Article pubs.acs.org/Langmuir
Light-Induced Switching of Surfaces at Wetting Transitions through Photoisomerization of Polymer Monolayers Jonas Groten, Christine Bunte, and Jürgen Rühe* Department of Microsystems Engineering, Chemistry and Physics of Interfaces, University of Freiburg, 79110 Freiburg, Germany S Supporting Information *
ABSTRACT: We report on a method to generate surfaces whose wettability can be reversibly switched between a superhydrophobic and Wenzel state or a Wenzel and superwetting state just by a short UV or VIS irradiation. To achieve this, we generate a silicon surface with a nanoscale roughness (“black silicon”) and attach a polymer monolayer to it. The polymer contains a fluorinated azobenzene moiety which can be switched between the cis and trans state depending on the wavelength of the light used during illumination. The surface energy of the polymer coating is carefully adjusted to the energy value which separates distinct wetting regimes of the nanorough surface. This coupling of light induced switching to a transition of the wetting regimes can cause changes in the water contact angle as high as Δθ = 140° in the advancing CA or more than 175° in the receding CA even when the surface energy is changed only in a rather small range. Short irradiation times with UV or VIS light are enough to change the roll-off angle from 90°) are used for monolayer generation, a superhydrophobic regime is obtained where drops stay on top of the roughness features corresponding to a low CA hysteresis and low roll off angles when the substrate is tilted. The wetting behavior of the surfaces depends only on the association of the surface energy of the polymer with the corresponding wetting regime and varies only very slightly within the wetting regimes. However, at certain surface energies of the polymers the CA changes rather abruptly, creating the transitions between the different wetting regimes. In order to gain insight into the interplay between roughness and photochemical switching, carefully tailored copolymers carrying azobenzene groups are used to tune the surface polarity close to the wetting transitions of the silicon nanograss surface. The CA before and after isomerization is measured as a function of the surface energy.
■
EXPERIMENTAL SECTION
Materials. 4-(Trifluoromethoxy)aniline (99+%) (1), Phenol (99%), sodium nitrite (99%), N,N-dimethylacrylamide (DMAA) (5), chloroform (CHCl3), N,N-dimethylformamide (DMF), dimethylsulfoxid (DMSO), and triethylamine (Et3N) were supplied by SigmaAldrich. DMF, DMSO, and triethylamine were dried with CaH2 and distilled under reduced pressure. All drying procedures were performed under a nitrogen atmosphere, and dried chemicals were B
dx.doi.org/10.1021/la302764k | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
Scheme 1. Synthesis of 2-[4-(4-Trifluoromethoxy Phenylazo) phenoxy]ethanamine (OCF3−AZO) (4)
Scheme 2. Synthesis of Surface Attached Monolayers of Azobenzene Group Containing Copolymers 10a−c
were treated by oxygen plasma in order to remove residues from the passivation before further processing. Generation of Polymer Monolayers. In order to attach the polymer monolayers to the substrate surfaces first surface-attached benzophenone silane monolayers were generated from (4-(3′ chlorodimethylsilyl) propyloxy benzophenone) (8) in toluene solution.24,27 Triethylamine (Et3N) was used as a catalyst and acid scavenger. Subsequently, roughly 100 nm thick polymer layers were generated by spin-coating of 7a−c from 10 mg/mL chloroform solutions (P(MAC2-AE) 7c in DMF). The spinning speed was 2000 rpm and the samples were dried by spinning for 60 s. Irradiation for 5 min at 254 nm (Stratalinker from Oxygene) led to the covalent attachment of a polymer monolayer to the silicon surfaces. Any residual, nonbound polymer was finally removed by rinsing the samples with the proper solvent and extraction in an ultrasonic bath for 20 min (Scheme 2). The fluorinated azobenzene group compound OCF3−AZO 4 was reacted with surface-attached active ester polymer monolayers 9a−c through an aminolysis reaction.28 The samples were placed in solutions with 20 mg/mL OCF3−AZO in DMF and 50 μL/ml triethylamine for 15 min. Subsequently the samples were rinsed with a large excess of DMF to remove any nonreacted OCF3−AZO from the surface (Scheme 2). The aminolysis reaction can also be done in solution. To this, 107 mg of 9a or 57 mg of 9b and 32 mg (0,1 mmol) were dissolved in DMF containing 50 μL of triethylamine. The solution was stirred for 15 min and subsequently precipitated in diethylether and then purified by repeated precipitation from chloroform. Sample Characterization. Contact Angles (CAs) were measured using an OCA20 measurement system from Dataphysics GmbH, Germany. Static contact angles were determined using drops of deionized water having a volume of 2 μL. Advancing/receding CAs were measured at a dosing/retraction speed of 0.1 μL/s. In order to measure the roll off angles the contact angle system OCA20 from Dataphysics is equipped with a rotatable sample stage. Samples are rotated from a horizontal position until the drop rolls off the surface. After roll off, the angle between a horizontal line and the sample position was measured using the OCA20 software. The UV irradiation process was carried out with an Oriel 69910 arc lamp from Newport, which was operated with an I-line filter cutting off light with wavelengths other than 365 nm. The intensity at 365 nm was 47 mW cm−2. Visible light illumination was done with a bright
Table 1. Composition of DMAA:MAC2-AE Copolymers name
DMAA % content
MAC2-AE % content
Mw g/mol
Mn g/mol
7a 7b 7c
95 90 0
5 10 100
100 000 110 000 145 000
24 000 28 000 54 700
LED Multihead lamp from HS Vision, Germany. The intensity at 430 nm was 32 mW cm−2. UV−vis spectra were recorded with a 50 Cary Bio from Varian. Layer thicknesses were measured with an imaging ellipsometer (EP3 from Nanofilm). 1H NMR and 13C NMR spectra were recorded on a Bruker 250 MHz instrument. The environmental scanning electron microscope (ESEM) images of the black silicon surface were made with an ESEM 2020 from Electroscan at a water vapor atmosphere of 12 mbar and an acceleration voltage of 25 kV. C
dx.doi.org/10.1021/la302764k | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
the trans and cis state by irradiation with different wavelengths. When a sample coated with a layer of OCF3−AZO is irradiated with UV light at a wavelength of λ = 365 nm, which roughly corresponds to the energy gap of the π − π* (S2 state) transition (λmax = 349 nm), the azobenzene derivate converts from the trans to the cis isomer. Consequently, the absorption band at 349 nm decreases and shifts slightly to 315 nm, in agreement with theoretical calculations.15 The band at 435 nm, which corresponds to the n−π transition (S1 state), increases (Figure 2a). However, as the absorption bands of the two
Figure 1. Environmental scanning electron microscope (ESEM) image of a black silicon surface coated with a monolayer of polymer 10c. The inset shows a high magnification image.
■
RESULTS AND DISCUSSION Synthesis. The black silicon samples were prepared by an anisotropic etching process using C4F8 as a passivation gas and SF6 plus O2 (10:1 ratio) for etching. Due to overpassivation a needle like surface structure with peaks in the nanometer range is obtained, commonly referred to as black silicon or silicon nanograss.23 In our case, the height of the spikes was on average 10 μm and the angle of opening less than 10°. The spikes were randomly arranged, and the spacing did not vary significantly from place to place. On average, a density of 25 spikes per 100 μm2 was determined (Figure 1). In order to control the surface energy independently from the topography, samples of silicon nanograss and, for reference purposes, flat silicon samples were coated with P(DMAAMAC2-AE) copolymers 7a−c containing different ratios of active ester (MAC2-AE) 6 to N,N-dimethylacrylamide (DMAA) 5 as listed in Table 1. To this end, the polymers were photochemically attached to self-assembled monolayers of the benzophenone silane 8 (Scheme 2) generated from a dilute (∼10−3 M) solution of 8 in toluene. A few drops of triethyl amine are added to bind the resulting HCl and act as a catalyst.24 First rather thick layers (50 and 200 nm) of the respective polymers are deposited by spin coating. Illumination with 254 nm for 5 min leads to a covalent attachment of polymer molecules contacting the surface by reaction of the benzophenone with a C−H bond in the polymer backbone. During extraction of the samples in a proper solvent all polymer except the covalently attached monolayer can be removed. To allow light induced surface energy variations a fluorinated azobenzene is introduced through aminolysis of the active ester groups which are present in the monolayer. The layer is swollen in DMF during the reaction. The synthesis of 2-[4-(4trifluoromethoxyphenylazo)phenoxy] ethanamine 4 (OCF3− AZO) was done in three steps (Scheme1). At first 4-(4trifluoromethoxy phenylazo) phenol 2 was synthesized by the diazocoupling reaction between 4-(trifluoromethoxy) aniline 1 and phenol. In the second step, the BOC protected amino group was introduced 3 through a substitution reaction with 2(boc-amino)ethyl-bromide and in the third step the protection group was removed using trifluoro acetic acid. Photoisomerization of the Azobenzene Derivates. The azobenzene units in the polymers can be switched between
Figure 2. (a) UV−vis spectrum of the trans (−−−) and cis isomer (- -) of the OCF3−AZO. (b) Switching kinetics of the band at 349 nm, when illuminated with visible light (gray ●) and UV light (■). (c) Switching kinetics from the cis to the trans state for the band at 349 nm under different light conditions; bright visible light (−■−), day light (gray −●−) and dark storage (−◆−).
isomers overlap, no photoinduced generation of a pure isomer is possible and even after prolonged isomerization, a significant number of molecules remains unchanged. The azobenzene can be switched back to the trans isomer when irradiated with visible light. As a result the band at 349 nm recovers and the 435 nm band vanishes. D
dx.doi.org/10.1021/la302764k | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
Figure 3. Contact angle of water on flat (a−c) and nanograss covered silicon surfaces (d−f). All surfaces are modified with surface attached monolayers composed of: (a,d) polymer 10a with 5%, (b,e), polymer 10b with 10% and (c,f) polymer 10c with 100% OCF3-AZO groups. Advancing CAs are represented by (▲), receding CAs by (▼). Samples are cyclically illuminated with UV light at 365 nm or with visible light for 1 min. (g,h,i) Schematic depiction of the contact angle situation for the different substrates and coatings in (a) to (f); simultaneously, the switching regimes induced by photoisomerization on a flat surface are shown as gray bars together with the wetting transitions of silicon nanograss (black lines) according to Dorrer23 (comp. also Figure 4).
The advancing CA changes from 72 to 60° for polymer 10a (Figure 3a), from 82 to 68° for polymer 10b (Figure 3b), and from 102 to 89° for polymer 10c (Figure 3c), As expected with increasing OCF3−Azo contents the polymers become more hydrophobic and the CAs are shifted to higher values. It should be noted, however, that the absolute value of the CA changes (the difference of the advancing CA between the two isomerization states), is in all cases around 12−14° regardless of the details of the polymer composition. Hence, the amount of OCF3−AZO incorporated in the polymers only varies the absolute value of the surface energy but not the extent of surface energy variation during illumination. In contrast to this all polymers have a receding CA at around 20° and this value is almost not influenced by the illumination process. The large contact angle hysteresis (i.e., difference between the advancing and receding contact angle) on these very smooth surfaces is probably due to rearrangements in the polymer layers. They can be caused by interaction of water with the amphiphilic polymer as this contains strongly nonpolar (OF3-substituted aromatic rings) and polar groups (amide moieties). Photoisomerization on Silicon Nanograss Surfaces. The wetting properties of black silicon surfaces coated with monolayers of polymers showing a different polarity has been described by Dorrer et al.23 In their study, the advancing and receding CAs of water drops on the coated nanograss surfaces were measured and plotted in Figure 4 as a function of the advancing contact angle of the surface coating on a planar
To get a better understanding of the photoisomerization process the switching kinetics of the azopolymer a layer of 10a spin coated on a silica glass slide was examined. The isomerization process is complete after 20 s for both directions (trans−cis and cis−trans) (Figure 2b). The stability of the cis state at room temperature, in the dark and at daylight was additionally tested. Full reisomerization at room temperature in the dark takes place within one week (due to thermal switching) or within 30 min at daylight. (Figure 2c). Photosiomerization on Flat Surfaces. To study the switching properties of azobenzene polymer monolayers of varying monomer composition 10a−10c were prepared on almost ideally flat surfaces (polished silicon wafers, rms r = 0.3 nm) (Table 1). The surface energy and accordingly the water contact angle (CA) is varied to some extent by switching the azobenzene from the trans to the cis state. In this study, only wetting by water is studied as this is for any practical application the most relevant liquid. It should be noted, however, that the water contact angle and the surface free energy are directly correlated for ideally flat surfaces. For azobenzenes it is known that the change in the wetting behavior upon irradiation is a consequence of the difference in the dipole moment between the trans and the cis isomer of the azobenzene derivate.15 Advancing and receding CAs of water on the planar silicon substrates are measured for every polymer in 5 cycles of subsequent illumination with 365 nm UV light and light in the visible range for 1 min. For the polymers 10a−10c, a variation of the CA can be clearly observed (Figure 3a−c). E
dx.doi.org/10.1021/la302764k | Langmuir XXXX, XXX, XXX−XXX
Langmuir
Article
properties of the polymers behave very different during photoisomerization: 1 Switching at the transition Wenzel-type wetting− superwetting (Figure 3d). If a black silicon surface modified with 10a is irradiated with UV light for 1 min, then the advancing CA changes from 134 to 0° and the receding CA stays at 0°. The large contrast in the advancing CA clearly indicates a transition from the Wenzel-type wetting regime to the superwetting regime. The receding CA stays at 0° after UV light irradiation as expected for both wetting regimes. Illumination of the dried surface with visible light for 1 min and then placing a new drop on the surface, leads to a full recovery of the advancing CA to 134°, so the surface is switched back from the superwetting regime to the Wenzel regime. The switching can be executed repeatedly in the same way and the wetting properties are reversibly switched between the two wetting regimes (Figure 3d). 2 Switching within the Wenzel regime (Figure 3e). For the coating with polymer 10b the advancing CA varies slightly around 126° no matter whether the sample was irradiated with UV light or visible light. In addition the receding CA stays at a value of 0° characteristic for the Wenzel-type wetting regime. Although photoisomerization of the polymer on a flat surface causes a small, but significant contact angle change of ΔCAflat = 15° (Figure 3b) there is no significant change of the CA caused by photoswitching of the same polymer monolayers on the black silicon. 3 Switching between the supherhydrophobic and the Wenzel-type wetting regime (Figure 3f). For a surface coated with the polymer 10c, the advancing CA changes from 179 to 154° and the receding CA changes from 166 to 0° after illumination with UV light. This is expected since the difference in the CA of the flat surface matches exactly to the transition between the superhydrophobic and the Wenzel type wetting regime on coated nanograss. As expected from the wetting behavior of the black silicon this leads to an extreme decrease of the receding CA. Again, the surface can be dried and switched back to the superhydrophobic state by irradiation with visible light for 1 min (Figure 3f). For very hydrophobic surfaces, the roll-off angle is also a characteristic parameter. For many applications the most important value. Due to the large change in the receding CA the roll off angle for samples coated with 10c changes also strongly. In Figure 5, the roll-off angle is shown after subsequent illumination steps with UV and visible light. After UV-irradiation the receding CA of the illuminated sample drops to 0° and hence the drops are strongly sticking to the surface even when the substrate is tilted to 90°. Upon illumination with VIS light, the roll off angle on this surface drops again to values
