Probing the Molecular Ordering and Thermal Stability of Azopolymer

Sep 26, 2016 - Polyelectrolyte layer-by-layer (LbL) films have many applications, but several parameters and procedures during film fabrication determ...
0 downloads 9 Views 4MB Size
Download PDF
Article pubs.acs.org/Langmuir

Probing the Molecular Ordering and Thermal Stability of Azopolymer Layer-by-Layer Films by Second-Harmonic Generation Heurison S. Silva† and Paulo B. Miranda*,‡,§,∥ †

Universidade Federal do Piauí, Campus Universitário Ministro Petrônio Portella, Bairro: Ininga, CEP, 64049-550 Teresina, PI, Brazil Instituto de Física de São Carlos, Universidade de São Paulo, CP 369, São Carlos, SP 13566-590, Brazil § Center for Nano Science and Technology (CNST@POLIMI), Istituto Italiano di Tecnologia, Via Pascoli 70/3, Milan, MI 20133, Italy ∥ Dipartimento di Fisica, Politecnico di Milano, Piazza Leonardo da Vinci 32, Milan, MI 20133, Italy ‡

ABSTRACT: Polyelectrolyte layer-by-layer (LbL) films have many applications, but several parameters and procedures during film fabrication determine their morphology and molecular arrangement, with important practical consequences. Here we have used optical second-harmonic generation (SHG) to investigate the molecular ordering of LbL films containing the anionic azopolymer PS-119 and the cationic polyelectrolyte PAH. We show that spontaneous drying leads to laterally homogeneous and isotropic films, while the opposite occurs for nitrogen-flow drying. The effect of film thickness and pH of the assembling/rinsing solutions on the molecular ordering was also investigated. The optical nonlinearity tends to significantly decrease for thicker films (∼10 bilayers), and a slight alternation of SHG intensity for films with odd or even number of layers (complete vs incomplete bilayers) was also observed, which results from the reorientation of azopolymer groups in the last layer after adsorption of an additional PAH layer. We propose a qualitative electrostatic model to explain the pH dependence of film growth and azopolymer orientation, which is based on changes of the charge density of the substrate and PAH and on different ionic screening of electrostatic interactions at various pH values. We also found that the nonlinear response presents a gradual and significant reduction upon heating, which is inconsistent with a glass transition temperature for these ultrathin LbL films. The thermal stability is improved with a combination of low ionic strength and higher charge density of the polyelectrolytes and substrate, which promotes better interlayer complexation. The SHG signal is recovered upon cooling, although for some conditions the molecular arrangement became anisotropic after a heating/cooling cycle. Such detailed information about the structural order of thin nonlinear optical azopolymer LbL films demonstrates that SHG is a powerful technique to probe the film structure at the molecular level, with important consequences for their applications in optical devices. studying polymer films of nanometric thickness. These ultrathin films could have quite different properties with respect to thick (“bulklike”) films. Furthermore, for thick films their surface properties may also be very different from the bulk.16 Even though these thermal techniques can be used to probe freestanding films (usually at least a few micrometers thick), they usually cannot be applied to films adsorbed on solid substrates or at liquid interfaces, such as layer-by-layer (LbL), Langmuir− Blodgett (LB) films, or Langmuir films. Lutkenhaus et al. have used an alternative approach for thermal characterization (DSC and TGA) of LbL films fabricated with PEO/PAA17 and PAH/ PAA.18 Very thick films (∼100 or 200 layers, with an average thickness per bilayer of ∼80 nm) were assembled on an Teflon substrate and subsequently removed to be investigated by DSC and TGA. They found that for PEO/PAA films (assembled by

1. INTRODUCTION Self-assembled electrostatic layer-by-layer (LbL) films were initially described by Decher et al. in the 1990s.1,2 Since then, many applications have been found for these films, such as organic diodes,3,4 optical storage,5 biosensors,6,7 and drug delivery,8,9 to name a few. However, it is known that many parameters and procedures during film fabrication (concentration, polyelectrolyte and substrate types, pH, ionic stretch, temperature, drying, etc.)8,10−14 determine their morphology and molecular ordering, with important consequences for their applications. In particular, optical storage and nonlinear optical applications usually require these films to have good thermal stability over long periods of time. The thermal behavior of polymers is usually studied by experimental techniques such as differential scanning calorimetry (DSC), thermal gravimetric analysis (TGA), and dynamic mechanical analysis (DMA). Although these techniques are very useful to characterize the thermal properties of bulk materials and free-standing films,15 they are not suitable for © XXXX American Chemical Society

Received: July 6, 2016 Revised: August 20, 2016

A

DOI: 10.1021/acs.langmuir.6b02486 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

SHG signal. SHG measurements as a function of polarization and sample azimuthal angle investigate how their molecular organization and anisotropy depend on fabrication conditions, such as the number of layers and the pH of the self-assembly solution. We also attempt to determine the Tg for these LbL films (if it indeed exists) by probing the effect of heat treatment on molecular ordering and their thermal stability.

secondary interactions) a glass transition temperature (Tg) could be detected, but not for PAH/PAA (where the assembly is based mainly on electrostatic interaction). However, this approach has two drawbacks: those films are not really ultrathin, since they are typically 2−8 μm thick, and the methodology is not very sensitive the influence of the substrate on the conformation of the initial layers, since their contribution is negligible in comparison to that of the other hundreds of layers. Therefore, it remains very challenging to determine the Tg of nanometric polymer films on a substrate, such as LbL films with thicknesses of only a few bilayers. Ferreira et al. have studied the thermal stability of the photoinduced birefringence of PAZO/PAH LbL films at temperatures up to the bulk Tg. From thermogravimetry and FTIR spectroscopy they have confirmed that no chemical changes occurred. They reported that films fabricated at higher pH, with PAH weakly ionized, are more stable at high temperatures. However, another interesting result is that the maximum birefringence decreases with the temperature, as expected due to enhanced chain flexibility and mobility with temperature.19 Ellipsometry and Raman spectroscopy20 are other optical techniques may also be used to probe the thermal behavior of thin polymer films. Although they can determine the Tg of the films, in good agreement with the Tg found by other techniques, these techniques do not provide information on the changes in the molecular arrangement across the glass transition. In order to fully characterize such thin polymer films, it is also desirable to probe not only their molecular ordering and thermal behavior but also how they change with the fabrication parameters. In particular, LbL thin films containing azopolymers, which have potential applications in optical storage5 and nonlinear optical devices,21,22 may have their molecular organization and thermal stability probed by nonlinear optical techniques such as second-harmonic generation (SHG). SHG is very useful to study and characterize nanometric structures, e.g., structural properties of nanopillars,23 study of kinetics of adsorption,24 or molecular ordering in azopolimeric thin films as a function of adsorption parameters or when the sample is heated, since it is sensitive to polar alignment of azochromophores groups.25−27 As the glass transition is associated with a change in molecular order and/or dynamics, second-order nonlinear optical techniques (such as SHG) are therefore good probes for Tg.28−30 At higher temperatures, the molecules acquire a progressively more random and dynamic conformation due to increased thermal motion, resulting in lower orientational ordering and consequently a reduced SHG signal. Thus, for a polar-ordered molecular thin film, a sudden change of the SHG signal should be observed at a phase transition.31 With respect to azopolymer LbL films, Lvov et al.32 measured the SHG signal for PDDA/ PAZO LbL films during heating and observed a significant drop of the SHG signal at ∼120 °C, implying that these films were not thermally stable, since their ordering was nearly destroyed beyond that temperature. In contrast, some reports have described LbL films showing thermal stability around 20%.33 Here we extend our previous study about molecular ordering of azopolymer LbL films27 to investigate how the fabrication parameters influence their molecular ordering and thermal stability. We will emphasize on probing qualitative changes in the molecular orientation of these LbL films by SHG, including molecular reorientation during LbL azopolymer adsorption, which will be directly probed by phase measurements of the

2. METHODS AND EXPERIMENTAL SECTION Second-harmonic generation (SHG) is a second-order nonlinear optical process whereby a high-intensity (pulsed) laser beam with frequency ω is incident on the sample and generates a reflected beam at the second-harmonic frequency, 2ω. The theory is well established, and we refer the reader to the literature for a detailed treatment.30,34−36 Here, we will briefly describe only the essential features for interpretation of results. The detected SHG signal is proportional to square of the incident laser intensity I at the fundamental frequency ω and to the square of the effective second-order susceptibility of the 30 sample, χ(2) eff : (2) 2 2 S(2ω) ∝ |χeff |I

(1)

As a polar third rank tensor, χ(2) changes sign under the inversion (2) (2) = −χ−i−j−k ), but in centrosymmetric media all operation (χijk directions are equivalent and χ(2) should remain unchanged (χ(2) ijk = (2) ). Thus, under the electric dipole approximation, second-order χ−i−j−k nonlinear optics is forbidden in media with inversion symmetry. However, processes like SHG are possible at interfaces between two different media or in the bulk of noncentrosymmetric materials. Therefore, SHG spectroscopy is a powerful tool for investigating surfaces and interfaces and is also very sensitive to the average molecular orientation of dipolar organic moieties. Here we will use it to probe the arrangement and molecular ordering of thin layer-by-layer polyelectrolyte films. In our experimental setup (see below) the SHG frequency 2ω is resonant with an electronic transition of the azogroups of one of the polyelectrolytes, so that the generated SHG signal will be dominated by these moieties. Hence, we will probe with SHG the average orientation of azochromophores integrated in the whole LbL film (for films thinner than the coherence length, ≲100 nm). In order to study the molecular orientation of self-assembled azopolymer films, LbL films of the cationic PAH (poly(allylamine hydrochloride), Aldrich, Mw = 15 000) and the anionic azopolymer PS-119 (Aldrich, Mw unknown) were prepared from solutions with 1 mg/mL concentration. PAH and PS-119 were used as received without further purification. pH values were adjusted to either 3.5, 7.0, or 10.0 by addition of HCl (Quemis, 37% analytical grade) or NaOH (Aldrich, electronic grade, purity of 99.99%) solutions. All solutions were prepared using ultrapure water (Milli-Q), with resistivity higher than 18 MΩ·cm. Substrates used were BK7 glass windows (10 × 30 mm2, 4 mm thick). They were cleaned for 30 min with sulfuric acid and hydrogen peroxide (H2SO4 + H2O2) solutions at 7:3 proportions (“piranha solutions”), followed by extensive washing with ultrapure water and drying with nitrogen flow. All films were fabricated by the electrostatic self-assembly technique.1,2 The adsorption time for each layer was 3 min, followed by rinsing in a washing solution (pure water pH adjusted to the same as the polyelectrolyte solutions). All films were prepared with spontaneous drying at the end of the fabrication procedure in order to retain its homogeneity, like previously reported for PAH/PSS films.14 This drying by slow water evaporations process yields more uniform films and better molecular complexation by electrostatic interaction,13 in contrast to the nitrogen drying process.27 Figure 1a shows the chemical structures of the polyelectrolytes used in this work. Figure 1b represents the absorbance of films at 475 nm (near the maximum of the absorption spectrum, shown in the inset) as a function of the number of bilayers, for different pH values used in the fabrication. We can see that films grow in a linearly fashion, suggesting that the same amount of azopolimer is adsorbed at each bilayer. Adsorption is more efficient at pH 10.0, since PAH is weakly charged in the basic environment, resulting in a more coiled B

DOI: 10.1021/acs.langmuir.6b02486 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (2) 2 (2) iϕref (2) iϕfilm 2 |χeff | = ||χref |e + |χfilm |e | (2) 2 (2) 2 (2) (2) = |χref | + |χfilm | + 2|χref ||χfilm | cos(Δϕ)

(2)

where Δϕ = (ϕfilm − ϕref) = (2π/λ)Δl is the phase difference between two contributions, and Δl is the difference of optical path length due to a compensator inserted in the detection beam path (fused silica window). Figure 2 sketches the modified experimental setup that we

Figure 1. (a) Scheme of the chemical structures of PAH and PS-119. (b) UV−vis absorbance at 475 nm for PAH/PS-119 films at several pH values. The inset shows the absorption spectrum of PS-119, and an arrow indicates the SHG wavelength.

Figure 2. Experimental arrangement for SHG phase measurements. used for SHG phase measurements. The optical path difference Δl = Δnd/(cos θ) for the SHG and pump beams from the sample to the detector can be adjusted by the angle θ of the compensator and also depends on the compensator thickness (d = 3 mm in our setup) and the difference in refractive indices of the fundamental and SHG beams, Δn. Hence, the phase difference between the two SHG signals of wavelength λ is given by Δϕ = (2π/λ)(Δnd/(cos θ)). Measuring the SHG signal as a function of the compensator angle θ yields an interference pattern which depends on ϕfilm. If the average molecular orientation of chromophores in the LbL flips from up to down, ϕfilm changes by 180° and the SHG interference pattern changes accordingly. Therefore, the SHG phase measurements allow us to compare the relative molecular orientation (up or down) among different samples. SHG intensities as a function of temperature for the SP polarization combination were recorded in order to probe the thermal behavior and stability of the polyelectrolyte films. They were placed in a temperature-controlled sample cell, and the temperature was varied from room temperature (∼25 °C) until 190 °C, with ∼0.34 °C/min heating rate.

conformation of PAH molecules (also increasing the PAH layer thickness and roughness). Also in this pH the high charge density on the substrate promotes PAH adsorption and creation of more sites for PS-119 adsorption. This behavior is in agreement with other reports.37,38 For example, such pH tuning of the film thickness was previously described by Zucolotto et al.39 They reported that it was possible to vary the thickness per bilayer by more than 1 order of magnitude (from 10 to 240 Å) only by varying the pH from 4 to 10. They also reported that at high pHs the PAH is only partially charged, adopting a coiled conformation. PS-119 is strongly active in second-harmonic generation due its electronic resonance at the SHG wavelength of 532 nm, when excited by a 1064 nm laser beam, while PAH is optically inactive. Thus, we are probing only azo-groups of the PS-119 polyelectrolyte, while PAH is used only for the LbL film assembly. This facilitates the interpretation of the SHG results. For SHG studies, we used a pulsed Nd3+:YAG laser to excite the samples (30 ps pulses, repetition rate 20 Hz, wavelength 1064 nm, 2.0 mJ pulse energy at the sample). Polarizers are used to determine the polarization combination, which will be denoted by two letters indicating the polarization of the input and output beams, respectively (for example, SP means S-polarized input and P-polarized SHG signal). The area of input beam on the sample was approximately 2 mm2. The angles of incidence/reflection were 60°. For measurements as a function of the sample azimuthal angle Ω (with respect to the LbL dipping direction), it was positioned at computer-controlled rotation/ translation stage, and the incident laser spot was aligned to cross the rotation axis at the sample surface, so that we probed the same region as the sample azimuthal angle was scanned. As a complex number, the effective second-order susceptibility of (2) iϕ the film can be written as χ(2) eff = A/(2ω − ωvg + iΓ) = |χeff |e , where | is its modulus and ϕ is the phase. Since this phase is related to the |χ(2) eff relative orientation (up or down) of molecules in the film, it is often necessary to experimentally measure it to gain further insight into the molecular order of the sample. This phase measurement is usually performed by interference between the SHG signal from sample (polymer films) and that from a nonlinear reference material, such as crystalline quartz or zinc sulfide, ZnS.40,41 In this procedure, we measure an effective χ(2) eff which is the sum of that from the reference (2) iϕref iϕfilm + |χ(2) . As the detected SHG and from the film: χ(2) eff = |χref |e film|e , we have signal is proportional to square of χ(2) eff

3. RESULTS AND DISCUSSION 3.1. Molecular Orientation. SHG signal was recorded as a function of azimuthal angle Ω at SP (S in and P out SHG signal) and SS (S in and S out SHG signal) polarization combinations for PAH/PS-119 films, varying the number of monolayers and the pH of the assembly solutions. It is known that for samples with in-plane isotropic ordering and net polar alignment along the surface normal the SP polarization combination is allowed and the SHG signal is usually intense, while that for SS polarization combination is only nonvanishing if the sample is anisotropic along the surface plane.27,36 Absence of signal for the SS polarization is indicative of an in-plane isotropy at a length scale smaller than the SHG wavelength (∼0.5 μm). As an initial assessment, we investigate the homogeneity of the films. Figure 3a shows SHG measurements as a function of azimuthal angle Ω for a one-bilayer film fabricated at pH 10.0, obtained at three different spots. These measurements show that the films are homogeneous (on the millimeter length C

DOI: 10.1021/acs.langmuir.6b02486 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

clear trend of increase or decrease in SHG intensity as the thickness increases (at least for films of up to four bilayers). The samples fabricated at other pH values remain isotropic, but the thickness dependence of SHG intensity has a different behavior than for pH 7.0. For better visualization and comparison among different sets of samples, Figure 5 displays the average intensities of the azimuthal SHG scans with SP polarization as a function of the number of layers, for samples fabricated with three different pH values. As can be seen in Figure 5, the effective value of χ(2) (which is proportional to the square root of the SHG signal) does not increase linearly with the number of bilayers, in contrast to the adsorbed amount of PS-119, as shown in Figure 1. For pH 3.5 we observe an initial reduction of SHG intensity from 1 to 2 bilayers, and then an increase of the signal with thickness up to 10 bilayers, but it is again reduced for thicker films, being significantly lower with 30 bilayers. Moreover, similar to the films fabricated at pH 7.0, it is also noted an alternating intensity upon adsorption of each polyelectrolyte layer (complete vs incomplete bilayers, or integer vs half-integer n), at least for the initial few bilayers. For pH 10.0, the SHG signal is initially much smaller than for the other pH values, rapidly decreases for the first few bilayers, and then remains low for films with up to 20 bilayers. We now discuss this thickness dependence of the SHG signal. The data in Figure 5 show that for the initial few bilayers the SHG signal grows with the number of layers for low pH (3.5), it is reduced for high pH (10.0), and remains unchanged for almost neutral pH (7.0). However, the absorbance measurements in Figure 1 indicate that for each pH value the adsorbed amount of PS-119 per bilayer is constant, and the film growth rates are not so different among these three pH values. Thus, we conclude that the azopolymer chains do not remain with the same degree of ordering as the film grows, and their molecular ordering varies considerably with the pH of the assembling solution. There are reports in the literature of a linear increase of χ(2) for azopolymer films with the number of bilayers, even for more than 10 bilayers.43−46 Specifically, for the films of PAH/PS-119 at pH 3.5 some authors43,44 report a linear growth of χ(2) for films up to 100 bilayers thick. In contrast, as described above and shown in Figure 5, we observe a very complex behavior of the SHG signal with the number of bilayers, for LbL films fabricated with the same polyelectrolytes, in disagreement with refs 43,44, and 46. We note that for films fabricated at other pH values the increment of χ(2) per bilayer was also not constant. This behavior was reproduced in different sets of samples manufactured with freshly prepared solutions. A similar behavior was observed for PDDA/PAZO LbL films by Lvov and co-workers,32 where they reported an increase of χ(2) until five bilayers, but a significant reduction for thicker films. At this moment we do not find any explanation for this discrepancy between our experimental results and those of refs 43, 44, and 46. However, we have previously reported changes in the molecular conformational of LbL films after adsorption of a subsequent layer,14 indicating that the polymer layers remain quite “flexible” even in an LbL assembly. Therefore, it would be very surprising to observe a linear increment of χ(2) per bilayer because that would mean that each PS-119 layer had an average ordering identical to the previous ones (and in the same direction), especially considering additional effects like interpenetration of layers and increased roughness with the number of layers, as has been observed for films of PAH/PAA,47 POMA/PVS,48 and PAH/Ma-coDR13.10,49 Specifically, the decrease in SHG signal at large

Figure 3. (a) SHG measurements in three different points of (PAH/ PS-119)1 film fabricated with spontaneous drying (pH 10). (b) SP SHG measurement from two different (PAH/PS-119)10 films (pH 10) fabricated with spontaneous drying (△) and nitrogen flow (○).

scale), in agreement with our previous results for PAH/PSS films fabricated with spontaneous drying process.13 The opposite was observed for LbL films of another azopolymer, PAH/Ma-co-DR13, which were dried by nitrogen flow, and the presence of orientational domains and inhomogeneity were found by SHG and BAM (Brewster angle microscopy) measurements.27,42 For investigating if the anisotropy either is a characteristic of that particular azopolymer or is linked to the fabrication procedure, we performed SHG measurement for two different (PAH/PS-119)10 films fabricated at same pH 10.0, but one with nitrogen flow drying and another by spontaneous drying for 40 h. As displayed in Figure 3b, nitrogen flow drying generates an in-plane preferential ordering, but not exactly along the dipping direction (Ω = 0). On the contrary, spontaneous drying does not generate any anisotropy (and results in a weak net polar ordering; see discussion below) for the (PAH/PS-119)10 film. SHG measurements as a function of the azimuthal angle Ω for PAH/PS-119 films prepared at (neutral) pH 7.0 are shown in Figure 4. Their molecular arrangement is isotropic in the sample plane, since there is a strong signal with SP polarization that is independent of sample orientation. Furthermore, signals with SS polarization are nearly zero for all films, since such polarization combination is forbidden for isotropic films. However, the SP signal changes with the number of layers, and there is a slightly alternation of the SHG intensities as the film is terminated by either PAH or PS-119, but there is no D

DOI: 10.1021/acs.langmuir.6b02486 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 4. SHG intensity as a function of sample azimuthal angle for films fabricated at pH 7.0 and with increasing number of layers.

electrostatic interactions that lead to a highly extended polyelectrolyte conformation and LbL films with thin and strongly interacting layers. This explains why the adsorbed amount per layer of PS-119 is the lowest of the three set of samples (see Figure 1b). The trend of nearly constant SHG intensity vs number of layers, with a slight alternation in intensity for complete vs incomplete bilayers, is similar to our previous observation with SFG spectroscopy for LbL films of PAH/PSS fabricated at this pH,14 so that we propose a similar explanation. The strong electrostatic interaction among successive layers induces a reconfiguration of the PS-119 anionic layer sandwiched between two cationic PAH layers into a nearly centrosymmetric arrangement that does not contribute to SHG. The situation is different for the first PS-119 layer, which “feels” the substrate because it interacts with one PAH layer that is in contact with a substrate of lower negative charge, and for the last layer of PS-119, which does not have PAH on top (for complete bilayers). Therefore, only the first (in the case of incomplete bilayers) or both the first and last layers (for

number of layers for acidic and basic pH implies that the adsorption of the last layers is affected by the average ordering of the already adsorbed LbL film, being adsorbed in opposite orientation to the average of the previous layers. If the first few layers were ordered and the following ones were disordered (with azo groups pointing up or down with equal probability), the SHG signal should initially increase with film thickness and then saturate at a constant value for thicker films. In order to rationalize the pH dependence of the molecular ordering for these LbL films, we should keep in mind that PAH is a weak polyelectrolyte whose charge density is quite sensitive to pH, and similarly for the charge density of the glass surface. For pH 7.0, ∼20% of the surface SiOH groups of the glass surface are deprotonated,50 yielding a moderate negatively charged substrate, and PAH is ∼85% ionized.51,52 Furthermore, at this pH the ionic strength of the solutions, defined as the half-sum of products of molar concentration of each ion present in a solution (ci) by the square of their respective valence (Zi), I = (1/2)∑iciZi2, is quite low (∼7 μM), favoring long-range E

DOI: 10.1021/acs.langmuir.6b02486 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

chains more coiled. Therefore, the LbL films at pH 3.5 are slightly thicker than at pH 7.0, with an increase in roughness as well as a reduction of electrostatic interactions caused by the increase of ionic strength. Thus, at pH 3.5 the adsorbed amount of PS-119 per layer is higher than at pH 7.0, as shown in Figure 1b, but not as high as at pH 10, due to the very low charge density of the substrate. However, as shown in Figure 5, at pH 3.5 the trend of the SHG signal with the increasing number of bilayers is very complex. While the SHG signal is quite significant for the first bilayer (PAH/PS-119), indicating that the PS-119 chromophores should be oriented on average toward the cationic PAH layer, it nearly vanishes for the film with three layers, (PAH/PS119)/PAH. Thus, the third layer (PAH) modifies the orientation of the previously adsorbed PS-119 layer, leading to a nearly symmetric configuration of the chromophores along the z-direction of film growth. This indeed should be expected because the two PAH layers of the PAH/PS-119/PAH film are highly and equally charged, exerting practically the same influence on the central layer for electrostatic ordering. Upon completing the second bilayer, the second PS-119 layer is adsorbed with only a slight preferential alignment (presumably toward the PAH layer), since there is only a minor increase of SHG intensity. The additional PAH layer of the 2.5 bilayer film again induces reordering of this second PS-119 layer, reducing once more the SHG intensity. Finally, the third complete bilayer now presents a considerable SHG signal, indicating the ordered adsorption of the third PS-119 layer. After this initial complex growth of the film, where the substrate influence plays a major role, it seems that the additional bilayers partially add to the average orientation of the chromophores, and the SHG susceptibility increases sublinearly with thickness. However, this coherent growth is sustained only up to ∼10 bilayers, from when the additional PS-119 layers start to adsorb on average with the opposite orientation, leading to a reduction of SHG intensity, similarly to the case of pH 10. To gain further insight into the orientation of the azo groups of PS-119 for the very thin films fabricated at pH 3.5, we performed a direct measurement of the phase of χ(2) for the SP polarization combination. The experimental setup for that was described in the Experimental Section and used as a reference sample a thin film of zinc sulfide (ZnS) sputtered on a glass substrate. As shown in Figure 6, the phase of the SHG signal from the sample is always the same for films formed at pH 3.5 with an integer number of bilayers (complete bilayers), indicating that the average orientation of the azopolymer chromophores is the same. For these films the last layer is always the azopolymer, adsorbed over a layer of PAH that is highly positively charged, and in particular for the first bilayer, it is expected that the preferred arrangement of the azo groups of PS-119 is toward the highly charged layer of PAH. Therefore, this behavior is preserved for all complete bilayers finished with PS-119, with the average orientation of chromophores pointing toward the substrate. Since the SHG signal nearly vanishes for the film with 1.5 bilayers, PAH/PS-119/PAH, it indicates a rearrangement of the PS-119 chromophores toward both layers of PAH, as observed previously for PAH/PSS films.14 For films with 2 and 2.5 bilayers, whose signals are very small, the phase measurement is difficult to perform (signals were scaled for better visualization, as indicated in the graph legend of Figure 6) but confirms the reordering of the azo groups as discussed above. For the 2.5 bilayer film (PAH/PS-119)2/PAH, which is again terminated by a layer of PAH, the PS-119 chromophores

Figure 5. SHG intensity and its square root (right axis) as a function of number of bilayers (n) for the LbL films (PAH/PS-119)n fabricated at pH 10.0, 7.0, and 3.5. The semi-integer n represents films whose previous bilayer has an additional adsorbed PAH layer on top. Error bars are the standard deviation of 73 measurements.

complete bilayers) contribute to the SHG signal, so that its intensity does not grow with thickness and shows an alternation between complete and incomplete bilayers. At pH 10.0, the glass is heavily charged (∼60%),50 but the PAH is only about 30% ionized,51,52 thus forming considerably thicker layers than at pH 7.0 or 3.5 due to a more coiled conformation. This combination of charged density values and high ionic strength (∼mM) resulted in a more efficient adsorption of PS-119, according to Figure 1. This may result from thicker PAH layers, which offer more sites for PS-119 complexation and adsorption, and also thicker PS-119 layers due to the higher ionic strength. However, under these circumstances the SHG intensity decreases for thicker films. Although the adsorbed amount per layer is large, the more coiled PAH and PS-119 layers with weaker (screened) electrostatic interactions may lead to a more mobile and disordered conformation for PS-119, which favors the polymer reorientation. Therefore, the preferential orientation of the first PS-119 layer induced by the presence of the substrate is gradually canceled by the rearrangement of subsequent PS-119 layers that point in the opposite orientation, decreasing the SHG signal. At pH 3.5, the glass substrate is only slightly charged (