Photoluminescence and lasing in a high-quality T-shaped quantum wires M. Yoshita, Y. Hayamizu, Y. Takahashi, H. Itoh, and H. Akiyama Institute for Solid State Physics, Univ. of Tok yo and CREST, JST L. N. Pfeiffer and K. W. West Bell Laboratories, Lucent Technologies FOPS, at Stanley Hotel, Estes Park, CO, USA (2004 aracterization of T-wires, single-T-wire lasers AFM, PL, PLE, PL scan, Lasing citon, biexciton, and plasma in our best-quality single T-wir PL, Absorption/Gain spectra measured by Cassidy’s method citon Mott transition picture does not work. New picture is n
31
Embed
Photoluminescence and lasing in a high-quality T-shaped quantum wires M. Yoshita, Y. Hayamizu, Y. Takahashi, H. Itoh, and H. Akiyama Institute for Solid.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Photoluminescence and lasing in a high-quality T-shaped quantum wires
M. Yoshita, Y. Hayamizu, Y. Takahashi, H. Itoh, and H. Akiyama
Institute for Solid State Physics, Univ. of Tokyo and CREST, JST
L. N. Pfeiffer and K. W. WestBell Laboratories, Lucent Technologies
FOPS, at Stanley Hotel, Estes Park, CO, USA (2004.8)
1. Characterization of T-wires, single-T-wire lasers
AFM, PL, PLE, PL scan, Lasing
2. Exciton, biexciton, and plasma in our best-quality single T-wire
PL, Absorption/Gain spectra measured by Cassidy’s method
3. Exciton Mott transition picture does not work. New picture is needed!!!
Cleaved-edge overgrowth with MBE
In situCleave
(001) MBE Growth (110) MBE Growth
[110]
[001]
GaAssubstrate
600oC 490oC
by L. N. Pfeiffer et al., APL 56, 1679 (1990).
490oC GrowthHigh Quality
T-wire
Interface control by growth-interruption annealing
(by M. Yoshita et al.
JJAP 2001)
Atomically flat
interfaces
600oC Anneal
armwell6nm
stemwell14nm
(Akiyama et al. APL 2003)
PL and PLEspectra
1D free exciton
small Stokes shift
1D continuum states
armwell
stemwell
T-wire
E-field
// to wire
_ to wire// to arm wellI
E-field
PLE
Absorption= 80-90 /cm (=5x10-4), or T= 1-2% @ L=0.5mm cavityfor single T-wire
Cavity length 500 m
Probability of Photon
Probability of Electron
Single quantum wire laser
=5x10-4
Scanning micro-PL spectra
ContinuousPL peak over 20 m
PL width < 1.3 meV
scan
T=5K
T-wire T-wirestem well stem well
500m gold-coated cavity
Threshold 5mW
(Hayamizu et al, APL 2002)
Lasing in a single quantum wire
Excitation power dependence of PL
Single quantum wire T=4K
M. Yoshita, et al.
Free exciton
Biexciton+Exciton
Electron-hole plasma
Den
sit
y
Single quantum wire T=4K
n1D = 2.6 x 104 cm-1 (rs = 30 aB)
n1D = 1.7 x 105 cm-1 (rs = 4.6 aB)
n1D = 1.2 x 106 cm-1
(rs = 0.65 aB)
n1D ~ 102 cm-1
aB ~13nm
EB =3meV
•No peak shift
•Gradual & symmetric broadening
Single quantum wire T=4K
Biexciton
Plasma
PL from 1D-continuum band edge
▼ plasma band edge (low energy edge of plasma PL) starts at biexciton energy and shows red shift.
exciton band edge, (onset of continuum states)exciton ground and excited states show no shift.
M. Yoshita, et al., submitted to PRL, but
Lasing & many-body effects in quantum wires
E. Kapon et al. (PRL’89) Lasing in excited-states of V-wiresW. Wegscheider et al. Lasing in the ground-state of T-wires, no energy shift, (PRL’93) excitonic lasingR. Ambigapathy et al. PL without BGR, strong excitonic effect in V-wires (PRL’97) L. Sirigu et al. (PRB’00) Lasing due to localized excitons in V-wiresJ. Rubio et al. (SSC’01) Lasing observed with e–h plasma emission in T-wiresA. Crottini et al. (SSC’02) PL from exciton molecules (bi-excitons) in V-wiresT. Guillet et al. (PRB’03) PL, Mott transition form excitons to a plasma in V-wiresH. Akiyama et al. Lasing due to e–h plasma, no exciton lasing in T-wires
(PRB’03)
F. Rossi and E. Molinari (PRL’96)F. Tassone, C. Piermarocchi, et al. (PRL’99,SSC’99)S. Das Sarma and D. W. Wang (PRL’00,PRB’01)
Theories
“1D exciton Mott transition”
eg. D. W. Wang and S. Das Sarma, PRB 64, 195313 (2001).
・ reduction of exciton binding energy
・ red shift of the band edge (band-gap renormalization (BGR))
Physical picture of 1D exciton–plasma transition
Increase of e–h pair density causes
the exciton Mott transition Our PL results show
band edge
exciton level
no energy shift of the exciton band edge
plasma low-energy edges appear at the bi-exciton energy positions, and show BGR
no connection, but coexistence of two band edges
no level-crossing between the band edges and the exciton level
θee
eE
ll
l
I22
2
sinR4)R1(
R)1(A)(
B. W. Hakki and T. L. Paoli JAP. 46 1299 (1974)
11
R1
ln1pp
l
min
sum/FSR
I
Ip
R :Reflectivity
c
Eln
: Absorption coeff.
D. T. Cassidy JAP. 56 3096 (1984)
Absorption/gain measurement based on Cassidy’s analysis of Fabry-Perot-laser emission below threshold
Free Spectral Range
Point
Absorption Spectrum by Cassidy method
Excitation Light : cw TiS laser at 1.631eV
WaveguideEmission
Polarizationparallel toArm well
Spectrometer with spectral resolution of
0.15 meV
Cassidy’s Method
Single wire laser, uncoated cavity mirrors
Excitation Light : cw TiS laser at 1.631eV
WaveguideEmission
Polarizationparallel toArm well
Stripe shape
Spectrometer with spectral resolution of
0.15 meV
Spontaneousemission
Measurement of absorption/gain spectrum
Cassidy’s Method
8.3mW
Absorption/gain spectrum (High excitation power)
Electron-Hole Plasma
EFEEBE
Gain
Absorption
Hayamizu et al. unpublished
8.3mW
1. Exciton peak and continuum onset decay without shift.
2. Gap between exciton and continuum is gradually filled.
3. Exciton changes to Fermi edge
Electron-Hole Plasma
ExcitonHayamizu et al. unpublished
Conclusions
Exciton-Mott-transition picture does not work. New picture is needed.
1. As e-h density is increased, exciton peak and exciton band edge (the onset of continuum states above excitons) decay with NO shift.
2. Exciton band edge does NOT connect to plasma band edge (the low-energy edge of plasma). They even co-exist. Therefore, these edges NEVER cross exciton peak.
3. The exciton-plasma evolution is NOT like an abrupt metal-insulator transition, but a gradual crossover.
4. Lasing is caused by plasma gain, but the gain spectral shape is NOT proportional to 1D density of states, probably due to Coulomb interactions.
5. Exciton gradually changes to Fermi edge in plasma.6. Biexciton PL gradually changes to plasma PL without shift.
•The gain peaks appear
2meV below biexciton energy.
•Gain peaks have symmetric
shape and no similarity
to 1D Density of States.
Gain peaks of 20-wires laser
•The gain peaks are broadened
with slight red shifts.
(001) and (110) surfaces of GaAs
(001) (110)
[001]
[110]
[110]
[001]
(By Yoshita et al. APL 2002)
Theory1D exciton and continuum states
Intensity vs. excitation powerSingle quantum wire T=4K
Plasmars = 0.65 aB
Quadratic increase
BiexcitonSingle quantum wire T=4K
T. Guillet et al. (PRB’03) Mott transition form an exciton gas to a dense plasma in very-high-quality V-wire
eg. D. W. Wang and S. Das Sarma, PRB 64, 195313 (2001).
・ red shift of the band edge (band-gap renormalization (BGR))