Archana Samanta and
Romain Bordes*
Department of Chemical Engineering, Chalmers University of Technology, Gothenburg, Sweden. E-mail: bordes@chalmers.se
First published on 13th January 2020
In the development of embedding devices on textiles like sensors and detectors the controlled formation of a conductive coating remains a critical point. Although there are several approaches for imparting conductivity to any textile, the challenges remain in balancing the practical aspects of the coating procedure that affect the conductivity with the associated mechanical properties of the textile along with a feasible economic viability of the process. In this research we developed an approach to deposit uniform conductive graphene surface coatings on polyester (PET) fabric using graphene oxide (GO) particles. Instead of using pre-reduced graphene oxide (rGO), an approach of coating an aqueous dispersion of hydrophilic GO particles was adopted. Stable aqueous dispersions of GO were formulated, and the dispersion properties were characterized using DLS and zeta potential before coating them onto the PET textiles. This approach not only helped in developing an aqueous coating technique but also helped in avoiding the need of any organic solvents which might have been required for coating hydrophobic rGO moieties onto the textile substrates. The uniformity of the coating was analyzed using scanning electron microscopy (SEM). Later, the GO coated textiles were reduced via thermal and chemical approaches and their effects on the conductive and mechanical properties of the fabric were assessed and compared. The reduction efficacy was analyzed and compared using XPS. The conductivity and water adsorption properties were correlated to the uniformity and retainment of rGO on the surface of the conductive textiles.
In general, fabrication of conductive textiles can be done through two approaches; (a) weaving or knitting a textile with conductive yarn of copper/silver/carbon fiber11 or (b) depositing conductive particles onto a woven or knitted fabric via suitable coating process.12 The second process is more commonly considered because of economic viability and ease of processing. Micro or nano dimensional particles of silver,13 gold,14 carbon nano tubes,15 carbon black,16 graphite17 etc., have been used for imparting conductive coatings. Among these class of particles, graphene has been of interest because of its high thermal conductivity (∼3000 W m−1 K−1) and high electron mobility (15000 cm2 V−1 s−1) along with comparatively high Young's modulus of ∼1.0 TPa, and high specific surface area owing to its two-dimensional nano structures with sp2 bonded conjugated carbon system.18–20 Because of these properties graphene based coatings have attracted interest for usage in the fields of photosensors,21 gas sensors,22 wearable electronics23 and others.24
However, realizing a uniform graphene coating remains a challenge in practice, as graphene cannot be handled as other conductive particles that are typically dispersible in a solvent, quite often water. This latter feature enables application of conductive nanoparticles by dip coating process,25 where the solvent is later removed by evaporation assisted with heat treatment. Brush coating,26 screen printing,27 electrochemical deposition28 techniques are also used for producing conductive coatings. However, their usage is limited due to the multi-step process and difficulty in scaling up. Furthermore, these treatments have poor environmental footprint as they consume large amount of water. One alternative to graphene is indeed graphene oxide (GO). GO differs from graphene in the sense that it has several oxygen functionalities at the edge and basal planes, which enables forming a stable aqueous dispersion. These groups increase the material hydrophilicity and enable water dispersibility. But at the same time they interrupt the sp2 carbon conjugation, resulting in meager conductivity;29 an issue that is typically remediated by in situ reduction. Thermal reduction and chemical reduction are the two mainly used process for reducing graphene oxide to its conductive form.30 Thermal reduction process is the simplest and most convenient technique to form highly conductive reduced graphene oxide (rGO).31 Alternatively, chemicals can be applied to GO once coated onto the desired substrate, through the addition of a suitable reductant. The sample are then washed thoroughly to get rid of the excess of reductant. Various chemical agents such as sodium hydrosulphite,32 hydrazine,33 sodium borohydride,34 oxalic acid,35 hydriodic acid36 and others have been used for chemically reducing graphene oxide. Green alternatives, such as glucose,37 citric acid,19 ascorbic acid,38 bovine serum albumin,39 polyphenols40 (present in green tea and caffeic acid) etc. are being used either solely or in combination with other chemical reductants or temperature to reduce graphene oxide.41
An investigation of the recent studies involving applications of graphene for conductive textiles, showed that most of these processes emphasized on chemical reduction via dip coating process, for example coating of graphene oxide and AgNO3 onto PET textile and its subsequent reduction with hydrazine42 or in another study Sahito et al. coated positively charged cotton fabrics with GO and reduced it using hydrazine hydrate vapors.43 Berendjchi et al. immersed the PET fabrics in GO dispersion at 70 °C for 1 h and then cured at 80 °C for 60 min. These samples were then immersed in Na2S2O4 solution at 90 °C for 60 min. And were then coated with poly(pyrrole) for better conductivity.44 In another study polypropylene non-woven webs were coated with GO dispersions in DMF which was then reduced using hydroiodic acid. Generally, a combination of dip coating with assisted chemical reduction process is adopted to fabricate conductive textiles.
In the present study we suggest an alternative that capitalizes on spray coating of dispersion of graphene oxide. GO coated textiles are reduced using thermal reduction, chemical reduction using NaBH4 and green reduction approach using L-ascorbic acid which is a nontoxic mild reducing agent45 and a comparative analysis of the reduction methods affecting their properties is studied in details. The motivation is to develop a continuous, energy efficient, economically viable and environmentally friendly coating methodology which can be easily adopted easily by textile industries for large scale fabrication of conductive coatings.
Conductivity surface conductivity of the samples was determined using a homemade 4 probe devices. Samples were placed on a flat insulating surface and the probes were contacted with the samples to get the readings on a digital screen. Sample dimensions were 2 cm × 5 cm. Surface conductivity (mS m−1) was calculated by dividing length of the tested sample with the product of measured resistance and area of cross section of the textile which was around 1.4 × 10−6 m2.
Dynamic light scattering was performed on Beckman Coulter, N4 plus sub-micron size analyzer to determine the particle size of the dispersions after sonication. Zeta potential was measured using a Zeta PALS zeta potential analyzer from Brookhaven instrument corporation with a red He–Ne laser (λ = 633 nm). Contact angle was measured using an Attension Theta optical tensiometer (Biolin, Finland). A drop of 2-5μl of water was applied to measure the contact angle and the time of absorption for the dried rGO samples.
X-ray Photoelectron Spectroscopy (XPS) was carried out to evaluate the graphene coating and chemical reduction. XPS was conducted using a PHI VersaProbe III Scanning XPS Microprobe from Physical Electronics; X-ray excitation was provided by Al Kα radiation with photon energy hν = 1486.6 eV. A sample area of approximately 400 × 500 μm was scanned, with an indicative information depth of about 4–5 nm. XPS measured the number of electrons and their corresponding kinetic energies which were released when the textile was irradiated with the X-ray beam. These information was then corelated to determine the binding energy spectra and composition of the coating at parts per thousand range.47
Tensile properties were measured using Instron 5565 instrument. Samples were preconditioned at 65% relative humidity, 20 °C for 12 hours before testing. A strain rate of 500 mm min−1 was applied to fracture the samples that had a 2 cm × 5 cm as dimension. An average of 5 readings is reported.
DLS was used to identify the average size of the GO aggregates in water. Though the GO are present in form of flakes, the value from DLS provided an indication of the overall GO aggregated dimensions in the dispersions and the values are reported in Table 1. As can be seen, the average particle size increases slightly with increase in GO concentration. Further, usage of 1:1 weight fraction of the surfactant with respect to GO resulted in reducing the particle size by preventing their aggregation. However, the trend is similar in both cases, i.e. increase in GO concentration results in formation of larger particles in the dispersions. This can be caused by remnants of aggregated GO particles at higher GO contents as the dispersion and sonication period was kept constant in all cases. As the presence of Ethylan 1008 helped in achieving lower sizes of GO dispersions, it was considered that in these dispersions presence of smaller GO moieties can result in uniform film coatings compared to the non-surfactant ones with same GO content and can potentially result in better conductivity of GO sheets by imparting better connectivity in the final product.
Sl no. | Concentration of GO (wt/wt%) in aqueous dispersion | Surfactant used | Particle size DLS (nm) |
---|---|---|---|
1 | 0.01 | — | 406 ± 4 |
2 | 0.1 | — | 471 ± 9 |
3 | 0.5 | — | 512 ± 14 |
4 | 0.01 | Ethylan 1008 | 385 ± 7 |
5 | 0.1 | Ethylan 1008 | 418 ± 9 |
6 | 0.5 | Ethylan 1008 | 484 ± 12 |
Zeta potential of the GO particles was measured to check if their oxidation degree would ensure electrostatic stabilization dispersion. Irrespective of the concentration, the zeta potential remains constant with values −37 mV and −40 mV for GO dispersions with and without surfactant respectively, which provided sufficient stabilization. As expected, the use of non-ionic surfactant did not influence the zeta potential values, since the surfactant is meant to provide steric stabilization. As the surfactant free dispersions had reasonable colloidal stability compared to the ethylan GO dispersions, an attempt was also made to study and compare their film formation tendency and conductive properties.
The purpose of this study was to utilize the hydrophilic characteristic of GO to form a direct aqueous dispersion, coat them to the textiles and then reduce the GO layer. This is because of better dispersibility of GO compared to its reduced form in aqueous media. Therefore, we sprayed 1 ml of GO dispersions with various GO concentrations onto a 10 × 10 cm2 of textiles and the dry weight increase was calculated from the weight change in the dried coated the samples. Summary of the samples details is given in Table 2.
Code | Conc. of GO wt%/application process | Weight increasea (%) | Vol. of spray (ml) per 10 × 10 cm2 | Reduction process | Surface conductivity (mS m−1) | Contact angle of the fabricc (°)/time of absorption (s) |
---|---|---|---|---|---|---|
a % Dry weight increase after coating with GO.b Samples containing Ethylan 1008, ascorbic and NaBH4 were washed thoroughly with deionized water after the drying of the sprayed formulations.c Surface resistance was too high to be measured by the instrument. | ||||||
SE1 | 0.01 spray, Ethylan 1008b | 0.02 | 1 | 200 °C, 2 h | —c | 65° ± 11, 95 ± 14 s |
SE2 | 1.1 spray, Ethylan 1008b | 0.17 ± 0.02 | 1 | 200 °C, 2 h | —c | 68° ± 7, 135 ± 17 s |
SE3 | 0.5 spray, Ethylan 1008b | 0.92 ± 0.04 | 1 | 200 °C, 2 h | 0.31 ± 0.07 | 74° ± 8, 152 ± 11 s |
SE4 | 0.5 spray, Ethylan 1008b | 2.3 ± 0.06 | 3 | 200 °C, 2 h | 0.37 ± 0.05 | 76° ± 10, 155 ± 8 s |
SE5 | 0.5 spray, Ethylan 1008b | 4.5 ± 0.05 | 5 | 200 °C, 2 h | 0.45 ± 0.11 | 76° ± 14, 159 ± 15 s |
S1 | 0.01 spray | 0.01 ± 0.01 | 1 | 200 °C, 2 h | —c | 69° ± 7, 110 ± 14 s |
S2 | 0.1 spray | 0.07 ± 0.02 | 1 | 200 °C, 2 h | 0.37 ± 0.09 | 71° ± 8, 158 ± 10 s |
S3 | 0.5 spray | 0.37 ± 0.08 | 1 | 200 °C, 2 h | 0.48 ± 0.10 | 71° ± 3, 900 ± 10 s |
S4 | 0.5 spray | 1.10 ± 0.07 | 3 | 200 °C, 2 h | 0.69 ± 0.07 | 75° ± 2, 1200 ± 7 s |
S5 | 0.5 spray | 1.83 ± 0.44 | 5 | 200 °C, 2 h | 2.38 ± 0.11 | 79° ± 8, 1500 ± 12 s |
SA1 | 0.5 sprayb | 0.37 ± 0.95 | 1 | L-Ascorbic acid | 0.48 ± 0.11 | 70° ± 5, 120 ± 9 s |
SA2 | 0.5 sprayb | 1.10 ± 0.75 | 3 | L-Ascorbic acid | 0.55 ± 0.08 | 73° ± 4, 660 ± 12 s |
SA3 | 0.5 sprayb | 1.83 ± 0.61 | 5 | L-Ascorbic acid | 0.83 ± 0.09 | 74° ± 2, 785 ± 12 s |
SN1 | 0.5 sprayb | 0.37 ± 0.89 | 1 | NaBH4 | 0.37 ± 0.06 | 68° ± 3, 109 ± 7 s |
SN2 | 0.5 sprayb | 1.10 ± 0.94 | 3 | NaBH4 | 0.48 ± 0.08 | 77° ± 3, 133 ± 9 s |
SN3 | 0.5 sprayb | 1.83 ± 0.45 | 5 | NaBH4 | 0.69 ± 0.07 | 64° ± 4, 186 ± 17 s |
B1 | 0.5 brush | 0.37 ± 0.12 | 1 | 200 °C, 2 h | 0.41 ± 0.07 | 65° ± 8, 720 ± 10 s |
B2 | 0.5 brush | 1.10 ± 0.08 | 3 | 200 °C, 2 h | 0.53 ± 0.04 | 66° ± 5, 840 ± 7 s |
B3 | 0.5 brush | 1.83 ± 0.52 | 5 | 200 °C, 2 h | 1.28 ± 0.09 | 72° ± 7, 1055 ± 15 s |
PETN | PET neatc | NA | NA | NA | c | 110° ± 2, 62 ± 4 s |
Two sets of textiles with and without surfactant were coated with similar concentration of GO as depicted in Table 2. In both cases, textiles coated with 0.01 wt% of GO dispersion showed poor interconnectivity of GO layer on the surface, as shown by the SEM images, Fig. 1. A volume of 1 ml was kept constant as this volume was found to provide enough wettability of fabric coating. GO concentration was increased to 0.1 wt% and then to 0.5 wt%. Application of 0.5 wt% resulted in better coverage of the fabric but non-continuity of coating was still prevalent. Above 0.5 wt%, the dispersibility of GO was affected and sedimentation occurred. To overcome this the volume of the formulation was increased from 1 ml to 3 ml and then to 5 ml per 10 × 10 cm of the textile to increase dry weight% values to from 0.37% to 1.10% and 1.83%, respectively. This resulted in the formation of a continuous uniform coating of GO on the surface of the fabric as shown by SEM. With 1.83% a well-connected GO layer was formed on the surface of the textiles even without the surfactant (Fig. 1e and g). Samples containing surfactant were washed with deionized water to remove excess surfactant from the surface which otherwise would have impacted the surface conductivity. The SEM image represented in Fig. 1h, shows removal of GO layers from the surface in the washed samples, implying that though the presence of surfactant would help stabilizing the GO dispersion it could induce a loss of material upon washing.
For comparison, brush-coated samples with similar GO content and volume were prepared. SEM image of brush coated sample with highest GO content i.e. of 1.83% is shown Fig. 1f. As can be observed the spray coated sample has more uniform GO film as compared to the brush coated sample with same GO content. This gives an indication that during brush coating the GO distribution is not uniform, probably the GO moieties are unable to spread uniformly onto the contours of interlacement points of warp and weft by the mechanical act of brush coating.
The efficiency of the reduction process of the GO spray coated textiles on GO coated PET samples was evaluated by XPS analysis, as reported in Table 3, while the effect of reduction on the coating integrity was evaluated by SEM, Fig. 2 and S1 (ESI†). Reduced GO forms a conjugated structure with reduced number of oxygen groups compared to its oxide state. Samples with 1.83% dry GO content, reduced via various aforementioned techniques which showed better surface coverage were analysed for this study. The percentage of oxygen reduces with the reduction process. GO coated samples reduced by NABH4 and ascorbic acid showed similar oxygen content reduction which was slightly lower than the thermally reduced PET sample. A maximum of 13% reduction of oxygen content was observed in thermally reduced GO coated PET sample. This data is also in agreement with the conductivity value which is highest with thermally annealed GO coated PET fabric at a dry GO content of 1.83%. The reduction process also affected the coating integrated and in particular through chemical approach as the samples were put under wet condition for a long period.
Sample | Atomic content (%) | |
---|---|---|
Carbon (C) | Oxygen (O) | |
PET neat | 67.95 | 30.66 |
PET GO thermally reduced | 77.18 | 21.99 |
PET GO ascorbic acid reduced | 72.90 | 24.82 |
PET GO NaBH4 reduced | 72.89 | 24.11 |
PET coated non-reduced | 68.02 | 31.19 |
The continuity of the GO layers in S5 samples was retained even after annealing at 200 °C for 2 h and later washing the annealed samples (Fig. 2a and b). Annealed B3 sample showed poorly connected corners and contours, associated with brush coating process. Spray coated samples with this concentration reduced chemically by NABH4 and ascorbic acid are shown in Fig. 2b and c respectively. Patches of non-continuous GO regimes could be observed, which presumably result from the removal of reduced GO layer during the dip immersion and washing stages of the reduction process. This hypothesis is validated by the conductivity data which showed poor surface conductivity of NaBH4 and ascorbic acid reduced samples due to formation of vacant spaces resulting in low connectivity of GO sheets.
The cross sections of the textile showed a complete non-bridged/non-covered layer signifying that the coating is only at the surface and it do not cover the cross section, Fig. 2f and g. This implies that the samples can suitably be used for e-textile applications without posing a potential threat for inherent sort circuiting as the two surface coated layers are well separated from each other. It was also observed that the spray coated sample had rGO layer more firmly attached to the textile surface compared to the brush coated samples (indicated by arrows in Fig. 2f). Poor penetration tendency of GO flakes in the brush coated samples can result in formation of a surface skin in these samples. The conductivity of the samples was retained after 10 washing cycles (samples were stirred under de-ionized water containing 1 wt% of Ethylan 1008, a non-ionic surfactant at 40 °C for 15 minutes and were washed under running water), Table S1 (ESI†) indicating that the coating had good wash fastness.
The abrasion resistance analysis of these coated samples was performed for 30000 cycles on a Martindale abrasion resistance tester where we noticed a drop of surface conductivity over 10-fold magnitude. We are currently working to improve the abrasion resistance of these coatings via surface modification mechanism. However, these abrasion conditions are well beyond the standards.
The reducing method also appeared to be central in the realization of conductive textile. Thermal annealing approach resulted in better conductivity compared to chemical reduction process of GO coated on textiles using sodium borohydride or ascorbic acid, as well as in a higher integrity of the resulting rGO layer. The structural defects induced by the wet nature of the chemical reduction turned out to impact negatively the mechanical properties of the coated PET. Even though the tensile modulus of the coated reduced samples was improved compared to neat textiles, it was for the thermally reduced rGO that the mechanical properties of PET were observed to be enhanced.
With this study, we demonstrate that adoption of the spray coating technique combined with a thermal annealing could be an efficient way to produce GO coated conductive textiles, while minimizing energy, and resources consumption.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9ra09164e |
This journal is © The Royal Society of Chemistry 2020 |