An investigation of a wash‐durable solar energy harvesting textile

This work demonstrates a novel and sustainable energy solution in the form of a photovoltaic fabric that can deliver a reliable energy source for wearable and mobile devices. The solar fabric was woven using electronic yarns created by embedding miniature crystalline silicon solar cells connected with fine copper wires within the fibres of a textile yarn. This approach of integrating solar energy harvesting capability within the heart of the textile fabric allows it to retain the flexibility, three‐dimensional deformability, and moisture and heat transfer characteristics of the fabric. In this investigation, both the design and performance of the solar cell embedded yarns and solar energy harvesting fabrics were explored. These yarns and resultant fabrics were characterised under different light intensities and at different angles of incident light, a critical factor for a wearable device. The solar cell embedded yarns woven into fabrics can undergo domestic laundering and maintained ~90% of their original power output after 15 machine wash cycles. The solar fabric embedded with 200 solar cells demonstrated here (44.5 mm × 45.5 mm active area) was capable of continuously generating ~2.15 mW/cm2 under one sun illumination and was capable of powering a basic mobile phone. The power generation capability and durability of the solar energy harvesting fabric proved its viability to power wearable devices as an integral part of regular clothing.

The emergence of wearable devices, for applications including non-invasive health care monitoring, 1 sports, 2 learning assistance, 3 and entertainment, 4 has been catalysed by the miniaturisation of electronic components and low power-consuming devices. Despite the unprecedented interest and potential foreseen, the Achilles heel of many wearable and electronic textile (E-textile) devices are their energy requirement, which is a major hurdle to the wider adoption of E-textiles. 5,6 Most commercially available wearable systems are powered by standard solid coin cells, pouch cell, cylindrical cell, or prismatic cell batteries of alkaline, nickel metal hydride (NiMH), lithium-ion (Li-ion), or lithium-ion polymer (LiPo) type. 7 These batteries are typically attached to the garment after assembly or embedded in a removable module, making the systems bulky and cumbersome to use. 8 Smaller and lighter batteries require frequent recharging. In light of these drawbacks, many have proposed integrating energy harvesting capability into clothing, 9,10 to power wearable devices using ambient energy available from the surrounding.
Amongst the other energy harvesting technologies explored for Etextiles, such as tribo-electric, 11,12 piezo-electric, 13,14 thermoelectric, 15 or electromagnetic induction, 16 solar energy harvesting has been one of the most investigated avenues due to the abundance of solar energy 17 and the maturity of PV technologies. 18 Various approaches to integrate solar energy harvesting capability into textiles have seen a rapid growth during the last two decades [19][20][21] : The first attempts were to superficially attach rigid 22 or flexible 23 solar panels onto fabrics; these were limited to functional clothing and futuristic fashion prototypes and were far from the appearance, feel, and durability of regular clothing. Printing, 24 laminating, 25 or coating 26 of organic PVs (OPVs), 27 and hybrid PV such as dye-sensitised solar cells (DSSC) 28 and perovskite SCs 29 onto textile substrates have dominated most of the textile-based PV research. Thin film PV laminates such as copper indium gallium diselenide (CIGS) 30 and amorphous thin-film silicon (a-Si, TF-Si) 31 have also been extensively explored for textile applications. In addition to the inherent flexibility, a preference towards OPVs, hybrid PVs, and thin film PVs have emerged due to their affordability and recent improvements in efficiency; however, it is important to note that these solutions are still incapable of matching the efficiencies of inorganic cells 32 . Despite being inherently flexible, the monolithic, large-area, nonporous structure of these PV systems significantly changes the appearance and feel of the textile and restricts the shear behaviour and air permeability of the textile substrate, making them uncomfortable and less appealing to the wearer.
An alternative solution is to weave PV-coated wires 33,34 or flexible PV tapes 27,35 of DSSC type, 36,37 perovskite type 38,39 or OPV 40,41 into fabrics, which has also been widely explored for wearable applications.
In principle, the PV layers of these wires or tapes were similar to the coated, laminated or printed films discussed earlier, with the small cross sections and large aspect ratios of these wires or tapes allowing them to be woven into fabrics. Fabrics woven with PV-coated wires or flexible PV tapes showed improvements in the shear behaviour and breathability; however, they still looked and felt significantly different from normal textile fabrics.
In general, the majority of PV-textile systems have not reported the compatibility with water or washing. Two recent studies on elastomer-coated organic PV 42 and textile-based polymer SCs 43 show some evidence of durability to detergent-water mixtures. In these studies, free-standing organic PV cells enhanced with liquid barrier laminates were tested inside a small beaker of stirred water-detergent mixture for 10 to 30minute cycles. However, these mild test conditions are far from the rigorous hydro-mechanical agitation undergone by regular clothing in a domestic washing machine.
The solar energy harvesting fabric demonstrated here was constructed by weaving textile yarns embedded with miniature SCs (solar-E-yarns). To achieve a drapable and soft fabric that can endure machine washing, the shear behaviour and a low bending rigidity of the structure had to be maintained. Therefore, the rigid PV elements were deployed in a discontinuous fashion within the fabric in yarn form by employing electronic yarn technology. 44 Electronic yarns have previously seen small electronic chips integrated into the core of a textile yarn to add electronic functionality including illumination, 45 temperature sensing, 46 vibration sensing, 47 light sensing, 48 and acoustic sensing. 49 The E-yarns with embedded SC (solar-E-yarns) were realised in three steps. First, miniature (1.5 mm × 3.0 mm × 0.2 mm) SCs were soldered in parallel onto two multistrand copper wires before being individually encapsulated within clear, cylindrical resin micropods.
The solar-cell-micropod filament containing the encapsulated cells was then covered by a fibrous textile sheath to give the final solar-E-yarn a textile feel and appearance. The discrete micropods and the fibrous sheath provided the solar-E-yarns with a low bending rigidity and a high degree of porosity enabling the transfer of moisture and heat through the E-yarns, and resultant fabrics. This feature was crucial to prevent discomfort to the user caused by thermal and sweat build-up, especially during warm and sunny conditions. 50 The textile sheath also allowed the yarn to take any colour without significantly altering its optoelectronic performance. This method is applicable for embedding different types of SCs (including organic, 51  The integration of SCs within the E-yarn structure (first the micropod and then the fibrous sheath) will have an effect on the embedded SC performance, 48 and therefore, this was important to analyse and understand. Voltage and current outputs have been presented at each stage of the E-yarn production process. Further data have been presented for solar-E-yarns with different colours of outer covering fibres. Similarly, the performance of the solar energy harvesting fabric was assessed under different incident light intensities, with further experiments conducted when the fabric underwent mechanical deformation. Since the solution was mainly intended for outdoor wearable applications, the evaluations were conducted between 100% and 25% of one sun intensities.
A critical factor to understand for a solar energy harvesting solution for a wearable application is how the angle of incident light effects the functionality of the device: ultimately a wearer of a solar energy harvesting device would move relative to the light source (i.e. the Sun). The performance as a function of the incident angle of the light has been characterised for solar-E-yarns both free-standing (the solar-E-yarn on its own and not embedded within a fabric) and in fabric form (woven into fabrics).
The solar-E-yarns woven into fabrics were assessed for their wash durability, with both hand washing and machine washing explored (in a domestic washing machine with a 2kg washing load). Resistance to abrasion was also evaluated. Solar-E-yarns were also assessed under wet conditions. Finally, the ability of the solar energy harvesting fabric demonstrator to charge various energy storage devices (such as batteries and supercapacitors) was assessed.
This unique approach of embedding solar energy harvesting capabilities within textiles will revolutionise the way in which wearable and mobile electronic devices will be powered in the future. This technology allows for the creation of a solar energy harvesting fabric where the end user will not have to compromise on reusability, appearance, or comfort.

| Creating solar cell embedded fabrics
To construct the SC embedded demonstrator fabric, 20 solar-E-yarns with 10 SCs per yarn were used: A woven fabric was created using these solar-E-yarns inserted in the weft direction ( Figure 2B,C). A table top weaving loom (four shafts, 24″ width; Harris Looms, UK) was prepared with 12-cm-wide sheet of warp yarns (~10 yarns per centimetre) using white cotton yarns (38.9 × 2 tex; Elton Vale Yarns Ltd, UK). The warp yarns were threaded to achieve a four by one shedding pattern (a basket weave), as shown in Section S2. The solar-E-yarns were inserted in such a manner that the photoactive side was fully exposed on the front surface of the fabric. Cotton yarns (same type used for the warp) were used as weft yarns to fill the gaps between solar-E-yarns. In the case of demonstrator fabric ( Figure 2B), one cotton weft yarn was inserted between each of the solar-E-yarns.
To assess the effect of incident angle, fabrics with different spacings between the solar-E-yarns were made. A gap of~1.0 mm was achieved by inserting three cotton weft yarns between solar-E-yarns, and~3.0mm gap was achieved by inserting seven cotton weft yarns between the solar-E-yarns. When required, knit braided yarns (with packing yarns and the knit braid only but without a solar-micropod filament) were used as the weft yarns along with cotton yarns, to add extra length to the fabric samples.
Two minimodules, each consisting of 10 solar-E-yarns, were created by connecting 10 solar-E-yarns in series, as shown in the circuit diagram in Figure 2D. The SC embedded fabric demonstrator ( Figure 2E) was realised by wiring the minimodules in parallel (see Section S4 for further information). The SC embedded region of the fabric (the photoactive area) had a footprint of 44.5 mm × 45.5 mm ( Figure 2C).

| Characterisation of solar cell embedded yarns and fabrics
For electrical characterisation of the solar-E-yarns and demonstrator fabric, two light sources were employed. For the majority of the experiments in this work, ABA type solar simulator (LSH-7320, Newport Corporation, UK) was used, with the exception of the incident angle varied measurements for the free-standing solar-E-yarns. For this experiment, a bespoke optical test rig with a tungsten halogen lamp (described in Section S3) was employed.
Electrical measurements were taken using a high precision digital multimeter (Model 34410A 6 ½, Agilent Technologies LDA UK Limited, UK). Unless otherwise stated, one sun intensity (100 mW/cm 2 with an AM1.5 G spectrum) was selected on the solar simulator for measurements. All fixed angle measurements were conducted at a temperature maintained at 25 ± 1°C using a feedback controlled cooling system (described in Section S3).
For I SC measurements of solar fabrics under different incident light angles, a rotary sample holder was devised, with a goniometer (with 5°i ncrements). The rotary sample holder was mounted horizontally onto a vertical pole using an axel fixed through the rotary axis of the holder, allowing the angle of the sample holder to be varied relative to a horizontal plane.
To generate IV curves, a simple decadic resistor network (1 Ω -100 MΩ) was built using fixed resistors (RS Components, UK). For each data point, the voltage and the electrical resistance across the resistor network was measured to calculate the corresponding current and power values. The maximum power point was realised to the accuracy of ±10 Ω.
All metallic components used in the experimental setups were covered with black nonreflective coatings or tapes. Additional details of the instruments and methods used for characterisation are included in Section S3.

| Liquid moisture management test
A control fabric woven using knit-braided yarns without solarmicropod filaments (all other material and process parameters remaining identical to the solar-E-yarns) was prepared (using an identical woven structure, process parameters, and additional cotton yarns similar to the demonstrator fabric) for a comparison test.  Laundering for Fabrics and Apparel. 58 The washing and rinsing were conducted with 50 ± 2°C tap water (recorded using a digital temperature meter).

| Durability testing
The third set of solar-E-yarns was subjected to abrasion testing using an abrasion tester (902 Mini Martindale, James Heal Ltd, England) for 6000 abrasion cycles according to BS EN ISO 12947-2:2016. 59 Microscopic images were taken before the start of the tests and after every 1000 cycles while the test sample was fixed to the abrasion tester.
A fourth set of five solar-E-yarns were prepared with enamelled seven-strand copper wires with a nylon sheath (BXL2001, OSCO Ltd, UK), instead of the multistrand copper wire (used in previous experiments). These were woven into a fabric similar to those used for the other durability tests. These solar-E-yarns were characterised after soaking with, and immersing in, tap water at room temperature (~20°C-25°C) for 30 minutes.
The solar-E-yarns in the fabric samples were characterised for output current and voltage under standard one sun (100 mW/cm 2 , AM 1.5 G spectrum) illumination before and after they have been subjected to the above described durability testing.
Further details of the sample preparation and test conditions are provided in Section S4.

| Solar cell embedded yarns
The fabrication of solar-E-yarns involves three steps, soldering, encapsulation, and covering in fibres. After each step of production, the electrical characteristics of the SCs were determined, and the results are shown in Figure 3A,B. The I SC , V OC , P MAX , and fill factor (FF) values derived form the curves ( Figure 3C(i)-(iv)) showed the changes in optoelectronic output due to the fabrication process. The linear relationship between light intensity and I SC explains the clear change in I SC during the yarn fabrication process. On the other hand, V OC showed a modest change due to its logarithmic relationship with light intensity. 60 Therefore, I SC can be considered as the parameter, representative of the amount of light flux received by the embedded SC.
When the SCs were encapsulated within the resin micropods, the I SC and P MAX values increase by 18.3% and 21.7%, respectively, due to the convergent (lensing) and light trapping effects by the micropod.
A previous study 48  After covering the solar-micropod filament with a fibrous sheath, the I SC and P MAX values decreased by 29.3% and 32.5% (relative to When red-and black-coloured fibres were used for the fibrous sheaths (see Figure 4A), solar-E-yarns showed 89.4 ± 5.8% and 77.7 ± 1.2% of the normal solar-E-yarn (white sheath) I SC value, respectively ( Figure 4B). These experiments proved the viability of creating coloured solar-E-yarns without significantly compromising their performance. This also suggested that the light penetration into the resin micropod predominantly occured through the spaces between the fibres.
At one sun intensity, I SC , V OC , FF, and power density values of 14.14 ± 0.05 mA, 5.14 ± 0.02 V, 0.598 ± 0.004, and 2.146 ± 0.014 mW/cm 2 were observed, respectively (see Figure 5B,C). Overall, the I SC and P MAX values showed a close linear relationship with light intensity level, indicating a behaviour equivalent to a typical c-Si SC network. 60 The P MAX per solar-E-yarn when woven into fabric was 217.3 μW,~4.5% lower than the average P MAX values of 20 individual solar-E-yarns (227.5 ± 17.5 μW), owing to current/voltage mismatches caused due to cell-to-cell variations 63  The I SC after bending significantly differed from the I SC before deformation (14.14 mA), due to the lower surface area exposed to the light source (less than 30% of the photoactive area was exposed), which is also curved (incident angle varies across the exposed area). The modest reduction in I SC after draping was due to the curvature of the photoactive area. There was an insignificant change in I SC after shear deformation. These results indicated that the changes were likely caused by the change in incident angles, not due to electromechanical effects within the cells and cell network. The measured I SC returned to normal after deformation in all cases. These results provide evidence of the viability of the solar-E-yarns for wearable applications where the clothing will be exposed to different levels of sunlight and has to undergo various mechanical deformations during its regular use.
The lengths of yarn used in the demonstrator shown in this work included a string of 10 SCs each; however, practically, this could be changed considering the required current, voltage outputs, and anticipated partial shading for a specific application. The E-yarn technology can be employed to embed any type of small semiconductor device (as demonstrated previously in the literature 45,47,48,64 ) including bypass diodes. Therefore, a bypass diode can be included at the end of each yarn (cell string) to minimise the potential effects of partial shading.
While it may be vital to include bypass diodes to mitigate the adverse effects of partial shading, more specific information on a given application would be required to engineer an optimised E-yarn, hence this was beyond the scope of this paper.
The fabric demonstrator showed similar liquid moisture management characteristics (saturated water capacity of 17.43 g) to a control fabric sample (made from yarns without solar-micropod filaments, all other parameters remained constant), with a saturated water capacity of 17.29 g. The moisture transfer behaviour is crucial for the thermal comfort of the wearer, especially for activewear applications where the wearer generates sweat and heat while directly being exposed to warm conditions.  In all of the cases explored, the solar-E-yarns within woven structures showed higher I SC values than the corresponding free-standing values for small incident angles and lower I SC values for higher incident angles. It is believed that this effect was due to the significant albedo effect 66 (light diffused by the surrounding and background) from the surrounding fabric at smaller incident angles, as shown in Figure 6C.

| Effects of change in the incident angle of light
When the incident angle increased, the adjacent solar-E-yarn started to shade part of the incident light. By studying the cross-sectional geometry of the woven structure, it was clear that the angle at which the direct shading from yarns started to occur increased with the gaps size, which supported the experimental results. When the incident angle increased, the light flux to the solar-E-yarn reduced due to direct shading (from neighbouring solar-E-yarns) as well as a diminished albedo effect (as the gaps between the solar-E-yarns were also shaded); this is illustrated in Figure 6E.
The impact of the albedo effect due to the surrounding fabric was investigated empirically ( Figure 6D)  while closely packed solar-E-yarn designs would give the best performance for applications with more predictable or small incident angles, or where the available surface area is limited. Additionally, a distributed yarn design would result in improved drapability; however, a detailed study of drapability was beyond the scope of this work.

| Durability testing
The wash durability of sets of five solar-E-yarns woven into fabrics was assessed using two washing methods. The first method used a domestic washing machine where one set of solar-E-yarns was washed with detergent inside of a wash bag in a 2-kg wash load and line dried. For the second method, a set of solar-E-yarns was hand washed with a detergent and line dried. Five solar-E-yarns woven into a fabric showed only a 5.6% reduction in I SC after 6000 abrasion cycles ( Figure 9A). Microscopic images confirmed that the solar-E-yarn surface fibres were redistributed due to abrasion, with an insignificant amount of surface fibre breakages ( Figure 9B): this explains the change in performance after the abrasion cycles.
The results of the abrasion test confirmed that the solar-E-yarn embedded fabrics can withstand the washing and drying, mechanical rubbing, and wearing undergone by regular clothing without significantly altering their energy conversion capability. In addition, the results provided evidence of continuous functionality of the fabric under wet conditions, which is beneficial for outerwear.

| Demonstrating power generation capability
The ability of the SC embedded fabric demonstrator to charge three types of electrical device was explored, as ultimately the SC embedded fabric was designed to charge wearable devices. A 47-mF (5.5-V) supercapacitor (KEMET Electronic Components, USA) was charged using the solar fabric (under 100%, 75%, 50%, and 25% of one sun illumination); the supercapacitor reached its maximum voltage within 15 and 60 seconds under 100% and 25% sun intensity, respectively ( Figure 10A).
Under 100% sun intensity, the fabric woven with solar-E-yarns was able to charge a 15-mAh (3.7-V) Li-ion battery to~3.7 V within  performance. The study has revealed that in fabric form, larger gaps between SC embedded yarns will enhance the performance across a wider range of incident angles due to the albedo effect and reduced shading. This meant that the SC embedded yarn distribution within a fabric can be varied to achieve the optimum balance between angle independence and power density for a given application. The experimental results on power generation performance under various lighting conditions, for wash durability, moisture management behaviour, and conformability to three-dimensional shapes, validate the utility of the solution for regular clothing applications. These attributes will enable these solar fabrics to feature in future wearable electronics and electronic textiles to provide a continuous supply of power, without having to compromise on comfort, aesthetics, or wash durability.