Tracking the Evolution of Polymer Interface Films ... - ACS Publications

Sep 8, 2016 - calculated model of big “U” conformation is provided in Figure. 1g. ..... Foundation of China (grants 21303200) and Youth Innovation...
0 downloads 0 Views 5MB Size
Article pubs.acs.org/Langmuir

Tracking the Evolution of Polymer Interface Films during the Process of Thermal Annealing at the Domain and Single Molecular Levels using Scanning Tunneling Microscopy Xiao-Ling Duan,† Hua-Jie Chen,‡ Jian-Yao Huang,‡ Zhi-Fei Liu,† Jin-Kuo Li,† Zhi-Yong Yang,*,† Wei-Feng Zhang,*,‡ and Gui Yu*,‡ †

School of Chemistry and Chemical Engineering, University of Chinese Academy of Sciences, 19A Yuquanlu, Beijing 100049, P. R. China ‡ Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, P. R. China S Supporting Information *

ABSTRACT: Structural evolution of polymer (NTZ12) interface films during the process of annealing is revealed at the domain and single molecular levels using the statistical data measured from scanning tunneling microscopy images and through theoretical calculations. First, common features of the interface films are examined. Then, mean values of surface-occupied ratio, size and density of the domain are used to reveal the intrinsic derivation of the respective stages. Formation of new domains is triggered at 70 °C, but domain ripening is not activated. At 110 °C, the speed of formation of new domains is almost balanced by the consumption due to the ripening process. However, formation of new domains is reduced heavily at 150 °C but restarted at 190 °C. At the single molecular level, the ratio of the average length of linear to curved backbones is increased during annealing, whereas the ratios of the total length and the total number of linear to curved skeletons reaches a peak value at 150 °C. The two major conformations of curved backbones for all samples are 120° and 180° bending, but the ripening at 150 °C reduces 180° folding dramatically. Molecular dynamic simulations disclose the fast relaxing process of curved skeletons at high temperature.



INTRODUCTION

The oligomers and polymers containing thiophene and its derivative units exhibit promising applications in polymer electronics. Therefore, their interface structures on different kinds of substrates are investigated extensively using scanning tunneling microscopy (STM) whose ability of obtaining submolecular-resolution images on the surface has been well proved.13−19 The oligothiophenes bridged with trimethylene form ordered patterns on a graphite surface but form disordered structures on a gold surface, and the systematic results indicate that molecules with either long oligothiophene backbone or long bridge part are more flexible and have more versatile conformations on the surface.20,21 STM studies show that most of the polythiophene analogues assemble into shortrange ordered structures on a graphite surface, and the folding of “U” shape is observed frequently.22−27 In contrast to the closely packed first layer, the second layer is usually constructed with isolated polymers curved at random angles. However, it is reported that when bromine is substituted at the end of side alkyl chains, the molecules tend to form a well-organized second layer.28 Furthermore, microphase separation of block copolythiophenes on the graphite surface is also disclosed by STM investigations.29 Besides film structures, electronic

Polymer electronic devices, including solar cells, field-effect transistors, and light-emitting diodes, are paid increasing attention because of their distinguishable featuresversatile raw materials, low cost, low weight, good flexibility, etc.1−3 With great efforts devoted in recent years, remarkable progress has been made in these areas. From the results, one can recognize that the device performance relies not only on the intrinsic properties of conjugated polymers but also on the structure of polymer films, especially the structure of the several molecular layers near the interface of the electrode or other functional components.4−7 Therefore, various methods have been developed to optimize the whole film and interface structures to advance the polymer electronics research.2,4,8−10 Polymer films need to be precisely characterized to extract the structural information in detail. X-ray diffraction techniques can provide a macroscopic arrangement of the polymer in the whole layer, whereas atomic force microscopy (AFM) and scanning electron microscopy disclose the film topography clearly.11,12 But characterization is quite difficult for the interface layers near electrodes because currently we have very few tools that can meet the critical demand of imaging the interface structure on the surface of device electrodes. Alternatively, interface structures are investigated on different substrates. © XXXX American Chemical Society

Received: June 7, 2016 Revised: August 3, 2016

A

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

mechanisms and procedures of the interface film are still valuable references for understanding the influence of thermal annealing on the structures of polymer films and on the performance of polymer devices.

properties of a single polythiophene chain are inspected using scanning tunneling spectroscopy (STS).30 The obtained results imply that charge transfer from the graphite surface to the bottom molecules is negligible, so the energy gap shown in the STS spectra reflects the intrinsic properties of polymers. However, the gap range in the STS curves collected at the double-layer areas displays an unexpected widening for reasons that are not fully understood.30 Graphite is an inert surface, usually interacting with adsorbing substances very weakly, whereas for metal substrates such as gold, the molecule− substrate interaction is much stronger. For this reason, the oligomers and polymers of thiophene units are totally disordered on a bare gold surface, and the gap measured here by STS is higher than that of bulk materials.31 Instead of depositing presynthesized polymers on the substrate, block copolymers containing two kinds of thiophene units are fabricated on an iodine-modified gold surface, and the block of respective monomers is distinguished clearly by its unique contrast in STM images. 32 The derivatives of poly(phenylenevinylene), poly(phenyleneethynylene), and polyfluorene are also studied using STM, and similar backbone foldings are observed as well.33−35 In a recent report, one kind of heterogeneous bilayer structure was constructed successfully by t h e d er i va t i ve o f p h t h a lo cy a n in e a n d p o ly (phenyleneethynylene). The top layer of phthalocyanine is found to be affected greatly by the structure of the bottom poly(phenyleneethynylene) layer.36 The interface structures of the assorted conjugated polymers have been studied using STM/STS, and the obtained results provide detailed information on the aspects of backbone conformations, arrangements in the first and second layers, electronic properties of polymer chains, and so on. However, few reports have attempted to learn the evolution of domain structures and molecular conformations during the process of thermal annealing, which has been proved in numerous research studies of polymer devices as an ultraeffective and the most widely used treatment to improve device performance.3,5,11 In the present study, the polythiophene derivative (NTZ12, Scheme 1) is deposited on the surface of highly



EXPERIMENTS AND METHODS

NTZ12 was synthesized as previously described.37,38 Then, it was dissolved in chlorobenzene (Sigma-Aldrich Co.) with the concentration around 0.1 mg mL−1. A drop of the solution was deposited on the freshly cleaved HOPG (SPI Co.) surface to prepare the samples for STM experiments. After the solvent vaporized completely, the sample was observed using STM at room temperature or annealed at different temperatures (70, 110, 150, or 190 °C) for 30 min under an Ar environment and cooled down to room temperature naturally before STM characterization. The annealing temperature in device researches is usually around 200 °C. Our highest temperature was set at 190 °C, a little lower than 200 °C, to avoid possible desorption of short-chain components on the inner graphite substrate. To learn the evolution of the interface film, three intermediate temperature points were added at equal intervals between room temperature and 190 °C. All STM experiments were performed on a Picoscan 2100 (Agilent Co.) scanning tunneling microscope with a mechanically cut Pt/Ir (90%/10%, Goodfellow Co.) tip. A bias was applied to the tip. The STM image was exported to the ImageJ software to determine the area of each domain, surface-occupied ratio, length of each linear and curved backbone, and the folding angle of each curved backbone in the image. Detailed measurement processes and calculating methods of these parameters are illustrated in Figure S1 and Table S1, and the corresponding explanations are provided as Supporting Information. Consequently, a large amount of raw data were obtained by measuring many STM images and further analyzed statistically using the Origin software. For each temperature sample, at least 100 STM images obtained from different areas were inspected. Semicontact AFM imaging is conducted on an NTEGRA Prima (NT MDT) microscope. Density functional theory (DFT) calculations were performed using the AM1 semiempirical method embedded in Gaussian 09, Revision E.01.39 To make the computing cost acceptable, the side chains of NTZ12 were reasonably simplified as a methyl or methoxy group and the substrate was not involved either. Therefore, optimization was carried out for the molecules in the gas phase. Molecular dynamic (MD) simulations were carried out using a canonical NVT ensemble (constant number of atoms N, volume V, and temperature T) and a universal force field. The circular conformation of the polymer with five repeat units was first optimized using the DFT method (Figure S2a). To avoid very large computing systems, side alkyl chains were simplified as a methyl or methoxy group. Single-layer graphene was used as the model substrate of graphite (Figure S2b). To stabilize the graphene sheet during simulations, hydrogenation was applied to the back side of the graphene sheet.

Scheme 1. Illustrative Structure of Polythiophene Derivative (NTZ12)



RESULTS AND DISCUSSIONS The results and related discussions are divided into three aspects: common features of the NTZ12 interface film, evolution of NTZ12 domains, and evolution of single molecular conformations. Common Features of the NTZ12 Interface Film. Concentration of the NTZ12 solution is set at the same level as that of device fabrication rather than the very dilute one designed for STM investigation because we attempt to keep our experimental conditions as close as possible to those of device research. Although the solution we used to prepare STM samples is pretty dense, the domains observed in the room temperature samples are unexpectedly small and the surfaceoccupied ratio is very low, as shown by the typical STM image in Figure 1a. On the whole, the structure shows more

oriented pyrolytic graphite (HOPG) and annealed at different temperatures. STM experiments and theoretical calculations are carried out to track the development of the interface structure at the domain level and at the single molecular level. On the basis of the statistical data for the sample at various temperatures, the changing tendency of the respective parameter, including surface-occupied ratio, average domain size, average domain density, ratio of different conformations, and angle distribution of curved backbones, is obtained quantitatively. Then, the transforming mechanism at each temperature stage and the whole evolution pathway of the NTZ12 surface structure are proposed with the assistance of theoretical calculations. Although the substrate is not exactly the same as the electrodes of polymer devices because of the limitation of the STM technique, the disclosed transfer B

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 1. Large-scale (a) and small-scale (b) STM images of the NTZ12 interface structure on HOPG at room temperature; large-scale (c) and small-scale (d−f) STM images of the NTZ12 interface structure on HOPG after annealing at 190 °C; (g) the optimized model of circle (top), big “U” folding (middle), and small “U” folding (bottom). Tunneling conditions: (a) Vbias = 0.938 V, Itip = 1.08 nA; (b) Vbias = 0.938 V, Itip = 1.159 nA; (c) Vbias = 0.986 V, Itip = 1.316 nA; (d) Vbias = 0.892 V, Itip = 1.239 nA; (e) Vbias = 0.917 V, Itip = 0.826 nA; for the inset: Vbias = 0.872 V, Itip = 1.069 nA; and (f) Vbias = 0.922 V, Itip = 0.441 nA.

construct this ring. Additional energy of 9.1 kcal is required to transfer the most stable nonbending backbone to this circle, which is larger than 2.4 kcal for forming the small “U”-shaped structure, because more bonds have to be rotated but can still be activated in our experimental conditions. However, the higher energy barrier makes the ring structure less observed than the small “U”-shaped bending. In bigger-sized circles, it can be expected that the average torsion for each single bond will be reduced. Because of the extremely small domain and the low surfaceoccupied ratio in the room-temperature sample, NTZ12 is easily disturbed by the STM tip, thus making it very difficult to acquire a high-resolution image. The STM image of the sample after being annealed at 190 °C (Figure 1c) demonstrates that both the surface-occupied ratio and the ordering of the whole interface structure are improved greatly. Meanwhile, molecular domains grow bigger in size, and most of the defects in domains are polished during the high-temperature treatment, which leads to the regularity of the interface film enhanced in the long range. NTZ12 in large-ordered domains is stable enough to undergo repeated scanning to get high-resolution images. Figure 1d shows such an image exhibiting wavelike backbones and the perpendicular orientation of the interdigitated alkyl chains. Although the treatment at 190 °C elevates the content of the linear backbone considerably, folding backbones can still be found in this sample and they even display more complicated bending styles. One leg of hairpin folds roughly 125°, as denoted by the dashed line in Figure 1c. The inset in Figure 1c is the zoomed-in picture of the area enclosed within the black square. In this zoomed-in picture, it can be seen that a bigger hairpin skeleton embraces an elongated ring with two small hairpin caps. The distance between the two legs of the bigger “U” folding is determined to be approximately 8.0 nm. The calculated model of big “U” conformation is provided in Figure 1g. In this model, the distance between the two ends is 8.3 nm, roughly the same as that of the experimental one, with 7 repeat

characteristics of disordering, although not in a completely chaotic state. The backbone of NTZ12 is clearly imaged, many of which are in folding conformations. In the medium-scale image shown in Figure 1b, assorted folding styles can be seen. The hairpin, or “U”-shaped, bending is most generally observed in our research, similar to the previous reports on the polythiophene derivatives.19 The distance between two parallel legs of most “U” configurations is roughly equal to the length of dodecyl alkyl chains (around 2.3 ± 0.2 nm), indicating that the side chains of neighbor backbones adsorb onto the surface in an interdigitated style. The bending part of this “U” shape is around 3−4 nm, about 2 repeat units. Using the DFT calculations, a model for this kind of “U” folding is proposed as in Figure 1g (bottom, embedded small one). Its energy is only 2.4 kcal higher than that of the most stable nonfolding structure. This energy difference is close to that of the reported data, obtained by rotating adjacent thiophene units.19,22 In this model, the two ends of the backbone are around 3.3 nm, which is bigger than the experimental value. Compressing this model to the size of 2.3 nm will produce more distortion energy. Under experimental conditions, the distortion energy could be neutralized by the close packing of side alkyl chains and the strong interaction between the side alkyl chains and the graphite substrate. However, these two kinds of stabilizing factors do not exist in the modeling process because neither the substrate nor the side chains are involved in DFT calculations to make the computing cost acceptable. Thus, the model relaxes into one of bigger size bearing lower energy. More complicated conformations, including roughly 90° and 120° angles and deformed circles, are also found. In Figure 1b, several black lines along the direction of backbones are used to mark out the nearby folding style. As for deformed circles, very few have been reported and are particularly interesting. The diameter of the smallest circle observed in our experiments is about 4 nm. The optimized model proposed at the top of Figure 1g demonstrates that 6 repeat units are needed to C

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 2. (a−e) Large-scale STM images of the NTZ12 interface structure on HOPG at room temperature and after annealing at 70, 110, 150, and 190 °C, correspondingly. Tunneling conditions: (a) Vbias = 0.938 V, Itip = 1.08 nA; (b) Vbias = 0.916 V, Itip = 0.514 nA; (c) Vbias = 0.837 V, Itip = 1.182 nA; (d) Vbias = 0.919 V, Itip = 0.699 nA; and (e) Vbias = 0.986 V, Itip = 1.316 nA.

Table 1. Domain Parameters for Each Temperature Sample average surface-occupied ratio (%) average size of a single domain (nm2) average length of linear NTZ12 (nm) number of molecules per domain domain density (per 100 × 100 nm2)

room temperature

70 °C

110 °C

150 °C

190 °C

29 ± 11 133 ± 86 7.3 ± 2.4 8 22

39 ± 11 137 ± 86 7.7 ± 2.5 8 28

64 ± 16 211 ± 142 8.7 ± 3.0 11 30

66 ± 16 283 ± 217 9.3 ± 3.6 13 23

83 ± 10 354 ± 316 10.2 ± 4.6 15 23

room temperature and at 190 °C. The change in structure is analyzed at two levels: domain level and single molecular level. Evolution of NTZ12 Domains. To have an overview of the trends of structure transformation, we list the typical STM images of the room temperature sample, taken after they were annealed at 70, 110, 150, and 190 °C (Figure 2a−e) although the images of the room temperature sample and the 190 °C sample have already been given in Figure 1. Overall, it can be found that the surface-occupied ratio and the size of single domain increase with the rising in temperature, and the structure of the whole interface film transforms into a more ordered arrangement. The topography features revealed by the AFM images in Figure S3 indicate the improvement in crystallinity as well. To explain the process of domain change quantitatively rather than only qualitatively, four parameters are used to characterize the features of NTZ12 domains. They are (1) average surface-occupied ratio (%), which reveals how much of the whole surface is in direct contact with the molecules and is a valuable parameter for device research because it is closely linked to the charge transfer in a molecule−metal interface; (2) average size of a single domain (nm2), which is an indicative factor of the ripening process and plays an important role in improving carrier mobility; (3) domain density (per 100 × 100 nm2), which discloses the net result of domain consumption and formation; and (4) the number of molecules per domain, which verifies the ripening process from another side. Mean values of the surface-occupied ratio and the size of the single domain for each temperature sample are determined by

units involved. The energy barrier of this big “U” folding is determined to be 4 kcal, which is rather low. Besides the fruitful single-bending styles detected in the room temperature samples, various hybrid bendings formed by connecting several folding fashions together are also observed, with some of them traced out by dashed lines (I−IV) in Figure 1e: a deformed ring combined with a hairpin (I in the inset in Figure 1e) or a semicircle (II), a ring with an inner separator, small part of backbone circled by the other part of molecular chain (III), and zigzag folding (IV). In fact, the bending of NTZ12 is far more diversified than those mentioned above, and there are too many to list completely. Only some of the typical styles are given in Figure 1 to show the distinguishable flexibility of NTZ12. In addition to the enriched foldings, merging of two conjugated backbones is observed occasionally, appearing as a piece of broader stripe (pointed out by the white arrow in Figure 1f). After being treated at 190 °C, the interface film of NTZ12 shows distinctive changes in comparison with the room temperature samples. We attempt to track the structure derivation in this annealing process and explain the changing tendency of the respective parameter quantitatively, which will help us understand the influence of thermal treatment on the interface structure and may provide some clues to infer the reason why thermal treatment can improve the performance of polymer devices. Therefore, we investigated further the corresponding samples annealed at 70, 110, and 150 °C using STM and considered these results together with those at D

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir Scheme 2. Illustrative Transformation Process of NTZ12 Domains

Table 2. Parameters of Single Molecular Conformations at Each Temperature average average ratio of ratio of ratio of

length of linear NTZ12 (nm) length of curved NTZ12 (nm) average length of linear to curved NTZ12 total number of linear to curved NTZ12 total length of linear to curved NTZ12

room temperature

70 °C

110 °C

150 °C

190 °C

7.3 ± 2.4 12.0 ± 4.1 0.61 1.0 0.6

7.7 ± 2.5 12.5 ± 4.4 0.62 1.3 0.8

8.7 ± 3.0 13.6 ± 5.0 0.64 2.6 1.6

9.3 ± 3.6 13.4 ± 4.9 0.69 5.8 4.0

10.2 ± 4.6 14.5 ± 6.8 0.71 2.9 2.0

number of molecules per domain being equal to that at room temperature. In one word, the major change in the NTZ12 interface film during the treatment at 70 °C is the increase in the surface-occupied ratio by the formation of new domains through interlayer molecular transfer, as shown in the first step of Scheme 2. Surface ripening is not activated at this temperature. After heating the sample at 110 °C, both the surfaceoccupied ratio and domain area demonstrate an apparent enlargement whereas the domain density almost stays in the original level of 70 oC. The linear part of the NTZ12 backbone is longer with reference to that at 70 °C, suggesting that the molecules are equipped with enough energy at this temperature to adjust their adsorbing conformations. The increase in the number of molecules per domain implies the start of the ripening process, which would diminish the domain density, no matter whether the growing mechanism is Ostwald ripening or dynamic coalescence. However, the mean domain density increases a little rather than decrease, which indicates that the formation of new domains through the interlayer molecular transfer is slightly faster than the consumption of the ripening process. The essential derivation features at this stage are illustrated in the second step of Scheme 2. On increasing the annealing temperature further to 150 °C, the surface-occupied ratio develops very less but the domain area follows the tendency of rapidly growing continually, and hence the domain density drops down dramatically. The tiny increase in the surface-occupied ratio denotes that very less NTZ12 is transported to the interface layer from the above layers; thus, the speed of producing new domains is reduced heavily. Then the decline in domain density is inevitable because some domains grow into bigger ones by consuming others. The disrupted supply of NTZ12 is probably because of the depletion of molecules in the nearby layers. The data in Table 2 show that the growth of domain area relies mainly on the increased number of its molecules because the linear length of NTZ12 at this temperature is extended rather small, which from the other side verifies that the decline in domain density is the result of sacrificing some domains to support the growth of

averaging a large amount of raw data measured from hundreds of STM images (Figure S1, Table S1, and their corresponding explanations in the Supporting Information). Domain density (per 100 × 100 nm2) is calculated as average surface-occupied ratio × (100 × 100 nm2) × (average size of a single domain)−1, meaning the average number of domains per 100 × 100 nm2. Furthermore, we realized that the ordered domain is mainly constructed with NTZ12 in a linear conformation and obtained the statistical length of linear NTZ12 as well. Then, the average number of molecules per domain is computed through the following equation: number of molecules per domain = average size of single domain × (average length of linear NTZ12)−1 × (2.3 nm)−1 (the above-mentioned data of the distance between two neighbor NTZ12 backbones). The statistical data of these four parameters and the statistical length of linear NTZ12 for each temperature sample are listed in Table 1. The average value of the surface-occupied ratio is quite low at room temperature, with only one-third of the graphite surface being covered. Meanwhile, the estimated number of molecules per domain implies that a single domain contains only eight NTZ12, with an average length of 7.3 nm. After annealing at 70 °C, the surface-occupied ratio increases whereas the domain area is almost as large as that of the room temperature sample, which suggests that the increase in the surface-occupied ratio should be attributed to the increase in the domain number. Then, one question arises naturally: Where do these additional new domains of NTZ12 come from? No solution remained on the graphite in our STM experiments, excluding the possibility of supplying molecules from the bulk solution. Therefore, the only reasonable way is transfer from the above layers, most probably from the second and third layers, which are closest to the first layer. On the other hand, the speed of domain consumption by the ripening process, either through Ostwald ripening or dynamic coalescence, is almost zero because the growth of the domain area is negligible. These analyses also reveal that the thermal energy of 70 °C is high enough for the interlayer molecular transfer but not enough for the activation of the ripening process. At 70 °C, average length of the linear NTZ12 and average size of a single domain are almost the same as those at room temperature, also leading to the average E

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

maximum at 150 °C from a small value at room temperature and dropping down again at 190 °C. Discussions on the results of average surface-occupied ratio and domain density have pointed out that at 150 °C, the formation of new domains is trivial and the major annealing result in this stage is the ripening of interface domains. Thus, many of the curved polymers are converted into linear ones, leading to the peak ratio of the total length and number. At 190 °C, the ratios of the total length and number are decreased sharply by the newly formed domains because both the content and the length of linear skeletons in the new domains are lower than in the annealed ones. In short, the dominant movement of NTZ12 backbones during the heating process is the conversion of kinetically controlled folding conformation to a thermally stabilized linear one. At the same time, the average length of the linear backbone increases as well, but not as much as the total number does. MD simulations at room temperature (298 K) and at 110 °C (383 K) are performed to understand the process of conformation transition at different temperatures. The snapshots of room temperature MD simulation in Figure S4 demonstrate that the molecular configuration oscillates around the half-circle conformation and that full relaxation to linear skeleton is not observed during the time scale of our calculations, which indicates that the energy barrier of conformation transformation cannot be overcome easily at room temperature. At 110 °C (383 K), relaxation of bending skeleton to quasi-linear one is quickly achieved as shown in Figure S5, which supports the experimental result that the total number and total length ratios of linear to curved NTZ12 increase sharply from 70 to 110 °C quite well. Furthermore, we examine the curved backbones in more detail. The bending angle as the key parameter of folding style is surveyed for all samples, distributed as shown in Figure 3. It

other domains. The third step in Scheme 2 illustrates the change of this stage. Comparing the data taken at 190 °C annealing with those taken after 150 °C annealing, it can be found that the essential features of structure derivation during the stage of 150−190 °C are similar to those of the 70−110 °C region: surface-occupied ratio, size of single domain, length of linear NTZ12, and number of molecules per domain increase noticeably, whereas domain density remains unchanged because of the balanced speed of formation and consumption of NTZ12 domains. In this process, NTZ12 may be transferred from more faraway layers to the interface layer. The altering characteristics of this process are summarized in the last stage of Scheme 2. Comparison of the data of room temperature samples to those of the highest annealing temperature shows that the overall increasing tendency of the surface-occupied ratio, the single domain area, and the average length of linear NTZ12 is rather straightforward. Only the statistical data of the domain density show an volcanic relationship with the annealing temperature. Meanwhile, the parameters of each stage disclose the detailed process of the structural transformation. The illustrative diagram combined with the essential changing characteristics of each stage is provided in Scheme 2 to show the structure derivation visually. Evolution of Single Molecular Conformations. After analyzing the structure evolution at the domain level, we attempt to understand the transition of NTZ12 at the single molecular dimension because the STM images resolve polymer backbones clearly even in the hundreds of nanometers (see in Figures 1 and 2). We characterize the single molecular conformation in two aspects: (1) ratios of linear to curved NTZ12 including average length ratio, total length ratio, and total number ratio, which disclose the statistical contribution of linear and nonlinear molecules in the structure of the interface film; and (2) folding angle distribution, which displays the distribution of the bending skeletons. The process of obtaining statistical values of these parameters is illustrated in Figure S1, Table S1, and the corresponding explanations in the Supporting Information. To avoid the influence of different numbers of counted STM images, for all the parameters except the average length, only the relative ratio of linear to curved NTZ12 is shown in Table 2 and discussed. From Table 2, it can be seen that a single curved NTZ12 is always longer than the linear one and both of them demonstrate a larger value at higher temperatures except for the length data of curved polymers at 150 °C. The slowly increasing ratio of the average length of linear to curved backbone discloses that the length of linear NTZ12 increases a little faster during the heating treatments. The ratios of the total number and total length of linear to curved NTZ12 reflect the contribution of these two kinds of conformations macroscopically. The statistical data of the total number ratio suggest that the total number of linear and curved polymers at room temperature is roughly the same. However, the ratio of total length suggests that all curved NTZ12 are larger than the linear ones at both room temperature and 70 °C, until at 110 °C when the latter exceeds, which indicates that a large part of NTZ12 backbones are trapped in a kinetic state (curved skeleton) at room temperature and at 70 °C whereas at high temperatures, more polymers are converted into thermally stable ones (linear skeleton). From the ratios of the total number and total length in Table 2, it can be found that these two factors demonstrate a similar trend, achieving the

Figure 3. Folding angle distribution of curved NTZ12.

can be observed very easily that in all samples, the relative frequency is quite low for 60° whereas it is rather high for 120° or 180° (hairpin conformation). Figure 3 shows that the frequencies for 60°, 90°, and 120° have a similar tide, increasing from room temperature to the largest value at 150 °C and then dropping abruptly at 190 °C. This tendency follows the variation in the total number and total length ratios of linear NTZ12 to curved one: increased by the ripening process but reduced by the formation of new domains. For the relative frequency at 180°, the case is reversed completely in comparison with that of 60°, 90°, and 120°, shrinking from room temperature to the smallest level at 150 °C then going up rapidly at 190 °C. The changing tendency of frequency at 180° F

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir indicates that the hairpin conformation is the result of a kinetic one and its content is cut down gradually to the valley position at 150 °C because of the ripening effect whereas formation of new domains at 190 °C counteracts the influence of the ripening process and causes the relative frequency to rise again. For the relative frequency at 150° bending and circle, the change in direction is not straightforward.

reduces at 180°, suggesting the energy state of 120° folding is more favorable than that of 180° folding. In this report, essential features of the structural transformation of the NTZ12 interface film at each annealing stage are proposed on the basis of the detailed statistical data extracted from a large number of STM images. The revealed intrinsic procedures are greatly helpful in learning the evolution of interface structures during the annealing of polymer devices, understanding how the performance improves through heat treatment, and guiding the parameter determination of postfabricating processes in the development of high-fulfillment polymer devices.



CONCLUSIONS The changing path of the interface domains and single molecular conformations of NTZ12 polymers during the annealing process is revealed using STM and theoretical calculations. On the one hand, the statistical data of the surfaceoccupied ratio, size of single domain, domain density, and the number of molecules per domain disclose the evolution route of NTZ12 domains. For all investigated samples, NTZ12 adsorbs onto the surface with the interdigitated side alkyl chains and shows varied folding styles. In the room temperature sample, the surface-occupied ratio is unexpectedly low. After annealing at 70 °C, the significant result is the formation of new domains because both surface-occupied ratio and domain density increase visibly whereas the size of a single domain remains as large as that of the room temperature sample. Interlayer molecular transportation supplies the molecules needed for the formation of new domains. The ripening process is not activated at this step; therefore, very few domains are consumed. The surface-occupied ratio and domain size increase largely at 110 °C, but the domain density is almost the same as that at 70 °C. Therefore, the production speed and the consumption speed of NTZ12 domains are balanced at this stage, leading to an extremely slow increase in the domain density. At 150 °C, the increase in the surface-occupied ratio is very less, whereas the increase in the domain size is obvious. As a result, there is a decrease in the domain density. We inferred that the depletion of molecules in the nearby layers is responsible for the small increase in the surface-occupied ratio and formation of new domains. So, some domains grow into bigger ones at the cost of other domains. The surfaceoccupied ratio starts expanding again at 190 °C whereas the domain density stays in the previous level, which indicates that the formation of new domains is reactivated and compensates for the consumption due to the ripening effect. The molecules in the layers farther from the interface film may be transported to the interface to support the formation of new domains. On the other hand, the mean values of the single molecular structure parameters, including the length of linear and curved polymer backbones, ratios of total length and total number of linear to curved NTZ12, and the angle distribution of curved skeletons, disclose the transformation route of the backbone conformation. The ratio of average length of linear to curved backbone is increased during the process of annealing because high temperature treatments provide enough energy for backbones relaxing from kinetically trapped style to thermally stabilized one. MD simulations confirm the quick relaxation of curved backbones at high temperature as well. The total length and total number ratios of linear to curved NTZ12 increase for the samples from room temperature to the maximum at 150 °C and descend at 190 °C, which is in agreement with the dominated process of 150 °C-ripening between domains and reduced formation of new domains. The two main arrangements of curved backbones are 120° and 180° foldings. Ripening at 150 °C promotes the relative frequency at 120° but



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.6b02139. Explanations on the statistical analysis of STM images using the ImageJ software, calculation of the parameters mentioned in the main text, AFM images, and the results of molecular dynamic simulations (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: +86-10-69672546 (Z.-Y.Y.). *E-mail: [email protected] (W.-F.Z.). *E-mail: [email protected] (G.Y.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work is supported in part by the National Natural Science Foundation of China (grants 21303200) and Youth Innovation Promotion Association of Chinese Academy of Sciences (2015360).



REFERENCES

(1) Bujak, P.; Kulszewicz-Bajer, I.; Zagorska, M.; Maurel, V.; Wielgus, I.; Pron, A. Polymers for Electronics and Spintronics. Chem. Soc. Rev. 2013, 42, 8895−8999. (2) Ma, H.; Yip, H.-L.; Huang, F.; Jen, A. K.-Y. Interface Engineering for Organic Electronics. Adv. Funct. Mater. 2010, 20, 1371−1388. (3) Beaujuge, P. M.; Fréchet, J. M. J. Molecular Design and Ordering Effects in π-Functional Materials for Transistor and Solar Cell Applications. J. Am. Chem. Soc. 2011, 133, 20009−20029. (4) Yao, Y.; Dong, H.; Hu, W. Ordering of Conjugated Polymer Molecules: Recent Advances and Perspectives. Polym. Chem. 2013, 4, 5197−5205. (5) Holliday, S.; Donaghey, J. E.; McCulloch, I. Advances in Charge Carrier Mobilities of Semiconducting Polymers Used in Organic Transistors. Chem. Mater. 2014, 26, 647−663. (6) Sun, X.; Di, C.-A.; Liu, Y. Engineering of the Dielectric− Semiconductor Interface in Organic Field-Effect Transistors. J. Mater. Chem. 2010, 20, 2599−2611. (7) Graetzel, M.; Janssen, R. A. J.; Mitzi, D. B.; Sargent, E. H. Materials Interface Engineering for Solution-Processed Photovoltaics. Nature 2012, 488, 304−312. (8) Miozzo, L.; Yassar, A.; Horowitz, G. Surface Engineering for High Performance Organic Electronic Devices: The Chemical Approach. J. Mater. Chem. 2010, 20, 2513−2538. G

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Poly[(S)-3-(3,7-dimethyloctyl)thiophene]. Adv. Mater. 2005, 17, 708− 712. (28) Bocheux, A.; Tahar-Djebbar, I.; Fiorini-Debuisschert, C.; Douillard, L.; Mathevet, F.; Attias, A.-J.; Charra, F. Self-Templating Polythiophene Derivatives: Electronic Decoupling of Conjugated Strands through Staggered Packing. Langmuir 2011, 27, 10251− 10255. (29) Willot, P.; Teyssandier, J.; Dujardin, W.; Adisoejoso, J.; De Feyter, S.; Moerman, D.; Leclère, P.; Lazzaroni, R.; Koeckelberghs, G. Direct Visualization of Microphase Separation in Block Copoly(3alkylthiophene)s. RSC Adv. 2015, 5, 8721−8726. (30) Scifo, L.; Dubois, M.; Brun, M.; Rannou, P.; Latil, S.; Rubio, A.; Grévin, B. Probing the Electronic Properties of Self-Organized Poly(3dodecylthiophene) Monolayers by Two-Dimensional Scanning Tunneling Spectroscopy Imaging at the Single Chain Scale. Nano Lett. 2006, 6, 1711−1718. (31) Liu, Y.-F.; Krug, K.; Lee, Y.-L. Self-Organization of TwoDimensional Poly(3-hexylthiophene) Crystals on Au(111) Surfaces. Nanoscale 2013, 5, 7936−7941. (32) Sakaguchi, H.; Matsumura, H.; Gong, H.; Abouelwafa, A. M. Direct Visualization of the Formation of Single-Molecule Conjugated Copolymers. Science 2005, 310, 1002−1006. (33) Lei, S.-B.; Wan, L.-J.; Wang, C.; Bai, C.-L. Direct Observation of the Ordering and Molecular Folding of Poly[(m-phenylenevinylene)co-(2,5-dioctoxy-p-phenylenevinylene)]. Adv. Mater. 2004, 16, 828− 831. (34) Lin, Z.-Q.; Shi, N.-E.; Li, Y.-B.; Qiu, D.; Zhang, L.; Lin, J.-Y.; Zhao, J.-F.; Wang, C.; Xie, L.-H.; Huang, W. Preparation and Characterization of Polyfluorene-Based Supramolecular π-Conjugated Polymer Gels. J. Phys. Chem. C 2011, 115, 4418−4424. (35) Lei, S.-B.; Deng, K.; Yang, Y.-L.; Zeng, Q.-D.; Wang, C.; Ma, Z.; Wang, P.; Zhou, Y.; Fan, Q.-L.; Huang, W. Substituent Effects on Two-Dimensional Assembling and Chain Folding of Rigid-Rod Polymer Poly(p-phenyleneethynylene) Derivatives on the Solid/ Liquid Interface. Macromolecules 2007, 40, 4552−4560. (36) Lei, S.; Deng, K.; Ma, Z.; Huang, W.; Wang, C. Templated Assembling of Phthalocyanine Arrays Along a Polymer Chain. Chem. Commun. 2011, 47, 8829−8831. (37) Nehls, B. S.; Füldner, S.; Preis, E.; Farrell, T.; Scherf, U. Microwave-Assisted Synthesis of 1,5- and 2,6-Linked NaphthyleneBased Ladder Polymers. Macromolecules 2005, 38, 687−694. (38) Cheng, C.; Yu, C.; Guo, Y.; Chen, H.; Fang, Y.; Yu, G.; Liu, Y. A Diketopyrrolopyrrole-Thiazolothiazole Copolymer for High Performance Organic Field-Effect Transistors. Chem. Commun. 2013, 49, 1998−2000. (39) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision E.01; Gaussian, Inc.: Wallingford, CT, 2009.

(9) Dong, H.; Jiang, L.; Hu, W. Interface Engineering for HighPerformance Organic Field-Effect Transistors. Phys. Chem. Chem. Phys. 2012, 14, 14165−14180. (10) Diao, Y.; Shaw, L.; Bao, Z.; Mannsfeld, S. C. B. Morphology Control Strategies for Solution-Processed Organic Semiconductor Thin Films. Energy Environ. Sci. 2014, 7, 2145−2159. (11) Tsao, H. N.; Müllen, K. Improving Polymer Transistor Performance via Morphology Control. Chem. Soc. Rev. 2010, 39, 2372−2386. (12) DeLongchamp, D. M.; Kline, R. J.; Fischer, D. A.; Richter, L. J.; Toney, M. F. Molecular Characterization of Organic Electronic Films. Adv. Mater. 2011, 23, 319−337. (13) Lei, T.; Wang, J.-Y.; Pei, J. Design, Synthesis, and Structure− Property Relationships of Isoindigo-Based Conjugated Polymers. Acc. Chem. Res. 2014, 47, 1117−1126. (14) Mali, K. S.; Adisoejoso, J.; Ghijsens, E.; De Cat, I.; De Feyter, S. Exploring the Complexity of Supramolecular Interactions for Patterning at the Liquid−Solid Interface. Acc. Chem. Res. 2012, 45, 1309−1320. (15) Wan, L.-J. Fabricating and Controlling Molecular SelfOrganization at Solid Surfaces: Studies by Scanning Tunneling Microscopy. Acc. Chem. Res. 2006, 39, 334−342. (16) Chen, T.; Yan, H.-J.; Yang, Z.-Y.; Wang, D.; Liu, M.-H. Assembling Structures of Barbituric Acid Derivatives on Graphite Surface Investigated with Scanning Tunneling Microscopy. J. Phys. Chem. C 2012, 116, 19349−19354. (17) Huang, W.; Zhao, T.-Y.; Wen, M.-W.; Yang, Z.-Y.; Xu, W.; Yi, Y.-P.; Xu, L.-P.; Wang, Z.-X.; Gu, Z.-J. Adlayer Structure of ShapePersistent Macrocycle Molecules: Fabrication and Tuning Investigated with Scanning Tunneling Microscopy. J. Phys. Chem. C 2014, 118, 6767−6772. (18) Yang, Z.-Y.; Tao, Y.; Chen, T.; Yan, H.-J.; Wang, Z.-X. Hydrogen Bonding Network of Truxenone on a Graphite Surface Studied with Scanning Tunneling Microscopy and Theoretical Computation. Phys. Chem. Chem. Phys. 2013, 15, 2105−2108. (19) Xu, L.; Yang, L.; Lei, S. Self-Assembly of Conjugated Oligomers and Polymers at the Interface: Structure and Properties. Nanoscale 2012, 4, 4399−4415. (20) Yang, Z.-Y.; Zhang, H.-M.; Pan, G.-B.; Wan, L.-J. Effect of the Bridge Alkylene Chain on Adlayer Structure and Property of Functional Oligothiophenes Studied with Scanning Tunneling Microscopy and Spectroscopy. ACS Nano 2008, 2, 743−749. (21) Yang, Z.-Y.; Zhang, H.-M.; Yan, C.-J.; Li, S.-S.; Yan, H.-J.; Song, W.-G.; Wan, L.-J. Scanning Tunneling Microscopy of the Formation, Transformation, and Property of Oligothiophene Self-Organizations on Graphite and Gold Surfaces. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 3707−3712. (22) Mena-Osteritz, E. Superstructures of Self-Organizing Thiophenes. Adv. Mater. 2002, 14, 609−616. (23) Mena-Osteritz, E.; Meyer, A.; Langeveld-Voss, B. M. W.; Janssen, R. A. J.; Meijer, E. W.; Bäuerle, P. Two-Dimensional Crystals of Poly(3-alkyl-thiophene)s: Direct Visualization Polyer Folds in Submolecular Resolution. Angew. Chem., Int. Ed. 2000, 39, 2679−2684. (24) Keg, P.; Lohani, A.; Fichou, D.; Lam, Y. M.; Wu, Y.; Ong, B. S.; Mhaisalkar, S. G. Direct Observation of Alkyl Chain Interdigitation in Conjugated Polyquarterthiophene Self-Organized on Graphite Surfaces. Macromol. Rapid Commun. 2008, 29, 1197−1202. (25) Peeters, H.; Couturon, P.; Vandeleene, S.; Moerman, D.; Leclère, P.; Lazzaroni, R.; De Cat, I.; De Feyter, S.; Koeckelberghs, G. Influence of the Regioregularity on the Chiral Supramolecular Organization of Poly(3-alkylsulfanylthiophene)s. RSC Adv. 2013, 3, 3342−3351. (26) Grévin, B.; Rannou, P.; Payerne, R.; Pron, A.; Travers, J. P. Multi-Scale Scanning Tunneling Microscopy Imaging of Self-Organized Regioregular Poly(3-hexylthiophene) Films. J. Chem. Phys. 2003, 118, 7097−7102. (27) Koeckelberghs, G.; Samyn, C.; Miura, A.; De Feyter, S.; De Schryver, F. C.; Sioncke, S.; Verbiest, T.; de Schaetzen, G.; Persoons, A. Polar Order in Spin-Coated Films of a Regioregular Chiral H

DOI: 10.1021/acs.langmuir.6b02139 Langmuir XXXX, XXX, XXX−XXX