High mobility 4H-SiC MOSFET using a combination of counter-doping and interface trap passivation

Transcription

1 1 SiC MOS Workshop August 14 th 2014, University of Maryland High mobility 4H-SiC MOSFET using a combination of counter-doping and interface trap passivation Aaron Modic 1, A. Ahyi 1, G. Liu 2, Y. Xu 2,P. Xu 1, Y. Zhou 1, C. Jiao 1, Y. Zheng 1, T. Isaac-Smith 1, M. C. Hamilton 1, L. C. Feldman, 2 and S. Dhar 1 1. Auburn University, Auburn, AL 2. Rutgers University, Piscataway, NJ

2 Max. mobility (cm 2 V -1 s -1 ) 2 Motivation: Interface traps may not be the only mobility limiting factor! E fm Interface traps E c E f E v Phosphosilicate glass (PSG) (a-face) NO (a-face) N plasma (Si-face) PSG process A etal Oxide SiC No. of interface traps (10 11 cm -2 )

3 Counter-doping Metal Dielectric Liu et al., IEEE Elec. Dev. Lett. (2013) (Rutgers University Collaboration) Surface with ultra-thin counter-doped n-type 3 p-type Semiconductor n-type counter-doped layer is normally off due to depletion from backside p-n junction since it is thin Positive gate voltage allows for carriers, forming a channel between source and drain Higher carrier density at same applied E-field Less traps filled at V T due to less band-bending More Coulomb screening E-field SR for same carrier density

4 4 Threshold reduction due to counter-doping Counter-doping charge N d.x d =10 12 cm -2 Metal Dielectric N d x d p-type Semiconductor Courtesy: Dr. Yuming Zhou (AU) Threshold reduction is less as counter-doped layer thickness is decreased

5 Surface counter-doping Fiorenza et al. APL, 103, (2013) 5 Nitridation and phosphorus treatments lead to unintentional thin counter-doped layers

6 6 Arsenic and Antimony Motivation N,P: trap passivation and counter-doping As, Sb: counter-doping only

7 7 Sb and As surface doping process 60keV 2.2E13 cm -2 As +

8 Sb CONCENTRATION (atoms/cc) C,Si,O INTENSITY (arbitrary units) 8 Antimony profile 1E+19 9E+18 8E+18 O-> C-> Si-> 1E+02 Antimony pile-up at interface 7E+18 6E+18 SiO 2 SiC 1E+01 Antimony is lost during oxidation 5E+18 4E+18 Sb Activated antimony percent unknown 3E+18 FWHM ~10nm 1E+00 2E+18 1E+18 0E+00 1E DEPTH (nm) SIMS detection limit is ~5x10 17 atoms/cm 3

9 Field-Effect mobility (cm 2 V -1 s 1 ) 9 High field-effect mobility Sb+NO MOSFET p-well doping: 6x10 15 cm -3 Process Threshold Voltage (V) NO 2 Sb 1.5 Sb+NO T= 300 K Sb+NO Sb NO Oxide Field (MV/cm)

10 10 Temperature effects on mobility Sb only Sb+NO (1) Sb : Counter-doping (2) NO: Trap passivation

11 11 Sb does not passivate interface traps Sb only NO ΔV FB =1.5V ΔV FB =9V Sb mobility pure counter-doping effect

12 CCDLTS (mv) CCDLTS Spectra: Sb-Implanted and NO-30 Samples Contributed by A. Basile, P. Mooney Simon-Fraser University O1+B1 O1 O2 CCDLTS spectra for an Sb-implanted 4H-SiC MOS capacitor (black) and a capacitor NO-annealed for 30 min (red) showing near-interface oxide traps O1 and O2 in both devices A signal from near-interface traps on the SiC side of the interface (B1) overlaps the O1 trap signal in NO-annealed capacitors. 40 Sb-implanted Sb+NO NO T (K) The absence of the B1 traps in Sb-implanted devices accounts for the lower CCDLTS signal at 100K. CCDLTS measurements confirm that Sb does not passivate near-interface oxide traps

13 13 High temperature mobility Sb+NO μ SR μ SP Sb+NO max µ FE decrease with increased T attributed to phonon scattering

14 #

15 15 Summary Sb and As counter-doping leads to high transconductance devices Sb process allows separation of counter-doping and trap passivation effect on mobility Sb and As counter-doping impact on DMOSFET? Process has a potential advantage for improvement of sub-threshold slope and threshold voltage control

16 16 Thank you This research supported in part by: II-VI Foundation NSF US Army Research Laboratory

17 Back up Slides 17

18 18 Future Work Quantification of As and Sb as a function of oxide thickness using SIMS (Rutgers) Sb+NO on heavier doped p-well (Yongju, Auburn) Activated percent of Ab and As? Use of C-ψ s analysis and simulation of CV curves for non-uniformly doped samples (Purdue?) Determination of Sb ionization energy in 4H-SiC (Simon- Fraser University)

19 19 High temperature effects on mobility Sb+NO NO High temperature reduces mobility more with counter-doping than passivation Phonon scattering effect with high carrier concentration?

20 20 Low temperature mobility Sb only Sb+NO

21 Id Sb+ NO? 5V N O Vgs Since all the methods of present study result in similar effective mobility at operating Eox ~4MV/cm, it is better to increase threshold voltage Vt to about 5V, high enough to provide safe normally-off operation at elevated temperatures and immunity from dv/dt related false turn-on; This will enable a unipolar gate drive, 0V +25V, without the need for negative turn-off bias and eliminating concerns over any NBTS-related Vt instabilities ; Better transconductance is desirable than with NO-process for faster switching

22 Id Sb+ NO? 5V NO Vgs Since all the methods of present study result in similar effective mobility at operating Eox ~4MV/cm, it is better to increase threshold voltage Vt to about 5V, high enough to provide safe normally-off operation at elevated temperatures and immunity from dv/dt related false turn-on; This will enable a unipolar gate drive, 0V +25V, without the need for negative turn-off bias and eliminating concerns over any NBTS-related Vt instabilities ; Better transconductance is desirable than with NO-process for faster switching

23 Antimony I-V Breakdown 23

24 I d -V g threshold E ox comparison 24

25 25 Acknowledgements Physics Department, Auburn University A. C. Ahyi, T. Isaacs-Smith, J. R. Williams, S. Dhar Institute for Advanced Materials, Devices and Nanotechnology, Rutgers University Y. Xu, G. Liu, L. C. Feldman Electrical and Computer Engineering, Purdue University S. Swandono, D. Morisette, J. A. Cooper