1. Introduction

Transition metal-doped chalcogenides (TM:II-VI) are promising gain media for high-power optically pumped middle infrared (mid-IR) lasers. These lasers are at the forefront of optical research and have applications that include medical surgeries and diagnostics, sensing, industrial process control, material processing, and free-space optical communication. Optically pumped lasers based on Cr:ZnSe operate over the 1.88 to 3.3 µm spectral range and provide up to 140 W of power with up to 65% efficiency in the CW regime [1,2,3]. In mode-locked operation, these lasers can provide 115 GW of peak power with pulses as short as 34 fs [4,5]. Despite these advances in Cr:ZnSe mid-IR lasers, they are currently dependent on robust pump sources and are complex and bulky. In addition to effective mid-IR lasing under optical excitation, these crystals, being wide-band semiconductors, hold the potential for direct electrical excitation [6]. The electron impact excitation mechanism of doped ions under a high electric field enables a simple experimental design. The middle-infrared (mid-IR) electroluminescence of Cr:ZnSe samples under electrical impact excitation was reported by several research groups [6,7]. However, there are no publications where Cr:ZnSe lasing under electrical excitation was reported.

The formation of n-type conductivity (σ) in ZnSe crystal by thermal annealing of the ZnSe samples in Zn-Al alloy has been well documented for many decades. It can provide highly conductive samples with resistivities as low as 0.046 Ω-cm [8]. In our previous publications [9], we found that thermal annealing of Cr:ZnSe crystals in a melt of Zn-Al alloy results in crystals’ purification and significantly decreases chromium concentration in the samples. In this article, we report on optimizing the doping technique for providing n-type conductivity to Al:Cr:ZnSe while preserving the chromium concentration.

2. Sample preparation and experiment setup

The sample fabrication procedures can be summarized as follows. We used a two-stage fabrication process to avoid chromium purification in Cr:ZnSe samples during annealing in the melt of Al-Zn alloy. In the first stage, the conductive Al:ZnSe samples were fabricated by annealing undoped ZnSe samples in a melt of Zn-Al alloy. In the second stage, we doped conductive Al:ZnSe samples with chromium using thermal diffusion from a thin chromium film deposited on the samples’ surfaces.

For the first fabrication stage, non-conductive polycrystalline ZnSe samples that were 10 × 10 × 2.1 mm³ in size were cut from commercial II-VI workpieces grown by II-VI Incorporated using chemical vapor deposition and prepared for Al doping. The preparation protocol included the samples being cleaned with acetone, methanol, and deionized water and sealed in evacuated to 25 mTorr quartz ampoules together with 3 grams of Fisher Scientific reagent grade granular zinc and 15 mg of Fisher Scientific 99.999% purity Al. According to previous doping experiments [8], this 0.5% mass ratio of Al to Zn is optimal for ZnSe conductivity. The ampoules were annealed for five days at 1050 C with the ZnSe samples submerged in the melt of Zn-Al alloy. After annealing, the ampoules were rotated in a vertical plane at 180 degrees to remove the Zn-Al melt from contact with the ZnSe sample. The ampoules were further annealed in Zn-Al vapors for 12 hours before turning off the furnace and cooling off the samples to room temperature. It should be noted that while the samples were prepared under identical conditions, slight differences in ampoule geometry caused variations of Al homogeneity in Al:ZnSe. Due to the low density of the samples, small parts of the samples may not have been completely submerged in the melt resulting in different conductivity levels. Approximately 2 mm was polished off from each sample edge to negate the higher dopant concentration at the edges. The samples’ 8 × 8 mm facets were finely polished, and the transmission spectra were taken with UV-VIS-NIR (Shimadzu UV-3600i) and FTIR (Shimadzu IR Affinity) spectrophotometers. After that, indium contacts were ironed on each sample's 8 × 8 mm facets using a soldering iron. The samples were further annealed for 2 minutes at 250 C to obtain ohmic contacts. The samples then had their resistivity measured using an oscilloscope, pulse generator with variable signal amplitude up to 50 V, and loading resistor connected, as shown in Fig. 1. The voltage signal was applied with a 1.28 ms pulse duration and a repetition rate of 104.2 Hz.

 

Fig. 1. Circuit for measuring sample conductivity.

Download Full Size | PDF

After the data was taken for these Al:ZnSe samples, their contacts were removed using ferric chloride before polishing the samples smoothly. After contact removal, polishing, and cleaning, the final thickness was 1.3-1.5 mm. They were then doped with Cr via post-growth Cr diffusion from the thin metal film. A 110 nm Cr film was deposited on each of the 8 × 8 mm facets of cleaned samples, and the samples were further evacuated, sealed, and annealed at 1050 C for 3 days.

The samples had their transmission spectra and conductivity measured again. The samples were further studied using Electron Paramagnetic Resonance (EPR) spectroscopy to determine the charge state of Cr. The samples were cut in half to preserve part of the sample for future measurements. One-half of each sample was cut and polished to a maximum diameter of 2.8 mm to fit in the EPR spectrometer. The EPR spectra were recorded at room temperature at X-band (9.4 GHz). The valence states of the Cr centers were identified through EPR parameters, such as the principal g-values and hyperfine coupling constants, and the cubic splitting parameters described in many other well-known EPR experiments [10,11,12,13]. Chromium concentration was calculated assuming a uniform distribution of impurities. This study characterized three samples labeled Ia, IIa, and IIIa after annealing in the Zn-Al alloy and Iac, IIac, and IIIac after chromium diffusion, respectively.

3. Results and discussions

Figure 2 shows the I–V characteristics of the n-type Al:ZnSe samples after annealing ZnSe samples in the melt of a Zn-Al alloy. For better figure visibility, the current of sample IIIa was scaled up 10 times. As one can see from the figure, the ohmic contact was formed in samples with the smallest conductivity (IIIa). The other two samples reveal the Schottky barrier effect. We used a linear regression fit on the voltage range larger than the Schottky barrier to estimate the samples’ resistivity. The resistivities (ρ) of Ia, IIa, and IIIa were measured to be 10.8 Ω-cm, 35.2 Ω-cm, and 992 Ω-cm, respectively, accurate to +/- 11%.

 

Fig. 2. I-V characteristics of Al:ZnSe (□- Ia; Δ-IIa; ★- IIIa), the dashed lines show linear fit. Plot IIIa was scaled up 10 times for better visibility.

Download Full Size | PDF

The electrical characteristics of the fabricated samples are summarized in Table 1. The carrier concentration (n_e) was estimated using free electron mobility at 300 K of 200 cm²V^-1s^-1 [14].

Table 1. Physical characteristics of the fabricated sample

View Table

The room temperature transmission of the n-type Al:ZnSe samples polished simultaneously is depicted in Fig. 3(A). The spectra displayed a broad absorption band in the mid-IR region. Figure 3(B) shows the absorption coefficient of the Al:ZnSe samples in the spectral range of the luminescence of Cr²⁺ ions at the ⁵T₂ → ⁵E laser active transition. The dashed line shows the spectral shape of the emission cross-section at this transition. If we consider a typical CW Cr:ZnSe laser with 55% reflectivity of output coupler and 4 mm length gain element, then we could estimate the gain coefficient in the Cr:ZnSe as $={-} \textrm{ln}(R )/2L \approx 0.75\; c{m^{ - 1}}$. As shown in Fig. 3(B), the maximum gain coefficient will be bigger than the passive losses in the crystal Ia with a resistivity of ∼10.8 Ω-cm.

 

Fig. 3. Transmission (A) and absorption (B) spectra of Al:ZnSe samples after annealing ZnSe in Al/Zn melt. Dashed-line shows the spectral shape of Cr²⁺ ions emission cross-section.

Download Full Size | PDF

Figure 4(A) shows the plot of absorption coefficients vs the optical frequency (in wavenumbers) of the annealed samples in the mid-IR spectral range. The spectral dependence is nearly the same as α ∼ν (cm^-1)^-2.8 (shown by the dashed line). It is close to this dependence reported earlier [15] for Al:ZnSe samples. Finally, the dependencies of the absorption coefficients at 2000 nm, 2500 nm, and 3000 nm on samples conductivity are depicted in Fig. 4(B). From this figure, we could suggest that the absorption coefficients are proportional to the carrier concentration at 2500 and 3000 nm. It indicates that carrier absorption is the dominant loss mechanism at these wavelengths. A strong drop of the absorption coefficient at 2000 nm compared with absorption at other wavelengths in crystal Ia with conductivity ∼0.1 Ω^-1-cm^-1 indicates that losses due to carrier absorption are not dominant at 2000 nm in this crystal.

 

Fig. 4. A) Mid-IR absorption spectra of Al:ZnSe samples after annealing ZnSe in Al/Zn melt. The dashed line shows k ∼ν^{- -2.8} dependence: B) the dependencies of the absorption coefficients at 3000 nm, 2500 nm, and 2000 nm on sample conductivity.

Download Full Size | PDF

After optical and electrical characterizations, samples Ia, IIa, and IIIa were subjected to chromium diffusion according to the procedure described above in section 2. The typical chromium concentration used for laser applications is between 2 × 10¹⁸ - 2 × 10¹⁹ cm^-3. Therefore, the target concentration for the fabrication procedure was 1.3 × 10¹⁹ cm^-3. After samples Ia, IIa, and IIIa were doped with Cr, the absorption spectra of the Cr:Al:ZnSe samples (renamed Iac, IIac, and IIIac) were measured over 600-2500 nm spectral range, as depicted in Fig. 5. As shown in the figure, the mid-IR absorption of the free carriers was significantly reduced to the level below the detection limit. Two absorption bands with maxima at 1750 nm and 820-830 nm could be identified in the spectra. The strongest absorption band with a maximum of ∼ 1750 nm is the Cr²⁺ absorption band at the ⁵E → ⁵T₂ laser active transition. The nature of the absorption around 820-830 nm visible only in the Iac crystal will be discussed later. Using its well-documented absorption cross-section [16], the concentrations of Cr²⁺ in the samples were estimated and summarized in Table 1. As shown in the Table, the Cr²⁺ concentrations in the samples were significantly below target concentrations, decreasing in crystals with initially higher conductivity. The resistivities of the IIac and IIIac samples were too high to measure in our experimental setup. Only the Iac sample with initial resistivity of ρ=10.8 Ω-cm shows measurable resistivity with ρ>100 KΩ-cm. One possible explanation for these results is the formation of the Cr⁺ valence state from Cr²⁺, which is a deep electron trap. It should be noted that a typical Cr⁺ concentration in a Cr:ZnSe crystal fabricated by thermal diffusion is less than 1 × 10¹⁴ cm^-3.

 

Fig. 5. Absorption spectra of Al:Cr:ZnSe samples.

Download Full Size | PDF

To determine whether the reduction of Cr absorption in Cr/Al co-doped samples is associated with the change of the Cr valence state, the samples were further studied by EPR spectroscopy. Figure 6 shows the RT EPR spectra of the Iac, IIac, and IIIac samples. The figure shows that all three samples reveal a strong EPR signal from Cr⁺ ions. To estimate Cr⁺ concentrations in the studied samples, we compared the EPR signal from the samples with a calibration standard of the weak pitch with a known number of spins. The number of Cr⁺ ions was determined according to the equation [17]

 

Fig. 6. RT EPR spectra of the Iac, IIac, and IIIac samples. A signal centered around G = 3357 is indicative of Cr⁺.

Download Full Size | PDF

The EPR measurements determined Cr⁺ concentration as shown in Table 1. Of note is that the Cr⁺ concentration is proportional to the sample conductivity. Samples with an initial higher conductivity had a high Cr⁺ and vice versa. This can be explained by charge transfer due to the self-compensation mechanism [17]. The sum of Cr⁺ and Cr²⁺ concentrations is close to the target concentration of chromium within the accuracy of the calibration of the EPR spectrometer (50%). As we mentioned above, the highest concentration of the Cr⁺ valence state was measured in Iac crystal; therefore, the absorption band at 820-830 nm could be assigned to the ⁶A₁→⁴T₁ spin forbidden transition of Cr⁺ ions.

It should be noted that such a high level of Cr⁺ concentration has not been reported so far. Doping Al:ZnSe with Cr in the manner discussed eliminates Cr purification that occurs when doping Cr:ZnSe with Al [9]. The Al doping process is the major cause of Cr purification in Cr:ZnSe crystals because Cr is more soluble in the Al-Zn alloy than in the crystal. When we switched the order to have the Cr doping last, we were able to preserve the initial Cr concentration. In addition, using a similar method for the fabrication of Al:V:ZnSe samples should result in the formation of a large concentration of V⁺ ions. This valence state with a ⁵D [12] ground state shows an energy structure similar to Cr²⁺ with a luminescence band around 2700 nm [18]. The tunable lasers based on this material could cover the important spectral range centered around 3000 nm.

4. Conclusion

Our research aims to study physical processes in the conductive Cr:Al:ZnSe crystals for developing Cr²⁺ mid-IR lasers under impact electrical excitation. The current paper reports on optical, electrical, and EPR spectroscopy of n-type conductive Al:ZnSe and Al:Cr:ZnSe samples. The absorption coefficients in Al:ZnSe crystals associated with free carriers’ absorption at 2.5 µm were measured to be 4.9 cm^-1, 2.9 cm^-1, and 0.7 cm^-1 for samples with resistivities 10.8 Ω-cm (Ia), 35.2 Ω-cm (IIa), and 992 Ω-cm (IIIa), correspondingly. A co-doping of n-type conductive Al:ZnSe samples by chromium ions using thermal diffusion results in a high concentration of chromium in the one plus valence state and a reduced electrical conductivity. The maximum value of Cr⁺ concentration (4 × 10¹⁸cm^-3) was measured for Al:ZnSe crystal with initial resistivity of 10.8 Ω-cm. Comparing the electrical carrier concentration with the concentration of Cr⁺ ions reveals a strong influence of the self-compensation mechanism on the chromium valence state in studied samples. The absorption band 820-830 nm in the crystal with a high concentration of Cr⁺ ions was assigned to the ⁶A₁→⁴T₁ spin forbidden transition of Cr⁺ ions. The electrical excitation of Cr²⁺ ions via impact excitation mechanisms and the impact ionization of Cr⁺ ions will be the subject of future research. In addition, the Al:V:ZnSe co-doped crystals were proposed as promising media for mid-IR lasers operating around 3 µm spectral range.