by | Professor Dr. Saleem H. Zaidi | Professor Dr. Kamaruzzaman Sopian, Solar Energy Research Institute, University Kebangsaan Malaysia
The ability and infrastructure to generate, distribute, and efectively consume electricity determine a society’s educational, social, and economical well being.
For example, near universal access to electricity, through transmission grids connected to large scale power plants, has been largely responsible for economic prosperity of the industrialized world. This centralized electricity distribution model is expected to be replaced by small scale de-centralized micro- grids connected to an array of distributed energy generators.
Continuing maturation of renewable energy generation and distribution technologies has been the driving force behind this paradigm shif. A critical feature of this paradigm shif lies in the concept of energy generation at the point of use. For third world countries, this paradigm shift offers unique opportunity to achieve economic progress in much the same way as telecommunication has evolved from telephone lines to wireless.
Crystalline silicon photovoltaic (PV) technology will play a leading role in this transition to small-scale distributed networks. Silicon (Si) forms almost 26 % of the Earth’s crust, it is uniformly distributed across the globe, and is free from economics of resource-depletion. Silicon is also the most dominant component of integrated circuit (IC) electronics; hence the IC-based infrastructure advances will continue to be accessible to the PV industry. Malaysia has been blessed with sunlight, silicon, and an infrastructure of technologies spanning the photovoltaic supply chain.
Unfortunately, a majority of these technologies are dominated by multinational companies focused on exploitation of cheap labour and natural resources; almost all finished products are exported. There is an urgent need for the Malaysian economic planners to take ownership of this energy sector, not only for the benefit of their own country, but also to become regional and global leaders in the field of renewable energy.
This article examines crystalline Si photovoltaics in terms of its physics, technology, and economics with the aim of developing self-supported, globally-distributed indigenous PV industries. While this goal is realistic and achievable, its success is critically dependent on implementation of energy infrastructure based educational policy. Without appropriate educational infrastructure, transition of emerging economies from carbon-based fuels to Si-based resources will not be possible.
Complete realization of human potential is critically dependent on access to energy1. Energy is required for basic human necessities including food, transport, industrial manufacturing, and medicine. Worldwide disparities in energy usage are reflected in drastic economic inequalities; such inequalities lead to incalculable loss to our civilization. Therefore, the supreme moral obligation and responsibility of our generation lies in creation of a global civilization in which the playing field is level for all humans. This challenge is viewed through the prism of energy.
To meet global energy requirements, an energy source must satisfy three conditions: (a) economically (and environmentally) sustainable, (b) inexpensive, and (c) technologically flxible to become indigenous. Based on these considerations, crystalline Silicon (c-Si) based generation of electrical energy mediated by photovoltaic process2 has been identified as the most logical solution. Si forms approximately 26 % of the Earth crust3, and combined with free energy from the sun, offers an unlimited energy resource uniformly distributed across the globe. Maturation of the Si PV technology combined with advances in distributed energy generation and microgrids have helped create a paradigm shift away from conventional approaches to energy generation, distribution, and consumption. As this paradigm shift develops and universally adapted, the long-cherished goal of economic equality across the globe will become a reality.
ELECTRICITY GENERATION AND GDP
A brief analysis of electricity generation and its relationship to gross domestic product (GDP) is per capita daily energy usage (total electricity generation per year/population) 6. Perhaps the most significant feature of this data is the realization that a majority of world’s population has practically no access to energy. By plotting GDP as a function of per capital energy usage (Fig. 1, right), relationship between energy usage and economical development can be better understood; countries with highest energy usage appear to have the highest incomes. The principal features of the plotted data in Fig. 1 have been summarised below.
i. Three orders of magnitude generation gap across the world; most countries in Africa barely produce any electricity (Fig. 1, left),
ii. Developed countries could be categorised with per capita energy usage in ~ 600-1000 W,
iii. Almost four orders of GDP variation as a function of energy usage, and
iv. Countries with the highest population generate lowest energy.
In order to better understand the relationship between per capita energy usage and GDP, data in Fig. 1 is plotted in Figs. 2 & 3 as a function of per capita energy usage over a broad energy range. The vertical bar represents total population of countries in the per capita range selected. The principal features of the data in Figs. 2 & 3 have been summarised below.
i. Per capita energy generation is less than 100 W for approximately 3000 million people (Fig. 2, left),
ii. China (Fig. 2, right) is poised to join advanced countries with per capita energy usage just below 400 W,
iii. Developed countries with lower populations produce far more energy than poorer countries with significantly larger (~ 10-20 times) populations (Fig. 3, left), and
iv. Indian subcontinent as well as the African continent supports the largest population generating the lowest per capita energy.
Table-1 selectively reproduces per capita energy and GDP statistics for representative countries across the entire range potted in Figs. 1-3. As noted above, economic disparity as a function of energy generation and GDP produced is huge. For example, the ratio of energy generation between USA and Congo is ~ 136; almost ½ of the GDP ratio of ~ 229. The higher energy generation does not always lead to higher per capita GDP as observed in the case of Russia. Similarly, some countries such as Mexico have higher per capita GDP despite much lower energy usage. Such discrepancies may be attributed to higher revenues generated from oil export. In most cases, higher energy generation capacity leads to higher GDP. Per capita energy usage of ~ 500-600 W is generally associated with a developed country; see for example Italy with GDP of ~ US $ 37,000.00.
Therefore, creating worldwide level playing field in energy generation will require approximately per capita energy of 600 W/day for a population of about 3000 million, i.e., ~ 1.5 E12 W.
Assuming conventional carbon-based energy generation resources, greenhouse emissions have been estimated for 1E6Wh range in table-27. Therefore, carbon-based energy generation of ~ 1.5E12 W will add green house emissions of ~ 1-2E9 lbs per hour. Such huge influx will lead to catastrophic weather changes with global impact. Therefore, even apart from moral considerations, our survival requires transition from carbon-based to Si-based energy generation resources; PV energy generation does not produce any green house emissions. Earth receives daily sunlight of ~ 174E15 W, therefore, less than 1 % conversion of this energy will be sufficient to meet all our energy requirements.
* Energy utilised in creation of Si solar cells and panels has been neglected in this calculation.
ELECTRICITY TRANSMISSION
Electricity generation and transmission model in developed countries is simply described by the schematic diagram in Fig. 4. Large (100’s MW) electricity generation plants based either on fossil fuels, nuclear, or hydro sources, use transmission grids extending to hundreds of miles to deliver electricity to vast regions. This model traces its evolution to the early days of electricity generation and distribution in which small scale (100’s KW) of electricity generation plants served small neighbourhoods through transmission grids extending to a few miles (Fig. 5); in the early days, it was more expensive to produce large power plants. With time, it became cheaper to produce large scale power plants, while the cost of large scale transmission grids was low. In today’s world, this system of electrical transmission is becoming increasingly redundant due to several factors including high resistive losses, high cost of grids (both environmental and economical) 12, long lead time of large scale power plants, and increasing availability of distributed energy resources (both renewable and carbon based) 14. This is, perhaps, the most important paradigm shift of the 21st century, and if channelled appropriately, will redefine the way electricity is generated, distributed, and utilised.
In developed countries, the transition from macro-grids to micro-grids is inhibited by the presence of fully-developed transmission grids aided by the monopoly of utility companies that control the grid as well as generation15. Transition will be slow, and is expected to take the form described in Fig. 6, where the grid will be slowly augmented with renewable energy sources. This situation will hold until some such time where the carbon foot print of large scale power plants will become too high to sustain for environmental and economic reasons. When this phase is reached, the macro grids will be divided into small scale distributed energy resources focused on energy generation at the point of use (Fig. 7). In this distributed energy system, several energy generation resources (renewable, carbon based) are combined with storage systems (batteries, fuel cells) to deliver electricity to small scale residential and industrial communities through micro-grids extending to a few miles.
Therefore, advances in energy technologies combined with fossil-fuel depletion and greenhouse emissions have created the conditions for a paradigm shift ideally suited for economically disadvantaged countries. By following distributed energy distribution model, the cost of building macro grids is eliminated, and by generating energy at point of use, urban sprawl is eliminated with the added benefit of decentralised economic development.
CASE FOR CRYSTALLINE SILICON PV TECHNOLOGY
We are blessed with almost unlimited (~ 1000 W/ m2) source of energy in the form of sunlight. A number of solar or photovoltaic technologies have been competing to become the renewable energy source of choice by focusing either on effiency, or low manufacturing costs. In order to diffrentiate between competing PV technologies, a comparative analysis in terms of economics, resource availability, environmental impact, and social benefis has been carried out. Each PV technology sector has been examined in terms of its physics, start-up capital cost, and recurring production costs. For each technology, an analysis of available of materials and supplies has also been considered in order to determine sustainability under high volume manufacturing required to meet energy demands described above. Environmental impacts are examined in terms of greenhouse emissions, toxicity and recycling ability.
Thin-film PV technologies including cadmium tellurium (CdTe) and amorphous silicon (a-Si) appear to be highly desirable on account of their perceived low-costs. However, a deeper look reveals serious raw materials availability issues and efficiency limitations; toxic effects have also become a serious environmental concern. For instance, the concentration of Te in earth’s crust is the same as Pt21. Similarly, due to its toxicity, Cd is one the six elements banned by European community making it difficult to recycle CdTe panels. Th thin-fim a-Si PV technology has long suffered from low efficiencies and light induced degradation. In contrast, crystalline Si technologies have established a track record of performance dating back to almost 50 years. Crystalline Si-based PV technologies benefit from R&D advances in semiconductor integrated circuit (IC) manufacturing. Start-up costs depend on the manufacturing approach, and are often times substantially lower than thin-film technologies. Principal advantages of c-Si-based PV technology include: (a) higher (~ 14-22% range) efficiencies, (b) abundance of resource availability of key ingredients such as Si and Al; Al is ~ 9% of the Earth’s crust, (c) identification of pathways through integration with Ge28 and compound semiconductors to boost efficiencies, (d) natural division into technology sectors such as crystal growth, solar cell, and panel manufacturing, and (e) simple enough to transition into indigenous industries.
Global PV market, dominated by Si, has been growing at ~ 30 % over last twenty years (Fig. 8, left). This growth has largely been attributed to grid-connected installations in Europe and Japan, and has been sustained in large part by government subsidies. Respective market share of PV technology sectors for year 2009 has been plotted in Fig. 8 (right); crystalline Si, in its mono and poly Si formats, dominates the market.
Resource Comparison
Figure 9 plots concentrations of five key elements used in thin-film and crystalline PV technology. Concentrations of these elements in Earth’s crust have been plotted on linear (Fig. 9 a) and logarithmic (Fig. 9 b) scales in order to identify the differences clearly. Principal features of the plotted data in Fig. 9 have been summarised below.
i. Compared to Si, all other elements are negligible,
ii. Te, the key component of CdTe thin-film solar cell has a concentration approximately 1E10 lower than Si, and
iii. Other elements fare better but are still far (~ 1E6 lower) less abundant than Si.
Businesses as a matter of practice do not disclose their plans on securing raw material and supplies. Even so, the case of Te deserves a closer look in view of claims of its low cost32. It is an integral component of the solar panel. According to First Solar’s production estimates, it required approximately 13 tons of tellurium in 200733. Assuming that First solar CdTe captures 10% of the American 2008 installed capacity; it would require 7,260 tons of tellurium. At present, Te is produced as a byproduct of Cu mining. Impacts of such large increase in Te production in terms of costs in the metal mining and refining industry have not been understood and reported. Even if all Te supply is secured, the fact remains that it is one of the scarcest elements on the Earth; reliance on it to meet our energy requirements needs is at best short-sighted, and at worse, no better than our current reliance on fossil fuels.
Crystalline Silicon Supply Chain
Figure 10 schematically draws supply chain of the c-Si PV technology. The first step is production of metallurgical grade Si through chemical reactions between high-purity coal and silica (either in the form or quartz or sand). At high (~ 2000 °C) temperatures, silica (SiO2) reacts with carbon (C) to form SiC, SiO react with C to form Si and CO; liquid, metallurgical Si is thus extracted from the bottom of the furnace. Approximately, eleven to fourteen MWh of electricity energy is consumed in producing 1 ton of metallurgical grade silicon. Metallurgical Si (~ 98 % purity) must be further purified for semiconductor and photovoltaic applications. Most of these silicon purifications processes are based on chlorine. In this purification process, Si chemically reacts with chlorine to form trichlorosilane from which highly pure poly Si is synthesised. Other applications of metallurgical Si are in Al alloys and plastics.
Purified Si (99.99999999) also known as Si feedstock is the starting raw material for both IC and solar industries. Fig. 10 identifies both mono and poly Si processes for wafer manufacturing. High purity Si feedstock is placed in a quartz crucible. The entire assembly is melted and slowly raised as it rotates to form single crystalline cylindrical ingots based on CZ process (Fig. 11)43. Typical ingot diameters are in 4-12 inches with lengths over six feet44. These ingots are subsequently sliced into wafers for IC and solar cell manufacturing. In the poly Si wafer manufacturing process, large ingots or bricks of multi-crystalline (mc) orientations are solidified in a quartz crucible. These bricks are subsequently sliced to form mc-Si wafers for processing into Si solar cells; this is the only application of poly-Si wafers. Solar cells based on either mono or poly Si wafers are then packaged into solar panels. Solar panels are subsequently arranged into electricity generation systems; the simplest configuration has been drawn in Fig. 1246. In view of the intermittent nature of sunlight-based electricity generation, system design is based on the use of batteries as the constant source of electricity; solar panels are used to charge batteries. A dc to ac inverter is used to deliver power to customers.
Economic Benefits of investment in silicon
Silicon solar cell industry has been growing at average rate of ~ 20 % over last 20 years (Fig. 8,left). When fossil-depletion and green house effects are included in analysis, the solar cell-generated electricity requirements are expected to exceed all growth projections. Figure 13 plots the average prices of semiconductor materials in the solar manufacturing process. Materials cost vary over four degrees of magnitude varying from ~ $ 0.1/Kg for silica to ~ $ 1000/kg for electronic grade wafers. Therefore, through processing of an inexpensive, readily available material, extremely high and sustainable economic benefits can be realised.
Si mass and cost required for generating energy of 1.5 E12 W have been calculated in table-3. Energy generated per gram of Si is calculated for a 200-µm thick crystalline solar cell. Silicon in excess of 4.66 billion kg is needed. Assuming US $ 50/Kg, this comes out to be about US $ 233 billion for 15 % efficient solar cells. Approximately 10 million tons of quartz will be processed to produce this much Si48. Current purified Si production is ~ 5E9 kg/year; most of which is used for integrated circuit manufacturing. Therefore, a doubling of current purified Si would be enough to meet our energy requirements. Note that even one percent of the purified Si market will be worth about US $ 2 billion.
CREATION OF INDIGENOUS PV TECHNOLOGIES
Silicon solar cells, based on photovoltaic effect, were fist fabricated in the early 1950s49. Since then, the growth of PV technologies aimed at sunlight to electricity generation has been truly phenomenal. However, world-wide generation of PV electricity has yet to realise its early potential to meet our growing energy demands. Principal factors responsible for slow replacement of carbon-based fuels by Si-based fuels have been summarised below.
i. Stiff competition from fossil-fuel technologies operating at lower costs,
ii. High energy conversion and capital cost per kWh of PV technologies, and
iii. Inherent intermittent nature of sunlight requires investment in storage systems to provide continuous energy.
It is expected that the competition from non-Si resources will continue to inhibit PV growth; however, it will eventually be eliminated through depletion of fossil fuels and environmental degradation from greenhouse emissions. At the same time, it is necessary to understand the fundamentals of the PV technology in order to fully realise its potential.
Crystalline Si PV technology requires processing of Si wafers into solar cells. In contrast with ICs fabricated on Si wafers, a solar cell is a pretty simple structure; essentially a large area diode. In this basic configuration, sunlight enters the cell from the top front surface. An internal electrical field separates the light-generated current, which is collected by the metal contacts at the top and bottom of the cell and transferred to an external circuit. In order to simplify solar cell manufacturing, the authors have designed and implemented a simplified manufacturing process drawn in Fig. 14. In this simplified design, solar wafers are cleaned and textured to reduce surface reflection. Surface properties are evaluated through simple spectral reflection and minority carrier measurements. Textured wafers are diffused to form n-p junctions followed by screen printing and annealing to extract photo-generated current from the solar cell. Completed solar cell response is evaluated through LIV, spectral response, sheet resistance, and electrical contact resistance measurements. All of the characterization and manufacturing equipments have been designed and built using simple, low-cost methods easily adaptable for indigenous manufacturing. Figure 15 shows pictures of the front and back surfaces of the solar cells, its microstructure, and the simple LIV measurement system to evaluate solar cell efficiency. This approach has been used to fabricate solar cells in ~ 15 % efficiency range; higher efficiency is limited by the quality of the starting material.
A similar approach has been developed by the authors to manufacture solar panels. Figure 16 schematically draws principal steps in manufacturing of a solar panel. Solar cells are tabbed and connected into series using simple equipments shown in Fig. 17. The strings of solar cells are combined to form complete panel circuit. Simple current-voltage measurement under a suitable light source is used to evaluate suitability for lamination. Panel is laminated using commercially-available vacuum laminator. Post lamination testing for cracks or damages is carried out through a combination of light source and telescopic visual enhancement. Laminated panels are framed and LIV tested using simple semi-automated testing equipments. Figure 18 shows pictures of solar panels and their LIV measurements. Solar panels manufactured through this process have achieved ICE certification.
Integration of solar panels into an electricity generation system based on the principles described in Fig. 12 is straightforward. Figure 19 shows a schematic diagram of a simple 300 W panel-battery backup system designed to power a light source, an evaporative cooler, and a water fountain.
ENVIRONMENTAL IMPACT OF PV TECHNOLOGIES
Long-term sustainability of PV technologies will require evaluation of their impact on environment. A brief overview of impacts of large scale PV technologies on environment is briefly discussed below.
Cadmium Environmental Impact
Cadmium is the key element in CdTe thin film solar cells. It is considered hazardous for humans. The most dangerous form of occupational exposure to cadmium is through inhalation of fine dust and fumes, which can result in pneumonitis, pulmonary edema, and death62. Cadmium is also an environmental hazard. In the case of CdTe solar cells formed through the vacuum sputtering process, the residues left in the metal and vacuum pumps pose very serious health hazard for workers. Cadmium is one of six substances banned by the European Union’s restrictions on hazardous substances (RoHS) directive which bans hazardous substances in electrical and electronic equipment but allows for certain exemptions and exclusions from the scope of the law63. The supply and use of cadmium is restricted in Europe. Although there is no danger of Cd exposure during the operation of a CdTe solar panel, the recycling of non-functional panels would require stringent safeguards, i.e., they can’t simply be disposed into a landfill.
Tellurium compounds are considered to be mildly toxic, need to be handled with care, although acute poisoning is rare. Te is not reported to be carcinogenic
Water Usage
Excessive water use in PV manufacturing is considered a negative impact on environment. In Si PV technologies, it has been estimated that a million gallons of water are used for each MW of production; this figures includes all PV technology sectors. Extensive research efforts aimed at recycling, reduction of water usage, and in some cases, its complete elimination will help make Si PV technologies more environment friendly.
Economic Benefits of Indigenous Si PV Manufacturing
Comparative analysis of thin-film and c-Si PV technologies shows that while thin-film technologies have role to play, their long-term sustainability have role to play, their long-term sustainability in terms of supply chain, environmental impacts, and efficiency improvements is not clear. Their long-term track record in comparison with c-Si has also not been demonstrated either through accelerated testing or by other suitable means. Si PV technology appears to be the only long-term sustainable option. Investment in Si PV creates a technology infrastructure that spans over many fields and has the potential of becoming indigenous. The investment in thin-film PV, in contrast, leads to a centralised facility with all imported equipment, most supplies also imported, and the workforce is not trained in technology basics.
The solar cell production may be understood in terms of two extremes: the small scale cottgare industry or µ-solar franchies with output in 1-5 MW range and large scale macro-solar with output in ~ 20-100 MW range. Simplified methods for fabrication of solar cells can be adapted towards industrial manufacturing. This can be better understood through a simulation of PV output by making the following assumptions:
i. Mono-crystalline (100) wafers,
ii. 180-200-µm thickness,
iii. 16 % to 18% efficiency, and
iv. 24 hour, year round continuous operation.
Figure 20 plots the PV output as a fucntion of wafers processsed per hour for efficiencies in 16-18 % range. Several conclusions can be drawn including:
i. Output dependence on cell efficiency is critical only at extremely high outputs,
ii. High outputs > 20 MW require full automation due to very high wafer/hour processing, and
iii. Low outputs < 5 MW, simple, manual and semi-automated operations are possible due to considerably lower wafer/hour requirement.
The inset in Figure 20 also plots the enlarged version of the µ-solar franchise. While the operation of a µ-solar franchise is a combination of manual, semi-automated, and fully-automated equipments, the macro solar must necessarily depend on full automation to achieve high throughputs. Table-4 provides a qualitative comparison of µ-solar and macro solar franchises.
In almost every category except marketing, µ-solar franchise is more beneficial. Since, the starting capital costs for typical µ-solar franchises are 10-15 times lower than a macro franchise for the same output, this approach reduces the entry barrier for new players, and perhaps, most importantly, it makes it possible to make PV technology indigenous. This approach is modular in nature, therefore, capacity extension is simply a matter of adding additional lines, or creating another franchise.
The ability to grow Si wafers significantly impacts the success of µ-solar franchises since it is the most critical element in the process, and if not produced locally, it will be expensive. Therefore, any long-term investment in PV technology must include Si wafer growth in the first phase, and Si purification in the second phase. Considering the integrated circuit related applications of this technology, such investments will pay for themselves in very short time frame.
SUMMARY AND RECOMMENDATIONS
Growth of PV technologies will critically impact our future social and economic growth. It is apparent from the analysis carried out in this paper that c-Si PV technologies offer the only long-term sustainable option for renewable generation to meet our energy requirements. Si PV technology is particularly desirable when considering the following factors:
i. A solar cell is just a large area diode,
ii. Manufacturing model based on cottage industry approach, and
iii. Substantial reduction in energy conversion cost ($/W) is achievable.
However, in order for this vision of indigenous PV technologies to be realised, several steps must be implemented, most critical are listed below.
i. Investment in PV technology infrastructure (education, manufacturing, and infrastructure),
ii. Development of energy-based curriculum from school level,
iii. Investment in PV technology creates economic growth, and
iv. Government commitment to provide electricity for all citizens.
Conventional wisdom suggests that manufacturing scale is the key requirement for cost reduction. This model has worked very well in IC manufacturing industry where costs of computers have kept low while their functionalities have continued to grow. The economies of scale have worked in IC industry because of extremely high value-added cost of the integrated circuit devices. In contrast, the value of a Si wafer appreciates by a factor of 2-3 by the time a solar cell is completed. Therefore, it makes more practical sense to reduce start-up capital costs and make technology more affordable in order to reduce costs and incorporate human innovation. The large scale production model supported by multi-national companies will limit technology access through high entry barriers and limit innovation in order to meet production targets, In contrast, small scale franchises distributed across the land will lead to decentralised power generation, reduces migration from rural to urban areas, and help enhance social life and education at remote locations.
In the final analysis, when considering PV technology, following factors must be taken into account:
i. PV Technology is Special in its unlimited potential to solve our energy problems,
ii. It is needed most by poorest of the poor,
iii. As currently configured, it is expensive,
iv. To meet worldwide energy requirements, PV technology must become indigenous, and
v. PV technology must evolve into cottage/community industry.
As transition from carbon-based energy resources to Si-based energy generation is achieved in the 21st century, a Paradigm Shift as Fundamental as that from coal to oil in 20th Century will be realised. The paradigm shifts that most of us are already familiar include cell phones versus land-based line phones, and satellite TV systems.
