Dry textile electrode structures and their coated conductive materials
In this study, several different materials were used to create electrodes using screen printing on flat and raised, silver and carbon knitted electrodes. Then the impedance of these electrodes to a skin model was measured, and ECG was recorded by these electrodes, to compare their performances to gel electrodes as well as bare silver and carbon electrodes. Silver-plated nylon and carbon-contained nylon conductive yarns (Myant Inc., Canada) were knitted into four different structures, namely, flat textile electrodes made of silver yarn (FS Sample code), raised 3D textile electrodes made of silver yarn (RS Sample code), flat textile electrodes made of carbon yarn (FC Sample code) and raised 3D textile electrodes made of carbon yarn (RC Sample code) using a flatbed knitting machine (Stoll, Ruetlingen, Germany).
Each textile electrode has a circular structure with a diameter of 2 cm, matching the active area of the hydrogel electrodes. Figure 1 shows a schematic of knitted textile swatches and the physical appearance of the flat/raised textile electrodes. A knitted silver-plated nylon trace/cable was used as a connection line between the knitted flat/raised electrode area and an electrode snap located on the back of the sample (Fig. 1a and b). Figure 1c-f show images of the flat and raised textile electrodes made of conductive silver and carbon yarns, respectively.
Thereafter, different conductive pastes (20 types) were screen-printed on these 4 types of textile electrodes (namely FS, RS, FC, and RC, shown in Fig. 1c-f). Details of conductive pastes and formulations described in “Methods” section.
The performance of dry textile electrodes for biopotential monitoring is determined by a variety of factors (Table S1), e.g., textile electrode substrates and structures, conductive coating materials for screen-printing, applied pressures, and frequencies. In this section, we first explored the effect of electrode structures and coating conductive materials on the fabrication of dry textile electrodes. The dry textile electrodes were manufactured as flat or raised 3D structures (Fig. 1c-f). The flat structure allows for uniform pressure distribution, in line with the base fabric construction. Conversely, the raised 3D structures are protrusions from the base fabric construction leading to higher contact pressure, which could result in improved signal acquisition properties (10). The microscopic images of FS (Fig. 1c) and FC (Fig. 1e) are shown in Fig. 2a and b. The microscopic images of RS (Fig. 1d) and RC (Fig. 1f) are shown in Fig. 3a and b, respectively. We can find a similar microscopic morphology between flat and raised 3D textile electrodes made of silver yarn (Fig. 2a and b), but a significant morphology difference between flat and raised 3D textile electrodes made of carbon yarn (Fig. 3a and b). Thereafter, we screen-printed conductive materials (listed in Table 1 and Table S2) on these four types of textile substrates. Carbon-contained conductive coating materials (with/without IL) showed good printability on flat structures without obvious cracks formation (Fig. 2c and d, Fig. 3c-f). However, carbon-contained conductive coating materials with high concentrations of IL (e.g., 7.5%) showed cracks after screen printing on the flat textile substrates (Fig. 3g and h) and failed on the raised 3D textile substrates, which was excluded from implementation. The PEDOT:PSS + PDMS-contained conductive coating materials with/without CNTs or with a suitable concentration of IL (e.g., ≤ 5%) only showed printability without mechanical cracking on the flat textile substrates made of silver yarn. In contrast, the other combinations of PEDOT:PSS + PDMS-contained conductive coating materials listed in Table S2, all failed when printed on the four textile substrates. For example, pristine PEDOT:PSS and PEDOT:PSS + IL could not be printed due to the observed mechanical cracks. This poor printability is due to the lack of flexibility and stretchability of PEDOT:PSS (30), which can be strengthened by applying biocompatible additives such as PDMS. In this scenario, we observed that PEDOT:PSS + PDMS-contained conductive coating materials without or with proper concentration of IL (e.g., ≤ 5%) could be printed on the flat textile substrates made of silver yarn. With further increase of IL concentration (e.g., 7.5%), this coating was observed not printable on the flat textile substrates made of silver yarn.
Impedance of Dry Textile Electrodes
Electrode impedance has been considered as one of the main performance metrics in evaluating electrodes for recording high quality biosignals (6, 27, 28). We evaluated the impedance performance of dry textile electrodes with and without conductive materials (listed in Table 1) screen-printed on them. An agar skin model was utilized to mimic human skin for consistency in the testing of the electrodes and the avoidance of intra- and inter-subject skin impedance variations (31).
Effect of Frequency and Coating Materials
The skin-electrode impedance of different dry textile electrodes at a widely accepted frequency range (e.g., 1 Hz ~ 10 kHz (6, 8)) is shown in the bode plots (Fig. 4a, b and Figure S3). All the tests were conducted under 20 mmHg (a typical pressure for wearable ECG measurement (32, 33)). The contact impedance decreased with the increase of frequency, which follows the typical trend of the electrical impedance measured on skin(34).
We then compared the skin-electrode impedance of dry textile electrodes with different conductive coating materials. Flat textile electrodes made of silver yarn were selected for comparison purposes since all conductive materials listed in Table 1 could be successfully printed on this type of textile substrate without cracks formation. As shown in Fig. 4a, the flat textile electrodes made of silver yarn with carbon-contained coating presented a much higher impedance than those with PEDOT:PSS-contained coating, showing a similar trend as previously reported (6). In addition, the impedance decreasing rate highly depended on the coating conductive materials. Dry textile electrodes with carbon-contained coatings showed a much larger impedance change than that with PEDOT:PSS-contained coatings (Fig. 4a). Dry textile electrodes with PEDOT:PSS-contained coatings showed a comparable impedance level with that of standard gel electrodes (Fig. 4a and d). This could be explained by two facts: a) carbon has a much higher sheet resistance than PEDOT:PSS (6); b) PEDOT:PSS can provide enhanced ionic-to-electric coupling effect, leading to a better impedance match with human skin (28). In addition, we found that the impedance of dry textile electrodes with PEDOT:PSS-contained coatings can be further reduced through the adding of CNT or ionic liquid (Fig. 4c). The CNT-induced enhancement can be because CNTs suppress the phase separation of PEDOT:PSS. Namely, the nanotubes establish electrical interconnections between the separate PEDOT:PSS (conductive phase) islands being dispersed in the insulating PSS-phase, thereby enhancing the electrical conductivity (35, 36). On the other hand, ILs as a secondary dopant can induce the formation of a nano − crystallized fibrillar structure of PEDOT chains with extended planar confirmations and reduced π − π interchain distances (37, 38). There is no impedance difference among PEDOT:PSS-contained coating with various concentrations of PDMS containing CNTs or IL (Fig. 4c).
Effect of different conductive substrates (silver or carbon) and knitted structures (flat or raised 3D): In this study, we selected the dry textile electrodes with carbon-contained coatings for investigation because PEDOT:PSS + PDMS-contained coatings cracked on all dry textile substrates (RS, FC, and RC) except the FS structure. These coatings can only be printed on FS structure. In contrast, carbon-contained coatings could be printed on both flat and raised 3D structures without cracks formation due to the strength of covalent bonds between carbon atoms. All tests were conducted at 5 Hz because the ECG spectrum has been reported to present optimal signal amplitudes at the frequency of 4–5 Hz (39). All carbon-contained coatings showed a similar impedance level on those four types of electrode structures; only the FS structure with (carbon + 5%IL) coating showed relatively high value of impedance (Fig. 5a) but still within the acceptable impedance range (e.g., ≤ 10 kΩ ) of dry electrodes (6).
Dry textile electrodes have also shown a huge potential for epidermal electroencephalography (EEG) and electromyography (EMG) recording, in which tests were clinically conducted at higher sampling frequency ranges (e.g., 1 kHz) (34). Therefore, we investigated the impedances of dry textile electrodes at the frequency of 1 kHz as well. As shown in Fig. 5b, all carbon-contained coatings showed similar impedance levels on FS, RS, and FC structures. Only carbon-contained coatings on the RC structure showed a relatively higher impedance (Fig. 5b).
Effect of Pressure
Dry electrodes have shown great potentials for getting high-quality epidermal biopotential signals, including ECG, EMG, and EEG, in various conditions such as dry and wet skin, and during body movement (28). Since dry textile electrodes do not have an adhesive, a proper pressure must be applied to maintain the skin-electrode contact to improve the quality of recorded biopotential signals (40). Therefore, we investigated the effect of pressure on skin-electrode impedance. We selected flat textile substrates made of silver yarn since all the conductive coating materials listed in Table 1 could be successfully printed on this type of substrate without cracks. Pressures of 10 mmHg, 20 mmHg, and 30 mmHg were chosen, as they are within the optimal pressure range for compression applications of electrodes (32, 33). We found that the pressure had little effect on the impedance of dry textile electrodes with different conductive coating materials (Fig. 6). Thereafter, the pressure of 20 mmHg was chosen for the following ECG measurement. This pressure has been shown to deliver the best signal fidelity for ECG applications (40).
Long-term impedance stability: The impedance stability performance of the dry electrodes over time is another important concern for the real application of smart garments (34). We evaluated the deterioration of the dry textile electrodes by quantifying their impedance change within one month. The dry textile electrodes with IL/CNTs-contained coatings maintained their impedance well (Fig. 7). In contrast, dry textile electrodes with pure carbon coatings or PEDOT:PSS + PDMS coatings showed obvious impedance increases.
Demonstration of ECG quality
We evaluated the ECG quality on human skin using the electrodes shown in Fig. 7. Here, we selected dry textile electrodes in FS structure with carbon + 7.5% IL coating, FS structure with PEDOT:PSS + 12.5% (PDMS + 1% CNT) coating, and FC structure with carbon + 7.5% IL coating for ECG monitoring due to their relative good long-term stability (Fig. 7). The electrodes were placed on the subject’s arm as shown in Fig. 8a, along with two commercial disposable Ag/AgCl electrodes as the gold standard simultaneously recorded as the second channel, and a third Ag/AgCl electrode as the driven ground electrode. The pressure was applied at 20 mmHg through a stretchable compression band and measured with a PicoPress® sensor. Although the dry textile electrode in FS structure with PEDOT:PSS + 12.5% (PDMS + 1% CNT) coating showed good long-term stability and an extremely low impedance (~ 50 Ω, Fig. 7b), it did not present the best ECG signal quality (Fig. 8d). In contrast, the dry textile electrode in FS structure with carbon + 7.5% IL coating showed a relatively high impedance (~ 1500 Ω, Fig. 7a), but it presented a similar ECG spectrum (Fig. 8c) as the standard gel electrode (Fig. 8b). Conversely, the dry textile electrode in FC structure with carbon + 7.5% IL coating showed a relatively high impedance (~ 1500 Ω, Fig. 7a), but it presented an ECG spectrum (Fig. 8e) with poor signal quality. In addition, we selected dry textile electrodes in FC structure with carbon coating as the representative electrodes without good long-term stability. We noticed that with a high impedance (~ 3000 Ω, Fig. 7a) and a poor long-term stability (with a ~ 75% impedance increase in 30 days, Fig. 7a), the dry textile electrode in FS structure with carbon coating showed a better ECG spectrum (Fig. 8f) than the FS structure with PEDOT:PSS + 12.5% (PDMS + 1% CNT) coating (Fig. 8b).
We further evaluated ECG signals based on the signal quality metrics of pSQI values, R-peak amplitudes, peak-to-peak amplitudes, and SNR values (Fig. 9). Although the impedance of dry textile electrodes varies based on the coating conductive materials, their pSQI values, R-peak amplitudes, and peak-to-peak amplitudes are all relatively stable and independent of the contact impedance. The relatively similar pSQI values indicate that the QRS complex was well-captured by all dry textile electrodes. We found that the dry textile electrode in FS structure with PEDOT:PSS + 12.5% (PDMS + 1% CNT) coating has a comparable impedance level to the standard gel electrodes but a significantly lower signal to noise ratio (SNR) value. In contrast, the dry textile electrode in FS structure with carbon + 7.5% IL coating showed a significantly higher impedance level than the standard gel electrodes but a comparable SNR value (Fig. 9). Therefore, in dry contact textile electrodes, lower impedance was not correlated to improvements in ECG signal performance, which is contrary to previous findings (6, 27, 28).