cryst. Solids 209(1-2), 200–203 (1997). [CrossRef]