Open Access
Add to BibSonomyAbstract
This paper describes a technique using a hot isostatic pressing (HIP) for the diffusion of transition metal ions into chalcogenide laser host crystals. Thin layers of chromium metal are sputtered onto the surface of zinc selenide and zinc sulfide crystals before treatment in a HIP chamber. The transmissivities, excited state lifetimes, and diffusion rates are measured for various dopant concentrations. Efficiency, spectral output, and tuning data are also measured for a Cr:ZnSe laser. The diffusion rate of 5.48 × 10−8 cm2/s is two orders of magnitude faster than other techniques reported in the literature, and the sub 140 pm measured linewidth is more than 350 times smaller than what is typical of commercially available crystals. Preliminary results for Fe:ZnSe, Co:ZnSe, and Ni:ZnSe are presented as well.
© 2016 Optical Society of America
1. Introduction
Mid infrared lasers, broadly defined as those having wavelengths in the range of 2 to 5 μm, have numerous applications in the scientific and technological community including spectroscopy, metrology, optical communications, photochemistry, biology and medicine, as well as military applications [1]. One important class of mid infrared lasers that has been studied extensively over the last twenty years is transition metal doped chalcogenide lasers [2]. These lasers consist of a divalent transition metal ion (such as Cr2+ or Fe2+) doped into a II–VI chalcogenide host crystal (such as ZnSe, ZnS, CdTe, etc.). One of the reasons these lasers are important is that strong electron-phonon coupling of the transition metal ions leads to significant broadening of the gain bandwidth, allowing for broadband tuning of the laser output (over several hundred nanometers) as well as the potential for ultra short pulse generation. A succinct review of this laser technology can be found in References [3] and [4].
Thus far, several techniques have been explored for the manufacture of these transition metal doped II–VI laser gain crystals. These techniques include self-seeded physical vapor transport [5], Bridgman growth [6], high temperature solution growth [7], hot-pressed ceramics [8], and post growth diffusion of active ions [9–11]. While much effort has been placed on developing post growth diffusion technology (as evidenced by their commercial availability), this technique is not without its drawbacks as the diffusion rate can be quite slow [10] and the spectral output of these lasers are broadband (indicative of inhomogeneous broadening) [12]. While techniques can be used to force single frequency operation of these lasers, such as the inclusion of intracavity diffraction gratings and etalons [13], so far no free running, narrow linewidth, diffusion doped, transition metal II–VI laser has been demonstrated.
In this report, a new technique for post growth diffusion of transition metal doped II–VI materials is presented using sputter deposition and Hot Isostatic Pressing (HIP) [14]. Traditionally, HIP treatment has been used for metallurgical applications such as upgrading metal castings, densifying sintered components, and consolidating powders. However, there do exist optical applications as it has also been shown that HIP treatment can grow grain sizes in undoped polycrystalline ZnSe [15] as well as removing defects in doped Cr:ZnSe crystals [16]. Furthermore, ceramic laser crystals of Nd:YAG have also been manufactured using HIP [17, 18]. Here, it is shown that using a HIP chamber for the forced diffusion of ions (as opposed to a post diffusion HIP treatment) results in a laser gain crystal whose ions diffuse faster than previous studies, and whose free running spectral bandwidth is orders of magnitude smaller than materials manufactured using traditional diffusion methods performed at low pressure under vacuum.
The report is divided up as follows. Section 2 explains the process of manufacturing diffusion doped Cr:ZnSe and Cr:ZnS samples using sputter deposition and HIP treatment. Section 3 describes the optical characterization (absorption, lifetime, diffusion rates, etc.) of the HIP diffused samples. Section 4 describes a laser setup demonstrating the use of these HIP diffused samples as laser gain media with results for slope efficiency, spectral output, and tuning performance. Section 5 previews the preliminary results of expanding this technique to Fe:ZnSe, Co:ZnSe, and Ni:Znse, followed by a concluding Section 6.
2. Crystal Diffusion
Post growth diffusion of chromium ions into II–VI chalcogenide materials (namely ZnSe and ZnS) was accomplished via a two step process of sputter deposition and hot isostatic pressing. For the sputtering phase, polycrystalline CVD grown crystals of ZnSe and ZnS were purchased from a commercial supplier and used as substrates. These samples were 12.5 mm in diameter and 2.0 mm thick with the two round faces polished to an optical quality. Images of the undoped polycrystalline ZnSe and ZnS samples can be seen in Fig. 1. The samples were loaded into a Denton Discovery 18 sputtering system where a thin film of metallic chromium was sputtered onto the the top round surface of each sample. For this process, the chromium target was DC sputtered at a power of 300 Watts. The background pressure inside the sputtering chamber was 3.0 × 10−8 Torr. The chamber was buffered with argon gas at a rate of 27.2 sccm resulting in a pressure of 3.00 mT. For these parameters, the deposition rate was experimentally determined to be 6.3 Å/s.
Fig. 1 Top Row: Photographs of the undoped zinc selenide (left) and zinc sulfide (right) substrates. Bottom Row: Photographs of the chromium doped zinc selenide (left) and zinc sulfide (right) substrates after processing. During the process, 3000 Å of metallic chromium was sputtered onto the top surface before the application of a hot isostatic press treatment at 1050 °C and 30,000 PSI for two hours. After processing, the crystals were polished and the characteristic red and green colors associated with Cr:ZnSe and Cr:ZnS can be seen.
To determine the thickness of the chromium film to be sputtered onto the surface of the samples, a calculation was made factoring in the desired concentration and diffusion depth of the dopant into ZnSe/ZnS substrates. Commercial laser samples of Cr:ZnSe and Cr:ZnS have a typical doping concentration n ≤ 1.00 × 1019 cm−3. For a uniform concentration and a targeted diffusion depth of 250 μm, this corresponds to approximately 6.14 × 1017 chromium ions in the 12.5 mm diameter samples. Therefore, by knowing the density of chromium (ρ = 7.19 g/cm3), and its molar mass of 52.0 g/mol, it was calculated that a thickness of 600 Å should be deposited on the samples. Because it was unclear if 100% of the chromium metal would diffuse into the samples, five different thicknesses (600, 1500, 3000, 4500, and 6000 Å) were deposited onto five different crystals of both ZnSe and ZnS.
Once the samples had been prepared with various thickness of chromium metal sputtered on the surface, they were individually wrapped in molybdenum foil and placed inside the HIP chamber for diffusion. The HIP chamber utilized for these experiments was quite large relative to the crystal size (6 inches in diameter by 14 inches tall), allowing for the bulk processing of numerous crystals in a single experimental cycle. Once in place, the chamber was sealed and the temperature and pressure were gradually raised over the course of two hours to their final values of 1050 °C and 30,000 PSI respectively. Inert argon gas was used as a buffer for the isostatic process. Once the samples reached their final temperature and pressure, they were allowed to “soak” for two hours before temperature reduction and gas ventilation over an additional two hours.
As a side note, it is known that zinc sulfide at standard pressure undergoes a crystal phase transition from the cubic sphalerite (zinc blende) structure to the hexagonal wurtzite structure at a temperature of 1020 °C [19]. While the pressure dependence of this phase transition has not been investigated in depth, the two structures have different bond lengths [20], resulting in a shift in the absorption peak of the Cr2+ ions due to differences in the crystal field. Such a shift was not observed in the absorption measurements described in Section 3 below. While this lack of shift could result from the absence of a crystal phase transition or a transition to wurtzite structure and back, the exact route is inconsequential as Cr2+ ions ultimately end up embedded in the cubic sphalerite crystal structure.
Next, the samples were removed from the HIP chamber and re-polished. During the HIP process, an exchange reaction takes place whereby zinc ions are replaced by chromium in the crystal. Because of this, the HIP treated samples have traces of zinc, chromium, and other chemical compounds thereof that need to be removed prior to optical analysis. For Cr:ZnSe and Cr:ZnS, the visual appearance of the post processed samples visually changed from the undoped substrates as can be seen in Fig. 1 for the 3000 Å deposition. For zinc selenide, the yellow substrate transformed to the characteristic red color of Cr:ZnSe and for zinc sulfide, the clear substrate appears green, corresponding to Cr:ZnS doping.
3. Characterization
While it visually appeared as though the HIP process was successful in post growth diffusion of undoped substrates with Cr2+ ions, the presence of red and green coloration alone is not a sufficiently convincing argument. Therefore, a series of measurements were performed to characterize the samples and verify the presence of the proper ionization state of chromium. The first measurements involved looking at the transmissivity of the samples from 1000 to 2500 nm using a CARY 5000 spectrophotometer and from 2.5 to 25 μm using a Nicolet 6700 FTIR. Previous spectroscopic studies of Cr:ZnSe indicate that the samples should demonstrate a dip in the transmissivity (corresponding to an absorption peak) at ~1780 nm and a broad absorption bandwidth spanning ~350 nm [21]. This absorption corresponds to an electric dipole transition from the 5T2 ground state to the 5E excited state. The normalized transmissivity of our HIP samples of Cr:ZnSe are shown in Fig. 2 for samples having initial sputtered thickness of 1500, 3000, 4500, and 6000 Å. Note, only the range from 1000 to 2500 nm is shown, as no absorption features were noted at longer wavelengths. From the figure, it can be seen that the HIP treated samples do have a characteristic dip in transmission at 1780 nm and a broad absorption bandwidth, as would be expected for Cr2+ doped into zinc selenide. It should be noted that no absorption was measured for the sample having 600 Å of sputtered chrome. Furthermore, as the initial sputtered thickness of chromium increased, the transmission decreased (peak absorption increased). This corresponds to an increase in the chromium concentration that diffused into the samples during the HIP process and indicates a concentration dependence on the initial sputtered chrome thickness.
Fig. 2 The normalized transmissivity of HIP treated Cr:ZnSe as a function of wavelength for various initial sputtered thickness of chromium. The dark blue, green, red, and light blue curves represent sputtered thickness of 1500, 3000, 4500, and 6000 Å respectively. Note the broadband absorption dip centered at 1780 nm corresponding to Cr2+ doped into ZnSe.
For the chromium doped zinc sulfide samples, transmissivity measurements were again taken on the samples from 1000 to 2500 nm using the CARY 5000 spectrophotometer and from 2.5 to 25 μm using a Nicolet 6700 FTIR. Because of the change in crystal lattice parameters for zinc selenide vs. zinc sulfide, it was expected, again from previous spectroscopic studies, that the 5T2 to 5E absorption peak (transmissivity dip) would blue shift to ~1680 nm while maintaining an absorption bandwidth of ~350 nm [21]. The normalized transmissivity of the HIP treated Cr:ZnS samples are shown in Fig. 3 for samples having initial sputtered thickness of 1500, 3000, 4500, and 6000 Å. As before, only the range from 1000 to 2500 nm is shown, as no absorption features were noted at longer wavelengths. From the figure, it can again be seen that the samples have a characteristic dip in transmission at 1680 nm and a corresponding broad absorption bandwidth, as would be expected for Cr2+ doped into zinc sulfide. As with the Cr:ZnSe samples, no absorption was measured for the sample with 600 Å of sputtered chrome. Additionally, a dependence on initial chromium sputtered thickness was observed as the absorption was maximized for the 6000 Å sputtered sample.
Fig. 3 The normalized transmissivity of HIP treated Cr:ZnS as a function of wavelength for various initial sputtered thickness of chromium. The dark blue, green, red, and light blue curves represent sputtered thickness of 1500, 3000, 4500, and 6000 Å respectively. Note the broadband absorption dip centered at 1680 nm corresponding to Cr2+ doped into ZnS.
For further verification that the crystal structure remained in the sphalerite phase after the HIP process, x-ray diffraction was performed on one of the polycrystalline Cr:ZnSe samples. While the data is not reproduced here, the measured diffraction peaks confirmed that the zinc selenide remained in the cubic sphalerite (zinc blende) configuration.
Next, the excited state lifetime of each sample was measured. For this experiment, a Q-switched thulium doped fiber laser was used as a pump source. This laser, manufactured by AdValue Photonics, produced 180 ns pulses with a pulse repetition frequency of 20 kHz. The average power output from the laser was 5 Watts and it operated at a wavelength of 1980 nm. To measure fluorescence from the samples, an extended range InGaAs photodiode was used in conjunction with a long-pass filter having a cut-on wavelength of 2200 nm to block residual pump light. For each sample, 1000 measurements of the fluorescent decay were taken and averaged. A representative curve of the average fluorescent decay for the 6000 Å Cr:ZnSe sample is shown in Fig. 4.
Fig. 4 An averaged measurement of the fluorescent emission of Cr:ZnSe for a 6000 Å initial sputtered thickness of chromium when excited with a resonant 180 ns pulse. The red curve represents an exponential fit of the decay having a lifetime of 4.18 μs.
Once the fluorescence had been measured for the eight samples doped with Cr2+ ions (namely 1500, 3000, 4500, and 6000 Å sputtered ZnSe and ZnS), the data sets were fitted with an exponential decay curve to determine the 1/e lifetime, τ, for each sample. For the representative measurement of the 6000 Å Cr:ZnSe shown in Fig. 4, the lifetime was determined to be 4.18 μs. A summary of the extracted lifetimes of all eight samples is reported in Table 1. From the Table, it can be seen that for increasing concentration (corresponding to greater initial sputtering thickness), τ decreases from 5.44 μs to 4.18 μs for Cr:ZnSe and from 5.00 μs to 3.53 μs for Cr:ZnS. This decrease in lifetime for an increasing concentration is indicative of concentration quenching in the samples.
Table 1. The measured lifetimes τ and average densities of Cr:ZnSe and Cr:ZnS sputtered with various thicknesses of metallic chromium prior to HIP treatment.
In order to determine the average concentration of Cr2+ ions in the ZnSe substrates, an additional sample was prepared by sputtering 5000 Å of chromium onto the 7 mm × 10 mm face of a 7 mm × 10 mm × 5 mm thick ZnSe crystal. The chromium was diffused into the substrate via the HIP process as described above. A determination of the chromium diffusion length into the substrate was performed by first capturing a side image of the sample using a Nikon AZ100 microscope. This captured image is reproduced in Fig. 5. From this image, an optical analysis was performed in MatLab whereby the normalized intensity of the red component of the image, corresponding to the diffusion of Cr2+ ions, was extracted as a function of distance into the crystal. This intensity data is also reported in Fig. 5. It should be noted that due to exposure saturation and losses due to scattering from the chamfered edges of the crystal, the first 80 μm of diffusion data is artificially restricted to unity.
Fig. 5 Top: A cross section image of Cr:ZnSe diffusion via the HIP process. Bottom: The concentration distribution as a function of position within the Cr:ZnSe crystal. This distribution has been fit with a diffusion equation to extract the diffusion length of chromium into the zinc selenide substrate (see text). Note the horizontal scale is the same for both figures.
From this data, a fit using Fick’s law can be made of the form
where C0 is proportional to the surface concentration at the crystal boundary, D is the diffusion rate of the dopant into the substrate, and t is the diffusion time, here taken to be the 7200 second soak time of the HIP process. From the fit, the diffusion rate, D, was found to be 5.48 × 10−8 cm2/s. It is important to note that this value for the diffusion rate is two orders of magnitude larger than previously reported for post growth diffusion under vacuum [10]. This is an important result as diffusion traditionally has been a time intensive process which may take weeks to diffuse several millimeters of chromium into zinc selenide crystals.Knowing the diffusion rate, it now becomes possible to determine the diffusion length, ℓ, of the Cr2+ ions using the relation
For the test sample, the diffusion length is calculation to be ℓ = 397 μm. With this diffusion length, an average concentration, , can be calculated using the Beer-Lambert relation
where T is the measured transmissivity of the sample and σ is the absorption cross section. For Cr:ZnSe at the peak absorption wavelength of 1780 nm, σ is reported to be 1.1 × 10−18 cm2 [21]. Using these values in conjunction with the measured transmissivity at that wavelength, it is now possible to calculate the average concentration for the four Cr:ZnSe samples. These values, ranging from 4.77 × 1017 cm−3 to 6.08 × 1019 cm−3, are also reported in Table 1. Because these concentrations are average values, it should be noted that the peak concentration of each sample is much higher. From this data, it is also important to note that while the effects of concentration quenching on the excited state lifetime are present as the average concentration increases from 4.77 × 1017 cm−3 to 6.08 × 1019 cm−3 (corresponding to a reduction of the excited state lifetime from 5.44 μs to 4.18 μs), the rate of concentration quenching in these samples is lower than has been previously been reported [22], where the lifetime had reduced to less than 1 μs for a concentration of 7.1 × 1019 cm−3.While in principle, the same measurement process could be repeated for Cr:ZnS to determine the average concentrations for the remaining four samples reported in Table 1, it was discovered that the diffusion rate of chromium into ZnS was sufficiently slow that the diffusion length could not be accurately determined for a typical 7200 second soak using this technique. Therefore, quantitative data corresponding to the average concentration of the Cr:ZnS samples is left unreported.
4. Laser Demonstration
Having characterized the optical parameters of the HIP diffused crystals, an attempt was made to utilize the 7 mm × 10 mm × 5 mm thick ZnSe crystal as an active gain element in a laser cavity. For this experiment, a z-cavity configuration was used shown schematically in Fig. 6. This cavity consisted of two curved mirrors (M1) having a focal length of 2.5 cm coated to be anti-reflective for the 1908 nm pump wavelength and highly reflective from 2000 nm to 3000 nm (covering the gain bandwidth of the Cr:ZnSe laser crystal). The mirrors were separated by a distance of 5 cm such that the two legs of the cavity (from M1 to the output coupler, OC, and from M1 to M2) were nominally collimated. To this end, mirror M2 was a flat mirror (infinite radius of curvature) and also coated to be highly reflective from 2000 nm to 3000 nm, akin to mirrors M1. The output coupler was also flat with an infinite radius of curvature and was chosen to have a reflectivity of 80% over the output bandwidth of the Cr:ZnSe laser crystal.
Fig. 6 The experimental layout of the z-cavity resonator. The Cr:ZnSe crystal is placed at Brewster’s angle between two dichroic mirrors, M1, that nominally collimate the beam between M1 and M2 as well as between M1 and the output coupler, OC. The pump laser light is mode matched to the resonator via the lens L1 before being injected into the resonator cavity.
Because the HIP treated Cr:ZnSe did not have an anti-reflection coating, the crystal was inserted into the cavity at Brewster’s angle to minimize losses due to Fresnel reflections. The pump laser was a 40 Watt continuous wave linearly polarized thulium doped fiber laser manufactured by IPG Photonics operating at 1908 nm. A lens, L1, was used to mode match the pump beam to the cavity. Due to the doping and optical path length of the crystal, not all of the pump light was absorbed and exited the cavity through the dichroic mirror M1. By factoring in the amount of pump light that was not absorbed, a slope efficiency measurement of the output power as a function of absorbed pump power could be made for this laser. Fig. 7 shows the result of that measurement. A linear fit was applied to the slope efficiency data for the first five data points above the lasing threshold. From this fit, it was found that the laser operated with an optical-to-optical slope efficiency of 56.2% and had a threshold of 2.88 Watts. As the absorbed pump power was increased beyond 4 Watts, the output power began to deviate from a linear relationship, corresponding to thermal lifetime effects in the laser crystal [2]. In all, a maximum output power of 1.33 Watts was observed corresponding to 6.16 Watts of absorbed pump light.
Fig. 7 The output power of the laser as a function of absorbed pump power. The red line represents a linear fit of the first five data points above threshold. From this fit, a threshold power of 2.88 Watts is calculated with an optical-to-optical slope efficiency of 56.2%. A maximum power of 1.33 Watts was measured corresponding to 6.16 Watts of absorbed pump light before thermal effects begin to dominate.
Upon successful laser demonstration of the HIP diffused Cr:ZnSe crystals, the spectral output of the laser was measured. As mentioned before, previous studies of Cr:ZnSe lasers using commercially available post growth diffusion doped laser crystals results in large spectral outputs (on the order of 50 nm) indicative of inhomogeneous broadening in the laser gain medium [12]. An example of this is shown in Fig. 8 where the normalized broadband spectral output of a Cr:ZnSe laser using commercially purchased crystals (diffusion doped at significantly lower pressure) was measured. From the plot, it can be seen that such lasers have a complex spectral output spanning several tens of nanometers in bandwidth. For single frequency operation of these lasers, additional intracavity elements, such as diffraction gratings and etalons, need to be included [13].
Fig. 8 The broadband spectral output of a laser using commercially available post growth diffusion doped Cr:ZnSe gain media. Note the bandwidth is quite large, spanning tens of nanometers, indicative of inhomogenous broadening within the gain medium.
To measure the spectral output of the HIP diffused samples, the full 1.33 Watts of output power from the z-cavity was attenuated using infrared neutral density filters and coupled into an optical spectrum analyzer (OSA, Thorlabs model OSA205). This analyzer measured the spectral output from 1.0 to 5.6 μm at a resolution of 140 pm. The resulting measured free running spectral output from the laser cavity is shown in Fig. 9. Remarkably, the broadband, inhomogenously broadened output that is typical of commercially available samples had collapsed to a single frequency, narrow bandwidth having a linewidth that was resolution limited by the OSA. This corresponds to a reduction in the spectral output by at least a factor of 350 given the resolution limit of the spectrum analyzer.
Fig. 9 The spectral output of the HIP diffused Cr:ZnSe laser crystal as measured on an OSA. The linewidth of the output was measured to be 140 pm, corresponding to the resolution limit of the detector.
Next, an experiment was performed where the laser was tuned to determine if the linewidth of the free running laser remained narrow across the entire range of the gain bandwidth. For this measurement, the folding mirror, M1, was replaced with a blazed optical grating in Littrow configuration, having 300 grooves/mm and a blaze wavelength of 2.5626 μm. As the grating was rotated, the wavelength was measured using the OSA and the output power of the laser was recorded. Fig. 10 shows the results of this measurement as the laser is tuned from 2300 to 2700 nm in 40 nm steps. The peak heights have been scaled relative to the maximum output power in order to approximate the gain bandwidth envelope of the Cr:ZnSe material. From this figure, it can be seen that not only can the output wavelength of the HIP treated Cr:ZnSe crystals be continuously tuned over a 400 nm range (typical of other Cr:ZnSe lasers), the spectral output remains narrow as well.
Fig. 10 The spectral output of the Cr:ZnSe laser as a function of wavelength when tuned using an intracavity diffraction grating. The power output has been normalized and scaled relative to the peak power at the center of the gain envelope. Note that the spectral narrowing corresponding to a homogeneously broadened gain media remains in tact over the entirety of the tuning range from 2300 to 2700 nm.
To determine the root cause for the narrowing of the output spectrum observed for the HIP diffused Cr:ZnSe laser, an additional HIP run was performed where a commercially purchased, diffusion doped Cr:ZnSe sample was treated. This crystal was subject to the same temperature, pressure, and soak time as the other samples whose performance were characterized above. However, for this experiment, the doped crystal did not have any additional chromium sputtered on its surface. Once the HIP process was complete, the sample was placed at Brewster’s angle into the z-cavity depicted in Fig. 6. Light from the laser was again coupled into an OSA and the resulting spectrum is shown in Fig. 11. Remarkably, after HIP treatment, the spectral bandwidth of the commercial sample had also reduced to the 140 pm resolution limit of the OSA. This is significant as it demonstrates that the spectral narrowing reported here results from the HIP process, most probably because HIP acts to remove defect centers from the crystal [16], eliminating the most likely origin of the inhomogenous broadening noted in Fig. 8.
Fig. 11 The spectral output of the HIP treated commercial Cr:ZnSe laser crystal as measured with an OSA. The linewidth of the output was measured to be 140 pm, corresponding to the resolution limit of the detector. Contrasting this spectra with that of a non-HIP treated sample, the reduction in linewidth most probably results from the removal of defect centers within the crystal via the HIP process.
It should be noted that a similar attempt was made to demonstrate lasing in the Cr:ZnS sample when inserted into the laser cavity at Brewster’s angle. Unfortunately, this attempt was unsuccessful for two reasons. First, as discussed above, the absorption peak of Cr:ZnS is blue shifted by approximately 100 nm relative to Cr:ZnSe, resulting in a decrease in the absorption cross section at the 1908 nm pump wavelength. Second, also noted above, the diffusion rate of Cr2+ ions into ZnS is slower than that of ZnSe, resulting in a significantly smaller gain volume. Future studies beyond the scope of this paper will address these issues by utilizing a different pump laser and increasing the soak time of the HIP process in order to increase the diffusion depth.
5. Additional Transition Metals
While the results presented thus far have focused on chromium doped into zinc selenide and zinc sulfide substrates, there is no aspect of the sputtering and HIP diffusion that is specific to that particular transition metal and/or host. To this end, another series of experiments was conducted whereby the dopant ion choice was expanded to included the following: titanium, vanadium, manganese, iron, cobalt, nickel, copper, zirconium, niobium, molybdenum, ruthenium, rhodium, palladium, silver, tantalum, tungsten, rhenium, iridium, platinum, and gold. In addition, alloys of nickel/chromium, and titanium/tungsten were also investigated. For each of these metals, a 5000 Å thick film was sputtered onto the surface of both undoped zinc selenide and zinc sulfide crystals. The samples were then processed in the HIP chamber for a 2 hour soak, as before, prior to being removed and polished. While the complete results of this experiment will be discussed in a future publication, preliminary data relating to the room temperature transmission of Fe:ZnSe, Co:ZnSe, and Ni:ZnSe are shown in Fig. 12, Fig. 13, and Fig. 14 respectively. For each measurement, the transmission from 0.2 μm to 2.5 μm was performed using the CARY 5000 spectrophotometer while transmission measurements from 2.5 to 25 μm were performed using a Nicolet 6700 FTIR as described above. For the Fe:ZnSe sample, a broad absorption feature corresponding to an excitation from the 5E ground state to the 5T2 excited state can clearly be seen centered at 3.0 μm. Similarly, for Co:ZnSe, three absorption bands from the 4A2(F) ground state to the 4T1(P), 4T1(F), and 4T2(F) excited states can be seen centered at 0.7 μm, 1.6 μm, and 3.0 μm respectively. Finally, for Ni:ZnSe, three absorption bands from the 3T1(F) ground state to the 3T1(P), 3A2(F), and 3T2(F) excited states can be seen centered at 0.75 μm, 1.1 μm, and 2.0 μm respectively. All of these absorption features indicate that this technique of HIP treatment for the post growth diffusion of dopant ions has broad application to laser crystal production.
Fig. 12 The transmission of Fe:ZnSe prepared via sputter deposition and HIP diffusion from 0.2 μm to 7μm. The dip in transmission around 3.0 μm indicates the presence of Fe2+ ions.
Fig. 13 The transmission of Co:ZnSe prepared via sputter deposition and HIP treatment from 0.2 μm to 5μm. The location and relative depths of the multiple absorption features indicates the presence of Co2+ ions.
Fig. 14 The transmission of Ni:ZnSe prepared via sputter deposition and HIP treatment from 0.2 μm to 5μm. The location and relative depths of the multiple absorption features indicates the presence of Ni2+ ions.
6. Conclusion
In this paper, a new technique for the manufacture of transition metal doped chalcogenide laser crystals has been presented. Starting with commercially available undoped zinc selenide and zinc sulfide substrates, various thicknesses of chromium metal were sputtered onto the surface of the crystals. The coated crystals were then placed inside the chamber of a hot isostatic press where the sputtered chromium diffused into the substrates in a non-reactive argon environment at a temperature of 1050 °C and a pressure of 30,000 PSI. After two hours of soak time, the crystals were removed from the chamber and polished. The transmission of the doped zinc selenide and zinc sulfide crystals was measured with a spectrophotometer indicating absorption features corresponding to the presence of Cr2+ ions in the substrate. Furthermore, a correlation between the initial sputtered thickness and the final concentration was established. Measurements of the excited state lifetime were made for each sample and the diffusion depth of chromium into zinc selenide was measured for the two hour HIP treatment. From this information, an average concentration was reported for each sample. With this data it was shown that the rate of concentration quenching in these samples is less than what had previously been reported.
To demonstrate the viability of these samples as laser gain media, a HIP diffused Cr:ZnSe sample was placed at Brewster’s angle inside of a z-cavity resonator with an 80% reflective output coupler and pumped with a 1908 nm thulium fiber laser. Continuous wave lasing was observed in the cavity having a threshold of 2.88 Watts and an optical-to-optical efficiency of 56.2%. The maximum output power measured was 1.33 Watts before thermal quenching effects began to dominate. The spectral output of the free running laser was measured using an OSA and it was found to be resolution limited by the OSA to less than 140 pm. This represents a greater than 350 times reduction in the inhomogeneously broadened linewidth observed in traditionally diffused Cr:ZnSe crystals at lower pressures. Furthermore, it was shown that the laser was tunable from 2.3 to 2.7 μm all the while preserving the narrow spectral output.
In addition, HIP treatment of a traditional diffusion doped sample was performed. The resulting laser demonstrated a reduction in the spectral bandwidth of the output from tens of nanometers to less than the 140 pm resolution limit of the OSA. This clearly indicates HIP treatment dramatically improves the laser characteristics by eliminating the sources of inhomogeneous broadening in the material. Although current commercially produced material can be converted from inhomogenously to homogenously broadened through the HIP process, HIP also provides a dramatic increase in the doping diffusion rate. This makes HIP an ideal one-step process for simultaneously doping and removing underlying defect structure in the crystal substrate.
Future studies of this technique will expand upon the preliminary results presented in Section 5 regarding HIP diffusion of other transition metal doped chalcogenide samples. While promising results for room temperature absorption of Fe:ZnSe, Co:ZnSe, and Ni:ZnSe were presented, much spectroscopic work remains to be done on temperature dependent absorption, fluorescence, lifetime, as well as laser operation of not only these three, but all the transition metal samples manufactured thus far using this technique.
References and links
1. M Ebrahim-Zadeh and I. Sorokina, Mid-infrared Coherent Sources and Applications (Springer, 2005).
2. L. D. DeLoach, R. H. Page, G. D. Wilke, S. A. Payne, and W. F. Krupke, “Transition metal-doped zinc chalcogenides: spectroscopy and laser demonstration of a new class of gain media,” IEEE J. Quantum Elect. 32(6), 885–895 (1996). [CrossRef]
3. T. J. Carrig, “Transition-metal doped chalcogenide lasers,” J. Electron. Mat. 31(7), 759–769 (2002). [CrossRef]
4. S. B. Mirov, V. V. Fedorov, D. V. Martyshkin, I. S. Moskalev, M. S. Mirov, and V. P. Gapontsev, “Progress in mid-IR Cr2+ and Fe2+ doped II–VI materials and lasers,” Opt. Mater. Express 1(5), 898–970 (2011). [CrossRef]
5. C.-H. Su, S. Feth, M. P. Volz, R. Matyi, M. A. George, K. Chattopadhyay, A. Burger, and S. L. Lehoczky, “Vapor growth and characterization of Cr-doped ZnSe crystals,” J. Cryst. Growth. 207, 35–42 (1999). [CrossRef]
6. M. E. Doroshenko, P. Koranda, H. Jelinkova, J. Sulc, M. Nemec, T. T. Basiev, V. K. Komar, A. S. Gerasimenko, and V. M. Puzikov, “Cr:ZnSe prism for broadly tunable mid-infrared laser radiation generation,” Laser Phys. Lett. 4(7), 503–506 (2007). [CrossRef]
7. O. O. Adetunji, N. Roy, Y. Cui, G. Wright, J.-O. Ndap, and A. Burger, “Growth of Cr- and Co- doped CdSe crystals from high-temperature selenium solutions,” J. Electron. Mat. 31(7), 795–798 (2002). [CrossRef]
8. A. Gallian, V. V. Fedorov, S. B. Mirov, V. V. Badikov, S. N. Galkin, E. F. Voronkin, and A. I. Lalayants, “Hot-pressed ceramic Cr2+:ZnSe gain-switched laser,” Opt. Express 14(24), 11694–11701 (2006). [CrossRef] [PubMed]
9. J.-O. Ndap, K. Chattopadhyay, O. O. Adetunji, D. E. Zelmon, and A. Burger, “Thermal diffusion of Cr2+ in bulk ZnSe,” J. Cryst. Growth 240, 176–184 (2002). [CrossRef]
10. U. Demirbas, A. Sennaroglu, and M. Somer, “Synthesis and characterization of diffusion-doped Cr2+:ZnSe and Fe2+:ZnSe,” Opt. Mater. 28, 231–240 (2006). [CrossRef]
11. U. Hommerich, I. K. Jones, E. E. Nyein, and S. B. Trivedi, “Comparison of the optical properties of diffusion-doped polycrystalline Cr:ZnSe and Cr:CdTe windows,” J. Cryst. Growth 287, 450–453 (2006). [CrossRef]
12. P. A. Berry and K. L. Schepler, “High-power, widely-tunable Cr2+:ZnSe master oscillator power amplifier systems,” Opt. Express 18(14), 15062–15072 (2010). [CrossRef] [PubMed]
13. G. J. Wagner, B. G. Tiemann, W. J. Alford, and T. J. Carrig, “Single-frequency Cr:ZnSe laser,” in Advanced Solid State Photonics, (Optical Society of America, 2004), paper WB12.
14. H. V. Atkinson and S. Davies, “Fundamental aspects of hot isostatic pressing: an overview,” Metal. and Mater. Transactions A 31A, 2981–3000 (2000). [CrossRef]
15. H. Namba, H. Osaka, K. Kamon, and F. Higuchi, “Process for preparing ZnSe single crystal,” U.S. Patent 4,584,053 (1986).
16. M. Chen, H. Cui, W. Li, H. Kou, J. Li, Y. Pan, and B. Jiang, “Reparative effect of diffusion process on host defects in Cr2+ doped ZnS/ZnSe,” J. Alloy. Compd. 598, 124–128 (2014). [CrossRef]
17. A. Ikesue and K. Kamata, “Microstructure and optical properties of hot isostatically pressed Nd:YAG ceramics,” J. Am. Ceram. Soc. 79(7), 1927–1933 (1996). [CrossRef]
18. M. Suarez, A. Fernandez, J. L. Menendez, M. Nygren, R. Torrecillas, and Z. Zhao, “Hot isostatic pressing of optically active Nd:YAG powders doped by a colloidal processing route,” J. Eur. Ceram. Soc. 30, 1489–1494 (2010). [CrossRef]
19. B. J. Skinner and P. B. Barton Jr., “The substitution of oxygen for sulfur in wurtzite and sphalerite,” Am. Min. 45, 612–625 (1960).
20. M. Tsuboi, “Optically active lattice vibrations of zinc blende type and wurtzite type crystals,” J. Chem. Phys. 40(5), 1326–1335 (1964). [CrossRef]
21. I. T. Sorokina, “Cr2+-doped II–VI materials for lasers and nonlinear optics,” Opt. Mater. 26, 395–412 (2004). [CrossRef]
22. A. Burger, K. Chattopadhyay, J.-O. Ndap, X. Ma, S. H. Morgan, C. I. Rablau, C.-H. Su, S. Feth, R. H. Pag, K. I. Schaffers, and S. A. Payne, “Preparation conditions of chromium doped ZnSe and their infrared luminescence properties,” J. Cryst. Growth 225, 249–256 (2001). [CrossRef]
