Open Access
Add to BibSonomy
Abstract
The development of higher-power InGaN-based diode lasers facilitates their application to optical pumping of Ti:sapphire lasers. Recent diode-pumping results highlight some unexpected behavior, specifically with 450-nm-wavelength devices. To better understand this we have measured and characterized, over a wide range of doping levels, the absorption properties of Ti:sapphire crystals. We find significant changes in the spectral shape of the pumping band in Ti:sapphire with increased doping, and explain the results in terms of absorption due to pairs of Ti3+ ions. Our subsequent discussion attempts to explain prior data, and also provides guidance on optimizing designs for InGaN-diode-pumped Ti:sapphire lasers.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Corrections
18 April 2019: A typographical correction was made to Ref. 13.
1. Introduction
Lasers based on Ti:sapphire (Ti3+:Al2O3) crystals remain among the most widely used devices for scientific and medical-research applications. Demonstration of the first laser operation from the material [1–3] was based on earlier publications showing broadband absorption [4] centered in the blue-green wavelength region (peaking at 490 nm) and associated broadband emission at near-IR wavelengths [5]. Lasing was made possible by optical pumping in the blue-green, 4A2→2E, 3d-3d absorption band, referred to in the following as the pump band.
The development of high-power, InGaN-based semiconductor diode lasers has opened up new options for pumping of Ti:sapphire lasers. Compared to, say, pump sources based on frequency-doubled, solid state lasers or optically pumped semiconductor lasers, InGaN diode lasers can provide greatly reduced complexity, size, and cost for Ti:sapphire systems.
At present, the highest power and highest efficiency, InGaN diode lasers operate in the 450-nm wavelength region. Effects observed with pumps in this region include reduced efficiency compared to longer-wavelength sources (beyond that explained from the larger photon deficit), an increased (and reversible) crystal loss [6–8], and the need to operate with crystals at cryogenic temperature to obtain laser operation [9]. In some work, there was no evident increase in crystal loss [10,11].
An earlier attempt at understanding these phenomena employed Gaussian fits to the pump-band absorption spectrum, which showed qualitatively that the pump-band absorption should effectively vanish at short wavelengths where there was still measurable absorption, and thus blue-near-UV-region absorption was not solely due to that expected from the pump band [12]. In the following, we extend that work to more fully examine absorption in Ti:sapphire for a number of crystals with widely varied doping levels. A more complete description and analysis of this new study has been accepted for publication [13], and includes coverage of other aspects of Ti:sapphire absorption, ranging from 190 to 2500 nm. Here we primarily consider the implications of our work to optical pumping by InGaN lasers, and refer to the more detailed paper at various points for more in-depth discussions of Ti:sapphire spectroscopy.
We find significant changes in the spectral properties of Ti:sapphire absorption in the blue-near-UV region with increasing doping. We note that this was apparent in some early data on absorption spectra as a function of doping level [14], as well in a more recent publication [15]. More importantly, we acknowledge prior work that has perhaps not received proper recognition, showing a similar change in the spectral shape of the near-UV-blue absorption with increased doping, and making an assignment of this feature to a different species than isolated Ti3+ ions. Specifically, the authors speculate the absorption is due to a complex of a pair of Ti3+ ions and a F center (oxygen vacancy with two trapped electrons) [16,17], or a single Ti3+ ion paired with a F2+ center [18], the latter center being paired oxygen vacancies with three trapped electrons. One of our contributions in this paper is to characterize the spectral lineshape for this change in absorption, and show an explicit dependence of the absorption intensity with Ti3+ concentration. Following some recent theoretical models (see [13]), we present a hypothesis that the entity responsible for the absorption change is pairs of Ti3+ ions. We provide guidance on the interaction between doping level and the nature of pump absorption as a function of wavelength, of particular use in design efforts to employ InGaN diodes as Ti:sapphire-laser pump sources.
2. Crystals examined
As detailed in [13], we obtained Ti:sapphire crystals from a number of sources, some of which were older material available at MIT Lincoln Laboratory, and grown there as well as at Crystal Systems and Union Carbide. We also characterized newly grown crystals, one set from Crytur Ld. (Turnov, Czech Republic), which featured high doping levels, with material grown by the Czochralski technique in tungsten crucibles. Other new crystals were grown at Northrop Grumman Synoptics (Charlotte, NC, USA), also by the Czochralski technique, which generally involved melting Al2O3 and Ti2O3 in an iridium crucible under a growth atmosphere of argon.
Table 1 is a list (in the interest of brevity) of just the samples specifically called out in in our data plots, a subset of all the 31 samples we studied [13], showing the labels we employed, the source, sample thickness (along the measurement direction, as determined by a precision micrometer), and the pi (E||c)-polarized absorption strength (cm-1) at 490 nm. This strength provides an indication of the Ti3+ concentration averaged along the sample length. In terms of wt. % Ti2O3, the sample doping levels ranged from 0.04 (SY1b) to 0.29 (CT1a). We discuss below the complications in relating 490-nm absorption to doping level. For labels, the first two letters of the sample names indicate the source, then a number that increases with the measured concentrations. Sample names differing by letters at the end were from the same growth run, in some cases from different portions of the crystal boule.
Table 1. List of samples in data plots
For absorption measurements we employed a Perkin-Elmer (PE), double-beam, Lambda 1050 UV/Vis spectrophotometer. In order to make polarized measurements, we inserted a PE Model B050-5284 Glan-Thompson polarizer in the sample beam. For reduction of the absorption data, we corrected the transmission measurements to account for the Fresnel losses of the samples, employing a three-term Sellmeier equation for the refractive index of Al2O3, valid from 200 to 5000 nm [19].
3. Experimental data on absorption
Figures 1 and 2 present polarized, 300-700-nm absorption data for a variety of samples with different doping levels, both actual values as well as data normalized to the value at 490 nm. The emergence of additional absorption feature peaking around 375 nm is apparent with increasing doping level. The more profound effect on the shape of sigma (E⊥c) spectra indicates the feature does not have the same polarization characteristics as that of the pump band. Sample absorption data at long wavelengths includes the sample-dependent, short-wavelength tail of the 800-nm-peak “IR” band, discussed in detail in [13]. The strength of this band is used in the so-called Figure of Merit (FOM) characterization of commercial materials, which is the ratio of pump-band absorption to peak IR absorption. For this paper we use a FOM definition of the ratio of absorption at 514.5 nm to 820 nm, for pi-polarized light.
Fig. 1. (a) Pi-polarized absorption spectra vs. wavelength from eight samples. (b) Same data as for (a) but normalized to absorption at 490 nm.
Fig. 2. (a) Sigma-polarized absorption spectra vs. wavelength from five samples. (b) Same data as for (a) but normalized to absorption at 490 nm.
4. Data analysis
As a start to analyzing the data shown in Figs. 1 and 2 we seek to distinguish absorption due to the pump band from the absorption that emerges with increasing Ti3+ doping level, which we refer to as “residual absorption.” Our approach is to best-fit the appropriate spectral model for the pump-band lineshape to the measured absorption, as a function of energy in wavenumbers (cm-1), not wavelength, then use that model to subtract the pump-band absorption from the sample data to show the residual absorption. As detailed in [13], we note that the very strong coupling of the electronic level to the lattice supports the use of Gaussian functions to approximate the pump-band lineshape. The Jahn-Teller splitting of the 2E upper level, leads to two, summed Gaussian bands (“high band” peaked near 489 nm and “low band” peaked near 557 nm) providing an excellent representation for double-peaked absorption lineshape of the pump band. The deviation between data and fit is ± 0.2% from 480 to 700 nm for sample UC1a, chosen because of its relatively weak residual absorption and very high FOM, > 700. As expected, the fit starts to deviate at shorter wavelengths.
We employed the fitting parameters (peak positions and linewidths derived from UC1a) to model the pi-polarized data from 16 different samples, and sigma-polarized data for 10 different samples, both from our complete set of crystals. We determined the best fit in terms of minimized difference from the data at pump-band wavelengths > 480 nm and subtracted the fit to generate residual absorption. Our pi-polarized results for 10 representative samples appear in Fig. 3, for both linear and logarithmic plots, the latter to provide a better view of the changing spectral shape from low to high doping levels. We show sigma-polarized results in Fig. 4 for a subset of 7 of the samples shown in Fig. 3. Data from samples not shown had similar spectral characteristics and dependence on doping level.
Fig. 3. (a) Pi-polarized residual absorption vs. wavelength from ten samples. (b) Same data as for (a) but plotted on logarithmic scale.
Fig. 4. (a) Sigma-polarized residual absorption vs. wavelength from seven samples. (b) Same data as for (a) but plotted on logarithmic scale.
The rising UV edge of the residual absorption is the long-wavelength tail of an intense absorption band peaking at 268 nm, discussed in more detail [13] as the “E” band. In that work we estimate the intensity of the tail and show the residual absorption lineshapes with the tail subtracted. We find the short-wavelength limit of residual absorption to be about 300 nm. Our data in Figs. 3 and 4 shows that, while the peak strength of residual absorption is comparable for both polarizations, the spectral shape is notably different, with sigma-polarized absorption extending out to much longer wavelengths. Also, it is apparent that the spectral shape of the residual absorption varies with doping, with the peak value gradually shifting from 440 to 400 nm with increasing doping.
As an attempt to determine how the overall strength of the residual absorption varies with doping level, in Fig. 5 we plot, for 17 samples, the pi-polarized, residual-band absorption at 400 nm as a function of the fitted high-band peak absorption in the sample. The latter represents a good measure of the Ti3+concentration, unencumbered by any residual absorption in the 490-nm-region. The high-band peak absorption is about 98% of the measured 490-nm, pi-polarized absorption for lightly doped samples, with a slight decrease to 97% at high doping. This is in contrast to sigma-polarized data at 490 nm, where, as is evident from Fig. 4b, the residual-band absorption has a larger effect on 490-nm-region absorption, even shifting the peak absorption to shorter wavelengths at high doping levels.
Fig. 5. Plot of the ratio of pi-polarized absorption for the residual absorption at 400 nm as a function of the high-band peak absorption coefficient. Data is for 17 samples, sorted and color coded for high-FOM samples from Synoptics (SY) and for all other samples, with the latter further sorted into low and high FOMs. We include linear fits (forced zero origin) to the high-FOM sets of samples, along with fitting parameters and associated R-squared values.
We distinguish data for three sample sets, annealed Synoptics samples (all but SY1) with a high FOM, other samples with a low FOM (< 30) and other samples with a higher FOM (>30). The high-FOM sets can be well-modeled with a linear dependence of the 400-nm peak values on Ti3+ doping levels, i.e. the residual absorption strength at 400 nm grows as the square of the Ti3+ concentration, with some apparent dependence on material growth and annealing details. We discuss the implications of this below.
5. Discussion and implications
5.1 Ti3+concentrations
Above, we have discussed Ti3+-ion concentrations in terms of measured optical absorption levels. We can connect these to actual ion concentrations based on assumptions about the cross sections for the pump band. We use a cross section, of 9.3 × 10−20 cm2 at 490 nm, determined by Aggarwal et al. [20] through magnetic susceptibility measurements. This implies a concentration (nTi) of 1.1 × 1019 Ti3+ ion/cm3 for a peak high-band absorption coefficient of 1 cm-1. Thus, our samples had concentrations ranging from about 1 × 1018 (SY1) to 9.4 × 1019 (CT1) cm-3. This corresponds, based on an Al3+ ion density (nAl) of 4.7 × 1022 cm-3, to molar Ti3+ percentage doping levels from 0.0021% to 0.2%.
In terms of weight (wt.) % Ti2O3 doping levels, we calculate that this value is 1.41 times the molar percentage level. Commercial literature values [21] that connect peak 490-nm absorption and wt.% Ti2O3 doping appear to employ the 490-nm cross section value (6.5 × 10−20 cm2) from [3], where the calculated absorption for 0.1 wt. % Ti2O3 doping would be 2.16 cm-1. The Aggarwal cross section yields a value of 3.09 cm-1, which we believe to more appropriate. In general, given the literature uncertainty in specifying percentage doping (molar %, wt. % Ti, Ti2O3 and even TiO2) we suggest that specifying material in terms of measured 490-nm, pi-polarized absorption is the least ambiguous method.
5.2 Square-law behavior and pair model
We have been able to benefit in our overall work from having a wide variety of samples with differing Ti3+ doping levels, and to a lesser extent, varying crystal-growth conditions. The most significant result of this, detailed in [13], is a clear indication that a major portion (in some cases essentially all) of the intensity of the residual, E-band and IR-band absorption features in Ti:sapphire is related to the square of the Ti3+ concentration. We find some variations in this, which we attribute to the presence of cation (aluminum) vacancies (VAl3-), which provide charge compensation for Ti4+ ions located at Al3+-ion sites. Such ions are always present in Ti:sapphire, with a concentration level strongly depending on the crystal-growth method and subsequent post-growth annealing. The latter seeks to improve material FOM by reducing processes that convert Ti4+ ions to the 3+ state.
There can be several physical explanations for the square-law behavior, and we choose to assume the simplest, that the square-law features result from pairs of Ti3+ ions, since we expect the concentration of pairs to scale as the square of the concentration of Ti3+ in the crystal. Upon making this assumption, we can estimate the concentration of pairs by assuming a random distribution of Ti3+ ions. Based on this we have calculated the peak cross section for the residual, pi-polarized absorption to be about 6 × 10−18 cm2. This in turn, combined with estimates of the spectral linewidth (6250 and 8325 cm-1 for pi and sigma polarizations, respectively), allows us to estimate the oscillator strength for residual absorption to be about 3 × 10−2, compared to the pump-band oscillator strength of about 3.8 × 10−4. The relatively high value for the residual-band oscillator strength strongly suggests the transition is more complex than between purely 3d-like levels, and may involve the conduction (3s-like) and/or valence (3p-like) bands of the sapphire crystal. Our working hypothesis is that the residual band results from a charge-transfer transition, where an electron from the valence band localizes on a Ti3+-pair site. We also hypothesize that the intensity of the residual band is enhanced in low-FOM samples, where we expect that the relatively high VAl3- concentrations perturb the absorbing species and further enhance the absorption strength.
We do not have an overall theory to explain all of the spectral features of the residual band. The spectral shape changes with increasing peak intensity, and we are not able to model the spectrum as the sum of a limited number of Gaussian-lineshape bands. Given the expected distribution of pair sites with different ion spacings, along with the energy spread of electrons in the valence band, which is small, but non-zero, we would expect the more complex absorption spectral shape we measure, rather than a simple Gaussian band.
5.3 Effect of residual absorption on optical pumping
The practical effect of residual absorption is that it overlaps with the Ti:sapphire laser pump band, and can impact laser operation. As we noted in the Introduction, recent experiments with blue-wavelength pump sources for Ti:sapphire lasers show unexpected low efficiencies as well as an apparent increase in crystal loss.
We can estimate the fraction of pump light that is absorbed directly by the pump bands through calculating, as a function of wavelength, the ratio of the expected pump absorption, as determined by the fitted Gaussian-pump-band models discussed in Section 4, to the measured total absorption. We term this ratio the “pump-band fraction” and plot it in Fig. 6 for pi- and sigma-polarized light as a function of pump wavelength. It falls to zero around 400 nm, the extreme of the pump band, and depends on the doping level in the region where absorption is a mix of that from the pump and residual bands. In addition, since, compared to the pi polarization, sigma-polarized absorption has a larger reduction in strength for the pump band compared to the residual band, we note the generally lower pump-band fraction for sigma-polarized light. At high doping levels this value is less than unity even for wavelengths at the peak of the pump band, since sigma-polarized residual absorption (Fig. 4) extends well into the pump band.
Fig. 6. (a) Pump-band fraction as function of wavelength for pi-polarized light for a number of samples. (b) Same data as for (a) but for sigma-polarized light.
Based on the calculation results in Fig. 6 and knowledge of the absorption properties of the samples, we can develop a relationship between the pi-polarized absorption at the peak (490 nm) and the pump-band fraction, with wavelength as a parameter. Referring back to Fig. 5, we note that the relation between the residual-band absorption strength and 490-nm absorption falls into two general categories, high-FOM samples from Synoptics (i.e. all but SY1) and those from all other sources. At low 490-nm absorption levels the relationship is similar for all high-FOM material but starts to diverge at higher absorption. Accordingly, we have developed two sets of relationships, one for high-FOM Synoptics samples and one for other high-FOM crystals. Figure 7 plots pi-polarized pump fractions for the calculated pump fraction from high-FOM Synoptics samples, while Fig. 8 provides results for the other high-FOM material. Figures 9 and 10 present similar calculations for sigma-polarized light. In all cases we show data points for the samples, with straight lines between them as an aid in extrapolation. Variations in results from samples with very similar doping levels likely indicates the uncertainty in our calculated pump-band fittings.
Fig. 7. Pi-polarized, pump-band fraction as function of sample pi-polarized 490-nm absorption coefficient, with wavelength as a parameter. Data is for high-FOM samples from Synoptics.
Fig. 8. Pi-polarized, pump-band fraction as function of sample pi-polarized 490-nm absorption coefficient, with wavelength as a parameter. Data is for samples other than from Synoptics.
Fig. 9. Sigma-polarized, pump-band fraction as function of sample pi-polarized 490-nm absorption coefficient, with wavelength as a parameter. Data is for high-FOM samples from Synoptics.
Fig. 10. Sigma-polarized pump-band fraction as function of sample pi-polarized 490-nm absorption coefficient, with wavelength as a parameter. Data is for samples other than from Synoptics.
For pi-polarized absorption, it is evident for all samples that pump wavelengths longer than 500 nm are almost totally absorbed by the Ti:sapphire pump band, while at, say, 450 nm, the fraction varies from 83% to 62% from low to high doping levels. The data for sigma-polarized absorption indicates that even at 500 nm, while the pump-band fraction is close to unity at low doping, it drops to 80% for high doping, while the 450-nm fraction ranges from 70 to 28%.
5.4 Analysis of past data on optical pumping
We rely on prior work to consider the effects of a less-than-unity pump fraction on laser operation. Wong et al. [22] excited the E band with a spectrometer-filtered xenon-arc lamp and observed Ti3+, 2E→2T2, 800-nm emission for samples at low temperature, claimed to be due to transfer of energy from a Ti3+-bound exciton to the 2E level. They observed that the 800-nm emission was “quenched strongly above 120 K,” which they attributed to thermally excited, non-radiative decay of the bound exciton. Although our data and analysis [13] shows strong evidence that the E-band is due to Ti3+-pair-related absorption (possibly exciting an exciton bound to the pairs), it is reasonable to assume that pairs, when excited, could undergo similar energy-transfer and non-radiative decay. In any case, at room temperature, one would expect that essentially all of the light absorbed by the E band would be converted directly to heat.
One can question whether the same behavior applies to light absorbed by the residual band. We were not able to find any references to residual-band excitation spectroscopy. In the following, we discuss a series of publications that indirectly support the concept that residual-band excitation leads to non-radiative decay, rather than to excitation of the Ti:sapphire upper laser level.
Photopyroelectric measurements [23] on room-temperature Ti:sapphire crystals, whose absorption spectra showed strong evidence of the residual band, indicate a sharp increase in the non-radiative energy conversion efficiency at wavelengths shorter than 450 nm. The measurement system used employed nominally unpolarized light [24], so further comparison with our models is difficult. We estimate the measured crystals had 490-nm, pi-polarized absorption falling in the 1-2 cm-1 range.
Considering more recent work on the use of blue-wavelength diodes as pump sources for Ti:sapphire lasers, we first discuss publications that showed both a lower-than-expected efficiency, and also an apparent, time-dependent change in Ti:sapphire crystal loss. Roth et al. [6] noted that 452-nm, diode pumping of a cw Ti:sapphire laser produced output powers that were lower than model predictions that otherwise worked well for 532-nm pumping. In addition, the 452-nm-pumped laser power dropped from 60 mW to 19 mW over a period of a few minutes. In experiments with a line-tunable, argon-ion laser as a pump source, the authors noted slight increases in the threshold and decreases in the slope efficiency for wavelengths from 514.5 to 476.5 nm, as expected with the change in wavelength. There was a substantial increase in threshold (about 1.75) and a slope efficiency reduction of 62% when the pump wavelength changed from 488.0 to 457.9 nm.
The 5-mm-long Ti:sapphire crystal in [6] had “0.25%” doping, and the authors assumed that led to a 532-nm fractional absorption of 87% based on pi-polarized absorption coefficients of 4.1 and 5.4 cm-1, for 532 and 490-nm wavelengths, respectively, derived from commercial literature [21]. From our data, this level of absorption is between that of samples SY6a and SY7a. At 457.9 nm, we estimate the absorption would be around 3.4 cm-1, for a crystal transmission of 82% and, from Fig. 7, the pump fraction would be 85%. At 488 nm the fraction increases to >98%. Based on a model for laser threshold [25], we expect the threshold to vary inversely with the product of the fractional absorption of the pump and the pump wavelength. If we assume that the residual absorption does not contribute to pumping of the upper laser level, we then further multiply this product by the pump-band efficiency. Based on this, we calculate a 1.4x increase in threshold from 488- to 457.9-nm pumping, compared to the measured factor of 1.75. The much larger observed increase is consistent, as the authors noted, with the additional effect of higher loss induced by the pumping process.
Other work [8] compared Ti:sapphire laser operation with three different sets of pump diodes operating at 520, 478 and 451 nm, and with three different laser crystals. Here the authors measured the absorption coefficients for their crystals, making it easier to estimate parameters such as the 490-nm absorption coefficient and the pump fraction. The two longer-wavelength devices had about the same beam parameters, while the 451-nm devices had much poorer beam quality, and the authors calculated “mode-matching efficiencies” to correct for the difference. The paper presents extensive data on the laser thresholds and slope efficiencies for all three pump-diode sets, taken for a variety of output-coupler transmissions. We have analyzed this data and calculated the ratios of both thresholds and slope efficiencies, for the same output couplers, between operation at 478 and 520 nm, as well as operation at 451 and 520 nm. We took the averages of these ratios over the set of data for different output couplers. The paper presented thresholds and slopes in terms of absorbed power, so we calculated the theoretical wavelength scaling of the thresholds as the inverse of the product of the wavelength, pump-band efficiency, and the “mode-matching efficiency.” The scaling of the slopes is directly proportional to this product.
Table 2 lists the parameters we used for our calculations on the three crystals (lengths in parentheses) and our calculated ratios (Theory) compared to the average measured ratios for both the threshold and slope. For the latter we show the inverse of the slope ratio, to directly compare with our calculated ratio. The authors noted discrepancies in their analysis in trying to determine consistent crystal losses by two different methods, both based on variations on laser performance resulting from varying the output coupling. One was Findlay-Clay [26], considering threshold changes and the other, Caird [27], considering slope-efficiency changes. We note a similar issue in our results. We would expect the comparison of thresholds and slopes to provide the same ratios, but the actual data showed considerable variance. The best agreement amongst the results was for sample “C”, notably in comparing data and theory with 450- and 520-nm pumps, less so for sample “B” and the worst for sample “A,” which also had the highest doping level. With two main exceptions, we can explain the reduced performance of 450-nm-pumped lasers compared to 520-nm pumped lasers reasonably well by assuming that a fraction of the pump power is lost to absorption by the residual band. We note that this loss can appear to look like a pump-wavelength-dependent increase in cavity loss, in that it leads to a reduction in expected output power, and creates challenges if not included in Findlay-Clay and Caird analyses,
Table 2. Data and analysis for diode-pumped Ti:sapphire lasers [8].
In contrast to [6], the work reported in [8] did not show any change in laser output with time for 451-nm pumping alone but did see a slight reduction in power over time when the laser was pumped by both 520- and 451-nm diodes. When the 451-nm pump was turned off the laser output recovered to the value obtained with just 520-nm pumping. We note other work where the authors did not observe any degradation in diode-pumped Ti:sapphire power with 445-nm pump lasers [10]. Based on measurements [10], the crystal used had about a 3.4 cm-1 absorption coefficient at 490 nm. Other work with around 450-nm pumping [11] described long-term operation of the Ti:sapphire in a multiphoton imaging application, and did not cite any power reduction.
The pump-induced and reversible sample losses are a signature property of transient color centers, and the variations in behavior amongst different crystals indicate that the formation of such color centers are likely related to impurities other than Ti3+ ions. We note that one color center identified in sapphire, the “V” center [28], is attributed to holes trapped at cation vacancies (VAl3-), As we have noted, cation vacancies are expected in Ti:sapphire crystals, with a density depending on the Ti4+ concentration level. The V centers have a high oscillator strength, estimated to be about 0.15, and induce broadband absorption starting at wavelengths shorter than 600 nm. The centers exhibit a spectral dependence in the form of a band with broad peaks at 410 and 260 nm, and a 50-% absorption point around 500 nm on the long-wavelength end. If the residual band is the result of a charge-transfer transition, it would create holes that could migrate to (VAl3-) sites and create V centers. The added loss in the Ti:sapphire pump band would give the appearance of higher crystal losses, in the same way that the residual band does by lowering pumping efficiencies. It is evident from published laser results on output-power recovery that light at wavelengths longer than 450 nm dissociates the color centers. This is also consistent with V-center formation and their subsequent absorption of a fraction of the pump light, which then must lead to moving holes away from the vacancy and back to the valence band. Further sample forensics, particularly measurements of Ti4+ concentration levels, might help to explain the variation in results for different samples.
Another result [9] noted that Ti:sapphire optical pumping by high-power, 450-nm, unpolarized, fiber-coupled diodes produced very weak fluorescence at 300 K, but fluorescence significantly increased as the crystal temperature was lowered. Although room-temperature fluorescence was also weak with pi-polarized, 532-nm pumping at the same 10 W of absorbed pump power, its increase with reducing temperature was much more rapid than for 450-nm pumping. Further analysis is complicated by the difference in both beam properties and sample absorption for the 450- and 532-nm pumps. However, the results are consistent with the expected improvement with reduced temperature in energy transfer from power absorbed by the residual band to excitation of the upper laser level.
To summarize our data analysis considering the impact of residual absorption on optical pumping of the Ti:sapphire laser:
- 1) Models assuming that, at room temperature, light absorbed by the residual band generates heat rather than excitation of the upper-laser level do explain some, if not all of the reduction in efficiency observed with 450-nm pump sources compared to longer-wavelength pumps. They also explain results from photopyroelectric measurements.
- 2) The observation, in some laser work, of pump-induced and reversible loss is consistent with the concept of transient color centers created through absorption of light by the residual band. This provides support for our theory that residual absorption is a charge-transfer process associated with the creation of holes, which can subsequently be trapped at other impurity sites. A possible mechanism is trapping at cation (aluminum) vacancies, and the creation of “V” color centers that induce additional absorption in the Ti:sapphire pumping region and lead to reduced laser efficiencies. The ability of long-wavelength pump light (i.e. free of residual absorption) to anneal away the added loss is also consistent with pump-light absorption due to V-center formation.
- 3) At sufficiently low temperatures, there is evidence that that light absorbed in the residual band can lead to pumping of the upper laser, as the rate for non-radiative decay of excitation into the residual band reduces and the transfer rate of energy to the upper laser level increases.
5.5 Guidelines for InGaN diode pumping of cw Ti:sapphire lasers
The large linewidth of the Ti:sapphire gain medium requires relatively intense pumping to reach threshold, and this has typically been accomplished through the use of longitudinally pumped designs and diffraction-limited lasers as pump sources. Even with high-beam-quality pump lasers, the pump mode sizes needed for sub-W threshold powers are small enough that diffraction of the pump light in the crystal is possible and will lead to a reduction in average intensity and hence increased threshold power. A good design strategy to minimize threshold involves use of heavily doped Ti:sapphire crystals that absorb pump light in a length comparable to or shorter than the pump confocal distance [25].
In common with other semiconductor lasers, presently available, diffraction-limited, InGaN-based devices have limited output powers, around 100 mW, comparable to the lowest observed threshold powers for cw Ti:sapphire lasers [25]. Operation well over threshold requires the use of multimode diode lasers and puts even more emphasis on the use of heavily doped laser material. In terms of the choice of diode wavelength, we plot in Fig. 11 the present status of electrical-optical efficiencies for a variety of devices. The general drop in efficiency with increasing wavelength is the result of unfavorable InGaN material trends due to the increasing level of indium concentration, as detailed in Refs. [29] and [30]. The low efficiencies at greener wavelengths impact maximum diode power output, due to high diode-chip heating. Notably, the highest reported output powers, 1.2 and 5.2 W at 530 and 465 nm, respectively [30], also correspond to the highest efficiencies shown in Fig. 11.
Fig. 11. Data on electrical-optical efficiency of InGaN diode semiconductor lasers as a function of diode wavelength. Data on commercial single-mode (SM) devices from Nichia and Osram and multi-mode (MM) devices from Nichia are from manufacturer’s data sheets. Published reports provided data on Osram MM [29] and Sony MM [30] diodes. Straight lines through the points are only an aid to viewing, not a theoretical model.
Based on the present status of InGaN diodes, despite the fundamental increase in quantum defect, for Ti:sapphire lasers one would be tempted to use 1) high-power 450-465-nm-wavelength diode lasers and 2) the highest laser-crystal doping levels. Past experimental data, and our analysis of the effects of residual absorption, show that 2) leads to increased residual absorption, while 1) leads to a reduced pump fraction and the possible creation of absorbing color centers. A revised strategy is to use diodes with a wavelength of about 490 nm or longer, combined with high doping levels. A less desirable alternative is to reduce the doping level with bluer pump diodes, and use material with a low level of Ti4+ impurities. The latter can be monitored through UV absorption measurements. Another element of the design strategy is to employ only pi-polarized pump light, which rules out the use of polarization-beam-combined diode lasers.
The commercial market for InGaN diodes is driven to a large extent by display applications, hence the emphasis in recent development efforts on devices either in the 450-465 or 520-530-nm wavelength regions. In an ideal case, 490-nm-region diodes would be a good compromise between diode efficiency and power and pumping effectiveness. At this writing, low-power, SM devices are available at 488 nm, and future commercial efforts may include development of MM devices. Figure 11 shows that electrical efficiencies for this wavelength could exceed 20%, with a future 30% possible, and output powers could exceed 3 W. In the absence of any targeted development efforts on this wavelength, use of 520-530-nm devices will be a good practical option, as evidenced by recent work [31].
6. Summary and conclusions
We have observed a significant change in the spectral shape of Ti:sapphire absorption in the blue-UV spectral region with increased doping. Through subtraction of the expected lineshape for 3d-3d-transitions of isolated Ti3+ ions, we determine both the spectral shape and dependence on Ti3+ concentration of a polarization-dependent, “residual absorption” spanning from about 550 to 300 nm. We find the intensity of the residual absorption generally follows the square of the Ti3+ concentration. Our subsequent analysis of the absorption data, with a more detailed version accepted for publication [13], hypothesizes that the square-law-dependent features are due to pairs of Ti3+ ions.
We conclude that a wavelength- and doping-dependent fraction of the light absorbed in what might be considered the optical pumping wavelength region of Ti:sapphire is due to residual absorption. Based on laser data and other measurements, this fraction, at least at room temperature, does not contribute to excitation of the upper laser level but is converted directly into heat. An appropriate strategy on this basis is to use pi-polarized pump wavelengths of 490 nm, or longer.
Funding
Office of the Assistant Secretary for Research and Technology (OST-R) (FA8702-15-D-0001).
Acknowledgements
Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Assistant Secretary of Defense for Research and Engineering. For the MIT Lincoln Laboratory work we acknowledge the skilled absorption measurements by Jonathan Wilson with support from Peter O’Brien, and sample polishing by Patrick Hassett. We benefitted from editing and comments by T.Y. Fan, R. Aggarwal and A. Sanchez. At Crytur, we cite the efforts of Karel Bartos, Jan Polak and Martin Klejch. Finally, we acknowledge work by Adam Lindsey, Chris Oles, Patricia Cajas and Mario Lopez at Synoptics.
References
1. Solid State Research Report (Lincoln Laboratory, MIT, 1982:3), pp. 15–21.
2. “Titanium-doped sapphire: A new tunable solid state laser,” in Physics News in 1982, P. F. Schewe, ed. (American Institute of Physics, 1982).
3. P. F. Moulton, “Spectroscopic and laser characteristics of Ti:Al2O3,” J. Opt. Soc. Am. B 3(1), 125–133 (1986). [CrossRef]
4. D. S. McClure, “Optical spectra of transition-metal ions in corundum,” J. Chem. Phys. 36(10), 2757–2779 (1962). [CrossRef]
5. B. F. Gachter and J. A. Koningstein, “Zero phonon transitions and interacting Jahn-Teller phonon energies from the fluorescence spectrum of α-Al2O3:Ti3+,” J. Chem. Phys. 60(5), 2003–2006 (1974). [CrossRef]
6. P. W. Roth, A. J. Maclean, D. Burns, and A. J. Kemp, “Directly diode-laser-pumped Ti:sapphire laser,” Opt. Lett. 34(21), 3334–3336 (2009). [CrossRef]
7. P. W. Roth, D. Burns, and A. J. Kemp, “Power scaling of a directly diode-laser-pumped Ti:sapphire laser,” Opt. Express 20(18), 20629–20634 (2012). [CrossRef]
8. R. Sawada, H. Tanaka H, N. Sugiyama, and F. Kannari, “Wavelength-multiplexed pumping with 478- and 520-nm indium gallium nitride laser diodes for Ti:sapphire laser,” Appl. Opt. 56(6), 1654–1661 (2017). [CrossRef]
9. S. Backus, M. Kirchner, R. Lemons, D. Schmidt, C. G. Durfee, M. Murnane, and H. Kapteyn, “Direct diode pumped Ti:sapphire ultrafast regenerative amplifier system,” Opt. Express 25(4), 3666–3674 (2017). [CrossRef]
10. C. G. Durfee, T. Storz, J. Garlick, S. Hill, J. A. Squier, M. Kirchner, G. Taft, K. Shea, H. Kapteyn, M. Murnane, and S. Backus, “Direct diode-pumped Kerr-lens mode-locked Ti:sapphire laser,” Opt. Express 20(13), 13677–13683 (2012). [CrossRef]
11. A. Rohrbacher, O. E. Olarte, V. Villamaina, P. Loza-Alvarez, and B. Resan, “Multiphoton imaging with blue-diode-pumped SESAM-modelocked Ti:sapphire oscillator generating 5 nJ 82 fs pulses,” Opt. Express 25(9), 10677–10684 (2017). [CrossRef]
12. A. J. Maclean, P. W. Roth, D. Burns, A. J. Kemp, and P. F. Moulton, “Pump Induced Loss in Directly-Diode Laser Pumped Ti:Sapphire Lasers,” in Lasers, Sources and Related Photonic Devices, OSA Technical Digest Series (CD) (Optical Society of America, 2010), paper AWB16.
13. P. F. Moulton, J.G. Cederburg, K.T. Stevens, G. Foundos, M. Koselja, and J. Preclikova, “Characterization of absorption bands in Ti:sapphire crystals,” Opt. Mater. Express 9(5), 2216–2251 (2019). [CrossRef]
14. R. E. Fahey, A. J. Strauss, A. Sanchez, and R. L. Aggarwal, “Growth of Ti:Al2O3 crystals by a gradient-freeze technique,” in Tunable Solid State Lasers II, A. B. Budgor, L. Esterowitz, and L. G. DeShazer, eds. (Springer, 1987), pp. 82–88.
15. A. Nehari, A. Brenier, G. Panzer, K. Lebbou, J. Godfroy, S. Labor, H. Legal, G. Chériaux, J. P. Chambaret, T. Duffar, and R. Moncorgé, “Ti-Doped Sapphire (Al2O3) Single crystals grown by the Kyropoulos technique and optical characterizations,” Cryst. Growth Des. 11(2), 445–448 (2011). [CrossRef]
16. V. S. Konevskii, E. V. Kryvonosov, L. A. Lytvynov, and M. I. Shakhnovich, “Optical absorption of Ticor,” J. Appl. Spectrosc. 50(4), 427–430 (1989). [CrossRef] We note the authors use of Ticor for Ti:sapphire could lead to the article not appearing in literature searches.
17. E. V. Kryvonosov and L. A. Lytvynov, “Properties of Ti-sapphire as a laser material,” Crystallogr. Rep. 57(7), 967–973 (2012). [CrossRef]
18. S. V. Nizhankovskii, N. S. Sidel’nikova, and V. V. Baranov, “Optical absorption and color centers in large Ti: sapphire crystals grown by horizontally directed crystallization under reducing conditions,” Phys. Solid State 57(4), 781–786 (2015). [CrossRef]
19. I. H. Malitson and M. J. Dodge, “Refractive index and birefringence of synthetic sapphire,” J. Opt. Soc. Am. 62, 1405 (1972). The reference is to a paper abstract, with no numerical data. The latter appears in Handbook of Optical Materials, M. J. Weber, ed. (CRC Press, 2003), Chapter 1, p.75.
20. R. L. Aggarwal, A. Sanchez, R. E. Fahey, and A. J. Strauss, “Magnetic and optical measurements on Ti:A12O3 crystals for laser applications: concentration and absorption cross section of Ti3+ ions,” Appl. Phys. Lett. 48(20), 1345–1347 (1986). [CrossRef]
21. http://eksmaoptics.com/out/media/TiSapphire_Laser_Crystals_Brochure.pdf
22. W.C. Wong, D. S. McClure, S.A. Basun, and M.R. Kokta, “Charge-exchange processes in titanium-doped sapphire crystals. I. Charge-exchange energies and titanium-bound excitons,” Phys. Rev. 51(9), 5682–5692 (1995). [CrossRef]
23. M. Grinberg and A. Mandelis, “Photopyroelectric-quantum-yield spectroscopy and quantum-mechanical photoexcitation-decay kinetics of the Ti3+ ion in Al2O3,” Phys. Rev. 49(18), 12496–12506 (1994). [CrossRef]
24. A. Mandelis, J. Vanniasinkam, S. Budhudu, A. Othonos, and M. Kokta, “Absolute nonradiative energy-conversion-efficiency spectra in Ti:Al2O3 crystals measured by noncontact quadrature photopyroelectric spectroscopy,” Phys. Rev. 48(10), 6808–6821 (1993). [CrossRef]
25. J. Harrison, A. Finch, D. M. Rines, G. A. Rines, and P. F. Moulton, “Low-threshold, cw, all-solid-state Ti:Al2O3 laser,” Opt. Lett. 16(8), 581–583 (1991). [CrossRef]
26. D. Findlay and R. A. Clay, “The measurement of internal losses in 4-level lasers,” Phys. Lett. 20(3), 277–278 (1966). [CrossRef]
27. J. A. Caird, S. A. Payne, P. R. Staver, A. J. Ramponi, L. L. Chase, and W. F. Krupke, “Quantum electronic properties of the Na3Ga2Li3F12:Cr3+ laser,” IEEE J. Quantum Electron. 24(6), 1077–1099 (1988). [CrossRef]
28. B. D. Evans and L. S. Cain, “A cation vacancy center in crystalline Al2O3,” Radiat. Eff. Defects Solids 134(1-4), 329–332 (1995). [CrossRef]
29. A. Avramescu, T. Hager, S. Bernhard, G. Brüderl, T. Wurm, A. Somers, C. Eichler, C. Vierheilig, A. Löffler, J. Ristic, J. Müller, T. Sönke, H. König, and U. Strauß, “High power blue and green laser diodes and their applications,” 2014 IEEE Photonics Conference, San Diego, CA, 2014, pp. 457–458.
30. M. Murayama, Y. Nakayama, K. Yamazaki, Y. Hoshina, H. Watanabe, N. Fuutagawa, H. Kawanishi, T. Uemura, and H. Narui, “Watt-class green (530 nm) and blue (465 nm) laser diodes,” Phys. Status Solidi A 215(10), 1700513 (2018). [CrossRef]
31. K. Gürel, V. J. Wittwer, M. Hoffmann, C. J. Saraceno, S. Hakobyan, B. Resan, A. Rohrbacher, K. Weingarten, S. Schilt, and T. Südmeyer, “Green-diode-pumped femtosecond Ti:Sapphire laser with up to 450 mW average power,” Opt. Express 23(23), 30043–30048 (2015). [CrossRef]
