MTLE-6120: Advanced Electronic Properties of Materials Optical - - PowerPoint PPT Presentation

mtle 6120 advanced electronic properties of materials
SMART_READER_LITE
LIVE PREVIEW

MTLE-6120: Advanced Electronic Properties of Materials Optical - - PowerPoint PPT Presentation

1 MTLE-6120: Advanced Electronic Properties of Materials Optical properties of Materials Reading: Kasap: 9.1 - 9.18 2 Maxwells equations in free media Linear response of materials described very generally by ( ) and ( )


slide-1
SLIDE 1

MTLE-6120: Advanced Electronic Properties of Materials Optical properties of Materials

Reading:

◮ Kasap: 9.1 - 9.18

1

slide-2
SLIDE 2

Maxwell’s equations in free media

◮ Linear response of materials described very generally by ǫ(ω) and µ(ω) ◮ Maxwell’s equations in the absence of free charges and currents

∇ · (ǫ(ω) E) = 0 ∇ × E = iω B ∇ ×

  • B

µ(ω) = −iω(ǫ(ω) E) ∇ · B = 0

◮ Substitute second equation in curl of third equation:

∇ × (∇ × B) µ(ω) = −iωǫ(ω)∇ × E = ω2ǫ(ω) B −∇2 B = ω2ǫ(ω)µ(ω) B using ∇ · B = 0

2

slide-3
SLIDE 3

Electromagnetic waves

◮ Write Maxwell’s equation in linear media with no free charge or current as:

v2(ω)∇2 B = −ω2 B v2(ω)∇2 E = −ω2 E where v(ω) ≡ 1/

  • ǫ(ω)µ(ω)

◮ This is exactly the equation of a wave with speed v

v2∇2f = −ω2f v2∇2f = ∂2f ∂t2 where f is either E or B

◮ Solution to the wave equation with |

k| = ω/v: f( r) = fei

k· r

f( r, t) = fei(

k· r−ωt)

3

slide-4
SLIDE 4

Polarization

◮ EM wave propagating along x:

  • E(

r, t) = E0ˆ yei(kx−ωt)

  • r
  • E(

r, t) = E0ˆ zei(kx−ωt) two field directions ⇒ two independent linear polarizations

◮ Linearly combine to

  • E(

r, t) = E0 ˆ y + iˆ z √ 2 ei(kx−ωt)

  • r
  • E(

r, t) = E0 ˆ y − iˆ z √ 2 ei(kx−ωt) ⇒ two independent circular polarizations (left/right)

◮ Photons: circular polarizations have spin angular momentum ±

along propagation direction (photons are spin s = 1 particles)

4

slide-5
SLIDE 5

Wave speed and refractive index

◮ Electromagnetic waves satisfy dispersion relation ω = v|

k|

◮ EM wave velocity v(ω) = 1/

  • ǫ(ω)µ(ω)

◮ In vacuum, ǫ(ω) = ǫ0 and µ(ω) = µ0, so that speed of light in free space

c = 1 √ǫ0µ0

◮ In materials, speed of light usually specified by refractive index

n(ω) ≡ c v(ω) =

  • ǫ(ω)µ(ω)

ǫ0µ0

5

slide-6
SLIDE 6

Refractive index: frequency dependence

◮ Most materials are non-magnetic µ ≈ µ0 ◮ Therefore n2(ω) ≈ ǫr ≡ ǫ(ω)/ǫ0 ◮ Several contributions to relative permittivity:

n2 ≈ ǫr = 1 + χ(free e−)

e

+ χ(bound e−)

e

+ χ(ions)

e

+ χ(dipoles)

e ◮ Frequency dependence (Drude-Lorentz model):

n2(ω) = 1 − ω2

p

ω2 + iω/τ +

  • α

χ0αω2

ω2

0α − iγαω − ω2

where α includes bound e−, ions and dipoles

◮ First Drude term proportional to free carrier density:

◮ Dominates for metals ◮ Present to varying degrees for doped semiconductors

◮ Remaining terms contribute in all materials

6

slide-7
SLIDE 7

Refractive index: wavelength dependence

◮ Frequency dependence (Drude-Lorentz model):

n2(ω) = 1 +

  • α

χ0αω2

ω2

0α − iγαω − ω2

(Drude term captured by setting χ0ω2

0 → ω2 p, ω0 → 0 and γ → 1/τ) ◮ Equivalently in terms of wavelength λ = 2πc/ω:

n2(λ) = 1 +

  • α

χ0αλ2 λ2 − iδαλ − λ2

where λ0 = 2πc/ω0 and δ = 2πcγα/ω2

◮ Can use a number of α in the empirical Sellmeier relation instead:

n2(λ) = 1 +

  • α

Aαλ2 λ2 − iδαλ − λ2

α

7

slide-8
SLIDE 8

Complex refractive index

◮ The n2 we have been working with so far has been implicitly complex

(due to the τ and γ’s)

◮ Customary to call complex refractive index N = n + iK

(where n and K are real)

◮ Wave speed v = c N = c n+iK is complex ◮ Wave vector k = ω v = ω c (n + iK) = k0(n + iK) is complex ◮ What do the imaginary parts here mean? ◮ Wave function (fields) ∝ eikx = ei(nk0)x · e−(Kk0)x ◮ Wave intensity ∝ e−(2Kk0)x (proportional to |field|2) ◮ ⇒ Wave attenuates as e−αx with absorption coefficient α = 2Kk0

8

slide-9
SLIDE 9

Kramers-Kronig relations

  • 3
  • 2
  • 1

1 2 3 4 5 0.5 1 1.5 2 α [q2/k] ω [(k/m)1/2] Re(α) Im(α)

◮ Real and imaginary parts of ǫ(ω) (and hence n(ω)) not independent ◮ Causality ⇒ response functions (eg. χ(ω))

must be complex analytic functions (χ(z))

◮ Leads to the Kramers-Kronig relations:

Reχ(ω) = 1 π P +∞

−∞

dω′ Imχ(ω) ω′ − ω , Imχ(ω) = −1 π P +∞

−∞

dω′ Reχ(ω) ω′ − ω

◮ Absorption at ω0 affects n at all ω

9

slide-10
SLIDE 10

Absorption mechanisms

Every polarizability term has a corresponding absorption mechanism:

◮ Free electrons: resistive loss in Drude term ◮ Bound electrons: absorption excites electronic transitions ◮ Ions: absorption excites phonons ◮ Dipoles: absorption excites molecular rotations ◮ Simple model so far (Lorentz) assumed single resonant frequency ω0 ◮ Each of the above processes have a band of absorption frequencies instead

10

slide-11
SLIDE 11

Lattice absorption: Restrahlen band

◮ Optical phonons in ionic crystals: charge oscillations ⇒ interact with EM

(hence the name ‘optical’ phonons)

◮ High DOS for transverse and longitudinal optical phonons ◮ Absorption peaks at corresponding wavelengths: Restrahlen band ◮ Typically in the mid-infrared part of the spectrum

11

slide-12
SLIDE 12

Electronic absorption: bound model

◮ Simple model for inuslators so far: electrons bound by springs ◮ In reality: electrons delocalized in insulators / semiconductors too ◮ Only no net current in completely filled (or empty) bands ◮ Low frequency response is effectively like bound electrons (because vd = 0) ◮ Bound model obtained response of form:

χ(ω) = χ0ω2 ω2

0 − iγω − ω2 ◮ Interpretation of behavior near ω ∼ ω0 needs revision ◮ Energy absorbed from light must be taken up by the electrons ◮ How can you increase the electronic energy of insulators? ◮ Move an electron from the valence to conduction band (leave hole behind) ◮ What is the minimum frequency that can do this? ω = Eg ◮ Effectively ω0 now corresponds to Eg/

12

slide-13
SLIDE 13

Optical excitation of electrons

◮ Light excites electron in valence band (initial energy Ei < 0) to conduction

band (final energy Ej > Eg)

◮ Energy conservation: Ei + ω = Ej ⇒ ω = (Ej − Ei)/ > Eg/ ◮ Momentum conservation ki + ω/c = kj ⇒ kj − ki = ω/c = 2π/λ ◮ Typical Eg ∼ 1 − 10 eV with λ∼ 100 nm - 1 µm (NIR, visible, UV) ◮ Typical ki, kj ∼ 2π/a with a < 1 nm ◮ ⇒ photon momentum is negligible in electronic processes ◮ Effectively ki = kj (vertical transitions) for optical absorption

13

slide-14
SLIDE 14

Direct absorption

◮ Direct band-gap semiconductors (eg. GaAs): VBM and CBM at same k ◮ Possible to have Ej − Ei = Eg with ki = kj ◮ Direct absorption: light produces electron-hole pair ◮ Allowed here for ω ≥ Eg/

14

slide-15
SLIDE 15

Indirect absorption

E in eV

6 −12 L Λ Γ Δ Χ Σ Γ −6

Phonon Photon

◮ Indirect band-gap semiconductors (eg. Si): VBM and CBM at different k ◮ Not possible to have Ej − Ei = Eg with ki = kj ◮ Indirect absorption: light produces electron-hole pair ± phonon ◮ Conservation: Ej − Ei ± ωph = ω, kj ± kph = ki ◮ Negligible energy in phonon, negligible momentum in photon ◮ Direct absorption becomes possible above interband threshold Et

15

slide-16
SLIDE 16

Absorption in metals (plasmon decay)

◮ Direct interband transitions from d-bands to Fermi level ◮ Indirect transitions possible in two ways:

◮ Get momentum from phonons (lattice vibrations) ◮ Geometry effect in nanostructures: uncertainty principle / momentum from

surface

◮ Additionally resistive losses at all frequencies (dominates at low frequency)

16

slide-17
SLIDE 17

Excitons

◮ Photons absorbed to produce electron-hole pairs for ω ≥ Eg ◮ This assumes electrons and holes don’t interact with each other ◮ Instead they attract with potential V (r) ∼ −e2/(rǫr) ◮ Corresponding binding energy Eb = meff mǫ2

r Ryd (typically ∼ 0.01 − 0.1 eV)

(similar to the donor/acceptor level estimate)

◮ Here meff ∼ m∗ em∗ h/(m∗ e + m∗ h) is reduced mass of electron-hole pair ◮ Exciton: bound pair of electron and hole ◮ Can absorb photons when ω ≥ Eg − Eb (slightly smaller than band gap)

17

slide-18
SLIDE 18

Wave packets

x0 = ω’(k0)t ei(k0x - ω(k0)t)

◮ Velocity so far is phase velocity v(ω) = ω k = c n(ω) ◮ Wavepackets travel with group velocity vg = ∂ω ∂k ◮ Same concept for all waves; only difference: dispersion relation ω(k)

18

slide-19
SLIDE 19

Group velocity and group index

◮ Group velocity

vg = ∂ω ∂k = 1 ∂k/∂ω = 1 ∂(ωn(ω)/c)/∂ω = c n(ω) + ωn′(ω) = c n(λ) − λn′(λ) since ω ∝ λ−1

◮ Correspondingly group index:

ng ≡ c/vg = n(ω) + ωn′(ω) = n(λ) − λn′(λ)

◮ For a non-dispersive medium (constant n), vg = v and ng = n

19

slide-20
SLIDE 20

Optical energy density and power flux

◮ Energy density in electric field = ǫE2 2 ◮ Energy density in magnetic field = B2 2µ ◮ Maxwell’s equation ∇ ×

E = −∂ B/∂t ⇒ | E| = v| B| =

| B| √ǫµ ◮ Therefore energy densities in

E and B are equal

◮ Net energy density = ǫE2 ◮ Wave moves with velocity v, so power flux is vǫE2

= v2ǫEB = 1

µEB = EH ◮ More generally, power flux given by Poynting vector

E × H

20

slide-21
SLIDE 21

Interface between media

◮ In material with index n1, light incident towards interface with n2 ◮ Wavevectors: incident

k1, reflected k′

1 and transmitted

k2

◮ At boundary,

E, D⊥ and B continuous (non-magnetic)

◮ Frequencies: common ω for all three (phases match for all t) ◮ Phase of fields must match at all x and y ◮ y-matching implies all three wavevectors in same plane ◮ x-matching implies that for all x:

eik1 sin θ1x = eik1 sin θ′

1x = eik2 sin θ2x

⇒ k1 sin θ1 = k1 sin θ′

1

= k2 sin θ2

◮ Therefore θ1 = θ′ 1 and

sin θ2 sin θ1 = k1 k2 = ω/v1 ω/v2 = c/v1 c/v2 = n1 n2 (Snell’s law)

21

slide-22
SLIDE 22

Fresnel’s equations: E in plane ()

◮ Matching conditions:

ǫ1(E1 sin θ1 + E′

1 sin θ′ 1) = ǫ2E2 sin θ2

(Normal) E1 cos θ1 − E′

1 cos θ′ 1 = E2 cos θ2

(Tangential)

◮ Solve for E′ 1 and E2:

r ≡ E′

1

E1 =

ǫ2 sin θ2 ǫ1 sin θ1 cos θ1 − cos θ2 ǫ2 sin θ2 ǫ1 sin θ′

1 cos θ′

1 + cos θ2

⇒ r = n2 cos θ1 −

  • n2 − sin2 θ1

n2 cos θ1 +

  • n2 − sin2 θ1

t = 2n cos θ1 n2 cos θ1 +

  • n2 − sin2 θ1

where n = n2/n1

22

slide-23
SLIDE 23

Fresnel’s equations: E normal to plane (⊥)

◮ Matching conditions:

1 v1 (E1 cos θ1 − E′

1 cos θ′ 1) = 1

v2 E2 cos θ2 (B tangential) E1 + E′

1 = E2

(E Tangential)

◮ Solve for E′ 1 and E2:

⇒ r⊥ = cos θ1 −

  • n2 − sin2 θ1

cos θ1 +

  • n2 − sin2 θ1

t⊥ = 2 cos θ1 cos θ1 +

  • n2 − sin2 θ1

where n = n2/n1

23

slide-24
SLIDE 24

Reflection and transmission coefficients

◮ Power flow = vǫ|E|2 = (c/n)n2|E|2 ∝ n|E|2 ◮ Therefore, R = |r|2 and T = n|t|2 (for each of and ⊥)

0.2 0.4 0.6 0.8 1 10 20 30 40 50 60 70 80 90 R θ [o] n = n2/n1 = 1.5 || | _ 0.2 0.4 0.6 0.8 1 10 20 30 40 50 60 70 80 90 R θ [o] n = n2/n1 = 1/1.5 || | _

◮ R⊥ increases monotonically with incidence angle ◮ R goes through zero at Brewster angle θB = tan−1 n ◮ R = 1 beyond θc = sin−1 n when n < 1 (total internal reflection) ◮ Structures on wavelength scale: interference effects

  • eg. anti-reflective coatings (see section 9.7)

24

slide-25
SLIDE 25

Scattering

◮ Index change due to impurity, defect or particles

(in a suspension / composite)

◮ Scattered light output in all directions

(cannot separate reflected / transmitted)

◮ Contributes to attenuation α (along with absorption) ◮ Rayleigh scattering: scatterer size ≪ λ (wave limit) ◮ Scatterng cross section ∝ λ−4 ◮ Angular distribution ∝ (1 + cos2 θ) ◮ Mie theory: general case of spherical particles <, ∼, > λ ◮ Used to determine particle sizes from diffraction measurements ◮ Scatter size ≫ λ (particle limit) ◮ Scattering cross section ∼ πr2 (constant in λ) ◮ Why is the sky blue, but clouds white/grey?

25

slide-26
SLIDE 26

Optical fibers

◮ Higher n core surrounded by

lower n sheath

◮ Channel light using total

internal reflection

◮ Pulse shape distortion due to

angle distribution

◮ Rectify / minimize using

graded index / single mode fibers

◮ Attenuation: scattering at low

λ, lattice absorption at high λ

◮ Alternate glasses like heavy

metal fluorides (expensive) eg. ZBLAN (Zr-Ba-La-Al-Na-F)

26

slide-27
SLIDE 27

Anisotropic media

◮ In crystals, n2 = ǫ is a tensor ◮ In general, values n1, n2, n3 along principal directions ◮ Isotropic: n1 = n2 = n3 eg. cubic ◮ Uniaxial: n1 = n2 = n3 eg. tetragonal, hexagonal ◮ Biaxial: n1 = n2 = n3 eg. orthorhombic, triclinic ◮ Optic axis: direction along which n⊥ is isotropic ◮ Uniaxial: single optic axis (direction 3 above) ◮ Biaxial: two optic axes (not along 1, 2 or 3) ◮ Consequence: two polarizations have different n for same direction ◮ Refracted waves in different directions (birefringence) ◮ Commonly used material: calcite (rhombohedral) ◮ Polarizers, half-wave and quarter-wave plates (Read 9.15 - 9.17)

27

slide-28
SLIDE 28

Luminescence

◮ Reverse of electronic absorption: electron de-excitation emits light

(de-excitation includes electron-hole recombination in semiconductors)

◮ Electrons in excited states for multiple reasons:

◮ Optical excitation: photoluminescence (PL) ◮ Electronic excitation: cathodoluminescence ◮ Thermal excitation: thermoluminescence

◮ Includes fluorescence and phosphorescence

(slowed by forbidden transitions)

◮ Phosphors: materials used for their luminescence ◮ Example: Cr3+ in corundrum Al2O3 emits in red (ruby) ◮ Traps / dopants in semiconductors: radiative recombination ◮ Applications:

◮ Defunct display technologies (CRT) ◮ Down-conversion: white LEDs generate red components from blue ◮ Time-resolved PL: analysis of trap levels and lifetimes

28

slide-29
SLIDE 29

Stimulated emission

◮ Consider two level system with populations N1 and N2 ◮ Incident light with intensity I: absorb to excite 1 → 2

− ˙ N1 = ˙ N2 = B12IN1

◮ Emission in presence of light:

˙ N1 = − ˙ N2 = AN2

  • spontaneous

+ B21IN2

stimulated ◮ Einstein’s theorem: B12 = B21 for detailed balance / microscopic

reversibility

◮ Basically, matrix element for forward and reverse process ◮ Net result:

˙ N1 = − ˙ N2 = AN2 + BI(N2 − N1)

◮ Important: stimulated emission in same quantum state as incident

(Bosonic effect)

29

slide-30
SLIDE 30

Lasers

◮ Basic idea: a synchronized

photoluminescence using stimulated emission

◮ Population inversion (N2 > N1)

by optical pumping

◮ Need long-lived metastable state

to hold N2 (small A)

◮ Resonant cavity: light reflects

back and forth

◮ Stimulated emission enhances

intensity (amplification)

◮ Light output through slightly

transmitting mirror

◮ Coherence: light with single

wavelength and phase (line narrowing)

30

slide-31
SLIDE 31

Electro- and magneto-optic effects

◮ Nonlinearity of ǫ and hence n with respect to

E n( E) = n0 + a1 · E

Pockel

+ E · ¯ a2 · E

  • Kerr

+ · · ·

◮ Kerr effect: present in all materials to varying degrees ◮ Pockel’s effect: present only in materials that are non-centosymmetric

(symmetry breaking needed to give a1 a direction)

◮ Note: n must be tensorial for such materials ◮ Similarly, magneto-optic effect: nonlinearity in

B

◮ Faraday rotation: magnetic field along propagation direction rotates

polarization

◮ Relative permeability of iron ∼ 2 × 105 ⇒ n = √ǫµ ∼ 450 ◮ Relative permeability of ferrite ∼ 640 ⇒ n = √ǫµ ∼ 25 ◮ Much higher than any dielectric: why are these not used in optics?

31