Damp Heat Degradation of CIGS Solar Modules

(1)

Master Level Thesis

European Solar Engineering School No. 222, June 2017

Damp Heat Degradation of CIGS Solar Modules

Title

Master thesis 15 hp, 2017 Solar Energy Engineering Author:

Jose Cano Garcia Supervisors:

Désirée Kroner Dorrit Roosen-Melsen

Examiner:

Ewa Wäckelgård Course Code: MÖ3031 Examination date: 08-06-2017

Dalarna University Energy and Environmental

Technology

(2)

(3)

Abstract

Due to the short period that some photovoltaic technologies have taken part on the solar energy market, it is crucial to evaluate the long term stability of solar cells belonging to those technologies in order to ensure a minimum lifetime of their performance.

Accelerated degradation tests are thus carried out to achieve such goals.

The present study analyzes the encapsulation effects on co-evaporated manufactured Copper Indium Gallium Selenide (CIGS) solar cells under damp heat conditions, consisting in 85 °C and 85 % relative humidity, during an approximated period of 1000 hours. The experimental procedure has been carried out at Solliance Solar Research facilities.

Since the encapsulation packages play a critical role as a protection to achieve long term stability of the solar cells and modules, several packaging structures and materials has been taken into study. Thus, eighteen types of mini modules were manufactured including different combinations of encapsulants, front sheet foils, thin film protective barriers and CIGS cells from different manufacturers. The design of these mini modules and the manufacturing process to obtain them is also presented in this work. Various characterization techniques were carried out in order to acquire the required information about the solar cells and encapsulants performance along the damp heat degradation process.

The results exposed that encapsulation packages including thin film barriers between the encapsulant and the front sheet foil allowed a longer solar cell lifetime due to their remarkable protection against moisture ingress. Moreover, the degradation of the molybdenum layer included in the CIGS cells was found as principal cause of efficiency decrement and end of performance of solar cells protected by regular encapsulant and front sheet foils. Some other findings in relation with the evaluated components are shown along the present study.

(4)

Acknowledgment

I would like to thank all those I met during my period at Dalarna University, from teachers to friends, who made me grow professionally and personally and made me feel like at home.

To all TNO and ECN employees that gave me the opportunity to develop this work and also taught me much more than I could have ever expected prior to my arrival to Solliance, specially to my supervisor Dorrit, and to Mirjam, Monique, Henk and Klaas.

To my Solliance and SEAC colleagues for their friendship and help during my stay at the Netherlands.

And above all, of course, to my family. Nothing would be possible without their unconditional love and support.

(5)

Nomenclature

AlOx Aluminium Oxide CdS Cadmium Sulfide

CIGS Copper Indium Gallium Selenide. It may refer to (a) thin-film solar cell or (b) absorber layer included in solar cell

Eff Efficiency

EVA Ethylene Vinyl Acetate FF Fill Factor

i-ZnO intrinsic Zinc Oxide

Jsc Short circuit current density JV Current density-Voltage

Mo Molybdenum

OCP Organic Coating for Planarization

PECVD Plasma-Enhanced Chemical Vapor Deposition

PET Polyethylene Terepthalate PIB Polyisobuttylene

PO Polyolefin

RH Relative Humidity (also referred as Saturation degree) Rs Series resistance

Rsh Shunt resistance (or Parallel resistance) SiN Silicon Nitride

SLG Soda Lime Glass

TCO Transparent Conductive Oxide (i-ZnO + ZnO:Al layers) Voc Open circuit voltage

WVTR Water Vapor Transmission Rate ZnO:Al Aluminium-doped Zinc Oxide

(7)

1 Introduction

Since the early 19th century when the photovoltaic effect was observed for the first time in history to the present days the study, development and implementation of photovoltaics devices have experimented an incredibly growth.

Technologies of the solar cells are usually divided into three groups or generations. First solar cells generation consist in monocrystalyne silicon and polycrystaline silicon cells which are based on silicon wafers and have a normal performance between 15 - 20 %.

These solar cells are the dominant in the current market and are typically installed in rooftops or solar power plants. The main benefits of these solar cells are their good performance and their high stability, proven along years of usage. However, the disadvantages for some purposes are that they are rigid and also require a lot of production energy. (Solar facts and advice, 2013)

The second generation solar cells, the so-called thin film solar cells at first point, are based on the technologies of amorphous silicon, CIGS and CdTe, which normally have a performance around 10 – 15 %. Second generation solar cells have a much lower material consumption since the material thickness varies from few nanometers to tens of micrometers. The main advantages of these cells in comparison with first generation are that cell structure is less susceptible when it is shaded and it offers bigger design flexibility for a large variety of applications. Moreover the color allows easy integration into most exterior designs, it has a lower price as a result of a lower material usage and it has a shorter payback time in terms of energy which provides production and resource savings.

Solar cells of third generation use organic materials as small molecules or polymers and high performance experimental multi-junction solar. These technologies are still into development stage and they have very low commercial applications due to a very high price of production. (Solar facts and advice, 2013)

The present study focuses on the research of second generation CIGS solar cells. The emergence of the third generation cells doesn’t stand for a decrease in the research and development of the previous generation technologies.

Due to economic and energy points of view the lifetime expectations for a solar panel must meet 20 or 25 years as minimum. Thus, installations of technologies developed mostly in recent years have not been able to be proven in-situ in order to provide reliable information about the lifetime characteristics. To obtain proper values for lifetime expectation of a CIGS solar cell the degradation mechanisms have to be identified and analyzed. The next step is to recreate in some extent the fore mentioned mechanisms over the cell and analyze the consequent behavior of the cell. At that point it is needed to concern about what can be changed in the manufacturing process of the cell in order to obtain a better response when the degradation mechanisms reach the cell surface and what encapsulation and type of materials are most suitable in order to avoid the deterioration agents penetrate enough to affect the proper performance of the solar cell.

There are several factors that lead to degradation of CIGS cells. Along this work the degradation due to damp heat at the encapsulation of the cells into solar modules will be under study. The structure of a typical CIGS solar cell, shown in Figure 1, is based on a soda lime glass (SLG), polyimide or metal foils substrate for mechanical stability; a molybdenum (Mo) back contact; the CIGS absorber and Cadmium Sulfide (CdS) buffer

(8)

for the photovoltaic effect generation; an intrinsic Zinc Oxide (i-ZnO) not conductive thin layer and an Aluminium-doped Zinc Oxide (ZnO:Al) front contact layer whit an antireflection coating. Both i-ZnO and ZnO:Al layers shape the socalled transparent conductive oxide (TCO). In order to create solar panels the CIGS solar cells acquire some extra frontal layers for protection and for an improvement in the mechanic properties. The mentioned layers consist in frontal protection, encapsulant and barrier thin film. The frontal protection is commonly glass or a front sheet made of a transparent polymeric film in case of flexible CIGS panels. The encapsulant is a transparent polymeric layer providing adhesion between the cell and the frontal protection with minimal gas inclusion. The barrier thin film is mostly used for flexible CIGS solar panels and consists in a transparent thin film providing water, oxygen, etc. barrier function.

Figure 1. Structure of a typical CIGS cell

In order to be able to develop a further study of the damp heat deterioration process over a flexible CIGS cell it is necessary to focus at a fist point in the cells deposited on a glass substrate since the qualitative requirements of a back protection are better accomplished by the use of glass instead of polymeric film. Thus, an analysis of what type of materials used as encapsulants and as thin film barriers present a greater protection against the deterioration due to damp heat and what type of encapsulation structure according to the position that thin film barrier takes places in the encapsulation will be developed in the present work.

The elaboration of the analyzed CIGS solar cells and modules, degradation procedure, necessary methodology and facilities in order to satisfactorily obtain the cell performance before and after degradation and the collected data from analysis are provided by the Netherlands Organization for Applied Scientific Research (TNO). The company elaborates the mentioned research process within the Solliance Solar Research facilities sited in Eindhoven where the current study is developed.

(9)

Aims 1.1

The overall aim of this thesis is to evaluate the CIGS solar modules performance after accelerated processes of degradation by damp heat in order to obtain valuable information of what encapsulant procedures and materials are most suitable for a greater lifetime of the CIGS modules. The study comprises a stage of research within a greater research project than involves stages from CIGS solar cells manufacture by different deposition techniques to final performance of flexible CIGS solar panels. The study is based on the stage of degradation of encapsulated cells into small CIGS panels, procuring measured values of the cell behavior.

The research does not allow a real calculation of the solar cell lifetime. It mainly provides information about the impact of the deterioration applied over the CIGS panel layers and to have a better understanding of how to improve the elements and structure used for CIGS modules.

Method 1.2

In order to obtain the desired information about the modules performance and to achieve the aims previously defined several steps are required:

 Literature study on CIGS cells deterioration, life prediction for CIGS solar modules, thin-film barriers and damp heat stability of encapsulated CIGS cells among other related topics to gain a thorough understanding of the required work.

 Design of the adequate layout of the CIGS cells and modules in order to obtain the desired goals of research.

 Manufacturing and encapsulation of CIGS solar cells into small modules by following different patterns of encapsulation structure and by the use of different materials.

 Enforcement of accelerated process of degradation by damp heat during predefined periods.

 Several test methods in order to obtain valuable data about the changes in the solar cell samples behavior in comparison with previous performance to the deterioration process. The tests are carried out by the execution of electric measurements obtaining IV curves, efficiency, open circuit voltage, current density, fill factor, shunt resistance and series resistance at predefined irradiation, light transmission and thermography measurements, etc.

 Repetition of degradation and testing in order to acquire performance analysis in several steps of the degradation process.

 Processing and analysis of data obtained by the measurements.

 Elaboration of results according with the information obtained during the execution of the method and the obtained knowledge through the process.

(10)

2 Background

According with the literature studied about the CIGS degradation, one can postulate an empirical kinetic model to fit the degradation rate by assuming that the rate of degradation is proportional to the concentration of water at the surface of the cell:

𝑅_𝐷 = 𝑘_𝑜𝑒⁽

−𝐸_{𝑎,𝑑𝑒𝑔} 𝑅𝑇 )

[ 𝑅𝐻_{𝑐𝑒𝑙𝑙} 1 − 𝑅𝐻_{𝑐𝑒𝑙𝑙}+ ɛ]

Equation 1 where 𝑅_𝐷 is the degradation rate, k_o is a rate constant, −𝐸_{𝑎,𝑑𝑒𝑔} is the thermal activation energy of degradation, 𝑅 is the universal gas constant, ɛ is an small constant to prevent infinite rate at 100 % relative humidity (RH), and 𝑇 is the temperature of the module.

(Coyle, 2011)

The humidity that the cell experiences at any time is simply

𝑅𝐻_{𝑐𝑒𝑙𝑙} = 𝐶_𝐸 𝑆_𝐸

Equation 2 where 𝐶_𝐸 is the concentration of water in the encapsulant (g/cm³), and 𝑆_𝐸 is the solubility of water in the encapsulant (g/cm³). (Coyle, 2011)

Unfortunately many of the required values on these equations are missing for present study, as the rate constant, thermal activation energy of degradation or solubility of the thin film barriers. Thus, the calculus of the degradation rate at the different stages of the degradation process was not able to be achieved. For this reason the content of the present thesis is not completely consider as a quantitative but as a qualitative study. However, the theoretical framework provided by the equations allows a proper understanding of the degradation process that takes place at the cell level in relation with the encapsulant degree of saturation.

A previous experiment with same aims as the present study was carried out at Solliance facilities. After the detailed analysis of the processed data extracted from the obtained results it was determined the necessity of designing a new experimental layout in order to avoid some flaws noticed along the experiment, as the moisture ingress through the edge or an incomplete protection of cell areas by the thin film barriers. The new planed layout was designed to prevent the previously observed defects and, especially, to ensure that the moisture ingress occurs through the front sheet foil surface and not due to the diffusion from the edges of the mini module.

(11)

3 Experimental procedure

The types of mini modules, according to their characteristics, and their manufacturing process are fully described in this chapter. First of all, test and analyzing procedures are described.

Degradation and analysis techniques 3.1

The components of the mini modules under study were exposed to a standard degradation process along the experiment. Diverse analysis techniques were carried out during the time that the experiment took place in order to obtain valuable information about the degradation process of the module components

3.1.1. Degradation test conditions

Standard damp heat is a type of accelerated lifetime test for PV technologies, consisting in exposure to 85 °C and 85 % RH conditions. Damp heat test is one of the tests included in the certified qualification IEC 61646 for long term reliability of thin film PV modules.

According to IEC 61646, 1000 hours of damp heat exposure should represent 25 years of field exposure in Miami. Nevertheless, similarities between real field conditions and accelerated tests are still under debate. (Theelen, 2015)

3.1.2. Analysis techniques

Current density-voltage (JV)

Measuring the current-voltage (IV) curve is the most standard measurement for a solar cell. Usually a cell is illuminated, a voltage is applied and the current is measured. Due to the experimental character of the present study and the fact that the active area of the cells under study does not have the exact same dimensions, current density values are found more valuable for a proper cell performance comparison than current values.

JV measurements were obtained using a solar simulator and controlled by using a JV runner software. The basic operating properties of the cells were obtained, including current density, voltage, fill factor, series resistance, shunt resistance and efficiency. The voltage was varied from -0.5 to +0.8 V in steps of 0.002 V and the current was measured, with a 1 ms delay between each reading.

Lock-in thermography

The lock-in thermography technique means that the heat sources in an object are active in a periodically pulsed manner, the object is imaged by an infrared camera, and all images captured in a certain acquisition time and sent to a computer. The object is heated by a Light Emitting Diode (LED) panel.

Lock-in thermography is very useful for detecting shunts in a solar cell. Since shunts in a solar cell heat up easily than the rest of the solar cell area, they should be visible with this

(12)

set up. For consistency all measurements where done with a lock in frequency of 25 Hz, a camera frequency of 100 Hz and a measurement time of 500 ms.

Photography

The photographs are one of the simplest characterizations that have been performed. It is quick and very useful for documenting the evolution of the aspect of the samples during the degradation. Photographs were taken by the use of a scanner.

Resistance

Resistance values are easily measured by the use of a simple handheld multimeter with accuracy equal to ±0.1 %. The four pads of each sample are connected to the multimeter terminals.

Light Transmission

Light transmission provides the percentage values of light transmitted through a material at certain wavelength. At first, the laser beam provides a baseline measurement. The modules are introduced and the laser beam provides transmission values by comparing the obtained light transmission through the materials with the baseline values.

To ensure a proper comparison between the transmission values of the encapsulants, all measurements have been done at the same spot of the modules surface. The measurements were taken for wavelengths at range of absorption of CIGS material, from 300 nm to 1100 nm. Since the stack is attached to a SLG substrate the measuring beam provides the values of the light transmission through a package consisting in foil, thin film barrier (in case it is included), encapsulants and glass.

Materials and structures under evaluation 3.2

In order to evaluate the performance of different encapsulants and thin-film barriers as a protection for CIGS cells under damp heat degradation, 27 modules were manufactured using different materials and following all possible structures and combinations.

To study the degradation mechanisms in different layers of a typical CIGS cell, each module includes:

 3 solar cells

 1 cell totally cover by the TCO layer

 1 Mo layer surface

Cells are not interconnected and their performances are analyzed independently. Modules including cells/surfaces obtained from Solliance facilities were manufactured twice with the same structure to find out if the results are repeatable.

(13)

Materials under study are:

 CIGS cells manufactured at Nantes University laboratories

 CIGS cells manufactured at Solliance laboratories

 Polyethylene Terepthalate (PET) front sheet foil (100 µm)

 Ethylene Vinyl Acetate (EVA) encapsulant (200 µm)

 Polyolefin (PO) encapsulant (200 µm)

 Organic Coating for Planarization + Silicon Nitride (OCP+SiN) thin-film barrier on PET

 SiN thin-film barrier (150 nm)

 Aluminium Oxide (AlOx) thin-film barrier (20 nm) on PET (125 µm)

Table 1 shows all manufactured modules and their characteristics.

Table 1. Module number and components Module

number CIGS

manufacturer Encapsulant Thin-film

barrier type Thin-fim barrier location

1 Solliance EVA

1(2) Solliance EVA

2 Solliance PO

2(2) Solliance PO

3 Nantes EVA

4 Nantes PO

5 Solliance EVA SiN Deposited over cell surface

5(2) Solliance EVA SiN Deposited over cell surface

6 Solliance PO SiN Deposited over cell surface

6(2) Solliance PO SiN Deposited over cell surface

7 Nantes EVA SiN Deposited over cell surface

8 Nantes PO SiN Deposited over cell surface

9 Solliance EVA OCP + SiN Between front sheet and encapsulant 9(2) Solliance EVA OCP + SiN Between front sheet and encapsulant 10 Solliance PO OCP + SiN Between front sheet and encapsulant 10(2) Solliance PO OCP + SiN Between front sheet and encapsulant 11 Nantes EVA OCP + SiN Between front sheet and encapsulant

12 Nantes PO OCP + SiN Between front sheet and encapsulant

13 Solliance EVA AlOx Between front sheet and encapsulant

13(2) Solliance EVA AlOx Between front sheet and encapsulant

14 Solliance PO AlOx Between front sheet and encapsulant

14(2) Solliance PO AlOx Between front sheet and encapsulant

15 Nantes EVA AlOx Between front sheet and encapsulant

16 Nantes PO AlOx Between front sheet and encapsulant

17 Solliance SiN Deposited over cell surface

17(2) Solliance SiN Deposited over cell surface

18 Nantes SiN Deposited over cell surface

(14)

Modules layout 3.3

The structure of the mini modules consists of:

 A 100x85 mm SLG substrate

 Polyisobutylene (PIB) edge sealant in each side

 An encapsulant layer covering the glass substrate

 An square shaped ring made of encapsulant surrounding the cells/samples

 Cells and samples surrounded by the encapsulant ring, on the middle of the mini module.

 The required encapsulant depending on the module characteristics

 The required PET foil depending on module characteristics.

Figure 2 illustrates the modules final aspect, including a red dashed line indicating the cross-section for the view of Figure 3, where it is possible to observe the inner layers. As it can be observed the contact wires make their way to the outside by passing between the two edge sealant layers.

A first look at the cross section reveals that there are differences in the height of the materials which would lead to not obtain a flat surface on the top of the modules.

Nevertheless the lamination process carried out at the oven melt down the polymer materials and they reach the same height with respect to the glass substrate surface.

Figure 2. Final aspect of a mini module. From left to right: TCO sample, three solar cells and Mo sample

(15)

Figure 3. Overview of module cross section from red dashed line from Figure 2. The soda lime glass (SLG) substrate does not allow moisture diffusion through the back side. The PIB edge sealant and the thick encapsulant layer avoid the moisture ingress through the edges. Moisture only reaches the sample surface (in dark blue over the SLG at the up-right part of the figure) by diffusing through the PET and the thin encapsulant layer.

Modules 17, 17(2) and 18 were planned to contain cells and surfaces protected by SiN thin film barrier deposited on them and to not be protected by any encapsulant or foil. Their structure thus is different to the previously explained modules. To obtain a good mechanically stable structure for these 3 modules, the cells/surfaces are placed on a layer of 200 µm thick encapsulant which is at the same time placed on a 100x100 mm SLG substrate. The sides of the cells and surfaces are protected with four dummy glasses which also allow the pads to lay on them. The pads are protected by other dummy glasses over them, with an encapsulant layer in between bottom and top glasses to provide adhesion and protection. Once the modules are laminated in the oven they obtain a final look as it is shown in Figure 4, with a red dashed line indicating the cross-section for the view of Figure 5.

(16)

Figure 4. Final aspect of modules with non-encapsulated samples

Figure 5. Overview of module with non-encapsulated samples cross section from red dashed line from Figure 4. Encapsulant is only used to provide adhesion between the SLG pieces.

Modules fabrication process 3.4

Several steps were followed in order to obtain the aforementioned mini modules. The procedure for their obtainment is fully described in this section.

3.4.1. Cells shape obtainment per slide

The solar cells and samples which were included in the manufactured modules were extracted from CIGS slides provided by Solliance and Nantes laboratiories. The slides have different dimensions depending on the manufacturer. All of the slides have rectangular

(17)

1 mm thick SLG rectangular plate. The first layer consists of a 19 mm height Mo layer.

Then the whole area of the slide is covered by the rest of the solar cell layers (CIGS+CdS+TCO). Cells and samples are extracted from these slides by the further explained process. Nantes University also provides slides consisting of only the Mo layer on the SLG.

Nantes slides

Nantes provides CIGS slides with 75x25 mm and Mo slides with 100x25 mm. The areas close to the edges on the CIGS slides don’t have homogeneity enough to be used as cells.

To provide shape to the future cells and samples some areas of the slides, in light blue in Figure 6, were removed by the use of laser.

Figure 6. Nantes CIGS slides after laser ablation. Light blue areas became bare SLG.

6 cells/surfaces per each slide were able to be obtained. At the end, 9 modules were produced including cells/surfaces manufactured by Nantes. Each module includes 3 cells, 1 cell totally covered by TCO and 1 Mo layer surface, leading to 45 cells/surfaces in total.

8 slides were provided from Nantes University (obtained from two different runs of the production line #1600 and #1601), resulting in 48 cells.

From the 45 required cells/surfaces, 27 are cells, 9 are cells totally covered by TCO and 9 are Mo layer surfaces. Mo surfaces were not extracted from Nantes CIGS slides but from the Nantes Mo slides. Mo surfaces have the same size as the rest of cells/surfaces but it was possible to extract 10 from each slide since Mo slides dimensions are 100x25 mm as Figure 7 shows.

Figure 7. Nantes Mo slide after laser ablation. Mo layer was removed at light blue areas.

(18)

Solliance slides

The dimensions of Solliance CIGS slides are 100x25 mm and, as the Nantes slides, areas close to the edges don’t have homogeneity enough to be used as cells. By the use of laser the areas shown in light blue on Figure 8 were removed.

Figure 8. Solliance CIGS slide after laser ablation

It is possible to obtain 8 cells from each slide. There must be 18 modules containing cells/surfaces manufactured by Solliance. Each module has 3 cells, 1 cell totally covered by TCO and 1 Mo layer surface, leading to 90 cells/surfaces in total. It means a minimum of 12 slides, resulting in 96 cells, were needed. Due to the necessity of covering some of the slides by a SiN thin film barrier, as it will be explained later on, and the high probability of failure during the cutting process into individual cells/surfaces, 18 slides manufactured at Solliance laboratories were provided.

From the 90 required cells/surfaces, 54 are cells, 18 are cells totally covered by TCO and 18 are Mo layer surfaces.

The process to obtain the Mo surfaces is done by scratching all the layers placed over the Mo layer. Important is to notice that there is a difference in shape between the Nantes Mo surfaces and the Solliance Mo surfaces. Solliance Mo layers are obtained once the rest of layers are scratched resulting in samples of 18x8 mm. Since Nantes Mo is provided as Mo layer slide, the samples size result is 23x8 mm.

3.4.2. Thermography measurement

After the laser ablation, thermography images were taken in order to find possible shunts in the rectangular shapes before converting them into cells, especially in the edges of the areas which became the active area. The results of the thermography can allow to decide which surfaces should remain as TCO samples and which should be selected to become cells and Mo surfaces, keeping the most shunted as TCO samples. In order to optimize the cells selection further JV measurements alongside the thermography images were taken into account.

Lock in Thermography showed some shunted spots in the edges of the rectangular shapes,

(19)

shows the thermography image of one CIGS slide manufactured by Nantes after laser ablation. It is possible to observe the shunts as white lines and red spots placed at the edges of the rectangles.

Figure 9. Thermography image of Nantes CIGS slide after laser ablation. Warm colours as red or white, visible at the edges and corners of the rectangular shapes, indicate a higher temperature in contrast with the rest of the areas, in blue. The higher temperatures evidence shunts.

3.4.3. Cells, TCO samples and Mo samples obtainment

The surfaces which were not kept as TCO samples were converted into cells. In order to obtain the cells, a 13 mm long surface of CIGS+CdS+TCO layers were removed by scratching by the use of a blade, leaving only the Mo layer as the back contact of the cell.

The resulting appearance is shown in Figure 10(a), where the light blue consists in CIGS+CdS+TCO and represents the top contact, the dark blue consists in Mo+CIGS+CdS+TCO and represents the active area, and the grey area consists in Mo and represents the back contact.

In order to achieve a reliable JV curve measurement two small drops of silver paste were deposited on two cells, one drop on the Mo back contact and another on the TCO front contact of the cells. The silver paste allows a proper electric contact in order to obtain a JV curve. The measurement confirmed an extremely low electric performance due to the shunted spots. To supress this effect a small amount (around 0.5 mm) of TCO+CdS+CIGS placed in the edges of the active area was removed by scratch. This led to a final aspect shown in Figure 10(b). A new JV measurement showed a correct electric performance of the cells. Thus, the scratch of the active area edges was executed for the rest of the cells.

(20)

Figure 10. CIGS solar cell appearance (a) after accessing to back contact and (b) after active area edges scratch

By the use of silver-plated nickel wire between the equipment measuring pads and the back and front contacts of the cells, a first JV measurement was carried out to have information of how the cells perform before the contact wires are placed. This measurement is not completely reliable since the contact between the wire and the cell is not properly done but it allows a first approach to compare the performance among cells. The Solliance cells with worst performance were completely scratched into Mo surfaces to be used as Mo samples.

3.4.4. SiN deposition over cells surface

To have the future possibility of making electric contact with the front and back contact areas it is needed to provide a mask to each cell/surface previous to the SiN barrier deposition. These masks were made by kapton tape and placed in the cleanroom after a cleaning of the cells/surfaces by rinsing them with isopropilene and drying them out with a nitrogen pistol. Figure 11(a) shows the masks (squares in dark brown color) over a cell.

The masks were positioned by hand so their dimensions and position in the figure are only an approximation.

The SiN deposition over the cells was executed by PECVD at Holst Center facilities. Due to any possible failure during the cutting into individual cells process may occur, a larger number than required of cells/surfaces were delivered. Thus 3 Nantes slides (containing cells and TCO surfaces), 1 Nantes Mo slide and 7 Solliance slides (containing cells, TCO and Mo surfaces) were selected for this purpose. The thickness of the thin film barrier layer is approximately 150 nm.

3.4.5. Cutting

Due to CIGS manufacturers cannot guarantee that all slides have a similar electric performance and long term stability it was necessary to cut the cells surfaces into individual samples and randomize them into the different types of modules. Otherwise the

(21)

performance of some cells could be attributed to the module characteristics instead to the actual cells slide of origin. The cutting process was done by the use of diamond blade. The final look of every cell sample is shown in Figure 11(b).

Figure 11. (a) Cell with masks for SiN deposition and (b) final look of cell samples

3.4.6. Gold (Au) contact deposition

To ensure a proper electric contact with the wires, not incorporated to the samples yet, a Au layer of 60 nm was deposited on all cells/surfaces by the use of a Au sputtering tool. A mask made by a 200 µm thick PET foil was used on the top of each samples to allow the gold be deposited at the desired place. Since the mask was placed by hand the size of the deposited area varies from one sample to another. Nevertheless they keep a similar look as exhibit at Figure 12(a), except for Solliance Mo samples which have an appearance as shown in Figure 12(b) due to their smaller surface area once the rest of the layers were scratched.

Figure 12. Appearance after Au layer deposition of (a) cells and (b) Solliance Mo samples

(22)

At this point there were still more cells than required for the experiment so a new JV measurement was taken in order to decide which cells have the best performance and consequently be selected to be part of the experiment. In this case reliable JV measurements were carried out since the gold layer allowed a proper contact between the equipment measuring pads and the cells.

3.4.7. Contact pads obtainment

For the electric measurements, once the cells are encapsulated, four flat silver-plated nickel wires were needed to be added to the cell. The wires, also referred as pads, have a thickness of 50 µm and 2 mm width. Since the wires have to pass through the edge sealant layers leading to a lower adhesion between layers, the width of the wires was reduced to 1 mm by cutting them in half. Each of them was cut with a length of 5 cm.

3.4.8. Electric conductive adhesive and wires placement

Parts of the Au surfaces were covered by a small drop of an electrically conductive adhesive epoxy in order to obtain mechanical and electrical adhesion between the cell and the pads. For such deposition it was used a mask made by stain steel, leaving a thin layer of 1x2.5 mm. The epoxy used was EMS DB-1588-4. The contact wires were placed on the conductive adhesive with the help of small pieces of kapton tape, making the top of the wire match the conductive epoxy surface, and introduced in the oven for the curing process. Cure schedule of the product is 30 minutes at 120 °C. Once the conductive epoxy was cured the cells obtained a look as shown in Figure 13.

Figure 13. Overview of a cell with contact pads (tilted image in comparison with previous figures) Since the cells and samples were ready for the lamination into modules, the first JV measurement included as a part of the experiment results was carried out. Additionally, a first measurement of the resistance of the TCO and Mo samples was obtained.

3.4.9. Glass substrate and encapsulant ring obtainment

Plates of 1 mm thick SLG cut with the dimensions of 100x85 mm were used as a main substrate for each module to lay the samples on.

The cells and surfaces, which were placed together on the center of the glass substrate, are surrounded by a square shaped ring made by EVA or PO, depending on the module characteristics. Dimensions of the encapsulant ring are exhibit at Figure 14.

(23)

Figure 14 Dimensions of encapsulant square shaped ring

The encapsulant ring must have the same thickness than the CIGS cells. Due to the glass used as a substrate for CIGS cells is 1 mm thick and the encapsulants thickness is 200 µm, the mentioned rings were made by a stack of 5 layers of encapsulants, reaching thus the same thickness as the glass.

3.4.10. Edge sealing and lamination

The edge sealing was done by the use of the tape HelioSeal® PVS 101, provided by Kömmerling. This PIB tape is 0.5 mm thick and was positioned on the main substrate glass, along the external edge of the encapsulant ring, as indicated in Figure 15. An encapsulant layer was placed between the glass substrate and the encapsulant ring. The SLG substrate thus was previously prepared clean, dry and free of dust and grease.

Figure 15. Edge sealant and encapsulant ring

(24)

The lamination process was carried out at the oven at 145 °C during 11 minutes. Figure 16 shows the final aspect of Module 2 after the manufacturing process.

Figure 16. Final aspect of Module 2. From left to right: TCO sample, three solar cells (cell a, cell b, cell c according to the position on the module) and Mo sample

After the lamination of module 17, shown in Figure 17, some cracks in the upper glasses were observed, probably due to the pressure created on the sides during the vacuum process on the lamination oven. To avoid this problem in modules 17(2) and 18 the upper dummy glasses where replaced by PET, which still allows enough protection to the pads.

(25)

Figure 17. Final aspect of Module 17. From left to right: TCO sample, three solar cells (cell a, cell b, cell c according to the position on the module) and Mo sample

Nantes slides were provided in two production line runs, #1600 and #1601, of four slides each. The name of each cell/surface was written under each of them by scratching the glass surface. The names were assigned in accordance with the run, number of slide and position, being the number of slide a digit between 1 and 4 and the position that occupy in the slide a character between A and F, in alphabetic order.

The name assigned to Solliance cells followed as well a similar pattern, taking in account that there are 19 Solliance slides.

Cells and samples included in each module were selected randomly, trying to gather cells from different slides in each module. The final disposition due to the randomization of cells and surfaces on each module it is shown in Table 2.

(26)

Table 2. Module number and names of samples included

Module TCO Cell A Cell B Cell C Mo Module TCO Cell A Cell B Cell C Mo 1 CO 30 E CO 11 B CO 14 E CO 18 A CO 13 A 10 CO 30 F CO 05 F CO 13 G CO 14 F CO 13 H 1(2) CO 33 D CO 05 C CO 07 F CO 30 G CO 32 B 10(2) CO 30 H CO 07 D CO 11 A CO 17 F CO 11 D 2 CO 33 H CO 05 A CO 13 C CO 30 C CO 05 D 11 1600-2 E 1600-3 B 1601-2 D 1600-1 D MoN 02 2(2) CO 30 B CO 07 E CO 13 D CO 18 H CO 07 C 12 1601-2 E 1600-2 C 1600-3 D 1601-3 C MoN 06 3 1601-3 A 1600-2 D 1600-3 C 1601-2 B MoN 04 13 CO 14 C CO 11 H CO 17 A CO 17 C CO 11 G 4 1600-1 E 1600-2 A 1600-1 B 1601-3 D MoN 03 13(2) CO 33 G CO 07 H CO 11 F CO 33 E CO 07 A 5 CO 10 F CO 19 C CO 12 H CO 16 G CO 03 C 14 CO 14 A CO 05 E CO 17 G CO 18 E CO 33 A 5(2) CO 10 G CO 12 G CO 16 A CO 19 D CO 03 A 14(2) CO 33 B CO 13 B CO 17 B CO 30 D CO 32 C 6 CO 10 D CO 12 E CO 16 H CO 19 H CO 20 F 15 1600-3 E 1600-1 C 1600-2 B 1601-2 F MoN 07 6(2) CO 10 A CO 08 C CO 12 F CO 19 A CO 03 D 16 1600-3 F 1600-1 A 1601-2 C 1601-3 E MoN 05 7 1601-1 F 1601-1 A 1601-4 A 1601-4 C Mo 03 17 CO 03 E CO 16 B CO 16 E CO 12 A CO 03 B 8 1601-1 E 1600-4 A 1601-4 B 1601-4 E Mo 01 17(2) CO 03 F CO 12 D CO 16 D CO 16 F CO 20 A 9 CO 30 A CO 05 B CO 15 A CO 18 G CO 13 F 18 1601-1 C 1600-4 E 1601-4 D 1601-4 F Mo 02 9(2) CO 14 H CO 05 H CO 07 G CO 33 F CO 18 F

After all modules were manufactured the first complete characterization was carried out, consisting of:

 Photography of modules

 JV curve of cells

 Transmission of module encapsulant

 Resistance of TCO and Mo samples

Once this characterization was finished the modules where finally prepared to start the damp heat degradation process by introducing them into a climate chamber at 85 °C and 85 % RH.

The modules where extracted from the climate chamber several times to develop characterization at different stages of the degradation process. Due to the already observed trends it is important to carry out more frequent measures at the early hours of the experiment and less frequent after a certain time of the degradation process. The selected time steps of degradation were: 5 hours, 20 hours, 36 hours, 74 hours, 140 hours, 230 hours, 641 hours, 800 hours and 1032 hours. During every degradation step were measured the JV curve parameters of cells, the resistance of the TCO and Mo layers, and scanning images of the modules were taken. On another hand and due to expected small changes in the values, the measurement of light transmission through encapsulants was carried out only before the degradation and at 230 hours, 800 hours and 1032 hours.

(27)

4 Results

As aforementioned, in order to study the repeatability of the obtained results all modules including cells or surfaces manufactured by Solliance laboratories and same material structure have been manufactured twice. Due to the big amount of cells and surfaces under analysis, figures and tables along this section will show, in some cases, values of one of the modules which have been duplicated, for an easy and summarized exhibition of the results.

Along this section several graphs of the main electric parameters of the cells will be presented. It is needed to explain that the values of the degradation time, at the x axis, start at -10 hours. This point represents the moment before the lamination process and 0 hours represent the moment after lamination and before the introduction into the climate chamber.

All the processed data obtained from the measurements of the cells and samples along the present study are exhibit at the Appendix section.

Effect of lamination on cells 4.1

Before the required lamination process in order to obtain the mini modules was developed, the cells were characterized by the measurement of their JV performance. The obtained data allowed a first comparison between the electric performance of the cells before and after the lamination, since a JV measurement was carried out after the lamination process and before the damp heat degradation treatment. These values provide how the addition of encapsulant materials, PET, thin fil barriers and the lamination process may affect to the electric behavior of the cells. Table 3 shows the change in percentage of the most important electric parameters after the lamination process. These values represent the average of the three cells included in each module. Although values vary from cell to cell, even showing an improvement in some cases, in overall current density of most of the cells decreased between 0 and 4 % from their initial values. Voc values were also affected but in a lower percentage. Nevertheless most of FF values experienced in overall an improvement between 0 and 2.5 %.

(28)

Table 3. Average percentage change of Voc, Jsc, FF and Eff of cells after lamination process

Module ΔVoc (%) ΔJsc(%) ΔFF (%) ΔEff (%)

1 -0.5 0.3 0.3 0.0

2 -0.9 -2.8 0.7 -3.0

3 -0.5 -2.9 -0.1 -3.5

4 -0.1 -2.5 0.5 -2.0

5 0.7 -1.4 2.5 1.8

6 -0.5 -2.1 2.3 -0.3

7 -0.2 -1.5 0.8 -0.9

8 -0.5 -3.2 1.6 -1.9

9 -0.7 -2.9 1.0 -2.7

10 -0.4 -1.4 1.8 0.0

11 0.1 -3.9 -0.1 -3.8

12 -0.5 -3.0 0.9 -2.6

13 -1.3 -3.3 -0.5 -5.0

14 -1.7 -1.2 1.0 -1.9

15 -0.5 -1.3 0.2 -1.6

16 -0.4 -0.4 0.3 -0.5

17 0.0 -2.7 2.2 -0.6

18 -0.5 -1.1 0.5 -1.1

There was no observed correlation between the changes in these parameters and the materials used or the origin of the cells. At this point it is important to mention that measurements of the same cell at different occasions but in the same stage of degradation may provide slight differences values due to the uncertainty of the measurement equipment and human related factors.

J

_sc

early values 4.2

An unexpected phenomenon was observed during the first hours of degradation in all samples under study. JV measurements showed a decrement of the Jsc values between hour 5 and hour 20 of the damp heat process. This decrement varies between 4 and 10 % of the absolute value, depending on the sample, which results in a large decrease of the efficiency. The fact that the decrease was observed in all samples, at the exactly same time period and in such a large amount led to think that it is certainly an unnatural performance, according to literature and previous experiments (Theelen, 2015). After 20 hours of degradation the Jsc values of every samples become steady and only experienced further changes once the degradation of the cells is taking place. Figure 18(a) and Figure 18(b) show the Jsc values of the 150 first hours of degradation of the cells included in Module 1 and Module 16 as an example of the common Jsc performance of all cells.

(29)

Figure 18. Jsc values of first 150 hours of degradation of the three cells included in (a) Module 1and (b) Module 16

The best hypothesis that has been found is that arrangements in the JV measurement equipment, which took place between hours 5 and 20, affected the calibration and measuring process. Since this decrease did not occur due to degradation it is important to take into account its effects on the cells efficiency.

Modules without thin-film barrier 4.3

The first set of mini modules consist in the basic encapsulation of EVA or PO with a PET foil. All modules presented a big decrement in efficiency during the first 20 hours of degradation. This effect is observed in a larger scale for the cells manufactured at Solliance lab than those manufactures at Nantes facilities. Previous studies showed that for co- evaporated CIGS cells alkali-rich, as it is the case of the present samples, but not encapsulated, the most common parameter affecting the decrement in efficiency at the early hours of damp heat degradation is the Voc (Theelen, 2015). Cells packaged without thin film barriers, as it is the case, experienced a large decrement in Voc values during the first 20 hours of degradation.

Beside the degradation during the early hours aforementioned, for Solliance manufactured cells encapsulated with EVA it is possible to observe a degradation trend in efficiency starting after 140 hours of damp heat which lead to a later end of performance of the cells.

Figure 19(a) illustrates the efficiency trend of cells in Module 1. This end of functioning appears the earliest after 480 hours of degradation and the latest after 800 hours, depending on the cell. The main cause of the efficiency decrement is due to an increase of series resistance. Cells included in Module 1 showed a decrease in shunt resistance which start to affect the cells performance mostly after 230 hours of degradation. Although Module 1(2) also showed shunts in some cells, the decrease of this parameter was less pronounced than in Module 1. The optical analysis revealed that the reason for the end of electric performance of all these cells is the degradation of a line on the Mo layer. This line is placed right along the location of the edge of the active area, as it can be observed in all cells exposed on Figure 20, indicated within a red circle in one of them. Once the degradation of this specific line of Mo is completed the electric current has no possible path to reach the contacts and thus the electric measurement of the cells becomes impossible. This effect has been observed as well for Nantes cells encapsulated with EVA.

Thus, beside the differences in the Mo sputtering conditions from both manufacturers, the

(30)

degradation of EVA leads to a faster corrosion of the Mo layer than the rest of the layers and more specifically at the aforementioned location. At the late stage of damp heat exposure it was also observed a higher degree of Mo degradation around the conductive epoxy edge. This effect was found as well more pronounced for the cells encapsulated with EVA than with PO. The Mo samples at these modules showed a slow increment in resistance followed by a faster increase after 480 hours of degradation. This effect has been found in all Mo surfaces for all samples included in modules not protected by a thin film barrier.

Figure 19. Efficiency curves of cells included at (a) Module 1 which includes Solliance cells, EVA and PET, (b) Module 2 which includes Solliance cells, PO and PET, (c) Module 3 which includes Nantes cells, EVA and PET and (c) Module 4 which includes Nantes cells, PO and PET. Legends at the right of each graph indicate the module number and the position of the cells on the module.

Figure 20. Back side view of cells included in (a) Module 1 and (b) Module 3 at 800 hours of degradation. Red circle indicates the degraded line of Mo along the active area edge, visible in the rest of cells For modules 2 and 2(2), consisting in Solliance cells encapsulated with PO, the degradation of Mo is also visible but in a much lower degree, indicating that, even when the effect will eventually occur, the required time for it is longer. The same corrosion lines along the

(31)

active area edge started to take shape, slower than in modules encapsulated with EVA, and in one case caused the end of performance of one cell. In general terms the degradation of Solliance cells encapsulated with PO is less pronounced than for EVA encapsulated cells.

Most of them still performing at 1032 hours of damp heat although at low efficiency values. The main reason for the decrement in efficiency at these samples is the constant increment of series resistance and in some cases the creation of shunt paths in the active area. Nevertheless the fill factor is affected by the shunt resistance at later stage than the series. Figure 19(b) shows the efficiency values of Module 2 along the experiment. Some cells included in these modules presented some delamination spots placed at the upper part of the cell, which layers consist of TCO, CdS and CIGS. This effect was observed already at 74 hours of degradation on a cell included in Module 2 and later on most of cells at Module 2 and 2(2). The delamination of these layers has been found not exclusive for these two modules. Figure 21 illustrates the differences in visual degradation in comparison with Figure 20, including a red circle indicating the delamination signs in the front contact.

Figure 21. Back side view of cells included in Module 2 at 800 hours of degradation. Red circle indicates an advanced stage of delamination of the front contact layers

Nantes manufactured cells encapsulated with EVA, corresponding to Module 3, showed a similar behavior among each other and a good performance since the efficiency degradation trend is not as pronounced as the in the previously analyzed cells. The slight increment of the series resistance was the main parameter influencing the change in efficiency. As it was mentioned earlier the degradation of the Mo line along the edge of the active area was the reason of a sudden end of electric performance of the cells, in this case after 480 hours of damp heat for one of the cells and after 641 hours for the other two cells, as it can be observed in the efficiency curves shown in Figure 19(c).

Two out of the three cells included in Module 4, manufactured in Nantes and encapsulated with PO, present an irregular and unexpected electric behavior. Efficiency values of cells at Module 4 are presented in Figure 19(d). Cell placed at position 4c showed an incredibly fast decrement in efficiency from an early stage of the degradation process. The reason of this performance is a rapid creation of shunt paths in the cell. The effect of the decrement in shunt resistance led a fast reduction of the Voc values, resulting in an end of electric performance of the cell after 480 hours of degradation. The irregular performance of cell positioned at 4a is driven by a large decrease and eventual recover of shunt resistance.

Thus, for a better analysis of the encapsulant effect on the degradation, the performance of cell positioned at 4b gives a more accurate representation, since the creation of shunt paths which is related to manufacture defects was not observed. The mentioned cell has a decrease of Voc at the early hours of degradation and, once it stabilizes, the efficiency is affected by the series resistance. Anyways the overall performance is the best among all

(32)

cells encapsulated with EVA or PO and PET since efficiency remains at 9 % after 1032 hours of degradation.

Modules with SiN deposited on cells 4.4

In overall, cells protected with SiN barrier deposited on top of the cell performed slightly lower initial values than the cells without the barrier. This might be caused due to the PECVD process required to provide the SiN layer. Since the thin film barrier was deposited before the gold layer (which is needed for a proper electric contact), it was not possible to carry out a reliable JV measurement of the cells before and after the SiN deposition. Thus a comparison between the efficiency values before and after the PECVD process is missing and it is not possible to certainly attribute the lower efficiency values to this process.

Figure 22 shows the efficiency values of module containing cells protected by SiN barrier on cells. After a first look at the efficiency of all modules including SiN deposited on cells manufactured by Solliance it is easy to identify that all of them contain one cell which electric performance don’t follow the same trend as the rest of analysed cells. These cells experienced a large decrement of the fill factor at the early hours of degradation due to a combination of series resistance and shunt resistance bad performance, leading to a drop in efficiency around the half of their initial values in the first 200 hours. They have in common that all of them where extracted from the same slide: CO19. The differences in the performance of these cells from the rest might be caused by a poorer manufacturing process or any possible damage occurred during the SiN barrier sputtering or the treatment followed until the obtainment of the modules.

Figure 22. Efficiency curves of cells included at (a) Module 5which includes Solliance cells with SiN, EVA and PET, (b) Module 6 which includes Solliance cells with SiN, PO and PET, (c) Module 7 which includes Nantes cells with SiN, EVA and PET and (d) Module 8 which includes Nantes cells with SiN, PO and PET.

(33)

Beside the performance of the mentioned cells, it was observed that the trend of efficiency decrement of SiN covered cells was similar to the cells not protected with a barrier, except for Solliance cells encapsulated with EVA. Since many of the cells which were not protected with a barrier had an early end of performance, the compared trend between them and the SiN covered cells refers to the degradation trend until 480 hours of degradation. This similarity was observed in both the efficiency values at that time and the percentage of total efficiency decrease from degradation start point. Not only the efficiency values were alike for these two types but also the series resistance also had a similar increasing rate. In order to have a better understanding of the similar increase rate of Rs, the TCO and Mo surfaces included in these modules were observed. The resistance values of TCO and Mo samples for both cases, non-barrier protected and samples with SiN barrier deposited on top, are presented at Figure 23. TCO samples experienced a constant decrement in their conductivity from the beginning of the degradation process, shown as increment of their resistance. The resistance values have a similar starting range, between 10 and 20 ohms, for both cases, and end up within a similar range of resistance values once the experiment is finished at 1032 hours. For the Mo samples it is possible to see that in overall, besides some exceptions, the resistance values start at a similar range and remain steady during the first 480 hours. After this period, the resistance of the samples started to increase for both types, the samples not protected with barrier and those which have a SiN thin film barrier on top. Nevertheless, the large initial decrement of Voc for the SiN barrier protected samples extended to the first 74 hours instead of the only 20 hours for the non-barrier protected.

(34)

Figure 23. Resistance curves of (a) TCO samples and (c) Mo samples at modules non-barrier protected, and resistance curves of (b) TCO samples and (d) Mo samples protected with SiN barrier.

Important is to point out that the SiN barrier avoided the Mo degradation line along the active area observed in the non-barrier protected cells. This protection prevented from an early end of performance. Nevertheless images showed a large Mo degradation underneath the conductive adhesive epoxy, starting at first along its edge, shown in Figure 24. This effect is easily observed at 480 hours of degradation and it was detected earlier in modules packaged with EVA than in modules packaged with PO. Moreover EVA encapsulated samples shown a faster degradation of the mentioned Mo areas. Of course the encapsulants of all the mentioned modules must have reached the same degree of moisture saturation at the same damp heat exposure time. Thus, it seems that the SiN barrier sputtered on cells avoid the sudden end of performance and provided a slower degradation rate at the early stage of the damp heat process but after 100 hours the barrier seems to have no longer protection effect.

(35)

Figure 24. Back side view of Module 5 at 1032 hours of degradation. Red circle indicates the degradation of Mo matching the area where the conductive epoxy is placed, visible in the rest of cells and the Mo sample.

In overall, the main cause of efficiency lost in the cells covered by SiN was the increase of series resistance, which became more pronounced after 480 hours of degradation in many cells. Cells obtained from slide CO19 have shown low values in shunt resistance at early stage of degradation.

Modules with SiN barrier deposited on cells and not 4.5

encapsulated

Cells only protected by SiN deposited on top and without any type of encapsulant or foil experienced a short lifetime under damp heat conditions. The common pattern showed a large decrement of efficiency for the first 74 hours of degradation, driven mainly due to the Voc decrement, and followed by a less pronounces efficiency degradation, mainly driven by the increment of the series resistance. Six out of the nine cells presented an abrupt end of performance before 480 hours of degradation. Rest of the cells also experienced the sudden end of performance at different stages of the degradation process.

The efficiency values of cells at Module 17(2) and Module 18 are presented in Figure 25.

Figure 25. Efficiency curves of cells included at (a) Module 17(2) which includes Solliance cells with SiN and (b) Module 18 which includes Nantes cells with SiN

Molybdenum resistance measurements of the Mo samples included in the modules revealed a very rapid decrease of conductivity from the early stage of degradation, showed in Figure 26(a). Unlike the non-barrier protected cells the end of electric performance does not occur due to Mo degradation along the active are edge in this case. As it was observed in the cells with SiN barrier deposited on top and encapsulated, the Mo start degrading first around the conductive epoxy edge and, in this case, much earlier and at a higher rate.

(36)

Visual observation revealed that once the Mo is completely degraded underneath the conductive epoxy area the photocurrent is not able to be collected at the pads resulting in the end of electric performance of the cells. Figure 26(b) shows the mentioned effect on cells included at Module 17(2) at 480 hours of degradation.

Figure 26. (a) Resistance values of Mo samples protected by SiN barrier on top and non-encapsulated and (b) back side view of cells included in Module 17(2) at 480 hours of degradation

Since there are not encapsulants or foil the cells are exposed to 85 °C and 85 RH % from the beginning of the degradation process. This explains the earlier occurrence of the Mo degradation. According with theory studied in literature the EVA and PET reach the 85 % RH of the climate chamber after approximately 70 hours (Annigoni, et al., 2002).

Thus it would have been expected that performance of the cells protected by SiN barrier on top and encapsulated had been similar to the non-encapsulated cells with an approximated 70 hours lag in terms of degradation effects. It is unknown yet why the obtained values in this case do not match the expected results, since encapsulated cells still have an electric performance at the end of the study.

Modules with thin-film barriers on PET 4.6

As it was stated previously, there are two types of thin film barrier deposited on the PET foil under study: OCP+SiN and AlOx. Unlike the SiN deposited barrier on the cell surface, these barriers are placed in the package between the PET and the encapsulant.

This means that the moisture diffusion through the PET occurs at the same speed as the previous analysed foils and the low water-vapor transmission rate (WVTR) of the barrier protects the encapsulant from a fast diffusion of moisture. Thus, the percentage of relative humidity that the cells are exposed to is equal to the saturation degree of the encapsulant, as it happened in the non-barrier protected cells (see Equation 2), but in this case it is expected to increase at a much lower speed due to the low diffusion time provided by the barrier (Coyle, 2011).

Cells packaged with both types of barriers on PET showed a similar performance.

Efficiency values exhibited, in overall, that AlOx barrier protected cells had a slightly better performance than SiN barrier protected cells. Figure 27 shows the average efficiency decrement of the three cells of each module during the 1032 hours of degradation for both types of barrier protection. Solliance cells protected by SiN barrier deposited on top had a range of efficiency values varying from 0 to 6 % at the end of the experiment.

Damp Heat Degradation of CIGS Solar Modules

Master Level Thesis

European Solar Engineering School No. 222, June 2017