Nanosecond laser textured superhydrophobic metallic surfaces and their chemical sensing applications

This work demonstrates superhydrophobic behavior on nanosecond laser patterned copper and brass surfaces. Compared with ultrafast laser systems previously used for such texturing, infrared nanosecond fiber lasers offer a lower cost and more robust system combined with potentially much higher processing rates. The wettability of the textured surfaces develops from hydrophilicity to superhydrophobicity over time when exposed to ambient conditions. The change in the wetting property is attributed to the partial deoxidation of oxides on the surface induced during laser texturing. Textures exhibiting steady state contact angles of up to ∼152◦ with contact angle hysteresis of around 3–4◦ have been achieved. Interestingly, the superhydrobobic surfaces have the self-cleaning ability and have potential for chemical sensing applications. The principle of these novel chemical sensors is based on the change in contact hemical sensor angle with the concentration of methanol in a solution. To demonstrate the principle of operation of such a sensor, it is found that the contact angle of methanol solution on the superhydrophobic surfaces exponentially decays with increasing concentration. A significant reduction, of 128◦, in contact angle on superhydrophobic brass is observed, which is one order of magnitude greater than that for the untreated surface (12◦), when percent composition of methanol reaches to 28%. © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license


Introduction
Controlling surface wettability has attracted increased research attention due to the wide range of applications ranging from domestic innovations such as self-cleaning glass to hightechnology fields like microfluidic control, corrosion resistance, water-oil separation and drag reduction [1][2][3][4]. New applications are continuously being discovered, for instance, the use of low wetting surfaces for the realization of microdroplet lasers [5,6]. Recently, the developments in designing and preparing novel surfaces with special wettability have opened opportunities to employ these surfaces for liquid transportation and smart sensing devices [7,8].
Today, superhydrophobic surfaces on glasses, semiconductors, polymers and metals can be generated by various methods, including lithography, plasma treatment, chemical deposition, colloidal assembling, templating, and ultrafast laser surface texturing [9]. Among these options, laser patterning is a promising method due to the potentially high fabrication speeds, low waste and maskless single-step processing [10][11][12][13]. This technology is also highly suitable for producing superhydrophobic surfaces with selective area patterning [14,15].
Superhydrophobic metallic surfaces based on laser texturing have been investigated previously [16,17]. However, to date, most of the reported works rely on high-cost, complex ultrafast (femto-/pico-second) lasers [16][17][18][19][20][21][22][23]. For these lasers, the pulse energy is relatively low so high pulse overlapping is required for creating the surface modification and therefore processing rates are slow. For real industrial applications, the ability to create superhydrophobic metallic surfaces using compact, robust and cost-effective alternatives with fast processing times, such as nanosecond fiber lasers, would be highly desirable.
Copper and brass are widely used for electronic components and industrial devices due to their high electrical and thermal conductivity. However, they are easily affected by environmental conditions, such as high humidity leading to corrosion. Recently  has been shown these problems can be prevented on these materials by creating a superhydrophobic surface [24].
In this paper, the superhydrophobic behavior of copper and brass surfaces achieved by infrared nanosecond laser surface texturing is presented. The superhydrophobic surfaces demonstrate the self-cleaning properties and have potential for liquid chemical sensing applications.

Materials
The experiments detailed in this paper were performed on commercially available copper (purity 99.9%) and brass sheets (RS Components, CZ121M, 58% Cu, 39% Zn, and 3% Pb) with thicknesses of 0.45 and 0.6 mm, respectively. Prior to laser radiation, the samples were cleaned with isopropanol.

Surface laser irradiation
Hierarchical structures on the sample surface were fabricated using a SPI fiber laser (20 W EP-S) system including a galvanometer scanning system to move the beam over the surface as shown in Fig. 1 [25]. Arrays of microgroove structures were directly written on the sample surface. The distance between adjacent microgrooves (pitch size, d) was kept constant. The sample was irradiated, normal to the surface, by the fiber laser with nominal beam spot of about 21 m, wavelength of 1064 nm and pulse width of ∼220 ns. A scanning speed of 75 mm/s and repetition rate of 25 kHz were used. This processing rate is about two orders of magnitude higher compared with that using femtosecond lasers [16,18]. Three different fluences were applied separately to the work piece for copper (75, 84, 93 J/cm 2 ) and brass (55, 65, 75 J/cm 2 ).

Surface analysis
The samples were left in air, under ambient conditions before studying their surface wettability by measuring the contact angle. A ∼5 L droplet of deionized water was dispensed onto the sample surface using a syringe while an image was captured by a camera (Unibrain 1394) combined with a 12× magnification system. The contact angle (Â) was then determined by analyzing the droplet images using the software FTA32 (version 2.0). The contact angle hysteresis ( Â) was estimated by comparing the advancing and receding contact angle. The rolling-off angle was estimated as the tilted superhydrophobic surface on which deposited droplets start to rolling downwards. In order to study the surface morphology of the laser textured samples, scanning electron microscopy (SEM) and optical microscopy (Leica DM6000M) were used. The chemical composition of the surface was investigated by energy-dispersive X-ray (EDX) measurements.

Chemical solution for sensing demonstration
Methanol (99.5% purity) was dissolved in deionized water to form methanol solutions with various concentrations. ∼5 L droplets of these solutions were dispensed on laser textured brass surfaces that had previously been exposed to ambient conditions for 38 days. The contact angle was determined as described in Section 2.3.

The effect of laser irradiation on surface structure
Prior to fabricating the microgroove structures, the effect of individual pulses on the surface was studied. Fig. S1 shows the nearcircular craters induced on the copper surface by single pulses of laser radiation. The craters are formed due to the rapid heating of the copper, resulting in melting and evaporation (ablation), at the location of the laser radiation. Most of the material from the center of the crater was redeposited in the area surrounding the crater. The total amount of removed material depends on laser fluence.  S3c) overlapping. With a repetition rate of 25 kHz, the microgroove structure was formed with a significant amount of recast material around the perimeter of the feature. Fig. 2a shows a SEM image of the completed microgrooves on brass, irradiated with 65 J/cm 2 , demonstrating the effect of the laser radiation on creating regular microstructures with micro-roughness on the sample's surface. Depth of the microgrooves on copper and brass surfaces, measured using optical microscopy, varied from around 10 to 25 m, depending on laser fluence. Fig. 2b shows a threedimensional image of the structure on brass surface irradiated with 65 J/cm 2 indicating that the depth of microgrooves was about 20 m. It is noted that, due to high pulse overlap, the depth of the channels is larger than that of the craters created by a single pulse.

Time effects on surface wettability
The laser treated surfaces becomes hydrophilic immediately after fabrication (see Supporting Video 1 for more information). However, the surface wettability decreases over time, as indicated by the increase in contact angle. The first measurement after processing was made after the samples were left under ambient conditions for 3 days. Fig. 3 plots the evolution of contact angle with time for copper and brass samples with d = 75 m, irradiated by three different fluences. Each data point presented is average value of three individual measurements. It can be seen that the contact angle exhibits a sharp increase during the first ten days, which then slows to a gradual growth before finally reaching a steady state. The steady state was reached after about 11 days for copper and brass samples irradiated with 75 and 55 J/cm 2 , respectively. However, it took much longer, approximately 30 days, for copper irradiated with 93 J/cm 2 .
Similar to ref. [16], the development of the contact angle over time is found to be well-fitted (as shown in Fig. 4a) with an exponential growth function expressed as: Â = Â eq (1 − exp(−t/ˇ)), where Â eq and ˇ are equilibrium (steady state) contact angle and time constant, respectively. It is noted that the value of Â eq was selfestimated for reducing standard error of ˇ (which indicates the time when the contact angle reaches 63% of its maximum). Fig. 4b and c plots Â eq and ˇ for copper and brass samples, respectively. It can be seen that the equilibrium contact is slightly above 150 • for all samples except the copper irradiated by 93 J/cm 2 . Samples irradiated with the lowest fluences (55 J/cm 2 for brass and 75 J/cm 2 for copper) develop from hydrophilic to superhydrophobic much faster than samples treated by higher fluences, which has been similarly observed in other metals [16]. For copper, the time constant was 3.7 days under 75 J/cm 2 . This value significantly increased to 13.5 days for samples irradiated with 93 J/cm 2 , an increase of nearly 265%. The evolution process was faster for brass samples, which took 2.9 and 5.6 days to reach 63% of steady state contact angle for fluences of 55 and 75 J/cm 2 , respectively.
Thermodynamic modeling has proven that the level of superhydrophobicity depends on groove depth and pitch size [26]. In this work, due to increased groove depth and strong modification of surface chemistry, the surface wettability was found to increase with laser fluence. To investigate the effect of pitch size on the It has been demonstrated that, although the laser textured copper and brass surfaces become hydrophilic directly after fabrication, the wettability changes to hydrophobic over time. This effect is highly reproducible. Fig. S4 compares the development from hydrophilicity to hydrophobicity for two different set of brass samples with d = 75 m and irradiated with 75 J/cm 2 (the data has been presented in Fig. 3b and Fig. 5). The result indicates very similar evolution trends.

Superhydrophobic behavior
A high contact angle (>150 • ) is only one important characteristic of a superhydrophobic surface. Another vital factor is small contact angle hysteresis or rolling-off angle (not exceeding 10 • ) [27]. This value can be measured by comparing advancing and receding angles of a droplet on a tilted surface [9]. Alternatively, it can be determined by subtracting the two contact angles of growing and shrinking droplets [16]. Fig. 6 shows dynamic of contact angle on a brass surface irradiated by 55 J/cm 2 after it reaches steady stage (see Supporting Video 2 for demonstration of the process). First, the droplet was suspended from a syringe (Fig. 6a) and then made to contact the surface by lifting up the sample with a translation stage (Fig. 6b). The droplet did not wet the surface and exhibits sliding on the surface when the substrate reached a certain height (Fig. 6c). The droplet then deformed when the surface is lowered (Fig. 6d and e) and finally departed from the surface (Fig. 6f). This behavior is found to be the same for all textured copper and brass samples. It indicates that there was negligible adhesion between the droplet and the surface and, therefore, suggests Â is small. Actually, the contact angle before the droplet begins sliding on (Fig. 6b) and departing from (Fig. 6e) the surface can be considered as maximum and minimum static contact angles, and these values should be close to advancing and receding contact angles, respectively. With this consideration, it is estimated Â for all samples was found to be in a range of 3-4 • , which meets the requirement of a superhydrophobic surface.
The obtained low value of Â is supported by rolling-off angle measurements. Fig. 7 shows a time sequence of a droplet rolling on a 2.9 • tilted hydrophobic brass surface irradiated with 55 J/cm 2 (see Supporting Video 3 for a demonstration of the process). The droplet was suspended then dropped from a syringe and started rolling down the incline on contact with the surface with rolling speed of about 8 cm/s. It was also found that rolling-off angle for many of the samples could be as small as 2 • . The potential to create such laser textured superhydrophobic surfaces on copper or brass could  enable the fabrication of devices which are highly water resistant, potentially improving corrosion resistance [24].

Explanation of change in wettability
Wettability of rough surfaces has been under investigation for decades [28,29] and the change in wettability of metallic surfaces due to laser texturing and after exposing to the air has been previously reported [16]. Several mechanisms have been proposed such as the decomposition of carbon dioxide into carbon with active magnetite, adsorption of organic matter from the atmosphere and creation of hydrophobic groups [16,30,31]. In our work, directly after fabrication, the surfaces exhibit hydrophilic behavior with very small contact angles. This effect is attributed to the modification of both surface morphology and surface chemistry during the laser processing. Roughness was introduced via ablation (vaporization) which increases the wettability accordingly to the Wenzel equation [28].
where, r > 1 is the roughness factor. Â w and Â f are contact angles on rough and flat surfaces, respectively. Eq. (1) also predicts the change in surface chemistry, which induces Â f , would lead to a change in Â w (Wenzel contact angle on a rough surface) as follows [32]: formed by the reaction between copper and zinc with oxygen. The formation of CuO (a hydrophilic material [33]), in conjunction with the roughness, should increase the wettability of laser textured surfaces and therefore explains the high wetting behavior observed on the surfaces shortly after processing with laser radiation. However, when exposed to ambient conditions the wettability of the laser textured surfaces changed with time. This phenomenon is ascribed to modification of the surface chemistry as there is no change in surface morphology over time. Even though the adsorption of organic matter from the atmosphere [31] is a considerable factor, measurements suggest that the deoxidation proposed by Chang et al. [33] should effectively contribute to the development from hydrophilic to superhydrophobic. Fig. 8 indicates decrease of the oxygen component (from 9.0 to 4.5%, 19.8 to 16.2%, and 20.2 to 19.5%, for fluences of 55, 65 and 75 J/cm 2 , respectively) after leaving the samples under ambient conditions for 38 days. This observation is understandable by the partial deoxidation of CuO into Cu 2 O (2CuO ↔ Cu 2 O + 1/2O 2 ) [33]. It is suggested that the deoxidation of the CuO is a heterogeneous nucleation and growth mechanism that initiates at defects, therefore the surface will consist of different regions of different wettability (CuO and Cu 2 O) that are of a scale much smaller than the droplet. The mole ratio of Cu 2 O to CuO increases with time, and as Cu 2 O is a hydrophobic material [33], the deoxidation process causes the change in wetting behavior. It is noted that we did not observe a clear relationship between the surface wettability and carbon content, as discussed previously [16]. When the contact angle reaches a steady state, the surface wetting becomes incomplete, suggesting a transition from Wenzel to Cassie-Baxter regime. The contact angle of droplets now can be described by the general Cassie-Baxter equation [29].
where f 1 and f 2 are local densities, function of surface (x, y) coordinates. f 1 + f 2 = 1. Â 1 and Â 2 are contact angles corresponding to Cu 2 O (suppose that the Cu 2 O is dominant to CuO) and air. It is noted that the deoxidation mechanism also provides an explanation for the observation that samples irradiated with a lower fluence become superhydrophobic more quickly as it is expected that for low incident energy the CuO layers are thinner which make them transition to Cu 2 O more readily.

Self-cleaning effect
Self-cleaning is an important application of superhydrophobic surfaces. This effect is exploited in nature in many plants because it helps to wash away contaminants and acts as a defense against pathogens [34]. For artificial superhydrophobic surfaces self-cleaning behavior is also has significant benefit for practical applications [1]. Consequently, the nanosecond laser surface texturing process developed for copper and brass in this work is a viable route to achieving this property. In Supporting Video 4, the self-cleaning effect on a 6 • tilted superhydrophobic brass surface is demonstrated. CaCO 3 powder served as a contaminant on the surface [24]. Water droplets were dropped on to the top part of the surface where they rolled quickly down to the bottom, taking away the fine particles. The dusty surface became clean after depositing several of droplets. In comparison, the unprocessed surface (central region of the sample) did not exhibit the self-cleaning ability and droplets remained on the surface due to strong adhesion.

Chemical sensing applications
Superhydrophobic surfaces have also been considered for sensing applications such as thermal and pH sensors [35,36]. In this work, the potential of using the nanosecond laser textured superhydrophobic surfaces for a novel chemical sensing application is demonstrated.
It is well known that the wettability of a surface is not only dependent on the surface morphology and chemistry but also on the liquid [37]. Due to the smaller surface tension for less polar liquids compared with water, the contact angle of water droplets can be much higher than droplets of less polar liquids (such as octane and methanol) on the same surface [37]. It therefore follows that, for solutions which are mixtures of water and a less polar liquid, the contact angle should decrease with increase of concentration of the less polar liquid. As a result, a solution concentration can be determined by simply monitoring the contact angle and hence provides a mechanism to develop chemical sensors. Fig. 9 compares the contact angle of a methanol/water solution on an unprocessed as-received brass substrate and superhydrophobic brass surfaces irradiated with 55 and 75 J/cm 2 and left in ambient conditions for 38 days. Again, each data point presented is an average value of three individual measurements. It is clear that the contact angle decreased with increasing methanol concentration. The relationship was well fitted with a linear function for the unprocessed substrate which exponentially decays for the laser textured surfaces. When the concentration rose from 0 to 15%, a slight reduction in the contact angle for all samples is expected. However, at higher concentrations, a significant decrease in contact angle was observed for the nanosecond laser textured samples. When the concentration is increased from 0% to 44%, the contact angle drops sharply from 150 • to 40 • for a sample irradiated with 55 J/cm 2 , a major decrease of 110 • compared with 24 • (from 70 • to 46 • ) for the unprocessed sample. The sample irradiated with 75 J/cm 2 demonstrated the highest sensitivity. There was a significant decrease of 128 • in contact angle (from 153 • to 25 • ), which is one order of magnitude higher compared with only 12 • for the unprocessed surface, when concentration of methanol reaches 28%. The drop in contact angle of the methanol solutions due to surface tension change is enhanced by the roughness factor. It is noted that the sensing process does not appear to be completely reversible. Substrates used for the sensing measurement lose their superhydrophobic properties after the experiment. The methanol solution seems to chemically modify the hydrophobic surface, thus resulting in lower contact angles and it appears to be a one-way reaction. The change is not entirely permanent; the surface wettability decreases when exposed to the air afterwards, but the development of this process is extremely slow compared with that of new laser structured surfaces.

Conclusions
It has been shown that textured superhydrophobic surfaces on copper and brass can be achieved by using a compact and cost effective nanosecond fiber laser. Directly after fabrication, the samples exhibit hydrophilic behavior but gradually become superhydrophobic over time with the maximum contact angle being achieved after ∼11 days. The hydrophilic property is ascribed to the laser induced roughness. However, this role of surface morphology changes over time, which results in the surfaces becoming superhydrophobic. The change in wettability is due to surface chemistry, in which partial deoxidation of CuO into Cu 2 O on the surface is likely a key factor. The steady state contact angle is about 152 • with low contact angle hysteresis of 3-4 • . The superhydrophobic surfaces have been demonstrated to exhibit self-cleaning behavior and show potential for liquid chemical sensing applications. The sensor's sensitivity is defined as the decrease of contact angle with increase of chemical concentration. The sensitivity for superhydrophobic surfaces is one order of magnitude higher compared with that for unprocessed surface. Overall, this work demonstrates a cost effective, direct write, environmentally friendly (low waste) and fast processing method to modify wettability of common metals for multifunctional applications.