SEARCH

SEARCH BY CITATION

Keywords:

  • LB films;
  • surfaces;
  • surfactants

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

Thermodynamic analyses of surface pressure-area (Π-A) isotherms and Brewster angle microscopy (BAM) reveal that poly(ε-caprolactone) (PCL) with a weight average molar mass of Mw = 10 kg mol−1 and polydispersity index of Mw/Mn = 1.25 and poly(t-butyl acrylate) (PtBA, Mw = 25.7 kg mol−1; Mw/Mn = 1.07) form compatible blends as Langmuir films below the dynamic collapse transition for PCL at Π = 11 mN m−1. For PCL-rich blends, in situ BAM studies reveal growth of PCL crystals for compression past the PCL collapse transition. PCL crystals grown in the plateau regime of the Π-A isotherm exhibit a dendritic morphology presumably resulting from the rejection of PtBA from the growing PCL crystals and hindered diffusion of PCL from the surrounding monolayer to the crystal growth fronts. The ability to transfer the PCL dendrites as Langmuir–Schaefer films onto silicon substrates spincoated with a polystyrene layer facilitates detailed morphological characterization by optical and atomic force microscopy (AFM). AFM reveals that the dendritic branching occurs along the {100} and {110} sector boundaries and is essentially independent of composition. AFM also reveals that the average thickness of PCL dendrites formed at room temperature (22.5 °C), ∼7–8 nm, is comparable with that of PCL crystals grown from single-component PCL Langmuir films and spincoated thin films. In contrast, for PtBA-rich blend films PCL crystallization is suppressed. These findings establish PCL blends as an ideal system for exploring the interplay between chain diffusion and crystal growth in a two-dimensional confined geometry. © 2007 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 45: 3300–3318, 2007


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

Thin film crystallization of semicrystalline polymers has drawn significant attention for understanding the properties of polymeric materials. Decreasing film thickness dramatically alters molecular mobility,1, 2 glass transition temperatures,3, 4 and segmental orientation5, 6 of semicrystalline polymers. These factors influence the transport of chain segments to the growth front of a crystallizing lamella, resulting in growth rates, degrees of crystallinity, morphologies, and melting behavior that differ from bulk crystals.7–29 Moreover, the reduction in dimensionality in going from a three-dimensional (3D) to quasi-two dimensional (2D) system further limits the amount of crystallizable material in the vicinity of the growth front. Indeed a number of previous studies have reported the observation of diffusion-limited crystal growth in sufficiently thin polymer films.7–29 These previously observed diffusion-limited morphologies include dense-branched morphologies, dendrites, and fractal structures.

One of the most heavily studied crystallizable polymers for thin film investigations is poly(ethylene oxide) (PEO). While single-crystal and spherulitic PEO morphologies are typical for bulk films, dendritic crystals have been observed as the PEO film thickness approaches ∼10 nm.12–17 Furthermore, blending amorphous poly (methyl methacrylate) (PMMA) with PEO has been used to “tune” the crystal morphology of PEO in mixed thin films.17–21 The blend morphology was found to evolve from spherulites, to needles, and dendrites as the composition changed from 90 to 30 wt % PEO.18 Addition of a small amount of clay (5 wt %) to PEO/PMMA blends yielded seaweed dendrites (50 wt % PEO), symmetric dendrites (30 wt % PEO), and fractal dendrites (20 wt % PEO).20, 21 Moreover, previous studies also suggest that crystallization temperature, blend composition, and PMMA molar mass play similar roles as far as morphological control is concerned because these parameters enable the tuning of the relative rate of crystal growth and diffusion.18

Thin film polymer crystallization studies have not been limited to PEO.22, 29 Taguchi et al. extensively investigated the crystal growth of isotactic polystyrene (it-PS) in ∼11-nm thick films.29 Dendritic crystals with sixfold symmetry and compact seaweed morphologies were observed with decreasing crystallization temperature. Similar diffusion-limited morphologies have also been reported for other polymer systems such as poly(styrene-block-ethylene oxide),16 poly (trifluoroethylene),30 poly(ethylene terephthalate),31 and poly(S-lactide).32

Another system, poly(ε-caprolactone) (PCL), is frequently used as a model semicrystalline polymer and will be used in this study. PCL is a biocompatible polyester with a bulk glass transition temperature of Tg = −60 °C and a melting point of Tm = 50 °C. PCL-based systems have attracted widespread interest because of their potential applications in controlled-release drug delivery33–35 and as scaffolds for tissue engineering.36–44 These applications could benefit from a better understanding of PCL's degradation mechanisms and crystallization behavior.45, 46 Crystallization studies of PCL and PCL-based polymer blends in thin film geometry have recently been reported.7–10 The growth rate of PCL crystals was found to be roughly one-half the bulk growth rate at crystallization temperatures of 50 and 54 °C for 15-nm thick films.9 In contrast, 6-nm thick films exhibit drastically slower growth rates which are comparable at both temperatures. As in the PEO case discussed above, diffusion-limited growth of PCL crystals are observed in 6 nm or thinner spincoated films on silicon substrates.9, 10

For previous studies on thin film crystallization, the polymer films are usually prepared on solid substrates by spincoating. On solid substrates, the restricted cooperative segmental motion directly affects transport processes near the crystal/melt interface and consequently the crystallization rate and morphology.6 However, the air/water (A/W) interface can be used to minimize these effects and provide a model surface for studying crystallization in thin films.47, 48 Whereas, the preparation of homogeneous ultrathin films approaching monomer-level thickness by spincoating is problematic, monolayers at the A/W interface for polymers with an appropriate balance between hydrophilic and hydrophobic character selfassemble into monomer thick layers. As a consequence, crystallization within a Langmuir film where polymer chains fold into the air presumably yields the thinnest possible stable crystals at room temperature. Furthermore, although previous studies of dendritic crystallization in thin films allow for in situ observations of the growth process and provide morphological parameters such as the branching angle, the thickness, the width, and the overgrowth of dendritic fingers, most of these studies have been only focused on the selfreorganization of polymer chains and their attachments to the growth fronts in terms of the kinetic pathway. For thin and ultrathin polymer films on solid substrates, experimental attempts to quantitatively understand the phase transformations from the amorphous phase to an ordered crystalline phase are still rare due to technical complications. Without a question, it is currently impossible to simultaneously obtain the “2D” phase diagram and the associated morphological changes during crystallization and melting processes. Here, this presentation aims to advance the capabilities of Langmuir film techniques for studying the dendritic crystallization of PCL-based blend systems in the context of recent relevant publications.47, 48 While the observation of dendritic patterns in supersaturated Langmuir monolayers is well documented for small amphiphilic molecules such as dioctadecylamine (DODA),49 ethyl palmitate (EP),49 ethyl stearate (ES),49 diacetylene 10, 12-tricosadiynoic acid,50D-myristol alanine,51, 52 and N-dodecylgluconamide,53 to the best of our knowledge, similar observations for polymers have only been first reported for PCL-based systems.47, 48 In these previous studies, we reported that PCL Langmuir films formed at low surface pressures exhibit the nucleation of crystals at room temperature (T = 22.5 ± 0.1 °C) from a meta-stable (supersaturated) monolayer at surface pressures (Π) just below the dynamic collapse pressure of ΠC, PCL = 11 mN m−1.47, 48 The growth of crystals in the subsequent plateau regime of the surface pressure-area per repeating unit (monomer for short, Π-A) isotherm was observed by in situ Brewster angle microscopy (BAM) studies. Electron diffraction studies on Langmuir–Schaefer (LS) films suggest that the lamellar crystals are oriented with the polymer chain axes perpendicular to the substrate surface, while atomic force microscopy (AFM) reveals a crystal thickness of ∼8 nm.47 In addition, a molar-mass-dependent collapse transition for PCL monolayers was observed in both dynamic compression and “equilibrium” addition experiments.48 Molar-mass-dependent morphological features of PCL crystals and their subsequent melting and respreading were also observed by in situ BAM studies during hysteresis experiments.48

In the present study, PCL is blended with amorphous poly(t-butyl acrylate) (PtBA). The choice of PtBA reflects the fact that PtBA can be spread to form stable liquid-condensed Langmuir monolayers (monomer thick) at the A/W interface, and PtBA monolayers have the appropriate viscoelastic properties to form Langmuir–Blodgett films.54, 55 The compatibility and rigidity of these blends in the monolayer regime is analyzed through the Wilhelmy plate technique and BAM. Morphological changes arising from the enhanced rigidity and hindered transport of PCL to the crystal growth fronts during dynamic compression experiments, and the melting and respreading of crystals during dynamic hysteresis experiments are examined in situ by BAM. A more extensive analysis of PCL dendrites formed in Langmuir films is obtained from optical microscopy and AFM studies performed on LS-films prepared on silicon substrates coated with polystyrene. Collectively, these measurements provide insight into the crystallization of polymers in a quasi 2D constrained geometry.

EXPERIMENTAL

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

Materials

PtBA (weight average molar mass, Mw = 25.7 kg mol−1; polydispersity index, Mw/Mn = 1.07) and PCL (Mw = 10 kg mol−1; Mw/Mn = 1.25) were purchased from Polymer Source, and were used without further purification. A series of spreading solutions were prepared by dissolving the PCL/PtBA blends in chloroform (∼0.5 mg g−1, HPLC grade) at room temperature.

Isotherm Studies

The chloroform solution was spread onto the surface of ultrapure water (18.2 MΩ, Milli-Q Gradient A-10, Millipore, <5 ppb organic impurities) using a Hamilton gas-tight glass syringe in a standard Langmuir trough (700 cm2, 601BAM, Nima Technology). The spreading solvent was allowed to evaporate by waiting for a suitable period of time (∼20 min) before the isotherm measurements were started. The trough was maintained at T = 22.5 ± 0.1 °C in a Plexiglas box with a relative humidity of 70–75%. To investigate the thermodynamic properties of PCL/PtBA mixtures at the A/W interface, surface pressure-average area per repeating unit (or monomer for short, Π-〈A〉) isotherms were measured. Π was determined by the Wilhelmy plate technique using a completely wetted filter paper plate. For compression and hysteresis loop experiments, the mixed surface layer was compressed at a rate of 8 cm2 min−1 to an arbitrary 〈A〉 and was immediately expanded at the same rate back to the initial trough area of 690 cm2.

Brewster Angle Microscopy

Brewster angle microscopy (BAM) (MiniBAM, NanoFilm Technologie GmbH, Linear resolution of at least 20 μm) studies were carried out simultaneously during hysteresis experiments to monitor the morphological transitions in the Langmuir film blends. BAM micrographs were taken with a charge coupled device (CCD) camera. The BAM images presented in this article are 2.0 × 2.4 mm2 in size unless specifically noted and were cut from the original 4.8 × 6.4 mm2 images using ACDSee 9 Photo Manager. BAM micrographs utilize an optimal brightness rather than an absolute intensity scale. While this feature provides the best optical contrast, care must be taken when comparing different images. Bright domains in one picture may be dark in the next image if a brighter object moves into the field of view. The Langmuir trough, BAM, and Plexiglas box rest on a floating optical table to minimize vibrations.

Langmuir–Schaefer Films

Dendritic PCL crystals grown in Langmuir films of PCL/PtBA blends were transferred by the Langmuir–Schaefer (LS) method onto silicon substrates possessing a spincoated polystyrene layer (1 wt % PS in toluene). The spincoated polystyrene layer was used to enhance interactions between the hydrophobic surface of PCL crystals and the solid substrates and preserve the morphological features of crystals grown at the A/W interface. The silicon wafers (100) used for spincoating were obtained from Waferworld. The silicon substrates were boiled in a 1:1:5 (by volume) solution of ammonium hydroxide (28%): hydrogen peroxide (30%): ultrapure water (18.2 MΩ, Milli-Q gradient A-10, Millipore, <5 ppb organic impurities) for 1.5 h. After the substrates were rinsed with copious amounts of Millipore water and dried under nitrogen, they were immersed in a 30:70 (by volume) solution of hydrogen peroxide (30%) and concentrated sulfuric acid for ∼3 h, rinsed with Millipore water, and dried under nitrogen. The clean substrates were then hydrophobized by treatment with HF (J. T. Baker, 1:7 buffered etch oxide) and NH4F (J. T. Baker, 40%) solutions. A compression rate of 8 cm2 min−1 was used to obtain the desired transfer pressure. The transferred LS-films were immediately used to conduct optical microscopy studies. The transferred LS-films were then stored in air for at least 24 h at room temperature before performing optical microscopy and AFM studies. Storage of the films has no effect on the morphological features as similar features were observed by optical microscopy for freshly transferred LS-films and LS-films stored for at least 24 h.

Optical Microscopy and Atomic Force Microscopy Studies

The morphological studies were performed using an optical microscope operating in reflection mode (Axiotech Vario 100 HD, Carl Zeiss) and an AFM operating in tapping mode with a Nanoscope III AFM (Dimension 3000 scope with a Nanoscope IIIa controller, Digital Instruments, set-point of ∼0.6). The spincoated polystyrene layer between the silicon substrate and the LS transferred PCL crystals is thick enough to allow us to see 7- to 8-nm thick crystals with visible light. The contrast mechanism is an interference effect arising from differences in the optical path lengths between the PS layer and the PS layer + PCL crystal for visible light reflected from the film/air and film/substrate interfaces. AFM images are reported without any image processing except flattening.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

Compression Isotherm Studies for PCL and PtBA Langmuir Films

Figure 1 shows Π-A isotherms for PCL and PtBA at the air/water (A/W) interface at a temperature of T = 22.5 °C. The isotherms for the single component PCL and PtBA Langmuir films are consistent with previous studies.47, 48, 56–60 Starting with PCL, the monolayer exhibits a gentle change in Π from 0 at A = 140 Å2 monomer−1 (not shown) to ∼11.3 mN m−1 at A = 20 Å2 monomer−1. The gentle rise in Π is characteristic of a compressible film, and the isotherm can be used to calculate the 2D analog to the bulk modulus, the static dilational elasticity, εs = κ−1 = −A(∂Π/∂A)T, where κ is the 2D analog of the bulk isothermal compressibility.61 εs values for PCL as a function of A are provided in the inset of Figure 1. As seen in Figure 1, εs exhibits a maximum value, εs,max,PCL = 14 mN m−1 at A = 37 Å2 monomer−1. εs,max,PCL is consistent with PCL forming a liquid-expanded (LE) monolayer.47, 48, 62, 63 For A < 20 Å2 monomer−1, the PCL isotherm exhibits a kink in the isotherm followed by a plateau that is indicative of film “collapse.” The kink yields a dynamic collapse pressure of ΠC,PCL = 11.3 mN m−1. As previously reported, “collapse” of the PCL monolayer during dynamic compression experiments corresponds to the nucleation and growth of crystals from a supersaturated monolayer.47, 48

thumbnail image

Figure 1. Π-A compression isotherms for Mw = 10 kg mol−1 PCL (○) and Mw = 25.7 kg mol−1 PtBA (+) at the A/W interface. The isotherms were obtained at T = 22.5 °C and a compression rate of 8 cm2 min−1. The inset shows the εs-A plot deduced from the Π-A isotherms.

Download figure to PowerPoint

Switching to the PtBA Π-A isotherm, PtBA initially shows a gentle rise in Π until ∼2 mN m−1 at A = 32 Å2 monomer−1 that is similar to PCL. However, further compression of the PtBA film causes Π to rise much more sharply for PtBA than for PCL indicating that PtBA is substantially less compressible. In fact, the maximum static dilational elasticity for PtBA, εs,max,PtBA = 110 mN m−1 at ∼23 Å2 monomer−1, is consistent with a liquid condensed (LC) monolayer.59, 62, 63 For A < = 21 Å2 monomer−1, the PtBA monolayer undergoes collapse into multilayer structures at a dynamic collapse pressure of ΠC,PtBA = 23 mN m−1.59, 60 On the basis of Figure 1, the addition of PtBA to PCL should enhance the rigidity of the resulting Langmuir film, and the ΠC,PCL value means that any analysis of monolayer compatibility needs to be restricted to Π < ΠC,PCL.

Compression Isotherm Studies for PCL/PtBA Langmuir Film Blends

Figure 2 contains Π-〈A〉 isotherms for PCL blends with PtBA at various mole fractions of PtBA, XPtBA. XPtBA values in this article are calculated on the basis of the molar mass of the repeating units rather than the molar masses of the entire polymer chain.64–66 Examination of Figure 2 clearly reveals two important regimes: Π < Πc,PCL and Π > Πc,PCL. For the higher Π region, Πc,PtBA is essentially constant and one is simply observing the effect of diluting PtBA with PCL that has already undergone “collapse”—crystallization or the formation of amorphous multilayer domains depending on XPtBA (to be discussed below). In the lower Π region, one observes that the blends show a systematic shift from single-component PCL behavior to single-component PtBA behavior as XPtBA increases. This region of the film corresponds to the Langmuir monolayer and can be used to determine the excess Gibbs free energy of mixing, ΔGexcess.

thumbnail image

Figure 2. (A) Π-〈A〉 compression isotherms for various PCL/PtBA blends at the A/W interface. The isotherms were obtained at T = 22.5 °C and a compression rate of 8 cm2 min−1 and correspond to XPtBA = 0 (PCL, ○), 0.19 (▪), 0.26 (□), 0.44 (•), 0.75 (⋄), 0.89 (▴), and 1.0 (PtBA, +). The inset highlights the essentially composition independent value of the dynamic ΠC,PCL. The arrows on (A) indicate the trend of increasing XPtBA. (B) The same isotherms in (A) are expanded to clearly show the monolayer regime.

Download figure to PowerPoint

To determine ΔGexcess for a binary mixture of PCL and PtBA in the monolayer regime, one starts with the area additivity rule for an ideal mixture64–66:

  • equation image(1)

where XPCL and XPtBA, and APCL(Π) and APtBA(Π) correspond to the mole fractions and areas per monomer of the pure components at specific Π values, respectively. Equation 1 then allows one to calculate the area change upon mixing, the 2D analog of the change in volume upon mixing in 3D:

  • equation image(2)

where 〈A(Π)〉 corresponds to the experimental values in Figure 2. Figure 3(A) shows a comparison of 〈Amix,ideal(Π)〉 with 〈A(Π)〉 as a function of XPtBA for different Π from 2 to 10 mN m−1. As seen in Figure 3(A), all mixtures show strongly negative deviation from ideal mixing at all XPtBA and Π. ΔGexcess can be obtained from eq 2:

  • equation image(3)
thumbnail image

Figure 3. (A) 〈A(Π)〉 obtained from the isotherms in Figure 2 for various PCL/PtBA blends at Π = 2 (+), 3(○), 4(•), 5(□), 6(▪), 8(▵), and 10(▴) mN m−1. The dotted lines correspond to 〈Amix,ideal(Π)〉 from eq 1 in the order of lowest to highest Π from the top to the bottom of the plot. Negative deviation from the area additivity rule is observed for all Π. (B) ΔGexcess obtained from eq 3 for the same Π values used in (A). The dotted lines are only provided to highlight the trend of increasing ΔGexcess with increasing Π, and the dotted horizontal line is provided to accentuate where ΔGexcess = 0.

Download figure to PowerPoint

Figure 3(B) shows ΔGexcess versus XPtBA at the same Π as in Figure 2. For all Π and XPtBA, ΔGexcess is strongly negative indicating that PtBA and PCL are compatible in the monolayer regime. This conclusion is supported by BAM images which are homogeneous for Π < ΠC,PCL.

The isotherms in Figure 2 can also be evaluated by considering the monolayer to be a 2D semidilute solution. During the 1980s, researchers in the field noted that in the semidilute monolayer regime, Π was molar mass independent and scaled with surface concentration as Π = Γz or Π = A−z.67 More recently, Esker et al. noted that if one assumed Π = 0 at the start of the semidilute regime, one obtained, εs = zΠ.61 Hence, the initial slope of an εs-Π plot, provides insight into interactions between water and the polymer through good solvent scaling predictions (z = 3 for a mean field model and z = 2.86 for numerical models)68, 69 where chains have a solvent swollen 2D conformation, and predictions for theta solvent conditions (z approaches infinity in a mean field model with predictions of z = 8–101 from different numerical methods)70–72 where the excluded volume of the 2D conformation matches the unperturbed 2D chain conformation. These predictions were consistent with the measurements of dynamic dilational elasticity obtained by Esker et al. from surface light scattering.61 Li et al. recently showed that the A/W interface is a good solvent for PCL through this analysis scheme.47, 48 For PtBA, the similarity of the structure and shape of the Π-A isotherm with poly(t-butyl methacrylate) would suggest that PtBA should be close to the theta solvent limit.

Figure 4(A) shows εs versus 〈A〉 for PCL/PtBA blends at various compositions. In contrast to the pure components (inset of Fig. 1), the blends show two elasticity maxima, εs,max, in two distinct regions εs,max,1 for A > = 22 Å2 monomer−1 (Π < ΠC,PCL, monolayer) and εs,max,2 for A < = 22 Å2 monomer−1 (Π > ΠC,PCL). Because of the limited compression ratio of our Langmuir trough, it was not possible to obtain both maxima for the two lowest XPtBA in a single experiment. The region corresponding to the Langmuir monolayer state for the blends is expanded as the inset in Figure 4(A). Table 1 summarizes the values of εs,max,1, and εs,max,2 and the 〈A〉 values where they occur, 〈As,max,1)〉 and 〈As,max,2)〉, respectively, for each XPtBA. For A < = 22 Å2 monomer−1 (Π > ΠC,PCL) the PCL component has “collapsed”—crystallization or the formation of amorphous multilayer domains depending on XPtBA. In this regime the observed behavior arises from PtBA that remains at the A/W interface following the “collapse” of PCL. Nonetheless, the PCL plasticizes the PtBA that remains at the A/W interface leading to progressively smaller εs,max,2 at smaller 〈As,max,2)〉 as shown in Figure 4(A) and summarized in Table 1. As this region of the isotherm is not the focus of this article, it will not be discussed further.

thumbnail image

Figure 4. (A) εs versus 〈A〉 curves for binary PCL/PtBA blends at the A/W interface. εs was calculated from the isotherms in Figure 2. The inset of (A) focuses on the behavior in the monolayer, Π < ΠC,PCL (PtBA has been excluded for clarity). (B) εs versus Π for binary PCL/PtBA blends. The lines are theoretical curves, εs = zΠ, for good solvent conditions (z = 2.86, solid line),69 the least extreme numerical value for theta solvent conditions (z = 8, dashed line)70 and the most extreme numerical value reported for theta solvent conditions (z = 101, dotted line).72 The symbols in (A) and (B) correspond to XPtBA = 0 (PCL, ○), 0.19 (▪), 0.26 (□), 0.44 (•), 0.75 (⋄), 0.89 (▴), and 1.0 (PtBA, +). (C) εs,max versus XPtBA (open symbols, left-hand axis) obtained from the inset of 4A and z versus XPtBA (filled symbols, right-hand axis) obtained from the slope of (B) in the monolayer regime, Π < ΠC,PCL. The right- and left-hand axes have been scaled to ensure that z and εs,max essentially match for the single-component films.

Download figure to PowerPoint

Table I. Elasticity Maxima for PCL/PtBA Blends at the A/W Interface
XPtBAAs,max,1)〉 Å2 monomer−1εs,max,1 mN m−1As,max,2)〉 Å2 monomer−1εs,max,2 mN m−1
  • a

    Not applicable.

  • b

    While only one maximum is observed for pure PtBA films, the values are listed twice as the PtBA component is responsible for the second peak in the blends.

0.003715N/AaN/Aa
0.193316≤5≥10
0.263216≤5≥27
0.442918844
0.7526231459
0.8925331957
1.00b2311023110

Focusing on the Langmuir monolayer regime, A > = 22 Å2 monomer−1 (Π < ΠC,PCL), there is a progressive increase in εs,max,1 and decrease in 〈As,max,1)〉 as XPtBA increases [inset of Fig. 4(A) and Table 1]. As shown in Figure 4(B), the increase in εs,max,1 is coupled with a change from good-solvent scaling behavior to poorer (near theta) solvent scaling behavior as the PtBA content of the binary blend films increases. Figure 4(C) shows that neither εs,max,1 nor z increase linearly over the entire range of XPtBA, and that z is more sensitive to XPtBA than εs,max,1 at small XPtBA. For the case of low XPtBA, the response is nearly linear and is consistent with adding rigid, nonswollen material to the film. Hence the principle effect, rising modulus, is similar to the case of adding a compatible filler to a soft polymer. In contrast, at high XPtBA the PCL and its attendant water can be thought of as a plasticizer. During compression, water must be preferentially squeezed out of the PCL before PCL is squeezed out of the monolayer. These two processes, along with the response of PtBA to compression, apparently combine to yield the nonlinear effects on z and εs.

Having established that PCL/PtBA blends form compatible Langmuir monolayers, and that the rigidity of the film increases with increasing PtBA content, it is necessary to return to the discussion of Figure 2. The inset of Figure 2 shows that the dynamic ΠC,PCL is essentially independent of XPtBA. This feature of the blend system is consistent with Li et al.'s observation that ΠC,PCL corresponds to the “crystallization” pressure for PCL Langmuir films at the A/W interface at 22.5 °C during dynamic compression experiments.47, 48 In bulk polymer systems, melting point depression is weakly dependent on composition because of the higher molar masses relative to small molecule systems. Hence, a weak composition dependence of ΠC,PCL on XPtBA can be expected. In the remainder of this article, we will focus on how the PtBA component affects the crystallization of PCL through BAM, optical microscopy, and AFM studies.

Morphological Studies of PCL Crystallization/Melting in Dynamic Hysteresis Experiments

Li et al. recently reported that PCL undergoes crystallization in dynamic compression Π-A isotherms for Π > ΠC,PCL and that both the formation and melting of these crystals in dynamic hysteresis experiments are molar mass dependent.47, 48 In these crystals, the polymer chain axis is perpendicular to the A/W interface and the lamellar thickness is on the order of 7–8 nm. More recently, work from the Duran group showed that PCL-PEO copolymers can also undergo crystallization in Langmuir films.58 Figure 5 shows representative dendritic PCL crystals grown during a dynamic compression experiment in a Langmuir film [Fig. 5(A)], and also shows the melting and respreading of these crystals during expansion of the same monolayer [Fig. 5(C–E)]. The Π-A isotherm in Figure 5 highlights where the BAM images [Fig. 5(A–E)] were recorded.

thumbnail image

Figure 5. (A–E) BAM images for PCL crystals grown from single-component PCL Langmuir films at T = 22.5 °C and a compression rate of 8 cm2 min−1. The letters on the Π-A isotherm highlight the concentrations where the BAM images were obtained during compression (A and B) 8 Å2 monomer−1, and expansion (C–E) 8.2, 17.3, and 27.2 Å2 monomer−1, respectively. Solid-like domains appear bright in the 1.28 × 0.96 mm2 image (A) and the 2.0 × 2.4 mm2 BAM images (C–E). The 0.56 × 0.56 mm2 BAM image (B) corresponds to an isolated crystal and the principle axes and crystal faces have been added to the crystal. (F) represents an optical microscopy image of an LS-film transferred during expansion at A = 17 Å2 monomer−1, a value comparable to the BAM image in (D). (G) An idealized depiction of the PCL dendrites observed during melting drawn with an orientation comparable to the BAM image in (B) and the optical microscopy image in (F).

Download figure to PowerPoint

The rationale developed by Mareau and Prud'homme for PCL crystals grown in spincoated thin films on solid surfaces9 suggest that the four symmetrically distorted sectors of PCL crystals in Figure 5(A) are of the {110} type while the remaining two are {100} sectors [Fig. 5(B)].73 In the absence of electron diffraction confirmation, this sector assignment is used in the following. For PCL single crystals grown in 30 nm spincoated films at 54 °C, the growth striations are perpendicular to the growth front in {100} sectors, while in {110} sectors, the striations are nearly parallel to the boundary line between of {110} and {100} sector.9 The diffusion-limited growth in 6-nm films distorts lamellar growth in presumably {110} sectors, indicating that molecular diffusion from a limited material reservoir is more spatially constrained.9 The striations in the four {110} sectors make an angle to the boundary line between {110} and {100} sectors.9 Meanwhile, the striations in two presumably {100} sectors still appear more or less perpendicular to the growth front and have greater contact areas to surrounding polymer melt.9 The morphological assignment used for PCL crystals grown at the A/W interface can be further rationalized by BAM studies during film expansion. During the early stages of expansion, the two presumably {100} sectors immediately start to melt and the PCL chains respread [Fig. 5(C)]. This behavior is also consistent with observations in single crystals grown in spincoated films. Mareau and Prud'homme attributed faster melting to relatively poor organization in {100} sectors.9 The lower melting point of the {100} sectors relative to the {110} sectors has also been observed for single crystals of poly(L-lactide)74 and linear polyethylene75 with similar orthorhombic unit cells. The lower melting point of {100} sectors has been attributed to lesser regularity in the folding75 (along alternating [110] and [ equation image] directions in the {100} sectors) and to larger lattice stress.76, 77 As the crystallized films are further expanded, the dendritic nature of these crystals is more clearly revealed [see the dendritic arms in {100} sectors on Fig. 5(D)]. Similar morphological features are also observed in the optical micrograph [Fig. 5(F)] obtained for a LS-film transferred at comparable A during expansion. The schematic in Figure 5(G) highlights the assignment of the presumably {100} and {110} sectors for the same orientation as Figure 5(B). In the schematic, the solid lines correspond to the orientation of the dendritic sidebranches. Further expansion of the film leads to nearly complete melting and respreading of PCL in the {100} sectors [Fig. 5(E)]. Four trunks, corresponding to the four presumable {110}/{100} sector boundaries, are the last features to melt as seen in Figure 5(E).

Morphological Studies of PCL Crystallization/Melting and Respreading in XPtBA = 0.19 PCL/PtBA Blends During Dynamic Hysteresis Experiments

For the Mw = 10 kg mol−1 PCL used in this study, the dendritic nature of the crystals formed during dynamic hysteresis experiments at Π > ΠC,PCL is only clearly evidenced during film expansion.47, 48 In contrast, PCL-rich blends with PtBA exhibit dendritic features even during compression. Figure 6 shows representative BAM images for PCL crystals grown during compression and melted and respread during the expansion step of a dynamic hysteresis experiment for a XPtBA = 0.19 PCL/PtBA blend. As noted in the discussion of PCL/PtBA compatibility, all BAM images are homogenous in the monolayer regime prior to compression into the metastable (supersaturated) monolayer regime. Upon compression into the supersaturated monolayer regime, Π close to ΠC,PCL = 11 mN m−1, small bright domains first appear (not shown here). Considering the limited resolution of the BAM image (features >20 μm can be resolved) we can only conclude that nucleation has started. With further compression of the film through the collapse transition for PCL [Fig. 6(A)] and into the plateau region of the Π-〈A〉 isotherm, these nuclei grow into larger crystalline domains [Fig. 6(B–D)]. Even with the limited resolution of the BAM images, it is clear that these crystals exhibit dendritic structures [Fig. 6(B) through (D)]. As the PCL crystals grow larger, six-arm dendritic morphologies are observed [Fig. 6(C) and (D)]. In addition, the backgrounds of the images appear to contain very small aggregates that are close in size to the resolution limit of the BAM. The nature of these features will be more clearly revealed in subsequent discussion of optical microscopy and AFM studies on LS-films prepared in these regions. Continuing with Figure 6, the film is expanded immediately after the compression step by opening the barriers at a constant expansion rate of 8 cm2 min−1. As seen in Figure 6(E), some of the PCL dendritic crystals are broken, possibly as the result of crystal impingement during the final stages of compression. Further expansion of the film reveals a plateau in the Π-〈A〉 expansion isotherm. This feature is similar to the behavior reported for pure PCL47, 48 shown in Figure 5. The plateau has been attributed to the melting and respreading of the PCL crystals in a process where the unstable edges diffuse back to the monolayers upon further expansion. During this process, the bright domains become smaller [Fig. 6(F)] and their structural details start to disappear in the BAM images as shown in Figure 6(G). Eventually, these domains become smaller than the 20-μm linear resolution of BAM, and cannot be detected at the end of the expansion step.

thumbnail image

Figure 6. BAM images for a XPtBA = 0.19 PCL/PtBA blend film obtained at T = 22.5 °C and a compression and expansion rate of 8 cm2 min−1. The letters on the isotherm indicate where the BAM images were taken during dynamic hysteresis experiments and correspond to compression (A) 18, (B) 16, (C) 12, and (D) 10 Å2 monomer−1 and expansion (E) 9.0, (F) 11, and (G) 32 Å2 monomer−1, of the film. Solid-like domains appear bright in all of the 2.0 × 2.4 mm2 BAM images.

Download figure to PowerPoint

Another interesting feature of Figure 6 is that in contrast to the case of the single component PCL film, we do not observe a preferential melting and respreading process in different sectors. One could argue that PCL dendrites formed in blends with PtBA possess uniformly poorer chain organization in all sectors than PCL crystals grown in single-component monolayers. Alternatively, in the case of PCL/PtBA blends, the respreading of PCL chains, hindered by the rigid PtBA-rich monolayer, may be the rate limiting step for the morphological change associated with melting.

To better understand the morphological features of PCL dendrites in Figure 6, optical microscopy and AFM were used to obtain higher resolution images of PCL crystals transferred to silicon substrates coated with a PS layer as LS-films. Figure 7 shows a representative optical micrograph of a single layer LS-film transferred for a XPtBA = 0.19 PCL/PtBA blend, while Figure 8 shows higher resolution AFM images of the sidebranching. Here, the power of using a PS film to allow the imaging of LS films by shifting interference effects into the visible wavelength range is on full display. First, Figure 7 shows the same six-arm dendritic morphology observed in BAM indicating that LS-transfer is suitable for retaining the crystal morphology seen at the A/W interface. Figure 7 clearly reveals that the PCL dendrites possess four mirror symmetric dendritic trunks that presumably form along the four {110}/{100}sector boundaries. Second, the resolution is sufficient to resolve the differences in branching angles (upper insets of Fig. 7) thereby allowing one to assign the {100} and {110} sectors as done in the lower inset. The four mirror-symmetric trunks that define the boundaries between the {110} and {100} sectors have an angle of α less than 90°. Exact measurements of α are complicated by sidebranching in the {100} sectors which causes the arms to bend back to the midline of the {110} sectors as seen in Figure 7. On the basis of solution crystallization studies of linear polyethylene, we know that the angle α relates to the relative rates of growth along {100} and {110} sectors and is not related to the symmetry of the growing crystal.77–81 The angle α varies with crystallization temperature, polymer molar mass and solvent quality. Side branches on the two trunks that bisect the {100} sectors experience symmetric growth fields and show a common branching angle of γ = 49.3° ± 4.6° [analysis in Fig. 7(B)]. The side branches on the four arms that define the {110} sectors exhibit asymmetric branching angles of θ = 77.2° ± 2.4° and β = 51.3° ± 3.2° [analysis in Fig. 8(A′)]. For sidebranches into the {110} sectors, the branching angle, θ, is nearly perpendicular to the main growth front, as expected based on the crystal structure of PCL.

thumbnail image

Figure 7. (A) A representative optical microscopy image of a single layer LS film prepared from a XPtBA = 0.19 PCL/PtBA blend film. The LS film was transferred onto a PS coated silicon substrate at 〈A〉 = 10 Å2 monomer−1 during compression in the plateau regime. The inset in the lower right-hand corner shows the assignment of the {100} and {110} sectors with the same orientation as the dendritic crystal in the figure. The inset in the upper left-hand corner shows an enlargement of the symmetric side branching in the {100} sector with a common angle, γ. The inset in the lower left-hand corner shows how θ and β can be resolved by optical microscopy along the boundary between a {110} and {100} sector. The white arrows represent regions where the growth of branches is inhibited by side branches from faster growing (earlier forming) branches. (B) The average geometric angle γ versus the number of sidebranches corresponding to the two dendritic trunks in the two {100} sectors in Figure 7(A). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

thumbnail image

Figure 8. Representative AFM images of a single layer LS film for a XPtBA = 0.19 PCL/PtBA blend. The LS film was transferred onto a PS coated silicon substrate at 〈A〉 = 10 Å2 monomer−1 in the plateau regime during compression. (A) a 55 × 55 μm2 height image, (A′) the average geometric angles θ and β versus the number of sidebranches corresponding to the dendritic trunk in the dotted box of (A) that divides a {100} sector from a {110} sector, (B) a 6 × 6 μm2 height image, and (B′) a cross-sectional analysis for the solid line in (B). The average thickness of the lamellar crystals is 7.1 nm. Z-scales for images (A) and (B) are 0–80 nm and 0–60 nm, respectively.

Download figure to PowerPoint

Figures 7 and 8 display two other significant features. The first is that optical microscopy (Fig. 7) confirms the tenuous conclusion of the low resolution BAM images that the background has small round structures which could be amorphous 3D aggregates. The other is that the crystal thickness from Figure 8 of 7.1 nm is comparable with values for single-component PCL samples crystallized at room temperature either in Langmuir monolayers at the A/W interface47, 48 or in films spincoated on silicon substrates.9

Effect of XPtBA on PCL Crystal Morphology in PCL/PtBA Blends During Dynamic Hysteresis Experiments

Qualitatively, dynamic hysteresis loops for PCL/PtBA blends with increasing XPtBA are similar to Figure 6. The only remarkable feature on the Π-〈A〉 hysteresis loop is a decrease in the area inside the hysteresis loop with increasing XPtBA. This feature is attributed to the inhibition of PCL crystallization during the compression step, a conclusion that is supported by BAM images. To illustrate this point, representative Brewster angle and optical microscopy images for crystals formed during compression at 〈A〉 = 10 Å2 monomer−1 are provided in Figure 9. The BAM images, Figures 9(A) through 9(C) show that both the size and number density of dendritic structures decrease with increasing XPtBA. In all three cases, the dendritic crystals exhibit four mirror symmetric {110} sectors and two mirror symmetric {100} sectors following the rationale of Mareau and Prud'homme.9 Further analysis of the dendritic crystals by AFM, Figure 10, reveals that the characteristic angles β, γ, and θ, along with the lamellar thickness are independent of XPtBA. These values are summarized in Table 2 for PCL and PCL/PtBA blends. With increasing XPtBA, the supersaturation is also diminished, and the diffusion is presumably more hindered in accord with the enhanced rigidity of the blend (Fig. 4). As a result, the diffusion length will be smaller, with an attendant decrease in the growth rate as seen in Figure 9, and the ultimate quenching of crystallization as seen in Figure 11. Once PCL crystallization is suppressed, only small clumps of small round structures are observed in AFM images (Fig. 11). A reasonable conclusion is that these structures may be similar to the 3D aggregates seen in Figure 7.

thumbnail image

Figure 9. PCL crystal morphologies for PCL/PtBA Langmuir film blends with different XPtBA at the A/W interface. The images on the left are BAM images obtained at T = 22.5 °C and an expansion rate of 8 cm2 min−1 for PCL/PtBA blend films of (XPtBA, 〈A〉/Å2 monomer−1): (A) (0.19, 10), (B) (0.26, 12), and (C) (0.44, 10). Solid-like domains appear bright in all of the 2.0 × 2.4 mm2 BAM images. The images on the right, A′, B′, and C′ are optical microscopy images for LS films transferred to PS coated silicon wafers at the same conditions as A, B, and C, respectively. The white arrows represent regions where the growth of branches is inhibited by side branches from other faster growing (earlier forming) branches.

Download figure to PowerPoint

thumbnail image

Figure 10. (A) AFM height image of a LS film transferred for a XPtBA = 0.26 PCL/PtBA blend film at 〈A〉 = 14 Å2 monomer−1 (z-scale: 0–60 nm); (A′) Crosssectional analysis for AFM image A (from the solid line on A). The average thickness of the lamellar crystals is 7.1 nm; (A″) The average geometric angles θ and β (defined by the dotted lines on A) versus the number of sidebranches corresponding to the dendritic trunk in image A along a {100} and {110} sector boundary. (B) AFM height image of a LS film transferred for a PCL/PtBA XPtBA = 0.44 blend film at 〈A〉 = 13 Å2 monomer−1 (z-scale: 0–60 nm); (B′) Cross-sectional analysis for AFM image B (from the solid line on B). The average thickness of the lamellar crystals is 7.5 nm; (B″) The average geometric angles θ and β (defined by the dotted lines on B) versus the number of sidebranches corresponding to the dendritic trunks in image B along the {100} and {110} sector boundary. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

thumbnail image

Figure 11. AFM phase images of LS films for PCL/PtBA blends obtained at (XPtBA, 〈A〉/Å2 monomer−1, Π/mN m−1): (A) (0.44, 13, 12), (B) (0.44, 13, 12); (C) (0.75, 15.6, 14); and (D) (0.89, 20, 14). Softer materials appear darker in all of the 10 × 10 μm2 AFM images with z-scales of 0°–40°. Arrows on the figure highlight features discussed in the text. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Table II. θ, β, and γ Values for PCL Dendrites Grown from PCL-Rich PCL/PtBA Belnds
XPtBA0.190.260.44
θ77.2 ± 2.479.5 ± 3.183.0 ± 3.7
β51.3 ± 3.250.9 ± 2.848.6 ± 3.6
γ49.3 ± 4.648.4 ± 3.752.5 ± 6.2

Mechanism for Dendritic Crystallization of PCL in PCL/PtBA Blends During Compression

Dendritic sidebranches normally reflect the symmetry of the crystal structure82 and grow along crystallographic angles.18, 21, 22, 29 In general, the diffusion field, either a concentration or temperature gradient at the growth front, causes an interfacial instability in the otherwise flat growth front, leading to the formation of a protrusion that ultimately gives rise to the formation of dendritic structures.83–85 In single-component thin films, it is often assumed that the gradient of the film thickness around the growth front is responsible for dendritic crystallization.29 For blend systems, the compositional instabilities can arise from a nonuniform compositional distribution in blend systems via phase separation, or from diffusion-limited local depletion of crystallizable material and the accumulation of noncrystallizable material that is rejected from the growing crystal at the growth front. Nevertheless, both compositional dendrites and thermal dendrites grow under transport-limited conditions.86–88 Although dendritic branching for PCL crystals has recently been observed in spincoated films,9 this article is the first report showing the formation of dendritic polymer crystals during compression in Langmuir monolayers. As noted by Volhardt et al.53 in studies of dendritic domains in Langmuir films of small molecules, the A/W interface is particularly attractive for studying diffusion-limited pattern formation, because the aqueous phase serves as a large heat sink relative to the single molecule thick films that is capable of quickly dissipating any latent heat thereby ensuring isothermal conditions.

The growth of dendritic structures during the compression of PCL/PtBA blends is comparable with that of compositional dendrites formed in multi-component systems such as PEO/PMMA.18–21 Starting from a stable nucleus, Mullins-Sekerka instabilites83–85 manifest themselves as protrusions along the four {110}/{100} sector boundaries and along the {100] growth front of the crystals. Crystal growth is the fastest at the tips of these apexes because the amorphous component is easily rejected from the apexes of the protrusions and the tips experience “virgin” melt. Thus, the semicrystalline component of the blend can preferentially attach on the growth front around the apexes of protrusions. Meanwhile, the accumulation of amorphous components in the vicinity of the protrusions reduces the supersaturation of polymer melt and prevents further growth in the immediate vicinity of the trailing edge of the protrusion. As a result, additional growth behind the leading edge of the tip manifests itself locally as a new protrusion where the supersaturation of crystallizable material has returned to “normal.” Once these sidebranches form they compete with each other in a limited reservoir of crystallizable polymer. As a result, faster growing branches can cut-off neighboring arms by locally depleting the crystallizable component as indicated by white arrows in Figures 7 and 9. This process, called “kinetic coarsening,” leads to different lengths of sidebranches.89 The deviations in branch length are more clearly seen in Figure 8(A). This process may also be responsible for the pronounced bend in the four principal trunks that form the boundaries between the {110} and {100} growth sectors as seen in Figures 6 and 7. The growth rates of these “secondary” and “tertiary” sidebranches are determined by the interplay of the surrounding diffusion fields, the degree of undercooling, and chain folding with respect to the growth front of the dendritic tips. The net results are the morphologies seen in Figures 5 through 11.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

Π-A isotherm and BAM studies reveal that PCL is ideal for studying polymer crystallization within Langmuir monolayer blends. Langmuir films represent the thinnest possible uniform films for PCL and exhibit relatively high compressibility (the A/W interface is a good solvent). PCL crystallization takes place at low surface pressure (Πc,PCL = 11 mN m−1) and room temperature. Crystallization of PCL is particularly amenable to blend studies because the isotherms are stable within the monolayer regime. The PCL/PtBA system exhibits the formation of dendritic crystals during compression of a homogeneous monolayer, a first for a polymer system at the A/W interface. These compositional dendrites exhibit morphologies that are comparable with diffusion-limited structures observed in polymer thin films on solid substrates. Moreover, the rheological properties of this blend are suitable for transferring the structures as LS films. Analysis of these LS films by optical microscopy and AFM in the context of previous electron diffraction and TEM studies by Mareau and Prud'homme9 suggest that these dendrites possess four symmetrically distorted {110} sectors and two {100} sectors. Within the {100} sector, the side branching angle is γ = 49° and the branching is symmetric. In contrast, the side branching along the four trunks defining the boundaries between the {100} and {110} sectors is asymmetric with θ = 80° on the {110} side, and β = 50° on the {100} side of the branch. Moreover, the crystal thickness, ∼7–8 nm, is also comparable to the threshold film thickness below which dendritic structures are observed in PCL spincoated films on solid substrates.9, 10 Hence, this study shows that Langmuir films provide a suitable framework for exploring dendritic crystal growth in thin films under ambient conditions where polymer degradation is not an issue by allowing one to vary both Π and T.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES

The authors thank Brian C. Okerberg, Jianjun Deng, and Zhenyu Huang for useful discussions, Mr. S. McCartney for his training assistance with AFM measurement, and the National Science Foundation (CHE-0239633) for financial support.

REFERENCES AND NOTES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES AND NOTES