Solvent-Induced Dewetting on Curved Substrates ... - ACS Publications

Aug 26, 2015 - Mu-Huan Chi , Zhi-Xuan Fang , Hao-Wen Ko , Chun-Wei Chang , and Jiun-Tai Chen. The Journal of Physical Chemistry C 2016 120 (50), ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Macromolecules

Solvent-Induced Dewetting on Curved Substrates: Fabrication of Porous Polymer Nanotubes by Anodic Aluminum Oxide Templates Mu-Huan Chi,† Chun-Wei Chang,† Hao-Wen Ko,† Chun-Hsien Su,‡ Chih-Wei Lee,† Chi-How Peng,‡ and Jiun-Tai Chen*,† †

Department of Applied Chemistry, National Chiao Tung University, Hsinchu, Taiwan 30010 Department of Chemistry and Frontier Research Center on Fundamental and Applied Sciences of Matters, National Tsing Hua University, Hsinchu, Taiwan 30013



S Supporting Information *

ABSTRACT: Dewetting of polymer thin films triggered by thermal or solvent annealing on flat substrates have been widely studied. Research on dewetting of polymer thin films on curved substrate, however, has been rarely investigated and remains a great challenge, especially for dewetting induced by solvent annealing. In this work, we study the dewetting of polymer thin films on curved substrates by annealing poly(methyl methacrylate) (PMMA) nanotubes in cylindrical nanopores of anodic aluminum oxide (AAO) templates via dimethylformamide (DMF) vapor treatment. After solvent annealing, dewetting of the curved PMMA thin films is induced, resulting in the formation of porous PMMA nanotubes. By changing the experimental conditions such as annealing time, solvent quality, and molecular weight of polymer, the hole sizes of the PMMA nanotubes can be controlled or even prevented. To demonstrate the universality of this method, porous nanotubes of polystyrene (PS) and fluorescent pyrene-ended PMMA (Py-PMMA) are also fabricated. This work not only provides a better understanding of dewetting of polymer thin films on curved substrates but also offers a promising new way to prepare porous polymer nanotubes with controllable hole sizes.



where A is the effective Hamaker constant, h is the film thickness, SP is the polar component of the spreading coefficient (short-range polar interactions), and l is the decay length describing the range of these interactions.7−9 There are four classes for the form of the excess intermolecular interaction free energy. In type I systems, both long-range and short-range interactions are attractive (A is positive and SP is negative). In type II systems, long-range interactions are attractive and short-range interactions are repulsive (A and SP are positive). In type III systems, long-range interactions are repulsive and short-range interactions are attractive (A and SP are negative). For these three systems, unstable polymer thin films can be observed. In type IV systems, both long-range and short-range interactions are repulsive (A is negative and SP is positive), and stable polymer thin films can be obtained.7 In the past, the dewetting of polymer thin films driven by thermal or solvent treatments on flat substrates has already

INTRODUCTION Polymer thin films have been intensively studied because of their applications in coatings, surface modifications, paintings, adhesives, and membranes.1−5 To maintain the functions of polymer thin films, it is critical to control their stabilities. One of the most important factors on the stabilities of polymer thin films is the wetting behavior. When an amorphous polymer thin film is annealed at temperatures higher than the glass-transition temperature (Tg) of the polymer, wetting or dewetting of the polymer chains can occur to reach more energy stable states. The wetting behavior of polymer thin films can be determined by considering the attractive forces between different components, which can be divided into long-range (dispersion interactions) and short-range (polar interactions) interactions.6 If the attractive forces between air and substrates are larger than those between substrates and polymer films, dewetting can occur. The excess intermolecular interaction free energy (ΔG), which can be used to predict the wetting or dewetting of polymer thin films, is defined as the following: 2

p

ΔG = −A /12πh + S exp( −h/l) © 2015 American Chemical Society

Received: June 4, 2015 Revised: August 12, 2015 Published: August 26, 2015

(1) 6241

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules

Figure 1. Schematic illustration of the formation of porous polymer nanotubes in the nanopores of AAO templates.

been widely studied.10−16 The dewetting of polymer thin films on curved substrates, however, is rarely investigated.17,18 For the development of portable and curved displays or other consumer electronics with ergonomic design, the stability of polymer thin films on curved surfaces is a critical issue and should be thoroughly addressed. In this work, we study the dewetting of polymer thin films coated on curved substrates using the cylindrical nanopores of anodic aluminum oxide (AAO) templates. The AAO template is one of the most commonly used templates for preparing versatile nanostructures such as nanospheres, nanopillars, nanotubes, and nanowires.19−29 AAO templates with cylindrical nanopores are prepared by electrochemical anodization processes of aluminum foils.30 The pore diameters, pore lengths, and pore-to-pore distances can be controlled using different anodization conditions, such as the applied voltage, the working time, the type of the electrolyte, and the electrolyte concentration.31 Using the AAO template, polymer solutions, melts, or solvent-vapor-swollen polymer films can be infiltrated into the nanopores by capillary force.32−35 After solidification of polymers and selective removal of the AAO templates, various polymer nanostructures can be obtained. Among different nanostructures, polymer nanotubes prepared by the AAO templates can be regarded as curved polymer thin films coated on curved substrates, which serve as a great model system for studying the dewetting behavior. For the polymer, we choose poly(methyl methacrylate) (PMMA) as a model material because it is one of the most commonly studied polymers in dewetting phenomena of polymer thin films on flat substrates. In addition, most physical and thermal properties of PMMA have already been well studied, enabling it a suitable material for comparing experimental results. Moreover, PMMA possesses excellent optical properties and good chemical etching resistance. Therefore, more applications can be expected from the dewetted PMMA nanotubes. Here, PMMA nanotubes are first prepared in the nanopores of AAO template by a melt wetting method.32 Because of the high surface energy of the aluminum oxide wall, dewetting could not be induced by

thermal annealing. Therefore, we anneal the polymer nanotubes by solvent annealing. After the PMMA nanotubes are annealed in N,N-dimethylformamide (DMF) vapor, dewetting occurs and porous PMMA nanotubes can be obtained. By changing the annealing conditions, such as solvent qualities, and molecular weights of polymers, the morphologies of porous PMMA nanotubes can be controlled. This work not only gives a deeper understanding of dewetting of polymers in curved and confined environments but also provides a novel method to fabricate porous polymer nanotubes. The porous polymer nanotubes prepared here may have potential applications in areas such as sensors, catalysts, drug delivery, and separations.36,37 In addition, these porous nanotubes may also be applied to advanced technological applications such as supercapacitors, gas storage, and electrodes for solar cells.38−40



EXPERIMENTAL SECTION

Materials. Poly(methyl methacrylate) (PMMA) with weightaverage molecular weights (Mw) of 30 kg mol−1 (PDI = 1.25), 68.5 kg mol−1 (PDI = 1.11), and 183 kg mol−1 (PDI = 1.7) were purchased from Polymer Source, Inc. Polystyrene (PS) with a weight-average molecular weight (Mw) of 78.5 kg mol−1 (PDI = 1.05) was purchased from Polymer Source, Inc. Pyrene-ended poly(methyl methacrylate) (Py-PMMA) was synthesized by atom transfer radical polymerization (ATRP) with a weight-average molecular weight (M w ) of 38.5 kg mol−1. Anhydrous N,N-dimethylformamide (DMF) was obtained from Tedia Company, Inc. AAO membranes (Anodisc 13) with average pore diameters of ∼225 nm and thicknesses of ∼60 μm were purchased from Whatman. Fabrication of Polymer Nanotubes by the Melt Wetting Method. Polymer nanotubes were fabricated by the melt wetting method.32 A PMMA or PS solution (10 wt % in ethyl acetate) was first dropped onto a glass substrate, followed by a drying process at ∼50 °C in a fume hood. The dried polymer film was then annealed at 200 °C for 4 h to flatten the film surface and release the residual stress in the polymer film. An AAO template was placed on top of the polymer film, and the sample was annealed at 230 °C for 1 h. After a cooling process, polymer nanotubes in the nanopores of the AAO template can be prepared. The AAO template can be removed by 5 wt % NaOH(aq) and filtrated to obtain the released polymer nanotubes. 6242

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules Fabrication of Porous Polymer Nanotubes by SolventAnnealing-Induced Dewetting of Polymer Nanotubes. The solvent vapor annealing processes were conducted in a sealed glass chamber. The residual moisture or solvents in the glass chamber were first removed by a dryer and purged by nitrogen. The nanotubecontaining templates were placed in the sealed chamber, which has an open bottle of DMF. The samples were annealed for 0−30 h. After the samples were dried and the templates were removed selectively by 5 wt % NaOH(aq), porous polymer nanotubes were fabricated. Structure Analysis and Characterization. A scanning electron microscope (SEM, JEOL, JSM-7401F) with an accelerating voltage of 5 kV was used to investigate the morphologies of the polymer nanostructures. The polymer nanostructures were dried in vacuum desiccators and coated with 4 nm of platinum before the SEM measurement. A transmission electron microscope (TEM, JEOL, JEM2100) with an accelerating voltage of 200 kV was also used to investigate the morphologies of the polymer nanostructures. For our regular examination of the polymer samples, the times of focusing and acquisition in the SEM and TEM measurements are less than 30 s. The effect of electron beams on the polymer samples are discussed in the Results and Discussion section. The photoluminescence (PL) spectra of the samples were measured using a Spex Fluorolog-3 spectrofluorometer (Jobin-Yvon, Inc.) equipped with a 450 W Xe lamp. Quantitative analysis of the SEM data was conducted using ImageJ software. For the minimum number of pores considered for the statistical analyses to evaluate the mean pore diameters, we have used at least 47 pores for most of the calculations, and one of the data points is obtained by measuring up to 141 pores. For one specific experimental condition (water content of DMF ∼ 9.95 wt %), however, the pore densities on the nanotubes are too low, and only 11 pores are measured. For the minimum number of pores considered for the statistical analyses to evaluate the mean pore areas, we have used at least 33 pores for most of the calculations, and one of the data points is obtained by measuring up to 85 pores. For one specific experimental condition (water content of DMF ∼ 9.95 wt %), however, the pore densities on the nanotubes are too low and only 17 pores are measured.

times. After polymer chains are swollen by the DMF molecules, the PMMA thin films dewet on the surface of nanopores because of the stronger interactions between DMF and alumina surfaces. Similar to the dewetting of polymer thin films on flat substrates, holes are formed at the interfaces between the AAO nanopores and swollen polymers. After the solvent and templates are removed, porous PMMA nanotubes can be fabricated. To investigate the morphologies of the nanostructures in every step, both SEM and TEM measurements are conducted. After the AAO templates are removed by NaOH(aq) selectively, the PMMA nanotubes prepared by the melt wetting method can be obtained, as shown in Figure 2a. The average



RESULTS AND DISCUSSION To study the dewetting of polymer thin films on curved substrates, we use the polymer nanotubes confined in the nanopores of AAO templates as a model system. The schematic illustration and experimental process of the solvent-vaporinduced dewetting process of polymer nanotubes inside AAO nanopores are shown in Figure 1 and Figure S1, respectively. First, PMMA nanotubes in AAO templates are prepared by the melt wetting method.32 Briefly, a PMMA film coated on a glass substrate is covered by an AAO template. The AAO template contains an average pore diameter and length of ∼225 nm and ∼60 μm, respectively, as shown in Figure S2. After the sample is annealed thermally at temperatures higher than the glass transition temperature (Tg) of PMMA (∼110 °C), the polymer chains are softened and obtain enough mobilities to infiltrate the AAO nanopores via capillary force. The infiltration process is driven by the reduction of the total surface and interfacial energies. The AAO walls possess high surface energy (∼1340 mJ m−2) and can be wetted by the PMMA melts easily.41,42 In this work, the PMMA nanotubes are prepared by annealing the samples at 230 °C.22 The annealing temperatures are much lower than the unzipping temperature of PMMA (>300 °C). Therefore, the unzipping process of PMMA is ignored during the annealing experiments.43−46 The polymer nanotubes in the nanopores of AAO templates can be regarded as polymer thin films coated on curved alumina surfaces. Subsequently, the PMMA nanotube-containing AAO templates are annealed with DMF solvent vapor in a sealed chamber for different annealing

Figure 2. PMMA (Mw = 68.5 kg mol−1) nanotubes fabricated by the melt wetting method at 230 °C for 1 h: (a, b) SEM images with different magnifications and (c, d) TEM images with different magnifications.

diameter and length of the nanotubes is ∼264 nm and ∼45 μm, respectively. This result shows that the PMMA melts wet the nanopores of the AAO templates completely. Some PMMA nanotubes with shorter lengths are also observed because of the defects in the nanopores or the destruction caused by the filtration process. The top view SEM image shows the open ends of the PMMA nanotubes, as shown in Figure 2b. The morphologies of the PMMA nanotubes can also be investigated further by TEM measurements. From Figure 2c, the hollow nature of the PMMA nanotubes can be confirmed. From the magnified TEM image (Figure 2d), the wall thickness of the PMMA nanotubes can be measured to be ∼20 nm. Subsequently, the nanotube-containing AAO templates are annealed by DMF vapor for different times. To control the vapor pressure of DMF, all experiments are conducted at 30 °C. The PMMA nanotubes with holes on the nanotube surfaces are formed after the solvent annealing process, as shown in Figure 3a. The nonuniform size of the holes is probably due to the merging process of nearby holes or the randomly occurred nucleation process of the hole formation. For the holecontaining surfaces, there are two possibilities, which are related 6243

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules

found for the structure of porous rods. Second, the inner part of the porous tubes can be seen through from the front holes, while the porous rods can not be seen through. Third, the contrast between the hole and the wall of tubes under SEM measurement is higher in porous tubes than that in porous rods, where less contrast variation is observed between the dimple and the rod surface. Both SEM and TEM images can be used to determine the morphology of the PMMA nanostructures. In the SEM image (Figure 3c), the dashed red circles indicate the locations of two of the holes on the nanotubes. From the hole indicated by the upper right circle, the sharp edge can be observed. From the hole indicated by the lower left circle, the hole can be seen through. In the TEM image (Figure 3d), the tubular structures can also be confirmed. Therefore, both the SEM images and the TEM image confirm that the porous PMMA nanostructures are porous tubular structures, as shown in the left structure in Figure 3b. After swollen by the DMF vapor, the PMMA chains obtain mobilities. At the same time, the composition of the atmosphere is also changed. Because of the stronger interactions between the DMF and aluminum oxide (hydrogen bonds), swollen PMMA chains dewet spontaneously on the surfaces of the AAO nanopores. Similar to the dewetting on flat substrates, the polymer thin films dewet on curved substrates and form holes randomly at the interfaces between polymer films and the alumina walls. Therefore, porous PMMA nanotubes with randomly distributed holes can be obtained. To study the formation mechanism of the porous PMMA nanotubes further, the annealing processes are also conducted with different annealing times and solvent qualities. For different annealing times, the samples are annealed in pure DMF vapor at 30 °C for 0−30 h. As shown in Figure 4a−f, the holes on the nanotubes start to form after a critical annealing time (∼10 h), and the sizes of the holes increase with the annealing time.

Figure 3. Porous PMMA (Mw = 68.5 kg mol−1) nanotubes fabricated by annealing the nanotube-containing AAO templates in DMF vapor at 30 °C for 30 h: (a) SEM image with a lower magnification, (b) graphical illustrations of a porous tube (left) and a porous rod (right), (c) SEM image with a higher magnification where the dashed red circles indicate the holes, and (d) TEM image.

to whether the inner nature of the nanostructures is hollow or not, as shown in the illustration (Figure 3b). The holes can be either on porous tubes (hollow) or on porous rods (solid). There are three major differences between these two structures. First, the cross section of the nanotube wall can be observed directly on the edge of the holes for the structure of porous tubes while only sharp edges with gradient thickness can be

Figure 4. Porous PMMA (Mw = 68.5 kg mol−1) nanotubes fabricated by annealing the nanotube-containing AAO templates in DMF vapor at 30 °C for different annealing times: (a) 0, (b) 9, (c) 12, (d) 18, (e) 24, and (f) 30 h. 6244

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules

Figure 5. Porous PMMA (Mw = 68.5 kg mol−1) nanotubes fabricated by annealing the nanotube-containing AAO templates at 30 °C for 24 h in DMF/water vapors. DMF/water mixtures with different water contents are used: (a) 1.30, (b) 3.26, (c) 5.09, (d) 6.38, (e) 9.95, and (f) 11.87 wt %.

Figure 6. (a) SEM image and illustration of the porous PMMA (Mw = 68.5 kg mol−1) nanotubes. The hole diameter and hole area are indicated in the illustration. (b) Plot of the average hole diameter versus the annealing time. (c) Plot of the average hole area versus the annealing time. In (b, c), the nanotube-containing AAO templates are annealed in DMF vapor at 30 °C. (d) Plot of the average hole diameter versus the water content of the annealing solvent (DMF/water mixture). (e) Plot of the average hole area versus the water content of the annealing solvent (DMF/water mixture). In (d, e), the nanotube-containing AAO templates are annealed in DMF/water vapors at 30 °C for 24 h.

Quantitative analysis is also performed to study the formation mechanism of the holes further. From the SEM images, the average hole diameter and hole area of the porous nanotubes prepared by different experimental conditions can be measured. For the convenience of the measurement, the hole diameter is defined as the length of the hole parallel to the tube axis, as shown in Figure 6a. The diameters of the holes along the peripheral axis are difficult to be measured from the SEM

The sizes of the holes are also affected by the solvent quality. Figure 5a−f shows the SEM images of the PMMA nanostructures by annealing the nanotube-containing AAO templates in DMF vapors containing different amounts of water (0−11.87 wt %) for 24 h. As the amount of water in DMF increases, the size of the holes on the nanotubes increases initially and decreases later after a critical amount of water (∼5 wt %). 6245

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules images, and we assume that the holes on the PMMA nanotubes have an ideal round shape. This assumption may not be correct because the dewetting rates of the PMMA films may not be the same at different directions caused by the curvatures of the wall surfaces of AAO nanopores. The hole area is defined as the area of the projected region in the SEM image, as shown in Figure 6a. The real hole area is larger than the hole area defined here because the dewetted region has a curved disk-like shape. Considering that it is difficult to measure the curved dewetted region, we use the projected area in the SEM image as the hole area. Even though the real values of sizes of the holes may be underestimated, we believe that the trend of the dependence of the sizes of the holes on the experimental conditions can still be revealed. Using the newly defined hole diameter and hole area, we conduct the quantitative analysis for the samples prepared at different conditions. Figure 6b−e shows the data from the samples annealed at different annealing times and amounts of water in DMF. After the samples are annealed longer than ∼10 h, both the diameter and area increase with the annealing time, as shown in Figures 6b and 6c, respectively. In the solvent annealing process, the polymer chains in the curved polymer films are swollen by the solvent molecules. For DMF, the boiling point (∼153 °C) is relatively high and the vapor pressure (∼5.30 Torr at 30 °C) is much lower than those of most organic solvents at room temperature.47 Hence, longer times are required for DMF vapor to evaporate and reach equilibrium in the annealing chamber. In addition, the swelling process of PMMA chains by DMF is slow due to the lower solubility.48 Therefore, holes start to form after a long annealing time. In the beginning, the nucleation sites of the dewetting process are formed randomly. With longer annealing time, the nucleation regions (holes) grow to decrease the total contact area between alumina wall and swollen PMMA films. The results are similar to those of dewetting of polymer thin films on flat substrates.13 In addition to the annealing time, the amount of water in DMF (solvent quality) also affects the sizes of the holes of the porous PMMA nanotubes. In ambient condition, DMF is miscible with water and can absorb moisture continuously. Therefore, the amount of water in DMF may have a great effect on the formation of the polymer nanostructures. Here, we mix different amount of water into pure DMF intentionally to discuss the effect of solvent quality on the formation of holes on the porous PMMA nanotubes. As shown in Figure 6d,e, both average hole diameter and area initially increase with the amount of water. After the amount of water is higher than ∼5 wt %, the average hole diameter and area decrease. These results can be explained by considering the vapor pressure of DMF, solubility of PMMA in DMF, and the surface tension of atmosphere. From the Raoult’s law ° χwater + PDMF ° χDMF Ptotal = Pwater

At lower water contents in DMF, the reduction of vapor pressure of DMF (ΔPDMF) and solubility of PMMA in DMF by water is small. The composition of atmosphere, however, changes largely because Pwater ° is much larger than PDMF ° (χwater in atmosphere is equal to P°water χ water/(P°water χ water + P°DMF χ DMF)), resulting in the change of the surface tension of the atmosphere. The change of the surface tension of the atmosphere seems to accelerate the dewetting process, probably because of the stronger interaction between the water and the AAO walls. As a result, the degree of dewetting increases with the amount of water in DMF at lower water contents. At higher water contents in DMF, the reduction of vapor pressures of DMF (ΔPDMF) and solubilities of PMMA in DMF by water starts to dominate the dewetting. The vapor pressure of DMF and solubility of PMMA are significantly reduced when the amount of water in DMF is higher than ∼5 wt %. Therefore, the PMMA chains are difficult to be swollen, and the chain mobilities decrease significantly. Although the driving force of dewetting becomes larger because of the change in the surface tension of the atmosphere, the effect of the reduced chain mobilities dominates. As a result, the degree of dewetting decreases with the amount of water in DMF at higher water contents. We propose that there are two possible mechanisms of the hole formation on porous PMMA nanotubes: the dewetting mechanism and the growth mechanism of water droplets. In the dewetting mechanism, as shown in the upper part in Figure 7,

Figure 7. Two possible mechanisms of the hole formation on porous PMMA nanotubes. The upper part illustrates the dewetting mechanism, where the sizes of the dewetted holes increase with the annealing time. The lower part illustrates the growth mechanism of water droplets, where the sizes of the water droplets increase with the annealing time. In the growth mechanism of water droplets, holes are formed on the PMMA nanotubes after the solvent and water are removed.

(2)

where Pwater ° is the vapor pressure of water at 30 °C (∼31.84 Torr),49 χwater is the molar ratio of water in the water/DMF mixture, P°DMF is the vapor pressure of DMF at 30 °C (∼5.30 Torr),47 and χDMF is the molar ratio of DMF in the water/DMF mixture (χDMF = 1 − χwater). The reduction of vapor pressure of DMF in the water/DMF mixture can be derived as

the sizes of the dewetted holes increase with the annealing time. In the growth mechanism of water droplets, as shown in the lower part in Figure 7, the sizes of the water droplets increase with the annealing time and holes are formed on the PMMA nanotubes after the solvent and water are removed. For the dewetting mechanism, we can estimate the excess intermolecular interaction free energy (ΔG). After the PMMA chains are swollen by solvent molecules, the system consists of three main components, including alumina walls, DMF vapor/

° − PDMF ° χDMF = PDMF ° (1 − χDMF ) = PDMF ° χwater ΔPDMF = PDMF (3) 6246

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules

rich and the solvent-rich phases, the domains have an asymptotic power law growth behavior. For the coalescence, which is caused by the impinging of two droplets by translational diffusion, a power law relation between the domain size and the time is usually followed. For the hydrodynamic flow, which is caused by the gradient of pressure, the growth rate of coarsening is linear in time.56 After the solvent are removed using a vacuum pump, the domains of water droplets form holes on the surfaces of PMMA nanotubes. To confirm the formation mechanism of the porous nanotubes, other solvents, including toluene, tetrahydrofuran (THF), acetone, and 1,2-dichlorobenzene (ODCB), are also used to anneal the PMMA nanotubes. The solvents can be categorized into nonpolar solvent (μtoluene ∼ 0.36 D), medium polar solvent (μTHF ∼ 1.75 D), and high polar solvent (μacetone ∼ 2.88 D and μODCB ∼ 2.50 D).57−59 Among these solvents, acetone and THF can form hydrogen bonds and are miscible with water while toluene and ODCB are insoluble in water. For samples annealed in toluene vapor, the interactions between the solvent molecules and the surfaces of the AAO nanopores are relatively weak. Therefore, no hole is formed on the surfaces of the PMMA nanotubes by annealing the sample in toluene vapor for 24 h, as shown in Figure 8a. For samples annealed in

air mixtures, and swollen PMMA molecules. By assuming that the PMMA film at the dewetting site (hole) is replaced by DMF molecules after the dewetting process, we can compare the ΔG of the initial state (PMMA film) and the final state (DMF wetting film). Because the content of DMF in air is low (∼1.77 wt % calculated from the vapor pressure of DMF at 30 °C), the DMF vapor/air mixtures are assumed to be pure air in the calculation.47 Although the wall surface of AAO nanopore is composed of γ-Al2O3, here we use the data of α-Al2O3 for approximation. Additionally, the data of DMF are taken from that of formamide. The effective Hamaker constant, A132 (AAO substrate: 1; air medium: 2; and PMMA or DMF film: 3), can be calculated by the equation A132 = ( A11 −

A33 )( A 22 −

A33 )

(4)

where Aii are the Hamaker constants of each component.13,50−52 The effective Hamaker constants for the initial and final states can be calculated to be ∼−3.72 × 10−20 and ∼−3.93 × 10−20 J, respectively. The polar component of spreading coefficient can be calculated from the equation Sp = γ12p − γ13p − γ32p = γ1p − 2( γ1+ −

γ3+ )( γ1− −

γ3− ) − γ3p

(5)

γpij

where are the polar components of the interfacial tensions, γpi are the polar components of the surface tensions, γ+ are the electron acceptor (proton donor) parameters, and γ− are the electron donor (proton acceptor) parameters.6,53−55 The values of Sp for the initial and final states can be calculated to be ∼5.22 × 10−3 and ∼−1.01 × 10−2 J m−2, respectively. Although the values of A132 and Sp of initial and final states are determined, the exact values of ΔG of both states cannot be calculated using eq 1 because the thicknesses (h) of those states are difficult to be measured. The thicknesses (h), however, should be much larger than the correlation length (l), and the values of ΔG caused by short-range interactions can be ignored (exponential decay); i.e., the long-range interactions dominate the values of ΔG. By assuming that the thicknesses of the initial and final states are equal, the value of ΔG is larger as the A132 becomes more negative. The larger value of the excess intermolecular interaction free energy of the final state than that of the initial state indicates that the swollen PMMA chains can be replaced by DMF molecules and can dewet on the surfaces of the AAO nanopores; i.e., the final state is preferred thermodynamically. This phenomenon can also be interpreted as the stronger attractive interactions between DMF molecules and alumina wall than that between PMMA film and alumina wall. We then consider the possibility of the growth mechanism of the water droplet, as shown in the bottom part of Figure 7. Because the interactions between water and alumina walls are stronger than those between swollen PMMA and alumina walls, the moisture in atmosphere (originating from DMF or the residual moisture in the annealing chamber) can diffuse through the swollen PMMA films and precipitate on the interfaces between the wall surface of AAO nanopores and the swollen PMMA films. Once the nucleus of water droplets is formed on the interfaces between the swollen PMMA and alumina walls, the domains of water droplets are enlarged with the annealing time. The growth mechanism of water droplets may result from Ostwald ripening, coalescence, and hydrodynamic flow.25 For the Ostwald ripening, which is driven by the reduction of the interfacial energy between the polymer-

Figure 8. PMMA (Mw = 68.5 kg mol−1) nanotubes without hole formation fabricated by annealing the nanotube-containing AAO templates in different solvent vapors at 30 °C for 24 h: (a) toluene, (b) tetrahydrofuran (THF), (c) acetone, and (d) o-dichlorobenzene (ODCB).

THF or acetone, hydrogen bonds are formed between the solvent molecules and the surfaces of the AAO nanopores. These two solvents can also absorb a considerable amount of moisture in air. After annealing the PMMA nanotubes by pure THF and acetone, no hole is formed on the PMMA nanotubes, as shown in Figures 8b and 8c, respectively. The results can be attributed to the weaker hydrogen bonds acceptor ability (lower pKHB) between the solvent molecules (THF and acetone) and the nanopores than those between the DMF molecules and the nanopores.60 For samples annealed in vapor 6247

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules

Figure 9. Porous PMMA nanotubes with different molecular weights fabricated by annealing the nanotube-containing AAO templates in DMF vapor at 30 °C for different times: (a) PMMA nanotubes (Mw = 30 kg mol−1) annealed for 15 h, (b) PMMA nanotubes (Mw = 68.5 kg mol−1) annealed for 15 h, and (c) PMMA nanotubes (Mw = 183 kg mol−1) annealed for 24 h.

beams and longer exposure times, however, it may be possible to observe shifting and shrinking of the nanotubes during the SEM and TEM measurements. To demonstrate the versatility of this work, we also apply this approach to fabricate porous nanotubes using polystyrene (PS), another commonly used polymer. Similar to the PMMA nanotubes, PS nanotubes with smooth surfaces can be first prepared by the melt wetting method at 230 °C for 1 h, as shown in Figure S5a. By annealing the nanotube-containing AAO templates in DMF vapor for 24 h, porous PS nanotubes are also formed, as shown in Figure S5b. The average hole diameters of porous PMMA and PS nanotubes prepared by the same conditions (in DMF vapor at 30 °C for 24 h) are ∼70 and 332 nm, respectively. It is interesting to observe that the sizes of the holes on the PS nanotubes are larger than those formed on the PMMA nanotubes, which may be attributed to the weaker interactions between PS and alumina walls than those between PMMA and alumina walls. After the nucleation sites are formed on the nanotubes, the dewetting rate of the swollen PS is expected to be faster than that of swollen PMMA, resulting in larger holes on the PS nanotubes. To demonstrate possible future works in functional materials, we also fabricate porous polymer nanotubes using pyreneended PMMA (Py-PMMA), a fluorescent material that is synthesized by atom transfer radical polymerization (ATRP). The SEM images and PL spectra of the Py-PMMA nanotubes before and after the hole formation are shown in Figure S6. From Figure S6a,b, it is demonstrated that porous nanotubes of fluorescent Py-PMMA can also be prepared by the solventinduced dewetting method. The PL spectra of the Py-PMMA nanostructures are measured by exciting those samples at 365 nm. The differences of the PL spectra are due to the better packing of the polymer chains after the solvent annealing process. Additionally, we have tried to use the porous PMMA nanotubes as etching masks to form holes on the wall surface of AAO nanopores. After the porous PMMA nanotubes are removed by methyl ethyl ketone (MEK) for 24 h, the etched AAO templates can be infiltrated with PS solutions and PS nanostructures with bumps can be formed, as shown in Figure S7. The PS nanostructures with bumps can also be fabricated by forming PS/PMMA core−shell nanostructures and removing the PMMA shells (Figure S8).

of ODCB, a high polar solvent that does not absorb water, no hole is formed on the PMMA nanotubes, as shown in Figure 8d. The result shows that solvents with higher polarity and stronger interactions, especially hydrogen bonds, between solvent vapor and alumina wall are necessary for the formation of porous PMMA nanotubes. These results demonstrate that the mechanism of the hole formation is less relevant to the growth of water droplets on alumina walls because holes are not observed when THF and acetone are used as annealing solvents. Therefore, we believe that the dewetting mechanism, as shown in the upper part of Figure 7, is the dominant mechanism on the formation of the holes on the PMMA nanotubes. For the samples annealed in DMF vapor, one way to suppress the formation of the holes on the nanotubes is increasing the amount of water in DMF, as demonstrated in Figure 6. Another feasible way is increasing the molecular weights of PMMA. Figure 9 shows the PMMA nanostructures by using PMMA with different molecular weights (Mw = 30, 68.5, and 183 kg mol−1). After the samples are annealed in pure DMF vapor at 30 °C for 15 h, PMMA nanotubes with different morphologies can be observed. As shown in Figure 9a,b, PMMA nanotubes with larger holes are observed for the samples with the lowest average molecular weight (Mw = 30 kg mol−1), while nanotubes with smaller holes are observed for the samples with the medium average molecular weight (Mw = 68.5 kg mol−1). For the samples with the highest average molecular weight (Mw = 183 kg mol−1), no hole is formed, and smooth PMMA nanotubes are obtained even after the samples are annealed for 24 h, as shown in Figure 9c. At higher molecular weights, DMF is harder to diffuse and swell the PMMA chains. Therefore, the PMMA chains obtain lower chain mobilities, and the formation of the holes on PMMA nanotubes is prevented. To clarify whether the dewetting phenomenon and the formation of the holes on the tube walls are related to the electron beams during the SEM and TEM measurements, we have carried out real-time SEM and TEM measurements on the nanotubes samples. For our regular examination of the polymer samples, the times of focusing and acquisition in the SEM and TEM measurements are less than 30 s. Therefore, we have exposed the samples up to 600 s in the real-time measurements to study the effect of electron beams. The SEM and TEM images of the PMMA nanotubes exposed to electron beams at different exposure times are shown in the Supporting Information (Figures S3 and S4). No hole is formed on the PMMA nanotubes during the electron beam irradiation, indicating that the dewetting and the formation of the holes are not caused by the electron beams. With higher-energy



CONCLUSION We study the dewetting of polymer thin films on curved substrates by annealing PMMA nanotubes in nanopores of 6248

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules

(16) Xu, L.; Shi, T. F.; An, L. J. J. Chem. Phys. 2008, 129 (4), 7. (17) Chen, D.; Zhao, W.; Wei, D. G.; Russell, T. P. Macromolecules 2011, 44 (20), 8020−8027. (18) Ruffino, F.; Grimaldi, M. G. J. Nanopart. Res. 2013, 15 (9), 17. (19) Martin, C. R. Science 1994, 266 (5193), 1961−1966. (20) Martin, J.; Maiz, J.; Sacristan, J.; Mijangos, C. Polymer 2012, 53 (6), 1149−1166. (21) Steinhart, M.; Wehrspohn, R. B.; Gosele, U.; Wendorff, J. H. Angew. Chem., Int. Ed. 2004, 43 (11), 1334−1344. (22) Zhang, M. F.; Dobriyal, P.; Chen, J. T.; Russell, T. P.; Olmo, J.; Merry, A. Nano Lett. 2006, 6 (5), 1075−1079. (23) Feng, X. D.; Mei, S. L.; Jin, Z. X. Langmuir 2011, 27 (23), 14240−14247. (24) Lee, C. W.; Wei, T. H.; Chang, C. W.; Chen, J. T. Macromol. Rapid Commun. 2012, 33 (16), 1381−1387. (25) Wei, T. H.; Chi, M. H.; Tsai, C. C.; Ko, H. W.; Chen, J. T. Langmuir 2013, 29 (32), 9972−9978. (26) Chi, M. H.; Kao, Y. H.; Wei, T. H.; Lee, C. W.; Chen, J. T. Nanoscale 2014, 6 (3), 1340−1346. (27) Chen, J. T.; Wei, T. H.; Chang, C. W.; Ko, H. W.; Chu, C. W.; Chi, M. H.; Tsai, C. C. Macromolecules 2014, 47 (15), 5227−5235. (28) Chu, C. J.; Chung, P. Y.; Chi, M. H.; Kao, Y. H.; Chen, J. T. Macromol. Rapid Commun. 2014, 35 (18), 1598−1605. (29) Ko, H. W.; Chi, M. H.; Chang, C. W.; Su, C. H.; Wei, T. H.; Tsai, C. C.; Peng, C. H.; Chen, J. T. Macromol. Rapid Commun. 2015, 36 (5), 439−446. (30) Masuda, H.; Fukuda, K. Science 1995, 268 (5216), 1466−1468. (31) Li, A. P.; Muller, F.; Birner, A.; Nielsch, K.; Gosele, U. J. Appl. Phys. 1998, 84 (11), 6023−6026. (32) Steinhart, M.; Wendorff, J. H.; Greiner, A.; Wehrspohn, R. B.; Nielsch, K.; Schilling, J.; Choi, J.; Gosele, U. Science 2002, 296 (5575), 1997. (33) Chen, J. T.; Lee, C. W.; Chi, M. H.; Yao, I. C. Macromol. Rapid Commun. 2013, 34 (4), 348−354. (34) Mei, S. L.; Feng, X. D.; Jin, Z. X. Soft Matter 2013, 9 (3), 945− 951. (35) Chang, C. W.; Chi, M. H.; Chu, C. W.; Ko, H. W.; Tu, Y. H.; Tsai, C. C.; Chen, J. T. RSC Adv. 2015, 5 (35), 27443−27448. (36) Dersch, R.; Steinhart, M.; Boudriot, U.; Greiner, A.; Wendorff, J. H. Polym. Adv. Technol. 2005, 16 (2−3), 276−282. (37) Wu, D. C.; Xu, F.; Sun, B.; Fu, R. W.; He, H. K.; Matyjaszewski, K. Chem. Rev. 2012, 112 (7), 3959−4015. (38) Cao, F.; Pan, G. X.; Xia, X. H.; Tang, P. S.; Chen, H. F. J. Power Sources 2014, 264, 161−167. (39) Rao, D. W.; Lu, R. F.; Meng, Z. S.; Wang, Y. H.; Lu, Z. L.; Liu, Y. Z.; Chen, X.; Kan, E. J.; Xiao, C. Y.; Deng, K. M.; Wu, H. P. Int. J. Hydrogen Energy 2014, 39 (33), 18966−18975. (40) Park, K. H.; Kim, S. J.; Gomes, R.; Bhaumik, A. Chem. Eng. J. 2015, 260, 393−398. (41) McHale, J. M.; Navrotsky, A.; Perrotta, A. J. J. Phys. Chem. B 1997, 101 (4), 603−613. (42) Chen, G.; Soper, S. A.; McCarley, R. L. Langmuir 2007, 23 (23), 11777−11781. (43) Manring, L. E.; Sogah, D. Y.; Cohen, G. M. Macromolecules 1989, 22 (12), 4652−4654. (44) Fragala, M. E.; Compagnini, G.; Torrisi, L.; Puglisi, O. Nucl. Instrum. Methods Phys. Res., Sect. B 1998, 141 (1−4), 169−173. (45) Rajkumar, T.; Vijayakumar, C. T.; Sivasamy, P.; Sreedhar, B.; Wilkie, C. A. J. Therm. Anal. Calorim. 2010, 100 (2), 651−660. (46) Ruffino, F.; Torrisi, V.; Marletta, G.; Grimaldi, M. G. J. Appl. Phys. 2012, 112 (12), 11. (47) Cui, X. L.; Chen, G. M.; Han, X. H. J. Chem. Eng. Data 2006, 51 (5), 1860−1861. (48) Evchuk, I. Y.; Musii, R. I.; Makitra, R. G.; Pristanskii, R. E. Russ. J. Appl. Chem. 2005, 78 (10), 1576−1580. (49) Wexler, A. J. Res. Natl. Bur. Stand., Sect. A 1976, 80A (5−6), 775−785. (50) Medout-Marere, V. J. Colloid Interface Sci. 2000, 228 (2), 434− 437.

AAO templates via solvent vapor treatment. After annealing the samples by DMF vapor, dewetting occurs and porous PMMA nanotubes can be fabricated. The sizes of the holes of the porous nanotubes can be controlled by changing the annealing times or the water contents in the DMF/water mixture. For example, the hole formation on PMMA nanotubes can be reduced by increasing the water content in the DMF/water mixture above 12 wt % or using polymers with higher molecular weights. To further confirm the mechanism of the hole formation, other annealing solvents such as toluene, THF, acetone, and ODCB are also used. This method can also be applied to PS, another commonly used polymer. In the future, we will use the porous polymer nanotubes as templates for preparing core−shell nanostructures or highly branched nanorods. In addition, we will apply this approach using functional materials such as conjugated polymers, biocompatible polymers, or polymer composite materials to fabricate porous nanostructures that can be used for drug delivery, photoelectric devices, and sensing applications.



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.5b01207. Experimental scheme and SEM images of AAO templates and PS nanostructures (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]; Tel 886-3-5731631 (J.-T.C.). Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the Ministry of Science and Technology of the Republic of China. REFERENCES

(1) Steward, P. A.; Hearn, J.; Wilkinson, M. C. Adv. Colloid Interface Sci. 2000, 86 (3), 195−267. (2) Hickner, M. A.; Ghassemi, H.; Kim, Y. S.; Einsla, B. R.; McGrath, J. E. Chem. Rev. 2004, 104 (10), 4587−4611. (3) Sangaj, N. S.; Malshe, V. C. Prog. Org. Coat. 2004, 50 (1), 28−39. (4) Ahmetli, G.; Sen, N.; Pehlivan, E.; Durak, S. Prog. Org. Coat. 2006, 55 (3), 262−267. (5) Zhao, C.; Li, L. Y.; Guo, M. M.; Zheng, J. Chem. Pap. 2012, 66 (5), 323−339. (6) Sharma, A. Langmuir 1993, 9 (3), 861−869. (7) Reiter, G.; Sharma, A.; Casoli, A.; David, M. O.; Khanna, R.; Auroy, P. Langmuir 1999, 15 (7), 2551−2558. (8) Vanoss, C. J.; Chaudhury, M. K.; Good, R. J. Chem. Rev. 1988, 88 (6), 927−941. (9) Sharma, A.; Jameel, A. T. J. Colloid Interface Sci. 1993, 161 (1), 190−208. (10) Reiter, G. Phys. Rev. Lett. 1992, 68 (1), 75−78. (11) Reiter, G. Langmuir 1993, 9 (5), 1344−1351. (12) Sharma, A.; Reiter, G. J. Colloid Interface Sci. 1996, 178 (2), 383−399. (13) Lee, S. H.; Yoo, P. J.; Kwon, S. J.; Lee, H. H. J. Chem. Phys. 2004, 121 (9), 4346−4351. (14) Mukherjee, R.; Bandyopadhyay, D.; Sharma, A. Soft Matter 2008, 4 (10), 2086−2097. (15) Xue, L. J.; Han, Y. C. Langmuir 2009, 25 (9), 5135−5140. 6249

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250

Article

Macromolecules (51) Fotland, P.; Askvik, K. M. Colloids Surf., A 2008, 324 (1−3), 22−27. (52) Otsuki, A.; Dodbiba, G.; Fujita, T. Mater. Trans. 2007, 48 (5), 1095−1104. (53) Holysz, L.; Chibowski, E. Langmuir 1992, 8 (2), 717−721. (54) Lopes, F. A.; Morin, P.; Oliveira, R.; Melo, L. F. Colloids Surf., B 2005, 46 (2), 127−133. (55) Vanoss, C. J.; Giese, R. F.; Good, R. J. Langmuir 1990, 6 (11), 1711−1713. (56) Song, S. W.; Torkelson, J. M. Macromolecules 1994, 27 (22), 6389−6397. (57) Redon, R.; Vazquez-Olmos, A.; Mata-Zamora, M. E.; OrdonezMedrano, A.; Rivera-Torres, F.; Saniger, J. M. Rev. Adv. Mater. Sci. 2006, 11 (1), 79−87. (58) Siclovan, O. P.; Zappi, G.; Soloveichik, G. L. ECS Electrochem. Lett. 2014, 3 (12), H41−H43. (59) Wannatong, L.; Sirivat, A.; Supaphol, P. Polym. Int. 2004, 53 (11), 1851−1859. (60) Kamlet, M. J.; Taft, R. W. J. Am. Chem. Soc. 1976, 98 (2), 377− 383.

6250

DOI: 10.1021/acs.macromol.5b01207 Macromolecules 2015, 48, 6241−6250