Final Report Summary - LARCIS (Large-Area CIS Based Thin-Film Solar Modules for Highly Productive Manufacturing)
The overall objective of the 'Large-area CIS based thin-film solar modules for highly productive manufacturing' (LARCIS) project was to develop advanced manufacturing technologies for copper, indium and selenium (CIS) thin-film solar modules both for the electrodeposition and coevaporation approach. This comprised transfer from laboratory scale to production level. Goal was to improve the manufacturing techniques for low-cost, stable, efficient and environmentally harmless CIS thin film solar modules on large area. This included work on the molybdenum back contact, the buffer layer, the CIS absorber and the quality and process control. Special emphasis was placed on the development of cadmium-free modules and of electrodeposition methods for CIS absorbers of relevant area.
Main objectives were:
- to develop a high quality back contact on large areas with high adhesion, high conductivity and reduced corrosion by humidity;
- to develop and provide a scalable, cost-effective Cd-free CBD buffer for large areas which is applicable both for the vacuum and the ED CIS approach;
- to develop and implement CIS co-evaporation methods which are suitable to enhance module efficiencies on large area in a production line;
-to demonstrate a working CIS production process by electrodeposition (ED) on an area of 30 x 30 cm2;
- to advance and apply non-destructive process and quality control methods during production and for finished modules.
TiN as well as ZrN have been found to strongly reduce or even inhibit Na diffusion. Na has therefore to be supplied either by a precursor layer or by a Na post treatment.
A high resistivity Mo layer with a high degree of intrinsic porosity is more prone to oxidation, but is less selenised in its as deposited or argon annealed state. A low resistivity Mo layer with a more closed structure is much more severely selenised in its un-annealed or Ar annealed state. However oxidation of the surface layer reduces the selenisation substantially, leaving the oxide almost intact.
Results have shown that even a thin Mo-layer of about 10 nm is sufficient to ensure same photovoltaic properties as with 'usual' thicknesses. Even though good reflective properties could be obtained for a ZrN layer, the losses, both electrical losses as well as optical could not completely be avoided. Calculations performed in another work package regarding the influence of efficiency for the cost structures in CIGS solar cell modules gave clear indications that the addition of an extra layer together with a small loss in efficiency could not make up for the reduction in CIGS material cost and increased throughput in the CIGS deposition step.
The TiN layers were effective barrier layers both for the copper and the sodium diffusion. For the final assessment of the new back contacts more work has to be done, especially regarding Na provision, film adhesion and patterning which exceed the purpose of this project.
The high potential of evaporated In2S3 buffer layers in an industrial process could be shown and the ability to form high efficiency devices and high quality junctions has been demonstrated. However, some obstacles still have to be addressed in view of an industrial implementation which mainly concerns the current collection and the cycle time.
The density for the cells prepared with 360 nm CIGS was about 22.5 mA/cm2. A 9 % relative improvement of the density would lead to 24.5 mA/cm2. This is substantially less than about 30 mA/cm2 which is normally achieved for a CIGS thickness of 1.5 µm.
Reduced thickness of the CIGS layer has been tried both for full sized modules and for module stripes. The density decreases for the modules with decreasing CIGS thickness, but in the case with the thinnest CIGS layer also the voltage and fill factor are affected negatively.
The influence of reduced thickness on solar cell efficiency was studied for two cases of Cu content bins as measured by XRF. The influence of the thickness reduction was observed in a reduced current density, but there was only a minor influence in the voltage and fill factor. In total, the efficiency was reduced by the thickness reduction. Best full size module at 1.4 micrometer CIGS thickness had an aperture area efficiency of 11.3 %, corresponding to a 76 W module.
The development of the CuInS2 technology has allowed the creation of NEXCIS which will improve the performance of CIS cell on a large scale. IRDEP enhances its activities of Ga insertion by modifying its procedures. At the laboratory scale of small area cells, the record efficiency is 11.4 % and 6.8 % on 30 x 30 cm2. The aims of this work package were:
- to increase these numbers to at least ? = 14 % and ? = 10 %, respectively;
- to demonstrate and improve the electrodeposition process on 30 x 30 cm2 in terms of reproducibility, robustness of the process and yield.
The first results using stack approach followed by annealing in sulphur atmosphere are obtained on a 5x5 cm2 cell. In the centre of the cell the performances are better than at the edge where the efficiency can be 3.3 %.
The understanding of bath stability, influence of Mo resistivity and mass transfer phenomena has given the opportunity for IRDEP to create a pre-pilot line production working with two approaches of electrodeposition: codeposition and stack approach and a process on CuInS2 cells. In the fourth period of LARCIS project NEXCIS has presented results on cells from 5x5 cm2 to 15x15 cm2 and development of electrodeposition on 30x60 cm2 scale. In spite of inhomogeneity problems the efficiencies are over 8 % on 15x15 cm2 and 6.2 % on module 15x15 cm2.
There was a separate work package concerning technological and economical assessment. The main objectives were:
- to assess the project development with respect to yield, material consumption, throughput and commercial module performance;
- to develop a common cost estimation methodology both for the co-evaporation and electrodeposition CIS approach;
- to estimate electrodeposition and co-evaporation cost production.
During the course of the project, a Cd-free technology was developed. According to the reported results, this process improves the efficiency of CIGS solar module. In addition, the results form graded CIGS process experiments also show improvement of the efficiency of CIGS solar module. Thinning down the CIGS layer in a solar module lowers the cost of the CIGS absorbing layer deposition but do not improve the efficiency of the CIGS. It was found that the electrodeposition of CIGS lowers the material and equipment costs for the absorber layer deposition. Applying a Cd-free layer on the electrodeposited CIGS improves the efficiency of the module.
Main objectives were:
- to develop a high quality back contact on large areas with high adhesion, high conductivity and reduced corrosion by humidity;
- to develop and provide a scalable, cost-effective Cd-free CBD buffer for large areas which is applicable both for the vacuum and the ED CIS approach;
- to develop and implement CIS co-evaporation methods which are suitable to enhance module efficiencies on large area in a production line;
-to demonstrate a working CIS production process by electrodeposition (ED) on an area of 30 x 30 cm2;
- to advance and apply non-destructive process and quality control methods during production and for finished modules.
TiN as well as ZrN have been found to strongly reduce or even inhibit Na diffusion. Na has therefore to be supplied either by a precursor layer or by a Na post treatment.
A high resistivity Mo layer with a high degree of intrinsic porosity is more prone to oxidation, but is less selenised in its as deposited or argon annealed state. A low resistivity Mo layer with a more closed structure is much more severely selenised in its un-annealed or Ar annealed state. However oxidation of the surface layer reduces the selenisation substantially, leaving the oxide almost intact.
Results have shown that even a thin Mo-layer of about 10 nm is sufficient to ensure same photovoltaic properties as with 'usual' thicknesses. Even though good reflective properties could be obtained for a ZrN layer, the losses, both electrical losses as well as optical could not completely be avoided. Calculations performed in another work package regarding the influence of efficiency for the cost structures in CIGS solar cell modules gave clear indications that the addition of an extra layer together with a small loss in efficiency could not make up for the reduction in CIGS material cost and increased throughput in the CIGS deposition step.
The TiN layers were effective barrier layers both for the copper and the sodium diffusion. For the final assessment of the new back contacts more work has to be done, especially regarding Na provision, film adhesion and patterning which exceed the purpose of this project.
The high potential of evaporated In2S3 buffer layers in an industrial process could be shown and the ability to form high efficiency devices and high quality junctions has been demonstrated. However, some obstacles still have to be addressed in view of an industrial implementation which mainly concerns the current collection and the cycle time.
The density for the cells prepared with 360 nm CIGS was about 22.5 mA/cm2. A 9 % relative improvement of the density would lead to 24.5 mA/cm2. This is substantially less than about 30 mA/cm2 which is normally achieved for a CIGS thickness of 1.5 µm.
Reduced thickness of the CIGS layer has been tried both for full sized modules and for module stripes. The density decreases for the modules with decreasing CIGS thickness, but in the case with the thinnest CIGS layer also the voltage and fill factor are affected negatively.
The influence of reduced thickness on solar cell efficiency was studied for two cases of Cu content bins as measured by XRF. The influence of the thickness reduction was observed in a reduced current density, but there was only a minor influence in the voltage and fill factor. In total, the efficiency was reduced by the thickness reduction. Best full size module at 1.4 micrometer CIGS thickness had an aperture area efficiency of 11.3 %, corresponding to a 76 W module.
The development of the CuInS2 technology has allowed the creation of NEXCIS which will improve the performance of CIS cell on a large scale. IRDEP enhances its activities of Ga insertion by modifying its procedures. At the laboratory scale of small area cells, the record efficiency is 11.4 % and 6.8 % on 30 x 30 cm2. The aims of this work package were:
- to increase these numbers to at least ? = 14 % and ? = 10 %, respectively;
- to demonstrate and improve the electrodeposition process on 30 x 30 cm2 in terms of reproducibility, robustness of the process and yield.
The first results using stack approach followed by annealing in sulphur atmosphere are obtained on a 5x5 cm2 cell. In the centre of the cell the performances are better than at the edge where the efficiency can be 3.3 %.
The understanding of bath stability, influence of Mo resistivity and mass transfer phenomena has given the opportunity for IRDEP to create a pre-pilot line production working with two approaches of electrodeposition: codeposition and stack approach and a process on CuInS2 cells. In the fourth period of LARCIS project NEXCIS has presented results on cells from 5x5 cm2 to 15x15 cm2 and development of electrodeposition on 30x60 cm2 scale. In spite of inhomogeneity problems the efficiencies are over 8 % on 15x15 cm2 and 6.2 % on module 15x15 cm2.
There was a separate work package concerning technological and economical assessment. The main objectives were:
- to assess the project development with respect to yield, material consumption, throughput and commercial module performance;
- to develop a common cost estimation methodology both for the co-evaporation and electrodeposition CIS approach;
- to estimate electrodeposition and co-evaporation cost production.
During the course of the project, a Cd-free technology was developed. According to the reported results, this process improves the efficiency of CIGS solar module. In addition, the results form graded CIGS process experiments also show improvement of the efficiency of CIGS solar module. Thinning down the CIGS layer in a solar module lowers the cost of the CIGS absorbing layer deposition but do not improve the efficiency of the CIGS. It was found that the electrodeposition of CIGS lowers the material and equipment costs for the absorber layer deposition. Applying a Cd-free layer on the electrodeposited CIGS improves the efficiency of the module.
