Imperceptible augmentation of living systems with organic bioelectronic fibres

The functional and sensory augmentation of living structures, such as human skin and plant epidermis, with electronics can be used to create platforms for health management and environmental monitoring. Ideally, such bioelectronic interfaces should not obstruct the inherent sensations and physiological changes of their hosts. The full life cycle of the interfaces should also be designed to minimize their environmental footprint. Here we report imperceptible augmentation of living systems through in situ tethering of organic bioelectronic fibres. Using an orbital spinning technique, substrate-free and open fibre networks—which are based on poly (3,4-ethylenedioxythiophene):polystyrene sulfonate—can be tethered to biological surfaces, including fingertips, chick embryos and plants. We use customizable fibre networks to create on-skin electrodes that can record electrocardiogram and electromyography signals, skin-gated organic electrochemical transistors and augmented touch and plant interfaces. We also show that the fibres can be used to couple prefabricated microelectronics and electronic textiles, and that the fibres can be repaired, upgraded and recycled.

Supplementary Video 1: Organic bioelectronic fibre tethering on different fingers Bioelectronic fibre tethering process on different fingers with a laser pointer being used to indicate the fibre deposition positions.The laser pointer is aligned with the plane of the needle and the rotating arm; thus, the laser spot would indicate the real time fibre deposition location.This could be helpful especially when repairing and upgrading the fibre structures ondemand, because the fibres are small and sometimes difficult to see with naked eyes.

Supplementary Video 2: Imperceptible fibre tethering
The compound leaves of Mimosa pudica plant would fold inward and droop when touched or shaken 1 ; thus, it is chosen as an exemplary target object to demonstrate that the fibre tethering process would induce minimal perturbance, and thus is imperceptible to the target objects.Bioelectronic fibres were first spun onto a leaf of the Mimosa pudica plant for around 90 seconds, the leaf did not fold or change in shape during the whole fibre deposition process.Afterwards, the same leaf was touched by fingers gently.Upon hand touching, the leaf folded immediately.

Supplementary Video 3: A chicken embryo's heartbeats with a fibre network
The day-3 chicken embryo was directly extracted from an egg, and the whole yolk with the embryo was placed inside a petri dish.Bioelectronic fibres were then spun onto the surface of the chicken embryo directly under room temperatures, and the process of fibre spinning lasted for around 5 minutes.Afterwards, the embryo was put back to a 37 °C incubator for 15 minutes before imaging the heartbeat.(Typical results from N>5 independent experiments on different embryos)

Supplementary Video 4: Fibre tethering and contact dynamics
This video shows the real-time fibre wetting process upon fibre tethering, where a bioelectronic fibre is spun onto a substrate made of a glass microscope slide with a round glass capillary placed on top.

Supplementary Video 5: Comparing fibre and commercial electrodes for fingertip-based ECG acquisition
This video shows the experimental configurations for acquiring ECG signals in real-time: Fibre and gel electrodes were placed onto the index and middle fingers respectively on the left hand as working electrodes, and a separate fibre electrode on the right hand index finger serve as a counter electrode.Supplementary Video 6: ECG monitoring with bioelectronic fibres under dry friction, humid, water-soaking, and mild heat conditions Clear ECG signals were measured from the bioelectronic fibres on a fingertip during dry friction wear, water soaking, ~ 90 % relative humidity, or mild heat at ~ 40 °C environments.For dry friction, a normal force of ~ 2 N was applied, and the surface of the plastic roller was moving at a speed of 4 cm/s.A reference friction experiment was presented where the simultaneous ECG measurements from the fibre electrode under friction were compared to unperturbed fibre and gel electrodes.For the mild heat environment, the fingertip with bioelectronic fibres was placed ~ 30 cm in front of a 100-watt infrared lamp.No apparent distortion to the fibre patterns or signal deterioration was observed during the above wearing processes.Supplementary Video 7: Augmented mist sensing on a dandelion An array of bioelectronic fibres could be tethered onto a dandelion to monitor environmental mist interactions.During the fibre spinning, a bug living on the dandelion is seemingly not disturbed.Both ends of this bioelectronic fibre array are connected to a multimeter, which monitors the resistance change of the fibres.When a water mist flows over the dandelion, it could trigger the resistance increase of the bioelectronic fibres.
Supplementary Table 1 | Comparing the imperceptible fibre augmentation with state-of-the-art on skin sensors and fibre printing techniques.The test condition is highlighted when the durability tests were performed in situ on the living skins of the target objects, because such test conditions best characterise the stability of the devices during actual application scenarios.Upon device attachment/deployment onto living skins, the living objects' posture and movement could influence the outcome of the device interface.During the wearing, living skins would naturally produce sebum and undergo trans-epidermal water loss, which could also influence the device performance.The device life cycle sustainability refers to the environmental impact of a device throughout its lifetime; it is evaluated by considering the criteria of 'Eco-material' (i.e.use of earth-abundant, biologically-derived or degradable raw materials), 'Reduce' (i.e.minimising total embodied material, energy usage and emissions during fabrication), 'Repair' (readily repairable or reconfiguration for other purposes), and 'Recycle' (of extracted materials, or re-use material constituents or components at the end of the device service life).The supporting data leading to the electromechanical performance summary are highlighted in grey.This criterion is set by considering human sensory system and functions of skin.In brief, a network/mesh is considered as fully skin imperceptible if it simultaneously fulfils the conditions of: (1) network/mesh opening between fibres (openness) greater than ~ 50 µm (c.f., the sweat gland pore size), but smaller than 1 mm (c.f., the fingertip receptor field); (2) width of individual fibres and thickness of the network/mesh smaller than ~ 10 µm (such that individual skin cells are mostly exposed through the open network, and the fingerprint ridge features are not compromised).
Supplementary Note 1: Estimation of power and materials consumption to create imperceptible augmentation devices via organic bioelectronic fibre tethering For the estimation of the power consumption of the base unit of the fibre spinning set-up, the fibre production and the solution feeding mechanisms are considered.For the fibre production mechanism, the rotating arm operates at 4-6 V with 120 mA current according to the manufacturer, corresponding to 0.6 W; and an Arduino UNO board was used to control the rotating arm, and the board usually operates with no more than 0.4 W 11 .For the solution feeding mechanism, the micro air pump operates at 45-60 mA according to the manufacturer, corresponding to 0.54-0.72W under 12 V voltage supply.Thus, one base unit of the fibre production and solution feeding mechanisms requires around 2 W power consumption.During the fibre spinning process, the flowrate of the solution feeding is adjusted at around 50-70 µL/hr (mean ~60 µL/hr), which corresponds to around 1 µL of solution per minute of fibre spinning.Considering typical bioelectronic fibre tethering for 2-5 minutes, corresponding to ~ 2-5 µL of solution being used.The solid matter concentration of all the solutions used is below 5 % (w/w); thus, the dry mass of the bioelectronic fibres is estimated to be ~ 0.1-0.3mg per device.When corresponding to the fibres used for per device, the fibre number density commonly used are ~ to for each device (NB. in most applications, the total fibre number used are ~100-300 fibres; and only in the OECT case, 15 fibres were used).It is to note that the solution usage and dry mass calculated here are the total embodied material usage (i.e., including solutions wasted within the needle and the rotating arm).Supplementary Note 2: Estimating the fibre deposition force Cantilever experiment was set up to characterize the dynamic fibre printing process.The cantilever was made of an elastic quartz micropipette-filling capillary (inner diameter of 100 μm and outer diameter of 164 μm; World Precision Instruments).The elastic behaviour of the cantilever (force at the free end tip versus tip deflection) had been characterized previously 12 .The cantilever deflection versus force could be estimated by Force(μN) = 9.46 × deflection (mm) where the force is causing deflection.During the experiment to characterise fibre deposition force, a single fibre was spun onto the free end tip of the cantilever, and the deflection of the tip was recorded through video and then measured that could be correlated to force.The bioelectronic fibre caused ~ 1.2 mm deflection of the cantilever, corresponding to less than 12 μN of deposition force.ii)-Theoretical analysis comparing the driving capillary force for spreading of a wet fibre versus the resisting elastic force Theoretical analysis is used to compare the driving capillary force for the spreading of a wet fibre versus the resisting elastic force.To conform to sub-millimetric reliefs, the fibre must wet the substrate so it can spread into the grooves.The driving capillary force for spreading of the fibre of radius R may be given by Fc ∼ γR, where γ is the liquid surface tension (assuming complete wetting).Concomitantly to spreading within a groove, say a "V-shaped" cut of depth h on the surface as shown in the figure below, the strain of the fibre could be shown as for it to reach a depth d (if the angle of the V-shaped groove is about π/3): So that the strain satisfies ε ∼ d/h.The dominating force resisting this elongation is thus elastic, Fe ∼ GR 2 ε, where G is the fibre's elastic (or storage) modulus.Balancing both forces, the fibre will reach a relative depth d/h ∼ γ/(GR).If γ ≫ GR, the fibre will completely conform to the jagged surface.On the other hand, the fibre will remain suspended if γ ≪ GR.
This analysis shows that the driving capillary force would dominate for the spreading of a wet fibre.Hence, for the bioelectronic fibre formulation, both experimental and theoretical evidence indicate that intimate contacts over hundreds of microns of topographical features are expected to form on convex and solid structures.
Supplementary Fig. 4 | Schematic illustration of the fibre wetting analysis model.

Supplementary Note 5: Evaluating the fibre array tensile properties different relative humidity environments
An array of parallel suspended fibres ( ~ ) was first deposited on a frame with a gap distance of 20 mm, and the frame was mounted onto a vertical translational stage (Thorlabs MTS50-Z8), moving at a constant speed of 50 μm/s upwards.The rising suspended fibre array was intercepted in the middle by a weight placed on a balance (Ohaus Scout Portable Balance, 120 g Capacity, 0.001 g Readability).The set-up is schematically shown in Supplementary Fig. 5a.The reading on the balance reflects the forced induced by the deformed fibre array.The strain of the fibre array could be calculated through: where D is the displacement of the vertical translational stage and L is the original length of the fibre array.The whole set-up is placed inside a humidity-controlled chamber for adjusting the relative humidity (RH).
The tensile strength of the bioelectronic fibres would weaken at high humidity levels (i.e., RH ~ 70 %), and this enables fibres at defined regions to be easily erased off on demand upon wetting (as exemplified in Supplementary Fig. 5d).Such mechanical remodel-ability further offers the possibility for conveniently remodel the in situ circuit patterning.For example, the tethered bioelectronic fibre circuit could be selectively erased off to create connection openings.
In comparison, the cellulose-based fibres possess almost 10 times higher tensile strength than the bioelectronic fibres, and their mechanical properties are insignificantly affected by environmental humidities for the duration of the measurements (~ 10 minutes).Thus, the cellulose-based fibres could be used as a protective fibre layer on top of bioelectronic fibres as an approach to enhance the environmental stability of the system (see examples in Supplementary Fig. 19).
dry versus wet conditions.In the dry erasing test, the cotton bud was used as it is, and in the wet erasing test, the cotton bud was soaked with water for erasing.Each erasing was conducted by moving the cotton bud on the fingertip back and forth for 5 cycles, with a pressing force of ~ 1 N.
Supplementary Note 6: Breathable skin-gated OECT The design of skin-gated OECT with an array of bioelectronic fibres is schematically shown in Supplementary Fig. 6a.The variable permeability of the human skin epidermis layer 15 , in combination with the ionic solution properties of interstitial fluid, allow gating circuitry to form across human body surface for transistors that rely on interfacial charge exchange to operate 16 .
In the context of OECT, the skin thus constitutes the "electrolyte" component for the OECT architecture 17 .Supplementary Fig. 6b shows the OECT responses versus the number of fibres in the channel with exponential fitting curves, and the time constants τ of the fitting curves are shown in Supplementary Fig. 6c (exponential fitting and time constant τ are generally used to characterise the OECT response behaviour 18 ).It is seen that the OECT response would slow down with increasing number of fibres in the channel, especially when the fibre number exceeds 20.Thus, normally a fibre array of ( = to ) is used as the channel.With ~ 15 fibres in the channel, the time constant τ of the OECT response behaviour is ~ 12 s, and the response time could be estimated in the range of 60 s, similar to other patch-like OECTs of centimetre channel length 19 .Such a skin-gated low-frequency OECT could be useful for future clinical applications in measuring quasi-static continuous biological signals with long fluctuation time in hours, such as for hydration and circadian rhythm monitoring 20,21 .
As for the operational safety considerations, the applied gate voltage does not exceed 0.5 V, and the gate and channel currents have been measured as shown in Supplementary Fig. 6d.With 0.5 V gate voltage, the maximum power through the channel does not exceed 20 μW, which would not cause overheating.The gate current that passes through the human body does not exceed 1 μA, which is 3 orders of magnitude lower than the safety current limit for human body (muscle let-go current) 22 .In addition, such voltage and current levels used in this work are much lower than commercial deep brain stimulation (2.5-3.5 V 23 ) and transcutaneous electrical nerve stimulation (up to 30 V 24 and 80-100 mA 25 ), and these devices usually work in long-term manners either implanted or directly interfaced with human skins.Finally, the channel current does not exceed 50 μA.The gate voltage, channel and gate currents of the skingated OECT in this work compare similar to or lower with those reported human-interfaced OECT devices from literature 16,26,27 .Supplementary Fig. 7 | Rheological and surface tension characterisations of the fibre solution with reference composite solutions to identify the 'spinnable' solution property region for the orbital fibre spinning technique.a, Shear viscosity obtained from steadyshear measurements.b, Creep shear modulus (G) extrapolated from shear creep measurements with 1 Pa of step stress applied.c, Storage modulus (G') and d, loss modulus (G'') obtained from oscillatory shear-frequency sweep measurements under 1 % of shear strain.e, Creep shear modulus values at 0.5 s from (b) and storage modulus values at 2 Hz from (c).f, Surface tension measurements.g, The ratio of elastic modulus (G' at 2 Hz) over surface tension.Reference composite solution i to iii are PEO (8M Da, 2 % w/w), PEO (8M Da, 2 % w/w)+HA (0.5 % w/w), and PEO (4M Da, 2 % w/w), respectively (all dissolved in water).The rheological characterisations were performed with a parallel plate configuration (plate gap distance at 1 mm) by a Kinexus KNX2112 rheometer at 25 °C.
It is to note that in the histogram shown in Supplementary Fig. 7e, the defined creep shear modulus (G) and storage modulus (G') values for different solution formulations correspond to the measurement frequency or time point at 2 Hz or 0.5 s, respectively, and the values appear to be similar for the same solution composition.The reason of choosing 0.5 s (2 Hz) as the time point is because during in situ fibre tethering, it takes ~ 0.5 s for the fibres to be initiated and land on a target.Thus, the lower frequency storage modulus (<< 2 Hz) or longer time creep response (>> 0.5 s) of the solution is not relevant to the fibre formation stage, and the higher frequency storage modulus (>> 2 Hz) or shorter time creep response (<< 0.5 s) will be subjected to high degrees of measurement uncertainties.However, even if different frequencies (or time points) were chosen other than 2 Hz (or 0.5 s) to measure the modulus, the ranking of modulus among the solutions stays the same and does not affect the final conclusion.
Additional discussion on fibre 'spinnability': In the orbital spinning set-up, in order for a fibre to be successfully initiated by mechanical stretching and then to span several centimetres in distance between the nozzle and the rotating arm, molecular chains within the solution must exist in a long-range, weakly percolated elastic state.Such percolation ability of the polymer chains is indicated by the elastic storage modulus (G') of the solution (i.e., a bulk solution effect).Therefore, G' is considered the most appropriate indicator for characterising the fibre 'spinnability', instead of the solution's flowability (i.e., shear viscosity).
The countering effect for fibre formation is the solution surface tension () (i.e., surface effect), which could result in beads-on-string structures or even solution fibre breakage.The ratio of G' to  could be used as an indicator to characterise the solution 'spinnability' under this technique.If the ratio is small, surface effects of the solution (i.e., surface tension) will play a more prominent role than the bulk solution effect (i.e., elastic storage modulus), and the fibre will not be 'spinnable'.For example, solutions with a very low G' cannot be used to form fibres with this technique.
Taken together, the results show that the PEDOT:PSS (and its additives) and PEO components in the bioelectronic fibre solution naturally help to reduce the surface tension of the solution.Further, incorporating hyaluronic acid (HA) into long-chain polyethylene oxide (PEO) could be useful in promoting a robust chain percolation (as indicated by increased shear modulus).However, excessively strong molecular chain connection would also make the solution undrawable likely due to elastic re-coil and hindered chain sliding.Therefore, a region of 'spinnability' was indicated in Supplementary Fig. 7g.Supplementary Fig. 11 | Bioelectronic fibres interfaced with chicken embryos, and biocompatibility test.a, Magnified photos showing the bioelectronic fibres deposited on a chicken embryo.b, A chicken embryo without fibres.c, Biocompatibility test of bioelectronic fibres via chicken embryo development study for 24 hours.The embryos showed a normal rate of somitogenesis (a hallmark measure for developmental speed) at 0.63±0.08somites per hour (n=4), which is indistinguishable from control groups without fibres (0.62±0.07, n=5).In addition, the embryos showed normal axis turning, as indicated by the thoracic flexure and turning of the brain, and developed largely normal morphology at 24 hours post printing.Fibres were patterned at a spacing of ~ 100 μm all over the surface of the chicken embryo, but due to the small fibre sizes, they are not clearly visible under the specified microscopic setting.The feature size, or width (w), of the deposited fibres would be affected by the contact states.For fibres in the surface-adhesion state, the cross-sectional shape would become semi-circular, and the average fibre width is 3.2±1.4μm.This is because the freshly initiated fibre remains semi-wet when deposited onto the surface of the target object and the surface wetting would result in the fibre spreading out.Suspended fibres would usually have a circular cross-sectional shape 39 , and the average diameter is 1.4±0.4μm.
We could model the bioelectronic fibre electrode on skin as a simplified circuit shown in the above Fig.a.The impedance/resistance of the entire circuit:  ~  +  +  , where  is the fibre array skin/contact interfacial impedance,  is the fibre array axial impedance, and ZB is the impedance of bulk skin and body.
Since for N~180, DC axial resistance of the fibre array around the entire fingertip is estimated to be ~5 kΩ (c.f.Supplementary-Table 2 shows 10 kΩ for N~100), while Where N is the number of fibres (N=kt, k is the fibre deposition rate, and t is the deposition time), A is the typical area of contact between each fibre and skin,  is the typical impedance of a single fibre at the interfacial layer.It is seen in e that when both working and counter fibre electrodes are soaked in the same water bath, the surface impedance and ECG amplitude decrease evidently (but still clearly visible).This is because the water bath 'short-circuit' the working and counter electrodes (the current loss between the working and counter electrodes via the water bath).This test demonstrates that the fibre electrodes could fully replace the gel electrodes for biopotential sensing, and ECG sensing could still carry out even when the exposed fibre electrodes are 'short-circuited'.Nonetheless, the performance retained after the fingers being removed from the water bath and dried.
Supplementary , where F is the peeling force in Newtons, N is the number of fibres deposited, and 3×10 -6 m (or 3 μm) is the assumed mean width of each bioelectronic fibre, w).c, Microscopic images showing the local shear fractures of a bioelectronic fibre array on a fingertip undergoing excessive dry frictions (scale bar, 500 μm).
Additional discussion on the adhesion properties of the bioelectronic fibres: the maximum recorded peeling force per fibre width is ~ 15 N/m in both cases, which are comparable to hydrogel and film-based skin electronics from literature 2,[45][46][47][48] .The photos in Supplementary Fig. 18a show that the peeling process ultimately resulted fibre breakage from the skin models instead of noticeable delamination, when the peeling force is approaching the bioelectronic fibres' tensile limit (~ 20 N/m under dry ambient room conditions, Supplementary Fig. 5).
beyond its threshold region, causing an ON-OFF switch.In practice, if the reconfigured fibre path changes in length, customising N in the fibre array could make sure V* does not exceed ~ 13 V to power the LED in the threshold status.
According to the manufacturer's data sheet, a voltage decrease from ~ 1.9 ± 0.1 V to ~ 1.7 V across the LED could give it a noticeable ON to OFF switch (c-i).In other words, when the LED light is powered at its threshold LED ON status, ~ 0.2 V reduction in voltage across the LED would be sufficient to cause the ON-off switch (c-ii).
In terms of the detection limit, as shown in a-i, ammonia exposure (as low as ~ 1.7 % in water) could cause over 40 % increase in fibre resistance ( > 1.4,where R0 is the fibre array resistance prior to ammonia exposure).In this case, the LED threshold turn-off could happen with minimal V* ~ 2.4 V (i.e., [pre-exposure] V* = 1.9 V (LED) + 0.5 V (fibre), versus [postexposure] V* = 1.7 V (LED) + 0.5 V × 1.4 (fibre)).In the experiments, N could always be customised to tune the fibre array resistance R so that the driving voltage V* was above 5 V, but smaller than 13 V (the upper bound is set to avoid high voltages and also considering electric field working range determined in Supplementary Fig. 23).
Addition notes on overall fibre path resistance and operation voltage after reconfiguration: the overall fibre path resistance R is associated with the fibre array formed by N fibres of fixed length connected in parallel format; thus, theoretically,  ∝ .Circuit reconfiguration could change the fibre path length (thus L).However, at the circuitry level, when the reconfigured fibre path differs in length L, the circuit level fibre array resistance R could be compensated by customizing the number of fibres N.
Supplementary Fig. 25 | Bioelectronic fibre network stretchability characterisations.a, A photo showing the stretching experiment with bioelectronic fibres deposited on an elastomer film (Ecoflex 00-30), and microscopic images showing the parallel and orthogonal fibre network configurations.Ecoflex elastomer was used as the substrate because it possesses high stretchability 50 and could be regarded as a 'skin-like' material 51 .Parallel fibres ( ~ ) and orthogonal fibre networks ( ~ in each direction) were deposited on elastomer substrates (~ 1 mm thickness, Ecoflex 00-30).A lab-designed tensile rig was assembled with 3D printed clamps attached onto a motorised stage (Thorlabs MTS50-Z8), and a multimeter (Keysight 34465A) was used to record the resistance.b, Cyclic stretching tests for parallel fibres and fibre networks.

Supplementary Fig. 1 |
Single fibre deposition force measurement.a, Sketch showing the experimental set-up where a single fibre is deposited onto the free end tip of a suspended elastic cantilever, causing the deflection of the cantilever.b, Overlayed images showing the cantilever deflection due to the deposition of a single bioelectronic fibre.Supplementary Fig. 3 | Fibre surface tethering and contact.a, Images showing the timedependent wetting process for a bioelectronic fibre (scale bars, 500 μm).The zoom-in images compare a bioelectronic fibre undergoing continuous wetting for ~ 3.5 s after in-air spinning, and a previously deposited dried fibre with much thinner diameter.b, Plots of fibre widths versus time upon fibre tethering on a glass surface for the two solutions.c, Plots of fibre solutions' normalised contact line length ( ( ) ) versus time for the two solutions; in particular for the bioelectronic fibre solution, treated glass surfaces with various hydrophilicities are tested.Ct is the time-dependent contact line length, and R is the radius of the curved glass surface, as defined in (a); and ϴ in (c) indicates the water contact angle for the treated glass surface.
(The orange brackets indicate newly formed somites and the orange arrows indicate thoracic flexure, scale bar = 1 mm).Supplementary Fig. 13 | SEM images and histograms showing fibre feature width (w) of bioelectronic and pH-responsive fibres in (a) surface-adhesion, and (b) suspension states.
Fig. b(i) above shows  (f=0.1Hz)~ 10 3 kΩ, thus in this setup,  ≪  +  (where  corresponds to only the ~mm distance between the metal contact and where the fibres start to tether the skin).Therefore, we can assume  ~  + , For explaining the relationship between the impedance  (f=1kHz) versus deposition time and the fitting curve in Fig. b(ii) above,  (1kHz) ~   =

Fig. 18 |
Surface adhesion characterisations and failure mechanisms of the bioelectronic fibre arrays on typical living structure surfaces under dry conditions.a, The process of a modified 90-degree peeling test (derived from the ASTM D2861 90-degree peeling test): an array of bioelectronic fibres ( ~ ) being peeled off from a piece of porcine skin or an orchid leaf (scale bars, 2 mm).The fibre number density of ~ (thus average fibre spacing ~50 μm) was used in the peeling test because it resembles the fibre number density used for on-skin biopotential monitoring (i.e., = for ECG sensing from a fingertip).b, Peeling force per unit fibre width measured from the peeling experiment (estimated through × •

Table 2 |
Formats of bioelectronic fibre arrays designed for various applications.A summary of the bioelectronic fibre number density ( ), fibre network opening (average fibre spacing, ), and estimated transparency in various bioelectronic applications demonstrated in this work, in comparison to the size of biological features on human skins.
Order of references in the abovetable: A. Jiang, Zhi, et al.Nature Electronics 5.11 (2022): 784-793. 2 B. Lee, Sunghoon, et al.Science 370.6519 (2020): 966-970. 3C. Miyamoto, Akihito, et al.Nature nanotechnology 12.9 (2017): 907-913. 4D. Matsuhisa, Naoji, et al.Nature materials 16.8 (2017): 834-840. 5E. Zhu, Zhijie, et al.Advanced Materials 30.23 (2018): 1707495. 6F. Kim, Kyun Kyu, et al.Nature Electronics 6.1 (2023): 64-75. 7G. Chang, Huibin, et al.Science 377.6602 (2022): 180-185. 8H. Qin, Zhao, et al.Nature communications 6.1 (2015): 7038. 9I. Yan, Wei, et al.Nature 603.7902 (2022): 616-623. 10Supplementary of dry surface friction on the ECG signal acquisition.Bioelectronic fibre electrodes were tethered on the index and middle fingers, and a commercial hydrogel electrode was worn on the ring finger of the same hand.Only the fibre electrode on the index finger was subjected to friction with a polylactic acid plastic drum (surface friction speed ~ 4 cm/s and normal force ~ 2 N).b, ECG signals collected simultaneously from the three electrodes on fingers, with Pearson correlation coefficients (P) comparing the signal correlations between the fibre electrode subjected to friction and the other unperturbed electrode.c, Vaseline® is applied on the bioelectronic fibre electrode to simulate the effect of grease.d, The fingertip with bioelectronic fibre electrode was placed ~ 30 cm in front of an 100 watt infrared lamp (~ 40 °C).e, Contact impedance and ECG sensing performance negligibly change when only left index finger (working electrode) is soaked in water, and dried after removal from water.f, Contact impedance and ECG sensing performance when both left (working electrode) and right (counter electrode) index fingers are soaked in the same water bath.(The mean peak value is calculated from the mean R peaks in the ECG signals.For impedance values, data are presented as mean values ± standard deviation, and 3 independent measurements have been conducted for each condition).