Characterization of HgCdTe MWIR Back-Illuminated Electron-Initiated Avalanche Photodiodes (e-apds)

Transcription

1 Draft, version 2.0, 24 Oct 2007 Characterization of HgCdTe MWIR Back-Illuminated Electron-Initiated Avalanche Photodiodes (e-apds) M. B. Reine, J. W. Marciniec, K. K. Wong, T. Parodos, J. D. Mullarkey, P. A. Lamarre, S. P. Tobin, R. W. Minich and K. A. Gustavsen BAE Systems Lexington, Massachusetts M. Compton and G. M. Williams Voxtel Inc. Beaverton, Oregon th U.S. Workshop on the Physics and Chemistry of II-VI Materials Baltimore, Maryland October 30 November 1, 2007

2 Acknowledgements Voxtel, Inc. Beaverton, Oregon Colleagues at BAE Systems: Gary R. Woodward Dr. Paul LoVecchio Funding from BAE Systems

3 At last year s Workshop Large-area back-illuminated 4x4 MWIR HgCdTe e-apd arrays 1.0x1.0 mm² array area largest HgCdTe e-apds yet reported Data at 160 K substantiated previously reported features of HgCdTe e- APDs: Gain increases exponentially with reverse bias voltage Gain-versus-bias curves are quite uniform from element to element Gain increases exponentially with cutoff wavelength for same bias voltage Spot scan data demonstrated spatially uniform gain, with hint of edge enhancement C(V) data reveal unconstrained depletion region, with N D as expected M. B. Reine, J. W. Marciniec, K. K. Wong, T. Parodos, J. D. Mullarkey, P. A. Lamarre, S. P. Tobin, K. A. Gustavsen and G. M. Williams, HgCdTe MWIR Back-Illuminated Electron-Initiated Avalanche Photodiode Arrays, presented at the U.S. Workshop on the Physics and Chemistry of II-VI Materials, Oct , 2006, Newport Beach, California. Jour. Electronic Materials 36, (2007).

4 HgCdTe Back-Illuminated Electron-Initiated APD Arrays 250 µm Unit Cell CdTe Passivation Indium Bump N-Contact Metal P-Contact N + HgCdTe Layer N - HgCdTe W(V) P-HgCdTe CdZnTe Substrate Antireflection Coating IR Radiation Bump-mounted onto fanout board with nearby preamps (or onto ROIC chip for e-apd Focal Plane Arrays) Planar n + -n - -p junction architecture based on well-developed single-layer LPE HgCdTe on CdZnTe High fill factor, high quantum efficiency, large area with spatially uniform response and gain Vertical collection of photocarriers bandwidth limited by vertical drift-assisted electron diffusion through thin p-type absorber layer No compromise between bandwidth and fill factor, in contrast to the p-around-n lateral collection geometry of DRS and SELEX/Southampton

5 Array Design & Fabrication P-type HgCdTe films grown by horizontalslider LPE from Te-solution 4x6 cm² (111)B CdZnTe substrates Hg vacancy acceptors, N HgV = 2e16 cm -3 Counterdoped with Indium during LPE growth at 4.5e14 cm -3 4x4 arrays with 250x250 µm² pixels 250x250 µm² unit cell with ground grid Ground grid on some arrays Bump mounted on fanout board Characterized in back-illuminated mode No antireflection coating (21% reflection loss at CdZnTe surface)

6 Dark current and photocurrent are gain-multiplied at higher temperatures Current (abs. value) (A) 1E-4 1E-5 1E-6 1E-7 1E-8 Dark Current + Photocurrent Dark Current Gain Max gain=648 at V n-on-p HgCdTe e-apd 163-A1, El 54 λ CO =4.06 µm T=160 K Area=250x250 µm² 1E+5 1E+4 1E+3 1E+2 1E+1 Gain 1E Bias Voltage (V) 1E+0 M(V) = I I PHOTO PHOTO (V) (V = 0) = I BB I BB (V) I (V = 0) I DARK DARK (V) (V = 0)

7 Gain is spatially uniform for large-area e-apds 1E Relative Response 1E-8 1E-9 V=-5 V V=0 163-A1, El 64 T=79 K With filter 0V µm -5V µm Ratio (Gain) 1E-10 Average gain= E Distance (µm) Well-behaved spatial response Hint of edge-enhanced gain First reported spot scan data for avalanche gain in HgCdTe e-apds 1

8 10.0 El. 54 slope=-1/2 Capacitance (pf) 1.0 Voxtel 163-A1, El 54 T=80 K Junction area = 200x200 µm² Corrected for stray capacitance N D (avg)=4.5e14 cm -3 Capacitance Data are Well Behaved Voltage dependence close to that for ideal abrupt junction: C(V) ~ V -1/2, W(V) ~ V 1/2 Depletion Width (µm) Voxtel 163-A1, El. 54 T=80 K Junction area = 200x200 µm² Corrected for stray capacitance N D (avg)=4.5e14 cm -3 Bias Voltage (V) Bias Voltage (V) 54 slope=1/2 Donor concentration deduced from C(V) agrees well with other data: 4.5e14 cm -3 (avg) from C(V) 4.3e14 and 5.2e14 cm -3 from Hall data on annealed pieces 4.8e14 cm -3 for indium from SIMS First reported C(V) data for HgCdTe e-apds

9 New Characterization Data for this Workshop Third film with different cutoff wavelength Expanded temperature range: 80 K 200 K Gain-normalized dark current data at 80 K Temperature dependence of M(V) gain characteristics Dependence of measured gain on diode area Data for F(M) 1 Observations on abrupt breakdown seen in all devices

10 Gain versus voltage data agree well with Beck s HgCdTe e-apd model A1 λ CO =4.06 µm Els 49, 51, 54, 62 n-on-p e-apds Voxtel Test Chips, Lot 1 T=160 K Area=250x250 µm² 100 Gain A2 λ CO =3.54 µm Els 36, 38, 48, 88 Beck model : M(V) = 1+ 2 V th = 6.8 E [2(V V G th ) / V th ] Bias Voltage (V) Beck model based on p-around-n HgCdTe e-apds at 77 K with cutoff wavelengths of µm. See: M.A. Kinch, J.D. Beck, C-F Wan, F. Ma and J. Campbell, "HgCdTe Electron Avalanche Photodiodes," J. Electronic Mat. 33, 630 (2004)

11 New M(V) Data for Third Film Agree well with Beck Model G, λ CO =4.29 µm Els 32, 34, 40, 44, 85, 89 n-on-p e-apds Voxtel Test Chips Lots 1, 2 T=160 K Area=250x250 µm² A1, λ CO =4.06 µm Els 49, 51, 54, 62 Gain 177-A2, λ CO =3.54 µm Els 36, 38, 48, Beck model : M(V) = 1+ 2 V th = 6.8 E [2(V V G th ) / V th ] Bias Voltage (V)

12 Gain-normalized dark current increases gradually with bias voltage at higher temperatures I DARK and I DARK /M (abs. value) (A) 1E-4 1E-5 1E-6 1E-7 1E-8 1E-9 I DARK I DARK /M Voxtel n-on-p Test Chip 163-A1 λ CO (160 K)=4.06 µm T=160 K Elements 49, 51, 54, 62 Area=250x250 µm² I SAT =kt/er 0 T=160 K R 0 A=2860 ohm-cm² R 0 =4.6e6 ohm 1E Bias Voltage (V) Average gain-normalized dark current = 24.4 na at V Average gain-normalized dark current density = 39 µa/cm² at V

13 Low Dark Current in MWIR HgCdTe e-apds at 80 K 1E-4 1E-5 F/5, with BB 163A1-B1, T=80 K λ CO =4.23 µm at 80 K 250x250 µm² 1E+9 1E+8 1E-6 1E+7 Current (abs value) (A) 1E-7 1E-8 1E-9 1E-10 1E-11 FOV=0 F/5, Dark 1E+6 1E+5 1E+4 1E+3 1E+2 Gain 1E-12 Gain 1E E+1 1E+0 Voltage (V)

14 Low Dark Current in MWIR HgCdTe e-apds at 80 K 1E-4 1E-5 F/5, with BB 163A1-B1, T=80 K λ CO =4.23 µm at 80 K 250x250 µm² 1E+9 1E+8 1E-6 1E+7 Current (abs value) (A) 1E-7 1E-8 1E-9 1E-10 1E-11 J DARK /M J DARK /M FOV=0 F/5, Dark 1E+6 1E+5 1E+4 1E+3 1E+2 Gain 1E-12 Gain 1E E+1 1E+0 Voltage (V)

15 Low Dark Current in MWIR HgCdTe e-apds at 80 K Cutoff Format A OPT I DARK J DARK Gain J DARK /M Reference (µm) (cm²) (A) (A/cm²) (A/cm²) DRS 4.3 8x8 array 3.E E-09 Beck, Proc SPIE 5564, 44 (2004) µm pitch 5.E E x8 array 2.8E E-07 Beck, Proc SPIE 5564, 44 (2004) E E-07 CEA/LETI E-06 8.E E E-07 Rothman, Proc SPIE 6542 (2007) BAE Systems x250 µm² 6.25E E E E II-VI Workshop E E E-07 T = K V = V

16 Low Dark Current in MWIR HgCdTe e-apds at 80 K 1E-5 1E-6 HgCdTe e-apds T = 80 K J DARK /M (A/cm²) 1E-7 1E-8 BAE, this Workshop DRS, 2004 CEA/LETI, E Cutoff Wavelength (µm)

17 Gain Test Chip 235G λ CO =4.29 µm at 160 K Els 32, 34, 40, 44, 85, 89 Area=250x250 µm² T=160 K Beck model Extraordinary element-to-element uniformity of gain versus voltage (max-min)/avg = 6.7% at V Voltage (V) Test Chip 235G λ CO =4.29 µm at 160 K Els 32, 34, 40, 44, 85, 89 Area=250x250 µm² T=160 K Gain 200 (max-min)/avg = 6.7% at V Voltage (V)

18 Gain Decreases with Increasing Temperature 1E+3 1E+2 T=80 K T=120 K 163A1-B1, El 8 250x250 µm² F/5 4 Oct 2007 Gain T=200 K 1E+1 T=160 K 1E Voltage (V)

19 Gain Decreases with Increasing Temperature 1E G Elements 32, 44, 85 Area = 250x250 µm² F/5 18 Oct E+2 T=80 K Gain T=120 K 1E+1 T=160 K T=200 K 1E Voltage (V)

20 E G (x,t) Accounts for Some of the Decrease in Gain at Higher Temperature Gain at -9.0 V G, El 32, Lco=4.53 µm at 80 K 163A1-B1, El 8, Lco=4.23 µm at 80 K DRS, Lco=2.63 µm at 300 K Beck model, x fixed Temperature (K)

21 Measured Gain Is Larger for Smaller-Area Devices 1E+3 Test Chip 235G λ CO =4.29 µm at 160 K Els 4, 8, 12, 24 Circular diodes T=160 K Circular diodes, widely spaced 1E+2 R J = µm Gain 1E+1 40 µm dia, El 4 60 µm dia, El µm dia, El µm dia, El 24 1E Voltage (V)

22 Photocurrent at V=0 has usual dependence on diode area 1E-5 Curve: least-squares fit to: I PH = I PH1D (1+L OPT /R J )² with L OPT = 12.0 µm A J =200x200 µm² Photocurrent at V=0 (A) 1E-6 1E-7 1E-8 slope=1 235G λ CO = 4.54 µm at 80 K Circular, R J = µm T = 80 K, F/5 1E-5 1E-4 1E-3 Junction Area (cm²)

23 Two-gain model fits dependence of measured gain on diode area M(-7.0 V) data Best fit, Lopt=12.0 µm 60 Gain at -7.0 V Curve is best fit to : M M = with M 1D 1D R 2 J + M = 45, E (R J M [(R + L E J + L OPT = 85, ) OPT 2 L ) 2 OPT R 2 J ] = 12.0 µ m 235G λ CO = 4.54 µm at 80 K Circular, R J = µm T = 80 K, F/5 0 1E-5 1E-4 1E-3 Junction Area (cm²)

24 F(M) 1 for MWIR n-on-p HgCdTe e-apd at 196 K A1, El 80 λ CO =4.06 µm at 160 K 250x250 µm² T = 196 K V = -0.1 V to -9.0 V Voxtel data k=0 k=0.025 El 80 F(M) 2 1 F(M) normailzed to 1 at V=-0.1 V Gain Data taken at Voxtel Inc., Beaverton, Oregon HP8447D wide-bandwidth preamp, HP8566B Spectrum Analyzer, HP8970B Noise Figure Meter λ = 1550 nm Frequency =?????????????

25 Observations on Abrupt Breakdown Abrupt breakdown of dark current seen in all diodes at large bias (unless something else happens sooner) Breakdown present in good diodes and leaky diodes Breakdown limits the achievable gain V B increases with increasing temperature in a given diode V B increases with E G for similar device design & processing V B can vary somewhat from element to element within a film Some diodes recover immediately after breakdown, some are permanently shorted, seemingly dependent on the film Current, absolute value (A) Current (absolute value) (A) 1E-4 1E-5 1E-6 1E-7 1E-8 1E E-3 1E-4 1E-5 1E-6 1E-7 1E-8 1E-9 80 K 200 K 160 K 120 K Voltage (V) 177A2 λ CO =3.54 µm at 160 K T = 173 K Voxtel Data 163A1-B1, El 15 λ CO =4.05 µm at 160 K 250x250 µm² FOV=0 350 µm dia. 250 µm dia. 150 µm dia. 100 µm dia. 60 µm dia. 40 µm dia Voltage(V)

26 Summary & Conclusions Gain data for third film substantiate key features of HgCdTe e-apds: Gain increases exponentially with reverse bias voltage Gain-versus-bias curves are quite uniform from element to element Gain increases exponentially with cutoff wavelength for same bias voltage Low gain-normalized dark current density at 80 K Gain decreases at higher temperatures, only partly due to E G (x,t) Gain is larger for smaller area devices Two-gain model fits dependence on area F(M) 1 at 196 K Abrupt breakdown appears technological, not fundamental