Photovoltaic Devices. Content

Transcription

1 Photovoltaic Devices EPFL IMT - PVLab Highlights in Microtechnology, 2014 Content Applications overviews Brief history Basics of photovoltaics (PV) Physical principle and performance limits Multi-junction devices PV technologies Si thin-film technologies Module technology Photodiodes and microsystems 2 1

2 Applications, History and Motivation Photovoltaic effect Edmund Becquerel, the French experimental physicist, discovered (in 1839, at the age of 19) the photovoltaic effect while experimenting (in his father s laboratory - Antoine-César Becquerel s laboratory) with an electrolytic cell made up of two metal electrodes placed in an electricityconducting solution--generation increased when exposed to light. E. Becquerel, 'On electron effects under the influence of solar radiation.' Comptes Rendus 9,

3 Brief history 1839 : Discovery of photovoltaic effect by Becquerel 1877 : Observation of the photovoltaic effect in solid selenium by W.G. Adams and R.E. Day 1883 : Description of the first solar cells made from selenium wafers by Charles Fritts (with efficiency < 1%) 1888 : Edward Weston receives first US patent for "solar cell" 1905 : Albert Einstein s paper on the photoelectric effect 1916 : Experimental proof of Einstein s theory on photoelectric effect by Robert Millikan 1922 : Einstein wins Nobel prize for 1904 paper on photoelectric effect 1941 : First silicon photovoltaic cell developed by Ohl 1954 : First high-power silicon PV cell (achieving 6%) by Bell Labs. The New York Times forecasts that solar cells will eventually lead to a source of "limitless energy of the sun" : US Vanguard I space satellite powered by a PV array 1963 : Viable photovoltaic module of silicon solar cells produced by Sharp Japan installs a 242-watt PV array on a lighthouse, the world's largest array at that time. 1970s : PV costs driven down by 80%, allowing for applications such as offshore navigation warning lights and horns lighthouses, railroad crossings, and remote use where utility-grid connections are too costly; start of mass production : World-wide PV production > 30 GW/year Approx. surface : 200 km 2 (1/3 of lake of Geneva surface) Equivalent annual energy production of 4 nuclear power plants 5 Devices and Applications Solar cells Photodiodes Flat panel imagers (a-si:h) Particle detectors (X,, b, g) Light sensors Position detectors Spatial modulators (optical valves) Bio-chips Smart windows.. 6 3

4 Building integration of photovoltaics 7 Building integration of photovoltaics Integration with partial shadowing 8 4

5 PV fields 9 Solar energy potential Yearly solar energy irradiation on Earth x World energy consumption One hour of solar irradiation on Earth surface Yearly World energy consumption Irradiation (sea level, full sun): 1 kwh/m 2 World primary energy consumption : Energy: >11 GToe Power > 13 TW Toe : Ton oil equivalent 10 MWh 10 5

6 Potential of PV electricity generation Sunlight hitting the dark discs could power the whole world 18 TW (averaged) electrical power for solar cells with 8% efficiency covering the area of the black dots! 11 Basics of PV 6

7 PV device - Physical principle step process : 1) Absorption of a photon in a semiconductor to create a free electron-hole pair 2) Separation of the electron-hole pair thanks to the internal electric field of a diode and carrier collection 13 Solar spectra Black body at 6000 K AM0 AM1.5 solar radiation reaching the Earth and absorption bands due to the atmosphere. Typical clear sky absorption and scattering of incident sunlight. Source PVCDROM 14 7

8 Air mass Air Mass : Attenuation (through the atmosphere) AM 1 cos where is the angle between the sun beam and the vertical direction. AM0: solar spectrum (outside of the atmosphere) AM1: solar spectrum on earth surface when the sun is at the vertical direction AM1.5: angle of 45 of the sun with respect to the vertical direction. AMx defines both the spectrum and the radiation intensity. AM mw/cm 2 AM1.5G (Global), AM1.5D (Direct) Efficiencies of PV systems are usually calibrated at this intensity. 15 Spectral change of illumination Daily (and seasonal) changes of spectral irradiation density Spectral irradiaton density [W/m 2 nm AM0 (data) AM1 AM1.5 (data) AM2 AM AM0: space AM1: equator AM1.5 (central Europe, equinox, noon, 48 above horizon) 30 above horizon 20 above horizon Wavelength [nm] Differences mostly in the visible part of the spectrum Maximum irradiation (normal surface): 1 kw/m 2 Yearly irradiation (Switzerland): kwh/m 2 /year (dependent on the local climatic conditions) 16 8

9 Forbidden gap Ge Si Cu S InP GaAs 2 2 E c E v E g = E c - E v ( gap ) a-ge:h CdTe AlSb Conduction band Valence band a-si:h Cu O Se GaP CdS Every semiconductor has a fixed gap E g, E g is temperature dependent. 2 categories of semiconductors: Simple elements (Si, Ge, C, Se...). Alloys and compounds semiconductors: III-V type (GaAs, AlSb, InP), II-VI type (CdS, ZnSe) or others (Cu 2 O, CuInSe 2, organics) E g [ev] gap E g of various semiconductor material at T=300 K 17 Electron-hole pair creation E c E = ph hv = E g E v E g Conduction band Valence band Optical absorption Three different situations: 1. E ph = E g E c E = ph hv > E g E v E g Conduction band Valence band 2. E ph > E g - Impact ionization (if E ph» E g or in case of high electric field) is negligible in the case of solar cell E c E = hv < E ph g E v E g = E c- E v Conduction band Gap states due to defects or impurities Valence band 3. E ph < E g No absorption Absorption only in presence of gap states (defects or impurities); 18 9

10 Absorption process High band gap (e.g. a-si) Low band gap (e.g. mc-si) Photons with energy less than the band gap are not absorbed Photon energy in excess of band gap is lost to thermalization 19 Bandgaps Direct bandgap semiconductors: GaAs, InP, High absorption coefficient Indirect bandgap semiconductors: Si, Ge, AlSb, GaP Phonon must be absorbed or emitted Lower probability than direct excitation (without phonon interaction) lower absorption coefficient Non direct bandgap: Disordered semiconductors 20 10

11 Spectral conversion efficiency c-ge (crystalline Germanium), E g = 0.66 ev (small gap) Incident photons Incident photons Ge Si Almost all photon are absorbed 40% s c-si (crystalline Silicon), E g = 1.12 ev (medium gap) Une big part of the photons are Absorbed; IR photons are not. 48% s Cu 2 O (Cupper oxide), E g = 2.1 ev (big gap) Spectral conversion efficiency S (physical limit of the conversion efficiency) S or S E g Energy content of the solar spectrum per unit of time where f is the photon flux f ( g fq)(e / q) Energy content of thesolar spectrum per unit of time Incident photons Cu 2O Une small part of the photons are absorbed Remarks: (f q) corresponds to an electrical current (E g / q) corresponds to an electric potential 28% s 21 Spectral conversion efficiency Spectral conversion efficiency s Ge CIS c-si GaAs / CdTe a-si:h Se E g [ev] spectral conversion efficiency s as a function of gap E g under AM1.5 illumination S depends on light spectrum semiconductor gap Optimum bandgap for each spectrum Combination of semiconductors with different gaps may increase S Incident photons Combination of 2 semiconductors with different gaps: With these 2 materials: s >60%. Cu2O (large Eg) Si (small Eg) 1. High energy photons are absorbed 2. Low energy photons are absorbed 22 11

12 Shockley-Queisser limit Fundamental efficiency limit of a (single junction) solar cell Considerations Blackbody radiation (radiation loss, depends on solar cell temperature) Recombination (only radiative, recombination of free holes and electrons) Spectral loss (no absorption & thermalisation) Maximum : 33.7% for AM1.5 Wikipedia 23 Device operation 12

13 Electron-hole pair separation p-n diode: carriers collection by diffusion process crystalline solar cells Crystalline Si solar cell p-i-n structure carriers collection by drift amorphous solar cells photodetectors (crystalline or amorphous) a-si:h p-i-n solar cell Ag Ag SiO 2 n+ Glass SnO 2 p-type a-sic:h p++ diffusion - + p-type sc Si µm p++ Al - intrinsic a-si:h µm n-type a-si:h Al or Ag drift + M. Zeman, Delft University 25 Band diagram Crystalline Si solar cell a-si:h p-i-n solar cell n + p p + p i n E C Energy [ev] E V E F Energy [ev] E F E C E V Depth [μm] 0 Depth [nm] 300 M. Zeman, Delft University 26 13

14 What is a solar cell: First approximation In darkness: Solar cell = diode I V I R V qv I I0 exp( ) 1 kt Basic equivalent circuit I-V (current voltage) curve Ideal diode equation Under illumination = A diode + a current source in parallel with current I L I I L I D V qv I I L I D I L I0 exp( ) 1 kt I -I L V Power can be delivered to a load! 27 Solar cell I-V curve Analyses of I-V curve I -V max V oc V I L I D I Short-circuit Conditions I = -I sc, V=0 -I max P max I L I D I V Open-circuit Conditions I=0, V = V oc, -I sc Power P max I L I D I R M Ideal power dissipation on V R M =V max /I max 28 14

15 Equivalent circuit of a p-n solar cell I L I I D p V I R p s R s R R oc I dv di V oc Strictly applicable to p-n junction (c-si). Approximately applicable to thin-film solar cells The series resistance R s characterizes the resistance of the metallic connections the resistance of the metal-semiconductor interfaces (ohmic contacts); the resistance of the semiconducting layers. The parallel resistance (or shunt) R p characterizes: the leakage currents at the junction periphery the impurities in the junctions additional (with respect to the ideal diode) recombination. R R p Isc dv di V 0 R s series resistance (must be small) R p parallel resistance (must be high) I qv IR nkt s I L I0 exp( ) 1 V IRs R p 29 Current generation/dissipation Photovoltaics CD-ROM 30 15

16 Device performance limits Ideal p-n junction and junction formation p-type Free holes Excited acceptor states n-type Free electrons Excited donor states Depletion region = Space charge region Diffusion of holes Depletion region Diffusion of electrons In the depletion region : Diffusion currents = drift currents Ideal diode Total depletion Quasi neutral bulk region and space region with fully ionized donor/acceptor states Sharp interfaces n-region with n»p p-region with p»n Drift process negligible for minority carriers Very thin depletion region 32 16

17 Thermal equilibrium diagram E Drift p-type Diffusion qf i E c E F Diffusion Drift n-type E i E v x p x n x p-bulk region Depletion/spacecharge region n-bulk region In thermal equilibrium: drift current = diffusion current 33 Diode equation (Shockley) Ideal I(V) characteristics in the dark (diffusion model) Solving the transport, continuity and Poisson equations (for an ideal p-n diode): qv I dark I0 exp 1 kbt with where 2 2 qdnn qdpn i i I0 A LnnA LpnD k : the Boltzmann constant q: the elementary charge T: the absolute temperature A: the junction area n A, n D : the dopant densities n i : intrinsic carrier density D p, D n : diffusion constants of the minority carriers in the p, resp, n region L p, L n : diffusion length of the minority carriers in the p, resp, n region 34 17

18 Empirical expression The theoretical expression of the saturation current I 0 comprises parameters that have to be determined experimentally It can be replaced by a more simple empirical expression where the minimum saturation current I 0 is given by: E 8 g I exp 0 k T B ma in cm This empirical expression is an important basis for the calculation of the theoretical maximum of V oc, FF et When generation and recombination in the depletion region is not neglected: 2 I dark qv qv I 0 exp 1 I w exp 1 kbt 2kBT Additional term due to the recombination, which depends on the gap, carrier lifetimes, doping concentrations, applied voltage, etc. 35 Ideality factor We can replace I dark 0 w qv qv I exp 1 I exp 1 kbt 2kBT with I dark 0 qv I exp 1 nkbt where n is the ideality factor which varies between 1 and 2 : n =1 : ideal case n= 2 : recombination current is dominant. a-si p-i-n case: 1.3 < n <

19 Superposition principle "Superposition" principle of the photogenerated current I L I Dark I(V) = I dark (V) - I L I L -V max V oc V Illuminated Principle valid for a p-n diode in ideal conditions (in general not valid for thin film diodes) -I max Maximum power -I sc 37 Short-circuit current I sc Experimental data Assuming R s = 0 et R p = cell thickness sufficient for absorbing all useful light total collection of the carriers no recombination. Ideal illuminated diode in short-circuit condition I sc =I L Theoretical maximum of I sc given by : solar spectrum semiconductor gap: I sc = fq, with f: number of photogenerated "electronhole" pairs per unit time q: elementary charge Theoretical maximum short-circuit current under AM1.5 illumination Effect of the temperature: I sc only very weakly temperature dependent (E g decreases when the temperature increases)

20 Open-circuit voltage V oc The voltage supplied by a cell directly depends on gap E g From the empirical expression of the ideal p-n diode V oc k B T q ln 1 I L 1, ma/cm 2 E g q V oc is a quasi linear function of the gap Open-circuit voltage Voc as a function of the gap for a AM1.5 illumination with the approximations: (1) V oc 2 E g 3 q (2) V oc k BT q ln 1 I L 1, ma/cm 2 E g q where the values I L are obtained from the Fig. 7.26, (3) theoretical limit of Kiess [Kiess, 1995] Experimental data Limits 2 Eg E g /q is an upper bound for V oc : Voc 3 q V oc increases if I L increases advantage of cells working with light concentration (higher efficiencies). Temperature effect V oc decreases almost linearly when the temperature increases (0,4% per degree Celsius for c-si). 39 Fill-factor and cell power I Fill- factor FF defined as -V max V oc V ImV FF I V sc m oc -I max -I sc Maximum power P max =FF I sc V oc where I m, V m are the current and voltage when the cell output power is maximum (maximum power point) 40 20

21 Fill-factor FF FF is only a function of the opencircuit voltage V oc (in the diffusion model of the p-n diode). voc lnvoc 0.72V FF voc 1 V with oc voc nk T / q valid for v oc >10, n: ideality factor B Experimental data semi-empirical value of Fill-factor FF as a function of the gap for two values of the ideality factor n and for an AM1.5 illumination. FF is a very sensitive indication of the solar cell quality! For a quality solar cell: 0.7 < FF < 0.8 (a-si:h) 0.8 < FF < 0.85 (c-si) Temperature effect: FF decreases with an increase of the temperature (due to the change of V oc with temperature). 41 Efficiency The efficiency is defined as: P max P inc with P max = I m V m = FF I sc V oc The total solar cell efficiency is given by: p. / n s Experimental data maximum "theoretical" limit as a function of gap E g under AM1.5 illumination p/n being the efficiency of the carrier separation (second step) Voc p / n FF Eg / q and s the spectral conversion efficiency. Practically = I sc [ma/cm 2 ] V oc [V] FF / 100 [mw/cm 2 ] 42 21

22 Efficiency dependences rel Intensity : 100 mw/cm 2 100% 75% 50% Effect of temperature For c-si : decreases by %/ C For a-si : decreases by %/ C Decrease is less important for high bandgap materials. 25% T = 25 C 14% Si-Mono 12% Si-Poly 6% 0 0,5 1 1,5 T[ C] Sun 2 Effect of the light intensity decreases with a decrease of the light intensity (I L /I 0 ration of V oc, and loss in R p ). decreases at high values of the intensity (loss in R s, increase of the temperature, diffusion model no more valid for diode model under high injection). 43 Quantum efficiency or front dead layer or insufficiently thick diodes - + PVCDROM 44 22

23 Multi-junction devices Monolithic tandem n + 2,50 Map of max. efficiency (AM1.5) 0 2,25 0, , ,00 0,1298 p + n + E g bottom [ev] 1,75 1,50 1,25 0,1730 0,2163 0,2595 0,3028 0,3460 1,00 0,75 0,50 0,50 0,75 1,00 1,25 1,50 1,75 2,00 2,25 2,50 E g top [ev] Two terminal device Current matching required, narrower region of high efficiency compared to four terminal devices 46 23

24 Tandem equivalent circuit top cell bottom cell RL Top cell: I top = I top (V top ) Bottom cell: I bot = I bot (V bot ) Tandem cell: I = I top = I bot V(I) = V top (I) + V bot (I) I min (I sc,top, I sc,bot ) V oc = V oc,top + V oc,bot 16 Simple cell (single junction) 12 I [ma/cm 2 ] 8 Tandem cell V [V] typical I(V) curves of a single-junction and tandem a-si:h cells 47 PV technologies 24

25 PV technologies Bulk technologies: Monocrystalline silicon Polycrystalline silicon Ribbon silicon GaAs (space application, concentrators), InP (space application) Thin-film technologies: Amorphous silicon Micro-/nano-crystalline silicon. micromorph CdTe (Cadmium telluride) CIGS (Cupper Indium Gallium Diselenide) CIS (Cupper Indium Diselenide, Cupper Indium Sulfide) Others: Dye cells, Graetzel Organic cells, polymers 49 Conversion efficiency vs. time Commercial cells exhibit efficiencies reduced by 20-60% 50 25

26 PV Development Installed capacity (2013) PV installed: > 134 GW, 2013 : GW (fastest growth in Asia) World share of PV in electricity production : 0.85% EIA PVPS Report 51 c-si standard technology Screen printed solar cells Buried Contact Solar Cells PVCDROM 52 26

27 High efficiency c-si cells Sunpower back contact solar cells 24.2% efficiency UNSW PERL (passivated emitter with rear locally diffused) cell 25.0% efficiency 53 HIT cell Band diagram of the HIT cell: (a) ideal double heterojunction; (b) estimated HIT structure: a-si intrinsic layers (~ 5 nm) as super-passivation layers, as well as a-si p+ emitter layer (~ 10 nm) on the front and a-si n+ contact layer (~ 15 nm) on the rear. ITO transparent contact Best lab efficiency (Panasonic/Sanyo): 25.6% (100 cm 2 ), V oc > 700 mv Highest efficiency for c-si cells Thermal coefficient: %/ C for V oc, %/ C for (lower than that of c-si!) 54 27

28 Si thin-film technologies Thin-film silicon characteristics a-si:h Medium gap (1.75 ev) Short-range order High hydrogen content, alloy with H (10-20%) Low defect density (H passivation) Very high resistivity, semiisolating (>10 10 cm) Deposited mainly by PE-CVD Various substrates possible Large area deposition µc-si:h (for PV applications) Low gap material (1.1 ev) few nm grain size (nano-crystalline silicon) Short-medium range order Medium hydrogen content (few %) Low defect density (H passivation) Presence of a-si:h incubation layer, a-si:h inter-grain tissue High resistivity ( cm) Deposited mainly by PE-CVD Various substrates possible Large area deposition 56 28

29 a-si:h optical absorption Absorption coefficient [1/cm] Wavelength [micrometers] Subbandgap defects a-sige:h a-si:h p-type a-sic:h c Si Photon energy [ev] Absorption coefficient «non-direct semiconductor» In visible part of spectrum 70 x higher than c-si Thin film (about 1 micron) absorbs 90% of usable solar energy Optical band gap Tunable Depends on H content a-si:h based materials alloys C increases E g Ge decreases E g 57 p-i-n solar cell 100 Å µm 200 Å Anti-reflection coating Glass superstrate Transparent Conductive Oxide p-type Intrinsic a-si:h n-type Back contact - + p-i-n structure necessary for disordered semiconductors Doped layer are very defective (dead layers) Doped layers only serve for electric field build-up through the intrinsic layer The intrinsic layer is the only active layer. Light enters the cell through the p-layer ( window layer ) Minimization of hole collection path (low hole mobility) Higher bandgap for p-layer (low optical loss) Electrically thin diode High electrical field minimal recombination in i-layer Optically thick diode maximum absorption of light Minimal thickness (processing time) Light trapping scheme 58 29

30 Light trapping for thin film silicon cells Glass Front ZnO p-i-n ZnO Back reflector Rough transparent conductive oxides (TCO) Back reflector Tin Oxide (SnO 2 ) deposited by AP-CVD Roughness depends on deposition process Zinc Oxide (ZnO) deposited by LP-CVD Roughness depends on film thickness and temperature 59 Multi-junction device for higher efficiency Tandems a-si:h / a-sige:h a-si:h / mc-si:h (mircomorph) a-si:h / a-si:h Triples a-sic:h / a-si:h / a-sige:h a-si:h / a-sige:h / a-sige:h a-si:h / mc-si:h / mc-si:h a-si:h / a-si:h / mc-si:h a-si:h / a-sige:h / mc-si:h a-si:h / mc-si:h / mc-ge:h Issues a-sige:h stability mc-si:h processing a-sic:h & a-sige:h stability a-sige:h stability Current matching (thick bottom) Current matching, very thin top a-sige:h stability mc-ge:h processing (commercially available devices in red) 60 30

31 Examples Back contact µc-si:h Bottom cell (1.1 ev gap) a-si:h Top cell (1.75 ev gap) TCO Glass Micromorph Tandem devices comprising Introduced by IMT (EPFL) Advantages: Better usage of solar spectrum Higher efficiency and stability Intensity [kw/m 2 /µm] µc-si:h a-si:h Wavelength [nm] Light UniSolar monolithic triple junction, flexible modules on steel: -stable cell efficiency: 13.1% -module efficiency ~8% ~1.8 ev ~1.6 ev ~1.4 ev 61 PV modules 31

32 Module serial connection Bulk technologies V 0.6 V 0.6 V 0.6 V Thin-film technologies + - Metall p-i-n TCO Substrate (glass, polymer or isolated metal) Monolithic interconnection for thin-film cells 63 Monolithic interconnection Deposition of TCO on glass Deposition of back contact Laser scribing of TCO Laser scribing of back contact Deposition of a-si diode Testing of finished module Laser scribing of a-si:h diode Active area Dead area 64 32

33 PV with thin film Si 65 Microrobots Solar modules for micro-systems High voltage modules Up to 180 V from 3x3 mm 2 Output voltage [V] mv/segment First module generation Second module generation Number of segments 66 33

34 a-si photodiodes Thin-film Solar cells vs Photodiodes hn p i n A Using primary photoconductivity (device with blocking contacts) given by the collection of photo-generated carriers I ph =Ge with : quantum efficiency, 1 (a-si:h cells) blocking contact: diode characteristic Operation conditions for a photodiode Current density dark illuminated I ph Operation conditions for a solar cell Voltage 68 34

35 Fluorescence detector, microfluidic A Schematic crosssectional view of a microlens and an integrated a-si:h fluorescence detector, forming a compact platform where microfluidic CE device is mounted. B Optical micrograph of the top view of the integrated fluorescence detector. Kamei, MRS Vertical Integration of Detector Arrays Front electrode a-si:h photodiode layer Back electrode Insulation layer CMOS circuit Technology known as Thin-Film on ASIC (TFA) Thin-Film on CMOS (TFC) Above IC technology Elevated diode Benefits High fill-factor, high sensitivity No trade off between sensitivity and circuit complexity No dead area 70 35

36 TFA imagers Visible light X-ray detection Metal-i-p a-si:h diodes deposited on unpassivated chip 64x64 pixels 38.4 µm size, 40 µm pitch 92% fill factor Pixels connect to charge integrators Dark current as low as 20 pa/cm2 Noise (fixed patterned noise Sensitivity 56 V/(µJcm2) Dynamic Range: >60 db 4x4 mm 2 ASIC 48 pixels with 150 µm size and 380 µm pitch connected to active feedback amplifiers 15 µm thick a-si:h diode N. 71 PV ressources

37 Bibliography Books S.M. Sze, Semiconductor Devices Solar Cells: Operating Principles, Technology and System Applications, M. Green, University of New South Wales Silicon Solar Cells: Advanced Principles and Practice, M.A. Green, University of New South Wales A. Shah, Thin-film silicon solar cells, EPFL Press, J. Kanicki, Amorphous and Microcrystalline Semiconductor Devices, 1991, Artech House R.A. Street, Hydrogenated Amorphous Silicon, Cambridge University Press, A. Ricaud, Photopiles solaires, PPUR, 1997 A. Goetzberger, Sonnenergie: Photovoltaik, Teubner, Other Photovoltaics: Devices, Systems and Applications CD-ROM by C. Honsberg and S. Bowden, University of New South Wales, 2 nd edition online at :