Article pubs.acs.org/Langmuir
Accelerated Amidization of Branched Poly(ethylenimine)/Poly(acrylic acid) Multilayer Films by Microwave Heating Kehua Lin,† Yuanqing Gu,‡ Huan Zhang, Zhe Qiang, Bryan D. Vogt,* and Nicole S. Zacharia* Department of Polymer Engineering, University of Akron, Akron, Ohio 44325, United States S Supporting Information *
ABSTRACT: Chemical cross-linking of layer-by-layer assembled films promotes mechanical stability and robustness in a wide variety of environments, which can be a challenge for polyelectrolyte multilayers in saline environments or for multilayers made from weak polyelectrolytes in environments with extreme pHs. Heating branched poly(ethylenimine)/ poly(acrylic acid) (BPEI/PAA) multilayers at sufficiently high temperatures drives amidization and dehydration to covalently cross-link the film, but this reaction is rather slow, typically requiring heating for hours for appreciable cross-linking to occur. Here, a more than one order of magnitude increase in the amidization kinetics is realized through microwave heating of BPEI/PAA multilayers on indium tin oxide (ITO)/glass substrates. The cross-linking reaction is tracked using infrared spectroscopic ellipsometry to monitor the development of the cross-linking products. For thick films (∼1500 nm), gradients in cross-link density can be readily identified by infrared ellipsometry. Such gradients in cross-link density are driven by the temperature gradient developed by the localized heating of ITO by microwaves. This significant acceleration of reactions using microwaves to generate a well-defined cross-link network as well as being a simple method for developing graded materials should open new applications for these polymer films and coatings.
■
INTRODUCTION
While lengthy cross-linking times may be acceptable for fundamental research, any commercial application would require significant reductions in reaction time. One example of polymer technology that has been commercially enabled by reductions in reaction time is patterning of photoresists used in the microelectronics industry.20 Initial photoresist platforms were based on the direct chemical conversion of light to induce differential solubility for patterning,21 but catalyst-driven chemical amplification22 has allowed for further miniaturization of this technology due to significant improvements in the photon efficiency and throughput.23 This photochemistry concept can be incorporated into custom synthesized polyelectrolytes to rapidly and selectively cross-link PEMs.24 Alternatively, a small molecule photosensitive cross-linker can be diffused into the PEM and subsequently photoexposed to cross-link the film.25,26 Similarly, the need for fast directed selfassembly of nanostructures27 has led to the development of a wide variety of rapid annealing technologies, including rapid thermal processing,28 photothermal treatment,29 laser spike annealing,30 and microwave annealing.31,32 These rapid processing techniques allow processing at temperatures above the degradation temperature of the polymer if the duration is sufficiently short.33 Among these, microwave treatment has been shown to effectively enhance the reaction kinetics to generate metal oxides in inorganic−organic hybrid films.34
Polyelectrolyte multilayers (PEMs) are thin films or coatings created by the directed complexation of oppositely charged polyelectrolytes, often assembled by sequential deposition steps in a process known as layer by layer (LbL).1,2 These materials have been proposed for potential applications as barrier materials,3 water purification materials,4,5 antireflection coatings,6 and coatings to modulate surface wettability7 and for a range of biomedical devices8 such as patches for drug delivery9 or coatings for stents or microneedles10 for ultrasonic drug delivery. However, in some of the aforementioned applications, the reversible nature of the ionic cross-links that comprise these films may not be suitable. For example, PEMs are commonly swollen significantly in water or even dissolved in salt solutions because of effective charge screening, including in phosphatebuffered saline (PBS) solutions that represent physiologically relevant environments.1112,13 This reversibility of the ionic cross-links may be particularly relevant in biomedical applications9 or salt water treatment,14 for which PEMs are often proposed. For environments like these, some degree of covalent cross-linking may be desirable to provide sufficient mechanical stability and robustness.15 PEMs containing amine and carboxylic acid groups can be cross-linked through amidization, by methods such as heating the multilayers,16−19 which has limited efficiency in the solid state, or EDC coupling,16 which also has limited efficiency because it is a heterogeneous reaction. As a result, cross-linking even thin films can take more than 10 h. © 2016 American Chemical Society
Received: May 31, 2016 Revised: August 20, 2016 Published: August 22, 2016 9118
DOI: 10.1021/acs.langmuir.6b02051 Langmuir 2016, 32, 9118−9125
Article
Langmuir
borosilicate tube. The power, set temperature, and time of exposure were controlled during microwave heating. The temperature of the substrate was confirmed independently using a thermal camera (Testo model 875-1). A thermal image of the PEM film during microwave heating is shown in Figure S1. The conventional heating process was performed using a preheated muffle furnace (Ney Vulcan model 3130) as a control for comparison. For conventional heating, the sample was heated by the furnace for a given duration (total heating times of 30, 90, 150, 210, 270, 330, 390, 540, 660, and 780 min at 180 °C and 40, 70, 90, 150, 210, 270, 330, 510, 630, and 750 min at 200 °C). When the films were removed from the furnace, the samples were quickly placed onto an aluminum block at room temperature to quench the reaction. For microwave heating, the sample was heated by microwave for a given duration (total heating times of 1, 2, 5, 9, 15, and 23.5 min at 180 °C and 1.5, 3, 6, 12, and 20 min at 200 °C). When the films were removed from the microwave, the reactions were quenched in the same manner as conventional heated samples, although the temperature decrease is generally quite rapid upon cessation of the microwave energy.35 Characterization. The thickness of the (BPEI/PAA)n films was determined by an UV−visible−NIR ellipsometer (VASE, M-2000, J. A. Woollam Co.) using 246−1690 nm light with incidence angles of 65°, 70°, and 75°. For fitting of these ellipsometric data, a Cauchy layer was used to describe the optical constants of the BPEI/PAA film.36 An IR ellipsometer (Mark II, J. A. Woollam Co.) was used to determine the film thickness and optical constants over the wavelength range of 1.7− 30 μm. The IR ellipsometry measurements were performed ex situ as the cross-linking rate is negligible at ambient temperature. The samples were measured by the IR ellipsometer prior to heating to obtain the initial state of the films. The IR spectra of the PEMs were recorded with a spectral resolution of 2 cm−1 at an incident angle of 75° and were fit using WVASE software (J. A. Woollam Co.). Figure S2 shows a representative fit of the ellipsometric data for the ITO/ glass substrate, which is then used to facilitate the fitting of the ellipsometric data for BPEI/PAA PEMs on these substrates (Figure S3). The optical constants were modeled using Gaussian oscillators to describe the IR absorption by the molecular vibrations of the PEM films. These oscillators can be directly associated with the chemistry of the film. For the amidization reaction, oscillators (r, reactant) at 1500 and 1560 cm−1, corresponding to -NH3+ and -COO−, respectively, provide direct evidence of the reaction. As the film becomes denser during the cross-linking reaction, the oscillator at 1400 cm−1 associated with the -CH2- group was used to normalize the changes in area of the other oscillators as this group does not participate in the cross-linking reaction. Additionally, there are three distinct oscillators associated with the formation of the amide I bond in the cross-linking product at 1620.7, 1637.7, and 1668.6 cm−1. These three oscillators are associated with the different chemical environment around the amide I bond. To determine the cross-linking conversion, the sum of the area of these three amide I oscillators (p, product) normalized by the -CH2- peak area (i, invariant) provides a simple expression for conversion
In this work, we report on the efficacy of microwaves in accelerating the amidization of PEM films made from branched poly(ethylenimine) and poly(acrylic acid) (BPEI/PAA) on ITO substrates. The ITO substrates efficiently absorb the microwave energy to rapidly reach (77.5% conversion is obtained in 1.5 min. With microwave heating, the plateau conversion (84.2%) is similar to that for conventional heating but is obtained within only 5 min of heating. This represents a more than one order of magnitude increase in the amidization kinetics with microwave heating. Even when the cross-linking temperature is decreased to 180 °C, the same plateau conversion is obtained within 10 min when using microwaves. As reaction rate constants generally double for every 10 °C increase in temperature,39 this significant difference in the cross-linking kinetics suggests that microwave heating of the ITO substrate only partially describes the microwave interactions that promote cross-linking. Microwave processing of polymer films tends to rely solely on substrate heating,40 but here the order of magnitude difference in kinetics is more consistent with prior arguments associated with the “microwave effect”.41,42 Water, which is a reaction byproduct, is known to have a large microwave cross section.43 The components of this PEM system are hydrophilic, and in these films, water is known to be present at a level as high as 10 wt % after fabrication, even after vacuum drying.38 The bound water in these PEMs is difficult to dissociate from the amines and carboxylic acids even upon heating to relatively high temperatures. Microwaves interact strongly with water, so water could be driven from the PEMs more effectively with microwaves to then provide an improved driving force for the amidization by Le Châtelier’s principle. Additionally, significant acceleration of reactions has been previously demonstrated using microwaves when functional groups involved in the reaction include amine44 or carboxylic acid.45 The localized absorption of microwave energy by these functional groups may facilitate these reactions. However, heating the substrate from the ITO is critical as the reaction in the microwave does not proceed when the same PEM is coated on glass and subjected to the same microwave conditions (Figure S6). This result demonstrates that microwave heating of the ITO and subsequent heat transfer to the PEM is important for driving this amidization reaction, but there is a secondary effect associated with the interactions of the microwaves with the reactants (amines/carboxylic acids) that provides extra energy locally and also additional microwave heating of the residual water as well as the evolved water in the film to promote its removal from the PEM. In addition to changes in the chemistry of these films, cross-linking also tends to increase the surface roughness of the (BPEI/PAA)6 film (Figure S7). The as-cast film exhibits a root-mean-square (rms) surface roughness of 3.1
(3)
where the subscript t refers to the area of the oscillator at time t while the subscript 0 refers to the initial area for the oscillator prior to the reaction. However, the limited intensity of this oscillator (r) leads to uncertainty in the conversion as the ellipsometric data can be fit almost as well as without this oscillator at high conversions. In these cases, the absorption associated with the amidization products is more prominent (oscillators shaded blue in Figure 2B), thus providing a more accurate method for calculating conversion using eqs 1 and 2. For all conversion data reported here, the latter method has been used to calculate the conversion in these films. Figure 3 illustrates the time-dependent evolution of the conversion based on the appearance of the amide I bond from the IR ellipsometry data. As the oscillator fits of the ellipsometric data are all independent of each other, this provides a simple route for checking the conversions by comparing that determined from the appearance of the amide I bond and the conversion determined by the consumption of the limiting reactants (Figure S5). The conversion is nearly indistinguishable between these methods, which provides confidence in the methodology used to analyze the ellipsometric data. For conventional heating in a furnace at 200 °C, the conversion increases rapidly to 30.6% in the first 30 min but then slowly increases over the next 420 min to a plateau of 9121
DOI: 10.1021/acs.langmuir.6b02051 Langmuir 2016, 32, 9118−9125
Article
Langmuir
cross-linked film.47 Similar EIS measurements on noncovalently cross-linked PEMs show significantly lower impedance (nearly 2 orders of magnitude) and semi-infinite diffusion of ions through the membrane to allow determination of the ion diffusion coefficients. For all the (BPEI/PAA)6 films examined, semi-infinite diffusion of ions is not observed (lack of a sloped straight line at low frequency) even with the measurements extended down to 0.01 Hz, and therefore in this case, EIS is used to elucidate only information about the charge transfer resistance associated with diffusion of the ion through the PEM films. Via examination of films with similar conversion, the highfrequency behavior is not impacted by the selection of the methodology for cross-linking, with the high-frequency intercept associated with the ohmic resistance invariant across all samples (Z′ ∼ 2−5 Ω). This behavior is due to the similarity of the resistance of the electrolyte. At intermediate and low frequencies, there is a marked depression in the curvature in the spectra for the cross-linked films using microwave heating. These more depressed pseudosemicircles in Nyquist plots have been associated with the increased roughness of the surface or inhomogeneity of coatings.48−51 The cross-link network that forms from microwave heating as determined by atomic force microscopy (AFM) (Figure S7) is rougher than that obtained from conventional heating. The EIS spectra of cross-linked PEM films were fit using an equivalent circuit to quantitatively understand any differences between conventional and microwave heating for cross-linking as shown in Figure S8. As the conversion increases, the overall resistivity increases for both conventionally heated and microwave-treated PEMs. At approximately 30% conversion, the resistivity of the microwave cross-linked sample is around 113 Ω, which is slightly higher than that after conventional thermal cross-linking (107 Ω). This difference in resistivity between microwave and conventional heating increases with an increase in conversion. At 74% conversion by microwave heating, the resistivity of microwave heating samples is increased to approximately 168 Ω, while it is only 144 Ω for the conventional cross-linked films, even though the conversion is slightly higher (77%). The higher resistivity and higher surface roughness upon microwave heating may be attributed
nm, while after the plateau conversion has been reached, the rms roughness increased to 5.8 and 9.5 nm for conventional and microwave heating, respectively. This difference in the surface morphology suggests that the cross-linked network that forms upon amidization may depend on the heating protocol. Defects in chemical cross-linked hydrogel networks, which can be tuned by kinetically controlling the cross-linking reaction, can impact mechanical properties.46 Similarly, the significant increase in the cross-linking rate of the (BPEI/ PAA)6 films may be expected to have an impact on the topology of the network that forms. To determine whether there are any differences in the structure imposed by the two different heating mechanisms, the permeability of ions through the cross-linked (BPEI/PAA)6 film was examined by EIS. EIS allows the dynamics of electrochemical processes to be elucidated through the application of a small-amplitude oscillatory potential to the system at equilibrium. Orders of magnitudes in times scales are probed with EIS by changing the ac frequency of the perturbation. The impedance provides a measure of the resistance in the system. The Nyquist plot in Figure 4 shows how the impedance of (BPEI/PAA)6-coated
Figure 4. Nyquist plots of the impedeance associated with (BPEI/ PAA)6 cross-linked to different extents by conventional (○) and microwave heating (×). Nominal conversions of 30, 65, and 80% are examined to compare conventional and microwave heating.
electrodes is affected by cross-linking using either microwave or conventional heating. The effective resistivity (impedance) of the (BPEI/PAA)6 film increases as the conversion (associated with the reaction to amide) increases for both heating methods, which is consistent with decreased ion mobility through the
Figure 5. Comparison of the ellipsometric angle, Ψ, of thicker (BPEI/PAA)17 after cross-linking using (A) conventional heating at 180 °C for 510 min and (B) microwave heating at 180 °C for 5 min. These conditions should yield similar conversions based on (BPEI/PAA)6 data in Figure 3. The black dashed line is the best fit of the data using the same model that was used for the (BPEI/PAA)6 film with significant error (as evidenced by the residual of >1) in the fit for the microwave-heated sample. For these samples, a graded model improves the fit (blue dashed line) with the residual shown by the blue line. (C) Conversion profile from the best fits for both (empty symbols) conventional heating and (filled symbols) microwave heating to film thickness. 9122
DOI: 10.1021/acs.langmuir.6b02051 Langmuir 2016, 32, 9118−9125
Article
Langmuir
the sample. The low thermal conductivity of polymers limits the heat transfer from the surface of the ITO through the film thickness, which results in a gradient in cross-linking density. Additionally, the heat is generated only locally at the ITO with microwaves. The local environment around the sample remains cool with microwaves (Figure S1), while the air around the sample is heated significantly when we attempt to generate a gradient with a hot plate. It should be noted that the improvement in the fit of the ellipsometric data for the thin films [(BPEI/PAA)6] with inclusion of a gradient is negligible, so the gradients in the cross-link density can be resolved only in thicker films. The difference between conventional and microwave heating is distinguished upon examination of the influence of film thickness on conversion. Figure 6 illustrates the time-depend-
to the localized heating of the hydroxyl groups by the microwaves that could provide different energy landscapes for the reaction across all potential cross-linking sites. Considering the heating during microwave processing is commonly localized at the inorganic substrates rather than within the polymer film,34,40 the heat transfer through the BPEI/PAA PEM could lead to a temperature gradient in a sufficiently thick film that may have an impact on the crosslinking reaction at different depths into the film. Figure 5 illustrates the ellipsometric angle, Ψ, of a (BPEI/PAA)17 film after cross-linking to approximately 60% conversion with both conventional and microwave heating [based on conditions for the (BPEI/PAA)6 film]. The (BPEI/PAA)17 film is approximately 1800 nm thick prior to cross-linking. The same oscillator model as used previously for the thinner PEM can adequately fit the IR ellipsometric data for conventional heating (Figure 5A). The residual, which is the difference between the predicted value from the model and the measured value, has an absolute value of
