Superhydrophobic Coatings as Anti-Icing Systems for Small Aircraft

: Traditional anti-icing/de-icing systems, i.e., thermal and pneumatic, in most cases require a power consumption not always allowable in small aircraft. Therefore, the use of passive systems, able to delay the ice formation, or reduce the ice adhesion strength once formed, with no additional energy consumption, can be considered as the most promising solution to solve the problem of the ice formation, most of all, for small aircraft. In some cases, the combination of a traditional icing protection system (electrical, pneumatic, and thermal) and the passive coatings can be considered as a strategic instrument to reduce the energy consumption. The effort of the present work was to develop a superhydrophobic coating, able to reduce the surface free energy (SFE) and the work of adhesion (W A ) of substrates, by a simplified and non-expensive method. The developed coating, applied as a common paint with an aerograph, is able to reduce the SFE of substrates by 99% and the W A by 94%. The effects of both chemistry and surface morphology on the wettability of surfaces were also studied. In the reference samples, the higher the roughness, the lower the SFE and the W A . In coated samples with roughness ranging from 0.4 and 3 µm no relevant variations in water contact angle, nor in SFE and W A were observed.


Introduction
Typically, the presence of tiny pieces of ice or supercooled liquid water in the clouds, which remain liquid below zero and suddenly turn to ice after the impact with the aircraft surfaces, are the main sources of ice deposition during a flight. The presence of ice on surfaces alters the airflow over the wing and tail, and then reduces the lift force that keeps the plane in the air. This potentially causes aerodynamic stall, a condition that can lead to a temporary loss of control of the aircraft. In order to prevent or reduce the ice formation or alternatively to remove the ice once it was formed, anti-icing and de-icing systems are usually adopted. Currently the thermal and pneumatic types represent the most largely employed systems. In details, the thermal system melts the ice accretion or prevents the ice from forming by the application of heat on the protected surface of the wing. The heat is generated by the hot air "bled" off the jet engine into piccolo tubes routed through wings, tail surfaces, and engine inlets. The spent bleed air is then exhausted through holes in the lower surface of the wing. Similarly, the electro-thermal systems use resistive circuits buried in the airframe structure to generate heat when a current is applied. The heat can be generated continuously to protect the aircraft from icing (anti-ice mode), or intermittently to shed ice as it accretes on key surfaces (de-ice), the latter being generally preferred due to the lower power consumption. Instead, the pneumatic de-ice boot consists of a rubber sheet bonded to the leading edge of the airfoil: when the ice builds up on the leading edge, an engine-driven pneumatic pump inflates the rubber boots, and then the ice is cracked and should fall off the leading edge of the wing. Additionally, some innovative systems are being developed, e.g., the electro-mechanical de-icing systems which use a mechanical force, generated for instance by actuators, to knock the ice off the flight surface, or the weeping wing system, which releases a glycol-based chemical onto the wing surface through small orifices which causes the detachment of ice. However, these methods can be inefficient, environmentally unfavorable, expensive and time consuming [1]. Thus, it would be advantageous if surfaces could passively prevent the ice formation and facilitate ice removal. In this contest, the superhydrophobic coatings, applied as usual paint, owing to their extraordinary water repellency, and not requiring additional energy consumption, can be viewed as excellent candidates for icephobicity in this area [2,3]. This is because, ultrahigh θrec, typical of nanostructured superhydrophobic surfaces, means low ice adhesion strength [4,5]. Subsequently, if supercooled liquid water freezes when it impacts the solid surface of an aircraft, the resulting ice can be shed, taking advantage of external forces, such as aerodynamic forces, to overcome ice-surface adhesion forces [6]. Evidently, the efficiency of these coatings as an anti-icing passive method largely depends on the different scenarios in which the ice may form [1]. In some cases, therefore, a combination of a traditional icing protection system and the superhydrophobic coatings could be seen as a strategic instrument able to assure a high efficiency in a wide range of environmental conditions.
In the last decade, a huge body of literature concerning the development of superhydrophobic and icephobic coatings able to delay the ice formation [7], or reduce the ice adhesion strength once formed, has been produced [8][9][10][11][12][13][14][15]. A shared strategy is to create surfaces with low surface free energy and, contemporary, micro-nanoscopic rough structures. In this regard, it was found that the icing probability was reduced to zero with particle diameters ranging between about 30 and 70 nm [16]. Some authors achieved this goal by using etching and/or high temperatures [9,11], or complex fabrication techniques [12][13][14][15][16][17], which are often limited by the substrate type and geometry that can be successfully coated [18].
The effort of the present work was to develop a superhydrophobic coating for metallic substrates with a simplified and non-expensive method, which could be employed as a usual paint able to prevent/reduce the formation of ice, especially on small aircraft [19]. The newly formulated superhydrophobic coating consists of nanostructured layers able to generate hierarchical micro/nano-structured roughness, and reduce the surface free energy, which as previously declared, are the two main factors useful to making a superhydrophobic surface. The effect of the substrates' roughness on the surface properties of coated and uncoated samples was also studied, and then correlated to the wettability of samples, in order to identify and separate the morphological and the chemical contributions.
This work is being developed in the framework of the Clean Sky 2 ongoing SAT-AM (More Affordable Small Aircraft Manufacturing) project, whose main goal is the investigation of new technologies for a future small aircraft able to fly with low fuel consumption, low noise levels and needing low quantities of raw material in its life cycle. The reference vehicle for this project is the M28 designed and manufactured by Consortium Partner Polskie Zakłady Lotnicze (PZL), Mielec (PL). It is a commuter category 19 passenger, twin-engine high-wing cantilever monoplane of allmetal structure, with twin vertical tails and a robust tricycle non-retractable landing gear, featuring a steerable nose wheel to provide for operation from short, unprepared runways where hot or highaltitude conditions may exist. The M28 is best suited for passenger and/or cargo transportation and is certified under EU CS-23 and USA FAR 23 requirements.

Materials and Methods
The developed superhydrophobic coating was prepared starting from the formulation described in a previous work [19] and slightly modified in order to improve its durability. In this regard, other details about the performed changes cannot be disseminated in this manuscript. Once prepared, several layers of the developed coating were applied with an aerograph on substrates representative of the M28 air intake. The dehumidified air was pressured at three bar; whereas the other application parameters, such as the layers' number, the curing temperature and the distance between the aerograph and samples were optimized in order to guarantee uniformity in the coating's thickness and weight.
The substrate samples, 5 × 5 cm 2 in dimensions, made of stainless steel 12H18N10T (same composition of the M28 air intake), and having roughness ranging from 0.7 to 4.6 µm, were provided by METROL (a partner of the Clean Sky project). The effect of the substrate roughness on the final properties of the coating in terms of wettability and adhesion of the coating on the substrate were studied.
Roughness of substrates before and after the application of the coating was measured by employing a SAMA SA6260 surface roughness meter. Roughness measures, performed according to the ISO 4288 [20], were reported as Ra, which represents the arithmetic average of the absolute values of the profile height deviations from the mean line. ID samples was correlated to the correspondent roughness values in Table 1. Contact angles (CA) were calculated according to Young's equation [21] (see Figure 1): where θ is the contact angle, LV γ and SV γ are the liquid and solid surface free energy, respectively, whereas SL γ is the solid/liquid interfacial free energy. The schematic illustration of the equilibrium among involved forces according to the Young equation is shown in Figure 1. The contact angle measurements were performed at room temperature in compliance with the ASTM D7490-13 [22] standard, using water (H2O) and diiodomethane (CH2I2). The Surface Free Energy (SFE) was calculated according to the Owens Wendt (OW) method [23,24], for which the surface energy of a solid is the sum of two components, a dispersive and a polar one. The dispersive component theoretically accounts for the Van der Waals and other non-site specific interactions between the surface and the applied liquid, whereas the polar component accounts for the dipole-dipole, dipole-induced dipole, hydrogen bonding, and other site-specific interactions [25]. The polar and dispersive components of the reference liquids are listed in Table 2. Tests were carried out by depositing ten drops of 3 µL of each liquid on the sample surface. According to the OW method [23], the interfacial solid/liquid energy can be evaluated as: which, combined with the Young equations, gives the following: Therefore, the polar (  θ , 2 θ are the SFE components of the two reference liquids and the corresponding contact angles between the solid and the liquids, respectively. Although the SFE's components can be known through the literature (Table 2), the contact angles have to be experimentally evaluated through several measures. That determines a stochastic nature of the contact angles and therefore of the solutions of the equation system (4).
So, to assess the SFE problem, a third reference liquid, i.e., the Formamide (HCONH2), was introduced in order to reformulate the initial problem in terms of the following nonlinear programming problem: or equally as: Finally, the Surface Polarity (SP) was calculated as: The Young Equation (1) is valid for smooth surfaces or substrate having very low roughness, but when the roughness is too high to be neglected, the equation should be corrected taking into account the actual surface morphology. Two different models can be useful to describe the wetting of textured surfaces, i.e., the Wenzel and Cassie-Baxter (Figure 2a,b respectively) models. In details, the Wenzel model [26,28] proposed a correction factor "r" for contact angle on rough surfaces, which is equal to the ratio of rough interfacial area over flat interfacial area under the droplet. Substituting the roughness ratio factor "r" in the Young's Equation (1), one can obtain the Wenzel's equation [28], i.e., θ r θ cos * = cos (12) where θ* and θ are the contact angles of a droplet on a rough surface and on the corresponding smooth surface, respectively. Wenzel's model assumes no air-trapping under droplet, which may not necessary be true. On the contrary, the Cassie-Baxter model [29] allows to estimating the contact angle on rough surface with air-trapping, through the following equation: where φS is the area ratio of liquid-air interface to the whole interface; θ* and θ are the contact angles with and without considering air-trapping. Cutting and tape test were carried out, according to the ASTM D 3359-09 standard [30], by making a grid incision with a specific cutter on the coated samples and then by applying an adhesive tape to cover the cut area. The test adhesive was vigorously removed and the involved area was inspected. According to the ASTM D 3359-09 [29], the test passes if the area involved in detached flakes of coating at intersections is less than 15%.
Sample morphology, measurements of nanoparticle size and of coating thickness, were carried out using a field emission SEM Tescan Mira3 (Tescan Orsay Holding, Brno-Kohoutovice, Czech). Samples were observed in as-realized conditions, without metallization on the observed surface.
Surface 3D morphology and the corresponding roughness were assessed by using the laser profilometer Ametek Talyscan 150. The scanning area for 3D surface reconstruction was 40 × 40 mm 2 (800 profiles with 5 microns resolution), acquired with a scanning velocity of 10,500 microns/s.

Substrates' Roughness of Substrates
Roughness of substrates measured with the SAMA SA6260 surface roughness meter, before and after the application of the coating, were compared in Figure 3. It highlights that, in general, the application of the coating reduces the original roughness. This especially for samples 1-6, for which the values of the coated samples' roughness vary in the range 0.4-0.6 µm, in spite of the substrates' roughness ranging from 0.7 to 1.5 µm, whereas, for samples 7-9, the application of the coating reduces the original roughness from 4-4.5 to about 3 µm.

Application of Coating
The aerograph setting parameters were varied and optimized in order to guarantee uniformity in terms of thickness and weight of the applied coating. Here, as an example, two different setting parameters, labelled as Procedure 1 and Procedure 2, are described; the corresponding schematics are shown in Figure 4. They differ mostly in the samples orientation with respect to the aerograph spray and in the drying temperature, whereas the layers' number, the coating formulation, and the substrates were the same. After the application, measures of contact angle with water at room temperature were carried out. It was found that Procedure 1 is able to produce samples with contact angle of 124°, whereas Procedure 2 gives surfaces with a water contact angle of 155° ( Figure 5). This difference is probably due to the different morphologies because of the diverse application procedures.   Hence, Procedure 2 was adopted to apply three layers of the developed coating on sample 1-9 with dimensions of 5 × 5 cm 2 . The Samples were weighed before and after the application of the coating, in order to assess a rough estimation of the coating's specific weight, which was plotted as a function of the sample roughness in Figure 6. It was found that the coating's specific weight was about 11 ± 1 g/m 2 . SEM cross-section images of sample 2 (Ra = 0.8 µm) and sample 7 (Ra = 4 µm), employed to measure the coating's thickness, are shown in Figure 7a,b, respectively. It seems that the two samples display some differences in both coating thickness and morphology. Sample 7 in Figure 7b, in fact, has a coating thickness higher than sample 2 shown in Figure 7b (see values in Table ), with a mean value of 190 µm vs. 127 µm. Moreover, sample 7 coating appears to be more uniform in thickness and morphology than the sample 2 coating, and with less amount of large and localized voids. Hence, it seems that the roughness is able to improve the coating's uniformity.

Contact Angle Measurements
The measures of the contact angles acquired with H2O and CH2I2 were reported as function of the substrate roughness in Figure 8 and Figure 9, respectively. Some representative drop pictures caught during the measures were overlapped. It should be noted that, for the reference samples, the water contact angle is almost constant with a substrate roughness up to 1.8 µm beyond that, as expected [31], it increases as the roughness increases, reaching a value of about 100-115° for the highest value of measured roughness, i.e., 4.6 µm. The increasing of the contact angles with the roughness of metallic substrates can be ascribed to the changes in morphology. Other authors [31,32] also observed this behavior with maximum values of the achieved contact angles at about 120°C [31]. Indeed, it was found that in order to achieve higher values of contact angles without modifying the surface chemistry, it is necessary to scale down the surface roughness into the sub-micron range, by creating specific and regular pillar-like structures having circular, square, triangle, crossed or fractallike shapes [33]. On the contrary, the combination of substrate roughness and low SFE can be seen as a synergic approach able to achieve contact angle mean value of 158° (Figure 8). Additionally, it was also found that both the original and the actual sample roughness (Figures 3 and 8) do not affect the achieved contact angle, since, in spite of the roughness value, the contact angle value does not change. The same uniformity in the achieved values of the contact angles was observed for measurements performed with CH2I2 in coated samples ( Figure 9). Here, the mean value of the contact angles in coated samples is 148°, regardless of the substrates (0.7-4.6 µm) and the actual roughness (0.3-3.2 µm) of the samples.

Surface Free Energy and Work of Adhesion
The solid SFE and the WA of both uncoated and coated samples were plotted in Figure 10a,b, respectively, as a function of the substrates' roughness. It was found that the solid SFE of the reference samples varies between 44 and 54 mN/m, for roughness values ranging from 0.7 to 2 µm, whereas for roughness of 4 µm the SFE becomes 38 mN/m (Figure 10a). Similarly, the WA varies between 108 and 122 mN/m, assuming a value of 69 mN/m for the sample at 4 µm in roughness. For samples having roughness higher than 4 µm, results cannot be achieved without taking into account the actual surface roughness, e.g., through Equations (12), (13), or more complex relationship, i.e., those described in ref. [33]. Indeed, it was found that roughness higher than 4 µm the Equations' system (4) did not give a solution, so no intersection of the two curves can be observed in graphs of Figure 11b. On the contrary, for lower values of the contact angles, the Equation system (4) reached solutions, as shown in Figure 11a as an example. The authors decided not to further investigate samples at high values of roughness, since values higher than 4 µm are out of interest for aeronautical applications.
On the other hand, the flat trend for SFE and WA was seen for coated samples (Figure 10b), since all values scatter around 0.38 mN/m for the SFE and around 6.1 mN/m for the WA. This result reflects the uniformity of the water contact angle values measured in spite of the original and actual roughness values: the latter changing between 0.3 and 3 m (Figure 3). These trends can be better  highlighted if results are reported as a function of the solid SFE ( Figure 12). For the reference samples (Figure 12a), it was found that, as expected, the higher the contact angle, the lower the SFE and the WA. Since this result was achieved by varying the roughness of substrates alone, it can be concluded that it can be ascribed to the surface morphology (MORPHOLOGICAL EFFECT). On the contrary, for coated samples, no changes in terms of water contact angle and WA, along with almost no variations in the solid SFE (Figure 12b) can be observed, in spite of the actual roughness of substrates. Therefore, it should be concluded that, for the analyzed samples, the chemical modification of the surfaces prevails on the morphological aspect (CHEMICAL EFFECT). The geometrical and chemical effects on the wettability of surfaces were also observed by Kam et al. [31]. They found that the maximum contact angle achieved with nano-microstructured metallic surfaces was 110°, whereas the combination of the surface geometry and its chemical modification with silanes allowed to additionally modify the wettability of surfaces, since the contact angles reached values higher than 140°.
Finally, in Figure 13 the polar and dispersive components of the surface free energy, and the WA assessed for the reference and the coated samples were compared. It highlights that the WA was reduced by 94%, the solid SFE by 99%, and the SP by 100%, since after the application of the coating, it changes from 34 to 0%.

3D samples' Morphology
The SEM images of coated samples 2 and 7 (Table 1), acquired at different magnifications (125-500,000×) are displayed in Figure 14 and Figure 15, respectively. It highlights that at the lowest magnification (125×), both samples exhibit characteristic sponge-like structures [15,19,32,[34][35][36], with the formation of air pockets, which enhance the superhydrophobicity. Samples differ in the density of the sponge-like structures. In fact, it appears that sample 7, having higher roughness (i.e., 2.8 µm)  generates air pockets that are more densely packed than the less rough sample 2 (Ra = 0.4 µm). In spite of this difference, both samples exhibit contact angles of 158° (see Figure 8). At the highest magnification (500,000×) (Figures 14 and 15), the hierarchical micro-and nano-structures appears, highlighting for both samples the dimension of the nanoparticles employed to formulate the coating, namely, about 30 nm in diameter. Therefore, in spite of the intrinsic roughness and the macroscopic density of the sponge-like structure, the generated air pockets made of multiscale roughness (microand nano-meter in dimension) guarantee the superhydrophobicity, i.e., contact angles of 158°.  The surface 3D profilometry of coated samples 2 and 7, having actual roughness of 0.4 and 2.8 µm, respectively, were reported in Figure 16a,b; whereas Figure 16c,d show the corresponding references, having roughness of 0.8 and 4.0 µm, respectively. In the insert, the pictures representative of the water contact angles were also shown. It highlights that all samples exhibit a macroscopic needle-like morphology deriving from the specific textured roughness, and there are no differences in the needle-like morphology [36] for each series of samples (compare Figure 16a with Figure 16c, and Figure 16b with Figure 16d). The reference with low roughness (see Figure 16c) does not exhibit the Cassie state, the surface area in contact with the water droplet is high, and consequently the water contact angle is low (50°). When the reference roughness increases to about 4 µm (see Figure 16d), the relative distance between the surface irregularities and the related height allow the contact angle to increase to 115°. The increase of the water contact angle due to the increased roughness alone can be ascribed to the MORPHOLOGICAL EFFECT. By comparing coated samples with the corresponding references, it highlights that, although the needle-like morphology and the corresponding roughness do not change, the water contact angle increases from 50° and 115°, respectively, to 158°, as a consequence of the chemical modification of the surface (CHEMICAL EFFECT). Finally, by comparing the two coated samples (see Figure 16a,b), it must be noticed that, in spite of the different actual roughness, i.e., 0.4 and 2.8 µm, for the samples 2 and 7, respectively, both of them exhibit the same water contact angle, namely, 158°. Hence, in conclusion, it seems that the chemical effect prevails on the morphological effect.  Figure 17 shows images related to the cutting and tape test performed on coated sample 7. In detail, Figure 17a shows the picture after the cut and Figure 17b shows magnification of the surface after the test. The shape of the water droplet in Figure 17c highlights that the superhydrophobicity was preserved after the test too. It must be noted that, according to the ASTM D 3359-09 [30], the test was passed, since quite no detached flakes of coating could be observed. This result was classified as 5B [30].

Low Temperature Wettability
A preliminary test about the wettability at low temperature of the developed coating was performed by freezing (at −27 °C) a few water droplets placed on both reference and coated sample 2 (pictures in Figure 18). It is interesting to note that the shape of the water droplets, and then of the wettability, does not change with freezing, and consequently, the coated samples display a reduced surface area in contact with the substrate if compared with the reference.

Conclusions
In the present work, the authors studied the wettability of a new developed superhydrophobic coating applied with a layer-by-layer approach as a common aeronautical livery, that can be potentially employed as a passive anti-icing system for aeronautical applications.
It was found that the wettability of the uncoated samples decreases as the roughness increases achieving a value of about 115° for Ra = 4 µm. The increase of the water contact angle due to the increased roughness alone can be ascribed to the morphology changes (MORPHOLOGICAL EFFECT). Instead, high values of contact angle, i.e., about 158°, were achieved only after the application of the coating, regardless of the substrate roughness, ranging from 0.4 to 3 µm (CHEMICAL EFFECT), so highlighting that the chemical modifications prevail on the substrate morphology.
The coating's application reduced the SFE and WA by 99% and 94%, respectively, with respect to the reference. Finally, a preliminary test about the wettability of the developed coating at low temperatures showed that the reduced wettability of coated samples was preserved also at −27 °C. In this regard, future research will be addressed to characterize the developed coating in severe environmental conditions simulating real flights.