Thermal management and thermal properties of high-brightness diode lasers

Transcription

1 Thermal management and thermal properties of high-brightness diode lasers Jens W. Tomm Max-Born-Institut für Nichtlineare Optik und Kurzzeitspektroskopie Berlin Max-Born-Str. 2 A, D Berlin, Germany Outline 1. Motivation 2. Experimental methodology and results obtained in Thermal wavelength shift 2.2 Micro-Raman spectroscopy 2.3 Thermoreflectance 2.4 Thermovision (thermal imaging) 3. Summary 1

2 1. Ultra high - power devices MAX-BORN-INSTITUT FÜR NICHTLINEARE OPTIK UND 1. Motivation Output power (W) New Press Release January 19, There is a general tendency towards higher power densities in diode lasers threshold x lower current density x higher thermal load per active volume unit 10 x higher Current (A) Output power (W) Current (A) 3. Increased temperatures accelerate device degradation! 2

3 2.1 Thermal wavelength shift Temperature dependence of the energy gap of the active region material. Emission wavelength (nm) t (s) Emission energy (ev) Pulsed operation f=1 khz τ=100 ns duty cycle 10-2 % I=30 A ( ) mevk T ( C) Averaged temperature along the whole Fabry-Perot cavity. T ( C) Determination of the intrinsic thermal tuning rate by a calibration measurement with low thermal load t (s) 3

4 Spatially resolved emission pattern of a regular In-packaged cm-bar T hs =22 C Local tuning rate T hs =58.5 C 4

5 Tuning rate (mev/k) MAX-BORN-INSTITUT FÜR NICHTLINEARE OPTIK UND Spatially resolved thermal tuning rate of a cm-bar temperature dependent In-packaged on Cu x (mm) Tuning rate (mev/k) AuSn-packaged on CuW x (mm) 5

6 Summary on 2.1 Method is the most established tool for the estimation of thermal properties of diode lasers. It quantifies the averaged bulk temperature along the cavity. But the thermal tuning rate can be interfered by packaging-induced strain. This results, e.g., in a temperature dependent tuning rate and an inhomogeneous tuning rate along a laser bar. Further general limits of the method: No genuine averaging along the laser axis. When changing the temperature, several key parameters are varied at the same time. Does not work for diode lasers made of material with a low de/dt (QD-laser, QCLs). 6

7 2.2 Micro-Raman spectroscopy Raman intensity (a.u.) 2000 IconstlphlphckTStaSTph = 4νexp excitation Stokes-peak 488 nm TO-GaAs < 1mW anti-stokes peak TO-GaAs wavenumber (cm -1 ) T ISt Ia-St νl temperature Stokes intensity anti-stokes intensity frequency of excitation light νph h c k frequency of phonons Planck s constant velocity of light Boltzmann constant S. Todoroki, M. Sawai, and K. Aiki, Temperature distribution along the striped active region in high-power GaAlAs visible lasers J. Appl. Phys. 58, pp ,

8 µ-raman-spectrometer DILOR-xy 8

9 Methodology: Calibration of the temperature measurement Two completely independent approaches based on the same spectra: 300 Example: Intensity ratio I AS :I S Intensity ratio T ( C) Raman line position (cm -1 ) 2. Line position T ( C) T Raman ( C) I (A) RW laser T R determined from intensity ratios (I AS :I S ) T R determined from wavenumber position ν St and ν AS 9

10 Methodology: pulsed measurement vs. cw Measured according to section 2.2 Micro- Raman spectroscopy Conditions: standard 808 nm single chip device repetition rate 20 khz pulse width 8 µs duty cycle 16% T ( o C) facet cw bulk cw facet pulsed Results: bulk cw 9.6 K/A pulsed 2.2 K/A facet cw 32.2 K/A pulsed 12.4 K/A 20 0 bulk pulsed I (A) Measured according to section 2.1 Thermal wavelength shift 10

11 Summary on 2.2 Method allows to determine the (absolute value of) the facet temperature. Probed region can be chosen arbitrarily (at the semiconductor) Dimensions of the probed region: < 1 µm depth ~ 100 nm Limits of the method: One needs to observe distinct Raman-lines from the region of interest (waveguide). If the lines are broadened, e.g. in a ternary or quarternary material no way for the application of the method. Large error bar: ± 10 K There are several parameters, which impact Raman line positions too. - Strain - Carrier concentration 11

12 2.3 Thermoreflectance Method allows to determine facet temperature distributions. Mapping technique. Probed region can be chosen arbitrarily (at the semiconductor) Caution: Each material requires individual thermal calibration No interference with strain effects. Dimensions of the probed region: < 1 µm depth ~ 100 nm Further general limits of the method: No absolute temperature determination. Calibration, e.g., by Raman is necessary. Not flexible regarding the operation regime. Lock-in measurement requires 50% duty cycle. No temporal resolution. 12

13 2.4 Thermovision (thermal imaging) Infrared Camera Thermosensorik CMT 384 M 384 x 288 FPA detector Spectral range: µm Maximum full-frame rate: 150 Hz NETD: < 25 mk Maximum sub-frame rate: > 5 khz Pixel size µm 2 IR-micro-objective (f/2.0) with nominal magnification of 2.5 was mounted on extension tubes of different lengths. Magnification of the imaging system is varied in the range Spatial resolution: A nominal lateral spatial resolution ~ 5 µm is achieved. Depth of field ~ µm Bulk properties are probed! 13

14 Temperature calibration Gray body 100 µm T=22 C, no lasing GaAs GaAs Cu bottom Cu top T=40 C, no lasing T ( C) Camera counts 14

15 Hot spot screening: Surface defect front view top view 15

16 Hot spot screening: Bulk defect front view top view Temperature ( o C) Lateral postion (µm) Temperature ( o C) ,0 0,2 0,4 0,6 0,8 1,0 Position along the laser axis [mm] 16

17 Hot spot screening: Hidden bulk defect Focal plane of the camera adjusted to the laser facet Focal plane adjusted to the defect (shifted by 300 µm) 17

18 Measurements at cm-bars 0 ms 7 ms 14 ms 35 ms 18

19 500 Multi-spectral thermal imaging: Camera signal (counts) Camera sensitivity (Counts/mW)) x (mm) NIR Wavelength (µm) I = 30 A cw T hs = 25 C MIR Temperature ( C) x (mm) Anna Kozlowska et al. Appl. Phys. Lett. 87, (2005). 19

20 Summary on 2.4 Method allows to determine the bulk temperature inside semiconductor lasers. Imaging technique. Probed region can be chosen arbitrarily. Almost all materials can be inspected. In absorbing materials surface temperature Dimensions of the probed region: < 5 µm depth? Further general limits of the method: Temperature calibration of transparent material is tricky. Lack of knowledge about absolute temperatures. Cavity effects are relevant if transparent materials are considered. 20

21 3. Summary Thermal properties of high-brightness diode lasers represent one bottleneck for applications (in particular if high brightness is achieved on cost of increased densities). Examples of results from are shown. Experimental methods are introduced and discussed - Thermal wavelength shift bulk -properties - Thermovision (thermal imaging) - Micro-Raman spectroscopy - Thermoreflectance surface -properties What is the best method? The methods complement each other. 21