Giovanni Bruno Istituto di Metodologie Inorganiche e dei Plasmi, IMIP-CNR, via Orabona 4, 70126 Bari, Italy
Energy Landscape Source: Scientific American
The green energy Everything that is old is new again Warwick Castle, Feb.14, 2010
Photovoltaic Energy: the Queen of the Green Energies I am in favor of any renewable energy product..but, I don't want "green" energy to mean more "green" from my wallet Who wants to pay more for green electricity?
Renewable energy in Italy Breakdown of the contribution of renewables to the national energy balance Wind Power 6.2% Geothermal 8.6% Hydroelectric 59.3% Photovoltaic 0.1% Solar Thermal 0.4% Biomass and waste 10.7% Biofuels 2.5% Other renewables 12.3%
Energy Poduction in Italy
World PV Producers The major producers of the world PV market 2006 (Source: PV News, April 2007)
Semiconductors materials for PV Solar cell materials and solar radiation spectrum (source/copyright:hahn-meitner-institut Berlin
PV-materials Market Market share of solar cell types sold during 2006
Si-based PV market Incidenza della scarsa disponibilità di silicio sul prezzo di mercato del fotovoltaico (Fonte: Solarbuzz, 2007)
Last News on Si-based PV market
Generation costs of PV electricity: Economic Perspective 1400 KWh/(KW.year) 3 KW (21.000 ) 4200 KWh/year 20 year (deprec.) 0.25 /KWh 1,2 /KWh 1,0 900 h/y 0.4 /KWh 0,8 0,6 0,4 0,2 1800 h/y 0.2 /KWh 0,0 1990 2000 2010 2020 2030 2040 [EPIA: Towards an Effective Industrial Policy for PV (RWE Schott Solar)
The key metric for PV : the cost per watt Photovoltaics is emerging as an important technology for the future of energy production. To fully realize this potential it will be necessary to reduce the manufacturing costs of photovoltaic cells while maintaining high working efficiencies Solar cell Materials and Processing Costs ($/m 2 ) $/W cost = Cell efficiency Manufact. Yield Insolation (W/m 2 ) The main routes in solar cell research target both an increase in cell efficiency and manufacturing cost reduction. One of the key elements in increasing cell efficiency is high photon absorption and its conversion into a maximum number of electron-hole pair With Si representing about 40 % of the cost/wp at module level, the amount of Si used per Wp has to be reduced drastically
New Concept for Next Generation Solar Cells: Why? Module cost ($/W) 100 10 1 1 1980 Benefits of scaling and standardization 1990 Historical Projected 2000 2010 Known developments 2015 10 100 1,000 10,000 100,000 Source: Dr.Richard Swanson SunPower, and SVC Cumulative production (MW) Needs revolutionary technology
EVOLUZIONE del FOTOVOLTAICO a base di SILICIO Si-based Semiconductor Materials Si 3 N 4 SiO 2 µc-si 1 3 a a Generazione: 2 a Generazione: Silicio cristallino nanostrutture sottili film a-si film sottili nc-si a-si:n Sputtering a-si: Si:Er a-si:c a-si: Si:Ge
I, II and III Generation Photovoltaics using silicon Cella PV-Plasmonica 100 0.10 US$/W 0.20 US$/W 0.50 US$/W a-si Area cella 0.8 cm 2 Silicio amorfo 1mt x 1mt EFFICIENCY (%) 80 60 40 20 nc-si np-ga II III 0 0 100 200 300 400 500 COST (US$/m 2 ) µc-si su PET np-au µc-si su PI I Thermodynamic Limit 2.50 US$/W Present limit 1.00 3.5 US$/W c-si Silicio cristallino Diametro 10cm
The first revolution in PV: from bulk c-silicon to a-si:h thin film 100 0.10 US$/W 0.20 US$/W 0.50 US$/W 80 Thermodynamic Limit EFFICIENCY (%) 60 40 20 0 III I II 0 100 200 300 400 500 COST (US$/m 2 ) 1.00 US$/W Present limit 3.50 US$/W
The first revolution in PV: from bulk c-silicon to a-si:h thin film In a landmark paper published in 1975, Spear and LeComber reported that plasma glow-discharge deposition of silicon, using mixtures of silane and either phosphine or diborane, enabled electrically active pentavalent (P) and trivalent (B) impurities to be incorporated into the films, making them n-type and p-type, respectively.. SiH 4 Si + 2H 2
STATE-of-the-ART: BEST RESEARCH-CELL EFFICIENCY EFFICIENCY (%) Source : www.nrel.gov/ncpv/thin-film/docs/kaz_best_research_cells.ppt Multi-junction Concentrators Three-junction Two-junction 40 35 30 25 20 15 10 5 Crystalline Si Cells Single crystal Multicrystalline Thin Si Thin Film Technologies Cu(In,Ga)Se2 CdTe a-si/a-ge (stabilize) nano-,micro-,poly-si Emerging PV Dye cells Organic Cells RCA ARCO (various technologies) Westing house North Caroline State Univ Solarex Kodak Boeing Matsushita Monosolar Univ of Maine RCA Boeing RCA RCA Boeing RCA RCA Spire Solarex Solarex Stanford ARCO AstroPower Varian AMETEK Spire Boeing UNSW Georgia Tech University So.Florida Photon Energy Lausanne Univ. UNSW RCA Kodak UC Berkley 0 1975 1980 1985 1990 1995 2000 2005 2010 YEAR Lausanne Univ. Kaneka Boeing Sharp United Solar UNSW Euro-CIS Japan Energy UCSB UNSW Georgia Tech AstroPower UNSW UNSW Cambridge Spectrolab / Spectrolab Spectrolab Cu(In,Ga)Se 2 United Solar Linz Univ Stuttgart Univ Princeton Siemens Spectrolab ECN The Nederlands / Spectrolab, Kaneka Univ Linz Photonics
STATE-of-the-ART: BEST RESEARCH-CELL EFFICIENCY EFFICIENCY (%) 40 35 30 25 20 15 10 5 Multi-junction Concentrators Three-junction Two-junction Crystalline Si Cells Single crystal Multicrystalline Thin Si Thin Film Technologies Cu(In,Ga)Se2 CdTe a-si/a-ge (stabilize) nano-,micro-,poly-si Emerging PV Dye cells Organic Cells ARCO (various technologies) Westing house North Caroline State Univ Solarex Kodak Boeing Matsushita Monosolar RCA Univ of Maine RCA Boeing RCA RCA RCA Boeing RCA RCA Spire Solarex Solarex Stanford ARCO AstroPower Varian AMETEK Spire Boeing UNSW Georgia Tech University So.Florida Photon Energy Lausanne Univ. UNSW Sanyo Kodak UC Berkley 0 1975 1980 1985 1990 1995 2000 2005 2010 Sharp United Solar YEAR UNSW Euro-CIS Boeing Japan Energy Lausanne Univ. UNSW Georgia Tech UCSB Kaneka UNSW UNSW AstroPower Cambridge Spectrolab / Spectrolab Spectrolab Stuttgart Univ Cu(In,Ga)Se 2 United Solar Princeton Siemens Linz Univ ECN The Nederlands Spectrolab / Spectrolab, Kaneka Univ Linz Photonics May 2009, 23% Evolution of Sanyo ACJ-HITTM solar cells efficiency on area of~100 cm2)(sanyo Courtesy [M. Tanaka, S. Okamoto, S. Tsuge, S. Kiyama, Proc. 3rd World Conf. on Photovoltaic Energy Conversion, Osaka, Japan, May 2003])
Why Silicon-based heterojuction solar cells (HIT)? by thermal diffusion @ 900 C by plasma deposition @ 200 C
Efficiency (%) Why Silicon-based heterojuction solar cells (HIT)? Excellent feature of temperature dependence: More output power even at high temperatures in summertime 20 19 18 17 16 15 14 1.6m x 0.85m =1.36m 2 240 Watt Eff.=17.6%(module) Eff. = 20.0%(cell) c-si -0.090% / C -0.045% / C HIT Current (Ampere) 13 10 20 30 40 50 60 70 Temperature ( C) 8 7 6 5 4 3 2 1 15.3% 16.3% 17.6% 18.9% 0 0 10 20 30 40 50 Voltage (Volts) AM1.5, 1 KW/m 2 75 C 50 C 25 C 0 C
How to Improve Efficiency and Reduce Costs of Bulk-Si PV Present technologies for improving efficiency and reducing the cost of solar cells Efficiency improvement technology Cost reduction technology High light trapping -Uniformly textured Surface -Double anti-reflective coating (ARC) -No or less shadow of contact Low recombination rate -Good-quality substrate -Passivation -Heterojunction Low series resistance -Thicker electrode -Fine contact pattern Takashi Tomita, SHARP Cor., Prog. Photovolt: Res. Appl. 2005; 13:471 479 -Thin substrate - Large cell size - Simple process - Non-vacuum technology - Low-temperature process - Making substrate directly from molten silicon
Thickness reduction implies efficiency reduction: how to solve this? 21 Efficiency (%) 20 19 18 17 16 15 14 100 200 300 Wafer thickness (µm) 400 300 200 100 Source: Dr.Richard Swanson SunPower, and SVC One of the most important considerations is the reduction in wafer thickness Mono-crystalline Multi-crystalline 2002 2004 2006 2008 2010 2012 50 µm target EC call
Heterojunction Solar Cell: The Sanyo record Light Trapping Si-wafer thickness=98µm
Next Generation of PV Initiatives Source: DOE Solar Energy Technologies Program, 2007)
Plasmonics Localized surface plasmon resonance (metal nanoparticles) ω lsp = ω p 1+ 2ε medium Collective oscillation of the conduction electrons free electron absorption Au bulk Surface plasmon resonance peak Au NPs [A. J. Haes et al. MRS Bullettin,]
ANCIENT ORIGIN OF PLASMON Licurgus Cup (glass, British Museum; 4th century AD) When illuminated from within, the Lycurgus cup glows red. The red color is due to gold nanoparticles (40nm) embedded in the glass, which have an absorption peak at around 520 nm
Plasmonic nanoparticles & p-n junction 1 Light can be converted to electricity via plasmon resonances in nanoparticles, by: far field effect which prolongs the optical path through the cell 2 a near field effect which locally enhance the energy conversion in the solar cell 3 a creation of energy rich charge carriers which are transferred to the solar cell. + ++ - -- h e + ++ - -- h e + ++ h e - --
INCREASED ABSORPTION by LOCALIZED SURFACE PLASMON RESONANCE Silicon does not absorb photons with energy below the bandgap of silicon, hence such photons cannot contribute to power generation Metal NPs exhibit pronounced absorption and scattering effects at optical frequencies due to the existence of collective electron excitations known as localized surface plasmon resonance. UV IR 5 4 c-si 500nmx500nm n-type a-si p-type Si 3 k 2 1 0 a-si:h 500 1.000 Wavelength (nm) 1.500
OUR SOLAR CELLS GEOMETRY Cell area = 1x1 cmxcm Active area = 0.8 cm 2 Number of fingers = 32 Finger interspace = 500µm Finger width= 100 µm Finger length = 4.85 mm Bus bar length = 10 mm Ag grid contact Au nanoparticles n-doped a(µc)-si Al-Ag contact light p-doped crystalline silicon 1 cm Maria Losurdo, Maria M. Giangregorio, Giuseppe V. Bianco, Alberto Sacchetti, Pio Capezzuto, Giovanni Bruno, Enhanced absorption in Au Nanoparticles/a-Si:H/c-Si heterojunction solar cells exploiting surface plasmon resonance, Solar Energy Materials and Solar Cells, 93, 1749-1754 (2009) ADVANTAGES Conventional methods of texturing are difficult to incorporate on the thin layer, and tend to degrade the quality of the semiconductor layer. Our approach: Involves very simple processing techniques and can be formed as the last stage of the device processing so there are no processing incompatibilities. The metal nanoparticles which are separated from the silicon by a thin spacer layer will not degrade the electrical characteristics of the device. Can be easily incorporated into other types of semiconductor substrates
OUR APPROACH and PROCESS 1) Native oxide removal by HF(1%) dipping 2) Loading into reactor; SiF 4 -plasma in situ cleaning and passivation 3) PECVD: SiF 4 -H 2 -PH 3 plasma deposition of n-type a-si:h 4) Sputtering of Au target switching to an Ar Plasma 5) Top contact grid evaporation Au Plasma
IN-LINE PROCESS MONITORING by SPECTROSCOPIC ELLIPSOMETRY EVOLUTION of the PLASMON PEAK Polarizer Photoelastic Modulator Analyzer Optical Fiber Detector Xe-lamp Monochromator Shutter 500nmx500nm <ε 2 > Ä_i 20 15 10 5 1 SPR Au NPs 3MLs Au 12 ML Au 18 ML Au a-si/c-si 2 3 4 5 Photon Energy (ev) 6 <ε>=<ε 1 >+i<ε 2 >= N 2 =(n+ik) 2 true extinction coefficient
TUNING Au PLASMON RESONANCE The XRD shows peaks at (2θ) 38.2 and 44.5 of (111) and (200) crystalline planes of Au and, therefore, the structure of Au NPs appears to be face centred cubic (111) 24nm Intensity (A.U.) (100) 2 20nm 36 38 40 42 44 46 48 50 2 θ (degrees) k 1 18nm 67nm 15nm 60nm 52nm 500 1,000 Wavelength (nm) 1,500
I-V CHARACTERISTICS of HETEROJUNCTIONS CURRENT (ma/cm 2 ) 35 30 25 20 15 10 5 (n)a-si:h/(p)c-si 50nm 15nm 12nm 30nm 20nm Au NPs 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 VOLTAGE (volts) Engineering Au nanoparticles diameter and density makes possible improvements in conversion efficiency J sc =33mA/cm 2 V oc =0.58V FF=0.65 η=12.5% J sc =23mA/cm 2 V oc =0.61V FF=0.68 η=9.5% J sc =22mA/cm 2 V oc =0.54V FF=0.64 η=8.9% J sc =19mA/cm 2 V oc =0.6V FF=0.65 η=8% J sc =21mA/cm 2 V oc =0.6V FF=0.45 η=6% J sc =7.9mA/cm 2 V oc =0.48V FF=0.47 η=2% 500nmx500nm HEIGHT (nm) 20nm 8 7 6 5 4 3 2 1 0 0.0 0.1 0.2 0.3 0.4 0.5 DISTANCE (µm) 15nm 12nm No Au NPs 30nm 50nm
Spettro Solare conversione con cella solare di silicio convenzionale plasmonica con nanoparticelle <50nm con nanoparticelle >100nm Converted
DEPENDENCE of ENHANCEMENT EFFICIENCY RATIO on Au NPs 1.6 Au MLs 8 15 18 20 25 60 (continuous film) η Au NPs /η a-si/c-si ENHANCEMENT RATIO 1.4 1.2 1.0 0.8 0.6 0.4 0.2 10 15 20 25 30 35 40 45 50 55 Au NPs DIAMETER (nm) both NPs size and density affect efficiency improvement importance of controlling the amount of Au Failure of device at large Au MLs is due to the inter-diffusion of Au and Au-Silicides formation (proven by XPS)
PHOTOCURRENT RESPONSE 20nm Au NPs/ (n) a-si:h/ (p) c-si (n) a-si:h/(p) c-si The presence of Au NPs on the cell yields an additional photocurrent collected Towards the yellow-green region above 2eV. In this spectral region the Au NPs have their plasmonic absorption.
OTHER METALS? [Appl. Phys. Lett. 73, (26) 1998)] Limit?: for Nps SPR ranging between 300-530nm H. R. Stuart and D. G. Hall, APL 69, 2327, 1996 S. Pillai et al, APL 88, 161102, 2006
CONCLUSIONS Solar Cell Configuration: Coupling Au metal NPs localized surface plasmon resonance to a- Si/c-Si heterojunction shows significant potential for enhancing solar cell absorption Process: Large area plasmonic solar cells could be fabricated in a potentially cost effective by combining PECVD of Si-based thin films with metal nanoparticle sputtering. Cost Reduction: Low cost plasmonic metal enhanced solar cells represent a viable solution to lower cost of photovoltaics
ACKNOWLEDGEMENTS The Pla.S.Ma. group (Plasmas for Semiconductor Materials Science) The 7 th FP European Projects NanoCharM (NMP3-CA-2007-218570) Multifunctional NanoMaterials Characterization exploiting Ellipsometry & Polarimetry NIM_NIL (NMP-2008-SMALL-2) Large Area Fabrication of 3D Negative Index Metamaterials by Nanoimprint Lithography