Thermal performance investigation of DQW GaInNAs laser diodes Jun Jun Lim, Roderick MacKenzie, Slawomir Sujecki, Eric Larkins Photonic and Radio Frequency Engineering Group, School of Electrical and Electronic Engineering University of Nottingham, Nottingham NG7 2RD M. Sadeghi, S.M. Wang, G. Adolfsson, Y.Q. Wei, A. Larsson Photonics Laboratory, Chalmers University of Technology, SE-41296 Göteborg, Sweden P. Melanen, P. Uusimaa Modulight Inc., Hermiankatu 22, FIN-3372 Tampere, Finland A.A. George, P.M. Smowton Cardiff School of Physics and Astronomy, Cardiff University, Queens Buildings, The Parade, Cardiff, CF24 3AA
Acknowledgements We gratefully acknowledge the support of the European Commission through the FP6 IST project FAST ACCESS (IST-4772)
Presentation Outline 1. Introduction 2. Device structure 3. Calibration of gain data 4. Description of laser simulator 5. Calibration of operating characteristic 6. Cladding doping concentration investigation 7. Conclusion
Dilute Nitride Lasers 7 Threshold current (ma) 6 5 4 3 2 1 15 μm 1 μm 5 μm 3 μm T = 181 K 215 K 266 K 243 K 1 2 3 4 5 6 7 8 9 1 11 12 Temperature ( o C) 35 µm RW lasers, uncoated facet Bandgap energy reduction Longer wavelength Large conduction band offset High T o Low-cost alternative to InP lasers for access networks 17 GHz maximum modulation bandwidth Characteristic temperature, T o = 181-266 K (2-7 C) Reference: Y.Q. Wei et al., Optics Express, Vol. 14, pp. 2753-2759, 26
Dilute Nitride Laser Structure 7nm Ga.61 In.39 N.12 As/GaAs DQW grown by Chalmers University Ridge width of 3.4μm, and etch depth of ~1.3μm processed by Modulight Inc. 1.3μm 3.4μm Reference: Y.Q. Wei et al., Appl. Phys. Lett., Vol. 88, 5113, 26.
Energy (ev) Bandstructure Calculation BAC conduction band model Use 4x4 k. p valence band mixing model Material parameters from: Vurgaftman & Meyer, JAP 89, 5815, 21 Band-offset ratio ΔE c /ΔE g =.7 Nitrogen level relative to VBM of GaAs, E N = 1.65 ev Adjusted interaction parameter, V MN = 2.15 x.4.35.3.25.2.15.1.5. -.5 CB1 CB2 N level -.1..2.4.6.8 k (1/Å) Energy (ev).12.1.8.6.4.2. -.2 HH1 HH2..2.4.6.8 k (1/Å)...2.4.6.8 Conduction band Valence band Transition strength Transition Strength.8.7.6.5.4.3.2.1 k (1/Å) CB1-HH1 CB2-HH2
Calibration of Gain Spectra Experimental gain measured using the segmented contact method Theoretical gain transformed from spontaneous emission spectra 3π 2 h 3 c 2 E ΔE g( E) = R ( E) 1 exp F 2n 2 E 2 spon kt Spontaneous emission spectra broadened using hyperbolic secant function 13 16 18 1 2 Broadening lifetime: τ in = (1.95 1 4 1 T ) n ( 1 ) s 6 4 6 4 6 4 Measurement Calculation Modal Gain (cm -1 ) 2-2 -4-6 Measurement Calculation.94.96.98 1. 1.2 1.4 Energy (ev) Modal Gain (cm -1 ) 2-2 -4 Measurement Calculation -6.92.94.96.98 1. 1.2 Energy (ev) Modal Gain (cm -1 ) 2-2 -4-6.92.94.96.98 1. 1.2 Energy (ev) T=3K T=325K T=35K
Description of 2D Laser Simulator The model solves the following equations: Poisson s Equation + ( ε ε φ) + q( p n + N N ) = r D A Continuity Equations n 1 = Jn ( Rnr + Rspont + R t q p t 1 = J q p ( R nr + R spont n cap + R ) p cap ) n t w p t w = = 1 q 1 q dj dx nw dj dx pw ( R ( R qw nr qw nr + R + R qw spont qw spont + R + R qw stim qw stim R R n cap p cap ) ) Photon Rate Equations (m modes, 1 eqn. per mode) ds m qw = vg ( Gm α ) Sm + βr dt spont The equations are cast into matrix form and solved simultaneously using Newton s Method.
Description of Laser Simulator Optical model Helmholtz s scalar 2D wave equation 2 Φ + 2 2 ( k( x, y) β ( ω) ) Φ = Solved using Rayleigh quotient iteration method Thermal model Lattice heat equation Heat sources: - Joule heat - Recombination heat H E q E q = c v J + J Joule n p Recomb nr g, w - Capture heat - Absorption heat H Cap H = R E = R ( E E ) H = v αφs net g gw Abs m g m hω
Calibration of Ridge Waveguide Laser Material parameters taken from the literature Layer d (μm) E g (ev) μ n (cm 2 /Vs) μ p (cm 2 /Vs) κ (W/mK) n r p-gaas.1 1.42248 139 42.2 46 3.41741 p-al.5 Ga.5 As (cladding) 1. 2.876 196 77.7 1.77 3.1515 p-al.5-.2 Ga.5-.8 As (graded).16 2.876-1.7233 89-2875 181-253 1.77-14.46 3.1515-3.3392 Ga.61 In.39 N.12 As (QW).7 1.797 5 483 5.8 3.6 GaAs (barrier & SCL).2 1.42248 78 49 46 3.41741 n-al.2-.5 Ga.8-.5 As (graded).16 1.7233-2.876 2875-89 253-181 14.46-1.77 3.3392-3.1515 n-al.5 Ga.5 As (cladding) 1. 2.876 196 77.7 1.77 3.1515 n-gaas.3 1.42248 1851 51 46 3.41741 Bandgap energies from Vurgaftman & Meyer, JAP 89, 5815, 21. Carrier mobilities and thermal conductivities from Joachim Piprek, Semiconductor Optoelectronic Devices, Academic Press, 23 & Vassil Palankovski, Analysis and Simulation of Heterostructure Devices, Springer, 24. Refractive indices from Afromowitz, Solid State Comm., 15, 59, 1974 and from Kitatani et al., Jpn. J. Appl. Phys., 37, 753, 1998 for AlGaAs and GaInNAs respectively.
Calibration of Material Parameters A good fit of threshold current vs temperature for different cavity lengths was achieved: σ IVBA ~2x1-16 cm 2 α i =1 cm -1 C CHHS = 1x1-28 cm 6 s -1 τ n = τ p =.5ns SRH recombination was found to dominate the threshold current. 7 6 Measurement Simulation 1 Threshold current (ma) 5 4 3 2 1 15μm 1μm 5μm 3μm Current (ma) 1 Total current SRH recomb. Auger recomb. Spont. recomb. 3 35 31 315 32 325 33 335 34 345 35 Temperature (K) 1 3 31 32 33 34 Temperature (K)
Calibration of Ridge Waveguide Laser 3.4 x 3 μm 2 RW laser To obtain agreement, Auger coefficient needs to be temperature dependent: C HSH C HSH ( T ) = C T n [ 1+ β ( T T ) ] 35K ( T ) = C T > < 35K C = 1x1-28 cm 6 s -1, β = 5x1-4, n = 2.5 and T = 35 K
Cladding doping concentration study p-clad and n-clad doping concentration set equal and varied from 1x1 17 to 4x1 18 cm -3 original doping concentration = 5x1 17 cm -3
Cladding doping concentration study L-I curves for different cladding doping concentration at 75 C.
Cladding doping concentration study Threshold current vs. cladding doping concentration Maximum output power vs. cladding doping concentration.3% improvement in threshold current at 3x1 17 cm -3 1% improvement in maximum output power at 1.5x1 18 cm -3
Conclusion Gain and spontaneous emission calculation calibrated to experimental data. Auger coefficient, SRH lifetime and absorption cross section fit to obtain agreement in threshold current vs. temperature and length. Self-heating model and temperature dependent Auger coefficient necessary to reproduce experimental L-I curves above threshold. An optimum doping concentration exists to reduce heat generation due to competition between Joule heating and FCA. A 1% improvement in maximum output power achieved using optimum cladding doping concentration.