Highly sticky surfaces made by electrospun polymer nano ﬁ bers

We report on a comprehensive study of the unique adhesive properties of mats of polymethylmethacrylate (PMMA) nano ﬁ bers produced by electrospinning. Fibers are deposited on glass, with varying of the diameter and the relative orientation of the polymer ﬁ laments (random vs. aligned con ﬁ guration). While no signi ﬁ cant variation is observed in the static contact angle ( (cid:1) 130 (cid:3) ) of deposited water drops upon changing the average ﬁ ber diameter up to the micrometer scale, ﬁ bers are found to exhibit unequalled water adhesion. Placed vertically, they can hold up water drops as large as 60 m L, more than twice the values typically obtained with hairy surfaces prepared by di ﬀ erent methods. For aligned ﬁ bers with anisotropic wetting behavior, the maximum volume measured in the direction perpendicular to the ﬁ bers goes up to 90 m L. This work suggests new routes to tailor the wetting behavior on extended areas by nano ﬁ ber coatings, with possible applications in adsorbing and catalytic surfaces, micro ﬂ uidic devices, and ﬁ ltration technologies.


Introduction
Wettability is one of the most important properties of solids, affecting their surface mechanics, tribology, resistance, and biocompatibility, and being governed by both the chemical composition and the morphology of the involved interface. 1,2In this respect, surfaces that have attracted a lot of attention in recent years are those exhibiting superhydrophobicity, namely an apparent static contact angle, q, formed by water drops greater than 150 inspired by many plants and insects, which provides new and versatile ideas for designing materials with self-cleaning and antifouling properties, and drag reduction. 35][6] On a lotus leaf, water drops roll off very easily even at inclination well below 10 , removing dust particles present on the surface (self-cleaning or lotus effect). 7In contrast, large water drops stick to rose petals even though they are tilted upside down (petal effect). 8Key features to achieve a specic superhydrophobic behavior involve both a proper chemical composition of the surface and an appropriate roughness at the micro/nanometer scale, 2,9,10 since in untextured surfaces q is generally below $120 , which is the value characteristic of uorinated materials. 1rious physical and chemical methods have been employed to realize either self-cleaning [11][12][13][14][15] or sticky 8,[16][17][18][19] superhydrophobic surfaces.1][22] To this aim, a high voltage is applied to the solution, which is extruded from a spinneret as an electried jet. 235][26][27][28][29][30][31][32][33][34][35] Their wettability has been mainly assessed by measuring the apparent contact angle (y150 ) and the roll off angle (#10 ), and explained in terms of the standard Cassie model with the drop contacting a composite landscape of trapped air and solid substrate. 1ere, we focus on a different property, studying highly sticky hydrophobic bers, and investigate a feature scarcely addressed [36][37][38] hitherto.The large adhesion is attributed to the water drop partially penetrating the surface texture according to a Cassie impregnating model. 1 We electrospin polymethylmethacrylate (PMMA) bers onto different substrates, varying the diameter and the relative orientation of the polymer laments (random vs. aligned conguration).Very large adhesion forces, capable to hold water drops as large as 90 mL are found.To better understand the role played by the substrate, the wetting data are compared with those obtained on freestanding mats.This study suggests new routes to produce coatings with tailored wetting properties which can easily cover extended surface areas.Potential applications of these ndings include the design and the fabrication of new adsorbing media, catalytic surfaces, delivery of uids with reduced or no volumetric loss, lab-on-chip architectures, and materials supporting water remediation.

Electrospinning
Fibers are electrospun from a chloroform solution of PMMA (120 000 g mol À1 , Sigma-Aldrich).The electrospinning apparatus consists of a 1 mL syringe tipped with a 27-gauge steel needle, mounted on a pump providing a ow rate of 0.5 mL h À1 (33 Dual Syringe Pump, Harvard Apparatus Inc.).The syringe needle is biased at 10 kV by a high voltage power supplier (EL60R0.6-22,Glassman High Voltage).Randomly oriented bers are deposited on (18 Â 18 mm 2 ) glass slides, positioned at 15 cm from the needle, whereas uniaxially aligned bers are realized by positioning the substrates on a disk (8 cm diameter, 1 cm thickness) rotating at 4000 rpm.Free-standing nonwovens made of nanobers, with overall thickness of about 60 mm, are also produced by depositing bers over a time up to three hours.
Fibers are inspected by atomic force microscopy (AFM, multimode head equipped with a Nanoscope IIIa electronic controller, Veeco), and by scanning electron microscopy (SEM, Nova NanoSEM 450, FEI) following the thermal deposition of 5 nm of Cr.AFM micrographs are acquired in tapping mode, using Sb-doped Si cantilevers with resonance frequency of 76 kHz.The average diameter (f) of the bers, calculated from at least 100 laments per each species imaged by SEM, is found to range from 0.6 to about 5.2 mm, with overall substrate coating thickness of about 10 mm.Fibers with diameters well below the micrometer-scale (down to f ¼ 0.6 mm) are spun by adding the organic salt tetrabutylammonium iodide (TBAI, Sigma-Aldrich) to the solution, at a concentration of 1% wt/wt relative to PMMA.In the following, samples labeled as R and A refer to randomly distributed and to aligned bers, respectively, whereas progressive numbers from 1 to 4 indicate increasing average diameters (1: 8 mm, and 4: f ¼ 4.6-5.2mm, respectively), as reported in Tables 1 and 2.

Wettability characterization
Measurements of the apparent contact angle (q) are performed depositing 1 mL drops of distilled water on samples positioned horizontally.Data are collected from at least ve drops placed in different positions, and the nal contact angle is expressed as the mean of the various values, with error given by the standard deviation which is representative of the surface uniformity.For aligned bers, drops are simultaneously viewed by two highresolution cameras from orthogonal directions: parallel and perpendicular to the bers orientation.The cameras mount telecentric lenses that guarantee good contrast and faithful imaging of the drop which is illuminated by two back-light collimated LED sources. 39The prole of each image is analyzed off-line by using a custom made program. 40For each drop, the apparent contact angle is deduced from a t of the prole.
Advancing (q A ) and receding (q R ) contact angles are measured through the dynamic sessile drop method by injecting and removing water from a drop having an initial volume of almost 1 mL and are dened as the threshold angles necessary to observe an expansion and a contraction of the contact line, respectively.For each surface, experiments are repeated in at least three different positions.From these data we can also extract the value of the contact angle hysteresis (Dq), i.e. the difference between the advancing and the receding contact angle.Again, for aligned bers both the advancing and the receding angles are simultaneously measured in the direction parallel and perpendicular to the ber length.
Finally, the maximum value of the volume (V max ) supported by the sample is measured by placing distilled-water drops of increasing volume on the horizontal surface and then tilting the surface vertically with a computer-controlled motor at a rate of $1 s À1 . 41Drops are deposited with a syringe pump (World Precision Instrument, Inc.) and the estimated uncertainty in V is about 3%.Drops having V larger than V max are found to fall down during the tilt, whereas drops with V # V max remain attached to the surface even though it is tilted fully upside down.For the aligned bers, this procedure is applied twice, namely with the bers either parallel or perpendicular to the gravity force.In spite of its conceptual simplicity, such a measurement is not common in published studies on sticky surfaces, which are mostly limited to determining contact and sliding angles.

Results
We have systematically varied the ber diameter to explore the wetting behavior and, more importantly, the water adhesion to our electrospun coatings and non-wovens.Exemplary bers from samples obtained with and without TBAI are shown in Fig. 1a and b (1-samples) and Fig. 1c and d (3-samples), respectively.All bers exhibit a smooth and uniform surface.In Fig. 1, micrographs of randomly oriented bers are reported in the top panels (Fig. 1a and c), whereas bers in the bottom images form arrays prevalently aligning along their longitudinal axis (le-right axis in Fig. 1b and d).

Wetting behaviour
The expectably different wetting behavior of random and aligned bers is shown in Fig. 1e-h, displaying a drop of greencolored water which assumes a circular contour with a contact angle signicantly larger than 90 on a random (R) sample, and a markedly elongated shape along the ber direction on an aligned (A) sample, with the contact angle depending on the orientation of the view.
The wettability characterization of R g -samples formed by ber mats deposited on glass is summarized in Table 1.For sake of comparison, and in order to evidence an eventual role played by the substrate underneath in determining the overall wettability properties, we have also characterized free standing mats with similar morphologies (R fs -samples).At least two (typically four) specimens are analyzed for each batch.Fig. 2 displays the resulting apparent contact angle (q) and contact angle hysteresis (Dq) behavior, which are not found to vary signicantly upon varying the ber size from 0.7 to 5.2 mm.Overall, q is comprised between 120 and 136 , increasing by more than 40 with respect to the value measured on a at PMMA surface.These results are in agreement with previous results on poly(vinyl butyral) (PVB) 28 and on uorinated polyimide bers. 38Compared with a smooth PVB surface prepared by spin coating, whose intrinsic contact angle is 60 , a mat of PVB bers of $0.6 mm in size exhibits a higher q $ 132 .This angle increases monotonically to 143 as the ber diameter is decreased to $0.1 mm.Similarly, randomly deposited uorinated polyimide bers with diameters ranged from $0.1 mm to $0.5 mm are found to increase the water contact angle to 143  from an intrinsic value of about 100 . 38However, their adhesive behavior is the opposite: PMMA ber mats are sticky, PVB and uorinated polyimide ber mats are highly water repellent, because PMMA is hydrophilic, while the two other materials are hydrophobic.The wetting behavior observed with the PMMA ber mats cannot be explained by the Wenzel model, which would predict a decreased q following texturing of the hydrophilic PMMA, in contrast to what we observe.Similarly, the drop cannot be in the Cassie regime, with air fully entrapped in the surface textures and the drop sitting on the composite solid-air interface, because this state is characterized by high contact angles but low hysteresis (sliding angles). 1 Hence, our measurements suggests the occurrence of a so-called Cassie impregnating state, intermediate between the two previous regimes, with water lling the large voids of the texture but not the small ones.This is supported by AFM topographic maps of randomly deposited PMMA bers (Fig. 3), highlighting a mat surface which consists of a complex 3D architecture (Fig. 3a) where a tangle of bers generates a multilayered structure with root mean square roughness of 0.8 mm.The network of bers forming the outward layer, in most direct contact with the liquid, covers about 50% of the total surface area.The corresponding height proles appear jagged with alternating peaks and valleys, with voids of various extension and peak-to-peak average distances roughly ranging between a few tens of nm and about 5 mm (Fig. 3b), depending on the ber size and specic arrangement/position within the mat.Such morphology exhibiting multiscale and multilayer texturing is highly suitable to promote hybrid impregnating states. 26,42ghly sticking behaviour Fiber surfaces present a similar apparent contact angle regardless of the fact that they are deposited on glass or free   standing.Also, deposited mats present a very large contact angle hysteresis Dq which makes them very sticky.This is conrmed by the maximum volume V max of the drops sustained by the surface tilted in a vertical position as shown in Fig. 2c.4][45][46] The sticky behavior of these superhydrophobic surfaces is wellexplained in terms of the Cassie impregnating state.The slightly lower V max value for free-standing samples might be related to minor topological differences in the electrospun bers, which is reliable given the observed scatter in the data corresponding to samples prepared in the same nominal conditions.More importantly, data for free-standing mats indicate that the presence of the hydrophilic glass substrate plays a minor role in the wetting and adhesive forces of deposited coatings, at least for the dense mats realized in this study.
The images in Fig. 4 provide further insight on the impregnation process.Fig. 4a shows the prole of a sessile drop deposited on a randomly oriented ber mat placed on glass.Fig. 4b and c display the same drop aer tilting the sample in the vertical position and upside down, respectively.At the beginning of plane tilting, the contact line is not completely pinned, and the front contact line moves downward due to the action of gravity and then gets pinned again, while the rear contact points are always pinned in the same position due to the very low receding contact angle.Fig. 4d is especially interesting, presenting the drop contour aer the sample is returned to the original horizontal position.It is evident an expansion of the contact area, which can increase up to almost 50%, and a corresponding decrease of the contact angle by about 20-30 , which suggests that the initial prole was in a metastable regime and that the tilting of the sample favors a better impregnation of the ber mat.We point out that, during plane tilting, the front angle increases and overcomes the advancing angle, while the rear angle decreases but it always remains higher than the receding angle.Consequently, only the front advances and the contact area becomes larger, implying a higher impregnation and adhesion with the surface.Further surface tilting does not alter this prole.Such an evolution is also observed with free-standing mats of nanobers.
We have also studied mats with aligned bers, whose wetting properties are summarized in Table 2 and in Fig. 2. As mentioned above, the shape of water drops deposited on these surfaces is elongated in the direction of the bers. 28,47While tending to expanding in every direction, a deposited drop nds randomly distributed energy barriers in its motion perpendicularly to the ber longitudinal axis, while it can move along the parallel direction.For such drops the apparent static contact angles in the direction perpendicular or parallel to the bers, q t and q k , are therefore different (Fig. 2d), with q t > q k as promoted by the relevant pinning phenomena.Dening dq ¼ q t À q || as the degree of wetting anisotropy, aligned bers on glass present dq $ 40 .The anisotropy is much less pronounced on the free-standing ber mats, possibly suggesting that the bers alignment is not so good as on the mats deposited on glass.
The anisotropy introduced by the aligned bers also greatly affects V max .When the bers are aligned horizontally (i.e., perpendicular to the direction of the acceleration of gravity), they are able to sustain larger drops than when they are aligned vertically (see Fig. 2f).In particular, on 4-A g samples with horizontally aligned bers we nd a V max value as high as 90 mL, hardly reported before.
Sticky surfaces have been used to transfer water drops from a superhydrophobic surface to a hydrophilic one without any loss or contamination. 16,43,48Fig. 5 shows a water drop having a volume of 44 mL placed on a PDMS reproduction of a lotus leaf obtained by nanocasting, 13 with a contact angle $150 and a sliding angle lower than 5 .Then, the drop is completely transferred to the ber mat by simply touching the drop with the sample 3-R g .The drop is then released from the ber mat to a glass slide.With a precision balance, we also quantify the transfer of water drops between different combinations of surfaces, with volume ranging between a few microliters and V max .From glass coated with ber mats to bare hydrophilic glass the transferred volume is more than 90% (practically the same) of the initial one.
The sticky behavior of coatings and surfaces based on electrospun polymer nanobers is to be based on relevant van der Waals forces, whose effect is enhanced by the high density of deposited non-wovens as well as by the large surface-to-volume ratio of the polymer laments.We point out that the multilayered geometry of electrospun bers is signicantly different from previously reported, gecko- inspired pillar architectures. 16,43Furthermore, the potentiality of electrospinning methods in view of realizing largearea samples is much higher than in other fabrication approaches, which opens interesting perspectives for various industrial applications.Fields beneting from the capability of delivering uids with reduced or no volumetric loss include analytical chemistry, microuidics, catalytic and ltration technologies.

Conclusions
We have studied the wetting and the unique adhesive properties of electrospun PMMA bers.The ber diameter has been intentionally varied in order to disentangle eventual ber size-effects affecting the overall surface wettability behavior.On randomly oriented bers, the apparent contact angle of deposited water drops is $130 , regardless of the bers size.More interestingly, very large adhesion forces capable to hold water drops as large as 60 mL, more than twice the characteristic values obtained with hairy surfaces, are achieved.Aligned bers present anisotropic wetting behavior, and a maximum volume of water drops retained in the direction perpendicular to the bers up to 90 mL.Measurements carried out on free-standing ber mats indicate that the presence of the glass substrate plays a marginal role in determining the above mentioned features.This work suggests electrospun polymers as very promising tool to tailor surface wetting behavior, through up-scalable production of nanobers which might exhibit modulated interactions with liquids.

Fig. 2
Fig.2(a) Contact angle, (b and e) contact angle hysteresis and (c) maximum supported volume measured on flat substrates, randomly oriented fiber mats (R) and aligned fiber samples (A).For A-samples, filled (open) symbols refer to the direction parallel (perpendicular) to the oriented fibers, as defined in the cartoons (d and f).For more details see Table1.
Fig.2(a) Contact angle, (b and e) contact angle hysteresis and (c) maximum supported volume measured on flat substrates, randomly oriented fiber mats (R) and aligned fiber samples (A).For A-samples, filled (open) symbols refer to the direction parallel (perpendicular) to the oriented fibers, as defined in the cartoons (d and f).For more details see Table1.

Fig. 3
Fig. 3 AFM topographic micrograph of a fiber mat (2-R g ).(a) 3D view.(b) 2D view.Scale bar: 20 mm.Vertical scale: 5 mm.The height profile is taken along the line highlighted in the planar view.

Fig. 4
Fig. 4 Typical drop profiles taken on randomly oriented fibers.(a) Sessile drop, (b) maximum volume held by the surface kept in the vertical position and (c) upside down, (d) the new drop contour after tilting the surface back to the horizontal position, exhibiting a lower apparent contact angle and a contact area larger than the initial one by about 40%.The scale is the same in all figures.

Fig. 5
Fig. 5 Transfer of a 44 mL water drop from a superhydrophobic water repellent sample to a glass hydrophilic surface through a glass coverslip coated with a fiber mat.(a) drop on the PDMS copy of a lotus leaf; (b) drop transfer from the lower superhydrophobic sample to the upper glass surface covered by a fiber mat; (c) drop pending from the fiber mat; (d) capillary bridge formed by the water drop between the upper fiber mat and the lower untextured glass slide; (e) drop deposited on the homogeneous glass slide.The scale is the same in all figures.

Table 1
Wettability characterization of the randomly oriented fiber mats.The g and fs subscripts indicate mats deposited on glass or free standing, respectively.Values measured on pristine bare glass and PMMA are also shown for comparison.See text for more details

Table 2
Wettability characterization of aligned fibers.Labels follow the scheme adopted for samples featuring fibers randomly oriented (see Table 1), where R is replaced by A, standing for aligned fibers.See text for more details AE 10 143 AE 11 31 AE 8 38 AE 7 103 AE 13 105 AE 13 44 AE 2 >58 AE 2