Fluorine-Free Super-Liquid-Repellent Surfaces: Pushing the Limits of PDMS

Methods for fabricating super-liquid-repellent surfaces have typically relied on perfluoroalkyl substances. However, growing concerns about the environmental and health effects of perfluorinated compounds have caused increased interest in fluorine-free alternatives. Polydimethylsiloxane (PDMS) is most promising. In contrast to fluorinated surfaces, PDMS-coated surfaces showed only superhydrophobicity. This raises the question whether the poor liquid repellency is caused by PDMS interacting with the probe liquid or whether it results from inappropriate surface morphology. Here, we demonstrate that a well-designed two-tier structure consisting of silicon dioxide nanoparticles combined with surface-tethered PDMS chains allows super-liquid-repellency toward a range of low surface tension liquids. Drops of water–ethanol solutions with surface tensions as low as 31.0 mN m–1 easily roll and bounce off optimized surface structures. Friction force measurements demonstrate excellent surface homogeneity and easy mobility of drops. Our work shows that fluorine-free super-liquid-repellent surfaces can be achieved using scalable fabrication methods and environmentally friendly surface functionalization.

For liquid flame spray ( Figure S1), a pilot flame was created by the process gases oxygen (2 L min -1 ) and methane (1 L min -1 ). A liquid feedstock consisting of tetraethyl orthosilicate (98 %, Sigma Aldrich) in isopropanol (Fisher Scientific) was dispersed into the flame. As a dispersion gas, oxygen was used at a flowrate of 5 L min -1 . The concentration of the precursor in isopropanol was 370 mg mL -1 . The particles were collected on glass substrates (Thermo Scientific) at a distance of 10 cm and 15 cm with respect to the burner unit for 2 min and 3 min, respectively.
A silica shell was added via chemical vapor deposition to enhance the mechanical stability of the coatings. The surfaces were placed in a desiccator together with tetraethyl orthosilicate (98 %, Sigma Aldrich, 1 mL in 2.400 cm 3 ) and aqueous ammonia solution (25 %, VWR 4 Chemicals, 1 mL in 2.400 cm 3 ). The reaction was allowed to proceed for 16 h at atmospheric pressure. Afterwards, the coatings were sintered for 3 h at 500 °C in air. Higher sintering temperatures result in a decrease in surface roughness ( Figure S2).

Figure S2
. SEM images of LFS films deposited onto silicon wafers exposed to a) no sintering, b) sintering at 500°C, and c) sintering at 1000°C. Sintering at higher temperatures results in an overall reduction in surface roughness.
For the samples prepared via spray coating, fumed silicon dioxide nanoparticles (Aldrich, 7 nm, SSA = 395 m 2 g -1 ) were dispersed in acetone at a concentration of 5 mg mL -1 . 5 mL of the dispersion were used to coat a microscope glass slide (Thermo Scientific) with an area of 76 x 26 mm. Samples were sprayed with a flow rate of 0.2 mL s -1 at 2 bar at a distance of 10 cm using a spray gun with a nozzle diameter with 0.5 mm. Afterwards, the samples were allowed to dry for 24 h. The candle soot-based particle surfaces were prepared according to a procedure Science. In contrast to the LFS particle clusters, the candle soot templated silica particles are hollow. The particle clusters are stabilized by a thin silica shell.  water and the water-ethanol solution, respectively. Drop impact was recorded using a high speed camera (Fastcam AX10, Photron) and a high magnification objective (2x, Mitutoyo) at a frame rate of 5000 frames per second.

Frictions measurements:
Furthermore, we measured the friction force F F between a drop and a surface using a dynamic adhesion force instrument (DAFI). The drop is pushed over the surface at constant velocity using a glass capillary. The friction causes a deflection of the capillary. According to Hooke's law, the force required to move the drop is equal to the deflection of the cantilever ∆x multiplied by the spring constant of the capillary k s : If the droplet friction is dominated by contact line friction, it has been shown that the force can be calculated using Furmidge's equation. The Furmidge equation is used to calculate the force of a drop just before motion.
In analytical expressions, the three-phase contact line is often approximated by an ellipse ( Figure S6).
where w is the drop contact width, γ is the surface tension of the drop, and the dimensionless factor k accounts for angular variations of the contact angle around the contact line, where k ≤ 1. θ is the contact angle, ϕ is the azimuthal angle, and r(ϕ) is the radius of the ellipse. 5 Notably, the components of the capillary force perpendicular to the direction of motion cancel out. Therefore, the width of the drop w appears in the equation of the static friction force (also termed lateral adhesion force) and not the length. This is in line with a recent study where we verified that the width and not the length of a drop determines the static friction force. In the study, we elongated a drop parallel or perpendicular to the direction of motion. In both cases, the length of the three-phase contact line and the contact area was kept constant within experimental accuracy. For comparison, we also measured the static friction force of a sessile drop of identical volume (isotropic). We observed that the maximum static friction force greatly differs while the dynamic friction force does not depend on the initial shape of the drop ( Figure S7). We observed this pronounced dependence of the static friction force on the shape of the drop for all investigated surfaces, i.e. for smooth as well as for rough surfaces. 9 .
According to equation 3, friction decreases with decreasing interfacial tension. It increases with increasing contact width. For drops on a hydrophobic surface, friction also increases for decreasing rear contact angles.
Friction forces were analyzed using water and a water-ethanol solution (20 wt% EtOH, γ = 37.7 mN m -1 ). 15 µL drops of the respective probe liquid were placed onto the surface and brought into contact with the sensor. Then, the surface was moved at a constant velocity of 0.5 mm s -1 over a distance of 25 mm. The deflection of the sensor was analyzed using a Matlab script from side-view videos and the friction force was calculated from equation 1. Since droplet friction is volume-dependent, the forces were normalized to the radius of the 15 µL droplet (r = 1.5 mm from ( 3V / 4 ) ⅓ ). The radius was chosen for normalization because the contact width perpendicular to the direction of surface motion is not accessible with a goniometer. Furthermore, this is in analogy to the normalization in colloidal probe atomic force microscopy. As a sensor a glass capillary (50 mm x 0.5 mm x 0.05 mm, CM scientific Ltd.) with a stiffness of 105 µN mm -1 was used. The sensor was calibrated by monitoring the deflection of the end of the glass capillary when a known load is applied. For each surface, three spots were analyzed using fresh probe liquid for each scan.
The measured friction forces (F exp ) were compared to values calculated according to equation 3 with k = 1 (F F , Table 1). Since the contact width w is not accessible with goniometry, the contact length along the direction of surface motion in the kinetic state was used. This may be due to the pronounced stick-slip motion that does not become apparent in the advancing and receding contact angle used for the calculation.
For the LFS and the spray coated surfaces, a decrease in the friction forces in line with a decrease of the probe liquid's surface tension is observed. In case of the candle soot surface, both the experimental and the calculated friction forces increase with decreasing surface tension due to a change in the solid-liquid interfacial area and the receding contact angle. 17 Table 2. Contact angle (θ) and roll-off angle (θ roll-off ) measurements conducted on different PDMS-functionalized surfaces. 6 µL droplets of water, water-ethanol solutions, diiodomethane (DI) and ethylene glycol (EG) were used as probe liquids. The surface tension of water-ethanol solutions was gradually reduced by increasing the ethanol fraction in steps of 5 wt% to a maximum of 35 wt%. Apparent contact and roll-off angles were recorded until the Cassie-to-Wenzel transition occurred. LFS: liquid flame spray; SSP: sprayed silica particles; CSC: candle soot coating.