Transcript Slide 1

Ultrafast processes in Solids
Two general categories of processes:
1) Electronic processes
How (fast) do “hot” electrons relax?
How (fast) do electrons and holes recombine?
How quickly is coherence lost?
2) Structural deformations/changes
Phonons
Melting
Electron-phonon interactions
Solids
Crystalline insulators/semiconductors
Metals
Molecular materials
Doped glasses and insulators
Ultrafast processes
Loosely classify solids into 5 categories:
Glasses (amorphous)
Most interesting
Least interesting
For electronic processes, concentrate on semiconductors and metals
Interact reasonably strongly with light
Relevant time scales (fs-ps)
For structural changes, crystalline materials are most interesting
Phonons only occur in crystalline material
Diffraction (used to study melting) only occurs in crystalline material
Bands
Energy
The discrete atomic states in isolated atoms transform into bands in a crystalline
solid as the interatomic distance decreases
free
atom
solid
Interatomic separation
Optical transitions typically correspond to an electron moving from one band
to another (up for absorption, down for emission).
When are such transitions allowed?
p anti-bonding
conduction band
p
s anti-bonding
Eg
p bonding
valence band
s
s bonding
Crystal
Molecule
Atom
Band structure
Taking the full crystal structure into account yields a complicated band structure
E(k)
depends on magnitude and direction of k
examples for two semiconductors shown at right
GaAs – direct gap = interacts strongly with light
Si – indirect gap = only absorbs light
Semiconductor/insulator: valence band full, conduction band empty (Fermi
level in the gap)
For direct gap materials, often only bands near gap need
to be considered
Energy
Metal: band half full (Fermi level in a band)
Conduction
Band
and approximated by parabolas
i.e., a free particle with effective mass
note that effective mass of electrons in valence band
is negative (faster it moves, the slower it goes)
k
Valence
Band
Energy
Optical transitions
Optical excitation adds the energy & momentum of a photon to that of an electron
Conduction
Band
Must be a state (in another band) at resulting energy and momentum
Photon momentum usually negligible
k
Vertical approximation
Valence
Band
equivalent to dipole approximation
Ignoring the Coulomb interaction:
Absorption spectrum due to density of states
Yields square root dependence on E
(in 3 dimensions)
Emission: opposite, subtract photon energy and momentum
Thus the poor emission of indirect gap materials
Absorption Coefficient
dipole moment approximately independent of k
Eg
Energy
Energy
Hot electrons
Conduction
Band
A short pulse tuned well above gap will create a non-thermal
distribution of electrons
Relaxation occurs in following steps:
k
Electrons relax to bottom of the conduction band, typically be
emitting phonons (particular optical phonons)
Valence
Band
Electrons then thermalize amongst themselves
Thermal distribution, but with a temperature much larger
than that of the crystal lattice
Electrons thermalize with lattice
Emission of acoustic phonons
Fermi-Dirac distribution
Electrons recombine with holes
Holes underwent similar relaxation to top of valence band
Lin, Schoenlein, Fujimoto and
Ippen, JQE 24, 267 (1988)
Excitons
What about the Coulomb interaction between the electron and hole?
Results in a bound state known as an “exciton”
Relative electron-hole coordinate is described by hydrogenic wavefunction
Binding energy
~ 4 meV (~ 44 K) in bulk GaAs
~ 10 meV in a GaAs quantum well
Optical
Absorption
~ 25 meV in wide-gap materials (GaN)
~ 104 reduction from hydrogen due to masses and
dielectric constant
Modifies the low temperature absorption spectrum
with
Coulomb
Exciton
and realistic
broadening
Why do we care about excitons?
Large oscillator strength
~ probability electron and hole are on same lattice site
Appreciable coherence time
-2
-1
0
1
Photon Energy - Gap Energy
2
Single particle  Pair
Can we draw an exciton “band” on the band diagram?
No – it is a single particle picture, and the exciton is a pair
Energy
Energy
But we can on a pair diagram
Conduction
Band
Unbound
e-h
pairs
Excitons
Light
k
Valence
Band
Center-of-mass quasi-momentum
Quantum Wells
Bulk
Energy
• Two valence bands
• Heavy Hole
• Light Hole
• Degenerate at k = 0
Energy
• Bulk GaAs:
Quantum Well
conduction
band
conduction
band
• Quantum confinement breaks degeneracy due to
mass difference
• Excitons form between holes in each valence band
– Dominant features in optics near band edge
Absorbance
1.4
k
heavy hole
valence band
1.4
HH exciton
1.2
1.2
1.0
1.0
0.8
0.8
0.6
0.6
LH exciton
0.4
0.4
0.2
0.2
0.0
0.0
1.535
kin plane
1.540
1.545
Photon Energy (eV)
1.550
light hole
valence band
heavy hole
valence band
light hole
valence band
Exciton formation
Free carriers can bind to form an excitons
(T.C. Damen et al, Phys. Rev. B 42, 7434 (1990))
Exciton formation II
However, it later became a topic of interest and controversy
Does a bound exciton actually have to form for luminescence to occur at exciton energy?
For a different interpretation see Chatterjee, et al.,
Phys. Rev. Lett. 92, 067402 (2004)
Exciton Ionization
It is also possible to watch excitons ionize
Use pump pulse with bandwidth less than binding energy
At room temperature, excitons will be thermally ionized
Use broadband probe to monitor spectrum
Exciton Dephasing I
Exciton Dephasing II
Exciton Dephasing: Negative Delay signal
Explanations ultimately included:
1) Local Fields
2) Excitation induced dephasing
3) Biexcitons
4) Excitation induced shift
Distinguishing these was problematic
Solution: 2D spectroscopy (!?!)
Excitonic Coherence
Excitons also have fairly long coherence time (10’s of ps)
For free carriers it is a few fs
Monitor using photon-echo/transient-four-wave-mixing
Decoherence due to
carrier scattering
phonon scattering
disorder effects
However, the coherent response is exquisitely sensitive to
many-body interactions
Exciton Wavepackets
It is also possible to form wavepackets of excitons
and magnetoexcitons
2D spectrum of exciton resonances
cb
-1/2
1/2
Co-circular polarized
excitation pulses
hh -3/2
-1/2
lh
3/2
1/2
Surprises:
• Presence of cross-peaks
• A cross peak is the strongest feature
• “Dispsersive” line shapes
• Lineshapes reveal microscopic interaction
mechanisms
[X. Li, T. Zhang, C.N. Borca, and S.T.C., Phys. Rev. Lett. 96, 057406 (2006)]
• Vertical stripes from continuum
[C.N. Borca, T. Zhang, X. Li, and S.T.C., Chem. Phys. Lett. 416, 311 (2005)]
Experiment – Theory Comparison
Only the full theory reproduces the experiment
T. Zhang, I. Kuznetsova, T. Meier, X. Li, R.P. Mirin, P. Thomas and S.T.C., Proc. Nat. Acad. Sci. 104, 14227-14232 (2007)
Metals
Electron dynamics in metals can be studied
difficult because band structure generally doesn’t allow vertical transitions
overcome using pump-probe photoelectron spectroscopy
Graphene
• Graphene is a layer of carbon atoms in hexagonal structure
• Graphite is many layers of graphene
• Unusual band structure: linear  electrons have zero effective mass
• Cooling of hot electons can be studied with ultrafast differential transmission
ARPES and High-Tc superconductors
• Angle resolved photo-emission spectroscopy (ARPES)
• Usually done at synchrotron
• Can be done with harmonic of ultrafast laser
• Basis for time-resolved ARPES
J.D. Koralek, et al., Rev. Sci. Instr. 78, 053905 (2007)
Time-resolved ARPES
Phonons
We discussed the observation of phonons in solids using impulsive stimulated Raman scattering
Melting
It is also possible to observe melting induced by a laser pulse
Change in reflectivity
Initial change in reflectivity due to e-h plasma
Transfer of energy to crystal causes non-thermal melting
Interpretation complex due to thin film effects
Diffraction
x-ray
electron
Melting observed with x-ray diffraction
Crystal germanium
Initial non-thermal melting
Acoustic effects: shifting of diffraction peaks
Electron Diffraction
Generate electron pulses as for molecule studies
Polycrystalline aluminum
Sufficient resolution to see pair correlations