5044
Langmuir 2008, 24, 5044-5051
Nanotexturing of Polystyrene Surface in Fluorocarbon Plasmas: From Sticky to Slippery Superhydrophobicity Rosa Di Mundo,*,† Fabio Palumbo,‡ and Riccardo d’Agostino†,§ Department of Chemistry, UniVersity of Bari, Institute for Inorganic Methodologies and Plasmas (IMIP)-CNR, Plasma Solution srl, Via Orabona 4, 70126 Bari, Italy ReceiVed January 8, 2008. In Final Form: February 12, 2008 In this work plasma etching processes have been studied to roughen and fluorinate polystyrene surface as an easy method to achieve a superhydrophobic slippery character. Radiofrequency discharges have been fed with CF4/O2 mixtures and the effect of the O2:CF4 ratio, the input power, and the treatment duration have been investigated in terms of wettability, with focus on sliding performances. For this purpose, surface morphological variations, evaluated by means of scanning electron microscopy and atomic force microscopy, together with the chemical assessment by X-ray photoelectron spectroscopy, have been correlated with water contact angle hysteresis and volume resolved sliding angle measurements. Results indicate that by increasing the height and decreasing the density of the structures formed by etching, within a tailored range, a transition from sticky to slippery superhydrophobicity occurs. A short treatment time (5 min) is sufficient to obtain such an effect, provided that a high power input is utilized. Optimized surfaces show a unaltered transparency to visible light according to the low roughness produced.
Introduction During the past decade hundreds of person-years have been spent in fabricating as well as in investigating the behavior of superhydrophobic surfaces. The enormous interest toward such materials arises mainly from industry demand on microfluidic devices (often referred to as lab-on-a-chip) and benchmark needs of self-cleaning surfaces. In microfluidic technology, surface effects are much more important than the volume/gravity ones and superhydrophobic domains are utilized to force the liquid front to other directions.1,2 The so-called self-cleaning surfaces are characterized by low adhesion both with water (water repellent character) and with contaminating particles (antiadhesive properties). On these surfaces, water forms spherical droplets, which roll off very quickly even at small inclinations; through such motion they can remove dirt particles whose adhesion to water is favored.3,4 Polystyrene is a very popular polymer; besides the numerous applications for fabrication of everyday life objects, its utilization in microfluidic devices production requires the developing of easy processes to achieve superhydrophobicity. Moreover, potential utilization as plastic for exterior uses could be enhanced by a self-cleaning surface. It is well-known that superhydrophobicity, namely, a property of surfaces with water contact angles (WCA) higher than 150°, can be only achieved by adding a texture to a hydrophobic surface, in other words by increasing its roughness. This effect is shown in nature by lotus, other plants leaves, and insects body. In particular, for lotus leaves a superimposition of a micro- and nanotexture has been reported to be relevant to superhydrophobicity.5
Two descriptions have been proposed for the dependence of the wetting behavior on surface roughness: the Wenzel and the Cassie-Baxter models. The Wenzel model holds for cases where the liquid remains in contact with the whole solid surface. The enhancement of hydrophobicity (but also hydrophilicity) is quantitatively described by the Wenzel equation:
cos θW ) r cos θE
(1)
where θW is the observed angle on the rough surface, θE is the equilibrium contact angle on the flat surface of the same chemical character, and r, with r g 1, is the ratio of the actual surface area to the projected area of the surface. This model is based on the idea that the increased surface area of a hydrophobic solid results in a raise of the contact angle, therefore a linear relationship, with a slope equal to r, is theorized.6 The Cassie-Baxter model holds for surfaces with a topography such that water cannot deeply penetrate and wet the whole surface, thus air is trapped into the grooves under the droplet, which is then suspended across the surface protrusions. In such a case, the droplet is in contact with a composite surface (solid and air) and the observed contact angle θCB, according to the CassieBaxter equation, is given by the linear combination:
cos θCB ) φs cos θs + φair cos θair
(2)
By considering the air area fraction as φair ) 1-φs, θair ) π, and assuming that the WCA of the solid fraction, θs, corresponds to that of the flat surface, θE, eq 2 can be rewritten as:
cos θCB ) φs cos θE - (1-φs)
(3)
* Corresponding author. E-mail:
[email protected]. Fax: +390805443405. † Department of Chemistry, University of Bari. ‡ Institute for Inorganic Methodologies and Plasmas (IMIP)-CNR. § Plasma Solution srl.
Therefore, in this model, the hydrophobic character of a rough surface is emphasized by the decrease of the solid-liquid contact area.7
(1) Tabeling, P. Introduction to Microfluidics; Oxford University Press: New York, 2006. (2) Thorsen, T.; Maerkl, S.; Quake, S. Science 2002, 298, 580. (3) Callies, M.; Que´re´, D. Soft Matter 2005, 1, 55. (4) Fu¨rstner, R.; Barthlott, W. Langmuir 2005, 21, 956-961.
(5) Cheng, Y. T.; Rodak, D. E.; Wong, C. A.; Hayden, C. A. Nanotecnology 2006, 17, 1359. (6) Wenzel, T. N. J. Phys. Colloid Chem. 1949, 53, 1466. (7) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546.
10.1021/la800059a CCC: $40.75 © 2008 American Chemical Society Published on Web 04/02/2008
Nanotexturing of Polystyrene Surface
Langmuir, Vol. 24, No. 9, 2008 5045
Beside the difference between the angle predicted by the two equations, which are thoroughly compared elsewhere,8,9 what mainly differentiates surfaces which fall in the Wenzel or in the Cassie-Baxter model, is the extent of the water contact angle hysteresis (WCAH), a concept not handled at all in their theories. WCAH is defined as the difference between the maximum and the minimum WCA values measured in a dynamic mode onto a surface, i.e., the maximum angle while the droplet is increasing its volume (advancing angle, θa) and the minimum angle just before the reduction of the contact area with the solid during the decreasing volume (receding angle, θr). WCAH is originated by the presence of heterogeneities (in topography and/ or chemical composition) which induce fluctuations in surface tension.10 For a drop at rest on a tilted plane, the contact angle is larger at front than at rear. Generally, the front and rear angles are considered equal to the advancing and receding angles measured in dynamic mode on a horizontal plane. The difference generates a force opposing the weight of the drop and, if the drop is small enough, will be able to balance it. Such a sticking capillary force, fc, has been defined as:
fc ) πbγlv(cos θr - cos θa)
(4)
where b is the radius of the solid-liquid contact and γlv is the liquid-vapor surface tension. The drop will remain stuck on the surface provided that :
πbγlv(cos θr - cos θa) g FgV sin R
(5)
where R is the angle of inclination, F the liquid density, g the gravity acceleration, and V the drop volume.11 As a consequence of such a phenomenon, a surface with high WCAH shows a sticking behavior while a surface with a low WCAH has a slippery character. Since hysteresis originates from defects, generally a high WCAH is expected for all rough materials. But what comes out from experience is that a surface in the Wenzel regime shows a high WCAH and therefore it is a sticky surface while a Cassie-Baxter surface has a low WCAH leaving the drop to roll off (slippery surface). Only qualitative explanations are given about this difference and mainly relate to the fact that when water is in contact with the whole surface profile (Wenzel regime), the structures act as obstacles to water motion and make it stable on the surface; it is possible to figure out that in such a context the higher the structures are the more the water is retained. On the other hand, for a nonwetted contact (Cassie-Baxter regime), the structures cannot hinder water motion as only their top surface is touched by water; in this case, the lower the contact with surface (φs) the less the drop is stable and its motion is more favored. Obviously, research efforts are mainly addressed toward a Cassie-Baxter surface as it is the one that gives significance to the concept of superhydrophobicity by effectively resulting in a water repellent behavior. As rationalized by D. Que`re`,12 who derived from eq 5 a separation between sticking and sliding drops, surfaces with low hydrophobic character can also show a sliding behavior provided that the hysteresis is very low; highly hydrophobic surfaces can be slippery also at relatively high hysteresis. (8) Patankar, N. A. Langmuir 2003, 19, 1249. (9) Que´re´, D.; Lafuma, A.; Bico, J. Nanotechnology 2003, 14, 1109. (10) De Gennes, P. G. ReV. Mod. Phys. 1985, 57, 827. (11) Furmidge, C. G. L. J. Colloid. Sci. 1962, 17, 309. (12) Que´re´, D. Rep. Prog. Phys. 2005, 68, 2495.
Several methods have been proposed to introduce a texture onto a material: various lithographic approaches,4,13-14 deposition of coatings with a structured topography,15 assembling of micro/ nanoparticles,16 and recently also the use of a natural template like the lotus leaf17 among others. Since those are generally time-consuming or multistep procedures, the research community is searching more simple and shorter procedures to produce surfaces with such performances. It is understood that plasma etching processes can increase roughness (and therefore the hydrophobic character) of polymers exploiting their heterogeneity and amorphous character. Generally, rough not ordered topographies are obtained. These processes do not consist uniquely of massive etching but chemical surface grafting is always present. Plasmas fed with oxygen and fluorine containing gases have been thoroughly investigated since the 80’s for the etching (ashing) of polymers utilized as resists in microelectronic fabrication. It was then understood that the variation of the O2:CF4 ratio allows to properly tune the density of fluorine/oxygen atoms within the discharge, which is an important property for controlling the surface and in turn the polymer etching.18 Some works have been published on the use of etching processes for surface roughening; generally a fluorocarbon-based chemistry, known for the intrinsic hydrophobic character, is superimposed to roughening. CF4 plasmas have been tested onto spin-coated polybutadiene,19 SF6 and then O2 plasmas on polydimethylsiloxane (PDMS) (followed by C4F8 deposition),20,21 CF4 and O2 plasmas onto polyethylene (PE),22 and O2 plasma followed by deposition of an organosilicon fluorinated compound onto polyethyleneterephtalate (PET).23 Superhydrophobic surfaces have been obtained in some cases. The present work focuses on the surface roughening of polystyrene (PS), which is particularly appealing for the potential impact in microfluidics. In particular, CF4-O2 fed discharges were investigated with the aim of tailoring a one-step and fast process, therefore compatible with low cost industrial fabrication needs. Attention was also paid as not to hinder intrinsic transparency of the polymer. The effect of the O2:CF4 ratio in the feed mixture, the input power, and the treatment duration have been evaluated in detail in terms of wettability performances by dynamic water contact angle and sliding angle measurements. Results have been correlated with chemical and especially with morphological characterizations with the particular purpose of better explaining WCAH variations.
Experimental Details Plasma Reactor and Operating Conditions. Plasma treatment experiments were run in a RF capacitively coupled parallel plate (13) Cao, L.; Hu, H.; Gao, D. Langmuir 2007, 23, 4310. (14) Yoshimitsu, Z.; Nakajima, A.; Watanabe, T.; Hashimoto, K. Langmuir 2002, 18, 5818. (15) Favia, P.; Cicala, G.; Milella, A.; Palumbo, F.; Rossini, P.; d’Agostino, R. Surf. Coat. Technol. 2003, 169/170, 609. (16) Han, J. T.; Kim, S.; Karim, A. Langmuir 2007, 23, 2608. (17) Sun, M.; Luo, C.; Xu, L.; Ji, H.; Oujang, Q.; Yu, D.; Chen, Y. Langmuir 2005, 21, 8978. (18) Egitto, F. D. et al. In Plasma deposition, treatment, and etching of polymers; d’Agostino, R., Ed.; Academic Press: Boston, 1990. (19) Woodward, I.; Schofield, W. C. E.; Roucoules, V.; Badyal, J. P. S. Langmuir 2003, 19, 3432. (20) Tserepi, A. D.; Vlachopoulou, M-E.; Gogolides, E. Nanotechnology 2006, 17, 3977. (21) Tsougeni, K.; Tserepi, A.; Boulousis, G.; Constantoudis, V.; Gogolides, E. Plasma Processes Polym. 2007, 4, 398. (22) Fresnais, J.; Chapel, J. P.; Poncin-Epaillard, F. Surf. Coat. Technol. 2006, 200, 5296. (23) Teshima, K.; Sugimura, H.; Inoue, Y.; Takai, O.; Takano, A. Langmuir 2003, 19, 10624.
5046 Langmuir, Vol. 24, No. 9, 2008
Figure 1. Schematic diagram of the plasma reactor.
plasma reactor (scheme in Figure 1). It consists of a cylindrical stainless steel vacuum chamber (internal diameter of 50 cm and height of 27 cm), which is pumped by a turbomolecular rotary pumping system. The pressure is monitored by a capacitive membrane transducer (MKS instruments baratron). The gas flow rates are controlled by electronic mass flow controllers (MKS instruments). The plasma is ignited with a 13.56 MHz power supply (Advanced Energy) through an impedance matching unit. The upper stainless steel electrode (10 cm diameter) is grounded while the bottom electrode (12 cm diameter, spaced 55 mm apart), on which the substrates are placed, is RF powered. The feed gas is admitted through a shower ring positioned close to the upper electrode. Polystyrene (PS) slices cut from PS Petri dishes (Bibby Sterilin) and ethanol sonicated were used as process substrates. Glow discharges were fed with mixtures of CF4 and O2, with a constant total flow rate of 60 sccm, and at a constant pressure of 60 mTorr. The flow rate of O2 was varied in the range 0-60 sccm, and therefore the O2 % in the CF4-O2 mixture was investigated in the range 0-100%. Also investigated were the effect of input power (50-300 W) as well as the treatment duration. Morphological Characterization. Scanning electron microscopy (SEM) was used to observe the morphology of samples. Specimens were gold metallized by sputter coating (Bio-Rad Polaron) and observed at 20 KV with a Cambridge SEM 360 scanning electron microscope. An image processing software (ImageJ, NIH) was utilized to evaluate the distribution of the top area of the structures from the acquired images. Atomic force microscopy (AFM) was utilized to have numerical indication of the root-mean-square (rms) roughness and maximum height of the structures. Images were acquired with a Thermomicroscope Autoprobe CP in noncontact mode using gold coated silicon and conical high resonance frequency probe tip. X-ray Photoelectron Spectroscopy (XPS). XPS analyses were carried out by means of Thermo Electron Corporation Theta Probe spectrometer with a monochromatic Al KR X-ray source (1486.6 eV) at a spot size of 400 µm corresponding to a power of 100 W at a take off angle of 37°. Survey (0-1200 eV) and high-resolution spectra (C 1s, O 1s, F 1s) were recorded at a pass energy of 200 and 150 eV, respectively. The C 1s signal for C-C(H) bonds, with a binding energy of 284.7 eV, was used as an internal standard for the correction of the samples charging. The atomic percentages reported have been calculated by using the sensitivity factors of the instrument from the high-resolution peaks. Best fitting of the C 1s XPS line shape has been performed by means of the instrument software (Avantage). For the C 1s XPS spectra, the following six components have been used:
Di Mundo et al.
Figure 2. Advancing (θa) and receding (θr) water contact angle of the polystyrene (PS) samples treated with CF4-O2 plasmas at 150 W for 5 min as a function of the O2% in the feed mixture. The solid straight line indicates the mean angle (θa + θr)/2 for untreated PS.
C-C(H) (C1, 284.7 ( 0.1 eV), C-CF/C-O (C2, 286.3 ( 0.2 eV), CdO (C3, 287.8 ( 0.2 eV), CF (C4, 288.7 ( 0.2 eV), CF2 (C5, 291.1 ( 0.2 eV), CF3 (C6, 293.2).24 The full-width at halfmaximum of each line shape was allowed to vary in the range 1.5-2.0 eV. Wettability Evaluation. Water contact angle (WCA) in dynamic mode was measured by means of a Rame`-Hart 100 goniometer. Advancing and receding angles were measured by depositing a droplet of 1 µL on the surface, then increasing the volume to 4 µL, and finally decreasing it. Advancing angles are the maximum angles observed during the droplet growth. Receding contact angles are the ones just before the contact surface reduction. Each WCA value has been averaged from measurements of four drops with an estimated maximum error of 3°. A CAM200 digital goniometer (KSV instruments) equipped with a BASLER A60f camera was used to evaluate the sliding angle of the samples. A volume resolved sliding angle evaluation was performed by measuring the sliding angle with water drops of different volume ranging from 1 to 20 µL. A distilled water drop was deposited on the specimen fixed to a tiltable plate, and then the plate was inclined slowly until the drop started to move. Angles have been digitally evaluated from the acquired image sequence through the instrument software (CAM200). Surface Transparency. A UV-vis spectrophotometer (Jenway 6505) was utilized to evaluate surface transparency variations by measuring the transmittance of the samples in the range 2001100 nm.
Results and Discussion Feed Mixture Effect. The advancing and receding WCA values of the PS samples treated with CF4-O2 plasmas at 150 W for 5 min with different O2 percentages in the feed mixture are reported in Figure 2. The WCA advancing and receding values for native PS are 94 and 92°, respectively. It is worth mentioning that untreated PS shows such a low hysteresis (2°) value only when freshly washed, otherwise much higher values are found; as mentioned above, WCAH depends on surface heterogeneity, which can be only increased by the presence of dirt particles. Generally higher WCAH are found for all the treated samples under these conditions (in the range 16-34°). An increase of absolute advancing and receding WCA values is observed when the O2 content in the feed is lower than 40%, and therefore (24) Beamson, G.; Briggs, D. High Resolution XPS of Organic Polymers; John Wiley & Sons: New York, 1992.
Nanotexturing of Polystyrene Surface
Figure 3. XPS F:C and O:C atomic ratios of PS samples treated with CF4-O2 plasmas at 150 W for 5 min as a function of the O2% in the feed mixture.
Figure 4. XPS C 1s signals of the 100% CF4 (a) and the 17% O2 (b) treated PS samples (150 W, 5 min).
an increase of hydrophobicity is found in this range. Interestingly, a maximum in the WCA trend is obtained for the 17% O2 mixture. The XPS F:C and O:C atomic ratios of samples treated under the same conditions are reported in Figure 3, where it can be observed that at the 17% O2 feed mixture the maximum F:C ratio (and also the maximum absolute fluorine percentage, with 44.2% fluorine and 49.4% carbon) is obtained. The oxygen atomic concentration in the same sample is 6.4%, just slightly higher than the ones found in the 100% CF4 feed (5.3%) and in the native PS (5.8%). The higher effectiveness in surface fluorination shown by the 17% O2 mixture with respect to the 100% CF4 can also be appreciated in Figure 4 (panels b and a, respectively), where the best fitted C 1s signals are compared. The high binding energy components are much more intense in the 17% O2 sample. Though unambiguous assignment of the best-fitting components is generally difficult when both fluorine and oxygen are present, the increase of C4, C5, and C6 is surely mainly due to the above reported fluorine-containing groups as the oxygen percentage is almost unchanged. This result could be likely due to an increased plasma activation of CF4, with scavenging of carbon and enrichment in fluorine atoms, when a small amount of O2 is introduced, as it is found in previous literature on CF4 containing plasmas.18 The higher fluorine content of the 17% O2 sample
Langmuir, Vol. 24, No. 9, 2008 5047
with respect to the 100% CF4 can only partially explain the higher water contact angle values. As highlighted above, the presence of a texture is necessary to achieve WCA values higher than 110-120° (as in the case of these surfaces), and it can also strongly affect the wettability difference. Coming back to Figure 2, some considerations should be done on surfaces obtained for O2 content in the feed higher than 20%. Clearly WCA decreases reaching hydrophilic values lower than the ones found in the native substrate when O2 % is higher than 40%. Beside eventual morphological implications, the WCA decrease can be safely ascribed to the reduced grafting of fluorine on the surface and to the increase of the O:C ratio, as highlighted in Figure 3. This result was not unexpected, since with increasing O2 dilution, the concentration of fluorinated species in the plasma decreases and the oxygen grafting of the sample prevails. The SEM images (as top view) of the PS samples treated under different feed conditions (with the corresponding rms roughness values) at 150 W for 5 min are reported in Figure 5 along with the one of the native sample (panel a). It is clear that under all feed conditions the plasma treatment is able to generate a texture on the surface. Density and distribution of the structures clearly vary with the feed mixture. In particular, a more significant texture is found for the oxygen containing feeds. rms roughness values found are not high, and in the case of the most hydrophobic sample (17% O2), it is 9 nm, whereas for the native polymer it is 3 nm. Since the sample obtained with an O2 content of 17% lead to the highest WCA, though not to the highest roughness, that mixture has been chosen for better optimization of the process. The hysteresis found (22°) for this surface is quite high and therefore, considering the contact angle theories, it should be concluded that its superhydrophobic character comes from a solid-water wetted contact (Wenzel regime) leading to sticking behavior. Power Effect. In Figure 6, advancing and receding WCAs for PS samples treated with a fixed CF4-O2 mixture (17% O2) in a 5 min process are reported as a function of input power. In the power range 50-150 W, it can be observed that both advancing and receding angles increase and in turn also the hydrophobic character (θa at 150 W is 155°), with a fairly constant hysteresis. In the range 150-275 W, mainly a reduction of hysteresis can be noticed (basically θr is increasing), while at 300 W, θa further increases (162°) and hysteresis falls to the very low value of 2°. As a matter of fact, the droplet used for the dynamic contact angle measurement (4 µL) was visibly unstable on this surface; under this experimental condition, the surface can be defined as superhydrophobic and slippery. Thus, in the 150-300 W range, a gradual transition from a wetted to a nonwetted contact occurs, and it appears complete only at 300 W. XPS response was found to not significantly vary from 150 to 300 W (similar to that reported in Figures 3 and 4); therefore the surface chemistry does not change once the O:C ratio is fixed. SEM images at 0° (i.e., top view, upper stripe) and at a tilting angle of 80° (lower stripe) for the treated surfaces from 150 to 300 W are reported in Figure 7a. The inspection of the figure shows that (upper stripe) by increasing the input power the structures formed by plasma etching become progressively bigger, more spaced, and therefore less numerous, and further (lower stripe) the structure height increases. In an attempt to get quantitative indication of the structures developed on the surface and deeper understanding of morphology effect on the WCAH, the top view images have been processed and the distribution areas, the mean area (as arithmetic mean), the density, and the total top area percentage of the structures have been calculated.
5048 Langmuir, Vol. 24, No. 9, 2008
Di Mundo et al.
Figure 5. SEM images of the PS native sample (a) compared to those treated at 150 W for 5 min under different feed conditions: 100% CF4 (b); 17% O2 (c); 100% O2(d).
Figure 6. Advancing (θa) and receding (θr) water contact angle values of the PS samples treated with a CF4-O2 fed discharge with 17% O2 for 5 min as a function of the power input.
A threshold in the gray scale was defined for every image in such a way that all the lighter distinguishable regions were taken into account for the calculation. Since these textured surfaces are not in the form of ordered arrays of equal geometrical units (square, circular pillars, and so on), this kind of calculation was necessary to evaluate the topography; the mean area and the density gives together the idea of what the lateral size and the clearance (or spacing) are for regularly patterned surfaces. The mean area and the density of the structures are plotted in Figure 8a; the density of the structures decreases fairly linearly with input power while mean area increases, showing a steeper rise in the high power range. The calculated total top area of the structures does not vary considerably under the different conditions (it remains in the range 15-24% of the image total area) in agreement with the opposite trends found for the mean area and the density. It is worth of consideration that, beside mean area, by increasing input power also the standard deviation of the areas increases (in other words, the structures distribution becomes broader at high RF power). Figure 8b reports the maximum values of the height of the structures obtained from AFM analysis; a progressive increase of the third dimension of the structures is observed. The steepest height increase is found to occur from 275 to 300 W, as also indicated in tilted SEM images (Figure 7a), where a
transition from a slight undulation to a pinnacles profile can be appreciated. The depth increase at 300 W can account for the discontinuity found in Figure 6 where hysteresis can be seen to pass from 8° at 275 W to 2° at 300 W. Thus, besides the 2D evolution, the input power increase seems to mainly affect the hysteresis reduction through the rising height of the structures. As the height increases, the air trapping in the grooves under the water drop is likely more and more favored until (for h ) 400500 nm, 300 W sample) the nonwetted contact (Cassie-Baxter regime) is reached, which is consistent with such a low hysteresis value. The rms roughness of the 300 W sample is 35 nm. As also reported in other published data, low roughness values can be sufficient to get superhydrophobic slippery surface as long as the proper balance between lateral size, spacing, and height is found.22 Treatment Time Effect. Figure 9a reports the advancing and receding WCA values for PS samples treated at 150 W as a function of treatment time. An increase of both θa and θr contact angles up to 5 min can be observed; beyond this time, θa remains quite constant, also at longer treatment duration (20 min), while the θr shows a very slight decrease, fairly enlarging the hysteresis gap. Thus, at this medium power value, a superhydrophobic slippery surface is not achieved for any duration of the process, and indeed the sticky character appears enhanced with increased time. SEM images of the 20 min sample are reported in Figure 7b; such views can be compared with the 150 W surface in Figure 7a which represents the starting point of the WCA plateau of Figure 9a, taking into account that the F/C and O/C ratios are almost constant. It is clear that the surface morphology has deeply changed; mean area is as high as 26 000 nm2, with a density of 8.6 µ-2 and a maximum height of 600 nm. In spite of the increased height and a total top area fraction similar to the slippery surfaces (21%), the wetting behavior is that of a superhydrophobic sticky surface. This fact likely depends on the too low density of the structures for that value of height; in other words, as also qualitatively evinced in the figure, the spacing of the grooves may be too high related to their depth and therefore water is allowed to fill the gaps. The rms roughness measured by AFM
Nanotexturing of Polystyrene Surface
Langmuir, Vol. 24, No. 9, 2008 5049
Figure 7. SEM images of the PS samples treated with a CF4-O2 fed discharge with 17% O2 for 5 min as a function of the power input (a), at 150 W for 20 min (b), and at 300 W for 4 min (c). For each sample, the 0° (top view) and the 80° tilted is reported.
Figure 8. Morphological features of the PS samples treated with a CF4-O2 fed discharge at 17% O2 for 5 min as a function of the power input: mean top area and density of the structures as calculated by imaging software elaboration of SEM top view images (a) and maximum height measured from AFM (b).
is the highest obtained (75 nm). Thus, by increasing the process duration at medium power input, it is possible to obtain a highly rough surface, but this is ineffective in terms of slipperiness. Moreover, the increased eight of the structures within a Wenzel regime reasonably accounts for the higher hysteresis found (higher drop stability on the surface).
Figure 9. Advancing (θa) and receding (θr) water contact angle values of the PS samples treated with a CF4-O2 fed discharge with 17% O2 as a function of treatment duration at 150 W (a) and 300 W (b).
Figure 9b reports advancing and receding WCA values of PS samples treated with the same feed mixture at the power of 300 W as a function of the process duration. A behavior similar to the one found for the investigation of the input power effect can be observed. During the first minute of treatment, both θa and θr increase, while in the range 1-5 min, θr increases more significantly and therefore hysteresis decreases. The optimum time is 5 min, which gives the already discussed slippery surface.
5050 Langmuir, Vol. 24, No. 9, 2008
Figure 10. Sliding angle vs drop volume for PS native and for those treated with a CF4-O2 fed discharge with 17% O2 at the power of 300 W for 4 and 5 min.
Anyway, a certain slippery behavior with drops used for WCA measurements has also been observed for the 4 min processed samples, characterized by θa of 160° and hysteresis of 8°. The mean area and density of the structures for this sample are 8180 nm2 and 27.4 µm-2, respectively (SEM images in Figure 7c), while the maximum height is about 300 nm. On comparison of this sample with the one obtained with 5 min treatment, it seems that the slightly lower slippery character indicated by the higher hysteresis (8° vs 2°) arises mainly from the lower height, which likely does not ensure a uniform nonwetting behavior. The reported results on the effects of power input and treatment indicate that both parameters are directly correlated to area, clearance, and height of the structures. However, the utilization of a high power input appears a more essential requirement to get low hysteresis (Figures 6 and 9). This fact is likely due to the higher directional character of the treatment toward the surface at higher power (more intense ion bombardment), which allows to get deeper grooves that are not too much spaced out. General Remarks. In order to better characterize the wettability behavior of the slippery surfaces, a volume resolved sliding angle has been measured for the 300 W treated samples at 4 and 5 min along with the native PS. Sliding angle (R)/drop volume (V) characteristics are reported in Figure 10. As a general consideration, since drop motion is a result of the competition of gravity and surface forces, higher tilting angles are necessary to allow the sliding of smaller drops as a consequence of their lower weight. It can be observed that the 5 min sample is much more slippery than the others since much lower angles are observed. Moreover, small drops (below 4 µL) cannot be deposited on this sample (surface tension is so low that water delivered from the syringe). On the other hand, though small drops can be deposited on the 4 min samples, no sliding can be observed even at 90° inclination (drop is stuck on the surface). Hence, the R/V characteristics allow to fully differentiate the wetting behavior of the two surfaces and to better appreciate the low WCAH difference. A behavior similar to the 4 min sample is observed for the native PS (freshly washed), which shows an hysteresis value of 2°. To be more precise, the kind of drop motion is very different on the two surfaces. On the superhydrophobic 4 min sample (but also the 5 min) the droplet shape is almost spherical touching the surface with a very low area and therefore it rolls off the surface with high speed. On the contrary, on native PS, the contact area of the drop is much higher and therefore its motion, which is in this case sliding properly, is significantly slower. It is worth mentioning that if these experiments are carried out by placing the drop on a previously inclined surface lower angles
Di Mundo et al.
Figure 11. Water contact angle hysteresis (WCAH) as a function of mean WCA, calculated as (θa + θr)/2, for PS samples treated at 150 and 300 W at different treatment times. The dashed curve represents the theoretical ∆θ(θ) function obtained from the numerical expansion of eq 5 in the text (ref 12).
are found (e.g., 5° for 5 µL drops for the 5 min sample) in agreement with recent fundamental studies about the method and are correlated to the metastable nature of the wetting phenomena, highly dependent on the initial state of the observation.25 Thus, wettability data collected in this study have given the chance of comparing very different typologies of hydrophobic surfaces, i.e., from slippery hydrophobic (freshly washed polystyrene) to slippery superhydrophobic (at various degree of slipperiness) to sticky superhydrophobic surfaces. In Figure 11, the WCA hysteresis (∆θ ) θa - θr) are plotted as a function of mean WCA defined as θmean ) (θa + θr)/2 for the surfaces produced at 150 and 300 W at different treatment durations (compare Figure 9) along with the PS native sample. It can be observed that we find exactly the Que´re´ prediction12 about the separation of a ∆θ(θ) diagram into a sticky domain, above the dashed curve, and a slippery domain, below the curve. The dashed curve represents the theoretical ∆θ(θ) function obtained from the numerical expansion of eq 5. The freshly washed PS surface falls in the slippery domain for medium θ and very low ∆θ (
