Influence of Wettability and Geometry on Contact Electrification between Nonionic Insulators

Contact electrification is an interfacial process in which two surfaces exchange electrical charges when they are in contact with one another. Consequently, the surfaces may gain opposite polarity, inducing an electrostatic attraction. Therefore, this principle can be exploited to generate electricity, which has been precisely done in triboelectric nanogenerators (TENGs) over the last decades. The details of the underlying mechanisms are still ill-understood, especially the influence of relative humidity (RH). Using the colloidal probe technique, we convincingly show that water plays an important role in the charge exchange process when two distinct insulators with different wettability are contacted and separated in <1 s at ambient conditions. The charging process is faster, and more charge is acquired with increasing relative humidity, also beyond RH = 40% (at which TENGs have their maximum power generation), due to the geometrical asymmetry (curved colloid surface vs planar substrate) introduced in the system. In addition, the charging time constant is determined, which is found to decrease with increasing relative humidity. Altogether, the current study adds to our understanding of how humidity levels affect the charging process between two solid surfaces, which is even enhanced up to RH = 90% as long as the curved surface is hydrophilic, paving the way for designing novel and more efficient TENGs, eco-energy harvesting devices which utilize water and solid charge interaction mechanism, self-powered sensors, and tribotronics.


Sample preparation
Silicon wafers covered with a 2 nm native oxide, and borosilicate glass (Mempax ® ) wafers were patterned with a hydrophobic fluorocarbon coating using a standard protocol in the MESA+ Institute for Nanotechnology of the University of Twente.Firstly, the wafers were primed by spin coating (4000 rpm, 30 s) HexaMethylDiSilazane (HMDS), followed by spin coating (4000 rpm, 30 s) a positive photoresist (Olin OIR 906-12) on the wafer.Hereafter, using a UV light source (350−450 nm), the photoresist on the wafer was illuminated through a mask with the geometrical patterns (EVG® 620 Mask Aligner).After the exposure, the resist was developed by placing the substrate for 60 s inside a beaker with the developer (OPD4262).The substrate was rinsed with DI water until the conductivity of the water reached 10 MΩ to remove all residues of chemical agents.Once the substrates had been dried, the CF x -layer (2 ≤ x ≤ 3) was deposited on the substrate by plasma polymerization of CHF 3 in a reactive ion etcher (RIE) system (25 sccm CHF 3 , 11W, 130 mTorr, 8 min., electrode temp.20 °C).Finally, using a lift-off process (substrate submerged in acetone sonicated for 15 min, followed by 15 min of sonication in IPA), all residues of the resist were removed from the substrates.Subsequently, the wafers were again rinsed with DI water to remove all chemical agent residues from the substrates, until the conductivity of the water reached 10 MΩ.

Probe and sample parameters
In Table S1 and S2 the material properties of the probes and samples are summarized.The radius of the colloidal probe is checked using a SEM (see Fig. S1).The spring constant is determined by using the thermal vibration method. 1 First the deflection sensitivity is determined on a sapphire sample and subsequently the spring constant is measured from the thermal tuning.Table S1: Various cantilever properties including the Young's modulus (Y ), poisson ratio (ν), spring constant (k), resonance frequency (f r , tip radius (r), relative permittivity (ϵ r ), particle density (ρ p ) and resistivity (ρ R ) .The spring constant is determined using the thermal vibration method. 1    (ii) through water in order to make water vapor.With the two manual valves, the relative humidity can be adjusted.A buffer chamber is inserted to make sure both the H 2 O and N 2 mix well.Just behind the buffer chamber, a relative humidity sensor is located.A similar sensor is placed behind the AFM chamber, to verify if the relative humidity remains the same throughout the whole process.

Acquirement of the F (D) curves
In Fig. S3 an example is given of the measured force (F ) as a function of time (t).The approach (t a ) and retraction (t r ) time are determined by the approach velocity (v a ).First the probe is out of contact and no force is exerted on the probe.At a certain distance from the surface, a total force is acting on the probe which consists of a combination of the van der Waals force (F vdW ), the electrostatic force (F e ) and the capillary force (F c ).When water is present on the surface or probe, the snap-in event is more pronounced compared to the other two forces (and depends also on the stiffness of the cantilever).From the snap-in event, the probe is into contact with the surface.The tip is pushed onto the surface till the load force (F L ) is reached. 8,9The load force remains constant through all the experiments.From this point the approach procedure is finished and the dwell time starts (t d ).During the dwell time, the force is kept constant.When the dwell time is over, the retraction sequence starts.
The probe is lifted till it snap-out of contact.From this moment the probe is no longer in contact with the surface, but long-range forces still act on the probe (such as F e ).When the probe is retracted even further away from the surface, the force acting on the probe is reaching zero and the procedure is over.
Figure S3: Example of a Force versus time curve.The contact time (t c ) consists of the dwell time (t d ) and the approach velocity ((v a ), which determines approach (t a ) and retraction time (t r ).The force applied to the colloidal probe during contact is given by (F L ) .
The contact time (t c ) is the time between the snap-in and snap-out event.It is influenced by the dwell time (t d ) and the approach velocity (v a ).The approach velocity determines the time between the snap-in event and the moment the maximum force is acting on the probe, and the time between the end of the dwell time and the snap-out event.
When a matrix measurement is performed, the probe touches the surface in a grid-like fashion.Both the CF x and the pristine surface are measured simultaneously (see Fig. S4).
On the interface between the CF x layer and the Si wafer, F (D) measurements are performed in a matrix format.In the inset of Fig. S4(b) an example is depicted.Every square is a measurement and when a force distance curve is performed, the probe moves to the next square till the whole area is covered.When finished, force-distance curves are collected on both the CF x layer as well as on the pristine Si surface.Note here, that only one measurement is performed per position.The median curve is than extracted from all the curves on the same surface.For single point measurements, a force-distance curve is performed multiple times on the same position.In this procedure, the history of measurements affects the result, as charging of the colloidal probe and the substrate changes the outcome of the force-distance curves.This procedure allows for studying the charging process as a function of the number of touches.

Electrostatic force extraction
In order to extract the contact electrification voltage (V CE ) from the measurements, the electrostatic force component needs to be extracted from the F (D)-curves.The electrostatic force acting on the probe and cantilever is described by Law and Rieutord 10,11 and is given by where ϵ 0 is the vacuum permittivity and g(D) the geometrical factor as a function of distance (D).The geometrical factor consists out of three components, the apex, cone and cantilever (for more information, see ref. 11 ).All three components contribute in different ways and have different distance dependencies.Due to the different components, it is regarded as a difficult task to quantify electrostatic forces.Therefore a useful approximation for all distances is with R the radius of the tip or colloidal particle.For probes with a large radius, such as colloidal probes, equation S2 is applicable.On short distances, both the electrostatic as well as the van der Waals force are present, however, typically the electrostatic force is of a much larger magnitude compared to the van der Waals force. 12 depicted in Fig. 1 in the main text, multiple forces constitute the colloid probe-surface interaction.In order to extract the V CE value, first the electrostatic component in the F (D) has to be determined.An example is shown in Fig. S5.Especially at high RH, the capillary force significantly affects the shape of the F (D)-curve, making it difficult to determine the starting point of the electrostatic force.In order to determine where the electrostatic force is more dominant compared to the capillary force, the derivative and second derivative of the F (D) are extracted.For low RH values, the first derivative is sufficient to determine the starting point, as this coincides with the highest value in the derivative.At higher RH values, no peak but a plateau is present in the dF (D)/dD-curve (see Fig. S5(a)).Therefore the minimum in the d 2 F (D)/dD 2 -curve is determined.This is the point where the slope changes in the F (D)-curve (blue triangle in Fig. S5(c)).The end point is placed at the position where the force reaches zero (red square in Fig. S5(c)).The data in between these points is then used to fit eq.S2, from which the V CE is extracted.
6 Snap-out distance, distance to zero force and indentation Besides the obtained V CE and F ad in the main text, several other parameters can be extracted from the obtained F(D) curves.The snap-out distance (D so ) is the distance at which the water meniscus snaps (i.e. the point where the electrostatic force starts to dominate, or the blue triangle in Fig. S5).In Fig. S6 the snap-out distance is plotted as a function of RH.
Only for the hydrophilic-hydrophilic material combination a D so is extracted, as the liquid bridge is not or barely present when a hydrophobic material is involved.Similar to the adhesion force, the snap-out distance is heavily dependent on the relative humidity.This is expected, because the capillary force dominates the adhesion force on hydrophilic-hydrophilic material combinations and the snap-out distance is also heavily dependent on the same force.As RH is increasing, more water is present on the surface, 13 and the capillary force and bridge formation is enhanced.
Another parameter extracted from the F (D)-curves, is the distance to zero force (D zf ) and is defined as the retraction distance at which no force is acting on the colloidal probe.
The dependence on the RH is shown in Fig. S7.For the hydrophilic-hydrophilic interaction, a small increase in D zf is observed, while for other material combinations D zf remains constant.
However, a clear difference in distance is observed between the measurements on CF x and on the pristine surface.On the latter, the distance to zero force is much smaller, indicating that long range forces are acting on a smaller length scale compared to the CF x layer.This is in agreement with the higher contact electrification voltage observed on the same layer.Indentation (δ) has the potential to measure local micromechanical properties of an interface, such as the hardness and elastic modulus. 14However, the interpretation of the results is complicated by the altering shape of the tip during the experiment and the small length scales.In Fig. S8 the extracted indentation values are plotted as a function of the RH.No clear trend is observed, but similar to D zf a clear difference is observed between the CF x coated surfaces and the pristine substrate.A larger indentation is observed on the CF x layer in agreement with the difference in Young's modulus (see Table S2).In general, also a higher indentation value is found for the measurements performed with a polystyrene colloidal probe (in agreement with the different Young's moduli found in Table S1).

Error Analysis
The uncertainty in the measurements is determined by calculating the standard deviation per measurement point in Figure S9 and Figure S10.Each measurement point in these Figures consists of 30-60 measurements.The adhesion and/or contact electrification voltage is determined from each curve.The standard deviation is visualized as the error bars in Figure S9 and Figure S10.In Figure S9, the error is significantly larger for the hydrophilic-hydrophilic material combination.This is most likely caused by the inhomogeneity in the water layer present on the colloid and the surface.6][17] In Fig. S11 the charge progress is monitored as a function of number of contacts (N c ).For all humidities and approach velocities, no additional charge is accumulated on the colloidal probe and on the surface.This is caused by the speed of the measurement compared to the obtained time constant of charging (Fig. 5b in the main text).During the measurement, the charge is already vanishing into the vapor phase.This process is faster then the measurement method and therefore no additional charging is observed.This agrees with the absence of electrostatic interaction in the approach curves (for instance Fig. 2  When the approach velocity is significantly increased, a sign of tribocharging is observed in the approach curve.An example is shown in Fig. S12.For slow approach velocities (red curve in Fig. S12), the force acting on the cantilever is zero till it snaps into contact.For the F (D)-curve measured with a fast approach velocity (blue curve in Fig. S12), an additional force is already interacting with the cantilever.Before the snap-in moment, the measured force is non-zero indicating an interaction between the surface and the cantilever.Because of the fast approach velocity and zero dwell time, the measurement takes approximately 0.125 s, which is faster than the obtained time constant of charging (τ d ≈ 0.5 s).Therefore, a small amount of charge can still be detected.However, the measurement is still too slow to observe the accumulation of charge (see Fig. S11).

Influence of the approach velocity
The influence of the contact time on the contact electrification voltage is shown in Fig. 5 and discussed in detail in the main text.However, also the approach velocity (v a ) is varied.Note that the approach velocity is equal to the retraction velocity discussed in the main text.In Fig. S13, V CE is plotted versus v a for different dwell times.The model described in eq.6 can be extended further to describe a particle impacting on a plane surface which is a good

Figure
Figure S4: (a) Topography image of the CF x coated zone on a Si wafer (10 × 10 µm, scale bar 2 µm).(b) Cross-section height profile along the white line in (a).The CF x layer is approximately 80 nm thick.Inset: A matrix indicating where the probe performed F (D) experiments.

Figure
Figure S5: (a) A F (D)-curve (black) at RH = 45%, t d = 2000 ms and v a = 400 nm/s and the derivative of the F (D)-curve (red).The same F (D)-curve (black) as in (a) and the second derivative of the F (D)-curve (blue).(c) The same F (D)-curve as in (a,b) with markers which indicate the begin (blue triangle) and end (red square) of the electrostatic force component.(c) The electrostatic part of the F (D)-curve in (b).The red line is a fit based on equation S2.

Figure S6 :
FigureS6: The Snap out distance (D so ) as a function of the relative humidity.A similar increasing trend with increasing RH is found as for the adhesion force (Fig.3of the main text).

Figure S7 :
Figure S7: The distance to zero-force (D zf ) as a function of the relative humdity for (a) the CF x coated surface and (b) the pristine surface.

Figure S8 :
Figure S8: Indentation (δ) as a function of the relative humdity for (a) the CF x coated surface and (b) the pristine surface.

Figure S11 :
Figure S11: The potential difference (V CE ) versus the number of contacts (N c ) for different relative humidities and approach velocities.No charging is observed between the different contact moments.

Figure S12 :
Figure S12: An example of a fast (blue, v a = 8000 nm/s) and a slow (red, v a = 160 nm/s) approach F (D)-curve.The fast curve contains a electrostatic component, while in the red curve the electrostatic component is absent.