Self-cleaning of Surfaces: the Role of Surface Wettability and Dust Types

The self-cleaning property is usually connected to superhydrophobic surfaces (SHSs) where the dust particles can be easily removed by the rolling motion of droplets. It seems that superhydrophobicity (its durability is questionable nowadays) is a necessity. However here, it is disclosed that self-cleaning can also be realized on an ordinary surface by droplet impinging. The effects of surface wettability and the types of dust particles are considered. The self-cleaning is realized by two steps: (1) the pickup of particles by the water-air interface of an impinging droplet, (2) the release of the impinging droplets from the surface. It can be observed that only the trailing edges of the droplets can pick up particles when the droplets recoil from the inclined surfaces. The hydrophilic surface can also achieve self-cleaning under some conditions. This interesting finding may be helpful for the successful implementation of self-cleaning with common surfaces.


Materials and Methods Materials
Tetraethoxysilane (TEOS) is purchased from Guangzhou Chemical Reagent Factory.
Hexamethyldisilazane (HMDS) is purchased from Aladdin and Adamas. Bisphe-nol Epoxy resin (E-44) and low molecule polyamide curing agent of commercial-grade is purchased from Yueyang petrochemical plant, China. Distilled water is prepared by a Purescience water purification system.

Preparation of Sample Surfaces with Different Wettability
The glass slides are used as hydrophilic surfaces. They are cleaned by ultrasound in a solution mixed with acetone and ethanol to remove the organic contaminants. After that, the slides were thoroughly rinsed with distilled water and dried under vacuum. Hydrophobic and superhydrophobic surfaces are prepared as the previous work 1 .Silica sol (SS) is prepared first. Epoxy resin (0.63 g) and polyamide (0.33 g) are dissolved in silica sol (9 ml) with magnetic stirring until an epoxy suspension (ES) is formed. The ES with different curing time is dripped onto the cleaned glass substrates and are evenly spread out. The resulting glass substrates are then dried at room temperature for one day. After the ethanol evaporates, epoxy resin particular films are formed on the substrates. The wettability of these films changes from hydrophobicity to superhydrophobicity as the curing time of ES prolongs.

Dust Particles
Two types of dust particles are used here. The first type of dust is collected from building site (D1). The second is self-prepared epoxy resin microspheres (ERMs). Their preparation method is as follows 1 . The curing time of ES are prolonged to 11h. The microspheres are collected by centrifuging the ES, then washed with ethanol several times and dried in 60 °C for 24h. The ERMs are hydrophobic with diameters in the range of 2-8 μm. The morphology of the two dust particles are shown in the following text. The dusty sample surfaces are prepared by the free settling method. First, the samples are placed horizontally on the bottom of a container. Then, the D1 particles or ERMs are scattered into air by a stirring device. They will settle on the sample surfaces by gravity for a long time. The free settling process is lasted for 24h.

Characterization
The morphology of sample surfaces and dust particles are observed by scanning electron microscope (SEM, S-3700N, HITACHI, Japan). Energy-dispersive X-ray spectroscopy (EDS), Fourier transform infrared (FTIR) spectrometry (BRUKER550, Germany) and X Ray Fluorescence (XRF, PANalytical Axios) are performed to measure the components of dust particles. Thermogravimetry (TG, NETZS, Germany) instrument is utilized to calculate the ignition losses of particles. Surface wettability measurements were performed with a Dataphysics OCA 20 contact angle system at ambient temperature. Advancing and receding contact angles of samples are measured both with volume addition/subtraction and by roll off methods. The drop volume is 6μL for the former method and 10μL for the latter method 2 .

Self-cleaning Experiments by Impinging Droplets
Self-cleaning experiments are performed by contaminating the samples, followed by impinging water droplets on these dusty surfaces, and observing the particle removal processes. The experiments are conducted in a room where the temperature and relative humidity (RH) are measured of 25±2°C and 45±5%, respectively. The impinging velocity of a droplet is controlled by changing its falling height from 0.003m to 0.044m. The inclined angles of sample surfaces are 45°. The droplet diameter is kept constant at 2.84 mm.
It's density is 997 g/mL and its surface tension is 0.072 N/m.
A high-speed camera (pco. Dimax HS1, CooKe) is used to observe how the impacting droplets carry away dust particles that are initially adhering to the surfaces. The resolution of the high speed camera is 1000 X 1000 pixels. Its recording rate is 5500 fps (frames per second) and the exposure time is 0.168 ms.

Sample Surfaces
The surface morphology of samples with different curing time are observed by SEM (Figure. S1). As the curing time prolonged, the microspheres gradually appear and grow up. The aged silica sol deposited on microsphere surfaces forms the multilevel structures. The wettability of sample surfaces is controlled by adjusting the reaction time of ES. The advancing θ A , receding θ R contact angles, and contact angle hysteresis (Δθ=θ A -θ R ) are measured 3 when the curing time of ES is changed from 1h to 24h (Figure. S2). As the curing time prolongs, the two angles increase and the contact angle hysteresis decreases, which means that the surfaces become more hydrophobic. For curing time less than 5h, the θ A of these samples are larger than 130°. However, their receding angles are less than 90°, which means that the hydrophobicity of these samples are poor. They can be regarded as ordinary hydrophobic surfaces. After 9h, the θ R of samples are larger than 135°, which can be regarded as a threshold for superhydrophobicity. Antonini 3 pointed out that the receding angle greater than 135° used as a threshold for superhydrophobicity will be more rational than the conventionally reported static contact angle greater than 150°. For the convenience of description, we define four kinds of surfaces according to the different wettability (see Figure. S2). The first is the hydrophobic surface (HS, the curing time is less than 5h) with a receding angle smaller than 90°. The second is high hydrophobic surface (HHS, the curing time is more than 5 h and less than 9 h) with a receding angle larger than 90° and smaller than 135°. The third is superhydrophobic surface (SHS, the curing time is larger than 9h) with a receding angle larger than 135°. The last one is the hydrophilic glass slide (GS) without coatings.

Dust Particles
The morphology of the two types of dust particles are observed by SEM. Figure. S3 shows that the shapes of dusts from building site are irregular, the ERMs are spherical. The dust particles can be distributed evenly on the sample surfaces by free settling method ( Figure. S3 c-d).
The components of dust particles are analyzed by EDS, FTIR and XRF. Table S1 shows the types of elements for D1 and ERMs by EDS. ERMs are the curing products of epoxy and polyamide curing agent.
They are organic particles which only contain elements of C, O and N (the element of N is not listed in Table   S1). These particles are hydrophobic, which has been proved in our previous work 1 . For D1 particles, their main elements are O and Si. "Others" in Table 1 refer to the sum of micro-constituents such as Ti, Zn, Sr, and Cl, etc.. FTIR is utilized to analyze the organic compounds of ERMs and D1 particles in Figure. Table S2. "Others" refer to the sum of micro-constituents such as TiO 2 , ZnO and SrO, etc..

Capillary Force
The dust removal experiments are conducted in a room where the temperature and relative humidity (RH) are measured of 25±2°C and 45±5%, respectively.
The liquid bridge is not formed if both the contacting surfaces are hydrophobic 5,6 . Thus, when hydrophobic ERMs particles contact with hydrophobic surfaces (HS, HHS, SHS), the capillary force is not considered.
For hydrophilic particles contacting with hydrophobic surfaces, the total adhesion forces between the two kinds of surfaces are found to be almost constant for all RHs 5,7 . This is because the capillary condensation is weak between surfaces, resulting in small capillary forces. Thus, when hydrophilic D1 particles contact with hydrophobic surfaces (HS, HHS, SHS), the capillary force can be ignored.
If the two contacting surfaces are both hydrophilic (e.g., D1 and glass surface), the effect of RH on the adhesion force is significant. There is no critical RH for capillary condensation, but the adhesion tends to rise steeply at RH > 60% 5,8 . Under the condition of low RH (45%), the capillary force is still small 9,10 . In addition, as shown in Figure 7a and Figure S12, the impinging drops cannot release themselves from glass surfaces. That's to say, D1 particles cannot be carried away by drops no matter whether their adhesion forces are large or small. The adhesion forces of D1 particles have no influence on the results of dust removal process.
Thus, under current conditions, only van der Waals force is considered.

Movie S1
The sliding-rolling removal process occurs on a SHS distributed with D1 particles. The We N is small (1.14).

Movie S2
The rebound removal process occurs on a SHS distributed with D1 particles. The We N is 16.8.

Movie S3
The retention of the drop impinging on a HS distributed with D1 particles. The We N is 8.39.

Movie S4
The rebound-wriggling removal process occurs on a HS distributed with ERMs. The We N is 8.39.

Movie S5
The rebound-tail rolling removal process occurs on a HS distributed with ERMs. The We N is 16.8.

Movie S6
The retention of a droplet impinging on a GS (hydrophilic surface) distributed with D1 particles. The We N is 16.8. Some particles redeposit on the cleaned area, which can be observed clearly form the movie.

Movie S7
The rebound-retention process of a droplet impinging on a GS (hydrophilic surface) distributed with ERMs. The We N is 8.39. The receding angle on the trailing edge is very large in the early recoiling stage, and a cleaned area appears after the drop. After that, the receding angle decreases dramatically due to the direct contact of the droplet with the hydrophilic GS. A small fraction of ERMs can be removed by the rebound part of the drop.