Carbon Nanofiber versus Graphene‐Based Stretchable Capacitive Touch Sensors for Artificial Electronic Skin

Abstract Stretchable capacitive devices are instrumental for new‐generation multifunctional haptic technologies particularly suited for soft robotics and electronic skin applications. A majority of elongating soft electronics still rely on silicone for building devices or sensors by multiple‐step replication. In this study, fabrication of a reliable elongating parallel‐plate capacitive touch sensor, using nitrile rubber gloves as templates, is demonstrated. Spray coating both sides of a rubber piece cut out of a glove with a conductive polymer suspension carrying dispersed carbon nanofibers (CnFs) or graphene nanoplatelets (GnPs) is sufficient for making electrodes with low sheet resistance values (≈10 Ω sq−1). The electrodes based on CnFs maintain their conductivity up to 100% elongation whereas the GnPs‐based ones form cracks before 60% elongation. However, both electrodes are reliable under elongation levels associated with human joints motility (≈20%). Strikingly, structural damages due to repeated elongation/recovery cycles could be healed through annealing. Haptic sensing characteristics of a stretchable capacitive device by wrapping it around the fingertip of a robotic hand (ICub) are demonstrated. Tactile forces as low as 0.03 N and as high as 5 N can be easily sensed by the device under elongation or over curvilinear surfaces.

presence of defects and are respectively the first and the second order relative to the breathing modes of sp 2 carbon bonds (1,2) . In single layer graphene, the 2D peak is constituted of a single component, while it presents many in the form of graphene platelets or graphite (3) . The G peak is correspondent to the E 2 g phonon at the Brillouin zone center (3,4) . The thickness of the employed GnPs is estimated to be > of 9 layers in a previous work (5) . The CnF used in this works are treated with a graphitizing procedure (heat process at ≈3000°C) to generate a catalyst free product and maximize conductivity in composites. The CnF spectrum presents the same main peak of GnPs indicating its nanocrystalline structure compared with amorphous carbon (6) . The ratio I(D)/I(G) is ≈ 0.1 and ≈0.6 respectively for GnPs and CnF, indicating a higher defect concentration for the nanowires.  Figure S2 presents the Thermogravimetric Analysis (TGA) of the pure elastomer (Nitrile), of the polymer blend sprayed on the Nitrile (Polymer) and of two conductive nanocomposites containing 10 wt.% of GnPs or CnF. The degradation of a material is quantified considering the weight loss versus the temperature as shown in Fig. S2a. All the samples are thermally stable until ≈ 230°C. Considering such value, the heat gun temperature was set at ≈180-200°C, thus avoiding thermal degradation while guaranteeing melting of the sprayed polymers. All the components (substrate and polymer blend matrix) degraded completely at ≈520 °C, preserving anyhow a residual percent mass around the 10% of its initial weight (7) . Interestingly, both the GnPs and CnF nanocomposites display a left-shift of the initial thermal degradation point. Such behavior is probably related to the nanofillers content which adsorb at the structural graphene/graphitic defects the -OH groups (8) .

Thermogravimetric Analysis
The first derivative of data in Fig. S2a is reported in Fig. S2b. Such graph highlights the degradation steps of the composites. Pure nitrile undergo a two steps degradation at ≈370-440 °C and at ≈440-530 °C (7) . The ink is composed by TPU (two degradation stages relative to the hard segment (between approx. 280 and 380 °C) and the soft segment (between approx. 380 and 440°C) (9) ) and HIPS (single degradation step between 330 and 450 °C (10) ). Therefore the composite without nanofillers (Polymer) presents a single degradation step which is the convolution of each degradation step presented above (Nitrile, TPU and HIPS). Introducing 10 wt.% GnPs or CnF, introduces peaks between 230 and 320 °C which are associated with the aforementioned defects of the graphitic material and the related -OH group.

Figure S 2:
Thermogravimetric analysis of the samples. a) Thermogravimetric weight loss measurements in nitrogen environment for various composites. b) The first derivative of the weight loss curve with respect to temperature.

Higher Magnification SEM images of the coatings Morphologies and Cross Sections
In Figure S3 we present SEM images of the details of the coating presented in the main text. Figure S3a and S3c display respectively the morphology and the cross section of the coating realized with 30 wt.% GnPs. Flakes of micrometric size are distinguishable in both the images (see black ellipse) and are distributed homogeneously through the sample. Figure

Fourier Transformed Infrared Spectroscopy
We investigated the Fourier Transformed Infrared Spectroscopy (FTIR) of the substrate (Nitrile) and of the HIPS, TPU and polymer blend sprayed on the nitrile (respectively H-Nitrile, T-Nitrile and Polymer). The results are displayed in Figure S4. The bare Nitrile shows the typical peaks of the rubber family: in particular the C≡N stretching vibrations of acrylonitrile at 2230 cm −1 and the band at 960 cm −1 relative to C-H wagging vibration of butadiene (11) . H-Nitrile exhibits the characteristic signature of polystyrene based materials, such as the aromatic ring bending between 700-800 cm −1 , C-H bending at ≈1440 and 1490 cm −1 and C-H stretching at ≈2840 and 2920 cm −1 (12) . T-Nitrile sample presents the peaks of TPU at ≈1720 cm −1 (hydrogen bonded carbonyl) and at ≈1733 cm −1 (relative to carbonyl) (13) .
Also the peaks at 2850 and 2930 cm − 1 (respectively the symmetric and asymmetric vibration of the C-H 2 group) and at 3330 cm − 1 (N-H group in urethane) are visible (14) . For both H-Nitrile and T-Nitrile samples, nitrile rubber chemical signature was not found. This is due to the thickness of the sprayed coatings (in the order of ten μm) which prevents penetration of the IR light to the elastomeric substrate.
The spectrum of the Polymer composites shows all the major peaks relative to the single polymer employed for the blend (HIPS and TPU) without important shifts. Again, the nitrile signal is absent due to the thickness of the sprayed coating. Adding the nanofillers introduces more noise and a decrease in the relative peaks intensity. Since no significant modification are observed, there appear to be no chemical interaction between GnPs or CnFs and the polymers in the nanocomposites.

Peel Tests for coating adhesion
Tape peel tests are accomplished to validate the adhesion strength of a coating on a substrate. We have applied tape peel adhesion tests to pure TPU-HIPS polymers, CnFs-and GnPs-based nanocomposite having the best electrical conductivity (30 wt.% filler concentration). In Figure S5 we display a typical peeling distance vs force graph obtained for the CnF-based sample. The adhesion force (resulting to be

Current-Voltage Curves
In Figure S6 we show typical current-voltage (I-V) trends obtained with our nanocomposites below and above electrical percolation. Below this threshold, the I-V curves display an highly hysteretic behavior characteristic of electrical insulator. We report only exemplificative results relative to the 0.5 wt.% CnF and 1wt.% GnPs based nanocomposite. After percolation a typical ohmic I-V response is reached. This time we showed as an example the 10wt.% CnF and GnPs loaded samples.  Figure S7 shows details on the cracking phenomenon observed on the GnPs based nanocomposites.

SEM images under stretching for 30 wt.% GnPs based Nanocomposites
Increasing the amount of GnPs leads to an earlier cracking behavior with elongation. We show details of the 30 wt.% GnPs loaded sample. At 15% elongation no cracks were present on the sample.
Afterwards large cracks were starting and at 60 % elongation the cracks are numerous on the nanocomposite surface.

Hysteresis Measurements
We evaluated the reversibility of deformation of our artificial skin materials through cyclic mechanical tests. All tests were performed on a Deben custom-designed dual-screw uniaxial testing machine.
Samples were stretched with the rate of 5 mm/min, until elongation , then deformation was released until tensile load was lower than 0.1 N, and the residual elongation was recorded. Each test had a duration of 10 cycles. From the cyclic stress-strain curves, the recovered deformation was calculated as ⁄ .
The stiffening effect due to either nanofiller is visible in the cyclic stress-strain curves (Fig. S8a)

SEM Images of the GnPs based samples released back to rest position (0)
In Figure S9 we show a sample loaded with 30 wt.% GnPs at rest length after 100% elongation.
The cracks formed during elongation are contacted back.

Bottom of the Stretchable Tactile Sensor
In Figure S11 is presented the bottom design of the sprayed tactile device. Such design can be obtained with both CnF and GnPs based conductive inks. We show a GnPs based coating 30 wt.% loaded.

GnPs-based Stretchable Tactile Sensor under Stretch
In Figure S12 is shown the GnPs-based stretchable tactile sensor under stretch. Cracks appear when stretched.

Figure S 12:
Device GnPs-based at zero elongation and 45 % stretch.