Textile-Integrated Conductive Layers for Flexible Semiconductor-Based Photovoltaic Structures

Abstract

This paper presents the results of research on conductive layers dedicated to flexible photovoltaic cells based on semiconductors integrated with a textile substrate. The presented work is part of a broader project aimed at producing flexible solar cells based on the CdTe semiconductor component and manufactured directly on textiles. The research focuses on the selection of textile substrates and contact materials, as well as the methods of their application. This study compares three types of fabrics (basalt, glass, and silicone fibers) and three metals (copper, molybdenum, and silver), evaluating their mechanical and electrical properties. During the experiments, flexible metallic layers with a thickness ranging from 160 to 415 nm were obtained. Preliminary experiments indicated that metallic layers deposited directly on textiles do not provide adequate conductivity, reaching the levels of several hundred Ω/sq and necessitating the introduction of intermediate layers, such as screen-printed graphite. The results show that molybdenum layers on basalt fabrics exhibit the lowest increase in resistance after dynamic bending tests. The obtained relative resistance changes in Mo layers varied from 50% to as low as 5% after a complete set of 200 bending cycles. This article also discusses current challenges and future research directions in the field of textile-integrated photovoltaics, emphasizing the importance of further technological development to improve the energy efficiency and durability of such solutions.

1. Introduction and Motivation: The Demand for Textile-Integrated Photovoltaics

The idea of photovoltaics based on a textile carrier seems attractive in numerous applications and thus was investigated in many aspects [1]. In several groups of popular and special applications like military, healthcare, life-saving services, and space exploration, it seems particularly advantageous over traditional electricity generation and is constantly being researched and developed [2]. However, the successful application of PV technology in textiles is a complicated and difficult process, which resulted in the limited range of commercially available products so far. For effective textile applications, several aspects of PV design and processing should be addressed.

Anybody interested in the practical applications of photovoltaics on flexible fabric substrates must remember that even though PV-generated energy is growing constantly, with a share of 68% in renewable generation in 2022 [3], it is still based on more than 90% of traditional rigid modules [4]. For fabric applications, these solutions are, unfortunately, inadequate. Regarding alternatives for rigid modules, also called the first generation by the M. Green classification [5], one may list inorganic flexible PV structures like amorphous silicon (a-Si), cadmium telluride CdS/CdTe solar cells, copper indium gallium selenide variants (CIGS), kesterite (CZTSSe) solar cells, antimuonium variants (Sb₂Se₃ and Sb₂S₃), and also copper-based structures like TiO₂/Cu; TiO₂/Cu₂O. All of them offer some extent of flexibility; however, each technology is not available at a high TRL level and is ready for mass production processes.

Apart from this group, organic devices are seen as naturally flexible structures suitable for bending. The most important representatives of this group are dye-sensitized solar cells (DSSCs), fully organic structures in the form of bulk heterojunction organic cells (BHJ), and perovskite-based modules (PSC). These types of cells are also often indicated as viable solutions for successive up-scaling processes [6]. All of them are functionally flexible, lightweight, and deposited on polymer substrates; however, their technological maturity is even lower than in the inorganic group, and stability issues have been constantly reported. Nevertheless, constant progress and intensive research efforts have been put into their further development, which boosts cell parameters every year [7].

Unfortunately, as the previous research of the authors indicated [8], the flexibility of the semiconductor structure is not sufficient for effective applications in elastic products. Durable, transparent contacts and encapsulation systems are equally important. Nevertheless, a growing number of successful implementations of flexible PV structures has given urgent prompting to the experiments aimed at textile applications.

For this kind of activity, it is necessary to consider the properties of the flexible cell substrate. Taking this into account, one may adopt polymer or metal foil material, which is embedded further on fabric, or direct structure deposition on the woven substrate. If we would like to consider transparent flexible polymers, we can distinguish the main groups including standard or high-temp PET foil (polyethylenterephthalat), poly(methyl methacrylate)—PMMA—, polydimethylsiloxane (PDMS) (e.g., Sylgard^®), or the Cyclo Olefin Polymer (COP) (e.g., ZEONEX^®). In the non-transparent group, one may distinguish polyester foils (e.g., MYLAR^®) and high-temp polyimide foils (e.g., KAPTON^® or UPILEX^®) [9]. All of them are flexible, lightweight, and mechanically durable; however, the processing temperatures vary from 100 °C in the case of standard PET up to 400 °C in the case of special polyamides. On the contrary, although metal foils are mechanically durable and thermally stable due to high weight, low flexibility, and corrosion problems, they are rarely considered as the PV substrates for textiles.

The current applications of PV-based textiles are numerous, from protective firefighter suits, military uniforms, medical equipment, and thermo-protective blankets through tent covers, yacht sails up to the packaging, RFIDs, and many more [10]. There is also a growing demand for power supply in the so-called smart textiles, which are currently equipped with a comprehensive set of biosensors, including cardiac monitoring, breathing status, skin temperature, or even sweat composition [11]. Normally, all those parameters are monitored by complicated stationary equipment, but for sports activity or rescue missions, all these devices may be incorporated into a regular garment and combined with a wireless connection to give invaluable information to the wearing person and their supervisors. Of course, all these systems demand a constant power supply, which, in many cases, may be delivered by wearable solar modules.

To fulfill all these demands, wearable photovoltaics must be fully compatible with the fabric according to the production and exploitation processes. Taking into account the most important factors, one may define the optimal production process for a PV module on fabric by a set of specific demands. These are the following:

• Low-temperature deposition and processing processes, or specific attachment post-production procedure since the most popular fabrics are not prepared to operate above the level of 100 °C;
• The possibility of easy and cheap upscaling from the lab-to-fab production level;
• Fast and smooth production procedure according to the standard roll-to-roll fabric treatment;
• True device flexibility, since any rigid add-on reduces comfort and limits the durability of PV operation.

All the mentioned demands are difficult to fulfill, especially taking into account the standard high-temperature processing of semiconductor layers and the deposition by the dedicated equipment in slow, energy-consuming, and clean procedures. Additionally, porous and unstable fabric surface poses a serious challenge to these process adaptations. In turn, during the exploitation course of PV on fabric, one may expect the following features:

• High durability according to the bending, stretching, and abrasion processes;
• Resistance to humidity in the normal exploitation and in the washing procedure;
• Resistance to low temperature and high insulation, which may pose a threat, especially to the polymer foil hermitization cover.

These parameters are normally tested according to the standards given by the textile industry, but for the PV modules, some special norms are currently under development.

The second critical issue, which should be addressed for successful device manufacturing, is the back contact, which is typically a “p” type base electrode. In this case, typically a thin metal layer is used, but conductive organic layers, carbon nanoform layers, and metal nanowires were also tested [12]. The specific material choice depends strongly on the solar cell type under consideration, but some general features should always be preserved:

• Low electrical resistance of the material;
• Good affinity with the base semiconductor material;
• High flexibility;
• Good adhesion to both—semiconductor and fabric;
• High resistance to degradation mechanisms (e.g., corrosion).

The above-listed demands generate several major challenges for scientists and engineers, especially taking into consideration a specific fabric construction and its behavior under mechanical stress. Additionally, there are often more sophisticated structures, manufactured as combined multi-layers, or 3D architecture materials proposed. In all these cases, obtaining the proper base contact, adjusted for a specific cell and a material, is a critical step for further device development.

Unfortunately, these problems have not been successfully covered yet. Thus, we decided to undertake this original technological investigation, focused exclusively on the manufacturing of proper flexible electrical contact films. In this work, specific experiments on selected metal contact layers, deposited by PVD (physical vapor deposition) on pre-treated fabric structures with screen-printed graphite layers, are described. For the first time, such a constructed contact system, dedicated to the CdS/CdTe solar cell structure on fabric, is explored. The technological and material experiments performed on a wide variety of textile substrates are supplemented with mechanical and electrical analysis and concluded with a discussion.

2. Current Development and Related Challenges of Wearable Photovoltaics

One of the most popular methods of integrating photovoltaic cells with a textile material is mounting them onto the surface of the fabric. This method has brought about many applications and gadgets. Such examples are backpacks with CIGS PV cells from two companies, Kingston’s Beam Backpack and Zenith Solar Backpack [13,14]. They both used solar cells with efficiencies of over 19% [14]. Another application idea that used this method is the winter outdoor jacket equipped with amorphous solar modules provided by Maier Sports. The maximum power generated by this jacket is claimed at the level of 2.5 W [15,16].

Significant progress in this field has been made by Sandia National Laboratories and their research conducted on non-organic microscale photovoltaic structures. Custom-made, small-scale photovoltaic cells feature a hexagonal shape, back contacts, and a thickness of about 14 μm, with widths ranging from 250 to 500 μm. With their optimized design, 14.9% energy conversion efficiency of such constructed solar cells has been achieved. However, despite the potential benefits for e-textiles, this innovative technology has not yet been utilized for this purpose [16,17].

In 2019, scientists from the Advanced Textiles Research Group at Nottingham Trent University demonstrated a revolutionary method of integrating small solar cells (dimensions: 1.5 mm × 3.0 mm × 0.2 mm) directly into textile yarn, which was then woven into fabric. The solar cells were mounted using flexible copper wires, allowing the creation of fabrics with standard textile properties. The fabrics also underwent washing tests in a standard household washing machine and retained approximately 90% of their original output power after 15 wash cycles. The photovoltaic fabric, with an active surface area of 44.5 mm × 45.5 mm, generated about 2.15 mW/cm², representing approximately 2.15% energy conversion efficiency, sufficient for charging a mobile phone and fitness monitor. Additionally, it was demonstrated that these fabrics are water-compatible [16,18].

In a 2019 study on textile-based organic photovoltaic (OPV) cells, a SiO₂/polymer composite layer was used as a protective encapsulation. The durability of these OPV devices was tested by immersing them in a detergent solution for 10 min after subjecting the device to 1000 bending cycles. These devices exhibited stability and achieved a power conversion efficiency of 7.26%, although they were not stretchable [16,19]. In 2016, Arumugam et al. implemented a fully spray-coated organic photovoltaic cell on a canvas fabric blended with cotton polyester by smoothing the fabric surface using a screen-printed interface layer. This design only achieved an efficiency of 0.02%, and resulted in the loss of device functionality after subjecting it to bending tests [16,20]. In 2017, Jinno et al. presented an OPV film that achieved a high conversion efficiency of 7.9%. This film was developed for wearable applications, being waterproof, washable, stretchable, and flexible. The washability of the OPV film was demonstrated by immersing it in a detergent and water mixture, which slightly reduced its efficiency. However, after prolonged immersion in distilled water, the device’s efficiency decreased by 46% [16,21].

Table 1. Comparison of energy conversion efficiencies of various types of photovoltaic cells.

Type of PV Cell	Year	Description	Efficiency (%)	Ref
CIGS (on backpack)	2018	PV cells mounted on the surface of a backpack	19.00	[14]
Non-organic microscale cells	2011	PV cells with a hexagonal shape, back contacts, and a thickness of about 14 μm, with widths ranging from 250 to 500 μm	14.90	[17]
Small solar cells (fabric)	2019	PV cells measuring 1.5 mm × 3.0 mm × 0.2 mm integrated directly into textile	2.15	[18]
OPV	2019	PV cells with a SiO2/polymer composite layer as protective encapsulation	7.26	[19]
OPV (sprayed on fabric)	2016	PV cells spray-coated onto a canvas fabric blended with cotton polyester	0.02	[20]
OPV (film)	2017	OPV film developed for wearable applications, being waterproof, washable, stretchable, and flexible	7.90	[21]

collects the data of the above-described cell types with the comparison of the obtained energy conversion efficiency.

The conclusions drawn from the above research indicate promising developments in the technology of flexible solar cells’ integration with textile fabric substrates. Despite the numerous technical challenges, significant progress has been made in this area. The studies pave the way for the utilization of photovoltaic textiles in various applications, ranging from smart clothing to sports accessories and electronic devices. However, further research and development efforts are necessary to improve the energy conversion efficiency and durability, and allow these innovative solutions for practical applications.

3. Flexible CdTe Solar Cells and Technologies for Fabric-Based PV Structures

There is a wide range of solar cell types that can be considered for manufacturing on textile substrates. Among the most popular technologies, there are dye-sensitized cells (DSSC), organic cells (OPV), and thin-film cells. They all have some potential for textile-integrated devices. However, due to the scope of the research presented in this article, we focus on cadmium telluride (CdTe)-based thin-film solar cells.

In terms of the photoconversion efficiency, CdTe-based solar cells achieve relatively high values. For example, in one study, a CdTe cell achieved an energy conversion efficiency of 22.1%, which is comparable to the efficiency of monocrystalline silicon, while maintaining cost advantage [22,23,24]. In

Table 2. Electrical parameters of typical CdTe solar cells with CdS and CdSe buffers [25].

Buffer Layer	Open-Circuit Voltage (V)	Short-Circuit Current Density (mA/cm2)	Fill Factor (-)	Conversion Efficiency (%)
CdS	0.847	24.7	0.70	14.6
CdSe	0.690	26.9	0.65	12.1

, the electrical parameters of CdTe solar cells with cadmium sulfide (CdS) and cadmium selenide (CdSe) buffer layers are presented.

Solar cells based on cadmium telluride are one of the most popular types of thin-film solar cells. They are commonly constructed as a “p-n” heterojunction, consisting of a p-doped absorber layer of CdTe and an n-doped window layer of CdS. The absorber material in CdS/CdTe solar cells is cadmium telluride, characterized by a direct bandgap of approximately 1.5 eV, resulting in its high absorption of sunlight [22,23,24]. Typically, thin-film CdS/CdTe deposition techniques include vapor-transport deposition (VTD) and close-spaced sublimation (CSS); however, chemical bath deposition (CBD) and physical vapor deposition (PVD) are also in use.

In addition to the traditional CdS/CdTe cells, nanowire solar cell (NSC) structures with graphite/Cu and ZnTe/Cu contacts have also been investigated. It has been shown that nanowire solar cells exhibit significantly higher reliability at high temperatures compared to their planar counterparts. Studies have demonstrated that NSC cells coated with graphite/Cu contacts improved their resistance to performance degradation by 2.75 times compared to the planar version of the cells. Nanowire structures combined with ZnTe contacts allowed for a decrease in performance degradation to only 4.5% after elevated temperature testing, which is 3.7 times more stable than in the case of planar CdS/CdTe cells [26].

As part of the research on CdS/CdTe structures, some experiments utilized the method of the gold-catalyzed growth of CdTe nanowires using metalorganic vapor phase epitaxy. Nanowires of CdTe have been grown at temperatures ranging from 485 to 515 °C on (111)B-GaAs substrates, allowing for the formation of vertical segments with a constant diameter. The results of these studies indicate the possibility of controlling the structure and properties of nanowires through an appropriate selection of growth conditions and catalytic materials. Low-temperature cathodoluminescence spectra recorded from individual nanowires showed well-resolved edge-band emission typical of CdTe with a dominant peak at 1.539 eV [27].

The absorber layers of CdTe usually grow on a high-quality transparent conducting oxide (TCO) layer, typically indium tin oxide (ITO), fluorine-doped tin oxide (FTO), aluminum-doped zinc oxide (AZO) [28], or amorphous indium zinc oxide (IZO) [29]. These layers can be deposited in the direct current magnetron sputtering at room temperature. The resistivity of ITO, FTO, AZO, or IZO films can be controlled over a wide range by varying the partial pressure of oxygen during the deposition process while maintaining a high optical transmittance of above 83% in the visible range.

The complete PV cell structure is enclosed with a metal-based back contact layer at the back of the cell. It is important to note that the active layers of CdTe are very thin, a few microns thick, which is about one-tenth of the diameter of a human hair. This makes CdTe cells not only efficient but also lightweight and flexible, making them ideal for applications requiring low mass and thickness.

The manufacturing of photovoltaic cells directly on fabrics is a process that must take into account the characteristic features of textile materials, such as folding, elasticity, and texture. However, to achieve satisfactory effectiveness, it is necessary to meet the requirements set for electrical conductors and semiconductor materials, such as porosity, flexibility, transparency, and stability. The coating process of fabric substrates with photovoltaic layers is extremely complex, and in some cases, requires high temperatures or specialized clean-room techniques.

There are two major strategies in the challenge of integrating photovoltaic cells with fabrics: producing photovoltaic fibers and creating fabrics from them, or adding photovoltaic layers to finished fabrics. In both cases, photovoltaic cells on fabrics require a conductive base that allows for the flow of electric charges. Although the surface texture is expedient to reduce light reflection losses and increase the light path through the cell, achieving a very thin and simultaneously continuous layer without its delamination is very difficult.

The main scope of the presented research is an experimental investigation of photovoltaic base contact layers deposited directly on textile substrates. All the parameters and settings applied to the experiments are compatible with the thin-film CdS/CdTe solar cell manufacturing technology. The choice of CdS/CdTe structure selection results from various factors, such as the semiconductor components’ stability over time, thin-film technology advantages, nearly perfect CdTe band gap value for photovoltaic energy conversion, and consequently, very high theoretical efficiency for CdTe-based PV structures. An additional aspect that prevailed in this case is long-term experience in the technology of CdTe-based photovoltaic devices of the research team from the Lodz University of Technology.

4. Materials and Methods of Obtaining Textile-Integrated Conductive Films

4.1. Substrate Fabrics Selection

The essential requirement to be met, in order to directly integrate thin-film CdS/CdTe semiconductor PV cell with a textile substrate, is the high-temperature resistance of the selected fabric [30]. Moreover, the structure of the substrate material must be suitable for the deposition of a continuous conductive layer constituting the base contact of the manufactured photovoltaic structure. Currently, there are a number of various types of fibers and textile materials meeting these conditions, and available on the market. The most-recognized materials are basalt or aramid fabrics (and their varieties such as Kevlar, Kermel, or Nomex); textile materials with the admixtures of glass, silicon, or calcium silicate fibers; and products based on ceramic fibers.

Table 3. Resistance of fabrics to continuous temperature load based on manufacturers’ data.

Fabric Type	Composition	Selected Manufacturers	Max Temp. (°C)	Ref
basalt fiber	thin, continuous fibers of 45–55% SiO2, 14% Al2O3, 10% CaO, 5–14% FeO, 5–12% MgO, 2–6% basic compounds (Na2O, K2O), and 0.5–2% TiO2	Mafic (San Jose, CA, USA), Basaltex (Wevelgem, Belgium), Technobasalt Invest (Kyiv, Ukraine), Smarter Building Systems LLC (Newport, RI, USA)	700	[31]
ceramic fiber	aluminosilicate fibers	3M (Saint Paul, MN, USA), Nippon Carbon (Tokyo, Japan), Zibo Jucos (Shandong, China), Norgpol (Ryki, Poland)	1260	[32]
silicon fiber	silica fibers containing approx. 98.9% SiO2	Ningguo BST (Anhui, China), Tespe (Bergamo, Italy), Texpack (Adro, Italy), Tectop (Jiangsu, China), McAllister Mills (Independence, VA, United States), AVS Industries (New Castle, DE, United States), Mid-Mountain Materials (Seattle, WA, USA), Norgpol (Ryki, Poland)	1000	[32]
calcium-silicate fiber	fibers of 60% SiO2, 23% CaO, 14% Al2O3, and 3% MgO, Fe2O3, and K2O	Zibo Jucos (Shandong, China), Norgpol (Ryki, Poland)	750	[32]
glass fiber	glass fibers containing calcium alumino-borosilicate and alkali, or magnesium aluminosilicate, or sodium-calcium-borosilicate	McAllister Mills (Independence, VA, USA), AVS Industries (New Castle, DE, USA), Texpack (Adro, Italy), JPS Composite Materials (Anderson, SC, USA), Tectop (Jiangsu, China), THS Industrial Textiles (Elland, UK), Montex (Maharashtra, India), Glass Fibre Europe (Brussels, Belgium), Norgpol (Ryki, Poland)	550	[32]
aramid/ kermel	aramid fibers produced as a result of the condensation reaction of paraphenylene diamine and terephthalyl chloride	DuPont (Wilmington, DE, USA), Beaver Manufacturing Company (Mansfield, GA, USA), Bally Ribbon Mills (Bally, PA, USA), Sinaora Advanced Materials (Hong Kong, China), Teijin Limited (Tokyo, Japan), BWF Protec (Hof-Gattendorf, Germany), Norgpol (Ryki, Poland)	350/400	[32]

lists the maximum operating temperatures of the selected types of fabrics provided by their manufacturers. Most of them can be made as dense, compact, and uniform structures, which is important to ensure the most possible uniformity of the subsequent layers of the photovoltaic cell.

The majority of the above-mentioned fabrics are nominally characterized by sufficiently high permissible temperature values according to the manufacturing technology requirements for flexible semiconductor-based PV structures. These types of fabrics are the subject of ongoing research conducted in order to evaluate their application in flexible textronic elements [33,34].

The above-described textile materials have been tested experimentally in order to verify their thermal properties. This study aimed at evaluating their actual temperature resistance over time allowing for the recrystallization processes of semiconductor photovoltaic structures (treatment of 60 min in 400 °C). The results do not include fabric made of calcium-silicate fibers due to the excessive porosity of its structure and low weave density, which makes it impossible to deposit semiconductor layers. Figure 1 shows the percentage of mass loss at elevated temperatures and extended time, as required by the recrystallization process of thin-film semiconductor-based photovoltaic structures.

The results presented in Figure 1 show, in some cases, significant discrepancies with the parameters provided by the manufacturers. This is especially visible in the case of ceramic materials. The values of maximum temperatures obtained for the materials based on aramid fibers and their derivatives can be considered consistent with the literature and catalog data. On the other hand, Thermo Glass (glass fiber) fabric showed surprisingly favorable thermal properties. In addition to the glass fabric, the best temperature resistance parameters are demonstrated by the textiles made of basalt and silicon fibers (Silicatherm), which withstand continuous temperature loads of 600 °C without any significant color changes or elasticity losses. For this reason, the three above-mentioned textile materials, made of basalt, glass, and silicon fibers, have been selected for further research conducted within this investigation. The structures of these fabrics are shown in Figure 2.

4.2. Investigated Contact Materials and Implemented Deposition Methods

Three selected types of fabrics have been studied in order to verify their long-term high-temperature resistance. This parameter is absolutely essential when considering the manufacturing technology of thin-film CdS/CdTe photovoltaic structures.

Conductive back contact materials have been selected according to the preliminary studies and experiments with the production technology of flexible CdS/CdTe. Three metals: copper, molybdenum, and silver have been investigated in further experiments, as they can be compatible with CdTe photovoltaic base in their target application, which was indicated in the previous research of the authors [30]. All the examined metal layers have been deposited on textile substrates using physical vapor deposition (PVD) technology. Because of the 3D structure of the textile materials, which causes the heterogeneity of their surface, all of the investigated samples had been pre-treated by the deposition of graphite layers that improved their uniformity. This treatment allowed us to achieve more continuous and homogeneous structures of electrical contact layers.

The printing paste for graphite layer deposition has been produced by homogenizing the functional phase material in the form of graphite flakes in a PMMA-based polymer matrix. The authors used Sonigraf graphite flakes MG 1596 with an average diameter of 10 μm, characterized by high electrical conductivity and a low degree of agglomeration [35]. Graphite platelets with PMMA matrix were rolled twice on a three-roll mill with a silicon carbide (SiC) roller. A highly flexible polymer matrix based on PMMA 350,000 MW has been developed by the authors and presented in previous publications [36]. The identical carbon composites have been screen-printed on all of the investigated textile substrates. During the printing process, a squeegee (75HS) forces the composite through a 77T nylon screen, reflecting the unmasked pattern. After the deposition process, the printed layers have been dried by evaporating the solvent in chamber dryers for 15 min at a temperature of 130 °C.

The screen-printed graphite layers are characterized by relatively high surface porosity. However, the advantage of this treatment is tightly filling the gaps between the fibers by graphite paste and creating applicable and continuous background structure for thin films of metal contacts. Microscopic images of the studied fabrics with the deposited layers of graphite are presented in Section 5 (Figure 3).

The following stage designed for this research involved thin-film metal layer deposition using physical vapor-assisted technologies. Pure particles of copper, molybdenum, and silver have been separately evaporated in a controlled laboratory environment in order to provide conductive films on fabrics with previously printed graphite background layers. The thicknesses of the evaporated layers have been measured with a DEKTAK profilometer on a silicon wafer control sample. These control samples intended for thickness measurements have been additionally included in each of the deposition processes.

The copper (Cu) contacts have been deposited in the PVD process through the target particle evaporation from the resistive source. Copper of a 99.99% laboratory purity has been used as a target material. The material has been evaporated in a vacuum chamber at an internal air pressure below 4 × 10⁻⁵ mbar for 12 min, achieving films of 415 nm thickness.

The molybdenum (Mo) layers have been obtained in another PVD technique using vacuum deposition with an electron gun. A molybdenum target with a laboratory purity of 99.95% has been evaporated in a vacuum chamber under the pressure of 5 × 10⁻⁶ mbar. A vapor deposition procedure of 18 min allowed for achieving layers of 160 nm thicknesses.

The silver (Ag) layers, analogously to the molybdenum films, have been produced in an electron gun-assisted vacuum deposition technique. Silver targets of the laboratory purity of 99.99% have been evaporated in a dedicated vacuum chamber under the pressure of 4.5 × 10⁻⁶ mbar for 10 min. These parameters provided continuous contact layers of 330 nm thickness.

For the above-described metal layers deposition, the BALZERS BAK 550 evaporator was used with the following technical parameters: chamber pressure up to 5 × 10⁻⁷ mbar, 3.8 kW thermal evaporation source, and 10 kV e-beam power supply, ion gun source included. In order to limit the area of metallization deposition, a mechanical mask cut in a polyimide Kapton foil was used.

Macroscopic photos of the prepared samples and their surface structure microscopic images are presented in Section 5.1 (Figure 4). The elaborated samples of the approximate dimensions of 4 cm × 8 cm have been subsequently subjected to the electro-mechanical durability tests. The assessment of their electrical properties, mechanical resistance, and electrical stability in dynamically changing conditions was the purpose of this examination.

Therefore, the samples have been tested in terms of their electrical properties, as well as mechanical strength and electrical stability under the influence of dynamic stress. They have been subjected to mechanical resistance examinations using dedicated l equipment allowing electrical measurements during a process of dynamic bending. The samples have been bent on 25 mm diameter cylinders for 200 cycles (one cycle means one bending and one straightening of the sample). The bending frequency was equal to 3 bandings per second. The electrical resistance values have been measured using the four-point probe testing station. The resistance changes have been read and automatically recorded during the dynamic bending process. In addition, before and after a series of mechanical bending cycles, the surface structures of the investigated layers have been visually inspected using an optical microscope. The results are described in Section 5.1.

5. Measurement Methodologies and Extracted Results

A very important factor in the technological processes of thin layer deposition is the ability to define and measure their parameters. Quality improvement of the obtained structures with simultaneous efficiency increase in the designed devices is the essential economic category in the contemporary techniques applied in flexible electronics.

5.1. Optical Imaging Assessment

The optical analysis has been performed using a standard stereoscopic laboratory microscope (DeltaOptical, Nowe Osiny, Poland) with the catalog magnification range from 8 to 400 times. Figure 3 shows the optical images of the three investigated fabrics with the graphite layers applied. The microscopic images illustrate how the graphite layer fills the spaces between the fibers and gives the structure continuity.

Figure 3. Photos of the samples illustrating the selected fabrics with screen-printed graphite layers (left), and their microscopic images (magnification: 40×) showing the border between the printed layer and textile substrate (right).

Figure 3. Photos of the samples illustrating the selected fabrics with screen-printed graphite layers (left), and their microscopic images (magnification: 40×) showing the border between the printed layer and textile substrate (right).

Figure 4 demonstrates, in both classical photography and the microscopical approach, the superficiality of particular fabric samples with the deposited contact layers of the selected metals. These conductive coatings are dedicated to perform as base electric contacts in textile-integrated photovoltaic structures.

Analyzing the basalt fabric with a copper contact, a uniform Cu vapor coverage of the fibers can be observed on the right side of the border. On the left side of the sample, there is a significant surface roughness caused by the previously applied graphite layer. Although the graphite layer appears to be uniform and evenly distributed in the macroscopic image, it can be noticed that it is very discontinuous and does not completely fill the gaps between the fabric fibers. Since copper effectively reflects light, the structure reveals cavities in the graphite layer. In the microscopic image showing the boundary of the graphite and copper layers on the Silicatherm fabric, the observed coverage of the graphite layer is more intensive and the gaps between the weft and the warp of the fabric are accurately filled. The metallic copper layer deposited on the fabric, without a previously printed graphite background layer, only covers the individual warp and weft fibers without creating a continuous layer. A similar situation is in the case of the Thermo Glass fabric, where the graphite layer creates a more continuous structure than the copper layer alone.

In the images illustrating the samples with molybdenum coatings, the boundaries between the area of the deposited metal and graphite layers are very thin. Graphite, in a granular form, fills the spaces between the weft and the warp of the fabric, while molybdenum covers the individual warp and weft fibers only on the surface. There are detectable areas not coated by contact in the spaces of interlacing, which constitutes a discontinuity of the layer. The image of the Mo layer on the Silicatherm fabric shows better coverage of molybdenum and graphite, both on the weft and the warp. However, an even layer of graphite, which would allow for the precise connection of the individual elements of the fabric structure, was not created. Conversely, the Thermo Glass fabric indicates decent graphite coverage, which provides a uniform structure, and good coverage of the fibers with molybdenum atoms.

The images of the silver layers present a fairly uniform coverage of the basalt fibers, as well as a rather loose coating of the fabric, which creates a rough structure. The graphite layer on the Silicatherm fabric sample is only superficial and shallow. It does not fill the gaps between the weft and the warp, but only covers the individual fibers separately on the surface. The graphite layer observed in the image has low roughness and is relatively dense. The silver layer covers the warp and weft surfaces accordingly, as well as the graphite layer. The silver atoms evenly cover the surface of the fabric fibers and the graphite layer on the Thermo Glass fabric. Layers of graphite form even coatings on the fabric surface, simultaneously filling the warp and weft interlaces.

Figure 4. Photos of selected fabric samples with graphite and metal layers (copper, molybdenum, and silver), as well as their microscopic images (magnification: 40×) with visible borders of graphite layers printed on textile substrates.

Figure 4. Photos of selected fabric samples with graphite and metal layers (copper, molybdenum, and silver), as well as their microscopic images (magnification: 40×) with visible borders of graphite layers printed on textile substrates.

5.2. Electrical Resistance Measurements

The four-point method is currently the most common way to measure the resistance of conductors and semiconductors. Its advantage is that the electrode/semiconductor contacts are not required to be ohmic contacts (metal/semiconductor contacts that do not have rectifying properties and do not have a barrier layer). The measurement probes in this method are either in linear or square arrangements.

The method using a linear arrangement of probes involves four metal needles spaced with equal distance from each other. Another typical technique to measure the resistivity of the semiconductor material is the method with a square arrangement of electrodes. In this case, the current is delivered through a pair of electrodes on one side of the square and the voltage is measured on the other pair of electrodes. Arranging the probes at the vertices of the square allows one to reduce the measurement error.

We used the four-point probe in the linear arrangement for the electrical resistance measurements of the investigated layers deposited on textile substrates. In order to achieve reliable results, several repetitions of the measurements have been performed by delivering current in combination (in both directions) through each pair of the four-point probe, and the average resistivity values are calculated and reported (

Table 4. The electrical resistance measured for the Cu, Mo, and Ag contact layers on the selected fabrics.

Electrical Resistance (Average Values) Rsq of the Contact Layers (Ω/sq); Accuracy ± 0.02%
Contact Layer	Basalt Fabric	Silicatherm	Thermo Glass
Cu	0.3	0.54	0.3
Mo	64.5	815.0	92.2
Ag	0.2	11.4	1.5

). The situation of the uniform distribution of the electric field near the electrodes is quite rare.

5.3. Controlled Dynamic Bending for Mechanical Endurance and Electrical Stability Tests

The dynamic bending method enables the testing of the mechanical resistance of layers according to the A-De Mattia method, compatible with the PN-EN ISO 7854 standard that regulates the determination of resistance to bending damage. The tested sample is placed on a movable roller and attached to a motor-driven wheel, which causes the sample to bend and straighten. The sample is mounted using metal plates which are connected to the electrodes of the multimeter connections. The multimeter records sequentially the resistance values during the bending process at the selected frequency.

The tested samples have been subjected to dynamic bending tests using the dedicated equipment described above. In this experiment, a cylinder of 25 mm diameter has been used in all the bending tests equally for each treated sample. The resistance testing procedure included 200 bending cycles for each sample (one bending cycle includes the bending and extension of the sample). The frequency of bending was set to 3 bends per second. The electrical resistance values have been recorded automatically using a laboratory Rigol DM3062 digital multimeter. The obtained results are presented in a graphical form showing the percentage change in the electrical resistance values during the process of dynamic bending. The measured values (Rm), recorded directly during the tests, are given in relation to the initial value (Ro) of the sample measured before the bending process. The changes in resistance have been calculated using the following relation:

x = (Rm − Ro)/Ro × 100%

where Ro—the initial value of the electrical resistance for the sample before bending, Rm—the measured value of the electrical resistance for during the bending process.

The electro-mechanical stability levels of the prepared samples of the metal layers deposited on textiles have been tested using the above-described methodology. Figure 5, Figure 6 and Figure 7 in Section 5.4 show the results of these experiments.

5.4. Numerical and Graphical Comparison of the Measured Results

After preparing the backgrounds and evaporating the conductive layers, surface resistance [Ω/sq] values have been measured for each sample using a RIGOL DM3058 multimeter and a four-point probe in a linear configuration. The results are summarized in Table 4, for copper, molybdenum, and silver, respectively.

The thicknesses of the layers were, respectively: 330 nm for Ag, 415 nm for Cu, and 160 nm for Mo (measurement method described in Section 5.2).

The initial values of the electrical resistance per square were different for each layer on each material. For silver, it was measured as 1.5, 11.4, and 0.2 Ω/sq, respectively, for Thermo Glass, Silicatherm, and basalt. The values for copper were 0.3, 0.5, and 0.3 Ω/sq according to the same order, and for molybdenum, respectively, 92.2, 815.0, and 64.5 Ω/sq.

Figure 5, Figure 6 and Figure 7 show the percentage change in the electrical resistance within the dynamic bending process for the three types of contacts mounted on the three types of textiles—basalt, glass (Thermo Glass), and silicon (Silicatherm) fabrics.

Figure 5. Electrical resistance changes during dynamic bending of copper layers deposited on three types of textiles (Thermo Glass, Silicatherm, and basalt).

Figure 5. Electrical resistance changes during dynamic bending of copper layers deposited on three types of textiles (Thermo Glass, Silicatherm, and basalt).

Figure 6. Electrical resistance changes during dynamic bending of molybdenum layers deposited on three types of textiles (Thermo Glass, Silicatherm, and basalt).

Figure 6. Electrical resistance changes during dynamic bending of molybdenum layers deposited on three types of textiles (Thermo Glass, Silicatherm, and basalt).

Figure 7. Electrical resistance changes during dynamic bending of silver layers deposited on three types of textiles (Thermo Glass, Silicatherm, and basalt).

Figure 7. Electrical resistance changes during dynamic bending of silver layers deposited on three types of textiles (Thermo Glass, Silicatherm, and basalt).

In the dynamic bending testing process, the resistance of the copper contact deposited on the basalt fabric increases during subsequent bends, occasionally even reaching values over 2500% higher than the initial measurement (Figure 5). Important changes begin after approx. 40 bends. Comparing these results to the corresponding microscopic image, we concluded that the small amount of graphite molecules initially filling the fabric structure eventually started to crumble, resulting in an insufficient continuity of the conductive layer.

In the case of the copper layer deposited on the silicon fabric (Silicatherm), the resistance also increases during subsequent bends, reaching a value 500% greater than the initial measurement. However, this increase is not as variable as for the previous sample (on the basalt fabric). This is most likely caused by the more precise and dense graphite coating possible for this fabric structure, resulting in a more durable continuity of the conductive layer. The graphite layer may also crumble and as the number of bends increases, the resistance values are systematically higher. In this case, however, the increase in resistance occurs already during the first bends and its highest values occur in the first phase of the tests. In the following stages of the process, the resistance values are subject to more stochastic laws of layer conduction while bending. As the greatest destruction of the graphite and copper layers occurs in the first phase of the test, this means that such structure is reasonably delicate and sensitive to mechanical strains.

For the glass fabric (Thermo Glass) coated with copper, the resistance increases gradually as the number of bends increases. Initial bends have a lower impact on the resistance increase than in the case of Silicaterm, although higher than for the basalt textile. Comparing the results of the dynamic bending tests with the microscopic image, we observe that the structure of the Thermo Glass fabric, which has a much denser weave than Silicatherm, may result in better bonding of the graphite layer between the weft and the warp, resulting in a smooth increase in resistance. Additionally, this fabric appears to be covered with a more uniform layer of graphite, which could translate into a significantly longer time required for the structure to crumble.

The results of the electro-mechanical durability measurements conducted for the molybdenum layers are presented in Figure 6. The resistance of the Mo sample deposited on the basalt fabric, in the initial phases of the test, changes only slightly, and even decreases, which is more of a stochastic effect. Nevertheless, the percentage increase in the resistance value after the last bending cycle is less than 100%, which indicates objectively high adhesion of the conductive layer to the fabric.

The samples of Mo deposited on the Silicatherm fabric illustrate that the thickness of the fabric weave can have an influence on the change in layer resistance during bending. We conclude that this non-isometric structure of the weave along and across the warp affects the resistance changes throughout the bending process. Remarkably, the resistance values in this case oscillate between +50% and −50% of the initial measurement, which means that its average value is constant. Thus, in this case, the electro-mechanical test confirmed good adhesion of the Mo layer to the textile fibers and the insignificance of the graphite layer crumbling impact on the overall structure conductivity.

The effect of the bending process on the electrical properties of Mo on the Thermo Glass fabric sample was even slighter. The increase in the resistance value during the entire test did not exceed 20% of the initial value. Moreover, perhaps resulting from some random arrangements, the resistance value even decreased in several stages of the measurement. Summarizing, the overall tendency of resistance increase is extremely minor for the tested molybdenum layers deposited on a glass fabric. This impressive result of the effective and stable conductivity is a consequence of the contributing elements. The important factors include the dense weave of the fabric, even and precise surface coverage with the graphite layer, and even adhesion of the molybdenum film.

The analogous measurement results of the silver contacts are visualized in Figure 7. The measured resistance of the Ag layer deposited on the basalt fabric systematically increased following the subsequent number of bending cycles. Relatively significant changes, reaching an increase of 200%, are observed in the first phase of the test after subjecting to several cycles. At the final stage, the resistance value of this sample reached 1800% of its initial value. This can be a consequence of the discontinuity of the layer and the low adhesion of silver to the basalt fibers.

The tests conducted for silver on the silicon fiber (Silicatherm) sample demonstrate relatively stable resistance values, with some deviations in high peaks followed by equivalent downhills. These impulse effects were probably caused by the anisotropy of the fabric structure. The peaks of highly increased resistance occurred randomly, starting at 400% and eventually reaching 1,500% of the resistance increase. Associating the resistance results with the macroscopic image of the sample, the explanation can be found in the considerable discontinuity of the structure, which creates significant breaks between the graphite layer and silver molecules. We estimate that the issue of increased distance between the Ag particles is reflected in these huge resistance jumps that occur spontaneously when the sample is in the maximum bending position. Then, during the straightening of the sample within the same cycle, the distance between the molecules is reduced again, and the structure undergoes reformatting resulting the resistance to decrease.

The results of the bending experiments conducted for the samples of Ag deposited on the glass fabric (Thermo Glass) show that the resistance changes smoothly as the number of bends increases. From the overall perspective, this sample can be assessed as relatively suitable in the context of this investigation. The adhesion of the conductive layer to the surface of the glass fabric is sufficient, resulting in an acceptable level of the resistance value increase (200% as a maximum in the final stage).

The synthesis of the presented outcomes demonstrates the high potential of the several proposed materials and technologies. The samples representing the molybdenum layers deposited on glass-based textiles (Thermo Glass) proved to be principally attractive for the development of textile-integrated photovoltaics. The glass textile indicated the most homogeneous surface structure which contributed to a stronger stability of all the layers deposited on these substrates. The molybdenum contacts revealed a dominant electrical stability considering all the investigated textiles; however, its initial conductivity is lower than the other materials.

6. Discussion, Conclusions, and Anticipated Research Progress

The presented research is aimed to provide an experimental and analytical verification of both material performance and technological possibilities in the advanced engineering of textile-integrated and semiconductor-based photovoltaics. To narrow the investigated area, our group focused on common photovoltaic structures designed and constructed in thin-film technology. In this study, the parameters adopted in the experiments followed the initially elaborated technology description guide of flexible CdS/CdTe solar cell manufacturing [34]. This paper synthesizes the comprehensive studies, undertaken as a first stage in this extensive research challenge. The investigations have been focused on the textile substrate selection with experiment-based suitability, as well as directed to the adjusted optimization of the technology and material engineering aiming on back contacts’ formation for a dedicated PV structure. The research comprised customized studies and meticulously designed experiments. The subsequent stages have involved several challenges including the deposition of the appropriately selected conductive layers on the dedicated textile substrates (of the required properties), the consequent measurement procedures supported by data processing, and multidimensional verifying analyses. After the preliminary studies, three fabric types (basalt, glass, and silicon fibers) and three elementary metals (copper, molybdenum, and silver) have been selected for this part of the investigation.

In the initial phase of the research, we have learned that thin metallic layers directly deposited on textile substrates, are insufficient to obtain the desired conductivity expected for the designed application. The reason lies in the discontinuity of the metal films caused by the complex structure of the textile substrate. Thus, an intermediate layer occurred necessary in order to provide textile-integrated metal structure with properties suitable for electrical contact. For this purpose, the screen-printed graphite layers were introduced as background structures between fabrics and metal films.

Primarily, the basalt fiber-based material has been subjected to be analyzed as a potential fabric substrate for the previously specified textile-integrated PV technology. The graphite layers deposited on the basalt fabrics are identically distributed on each investigated sample. Their surfaces are granular, porous, and discontinuous, and they fill the gaps between the warp and weft to a minor extent. The results of the electrical measurements showed that among the investigated contact materials, the silver layers are characterized by the lowest surface resistance. The analysis of the dynamic bending tests and simultaneous measurements of their impact on the electrical properties of the samples resulted in the following points of conclusion: The lowest overall resistance change, after the full procedure is completed, is reported for the molybdenum layer. The copper layer performance in this category indicates the highest value. Compared to molybdenum, the final increase in the copper layer resistance was several orders of magnitude higher, exceeding 20,000%, while for molybdenum, the change of only 100% was reported. Silver contact resistance on the basalt fabric increased in this test by over 15,000%. Therefore, with respect to the numbers, molybdenum contact was chosen as a suitable for the basalt fabric substrate. Although the initial resistance of the molybdenum contact is higher, compared to the other layers, its small changes as a result of repeated bending and straightening is the reason for choosing this contact as the most appropriate to be used on the basalt fabric.

Analyzing the parameters of the samples deposited on the silicon fabric (Silicatherm), only the optical observations indicate slight differences in the applied graphite layers. It best covers the weft and warp, as well as individual fibers; however, the twisted parts are left unevenly uncoated. This distribution, together with the differences in thicknesses for particular metal coatings, may explain the substantial varieties in the initial resistance values of all the tested materials. In the final stage of the bending durability test (after 200 bending cycles), the highest changes in the electrical resistance, reaching a total increase of 1,500%, are reported for the silver layer. On the contrary, the sample with the molybdenum contact deposited on this fabric shows only 50% of a maximum resistance change. The final resistance increase for the copper-coated samples is approximately 500% of its initial value. Considering the results collected in this section, molybdenum material occurs as the most suitable contact to be used with silicon fiber-based substrate. Although the initial conductivity of this layer is lower compared to other materials investigated in this study, its electro-mechanical stability in continuous repeated bending cycles is significant.

The last group to be analyzed are the samples deposited on the glass fabric (Thermo Glass) substrates. In this case, all the graphite layers fully fill the spaces between the warp and weft, densely and evenly covering the fiber surfaces. The results of the dynamic bending tests demonstrated an analogous relation as shown for the silicon fabric substrates. In the final stage of the mechanical test, after completing 200 bending cycles, the highest increase in the electrical resistance in relation to its initial value is observed for the silver-coated sample, and it reaches the maximum level over a 700% increase. Slightly less increase of 600% is reported for the sample with the copper layer, and only 34% for the sample with molybdenum. Similarly to the previously analyzed basalt and silicon fabrics, the molybdenum contacts were revealed to be the most suitable also for the glass fiber-based substrates. Its relatively high initial resistance is compensated by a minor increase in the process of bending and straightening over many cycles, especially when compared to the other conductive materials, such as copper and silver.

The electrical properties of all the investigated samples, calculated and shown as the trend lines, are collected in Figure 8.

The conclusions to be drawn from the presented results are expected to provide the codependent verification and critical selection of textile substrates and contact materials. For the proposed technology of textile-integrated photovoltaic cells, the Thermo Glass fabric is characterized by the most suitable properties that are needed to be met by the substrate material. All the samples deposited on this fabric showed the lowest changes in their electrical resistance during dynamic bending tests, which indicates that its structure is optimal for combining with materials such as graphite, silver, copper, and molybdenum. Its weave is dense with proportional warp and weft. This is a fabric characterized by the high adhesion to the applied layers both by screen printing and vapor deposition. Moreover, it is exceptionally stable during the high-temperature processes that are used in the production of thin-film Cds/CdTe photovoltaic cells. Considering the base contact material selection, the analysis indicates molybdenum as the most appropriate in this study. Its properties as a flexible contact layer significantly exceed those of the other investigated materials. Its electrical resistance changes only slightly during intensive dynamic bending tests, which can be properly controlled. Molybdenum is a relatively less expensive element compared to silver or gold, which makes it a good technological option.

As discussed in the paper, several factors influence the creation of an optimal flexible contact. The appropriate substrate material that can withstand technological processes under high temperatures is the initial challenge. Another issue is the proper layering technology with the background supportive layer as a key stabilizer for the subsequent layers. Graphite is a great candidate for this role as it is a low-cost material and can be successfully used widely in the production of flexible layers. Screen-printed graphite fills the spaces inside the fabric, i.e., between the weft and the warp, thus improving the uniformity of the conductive layer deposited afterwards. Finally, the use of a material characterized by stable electrical resistance in long-term mechanical loading such as bending and straightening is necessary.

To sum up the results of the presented investigation, several important conclusions may be drawn. Firstly, the proper choice of substrate fabric must suit the specific application and is critical for further contact construction. Subsequently, in some untypical applications such as highly porous fabrics, it is impossible to achieve an effective contact layer by a direct metal deposition. Some special sub-layers should be then applied for the proper material preparation. It ameliorates the adhesion and enables effective electrode functioning. Finally, the electrical and mechanical parameters’ optimization is typically a tradeoff, which should be considered according to a specific solar cell construction. One should remember that a significant increase in the series resistance, introduced by the lack of proper ohmic contacts, significantly reduces the fill factor (FF) of the obtained solar cell. However, on the other hand, flexible thin-film photovoltaic structures are not typically as demanding as highly optimized silicon devices. Nevertheless, the manufacturing of the proper base composition is a necessary step for the successful implementation of the photovoltaic structure on a textile substrate.

This work is a part of a broader project dedicated to obtaining functional and fully textile-integrated PV structures based on CdTe semiconductor components. The complementary tests of the metal layers deposited on the selected fabrics represent the first stage of the textile-integrated, thin-film flexible CdS/CdTe photovoltaic cells’ development. This study does not exhaust, but only opens up research on the exploration of material textile-integrated flexible photovoltaic cells. Subsequent experiments are systematically and continuously conducted by our research team.

There is also a significant added value of this article taking into consideration the extensive research already carried out in the present field of flexible textile-integrated photovoltaics. It is worth highlighting that these studies follow a particularly up-to-date research path, which is strongly future-oriented with the potential for multidisciplinary implications.