A lay-by-layer assembly process was used to prepare a superhydrophobic encapsulated MWCNTs/Bi2Te3 thermoelectric fabric. As illustrated in Fig. 1(a), MWCNTs dispersions were uniformly drop-coated on the cotton fabric substrate to construct a primary conductive network on the surface of the fabric. Then, Bi2Te3 particles were deposited upon the MWCNTs conductive networks via a PVD process. To minimize the decreasing of electrical conductivity during deformation, MWCNTs with high aspect ratio are appropriate for constructing effective conductive networks on flexible substrate. The inorganic Bi2Te3 particles decorated along the surface of MWCNTs can enhance the Seebeck coefficient of the coated cotton fabric. Thus, the conversion performance of the designed flexible thermoelectric fabric is expected to be improved. A double layer encapsulation was applied to fabricate the superhydrophobic thermoelectric fabric. A silicone with Shore 20A was coated onto the surface of Bi2Te3/CCF sample to obtain the first protection layer. Then, the second superhydrophobic layer was fabricated by spraying the dispersion of PDMS and PMMA in THF on the silicone layer. The low surface energy and the microscale roughness of PDMS/PMMA coating synergistically contributed to the superhydrophobic and self-cleaning effect of the thermoelectric fabric. It should be noticed that direct coating the PDMS/PMMA on the Bi2Te3 surface cannot obtain the superhydrophobic effect, and furthermore, it also reduces the performance of thermoelectric layer. The PDMS (Sylgard 184 Silicone) used for superhydrophobic coating has a hardness of 43A, which would lead to a severe cracking phenomenon on Bi2Te3 layer and the difficult spreading out of PDMS dispersions on the Bi2Te3 surface. By introducing the Shore 20A silicone as a middle protection layer, the cracking of Bi2Te3 layer was reduced significantly, and the PDMS superhydrophobic coating can be successfully formed.
The photograph of the prepared CCF, Bi2Te3/CCF, and P-Bi2Te3/CCF samples are shown in Fig.1 (b)-(d) respectively. The CCF conductive fabric is black and can be folded at any angle, which was attributed to the flexibility of cotton fabric substrate. After the thermal evaporation of Bi2Te3, the color of the Bi2Te3/CCF have turned to silver grey, suggesting that the Bi2Te3 powders have been successfully decorated on the surface of CCF, and the Bi2Te3/CCF remained flexible. Since the double layer superhydrophobic encapsulation is transparent and flexible, the P- Bi2Te3/CCF strip exhibits the same silver grey color and can be easily twisted without being damaged.
The SEM images of original cotton fabrics (Fig. S1(a, b)), CCF (Fig. 2(a-c)), Bi2Te3/CCF (Fig. (d-f)) and P-Bi2Te3/CCF (Fig.(g-i)) were observed, respectively. Fig. 2 (a) and (b) exhibit the surface structure of CCF, and the cross-section structure is shown in Fig.2 (c). It can be observed that MWCNTs with high aspect ratios were randomly connected and uniformly coated on the surface of cotton fabric substrate. Thus, the primary conductive network was formed. The surface and cross-section structures of Bi2Te3 coating layer are shown in Fig. 2 (d-f), indicating that Bi2Te3 particles were successively deposited on the surface of MWCNTs. The surface and cross-section structures of the PDMS/PMMA superhydrophobic coating are shown in Fig.2 (g-i). The microscale rough surface was achieved by distributing polymer particles in the coating layer, which contributed to the superhydrophobic performance of the encapsulation layer.
X-ray energy dispersive spectrum was used to express the surface elements of the prepared CCF, Bi2Te3/CCF and P- Bi2Te3/CCF samples, respectively. As shown in Fig. 2(j), compared with the original cotton fabric (Fig S2), the surface element species of the CCF sample remained the same, but the content of C element increased significantly, from 55.41wt% to 64.48wt%, which was attributed to the adhesion of MWCNTs. In contrast, as shown in Figure 2(k), the Bi and Te peaks appeared on the surface of Bi2Te3/CCF, and the contents of C, O, Bi and Te elements were 25.61wt%, 7.29wt%, 11.23wt% and 55.86wt% respectively, which indicates the successful coating of Bi2Te3 on the CCF surface. Figure 2 (l) shows the surface elements of P- Bi2Te3/CCF. The peaks of Bi and Te disappeared, replaced by the appearance of Si peaks, which indicates the Bi2Te3/CCF was well encapsulated by the superhydrophobic coating of silicone and PDMS. Furthermore, the elemental distribution of these samples showed that MWCNTs, Bi2Te3 and PDMS/PMMA were uniformly covered the fabric substrate.
To further investigate the thermoelectric properties of the prepared Bi2Te3/CCF, the electrical conductivity, Seebeck coefficient, and power factor at room temperature were observed and calculated respectively. The thermoelectric performances of Bi2Te3/CCF with different deposition times of Bi2Te3 powder are shown in Fig.3 (a) and (b). It can be observed that both electrical conductivity and Seebeck coefficient of the Bi2Te3/CCF samples were enhanced with the increased times of Bi2Te3 deposition. Obviously, Bi2Te3 as an inorganic semiconductor has excellent thermoelectric performance at room temperature. The more times of deposition would increase the coating contents of Bi2Te3 on CCF surface, which resulted in the enhancement of electrical conductivity and Seebeck coefficient. However, the electrical conductivity of the Bi2Te3/CCF sample was decreased at fifth times of deposition. This result implies that further increasing the number of Bi2Te3 deposition times cannot continuously increase the thermoelectric performance. Since the Bi2Te3 is a rigid semiconductor lack of flexibility, a thicker Bi2Te3 coating layer is prone to crack, which would lead to the breaking of electrical connections. The SEM images of Bi2Te3/CCF under different times of Bi2Te3 deposition (Fig. S3) proved that Bi2Te3 coating layers were gradually thickening with increased number of deposition times. Although the Seebeck coefficient of Bi2Te3/CCF-5 sample was still increased, the overall power factor of the sample was decreased due to the reduced electrical conductivity. Therefore, the optimal power factor of 0.15 μW·m−1·K−2 was achieved by Bi2Te3/CCF-4 sample.
The heat transfer performance of the sample Bi2Te3/CCF-4 was visualized using a temperature-sensitive paint which was red at room temperature and turned yellow when the temperature was above 40 °C. As shown in Fig. 3(c), the two ends of the sample with a length of 4.5cm and a width of 2cm were placed on the surface of two copper blocks, and then the left copper block was heated to 45 °C, while the temperature of the right copper block was kept at room temperature. The sample was placed in physical contact with the copper block so that the heat of the copper block can transfer laterally along the sample to change the color of the temperature-sensitive paint. After heating the left side of the sample for 2 hours, the temperature-sensitive paint on the right side of the sample remained red (Fig.3(c)). As shown in movie S1, the hot end of thermochromic paint was slowly extended about 0.3cm toward the cold end during the 2 hours heating. It demonstrated that the prepared Bi2Te3/CCF-4 sample has limited in-plane heat transfer, which is conducive to the heat-to-electricity conversion process.
In wearable application, flexible thermoelectric devices without proper encapsulation are easily damaged by external environment. Oxidation and pollution will reduce the performance of the devices. To overcome these obstacles for real application, superhydrophobic encapsulation of flexible fabric based thermoelectric materials was introduced. The evolution of surface hydrophilicity for the thermoelectric fabric Bi2Te3/CCF, Silicone coated fabric SC- Bi2Te3/CCF and superhydrophobic fabric P-Bi2Te3/CCF was investigated. As shown in Fig. 4 (a), the water contact angle (CA) was converted from 0.3° to 114.2° after encapsulation of the thermoelectric fabric Bi2Te3/CCF with silicone, resulting in a hydrophobic effect. The silica gel coating forms a translucent film on the surface of the thermoelectric fabric Bi2Te3/CCF (Fig. S4). It is worth noting that the hardness of silicone used here is 20A. The Bi2Te3 coating layer encapsulated with higher hardness silicone would be cracked severely. To further achieve the superhydrophobic effect, the dispersion of PDMS and PMMA in THF was sprayed onto the surface of 20A silicone encapsulated fabric. Hence, the surface roughness of the encapsulation layer could be increased, resulted in a superhydrophobic stain repellent coating. Generally, a surface with water contact angle higher than 150° is considered to have superhydrophobic property. As shown in Fig. 4 (a), the water contact angle of P-Bi2Te3/CCF was increased to 158.6°, indicating that the prepared encapsulation layer has superhydrophobic performance. P-Bi2Te3/CCF also showed low hysteresis with a small sliding angle (SA) of 6.5°, indicating the low surface adhesion (Fig 4 (b)). When other liquid droplets such as water, milk, and coffee were dropped on the surface of the P- Bi2Te3/CCF, the droplets steadily stood on the surface rather than permeating the thermoelectric fabric (Fig. S5(a)). Therefore, the stains in daily life could be easily removed from the surface of the encapsulated thermoelectric fabric with rinse (Fig. S5(b)). Besides, the encapsulated sample exhibits superior mechanical flexibility. As shown in Fig. 4(c) and (d), the P- Bi2Te3/CCF-4 sample can be easily bended and twisted. With the self-cleaning property and great flexibility, the thermoelectric fabric can be used in various daily wearable applications.
Meanwhile, the impact of the superhydrophobic encapsulation, mechanical deformation, and chemical solvents on the thermoelectric performance of P-Bi2Te3/CCF-4 sample were evaluated. As shown in Fig. 4(e), compared with the original Bi2Te3/CCF-4 sample without encapsulation, the Seebeck coefficient of the P- Bi2Te3/CCF sample was only slightly decreased from 20.26μv/k to 19.82μv/k, which indicated that the double layer superhydrophobic coating had little negative effect on the original thermoelectric performance. To investigate the mechanical durability, the P-Bi2Te3/CCF-4 sample was bended with a distorted radius of 2.0mm and twisted with a torsion angle of 360° respectively. After 100 bending-release cycles and 100 torsional-release cycles, P- Bi2Te3/CCF maintained excellent Seebeck coefficients of 18.62μv/k and 18.38μv/k, respectively, which indicated that the P- Bi2Te3/CCF had excellent mechanical flexibility and durability. To overcome the adverse impact of harsh environment, the thermoelectric fabric needs to have certain chemical resistance. To investigate the influence of chemicals on the Seebeck coefficient, P- Bi2Te3/CCF-4 sample was immersed in acid (pH = 2.0), alkali (pH = 12.0), ethanol, formaldehyde, and propanol respectively for 10 h. As shown in Fig. 4(f), the Seebeck coefficient of the P- Bi2Te3/CCF-4 sample after immersed with the above chemicals was still higher than 18.83μv/k, indicating that these chemicals have no obvious impact on thermoelectric performance of the sample. Hence, the prepared superhydrophobic thermoelectric fabric exhibits great mechanical flexibility and chemical resistance that could ensure long-term durability in daily life wearable situations, even in harsh environments.
To demonstrate the potential of the P-Bi2Te3/CCF thermoelectric fabric for wearable application, a fabric TEG with 5 legs was fabricated. As shown in Fig. 5(a), five rectangular P- Bi2Te3/CCF strips with the size of 3×0.6cm (L×W) were used as P-type legs and connected in series with copper wire. Copper foils were adhered to the two ends of P- Bi2Te3/CCF strips to improve the electrical connection stability between the copper wires and P-type legs. To evaluate the output performance of TEG, one side of the TEG was heated, and the other side was exposed to the ambient atmosphere. As the temperature difference increased from 5℃ to 30℃, the corresponding output voltage and output power were measured. As shown in Fig. 5(b), an output voltage of 1.8 mV and output power of 1.6 nW were obtained at a temperature difference of 30℃. Besides, the prepared wearable TEG can be folded and attached to the end of sleeves or socks, as shown in Fig. 5 (c). In this design, one side of the TEG can be in contact with the skin or placed at the inner layers of clothes, while the other side can be exposed to the environment. Generally, wearable devices are easily affected by the surrounding environment in practical applications. For example, when the TEG is placed on the wrist or ankle, it will be easily contaminated by sweat or rain. In this case, the superhydrophobic encapsulation can improve the durability and practical application value of thermoelectric fabrics.