Cost Drivers in the Photovoltaic Solar Industry
An Analysis of the Potential for Reducing Costs
Marcus Ejder Royal Institute of Technology
Richard T. Carlsen MG104X: Bachelor Assignment
Industrial Engineering and Management Councilor: Jan-Olof Svebeus
Report Summary
The demand for energy is increasing at an incredibly fast rate globally. Electrical energy, supplied through interconnected grids is a major constituent of this demand.
The electricity market, today, however, finds itself in a state of flux. Rising costs for conventional non-renewables accompanied with a growing awareness for the environment and the detrimental effects of our reliance on fossil fuels is leading to a paradigm shift in energy policy for governments, businesses and the public alike. There is now a growing desire to make clean, sustainable and renewable energy sources a larger part of our generation capacity. Solar photovoltaic (PV) power, through crystalline silicon (c-Si) technology, fulfills this list of criteria. In order to make solar PV a larger part of our electricity generation stack, however, it needs to be made more affordable and price competitive. There is therefore a strong incentive for the PV industry to reduce its costs.
This report seeks to identify the major cost drivers for the c-Si solar photovoltaic industry and to determine potential avenues for future cost reductions. An analysis of the conceivable magnitude of these cost reductions is also undertaken for the short term (2010-2015). The fundamental cost structure of a c-Si photovoltaic system is divided into the solar module (actual solar cell technology) and the so-called balance of systems (BOS) components, which are the additional electrical components and support structures required. The costs are reduced primarily through technological innovation, such as increasing solar cell efficiency, economies of scale benefits and through avenues that optimize productions processes. Increases in solar efficiency are found to represent the largest individual benefactor on costs, as it increases the amount of energy generated electricity for the same size solar cell. Segmenting the cost drivers into the various phases of the production value chain is useful. The following chart represents forecasted cost reductions for each of the five production phases (polysilicon refining, ingot/wafer processing, cell manufacture, module assembly and balance of systems installation). The total cost for a PV system is forecast to move from $3.65/W in 2010 to $2.94/W in 2015. This reflects a cost reduction of 19.5% in five years. Furthermore, it can be deduced from this chart that the largest cost cuts will stem from reductions in the module cost (77.5%), whereas only 22.5% of total cost reductions comes from the BOS component.
Forecasted Cost Progression of a c-‐Si PV solar System 2010-‐2015, $/W
Year 2010 2011E 2012E 2013E 2014E 2015E % of Total Cost Reduction
Polysilicon $0.43 $0.38 $0.29 $0.23 $0.20 $0.18 35.2%
Ingot/Wafer $0.46 $0.43 $0.41 $0.39 $0.36 $0.35 15.5%
Cell $0.36 $0.33 $0.31 $0.30 $0.28 $0.26 14.1%
Module $0.50 $0.48 $0.46 $0.44 $0.42 $0.41 12.7%
Total Module $1.75 $1.62 $1.47 $1.36 $1.26 $1.20 77.5%
Balance of
Systems $1.90 $1.87 $1.84 $1.80 $1.77 $1.74 22.5%
Total Cost $3.65 $3.49 $3.31 $3.16 $3.03 $2.94 100%
Sammanfattning
Den globala efterfrågan för energi ökar i en snabb takt. Elektricitet, av den typ som levereras via vårt elnät, är en viktig komponent av denna efterfrågan. Elmarknaden befinner sig dock i ett utdraget förändringstillstånd idag. Ökade kostnader för konventionella, oförnybara energikällor i kombination med växande medvetenhet om negativa miljöeffekter leder mot ett paradigmskifte i energipolitiken för regeringar, företag och allmänhet att följa. Det växer fram en önskan om att göra rena och förnybara energikällor till en större del av vårt samhälle. Solceller, via den kristallina kiselteknologin (c-Si), uppfyller samtliga dessa kriterier. För att göra solceller till en större del av vår elproduktion måste dessa dock bli mer prismässigt konkurrenskraftiga.
Det finns därför ett starkt incitament för industrin att minska sina kostnader och ta denna chans.
Denna rapport identifierar de största kostnadsdrivarna för den kristallina kisel- teknologin och söker vägar för framtida kostnadsminskningar inom dessa. En analys av kostnadsminskningarna presenteras även för nästkommande fem år (2010-2015), som uppdelas efter den grundläggande kostnadsstrukturen i teknologin. Denna är delvis modulkostnaden, vilket är kostnaden för att tillverka själva solcellen, och arbetet efter tillverkningen (såsom arbete, underhåll, ytterligare elektriska komponenter osv).
Kostnaderna reduceras här främst genom tekniska innovationer, som att öka effektiviteten i solcellen, stordriftsfördelar och genom att optimera produktionsprocesserna. Ökningar i effektivitet konstaterades här vara den största enskilda kostnadsdrivaren, då den ökar mängden genererad elektricitet för en viss bestämd cellstorlek.
Följande tabell representerar den prognostiserade kostnadsminskningen för fem olika produktionsfaser (råmaterial, ingot/wafer, cell, modul och efterarbete). Den totala kostnaden för ett solcellssystem förväntas minska från $3.65/W under 2010 till $2.94/W år 2015. Detta speglar en kostnadsminskning på 19.5% under fem år. Vidare kan man utläsa ur tabellen att den största kostnadsnedskärningen blir resultatet av minskningar i modulkostnaden, dvs. 77.5%, medans endast 22.5% kommer från efterarbetet.
Beräknad kostnad för en c-‐Si-‐solcell år 2010-‐2015, $/W
År 2010 2011E 2012E 2013E 2014E 2015E % av total kostnadsreduktion
Kisel $0.43 $0.38 $0.29 $0.23 $0.20 $0.18 35.2%
Ingot/Wafer $0.46 $0.43 $0.41 $0.39 $0.36 $0.35 15.5%
Cell $0.36 $0.33 $0.31 $0.30 $0.28 $0.26 14.1%
Modul $0.50 $0.48 $0.46 $0.44 $0.42 $0.41 12.7%
Total Modul $1.75 $1.62 $1.47 $1.36 $1.26 $1.20 77.5%
Efterarbete $1.90 $1.87 $1.84 $1.80 $1.77 $1.74 22.5%
Total Kostnad $3.65 $3.49 $3.31 $3.16 $3.03 $2.94 100%
Table of Contents
Ch. 1: Introduction and Principal Inquiry ... 1
1.1: Introduction ... 1
1.2: Principal inquiry ... 1
1.3: Report Layout ... 2
1.4: Report Methodology ... 2
1.5: Report Uncertainty ... 3
Ch. 2: Introduction to the Photovoltaics Market ... 4
2.1: The Solar Resource ... 4
2.2: Photovoltaic (PV) Electricity ... 5
2.2.1: The Photovoltaic Effect ... 5
2.2.2: Silicon ... 5
2.2.3: Semiconductor Electronics and Doping ... 5
2.2.4: The p-n-junction ... 6
2.2.5: Photovoltaic Electric Current ... 7
2.3: The Photovoltaic System and Production ... 8
2.3.1: The PV Module Production Value Chain ... 8
2.3.2: Balance of Systems (BOS) ... 11
2.4: Photovoltaics and the Market for Electricity ... 11
2.4.1: The Electricity Grid ... 11
2.4.2: The Generation Stack and Solar PV ... 12
2.4.3: Political Incentives and the Feed-In-Tariff (FIT) ... 13
2.5: The Benefits of Solar PV ... 13
Ch.3: Reducing Costs for PV systems ... 15
3.1: Introduction: Cost Structure and Trends ... 15
3.2: Technological Innovation: Cost Reductions ... 17
3.2.1: Efficiency ... 17
3.2.2: Quantum Efficiency ... 18
3.2.3: Optical Efficiency ... 19
3.2.4: Recombination Efficiency ... 20
3.2.5: Resistance Efficiency ... 20
3.2.6: Efficiency and Costs ... 20
3.3: Economies of Scale: Cost Reductions ... 21
3.4: The PV Value Chain: Total Cost Reductions ... 21
3.4.1: Polysilicon Costs ... 21
3.4.2: Ingot/Wafer Processing Costs ... 25
3.4.3: Cell Manufacturing Costs ... 27
3.4.4: Module Assembly Costs ... 28
3.4.5: Balance of Systems (BOS) Costs ... 28
Ch.4: Conclusion ... 29
4.1: Results ... 29
4.2: Looking Forward ... 30
4.3: Limitations, Uncertainty and Further Study ... 31
Reference Page ... 32
Attachment 1 - Efficiency and Costs Calculations ... 37
Attachment 2 - Polysilicon Calculations ... 38
Attachment 3 - Ingot/Wafer Processing Calculations ... 40
Attachment 4 - Cell Manufacturing Calculations ... 41
Attachment 5 - Module Assembly Calculations ... 42
Attachment 6 - Final Summation of Costs for entire PV System ... 43
Ch. 1: Introduction and Principal Inquiry 1.1: Introduction
In physics, energy is defined as “the property of matter and radiation, which is manifest as a capacity to perform work, such as causing motion or the interaction of molecules”1. Energy exists in a myriad of forms and our ability to harness and employ it has determined our capacity for survival since the beginning of time. Electrical energy, the existence and flow of electrical charge, has risen in the past century to become an integral part of our lives, its applications ranging from lighting and heating our houses to the rise of telecommunications and information technology. The global electricity market is expected to supply a breathtaking 20 000 terawatt-hours of power and represent a $2 650 billion industry in 20122.
In the industrialized world today, the overwhelming majority of the population is provided with electricity through centralized electricity grids. These centralized grids transmit electricity from an array of power sources, through high-voltage electricity cables, to our households and workplaces. Historically, fossil fuels, such as coal, complemented by nuclear power, have constituted the bulk of total electricity generation capacity (currently comprising approximately 70%3) due to an unchallengeable cost advantage. Today, however, rising costs for conventional non-renewables accompanied with a growing awareness for the environment and the detrimental effects of our reliance on fossil fuels is leading to a paradigm shift in energy policy for governments, businesses and the public alike.
There is now a growing desire to make clean, sustainable and renewable energy sources a larger part of our generation capacity. Galvanized by political support, enormous progress has been made to narrow the cost gap between the renewable sources and fossil fuels, through technological innovation, process improvements and scale. One of the most promising renewables candidates is solar photovoltaic (PV) energy, which is now close on the heels, in terms of price competitiveness, with the conventional energy sources. Crystalline silicon (c-Si) solar cells are, and always have been, the predominant technology within the solar photovoltaic industry, currently constituting 80% of global production capacity4.
1.2: Principal inquiry
This report, written under the Department of Integrated Production at the Royal Institute of Technology, seeks to identify the major cost drivers for the c-Si solar photovoltaic industry and to determine potential avenues for future cost reductions. An analysis of the conceivable magnitude of these cost reductions will also be undertaken for the short term. This inquiry will primarily be addressed by looking at the potential for cost reductions through technological innovation, process optimization and economies of scale.
1 Oxford Dictionaries (2011)
2 Economy Watch (2011)
3 Mercom Capital Group (2011) Mercom Solar Intelligence Report 2011-03-28
4 Lynn, P. (2010)
1.3: Report Layout
The report begins in Chapter 2 with an introduction of the fundamentals behind solar cell technology and the photovoltaic (PV) effect. This is followed by a walkthrough of the various phases of the PV production chain. The introductory chapter is concluded with an explanation of the competitive landscape and the cost reduction incentives for the industry.
In chapter 3, the various cost drivers for the industry are identified and the prospective avenues for future reductions are asserted. Technological innovation, taken to be synonymous with solar cell efficiency improvements, is discussed first by itself.
Subsequently, the general impact of economies of scale on costs is introduced. Next, in Chapter 3.5, the cost drivers and potential cost cuts for each phase in the production value chain are examined. The report is culminated in chapter 4 with the results of our study, a short discussion of the impact and context of these results and finally a review of some of the limitations and constraints of the report.
1.4: Report Methodology
This report has been written based on data and information from a variety of literary sources, industry reports and relevant websites, as well as from interviews with knowledgeable employees of c-Si solar PV firms5.
A forecast of the magnitude of cost cuts for each production phase in the value chain in the period 2010-2015 has been conducted. These forecasts include the total impact of all the introduced cost reduction measures (efficiency gains and economies of scale), in addition to process enhancements, which are introduced phase for phase. Certain cost reduction fields, such as system lifetime extension and improvements in industry standards and specifications have not been taken into consideration in this report.
Additionally, maintenance costs have been taken to be negligible due to their minor impact on the total cost picture.
The reason for discussing certain cost drivers individually first, and then including them in the total analysis in conjunction with process optimization later is to clearly illustrate the various drivers and their individual impact on costs. This methodology has been chosen due to the difficulty in obtaining cost data describing the impact of individual factors. As a conclusion, the forecast for each phase will be summated to arrive at a final cost figure ($/Wp) for 2015. Conjointly with this analysis, a case study, considering a 150MW PV facility will also be subjected to the forecasted cost cuts to elucidate the effect of the cost reductions in real monetary terms. This size facility (150 MW) has been chosen simply because it is a very typical facility size and will be used strictly as an example.
The relatively short time frame (2010-2015) has been chosen due to the immense pace of change in the young solar PV industry. In order to accurately reflect industry-wide accepted developments and their subsequent impact on costs, therefore, this five-year time frame has been chosen. Also, the availability of accurate forecasting data is increasingly difficult to come by beyond 2015. A longer time frame would thus significantly increase the uncertainty margin of our results.
5 See Reference List
The forecasts made in this report will take a conservative stance and omit the impact of certain, more speculative improvements. Analysis is based on the so-called “business as usual” improvements. Additionally, the data employed in the numerical analysis will be weighted averages taken over a wider portion of the total market, as opposed to individual corporate forecasts and predictions.
1.5: Report Uncertainty
At the outset, it is important to remark that the PV market today is a considerably fragmented and heterogeneous market, in terms of technology, production and applications. Massive funding is, and will continue to be, attributed to research and development in the entire PV value chain. In addition, external factors, such as development of alternative energy technologies (wind, biomass, geothermal, nuclear and fusion) may have a tremendous effect on the competitive landscape for solar PV.
The natural corollary, therefore, is a substantially large amount of uncertainty when prognosticating the future of the energy market, and of the photovoltaics industry.
Recent trends have shown a remarkable growth for the photovoltaics market, making it the fastest growing segment of the energy market (growth of more than 100% in 20106) and according to the Mercom Capital Group, “while we can argue when solar will become cost competitive with retail electricity, at the end of the day, solar is on a declining cost curve and the price of electricity is headed higher”7. This report aims to evaluate the nature and capacity for this transition.
6 Interview: Bjørseth, Alf
7 Mercom Capital Group (2011) Mercom Solar Intelligence Report 2011-03-28
Ch. 2: Introduction to the Photovoltaics Market 2.1: The Solar Resource
The sun transmits an inconceivable amount of energy to the earth in the form of electromagnetic radiation – ca. 10^17 W 8. To state the sun´s potential differently: “In about an hour the sun provides the earth with the present energy requirements of the entire human population for a whole year”9. If this astronomical potential could be harnessed in a cost-efficient fashion, it would meet our global energy needs 10 000 times over10. In Figure 1 below the total global potential for solar energy is depicted relative to the potential of other energy sources.
Solar potential
Figure 1: The potential of solar power dwarfs competing energy sources, where wind is firmly seated in second place.11
The sun´s power density (more commonly referred to as the level of insolation), defined as power per unit area perpendicular to the direction of the sun´s rays, is ca. 1366 W/m2 just above the earth´s atmosphere. This value is known as the solar constant. This figure diminishes by around 30% as the sun´s rays traverse through the atmosphere, however, and the accepted standard power density at sea level on a cloudless day at a temperature of 25°C is amounted to 1000 W/m2 (referred to as the peak-watt level (Wp), which is the customary norm used by the industry for laboratory testing and as a comparable). If not mentioned otherwise, this will be the assumed level of insolation for models and figures employed in this report as well.12
One of the governing characteristics of solar energy supply is its nature as an intermittent power source. Intermittency refers to the variability of its´ energy output beyond our direct control. The level of insolation varies greatly during a day/night, from season to season, as well as depending on weather conditions and cloud-levels. This intermittency is reasonably predictable today from a macroscopic perspective, but a natural restriction still exists on the time frame where solar energy may be harnessed.
With limited energy storage capabilities today, the nature and timing of energy
8 Lynn, P. (2010)
9 Lynn, P. (2010)
10 Bradford, T. (2006)
11 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
12 Lynn, P. (2010)
generation plays a fundamental role in determining the costs and mix of electricity generation sources used by utilities today. This will be discussed in more detail towards the end of the chapter.
2.2: Photovoltaic (PV) Electricity 2.2.1: The Photovoltaic Effect
The photovoltaic effect is a natural phenomenon that enables the direct conversion of light energy from the sun to electrical energy. Upon exposure to light, a voltage or a corresponding direct electric current is induced in a material. Stated simply, photons of light give their energy to valence electrons of the solar cell material, which are through an induced electric current, led into an external circuit. Semiconductor materials facilitate this effect most efficiently. A semiconductor is a material, which exhibits intermediate conductivity levels, between that of the conductors and insulators. The principal trait of semiconductors that have made them so especially useful is their extreme sensitivity to the presence of minuscule amounts of impurity atoms. Minute alterations to their molecular structure allow us to change and control their electrical characteristics to our great benefit. The most commercially successful semiconductor, used in the diodes and transistors of modern electronics, as well as in solar panels, is silicon.
2.2.2: Silicon
Silicon solar cells have been the backbone of the PV industry since the very beginning.
The technology has experienced an overwhelming development and reduction of costs in its lifetime which, coupled with strong future prospects, will make it hard to topple silicon from its golden pedestal.
Silicon (Si) is a chemical element with atomic number 14. It is the second most abundant element in the earth´s crust, following oxygen, but is rarely found as a free pure element. Most often, it is found as a silicon dioxide or other silicate in the form of quartz sand and dust. Furthermore, silicon is a stable element with a considerably long lifetime and very low levels of toxicity (essentially non-toxic). These attributes make silicon a prime contender for sustaining any future growth in the PV market.
2.2.3: Semiconductor Electronics and Doping
Silicon has a total of 14 protons in its nucleus, corresponding to 14 electrons, 10 of which are tightly bound to the nucleus. The remaining four electrons in the outermost energy shell, called valence electrons, are decisive for the facilitation of the photovoltaic effect. In a pure silicon crystal lattice, each of the four valence electrons forms a covalent bond with the electron´s from the four nearest neighbors, creating a three- dimensional tetrahedron structure (Can be seen in Figures 2 and 3)13.
13 Goetzberger, A., Hoffmann, V. U. (2005)
Covalent Bonding Silicon Tetrahedron Lattice
Figure 2: Schematic representation of Silicon´s covalent bonds.14
Figure 3: The three-dimensional view of a silicon diamond lattice.15
At low temperatures, the bonds tightly constrain the movement of electrons, but at higher temperatures or as a result of an external energy stimulus these bonds may be broken. In our case, this external stimulus is packets of photons, called quanta of light energy. The valence electron pairs of the crystal constitute what is called the valence band, whilst unoccupied atomic orbits of a higher energy level merge to form the conduction band. The difference in energy between these two bands is referred to as the band gap, where no electron is stable (which for silicon is approximately 1.1 eV)16. This is equivalent to the energy required to kick an electron out of its’ orbit17. Although silicon atoms are electrically neutral under normal conditions, a photon of light with a sufficient amount of energy may be absorbed by a valence electron, forcing it to jump across the band gap and into the conduction band where it is free to move without hindrance. If this occurs, a corresponding positive region, referred to as a hole, will be left in its place. Holes and electrons are known as charge carriers and their movement within the material will ultimately generate the required electric current. Without an electric field present, however, these holes and electrons will migrate at random through the material and eventually, through a process called spontaneous recombination, cancel each other out. In order to exploit the energy of the electrons and holes to power our electrical devices, we must therefore separate them from one another before they can recombine. This may be accomplished with an electric field induced by a p-n-junction.
2.2.4: The p-n-junction
Doping is a process where a small number of impurity atoms are diffused in to replace some silicon atoms in the crystal structure in order to increase the amount of charge carriers and enhance the material´s electrical characteristics. Phosphorous, with five valence electrons, is used to concoct the so-called n-type silicon. Phosphorous is considered an electron donor because it donates an extra electron, which is free to roam around the material. The net charge of the n-side after donation is therefore positive due to the loss of this electron. Boron, with 3 valence electrons is used to forge p-type silicon, and is referred to as an electron acceptor due to its one broken bond (hole). The net charge of the p-side eventually becomes negative as electrons fill the vacant holes.
When p-type and n-type silicon come into contact, they form the so-called p-n-junction.
Electrons and holes diffuse across the junction to replace one another until they have all recombined (favorable recombination). This leads to a net negative charge on the p-side
14 Atomic Scale Design Network (2011)
15 University of Chicago (2010)
16 Callister, W., Rethwisch, D. (2011)
17 Edwards, D. (2010)
and a net positive charge on the n-side, inducing an electric field in what is termed the depletion region, which is illustrated in Figure 4.
Depletion Region and Current Induction
Figure 4: Electrons diffuse across the p-n-junction to fill the holes on the p-side. When all the charge carriers are used up (depleted), a net difference in potential induces an electric current.18
This electric field will prevent spontaneous (unwanted) recombination and allow us to harness the electron excitation energy supplied by the sun. As photons of light strike the junction, electrons will jump from the valence band to the conduction band leaving positive holes as explained earlier. Due to the induced electric field, however, these holes and electrons will now flow in opposite directions generating a direct electric current.
2.2.5: Photovoltaic Electric Current
In order to construct a modern solar cell, therefore, we begin with a metallic substrate layer used as the back electrode. On top of this, a thin layer of boron-doped p-type silicon is placed which is subsequently complemented with an even thinner layer of diffused phosphorous-doped n-type silicon. The junction is placed near the top in order to increase the probability of photons reaching it. Thereafter, the cell is topped of with an antireflection coating (ARC) to reduce the amount of sunlight that reflects off the top surface. Finally, the front-surface contacts are added as thin bus bar and finger- configurations to reduce shadowing. An illustration of a completed solar cell connected to a load is depicted in Figure 5.
The Photovoltaic Effect
Figure 5: The p-n-junction functions like a diode, allowing current flow in only one direction. Electrons flow through the top contacts and into the external circuit, through the load (light bulb) and are finally met by holes in the bottom electrode.19
18 Wolff, C. (2011)
19 Sun NRG (2011)
2.3: The Photovoltaic System and Production
The Photovoltaic System is an all-inclusive term for the components required to generate electricity through solar cell technology, including the solar panels themselves, the support structures and all of the electrical components required to supply electricity from a PV power station. The PV system is usually separated into two fundamental divisions: the PV module, which is the actual solar cell technology framework, and the balance of systems (BOS), which is a collective term for all of the electrical components and support structures required.
2.3.1: The PV Module Production Value Chain 2.3.1.1: High Purity Polysilicon Production
The genesis of the c-Si solar cell begins with the processing of high quality quartz or sand deposits (SiO2) into metallurgical silicon. Subsequently, the silicon created is further purified by reacting it with hydrogen chloride, HCL, to form a gas called trichlorosilane (SiHCl3)20:
Si (s) + 3HCl (g) è SiHCl3 (g) + H2 (g)
A number of additional distillation processes of liquefied trichlorosilane are performed until the impurities are reduced to a mere few parts per billion (ppb). The final step of the refining process is referred to as the Siemens Process. In this process, the purified trichlorosilane is introduced into a chemical vapor deposition (CVD) reactor, where the gas is deposited under very controlled conditions onto a so-called monosilicon seed crystal (used as a material template):
SiHCl3 (g) + H2 (g) è Si (s) + 3HCl (g)
At this point, the silicon is purified to the required level of ca. 99.99999% for use in solar cells. In order to facilitate the photovoltaic effect successfully, this high level of purity is paramount. It is now referred as high purity polysilicon (or solar grade silicon)21.
2.3.1.2: Ingot Growth and Wafer Cutting
The next step in the production process is to transform this high purity polysilicon (defined as containing many crystals) into a single crystal monosilicon lattice structure.
This is most commonly done through the Czochralski process. A schematic illustration of the process is portrayed in Figure 6. The polysilicon is melted in a large quartz crucible to ca. 1500°C. Thereafter, a thin seed crystal, ca. 0.5 cm in diameter and 10cm in length, (monosilicon lattice) is dipped into the melt22. Under closely administered temperature gradients, seed crystal rotation and extraction rates it is possible to utilize the melt to grow (from the interface between seed and melt) a continuation of the same monosilicon crystallographic structure. The now expanded monosilicon single crystal structure is referred to as an ingot. This ingot, depicted in Figure 7, is typically 1-2 m in length and around 200mm in diameter23.
20 Chang Mai University Paper (2006)
21 Chang Mai University Paper (2006)
22 Victor Jones, R., Harvard University Paper (2001)
23 Victor Jones, R., Harvard University Paper (2001)
Czochralski Growth Silicon Ingot
Figure 6: Single crystal pulling through the Czochralski process.24
Figure 7: Monocrystalline Silicon Ingot.25
After growth of the silicon ingot, it is necessary to cut the ingot into physical shapes better suited for use as solar cells. The three primary slicing steps are visualized in Figure 8. The conventional method used today is to slice the ingot vertically using automated wire saws, forming very thin pseudo-circular single crystal plates, called wafers. A large web of tightly organized parallel steel wires (up to 1000 of them26) is lowered onto the ingot. Powerful engines drive the entire web at the same relative speed. A slurry (a mix of abrasive particles and coolant fluid) is secreted onto the ingot to facilitate slicing of the ingot by erosion rather than cutting. They are
typically manufactured with a thickness of ca. 200µm27.
Ingot/Wafer Slicing
Figure 8: Wire saws first cut off the ends of non-uniform diameter. Afterwards, the sides are flattened to give the wafer its shape and finally, the matrix of wire saws simultaneously segments the ingot into wafers.28
2.3.1.3: Cell Formation
The next phase in the production value chain involves the transformation of the wafer into a solar cell. In order to do this, four fundamental procedures are executed. First, the wafer´s top layer is completely flattened through a machining process. Thereafter, the cell is introduced into a diffusion oven, where impurities (boron and phosphorous) are added to form the previously discussed p-n-junction. Third, a so-called antireflective coating (ARC) is added to the top of the cell to reduce the amount of reflection and to see to it that the cell absorbs the maximum amount of sunlight. Finally, the electrodes that are used to channel the electron flow into the external circuit are infused into the cell. Thin coatings (usually made of silver), known as fingers and bus bars, are attached to the surface of the cell. The fingers collect the electric current generated, while the
24 Victor Jones, R., Harvard University Paper (2001)
25 The University of Bolton (2011)
26 Applied Materials (2011)
27 Lynn, P. (2010)
28 Applied Materials (2011)
thicker bus bars attach the fingers together and provide external connection points to other cells.” Additionally, the entire rear surface of the cell is layered with a reflective inner surface (usually aluminum) to enhance the chances of capturing additional photons29. Both Figures 9 and 10 portray finished solar cells.
Solar Cell Manufacture Monocrystalline Silicon Cell
Figure 9: A fully processed silicon solar cell.30
Figure 10: The c-Si solar cell receives its blue color due to the deposition of the ARC.31
2.3.1.4: The Completed PV Module
In the final phase of the production value chain, the individual solar cells are connected together in systems comprising several cells, referred to as solar modules. For high- power grid-connected systems each module may consist of up to 72 individual cells.
The number of cells used per module and their connection characteristics (parallel vs.
series) is chosen to generate the most favorable voltage/current levels. A monocrystalline silicon solar module with 72 cells and a surface area of 1.5m2 can yield up to 200Wp (peak watts of power)32. In order to provide the modules with the necessary mechanical strength, as well as weather and corrosion resistance, they are often encapsulated and sandwiched in a variety of materials (such as ethyl vinyl acetate, tedlar and aluminum framing), which are depicted in Figure 11. When several solar modules are interconnected, they are commonly referred to as solar arrays (See Figure 12).
Solar Module Layers Array of Solar Modules
Figure 11: A schematic cut-view of the various substrates and encapsulates used to preserve and protect the solar cells.33
Figure 12: A solar array, connecting its´ constituent solar modules and cells.34
The production of the solar module and its underlying technology is designated as the segment in PV electricity generation with the most potential for further cost reductions.
It is commonly asserted to represent ca. 50% of the total cost structure. The remaining
29 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
30 Wikipedia (2011)
31 Wikipedia (2011)
32 Lynn, P. (2010)
33 DRB-Squared LLC (2011)
34 Sciences (2011)
50% stems from the previously broached balance of systems (BOS) components35.
2.3.2: Balance of Systems (BOS)
Balance of systems is a collective term describing the various items required (in addition to the solar module) to complete a fully engineered installation of a grid- connected PV system. This will include a wide variety of static module/array mounting structures manufactured to support the modules securely, safely and at an angle towards the sun. These are usually made of aluminum or steel and exist in a number of variations depending on application and location. UV and water resistant cabling for direct current (DC) wires also fall under the BOS costs. The largest BOS component, in terms of cost, is the inverter unit. An inverter is an essential component of the grid- connection system. An inverter makes use of advanced electronics to convert DC to AC power at the correct frequency and voltage in accordance with electricity grid supply.
Furthermore, there are a multitude of other electrical devices required, such as a PV combiner unit (junction box for wires with surge protection), a protection unit (allowing for disconnection from the inverter) and an energy-flow metering system36.
2.4: Photovoltaics and the Market for Electricity 2.4.1: The Electricity Grid
The electricity grid is an interconnected system employed to transfer electrical energy from suppliers to consumers. The final product of a grid system is a steady supply of electricity. Because electricity cannot be stored cost-efficiently today, it must be continually generated and transmitted to customers. A customer does not receive a fixed quota of electricity, but rather expects to receive electricity whenever an electrical load, such as a light or an air-conditioner, is switched on. Ideally, therefore, the supply of electricity must exactly match the demand, which is subject to the individual will of the consumer. This demand will fluctuate excessively based on an overwhelming amount of factors. These may be variations in demand between different consumers, such as those typically between residential and commercial users, but also due to differing energy cultures and traditions (f. ex urban vs. rural users). Demand also fluctuates tremendously based on variances in time of day, geography, season, climate and weather conditions. Figure 13 depicts the annual electricity demand curve for the ECAR region in the northeastern United States in 2005.
Electricity Demand in 2005; ECAR-region, USA
Figure 13: This graph depicts a curve of the electricity demand over the course of a year. The peak of this demand arrives in the summer months (from 4001 to 6001 hours). The peak-load demand requires a maximum generation capacity connected to the grid of ca. 100 000 MW. The minimum amount of generation capacity that must be connected to the grid at all times (the base-load capacity) is around 40 000MW.37
35 Interview: Bjørseth, Alf
36 Lynn, P. (2010)
37 European Tribune Group (2005)
In order to match the supply of electricity with demand, the grid operators need to make accurate predictions. These forecasts are always based on models created from past empirical data. Electricity demand is, for example, highly correlated with temperature.
Demand will typically peak during mid-summer and winter months when users turn on their air-conditioners or heaters, respectively. It is important to note that electricity demand trends are highly dependent on geography and local weather conditions. One of the essential traits of a region is a measure of the “minimum amount of power that has to be supplied at any given time to a power grid”38. This level is labeled as base-load power. The power plants that supply base-load power typically have to be highly efficient, low-cost generators that take advantage of the fact that they need to continuously generate electricity at all times. They are also often highly expensive to shut down and restart. Peak-load power is at the other extreme and refers to the power that is generated when demand is at its highest and the entire capacity of available generators is required. The power plants that supply peak-load power have to be flexible enough to start up quickly to provide short-term electricity in peak demand periods.
Intermediate-load power is the term for electricity generated somewhere in between the two extremes.
2.4.2: The Generation Stack and Solar PV
In deregulated markets, prices of electricity are typically determined through a combination of auctions. Power producers will submit a price at which they are willing and able to supply electricity into the grid. Subsequently, grid operators will activate generators in order from the lowest to the highest price bids. The lowest bids will tend to constitute the base-load supply, while the higher bids will make up the peak-load generators as they require higher prices to cover investments costs. These auctions are referred to, as “non-discriminatory” because all accepted bidders eventually receive the same price, regardless of their bid39. This price, called the clearing price, will reflect the highest price bid that was activated. The resulting body of power producers that are activated are referred to as the generation stack. The generation stack of the eastern United States PJM region in July 2005 is illustrated in Figure 14. The x-axis represents cumulative electricity capacity in Megawatts and as can be seen, the ranking is made from low-cost to high-cost producers.
Generation Stack in 2005; PJM-Region, USA
Figure 14: The x-axis, representing cumulative capacity connected to the grid, can be interpreted as moving from base-load capacity (hydro, nuclear and coal) to intermediate capacity (wind, pumped hydro, gas) and finally to peak-load capacity, during which PV can be generated. This movement is highly correlated with their respective operational cost ($/MWh).40
38 Edwards, D. (2010)
39 Edwards, D. (2010)
40 Massachusetts Institute of Technology (2007)
Today, solar photovoltaics producers compete for peak-load power generation due to the comparatively higher cost/kWh. An advantage of solar PV is that peak-demand periods (such as in the middle of the day) coincide with the periods during which solar panels can most effectively generate electricity. The dynamics of this situation are currently undergoing a number of changes, however, as conventional electricity prices are rising and solar PV prices are decreasing. Further reductions in costs will allow solar PV to represent a larger portion of the current generation stack. These changing dynamics are largely due to government support schemes.
2.4.3: Political Incentives and the Feed-In-Tariff (FIT)
Government support measures, such as tax benefits, soft loans and feed-in tariffs have played a vital role for the emergence of PV as a competitive player among power producers. “The overall goal of government policy is to guarantee that citizens can get safe, secure, sustainable and low-carbon energy at affordable and competitive prices”41. In order for PV to be available at competitive and affordable prices, however, costs have needed to be reduced substantially. Without government subsidization, the idea is that market forces alone will be insufficient in making photovoltaics competitive within the accepted time frame.
Due to negative and potentially irreversible developments, such as global warming, non-renewable resource depletion, pollution, eco-system degradation, safety hazards and more, time is viewed as a luxury we do not have. Through government subsidization, therefore, PV has been molded into a positive investment opportunity, which in turn has stimulated financing opportunities and bolstered incentives for further cost reductions. One of the most successful and widely adopted subsidy schemes has been the feed-in-tariff (FIT), which was first employed by Germany. Through the FIT scheme, producers of PV electricity are given a long-term contract for guaranteed access to the grid at a certain premium price (above the market price). The delta in price will be absorbed by the end-consumers who received a slightly augmented electricity bill. In the early years after FIT implementation, the subsidized portion of the price could be considerably large. As costs have continuously reduced, however, the levels of subsidization have been reduced accordingly and the idea is that this will commence until PV producers can fend for themselves and FITs are no longer needed. In certain regions, this self-reliance has already become a reality42.
2.5: The Benefits of Solar PV
There are a number of unique benefits to adopting PV for electricity generation purposes. Other energy sources, such as coal, gas and nuclear, are finite resources, which have a limited supply lifetime. Solar power, on the other hand, is arguably unlimited in supply. Additionally, where fossil fuel plants pollute and release greenhouse gases, PV´s carbon footprint is comparatively non-existent (“the manufacturing of PV systems emits between 15-25 g of CO2-equivalent per kWh, to be compared with the average 600g that each kWh produced in the world emits”43). Also, although pollution levels of nuclear-generated electricity are not equally exorbitant, other adverse issues exist. These include difficulties regarding storage of radioactive
41 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
42 Interview: Carlsen, Raymond
43 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
waste, as well as safety/security matters, such as for instance, the unwanted use of technology for nuclear weapons or events such as the Chernobyl incident and the recent catastrophe at the Fukushima reactor in Japan. Solar panels, on the other hand, are completely safe as they require no fuel, have no moving parts and produce no noise.
They also provide energy security, as they will supply electricity at a known cost for ca.
25-30 years after installation44. Furthermore, since solar panels are small and relatively aesthetic, they do not require large production facilities at great distances from consumption. Solar PV is versatile and may be located wherever the sun shines (reducing transmission costs). The only real physical limitation of solar energy´s potential today is the lack of effective storage applications, which place a natural restraint on the applicable timeframe of PV due to intermittence. Despite this, a study has shown that PV could account for up to 20% of global electricity supply at today’s technology level, which indicates that there is abundant growth potential before intermittence becomes a capacity limitation45. All in all, however, if costs are reduced to the point where it is affordable and competitive, solar PV represents a unique, complete and sustainable solution with very few downsides.
44 European Photovoltaic Industry Association (EPIA), A.T. Kearney (2009)
45 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
Ch.3: Reducing Costs for PV systems 3.1: Introduction: Cost Structure and Trends
The costs of PV-generated electricity have been decreasing continuously over time and the transition to price competitiveness is just around the corner in many places. A cardinal concept and major ambition of the PV industry is to reach grid parity. Grid parity may be defined as the point in time at which the cost of unsubsidized photovoltaic electricity is equal to the cost of conventional electricity sources. It is important to understand that grid parity is not one single figure, but rather differs greatly from region to region based on the costs of the local generation stack. Grid parity has already been reached in certain regions of the world, such as southern California and Italy, where both insolation levels and electricity prices are high46. The levelized cost of electricity, depicted in Figure 15, is a weighted measure for average generation costs, often used to as a comparable between different energy sources. As can be seen from the following chart constructed by Goldman Sachs in January of 2011, although coal and nuclear still remain at a cost advantage, the spread between them and PV is not unmanageable.
The Levelized cost of Electricity in 2011: A Comparison of Energy Sources
Figure 15: The levelized cost for electricity ($/MWh), depicted in the y-axis, varies considerably for each energy source, which leads to the substantial cost ranges for the individual sources (such as gas peaking). As costs for solar PV fall, its column will be lowered, giving it more competitive muscles and a larger portion of the generation stack.47
Another important concept when viewing cost reduction trends and the potential for future reductions is the experience curve. The experience curve states that “for a certain increase in cumulated production, the price drops by a certain fraction”48. Over the past 30 years, the experience curve reveals that the “price of PV modules has reduced by 22% each time the cumulative installed capacity (in MW) has doubled”49. These are considerable reductions in prices and if solar PV commences along this same experience curve, its competitive muscles will soon be disquieting for competing energy producers.
Figure 16 depicts the historical progression of module prices in $/Wp.
46 The Goldman Sachs Group Inc (2011)
47 The Goldman Sachs Group Inc (2011)
48 Goetzberger, A., Hoffmann, V. U. (2005)
49 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
Solar Module Price ($/Wp) Trend from 1980-2010E
Figure 16: The price for a completely manufactured PV module has fallen from a price of ca. $21/Wp in 1980 to approximately $1.8/Wp in the end of 2010, a colossal reduction in price of over 90% in 30 years.50
These reductions in price have come as a result of more experience, technological developments, process optimization and volume increases among other things. The initial investment cost in modules and BOS components represents the overwhelming majority of costs for PV generated electricity during the system´s 25-30 year lifetime, as such costs as maintenance and control are virtually negligible. Typically, module cost reductions have outpaced BOS cost reductions (with the exception of inverters which have followed a similar learning curve to that of modules). This is usually because BOS components, such as labor costs and commodity prices (steel, aluminum, copper) have not depreciated over time. A typical cost structure for a PV system is illustrated in Figure 17. In this rooftop installation, module costs represent 60% of the total system cost, whereas the BOS component amounts to 40%51.
Cost Structure of a c-Si Rooftop PV system
Figure 17: This figure illustrates the various components of a PV system and their relative weights in terms of cost.52
In this chapter, the most promising fields for future cost reductions of PV solar systems will be outlined. Prospective developments within efficiency gains and economies of scale and their effect on costs will be discussed individually first. Subsequently, the impact of these advancements will be fused with the potential cost cuts through process optimization strategies. A complete analysis will be conducted jointly based on the various phases of the production value chain. Two separate numerical studies, based on
50 The Goldman Sachs Group Inc (2011)
51 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
52 Greenpeace International and European Photovoltaic Industry Association (EPIA) (2011)
the same forecasting data taken from a variety of sources, will be conducted in unison with our progression through the phases of the production chain. The first will forecast potential cost cuts within each individual phase in the period 2010-2015 and employ the dollars per watt figure ($/Wp). The second will utilize the same data to illustrate the effect of these reductions in real monetary terms for an example 150MW solar PV facility producing explicitly solar modules and BOS components with an initial investment cost of $547.5million. Conclusively, the forecasted cost reductions in each phase will be summated to arrive at a total cost figure for the entire PV system (module and BOS components) in 2015.
3.2: Technological Innovation: Cost Reductions 3.2.1: Efficiency
Arguably the most crucial factor for reduction of total system costs is through enhancement of module efficiency. Solar cell efficiency may be defined as the percentage of solar radiation that is converted into electricity. By increasing the cell efficiency, a solar cell will produce more electrical power per unit surface area. As a consequence, this will in turn decrease material costs and BOS costs per unit of electricity produced. An efficiency increase of 1% has been said to result in a reduction of up to 10% of the total system´s cost/Wp53. Costs measured in $/Wp are an effective portrayal of costs as they take into consideration the efficiency of cells. In this respect, PV producers can be interpreted as manufacturing a certain level of efficiency (at a certain cost), rather the more ambiguously defined “solar panel”.
This section will cover the fundamental avenues taken by PV corporations to improve their cell efficiencies and hence reduce their total system costs. An effective way of modeling improvements in efficiency is to first distinguish between the various loss mechanisms. If the entire amount of light energy radiated onto the surface area of a solar cell was converted to electrical energy, a cell would have an efficiency of 100%.
Unfortunately, however, this is not the case, due to natural constraints caused by fundamentals of light propagation, quantum theory and semiconductor science. Typical crystalline solar cell efficiencies are in the range 15-20% as can be seen in the following illustration.
Efficiency Loss Mechanisms
Figure 18: Approximately 80-85% of the total energy from the sunlight incident on the solar module is lost due to a
combination of quantum-, optical-, recombination- and resistance-losses, of which quantum is the most substantial.54
As depicted in Figure 18, efficiency losses can be categorized into quantum-, optical-, recombination- and resistance losses. Each of these loss mechanisms will be examined in more detail in conjunction with their respective capacities for improvement.
Reducing the prevalence of these losses will play a major role in reducing the future costs/W.
53 Mercom Capital Group (2011) Mercom Solar Intelligence Report 2011-04-11
54 Lynn, P. (2010)
