Engineering Graphene Flakes for Wearable Textile Sensors via Highly Scalable and Ultrafast Yarn Dyeing Technique

Multifunctional wearable e-textiles have been a focus of much attention due to their great potential for healthcare, sportswear, fitness, space, and military applications. Among them, electroconductive textile yarn shows great promise for use as next-generation flexible sensors without compromising the properties and comfort of usual textiles. However, the current manufacturing process of metal-based electroconductive textile yarn is expensive, unscalable, and environmentally unfriendly. Here we report a highly scalable and ultrafast production of graphene-based flexible, washable, and bendable wearable textile sensors. We engineer graphene flakes and their dispersions in order to select the best formulation for wearable textile application. We then use a high-speed yarn dyeing technique to dye (coat) textile yarn with graphene-based inks. Such graphene-based yarns are then integrated into a knitted structure as a flexible sensor and could send data wirelessly to a device via a self-powered RFID or a low-powered Bluetooth. The graphene textile sensor thus produced shows excellent temperature sensitivity, very good washability, and extremely high flexibility. Such a process could potentially be scaled up in a high-speed industrial setup to produce tonnes (∼1000 kg/h) of electroconductive textile yarns for next-generation wearable electronics applications.


S1. Methodology
represents the process sequence in order to manufacture graphene-based textile sensors in a scalable quantity. We engineer the graphene flakes and optimise the processing parameters in order to find the best formulation for our applications. We then use laboratory scale yarn dyeing machine in order to coat a batch of 100% cotton yarn. This could readily be scaled up into an industry scale yarn dyeing machine that could potentially produce tonnes (~1000 kg/hr) of graphene-based conductive yarns. We then integrate graphene coated yarn into knitted fabric structure which could then be connected to a self-powered RFID or a lowpowered Bluetooth in order to transmit the data wirelessly. The wearable garments thus produced and integrated with graphene textiles sensors can monitor the body temperature wirelessly and send to a mobile device.

Figure S1| Graphical representation. Engineering graphene flakes to wearable textile sensors via highly scalable and ultra-fast yarn dyeing technique. Illustration by Kazi Farhan
Hossain Purba and used with permission from the artist. 3

S2. Graphene Materials Synthesis and Characterisation
One of the major challenges with graphene's commercial development is its poor colloidal stability in common solvents. The dispersibility of the graphene derivatives could be arranged as follows: GO > rGO > G. 1 As GO is highly functionalized, therefore it provides higher yield, good colloidal stability and excellent dispersibility. 2 After reduction to rGO, the dispersibility becomes almost the same as G; 2 however the fact is that rGO disperses more effectively in polar solvents than G 3 may be due to the presence of some residual functional groups even after reduction. The dispersibility of rGO could be improved by introducing an energy barrier to aggregation, through either covalent or non-covalent interactions. 4 We synthesize graphene oxide using a modified Hummers method and chemically reduce GO to rGO by modifying our previously reported methods. 5, 6 We use ascorbic acid (AA) and sodium hydrosulphite (SH) as reducing agents and functionalise the surface of rGO flakes using PSS/PVA to have better dispersibility and prevent agglomeration. 7 We also exfoliate graphene-based inks (G) using a highly scalable microfluidization technique. 8,9 We use sodium dexoycholate (SDC) as a surfactant in order to disperse graphene flakes in water through non-covalent bonding. show that the mean lateral dimension of GO is ~5.85 µm. After reduction with sodium hydrosulphite (SH) and L-Ascorbic Acid (AA), the flake size is reduced to ~4.86 µm, may be due to the stresses during pre-mixing and centrifugation steps in post-washing cycles. We also obtain lower flake size for G flakes (~1.45 µm), may be again due to the extremely high shear rate (10 8 s -1 ) it is subjected to during the microfludization of graphite.   Figure S5b.

S3. Stability of Graphene Dispersions
Although the reduction of graphene oxide is a popular approach, a major problem associated with this route is that rGO aggregates in aqueous solution due to its hydrophobic nature, modification by forming non-covalent bonds is typically preferred. They are comparatively weaker than covalent bonds but multiple non-covalent bonds working in harmony can yield highly stable modifications. Moreover, these non-covalent bonds are easy to achieve over the entire graphene surface. On the other hand, covalent bonding is commonly preferred when stability and the strong mechanical properties of modified graphene are required.
Stankovich et al. 11 reported that the stable rGO dispersion can be prepared by the surface modification using poly (styrenesulfonate) (PSS). Since then, a significant amount of works [12][13][14][15] has been carried out to uncover the prospect of functionalised rGO. Wei et al. 12 used several modification media such as poly(styrenesulfonate) (PSS), polyelectrolyte, polyaniline and ionic liquid to prepare high quality, free-standing graphene films. The

S4. The Degree of rGO Reduction -Raman and XPS Analysis
We reduce GO to rGO chemically using Na 2 S 2 O 4 and L-AA. Figure S8 shows the change of the colour of GO dispersions to opaque black almost immediately, which is not the case for L-AA. This confirms faster and efficient reduction of GO with Na 2 S 2 O 4 , which is in agreement with previous studies. 16 All sp 2 bonded carbon show a common Raman shift where the G peak is at ~1580 cm -1 and the D peak is at around ~1360 cm -1 . 18 The D peak is the disorder band or defect band that shows a spectrum due to the breathing modes of six-atom rings and observed when defects are present. 19 Moreover, the increment in the intensity ratio (I D /I G ) indicated the generation of large number of sp 2 domain and formation of rGO from GO. 15 Table S1 shows mean of I D /I G ratio of rGO formulation that could be used to observe the degree of reduction with various reduction time and reducing agents.    Table S1 shows that a better reduction can be achieved (maximum I D /I G ~ 1.73) by using

S5. Temperature Dependence of rGO Resistances and I-V Curves
As discussed earlier, we make 48 formulations in order to optimize the GO reduction for both AA and SH using various reduction time (12,24,48, 72 hrs) and four stoichiometric ratios (1:0, 1:1, 1:5 and 1:10) between GO and polymers (PSS/PVA). We drop-cast graphene dispersions on Si/SiO 2 wafer using a micropipette and dry them at room temperature.
We then measure the conductivity of graphene inks from I-V curves using a 2-probe measurement at various temperature ranges (150K to 280K), Figure S11    As evident from the wash stability tests, GO provides better durability than G flakes due to the presence of residual functional groups, Figure S9 surface after few washing cycles. This may be due to the absence of the residual functional groups in G flakes as seen from the high resolution XPS spectrum. Figure S16c shows large amount of G flakes on the fibre surface, which are removed after 10 washing cycles with only few flakes left on the surface, figure S16d.