Final Report Summary - ALPINE (Advanced lasers for photovoltaic industrial processing enhancement)
Executive summary:
The aim of the ALPINE project it to push forward the research and development (R&D) of fibre laser systems for scribing of photovoltaic (PV) modules. The project consortium focused on a new high brilliance, high efficiency and premium beam quality laser based on Photonic crystal fibre (PCF)s. The all-around development cycle comprising of the beam source, beam delivery and manipulation, scribing processes, PV module validation have been demonstrated. The novel laser systems have been designed to fit the requirements for scribing innovative and flexible PV modules rather than standard ones based on crystalline silicon wafer. In particular, the two most appealing technological alternatives have been considered, that is cadmium telluride (CdTe) and Copper indium diselenide (CIS) technology of thin-film solar cells. Validation of the quality process has been successfully assessed. The project joined together the two exciting challenges of the laser development for advanced industrial processing, on one side, and solar energy exploitation, on the other side. Thus has constitutes a crucial opportunity to continue the innovation of European industries involved in material processing applications by employing laser technology and to consolidate PV manufacturing European position. New promising scientific and technological approaches have been investigated thus stimulating the continuous growth of the market in these strategic fields for European development.
Project context and objectives:
The aim of the ALPINE project it to push forward the R&D of fibre laser systems for scribing of PV modules. The project consortium focuses on a new high efficiency laser based on PCFs. The consortium intends to demonstrate the all-around development cycle comprising of the beam source, amplification stage, beam delivery and manipulation, scribing processes, advanced diagnostic and validation of modules. In addition the novel laser system must be designed to fit the requirements for scribing innovative and flexible PV modules rather than standard ones based on crystalline silicon wafer. In particular the two most appealing technological alternatives are considered, that is CdTe and CIS or Copper indium gallium diselenide (CIGS) technology of thin-film solar cells. Validation of the quality production process must be assessed as well.
The project joins together two exciting challenges, the laser development for advanced industrial processing, on one side, and solar energy exploitation through new materials, on the other side. Thus it constitutes a crucial opportunity to continue the innovation of European industries involved in material processing applications by employing laser technology and to consolidate PV manufacturing European position.
To achieve these aims the ALPINE consortium focuses on different targets, ranging from the laser development to the definition of new fabrication processes for PV modules, from the integration of PCF technology to the definition of validation protocol of solar modules, and many others. In particular, the WPs 1-3 are in charge for the laser prototype development, based on the new seed and pump laser diodes and on specifically designed PCFs.
The fourth WP is devoted to the production of new PV module on flexible substrates using different production technologies, such close-spaced sublimation, thermal deposition, thermal co-evaporation, electro-deposition. WPs 5-6 constitute a sort of contact point of the two previous activities, being in charge to integrate the laser prototype in an industrial scribing machine and to perform the scribing tests on the provided PV modules. WP7 validates the scribed thin film modules, both in substrate and superstrate configuration, for both CdTe and CIS, benchmarking results to the state of the art module production.
Project results:
This section provides, for each Work package (WP), a description of the main scientific and technological results achieved by the project.
WP1: Design, development and optimisation of fibre laser prototype
Objectives
The goal of this WP is to design and develop an optimised PCF based laser source for cost effective, high quality and fast processing of solar cells. The work is split in two directions: first we defined the best parameters for optimum solar cell processing. This step involves designing and prototyping of a versatile Master oscillator power amplifier (MOPA) high-power laser module working in the infrared (IR), green, and ultraviolet (UV). The MOPA device has been adjusted in terms of pulse duration and repetition rate in order to be a powerful developing tool for the solar cell process optimisation. In a second step, the results achieved in WP5 with the MOPA prototype have been used to specify and develop a high power Q-switched laser with parameter close to the optimum. This ended up with a cost effective high power fibre laser giving the best process performances.
The partners involved in WP1 are MULTITEL (WP leader), University of Parma, OCLARO, Eolite, LPKF, Quanta, and University of LJubljana.
Summary of progress towards objectives and details for each task
Task WP1.1: Development of MOPA prototypes
Two MOPA lasers based on a gain switched diode were built by MULTITEL for tests. An additional laser based on a mode-lock source and pulse picking, operating at 500 ps, down to 100 kHz has been done.
Task WP1.2: MOPA assembling and frequency conversion
The laser from MULTITEL has been delivered to Eolite for amplification tests. With the Eolite amplifier it was possible to obtain 8 W at 300 ps and 20 W for pulses between 1 and 10 ns. The laser and amplifier were sent together to LPKF for the tests in WP5. Frequency doubling has been tested by LPKF with another shorter pulsed commercial laser available. Implementation with the MOPA laser from ALPINE has been successfully done, as well.
Task WP1.3: PCF high power MOPA
The task consists in the optimisation of the MOPA structures and test of PCF for all fibred amplification.
1. A laser from MULTITEL combined with the rod amplifier from Eolite, whose fibre has been produced by NKT, has been assembled.
2. The rod solution has been used to make the solar cell scribing.
Tests with the flexible PCF have been done in the hundreds of picoseconds regime. The flexible fibres can reach less peak power than the rod due to non-linearity. Flexible fibres offer a more rugged laser architecture, as e.g. all-fibre solutions. Flexible fibres is an option for processes where less output power or larger repetition rate are needed.
Task WP1.4: New approach of Q-switching
A Single crystal photo-elastic modulator (SCPEM) device and the relative driver have been developed successfully by UNILJ. The device exhibits a short response time that makes it a good candidate for Q-switching.
Task WP1.5: Q-switch optimisation for SCPEM
Two devices from UNILJ were tested together with Eolite in a Q-switch laser cavity. The results that have been obtained are the following: pulse duration of 26 ns, repetition rates of 183 or 274 kHz and average output power 47 or 67 W respectively, depending on the used device.
Task WP1.6: High power Q-switch with frequency conversion
Eolite has developed a Q-switched high power laser that can work in the IR, green and UV. This is the first Q-switched fibre laser in the UV.
The consortium has done extensive tests on lifetime of the components and particularly of UV optics and crystal. Eolite has demonstrated over 250 h at 23 W average power in the UV. A green and IR version of the Q-switched laser has been delivered to the UniPR for scribing tests.
Significant results
The prototype from Quanta operates in the nanosecond regime from 10ns to few microseconds or more. They have been amplified to a few Watt and done with a linearly polarised output.
The laser from MULTITEL operates in a shorter pulse duration regime (300 ps to 12 ns) and, combined with the amplification module from Eolite, delivers 8 to 20 (depending on the pulse duration) at a repetition rate of 500 kHz.
These are limit values that we decided to respect to avoid any risk of damaging but the laser could work at higher power. It will be for the operator to take care not going beyond those limits. It is explained in the manual of operation.
Rod amplification and SHG tests have been performed: amplification tests (at 60 kHz) with a 50 micrometre core diameter rod-type fibre have been successfully performed.
The maximum output power was 6.86 W (with Raman effects appearing) for 50 W pump power. At 100 kHz we reached 9 W and we tested SHG with a Lithium triborate (LBO) crystal. We got 42 % conversion to the green. At 300 kHz we had 13 W after amplification.
Explain the reasons for deviations from annex I and their impact on other tasks as well as on available resources and planning.
The only deviation from annex 1 is related to the integration of a flexible PCF in the amplified picoseconds laser system. The reason for this deviation comes from the fact that the energy and peak power required for the solar cells micromachining is too high for the flexible PCFs. On the other hand, the results obtained with the ROD type amplifier are well satisfactory for the project.
In addition, MULT, SSE and UniPR contributed to scribing test activity. This action was not originally planned for these partners and was proposed to catch up delay in scribing tests. For this aim, MULT used also an in-house available system at 250 ps while SSE and UniPR used a fibre laser prototype developed by UNIPR and Quanta and a green/IR version of the Q-switched fibre laser developed by Eolite.
Explain the reasons for failing to achieve critical objectives and / or not being on schedule and explain the impact on other tasks as well as on available resources and planning.
A funding reallocation has been proposed and accepted by the Commission for taking into account the corrective actions for catching up the original planning about laser scribing tests.
WP2: Development of active optical components
Objectives
The purpose of this WP was developing active optical components, such as seed and pump lasers, for fibre based laser source. For cost efficiency and also for possibility of applying high speed modulation, the seed lasers were based on a Distributed feedback (DFB) 1 064 nm laser diode. The DFB laser diode design is best tailored for the narrow linewidth, low noise, and short pulse operation required for seeding of high power fibre lasers and for frequency conversion. We designed, developed, and manufactured a narrow linewidth seed laser module with 10xx nm DFB laser diode. The seed laser modules were characterised and optimised for high frequency operation. Additionally, multi-mode pump laser modules were adapted to include a dichroic filter in order to provide protection against fibre laser radiation traveling in the backward direction. Partners involved in WP2 are: OCLARO (WP leader), University of Parma, MULTITEL, Quanta, Eolite, LPKF, and University of Ljubljana.
Summary of progress towards objectives and details for each task
During the duration of ALPINE project we:
1. Successfully developed the building blocks of the manufacturing process for the DFB laser diodes (such as grating lithography, etching, and overgrowth).
2. Designed and manufactured highly efficient DFB lasers operating in 10xx nm spectral region.
3. Optimised DFB lasers for high frequency direct modulation.
4. Carried out optimisation of module package assembly for HF operation of DFB laser diodes.
5. Built laser module prototypes with narrow stripe DFB lasers and single mode polarisation maintaining fibre pigtails. CW operating powers of more than 200 mW at 400 mA out of the fibre is achieved.
6. Performed final test of seed laser modules optimised for ns pulse modulation using fast pulse driver at MULTITEL. Modulation of the DFB laser modules with subnanosecond pulses was demonstrated. The output spectrum modulated with the short pulses exhibited wavelength stabilisation even for such short pulses.
7. Developed single emitter multimode 976 nm pump modules with dichroic pump protection against back-reflected fibre laser emission at 1060 nm. The protection is realised by applying the dielectric coating to one of the discrete optical elements inside the module, which are coupling the light into the multimode fibre. We achieved approximately 40 dB suppression for 1 060 nm radiation.
8. Built several prototypes of single emitter 976 nm pump modules with dichroic pump protection.
9. The prototypes of both DFB seed lasers and pump modules were delivered to the project partners developing the fibre lasers (University of Parma, MULTITEL, Quanta, Eolite, and University of Ljubljana).
Task WP2.1: Device design of the DFB laser diode structure and development of manufacturing processes (months: 1-6)
One of the activities was designing the single mode DFB lasers. Bragg gratings were intended for the emission wavelength of 1 064 nm. For manufacturing the DFB laser at least 2 steps of epitaxy are necessary. During the first step so called laser bottom epistructure is grown, which is based on Oclaro's high-power single mode InGaAs / AlGaAs SQW laser diodes. Then the DFB grating should be defined and finally overgrown during the second epitaxial step.
At this stage of the project we considered both first and second order gratings defined using either e-beam or holographic lithography correspondingly. Gratings were etched using the Reactive ion etching (RIE) kit. The overgrowth step was optimised in order to preserve the gratings and at the same time to produce good structural quality material.
The narrow stripe ridges were later processed using the standard process well established for Oclaro's narrow stripe laser products.
Task WP2.2: Manufacturing and testing of the DFB laser diodes and packaging of LF modules (months: 6-12)
D2.1: Report on the CW performance of a DFB seed laser and delivery of initial samples in non-optimised (LF) package (month 12).
Here we exploited 2 grating designs for 1 060 nm wavelength: 1st order DFB defined using e-beam lithography and 2nd order DFB defined using holographic lithography. The wafers with gratings were processed and laser diodes for both designs were tested. We found that the first order DFB lasers and second order DFB lasers have shown very similar electro-optical performances.
In order to complete the task, several DFB lasers were packaged in low-frequency (LF) 'butterfly modules with single mode polarisation maintaining (PM) fibre pigtail. SM PM fibre is required for easy implementation of DFB seed modules into the MOPA fibre laser systems. The laser emission was coupled into the single mode polarisation maintaining fibre using anamorphic fibre lens. Power of >200 mW out of the fibre is achieved. Fibre coupling efficiency is approximately 75 %. The spectral characterisations of laser radiation out of the fibre have shown that the lasers are locked to the DFB wavelength for all investigated conditions and the side mode suppression ratio (SMSR) is > 50 dB at I > 250 mA, confirming the single longitudinal mode operation.
Task WP2.3: High-frequency packaging for ns pulsing of DFB laser diodes and delivery of pump module with dichroic pump protection (months: 1-18)
D2.2: Report on module package assembly optimised for HF operation and delivery of pump modules with dichroic pump protection (month 18).
For this task we have compiled an equivalent electrical circuit for Small signal modulated (SSM) forward biased laser operating above the threshold in butterfly package. Our model predicts possibility to modulate up to 500 MHz for our standard packaging scheme. We identified the ways to improve the modulation speed up to 1 GHz and higher.
The second part of the project was devoted to introduction of the fibre laser pump protection against fibre laser 1 060 nm radiation coupled back into the modules. The pump laser radiation with the wavelength in the range between 915-980 nm is coupled into the multimode fibre using Fast axis collimating (FAC) and Slow axis collimating (SAC) lenses. The high reflectivity coating for the wavelength > 1010 nm was deposited on one side of the SAC lens, closest to the fibre entrance. Such coating prevents the back-reflected fibre laser emission coming out of the fibre from focusing back into the laser waveguide and consequently from destroying the pump laser.
Using this design we built modules with protection from the fibre laser radiation. We experimentally determined the suppression of 1060 nm radiation by approximately 40 dB.
Task WP2.4: Complete pulsed operation characterisation, final reporting, and delivery of prototypes in high frequency capable packages (months: 19-24)
D2.3: Final performance report and prototype delivery (month 24)
During the final 6 month we focused on improving the capability of developed DFB laser devices for high-frequency modulation. The devices were thoroughly characterised using both SMM experimental set-up build at Oclaro and high current sub-nanosecond pulse modulation using the set-up at MULTITEL.
By changing the chip design and thus reducing the parasitic resistance we were able to extend the SSM bandwidth from approximately 1 up to 4 GHz.
As the next step the DFB lasers were assembled in the butterfly module package with single mode polarisation maintaining fibre pigtail and were tested using the sub-nanosecond current pulses at MULTITEL.
Current pulses adjustable between approximately 400 ps and 4 ns were applied to the laser module. The DFB laser in the module optimised for HF operation was capable of producing optical pulses shorter than 1 ns with the wavelength locked at approximately 1062 nm.
Significant results
1. manufacturing process of DFB lasers is established at Oclaro;
2. high performance 1 060 nm DFB lasers are developed;
3. High-speed modulation of the module packaging is demonstrated;
4. carried out optimisation of DFB laser design for ultrashort pulse modulation and successfully demonstrated wavelength stabilisation for the DFB laser packaged in the butterfly module driven with subnanosecond pulses;
5. approximately 40 dB suppression of the 1 060 nm radiation using dichroic filter is demonstrated experimentally;
6. single emitter pump modules are assembled with such filters for protection against the fibre laser radiation.
Explain the reasons for failing to achieve critical objectives and / or not being on schedule and explain the impact on other tasks as well as on available resources and planning. Development of laser pump modules with the dichroic filters was delayed (approximately 4 month) because of difficulties in designing stop band of the filter on the curved surface without affecting the lens optical throughput efficiency at the laser wavelength. Several design iterations were necessary to achieve the required result. Meantime we provided standard laser pump modules without protection to our partners and the pump protection was achieved using the external filter. Therefore it was no impact on the other tasks of project.
Resources were spent according to the project plan. However the actual spending was lower than planned. Therefore part of the funding for WP2 project was transferred to other WPs.
3. WP3: PCF design, production and testing
Objectives
The aim of WP3 is to develop active, Large mode area (LMA) fibres and interfacing technologies for Q-switched and MOPA laser configurations. The WP is closely linked to WP2 for seed- and pump sources and to WP1 for complete laser integration and system tests.
The goal is to leverage the stand-alone PCF technology to monolithic solutions enabling high-power pulsed fibre lasers to enter the solar cell market. Objectives for the WP are:
- optimising flexible LMA fibres for highest possible power performance;
- develop interfacing technologies that allow simple access to PCFs;
- produce non-flexible (rod-type) fibres enabling new laser tools for solar cell materials processing.
The partners involved in WP3 are: NKTP (WP leader), UNIPR, EOL, Oclar, UNILJ and MUL.
Summary of progress towards objectives and details for each task: The development in WP3 combines numerical simulations from the University of Parma, fibre drawing at NKT Photonics A / S (NKTP), and fibre laser expertise from Eolite and Multitel.
Task WP3.1 - Flexible active LMA fibre with conventional fibre pigtail (months: 1-12): The tasks in WP3.1 have been concluded and the results are described in deliverable report D3.1 which was delivered in time. The work was headed by NKTP with inputs from UNIPR and EOL.
Task WP3.2 - Development of modelling tools (months: 1-20): The tasks in WP3.2 have been concluded and the results are described in deliverable report D3.2 which was delivered in time. The work was headed by UNIPR with inputs from NKTP and EOL.
Task WP3.3 - Monolithic pump-signal combined, flexible active LMA fibre (months: 20-26): The tasks in WP3.3 have been concluded and the results are described in deliverable report D3.3 which was delivered with a delay of five months. The delay on the combiner was not critical for the ALPINE progress, as the fibre laser prototype for solar cell scribing in WP1 was made without it.
Task WP3.4 - Rod-type active LMA fibre (months: 22-30): The tasks in WP3.4 have been concluded and the results are described in deliverable report D3.4 which was delivered in time. The development is based on a patent pending rod design developed in the Seventh Framework Programme (FP7) project LIFT, whence an agreement was made to transfer the foreground knowledge. The results showed that thermal effects have a significant impact on rod performance. During month 30 progress meeting, it was thus decided to try and reallocate funds for an additional activity on an ROD type fibre incorporating thermal effects. This activity is described below.
Task WP3.5 - Rod type active fibre incorporating thermal effects (months: 30-34): This task has been added to the originally planned WP3 activity by the GA amendment. The tasks in WP3.5 have been concluded and the results are described in deliverable report D3.5 which was delivered in time. The new rod has significantly better performance than PCF#3 and NKTP expects to launch it as a commercial product later in 2012.
Significant results
The development up to the fabrication of PCF1 (D3.1) has resulted in an improved production method, which has been implemented in NKTPs top selling LMA commercial fibre product DC-200-40-PS-Yb. This has matured the fibre and it is expected to be transferred from R&D to production in 2012.
A major obstacle for wide spread commercialisation is that the fibre is difficult to use. Accordingly the pump / signal combiner to be developed in task 3.2 should increase the addressable market of the fibre significantly.
NKTP started developing a backward pumped combiner in 2010, and made a prototype with excellent performance. However, it was not possible to reproduce this combiner (see D3.3 for details). Independent development in WP1 by Multitel and Eolite showed that the flexible fibre platform was not able to handle sufficient power for making the ALPINE laser demonstrators. Thus, NKTP did not send the single good multiplexer to EOL, but instead sought for a reproducible combiner solution. This involved changing from a backward - to a forward pumped combiner and outsourcing part of the development to a third source. The details of the combiner are reported in D3.3 which was delivered in Feb-2012. Since then, NKTP has shown that the combiner works very well in a ps amplifier (80 MHz, 5 ps at 1 064 nm), where amplification up to 80W output power has been demonstrated.
The combiner was sent to EOL on 29 June 2012, where it will be tested in a ns amplifier in the last months of the project. Subsequently NKTP will sample key customers with starting from Q3 2012 and it is expected that it will be made generally commercially available in 2013.
The results from PCF2 (D3.2) have so far not resulted in a commercial product. It has been shown that the fibres can successfully suppress non-linear effects, such as ASE. Thus they potentially have better performance than conventional fibres for operation at 1 064 nm or higher wavelengths, and for pulsed systems with low repetition rate. So far it is unclear whether the improved performance is sufficient to justify the additional fabrication complexity. Accordingly, NKTP have per 1 January 2012 started a doctorate (PhD) project with the aim of maturing the fibres. It is still to goal to release these fibres as a commercial product in 2013. The ROD fibre developed for PCF3 (D3.4) showed large promise, but was difficult to reproduce and had issues with thermal effects. It was proposed as a corrective action to reallocate funds to NKTP and UNIPR to develop a fourth fibre in a new task 3.5.
The resulting rod fibre has been fabricated in early June 2102 and did indeed have the desired properties:
- wide single mode guidance region from 1 030 to 1110 nm;
- high power operation with a single mode output;
- 175 W output shown;
- excellent output power stability: over 200h test with 50 W output.
Thus, NKT Photonics is planning to launch it as a commercial product later in 2012 (for further details see report D3.5).
WP3 had furthermore led to two master projects at the University of Parma, one master project at The Technical University of Denmark (DTU) with NKT as co-supervisors. Furthermore DTU and NKTP have in January 2012 started a PhD project to further develop the fibre concept, which was used for D3.2. The work in WP3 has been disseminated in 5 journal publications and 13 talks at peer reviewed conferences. The dissemination has mainly been driven by UNIPR.
WP4: Thin film solar cell production
Objectives
The aim of this WP is to prepare and supply to all the other partners samples of solar cells of both CdTe / Cadmium sulfide (CdS) and CuInGaSe2 / CdS deposited on different substrates such as glass, polyimide and metallic foils. Results on the module characterisation and validation here performed are also part of WP7 activity and will not be reported once again in WP7 section. The partners involved in WP4 are: SSE (WP leader), NEXCIS, SSW, WS, UNIVR.
Summary of progress towards objectives and details for each task:
Within the WP, there are two main groups: one made up of industrial partners and the other are of research laboratories. The first part consists of SSE, NEXCIS and WS which respectively produce: CdTe PV modules based on rigid substrates, CIGS PV modules on rigid and flexible substrates, CIGS modules on glass substrates. The second part consists of SSE, SSW and University of Verona which respectively produce: CdTe, CIGS and CIS solar cells both on rigid (glass and ceramic) and flexible substrates, CIGS solar cells and modules both on rigid and flexible substrates and CdTe both on rigid and flexible substrate.
SSE
SSE has continued his research with respect to the CdTe-based cells in substrate configuration and on CIGS-based cells both on glass and on ceramic substrates.
1. CdTe modules at different stages of fabrication for P1, P2 and P3 laser scribing. 10x10 cm2 and 30x30 cm2 mini-modules were sent to Multitel.
2. CIGS based solar cells completely fabricated by sputtering and selenisation. SSE innovation consists in the use of the sputtering technique to deposit the starting precursors (In2Se3 or InSe e Ga2Se3 or GaSe) and a consequent selenisation in pure Se atmosphere.
With the innovative method of deposition of the composite precursors developed by SSE and described before an important result was obtained by using as a substrate ceramic commercial tiles. A PV conversion efficiency of 13.95 % was reached and this is one of the highest value ever obtained on ceramic substrates.
In this period SSE produced a lot of CdTe and CIGS samples used for the laser scribing test. In fact, by September 2011, SSE installed a machine for laser scribing in the Thin Film Laboratory (ThiFiLab) of the University of Parma.
In a first stage, the scribing machine was equipped with a conventional solid-state laser, which works both at 1064 and 532 nm. Later on the scribing system was equipped first with a fibre laser built up by the laboratory of Stefano Selleri at the University of Parma and secondly by a fibre laser produced by Eolite.
The first fibre laser installed worked only at 1 064 nm. Some results so far obtained on TCO samples normally used in CdTe/CdS solar cell fabrication are shown in the following. Results of comparison of a P1-scribing performed by a commercial solid-state laser and a P1-scribing carried out with the UNIPR laser show the great finesse difference of the fibre laser. Also P2 and P3 scribings tests, carried out on CdTe/CdS modules by using Eolite fibre laser, shown very clean and sharp edges and they are not affected by the thermal damage (HAS) usually observed in CdTe standard manufacturing.
CdTe mini-modules, with an area of 10x10 cm2, by testing a fibre laser scribing at a time were made. This allowed us to characterise in detail the single scribing carried out with new fibre lasers by finishing the module with traditional scribing.
With Eolite's fibre laser many tests on CIGS-based mini-modules (dimension of 10x10 cm2) sent from SSW have been performed. It has been thought to proceed as for CdTe-based minimodules. Experiments are underway and we have been successfully performed the P1 scribing (at 1064 nm) and P2 (at 532 nm). Some 10x10 cm2 CIGS-based minimodules equipped with P1 scribing directly made by SSW were tested in our laboratory for P2 scribing and then re-sent to SSW that will carry them out by running the P3 scribing in traditional way (mechanical scribing). In this way it will be clear the behaviour of the P2 scribing realised with the Eolite fibre laser at 532 nm on CIGS-based mini-modules in respect to standard P1 and P3 traditional scribing. The electrical characterisation of the different processing steps certifies in all cases an efficiency around 10 %.
NEXCIS
NEXCIS started to provide electrodeposited based CIS on glass at month 2 (2 months before the compromise acquired in the project) and CIGSe at month 12. During the last months NEXCIS has validated and increased the quality of the absorbers and devices that were scaled up to 60x120 cm2 size. Wafers of this size (and size smaller than this) have already been sent to laser partners, especially for testing of the automated machine. Moreover, complete modules have been sent to JRC for measurement of standard modules to be used as reference at NEXCIS. More samples have been sent to LPKF for final scribing and final modularisation at NEXCIS.
Here the results acquired in the project activity are reported. The best efficiency is 13.9 % (by Newport in the US) for a cell (approximately 0.5 cm2), medium efficiency is 12.2 for 95% of the wafer surface (for a 30x60 cm2 wafer), and 11.6 % for 100 %. An efficiency of 8.2 % for a 30x60 cm2 module was communicated. Today, efficiency for a pixel is certified at 14 % (by Newport in the US) for a cell (approximately 0.5 cm2), in case of internal measurements we have measured a cell at 14.9 % (under certification measurements), medium efficiency in a wafer is 13.2 for 95 % of the surface of the wafer, and 13.7 % for 100 %. There has being also a big reduction in the gap between wafer and modules, and 30x60 cm2 modules are now obtained in a repeatable way with efficiencies at about 10.7 %. The record efficiency for encapsulated modules is 12.3 % as certified by Certisolis.
However what it is more important, 60x120 cm2 wafers with a medium efficiency of 11.2 % (12 % for 90 % of the wafers) are obtained in our baseline, which will be providing in near future with modules 60x120 cm2 with efficiencies larger than 10.5 %.
The process portability on flexible metallic substrate has been fully demonstrated. However, due to more work needed with the right encapsulation material, metal and polymer modules are in stand-by due to reliability issues not concerned with NEXCIS process but, more in general, to lightweight thin film objects. In this sense NEXCIS has focus all activities in glass-glass solutions.
Pre-qualification of the 60x120 cm2 reactor for electro-deposition of Cu, In and Ga is already achieved. The same pre-qualification has been achieved for the furnace developed by NEXCIS. NEXCIS is at present working in validating the yield, throughput and economic issues to go towards industrial pilot line conditions, which is foreseen at the end of 2013. At present, more than 90 employees are working at NEXCIS facilities while at the beginning of the project less than 10 employees were working at NEXCIS.
WS
Wurth Solar was able to deliver CIGS modules on large glass substrates after any relevant processing appropriate for laser scribing tests (P1-P3, edge cleaning). Wurth Solar organised to be able to respond flexibly and quickly to partners with substrates as soon as required. WS has sent standard modules (P2, P3 mechanically patterned) to JRC for validation to obtain a reference for the laser scribed samples.
During the last months of the last year, WS was partially acquired by MANS GmbH. Only in late April 2012 it was clear that this partner was yet inside the consortium. For this reason WS has considerably reduced its activity during the last months of the project. WS sent mini-modules to UniPr for scribing test in M36.
SSW
SSW has continued to deliver CIGS modules with sizes of 10 cm x 10 cm as samples for the testing of patterning lasers to the partners LPKF, MULTITEL and University of Parma. These samples included modules ready for the application of all three patterning steps. Also, additional samples carrying an array of 81 test cells for the purpose of P3 quick tests in the laser labs have been delivered.
For the purpose of having reference samples with known electrical properties, one out of each set of nine samples was completed using conventional patterning at SSW. These references reached efficiencies of up to 15 %. This improvement of about 1 % compared to previously reported reference samples is due to improved deposition parameters for the sputtering of the zinc oxide (ZnO) front contact which lead to an increase of the open circuit voltage of the devices.
Furthermore, additional modules on flexible substrates - polyimide film with a thickness of 25 micrometre - have been fabricated and sent to MULTITEL for testing patterning steps P1, P2, and P3.
Flexible substrates made from steel foil covered by a barrier layer have been prepared. This silicon oxide barrier layer, which is about 1 micron thick provides the electrical insulation of the subsequently deposited Mo back contact layer and of the individual cells. Sample have been sent to MULTITEL for P1 scribing experiments.
UNIVR
In UNIVR labs, solar cells have been fabricated by depositing each single layer (except for the front contact), namely CdS, CdTe and back contact, by vacuum evaporation. For the preparation of the samples to be scribed by the laser partners we have prepared and optimised a complete fabrication process for deposition on glass and on polymers, so by applying a lower temperature process compared to the typical ones that use substrate temperatures above 500 degrees of Celsius.
Front-contact deposition
For Transparent conductive oxides (TCO), that make the front contact of the device, two different glass coatings have been used: a commercially available ITO film coated glass with a SiO2 barrier layer and a laboratory scale bi-layer of conductive ITO + thin insulating ZnO, deposited at high temperature. The first one is a commercially prepared ITO with a thickness of 180nm and a sheet resistance of 10 Ohm / square. The second one is a 400nm ITO+100nm intrinsic ZnO deposited by radio frequency-sputtering at a temperature of 400 degrees of Celsius with a sheet resistance below 5 Ohm / square. Due to its high substrate deposition temperature this layer is much more crystallised and stable; moreover because of the ZnO layer, which acts as a barrier to diffusion of impurities, indium diffusion is avoided. A comparison of the devices made on these two different substrates will outline some of the properties observed.
Window layer fabrication
CdS is evaporated in vacuum at a pressure of 10-6 mbar using direct current heating of a molybdenum crucible and deposited on the glass / TCO stacks at a substrate temperature of 150 degrees of Celsius with a deposition rate ranging from 0.15 to 0.45 nm / sec (controlled by a quartz crystal thickness monitor). After deposition, CdS was annealed by heating the stacks in vacuum for 30 minutes (in order to increase the stability of CdS to the subsequent depositions and CdTe activation treatment), a slight re-crystallisation of the grains was observed by AFM with an enlargement of the grain size from 50-100 to 100-200 nm.
CdTe deposition and post-deposition treatment
After deposition and treatment of the CdS layer, CdTe is deposited in the same chamber by heating a special in-house made graphite crucible at a substrate temperature ranging from 300 degrees of Celsius up to 340 degrees of Celsius and deposition rate of about 2 nm / sec. A typical CdTe layer for these devices is 3.5 / 4 micron thick with quite compact morphology and grain size of about 1 micron; we have observed that the morphology and the grain size is depending not only on the substrate temperature but also on the TCO. CdTe deposited on commercial ITO and ITO+ZnO shows a similar but not same morphology, grain size is typically around 1 micron for both but it results to be much more compact in the second case. Compared to the Close space sublimated (CSS) deposited CdTe grains (which have a much higher substrate temperature) are quite small and activation treatment is needed also for increasing the grain size.
It is known that for a good cell performance an activation treatment is necessary. Generally, in order to enhance the grain size and passivate the grain boundaries in the CdTe polycrystalline structure a layer of CdCl2 is deposited on top of the CdTe and then annealed in air. We have prepared two different activation processes: the standard CdCl2 treatment and the alternative treatment with chlorine containing gases.
Back contact
Back contact was made by vacuum evaporation of Cu/Au stacks. Subsequently an annealing of the back contact is applied at 200 degrees of Celsius in air for 20 minutes in order to diffuse Cu in CdTe and make an ohmic contact. In case of CdCl2 treatment, a bromine-methanol etching is applied in order to remove the CdCl2 layer and to prepare the surface for the Cu / Au deposition, however the removal of the CdCl2 in case of wet treatment was quite difficult. In case of freon treatment, the treated CdTe film is not etched since the surface is free of CdCl2 (re-evaporated at high temperature in vacuum) and a tellurium rich layer is made during the activation process.
Deposition of copper results to be very sensitive, a higher amount of copper improves the back contact also doping the bulk CdTe but reduces the shunt resistance (Rsh) resulting in a lower fill factor. The amount of copper has been tuned for the differently treated CdTe layers. A more conductive absorber would need less copper to improve the performance.
Several PV devices have been made with different parameters of both CdCl2 and freon treatment for CdTe deposited on the different stacks described before.
The highest performance was given by CdCl2 treated cells and with stronger CdS / TCO stacks, this means that the most performing TCO was the one made with high temperature deposited ZnO / ITO with enhanced ability and prevention of sodium and indium diffusion. Moreover only 2 nanometres of copper with 50 nanometres of gold are necessary for a good contact. With this configuration efficiencies exceeding 13.5 % are routinely obtained. If freon treatment is applied, much lower efficiencies are measured: highest efficiencies are obtained with 8 nm of copper, however a much lower Rsh gives lower fill factors and lower open circuit voltages.
If a 2 nm copper contact is applied on these freon treated cells a very low efficiency is performed. Anyway with more aggressive freon treatment (by higher freon partial pressures or higher temperature) bigger grain size is obtained but not higher efficiencies. Excess of freon pressure have resulted in shunted device. So the required copper amount is also depending on the different CdTe treated layer. We have registered that in case of freon treated CdTe devices four times the amount of copper is needed for a reasonable efficiency compared to the CdCl2 treated cells. This is consistent with the fact that activation by freon is much weaker than the one made with CdCl2 as reported above. On the other hand, using a lower amount of copper allows to have a better working cell and this results in a higher efficiency.
Cells made with CdCl2 treatment exhibit efficiencies from 10 to 13.5 %, with current in the range of 22 to 25 mA / cm2, Voc exceeding 830 mV and FF from 60 to 70 %. The relatively low Voc and FF are connected with relatively low Rsh. This can be explained by an excessive consumption of CdS into the CdTe layer. The best results were obtained with thicker CdS (more than 500nm) annealed in vacuum at 450 degrees of Celsius. Freon-treated cells show lower efficiencies than CdCl2 treated ones. The highest achieved efficiency for this treatment is 8.7 %. Although the freon treated cells would not need a post deposition etching, a good ohmic back contact was not easily reproducible. Many of our freon-treated cells have shown a kink on the positive part of the curve.
Transport properties
For device characterisation two different types of cells made with vacuum evaporation were taken into consideration: one activated by freon and one activated by CdCl2. Moreover cells made by close space sublimation at the University of Parma, were successfully activated by HCF2Cl gas, with efficiencies exceeding 10 % as a reference with the low temperature deposited cells.
A standard technique which is very often applied for determination of the doping level in semiconductor junctions is capacitance voltage (C / V) profiling. However, in the presence of deep levels free carrier concentrations determined by C / V profiles can be subjected to large errors. This can be overcome by drive level capacitance profiling (DLCP), as DLCP is a fully dynamical measurement giving undistorted doping distributions. For junctions without deep traps, adjusting their charge state to the DC bias, DLCP and C/V profiles coincide, so the difference between the C / V and DLCP profile gives a lower bound for deep defect concentration active in a given measurement conditions (i.e. temperature and frequency). The capacitance measurements were performed with an Agilent E4980A LCR meter controlled by LabView via computer. Doping concentrations in the space charge region (SCR) of CSS-freon, VE-CdCl2 and VE-freon samples were estimated to be 2x1014 cm-3, 6x1013 cm-3, 3x1013 cm-3, respectively. Deep defects contributing to capacitance-voltage profiles were also detected. The difference between the C / V and DLCP profile, indicating the lower bound for the concentration of deep defects following the DC bias sweep, is also the highest for the CSS-freon sample and amounts 4x1014 cm-3. For the VE-freon device the defect concentration varies from 1x1013 cm-3 at 2 micrometre up to 2.3x1014 cm-3 at distances larger than 2.5 micrometre. In the VE-CdCl2 sample we detected no large differences between the C / V and DLCP profiles around RT.
Flexible cells
The same optimised process has been applied to the flexible polymer substrates. Prior to this, a study of the different polymers have been made. Different polymers (some of them are not on the market but still in prototype form) were taken into consideration: DuPont (Kapton), Kaneka, UBE (Upilex) and Mitsubishi (Neopulim). For Kaneka and DuPont a smaller thickness must be used. Indeed, the 25 micron thick substrates are almost brown, cutting transparency below 400 nm for Kaneka and Kapton. Mitsubishi polymers, however, have a higher transmittance at low wavelengths even if they remain opaque (probably still need to reduce the thickness).
The polyamides have been tested with an annealing in order to check their temperature resistance to the deposition process, treatments were made for 30 minutes at 370 and at 450 degrees of Celsius in air. Flexible solar cells have been fabricated by using Upilex foils12.5 microns thick. For the CdTe deposition process and post-deposition treatment, the same process optimised on glass was used also for the polymer substrate.
Morphological and crystallisation parameters have been studied with atomic force microscopy and X-ray spectroscopy. CdTe on flexible substrates grows similarly as on glass, in terms of grain orientation and size. However some restrictions come from the limited transparency of the polymer substrate. Polymer looks yellow in colour.
Significant results
- The objectives for the CdTe-based cells made on substrate configuration have been achieved.
- CdTe and CIGS mini-modules have been made in traditional way for reference.
- CdTe and CIGS mini-modules in various stages of realisation have been made and sent to all the scribing partners who have requested them.
- High efficiency CIGS-based cells made on ceramic substrate have been realised. Some of them have been used for testing the laser scribing.
- The SSE scribing work is intensively continuing both on CdTe and CIGS samples. P1, P2 and P3 on CdTe minimodules were definitively developed and optimised. Fibre-laser scribed CdTe minimodules were made and sent to JRC for characterisation. Results are very encouraging in order to use this technology for solar cell processing.
- Low temperature vacuum evaporated high efficiency CdTe solar cells both on rigid and polymeric flexible substrate were realised and cells in different stage of fabrication were sent to all the scribing partners who have requested them.
- Successful scale-up from 15x15 to 30x60 cm2 on glass substrate and preliminary to 60x120 cm2 at NEXCIS.
- Today NEXCIS efficiency for a pixel is certified at 13.9 % (by Newport in the US), medium efficiency in a wafer is 12.2 for 95 % of the surface of the wafer, and 11.6 % for 100%.
- The process portability on metallic substrate has been fully demonstrated.
- Prequalification of the 60x120 cm2 reactor for electrodeposition of Cu, In, Ga is already achieved.
Considering the great amount of exchange of modules by partners, the consortium agreed to demonstrate the feasibility of the laser patterning process by producing a series of 30 cm x 30 cm mini-modules for CIGS and 10 cm x 10 cm mini-modules for CdTe. In this way all the scribing systems can participate to the effort to build PV modules fully scribed with fibre laser. Moreover these dimensions are sufficient to demonstrate the scalability at an industrial level of the scribing made with fibre laser.
A statement on the use of resources, in particular highlighting and explaining deviations between actual and planned person-months per WP and per beneficiary in Annex 1 (Description of Work).
The use of the allocated resources has been as planned, but WS whose reduced participation during the last months resulted in a reduced actual budget.
WP5: Laser scribing and beam handling
Objectives
The aim of WP 5 is to integrate the laser source developed in WP1 in a machine concept. That includes the beam handling and new concepts for the beam splitting. On that machine industrial like laser scribing of solar panels shall be achieved.
The partners involved in WP5 are: LPKF (WP-leader), Eolite, NEXCIS, SSW, WS, UNIVR, ES.
Summary of progress towards objectives and details for each task
Task WP 5.1 - First scribing machine prototype: The scribing machine was designed and built upon requirement specifications taking into account the different technologies to be tested (CdTe and CIGS) as well as production related issues like accuracy and throughput. As a result the machine is capable of processing substrate and superstrate configurations at high speed being flexible in providing a platform for different laser sources and optical configurations.
Task WP5.2 - Test of the MOPA laser prototype: The final MOPA test laser has been integrated into the ALPINE test system in Garbsen and tested with both wavelengths 532 nm and 1 064 nm at a pulse length of approx. 500 ps on all test materials provided by the partners SSE, University of Verona, Nexcis and SSW. Tests with the state of the art picosecond laser source and the beam shaping plate have made at Multitel and using the Q-switch laser from WP1.4 SSE performed further scribing test. Samples were sent back and forth for testing and further processing and final mini-module performance compared to those manufactured with standard scribing methods.
Task WP 5.3 - Beam shaping test and Q-switched laser integration: The first approach by LPKF has been to use a special beam shaping optic providing a quasi-top-hat profile in order to reach a higher depth of focus compared to the one achievable with exact top-hat intensity profiles. The result gives a good top-hat approximation with a depth of focus around +/- 400 mm at 532 nm. This result has been compared to a fibre-based top-hat generation done by Eolite.
Task WP5.4 - Final optimisation of the scribing machine: Components of the ALPINE test system have been optimised in their design aiming at highest throughput and accuracy. The main components influencing the tact time of scribing are the scribing and glass transporting axes. LPKF has worked on two different axis designs which are currently set-up on a test bench and are going to be evaluated regarding acceleration, speed, accuracy, heating taking into account large and heavy loads as 4 mm thick substrates of up to GEN5 sizes, as well as free flying optics. SSW has tested and compared scribed test modules from all participating partners to conventional scribed ones. Laser scribing has already proven to be a promising alternative to mechanical scribed CIGS modules.
Significant results
- Laser scribing machine capable of running of up to 2 m / s on the scribing axis.
- Laser scribing machine designed to perform film-side and glass-side scribing comprising a flexible exhaust system to process CdTe modules as well as CIGS modules.
- Design and test of a beam shaping module providing good top-hat approximation with an excellent depth of focus.
- MOPA laser has been integrated in ALPINE test system and extensive scribing tests and parameter improvements on all relevant material systems have been carried out.
- CIGS samples with P2 scribed at LPKF using enhanced parameters have been completed and analysed at SSW. The I-V curves indicate a good P2 scribe quality.
- P3 laser-scribed CIGS sample at MULTITEL was comparable with conventional mechanical scribing.
WP6: Automation
Objectives
The goal of this WP is to study, develop and test standard parameters in order to produce system integration guidelines and to realise systems to focus and guide the laser beam. The partners involved in WP6 are ES (WP-Leader), Quanta and LPKF.
Summary of progress towards objectives and details for each task
WP6.1: Standards for automation integration in scribing systems
The first task is addressed into developing rules for integration and experiences in the real industrial environment. The activity were dedicated to collect information and experiences with partners and to attend important European fairs to understand needs of the market and to identify issues to be faced. In this package were assessed important parameters for the laser integration like temperature effects (gradients and stability), laser scribing panel feeding, mechanical stress, systems integration issues. The integration of the protocol developed by Quanta to manage the device they developed in an industrial network has been studied and implemented.
WP6.2: Laser beam moving and focusing system
The WP has developed a suitable focusing head achieving a stable and reliable processing of the glass. Due to changes in the focal position the spot diameter changes accordingly and so the power density on the layer to be processed. Allowing a dynamic focusing of the laser beam brings to a higher stability of this parameter and consequently of the ablation rate. The core philosophy of the compensation system is a matrix camera and a focus system developed with the collaboration of all the WP partners.
The WP also developed a reliable moving system with adequate precision in order to explore and test the feasibility of a guiding system as well as a cost effective and less precise moving system with acceptable results. The moving system moves the sample under test below a fixed laser and guiding system which is composed by two linear axes in a table configuration. Details of focusing and moving systems are reported in deliverables D6.2 and D6.3.
Significant results
A practical , computer controlled high-end dynamic guiding system has been developed that allows for tracking of the glass scribed waviness on thin films solar cells.
A microprocessor controlled high-end dynamic focusing system has been developed that allows for fast tracking of the glass deformations on thin films solar cells.
WP7: Testing of PV devices and scribing process validation
Objectives
The objective of WP7 is to evaluate the quality of PV thin film modules patterning on devices realised in different configurations (substrate / superstrate) and on different substrate materials (glass / flexible). Both CIS and CdTe modules are researched. The idea is to define a set of suitable tests leading to correctly identify potential critical aspects both at cell level and at module level. Similar measurements are also carried out on state-of-the-art solar modules with conventional processed patterning and edge cleaning, in order to provide a benchmark against which to highlight process improvements by laser use.
The partners involved are JRC (WP leader), SSW and University of Verona as research centres, together with SEE, Nexcis, Wurth Solar and LPKF as industrial partners. Multitel has also been involved in the scribing work to prepare devices.
The agreed strategy of the consortium is to demonstrate the feasibility of the PCF laser patterning process by producing a series of demonstration devices (30 cm x 30 cm mini-modules for CIGS and 10 cm x 10 cm mini-modules for CdTe). It is noted that this device production and activity is partly in common with WP4.
To enable assessment of each of the scribing steps as well as the overall working device, each combination (that is CdTe or CIGS on glass or flexible substrate, substrate or superstrate configuration, scribing steps) has included, as a minimum, a group of 4 sample devices with different characteristics according to the scribing process P1, P2 or P3. Details can be seen in the deliverables D7.5 and D7.6.
The verification testing uses the methods described in Deliverables D7.3 and D7.4. The electrical performance tests on one of the four samples include as a minimum spectral response and current-voltage (IV) characteristics at Standard test condition (STC)s. These have been complimented by dark IV, thermography and LBIC depending on time and resources are available.
Summary of progress towards objectives and details for each task: The activities performed are clearly explained in deliverables D7.5 and D7.6 and summarised as follows:
Task WP7.2 - Validation of thin film modules realised in a substrate configuration (CIS) on glass and processed by means of advanced fibre lasers: SSW has prepared and characterised devices scribed at LPKF. This work is reported in WP4.
Task WP7.3 - Validation of thin film modules realised in a substrate configuration (CIS) on flexible substrates and processed by means of advanced fibre lasers: Devices have been prepared using polyamide and steel foil substrates. Scribing trials tried to cope with the waviness and flexibility of the substrate. The process is critical.
Task WP7.4 - Validation of thin film modules realised in a superstrate configuration (CdTe) on glass substrates and processed by means of advanced fibre lasers: SSE has prepared and characterised devices scribed at the SSE/University Parma facilities. JRC has validated the electrical performance of the devices. A series of devices has been prepared by University Verona. Results are reported in the updated version of D7.6.
Task WP7.5 - Validation of thin film modules realised in a superstrate configuration (CdTe) on flexible substrates and processed by means of advanced fibre lasers: A series of devices has been prepared by University Verona. The scribing process is still critical.
Significant results
Concerning the most suitable analytical methods to evaluate the scribing operations, the following have been identified as the most feasible for a correct check on the quality of laser scribing process:
- optical microscopy;
- scanning electron microscopy;
- electrical insulation test;
- I-V characterisation of solar cells;
- thermal imaging;
- LBIC mapping.
Measurements according to the first three methodologies have been performed just after laser scribing process, in order to rapidly identify potential faults;
- I-V characterisation and thermal imaging have been then implemented at a later stage, when a completed cell or module is assembled. Tests are also considered in the field of validation of thin film modules comprising IV curve STC indoor and outdoor, stabilisation cycles, spectral response, TCO and matrix.
In particular, referring to I-V curve measurements STC conditions, a dedicated research activity is ongoing at JRC, focusing on the measurements methods for improving the compliance among indoor and outdoor measurement results.
As indicated in the above table, an extensive effort is being made to produce demonstration devices using the laser patterning techniques developed in the project.
The performance verification of the following device combinations is already at an advanced stage:
- CdTe on glass substrate with PCF laser scribing (SSE and University Parma).
- CIS on glass substrate with PCF laser scribing (SSW with LPFK, SSW with Multitel / University Parma).
In additional the performance verification of CdTe on glass substrate produced by University Verona together with LPKF and Multitel scribing using ALPINE lasers is fully completed. Results are reported in the updated version of D7.6.
Results to date indicate that PCF laser scribed devices are very promising and laser technology is very effective for glass substrate modules. Flexible substrate ones proved to be more critical.
The WP7.2-7.5 activities are largely dependent on the availability of laser-patterned devices from other WPs. The supply of these has been delayed, largely due to problems with availability of lasers for the scribing operations. At the general assembly meeting on 28 September 2011 the consortium agreed to enforce the laser scribing activity (GA amendment n.3) to catch up the delay. In addition a detailed planning to complete the work in remaining months of the project was agreed at the executive meeting held 28 March 2012, and these actions allowed the first modules to be already fabricated and validated.
The overall objective of the WP has been maintained, i.e. to demonstrate the quality of patterning / scribing of CIS and CdTe thin film technologies on glass substrate using advanced pulse conditioned lasers, however successful tests on flexible substrates have not yet been achieved. Partners are still working on it, even after the project end date.
List of websites: http://www.project-alpine.eu
The aim of the ALPINE project it to push forward the research and development (R&D) of fibre laser systems for scribing of photovoltaic (PV) modules. The project consortium focused on a new high brilliance, high efficiency and premium beam quality laser based on Photonic crystal fibre (PCF)s. The all-around development cycle comprising of the beam source, beam delivery and manipulation, scribing processes, PV module validation have been demonstrated. The novel laser systems have been designed to fit the requirements for scribing innovative and flexible PV modules rather than standard ones based on crystalline silicon wafer. In particular, the two most appealing technological alternatives have been considered, that is cadmium telluride (CdTe) and Copper indium diselenide (CIS) technology of thin-film solar cells. Validation of the quality process has been successfully assessed. The project joined together the two exciting challenges of the laser development for advanced industrial processing, on one side, and solar energy exploitation, on the other side. Thus has constitutes a crucial opportunity to continue the innovation of European industries involved in material processing applications by employing laser technology and to consolidate PV manufacturing European position. New promising scientific and technological approaches have been investigated thus stimulating the continuous growth of the market in these strategic fields for European development.
Project context and objectives:
The aim of the ALPINE project it to push forward the R&D of fibre laser systems for scribing of PV modules. The project consortium focuses on a new high efficiency laser based on PCFs. The consortium intends to demonstrate the all-around development cycle comprising of the beam source, amplification stage, beam delivery and manipulation, scribing processes, advanced diagnostic and validation of modules. In addition the novel laser system must be designed to fit the requirements for scribing innovative and flexible PV modules rather than standard ones based on crystalline silicon wafer. In particular the two most appealing technological alternatives are considered, that is CdTe and CIS or Copper indium gallium diselenide (CIGS) technology of thin-film solar cells. Validation of the quality production process must be assessed as well.
The project joins together two exciting challenges, the laser development for advanced industrial processing, on one side, and solar energy exploitation through new materials, on the other side. Thus it constitutes a crucial opportunity to continue the innovation of European industries involved in material processing applications by employing laser technology and to consolidate PV manufacturing European position.
To achieve these aims the ALPINE consortium focuses on different targets, ranging from the laser development to the definition of new fabrication processes for PV modules, from the integration of PCF technology to the definition of validation protocol of solar modules, and many others. In particular, the WPs 1-3 are in charge for the laser prototype development, based on the new seed and pump laser diodes and on specifically designed PCFs.
The fourth WP is devoted to the production of new PV module on flexible substrates using different production technologies, such close-spaced sublimation, thermal deposition, thermal co-evaporation, electro-deposition. WPs 5-6 constitute a sort of contact point of the two previous activities, being in charge to integrate the laser prototype in an industrial scribing machine and to perform the scribing tests on the provided PV modules. WP7 validates the scribed thin film modules, both in substrate and superstrate configuration, for both CdTe and CIS, benchmarking results to the state of the art module production.
Project results:
This section provides, for each Work package (WP), a description of the main scientific and technological results achieved by the project.
WP1: Design, development and optimisation of fibre laser prototype
Objectives
The goal of this WP is to design and develop an optimised PCF based laser source for cost effective, high quality and fast processing of solar cells. The work is split in two directions: first we defined the best parameters for optimum solar cell processing. This step involves designing and prototyping of a versatile Master oscillator power amplifier (MOPA) high-power laser module working in the infrared (IR), green, and ultraviolet (UV). The MOPA device has been adjusted in terms of pulse duration and repetition rate in order to be a powerful developing tool for the solar cell process optimisation. In a second step, the results achieved in WP5 with the MOPA prototype have been used to specify and develop a high power Q-switched laser with parameter close to the optimum. This ended up with a cost effective high power fibre laser giving the best process performances.
The partners involved in WP1 are MULTITEL (WP leader), University of Parma, OCLARO, Eolite, LPKF, Quanta, and University of LJubljana.
Summary of progress towards objectives and details for each task
Task WP1.1: Development of MOPA prototypes
Two MOPA lasers based on a gain switched diode were built by MULTITEL for tests. An additional laser based on a mode-lock source and pulse picking, operating at 500 ps, down to 100 kHz has been done.
Task WP1.2: MOPA assembling and frequency conversion
The laser from MULTITEL has been delivered to Eolite for amplification tests. With the Eolite amplifier it was possible to obtain 8 W at 300 ps and 20 W for pulses between 1 and 10 ns. The laser and amplifier were sent together to LPKF for the tests in WP5. Frequency doubling has been tested by LPKF with another shorter pulsed commercial laser available. Implementation with the MOPA laser from ALPINE has been successfully done, as well.
Task WP1.3: PCF high power MOPA
The task consists in the optimisation of the MOPA structures and test of PCF for all fibred amplification.
1. A laser from MULTITEL combined with the rod amplifier from Eolite, whose fibre has been produced by NKT, has been assembled.
2. The rod solution has been used to make the solar cell scribing.
Tests with the flexible PCF have been done in the hundreds of picoseconds regime. The flexible fibres can reach less peak power than the rod due to non-linearity. Flexible fibres offer a more rugged laser architecture, as e.g. all-fibre solutions. Flexible fibres is an option for processes where less output power or larger repetition rate are needed.
Task WP1.4: New approach of Q-switching
A Single crystal photo-elastic modulator (SCPEM) device and the relative driver have been developed successfully by UNILJ. The device exhibits a short response time that makes it a good candidate for Q-switching.
Task WP1.5: Q-switch optimisation for SCPEM
Two devices from UNILJ were tested together with Eolite in a Q-switch laser cavity. The results that have been obtained are the following: pulse duration of 26 ns, repetition rates of 183 or 274 kHz and average output power 47 or 67 W respectively, depending on the used device.
Task WP1.6: High power Q-switch with frequency conversion
Eolite has developed a Q-switched high power laser that can work in the IR, green and UV. This is the first Q-switched fibre laser in the UV.
The consortium has done extensive tests on lifetime of the components and particularly of UV optics and crystal. Eolite has demonstrated over 250 h at 23 W average power in the UV. A green and IR version of the Q-switched laser has been delivered to the UniPR for scribing tests.
Significant results
The prototype from Quanta operates in the nanosecond regime from 10ns to few microseconds or more. They have been amplified to a few Watt and done with a linearly polarised output.
The laser from MULTITEL operates in a shorter pulse duration regime (300 ps to 12 ns) and, combined with the amplification module from Eolite, delivers 8 to 20 (depending on the pulse duration) at a repetition rate of 500 kHz.
These are limit values that we decided to respect to avoid any risk of damaging but the laser could work at higher power. It will be for the operator to take care not going beyond those limits. It is explained in the manual of operation.
Rod amplification and SHG tests have been performed: amplification tests (at 60 kHz) with a 50 micrometre core diameter rod-type fibre have been successfully performed.
The maximum output power was 6.86 W (with Raman effects appearing) for 50 W pump power. At 100 kHz we reached 9 W and we tested SHG with a Lithium triborate (LBO) crystal. We got 42 % conversion to the green. At 300 kHz we had 13 W after amplification.
Explain the reasons for deviations from annex I and their impact on other tasks as well as on available resources and planning.
The only deviation from annex 1 is related to the integration of a flexible PCF in the amplified picoseconds laser system. The reason for this deviation comes from the fact that the energy and peak power required for the solar cells micromachining is too high for the flexible PCFs. On the other hand, the results obtained with the ROD type amplifier are well satisfactory for the project.
In addition, MULT, SSE and UniPR contributed to scribing test activity. This action was not originally planned for these partners and was proposed to catch up delay in scribing tests. For this aim, MULT used also an in-house available system at 250 ps while SSE and UniPR used a fibre laser prototype developed by UNIPR and Quanta and a green/IR version of the Q-switched fibre laser developed by Eolite.
Explain the reasons for failing to achieve critical objectives and / or not being on schedule and explain the impact on other tasks as well as on available resources and planning.
A funding reallocation has been proposed and accepted by the Commission for taking into account the corrective actions for catching up the original planning about laser scribing tests.
WP2: Development of active optical components
Objectives
The purpose of this WP was developing active optical components, such as seed and pump lasers, for fibre based laser source. For cost efficiency and also for possibility of applying high speed modulation, the seed lasers were based on a Distributed feedback (DFB) 1 064 nm laser diode. The DFB laser diode design is best tailored for the narrow linewidth, low noise, and short pulse operation required for seeding of high power fibre lasers and for frequency conversion. We designed, developed, and manufactured a narrow linewidth seed laser module with 10xx nm DFB laser diode. The seed laser modules were characterised and optimised for high frequency operation. Additionally, multi-mode pump laser modules were adapted to include a dichroic filter in order to provide protection against fibre laser radiation traveling in the backward direction. Partners involved in WP2 are: OCLARO (WP leader), University of Parma, MULTITEL, Quanta, Eolite, LPKF, and University of Ljubljana.
Summary of progress towards objectives and details for each task
During the duration of ALPINE project we:
1. Successfully developed the building blocks of the manufacturing process for the DFB laser diodes (such as grating lithography, etching, and overgrowth).
2. Designed and manufactured highly efficient DFB lasers operating in 10xx nm spectral region.
3. Optimised DFB lasers for high frequency direct modulation.
4. Carried out optimisation of module package assembly for HF operation of DFB laser diodes.
5. Built laser module prototypes with narrow stripe DFB lasers and single mode polarisation maintaining fibre pigtails. CW operating powers of more than 200 mW at 400 mA out of the fibre is achieved.
6. Performed final test of seed laser modules optimised for ns pulse modulation using fast pulse driver at MULTITEL. Modulation of the DFB laser modules with subnanosecond pulses was demonstrated. The output spectrum modulated with the short pulses exhibited wavelength stabilisation even for such short pulses.
7. Developed single emitter multimode 976 nm pump modules with dichroic pump protection against back-reflected fibre laser emission at 1060 nm. The protection is realised by applying the dielectric coating to one of the discrete optical elements inside the module, which are coupling the light into the multimode fibre. We achieved approximately 40 dB suppression for 1 060 nm radiation.
8. Built several prototypes of single emitter 976 nm pump modules with dichroic pump protection.
9. The prototypes of both DFB seed lasers and pump modules were delivered to the project partners developing the fibre lasers (University of Parma, MULTITEL, Quanta, Eolite, and University of Ljubljana).
Task WP2.1: Device design of the DFB laser diode structure and development of manufacturing processes (months: 1-6)
One of the activities was designing the single mode DFB lasers. Bragg gratings were intended for the emission wavelength of 1 064 nm. For manufacturing the DFB laser at least 2 steps of epitaxy are necessary. During the first step so called laser bottom epistructure is grown, which is based on Oclaro's high-power single mode InGaAs / AlGaAs SQW laser diodes. Then the DFB grating should be defined and finally overgrown during the second epitaxial step.
At this stage of the project we considered both first and second order gratings defined using either e-beam or holographic lithography correspondingly. Gratings were etched using the Reactive ion etching (RIE) kit. The overgrowth step was optimised in order to preserve the gratings and at the same time to produce good structural quality material.
The narrow stripe ridges were later processed using the standard process well established for Oclaro's narrow stripe laser products.
Task WP2.2: Manufacturing and testing of the DFB laser diodes and packaging of LF modules (months: 6-12)
D2.1: Report on the CW performance of a DFB seed laser and delivery of initial samples in non-optimised (LF) package (month 12).
Here we exploited 2 grating designs for 1 060 nm wavelength: 1st order DFB defined using e-beam lithography and 2nd order DFB defined using holographic lithography. The wafers with gratings were processed and laser diodes for both designs were tested. We found that the first order DFB lasers and second order DFB lasers have shown very similar electro-optical performances.
In order to complete the task, several DFB lasers were packaged in low-frequency (LF) 'butterfly modules with single mode polarisation maintaining (PM) fibre pigtail. SM PM fibre is required for easy implementation of DFB seed modules into the MOPA fibre laser systems. The laser emission was coupled into the single mode polarisation maintaining fibre using anamorphic fibre lens. Power of >200 mW out of the fibre is achieved. Fibre coupling efficiency is approximately 75 %. The spectral characterisations of laser radiation out of the fibre have shown that the lasers are locked to the DFB wavelength for all investigated conditions and the side mode suppression ratio (SMSR) is > 50 dB at I > 250 mA, confirming the single longitudinal mode operation.
Task WP2.3: High-frequency packaging for ns pulsing of DFB laser diodes and delivery of pump module with dichroic pump protection (months: 1-18)
D2.2: Report on module package assembly optimised for HF operation and delivery of pump modules with dichroic pump protection (month 18).
For this task we have compiled an equivalent electrical circuit for Small signal modulated (SSM) forward biased laser operating above the threshold in butterfly package. Our model predicts possibility to modulate up to 500 MHz for our standard packaging scheme. We identified the ways to improve the modulation speed up to 1 GHz and higher.
The second part of the project was devoted to introduction of the fibre laser pump protection against fibre laser 1 060 nm radiation coupled back into the modules. The pump laser radiation with the wavelength in the range between 915-980 nm is coupled into the multimode fibre using Fast axis collimating (FAC) and Slow axis collimating (SAC) lenses. The high reflectivity coating for the wavelength > 1010 nm was deposited on one side of the SAC lens, closest to the fibre entrance. Such coating prevents the back-reflected fibre laser emission coming out of the fibre from focusing back into the laser waveguide and consequently from destroying the pump laser.
Using this design we built modules with protection from the fibre laser radiation. We experimentally determined the suppression of 1060 nm radiation by approximately 40 dB.
Task WP2.4: Complete pulsed operation characterisation, final reporting, and delivery of prototypes in high frequency capable packages (months: 19-24)
D2.3: Final performance report and prototype delivery (month 24)
During the final 6 month we focused on improving the capability of developed DFB laser devices for high-frequency modulation. The devices were thoroughly characterised using both SMM experimental set-up build at Oclaro and high current sub-nanosecond pulse modulation using the set-up at MULTITEL.
By changing the chip design and thus reducing the parasitic resistance we were able to extend the SSM bandwidth from approximately 1 up to 4 GHz.
As the next step the DFB lasers were assembled in the butterfly module package with single mode polarisation maintaining fibre pigtail and were tested using the sub-nanosecond current pulses at MULTITEL.
Current pulses adjustable between approximately 400 ps and 4 ns were applied to the laser module. The DFB laser in the module optimised for HF operation was capable of producing optical pulses shorter than 1 ns with the wavelength locked at approximately 1062 nm.
Significant results
1. manufacturing process of DFB lasers is established at Oclaro;
2. high performance 1 060 nm DFB lasers are developed;
3. High-speed modulation of the module packaging is demonstrated;
4. carried out optimisation of DFB laser design for ultrashort pulse modulation and successfully demonstrated wavelength stabilisation for the DFB laser packaged in the butterfly module driven with subnanosecond pulses;
5. approximately 40 dB suppression of the 1 060 nm radiation using dichroic filter is demonstrated experimentally;
6. single emitter pump modules are assembled with such filters for protection against the fibre laser radiation.
Explain the reasons for failing to achieve critical objectives and / or not being on schedule and explain the impact on other tasks as well as on available resources and planning. Development of laser pump modules with the dichroic filters was delayed (approximately 4 month) because of difficulties in designing stop band of the filter on the curved surface without affecting the lens optical throughput efficiency at the laser wavelength. Several design iterations were necessary to achieve the required result. Meantime we provided standard laser pump modules without protection to our partners and the pump protection was achieved using the external filter. Therefore it was no impact on the other tasks of project.
Resources were spent according to the project plan. However the actual spending was lower than planned. Therefore part of the funding for WP2 project was transferred to other WPs.
3. WP3: PCF design, production and testing
Objectives
The aim of WP3 is to develop active, Large mode area (LMA) fibres and interfacing technologies for Q-switched and MOPA laser configurations. The WP is closely linked to WP2 for seed- and pump sources and to WP1 for complete laser integration and system tests.
The goal is to leverage the stand-alone PCF technology to monolithic solutions enabling high-power pulsed fibre lasers to enter the solar cell market. Objectives for the WP are:
- optimising flexible LMA fibres for highest possible power performance;
- develop interfacing technologies that allow simple access to PCFs;
- produce non-flexible (rod-type) fibres enabling new laser tools for solar cell materials processing.
The partners involved in WP3 are: NKTP (WP leader), UNIPR, EOL, Oclar, UNILJ and MUL.
Summary of progress towards objectives and details for each task: The development in WP3 combines numerical simulations from the University of Parma, fibre drawing at NKT Photonics A / S (NKTP), and fibre laser expertise from Eolite and Multitel.
Task WP3.1 - Flexible active LMA fibre with conventional fibre pigtail (months: 1-12): The tasks in WP3.1 have been concluded and the results are described in deliverable report D3.1 which was delivered in time. The work was headed by NKTP with inputs from UNIPR and EOL.
Task WP3.2 - Development of modelling tools (months: 1-20): The tasks in WP3.2 have been concluded and the results are described in deliverable report D3.2 which was delivered in time. The work was headed by UNIPR with inputs from NKTP and EOL.
Task WP3.3 - Monolithic pump-signal combined, flexible active LMA fibre (months: 20-26): The tasks in WP3.3 have been concluded and the results are described in deliverable report D3.3 which was delivered with a delay of five months. The delay on the combiner was not critical for the ALPINE progress, as the fibre laser prototype for solar cell scribing in WP1 was made without it.
Task WP3.4 - Rod-type active LMA fibre (months: 22-30): The tasks in WP3.4 have been concluded and the results are described in deliverable report D3.4 which was delivered in time. The development is based on a patent pending rod design developed in the Seventh Framework Programme (FP7) project LIFT, whence an agreement was made to transfer the foreground knowledge. The results showed that thermal effects have a significant impact on rod performance. During month 30 progress meeting, it was thus decided to try and reallocate funds for an additional activity on an ROD type fibre incorporating thermal effects. This activity is described below.
Task WP3.5 - Rod type active fibre incorporating thermal effects (months: 30-34): This task has been added to the originally planned WP3 activity by the GA amendment. The tasks in WP3.5 have been concluded and the results are described in deliverable report D3.5 which was delivered in time. The new rod has significantly better performance than PCF#3 and NKTP expects to launch it as a commercial product later in 2012.
Significant results
The development up to the fabrication of PCF1 (D3.1) has resulted in an improved production method, which has been implemented in NKTPs top selling LMA commercial fibre product DC-200-40-PS-Yb. This has matured the fibre and it is expected to be transferred from R&D to production in 2012.
A major obstacle for wide spread commercialisation is that the fibre is difficult to use. Accordingly the pump / signal combiner to be developed in task 3.2 should increase the addressable market of the fibre significantly.
NKTP started developing a backward pumped combiner in 2010, and made a prototype with excellent performance. However, it was not possible to reproduce this combiner (see D3.3 for details). Independent development in WP1 by Multitel and Eolite showed that the flexible fibre platform was not able to handle sufficient power for making the ALPINE laser demonstrators. Thus, NKTP did not send the single good multiplexer to EOL, but instead sought for a reproducible combiner solution. This involved changing from a backward - to a forward pumped combiner and outsourcing part of the development to a third source. The details of the combiner are reported in D3.3 which was delivered in Feb-2012. Since then, NKTP has shown that the combiner works very well in a ps amplifier (80 MHz, 5 ps at 1 064 nm), where amplification up to 80W output power has been demonstrated.
The combiner was sent to EOL on 29 June 2012, where it will be tested in a ns amplifier in the last months of the project. Subsequently NKTP will sample key customers with starting from Q3 2012 and it is expected that it will be made generally commercially available in 2013.
The results from PCF2 (D3.2) have so far not resulted in a commercial product. It has been shown that the fibres can successfully suppress non-linear effects, such as ASE. Thus they potentially have better performance than conventional fibres for operation at 1 064 nm or higher wavelengths, and for pulsed systems with low repetition rate. So far it is unclear whether the improved performance is sufficient to justify the additional fabrication complexity. Accordingly, NKTP have per 1 January 2012 started a doctorate (PhD) project with the aim of maturing the fibres. It is still to goal to release these fibres as a commercial product in 2013. The ROD fibre developed for PCF3 (D3.4) showed large promise, but was difficult to reproduce and had issues with thermal effects. It was proposed as a corrective action to reallocate funds to NKTP and UNIPR to develop a fourth fibre in a new task 3.5.
The resulting rod fibre has been fabricated in early June 2102 and did indeed have the desired properties:
- wide single mode guidance region from 1 030 to 1110 nm;
- high power operation with a single mode output;
- 175 W output shown;
- excellent output power stability: over 200h test with 50 W output.
Thus, NKT Photonics is planning to launch it as a commercial product later in 2012 (for further details see report D3.5).
WP3 had furthermore led to two master projects at the University of Parma, one master project at The Technical University of Denmark (DTU) with NKT as co-supervisors. Furthermore DTU and NKTP have in January 2012 started a PhD project to further develop the fibre concept, which was used for D3.2. The work in WP3 has been disseminated in 5 journal publications and 13 talks at peer reviewed conferences. The dissemination has mainly been driven by UNIPR.
WP4: Thin film solar cell production
Objectives
The aim of this WP is to prepare and supply to all the other partners samples of solar cells of both CdTe / Cadmium sulfide (CdS) and CuInGaSe2 / CdS deposited on different substrates such as glass, polyimide and metallic foils. Results on the module characterisation and validation here performed are also part of WP7 activity and will not be reported once again in WP7 section. The partners involved in WP4 are: SSE (WP leader), NEXCIS, SSW, WS, UNIVR.
Summary of progress towards objectives and details for each task:
Within the WP, there are two main groups: one made up of industrial partners and the other are of research laboratories. The first part consists of SSE, NEXCIS and WS which respectively produce: CdTe PV modules based on rigid substrates, CIGS PV modules on rigid and flexible substrates, CIGS modules on glass substrates. The second part consists of SSE, SSW and University of Verona which respectively produce: CdTe, CIGS and CIS solar cells both on rigid (glass and ceramic) and flexible substrates, CIGS solar cells and modules both on rigid and flexible substrates and CdTe both on rigid and flexible substrate.
SSE
SSE has continued his research with respect to the CdTe-based cells in substrate configuration and on CIGS-based cells both on glass and on ceramic substrates.
1. CdTe modules at different stages of fabrication for P1, P2 and P3 laser scribing. 10x10 cm2 and 30x30 cm2 mini-modules were sent to Multitel.
2. CIGS based solar cells completely fabricated by sputtering and selenisation. SSE innovation consists in the use of the sputtering technique to deposit the starting precursors (In2Se3 or InSe e Ga2Se3 or GaSe) and a consequent selenisation in pure Se atmosphere.
With the innovative method of deposition of the composite precursors developed by SSE and described before an important result was obtained by using as a substrate ceramic commercial tiles. A PV conversion efficiency of 13.95 % was reached and this is one of the highest value ever obtained on ceramic substrates.
In this period SSE produced a lot of CdTe and CIGS samples used for the laser scribing test. In fact, by September 2011, SSE installed a machine for laser scribing in the Thin Film Laboratory (ThiFiLab) of the University of Parma.
In a first stage, the scribing machine was equipped with a conventional solid-state laser, which works both at 1064 and 532 nm. Later on the scribing system was equipped first with a fibre laser built up by the laboratory of Stefano Selleri at the University of Parma and secondly by a fibre laser produced by Eolite.
The first fibre laser installed worked only at 1 064 nm. Some results so far obtained on TCO samples normally used in CdTe/CdS solar cell fabrication are shown in the following. Results of comparison of a P1-scribing performed by a commercial solid-state laser and a P1-scribing carried out with the UNIPR laser show the great finesse difference of the fibre laser. Also P2 and P3 scribings tests, carried out on CdTe/CdS modules by using Eolite fibre laser, shown very clean and sharp edges and they are not affected by the thermal damage (HAS) usually observed in CdTe standard manufacturing.
CdTe mini-modules, with an area of 10x10 cm2, by testing a fibre laser scribing at a time were made. This allowed us to characterise in detail the single scribing carried out with new fibre lasers by finishing the module with traditional scribing.
With Eolite's fibre laser many tests on CIGS-based mini-modules (dimension of 10x10 cm2) sent from SSW have been performed. It has been thought to proceed as for CdTe-based minimodules. Experiments are underway and we have been successfully performed the P1 scribing (at 1064 nm) and P2 (at 532 nm). Some 10x10 cm2 CIGS-based minimodules equipped with P1 scribing directly made by SSW were tested in our laboratory for P2 scribing and then re-sent to SSW that will carry them out by running the P3 scribing in traditional way (mechanical scribing). In this way it will be clear the behaviour of the P2 scribing realised with the Eolite fibre laser at 532 nm on CIGS-based mini-modules in respect to standard P1 and P3 traditional scribing. The electrical characterisation of the different processing steps certifies in all cases an efficiency around 10 %.
NEXCIS
NEXCIS started to provide electrodeposited based CIS on glass at month 2 (2 months before the compromise acquired in the project) and CIGSe at month 12. During the last months NEXCIS has validated and increased the quality of the absorbers and devices that were scaled up to 60x120 cm2 size. Wafers of this size (and size smaller than this) have already been sent to laser partners, especially for testing of the automated machine. Moreover, complete modules have been sent to JRC for measurement of standard modules to be used as reference at NEXCIS. More samples have been sent to LPKF for final scribing and final modularisation at NEXCIS.
Here the results acquired in the project activity are reported. The best efficiency is 13.9 % (by Newport in the US) for a cell (approximately 0.5 cm2), medium efficiency is 12.2 for 95% of the wafer surface (for a 30x60 cm2 wafer), and 11.6 % for 100 %. An efficiency of 8.2 % for a 30x60 cm2 module was communicated. Today, efficiency for a pixel is certified at 14 % (by Newport in the US) for a cell (approximately 0.5 cm2), in case of internal measurements we have measured a cell at 14.9 % (under certification measurements), medium efficiency in a wafer is 13.2 for 95 % of the surface of the wafer, and 13.7 % for 100 %. There has being also a big reduction in the gap between wafer and modules, and 30x60 cm2 modules are now obtained in a repeatable way with efficiencies at about 10.7 %. The record efficiency for encapsulated modules is 12.3 % as certified by Certisolis.
However what it is more important, 60x120 cm2 wafers with a medium efficiency of 11.2 % (12 % for 90 % of the wafers) are obtained in our baseline, which will be providing in near future with modules 60x120 cm2 with efficiencies larger than 10.5 %.
The process portability on flexible metallic substrate has been fully demonstrated. However, due to more work needed with the right encapsulation material, metal and polymer modules are in stand-by due to reliability issues not concerned with NEXCIS process but, more in general, to lightweight thin film objects. In this sense NEXCIS has focus all activities in glass-glass solutions.
Pre-qualification of the 60x120 cm2 reactor for electro-deposition of Cu, In and Ga is already achieved. The same pre-qualification has been achieved for the furnace developed by NEXCIS. NEXCIS is at present working in validating the yield, throughput and economic issues to go towards industrial pilot line conditions, which is foreseen at the end of 2013. At present, more than 90 employees are working at NEXCIS facilities while at the beginning of the project less than 10 employees were working at NEXCIS.
WS
Wurth Solar was able to deliver CIGS modules on large glass substrates after any relevant processing appropriate for laser scribing tests (P1-P3, edge cleaning). Wurth Solar organised to be able to respond flexibly and quickly to partners with substrates as soon as required. WS has sent standard modules (P2, P3 mechanically patterned) to JRC for validation to obtain a reference for the laser scribed samples.
During the last months of the last year, WS was partially acquired by MANS GmbH. Only in late April 2012 it was clear that this partner was yet inside the consortium. For this reason WS has considerably reduced its activity during the last months of the project. WS sent mini-modules to UniPr for scribing test in M36.
SSW
SSW has continued to deliver CIGS modules with sizes of 10 cm x 10 cm as samples for the testing of patterning lasers to the partners LPKF, MULTITEL and University of Parma. These samples included modules ready for the application of all three patterning steps. Also, additional samples carrying an array of 81 test cells for the purpose of P3 quick tests in the laser labs have been delivered.
For the purpose of having reference samples with known electrical properties, one out of each set of nine samples was completed using conventional patterning at SSW. These references reached efficiencies of up to 15 %. This improvement of about 1 % compared to previously reported reference samples is due to improved deposition parameters for the sputtering of the zinc oxide (ZnO) front contact which lead to an increase of the open circuit voltage of the devices.
Furthermore, additional modules on flexible substrates - polyimide film with a thickness of 25 micrometre - have been fabricated and sent to MULTITEL for testing patterning steps P1, P2, and P3.
Flexible substrates made from steel foil covered by a barrier layer have been prepared. This silicon oxide barrier layer, which is about 1 micron thick provides the electrical insulation of the subsequently deposited Mo back contact layer and of the individual cells. Sample have been sent to MULTITEL for P1 scribing experiments.
UNIVR
In UNIVR labs, solar cells have been fabricated by depositing each single layer (except for the front contact), namely CdS, CdTe and back contact, by vacuum evaporation. For the preparation of the samples to be scribed by the laser partners we have prepared and optimised a complete fabrication process for deposition on glass and on polymers, so by applying a lower temperature process compared to the typical ones that use substrate temperatures above 500 degrees of Celsius.
Front-contact deposition
For Transparent conductive oxides (TCO), that make the front contact of the device, two different glass coatings have been used: a commercially available ITO film coated glass with a SiO2 barrier layer and a laboratory scale bi-layer of conductive ITO + thin insulating ZnO, deposited at high temperature. The first one is a commercially prepared ITO with a thickness of 180nm and a sheet resistance of 10 Ohm / square. The second one is a 400nm ITO+100nm intrinsic ZnO deposited by radio frequency-sputtering at a temperature of 400 degrees of Celsius with a sheet resistance below 5 Ohm / square. Due to its high substrate deposition temperature this layer is much more crystallised and stable; moreover because of the ZnO layer, which acts as a barrier to diffusion of impurities, indium diffusion is avoided. A comparison of the devices made on these two different substrates will outline some of the properties observed.
Window layer fabrication
CdS is evaporated in vacuum at a pressure of 10-6 mbar using direct current heating of a molybdenum crucible and deposited on the glass / TCO stacks at a substrate temperature of 150 degrees of Celsius with a deposition rate ranging from 0.15 to 0.45 nm / sec (controlled by a quartz crystal thickness monitor). After deposition, CdS was annealed by heating the stacks in vacuum for 30 minutes (in order to increase the stability of CdS to the subsequent depositions and CdTe activation treatment), a slight re-crystallisation of the grains was observed by AFM with an enlargement of the grain size from 50-100 to 100-200 nm.
CdTe deposition and post-deposition treatment
After deposition and treatment of the CdS layer, CdTe is deposited in the same chamber by heating a special in-house made graphite crucible at a substrate temperature ranging from 300 degrees of Celsius up to 340 degrees of Celsius and deposition rate of about 2 nm / sec. A typical CdTe layer for these devices is 3.5 / 4 micron thick with quite compact morphology and grain size of about 1 micron; we have observed that the morphology and the grain size is depending not only on the substrate temperature but also on the TCO. CdTe deposited on commercial ITO and ITO+ZnO shows a similar but not same morphology, grain size is typically around 1 micron for both but it results to be much more compact in the second case. Compared to the Close space sublimated (CSS) deposited CdTe grains (which have a much higher substrate temperature) are quite small and activation treatment is needed also for increasing the grain size.
It is known that for a good cell performance an activation treatment is necessary. Generally, in order to enhance the grain size and passivate the grain boundaries in the CdTe polycrystalline structure a layer of CdCl2 is deposited on top of the CdTe and then annealed in air. We have prepared two different activation processes: the standard CdCl2 treatment and the alternative treatment with chlorine containing gases.
Back contact
Back contact was made by vacuum evaporation of Cu/Au stacks. Subsequently an annealing of the back contact is applied at 200 degrees of Celsius in air for 20 minutes in order to diffuse Cu in CdTe and make an ohmic contact. In case of CdCl2 treatment, a bromine-methanol etching is applied in order to remove the CdCl2 layer and to prepare the surface for the Cu / Au deposition, however the removal of the CdCl2 in case of wet treatment was quite difficult. In case of freon treatment, the treated CdTe film is not etched since the surface is free of CdCl2 (re-evaporated at high temperature in vacuum) and a tellurium rich layer is made during the activation process.
Deposition of copper results to be very sensitive, a higher amount of copper improves the back contact also doping the bulk CdTe but reduces the shunt resistance (Rsh) resulting in a lower fill factor. The amount of copper has been tuned for the differently treated CdTe layers. A more conductive absorber would need less copper to improve the performance.
Several PV devices have been made with different parameters of both CdCl2 and freon treatment for CdTe deposited on the different stacks described before.
The highest performance was given by CdCl2 treated cells and with stronger CdS / TCO stacks, this means that the most performing TCO was the one made with high temperature deposited ZnO / ITO with enhanced ability and prevention of sodium and indium diffusion. Moreover only 2 nanometres of copper with 50 nanometres of gold are necessary for a good contact. With this configuration efficiencies exceeding 13.5 % are routinely obtained. If freon treatment is applied, much lower efficiencies are measured: highest efficiencies are obtained with 8 nm of copper, however a much lower Rsh gives lower fill factors and lower open circuit voltages.
If a 2 nm copper contact is applied on these freon treated cells a very low efficiency is performed. Anyway with more aggressive freon treatment (by higher freon partial pressures or higher temperature) bigger grain size is obtained but not higher efficiencies. Excess of freon pressure have resulted in shunted device. So the required copper amount is also depending on the different CdTe treated layer. We have registered that in case of freon treated CdTe devices four times the amount of copper is needed for a reasonable efficiency compared to the CdCl2 treated cells. This is consistent with the fact that activation by freon is much weaker than the one made with CdCl2 as reported above. On the other hand, using a lower amount of copper allows to have a better working cell and this results in a higher efficiency.
Cells made with CdCl2 treatment exhibit efficiencies from 10 to 13.5 %, with current in the range of 22 to 25 mA / cm2, Voc exceeding 830 mV and FF from 60 to 70 %. The relatively low Voc and FF are connected with relatively low Rsh. This can be explained by an excessive consumption of CdS into the CdTe layer. The best results were obtained with thicker CdS (more than 500nm) annealed in vacuum at 450 degrees of Celsius. Freon-treated cells show lower efficiencies than CdCl2 treated ones. The highest achieved efficiency for this treatment is 8.7 %. Although the freon treated cells would not need a post deposition etching, a good ohmic back contact was not easily reproducible. Many of our freon-treated cells have shown a kink on the positive part of the curve.
Transport properties
For device characterisation two different types of cells made with vacuum evaporation were taken into consideration: one activated by freon and one activated by CdCl2. Moreover cells made by close space sublimation at the University of Parma, were successfully activated by HCF2Cl gas, with efficiencies exceeding 10 % as a reference with the low temperature deposited cells.
A standard technique which is very often applied for determination of the doping level in semiconductor junctions is capacitance voltage (C / V) profiling. However, in the presence of deep levels free carrier concentrations determined by C / V profiles can be subjected to large errors. This can be overcome by drive level capacitance profiling (DLCP), as DLCP is a fully dynamical measurement giving undistorted doping distributions. For junctions without deep traps, adjusting their charge state to the DC bias, DLCP and C/V profiles coincide, so the difference between the C / V and DLCP profile gives a lower bound for deep defect concentration active in a given measurement conditions (i.e. temperature and frequency). The capacitance measurements were performed with an Agilent E4980A LCR meter controlled by LabView via computer. Doping concentrations in the space charge region (SCR) of CSS-freon, VE-CdCl2 and VE-freon samples were estimated to be 2x1014 cm-3, 6x1013 cm-3, 3x1013 cm-3, respectively. Deep defects contributing to capacitance-voltage profiles were also detected. The difference between the C / V and DLCP profile, indicating the lower bound for the concentration of deep defects following the DC bias sweep, is also the highest for the CSS-freon sample and amounts 4x1014 cm-3. For the VE-freon device the defect concentration varies from 1x1013 cm-3 at 2 micrometre up to 2.3x1014 cm-3 at distances larger than 2.5 micrometre. In the VE-CdCl2 sample we detected no large differences between the C / V and DLCP profiles around RT.
Flexible cells
The same optimised process has been applied to the flexible polymer substrates. Prior to this, a study of the different polymers have been made. Different polymers (some of them are not on the market but still in prototype form) were taken into consideration: DuPont (Kapton), Kaneka, UBE (Upilex) and Mitsubishi (Neopulim). For Kaneka and DuPont a smaller thickness must be used. Indeed, the 25 micron thick substrates are almost brown, cutting transparency below 400 nm for Kaneka and Kapton. Mitsubishi polymers, however, have a higher transmittance at low wavelengths even if they remain opaque (probably still need to reduce the thickness).
The polyamides have been tested with an annealing in order to check their temperature resistance to the deposition process, treatments were made for 30 minutes at 370 and at 450 degrees of Celsius in air. Flexible solar cells have been fabricated by using Upilex foils12.5 microns thick. For the CdTe deposition process and post-deposition treatment, the same process optimised on glass was used also for the polymer substrate.
Morphological and crystallisation parameters have been studied with atomic force microscopy and X-ray spectroscopy. CdTe on flexible substrates grows similarly as on glass, in terms of grain orientation and size. However some restrictions come from the limited transparency of the polymer substrate. Polymer looks yellow in colour.
Significant results
- The objectives for the CdTe-based cells made on substrate configuration have been achieved.
- CdTe and CIGS mini-modules have been made in traditional way for reference.
- CdTe and CIGS mini-modules in various stages of realisation have been made and sent to all the scribing partners who have requested them.
- High efficiency CIGS-based cells made on ceramic substrate have been realised. Some of them have been used for testing the laser scribing.
- The SSE scribing work is intensively continuing both on CdTe and CIGS samples. P1, P2 and P3 on CdTe minimodules were definitively developed and optimised. Fibre-laser scribed CdTe minimodules were made and sent to JRC for characterisation. Results are very encouraging in order to use this technology for solar cell processing.
- Low temperature vacuum evaporated high efficiency CdTe solar cells both on rigid and polymeric flexible substrate were realised and cells in different stage of fabrication were sent to all the scribing partners who have requested them.
- Successful scale-up from 15x15 to 30x60 cm2 on glass substrate and preliminary to 60x120 cm2 at NEXCIS.
- Today NEXCIS efficiency for a pixel is certified at 13.9 % (by Newport in the US), medium efficiency in a wafer is 12.2 for 95 % of the surface of the wafer, and 11.6 % for 100%.
- The process portability on metallic substrate has been fully demonstrated.
- Prequalification of the 60x120 cm2 reactor for electrodeposition of Cu, In, Ga is already achieved.
Considering the great amount of exchange of modules by partners, the consortium agreed to demonstrate the feasibility of the laser patterning process by producing a series of 30 cm x 30 cm mini-modules for CIGS and 10 cm x 10 cm mini-modules for CdTe. In this way all the scribing systems can participate to the effort to build PV modules fully scribed with fibre laser. Moreover these dimensions are sufficient to demonstrate the scalability at an industrial level of the scribing made with fibre laser.
A statement on the use of resources, in particular highlighting and explaining deviations between actual and planned person-months per WP and per beneficiary in Annex 1 (Description of Work).
The use of the allocated resources has been as planned, but WS whose reduced participation during the last months resulted in a reduced actual budget.
WP5: Laser scribing and beam handling
Objectives
The aim of WP 5 is to integrate the laser source developed in WP1 in a machine concept. That includes the beam handling and new concepts for the beam splitting. On that machine industrial like laser scribing of solar panels shall be achieved.
The partners involved in WP5 are: LPKF (WP-leader), Eolite, NEXCIS, SSW, WS, UNIVR, ES.
Summary of progress towards objectives and details for each task
Task WP 5.1 - First scribing machine prototype: The scribing machine was designed and built upon requirement specifications taking into account the different technologies to be tested (CdTe and CIGS) as well as production related issues like accuracy and throughput. As a result the machine is capable of processing substrate and superstrate configurations at high speed being flexible in providing a platform for different laser sources and optical configurations.
Task WP5.2 - Test of the MOPA laser prototype: The final MOPA test laser has been integrated into the ALPINE test system in Garbsen and tested with both wavelengths 532 nm and 1 064 nm at a pulse length of approx. 500 ps on all test materials provided by the partners SSE, University of Verona, Nexcis and SSW. Tests with the state of the art picosecond laser source and the beam shaping plate have made at Multitel and using the Q-switch laser from WP1.4 SSE performed further scribing test. Samples were sent back and forth for testing and further processing and final mini-module performance compared to those manufactured with standard scribing methods.
Task WP 5.3 - Beam shaping test and Q-switched laser integration: The first approach by LPKF has been to use a special beam shaping optic providing a quasi-top-hat profile in order to reach a higher depth of focus compared to the one achievable with exact top-hat intensity profiles. The result gives a good top-hat approximation with a depth of focus around +/- 400 mm at 532 nm. This result has been compared to a fibre-based top-hat generation done by Eolite.
Task WP5.4 - Final optimisation of the scribing machine: Components of the ALPINE test system have been optimised in their design aiming at highest throughput and accuracy. The main components influencing the tact time of scribing are the scribing and glass transporting axes. LPKF has worked on two different axis designs which are currently set-up on a test bench and are going to be evaluated regarding acceleration, speed, accuracy, heating taking into account large and heavy loads as 4 mm thick substrates of up to GEN5 sizes, as well as free flying optics. SSW has tested and compared scribed test modules from all participating partners to conventional scribed ones. Laser scribing has already proven to be a promising alternative to mechanical scribed CIGS modules.
Significant results
- Laser scribing machine capable of running of up to 2 m / s on the scribing axis.
- Laser scribing machine designed to perform film-side and glass-side scribing comprising a flexible exhaust system to process CdTe modules as well as CIGS modules.
- Design and test of a beam shaping module providing good top-hat approximation with an excellent depth of focus.
- MOPA laser has been integrated in ALPINE test system and extensive scribing tests and parameter improvements on all relevant material systems have been carried out.
- CIGS samples with P2 scribed at LPKF using enhanced parameters have been completed and analysed at SSW. The I-V curves indicate a good P2 scribe quality.
- P3 laser-scribed CIGS sample at MULTITEL was comparable with conventional mechanical scribing.
WP6: Automation
Objectives
The goal of this WP is to study, develop and test standard parameters in order to produce system integration guidelines and to realise systems to focus and guide the laser beam. The partners involved in WP6 are ES (WP-Leader), Quanta and LPKF.
Summary of progress towards objectives and details for each task
WP6.1: Standards for automation integration in scribing systems
The first task is addressed into developing rules for integration and experiences in the real industrial environment. The activity were dedicated to collect information and experiences with partners and to attend important European fairs to understand needs of the market and to identify issues to be faced. In this package were assessed important parameters for the laser integration like temperature effects (gradients and stability), laser scribing panel feeding, mechanical stress, systems integration issues. The integration of the protocol developed by Quanta to manage the device they developed in an industrial network has been studied and implemented.
WP6.2: Laser beam moving and focusing system
The WP has developed a suitable focusing head achieving a stable and reliable processing of the glass. Due to changes in the focal position the spot diameter changes accordingly and so the power density on the layer to be processed. Allowing a dynamic focusing of the laser beam brings to a higher stability of this parameter and consequently of the ablation rate. The core philosophy of the compensation system is a matrix camera and a focus system developed with the collaboration of all the WP partners.
The WP also developed a reliable moving system with adequate precision in order to explore and test the feasibility of a guiding system as well as a cost effective and less precise moving system with acceptable results. The moving system moves the sample under test below a fixed laser and guiding system which is composed by two linear axes in a table configuration. Details of focusing and moving systems are reported in deliverables D6.2 and D6.3.
Significant results
A practical , computer controlled high-end dynamic guiding system has been developed that allows for tracking of the glass scribed waviness on thin films solar cells.
A microprocessor controlled high-end dynamic focusing system has been developed that allows for fast tracking of the glass deformations on thin films solar cells.
WP7: Testing of PV devices and scribing process validation
Objectives
The objective of WP7 is to evaluate the quality of PV thin film modules patterning on devices realised in different configurations (substrate / superstrate) and on different substrate materials (glass / flexible). Both CIS and CdTe modules are researched. The idea is to define a set of suitable tests leading to correctly identify potential critical aspects both at cell level and at module level. Similar measurements are also carried out on state-of-the-art solar modules with conventional processed patterning and edge cleaning, in order to provide a benchmark against which to highlight process improvements by laser use.
The partners involved are JRC (WP leader), SSW and University of Verona as research centres, together with SEE, Nexcis, Wurth Solar and LPKF as industrial partners. Multitel has also been involved in the scribing work to prepare devices.
The agreed strategy of the consortium is to demonstrate the feasibility of the PCF laser patterning process by producing a series of demonstration devices (30 cm x 30 cm mini-modules for CIGS and 10 cm x 10 cm mini-modules for CdTe). It is noted that this device production and activity is partly in common with WP4.
To enable assessment of each of the scribing steps as well as the overall working device, each combination (that is CdTe or CIGS on glass or flexible substrate, substrate or superstrate configuration, scribing steps) has included, as a minimum, a group of 4 sample devices with different characteristics according to the scribing process P1, P2 or P3. Details can be seen in the deliverables D7.5 and D7.6.
The verification testing uses the methods described in Deliverables D7.3 and D7.4. The electrical performance tests on one of the four samples include as a minimum spectral response and current-voltage (IV) characteristics at Standard test condition (STC)s. These have been complimented by dark IV, thermography and LBIC depending on time and resources are available.
Summary of progress towards objectives and details for each task: The activities performed are clearly explained in deliverables D7.5 and D7.6 and summarised as follows:
Task WP7.2 - Validation of thin film modules realised in a substrate configuration (CIS) on glass and processed by means of advanced fibre lasers: SSW has prepared and characterised devices scribed at LPKF. This work is reported in WP4.
Task WP7.3 - Validation of thin film modules realised in a substrate configuration (CIS) on flexible substrates and processed by means of advanced fibre lasers: Devices have been prepared using polyamide and steel foil substrates. Scribing trials tried to cope with the waviness and flexibility of the substrate. The process is critical.
Task WP7.4 - Validation of thin film modules realised in a superstrate configuration (CdTe) on glass substrates and processed by means of advanced fibre lasers: SSE has prepared and characterised devices scribed at the SSE/University Parma facilities. JRC has validated the electrical performance of the devices. A series of devices has been prepared by University Verona. Results are reported in the updated version of D7.6.
Task WP7.5 - Validation of thin film modules realised in a superstrate configuration (CdTe) on flexible substrates and processed by means of advanced fibre lasers: A series of devices has been prepared by University Verona. The scribing process is still critical.
Significant results
Concerning the most suitable analytical methods to evaluate the scribing operations, the following have been identified as the most feasible for a correct check on the quality of laser scribing process:
- optical microscopy;
- scanning electron microscopy;
- electrical insulation test;
- I-V characterisation of solar cells;
- thermal imaging;
- LBIC mapping.
Measurements according to the first three methodologies have been performed just after laser scribing process, in order to rapidly identify potential faults;
- I-V characterisation and thermal imaging have been then implemented at a later stage, when a completed cell or module is assembled. Tests are also considered in the field of validation of thin film modules comprising IV curve STC indoor and outdoor, stabilisation cycles, spectral response, TCO and matrix.
In particular, referring to I-V curve measurements STC conditions, a dedicated research activity is ongoing at JRC, focusing on the measurements methods for improving the compliance among indoor and outdoor measurement results.
As indicated in the above table, an extensive effort is being made to produce demonstration devices using the laser patterning techniques developed in the project.
The performance verification of the following device combinations is already at an advanced stage:
- CdTe on glass substrate with PCF laser scribing (SSE and University Parma).
- CIS on glass substrate with PCF laser scribing (SSW with LPFK, SSW with Multitel / University Parma).
In additional the performance verification of CdTe on glass substrate produced by University Verona together with LPKF and Multitel scribing using ALPINE lasers is fully completed. Results are reported in the updated version of D7.6.
Results to date indicate that PCF laser scribed devices are very promising and laser technology is very effective for glass substrate modules. Flexible substrate ones proved to be more critical.
The WP7.2-7.5 activities are largely dependent on the availability of laser-patterned devices from other WPs. The supply of these has been delayed, largely due to problems with availability of lasers for the scribing operations. At the general assembly meeting on 28 September 2011 the consortium agreed to enforce the laser scribing activity (GA amendment n.3) to catch up the delay. In addition a detailed planning to complete the work in remaining months of the project was agreed at the executive meeting held 28 March 2012, and these actions allowed the first modules to be already fabricated and validated.
The overall objective of the WP has been maintained, i.e. to demonstrate the quality of patterning / scribing of CIS and CdTe thin film technologies on glass substrate using advanced pulse conditioned lasers, however successful tests on flexible substrates have not yet been achieved. Partners are still working on it, even after the project end date.
List of websites: http://www.project-alpine.eu
