The rapid increase in worldwide shipments of PV, the steep growth in manufacturing capacity and the currently positive political climate offer good prospects for the future of PV, but at the same time carry some risks. PV should not be charged with the sole responsibility of solving the climate problem, and, despite its massive long-term energy potential, it does not (yet) have to power the world alone.
The PV industry needs a stable political climate for continuous and sustainable growth, in both the old and the new senses of the word ‘sustainable’. Rapid changes in subsidy levels and conditions, or in political attitudes, can seriously harm stable growth.
In the past there were essentially only three sectors involved in the PV arena: government, the PV industry and the utilities. In recent years, three new sectors have become increasingly active: regional and local governments (municipalities), the building sector (architects, project developers, urban planners) and, last but not least, the financing institutions (banks, bilateral and multilateral agencies). Proper collaboration between these ‘old’ and ‘new’ sectors is a necessity to promote PV in a wide range of applications. This collaboration can take the form of joint ventures, or through covenants or agreements between the parties.
PV energy is an indigenous energy source, and can therefore contribute to reducing dependence on energy imports and increasing security of supply. Development of solar energy can actively contribute to job creation, mainly in small and medium-sized enterprises. Deployment of PV can be a key feature in regional development with the aim of achieving greater social and economic benefits. Proper but temporary financial mechanisms (green pricing, base-rate incentives …) should be implemented in the current transitional market.
The expected growth in energy consumption in many third-world countries in Asia, Latin America and Africa offers a great opportunity for solar energy. The modular character of most renewable technologies allows gradual implementation, which is easier to finance and allows rapid scale-up where required.
If one adds its growing local roles to the very large long-term potential of PV, and recognises the time it takes before a major new energy technology matures, it is fully justified that governments support programmes to develop and demonstrate PV technologies.
In conclusion, it is our view that the future of PV looks bright, if stable growth of the industry can be sustained, proper co-operation can be agreed between the growing number of players involved in the PV arena, and reliable but temporary financial support mechanisms are implemented.
= 1 short ton (US) = 0.90718 tonne
 The unit cost of PV electricity depends not only on the capital cost and lifetime of the system components, but also on the local insolation and the cost of borrowing money to finance the system. Energy costs and prices vary widely within and between countries. The costs and assertions in this section are baldly stated, but derive from the detailed costings and assessments of Chapters 15 and 17.
Becquerel’s observation was strictly speaking a photoelectrochemical effect, but its basis—the rectifying junction formed between two dissimilar electric conductors—is the same as that of the photovoltaic effect in purely solid-state devices.
 CdS also lives on in the paintings of impressionists such as Monet, whose favourite yellow pigment it was.
c-Si cells are always configured n-on-p because this best suits the properties of silicon, but some otheip-n cells are configuredp-on-л. These cells are also quite thick, because c-Si absorbs light relatively weakly. Most other cells are much thinner.
 In some semiconductors, particularly organic semiconductors, hole-electron pairs remain tight-bound, and are then referred to as excitons.
" The Fermi level is the energy for which the probability of a state being occupied by an electron is exactly one – half. In an intrinsic (undoped) semiconductor, the Fermi level fells in the middle of the forbidden gap. In a lightly doped semiconductor, the Fermi level remains within the forbidden gap but is near the majority-carrier band edge. In a heavily doped semiconductor, the Fermi level lies within the majority-carrier band.
AII the currents given the symbol і in Figs. 1.8-1.10 are strictly speaking current densities.
 Forward biasing a junction means applying a voltage across the device that lowers the band-bending barrier. Reverse biasing means applying a voltage in the opposite direction.
 An ideal isotropic cell is one in which electrons and holes are thermalised to the band edge, the only decay channel for excited states is radiative recombination, and light can enter the cell at all forward angles.
Gratzel cells, in common with other photoelectrochemical devices, will be dealt with in Volume 111 of this series.
 For simplicity we do not here account for a voltage drop due to the series resistance of a solar cell; Vj„ denotes the portion of the applied voltage that appears across the junction.
 For photon energies exceeding twice the band-gap energy Ug, it is possible to create two electron-hole pairs by impact ionisation. However, the number of photons with energies exceeding 2Ug can be neglected for silicon solar cells illuminated by terrestrial sunlight.
 These numerical values are sensitive to the solar spectrum used in the calculation, and to whether series connection is assumed or not.
 Modified closed-space sublimation As already mentioned, soda lime glass, the cheapest substrate, becomes soft above 550 C and has to be supported in order to avoid warping. A process in which the glass is supported by rollers during deposition was developed by Solar Cells Inc. (Meyers et al., 1993, Sandwisch, 1994). The CdTe is evaporated from semi-cylindrical troughs positioned above the horizontal substrate and the vapour is directed downward to the substrate by suitably structured shields. This has so far led to modules of 60 x 120 cm2 area, certified at NREL as up to 8.2% efficient.
 Much of the thermal waste is already recovered, including some of the radiative losses. However, our calculations suggest that there is great potential for significant increases in the magnitude of recovered energy.
 Efficiencies referred to as confirmed have been measured in independent laboratories. Many of those quoted are in Green et at. (1995 and 1997).
 There is no general agreement as to standard irradiance conditions for concentrators. In the US, the standard irradiance is often taken as 850 W rrf2 or 1000 W пГ2.
 STC (standard temperature and conditions) for unconcentrated terrestrial radiation are planc-of-array global irradiance 1000 Wm"2, cell temperature 25 C, and an air mass 1.5 solar spectral distribution.
 PTC conditions are plane-of-array global irradiancelOOO W nf2, ambient temperature 20 C, and 1 m s’ wind speed.
 Also known as the proton exchange membrane fuel cell.
 In summer 2000, BP Amoco underscored their commitment to clean energy and renewables by restyling themselves as bp, standing for ‘Beyond Petroleum’.
 Assuming that 1 toe generates 4,000 kWh, based on an average thermal efficiency of 33% for today’s stock of power stations (see the BP Statistical Review of World Energy (any year)). The world’s commercial energy demand in primary energy units is approximately 8.1 billion tons of oil-equivalent energy, or 350 EJ per year. The land area of the industrial and the developing countries is approximately 110 million km2, of which 38% is under crops and pasture.
 See Ahmed (1994) for a comprehensive survey of costs over the period 1972 to 1992 and a summary of cost projections in engineering studies, and the report of the US President’s Committee of Advisers on Science and Technology (1997) for a more recent statement on current costs. See also Anderson (1997). The papers in Johansson et al. (1993) provide reviews of the various technologies.
 See Anderson (1997) and World Bank (1996). The actual prices charged in developing countries are sometimes more than twice the cost figures quoted above, however, largely because of the high costs of establishing markets in rural areas. Small 40W systems in Africa have been found to cost $12-20/Wp (Ashford, 1998), though this often includes a lighting unit. A recent survey of quoted prices for grid and off-grid PV systems was conducted by the EA (1997), which found that $10/WP was the lowest of the range of prices found in OECD countries.
 See World Bank (1996) for an analysis of alternatives.
 Acker and Kammen (1996), Cabraal and Cosgrove-Davis (1995), Miller (1998), Adamantiades et. al. (1998) and World Bank (1996) discuss the issues listed in this paragraph and provide practical guidance.
 See for example the Project Implementation Review of the Global Environment Facility (1997) and Adamantiades etal. (1998).
 Masashi’s dissertation (1997). It should be added that the devices have been promoted through subsidies rather than through offering lower prices in off-peak periods.
 The term embedded generation refers to PV systems located at particular points in the distribution networks (generally substations), while distributed generation refers to grid PVs owned by consumers.
 See van Zee (1998), Nodfmann (1998) and Strong (1998), who together provide numerous examples and illustrations. The trade journals are replete with architecturally attractive examples.
 See for example the scenarios of the Shell Group of Companies (Jennings, 1995), IPCC (1995), Watson
(1996) and Johansson et al. (1993), World Bank (1992), and Anderson (1994).
 It levers approximately $4 additional private and public finance each $1 of its own finance. See GEF
(1997) and its Quarterly Operations Reports.
 An international review of policies on privatisation and reform in the electricity industry is provided by Bacon (1995).
 Myers (1998), World Bank (1992).
 See my own review, Anderson (1997).
 The arguments presented in this section can be readily understood without reference to the elementary mathematical treatment also presented; the latter may be useful however, as a means of estimating the level of tax-incentives or subsidies required for an optimum allowance for the external benefits discussed. Reference should also be made to the famous paper of Arrow (1962) and the review by Baumol (1995).
 See Baumol (1995).
 See DTI (1998) for a review.
 A review of the merits and limitations of the UK NFFO programme and a comparison with programmes in other countries is provided by Shepherd (1998).
1 wish to thank Katherine Kramer for this suggestion and our discussions about it.