Open Access
Add to BibSonomy
Abstract
SiO2 is a commonly used insulation layer for QCLs but has high absorption peak around 8 to 10 µm. Instead of SiO2, we used Y2O3 as an insulation layer for DC-QCL and successfully demonstrated lasing operation at the wavelength around 8.1 µm. We also showed 2D numerical analysis on the absorption coefficient of our DC-QCL structure with various parameters such as insulating materials, waveguide width, and mesa angle.
© 2017 Optical Society of America
1. Introduction
Quantum cascade lasers (QCLs) are capable of emitting wide-wavelength range of light from a few micro-meters to the terahertz region [1,2]. As the emission process is based on the intersubband transition, QCLs can control the emission wavelength by modifying the thickness of active layers. After the first QCL emission was reported in 1994, the performance of QCLs have been rapidly improved including room-temperature (RT) continuous-wave operation, watt-level output power or high wall-plug efficiency [3–5].
For the high performance operation, QCL devices are usually fabricated with buried hetero (BH) structure by Fe doped InP layers grown lateral to the active region [6,7]. The BH structure offers advantages in an efficient thermal management allowing QCLs to be operated in a CW current injection. Another approach of the QCL structure is a double channel (DC) structure. The DC-QCL is made by etching both side of the mesa and then, the mesa is covered by contact metals with an insulation layer. The fabrication process of the DC-QCL is cost effective since it does not require any regrowth process. CW operation was also demonstrated with thick Au electroplating for heat dissipation [8,9].
SiO2 is the most common material as an insulator for QCLs. It has a good insulation property along with a low extinction coefficient in the near infrared (IR) to mid IR region. In the wavelength between 8 to 10 µm however, SiO2 shows very high absorption coefficient which will degrade the performance of QCLs. Si3N4 can also be used at the wavelength above 8 µm though, it still has some losses. As an alternative to SiO2 and Si3N4, TiO2, As2S3 or GeSe were proposed to be used as the insulation layer and some results showed better performance in a theoretical study [10,11]. Y2O3 could be an another option as an insulator and it is already widely used material for HR or AR coatings in the QCL fabrication process [12]. The material absorption of the Y2O3 is very low up to 11 µm and has an excellent adhesion property with semiconductors as well as metals [13].
In this paper, we numerically compared the performance of Y2O3 insulated DC-QCLs and that of SiO2 or Si3N4. And we also demonstrated a DC-QCL with the Y2O3 insulation layer for the first time at the wavelength around 8.1.
2. Experimental result
In order to confirm the absorption characteristics of SiO2 Si3N4 and Y2O3 layer, we measured fourier transform infrared (FTIR) transmission of the two thin films. 100-nm-thick SiO2, Si3N4 and Y2O3 films were deposited on a semi-insulated (SI) InP substrates by plasma enhanced chemical vapor deposition (PECVD) (SiO2 and Si3N4) and electron beam (EB) evaporation (Y2O3), respectively.
Figure 1 shows the measured FTIR transmission spectra of SiO2, Si3N4 and Y2O3 films. Orange short dashed line is the Si3N4 film. Green alternate long and short dashed line is the SiO2 film and red solid line is the Y2O3 film. Black long dashed line is transmission spectra of a bulk SI-InP substrate as a reference data. Higher transmittance than the reference data is indicating an antireflection effect by the thin film and lower transmittance means absorbance by the thin film. The SiO2 film shows strong absorption in the wavelength between 8 to 10 µm. At the same wavelength range, Si3N4 also shows absorption though it is slightly smaller than SiO2. In a contrary, Y2O3 film showed low material absorption up to 11 µm except two absorption peaks at the wavelength around 2.9 and 7 µm which are water vibrational frequencies [14]. As described above for better absorption characteristics, Y2O3 could be an alternative to SiO2 or Si3N4 at the wavelength between 8 to 11 µm.
As a next step, we fabricated the DC-QCL with the Y2O3 insulation layer as illustrated in Fig. 2. For the active layers, we used similar layer design reported in [14]. A highly doped InP substrate was used for the QCL growth and therefore, a low-doped InP buffer layer was added as a low cladding layer. Epitaxial growth was performed by the molecular beam epitaxy method. The laser mesa was formed by Cl2 based inductively coupled plasma reactive ion etching to a depth of 6.8 µm. We used photoresist as a mask for the etching and we could obtained an angled mesa profile. After etching of the laser mesa, a 250-nm-thick Y2O3 film was deposited on top of the laser mesa by the electron beam (EB) evaporation for the insulation. As this is the first attempt to fabricate the Y2O3 based QCL, structural parameters such as thickness of the Y2O3 film need is not fully optimized yet.
Fig. 2 DC-QCL structure for fabrication and simulation model. (The layer sequence of one active cell starting from the injector barrier, given in nanometers, 4.3/1.7/0.9/5.4/1.1/5.3/1.2/4.7/2.2/4.3/1.5/3.8/1.6/3.4/1.8/3.0/2.1/2.8/2.5/2.7/3.2/2.7/3.6/2.5, InAlAs barriers are in bold face, InGaAs wells are shown in normal face, and underlined correspond to the n-doped layers (Si, 1.5 × 1017 cm−3) (Ref [14].).
In order to open the window for a metal contact, the Y2O3 film on the laser mesa was wet chemically etched by a HCl based solution as shown in Fig. 3(a). Some residues were remained on the surface of the InP after Y2O3 film removal. The residues were hard to remove in the HCl solution or other wet solutions so we used Ar plasma to clean the surface. Figures 3(b)–3(e) are scanning electron microscope (SEM) images of the surface after Y2O3 removal and Ar plasma treatment. When we used a lower plasma power, residues were not clearly removed or required longer time to be removed and when we used a higher plasma power, the surface was noticeably damaged by Ar radicals especially at the edges of the removed Y2O3 film. From the SEM images, we chose to use 50W plasma power for 5 minutes as a surface cleaning condition.
Fig. 3 (a) SEM image of Y2O3 film after wet etching by HCl solution, surface residual removal by Ar plasma after wet etching of Y2O3 (b) 30W, 10 min (c) 50W, 10 min (d) 70W, 5min (e) 100W, 5 min.
For contact formation, Ti/Au (25 nm/300 nm) was EB evaporated as a top contact and the same thickness of a bottom contact was also EB evaporated after thinning the substrate. Cross-sectional SEM images of a fabricated QCL is shown in Fig. 4. From the close-up images, InGaAs confinement layers are visible which are surrounding the core layer and the Y2O3 insulation layer is also visible between epitaxial layers and metals. After the fabrication process, the device was then cleaved by a length of 3 mm.
Fig. 4 Cross-sectional SEM image of the fabricated QCL (left) and a close-up of the sidewall (right).
Pulsed L-I curves (pulse width: 200ns, repetition rate: 20 kHz) of the fabricated QCL device as a function of the mount temperature is shown in Fig. 5(a). The laser width was 15 µm and both facets were uncoated. The threshold current was 0.93 A at the temperature of 10°C (threshold current density: 2.1 kA/cm2) and increased to 1.12 A at the temperature of 60°C (2.5 kA/cm2). Over 60 mW of a peak output power was achieved when the applied current was 1.5 A at the temperature of 20°C. Measured lasing spectra of the same QCL device as a function of the mount temperature is in Fig. 5(b). The peak wavelength was varied from 8.07 µm to 8.10 µm according to the temperature ranging from 10°C to 60°C which is correspond to 0.6 nm/K. The QCL device was only operated in the pulse mode though, CW operation would be available with more effective heat dissipation technics such as Au electroplating or AuSn epi-side down mounting on an AlN substrate. Nevertheless, we successfully demonstrated Y2O3 insulated QCL at the operating temperature up to 60°C.
3. Numerical analysis of DC-QCL: absorption coefficient
Absorption coefficient of QCLs can be estimated by the 2-dimensional finite element method (FEM) simulation of the actual QCL structure [15]. Overall layer design is as mentioned in chapter 2 (see Fig. 2). As a simulation parameter, the IR absorption of InP and InGaAs was simply calculated by drude model [16]. The whole core region was simplified as one material (n: 3.3, k: 0) and only Au was considered in the simulation as a contact metal (titanium was ignored in the simulation). Refractive indices of SiO2 (n: 0.48405, k:0.42273), Si3N4 (n: 1.986, k: 0.24464) and Y2O3 (n: 1.7, k: 0.006) were referred from [17] and [18].
Wavelength was set to 8.1 µm and only the fundamental mode was considered in the simulation. The minimum waveguide width was limited to 13 µm for the simulated waveguide structure since the InP contact layer became too narrow to be fabricated in our etching condition. The calculated absorption coefficient depend on the waveguide width is shown in Fig. 6. Si3N4 insulated QCLs showed the highest absorption coefficient than that of SiO2 or Y2O3. Although the material absorption of the Si3N4 film was lower than the SiO2 film, the high refractive index of the Si3N4 film caused more mode overlap on the insulation layer and the contact metal. SiO2 insulated QCLs showed higher absorption coefficient than that of Y2O3 for any waveguide width. The absorption coefficient of the SiO2 (Y2O3) insulated QCL was 12.1 cm−1 (8.7 cm−1) at 13 µm and 8.1 cm−1 (7.8 cm−1) at 25 µm width. The difference was even increased with narrower waveguide-width which is indicating larger optical mode overlap in the insulator region. Narrow waveguide-width is required for CW operation in terms of a heat generation.
Fig. 6 Electric field distribution of QCL (a) and waveguide width dependence of absorption coefficient (b). (open blue circles: Si3N4 insulated QCLs, open green circles: SiO2 insulated QCLs, filled red circles: Y2O3 insulated QCLs, open green triangles: SiO2 insulated QCLs with low doped substrate, filled red triangles: Y2O3 insulated QCLs with low doped substrate.)
Even though the QCLs were simulated with Y2O3 insulation, relatively higher absorption coefficient was obtained. It was affected by the highly doped (2 × 1018 cm−3) substrate and the buffer layer was too thin to suppress the mode overlap into the substrate. Thicker buffer layer more than 2 µm or an InP substrate with lower doping level 1 × 1017 cm−3 could improve the absorption coefficient (see filled circles in Fig. 6(b)). The absorption coefficient was as low as 1.9 cm−1 for 25-µm-width waveguide in the case of Y2O3 insulation layer and only 0.4 cm−1 was increased when the waveguide width was 13 µm. But still, the absorption coefficient was 5.3 cm−1 for the SiO2 insulated QCL with 13 µm waveguide width. More than twice the absorption coefficient was added due to the SiO2 layer instead of Y2O3.
For the further investigation, we simulated the absorption coefficient with a higher mesa angle as illustrated in Fig. 7(a). We assumed the higher mesa angle will promote the mode overlap on the active region so that the absorption coefficient could be further improved. Narrower waveguide width is also applicable which was limited in the previous structure. Figure 7(b) shows the waveguide width dependence of the absorption coefficient of QCLs with the higher mesa angle. For the comparison, the absorption coefficients of QCLs with the previous mesa angle as in Fig. 6(b) are also plotted in Fig. 7(b). QCLs with higher mesa angle showed better absorption coefficients compared to that of the lower mesa angle. The absorption coefficient of 2.3 cm−1 (lower mesa angle) was improved to 2.1 cm−1 (higher mesa angle) at the waveguide width of 13 µm and it was 2.4 cm−1 at the waveguide width of 10 µm. Additional losses such as scattering losses from the sidewall should be considered though, the DC-QCLs with Y2O3 insulation layer showed promising potential at the spectral range of 8 to 10 µm.
Fig. 7 Electric field distribution of QCL with higher mesa angle (a) and waveguide width versus absorption coefficient (b). (filled green diamonds: SiO2 insulated QCLs with higher mesa angle and low doped substrate, filled red circles: Y2O3 insulated QCLs with low doped substrate (data in Fig. 6(b)), filled red diamonds: Y2O3 insulated QCLs with higher mesa angle and low doped substrate.)
4. Conclusion
The absorption characteristics of SiO2, Si3N4 and Y2O3 thin films were measured by FTIR transmission spectroscopy. Y2O3 thin film was deposited by EB evaporation. SiO2 and Si3N4 thin film was deposited using PECVD. The SiO2 and Si3N4 thin film showed extremely high absorbance in the spectral range from 8 to 10 µm while the Y2O3 thin film showed negligible material absorption up to 11 µm. By using the Y2O3 film as an insulation layer, we successfully demonstrated DC-QCL at the wavelength around 8.1 µm. Over 60 mW of peak output power was achieved with 15-µm-wide QCL and we observed lasing characteristics up to 60°C.
Numerical analysis on the absorption coefficient of DC-QCLs showed apparent difference depending on the insulation layer. The Y2O3 insulated QCLs had better absorption coefficient than SiO2 or Si3N4 and the difference was increased with narrower waveguide width due to the larger mode overlap on the insulation layer. We also showed further improvements could be expected by using a low doped substrate instead of the highly doped substrate which is currently used and by introducing higher mesa angle to enhance the optical mode confinement in the active region.
Funding
Ministry of Trade, Industry and Energy (MI, Korea) (10053010).
References and links
1. J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, and A. Y. Cho, “Quantum Cascade Laser,” Science 264(5158), 553–556 (1994). [CrossRef] [PubMed]
2. B. S. Williams, “Terahertz quantum-cascade lasers,” Nat. Photonics 1(9), 517–525 (2007). [CrossRef]
3. J. Faist, F. Capasso, C. Sirtori, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, S. N. G. Chu, and A. Y. Cho, “High power mid-infrared (λ~ 5 µm) quantum cascade lasers operating above room temperature,” Appl. Phys. Lett. 68(26), 3680–3682 (1996). [CrossRef]
4. A. Lyakh, R. Maulini, A. Tsekoun, R. Go, C. Pflügl, L. Diehl, Q. J. Wang, F. Capasso, and C. K. N. Patel’, “3W continuous-wave room temperature single-facet emission from quantum cascade lasers based on nonresonant extraction design approach,” Appl. Phys. Lett. 95(14), 141113 (2009). [CrossRef]
5. 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]
6. A. Evans, J. S. Yu, J. David, L. Doris, K. Mi, S. Slivken, and M. Razeghi, “High-temperature, high-power, continuous-wave operation of buried heterostructure quantum-cascade lasers,” Appl. Phys. Lett. 84(3), 314–316 (2004). [CrossRef]
7. W. Metaferia, B. Simozrag, C. Junesand, Y. T. Sun, M. Carras, R. Blanchard, F. Capasso, and S. Lourdudoss, “Demonstration of a quick process to achieve buried heterostructure quantum cascade laser leading to high power and wall plug efficiency,” Opt. Eng. 53(8), 087104 (2014). [CrossRef]
8. S. Slivken, J. S. Yu, A. Evans, J. David, L. Doris, and M. Razeghi, “Ridge-Width Dependence on High-Temperature Continuous-Wave Quantum-Cascade Laser Operation,” IEEE Photonics Technol. Lett. 16(3), 744–746 (2004). [CrossRef]
9. Z. Liu, D. Wasserman, S. S. Howard, A. J. Hoffman, C. F. Gmachl, X. Wang, T. Tanbun-Ek, L. Cheng, and F. S. Choa, “Room-Temperature Continuous-Wave Quantum Cascade Lasers Grown by MOCVD Without Lateral Regrowth,” IEEE Photonics Technol. Lett. 18(12), 1347–1349 (2006). [CrossRef]
10. G. Rehouma, C. A. Evans, Z. Ikonic, D. Indjin, and P. Harrison, “Comparison of SiO2, Si3N4, As2S3, and Ge0.25Se0.75 dielectric layers for InP-and GaAs-based material systems for midinfrared quantum cascade laser waveguides,” Appl. Phys. Lett. 106, 053104 (2009).
11. S. Ferre, A. Peinado, E. Garcia-Caurel, V. Trinite, M. Carras, and R. Ferreira, “Comparative study of SiO2, Si3N4 and TiO2 thin films as passivation layers for quantum cascade lasers,” Opt. Express 24(21), 24032–24044 (2016). [CrossRef] [PubMed]
12. J. S. Yu, S. Slivken, A. Evans, L. Doris, and M. Razeghi, “High-power continuous-wave operation of a 6 µm quantum-cascade laser at room temperature,” Appl. Phys. Lett. 83(13), 2503–2505 (2003). [CrossRef]
13. D. F. Bezuidenhout and R. Pretorius, “The optical properties of evaporated Y2O3 films,” Thin Solid Films 139(2), 121–132 (1986). [CrossRef]
14. A. Wittmann, Y. Bonetti, J. Faist, E. Gini, and M. Giovannini, “Intersubband linewidths in quantum cascade laser designs,” Appl. Phys. Lett. 93(14), 141103 (2008). [CrossRef]
15. Y. H. Ko and J. S. Yu, “Analysis and design of waveguide structure for InGaAs/InAlAs quantum cascade lasers at λ~4.6-9.5 µm,” Phys. Status Solidi., A Appl. Mater. Sci. 208(12), 2900–2906 (2011). [CrossRef]
16. R. A. Soref and B. R. Bennett, “Electrooptical Effects in Silicon,” IEEE J. Quantum Electron. QE-23(1), 123–128 (1987). [CrossRef]
17. J. Kischkat, S. Peters, B. Gruska, M. Semtsiv, M. Chashnikova, M. Klinkmüller, O. Fedosenko, S. Machulik, A. Aleksandrova, G. Monastyrskyi, Y. Flores, and W. Ted Masselink, “Mid-infrared optical properties of thin films of aluminum oxide, titanium dioxide, silicon dioxide, aluminum nitride, and silicon nitride,” Appl. Opt. 51(28), 6789–6798 (2012). [CrossRef] [PubMed]
18. J. D. Traylor Kruschwitz and W. T. Pawlewicz, “Optical and durability properties of infrared transmitting thin films,” Appl. Opt. 36(10), 2157–2159 (1997). [CrossRef] [PubMed]
