**4.3 Production process flow**

Multiple chambers are used for deposition of different functional layers in the module production process. Optimizing the arrangement of chambers and controlling of the process flow are crucial to the production throughput and directly affect the panel production cost. There are mainly three types of process flows: batch process, continuous process and hybrid process. Characteristics of the three processes are compared in Table 2.

Batch process of film deposition is the most intuitive way of arranging deposition chambers. In this configuration, functional layers are deposited consequently onto batches of substrates. The typical batch processes are seen in Oerlikon's thin film production lines. An example is the Oerlikon KAI-20 1200 production system (Fig. 11a), which consists of two PECVD process towers, two load-locks, one transfer chamber and an external robot for glass loading from cassettes (Kroll et al. 2007). Each process tower is equipped with a stack of ten plasma-box-reactors where ten substrates are deposited simultaneously. The layers are processed in parallel at the same time in both stacks (2×10 reactors). The whole KAI-20 1200 PECVD production system shares one common gas delivery system including the mass flow controllers and one common process pump system. Engineering work has been put to ensure small box-to-box variations of deposition rates, layer thickness uniformities. The batch process normally requires small footprint, and is suitable for slow deposition that requires long process time (e.g., the absorbing i-layers). In fact the PECVD deposition of different p-, i-, and n- layers can be combined within the same chamber as long as dopant diffusion from the process chamber can be minimized. In most cases more than one chambers are used for the entire film stack, thus when they are moved between separate chambers the substrate manipulation and heating / cooling time has to be minimized to increase the process throughput.

352 Solar Cells – Thin-Film Technologies

for long panel lifetime. The most common encapsulation method for panels with the glass substrate is to use another piece of glass to cover the functional films. The gap between the two glass plates is filled with an epoxy (ethylene vinyl acetate, EVA, or polyvinyl butyral, PVB) film, which not only insulates the functional films against reactants like oxygen and moisture, but also mechanically strengthens the rigidness of the finished panel. Quality of the module encapsulation is directly associated with the failures of panels in the field. Judgment of the encapsulation properties includes low-interface conductivity, adequate adhesion of encapsulants to glass as a function of in-service exposure conditions, and low moisture permeation at all operation temperatures

The panel then passes through a laminator where a combination of heated nip rollers removes the air and seals the edges. The lamination film at the same time provides electrical insulation against any electric shock hazard. At the exit of the laminator conveyer, the modules are collected and stacked together on a rack for batch processing through the autoclave where they are subjected to an anneal/pressure cycle to remove the residual air and completely cure the epoxy. Finally, a junction box is attached to the cross bus wire and sealed on top of the hole of the back glass and is filled with the pottant to achieve a complete

The fully processed module is then tested for output power, *I*SC, *V*OC, and other characteristics under a solar simulator. Then it is labeled, glued to the supporting bars, and

Multiple chambers are used for deposition of different functional layers in the module production process. Optimizing the arrangement of chambers and controlling of the process flow are crucial to the production throughput and directly affect the panel production cost. There are mainly three types of process flows: batch process, continuous process and hybrid

Batch process of film deposition is the most intuitive way of arranging deposition chambers. In this configuration, functional layers are deposited consequently onto batches of substrates. The typical batch processes are seen in Oerlikon's thin film production lines. An example is the Oerlikon KAI-20 1200 production system (Fig. 11a), which consists of two PECVD process towers, two load-locks, one transfer chamber and an external robot for glass loading from cassettes (Kroll et al. 2007). Each process tower is equipped with a stack of ten plasma-box-reactors where ten substrates are deposited simultaneously. The layers are processed in parallel at the same time in both stacks (2×10 reactors). The whole KAI-20 1200 PECVD production system shares one common gas delivery system including the mass flow controllers and one common process pump system. Engineering work has been put to ensure small box-to-box variations of deposition rates, layer thickness uniformities. The batch process normally requires small footprint, and is suitable for slow deposition that requires long process time (e.g., the absorbing i-layers). In fact the PECVD deposition of different p-, i-, and n- layers can be combined within the same chamber as long as dopant diffusion from the process chamber can be minimized. In most cases more than one chambers are used for the entire film stack, thus when they are moved between separate chambers the substrate manipulation and heating / cooling time has to be minimized to

packaged. At this point, the full panel assembly is finished.

process. Characteristics of the three processes are compared in Table 2.

(Jorgensen et al. 2006).

module integrity.

**4.3 Production process flow** 

increase the process throughput.


Table 2. Comparison of three thin film solar module process flow types

Continuous deposition of the multilayer structure is realized in a roll-to-roll manner, which ensures stable chamber conditions for consistent film growth for large volume production. United Solar, Energy Conversion Devices (ECD), and Xunlight took this type of growth configuration. For example, the ECD 30 MW a-Si process line consists of nine seriesconnected chambers with gas gates that isolate dopant gases between chambers (Fig. 11b)

The hybrid-process system is designed to combine the advantages of batch and continuous processes. In this configuration, separated chambers are used like those in batch process, but individual substrates are fed into different chambers for optimal chamber utilization. Each substrate sees a queue of different process chambers like that in continuous process. Applied Materials configured its SunFab in the hybrid mode, where a group of several process chambers construct a functional cluster unit sharing a heating chamber and a center transfer robot (Fig. 11c). Each cluster is focused on a group of related functional layers (e.g., layers comprising a subcell in a multi-junction structure), and deposition of the multijunction stack is realized by going through clusters. In this configuration, each chamber can have flexible deposition time, and the flow of substrates and synchronization of chambers are controlled by artificial intelligence algorithm for optimal system throughput (Applied Materials 2010; Bourzac 2010). This process flow combines the advantages of small footprint, easy maintenance and high production throughput, and provides flexible system

There are a number of considerations to weigh when deciding among batch, continuous or hybrid processes, and some of the major reasons are listed in Table 2. Generally, small production volumes favor the batch process type while continuous process is more suitable for high volume production. Capital investment cost of a batch or hybrid process system is also usually lower than the continuous process because the same equipment can be used for multiple unit operations and can be reconfigured easily for a wide variety of panel structures, though the operating labor costs and utility costs tend to be high for the former two systems (Turton et al. 2008). The continuous configuration is also more favored for 'substrate' type solar cells on metal foil substrates in a roll-to-roll deposition (Izu and Ellison 2003). Though the comparisons in Table 2 generally holds true, it is also possible that the configuration works for one solar plant may not be the best choice of another, as each plant differs at production scale, materials supply, geological confinement and many other

In this chapter, the cost structures of a-Si/µc-Si solar modules has been described with analysis of the multilayer cell structure and module production. The monolithically integrated structure is described with explanations of layer functions. The industrial fabrication of large-area modules are introduced, including FEOL and BEOL process steps. Module costs around half of the total thin film PV system. We analyzed the factors affecting the module efficiency and cost in terms of energy consumption, equipment investment, spending on direct material, labor and freight cost. To probe strategies of efficiency improvement, we started from the introduction of the Si p-i-n junction structure and the front/back contacts, and discussed the light absorption and its enhancement with light trapping. The photocurrent generation is achieved by effective capture of the incident solar photons, and conversion into free electrons and holes by the build-in field of the p-i-n junction. Resistance loss during photocurrent collections is minimized by the conductive front and back contact layers. At the meantime, enhancing the light absorption within thin layers is achieved using band gap engineering of the absorbing layer and optical trapping of

Fabrication of large-area tin film solar panels are the key to increasing the production volume and reducing the \$/Wp of modules. State-of-the-art fabrication includes FEOL and

configuration for versatile panel fabrication.

practical characters.

the front/back contact layers.

**5. Conclusion** 

(Izu and Ellison 2003). The film deposition substrates are 2.6 km long, 36 cm wide, 127 µm thick stainless-steel rolls fed into the deposition system at constant speed. For quality assurance, online diagnostic systems are installed allowing for continuous monitoring of the layer thickness and characterization of the PV properties of the manufactured solar cells. A big advantage of the continuous process is that the substrate does not see the atmosphere during the process, and needs to be heated and cooled only at the beginning and last chamber, thus greatly saving the pumping time and energy cost. At the same time, all chambers continuously run at the optimized, stable states, thus depositing films with uniform and consistent properties. On the other hand, Since the deposition rate and thickness of each layer varies a lot (e.g., typical p-layers are < 20 nm while the µc-i layer is normally 1-2 µm), the deposition time in each chamber are very different. Limited by a constant substrate roll feeding speed, the chamber for growing i-layers are much longer than the doped layer chamber. In fact, this 30 MW system is 90 m long.

Fig. 11. Typical process systems used for Si thin film solar cell manufacturing. a) Batch process. Schematic side and top view of an Oerlikon KAI-20 1200 PECVD process system for a-Si deposition (Kroll et al. 2007). b) Linear Process. Schematic diagram of a United Solar Ovonic Corporation roll-to-roll a-Si:H alloy triple-junction solar cell processor (Yang et al. 2005). c) Hybrid (batch plus linear) process. Schematics of a Applied Materials SunFab thin film production line (Applied Materials 2010).

(Izu and Ellison 2003). The film deposition substrates are 2.6 km long, 36 cm wide, 127 µm thick stainless-steel rolls fed into the deposition system at constant speed. For quality assurance, online diagnostic systems are installed allowing for continuous monitoring of the layer thickness and characterization of the PV properties of the manufactured solar cells. A big advantage of the continuous process is that the substrate does not see the atmosphere during the process, and needs to be heated and cooled only at the beginning and last chamber, thus greatly saving the pumping time and energy cost. At the same time, all chambers continuously run at the optimized, stable states, thus depositing films with uniform and consistent properties. On the other hand, Since the deposition rate and thickness of each layer varies a lot (e.g., typical p-layers are < 20 nm while the µc-i layer is normally 1-2 µm), the deposition time in each chamber are very different. Limited by a constant substrate roll feeding speed, the chamber for growing i-layers are much longer

Grid

Triple-junction Cell Structure

> Al / ZnO N1 I1 N2 P1 I2 P2 N3 I3 P3 TCO

Stainless Steel

than the doped layer chamber. In fact, this 30 MW system is 90 m long.

8.6m

Moving Stainless Steel Web

Fig. 11. Typical process systems used for Si thin film solar cell manufacturing. a) Batch process. Schematic side and top view of an Oerlikon KAI-20 1200 PECVD process system for a-Si deposition (Kroll et al. 2007). b) Linear Process. Schematic diagram of a United Solar Ovonic Corporation roll-to-roll a-Si:H alloy triple-junction solar cell processor (Yang et al. 2005). c) Hybrid (batch plus linear) process. Schematics of a Applied Materials SunFab thin

(b)

N2 I2 P2 N3

N1 l1 P1 l3 P3

x10

x20 x20

6m

film production line (Applied Materials 2010).

(c)

x10

3.2m

(a)

The hybrid-process system is designed to combine the advantages of batch and continuous processes. In this configuration, separated chambers are used like those in batch process, but individual substrates are fed into different chambers for optimal chamber utilization. Each substrate sees a queue of different process chambers like that in continuous process. Applied Materials configured its SunFab in the hybrid mode, where a group of several process chambers construct a functional cluster unit sharing a heating chamber and a center transfer robot (Fig. 11c). Each cluster is focused on a group of related functional layers (e.g., layers comprising a subcell in a multi-junction structure), and deposition of the multijunction stack is realized by going through clusters. In this configuration, each chamber can have flexible deposition time, and the flow of substrates and synchronization of chambers are controlled by artificial intelligence algorithm for optimal system throughput (Applied Materials 2010; Bourzac 2010). This process flow combines the advantages of small footprint, easy maintenance and high production throughput, and provides flexible system configuration for versatile panel fabrication.

There are a number of considerations to weigh when deciding among batch, continuous or hybrid processes, and some of the major reasons are listed in Table 2. Generally, small production volumes favor the batch process type while continuous process is more suitable for high volume production. Capital investment cost of a batch or hybrid process system is also usually lower than the continuous process because the same equipment can be used for multiple unit operations and can be reconfigured easily for a wide variety of panel structures, though the operating labor costs and utility costs tend to be high for the former two systems (Turton et al. 2008). The continuous configuration is also more favored for 'substrate' type solar cells on metal foil substrates in a roll-to-roll deposition (Izu and Ellison 2003). Though the comparisons in Table 2 generally holds true, it is also possible that the configuration works for one solar plant may not be the best choice of another, as each plant differs at production scale, materials supply, geological confinement and many other practical characters.

#### **5. Conclusion**

In this chapter, the cost structures of a-Si/µc-Si solar modules has been described with analysis of the multilayer cell structure and module production. The monolithically integrated structure is described with explanations of layer functions. The industrial fabrication of large-area modules are introduced, including FEOL and BEOL process steps.

Module costs around half of the total thin film PV system. We analyzed the factors affecting the module efficiency and cost in terms of energy consumption, equipment investment, spending on direct material, labor and freight cost. To probe strategies of efficiency improvement, we started from the introduction of the Si p-i-n junction structure and the front/back contacts, and discussed the light absorption and its enhancement with light trapping. The photocurrent generation is achieved by effective capture of the incident solar photons, and conversion into free electrons and holes by the build-in field of the p-i-n junction. Resistance loss during photocurrent collections is minimized by the conductive front and back contact layers. At the meantime, enhancing the light absorption within thin layers is achieved using band gap engineering of the absorbing layer and optical trapping of the front/back contact layers.

Fabrication of large-area tin film solar panels are the key to increasing the production volume and reducing the \$/Wp of modules. State-of-the-art fabrication includes FEOL and

Fortunato, E., et al. (2007). Transparent Conducting Oxides for Photovoltaics. *MRS* 

Fthenakis, V. (2009). Sustainability of photovoltaics: The case for thin-film solar cells.

Green, M. A. (2007). *Third Generation Photovoltaics: Advanced Solar Energy Conversion* 

Hegedus, S. S. & Luque, A., Eds. (2003). *Status, Trends, Challenges and the Bright Future of* 

Izu, M. & Ellison, T. (2003). Roll-to-roll manufacturing of amorphous silicon alloy solar cells

Jäger-Waldau, A. (2007). Status and Perspectives of Thin Film Solar Cell Production.

Johnson, T. R., et al. (2003). Investigation of the Causes and Variation of Leakage Currents in

Jorgensen, G. J., et al. (2006). Moisture transport, adhesion, and corrosion protection of PV

Kambe, M., et al. (2009). Improved light-trapping effect in a-Si:H / µc-Si:H tandem solar

Kessels, W. M. M., et al. (2001). Hydrogenated amorphous silicon deposited at very high

Klein, S., et al. (2002). High Efficiency Thin Film Solar Cells with Intrinsic Microcrystalline

Kolodziej, A. (2004). Staebler-Wronski effect in amorphous silicon and its alloys. *Opto-*

Kroll, U., et al. (2007). Status of thin film silicon PV developments at Oerlikon solar.

Luft, W. & Tsuo, Y. S. (1993). *Hydrogenated amorphous silicon alloy deposition processes*. CRC

Mai, Y., et al. (2005). Microcrystalline silicon solar cells deposited at high rates. *Journal of* 

Meier, J., et al. (2010). From R&D to Large-Area Modules at Oerlikon Solar. *Materials Research Society Symposia Proceedings* Vol. 1245: pp. 1245-A01-02, ISSN 0272-9172 Meier, J., et al. (2007). UP-scaling process of thin film silicon solar cells and modules in

*Applied Physics* Vol. 97, No. 11: pp. 114913-114912, ISSN 2158-3226

*on Photovoltaic Energy Conversion, WCPEC-4*, pp. 1720-1723

*Electronics Review* Vol. 12, No. 1: pp. 21-32, ISSN 1230-3402

*Renewable and Sustainable Energy Reviews* Vol. 13, No. 9: pp. 2746-2750, ISSN 1364-

*Solar Electricity from Photovoltaics*. John Wiley & Sons Inc., ISBN 0-471-49196-9,

with in situ cell performance diagnostics. *Solar Energy Materials and Solar Cells* Vol.

*Processings of 3rd International Photovoltaic Industry Workshop on Thin Films*, EC JRC

Amorphous Silicon P-I-N Diodes. *Materials Research Society Symposium - Proceedings*

module packaging materials. *Solar Energy Materials and Solar Cells* Vol. 90, No. 16:

cells by using high haze SnO2:F thin films. *Processings of Photovoltaic Specialists Conference (PVSC), 2009 34th IEEE*, pp. 001663-001666, Philadelphia, USA, June 2009

growth rates by an expanding Ar-H2-SiH4 plasma. *Journal of Applied Physics* Vol.

Silicon Prepared by Hot Wire CVD. *Materials Research Society Symposia Proceedings*

*Processings of 22nd European Photovoltaic Solar Energy Conference*, pp. 1795-1800,

industrial pecvd kai systems. *Conference Record of the 2006 IEEE 4th World Conference* 

*BULLETIN* Vol. 32, No. 3: pp. 242-247, ISSN 0883-7694

Springer, ISBN 978-3540265627, New York

78, No. 1-4: pp. 613-626, ISSN 0927-0248

Vol. 762: pp. A7.7.1-A7.7.6, ISSN 0272-9172

89, No. 4: pp. 2404-2413, ISSN 2158-3226

Vol. 715: pp. A26.22, ISSN 0272-9172

Press, ISBN 978-0824791469, New York

Mehta, S. (2010). Thin film 2010: market outlook to 2015

Milan, Italy

Chippenham, Great Britain

Ispra, November 22-23, 2007

pp. 2739-2775, ISSN 0927-0248

0321

BEOL process steps. In the FEOL processes, glass substrates are subsequently coated with functional layers, i.e., the a-Si/µc-Si layers by PECVD, TCO and reflector layers are grown by PVD or CVD. The monolithically integrated module structure is achieved by laser scribing of individual layers. In the BEOL processes, the panels are cut and encapsulated. Electrical wiring are also finished in the BEOL steps. The batch, linear, and hybrid process flow schemes are compared with actual factory examples.

Thin film a-Si/µc-Si solar panels have been holding the largest market share among all produced thin film panels. The power conversion efficiency of these panels is likely to increase to above 12% in the near future, but not exceed that achieved in crystalline cells. Advantages such as large-area, low-cost fabrication, and demonstrated field performance, nevertheless, render a-Si/µc-Si thin film technology attractive for large-area deployment like in solar power plants. In particular with the uncertain elemental supply becomes an issue for CdTe and CIS cells that might impair the sustainability of those PV products (Fthenakis 2009), thin film a-Si/µc-Si is likely to have long-term potential for providing energy supply in an even larger scale. Improvements on efficiency and stability would continue to drive the research in this area, while panel manufacturing will continue to be optimized for achieving lower production cost and optimal \$/Wp.

#### **6. References**


BEOL process steps. In the FEOL processes, glass substrates are subsequently coated with functional layers, i.e., the a-Si/µc-Si layers by PECVD, TCO and reflector layers are grown by PVD or CVD. The monolithically integrated module structure is achieved by laser scribing of individual layers. In the BEOL processes, the panels are cut and encapsulated. Electrical wiring are also finished in the BEOL steps. The batch, linear, and hybrid process

Thin film a-Si/µc-Si solar panels have been holding the largest market share among all produced thin film panels. The power conversion efficiency of these panels is likely to increase to above 12% in the near future, but not exceed that achieved in crystalline cells. Advantages such as large-area, low-cost fabrication, and demonstrated field performance, nevertheless, render a-Si/µc-Si thin film technology attractive for large-area deployment like in solar power plants. In particular with the uncertain elemental supply becomes an issue for CdTe and CIS cells that might impair the sustainability of those PV products (Fthenakis 2009), thin film a-Si/µc-Si is likely to have long-term potential for providing energy supply in an even larger scale. Improvements on efficiency and stability would continue to drive the research in this area, while panel manufacturing will continue to be optimized for

Agashe, C., et al. (2004). Efforts to improve carrier mobility in radio frequency sputtered

Applied Materials, Inc. (2010). Applied Materials Solar, Last accessed 2011, Available from

Beck, N., et al. (1996). Mobility lifetime product---A tool for correlating a-Si:H film

Berger, O., et al. (2007). Commercial white paint as back surface reflector for thin-film solar

Beyer, W., et al. (2007). Transparent conducting oxide films for thin film silicon photovoltaics. *Thin Solid Films* Vol. 516, No. 2-4: pp. 147-154, ISSN 0040-6090 Bhan, M. K., et al. (2010). Scaling single-junction a-Si thin-film PV technology to the next

Borrajo, J. P., et al. (2009). Laser scribing of very large 2,6m x 2,2m a-Si: H thin film

Bourzac, K. (2010). Scaling Up Solar Power. *Technology Review* Vol. 113, No. 2: pp. 84-86,

Bruno, G., et al. (1995). *Plasma Deposition of Amorphous Silicon-Based Materials (Plasma-*

Droz, C., et al. (2000). Electronic transport in hydrogenated microcrystalline silicon:

level. *Photovoltaics International* Vol. 7: pp. 101-106, ISSN 1757-1197

*Materials Interactions)*. Academic Press ISBN 978-0121379407

aluminum doped zinc oxide films. *Journal of Applied Physics* Vol. 95, No. 4: pp. 1911-

properties and solar cell performances. *Journal of Applied Physics* Vol. 79, No. 12: pp.

cells. *Solar Energy Materials and Solar Cells* Vol. 91, No. 13: pp. 1215-1221, ISSN 0927-

photovoltaic modules. *Processings of Spanish Conference on Electron Devices, 2009*, pp.

similarities with amorphous silicon. *Journal of Non-Crystalline Solids* Vol. 266-269,

flow schemes are compared with actual factory examples.

achieving lower production cost and optimal \$/Wp.

http://www.appliedmaterials.com/technologies/solar

1917, ISSN 2158-3226

9361-9368, ISSN 2158-3226

402-405, 11-13 Feb. 2009

No. Part 1: pp. 319-324, ISSN 0022-3093

ISSN 1099274X

**6. References** 

0248


**17** 

*1Japan 2Canada* 

**Novel Deposition Technique for Fast Growth of** 

**for Thin-Film Silicon Solar Cells** 

*1Department of Functional Material Science & Engineering,* 

*2Current address: Advanced Photovoltaics and Devices,(APD) Group,* 

Jhantu Kumar Saha1,2 and Hajime Shirai1

*Edward S. Rogers Sr. Department of Electrical and Computer Engineering, University of Toronto,* 

*Faculty of Engineering, Saitama University,* 

**Hydrogenated Microcrystalline Silicon Thin-Film** 

The microcrystalline silicon material is reported to be a quite complex material consisting of an amorphous matrix with embedded crystallites plus grain boundaries. Although this material has a complex microstructure, its optical properties have a marked crystalline characteristic: an optical gap at 1.12 eV like c-Si. This implies the spectral absorption of µc-Si:H covers a much larger range than a-Si:H which posses an optical gap between 1.6 and

**2. Growth techniques of hydrogenated microcrystalline silicon** 

The growth of µc-Si:H material uses silane (SiH4) and hydrogen as source-gas. It is currently admitted that free radical precursors (SiHx)-SiH3 is suspected to favor the µc-Si:H growthand H-enhances crystalline growth by etching of looser a-Si:H tissue-were needed to attain microcrystalline growth. In order to obtain such reactive species, decomposition of the source-gases is necessary. At first, this was obtained by using PE-CVD at high temperatures (600°C). The use of low deposition temperatures of 200-300°C with a plasma present in the

. Compared to a-Si:H that absorbs light up to 800 nm, µc-Si:H absorbs light coming from a wider spectral range, extending up to 1100 nm . On the other hand, within its range of absorption, the absorption of a-Si:H is higher than that of µc-Si:H –due to the indirect gap of the latter. Therefore, the optical combination of these two materials takes advantage of a larger part of the solar spectrum (compared to a single-junction cell) and the conversion efficiency of the incident light into electricity can be consequently improved. Furthermore, the µc-Si:H solar cell is reported to be largely stable against light induced degradation and enhanced carrier mobility in contrast to amorphous silicon films counterpart. Consequently hydrogenated microcrystalline silicon is one of the promising materials for application to

**1. Introduction** 

thin-film silicon solar cells.

1.75eVi

