A sample of Fe:ZnSe fabricated by post-growth thermal diffusion was purchased commercially. The sample was cooled to 80 K using liquid nitrogen and used as the gain element in a Watt-class continuous-wave laser with an output wavelength centered at 4050 nm. The sample was removed from the laser and treated using a hot isostatic press (HIP) technique. The crystal was then re-placed in the laser resonator. After the HIP treatment, and with no other changes to the laser resonator, the slope efficiency of the laser increased by 1.5×. The spectral output was red-shifted to 4122 nm and the output linewidth was narrowed by nearly two orders of magnitude, resulting in a 36× increase in power spectral density. The shift in wavelength and the increase in power scaling performance is consistent with the activation of previously inactive iron impurities in the sample by the HIP treatment.
Many scientific, commercial, and military applications for high power mid-IR laser sources have emerged in recent years. Some such applications depend only on total power, but some sensing techniques require spectrally narrow sources of coherent mid-IR radiation. To the best of our knowledge, there currently are no multi-watt, spectrally narrow (<1 nm), efficient lasers in the 3 – 5 μm region of the electromagnetic spectrum despite a compelling need.
Quantum cascade lasers (QCLs) have become ubiquitous ; however, QCLs are incapable of producing energetic pulses since the lifetime of the carriers in QCL heterostructures is not long enough to store significant amounts of energy to be useful in a Q-switched regime. Sources using nonlinear optical physics, such as optical parametric oscillators (OPO) [2–4] or Raman lasers , can produce energetic pulses of radiation in this waveband. However, the large quantum defect associated with photon splitting gives rise to low laser efficiencies.
Mid-IR lasers based on transition metals doped into II–VI hosts are able to overcome some of the limitations of QCLs and OPOs. Such sources, especially those based on chromium and iron, are rapidly maturing and have been demonstrated over a wide spectral range and in many different pulse regimes [6–9]. In such sources, iron impurities substitute for the group II ion in II–VI semiconductors and assume a 2+ ionization state. The Fe2+ ion resides in a symmetric Coulomb potential which splits the 5D ground level of the ion into an upper 5T2 manifold and a lower 5E manifold. Optical transitions between these two manifolds have been used to create efficient laser devices in Fe:ZnS , Fe:ZnSe [11, 13], and Fe:CdMnTe .
Fe:ZnSe lasers have proliferated in the literature in the past decade with ever increasing average power and pulse energy [8,12,15–20]. When reported, the spectral width of the output of Fe:ZnSe lasers is characteristically 10 to 50 nm. Such large linewidths are typical of inhomogeneously broadened lasers, but theoretical considerations of the crystalline environment of the Fe2+ ions in zinc selenide predict homogeneous broadening .
Recently, along with other colleagues in the Air Force Research Laboratory, we have shown very narrow emission linewidths from both Cr:ZnSe [22–24] and Fe:CdMnTe , which are expected for homogeneously broadened laser gain media. These results show that the broad spectral linewidths previously reported are not an intrinsic characteristic of such lasers. Most significantly, Stites  and Barnes  et al. have shown that hot isostatic press (HIP) treatment can be used to create Cr:ZnSe laser gain media and also to narrow the spectral linewidth of a laser using a commercially produced Cr:ZnSe crystal made by (low pressure) post-growth thermal diffusion (PGTD). In this work, we examine the laser properties of a commercially available Fe:ZnSe sample before and after HIP treatment and report significant increases in efficiency and power spectral density.
2. Sample preparation and spectroscopy
The Fe:ZnSe polycrystal used in this work was fabricated by IPG Photonics using the PGTD technique. The sample measured 2.38 × 5.81 × 7.67 mm3 and the doping concentration was determined to be 7.5×1018 ions/cm3 by the vendor. The sample was anti-reflection (AR) coated from 2.7 – 5 μm on the 2.38 × 5.81 mm2 facets.
Spectroscopic analysis was performed on a similar sample fabricated in the same way. The sample was cooled to 80 K in a closed-cycle helium cryostat and pumped by an Er:YAG laser operating with 200 mW of output power at 2937 nm. The fluorescence was imaged onto the entrance slit of an SP-2500 spectrometer from Princeton Instruments used in combination with a liquid-nitrogen-cooled InSb detector and an SRS830 lock-in amplifier. The fluorescence spectrum was corrected for the system response of the collection apparatus using the blackbody compensation technique. The corrected spectrum was transformed into an emission cross-section σem using the Füchtbauer–Ladenburg equation:21].
The absorption spectrum of the sample was recorded using a Nicolet 6700 Fourier transform interferometer (FTIR). The integral reciprocity equation was used to calculate an approximate absorption cross-section from the absorption spectrum:Fig. 1.
For use in a laser resonator, the commercially purchased Fe:ZnSe sample was wrapped in indium foil and clamped to a copper cold finger in a liquid nitrogen dewar. The dewar was evacuated to approximately 1 mTorr and cooled with liquid nitrogen. The laser resonator was constructed using the familiar X-cavity geometry of Kogelnik  as shown in Fig. 2.
A CorActive continuous-wave Er-fiber laser emitting at 2937 nm was used to pump the Fe:ZnSe laser. The pump beam was measured by the vendor to have an M2 < 1.13 and was collimated using a CaF2 lens. The collimated pump beam was mode-matched to the resonator using lens L1 (f = 150 mm) to focus the beam through M1 and into the center of the Fe:ZnSe sample. Mirrors M1 and M2 were separated by their radii of curvature (100 mm) such the laser mode was collimated in the folded legs of the resonator. Mirrors M1, M2, and M3 were coated as dichroic elements with high-transmission at 2.937 μm and a high-reflection band from 3.7 – 5 μm. Windows W1 and W2 were AR coated from 3000 – 5500 nm with approximately 98% transmission at 4000 nm. The outcoupling mirror OC had a reflectivity of 50% at 4000 nm. Using LASCAD resonator modeling software, the 1/e2 beamwaist of the resonator mode was calculated to be elliptical with ws and wt approximately 90 μm and 120 μm respectively. Likewise, the pump beam was calculated to have ws = 90 μm and wt = 60 μm.
4. Method and results
The Fe:ZnSe laser resonator was aligned and optimized for maximum output power. The input pump power was measured immediately prior to W1, and the output power was measured immediately after the outcoupler M4. From these measurements, the threshold and slope efficiency of the laser were calculated to be 491 mW and 30.4% respectively. The output spectrum of the laser was recorded using a Thorlabs OSA205 mid-IR optical spectrum analyzer (OSA). The laser output was centered at approximately 4050 nm and had spectral content spanning a range of more than 100 nm. The output spectrum confirmed that no pump light was present in the output of the laser.
The sample was then removed from the apparatus and treated using the HIP technique of Stites . The temperature and pressure were linearly ramped from ambient conditions to 1050°C and 30,000 PSI over the span of one hour and held constant for two hours. The temperature and pressure were then linearly ramped back to ambient conditions over the span of one hour. The AR coatings were destroyed by the HIP process, so the sample had to be polished and re-coated.
The sample was then reintroduced to the laser cavity. No changes were made to the resonator other than slight adjustment of the tip and tilt of the outcoupler for optimization. Under identical conditions, the power and spectral measurements were repeated. The threshold and slope efficiency were calculated to be 497 mW and 46.8% respectively. The power scaling results from before and after HIP treatment are shown in Fig. 3. Clearly, the power scaling behavior of the laser is improved significantly by the HIP treatment.
The output spectrum of the laser was again recorded using the OSA. The output wavelength was found to be 4122 nm and the linewidth was measured to be < 1 nm, which is the resolution limit of the instrument. The laser spectra before and after HIP treatment are shown in Fig. 4. Note that the output wavelength of the laser was significantly redshifted after the sample had been HIP treated.
These results represent a significant improvement in the power scaling and spectral characteristics of the laser as a result of HIP treatment of the Fe:ZnSe sample. The threshold of lasing was effectively the same and the slope efficiency of the laser increased by a factor of 1.5×. Notably, the optical efficiency of the laser at maximum power changed from 27% to 41%. All of these improvements are attributable to an increase in the gain and/or a decrease in the passive losses of the Fe:ZnSe crystal.
In addition to these improvements in efficiency, the peak power spectral density also increased by a factor of 36× after the HIP treatment. This improvement suggests that the Fe:ZnSe crystal has been converted from an inhomogeneously broadened medium to a homogeneously broadened medium. Fundamentally, inhomogeneous broadening is due to the formation of different ‘groups’ of emitters in a laser gain medium. The nature of the groups formed in this crystal is not clear, but they represent some non-uniform degree of perturbation to the total Hamiltonian of the Fe2+ ion. Possible candidates for this perturbation include crystal defects as well as second-nearest neighbor effects from clusters of Fe2+ ions. The HIP treatment is hypothesized to remove these perturbations by healing crystal defects and uniformly redistributing the active ions. These changes collapse separate optically active groups into a single group, which then increases the gain of the crystal at the laser wavelength due to an increase in the fraction of population which strongly emits at the laser wavelength.
Admittedly, the redshift of the output is both interesting and somewhat surprising. One hypothesis is that the HIP treatment reduces the arrangement of the constituent ions to its most compact and lowest energy configuration. This reduction in lattice energy then results in a smaller energy splitting between the 5T2 and 5E manifolds. However, if a reduction in the energy of the Fe2+ ions were the only effect of the HIP treatment, we would expect a slight reduction in the output power due to the change in the quantum defect of the laser. Such an effect was not observed. On the contrary, the efficiency of the laser is drastically improved by the HIP treatment. This could be explained by a reduction in the scattering losses within the crystal.
This ‘perturbation’ hypothesis was easily tested using low temperature absorption spectroscopy. Fig. 5 shows the low-temperature spectral absorption coefficient of a sample of Fe:ZnSe produced by PGTD and a second sample produced by HIP. The spectra were taken using a Nicolet 6700, and an ARS closed-cycle helium cryostat was used to cool the samples to 11.5 K. The zero-phonon lines (ZPLs) of both samples reside at the same wavelengths. The spectrum taken from the HIP sample exhibits significant saturation effects on the ZPL corresponding to the Γ5 →γ1 transition at 3654 nm. This effect does not change the conclusion that the overall energy of the laser transition is not affected by HIP treatment.
However, the combination of power scaling improvements and the redshift suggest another hypothesis: not all the iron atoms in the sample were optically active prior to HIP treatment. If some of the iron atoms were not incorporated substitutionally, we would expect that these atoms would not have the proper valence or tetrahedral coordination to give rise the laser levels usually associated with Fe2+ ions in ZnSe and would be optically inactive at the pump and output wavelengths of the laser. We hypothesize that HIP treatment integrates these inactive atoms into the ZnSe lattice substitutionally, leading to an increase in the concentration of Fe2+ ions. The additional ions contribute to an effective increase in the total rate of stimulated emission in the crystal. The additional ions also contribute to increased re-absorption of the spontaneous emission that builds up to lasing in the resonator. The re-absorption only occurs on the blue side of the emission band; so, increasing the rate of re-absorption also introduces a spectrally dependent loss into the resonator which shifts the peak of the round-trip gain to longer wavelengths.
6. Conclusions and future work
In conclusion, we have reported improvement in the lasing characteristics of an Fe:ZnSe laser by HIP treatment of the gain crystal. The slope efficiency of the laser increased by a factor of 1.5× with the same threshold. HIP treatment also converted the Fe:ZnSe crystal from an inhomogeneously broadened laser medium to a homogeneously broadened medium and resulted in a 36× increase in the peak power spectral density.
Unfortunately, the single pass gain and passive losses of the Fe:ZnSe sample were not measured prior to HIP treatment. Future work will repeat this experiment in an amplifier configuration to determine the actual increase in optical gain and decrease in passive loss in the crystal. Alternatively, a full Rigrod analysis  of the laser before and after the sample is HIP treated would also determine the changes in gain and loss due to the HIP treatment. Future work will also include measurements of the lasing threshold and (unsaturated) absorption of the sample before and after HIP treatment.
Low-temperature absorption spectroscopy was used to determine that the spectral locations of the zero-phonon lines of Fe:ZnSe are not affected by HIP treatment. Since they are not, it is clear that the redshift observed in the laser wavelength after the laser crystal was HIP treated is not due to perturbations to the Hamiltonian, but rather to changes in the round-trip spectral gain of the crystal due to the re-absorption of spontaneous emission.
We acknowledge and thank the Sensors Directorate and the Air Force Office of Scientific Research (AFOSR) for funding this effort. This research was performed while Thomas R. Harris held an NRC Research Associateship award at the Sensors Directorate.
We thank Jacob Barnes of UES for performing the HIP treatment of the Fe:ZnSe samples. We also thank Gary Cook, Robert Bedford, David Tomich, and Rita Peterson of the Sensors Directorate, Glen Perram of the Air Force Institute of Technology, as well as Shekhar Guha of the Materials and Manufacturing Directorate and Sean McDaniel of Leidos for technical discussions. Additionally, we thank Pamela Evans for proofreading this manuscript.
References and links
2. L. E. Myers, W. Bosenberg, R. C. Eckardt, M. M. Fejer, and R. L. Byer, “Multigrating quasi-phase-matched optical parametric oscillator in periodically poled LiNbO3,” Opt. Lett. 21, 591–593 (1996). [CrossRef] [PubMed]
3. E. Cheung, S. Palese, H. Injeyan, C. Hoefer, J. Ho, R. Hilyard, H. Komine, J. Berg, and W. Bosenberg, “High power conversion to mid-IR using KTP and ZGP OPOs,” in “Advanced Solid State Lasers,” (Optical Society of America, 1999), p. WC1.
4. R. D. Peterson, D. Bliss, C. Lynch, and D. H. Tomich, “Progress in orientation-patterned GaAs for next-generation nonlinear optical devices,” in “Lasers and Applications in Science and Engineering,” (International Society for Optics and Photonics, 2008), pp. 68750D
5. T. T. Basiev, M. E. Doroshenko, V. V. Osiko, S. E. Sverchkov, and B. I. Galagan, “New mid-IR (1.5 – 2.2 μm) Raman lasers based on barium tungstate and barium nitrate crystals,” Laser Physics Letters 2, 237 (2005). [CrossRef]
7. T. J. Carrig, G. J. Wagner, A. Sennaroglu, J. Y. Jeong, and C. R. Pollock, “Mode-locked Cr2+:ZnSe laser,” Opt. Lett. 25, 168–170 (2000). [CrossRef]
8. J. W. Evans, P. A. Berry, and K. L. Schepler, “A Passively Q-switched, CW-pumped Fe:ZnSe Laser,” Quantum Electronics, IEEE Journal of 50, 204–209 (2014). [CrossRef]
9. S. Vasilyev, I. Moskalev, M. Mirov, V. Smolski, S. Mirov, and V. Gapontsev, “Ultrafast middle-IR lasers and amplifiers based on polycrystalline Cr:ZnS and Cr:ZnSe,” Opt. Mater. Express 7, 2636–2650 (2017). [CrossRef]
10. M. P. Frolov, Y. V. Korostelin, V. I. Kozlovsky, Y. P. Podmarkov, S. A. Savinova, and Y. K. Skasyrsky, “3 J pulsed Fe:ZnS laser tunable from 3.44 to 4.19 μm,” Laser Physics Letters 12, 055001 (2015). [CrossRef]
11. A. A. Voronov, V. I. Kozlovskii, Y. V. Korostelin, A. I. Landman, Y. P. Podmar’kov, Y. K. Skasyrskii, and M. P. Frolov, “A continuous-wave Fe2+:ZnSe laser,” Quantum Electronics 38, 1113 (2008). [CrossRef]
12. V. I. Kozlovsky, Y. V. Korostelin, Y. P. Podmar’kov, Y. K. Skasyrsky, and M. P. Frolov, “Middle Infrared Fe2+:ZnS, Fe2+:ZnSe and Cr2+:CdSe Lasers: New Results,” Journal of Physics: Conference Series 740, 012006 (2016).
13. J. J. Adams, C. Bibeau, R. H. Page, D. M. Krol, L. H. Furu, and S. A. Payne, “4.0 – 4.5-μm lasing of Fe:ZnSe below 180 K, a new mid-infrared laser material,” Opt. Lett. 24, 1720–1722 (1999). [CrossRef]
14. J. W. Evans, B. D. Dolasinski, T. R. Harris, J. W. Cleary, and P. A. Berry, “Demonstration and power scaling of an Fe:CdMnTe laser at 5.2 microns,” Opt. Mater. Express 7, 860–867 (2017). [CrossRef]
15. V. Fedorov, S. Mirov, A. Gallian, D. Badikov, M. Frolov, Y. Korostelin, V. Kozlovsky, A. Landman, Y. Podmar’kov, V. Akimov, and A. Voronov, “3.77 – 5.05 μm tunable solid-state lasers based on Fe2+-doped ZnSe crystals operating at low and room temperatures,” Quantum Electronics, IEEE Journal of 42, 907–917 (2006). [CrossRef]
17. J. W. Evans, P. A. Berry, and K. L. Schepler, “840 mW continuous-wave Fe:ZnSe laser operating at 4140 nm,” Opt. Lett. 37, 5021–5023 (2012). [CrossRef]
18. J. W. Evans, P. A. Berry, and K. L. Schepler, “A broadly tunable continuous-wave Fe:ZnSe laser,” Proceedings of SPIE 8599, 85990 (2013). [CrossRef]
19. H. Jelínková, M. E. Doroshenko, M. Jelínek, D. Vyhlídal, J. Šulc, M. Němec, V. Kubeček, Y. A. Zagoruiko, N. O. Kovalenko, A. S. Gerasimenko, V. M. Puzikov, and V. K. Komar, “Fe:ZnSe laser oscillation under cryogenic and room temperature,” Proc. SPIE 8599, 85990E(2013).
20. D. V. Martyshkin, V. V. Fedorov, M. Mirov, I. Moskalev, S. Vasilyev, and S. B. Mirov, “High Average Power (35 W) Pulsed Fe:ZnSe laser tunable over 3.8 – 4.2 μm,” in “CLEO: Science and Innovations,” (Optical Society of America, 2015), pp. SF1F–2.
21. J. W. Evans, T. R. Harris, B. R. Reddy, K. L. Schepler, and P. A. Berry, “Optical spectroscopy and modeling of Fe2+ ions in zinc selenide,” Journal of Luminescence 188, 541–550 (2017). [CrossRef]
22. R. W. Stites, S. A. McDaniel, J. O. Barnes, D. M. Krein, J. H. Goldsmith, S. Guha, and G. Cook, “Hot isostatic pressing of transition metal ions into chalcogenide laser host crystals,” Opt. Mater. Express 6, 3339–3353 (2016). [CrossRef]
23. S. A. McDaniel, P. A. Berry, K. L. Schepler, J. R. Macdonald, S. J. Beecher, and A. K. Kar, “Gain-switched operation of ultrafast laser inscribed waveguides in Cr:ZnSe,” Proc. SPIE 9342, 93420E (2015).
24. S. A. McDaniel, A. Lancaster, R. W. Stites, F. Thorburn, A. K. Kar, and G. Cook, “Cr:ZnSe guided wave lasers and materials,” Proc. SPIE, Volume 10082, 10082D (2017).
25. J. O. Barnes, R. W. Stites, G. Cook, S. A. McDaniel, D. M. Krein, S. Guha, and J. Goldsmith, “Doping transition metal ions into laser host crystals by hot isostatic pressing (HIP),” Proc. SPIE 10192, 101920A (2017). [CrossRef]
26. H. Kogelnik, E. Ippen, A. Dienes, and C. Shank, “Astigmatically compensated cavities for CW dye lasers,” Quantum Electronics, IEEE Journal of 8, 373–379 (1972). [CrossRef]
27. W. W. Rigrod, “Saturation Effects in High-Gain Lasers,” Journal of Applied Physics 36, 2487–2490 (1965). [CrossRef]