Polymer surface structures determined using ToF-SIMS

End-group distribution at the surface of flat-on lamellae

Polymer crystallization is an active research area in polymer physics. But many of its features remain poorly understood. In particular, the kinetic pathway from a disordered coiled state to a perfectly ordered state cannot be easily explained by a simple theory of first-order nucleation and growth owing to the connectivity of polymer chain segments. Recently, due to advances in atomic force microscopy (AFM), it has become possible to examine the morphology of a crystalline polymer surface at the molecular level (Li et al. 1999, 2001, Lei et al. 2002, 2003, Jiang et al. 2003a,b,c, Luo et al. 2003, Wang et al. 2004, 2008, Chan and Li 2005, Zhou et al. 2005). At such a surface, the polymer crystals have two basic orientations: flat-on and edge-on orientations. Figure 1 is a schematic diagram showing these two orientations (Wang et al. 2008). Wang et al. (2008) found that edge-on and flat-on lamellae tend to develop at the surface of a polymer film at low and high temperatures, respectively. At temperatures above the glass transition temperature (T_g) of a polymer, edge-on lamellae are mostly observed at the surface. As the temperature increases, more flat-on lamellae will appear at the surface. At very high temperatures near the melting point of the polymer, the surface is filled mostly with flat-on lamellae. The surface morphology of a semicrystalline polymer film was also found to be dependent on the film thickness. Flat-on lamellae tend to develop in ultra-thin polymer films. However, at large film thicknesses, the concentration of edge-on lamellae would increase even at very high crystallization temperatures because flat-on lamellae, which originate at the interface between the polymer and the substrate, take a long time to appear at the surface (Chan and Li 2005, Wang et al. 2008).

Figure 1

Schematic showing edge-on and flat-on lamellae. The end groups of the polymer are bromine. Reproduced with permission from Chan and Li (2005). Copyright 2005 Springer-Verlag.

The crystallization of poly(bisphenol-A-co-ether octane) (BA-C8) at room temperature, which has been studied using AFM (Li et al. 1999, 2001, Lei et al. 2002, Chan and Li 2005, Wang et al. 2008), started with the formation of a nucleus that grew in length to become a founding lamella, as shown in Figure 2A–F. The synthesis and the structure of BA-Cn polymers are shown in Scheme 1. When the founding lamella had grown to a certain length, it began to branch, producing many daughter lamellae and eventually leading to the formation of a lamellar sheaf, as shown in Figure 2P. It is important to point out that the lamellae observed were all in the edge-on orientation because crystallization occurred at about 20°C, which was near the T_g of the polymer.

Figure 2

A series of AFM phase images obtained on a BA-C8 film. (A) An embryo; (B) a short lamella (founding lamella) developed from the embryo shown in (A); (C–F) growth of the founding lamella; (G–P) branching and splaying apart of the subsidiary lamellae. Reproduced with permission from Lei et al. (2002). Copyright 2002 American Chemical Society.

Scheme 1

Synthesis and structure of BA-Cn polymers. Reproduced with permission from Chan and Li (2005). Copyright 2005 Springer-Verlag.

Figure 3A–C shows a condensed but complete picture of the crystallization process (Li et al. 2000). An amorphous surface, which was formed by spin coating the BA-C8 polymer solution on a silicon wafer, is shown in Figure 3A. The morphologies of the surface after crystallization at 25°C for approximately 10 h and 1 week are shown in Figure 3B and C, respectively. In addition to AFM, ToF-SIMS chemical imaging can also provide interesting information about the conformation of the polymer chains at the surface. The BA-C8 polymer chains were capped at both ends with bromine atoms, which could be easily detected by ToF-SIMS. Figure 3D–F show the chemical images of the BA-C8 films obtained by mapping the characteristic fragments of the end groups, the -⁷⁹Br^- and -⁸¹Br^- peaks, at mass-to-charge (m/z)=79 and 81. Figure 3D shows a homogeneous distribution of Br, suggesting that the polymer chains existed as random coils in the freshly prepared amorphous film. A spherulite of ∼10 μm in diameter can be observed in the AFM height image (cf., Figure 3B) after about 10 h. A comparison between the AFM topographic image (cf., Figure 3B) and the ToF-SIMS chemical image (cf., Figure 3E) reveals that the size of the spherulite in the former was similar to the sizes of the spherulites (the areas with higher Br concentrations) in the latter. After 1 week, the polymer film was fully crystallized with many impinged spherulites, as shown in Figure 3C and F.

Figure 3

Surface morphology and end-group distributions of the BA-C8 polymer films: (A, D) freshly prepared films; A, B and C are AFM height images. D, E and F are ToF-SIMS Br ion images. (B, E) after isothermal crystallization at 25°C for approximately 10 h; (C, F) after isothermal crystallization at 25°C for approximately 1 week. Reproduced with permission from Li et al. (2000). Copyright 2000 American Chemical Society.

It is interesting to note that the bromine concentration was the highest at the interspherulitic boundaries as shown in Figure 3F. The distribution and rearrangement of the bromine end groups at the boundaries might be driven by the expulsion of the end groups from the lamellar structure. The expelled end groups were oriented in the amorphous areas among the lamellae of a spherulite, as illustrated schematically in Figure 1. The AFM and ToF-SIMS results suggest that the segments and end groups of the BA-C8 polymer can rearrange their conformations at temperatures above the polymer’s glass transition temperature. They also suggest that the surface morphology and chemical structure are governed by the dynamic process of polymer chain arrangement. This dynamic process can be very slow for some polymer materials, as is the case for the BA-C8 polymer.

End-group distribution at the surface with a concentric ringed structure

Wang et al. (2006) showed that poly(bisphenol A hexane ether) (BA-C6) formed micron-sized concentric rings in thin films. The results of optical microscopy (OM) and AFM reveal that the banded structures, which were not caused by lamellar twisting, contained alternating ridge and valley bands of polymer crystals in the flat-on orientation. Figure 4 shows an AFM height image of a typical concentric ringed structure in a 190-nm-thick BA-C6 film. The formation of this ringed structure was caused by the periodic changes in the diffusion rate of the polymer in the thin film.

Figure 4

(A) AFM image showing a typical concentric-ringed crystalline structure formed in a BA-C6 (M¯w=12,400 g/mol) film 190 nm thick annealed at 80°C for 2 days. (B) The corresponding height profile. The scale bar in (A) represents 10 μm. Reproduced with permission from Wang et al. (2006). Copyright 2006 American Chemical Society.

Molecular fractionation is a direct consequence of the crystallization process. The segregation of molecules occurs in such a way that different molecular species crystallize at different stages. During the formation of the concentric-ringed crystalline structure, the chains with relatively lower molecular weights and lower surface energies migrate to the surface of the melt and preferentially migrate to the fold surface of the flat-on crystals, which have a much higher surface energy. To reach the surface of a ridge, the polymer chains must undergo longer periods of fractionation than do those that reach the surface of a valley. Thus, the surfaces of the ridges should have a higher concentration of low-molecular-weight chains than do the surfaces of the valleys because polymer chains with different molecular weights crystallize at different stages. The low-molecular-weight chains will be the last ones to crystallize.

To test the above hypothesis, ToF-SIMS imaging was used to determine the distribution of end groups. In the synthesis of BA-C6, an excess amount of 1,6-dibromohexane was used to ensure that Br was present at both ends of the polymer chains. Therefore, the surface with a high concentration of low-molecular-weight chains should show a high surface Br concentration. ToF-SIMS Br imaging was performed on the surface of a 140-nm-thick BA-C6 film annealed at 85°C for about 48 h. The AFM height profile obtained on this film indicated that the height difference between the first ridge and the next valley (the largest height difference in this structure) was about 220 nm and that the mean horizontal distance between the neighboring mature ridges and valley bands was about 4200 nm. Figure 5A and B show the ion images using the total intensity of all negative ions as well as the intensity of the ⁷⁹Br^- and ⁸¹Br^- ions at m/z=79 and 81, respectively. Figure 5B shows a higher Br concentration at the topmost surface of the ridges. In order to eliminate the topographic effect between the ridges and the valleys, the relative intensity between the characteristic ion (Br^-) and the fragment (CH₂^-) was used. The values of Br^-/CH₂^- at the surfaces of the valleys and ridges were calculated. Based on ten measurements, the average value of Br^-/CH₂^- at the surfaces of the ridges was about 3.7 times greater than that at the surfaces of the valleys. ToF-SIMS images showed higher concentrations of low-molecular-weight polymer chains on the surfaces of the ridges than in the valleys. This result indicates that enrichment of relatively small chains occurred at the topmost surface of the ridge bands.

Figure 5

Negative ToF-SIMS ion images. (A) Total negative characteristic fragments and (B) characteristic fragments of the end group ⁷⁹Br^- and ⁸¹Br^- peaks at m/z=79 and 81 amu, respectively. Reproduced with permission from Wang et al. (2006). Copyright 2006 American Chemical Society.

Lamellar orientation at the surface

As mentioned in an earlier section, there are two lamellar orientations at the surface of a semicrystalline polymer film: the edge-on and flat-on orientations. Lau et al. (2008) used ToF-SIMS to determine the lamellar orientation (flat-on versus edge-on orientation) at the surfaces of the BA-C8 polymer with a number-average molecular weight of 9100 g/mol. BA-C8, which consists of both rigid aromatic and flexible aliphatic components, was chosen because of its unique molecular structure. Figure 6 summarizes the possible sources of the positive and negative molecular ions from the chain structure of BA-C8.

Figure 6

Structural origins of the characteristic molecular ions of BA-C8. Reproduced from Lau et al. (2008) with permission of Elsevier.

The edge-on and the flat-on lamellae can be predominantly produced at low and high crystallization temperatures, respectively (Wang et al. 2008). If BA-C8 polymers films were crystallized at increasing temperature, the surface morphology of these films would change from the edge-on orientation to the flat-on orientation. Figure 7 is a schematic diagram showing the surface chain arrangement of the melt, the edge-on lamellae, and the flat-on lamellae of BA-C8.

Figure 7

Schematic diagram showing the chain conformation in (A) the melt, (B) the edge-on orientation of a lamella (the surface of the edge-on lamellae lies in the xy plane), and (C) the flat-on orientation (indicated with an arrow) of a lamella. Reproduced from Lau et al. (2008) with permission of Elsevier.

If the intensities of characteristic ions are measured as a function of temperature, a relationship between the intensities of these ions and the orientation of the lamellae can be established. Table 1 compiles the intensities of the positive and negative characteristic ions normalized with respect to the total ion intensity, RjT, given by:

where I_j and I_T are the intensity of ion j and the total ion intensity, respectively. On the surface with the edge-on lamellae, RjT, which was calculated to be 2.6 for j=245 (repeat unit ion), is much higher than that on the surface with the flat-on lamellae (RjT=0.9) and that on the amorphous surface (RjT=0.9). It is obvious that the lateral crystal planes of the edge-on lamellae were filled with a large number of BA-C8 repeat units lying parallel to the substrate surface, including the uncrystallized bridging stems in the interlamellar amorphous regions (cf., Figure 7). Consequently, the positive secondary ion yield of the repeat units at m/z=245 was enhanced by this crystal orientation. The results from the positive and negative ion spectra are in good agreement and reveal that ToF-SIMS data can be used to distinguish between the edge-on and flat-on lamellae.

Lamellar fold structure

The above results suggest that more ions from the rigid segments were detected on the flat-on lamellae, suggesting that the fold surface was composed of the rigid segments. With this information, Lau et al. (2011) studied the effects of crystallization temperature on the chain folding development of a BA-C8 polymer. Thin films of BA-C8 having a thickness of about 50 nm were prepared by spin coating cleaned Si wafers with a dilute polymer solution of chloroform (10 mg ml^-1) at a rotating speed of 3000 rpm. Such thin films were used because Wang et al. (2008) found that flat-on lamellae were produced in ultra-thin films or thicker films at high crystallization temperatures. AFM height and phase images of a BA-C8 polymer film annealed at 35°C and 75°C for 48 h indicated that the films were not fully crystallized. A mixture of flat-on lamellae and amorphous areas was formed on the surfaces. The samples annealed at these two temperatures were also analyzed by ToF-SIMS chemical imaging in the positive ion mode, and principal component analysis (PCA) was performed on the ion images to discriminate the conformation between the flat-on lamellae and the amorphous areas. Figure 8A shows a PCA loading plot of the film annealed at 35°C for 48 h. The positive and negative loadings were related to the amorphous and folding surfaces, respectively. Two particular outliers at m/z of 69 and 135, corresponding to C₅H₉⁺ and C₉H₁₁O⁺, respectively, were largely associated with the negative loadings. The other masses that partly contributed to the negative loadings included m/z of 43 (C₃H₇⁺), 57 (C₄H₉⁺), 81 (C₆H₉⁺), 107 (C₇H₇O⁺), 147 (C₁₀H₁₁O⁺), 161 (C₁₁H₁₃O⁺), 213 (C₁₄H₁₃O₂⁺), and 247 (C₁₇H₂₇O⁺). The ion C₁₇H₂₇O⁺ at m/z=247 is the intact repeat unit of the BA-C8 polymer. For the positive loadings, there were no distinctive outliers. Moreover, the masses with positive loadings, such as m/z of 39 (C₃H₃⁺), 51 (C₄H₃⁺), 63 (C₅H₃⁺), 115 (C₉H₇⁺), 128 (C₁₀H₈⁺), 189 (C₁₅H₉⁺) and 202 (C₁₆H₁₀⁺), were hydrogen deficient, whereas the negatively loaded hydrocarbon ions were those with the same formula of C_xH_y⁺ but with a higher degree of saturation. The difference in the ion fragmentation indicates that the conformations of the precursors of the amorphous areas and folding surfaces were different. The folding surfaces most likely contained many loose loops. If the folds preferably involve elongated short stems composed of partial fragments of a rigid segment and a flexible segment with an ether linkage, the precursors will eventually lead to a very high probability of producing the ions at m/z of 69 and 135 on the surfaces of the flat-on lamellae, as indicated in Figure 8B.

Figure 8

(A) Loading plot of the film surfaces of the BA-C8 polymer annealed at 35°C. (B) Proposed structural models of the fold stems in accordance with the fragmentation of the positive ions identified from the negative loadings in (A). Reproduced with permission of John Wiley & Sons, Ltd. (Lau et al. 2011).

Figure 9A shows a loading plot of the film surfaces of the BA-C8 polymer annealed at 75°C. The negative loading pattern is very similar to that shown in Figure 8A. The two loading patterns had similar sets of outliers. However, the relative contributions of some of these masses to principal component 1 were different. In the case of crystallization at the higher temperature, as shown in Figure 9A, the loading of the C₁₄H₁₃O₂⁺ ion (m/z=213) significantly decreased. Some other high-mass ions, such as C₁₇H₂₇O⁺ (m/z=247), which appear in the loading plot in Figure 8A, are supplanted in Figure 9A. Thus, the fold stem of the BA-C8 flat-on lamellae developed at the higher crystallization temperature was composed mainly of partial fragments of a rigid segment and a flexible segment with an ether linkage, as shown in Figure 9B. By combining the loading results of Figures 8A and 9A, Lau et al. (2011) concluded that the length of the fold stems decreased with the crystallization temperature. The folding surface produced at high temperatures contained shorter stems, as confirmed by the loading analyses using PCA. This is because at high temperatures, polymer chains can overcome the diffusion barrier arising from the entanglements to reach and attach themselves to the crystal growth front. The possibility of a remnant stem being trapped in the amorphous matrix is reduced during chain folding at high temperatures.

Figure 9

(A) Loading plot of the film surfaces of the BA-C8 polymer annealed at 75°C. (B) Proposed structural models of the fold stems in accordance with the fragmentation of the positive ions identified from the negative loadings in (A). Reproduced by permission of John Wiley & Sons, Ltd. (Lau et al. 2011).

Chain folding of a monodisperse oligomer

The folding conformations of an oligomer may be different from its polymer. In the previous sections, the fold conformation of the BA-C8 polymer was investigated in detail. ToF-SIMS was used to detect the changes of the fold conformations of a monodisperse (bisphenol-A-etheroctane) oligomer with three repeat units (BA-C8-3) that has a T_g of -15°C (Lau et al. 2009). At temperatures above 35°C, which is more than 50°C above its T_g, the lamellae were in the flat-on orientation. The chain folding of this oligomer was monitored as a function of crystallization temperature. Small-angle X-ray scattering (SAXS) spectra of this oligomer obtained at different temperatures, as shown in Figure 10, show the presence of four peaks. From the locations of these four peaks, we proposed four different structures. Structures 1, 2, 3, and 4 are associated with peaks 1, 2, 3 and 4, respectively. Fully extended lamellar (structure 1), ciliated-extended (structure 2), once-folded (structure 3), and ciliated-folded (structure 4) structures are shown in Figure 10. Based on the SAXS results, a schematic showing the structural transformation of the oligomer as a function of temperature is shown in Figure 11.

Figure 10

Scattering profiles of bulk crystallized films of BA-C8-3 samples obtained at different temperatures (A–E). The inserts show the proposed structures with measured lamellar thicknesses and a computer-generated model of a fully extended structure of a freely bound BA-C8-3 chain at its lowest energy state. The end-to-end distance of the model was determined to be 68.838 Å (6.88 nm). Reproduced with permission from Lau et al. (2009). Copyright 2009 American Chemical Society.

Figure 11

Schematic showing the structural transformation as a function of crystallization temperature (T_c). Reproduced with permission from Lau et al. (2009). Copyright 2009 American Chemical Society.

However, it was possible that peaks 3 and 4 were the second-order scattering of peaks 1 and 2 because the long periods of peaks 3 and 4 were approximately half of those of peaks 1 and 2. ToF-SIMS was therefore used to clarify this issue. ToF-SIMS was used to investigate the fold surfaces of the lamellae of BA-C8-3, with the objective of producing chemical evidence that could be used to support the hypothesis that peaks 3 and 4 were not the second-order scattering of peaks 1 and 2, respectively.

In our previous sections, the surfaces of the edge-on and flat-on lamellae of BA-C8 were differentiated by ToF-SIMS using ions that are characteristic of different polymer segments (Lau et al. 2008). The intensity ratios between any two characteristic ions from the spectra are based on the following equation: R_j/k=I_j/I_k, where I is the absolute intensity of an ion fragment at m/z=j or k. Three characteristic molecular ions at m/z=111, 135, and 325, which were chosen from the positive spectra, were characteristic of the octane segments, bisphenol-A segments, and cilium segments, respectively. Figure 12A presents a typical AFM image of a 250-nm-thick BA-C8-3 film crystallized at 60°C. The AFM results show that all samples prepared for the ToF-SIMS analyses had flat-on lamellae at the surface. The molecular structures of the selected ions used to calculate R_j/k are shown in Figure 12B and C. In Figure 12B, R_j/k represents the ratio of the concentration of the octane segments to that of the bisphenol-A segments, and in Figure 12C, R_j/k represents the ratio of concentration of the cilium segments to that of the bisphenol-A segments. At crystallization temperatures below 60°C, the R_j/k values decreased as the crystallization temperature increased, indicating that the concentrations of the cilium and octane segment on the crystal surfaces decreased due to the supplanting of the ciliated-folded structure by the once-folded structure and that the concentration of the once-folded structure increased with crystallization temperature. For crystallization temperatures above 60°C, the cilium concentration increased due to the supplanting of the once-folded structure by the ciliated-extended structure. Similar results were observed from samples of two different film thicknesses, indicating the absence of film thickness effects. If structures 3 and 4 did not exist, following the hypothesis stating that peaks 3 and 4 in the SAXS curves were the second-order scattering of peaks 1 and 2, respectively, then we should not expect that the surface chemical composition of the crystals be changed as a function of crystallization temperature. However, the ToF-SIMS results provided strong chemical evidence that the chemical nature of the crystal surfaces did change as a function of temperature. The agreement among SAXS and ToF-SIMS results clearly infers that lamellar folding changes occurred as the crystallization temperature was increased, following the sequence displayed in Figure 11.

Figure 12

(A) AFM height image of the surface morphology of a BA-C8-3 film (with an initial thickness of 250 nm) crystallized at a temperature of 60°C. (B) Plot of R_j/k (j=111 and k=135) versus the crystallization temperature (T_c) of films with thicknesses of 150 nm (□) and 250 nm (△). (C) The plot of R_j/k (j=325 and k=135) versus the crystallization temperature (T_c) of films with thicknesses of 150 nm (□) and 250 nm (△). Reproduced with permission from Lau et al. (2009). Copyright 2009 American Chemical Society.

Distribution of an amorphous component of a blend on the surface of the crystalline component

Blends of poly(ε-caprolactone) (PCL) and poly(vinyl chloride) (PVC) form banded spherulites as PCL crystallizes (Owen and Wendt 1969, Thomas and O’Malley 1979). The surface energies of PCL and PVC are 42.9 and 44.0 J/m², respectively (Owen and Wendt 1969). Based on their surface energies, PCL should be the major component on the surface of blends of PCL/PVC. Indeed, for the blends containing 50–75 wt% PCL, the PCL concentration was found to be much higher at the surface than at the bulk due to the fact that the surface energy of PCL is lower than that of PVC. However, surface enrichment of PVC was observed for the blend containing 90 wt% PCL. Based on the surface energy argument, it is not possible to explain the enrichment of PVC at the surface of the blend containing 90 wt% PCL. To unravel this mystery, Cheung et al. (2005) studied the mechanism of the surface segregation of PVC in the blend of PCL (90 wt%) and PVC (10 wt%) using AFM, XPS, and ToF-SIMS. The XPS result obtained on the film of this blend annealed at 45°C for 1 week showed that the surface contained 17 wt% PCL. Figure 13 shows a secondary electron micrograph of the blend obtained using a ToF-SIMS instrument equipped with a secondary electron detector (SED).

Figure 13

SED image of the blend of PCL and PVC grown at 45°C. Reproduced with permission from Cheung et al. (2005). Copyright 2005 American Chemical Society.

A few spherulites with diameters of approximately 500 μm were observed. The surface was corrugated. A small area near the center part of Figure 13 was chosen for ion mapping. O^- and Cl^- ions were chosen to represent PCL and PVC, respectively. Figure 14 is an ion map showing the distribution of O and Cl. The red and green areas represent the regions containing higher concentrations of O and Cl, respectively. Red and green concentric rings are clearly visible. Such periodic structures might be related to the surface morphology of the banded spherulites.

Figure 14

Ion map of the blend of PCL and PVC grown at 45°C. Reproduced with permission from Cheung et al. (2005). Copyright 2005 American Chemical Society.

Ridges and valleys have been shown to represent the edge-on and flat-on lamellae, respectively, in banded spherulites. Figure 15A is a height image showing a surface containing concentric ridges and valleys. Selected areas showing the interface between the ridges and valleys are marked with boxes b, c, and d. Figure 15B shows a higher magnification phase image of the interface between the ridges and valleys. It shows the transition from the edge-on lamellae to the flat-on lamellae when moving from the ridges to the valleys. Figure 15C and D are phase images showing edge-on and flat-on lamellae in the ridges and valleys, respectively. A careful examination of the ToF-SIMS and AFM results revealed that the Cl concentration in the valleys was much higher than that in the ridges (the surface edge-on lamellae), implying that the PVC concentration in the valleys (the surface of the flat-on lamellae) was higher. The surface energy of PCL is lower than that of PVC. The lower surface energy component of the blends is expected to segregate to the surface. However, it is not clear what type of surface the reported surface energy of PCL is for. The surface of the PCL could have been flat-on, edge-on, or a mixture of both. The surface energy of this PCL film with only edge-on lamellae was measured to be 41.9 J m^-2 (Cheung et al. 2005). It is obvious that the surface energy of the edge-on lamellae is lower than that of PVC, implying the absence of PVC on the edge-on lamellae (the ridges). However, it is well known that the surface energy of flat-on lamellae is 3–6 times higher than that of edge-on lamellae because of the presence of the folding surface (Muthukumar 2004). It would be logical to assume that PVC was the component on the surface of the flat-on lamellae (valleys). As a result of the segregation of the PVC to the surface of the flat-on lamellae, the PCL concentration as measured by XPS was reduced to a level that was lower than that of the bulk. The distribution of Cl on the surface of the PCL/PVC blend as concentric rings was caused by the change in the surface morphology as the PCL lamellae twisted during their radial growth. The PVC segregated on the surface of the valleys, which consisted of PCL flat-on lamellae. On the surface of the ridges, which consisted mainly of edge-on lamellae, PCL was the dominant species.

Figure 15

AFM images of the blend of PVC and PCL grown at 45°C. (A) Height image showing several bands of a spherulite. The areas marked with boxes b, c, and d are enlarged in Figure 15B–D, respectively. (B) Phase image showing the interface between a ridge and a valley at which edge-on and flat-on lamellae are interwoven together. (C, D) Phase images showing the morphologies of the ridges and valleys containing the edge-on and flat-on lamellae, respectively. Reproduced with permission from Cheung et al. (2005). Copyright 2005 American Chemical Society.

Thin-film stability

Ren et al. (2013) applied ToF-SIMS depth profiling to determine why spin-coated BA-C10 thin films were stable. Several solvents, including 1,2-dichlorobenzene (ODCB), chloroform (CHCl₃), tetradrofuran (THF), and 1,4-dioxane and dimethylformamide (DMF), were used. Silicon wafers, highly ordered pyrolytic graphite (HOPG), and mica were used as the substrates onto which the polymers were deposited by spin coating. The stability of these films was observed by OM.

Figure 16 presents a series of optical micrographs of BA-C10 films prepared with different solvents on silicon wafers. Both the optical micrographs (Figure 16A and B) and the surface profile (Figure 16C) show a smooth film surface, suggesting wetting of the polymer when ODCB was used as the solvent. Cellular droplets, ring-like droplets, and network structures were seen at the spinning center of the films when CHCl₃, THF, and 1,4-dioxane were respectively used. These observations strongly suggest a gradual increase in the dewetting instability of the polymer films, as shown in Figure 16D–O, indicating partial and complete dewetting of the polymer films.

Figure 16

Optical micrographs and surface profiles obtained with an optical profiler showing the surface morphologies of spin-coated BA-C10 films with (A–C) ODCB, (D–F) CHCl₃, (G–I) THF, (J–L) 1,4-dioxane, and (M–O) DMF as the solvents for the precursor solutions. The films were prepared by spin-coating silicon wafers with 30 mg ml^-1 polymer solutions. The scale bars represent 50 μm. Reproduced with permission of John Wiley & Sons, Ltd. (Ren et al. 2013).

Figure 17A and B are, respectively, the optical micrographs at the center and periphery of a spin-coated BA-C10 film on mica. Figure 17D and E are same as Figure 17A and B, except that HOPG was the substrate. Figure 17C and F present the surface profiles of the films as a result of the wetting and dewetting of BA-C10 when mica and HOPG were used as the substrates, respectively.

Figure 17

Optical micrographs and surface profiles obtained with an alpha step profiler showing the surface morphologies of spin-coated BA-C10 films with (A–C) mica and (D–F) HOPG as the substrates. The films were prepared by spin-coating the substrates with 30 mg ml^-1 polymer/CHCl₃ solutions. The scale bars represent 50 μm. Reproduced with permission of John Wiley & Sons, Ltd. (Ren et al. 2013).

Polymer end groups, which may influence the stability of polymer films, tend to segregate to the interface to lower the interfacial free energy (Yerushalmi-Rozen et al. 1994, Yerushalmi-Rozen and Klein 1995). ToF-SIMS depth profiling was conducted to determine the relationship between the stability of the polymer films and the concentration of the bromine end groups at the interface between the polymer film and the substrate. ToF-SIMS depth profiling was performed with a ToF-SIMS V instrument (Ion-ToF GmbH, Münster, Germany) equipped with a C₆₀⁺ ion source for abrasion and a bismuth liquid-metal ion source for analysis. A focused 10 keV C₆₀⁺ beam at an angle of incidence of 45°, rastered over an approximate area of 800 μm×800 μm at a current of 0.4–0.6 nA, was used to sputter through the polymer films. This process was monitored in real time by recording a depth profile over an analysis area of 150 μm×150 μm at the center of the sputtered region. The analysis was carried out with a 25 keV Bi₃⁺ beam operating at a pulsed current of 0.3 pA. Figure 18A shows a negative ToF-SIMS spectrum of a BA-C10 film prepared with CHCl₃ as the solvent and mica as the substrate. Figure 18B–F are the ToF-SIMS depth profiles of these films. The negative ion ⁸¹Br^- was used to represent the polymer end groups. The interface is usually defined as the midpoint of the intensity gradient of a representative substrate ion (Shard et al. 2007, Green et al. 2009). The intensity profiles of the Al₂O₃^-, SiO₂^-, and C₂^- ions were used to define the interface when mica, silicon wafer, and HOPG, respectively, were used as the substrates. An equation was developed to calculate the average intensity of an ion along the depth of a sample:

Figure 18

(A) ToF-SIMS negative ion spectrum of a BA-C10 film prepared with CHCl₃ as the solvent and mica as the substrate. (B–F) ToF-SIMS depth profiles of end-group ion Br^- and the substrate ions of spin-coated BA-C10 films. The pairs of solvent and substrate used in the spin coating process were (B) CHCl₃ and mica, (C) CHCl₃ and silicon wafer, (D) CHCl₃ and HOPG, (E) ODCB and silicon wafer, and (F) THF and silicon wafer. Reproduced with permission of John Wiley & Sons, Ltd. (Ren et al. 2013).

where I¯i is the average intensity of ion i along the depth, A_i is the peak area of ion i across the whole film, d is the entire sputter depth, I_i is the intensity of ion i at a particular depth, and Δd is the sputtering step size. In each of Figure 18B–F, a green level line was drawn to represent the average intensity of Br^- along the entire depth. The distribution of the bromine end groups along the depth of a sample can be readily identified by comparing the depth profile of Br^- with its average intensity line.

In order to relate the stability of the polymer films to the concentration of the bromine end groups at the interface, ToF-SIMS depth profiles of the films as shown in Figure 18B–D were examined. The concentration of the bromine end group at the interface of the samples prepared with CHCl₃ as the solvent and either silicon wafer, HOPG, or mica as the substrate was calculated using the following equation:

where χ_Brs is the relative concentration of the bromine end groups at the interface to that of the whole film, ABrI is the peak area of Br^- at the interface, and d^I is the interface width. Table 2 shows high values of χ_Br in the completely wetted BA-C10 films on silicon wafer with ODCB as the solvent and on mica with CHCl₃ as the solvent. As the value of χ_Br decreased, partial wetting and complete dewetting of the film were, respectively, observed on the silicon wafer and HOPG. These results clearly show that the stability of spin-coated BA-C10 films is related to the concentration of the end groups at the interface

Thin-film structures

Ren et al. (2012) applied ToF-SIMS three-dimensional imaging to determine the interior structure of spin-coated BA-C10 thin films in detail. Similar to the depth profiles above, 10 keV C₆₀⁺ with a raster area of 800×800 μm² was used for sputtering, while 25 keV Bi₃⁺ beam was used for ion imaging. With CHCl₃ as the solvent, the BA-C10 films on a silicon surface exhibited cellular patterns and radial striations at the center and the periphery of the spinning, respectively, as shown in Figure 19A (optical micrograph) and B (optical profilometry image). The two-dimensional surface height profile shows that the height of the droplets with respect to the valley (peak-to-valley distance) was about 100 nm on average.

Figure 19

Spin-coated BA-C10 films with CHCl₃ as the solvent. (A) Optical micrograph; (B) optical profilometry image; (C) ToF-SIMS total ion image of all sputter cycles stacked together; (D) ToF-SIMS images of the polymer and substrate ions at different sputter depths; and (E) cross-section views of the polymer ion (the thicknesses of polymer layers defined in Figure 20B are indicated). Reproduced with permission from Ren et al. (2012). Copyright 2012 American Chemical Society.

Positive ion ToF-SIMS depth profiling was carried out on an area where the droplets were present. An imaging area containing at least two to three droplets (100 μm×100 μm) was analyzed. Figure 19C shows the total ion image of all sputter cycles. The total ion intensity in this area containing the droplets was found to be lower than those in other areas. ToF-SIMS images of ions representing the polymer and the substrate at different sputter depths are shown in Figure 19D. C₉H₁₁O⁺ and SiOH⁺ were identified as the characteristic ions of BA-C10 and the substrate, respectively. Each image in Figure 19D is an integration of ten consecutive cycles, which are approximately 5 nm thick, so that they are bright enough to see. From the surface (0 nm) to the middle of the film (135 nm), the polymer ion C₉H₁₁O⁺ was homogenously distributed and no substrate signal was observed. Near the interface at a depth of 275 nm, the substrate ion appeared first at the centers of the droplets where the intensity of the polymer ion was lower. These results suggest that the effective thickness of the film where the droplets were located was smaller than that of the surrounding area. Based on the results of optical profilometry, shown in Figure 19B, which indicate that the droplets were 100 nm taller than the surrounding valleys, the only explanation to these unexpected results is that the droplets were hollow. In fact, if the droplets were solid, the substrate ion should appear first at the valleys. The vertical distributions (XZ and YZ cross-section views) of the polymer ion C₉H₁₁O⁺ along the lines shown in the XY view (all scans) are shown in Figure 19E. The apparent thickness of the polymer in the area surrounding the droplets was slightly larger than that at the droplets. This also explains why the total ion intensity (Figure 19C) and the polymer ion intensity (Figure 19E, the XY view) were found to be lower at the droplets. The cross-section results also confirm that the droplets were hollow.

Although the above results clearly show that the droplets were hollow, there were still two possibilities for the structure of the bubbles: (1) the bubbles sat on the substrate (Figure 20A) and (2) the bubble was sandwiched between two polymer layers (Figure 20B). To determine the exact structure of these bubbles, depth profiles were retrospectively reconstructed from the droplets and the valley areas (indicated by a red rectangle and a blue square, respectively, in Figure 19C) and the results are displayed in Figure 21. In the area surrounding the droplet (Figure 21A), the polymer ion (C₉H₁₁O⁺) intensity, after a dramatic decrease due to the surface transient effect (Lau et al. 2008), remains constant until the substrate is reached. The substrate ion, on the contrary, is absent until the polymer ion intensity starts to decrease. The decrease in the polymer ion intensity and the increase in the substrate ion intensity occur simultaneously, as is normally observed in the depth profiling of polymer thin films (Shard et al. 2007). From the profile, the film thickness at the valleys was estimated to be about 290 nm.

Figure 20

Schematic illustration of two possible locations of a bubble inside a thin film: (A) on the substrate and (B) between the top and bottom polymer layers. Reproduced with permission from Ren et al. (2012). Copyright 2012 American Chemical Society.

Figure 21

Depth profiles reconstructed retrospectively from the areas indicated in Figure 19C. (A) The valley area (the blue square in Figure 19C) and (B) the droplet area (the red rectangle in Figure 19C). The black dashed line indicates the interface point; the red dot-dashed line indicates the point at which the polymer ion intensity starts to decrease; the blue dot-dashed line indicates the point at which the substrate ion intensity starts to increase; and the purple dot-dashed line indicates the overlapping of the red- and blue-dotted lines. Reproduced with permission from Ren et al. (2012). Copyright 2012 American Chemical Society.

However, the depth profile of the polymer ion in the area where the droplet sits (Figure 21B) is clearly different from that of the polymer ion in the area surrounding the droplet (Figure 21A). The depth profile can be divided into two regions. In the first region, the distribution is very similar to that of Figure 21A. After a surface transient region, the polymer ion intensity remains constant until the depth of 90 nm is reached. In the second region (after 90 nm), the polymer ion intensity drops suddenly and then gradually decreases until the substrate is reached. This special polymer ion intensity distribution is a strong indication that the bubble is positioned between two polymer layers (cf., Figure 20B). Indeed, if the bubble had a structure shown in Figure 20A, there would be only one polymer layer on top of the bubble. As a consequence, a continuous depth profile of the polymer ion similar to that of the valley area (Figure 21A) would be expected. But the two-region depth profile shown in Figure 21B is different from that shown in Figure 21A. This unusual depth profile can be easily explained with the structure shown in Figure 20B. The first part of the depth profile corresponds to the top polymer layer of the bubble. Upon removal of this layer, due to a sudden change in the topology of the droplets from convex to concave (from point b to b′ in Figure 20B), the polymer ion intensity drops sharply. The second part of the depth profile corresponds to the bottom polymer layer of the bubble and its intensity is suppressed probably due to the shadow effect.

Morphology of polymer blends

ToF-SIMS has been used to study the morphology of polymer blends (Weng et al. 1998, Weng and Chan 2006, Miyasaka et al. 2008). Weng et al. (1998) studied the morphology of blends of ethylene-tetrafluoroethylene copolymer (ETFE)/poly(methyl methacrylate) (PMMA) using differential scanning calorimetry (DSC) and ToF-SIMS. Three T_g values, which were detected in these blends by DSC, indicated that there were three regions in these blends: a semicrystalline region containing mostly ETFE with a T_g close to that of pure ETFE, an amorphous region containing mostly PMMA with a T_g close to that of pure PMMA, and an amorphous region containing significant levels of ETFE and PMMA.

ToF-SIMS imaging was used to identify these three regions of the blends. To obtain a flat surface for ToF-SIMS analysis, the samples were prepared with a cryomicrotome at -100°C. O^-+OH^- and F^- were chosen to represent PMMA and ETFE, respectively.

The O^-+OH^- and F^- images for the blends (weight ratio of ETFE/PMMA=20/80) are presented in Figure 22. The upper and lower images are for the sample prepared by slow cooling and quenching, respectively. The O^-+OH^- and F^- images are complementary. For the slowly cooled sample, the ETFE particles (phase B) were dispersed in the PMMA matrix (phase A), as shown in the upper images of Figure 22. In most cases, the particles were well defined, with sizes ranging from about 1 to 20 μm. The spectra, which were reconstructed using data obtained from phases A and B, show features very similar to those of pure ETFE and PMMA. However, the intensities of the O^-+OH^- and F^- peaks were not zero in phases B and A, respectively, implying that there were small amounts of PMMA and ETFE in phases B and A, respectively. In addition, a small amount of a phase (phase C) that contained high levels of ETFE and PMMA could also be detected. Both PMMA and ETFE fragments were present in the spectrum reconstructed from phase C. These results are in agreement with the DSC measurements. For the quenched sample, the surface morphology was quite different, as shown in the lower images of Figure 22. Although the dispersion of the ETFE particles in the PMMA matrix was still observed, a much larger amount of phase C was detected. A line scan at the location marked with a blue line in the F^- image (the lower-right image of Figure 22) is shown in Figure 23. It is clear from the line scan that the F signal was not zero in phase A or very high in phase B. A small amount of phase C was also detected. In summary, the work of Weng et al. (1998) has demonstrated that ToF-SIMS is a powerful tool for studying the morphology of polymer blends.

Figure 22

Negative ToF-SIMS images obtained for two samples with a blend composition of ETFE/PMMA=20/80. The upper and lower images are for the slowly cooled and quenched samples, respectively. Imaged area is 200 μm×200 μm. Reproduced with permission from Weng et al. (1998). Copyright 1998 American Chemical Society.

Figure 23

Line scans for F^- (represented by the blue lines in the F^- images, as shown in Figure 22). Reproduced with permission from Weng et al. (1998). Copyright 1998 American Chemical Society.