YFR8010 Practical Spectroscopy

Transcription

1 YFR8010 Practical Spectroscopy Jüri Krustok Lembit Kurik Urmas Nagel (KBFI) 1 YFR8010 Practical Spectroscopy I 4 laboratory works: 1. Raman spectroscopy 2. Photoluminescence 3. µ-photoluminescence 4. External quantum efficiency of solar cell 2 1

2 YFR8010 Practical Spectroscopy Raman spectroscopy Professor Jüri Krustok 3 History n Discovered in 1928: India (Raman, Krishnan) and Soviet Union (Landsberg, Mandelstam) n In 1928 Raman and Krishnan published a paper in "Nature", but the explanation of the experiment was not correct. n Few month later in 1928 Landsberg and Mandelstam published a paper with correct explanation. n In Soviet Union this effect was called combination scattering. 4 2

3 Raman 5 Light and matter n If the light is directed on to matter we observe absorption, reflection and scattering. n SCATTERING can be: n Elastic (Rayleigh) λ secondary = λ primary n Nonelastic (Raman) λ secondary λ primary The wavelength of light changes with RAMAN scattering 6 3

4 Elastic (Rayleigh) scattering Intensity of scattering depends on the wavelength λ of primary light and on the scattering angle Θ Blue light is scattered more than red light Blue sky- Reyleigh scattering of sunlight on molecules of air! 7 Raman vs. Rayleigh scattering Raman scattering intensity is 10 7 times weaker than the primary light (laser) intensity 8 4

5 Raman and FTIR (infrared spectroscopy) FTIR Raman i. vibrational modes vibrational modes ii. change of dipole movement change in polarizability δ - 2δ + δ - iii. Excitation of the molecule Light as an electromagnetic field: to higher vibrational electron state disturbing of electron cloud near the bond 9 Physics of Raman effect n Raman spectra show very weak peaks with frequency ω and this frequency is a combination of primary light frequency ω 0 and the frequency of vibrational states of molecules (or vibrational modes of crystals) ω i. ω ω ± = 0 ω i 10 5

6 Raman theory A Excited electron levels Virtual levels A Laser photon is absorbed 11 Raman theory Rayleigh scattering energy does not change!! hν in = hν out A Raman scattering energy changes!! hν in <> hν out 12 6

7 Raman theory Anti-Stokes: E = hν + ΔE Stokes: E = hν - ΔE ΔE is related to the vibrational states of the molecule 13 Raman spectroscopy: wavenumbers ~ = ν 1 λ ν(cm ~ 1 4 ) = 10 λ(µm) 1 cm -1 = mev In spectra it means an energetic distance from laser line!! 14 7

8 Raman spectrum In typical Raman spektrum we have: 1. Rayleigh peak (ΔE=0) 2. Anti-Stokes peaks, where the energy is increased, i.e. ΔE<0 3. Stokesi peaks, where the energy is decreased, i.e. ΔE>0 In general we record Stokesi peaks 15 Molecules and crystals n Molecules vibrations of molecules n Crystals- Lattice vibration modes (phonons) n Lattice vibration modes depend on symmetry of the lattice, atomic masses and bond strengths between atoms 16 8

9 Example- chalcopyrite CuInSe 2 CuInS 2 CuGaSe 2 and others 17 Vibrational modes in chalcopyrite n In chalcopyrite A I B III C 2 VI crystal we have 8 atoms in unit cell. Accordingly we have 24 vibrational modes: 1A 1 + 2A 2 + 3B 1 + 4B 2 + 7E, From them we have 22 Raman active modes: 1A 1 + 3B 1 + 3B 2 (LO) + 3B 2 (TO) + 6E(LO) + 6E(TO). n A 1 always dominates spectra, because it is related to anion movement only n B 1 mode is related to cation movement and B 2 and E modes are related to movement of all atoms. 18 9

10 A 1 mode in chalcopyrite The most intensive peak in Raman spectra 19 A 1 peaks in chalcopyrites 20 10

11 Raman spectra of CuInS(Se) 2 21 The shape of Raman peaks Lorenz shape Gauss shape 22 11

12 Raman spectra Intensity BLUE laser Fluorescence GRN laser Raman shift RED laser 400 Wavelength Ramani peaks are very close to the laser line n The separation of peaks is better with red laser n Intensity of Raman peaks is proportional to [1/λ laser ] 4 n blue laser gives better signal! Fluorescence kills a weak Raman 23 signal Micro-Raman spectrometer Notch Filter Obj Lens CCD Detector Sample Laser 24 12

13 Notch filter It is used to remove scattered laser light 25 Long-wave-pass filter 26 13

14 Confocal microraman 27 Raman spectrometer HORIBA Jobin Yvon LabRam HR VI-143 room 28 14

15 Raman spectrometer (from behind) 29 Scanning Raman Micro scan from the surface of unknown pill 4 different peaks- 4 different colors 30 15

16 Diamond Raman cm -1 Zirconia (used to replace diamond) : 31 Raman of different polymers 32 16

17 Raman of human body 33 Raman of explosives 34 17

18 Fitting of Raman (and PL) spectra n Free FITYK software 35 Raman measurements n Measure different samples n Spectra fitting with Lorenzian function n Find Imax, Emax, W for each peak n Memory stick!!!! 36 18

19 YFR8010 Practical Spectroscopy PHOTOLUMINESCENCE Prof. Jüri Krustok 37 Photoluminescence in semiconductors 38 19

20 Recombination n Radiative recombinationphotoluminescence n Non radiative recombination- DARK! 39 Equipment for photoluminescence 40 20

21 PL setup laser 41 PL measurements Detectors: Photomultiplier tubes (PMT) : FEU-79 (visible range) R632- red and near infrared range Semiconductor detectors: Si- Visible and near infrared range (~1000nm) InGaAs- near infrared range (~1700 nm) PbS- infrared (~3000 nm) 42 21

22 PL measurement software in TUT- LabView project 43 PL bands in semiconductors: Deep defects Shallow defects Band-to-band Excitons E g E, ev 44 22

23 Photoluminescence in semiconductors n Recombination luminescence n Center luminescence 45 Edge emission 46 23

24 Edge emission 47 GaAs edge emission I( hν ) ~ A( hν E ) 1/ 2 g hν E exp( kt g ) 48 24

25 Edge emission is used in: n LED (Light emitting diode) n Semiconductor lasers 49 Excitons AX DX FE ~mev Exciton: electron and hole pair E g E (ev) 50 25

26 PL of excitons S.H. Song et al. / Journal of Crystal Growth 257 (2003) CdTe excitons 51 Excitons in CuInS

27 Photoluminescence of donor-acceptor pairs 53 Photoluminescence of donor-acceptor pairs n Probability of electron (on the donor defect) and hole (on the acceptor defect) recombination: W DA =W o exp(-2r/a o ) n The PL energy of each DA pair depends on the distance between donor and acceptor R: 2 e hν DA( R) = Eg ( ED + EA) + εr n In case of shallow levels we have a distribution of pairs: 4 4π 3 N ( R) = N N 4πR exp( N R ) DA D A 3 D 54 27

28 Photoluminescence of donor-acceptor pairs classic work! DAP in GaP Phys. Rev. 133, 1964, p. A Photoluminescence of donor-acceptor pairs DAP in GaP 56 28

29 Photoluminescence of donor-acceptor pairs t-shift GaN MRS Internet J. Nitride Semicond. Res. 6, 12(2001). R increases W decreases t increases E max decreases Closest pairs will recombine FASTER 57 Photoluminescence of donor-acceptor pairs j-shift R increases W decreases Phys. Rev. B, 6, 1972, 3072 I laser increases E max increases The recombination of DA pairs having larger distance will start to saturate, because the recombination probability W is 58 small 29

30 Deep donor-deep acceptor pairs -> DD-DA pairs. n DD-DA can be found in different materials. n Only closest pairs can excist between these deep defects, because of overlap of wave-functions. n It is possible to find an energy difference between closest pairs: ΔE mn = 2 e 1 ε rm 1 r n 59 DD-DA pairs as grown compensated c-band r 1 r 2 E D 0 D1 D2 r W D1 D2 E A 0 v-band 60 30

31 DD-DA pairs D6 CuIn 0.5 Ga 0.5 Se 2 D1 PL intensity (a.u.) CuGaSe 2 D6 AgInS D6 2 D1 D1 DD-DA pairs in different ternaries. Theoretical positions are marked with vertical lines D6 D1 CuInS E (ev) 61 DD-DA pairs 5 4 D5 0 b) a) CuInS 2 T=8K D4 0 DD-DA pairs can also bound excitons: PL intensity (a. u.) hω 1 hω 2 Deep excitons in CuInS E (ev) 62 31

32 DD-DA pairs in CdTe Different interstitial positionsdifferent PL bands J. Krustok, H. Collan, K. Hjelt, J. Mädasson, V. Valdna. J.Luminescence, v , pp (1997). 63 DD-DA pairs in AgInS 2 Chalcopyrite Orthorombic PL intensity (a. u.) D5 D6 D4 T=8K D3 D2 D1 c-agins 2 PL intensity (a. u.) D6 D4 D5 D3 D2 D1 o-agins 2 T=8K E (ev) E (ev)

33 Temperature quenching of PL PL ln (I) E T =114 ± 6 mev Fitting Experiment E T /T (1/K) Thermally liberated holes 65 Temperature quenching of PL q THEORY: q J. Krustok, H. Collan, and K. Hjelt. J. Appl. Phys. v. 81, N 3, pp (1997). I ( T ) I 0 = ϕ T + ϕ T exp 1 2 ( E kt ) T / 66 33

34 Shape of PL band n Theory Pekar, Huang, Rhys, jt. n The shape is not easy to calculate n Simplifications: n T=0 n only one phonon is present and not the phonon spectrum Then in is possible to find transition probability from each energy level of excited state to each energy level of ground state- all these probabilities follow Poisson distribution 67 r µ exp( µ ) Poisson distribution P( r) = r! 68 34

35 Shape of PL band n The theoretical shape of PL band is: m exp( S) S I ( E) = I0 δ ( E0 m! ω E) m! m The shape of each m-th small PL band depends mainly on interactions with acoustic phonons 69 Shape of PL band Int F(x) m=1 E 0 m=0 0-phonon peak S=0.47 m exp( S) S I ( E) = I0 δ ( E0 m! ω E) m! m Phonon replicas having Gaussian shape m=2 phonon energy PEKARIAN

36 Shape of PL band 5 D5 0 a) b) CuInS 2 PL band in CuInS 2 4 T=8K D4 0 PL intensity (a. u.) hω 1 hω 2 two phonons can be seen E (ev) 71 Shape of PL band Gaussian shape: I( E) 4ln2 exp W = I0 2 2 ( E E ) max If in Pekarian S ->, then we also will have a pure Gaussian shape: I( E) 2 ( E E S! ω) 4ln2 0 exp 8ln2S(! ω) = I

37 Shape of PL band PL intensity (a. u.) Lorenz shape Gaussian shape E (ev) 73 PL of heavily doped semiconductors n Most of ternaries and quaternary compounds are heavily doped, i.e. they have hudge amount of defects. n distance between defects is smaller than the Bohr radius of electrons and holes on defects-> heavy doping. n Edges of conduction and valence bands are curved because of potential fluctuations

38 PL of heavily doped semiconductors PL intensity T=12.5K Typical PL spectrum in CuInGaSe E, ev J. Krustok, H. Collan, M. Yakushev, K. Hjelt. The role of spatial potential fluctuations in the shape of the PL bands of multinary semiconductor compounds. Physica Scripta, T79, pp (1999). 75 Theory of heavily doped materials ε F n ρ c BT BB E g E c Theory: Shklovskii, Efros Levaniuk, Osipov ρ v E v 76 38

39 Edge emission of CuInGaSe 2 PL intensity (a. u.) K 4 15 K 20 K 30 K 40 K 3 50 K 60 K 70 K 80 K 95 K K 120 K 130 K 145 K K 175 K 190 K 210 K BB-band 240 K E (ev ) BT-band J. Krustok, H. Collan, M. Yakushev, K. Hjelt. The role of spatial potential fluctuations in the shape of the PL bands of multinary semiconductor compounds. Physica Scripta, T79, pp (1999). BB and BT bands 77 Temperature dependence of BB and BT bands hν max (ev) BT- band hν max ~ - 4.4kT hν max ~ 2.1kT T 1 =96K BB- band T 2 = 175K T (K) J. Krustok, J. Raudoja, M. Yakushev, R. D. Pilkington, and H. Collan. phys. stat. sol. (a) v. 173, No 2, pp (1999). J. Krustok, H. Collan, M. Yakushev, K. Hjelt. Physica Scripta, T79, pp (1999)

40 PL of heavily doped semiconductors 79 PL measurements n Measure PL spectra n Spectra fitting with Gaussian function n Find Imax, Emax and W for each PL band 80 40

41 YFR8010 Practical Spectroscopy Spectral Response of Solar Cells Professor Jüri Krustok 81 I-V Curve of Solar Cell 82 41

42 Spectral Response ideal solar cell Lost current 83 Spectral Response- theory EXTERNAL QUANTUM EFFICIENCY EQE KαL eff A E E g 1=2=E 84 42

43 Fiding E g 85 Finding E g EQE KαL eff A E E g 1=2=E 86 where constant A includes all en 43

44 Finding E g 2,7 2,4 2,1 SnS T=300K E g =(1,411±0,008) ev (hν*eqe) 2 (arb.units) 1,8 1,5 1,2 0,9 0,6 E g =(1,298±0,005) ev 0,3 0,0 1,25 1,30 1,35 1,40 1,45 1,50 1,55 1,60 E(eV) 87 What to do? n Measure EQE (E) spectrum for solar cell n Draw (EQE*E) 2 vs E curves n Calculate E g 88 44

45 Meeting point U06 first floor end of the building room 156 Phone: