We demonstrate cw output powers >290 mW into a nearly diffraction-limited (M2 ≈2.2) output beam from an interband cascade laser operating at λ = 3.6-3.7 μm at room temperature. The interband cascade laser was designed for nearly equal electron and hole populations in the active region with heavy electron-injector doping, and was processed into narrow ridges mounted epitaxial side down on a copper heat sink. A 15.7-μm-wide, 4-mm-long ridge with the back facet coated for high reflection (HR) and an anti-reflection-coated front facet produced 253 mW of cw output power at T = 25°C into a beam with M2 ≈2.7. Furthermore, corrugating the sidewalls of the ridge leads to a 20% improvement in the brightness. A 15.7-μm-wide, 0.5-mm-long ridge with an HR-coated back facet and an uncoated front facet exhibited a maximum cw wall-plug efficiency of nearly 15% at room temperature.
©2012 Optical Society of America
Interband cascade lasers (ICLs) combine a long-lived interband transition in the active region with cascading to reduce the parasitic voltage drop of these mid-infrared (mid-IR) lasers [1–4]. While both electrons and holes can be generated at a semimetallic interface (SMIF) within each stage, there appears to be a fundamental asymmetry in the properties of practical electron and hole injectors on both sides of the SMIF that causes the holes to greatly outnumber electrons in the active region of conventional ICL designs. The paucity of electrons results in inefficient utilization of carriers in the active region and an increase in the optical loss due to the larger free-carrier absorption cross section for holes. A straightforward fix is to introduce heavy n-type doping of the electron injector (carrier rebalancing), which has produced significant improvements in the threshold current and power densities and the slope efficiencies . The previous results include pulsed threshold current densities as low as Jth = 170 A/cm2 at room temperature (RT), continuous-wave (cw) wallplug efficiencies as high as 13.5%, drive powers as low as 29 mW at RT, and a maximum cw operating temperature of 107°C for a narrow ridge emitting at λ = 3.88 μm. It was further shown that the advantages of the ICL are maintained at least over the entire 3-6 μm wavelength range, with RT cw operation at threshold power densities of Pth ≈1 kW/cm2 at λ = 4.7 and 5.6 μm [5,6].
In this article, we report high cw output powers of several hundred mW from narrow ridges fabricated from ICLs with carrier rebalancing mounted epitaxial-side-down on a copper heat sink. A record high cw operating temperature is also demonstrated. We show that the degradation of beam quality with increasing ridge width is gradual, and that the brightness can be further enhanced by incorporating corrugations into the ridge sidewalls, which act to increase the loss of higher-order optical modes. The measured thermal resistance compares favorably with numerical simulations, indicating that the thermal bond itself accounts for a small fraction of the resistance.
2. Sample design and pulsed characterization
The design of the sample studied in this work is based on the concept of rebalancing the electron and hole densities in the active region [5,6] by heavily doping the middle four wells in the electron injector with a sheet doping density of 4.7x1012 cm−2 . The nominal design and the doping are identical to those found in . Some of the electrons generated by the doping are transferred into the active region, where they combine with the carriers generated at the SMIF to produce an electron-hole density ratio slightly larger than unity. The thickness of the two GaSb separate confinement layers sandwiching the active core was 500 nm, and that of the top and bottom n-InAs/AlSb claddings was 1.4 and 2.8 μm, respectively . Appropriate transition regions and graded doping profiles were introduced in order to reduce the operating bias and the internal loss .
The 5-stage ICL wafer was grown on n-GaSb (100) substrates by molecular beam epitaxy . While the nominal doping levels were the same as in the previous carrier-rebalanced ICLs , minor variations may have occurred because of the limited precision of the calibration procedure. In particular, the low-doped (≈1 x 1016 cm−3) n-GaSb SCLs, which contain 54% of the optical mode intensity, are particularly sensitive to fluctuations of the p-type background doping.
Contact lithography and wet chemical etching were first used to produce 150-μm-wide ridges with intentional lateral corrugation of the sidewalls to eliminate parasitic lasing modes. The uncoated cavities were cleaved to Lc = 2 mm length and tested at a pulse repetition rate of 3 kHz and width of 250 ns. The RT emission wavelength of 3.66 μm was close to the design value.
The RT pulsed threshold current density was 134 A/cm2, and the threshold voltage was 2.35 V, resulting in a threshold power density of 315 W/cm2. The RT slope efficiency was ηs = 338 mW/A per facet, which yields a differential quantum efficiency per stage of ηd = 39.9%. Assuming a facet reflectivity of R = 41%, estimated using the method given in , and an internal efficiency per stage of ηi = 80%, the internal loss can be estimated as αi = ln(1/R)(ηi/ηd-1)/Lc = 4.5 cm−1. Similar internal efficiencies were previously measured on other ICL wafers with low RT thresholds , although cavity-length measurements are not expected to be precise in this case because the internal loss is a function of cavity length. The main physical mechanism for a non-unity internal efficiency is likely to be unavoidable carrier heating in the active region. While the reasons for the low internal loss of the present sample are not known, one potential candidate is a reduction in the n-type doping and free-carrier absorption achieved in the GaSb SCLs. For example, a change in the electron density on the order of 1016 cm−3 would result in a loss variation on the order of 1 cm−1 . Given the different optical loss, the Auger model of the ICL threshold reported in  accounts for approximately two thirds of the 25% difference in the threshold current density with the best-performing sample reported in .
The threshold current and power densities display a super-exponential dependence for temperatures above 300 K, which is not represented if only a single value of T0 is provided. Similarly, the slope efficiency is poorly fitted with a simple an exponential form above RT. In order to accurately represent the important variation of these parameters with temperature for ΔT = T - 300 K > 0, we have fitted the normalized quantities ln(Jth(T)/Jth(300 K)), ln(Pth(T)/Pth(300 K)), and ηs(300 K) – ηs(T) to the same quadratic form: ΔT/T0 + (ΔT/T1)2. The following parameters were extracted over the 300-375 K temperature range: for Jth, T0 = 55 ± 2 K, T1 = 116 ± 5 K; for Pth, T0 = 57 ± 4 K, T1 = 87 ± 5 K; for η, T0 = 498 ± 28 K, T1 = 293 ± 18 K. The sample exhibits a more gradual degradation of the thresholds and efficiencies with temperature than those reported in . The lower loss in the present sample should contribute to that via the lower threshold gains, threshold carrier densities, and Auger rates at all temperatures.
3. Epitaxial-side down fabrication and continuous-wave performance
Narrow ridges were fabricated by photolithography and reactive-ion etching, using a Cl-based inductively coupled plasma (ICP) process followed by cleaning with a phosphoric-acid-based wet etch. The etching was designed to stop just beyond the active core of the device, in the bottom GaSb SCL. Ridge widths between 7.7 and 25.2 μm were fabricated. For some ridges, the sidewalls were nominally straight, while for others, a sidewall corrugation with a peak-to-valley amplitude of 1.4 μm and a period of 2.0 μm was introduced. A 200-nm-thick Si3N4 layer was deposited by plasma-enhanced chemical vapor deposition, and a top contact window was etched back using SF6-based ICP. Next, 100 nm of SiO2 was sputtered to block occasional pinholes in the Si3N4, with the photoresist covering the top contact window being removed afterwards. The ridges were metallized and then electroplated with ≈6 μm of Au. Various cavity lengths were cleaved and mounted epitaxial-side-down on a Cu heat sink C-mount with a 260-μm-thick tungsten copper (W0.9Cu0.1) alloy submount using Au-Sn solder. The heat sink was attached to a thermoelectric cooler. High-reflectivity (HR) coatings comprised of 200 nm Al2O3 topped by 100 nm Au were applied to the back facet of all the devices. In addition, anti-reflection (AR) coatings consisting of a λ/4 layer of Al2O3 were deposited on the front facets of selected devices. The micrographs in Fig. 1 show fabricated ridges with straight and corrugated sidewalls before the epi-down mounting.
Cw emission spectra for the 7.7-μm-wide, 4-mm-long device with HR/AR coated facets and straight sidewalls at a series of temperatures are shown in Fig. 2 . The emission wavelengths ranged from 3.6 μm at room temperature to 3.9 μm at 115 °C near the highest operating temperature. At all temperatures, the emission spectrum is comprised of multiple longitudinal modes, while the analysis of the far-field pattern described below indicates that the emission is almost entirely into a single lateral optical mode.
In order to characterize the quality of the thermal bond for the epi-down mounting, we measured the thermal resistance-area product as obtained from the light-current-voltage characteristics of the devices. The extracted values ranged from 3.3 to 5.2 K/(kW/cm2) for straight-sidewall ridges varying from 7.7 to 15.7 μm in width. These compare with the corresponding range of 5.1-8.6 K/(kW/cm2) for epi-up-mounted ridges of the same dimensions. The thermal resistance of the ridges was also computed numerically using the COMSOL finite-element software. The anisotropic thermal conductivity of the superlattice cladding was a fitting parameter and for simplicity taken to be the same as that of the active region. The variation of the thermal resistance with ridge width was fitted with a vertical conductivity of σ⊥ = 2 W/(m K) and an in-plane conductivity of σ|| = 10 W/(m K), although the results are much more sensitive to the vertical value. The absolute thermal resistances are fitted to <20%, which puts an upper limit on the contribution of the thermal bond to the resistance. The fitted value of σ⊥ is a little lower than the 2.7-3.3 W/(m K) range measured for a somewhat thicker InAs/AlSb SL , and represents by far the largest contribution to the observed thermal resistance.
Figure 3(a) shows the L-I-V characteristics of a 15.7-μm-wide, 4-mm-long, HR/AR laser ridge with straight sidewalls. The differential series resistance-area product measured at the highest current is ≈0.5 mΩ cm2. The threshold current density of 156 A/cm2 and threshold voltage of 2.3 V are comparable to the values observed for wide ridges in pulsed mode, which indicates minimal degradation due to the narrow-ridge processing, epi-down mounting procedure, or lattice heating (in the presence of HR and AR coatings with reflectivities close to 100% and 3%, respectively, the 4-mm-long narrow ridge is expected to have a mirror loss similar to the 2-mm-long uncoated ridge). In order to assess the beam quality of the output light, we employed the “effective M2” approach [11, 12]. In this method, the average absolute value of the far-field angle from the measured far-field profile, shown in Fig. 3(b) is multiplied by the standard deviation of the position from the calculated near-field profile and normalized to the value for the Gaussian mode. The absolute value rather than the standard deviation is used for the angle in order to minimize the impact of noise at large angles. It appears to be quite reliable for narrow ridges with uniform carrier injection, which in the case of ICLs is enabled by significant lateral current spreading in the superlattice-like active region. In this example the extracted M2 varies from 2.2 at a current of 0.4 A to 2.7 at the maximum I = 1.2 A where the output power is 253 mW. Even though several lateral modes appear to be excited, the observed beam quality is sufficiently close to the diffraction limit for a number of high-power applications.
While the beam quality can be improved by employing a narrower ridge, this improvement comes at the expense of output power. Figure 4(a) shows the cw output vs. current for a 7.7-μm-wide, 4-mm-long HR/AR cavity with straight sidewalls at several heat-sink temperatures. At RT, the threshold current density increases slightly to 184 A/cm2, primarily owing to the effect of carrier recombination at the sidewalls, while the maximum output power of 138 mW at the current of 0.6 A is not fully saturated. The M2 parameter computed from the measured far-field characteristics shown in Fig. 4(b) range from 1.1 at 0.2 A to 1.3 at 0.6 A, indicating emission primarily into a single lateral mode. Thus in spite of its lower maximum output power, the narrower ridge actually achieves 13% higher brightness, defined here as the ratio of output to M2, than the wider ridge from Fig. 3. Furthermore, this device operates in cw mode up to a maximum temperature of 118°C, which represents a new record for semiconductor lasers emitting in the 3-4 μm range.
The L-I and far-field characterizations indicate that the 7.7- and 10.8-μm-wide ridges displayed similar brightness values. To investigate the potential for further enhancement of the brightness, we also fabricated and tested ridges with sidewall corrugations. The corrugations should preferentially suppress higher-order optical modes because their scattering losses are greater than for the fundamental mode with peak intensity at the center of the ridge. However, the net effect on brightness depends on whether this mode selection outweighs the lower maximum power associated with a modest increase in the scattering losses for the fundamental mode.
The light-current characteristics for an 18.2-μm-wide ridge with corrugated sidewalls are displayed in Fig. 5(a) for several heat-sink temperatures, while the corresponding far-field profiles at T = 25°C are shown in Fig. 5(b). The beam quality varies from M2 = 1.3 at I = 0.4 A to 1.6 at I = 1.2 A. As expected, while M2 improves considerably by comparison to a similar ridge with straight sidewalls, the maximum power decreases to 205 mW. Figure 6(a) summarizes the L-I characteristics for the series of corrugated-sidewall ridges with widths varying from 13.2 to 25.2 μm, from which we can determine the dependence of brightness on width. A comparison of the far-field profiles for the 18.2-μm-wide ridge in Fig. 5(b) with those for the 25.2-μm-wide ridge in Fig. 6(b) implies that the widest ridge with maximum output power of 291 mW is also brighter, although only by 4% compared to the narrower ridge. The brightness of the 25.2-μm-wide corrugated-sidewall ridge also exceeds the best straight-sidewall result by 20%. The output power, M2, and brightness characteristics of all the ridges in this study are summarized in Table 1 .
While long cavities such as those discussed above produce the highest output powers, we also investigateed the limits on wallplug efficiency (WPE) for the epi-down-mounted ICL. The maximum WPE is obtained for much shorter cavities that optimize the balance between mirror loss and internal loss. Figure 7 displays the cw output power and WPE dependences on injection current for HR/U 0.5-mm-long and HR/AR 1-mm-long cavities with the same width of 15.7 μm . The shorter cavity reaches a maximum cw WPE of nearly 15% at the current of 76 mA where the output power from the uncoated front facet is 36 mW. This is the highest value ever attained for an ICL, and is a general record for the 3-4 μm spectral range. Several higher RT cw values have been reported for QCLs operating at λ = 4.8-4.9 μm [14, 15], with the highest thus far being 21%. An important difference is that typical QCLs reach their maximum WPE at a much higher output powers, owing to their larger number of stages and also their lower internal loss. On the other hand, the threshold power of 47 mW for the 0.5-mm-long ICL cavity is over an order of magnitude lower than the best reported values for QCLs [16–18].
We have shown that epi-down-mounted ICL narrow ridges at room temperature can produce hundreds of mWs of continuous-wave mid-IR power in a nearly-diffraction-limited output beam. Corrugation of the ridge sidewalls leads to a better beam quality for wider ridges, and to a corresponding increase of the brightness by more than 20% despite a slight decrease of the maximum output power. This results from preferential scattering loss for the higher-order lateral optical modes as compared to the fundamental mode. By fitting the ridge dependence of the thermal resistance-area product, we have shown that the thermal bond for epitaxial-side-down mounting contributes at most a small fraction of the observed thermal resistance. The main contribution appears to be the high vertical thermal resistance of the superlattice top cladding layer. Short-cavity ICLs have displayed cw room-temperature wallplug efficiencies close to 15%, which is only ≈6% lower than the record value for quantum cascade lasers. We expect significant further improvements in the high-power operation of ICLs to be realized once the device structure is fully optimized.
References and links
1. R. Q. Yang, “Infrared-laser based on intersubband transitions in quantum-wells,” Superlattices Microstruct. 17(1), 77–83 (1995). [CrossRef]
2. J. R. Meyer, I. Vurgaftman, R. Q. Yang, and L. R. Ram-Mohan, “Type-II and type-I interband cascade lasers,” Electron. Lett. 32(1), 45–46 (1996). [CrossRef]
3. I. Vurgaftman, C. L. Canedy, C. S. Kim, M. Kim, W. W. Bewley, J. R. Lindle, J. Abell, and J. R. Meyer, “Mid-infrared interband cascade lasers operating at ambient temperatures,” New J. Phys. 11(12), 125015 (2009). [CrossRef]
4. I. Vurgaftman, W. W. Bewley, C. L. Canedy, C. S. Kim, M. Kim, J. R. Lindle, C. D. Merritt, J. Abell, and J. R. Meyer, “Mid-IR type-II interband cascade lasers,” IEEE J. Sel. Top. Quantum Electron. 17(5), 1435–1444 (2011). [CrossRef]
5. I. Vurgaftman, W. W. Bewley, C. L. Canedy, C. S. Kim, M. Kim, C. D. Merritt, J. Abell, J. R. Lindle, and J. R. Meyer, “Rebalancing of internally generated carriers for mid-infrared interband cascade lasers with very low power consumption,” Nat. Commun. 2, 585 (2011). [CrossRef] [PubMed]
6. W. W. Bewley, C. L. Canedy, C. S. Kim, M. Kim, C. D. Merritt, J. Abell, I. Vurgaftman, and J. R. Meyer, “Continuous-wave interband cascade lasers operating above room temperature at λ = 4.7-5.6 μm,” Opt. Express 20(3), 3235–3240 (2012). [CrossRef] [PubMed]
7. C. L. Canedy, C. S. Kim, M. Kim, D. C. Larrabee, J. A. Nolde, W. W. Bewley, I. Vurgaftman, and J. R. Meyer, “High-power, narrow-ridge, mid-infrared interband cascade lasers,” J. Vac. Sci. Technol. 26(3), 1160–1162 (2008). [CrossRef]
8. J. Buus, “Analytical approximation for the reflectivity of DH lasers,” IEEE J. Quantum Electron. 17(12), 2256–2257 (1981). [CrossRef]
9. A. Chandola, R. Pino, and P. S. Dutta, “Below bandgap optical absorption in tellurium-doped GaSb,” Semicond. Sci. Technol. 20(8), 886–893 (2005). [CrossRef]
10. T. Borca-Tasciuc, D. W. Song, J. R. Meyer, I. Vurgaftman, M.-J. Yang, B. Z. Nosho, L. J. Whitman, H. Lee, R. U. Martinelli, G. W. Turner, M. J. Manfra, and G. Chen, “Thermal conductivity of AlAs0.07Sb0.93 and Al0.9Ga0.1As0.07Sb0.93 alloys and (AlAs)1/(AlSb)11 digital-alloy superlattices,” J. Appl. Phys. 92, 4994–4998 (2002). [CrossRef]
11. C. S. Kim, M. Kim, W. W. Bewley, J. R. Lindle, C. L. Canedy, J. A. Nolde, D. C. Larrabee, I. Vurgaftman, and J. R. Meyer, “Broad-stripe, single-mode, mid-IR interband cascade laser with photonic-crystal distributed-feedback grating,” Appl. Phys. Lett. 92(7), 071110 (2008). [CrossRef]
12. I. Vurgaftman, W. W. Bewley, R. E. Bartolo, C. L. Felix, M. J. Jurkovic, J. R. Meyer, M. J. Yang, H. Lee, and R. U. Martinelli, “Far-field characteristics of mid-infrared angled-grating distributed-feedback lasers,” J. Appl. Phys. 88(12), 6997–7005 (2000). [CrossRef]
13. W. W. Bewley, C. D. Merritt, C. S. Kim, M. Kim, C. L. Canedy, I. Vurgaftman, J. Abell, and J. R. Meyer, “Mid-IR interband cascade lasers operating with <30 mW of input power,” Proc. SPIE 8374, 83740H, 83740H-9 (2012). [CrossRef]
14. Y. Bai, N. Bandyopadhyay, S. Tsao, E. Selcuk, S. Slivken, and M. Razeghi, “Highly temperature insensitive quantum cascade lasers,” Appl. Phys. Lett. 97(25), 251104 (2010). [CrossRef]
15. Y. Bai, N. Bandyopadhyay, S. Tsao, S. Slivken, and M. Razeghi, “Room-temperature quantum cascade lasers with 27% wall plug efficiency,” Appl. Phys. Lett. 98(18), 181102 (2011). [CrossRef]
16. Y. Bai, S. R. Darvish, N. Bandyopadhyay, S. Slivken, and M. Razeghi, “Optimizing facet coating of quantum cascade lasers for low power consumption,” J. Appl. Phys. 109(5), 053103 (2011). [CrossRef]
17. F. Xie, C. G. Caneau, H. P. LeBlanc, N. J. Visovsky, S. Coleman, L. C. Hughes, Jr., and C.-E. Zah, “Continuous wave operation of distributed feedback quantum cascade lasers with low threshold voltage and small divergent angle for CO2 isotope sensor,” in Proceetings of SPIE Photon. West (San Francisco, 21–26 January 2012).
18. B. Hinkov, A. Bismuto, and R. Terazzi, Y. Bonetti, M. Beck, S. Blaser, and J. Faist, “Junction-up mounted, mid-infrared emitting, continuous-wave DB quantum cascade lasers with very low (< 900 mW) electrical dissipation at room temperature,” Conference on Lasers and Electro-Optics (San Jose, 6–11 May 2012).