Skip to main content

Non-vacuum processes for deposition of CI(G)S active layer in PV cells

Final Report Summary - NOVA-CI(G)S (Non-vacuum processes for deposition of CI(G)S active layer in PV cells)

Executive Summary:
Current production methods for thin film photovoltaics typically rely on costly, difficult to control (over large surfaces) vacuum-based deposition processes that are known for low material utilisation of 30-50%. NOVA-CI(G)S proposes alternative, non-vacuum ink-based simple and safe deposition processes for thin film CI(G)S photovoltaic cells. The low capital intensive, high throughput, high material yield processes will deliver large area uniformity and optimum composition of cells. The project objectives are to achieve competitive about 14% small area cell efficiency and to demonstrate the processes at high speed on rigid and flexible substrates while maintaining acceptably high efficiencies. The processes reduce cost of the CI(G)S layer by 75-80% in comparison to the evaporated CI(G)S, which translates into a 20-25% reduction of total module cost. Major scientific breakthroughs of the project include improved materials control in novel precursor materials by using nano-sized particles of specific chemical and structural characteristics and innovative ink formulation, to enable coating by simple processes while avoiding the use of toxic gases in subsequent process steps. This industry-led project constitutes the first essential step for a fully non-vacuum, roll-to-roll process aimed to achieve the solar module production cost substantially below 0,8 €/Wp.

- WP1 (Umicore as leader): Precursor materials and ink formulation
- WP2 (ZSW as leader): CI(G)S layer deposition and activation
- WP3 (TUC as leader): Large area deposition and annealing of CI(G)S layer
- WP4 (WERes as leader): Exploitation & Dissemination
- WP5 (Umicore as leader): Management

Developped technology will rely upon the use of inks prepared from particles, considering the following requirements and benefits:

- Typical particles we consider are partly based on metal oxygen-bearing derivatives
- Particles prepared using Umicore’s competences on wet-phase chemistry
- Inks prepared according to end-application specifications
- 2 materials systems selected for further optimization
- Move away from resort to H2Se or harsh selenization conditions

Many particle-based precursors were prepared and evaluated and it was decided to focus efforts in an optimization phase on 2 chosen promising precursor materials systems. These precursors can be prepared easily in a controlled fashion (while the technology being industrially scalable and exploited) and delivered on all typical quality control dimensions (particle composition & morphology, phase formation, formulation, deposition, conversion). Formulation efforts were also carried out to reduce the amount of carbon which needs to be removed during/before conversion to the CIGS absorber layer, improve ink stability and versatility. Currently available formulations prepared within the consortium can be printed using lab-scale standard equipment and can be converted, following a sequence of post-deposition steps, to a functional CIGS absorber layer with cell efficiency up to 8%.

Information can also be found at www.novacigs.info

For more information and further contact, please reach to fabrice.stassin@umicore.com
Project Context and Objectives:
The goal of NOVA-CI(G)S is to develop an ink based non-vacuum simple and safe deposition process of the CI(G)S absorber layer for highly efficient low cost solar cells. CI(G)S has high cost savings potential, coupled with the highest operating efficiency of all TF PV. In a CI(G)S device, the absorber layer is the single most scientifically challenging one and with the biggest impact on materials cost.

The project aims to demonstrate the following:

1. Non-vacuum deposition of the CI(G)S layer is possible whilst maintaining acceptably high efficiency
2. Non-vacuum deposition of CI(G)S layers can be realized at high speed (30m2/min) on rigid substrates whilst maintaining acceptably high efficiency on assembled cells
3. Non-vacuum deposition of CI(G)S layers can be realized at high speed (100m2/min) on flexible substrates replicating similar dried CI(G)S (pre-annealing) characteristics as for rigid substrates

The overall aim is for sure to make a significant impact on reducing the cost (€/Wp) of the CI(G)S layer and pave the way to addressing the device’s adjoining layers with a similar fully integrated low cost deposition process ultimately achieving a CI(G)S TF module cost substantially below 0,8 €/Wp. Non-vacuum deposited CIGS would then become more competitive versus the strongly positioned CdTe thin film PV technology.

The ambitious but realistic initial objectives of NOVA-CI(G)S are covering:

1. General CIGS layer deposition process performance

- Total CI(G)S layer cost (capital, process, materials) reduced by 75-80% (reference section 3.1) versus cost of vacuum technology
- Material utilization of more than 90% up from current 30-50%
- Elimination of hazardous materials

2. Cell & module performance (rigid substrates)

- Printed CI(G)S small area cell efficiencies of about 14%
- Module efficiencies of at least 70% of small area efficiencies
- Deposition productivity of 30m²/min, with structural characteristics equivalent to the optimized lab layer (prior to annealing)

3. Cell & module performance (flexible substrates)

- Printed CI(G)S small area cell efficiency 90% of cell efficiency on rigid substrates
- Module efficiencies of at least 70% of small area efficiencies
- Deposition productivity in the 50 – 100 m²/min range with structural characteristics equivalent to the optimized lab layer (prior to annealing)

As far as cell efficiency objectives are concerned, the objectives of reaching 14% on small area rigid substrates and 12-13% on small area flexible substrates remain relevant in the context of the scientific & technological challenges that need to be tackled by the different partners. Comparing to cell efficiencies recorded for vacuum-deposited CIGS, the objectives as stated herein can be considered a first step in the good direction with however a clear need to go in the future beyond the objectives described herein.

The shift to non-vacuum technology also has serious consequences in terms of dealing with issues as to resource efficiency and sustainability. Indeed, non-vacuum technology will enable a material utilization of more than 90% compared to current typical 30-50% for vacuum-based technologies. Although it is now agreed by many that indium should not face major supply risk in the near future, the reduction of material utilization is a clear advantage here. In term of susbtainability, one of the objectives of the project is to keep away from using toxic, harsh chemicals for the deposition and conversion/selenization steps. No resort will be made to compounds such as H2Se, hydrazine and its toxic derivatives, … One consequence of the choice of the 2 precursor materials systems developped at Umicore is that the source of selenium would actually be embedded (experiments conducted until now still rely upon the need for a separate selenization step until selenium inks are fully developped) already into the ink to discard the need for a separate selenization step or at least reduce strongly the amount of Se used in that step.

The work plan strategy is based on a step-by-step selection of technology combinations limiting complexity of the project, optimising resource investment and reducing technical risks. The main steps are: Initially start lab-scale deposition using rigid substrates only to develop and select precursors and deposition technologies. Upon reaching certain milestones, RTD lines for large-area deposition will be started, followed by tests on flexible substrates.

From the perspective of strategy, timeline and milestones of the NOVA-CI(G)S project, a few elements need to be stressed.

Initially, RTD will focus on rigid substrates only, using a standard soda-lime glass substrate without barrier layer. Such rigid substrates allow higher efficiencies (due to higher temperature annealing step), while simpler deposition reduces technical complexity. Secondly, focusing on one substrate type limits workload on testing ink and deposition technology combinations.

For deposition, ink-jet and gravure printing have been identified as 2 technologies that are representative of high throughput potential techniques (1 contactless medium-high throughput / 1 contact very high throughput). A brief comparative study will be kicked off early on and the most suitable technology to be used for the ongoing lab work will be determined in month 4 (MS1).

Right at the project start, and parallel to selecting the standardised lab based deposition technology mentioned above, doctor blade will be used to allow initial tests of 4 families of precursors and derived inks. The system has low productivity, but is simple, flexible and expected to deliver required control over layer deposition for preliminary testing. Further, it requires limited investment as WP2 partners have experience with the technology and equipment available.

Upon selection of the lab-scale deposition technology, a standardised laboratory set-up will be installed at Umicore (WP1) and ZSW and EMPA (both WP2). The standardised set-up will allow WP1 to make initial analyses of the deposited layer, so feedback-loops are short and the number of samples that are selected for further processing by WP2 remain limited. Using the standardised set-up at WP2 ensures verification and copy of WP1 results.

In an iterative process, ink routes will be further developed on the selected deposition technology. Interesting CI(G)S layers will be activated by traditional methods and laser annealing (by ZSW) and in parallel by furnace RTP (by EMPA), until halfway of the project.

At month 18, the goal is to have achieved 50% of the target efficiency for a small area cell printed on rigid substrate (MS2). At this moment, the work plans and focus will be revised to focus on the most promising approaches. The overall strategy of the project will be assessed if goals are not met. At month 18 as well, the 2 best precursor / technology combinations and derived inks will be selected for further improvement in the lab on rigid & flexible substrates (MS3). Parallel to that, a proof of concept study will be conducted for large-scale printing on rigid substrate (MS4) (study will aim at printing on a rigid substrate at a throughput of 30 m2/min a layer replicating the lab-based pre-annealed layer produced by WP2) as be the basis for selection in month 31 of either ink-jet or gravure deposition technique for the feasibility study for large-scale printing on rigid substrate (MS5). In light of results obtained until now, the MS2 will be postponed by 6 months until M24 and activities on flexible substrates could only, at best, start after M24. More details are provided in section 3.2.3 of the present report.

If 50% of the target efficiency for a small area cell printed on flexible substrate is achieved (MS6), the decision will be taken, at month 31, to proceed with proof of concept study (MS7) for R2R (ink-jet, gravure) large-scale printing on flexible substrate. Study will aim at printing on a flexible substrate at a throughput in the 50 – 100 m2/min range a layer replicating the lab-based pre-annealed layer produced by WP2. Over the duration of the project, the integration of annealing into large-scale printing will be examined (MS8).

Project Results:
The printing of CIGS consists of 5 steps being the preparation of the precursor powder, the preparation of an ink, the deposition of a wet ink layer, the post-deposition step of drying and debinding the layer, finally followed by the crystallization of the layer into an efficient CIGS absorber layer.

To record moderate efficiency, printed CIGS precursor materials face 3 technical challenges being:

1. The formation of phase pure CIGS ... which requires that the precursors transform completrely into CIGS, that the dispersants are completely removed, that the transformation process is carefully designed

2. The purity of precursors to be used and also the need for dispersants to be removed at low temperature

3. The microstructure of CIGS layers is to be highly dense and needs to feature large grains ... which require that the dried film with numerous pores and defects needs to be self-healing and that the precursor material generates liquid intermediate phase during the transformation into CIGS

Within the NOVA-CI(G)S project, these 3 technical challenges were partially met bringing our developments to an efficiency of our best cells measured in the range of 8-9 % with however sometimes issues around efficiency spread. The biggest challenge still to be solved is the ability to generate crack-free CIGS layers. Three other challegens are also the stability upon formulation and drying of the precursor materials, the need to select the right dispersant in the right amount with a preference for as little dispersant as possible obviously, and last but not least the need to obtain dense large grains to record high efficiency.

Over the course of NOVA-CI(G)S, all partners have gained a better insight into the high technological challenges to solve in order to obtain high efficiency of printed CIGS. It is believed that based on our strategies and developped precursors / inks, a reproducible efficiency (on rigid subtrates) of 10% can be obtained within a short timeframe. It is however uncertain that bridging the gap to 20% cell efficiency with the developped technology is feasible or attainable within a timeframe of a couple of years.
Potential Impact:
Technology risk is considered very high:

- Results until now are 8% efficiency at cell level and good insight in technological challenges
- It could be imagined with more time and resources to reach 10% cell efficiency in 2015
- It is however irrelevant since the efficiency to reach for CIGS would be 20% by 2020 in order to compete with CIGS produced by vacuum processes (this is highly unlikely)
- Printing as an alternative to vacuum processes has shown how challenging it is and similar conclusions are reached in other similar projects of which we are aware

Business risk is considered very high:

- Initial market assessment proved interesting although already considered risky
- Market for CIGS absorber materials (across all deposition technologies) was then estimated at close to 300 Meuro by 2016 with potential to double by 2020
- Penetration of printing was and is still impossible to predict for sure since it depends on reaching by printing similar efficiencies than with vacuum processes
- The market which was targeted was new CIGS production plants since the financial incentive (up to 20-25% CAPEX reduction, up to 20% material utilization, up to 15% decrease in module production cost) was not considered sufficient to initiate retro-fitting at current plants
- The market was then evaluated at 5 GWp market increase till 2020 and provided printing would dominate (we now know that it will never be the case in a 10-year timeframe), the maximum CIGS inks market could be 300Meuro+ in 2020 possibly leading to a profit pool of max 100 Meuro to be shared among all players
- The market for photovoltaics has however changed over the years
- CIGS efficiency has to grow further to possibly compete with crystalline silicon, this creates a moving target for printed CIGS which is impossible to catch up with present technology
- CIGS industry is also looking at standardizing production processes to compete on cost with crystalline silicon and printing of CIGS is not a priority for the industry
- CIGS market will grow slower than expected and only a few players (Asian players with mostly operations of module production in Asia) will call the shots with a strong impact on the profit pool

List of Websites:
www.novacigs.info
fabrice.stassin@umicore.com