These factors could uniquely advantage MENA as a global location of choice for CSP production and could strongly drive local manufacturing while 1 creating demand for installed capacity. The main objectives of this study are: to provide an overview of manufacturing processes for key CSP components as well as a cost analysis for CSP components and systems, and for CSP plants as a whole, including the potential for cost reduction to further assess the potential in the MENA region for building and developing a CSP component and equipment manufacturing industry, focusing on the five MENA CTF countries, but with a broader view to the MENA region to propose roadmaps and an action plan to help develop the potential of locally manufactured CSP components in the existing industry and for new market entrants
The term local manufacturing comprises both local industries and subsidiaries of international players established in a country to produce locally.
to analyze potential economic benefits of developing a CSP component manufacturing industry and CSP manufacturing at the construction site of new CSP plants
Figure S-1 Main Objectives of the Study
The CSP market environment: Positive trend
After twenty years of operation in the Solar Electric Generating System (SEGS) plants in California, the worldwide market growth of renewable energies gave CSP technology a new outlook in countries with high direct radiation. Beginning with the Spanish and US electricity markets, many projects are now under development. Electricity-producing CSP plants doubled their capacity with new installations since 2007; by the middle of 2010, a total of over 800 MW of CSP plants were in operation. Although the United States and Spain strongly dominate the CSP market, national support incentives for CSP has caused the market to boom over the past few years. Australia and countries in MENA and Asia are developing their first projects.

Figure S-2

a) Global CSP capacity existing by mid-2010 and projected through 3 b) MENA CSP capacity: projects under operation /construction and in planning phase
Source: Estela, 2010 The MENA CSP IP and its co-financing by the CTF play a vital role in stimulating CSP plans in the MENA region. Table S-1 shows the CSP projects in the MENA CSP IP pipeline as of October 2010. In total, for the five MENA CTF countries considered in this study, nearly 1.2 GW of CSP power plants are expected to be developed in the coming years.Error! Bookmark not defined.

[MW] Tower Parabolic Fresnel Dish & Stirling Total OperationalError! Bookmark

not defined.

Under construction 17 1,1 1,448
Planning phase 1,603 8,2,247 12,128
Total 1,664 10,2,250 14,409
Source: Sun & Wind Energy 2010 A recent study carried out by the European CSP industry association Estela and by AT Kearney (Estela, 2010) analyzed the latest cost reduction potential by interviewing the existing CSP industries regarding technology improvements and effects of economies of scale. The results are shown in table S-3. Overall the levelized cost of electricity (LCOE) could decrease by 4560 percent by 2025 according to AT Kearney. Economies of scale, efficiency increases, and technology improvements are the main drivers for this development. Many factors will contribute to these total technology and cost improvements by values of 15 to 25 percent including: an increasing number of plants being built in sustainable and reliable markets, competitive market mechanisms, including established and innovative CSP technologies further research and development.
On average, the expected annual cost decrease is about 34 percentplacing CSP between wind energy (with expected cost decreases of about 2 percent a year) and photovoltaic (PV) (with a cost reduction path exceeding 10 percent annually).
Planning phase: Projects are announced by project developers or owners. Pre-engineering is taking place, but real construction and all administrational authorizations have not been finalized yet.
Table S-3 Potential reductions in levelized cost of electricity to 2025
Technology improvements 2020 Source: Estela 2010
Reduction in total plant Levelized Cost of Electricity LCOE (2025) Economies of scale Efficiency increase Technology improvements Mirrors Mirrors flat Receivers parabolic 25% 25% 25%
45-60% 21-33% 10-15% 18-22% Steel structure 30% Storage tank 20% Molten salt 15%

The CSP value chain

An evaluation of the MENA regions potential for developing a home base for CSP requires a detailed analysis of the CSP value chain: the technologies and services, the production processes, and the main industrial players. It is further important to review the cost of CSP and the contributions from individual components of the CSP value chain. Based on the complexity level and the potential for local manufacturing, as well as the share of added value in the CSP value chain, a number of key components and services can be identified that are most promising: key components include mounting structures, mirrors, and receivers, and key services range from assembling and engineering, procurement and construction (EPC) to operation and maintenance (O&M). Single countries of the MENA region have already developed some production capabilities of secondary componentsincluding electronics, cables and pipingwhich might contribute to the local supply of future CSP projects, although their share in the overall value chain might be of minor importance. Figure S-3 shows the different components and services linked to the production and use of CSP, and their shares in the value chain.

Metal support structure

Experience New business opportunities for structural steel Low entry barriers
Increase of efficiency and size
Opportunities for MENA industries of manufacturing CSP components in the value chain
The report analyzes in depth the complexity and investment intensity of a selection of production processes to give a broad overview of which CSP components can be most easily adapted for local manufacturing by local or international industry, and would consequently have the highest potential for manufacture in MENA countries in the short- and mid-term. For each manufacturing process or service, barriers and bottlenecks can be identified that could impede local MENA industries entry to the CSP market in MENA. Table S-5 provides an overview of technical and economic barriers to manufacturing CSP components that will need to be minimized with special roadmaps and action plans if the greatest potential of MENA in CSP is to be realized.
Table S-5 Technical and economic barriers to manufacturing CSP components
Components Civil work EPC engineers and project managers Assembly Technical barriers
Low technical skills required

Financial barriers

Investment in large shovels and trucks

Quality

Market

Suppliers

Existing supplier structure can be used for materials Need to build up an own network

Level of barriers Low

Very highly skilled professionals: engineers and project managers with university degrees Logistic and management skills necessary Lean manufacturing, automation Investment in assemblybuilding for each site, investment in training of work force
Standard quality of civil Successful market players will provide works, exact works these tasks Quality management of Limited market of total site has to be done experienced engineers

Medium

Accuracy of process, Collector assembly has low fault production to be located close to during continuous large site output Low skilled workers High process know-how Low market for continuous high opportunities to sell this product to other quality industries and sectors Purity of white glass Large demand is required to build (raw products) production lines

This combination of factors could give MENA a unique advantage as a global location for CSP production and, while creating demand for installed capacity, it could strongly drive local manufacturing. The World Bank has mandated Ernst & Young, the Fraunhofer Institute for Solar Energy Systems (ISE), and the Fraunhofer Institute for Systems and Innovation Research (ISI) to investigate the potential for local manufacturing in the MENA region, which is the most promising area for its development due to the excellent solar conditions and the proximity to the potential export market for solar electricity in Europe. The main objectives of this study, Assessment of the Local Manufacturing Potential for Concentrated Solar Power (CSP) Projects for the World Bank, are as follows: The study should provide an overview of manufacturing processes for key CSP components as well as a cost analysis for CSP components and systems and for CSP plants as a whole, including the potential for cost reduction. It should further assess the potential in the MENA region for building and developing a CSP component- and equipmentmanufacturing industry, focusing on Morocco, Algeria, Tunisia, Egypt and Jordan (i.e., those countries that have already submitted CSP projects for financing by the CTF (see Error! Reference source not found.)), but with a broader view to the MENA region. An action plan should be proposed to develop the potential of locally manufactured CSP components in the existing industry and of new market entrants. Finally, the study should analyze potential economic benefits of developing a CSP component manufacturing industry and CSP manufacturing at the construction site of new CSP plants.
Analyzing the manufacturing processes of CSP components and systems provides a suitable basis for understanding the effects, including possible industrial development, that the CSP Scale-up Initiative will create in the MENA region. Other markets for renewable energies have already been shown to create local industry in the new innovative field of renewable energies; for example, the photovoltaic industry in Germany or the wind industry in Denmark. Emerging countries like China and India are also playing an interesting role by boosting their own renewable-energies industries. India, for example, is creating a powerful local wind industry, (with new jobs and 14 economic benefits for the country), that supplies the home market as well as the international wind power market. A similar development in the MENA region could be promoted by the action plan for CSP projects and their local manufacturing, as discussed in this report.

concept is based on several parallel tubes forming a multi-tube receiver, thereby increasing the width instead of using a secondary reflector. Compared to trough plants, commercial LFC technology is relatively novel. Several prototype collectors and prototype power plants have been installed in the past few years, but no fully commercial LFC power plants are yet in operation. Novatec, however, is currently building a commercial 30 MWel power plant in Spain. Several concepts with different geometric and design characteristics have been developed by a number of companies, see Table 2. The main differences between the Fresnel concept and the parabolic trough collector include: LFCs use cheap, flat mirrors (6-20 /m2) instead of expensive parabolic curved mirrors (2530 /m2); furthermore, flat glass mirrors are a standardized mass product. LFCs require less heavy steel material, using a metal support structure with limited or no concrete (making for easier assembly). On-site installation of LFCs is predicted to be faster. Wind loads are smaller for LFCs, which leads to easier structural stability, reduced optical losses, and less mirror-glass breakage. The receiver on LFCs is stationary, whereas the trough receiver moves with the entire trough system around the centre of mass. This necessitates flexible connections to the piping, which is technically challenging and maintenance intensive. The receiver is the most expensive component in both parabolic trough collectors and in LFCs; however, the mirror surface per receiver is higher in LFCs than in PTCs. The optical efficiency of LFC solar fields (referring to direct solar irradiation on cumulated mirror aperture) is lower than that of PTC solar fields due to geometric principles: In order to reach a certain solar concentration, the LFC mirrors are packed more densely than in PTC plants. The advantage of reduced mirror spacing is that it requires less land; the disadvantage is that mutual mirror shading and mirror blocking of the reflected sun-light occurs. Furthermore, the sun rays are not hitting the LFC mirror perpendicularly, which leads to cosine losses.
It is expected that the mentioned cost advantages will more than compensate for the efficiency drawbacks of LFC technology, but this will have to be proven in commercial plants. Linear Fresnel collectors seem to be more open for redesign and adaptation to local conditions. Local content is probably higher than for the parabolic trough due to the simpler components. All commercial Fresnel collectors use pressurized water / steam as an environmentally friendly heat-transfer fluid. A power plant with direct steam generation thus requires fewer heat exchangers than one using HTF thermal oil.

Figure 11 MENA CSP capacity: projects under operation/construction and in planning phase
There are, however, some threats to these developments, especially on two fronts: Due to the long-term impacts of the financial and economic crisis, a larger number of planned installations are not being realized. This could hamper the cost degression of the technology and its penetration in the MENA region. Other renewable energy sources show far greater dynamics: by the end of 2010, wind energy may have passed the 200 GW level of installed capacity, photovoltaic (PV) will reach 32 GW. Although CSP is seen as a complementary renewable option to wind and PV, there is also an increasing element of competition, especially with PV. Installed capacity (GW) Wind energy Photovoltaic (PV) CSP End 2009 159,2 22,9 Mid 2010 175,0 0,8 End 2010 200,0 32,0
Sources: World Wind Association 2010 (http://www.wwindea.org/home/index.php); Solarbuzz 2010 (http://www.solarbuzz.com/)
Higher figures have been forwarded in some countries, e.g., 2000 MW in Morocco. This figure only includes planned plants that are sufficiently well documented, e.g., through calls for tender. It is not always clear how large the CSP share in those plans could be.
1.2 Structure and characteristics of international players in the CSP value chain
1.2.1 The CSP core value chain
This section gives an overview of the existing CSP value chain. It will describe the international CSP market, the key players in completed and ongoing CSP projects, and the CSP component manufacturing industries in the main markets (Europe and the United States). The CSP core value chain consists of six main phases: Project Development Materials Components Plant Engineering & Construction Operation Distribution
There are also three cross-cutting activities, which are not directly part of the value chain, but rather serve a super ordinate function. They support the project from the beginning to the end or accompany the technology development and specifications over many years: Finance & Ownership Research & Development Political Institutions
In addition, these cross-cutting activities also offer prospects for local employment.
Figure 12 Basic structure of the CSP value chain including cross-cutting activities

Core value chain

Project Development
Concept Engineering Geographical Det erminat ion Det erminat ion of general requirement s

M at erials

Component s
Plant Engineering & Const ruct ion
EPC-Cont ract or: Det ailed Engineering Procurement Const ruct ion

Operat ion

Dist ribut ion
Element s of t he core value chain
Concret e St eel Sand Glass Silver Copper Salt Ot her chemicals
M irrors M ount ing St ruct ure Receiver HTF Connect ion piping St eam generat or / heat exchanger Pumps St orage Syst em Pow er Block Grid connect.
Operat ion & maint enance of t he plant

One part of the running cost of the plant is the insurance. The insurance cost is determined by what is to be assured and secured financially. Usually, 0.5-1 percent of the initial plant investment is paid as annual insurance cost. The larger portion of the running cost is the operation and maintenance (O&M) cost of the plant. Operation and maintenance costs of power plants that have been put into operation since the CSP renaissance in 2007 have not been made publically available. However, a very comprehensive study assessed the detailed structure of the O&M cost and activities of the Californian SEGS plants at Kramer Junction (Cohen, 1999). The following table summarizes the main findings of this study, which aimed at assessing and improving the O&M activities in the Kramer Junction power plants (SEGS III-VII).
Table 10 Summary of total annual costs for parts and material for solar field maintenance
Parts and Materials Unit Cost SF size = 500,000 m $ %replace m/unit $/m-yr Mirrors 100 0.0.25 Receivers 22 0.963 Sun Sensor 150 0.0.001 LOCs 200 0.0.002 Ball joints 2100 0.0.039 Hdr. Drive 6000 0.Miscellaneous assumed as 5% of total equipment costs above TF Pump Seals per year 500000 0.005 HTF Makeup 4,221,500000 0.084 Water demineralized water for mirror washing Nominal -KJ With 30% higher material costs and cheap water With only cheap water With only higher material costs
Note: KJ stand for Kramer Junction, the reference plant in California Source: Cohen, 1999
$K/yr 125 481.7 0.7 0.9 19.3 27.5 33.6 2.4 42.2 243.3 976.5 1034.9 810.1 1196.5
As can be seen in Table 10, replacement of receivers and mirrors is among the largest cost positions. The reason for this is in both cases glass breakage. The receiver glass tube breakage through thermal expansion problems may have been solved since then by adapting the thermal expansion of glass and metal (see section .1.3). Glass breakage in the solar field could also be significantly 1 reduced by stiffer collector sub-structures. Another cost position is water for mirror washing (see Sources: Cohen, 1999 and Novatec, 2010).
Figure 32 left: Conventional cleaning methods (Cohen, 1999), right: water-efficient cleaning using a cleaning robot of Fresnel collector by Novatec Biosol (Novatec, 2010)
Sources: Cohen, 1999 and Novatec, 2010 Looking at staffing, the study by Cohen presents different scenarios for nominal (=optimal) O&M procedures and reduced O&M activities, both for a developed country and for a developing country (see Table 11).

MENA Glass and mirror industry
The main output of the MENA glass and mirror industry is food and beverage glass, glassware, building, and automotive glass. The glass product that is of direct interest to CSP is float glass as, given appropriate quality, it can be transformed into flat mirrors (solar tower or linear Fresnel) and bent into parabolic mirrors. Float glass currently produced in MENA countries is used for building, and automotive and household mirrors. No flat or parabolic CSP mirror production has been identified in the countries covered by our study Float glass production capacities were scarce until recent years but are currently increasing in Algeria and Egypt. However, most of the regional demand is still supplied through imports. Figure 34 gives an overview of the companies that are active in the MENA market and the float glass lines that are currently in service or under construction (starting date in brackets when available). Out of the five countries covered by the CTF (CTF MENA countries), only Algeria and Egypt have float glass production capacities (Table 14).
Note that, despite ongoing CSP projects considered in Jordan, the Hashemite Kingdom does not appear as a primary target for international companies related to CSP. This is particularly due to the small size of the local market and to high energy prices. There are also a very limited number of mainstream industrial sectors that could step into CSP components manufacturing in Jordan. As a consequence, the present study displays limited information on this country, especially dealing with its manufacturing capabilities and potential.
Table 14 Float glass production capacities in CTF MENA countries (source: EY, based on interviews)
Name of company Egyptian Glass Company Country Current output Clear glass only Quality grades: Building, Silvering & Auto Glass with thickness varying between 2 mm and up to 19mm in varied sheet sizes up to jumbo size glass panes as large as 6000mm x 3660mm. products Shareholders JV : 55% Khalafi / 45% Guardian Citadel Group + Dubai Capital Technology transfer agreement with PPG In partnership with MM-ID & Ali Moussa JV with the Chinese CLFG Production capacities

1 oven of tons/year

Sphinx glass co.
1 oven of tons/year (starting date: Q1 2011)
Saint-Gobain Mediterranean Float glass/ CEVITAL

Float glass

1 oven of tons/year to be commissioned in sept. lines of 600, 700 and 900 tons/day

Algeria

The four float glass producers of CTF MENA countries are the Egyptian Glass Company, Sphinx Glass, Saint-Gobain and Cevital. Note that there are no float glass producers in Tunisia, Morocco and Jordan. In these countries, high energy price combined with low local demand for float glass are strong drawbacks for installing production units. As an example, the local demand in Tunisia is around 25 percent of the production of a profitable float glass plant (for which the minimum output can be estimated at approximately 150,000 tons/year).

Table 22 Barriers and problems expressed by the CSP industry
Often mentioned problems in countries of North Africa Payment of bills Political risks Corruption Security concerns Qualification and education of work force Problem of time scheduling
Political risks are critical for international industry. The industry ranks them among the most severe problems in North Africa. Political risks lead to several barriers to productivity: direct investments are limited and very expensive because of countries high risk ratings, and the full potential for local value generation is not tapped. The CSP market can only reach its potential with a foreseeable market development and lower debt costs. Companies are seeking long-term guaranties and long-term investment in the solar market. One company also recommended the creation of an arbitration court with international standards to secure payments for supplied components and long-term contracts. Security concerns are important for international firms. Companies often incur extra expenses for on-site security staff, if they send their employees into the countries of North Africa. Customs duties hinder business activities in the CSP market. For an integrated MENA market, intra-MENA countries logistics might be a relevant aspect if barriers of international trade come into the focus. Education and qualifications of the workforce are lower in North Africa, but this is not seen as a primary problem for international companies because training on the job is possible. Wages for local staff are 1/3 to 1/4 lower than in Europe, which helps offset lower education and efficiency. However for expert tasks the international companies still use their regular employees from Europe. In summary, the interviews indicated a different rating of risk and problems in different countries. Stable frameworks in combination with strong local business partners could facilitate growth. Further regulatory and legal support would help international companies build up local subsidies and joint ventures and consequently new factories and production capabilities.
Potential for local subsidies and local manufacturing by international companies

Sphinx Glass MENA Glass, through its fully-owned subsidiary Sphinx Glass, is a company established by Citadel Capital and a group of leading regional investors to pursue investments in the promising MENA glass industry, with an initial capital of $120 million. Sphinx Glass' greenfield plant is located in Sadat City, 70 km north of Cairo, and has licensed world-class production technologies from PPG Ind. Inc. The new state-of-the-art facility has a production capacity of 600 tons/day and sells to both local and export markets. Two hundred new jobs will be created initially. The construction phase has employed some 2,000 workers. The new plant is producing high quality clear, colored, and reflective float glass for use in both the automotive and construction industries. Glass sizes vary in thickness between 2-19 mm. Hisham El-Khazindar, Managing Director and Co-Founder of Citadel Capital, sees strong competitive advantages in Egypt in the global glass industry with a large supply of high-quality raw materials, local availability of natural gas, low labor costs and a geographic location that easily supports exports. The country is perfectly suited to become an important manufacturer and exporter of both float and container glass. Saint-Gobain The Saint-Gobain and iecam groups have agreed to jointly develop their flat glass (float) activities in Egypt and Russia, by carrying out two projects together. iecam will take a minority stake in Saint-Gobain Glass's investment project to build its first float glass production line in Egypt, alongside Saint-Gobain's local partner, MMID. In addition, a joint venture will be set up between Saint-Gobain and iecam for the construction of a float line in Russia, in the Republic of Tatarstan. A feasibility study for this investment was recently announced by Trakya Cam, a company of the iecam Group. Output from this plant will be sold in the building and automotive markets. Guardian industries Guardian Industries (GI) operates four float glass plants in Egypt, Saudi Arabia, EUA, and Israel. The plants are high tech and modern. The group has taken over 62 percent of the stakes of the Egyptian Glass Company (EGC), a former state-owned company, from the Egypt Kuwait Holding Co. With the same infrastructure, the company managed to increase the daily output from 400 to 500 tons. Float glass in Egypt is much cheaper than in the US (US: US$350-400/ton, Egypt: US$200/ton). However, GI has expressed concerns in interviews for the poor condition of infrastructure in the country and in the region as a whole. In the context of CSP, Guardian Industry is producing flat monolithic, bent monolithic, flat laminated and bent laminated mirrors in the United States and in Israel. Laminated mirrors have been developed for higher reflectivity and increased durability. Guardian industry is already experienced in CSP mirror manufacturing with annual production figures of 7.4 million square meters bent, 9.2 million square meters laminated, 20.4 million square meters mirrored Among other projects, Guardian Industries has already supplied the following solar fields:

NSFs production facilities include the latest CNC machines, laser cutting equipment and highly automated robots. In addition, NSF production is executed in line with ASME, AISC, BS, DIN and Euro norm quality certificates.
Electric and electronic components manufacturers
Chakira cables (Elloumi group) Chakira is one of the main cable manufacturers in Tunisia and in the MENA region. Chakira does not see any issue in supplying CSP plant with cables produced locally. Chakira is looking forward to diversify its production and is already producing PV cables. The company would be ready to invest for CSP cables as they know that they would get financial and technical support from the Ministry of Industry. Indeed, the Ministry has set up a fund aiming at supporting innovation in Tunisia. For information, the Elloumi Group entails an entity which is specialized in electric installations. This entity has performed the electric network of the first wind farm in Tunisia. Furthermore a subsidiary of Elloumi group is doing R&D electric vehicles (cables enabling high speed charge of vehicles). Arab British Dynamics, affiliate of the Arab Organization of Industrialization (AOI)
The Arab British Dynamics is state-owned and was established in 1978 as a joint venture of AOI and British Aerospace for design of defense systems. Since 1998 it is completely owned by the AOI. Their main products are defense products (wireless communication systems, rocket systems), aircraft harnesses and cables for navigation systems, gas- and oil burners, gas taps, medical equipment (e.g. hospital beds). Production lines are Computer Numerically Controlled (CNC) and are certified ISO standard 9000 & 14001 and awarded by national standard certificates. AOI believes a production of CSP mounting structures is possible but is not yet well informed about it. AOI might have the capabilities to enter into manufacturing of CSP components but currently the awareness about the technology is low and the focus is on other technologies (mainly PV). AOI-Electronics Factory AOI-Electronics Factory is also an affiliate of AOI and produces communication systems (military) and consumer electronics (LCD television screens, speakers, etc.). The company is interested in renewable energies and already did the design of a wind turbine control unit (but manufactured only 2 pieces in total). El Sewedy Power El Sewedys companies are active in a large variety of sectors: cables, electrical products, communications, transformers, meters, steel structures, wind turbines, EPC. El Sewedy is exporting to over 100 countries and operates 30 manufacturing plants in 16 countries. It is the 4th largest cable manufacturer worldwide. El Sewedy Power owns the biggest galvanization factory in the world. El Sewedy Power holds a 90% share in Spanish wind manufacturer M Torres and manufactures wind turbines and blades. Currently the company negotiates with the government the erection and support of a 300 MW wind park in Egypt. The deal is that El Sewedy Power gets the contract but produces the turbines 100% locally.

Price competitiveness is a particular strength for North Africa as industry representatives expect lower production costs in MENA compared to Europe. Quality control will be critical, however, and potential local companies will have to focus on this issue.
4.6 Conclusion of chapter 4
In the Transformation scenario (scenario C) with large market growth, the total potential of local manufactured added value of CSP plants will increase constantly and could reach almost 60 percent as average value for all CSP projects. The average share of local added value ranges from 20 percent in 2012 in scenario A to 60 percent in 2025 in scenario C. Assuming a slow market development, which can be the result of competition with other technologies or a lack of financial support for CSP plants, the local share would be limited to about 27 percent of the total investment for construction and components (scenario A). Also under the conditions of scenario B (No-replication scenario) the impact on local manufacturing is comparatively low, as most CSP components would remain imported, only construction, and project management, and basic engineering services might increase. Market demand is the main driver of local manufacturing because a large market attracts local companies to invest in new production lines or international investors to build up local subsidiaries. Large market demand (scenario C) stimulates the creation of a CSP industry in the MENA region. This development could increase the local share of some projects up to 70 percent. After 2025 the share of local manufacturing is assumed to increase further due to technology transfer and the knowledge acquired through realization of many CSP plants in the region. The level of local share influences economic impact and job impact of CSP development in the MENA region. Economic impact is strongly related with market size of CSP in the MENA region. 5 GW by 2020 in scenario C create a local economic impact of US$14.3 billion, compared to US$2.2 billion in Scenario B. In the year 2025 the number of permanent local jobs can rise up to between 64,000 and 79,000 (scenario C). In the construction and manufacturing sector there are 45,000 to 60,000 annual jobs created plus 19,000 annual jobs in operation and maintenance. Looking at the time horizon of the CTF projects (only until 2020), between 2011 and 2020 the following results arise: Within these ten years the cumulated total jobs of full-time equivalent (1-year) for construction, manufacturing and O&M will increase to over 180,000 in scenario C. That means 34,000 employees working in CSP industry permanently by 2020. By contrast, in scenario B 33,000 jobs (cumulated over 10 years) will have a lower impact on the local economy and technological expertise. In this scenario a permanent workforce of 4500 to 6000 local employees is created by the year 2020. Additional impacts for job creation and growth of GDP could come from an export opportunity of CSP components. Besides economic and social benefits, MENA countries could also increase local technical expertise in renewable energy technologies by following the path to invest in solar energy.

Wind turbine manufacturing in India
India entered the market for manufacturing of wind turbines in the mid 1990s Though other market players, such as Denmark, The Netherlands, Germany and the USA, had already consolidated their leading market positions for over 20 years, Suzlon, at present Indias largest manufacturer of wind turbines, achieved 8% of the global- and 52% of the Indian market share in 2006 (Lewis & Wiser 2007). An analysis of the development of the framework conditions in India and Suzlons strategy to acquire the technological know-how to become Indias leading wind turbine manufacturer can help to understand the processes of international technology transfer in the sector of renewable energies and to derive lessons for the development of strategies therefore.
The history of wind energy in India already began in the early 1980s. Facing a rapid population growth and economic rise accompanied by a strongly increasing demand for electricity, the Indian government started to promote a diversification and enhancement of the power sector. In 1982 the Ministry for Non-Conventional Energy Sources (MNES)38 (since 2006 Ministry of New and Renewable Energy, MNRE) was established with the aim of laying a stronger focus on the promotion of alternative and renewable energies (including wind) and to reduce the dependency on domestic coal resources. 1984 a national Wind Power Program was initiated which included resource assessment, incentive schemes, demonstration- and research activities in the field of renewable energy technologies (RETs) and the establishment of regional agencies for the promotion of RETs in the country. Several regularly updated handbooks on wind energy resources were published to communicate the large potential of the development of the technology in India. Nevertheless the private participation in the sector was still limited at this time and nearly all components for the wind power plants were imported. In 1991 the private power policy was announced which liberalized the wind sector and led to a substantial rise in installed wind capacity in India. By 1995 a change in tax policies made incentive packages less attractive and led to stagnation in the growth of the installed within the following years. Subsequently several Indian states introduced different support schemes to incentivize a further development of the wind energy sector. However, inconsistency and instability of the Indian incentive schemes represented a major restraint for the development of the wind sector since it created insecurity for potential investors (Bhattacharya 2009).
Figure 88 Growth of the Indian wind market - cumulative capacity 1999-2008 (Indian Wind Energy Outlook 2009)
In 1995 the National Guidelines for Clearance of Wind Power Projects were implemented which aimed at assuring grid compatibility of future wind power projects and offered further financial incentives like tax holidays and depreciations on required equipment for wind plants. Besides the general growth of the wind turbine market in India (cf. Figure 88), also these measures created conducive framework conditions for the development of a local manufacturing industry. Favorable customs and excise duties for parts and equipment for wind

Chondritic Meteorites and Their Components 2. Chondritic Components
The nature and abundances of the three major chondritic ingredientsrefractory inclusions, chondrules and matrix materialvary widely among the different chondrite groups. Table 1 lists the proportions of these ingredients in each group, the mean
Table 1. Concentrations of chondritic components in the chondrite groups.
Group Type Refract. Chondr. Chondr. Fe,Ni (vol.%)+ mean metal incls. (vol.%) (vol. %) diam. (mm) <0.<0.3 0.15 1.0 <0.01 0.1 1-5 0-5 Matrix Fall freq. Refract. Examples (vol.%) (%)^ lith./Mg rel. CI#

Carbonceous CI CM CO CV

1 1-2-3

0.5 1.6 0.5 0.6

1.00 1.15 1.13 1.35
CK 3-15 0.8 <0.0.2 1.21 CR 1-2 0.5 50-60 0.7 5-8 30-50 0.3 1.03 CH 3 0.1 ~70 0.1.00 CBa 3 <0.~<1.0 CBb 3 <0.~0.<1.4 Ordinary H 3-6 0.01-0.2 60-80 0.10-15 34.4 0.93 Dhajala L 3-6 <0.1 60-80 0.10-15 38.1 0.94 Khohar LL 3-6 <0.1 60-80 0.6 1.5 10-15 7.8 0.90 Semarkona Enstatite EH 3-6 <0.1 60-80 0.<0.1-10 0.9 0.87 Qingzhen EL 3-6 <0.1 60-80 0.<0.1-10 0.8 0.83 Hvittis Other K 3 <0.1 20-30 0.6 6-0.1 0.9 Kakangari R 3-6 <0.1 >40 0.4 <0.0.1 0.95 Rumuruti Sources of data: Scott & Krot (2003) and references listed therein. ALH = Allan Hills; QUE = Queen Alexandra Range. # Mean ratio of refractory lithophiles relative to Mg, normalized to CI chondrites. + Includes chondrule fragments and silicates inferred to be fragments of chondrites. ^ Fall frequencies based on 918 falls of differentiated meteorites and classified chondrites (Grady 2000). Includes matrix-rich rock fragments, which account for all the matrix in CH and CB chondrites.
Orgueil Murchison Ornans Vigarano, Allende Karoonda Renazzo ALH 85085 Bencubbin QUE 94411
chondrule sizes, and the abundances of Fe,Ni metal grains, which are located within chondrules or probably formed with them. All but two of the 15 chondrite groups fall into the ordinary, carbonaceous, or enstatite classes (Table 1). A critical step in understanding the origin of chondritic components was to identify the effects of metamorphism, aqueous alteration, shock, and brecciation in asteroids and to establish which chondrites could have been derived from a common source (Wood 1962; Zolensky & McSween 1988; Scott et al. 1989; Scott 2002). Van Schmus & Wood (1967) inferred that most chondrites had been heated in asteroids

Scott and Krot

and devised various mineralogical and chemical criteria to divide the chondrite groups into six metamorphic (or petrologic) types (Keil, this volume). These criteria and the development of the electron microprobe quickly led to the establishment of a small group of type 3 chondrites as the least equilibrated or metamorphosed of the ordinary chondrites and the precursors to the strongly metamorphosed types 4-6 (Dodd et al. 1967). In part because CI chondrites are closest in composition to solar composition, they were classed as type 1 and thought to be the primary material from which types 2-6 were derived. However, types 1 and 2 are now considered to be products of aqueous alteration on asteroids, and not pristine nebular aggregates. Type 3 chondrites in the ordinary and CO groups were further subdivided into 10 subtypes: 3.0 (least metamorphosed) to 3.9 (Sears et al. 1980; Scott & Jones 1990). Most of the mineralogical and chemical characteristics of pristine chondrules were then established by studying chondrules in the rare type 3.0 chondrites (see Brearley & Jones 1998). However, CV and enstatite chondrites have generally resisted subdivision because of insufficient samples and confusion over effects due to shock, hydrothermal alteration, and brecciation. The Allende meteorite, which for many years was virtually the sole source of CAIs and the most studied chondrite, was once thought to be the most pristine chondrite. But it now appears to be type >3.6 on the basis of the degree of disorder in the graphitizable carbon (Bonal, Quirico, & Bourot-Denise 2004). In many groups, especially the ordinary and R groups, there are chondrites resembling type 3 chondrites that are actually breccias composed of fragments of material with diverse metamorphic histories in a chondrule-rich matrix. Thus, some chondrites may have formed millions or even billions of years after their ingredients were made. Below, we review the mineralogical and chemical properties of the three major chondritic components based largely on studies of the few type 2 and 3.0 chondrites that show the least metamorphism, alteration, and brecciation [Allan Hills (ALH) A77307 and Yamato 81020 (CO3.0), Semarkona (LL3.0), Acfer 094, Adelaide, and Lewis Cliff (LEW) 85332 (ungrouped carbonaceous chondrites), and the CR2 and CH3 chondrites]. The chondrite groups are discussed in section 3, and the isotopic properties of the chondritic components in section 4. Detailed accounts of the mineralogy of all chondritic components are given by Brearley & Jones (1998). 2.1. Refractory Inclusions The numerous varieties of refractory inclusions are divided into two basic types: CaAl-rich inclusions (CAIs), which are composed of refractory Ca-Al-Ti minerals, and amoeboid olivine aggregates (AOAs), which are composed of forsterite and Ca-Al-Ti mineral aggregates (MacPherson 2003; MacPherson et al., this volume). The two types of refractory inclusions are closely related, although CAIs formed at higher nebular temperatures. Each of the 15 chondrite groups contains roughly equal proportions of CAIs and amoeboid olivine aggregates, but the concentration of refractory inclusions varies enormouslybetween 0.01 and 10 vol.% (Table 1). The primary tool for understanding the mineralogy of the refractory inclusions (and many other chondritic ingredients) is the equilibrium mineral stability diagram for the solar nebula (Fig. 1), which is calculated from the thermodynamic properties of minerals and gaseous species (Grossman & Larimer 1974; Yoneda & Grossman

CAIs and ferromagnesian chondrules, their isotopic and chemical compositions suggest that chondrules and refractory inclusions formed in distinctly different regions (Russell et al., this volume). Ca-Al-rich inclusions and amoeboid olivine aggregates have unique O-isotopic compositions and formed at diverse temperatures above 1350 K in a nebula of solar composition at pressures of ~10-4 bar. They were then cooled quickly so that lowertemperature gas-solid reactions were kinetically inhibited.
Figure 3. Sketch showing diverse varieties of rims on Ca-Alrich inclusions (a) and chondules (b). The outermost fine-grained rims of matrix material, which are nearly ubiquitous on chondrules and CAIs, were acquired after these objects cooledprobably as the chondritic components accreted together. The rims inside the matrix rims, which are less common (see text), contain minerals like those present in the enclosed CAI or chondrule and were acquired in the final stages of CAI or chondrule formation.
2.2. Chondrules Chondrules are round or irregularly shaped particles that were wholly or partly molten before they accreted (Fig. 4). They are the most abundant component in all chondrites except the CI, CM, and CK groups, which are dominated by matrix, and CB chondrites, which are dominated by metallic Fe,Ni (Table 1). Mean chondrule sizes in the groups vary between 0.05 and 1 mm, except for the CBa subgroup in which chondrules are typically 5 mm in size (Table 1). In addition to olivine and low-Ca pyroxene, which are the major minerals, chondrules commonly contain metallic Fe,Ni and troilite, FeS. Metallic Fe,Ni is stable above 1300 K in the canonical solar nebula with forsterite and enstatite (Fig. 1); troilite would form from Fe,Ni below 700 K (Grossman & Larimer 1974) or at higher temperatures in solid-gas fractionated systems (Wood & Hashimoto 1993). Some chondrules, called type I, which account for most chondrules in carbonaceous chondrites, have olivine and low-Ca pyroxene with Mg/(Mg+Fe) ratios >0.95 (>0.9 in ordinary chondrites). Other chondrules,

Chondritic Meteorites and Their Components 2.2.1. Origin of Porphyritic Chondrules
Porphyritic chondrules can be made in the laboratory by partly melting fine-grained dust of the appropriate composition and cooling at the appropriate rate to allow phenocrysts to form. However, just melting so-called dust-balls or dusty clumps without some evaporation and condensation cannot explain several observations. The most important is that the proportion of olivine to olivine plus pyroxene in porphyritic chondrules ranges from <1% to >99%. Since physical processes in the solar nebula cannot plausibly separate olivine and pyroxene dust and solid-state conversion of olivine to pyroxene is kinetically inhibited (Imae et al. 1993), it is rather unlikely that chondrules formed simply by melting aggregates of numerous dust grains. Another problem is that dustballs or compacted dustballs have not been identified in chondrites, except for matrix lumps that are not compositionally appropriate precursors. All of the possible fine-grained precursors that have been identified in chondrules (e.g., aggregational chondrules) have igneous textures. Finally, it is not clear that millions of dry, cold, silicate dust grains could stick together long enough for an aggregate to reach the mass of a chondrule (Wood 1996). Many features of chondrules suggest that collisions between molten, partly molten, or hot, solid grains were an integral part of chondrule formation that allowed growth of mm-sized particles from dust. Half the chondrules in CV chondrites and 10% of those in ordinary chondrites have distinct rims with igneous textures and grain sizes of ~5-10 m (Rubin 2000; Fig. 3b). These resulted from collisions of partly molten chondrules with dust or remelting of such accretionary rims (Krot & Wasson 1995; Rubin 2000; Hewins et al. and Jones et al., this volume). Other evidence for collisions during chondrule formation and multiple heating events comes from compound chondrules, which consist of two chondrules fused together (Wasson et al. 1995), and fragments of chondrules within chondrules (Fig. 6). Thus growth
Figure 6. Sketches showing how chondrules formed by repetitive processes involving heating and melting of dust plus collisions between solid particles and melted or partly melted objects. Such processes can account for the existence of fragments of chondrules and CAIs within chondrules and igneous rims and adhering chondrules on the exterior. Type I chondrule compositions were also modified by incorporation of refractory forsterites (Pack et al. 2004), gas-liquid equilibration (Krot et al. 2004b), loss of metallic Fe,Ni and possibly evaporation (Jones et al., this volume).

grains. Despite the heterogeneous nature of matrices, they are closer to solar composition than the bulk chondrites (see Huss et al., this volume). Studies of solar noble gases and irradiation tracks in regolith breccias show that the matrix rims on chondrules were not acquired in asteroidal regoliths (Nakamura et al. 1999; Metzler et al. 2004). The rims were probably formed when the chondrules and other ingredients accreted in a turbulent, dusty nebula (Cuzzi 2004). Because of its fine-grained nature, matrix was readily modified by heating and aqueous alteration on asteroids. Deciphering how and where the matrix minerals formedin asteroids or the nebulahas been highly controversial (Nuth et al. and Huss et al., this volume). Many matrices of C and O chondrites contain hydrated minerals that were once thought to have condensed at low temperatures from the solar nebula (e.g., Wright 2004). However, most authors now infer that ice accreted with matrix-rich chondrites and that hydrated minerals formed predominantly in asteroids by aqueous alteration (see also Ciesla et al. 2003). The matrices of many type 3 chondrites contain Fe-rich olivine that was once thought to have condensed at high temperatures in the solar nebula (e.g., Scott et al. 1988; Wright 2004). However, Fe-rich olivine, which is especially abundant in the Allende CV chondrite (Fig. 7a), has been shown to be a product of aqueous alteration in asteroids (Krot, Petaev, & Bland 2004c; Nuth et al., this volume). To understand what minerals were present in the fine-grained matrix materials that rimmed chondrules and refractory inclusions before they accreted into asteroids, we exclude matrices in those chondrites with chondrules and refractory inclusions containing hydrated minerals and other minerals that clearly formed after the components had accreted into asteroids as a result of alteration and metamorphism. This eliminates almost all of the ~4000 chondrites! Four remaining chondrites have matrices largely lacking hydrated minerals and Fe-rich olivine: these are Acfer 094 and ALHA77307, which are the least metamorphosed of the type 3.0 chondrites (Grossman & Brearley 2005), Adelaide, and Kakangari. Their matrices are largely composed of crystalline, Mg-rich silicates and amorphous, Fe-rich silicate with additional grains of metallic Fe,Ni, sulfides, refractory oxides, and presolar grains (Brearley 1989, 1993; Greshake 1997; see Scott & Krot 2005). We focus here on the silicates and oxides in matrices, but note that the organic materials in chondrite matrices and IDPs provide additional constraints on the thermal processing of dust in the nebula (Alexander, this volume). 2.3.1. Crystalline Mg-rich Silicates The crystalline silicates in the most pristine chondrites are grains of olivine and lowCa pyroxene with Fe/(Fe+Mg) ratios of <0.05, which are mostly 100-1000 nm in size (Fig. 7b). They tend to occur as isolated crystals in the amorphous matrix, but clusters and aggregates are also present. Olivine tends to be more abundant than pyroxene, except in Kakangari. Both silicates may have high Mn concentrations (0.6-2 wt.% MnO), which are also observed in the Mg-rich olivine in amoeboid olivine aggregates, suggesting that the matrix silicates are also condensates (see Petaev & Wood, this volume). Many low-Ca pyroxenes have very fine-scale intergrowths of orthorhombic and monoclinic structures indicative of cooling at 1000 K hr-1 from 1300 K. The general similarity of this rate to the cooling rate of chondrules suggests

collisions between molten asteroids (see Sanders & Taylor, this volume). The origins of chondrules are discussed further in sections 6 and 7. 2.5. Alteration and Mixing of Chondritic Components No chondrites are entirely free from the effects of metamorphism, alteration, shock and other impact effects, and many of the best-studied type 2 and 3 chondrites, some of which are listed in Table 1, were severely affected. As a result, chemical, isotopic and mineralogical data for chondritic components need to be carefully assessed (see e.g., Jones et al., this volume). To illustrate the fundamental importance of understanding geological processes on asteroids in elucidating the nebula properties of the components in chondrites, we mention some of the minerals that were once thought to be nebular products. Minerals such as nepheline (NaAlSiO4), sodalite (NaAlSiO4NaCl), Fe-rich olivine, and fayalite (Fe2SiO4), which are all present in CAIs in Allende and some other CV3 chondrites and are not stable in a solar nebula at high temperatures (Fig. 1), are now recognized to be secondary, asteroidal products. The identification of chondrites like ALHA77307 and Acfer 094 as the best guides to the nebular properties of components in carbonaceous chondrites required an understanding of the origin of the abundant secondary minerals in Allende and the mechanism whereby Allende could form from chondritic components like those in ALHA77307. The timely fall in 1969 of the large Allende meteorite provided valuable clues that improved our understanding of the early solar system, but progress would have been more rapid if several tons of a less altered CV chondrite had fallen instead (MacPherson 2003)! Table 1 would contain far fewer groups but for the wealth of new meteorites recovered in the past 30 years from Antarctica and various deserts around the world (Sears 2004). Nearly every one of the most pristine chondrites was recovered by people hunting for meteorites. Continued recoveries are important for finding new chondrite groups and identifying unrecognized effects of asteroidal processing. One of the most surprising features of chondrites is their wide variety of components. One might expect that accretion processes and subsequent impacts over 4 Gyr in the asteroid belt would have eliminated heterogeneities in nebular solids, except for those due to ambient temperature. Although there was extensive mixing in the nebula before the chondritic ingredients accreted at low velocities, and numerous asteroidal impacts at high velocity after accretion, the asteroids have preserved an amazing variety of distinct kinds of chondrites. Below, we discuss the properties of the chondrite groups and the nature and location of their parent bodies. 3. Properties of Chondrite Groups

have melted unless thermally buffered by significant amounts of ice and other volatiles. The parent bodies of the chondrites may therefore have accreted in the asteroid belt 2-4 Myr after CAI formation when 26Al was no longer a potent heat source (Hevey & Sanders 2005; Sanders & Taylor, this volume). Tungsten isotope data for iron meteorites support earlier formation of differentiated asteroids (Kleine et al. 2005), but more radiometric age data are needed. A knowledge of the accretion locations and times of individual chondrite groups (and the interplanetary dust particles) will surely allow considerable insights into the origin and accretion of the chondritic components, asteroids and the terrestrial planets. However, this must await further insights from the isotopic properties of chondritic components, and formation ages for chondrules and CAIs. 4. Isotopic Compositions of Chondritic Components
A wide variety of isotopic variations are found in chondrites, which provide key constraints on the origins of the chondritic components (e.g., Birck 2004; Podosek, this volume). These include (1) fractionation effects that are a linear function of isotope mass in CAIs (but not chondrules) that were produced during evaporation and condensation in the solar nebula, (2) nucleosynthetic effects preserved in presolar grains from diverse evolved stars, (3) isotopic variations due to decay of radioactive isotopes, (4) mass-independent effects in oxygen, which are ubiquitous in refractory inclusions and chondrules, and (5) spallation effects in rocks due to irradiation by energetic particles during transit to Earth and on asteroidal surfaces, and possible spallation effects in CAIs due to irradiation from the protosun. Here, we briefly discuss mass-related effects before focusing on the nucleosynthetic isotopic variations in presolar grains and the memories of these variations that are preserved in chondrites, isotopic variations due to short-lived isotopes, and the oxygen isotope effects. 4.1. Mass-dependent Isotopic Variations At equilibrium, mass-dependent isotopic differences between gas, solids, and melts are negligible at the high temperatures at which chondrules and CAIs formed: < 1-2 parts per thousand (). However, kinetic processes during condensation and evaporation can produce large mass-related isotopic effects, which provide important constraints on nebula conditions and timescales during high temperature processing (Davis & Richter 2003; Davis et al., this volume). For example, the range of massdependent variations for Ca isotopes is 20 larger in CAIs than in terrestrial, lunar and bulk meteorite samples (Niederer & Papanastassiou 1984). Maximum fractionation occurs when the starting phase is continually homogenized and back reaction between the gas and the condensed phase is minimized. Because diffusion is much more rapid in liquids than in solids, maximum fractionation requires a liquid, but other factors such as gas composition and pressure are also critical. Mass fractionation effects are relatively minor for chondrules implying that conditions were close to equilibrium during gas-liquid interactions, whereas igneous CAIs typically show large effects (Davis et al., this volume). For example, igneous type B CAIs show enrichments of heavy isotopes of Si and Mg as a result of evaporation from melts during brief reheating. A rare subset of igneous CAIs called FUN inclusions shows larger mass fractionation effects for many elements including O

than most other CAIs, e.g., 7-30 per atomic mass unit (amu) for Mg isotopes in FUN CAIs, cf. 0-11 amu-1 for normal CAIs (Davis & Richter 2003). The FUN inclusions also show large nucleosynthetic isotopic effects (Fractionation and Unknown Nuclear effects), but the reason for this coupling of mass and nuclear effects is not clear. 4.2. Nucleosynthetic Isotopic Variations As noted above, the isotopic deviations from terrestrial values in presolar grains due to nucleosynthetic effects are orders of magnitude larger than those preserved in CAIs, chondrules, and bulk matrix, which are measured in parts in 103 or 104 ( or units, respectively). Thus, presolar particles and those that formed in the solar system can be readily distinguished by their mass-independent isotopic deviations (Fig. 11). The largest isotopic deviations in solar-system particles are found in the FUN-type
Figure 11. Deviations in the Nd isotopic ratios in a FUN-type Ca-Alrich inclusion (EK1-4-1) and a presolar SiC grain normalized to 142Nd and a terrestrial standard. The isotopic variations in the presolar grain are 100 larger than those in the CAI and of opposite sign. The presolar grain contains predominantly s-process Nd, which is deficient in the CAI. (After Ott 1993).
CAIs and certain hibonite inclusions, which contain deviations in 50Ti, for example, of up to 10% (Ireland & Fegley 2000). The nucleosynthetic effects in normal CAIs are much smaller, e.g., variations of 5-30 units in 50Ti/46Ti (Niemeyer & Lugmair
1984), while chondrules have even smaller variations, e.g., a few -units or less in 50 Ti/46Ti (Niemeyer 1988). Bulk chondrites and matrix samples also show isotopic variations of a few -units in 50Ti/46Ti (Niemeyer & Lugmair 1984), 54Cr (Podosek et al. 1999) and Mo isotopes (Yin, Jacobsen, & Yamashita 2002). The effects in CAIs, chondrules, matrices, and whole chondrites probably reflect the gradual isotopic homogenization by thermal processing in the solar nebula of diverse presolar grains that differ in their thermal stability, grain size, and other properties. This implies that FUN inclusions predate other CAIs (Wood 1998; Sahijpal & Goswami 1998), and that CAIs predate chondrules. The nucleosynthetic isotopic effects preserved in chondrules and chondrites have not yet provided firm constraints on their formation locations or mechanisms, though they clearly emphasize that the high-temperature processing experienced by chondrules was brief and/or localized. 4.3. Oxygen Isotopic Compositions The discovery by Clayton and colleagues of oxygen isotopic variations in CAIs and chondrules that could not be attributed to mass-dependent fractionation transformed the study of chondrites and cosmochemistry (see Clayton 1993, 2003). It galvanized the search for presolar grains (although ironically they are absent in CAIs), led to the demise of models for chondrites that relied solely on a hot, gaseous, and homogeneous nebula, provided a new way to classify meteorites, and generated important constraints on the origin of chondritic components, and insights into the history of minerals, moons, and planets. A key property of oxygen that was important in preserving isotopic heterogeneities in the protosolar disk is that oxygen is the only element with three or more isotopes that exists in several forms both gaseous (CO and H2O) and solid (ice and rock) over a wide range of nebula temperatures (Wood 1981). Oxygen isotopic data are plotted on a graph of 17O/16O vs. 18O/16O with units of deviations in parts in 103 from standard mean ocean water (Fig. 12). Most chemical reactions partition isotopes according to their masses along a single line of slope 0.52. Terrestrial samples lie on such a line showing that the Earth was initially isotopically homogeneous. Martian meteorites lie on a parallel line just above the terrestrial line. However, refractory inclusions, chondrules, and matrix samples plot on or near a slope-1 line with refractory inclusions furthest from the terrestrial line (Fig. 12). Vertical deviations from the terrestrial line are defined by the parameter 17O = 17O - 0.5218O (Clayton 1993). Mixtures of two components lie on a straight line connecting the end-members. Although the O-isotopic variations among CAIs were initially attributed to presolar grains, a local origin is now favored as O-bearing stardust in chondrites is largely 17O-rich and CAIs do not show isotopic abnormalities in other elements (Clayton & Nittler 2004). Laboratory experiments by Thiemens & Heidenreich (1983) showed that certain gas phase reactions could lead to products that differed in 17O from the reactants. Although various mechanisms have been proposed involving the symmetry of minor molecules such as O3, O2, and CO2 (Thiemens 1999), the most plausible explanation for the chondrite variations invokes isotopic self-shielding during UV photolysis of the abundant molecule CO in an initially 16O-rich protoplanetary disk or parental molecular cloud with 17O ~-25 (Clayton 2002; Lyons & Young 2005; Yurimoto & Kuramoto 2004; Yin 2004; Krot et al. 2005a). Astronomi-

cal observations show that in molecular clouds, 12C17O and 12C18O can be preferentially dissociated, as UV photons that dissociate the vastly more abundant 12C16O cannot penetrate beyond the surface of the cloud (Federman et al. 2003). Atomic 17O and 18O can combine with hydrogen forming water ice, while the CO is enriched in 16 O. In this model, the bulk solar system has a 17O value of ~-25, like most CAIs. If the inner nebula becomes enriched in H2O because of meter-sized, ice-rich bodies that drift rapidly towards the Sun and evaporate (Cuzzi & Zahlne 2004), then the inner nebula and chondrules formed therein will be enriched in 17O and 18O.
Figure 12. Oxygen isotopic plots of 17O/16O vs. 18O/16O with isotopic compositions plotted as deviations from standard ocean water in parts per 103 (-units). (a) Terrestrial samples show mass-dependent isotopic variations and plot on the line labeled terrestrial fractionation whereas refractory inclusions in most pristine chondrites plot in the lower left of the diagram. Chondrules and matrices plot closer to the terrestrial line within the box. CAIs in the Allende chondrite scatter along a slope-1 line. (b) Enlarged view of boxed region in (a) showing ranges of O isotopic compositions of large chondrules in carbonaceous chondrites, enstatite chondrites, ordinary chondrites (OCs), and R chondrites. There was little mixing between the chondrules in these four kinds of chondrites. (After Rubin 2000.)
Although many details of the 16O-fractionation process are not fully understood (Lyons & Young, this volume), one prediction of the CO self-shielding model has been confirmed. Ion probe analyses of the surfaces of lunar grains of metallic Fe (or Fe,Ni) revealed implanted solar wind with near-CAI oxygen isotopic compositions (Hashizume & Chaussidon 2005). What can be inferred about the origin of CAIs and chondrules from their Oisotopic compositions? First, most CAIs formed in a relatively 16O-rich environment that was probably solar in composition and quite distinct from the 16O-poor environment in which chondrules formed. Exceptions include some 16O-poor CAIs in CH chondrites, all CAIs in CB chondrites, and some igneously formed CAIs in CV and CR chondrites that appear to have partially exchanged oxygen with an 16O-poor gas

when they were molten (Krot et al. 2005a). Minerals in these exceptional CAIs have oxygen isotopic compositions that plot in Figure 12 on the line marked Allende CAIs, which has a slope of ~1. Chondrules and matrix samples have oxygen isotopic compositions that lie close to the top of the Allende CAI line, near the terrestrial fractionation line (Fig. 12b). If the oxygen isotopic composition of the nebula became progressively more 16O-poor (Krot et al. 2005a), then chondrules formed after CAIs. Most individual chondrules in unaltered chondrites, like most CAIs, do not show mass-independent O-isotopic fractionations between, or within, minerals, except for 16O-rich relict grains. However, chondrules from one chondrite (excluding those in E and CB chondrites) commonly scatter along or near the slope-1 line with a spread of 17O values of up to 5 (e.g., Clayton et al. 1985, 1991; Scott & Krot 2001; Jones et al., this volume). This shows that they cannot be derived from a single source such as a molten body. The wide range of O-isotopic compositions of chondrules in all chondrites shows that many distinctly different batches of chondrules were manufactured in the nebula (Fig. 12). Oxygen isotopes therefore confirm the mineralogical evidence that chondrules in enstatite, ordinary, carbonaceous, and R chondrites come from distinctly different sources. In addition, carbonaceous chondrules come from several sources, e.g., those in CR chondrites are mineralogically and isotopically distinct from chondrules in CM and CO chondrites. Other distinct sources are needed for the chondrules in CH, CB, and K chondrites. Al-rich chondrules, which are present in many groups, tend to have overlapping 16O-rich O-isotopic compositions, consistent with an origin as mixtures of chondrule and CAI material (Krot et al. 2005b). To illustrate that chondrites consistently provide exceptions for every rule, we note that chondrules in CH chondrites, which have the widest range of 17O values (+5 to 35), extend along the slope-1 line to below the CAI field (Kobayashi et al. 2003; Yoshitake & Yurimoto 2004). A few CH chondrules also have heterogeneous O isotopic compositions. The 16O-rich CH chondrules may have formed after other chondrules when 16O-rich CO began to dominate the nebula gas. 4.4. Short-lived Radioactive Isotopes Proof that short-lived isotopes decayed in chondritic components comes from excesses of the daughter isotopes that are correlated with the abundances of a stable isotope of the parent element in cogenetic minerals, e.g., 26Mg/24Mg ratios in diverse minerals of igneous CAIs are correlated with 27Al/24Mg ratios showing that 26Al decayed to 26Mg in situ (Table 2). The inferred initial concentrations of relatively longlived isotopes like 53Mn and 182Hf appear to be compatible with steady-state interstellar abundances due to continuous galactic nucleosynthesis (Meyer, this volume). However, the inferred concentrations of 41Ca, 26Al, 60Fe and 36Cl are much higher than the steady-state abundances. These isotopes formed by nucleosynthesis in evolved stars or by spallation during energetic particle bombardment in the solar nebula or molecular cloud (Goswami et al., this volume). For 60Fe, a stellar origin is accepted, as energetic particle interactions are not a plausible source. However, 10Be, which was present in diverse kinds of CAIs (McKeegan, Chaussidon, & Robert 2000), is not made in stars, and inferred initial abundances of 10Be and 26Al are not correlated (Marhas et al. 2002). 10Be was probably acquired during spallation by pro-

drule formation period. Although the CAIs in each group have much in common, the mineralogical differences between groups are hard to explain. However, size sorting and alteration may account for some of the observed differences. If the abundance of refractory inclusions in the disk gradually declined, CV chondrites may have formed first, then the other matrix-rich C chondrites, followed later by O and E chondrites (Cuzzi et al. 2003). However, Al-Mg ages do not support later formation of ordinary chondrites: data for E chondrites are lacking (Kita et al., this volume). In addition to primitive nebular materials, chondrites also contain fragments of early-formed bodies that were aqueously altered or melted (section 2.4). Given a homogeneous 26Al distribution, it is almost inevitable that the parent asteroids of the chondrite groups accreted >2 Myr after CAIs formed, whereas the melted asteroids accreted <2 Myr after CAI formation. The abundance of altered chondritic rock fragments in chondrites suggests that carbonaceous asteroids that were subsequently altered hydrothermally accreted in the outer part of the asteroid belt while differentiated bodies at the inner edge of the main belt or inside 2 AU. We infer that both kinds of bodies formed from materials present in the chondrites, and, contrary to Sanders (1996), do not infer that most chondrules formed from molten asteroids. If any chondrites formed from colliding asteroids, it is likely to be the CB chondrites. As noted above, their properties are very different from those of other chondrites. 7. Nebular Heat Sources for Chondritic and Cometary Silicates
Although the earliest refractory condensates in CAIs and AOAs appear to have formed from rapidly cooling nebular gas, probably close to the protosun (Petaev & Wood, this volume), the heating mechanism is poorly understood (Shu et al. 1996; Wood 2004). The most plausible mechanism for making chondrule and CAI melts is shock wave heating in the solar nebula. However, the sources of the shock waves remain controversial (Ciesla, this volume). Possible sources include bow shocks generated by planetesimals (Hood et al., this volume), spiral arms and clumps in a gravitationally unstable disk (Boss & Durisen, and Boley et al., this volume), close passes by protosolar companions (Reipurth, this volume; Bally et al., this volume), and Xray flares (Nakamoto et al., this volume). The properties of chondrite matrices and anhydrous, porous chondritic IDPs suggest that thermal processing of interstellar dust was ubiquitous, brief, and localized, but limited in extent so that very small fractions of presolar materials survived. Annealing of presolar amorphous material either close to the protosun (Hill et al. 2001; Gail 2004) or by shocks (Harker & Desch 2002) is commonly thought to have generated cometary silicate crystals. However, the general absence of Mg-rich, Fepoor, amorphous silicate from which the Mg-rich crystalline silicates could have formed by annealing favors an origin by condensation. If chondrules formed by shock heating, as seems plausible, it is likely that dust was vaporized by the shock (Desch & Connolly 2002) and rapidly condensed so that the bulk composition of the dust was not grossly disturbed. Since forsterite crystals and amorphous silicates are present around many T-Tauri and higher mass Herbig Ae/Be stars (Bouwman et al. 2003; Honda et al. 2003), the processes that heated and mixed silicate dust in the solar nebula (Fig. 13) may also have operated in other protostellar disks.

Thiemens, M. H., & Heidenreich, J. E., III 1983, Science, 219, 1073 Tissandier, L., Libourel, G., & Robert, F. 2002, Meteorit. Planet. Sci., 37, 1377 Van Schmus, W. R., & Wood, J. A. 1967, Geochim. Cosmochim. Acta, 31, 747 Wadhwa, M., Amelin, Y., Bogdanovski, O., Shukolyukov, G. W., & Janney, P. 2005, Lunar Planet. Sci., 36, 2126 Wasson, J. T., & Rubin, A. E. 2003, Geochim. Cosmochim. Acta, 67, 2239 Wasson, J. T., Krot, A. N., Lee, M. S., & Rubin, A. E. 1995, Geochim. Cosmochim. Acta, 59, 1847 Weissman, P., Bottke W. F. Jr., & Levison, H. F. 2002, in Asteroids III, eds. W. F. Bottke, A. Cellino, P. Paolicchi, & R. P. Binzel (Tucson: Univ. Arizona Press), 669 Westphal, A. J., & Bradley, J. P. 2004, ApJ, 617, 1131 Wood, J. A. 1962, Geochim. Cosmochim. Acta, 26, 739 Wood, J. A. 1963, Icarus, 2, 152 Wood, J. A. 1981, Earth Planet. Sci. Lett., 56, 32 Wood, J. A. 1996, in Chondrules and the Protoplanetary Disk, eds. R. H. Hewins, R. H. Jones, & E. R. D. Scott (Cambridge: Cambridge Univ. Press), 55 Wood, J. A. 1998, ApJ, 503, L101 Wood, J. A. 2004, Geochim. Cosmochim. Acta, 68, 4007 Wood, J. A., & Hashimoto, A. 1993, Geochim. Cosmochim. Acta, 57, 2377 Wooden, D. H. 2002, Earth, Moon, & Planets, 89, 247 Wooden, D. H., Woodward, C. E., & Harker, D. E. 2004, ApJ, 612, L77 Wright, I. 2004, in An Introduction to the Solar System, eds. N. McBride, & I. Gilmour (Cambridge: Cambridge Univ. Press), 315 Yin, Q. 2004, Science, 305, 1729 Yin, Q., Jacobsen, S. B., & Yamashita, K. 2002, Nature, 415, 881 Yoneda, S., & Grossman, L. 1995, Geochim. Cosmochim. Acta, 59, 3413 Yoshitake, M., & Yurimoto, H. 2004, http://www.lpi.usra.edu/meetings/chondrites2004/pdf/9062.pdf Yu, Y., Hewins, R. H., Alexander, C. M. O'D., & Wang, J. 2003, Geochim. Cosmochim. Acta, 67, 773 Yurimoto, H., & Kuramoto, K. 2004, Science, 305, 1763 Zanda, B. 2004, Earth Planet. Sci. Lett., 224, 1 Zinner, E. K. 2003, in Meteorites, Comets, and Planets, ed. A. M. Davis, Vol. 1, Treatise on Geochemistry, eds. H. D. Holland, & K. K. Turekian (Oxford: Elsevier-Pergamon), 17 Zolensky, M. E., & Ivanov, A. V. 2003, Chem. Erde, 63, 185 Zolensky, M. E., & McSween, H. Y., Jr. 1988, in Meteorites and the Early Solar System, eds. J. F. Kerridge, & M. S. Matthews (Tucson: Univ. Arizona Press), 114 Zolensky, M. E., Nakamura, K., Gounelle, M., Mikouchi, T., Kasama, T., Tachikawa, O, & Tonui, E. 2002, Meteorit. Planet. Sci., 37, 737