MBE Growth of Terahertz Quantum Cascade Lasers Harvey Beere

Size: px

Start display at page:

Download "MBE Growth of Terahertz Quantum Cascade Lasers Harvey Beere"

Homer Wiggins
5 years ago
Views:

1 MBE Growth of Terahertz Quantum Cascade Lasers Harvey Beere Cavendish Laboratory J J Thomson Avenue Madingley Road Cambridge, CB3 0HE United Kingdom

2 People involved Harvey Beere, Chris Worrall, Josh Freeman, Sean Whelan, David Ritchie Jesse Alton Stefano Barbieri, Carlo Sirtori Université Paris 7 2

3 People involved Harvey Beere, Chris Worrall, Josh Freeman, Sean Whelan, David Ritchie Jesse Alton Stefano Barbieri, Carlo Sirtori Université Paris 7 3

4 Presentation Outline - Introduction - Issues associated with growth - Robustness of Active Region designs - Minor tweaks to Active Region - Transfer of structures between growth reactors - Summary

5 The Unipolar Semiconductor Laser CB CB Diode laser: material VB Quantum Cascade laser: layer thickness materials by design : band structure engineering through molecular beam epitaxy 1971: amplification from intersubband transitions is first postulated by R. F. Kazarinov and R. A. Suris, Sov. Phys. Semicond. p5, (Ioffe) 1994: QCL is first experimentally demonstrated in MIR by J. Faist et al. Science 254, p553 (Bell Labs) 2002: QC-lasers outperform other mid-ir lasers in many aspects R. Köhler et al. Nature 417, p156 (SNS Pisa/Univ Cambridge)

6 What Grower Sees Köhler et al., Nature 417, p156 (2002) 4.4THz chirped superlattice QCL AR x104 Repeats Al 0.15 Ga 0.85 As/GaAs

7 What Grower Sees Final QCL device: periods active region 12-18µm thick layers Some barriers ~6Å (~2MLs) 12-18hrs growth duration

8 What Grower Sees Final QCL device: periods active region 12-18µm thick layers Some barriers ~6Å (~2MLs) 12-18hrs growth duration PUSHING BOUNDARIES OF GROWTH TECHNIQUES - growth rate calibration - growth rate stabilty - inteface roughness } accuracy of layer thickness

9 Molecular Beam Epitaxy Precise semiconductor growth technique extensive range of source materials layer thickness to monolayer accuracy high degree of control on layer composition & doping level abrupt interfaces little interface diffusion

10 Theoretical growth rate tolerances 23 Optical transition energy (mev) Transition energy within +/- 0.5meV Gallium Flux Aluminium Flux Design transition energy is 18.0meV Percentage variation of growth flux from desire rates GaAs growth rate within +2% and 1% AlAs growth rate within +10% and 5% Beere et al., J Cryst Growth 278, p756 (2005)

11 X-ray spectra of THz laser Intensity (arb. units) A2565 Simulation ω / 2θ (arcsec) Clear satellite structure in good agreement with simulation Confirms excellent growth stability over 12 hour growth duration Thickness variation across 2 wafer is less than 2%

12 Terahertz Quantum Cascade Laser Single-mode emission from a Fabry-Perot Cavity Emission at 18.4meV: 4.44THz - Good agreement with design

13 Structure Reproducibility Intensity (arb. units) A2888 A2742 A2702 A2565 Simulation ω / 2θ (arcsec) Good agreement design thickness and measured value (-1% to +2%)

14 Structure Reproducibility A2565 A2702 A2742 A2888 Peak Power (mw) Current Density (Acm -2 ) Single-mode emission from a Fabry-Perot Cavity Emission at ~18.2meV: 4.4THz - Good agreement with design

15 Robustness of Active Region Design Barbieri et al., APL, vol. 85, p1674 (2004) THz bound-to-continuum QCL AR Injection barrier Energy (mev) mev ~12 mev 16 mev x90 Repeats Distance (nm) Al 0.15 Ga 0.85 As/GaAs

16 Active Region Robustness % -2.5% 0% +2.5% +5.0% Calculated Frequency (THz) Energy (mev) AR Period Length Calculated emission frequency vs total AR thickness Again ±0.5meV emission energy equates ~ ±2% thickness

17 Active Region Robustness Current (A) PULSED OPERATION 3mm x 0.25mm ridge Voltage (V) 2 1 4K 20K 40K 60K 70K 77K Peak power (mw) 80K Current density (A/cm 2 ) 90K 95K 0 Single plasmon waveguide V align ~2v: T max ~95K: P max ~90mW Series structure growths -5% to +5%

18 Active Region Robustness % -2.5% 0% +2.5% +5.0% Calculated Veeco Wafers Frequency (THz) Counts θ / deg AR Period Length Actual range (-3%, +5%) ~0.4THz (~2.5meV) Systematic route to tuning emission frequency Could possibly extend frequency range further (>5%)? Energy (mev)

19 Active Region Robustness II Worrall et al., Optics Express, 14, 171 (2006) 2.0 THz bound-to-continuum QCL AR 120 Al 0.1 Ga 0.8 As/GaAs F = 1.5 kv/cm Injection barrier 90 Energy (mev) g 16 mev ~8 mev 1 14 mev Distance (nm)

20 Active Region Robustness II Current (A) Frequency Voltage (V) Emission (Norm Arb) Energy (mev) 4K 67K Emission (mw) Current Density (A/cm 2 ) Single plasmon waveguide 3mm x 0.25mm ridge waveguide J th 103 Acm -2 V align ~1.8v : T max = 67K Output Power ~22mW f ~2.00THz

21 Active Region Robustness II Percentage Thickness Variation (%) Calculated Laser Emission Variation for different Active Region thickness Emission Frequency (THz) Emission Energy (mev) Active Region Thickness (Å) Similar linear trend emission frequency against AR thickness Limited range (~3%) explored to date ~0.06THz span Investigate extending frequency further! 7.86

22 Active Region Robustness II - Bound-to-continuum design very robust! - Possible systematic route to tuning frequency how far can we realistically exploit this method?

23 Minor Structure Variations V305 3mm x 250mm 2.0THz Reference Current (A) Frequency Voltage (V) Emission (Norm Arb) Energy (mev) 4K 67K Emission (mw) Current Density (A/cm 2 ) J th 103 Acm -2 V align ~1.8v : T max = 67K Output Power ~22mW f ~2.00THz

24 Minor Structure Variations V309 3mm x 250µm 2.0THz Thicker Injection Barrier Current (A) Reduced Leakage, lower threshold (increased useable current) 50 Voltage (V) Similar dymanic range compared to reference Same tempurature of operation as reference (12% thicker injector: 6Å) 4K 67K Emission (mw) Current Density (A/cm 2 ) J th 90 Acm -2 V align ~1.7v : T max = 67K Output Power ~22mW

25 Minor Structure Variations V309 (Thick Barrier) Laser Emission Spectra Emission (Norm Arb) mA 954mA 952mA 910mA Energy (mev) 1.99THz (8.25meV) Singlemode (Frequency identical to Reference)

26 Minor Structure Variations Voltage (V) V308 3mm x 250 µm 2.0THz Thinner Injection Barrier Current (A) Increased leakage, higher threshold current Device performance lower (12% thinner injector : 6Å) 4K 47K Current Density (A/cm 2 ) J th 133 Acm -2 (V align ~1.9A): T max = 47K Output Power ~7mW Emission (mw)

27 Minor Structure Variations 2.0THz Thin Injector Structure Laser Emission Spectra Emission (Norm Arb) V308 (Thick Inj) 1000mA 1030mA 1050mA 1100mA 1150mA 1200mA Energy (mev) 1.88THz (7.79meV) lower frequency compared 2.0THz reference

28 Minor Structure Variations Overlap strength between lower state and upper, injector 2THz Design 2THz Thinner injection Barrier Dipole (nm) Aligned Voltage (V) injector upper Splitting (mev) Dipole (nm) Laser action from the injector. Hence the longer wavelength Voltage (V) injector upper Splitting (mev) Thinner injector barrier no longer produces isolated upper and injector states STRUCTURE LASES FROM INJECTOR TO LOWER STATE LOWER FREQUENCY

29 Minor Structure Variations Voltage (V) V307 3mm x 250µm 2.0THz Higher Doping Current (A) Laser Action Two Terminal ONLY 38% increase in doping Current spreading is more uniform two terminal, increasing the effective overlap 4K 10K Device worked to 10K - IVs to 20K shown Current Density (A/cm 2 ) Device operation severely degraded Emission (arb)

30 Minor Structure Variations V307 (High Doped) Laser Emission Spectra Emission (Norm Arb) A 2.05A 2.00A 1.95A 1.90A Energy (mev) 1.94THz (8.04meV) Multimode: just above threshold 1.96THz (8.10meV) Multimode: just before NDR

31 Minor Structure Variations 2.0THz Higher Doping & Reference (3mm x 250µm) Current (A) % doping increase 92% current increase 3 Terminal Reference V305 (1.3 x10 16 ) Higher Doping V307 (1.8 x10 16 ) Voltage (V) Alignment Voltage Killed by doping? Emission (arb) x Current Density (A/cm 2 ) Higher current kills device operation

32 Minor Structure Variations 2THz Reference (V305) double-metal THz QCL devices Frequency / THz Voltage / V K 10K 20K 30K 40K 50K 60K 70K 80K 89K Pulsed LIVs - V305 DM 1.35mm, 50µm Power / a.u. Power / arb CW Spectra - V305 DM 1.35mm, 50µm 57mA 67mA 77mA 87mA 97mA Current / ma T max = 89K Wavenumbers / cm µm x 50 µm device Single plasmon waveguide T max ~67K 32

33 Minor Structure Variations Voltage / V 3 2 High doping (V307) double-metal THz QCL devices Pulsed LIVs - V307 2THz 1.02mm, 50µm 4K 20K 40K 60K 73K Power / a.u. Power / arb mA 100mA 110mA 120mA 130mA Frequency / THz CW spectra - V307 2THz High dope 1.02mm, 50µm Current / ma T max = 73K Wavenumbers / cm µm x 50 µm device Slightly degraded performance compared to single plasmon 33

34 Minor Structure Variations - QCL performance sensitive to structure - doping level critical (especially single plasmon waveguide) - ensure variations does not completely change design - Growth needs to be accurate!

35 Transfer of QCLs between growth systems Veeco (V305) 3mm x 250mm 2.0THz Reference Current (A) Frequency Voltage (V) Emission (Norm Arb) Energy (mev) 4K 67K Emission (mw) Current Density (A/cm 2 ) J th 103 Acm -2 V align ~1.8v : T max = 67K Output Power ~22mW f ~2.00THz

36 Transfer of QCLs between growth systems VG (A3847) 3mm x 250mm 2.0THz Reference Current Density (A cm -2 ) Voltage / V Normalized Emission Frequency (THz) Energy (mev) 4K 57K 16 8 Power (mw) Current (A) J th 82 Acm -2 V align ~1.6v : T max = 57K Output Power ~10mW f ~1.99THz

37 Transfer of QCLs between growth systems Percentage Thickness Variation (%) Calculated Laser Emission Variation for different Active Region thickness Emission Frequency (THz) Emission Energy (mev) Active Region Thickness (Å) Active regions different by 27Å (~2%) 0.05THz Variation due to growth or fabrication?

38 Transfer of QCLs between growth systems Percentage Thickness Variation (%) Calculated Laser Emission Variation for different Active Region thickness Emission Frequency (THz) Veeco wafers VG wafers Emission Energy (mev) Active Region Thickness (Å) VG wafers consistently lower frequency (~0.05THz)

39 Transfer of QCLs between growth systems % -2.5% 0% +2.5% +5.0% Frequency (THz) Energy (mev) Calculated Veeco Wafers VG Wafers AR Period Length VG wafers consistently higher frequency (~0.1THz) Observed frequency differences Barrier profile/thickness Growth interfaces

40 Transfer of QCLs between growth systems - Successful transfer of multiple AR designs - Similar performance levels (P, T, J th ) - Different frequency observed for same AR thickness - under investigation

41 Frequency Span Emission Wavelength (µm) µm 150µm 100µm 75µm 60µm Power (mw) T Operation With B Field Emission Frequency (THz) Over 60 working QCLs, incorporating in excess 30 different ARs Since 2002 frequency spans 0.95THz 4.8THz (300µm 62µm)

42 Summary - Highlighted issues associated with growth - Study of Active Region robustness presented - Minor tweaks to Active Region - Transfer of structures between growth reactors - Span of frequencies so far

High power and single frequency quantum. cascade lasers for gas sensing. Stéphane Blaser

High power and single frequency quantum cascade lasers for gas sensing Stéphane Blaser Alpes Lasers: Yargo Bonetti Lubos Hvozdara Antoine Muller University of Neuchâtel: Marcella Giovannini Nicolas Hoyler