Wearable Electronics and Smart Textiles: A Critical Review

2.1. Conductive Fibers

Initially, conductive threads were mainly used in technical areas: clean room garments, military apparel, medical application and electronics manufacturing [34]. Textile structures that exhibit conductivity or serve an electronic or computational function are called electro-textiles [35]. They can have a variety of functions, like antistatic applications [36], electromagnetic interference shielding (EMI) [37], electronic applications, infrared absorption or protective clothing in explosive areas [38].

The conventional process to produce metal fibers is wire drawing, a mechanical production process. This process is characterized by its various drawing steps, called coarse, medium, fine and carding train (Figure 2).

The drawing die, used to draw the fiber, consists of a steel mount with a core out of ceramics, carbide or diamond. The initial diameter of the metal wire varies depending on the material. For copper, for instance, it is usually is 8 mm, while for iron it is 5 mm. After drawing, the wire is annealed at temperatures ranging between 600 and 900 °C. Subsequently, they are quenched. The fine metal wire is then wrapped onto a revolving wire drawing cylinder [39].

2.2. Treated Conductive Fibers

Instead of attaching electronics to textile substrates, the yarns of the textile can be functionalized with electronics. Electrically conductive fibers can also be produced by coating the fibers with metals, galvanic substances or metallic salts. Coatings can be applied to the surface of fibers, yarns, or even fabrics to create electrically conductive textiles. Common textile coating processes include electroless plating, evaporative deposition, sputtering, coating the textile with a conductive polymer [42].

In [43] a method to fabricate fibers with different material layers and material structuring is presented. The fabrication process is based on the conventional preform-based fiber-processing, easily yielding kilometers of functional fiber during the process.

Another relevant work is to use the crossing yarns in a textile to fabricate a transistor [44,45]. A schematic of a yarn-based transistor is shown in Figure 4. The resulting transistor shows an on-off current ratio of more than 1000, operated with a gate voltage of 1.5 V.

Figure 3 represents two yarns coated with PEDOT:PSS, one serving as the gate contact for the transistor while the second serves as drain and source contact. At the crossing of the two yarns, an electrolyte is placed. A redox process at the interface between electrolyte and PEDOT:PSS turns the transistor on and off [44].

2.3. Conductive Fabrics

There are different ways to produce electrically conductive fabrics. One method is to integrate conductive yarns in a textile structure, e.g., by weaving. However, the integration of conductive yarns in a structure is a complex and seldom a uniform process as it needs to be ensured that the electrically conductive fabric is comfortable to wear or soft in touch rather than hard and rigid. Conductivity can be established with different thread types (Figure 5):

However, woven fabric structures can provide a complex network that can be used as elaborated electrical circuits with numerous electrically conducting and non-conducting constituents, and be structured to have multiple layers and spaces to accommodate electronic devices.

Researchers at the Electronics Department and the Wearable Computing Laboratory at the ETH in Zürich produced a plain woven textile structure consisting of polyester yarns that are twisted with one copper thread. Initially, they started with a standard design (Figure 6a), then the researchers design a hybrid fabric called PETEX (Figure 6b) [47]. It consists of woven polyester monofilament yarn (PET) with diameter of 42 μm and copper alloy wires with diameter 50 ± 8 μm (AWG 461). Each copper wire itself is coated with a polyurethane varnish as electrical insulation. The copper wire grid in the textile features a spacing of 570 μm (mesh count in warp and in weft is 17.5 cm⁻¹).

With the PETEX the ETH researchers introduced a new approach to Smart Textiles and in particular a new manufacturing method. The aim was the possibility to realize a custom textile circuit (Figure 7). The wiring structure among circuit components is established by connecting the fabric embedded copper wires. Cuts must be placed at specific locations in the wiring in order to avoid short-circuits between copper wires. In particular, the procedure is as follows [47]:

• coating removal on copper wires at defined intersections with laser ablation;

• cutting the wires avoiding the signal leakage with laser;

• creating the interconnection with a drop of conductive adhesive;

• adding mechanical and electrical protection with an epoxy resin deposition.

2.4. Conductive Inks

Interactive electronic textiles can also be produced by using conductive inks. First of all conductive inks must contain an appropriate highly conductive metal precursor such as Ag, Cu, and Au NPs and a carrier vehicle. Most of them are water based: water is the main ink component and to limit contaminants, it must be as pure as possible. These specialized inks can be printed onto various materials, among them textiles, to create electrically active patterns. Screen printing also makes integration with planar electronics simpler than with conductive yarn systems.

There are several technologies that can print conductive material on different substrate. Sheet-based inkjet and screen printing are best for low-volume, high-precision work.

Inkjets are flexible and versatile, and can be set up with relatively low effort [51]. Inkjets offer lower throughput of around 100 m²/h and lower resolution (ca. 50 μm). It is well suited for low-viscosity, soluble materials like organic semiconductors. With high-viscosity materials, like organic dielectrics, and dispersed particles, like inorganic metal inks, difficulties due to nozzle clogging occur. Because ink is deposited via droplets, thickness and dispersion homogeneity is reduced. Simultaneously using many nozzles and pre-structuring the substrate allows improvements in productivity and resolution, respectively [52].

For inkjet printing, the inks should respect the following requirements [53]:

▪

high electrical conductivity;

▪

resistance to oxidation;

▪

dry out without clogging the nozzle during printing;

▪

good adhesion to the substrate;

▪

lower particle aggregation;

▪

suitable viscosity and surface tension.

Inks may also contain additives which are used to tune ink properties or to add specific properties thus increasing its performance [54].

After inkjet printing of a metal NP-based ink, in order to form a conductive printed pattern, particles must be sintered to create continuous connectivity between them and obtain electrical percolation. Sintering is the process of welding particles together at temperatures below the corresponding bulk metal melting point, involving surface diffusion phenomena rather than phase change between the solid and the liquid [55]. For instance, with inks based on gold NPs (1.5 nm diameter), the melting temperature was experimentally found to be as low as 380 °C, while for inks based on silver NPs (15 to 20 nm in diameter), a complete sintering can be obtained down to 180 °C [54,56].

Screen printing is appropriate for fabricating electrics and electronics due to its ability to produce patterned, thick layers from paste-like materials. The screen printing procedure, a stencil process, comprises the printing of a viscous paste through a patterned fabric screen and is usually followed by a drying process. The method can be applied to flat or cylindrical substrates. Depending on the substrate materials and the requirements for the printed structures, a high temperature densification can also be necessary (organic substrate T < 150 °C. Glass, ceramic and metal substrate T > 500 °C) [57].

This method can produce conducting lines from inorganic materials (e.g., for circuit boards and antennas), but also insulating and passivating layers. Generally, the throughput is about 50 m²/h with a resolution lower than 100 μm. By optimizing of the process condition and material the resolution may decrease to 30 microns line/space on thin flexible substrates [58].

This versatile and comparatively simple method is used mainly for conductive and dielectric layers [59], but also organic semiconductors can be printed, e.g., for OPVCs and OFETs [60].

Recently the researchers of University of Southampton [61] have developed an innovative screen printed network of electrodes and associated conductive tracks on textiles for medical applications (Figure 9). A polyurethane paste is screen printed on to a woven textile to create a smooth, high surface energy interface layer and a silver paste is subsequently printed on top of this interface layer to provide a conductive track. Silver pastes have been printed on to non-woven textiles to create wearable health monitoring devices.

The researchers developed different bio-potential sensing systems with dry electrodes and conductive ink for signal traces in order to demonstrate that this technology could be used for biomedical application. In particular the application tests were: ECG, facial EMG (Figure 10) and forearm EMG.

2.5. Conductive Materials as Sensors

Conductive textiles that change their electrical properties as a result of the environmental impact can be used as sensors. Typical examples are textiles that react to deformations such as pressure sensors, stretch sensors and breathing sensors. On the other hand, with smart textiles we have the further possibility to make bio-potential sensors.

2.6. Planar Fashionable Circuit Board (P-FCB)

The P-FCB is one of the new technologies allowing implementation of a circuit board on a plain fabric patch for wearable electronics applications. It features a soft and flexible impression just as normal clothes.

The P-FCB substrate is fabricated using woven fabrics. The planar electrodes are deposited on the fabric patch directly by silk screening of conducting epoxy or by gold sputtering. First, the circuit board is silkscreen printed on the fabric patch. Then the IC is placed on the fabric and wire-bonded to the patterned electrodes. Finally, the IC is molded with non-conductive epoxy (Figure 16a) [89].

With these techniques, KAIST researchers also developed a multilayer circuit (Figure 16b). The previous research highlight how electrical and mechanical characteristics and valuable system design parameters are obtained, such as maximum power consumption, maximum current density, crosstalk between neighboring two lines, and durability [90].

2.7. Wearable Antenna

Thanks to the rapid progress on the fabrication of conductive textiles, a significant development of wearable antennas has begun, exploiting new flexible and conformable smart structures [91].

An antenna is essential, if the purpose is to develop a wearable and autonomous system. It allows one to transfer information from the sensors hosted inside the garment to a control unit or to monitor other electronic parameters.

A wearable antenna is thus the bond that integrates clothes into the communication system, making electronic devices less obtrusive. To achieve good results, wearable antennas have to be thin, lightweight, low maintenance, robust, inexpensive and easily integrated in radio frequency (RF) circuits. Planar structures, flexible conductive and dielectric materials are specific requirements for wearable antennas [92,93].

Several properties of the materials influence the behavior of the antenna. For instance, the bandwidth and the efficiency of a planar microstrip antenna are mainly determined by the permittivity and the thickness of the substrate [94,95]

In general, textiles present a very low dielectric constant that reduces the surface wave losses and increases the impedance bandwidth of the antenna. Therefore it is important to know how these characteristics influence the behavior of the antenna in order to minimize unwanted effects.

Another issue regards the movement of the body that can deform the spatial geometry of the antenna and affect its performance. When the textile fabric adapts to the surface topology it bends and deforms, causing changes to its electromagnetic properties and thus influencing the antenna performance [96]. Thus, a wearable FM antenna should be designed so as to be wider than the FM broadcast band (about 81–130 MHz) not to suffer from the detuning caused by the human body [97].

To summarize, the guidelines for a correct antenna project are shown below:

• choosing the correct positioning of the textile antenna [98,99];

• the textile antenna must be made with an accurate thickness stacking the different fabrics [99];

• the geometrical dimensions of the patch must remain stable [100];

• the connections between the layers must not affect the electrical properties and the connections with other parts of “e-garments” have to be stable and robust [101–103].

If some of these points are not followed, undesired effects may occur to the functioning of the device.

2.8. Stretchable Interconnection

Deformable electronic circuits, in particular in bio-medical application, are needed. E-textiles are the candidate technology for this scope but an important aspect regards the interconnections between components and devices. In most cases the approach used is to develop sinuous electroplated metallic wires in a stretchable substrate material [107].

Figure 17 presents a stretchable interconnection structure. This configuration, called “horseshoe-shape”, is able to accommodate large deformation in response to a mechanical stress, preserving the electrical properties [108].

Indeed, with high-frequency signal the interconnection have to be stretchable and therefore the substrate and the conductors have to be stretchable as well. Usually, a polymeric material is chosen as encapsulating substrate, because it can be stretched and also because it may be bio-compatible [109].

Interconnections made with these techniques coupled with conductive yarns can go beyond actual issues regarding the lack of robustness in e-textiles circuit.

3.1. Conductive Wires

Especially for clothing, tactile properties such as stretch, recovery, drape, shear and handle are quite important. For this reason the fibers that are used should be fine and fabrics should have a low weight per unit area (not more than 300 g/m²). These demands are inconsistent with the materials and geometries that are needed for a reasonable electrical conductivity, because the incorporation of elements such as metal wires within textiles increases stiffness and reduces elasticity [110].

Often in literature no distinction is made between metal wires and metal fibers. However, the company Sprint Metal (Hemer, Germany) defined metal fibers and wires according to their diameter. While a fine wire has a diameter between 30 μm to 1.4 mm, a metal fiber possesses a diameter of 2 to 40 μm [111].

The advantages of metal fibers are their strength, composition, biological inertness and ready availability in textile form at low costs. Due to its inertness it is not sensitive to washing or sweating.

However, they cannot provide uniform heating and their brittle characteristics can damage spinning machinery over time. Additionally, they are heavier than most textile fibers making homogeneous blends difficult to produce [112].

Using the ’conductive thread’ approach, no additional step after manufacturing of the fabric is required to establish conductivity. The conductivity of these conductive threads lies in the range of 10–500 Ω/m [113].

For instance, Figure 4a shows the copper wire twisted around the yarn and their locations are not precisely defined within the fabric because of their helical path around the yarn [114]. The conductance of a wire in the fabric features about 17.8 Ω/m.

3.2. Treated Coatings

The advantage of coatings is that they are suitable for many fibers types and they produce good conductivity without significantly altering existing key substrate properties such as density, flexibility and handle. Nevertheless adhesion between the metal and the fibers as well as corrosion resistance can lead to problems.

It is possible to make a conductive fabric through a coating treatment even though the main applications for these textiles are protection against static electricity charge and electromagnetic interference (EMI). The surface resistivity (Rs) of normal textiles goes from more than 10⁵ Ω/sq down to antistatically finished textiles and conductive textiles for EMI and heating applications, where Rs can be lower than 100 Ω/sq. It is usual to incorporate or impregnate organic antistatic agents [98].

At present the most important ways of producing conductive fibers are:

• incorporation of conductive fillers (carbon black, metal wires, graphite and metal powder or flakes of Al, Cu, Ni, Ag).

• blended fabrics using conductive metallic or polymeric fibers (such as polypyrrole, polyaniline) (Figures 16b and 18a) [115].

However, the surface resistance of polypyrrole-coated (C_pyrrole) polyester fabrics is another method to functionalize the textiles. Rs decreased marginally from 106 Ω/sq to 103 Ω/sq with increasing pyrrole concentration within 0.2 mg/mL until the concentration reached a value of about 0.4 mg/mL, above which the rate of decrease diminished. The decrease in surface resistance is due to the increase in the thickness of the conducting polymer layer [116].

The main advantage of functionalized yarns is that the textile properties of the yarns are preserved, in particular they remain flexible.

3.3. Conductive Fabric PETEX

Normally, a fabric made with conductive fibers has a structure according to Figure 19. If the aim is to exploit the conductive fibers to create an electrical circuit, we must take into account the irregular shape of the fabric. It is not possible to obtain a geometric repeatability of the position of the contact points between the conductive fibers.

The solution adopted by the ETH researchers is to develop a synthetic fabric and it consists of woven polyester monofilament yarn (PET) with diameter 42 μm and copper alloy wires with diameter 50 ± 8 μm (AWG461). Each copper wire itself is coated with a polyurethane varnish as electrical insulation. The wires in warp direction possess a higher tensile strength compared to the wires in weft direction. These differences support a reliable manufacturing process and minimize breakage.

Figures 18b and 20a below, show the differences between a normal electronic textile structure and PETEX [47].

3.4. MIT CAD Embroidery

This technique allows precisely specifying the circuit layout and stitching pattern in a computer-aided design (CAD) environment, from which any number of articles can be sewn under machine control. This process also allows control and integration of yarns with different electrical properties, for instance, different resistances. Embroidery offers advantages over knitting or weaving. Conductive thread and yarn embroidery can be accomplished on single or multiple layers of fabric or can be applied on various types of textile and apparel products in one step [39].

3.5. Conductive Inks

Direct printing of conductive tracks on a fabric is a versatile technique but sometimes shows its limits. Most conductive inks and pastes are based on silver filler and suffer from brittleness. During printing, one of the limits of the silver paste is the thickness. For instance, several passes are needed to achieve a layer thick 40 μm and the electrical resistance depends on the number of passes applied (Figure 21).

Another problem is the mechanical resistance. If you create a condition as shown in Figure 22a, the risk is to have a break in the conductivity. In particular the electrical conductivity changes both for the number of passes and for the number of creasing interactions (Figure 22b) [47].

The researchers of North Carolina Institute, developed a new method to control the durability of the printed circuits, which includes coating of the printed lines. The viscosity of the ink dictated the performance of the printed media during washing trials.

The results showed that higher viscosity inks stay on the surface, resulting in better short-term results. (Figure 23) The ink used was produced by Creative Materials and showed rather a smooth and lower impedance profile than other commercial conductive inks. This ink is based on silver particles and the lines printed on Evolon^® show 42 times higher resistivity than standard PCBs (0.50 mΩ/m). However, Evolon^® like structures with the highest viscosity ink represent the best solution in terms of durability and conductivity also after washing. This substrate-ink combination protected by a TPU shows the best compromise to become a low-cost electronic textile solution for wearable computing interconnects [117].

3.6. Stretchable Sensors

Usually, textile materials composed of fibers form complex networks of conducting parts that make multiple contacts. During deformation a number of mechanisms take place:

• the number of contact points changes;

• fibers are extended;

• fiber cross-section is decreased.

These factors listed above may compromise the stretch sensors qualities [118]. Polymers are interesting when developing strain sensors with large strain [1]. The stretch sensor developed by Mattmann et al. [119] consists of a mixture of 50 wt-% thermoplastic elastomer (TPE) and 50 wt-% carbon black particles and is fiber-shaped with a diameter of 0.315 mm. This sensor configuration was characterized using a strain tester and measuring the resistance (extension-retraction cycles): it showed a linear resistance response to strain, a small hysteresis, no ageing effects and a small dependence on the strain velocity. The attachment of the sensor to the textile is realized using a silicone film.

The graphs (Figure 24) below show the behavior of the sensor and the typical resistance vs. time plot. The sensor was cycled between 0% and 80% strain at a speed of 200 mm/min and waiting times at minimal and maximal strain of 2 min. In the lower plot the measured resistance is shown which varying between 2 kΩ and 19 kΩ. Instead, Figure 25 shows a linear rise in resistance when applying strain and only a small hysteresis.

Commercially available strain gauges reach a higher than 80% linearity but at a reduced working range of less than 1%. The sensor presented above has a high sensitivity of 1.25 kΩ/mm with a stable sensor properties while in continuous usage. Furthermore the sensor is washable.

A similar approach is that used by Shyr et al., in which they developed three different geometries of a strain-resistance sensors using Lycra fiber and carbon coated polymide fiber: flat, tubular and belt webbings. It was found that tensile hysteresis and contact resistance significantly influence the tensile elasticity and the resistance sensitivity of the webbings, respectively. Their results showed that all the webbings had a good linear elasticity relationship between tensile load and strain when stretched with the aim to reduce the tensile hysteresis, decreasing the friction among the conductive yarns [118]. The stability of resistance sensitivity and the elasticity are two relevant properties for e-textiles when used as a strain-resistance sensor.

3.7. Pressure Sensors

The pressure sensor developed by Meyer et al. [9] has a spatial resolution of 2 × 2 cm and an average error below 4% within the measurement range 0 to 10 N/cm². It is a pressure sensor based on capacitive behavior. The measured capacitance of a single electrode of the prototype with the textile spacer varies from 3.5 pF uncompressed to 5.8 pF at a pressure of 5 N/cm² for the textile spacer with 6 mm thickness. Figure 26 shows the capacitance vs. pressure with a hysteresis up to 30%. The high value of hysteresis is due to the dielectric material interposed between the two conductive faces capacitor.

An ideal pressure sensor must ensure repeatable measures over time with a low hysteresis value. Therefore, the textile pressure sensors are applicable in areas where high accuracy is not required.

3.8. Electrochemical Sensors

One of the common disadvantages of chemical/biochemical sensors is the need to calibrate the response from time to time due to either small changes in the activity of the sensor itself and/or the contamination of the sensor surface during operation. This has seriously limited the application of these sensors in continuous monitoring scenarios for long time. The major obstacles still remain, preventing the realization of practical, robust wearable sensors suitable for routine use by the general population.

Among these limitations, we have the lack of appropriate methods to seamlessly integrate electronics and wireless transmitters into the sensor package. The graph below (Figure 27) shows the performance of the wearable sweat sensor during an exercise trial. Reference pH measurements are taken along with heart rate and breathing rate data. The analysis were taken by placing the electrode on a reference fabric patch, as performed in previous studies [120].

3.9. Power Supply

In the following Table 4, an overview of different possible energy sources and the amount of energy that they can generate is given [120,121].

From these data, we can assume that the energy harvesting coming from our body is very small although it is always present. If the wearable technologies will take advantage of these energy sources, they must feature a low electrical consumption and a high efficiency harvesting system.

3.10. Planar Fashionable Circuit Board (P-FCB)

The measured sheet resistance of the conducting film on the P-FCB is 44 mΩ/m, as determined by the four-point probe method. Although no noticeable degradation has been observed, even after 50 laundry cycles, more durability examinations will be required with an automatic testing system.

The resistance measurement results of the proposed via and conductive adhesive and bandwidth of a P-FCB transmission line 15 cm long and 1mm wide are shown in Figures 26b and 28a. Its bandwidth is 80 MHz, and this is sufficient to deal with bio-signal processing. The analysis were performed with a vector network analyzer [122]. To get an average value, as many as 100 connections were formed and measured. The resistances of the proposed via and conductive adhesive measured 0.24 W and 0.34 W, respectively [123].

Mechanical strength of the P-FCB was tested through washing tests; after 50+ cycles, the electrical characteristics change was negligible [123]. As a result, the monitoring shirt can be easily made at low cost, and the wearability and flexibility are improved.

3.11. Wearable Antenna

Antennas for wearable applications need to be comfortable to wear and traditional technology with rigid substrates is not well suited to this.

In this sector the dielectric properties of the material that make up the antenna are more relevant. A wrong choice of material would compromise the performance of the device.

The constitutive parameter of dielectrics is the permittivity, ε, a complex value parameter and it usually expressed by Equation (1) below:

Table 5 below shows the dielectric constant ε ′_r and the ratio of the imaginary part to the real part, also known as tan δ = ε″_r/ε′r [94].

The low dielectric constant (between 1 and 2) reduces the surface wave losses that are tied to guided wave propagation within the substrate. Therefore, lowering the dielectric constant increases spatial waves and hence increases the impedance bandwidth of the antenna [125].

In addition, the moisture can affect the dielectric behavior of textiles. In particular, when water is absorbed by the fibres, an increase of the dielectric constant and loss tangent occurs. A textile cover or superficial treatment may then provide a sufficient protection from the humidity and varying climatic conditions [100].

Researchers of Tampere University of Technology, in collaboration with Patria, developed a robust textile antenna able to operate in different environmental conditions. They demonstrated that different functional textile layers can improve and optimize the antenna performance. In this occasion they succeeded to protect the antenna against moisture and others environmental factors [126].

The thickness of the dielectric material changes the bandwidth of the planar antenna. It is necessary to find a compromise between efficiency and bandwidth, well as to influences the geometric sizing of the antenna [125].

For the antenna design, the relevant parameter is the conductivity of the fabric (σ), which unit is Siemens per meter (S/m). It is expressed by this Equation (2):

Another aspect that influences the conductivity is the structure of the fabric. The current flow depends on the type of textile structure used and to minimize the conductive losses it is better to align the conductive paths with the current direction [98].

Finally, a correct choice of materials combined with textile structures can improve the performance of the textile antenna equaling the features guaranteed by traditional antennas.

3.12. Stretchable Interconnections

Stretchability of the electrical conductors is a requirement that can be satisfied with different designs. The design optimization is done using FEM modeling (Figure 29) and other tests. Gonzalez et al. evaluated the “horseshoe-shape” to be the optimal solution for this purpose [108]. In this way the damage in the conductor is significantly decreased with a larger deformability. They demonstrated that in case of multiple and narrow metallisations the stretchability of the circuit reaches 100% of the deformability.

Hu et al. used a different approach to obtain a stretchable conductive textile. With a simple “dipping and drying” process using single-walled carbon nanotube (SWNT) ink, they produced highly conductive textiles with conductivities of 125 S·cm⁻¹ and sheet resistances less than 1 Ω/sq. In this way, such conductive textiles preserve the same stretchability of the normal fabric. Indeed, the porous structure of the textiles facilitates accessibility from any electrolyte and such porous and stretchable conductors find a wide range of applications in e-textiles field [128].