I. Roy and
S. Hazra*
Saha Institute of Nuclear Physics, 1/AF Bidhannagar, Kolkata 700064, India. E-mail: satyajit.hazra@saha.ac.in
First published on 12th January 2017
The structures of the spin-coated poly(3-dodecylthiophene) [P3DDT] thin films along the out-of-plane and in-plane directions, and the influence of solvent vapor annealing and thermal annealing on their structures, which are of immense importance for their performance as semiconducting materials, were investigated using complementary X-ray reflectivity and atomic force microscopy techniques. The P3DDT films of different thicknesses (∼3.4 nm, i.e. ‘monolayer’ and ∼11 nm i.e. ‘multilayer’) prepared using different polymer concentration and solvents (chloroform and toluene) are found to compose of grain-like structures (of size ξ) of predominantly edge-on orientation (probably to maximize the coverage of the low energy methyl groups at both the interfaces) with some defects. Edge-on oriented grains are particularly prominent for the monolayer films (ξ ≈ 50 nm), where the interactions at both the interfaces are weak, but not in the case of the multilayer films (ξ ≈ 10 nm), where out-of-plane interaction between P3DDT molecules and evaporation rate of solvent also become important. Slow evaporation of toluene creates better ordering in that multilayer film. Solvent vapor annealing of the films although reduces the defects, forms small size grains with relatively less ordered structure due to near dissolution (or coil-like conformation) of the P3DDT at saturated vapor of good solvent and subsequent hindrance in the recovery of rod-like conformation and its π–π stacking under fast solvent evaporation. The change in the grain-size is only effective on the monolayer films, where the initial grain-size was large, while the change in the ordering is mainly effective on the toluene evaporated multilayer film, where the initial ordering was better. Further thermal annealing of the films only recovers the size of the grains and the ordering of the P3DDT molecules to a very small extent. However, thermal annealing in presence of solvent vapor on the as-deposited multilayer film show remarkable improvement of edge-on ordering. It is likely that the temperature first restrict the amount of solvent vapor and then such limited vapor pressure in presence of an intermediate temperature helps the reorganization of P3DDT molecules to form edge-on oriented structure.
The most widely studied member of the P3AT family is, perhaps, P3HT, which owes its popularity to its optimal alkyl chain length (6 carbon atoms). P3HT-based devices are shown to give rise to higher field-effect mobilities than devices based on P3ATs with longer or shorter alkyl chains. However, stronger photoluminescence26 and electroluminescence27 intensities are obtained from P3ATs with longer alkyl side chains, such as poly(3-dodecylthiophene) (P3DDT), than P3ATs with shorter alkyl chains. In spite of that, P3DDT is less studied due to its lower crystallizability, which is due to longer side chain interactions.13,15,23,28–32 However, the longer alkyl chains of P3DDT endow them with better solubility, which is conducive for preparing films with better continuity and smoothness,15 which, in turn, is beneficial for device applications. In fact, we are interested to have smooth and continuous edge-on oriented ordered film, which is only possible by preparing continuous smooth film first and then try to enhance the ordering by tuning the post annealing process keeping the previous two parameters (namely continuity and smoothness) almost intact to create long range ordering. Spin coating is the most common method for preparing smooth and uniform polymer films. However, it is known that the rapid evaporation of the solvent in the spin coating technique may result in non-equilibrium film structure with some residual strain and defects, which are hindrance to the proper device performances. The removal of such strain and/or defects are important, which can be done by solvent vapor annealing (SVA) and/or thermal annealing (TA). Annealing essentially enhances the mobility of the polymer chain, which can remove the strain and can also change the ordering and hence the device performance.33 For example, TA was found to rearrange polymer chains into a more ordered structure on slow cooling.34 On the other hand, SVA was found to enhance the ordering35,36 and even change in the orientation.37 However, no attempt has been made to understand the effect of SVA and TA on the structures of smooth P3DDT thin films.
In this paper we focus our attention in understanding the effect of SVA and TA on the structures of spin-coated P3DDT films of monolayer and multilayer thicknesses using complementary X-ray reflectivity (XR)15,38–40 and atomic force microscopy (AFM) techniques.41–43 The XR technique essentially provides an electron-density profile (EDP), i.e., in-plane (x–y) average electron density (ρ) as a function of depth (z) in high resolution, i.e., out-of-plane information of the films15,41–43 and is particularly preferential over other X-ray scattering techniques to get the information about the perfect edge-on ordering of a continuous smooth film. On the other hand, AFM technique provides topographic information from which one can infer about the in-plane structures of the films.41,43 Combination of these two techniques suggests that there is more in-plane structural modification compared to the out-of-plane structural modification after the SVA and TA of the P3DDT thin films, while simultaneous SVA and TA enhances the edge-on ordering quite appreciably. Possible reasons for such structural modification and its implication to the device properties are discussed.
All the films were characterized using XR and AFM techniques. XR measurements of the as-cast spin-coated P3DDT thin films and after SVA and also after TA were performed using a versatile X-ray diffractometer (VXRD) setup,15,41,43,48,49 to get statistically averaged electron density information. VXRD consists of a diffractometer (D8 Discover, Bruker AXS) with a Cu source (sealed tube) followed by a Göbel mirror to select and enhance Cu Kα radiation (λ = 1.54 Å) and further followed by slits to define the size (0.2 mm × 10 mm) of the beam. The diffractometer has a two-circle goniometer [θ(ω) − 2θ] with a quarter-circle Eulerian cradle as a sample stage. The latter has two circular (χ and ϕ) and three translational (X, Y, and Z) motions. The scattered beam was detected using a NaI scintillation (point) detector. Data were taken in the specular condition, i.e. reflected angle is equal to the incident angle, θ. Under such condition there exists a nonvanishing wave vector component, qz, which is equal to (4π/λ)sinθ with resolution 0.002 Å−1. The detailed top surface morphologies of the as-cast spin-coated P3DDT thin films and after SVA and also after TA were mapped using an AFM (5500 AFM, Agilent) technique.43 Topographic images were collected in a noncontact (or intermittence contact) mode to minimize silicon tip-induced damage of the soft film. Scans of different sizes and in different portions of the sample were carried out to get statistically meaningful information about the topography. WSXM software50 was used for AFM image processing and analysis. Power spectral densities (PSD) have been calculated from the AFM images to find out the in-plane correlation lengths of the samples.43
Fig. 1 XR data and analyzed curves for a as-cast SL-CF–P3DDT/OH–Si film and after SVA and TA treatments. Inset: corresponding analyzed EDPs. XR curves and EDPs are shifted vertically for clarity. |
Fig. 2 XR data and analyzed curves for a as-cast ML-CF–P3DDT/OH–Si film and after SVA and TA treatments. Inset: corresponding analyzed EDPs. XR curves and EDPs are shifted vertically for clarity. |
Fig. 3 XR data and analyzed curves for a as-cast SL-TL–P3DDT/OH–Si film and after SVA and TA treatments. Inset: corresponding analyzed EDPs. XR curves and EDPs are shifted vertically for clarity. |
Fig. 4 XR data and analyzed curves for a as-cast ML-TL–P3DDT/OH–Si film and after SVA and TA treatments. Inset: corresponding analyzed EDPs. XR curves and EDPs are shifted vertically for clarity. |
To get the quantitative information about the films and their small change after SVA and TA, all XR profiles have been analyzed using Parratt's formalism51 after incorporating roughness at each interface.15 It is necessary to mention that the simple 1-layer or 2-layer model, which mainly arises due to the amorphous structure of the film without or with low density interfacial layer, cannot fit the XR data (as shown in Fig. S1–S4 of the ESI†). Rather 3-layer (of low-high-low density due to predominantly edge-on structure) or m-layer (i.e. repetition of 3-layer structure) models are essential for the analysis (as shown in Fig. S1–S4 of the ESI†). It is clear that the better contrast between low and high density layers and/or their periodicity towards uniform value (of about 2.6 nm) suggest better or more edge-on ordering in the film. The best fit XR profiles along with the corresponding EDPs obtained from the analysis are shown in Fig. 1–4. The position of silicon oxide layer in the EDPs are indicated by the gray color. It can be noted that for the films prepared using TL solvent, a layer of large thickness but low density or coverage, on top of the regular single layer or multilayers, was necessary to improve the quality of the XR data fitting, which may be related to some patches in such films and can be neglected for further analysis. On the other hand, the average film thickness (D), the top surface roughness (σtop), the film-substrate interfacial roughness (σin) and the electron density difference between interfacial dip and next peak (Δρ), which are the important parameters to understand the evolution of the structure and/or ordering of the film, are extracted from the EDPs and tabulated in Table 1 for comparison.
Sample | Solvent | c [mg ml−1] | Treatment | D [nm] | Δρ [e Å−3] | σin [nm] | σtop [nm] | σAFM [nm] | ξ [nm] | σξ [nm] | σd [nm] |
---|---|---|---|---|---|---|---|---|---|---|---|
SL-CF–P3DDT/OH–Si | CF | 2/3 | As-cast | 3.4 | 0.11 | 0.44 | 0.20 | 0.26 | 48 | 17.0 | 30 |
SVA (CF) | 3.5 | 0.09 | 0.47 | 0.21 | 0.07 | 6 | 2.2 | 5 | |||
TA (130 °C) | 3.4 | 0.09 | 0.47 | 0.22 | 0.09 | 9 | 2.9 | 8 | |||
ML-CF–P3DDT/OH–Si | CF | 2 | As-cast | 13.0 | 0.06 | 0.45 | 0.65 | 0.27 | 12 | 3.5 | 6 |
SVA (CF) | 13.0 | 0.07 | 0.45 | 0.65 | 0.26 | 19 | 5.0 | 8 | |||
TA (130 °C) | 12.6 | 0.11 | 0.45 | 0.65 | 0.15 | 15 | 4.2 | 7 | |||
SL-TL–P3DDT/OH–Si | TL | 1 | As-cast | 3.7 | 0.16 | 0.39 | 0.17 | 0.28 | 48 | 14.0 | 30 |
SVA (TL) | 3.5 | 0.06 | 0.45 | 0.37 | 0.10 | 9 | 2.6 | 8 | |||
TA (130 °C) | 3.6 | 0.15 | 0.41 | 0.16 | 0.12 | 11 | 3.0 | 8 | |||
ML-TL–P3DDT/OH–Si | TL | 3 | As-cast | 11.2 | 0.22 | 0.36 | 0.50 | 0.14 | 9 | 2.6 | 5 |
SVA (TL) | 11.2 | 0.08 | 0.45 | 0.70 | 0.21 | 9 | 2.4 | 5 | |||
TA (130 °C) | 11.0 | 0.11 | 0.42 | 0.68 | 0.20 | 11 | 2.9 | 6 |
It is clear from the EDPs (Fig. 1) and the parameters (Table 1) that there is very little change in the SL-CF–P3DDT/OH–Si film after SVA and TA. Moderate value of Δρ in the as-cast film suggests moderate edge-on oriented ordering of P3DDT in near single layer, the value of which decreases, while the value of both the roughness (σin and σtop) increases slightly after annealing suggesting slight deterioration in the edge-on ordering. For the as-cast ML-CF–P3DDT/OH–Si film the value of Δρ is found very low, suggesting very little edge-on oriented ordering of P3DDT near the substrate and in the subsequent layers. After annealing such Δρ-value increases a little but no change occurs in any roughness value, suggesting slight improvement in the edge-on oriented ordering of P3DDT only near the film-substrate interface especially after TA. Thus for the P3DDT films prepared from CF solvent the effect of annealing in the edge-on oriented ordering of P3DDT near the film-substrate interface is found opposite for single- and multi-layer stack films. For the as-cast SL-TL–P3DDT/OH–Si film the value of Δρ is found quite high when compared with the SL-CF–P3DDT/OH–Si film, suggesting better edge-on oriented ordering of P3DDT in the film. After SVA, the roughness value increases and the Δρ-value decreases appreciably suggesting quite strong deterioration in the edge-on oriented ordering of P3DDT. Interestingly after TA, the ordering in the film is nearly recovered. Finally for the as-cast ML-TL–P3DDT/OH–Si film the value of Δρ is found highest and the value of σin is found lowest among all the films, suggesting the best edge-on oriented ordering of P3DDT near the film-substrate interface. Also better ordering is found throughout this film. After SVA, the roughness value increases and the Δρ-value decreases significantly suggesting strong deterioration in the edge-on oriented ordering of P3DDT. After TA, the ordering in the film is recovered but only to some extent. Thus for the P3DDT films prepared from TL solvent the effect of annealing in the edge-on oriented ordering of P3DDT near the film-substrate interface is found qualitatively similar for single- and multi-layer stack films.
In order to get the proper visualization about the height-fluctuation, bearing ratio (i.e. coverage as a function of height) estimated from the AFM images are presented in Fig. 5–8. The height-fluctuation can be quantified through roughness value (σAFM), which is tabulated in Table 1. Large height-fluctuation in the as-cast SL-CF–P3DDT/OH–Si film (as evident from both Fig. 5 and Table 1) decreases significantly after annealing. For the ML-CF–P3DDT/OH–Si film (see Fig. 6 and Table 1), the large height-fluctuation in the as-cast film remains almost same (though changes slightly in nature) after SVA but decreases appreciably after TA. The height-fluctuation of the SL-TL–P3DDT/OH–Si film is found similar to that of the SL-CF–P3DDT/OH–Si film, that is, large height-fluctuation in the as-cast film decreases significantly after annealing (see Fig. 7 and Table 1). For the ML-TL–P3DDT/OH–Si film (see Fig. 8 and Table 1), the nature is somewhat different. The small height-fluctuation in the as-cast film increases slightly after annealing.
Now to have a better idea about the in-plane information or the height–height correlation, the PSD52,53 for all the films, extracted from the AFM images of different scan size,43 are also plotted in Fig. 5–8. It can be noted that the PSD is the angular averaged radial distribution of the Fourier transformed AFM image, which can be expressed as
(1) |
(2) |
It is very clear from Table 1 that the in-plane correlation length or more specifically the value of ξ of the as-cast SL-CF–P3DDT/OH–Si and SL-TL–P3DDT/OH–Si films are quite large, which decreases considerably after SVA and recovers very little after TA. The decrease in the value of ξ after annealing is the indication of the deterioration in the ordering or crystallinity of the film. However, the decrease in the value of σAFM suggests that there is a strong decrease in the size and depth of the defects as well. Such decrease counterbalances to a large extent the apparent deterioration in the quality of the film. On the other hand, the value of ξ of the as-cast ML-CF–P3DDT/OH–Si and ML-TL–P3DDT/OH–Si films are quite small, which, after annealing, increases a little bit. However, for the ML-TL–P3DDT/OH–Si film, the value of σAFM increases appreciably, which effectively deteriorates the quality of the film, while for the ML-CF–P3DDT/OH–Si film, the roughness value decreases with the small increase in grain-size, which essentially suggests that annealing only improves the quality of this film.
It is known that the structure of a spin-coated thin film is strongly influenced by the nature of the solvent, the substrate surface and the deposition conditions. The surface free energy, polar–nonpolar (hydrophilic–hydrophobic) or electrostatic nature of a substrate surface are strongly influenced by the surface passivation, which can then control the substrate–polymer interactions.15,16,41,42,47,54,55 For example, the adsorption/interaction of P3DDT molecules on/with the weak-hydrophilic OH–Si substrate has been predicted weak,15 which helps the in-plane diffusion and organization of P3DDT molecules, overcoming the fast evaporation induced immobilization during spin-coating. It can be noted that the spinning of solution during such deposition conditions forces the polymer backbones to lie on the surface, which then try to arrange in predominantly edge-on oriented structures. The latter is because the edge-on configuration maximizes the coverage of the low-energy methyl groups at both interfaces and thus it is likely the lowest energy configuration state.56 The arrangement is prominent when monolayer thickness is formed (as evident from the large size predominantly edge-on oriented grains for the as-cast SL films), where interfacial interaction helps the polymer diffusion and ordering. On the other hand, when multilayer thickness is formed through spin coating, the evaporation rate (E) of the solvent becomes important. Fast evaporation of chloroform (E = 11.6)57 creates hindrance in the diffusion and thus in the organization of the P3DDT molecules, particularly above the first layer, which are not in direct contact with the substrate (low contrast between the low and high density layers in the EDP), while the slow evaporation of toluene (E = 2.24)57 allows sufficient time to the P3DDT molecules to diffuse and organize (relatively high contrast between the low and high density layers in the EDP). Accordingly, well ordered edge-on orientated structure is observed in the as-cast ML-TL–P3DDT/OH–Si film. However, the in-plane size or extent of the edge-on orientated grains decreases considerably for the multilayer films. This is probably due to out-of-plane interaction between P3DDT molecules, which creates hindrance in the extent of the in-plane ordering and thus in the grain-size.
The SVA of the films is carried out, by placing the as-cast films and a small beaker containing the solvent inside a covered container, i.e., at saturated vapor pressure for 2 h and then removing the cover. The saturated vapor is likely to increase the mobility of the polymer to a great extent, which is likely to destroy all the ordering. Evaporation of solvent after removal of the cover is unable to recover all the ordering. This is particularly noticeable for the monolayer thick films, where the large-size grains formed during spin-coating technique (due to laying of polymer backbones on the substrate and weak substrate–polymer interaction) are dewetted after SVA to form much small size grains. Further TA can not improve the deteriorated ordering to a large extent as it can not fully recover the size of the grains. For the multilayer films, where the grain-size of the as-cast film is small, it is well recovered after SVA, suggesting that normal evaporation of solvent after large mobility of the polymer during SVA, can form such small size grains. However, the grain–grain ordering is deteriorated which is particularly evident for the ML-TL–P3DDT/OH–Si film, where the ordering of the as-cast film was high. Further TA, although recovers some ordering through diffusion, the recovery is very small.
The observed decrease in the size of defects after annealing, which is expected to improve the electrical transport properties of the film, will certainly be nullified by the deterioration in the ordering of the film during annealing. The latter is due to the uncontrolled diffusion of polymers, particularly for the SVA, where a large intake of good solvent under saturated vapor pressure essentially tries to dissolve the polymer i.e. it dislodges/destroys the existing π–π stacking of the polymer and subsequently their rod-like conformation (to coil-like). Subsequent uncontrolled fast evaporation of the good solvent neither helps to improve nor even helps to recover the unfavorable23 rod-like conformation (from coil-like) of the polymer and their π–π stacking. As a result, small size grains and poor ordering are observed after SVA, which is not recovered even after further TA. Thus it suggests that the controlled annealing of the film is necessary for the relaxation of stress, the annihilation of defects (created during spin-coating) and the organization of molecules to obtain P3DDT films with better order and less defects for the improved device properties.
To verify this we have performed thermal annealing (at 130 °C) in presence of solvent vapor (SVTA) of an as-deposited ML-P3DDT film. The XR data and the analyzed profiles of the as-deposited ML-P3DDT film and after SVTA are shown in Fig. 9. The enhancement of the pseudo Bragg peak (at around qz ≈ 0.24 Å−1) and the appearance of the second order peak after SVTA are clearly evident. This corresponds to the improvement in the edge-on ordering of P3DDTs in the film after SVTA, which is clearly evident from the enhancement in the electron density contrast between high density polymer backbone and low density side chains of P3DDT (see the analyzed EDPs in the inset of Fig. 9). The drastic improvement of the edge-on ordering upon SVTA can be visualized as follows. First the temperature restrict or control the amount of solvent vapor. Such limited vapor pressure in presence of an optimum intermediate temperature [between melting of the alkyl side chains (53 °C) and polythiophene backbones (163 °C)]23 helps the reorganization to form edge-on oriented structure.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra25560d |
This journal is © The Royal Society of Chemistry 2017 |