Conversion from self-assembled block copolymer nanodomains to carbon nanostructures with well-defined morphology

Abstract

We developed a method to fabricate arrays of nanostructured carbon materials of high quality, via direct pyrolysis of poly(styrene-b-2-vinylpyridine) (PS-b-P2VP) block copolymer (BCP) as thin films without small organic molecules incorporated as carbon resources. Prior to pyrolysis, the nanodomains were cross-linked with UV radiation under nitrogen (UVIN). As upon pyrolysis thermal energy imparts mobility to nanodomains to overcome the constraints imposed by cross-linked chains, the free surface is inevitably covered with the PS block that has a smaller surface energy; unless favorable interactions with the P2VP block exist at the free surface, the outer layer of carbonaceous films is predominantly composed of the pyrolyzed PS block. To overcome this problem, films were capped with a SiO₂ layer after UVIN. The capping layer appears to have two advantages – increased areal yields and an improved morphological fidelity. As a result, arrays of nanostructured carbon of great quality were fabricated even at temperatures far above the decomposition temperatures of PS and P2VP blocks.

---

Introduction

Carbon nanostructures of zero, one, two or three dimensions (0D–3D) have aroused great interest because of their potential applications as catalysts for fuel cells, sensors and supercapacitors, to name a few areas.¹ Carbon nanostructures doped with nitrogen, for example, possess catalytic activities for some important electrochemical reactions such as oxygen reduction.^2–4 The growth of nitrogen-doped carbon nanostructures from carbon and nitrogen precursors with chemical-vapor deposition on catalytically active metals is a recent method to produce carbon nanostructures with a well defined morphology and excellent properties.²

Another strategy is to use a block copolymer (BCP) as a template to produce carbon nanostructures after pyrolysis at elevated temperatures, as self-assembly of a block copolymer offers access to ordered nanodomains with tunable dimension and structure through control of such molecular parameters as volume fraction and molecular mass.^5–16 Several literature reports describe small organic molecules as sources of carbon within BCP self-assembled templates that yield carbon nanostructures of all dimensionalities.^5–16 As BCP materials act only as templates, removal of these templates is necessary after pyrolysis. Recent reports have demonstrated that cross-linked self-assembled BCP can be pyrolyzed directly into carbon nanostructures with morphological fidelity without addition of small organic precursors, as BCP itself is an organic material that can be regarded as a source of solid carbon.^17–28 In particular, pyrolysis of BCP materials with one block composed of components containing nitrogen, such as poly(acrylonitrile) (PAN) or poly(vinylpyridine) (PVP), can directly produce carbon nanostructures enriched with nitrogen. For vinyl-based BCP, cross-linking is triggered with UV radiation in an inert environment.^23,24 In contrast, for BCP based on PAN, UV irradiation is unnecessary as PAN itself can trigger cross-linking during thermal pyrolysis after cyclization.^17–22,25 Nitrogen-enriched carbon nanostructures also demonstrate remarkable catalytic activities for oxygen reduction.^25–28 For comparison, the pyrolysis of self-assembled BCP nanodomains might present the two advantages that the fabrication of carbon nanostructures with a high yield is cost-effective, and that the self-assembly of mesoscopic structures can be precisely controlled with simple and cheap processes; both advantages are keys to the large-scale production of carbon nanostructures. Nevertheless, the mechanism of pyrolysis of BCP nanodomains remains poorly understood; as gaseous molecules inevitably grow through chain scission during pyrolysis, escape of these gaseous molecules causes a significant loss of mass. Retaining both a high yield and morphological fidelity of solid carbon nanostructures is thus still challenging.

In this work, we studied the pyrolysis of polystyrene-block-poly(2-vinylpyridine), PS-b-P2VP; their thin films provide a model system to understand the mechanisms of pyrolysis to form carbon nanostructures. The PVP block has been demonstrated to bind with metal ions through favorable interactions due to electrons in lone pairs in the plane of the pyridine rings; the PVP-containing BCP can serve as templates for the fabrication of nanoscale dots and wire of metal materials.^29–37 The control of spatial orientation, order and morphological diversity has been well addressed through annealing in variously selected solvent vapors for PVP-containing BCP.^38–44 Since the discovery of the synthesis of carbon nanostructures through a direct pyrolysis of PS-b-P2VP nanodomains,²³ many topics including metal/carbon nanocomposites have been widely studied,^45,46 but relations between the pyrolysis temperature and the solid carbonaceous residue have yet to be clarified. The objective of our work was to examine how the temperature of pyrolysis influences the morphological fidelity and thermal stability of the structural dimensions of carbon nanostructures fabricated with pyrolysis of PS-b-P2VP thin films comprising self-assembled nanodomains of three types. To form these nanodomains, annealing in a solvent vapor (SVA) was performed. Before pyrolysis at elevated temperatures, the nanodomains were stabilized with two methods – UV irradiation under nitrogen (UVIN) or this UVIN treatment combined with capping with a silicon dioxide layer on top of the thin films. The stabilization of nanodomains is important for solid carbonaceous residues and even retention of morphology when pyrolysis is performed at temperatures above the decomposition temperatures (T_d) of the two blocks. The latter method appears more effective than the former for the fabrication of carbon nanostructures with morphological fidelity.

Experimental section

Materials and sample preparation

Three PS-b-P2VP BCPs and homopolymer polystyrene (hPS) were purchased from Polymer Source Inc and used as received. Homopolymer poly(2-vinylpiridine) (hP2VP) and o-xylene were purchase from Aldrich and used as received. The PS-b-P2VP BCPs are briefly designated as P(S_m2VP_n) in which m and n represent the number-average molecular weight of each constituted block. The characteristics of the polymeric materials are shown in Table 1.

Table 1 Characteristics of PS-b-P2VP BCPs and PS and P2VP homopolymers^a

Sample	M^totaln (kg mol^−1)	M^PSn (kg mol^−1)	M^P2VPn (kg mol^−1)	Mw/Mn	fP2VP (%)	Morphology
a S, C⊥, and C‖ respectively denote truncated core–shell nanospheres, upstanding and lying nanocylinders.
P(S500002VP16500)	66.5	50	16.5	1.09	25	S, C⊥ and C‖
P(S485002VP70000)	118.5	48.5	70	1.13	60	S
PS45000	45	45		1.05
P2VP152000	152		152	1.05

---

We spin-coated 1.5 or 2.0 mass% solutions of PS-b-P2VP in o-xylene at 1000 rpm (60 s) onto cleaned Si wafer substrates to prepare thin films of different thicknesses. The substrates were cleaned in piranha solutions (30% H₂O₂/98% H₂SO₄: 3/7, v/v) for 40 min, rinsed with deionized water for 3 times, and subsequently dried under N₂ flow. For solvent annealing, thin films were then placed in a sealed jar (110 mL) with saturated chloroform or acetone vapor for 4 h at a maintained temperature 21 ± 1 °C. Chloroform (δ = 19 MPa^1/2) is a nearly neutral solvent for both PS (δ = 17.6 MPa^1/2) and P2VP (δ = 20.4 MPa^1/2)⁴¹ whereas acetone (δ = 20 MPa^1/2) is a highly selective solvent.⁴⁷ After specimens were treated with SVA for a certain period of time, then those were quickly removed from solvent vapor and dried at ambient temperature.

For cross-linking self-assembled nanodomains, the films were directly exposed to UVIN (UV lamp: a germicidal lamp of G20T10 20W light tube, SANKYO DENKI). The irradiation intensity from the UV lamp was gauged with a UVX Radiometer (UVP, LLC) at the wavelength λ = 254 nm and the dosage was 2.4 mW cm⁻² at a distance of 10 cm, which was the length used in all the experiments included ex situ atomic force microscopy (AFM) and X-ray photoemission spectroscopy (XPS). The UVIN dosage was controlled by varied time intervals (from 10 m to 9 h) of exposure to UV light. The UVIN-treated films were pyrolyzed in a one-zone diffusion furnace at temperatures in the range of 400–900 °C for 1 h in argon (Ar) gas to form carbon nanostructures. For deposition of a capping layer of silicon dioxide (SiO₂), the samples were loaded into a high vacuum electron-beam evaporation system. The base pressure in the deposition chamber was better than 5 × 10⁻⁷ torr. A 40 or 100 nm-thick amorphous SiO₂ (a-SiO₂) thin film (purity: 99.999%) was deposited onto the polymer film-coated Si substrate each at room temperature. The deposition rate was controlled at around 0.06–0.1 nm s⁻¹. During the deposition process, the film thickness and deposition rate were monitored in situ by a quartz crystal microbalance.

Apparatus and characterization

AFM images were collected by an atomic force microscope (AFM, Seiko SPA400) in tapping mode. Aluminum-coated silicon cantilevers were used with the force coefficient of 7.4 N m⁻¹, and the resonant frequency was 160 kHz (length: 150 μm, width: 26 μm, thickness: 300 μm). The images were monitored with a scan rate of 1 Hz and scanning pixels of 256 lines per frames. Carbon nanostructures and capping layers were characterized with a high-resolution field-emission scattering electron microscope (HR-FESEM, JEOL JSM-7600F) operated at 10 kV. A Raman microscope was used to observe the Raman inelastic scattering signals excited by a laser with the wavelength 532 nm and with laser power of 5–30 mW. The acquisition time of each Raman spectral curve was 30 s. Spectral resolution was usually set to 3 cm⁻¹ for this cooled CCD detection system (DV401-BV, ANDOR^TW) and the spectrometer fitted with a 1200 groove per mm grating. The frequency calibration was set by reference to the 520 cm⁻¹ vibrational band of a silicon wafer. The Raman data were obtained with a 100× microscope objective for Raman excitation/collection. The analysis of chemical components of micellar films was performed with a XPS system (VG Scientific ESCALab 250) with a Mg Kα radiation (max: 15 keV) and beam sizes of 650–120 μm. Thermogravimetric analysis (TGA) measurements were taken using a TGA-7 thermogravimetric analyzer (PERKIN ELMER). Samples (∼15 mg powder) were heated from room temperature to 700 °C at 10 °C min⁻¹. The film thickness was measured by Alpha-Step (KLA tencor Alpha-Step IQ). Grazing incident small angle X-ray scattering (GISAXS) measurements were implemented with a synchrotron X-ray radiation source at beamline BL23A1, National Synchrotron Radiation Research Center (NSRRC) at Hsinchu or with a Nano-Viewer (Rigaku) using Cu K_α X-rays at National Central University. For GISAXS measurements at NSRRC, the wavelength λ = 1.24 Å was used, and GISAXS patterns were collected by a two-dimensional detector (PILATUS 1M) with a pixel array of 1043 × 981. The scattering vectors in these GISAXS patterns were calibrated by a sample standard of silver behenate. To acquire scattering images with high signals-to-noise ratios, the angle of incidence of each X-ray beam was α_i = 0.2°. Those images were acquired with data recording time of 10–60 s. The in-house GISAXS measurement was also implemented with X-rays produced by a rotating copper-anode generator (Nano-viewer, Rigaku) operated at 1.2 kW in vacuum (40 kV and 30 mA) equipped with confocal max-flux optics. X-ray scattering images were collected with a 2D areal detector (Pilatus 100K of 83.8 × 33.5 mm²). Exposure duration of GISAXS measurement was 30 min for each image. The condition of incident angle was determined at 0.2°, which is a proper angle to acquire scattering images with high signals-to-noise ratio.

Results and discussion

Surface morphology of PS-b-P2VP thin films with various nanostructures

To make self-assembled nanodomains of three types, we used varied molecular masses (MM) with P2VP volume fractions 25% and 60%. Microphase-separated nanodomains of PS-b-P2VP in thin films were examined with AFM in tapping mode and SEM. Closely packed nanospheres of multi-layer thickness were prepared on spin coating from P(S₄₈₅₀₀2VP₇₀₀₀₀) (2.0 mass%) in o-xylene (Fig. 1a and b). The thickness of the as-spun P(S₄₈₅₀₀2VP₇₀₀₀₀) micellar film was approximately 102 nm. The corresponding GISAXS 2D pattern and 1D profile of the closely packed and dense layer of micelles with short-range order on a solid only reveals a discernible interference pattern at the low q_‖ region and fringes at the medium q_‖ region (Fig. 1c and d). Scattering features corresponding to the form factor of core–shell truncated spheres are absent from Fig. 1d. Our previous work demonstrated that, because o-xylene is effective for PS but poor for P2VP, PS-b-P2VP chains tend to form core–shell spherical micelles; at a small concentration in solution each such micelle comprises a swollen PS shell and a compact P2VP core.⁴⁸ Upon spin coating onto a solid substrate, the micelles become truncated along the substrate normal and thus appear as truncated nanospheres in the monolayer region.⁴⁸ When the micelles covered a substrate at large densities, the micelles would closely contact and tightly adhere with their neighbors. The PS shells tended to interpenetrate with their identical species to form a continuous matrix. As a result, the morphology of closely packed micelles resembled discrete spherical P2VP domains dispersed within the PS matrix.

Fig. 1 (a) A 2 × 2 μm AFM topographic image, (b) SEM image, (c) GISAXS 2D pattern, and (d) 1D in-plane profile for the micellar film prepared on spin coating from P(S₄₈₅₀₀2VP₇₀₀₀₀) (2.0 mass%) in o-xylene.

Cylindrical nanodomains were prepared with spin coating from o-xylene containing P(S₅₀₀₀₀2VP₁₆₅₀₀) (1.5 mass%) and subsequently with SVA. At such a concentration, the thickness of the P(S₅₀₀₀₀2VP₁₆₅₀₀) films as spun was approximately 90 nm, determined with an Alpha step. Before SVA, spin coating from o-xylene also produced spherical micelles with short-range spatial order due to a rapid removal of the solvent (Fig. S1†); the inter-micelle distance was 27.3 nm. Wang and coworkers reported that the orientation and spatial order of nanocylinders can be controlled with solvent selectivity and drying rate.⁴¹ After SVA, the P(S₅₀₀₀₀2VP₁₆₅₀₀) spherical micelles can further grow into nanocylinders with improved spatial order. According to this annealing strategy,⁴¹ P2VP nanocylinders having either perpendicular or parallel orientations embedded within the PS matrix were prepared on annealing P(S₅₀₀₀₀2VP₁₆₅₀₀) thin films respectively in vapors of trichloromethane and acetone (Fig. 2). For SVA in trichloromethane, the P(S₅₀₀₀₀2VP₁₆₅₀₀) thin film has nanocylinders with a perpendicular orientation and hexagonal packing in a plane (Fig. 2a), whereas for SVA in acetone P2VP nanocylinders tend to orient parallel to the substrate surface (Fig. 2d). The P2VP nanodomains appeared as shallow concavities of depth a few nanometers on the surface of the annealed film, as the upper ends of the standing P2VP cylindrical nanodomains and the half layer of the lying P2VP cylindrical nanodomains were lower than the surrounding PS matrix.⁴¹ Fig. 2b, c, e and f show GISAXS 2D patterns and 1D in-plane profiles, which exhibit Bragg diffraction truncated rods in a series, indicating that the nanocylinders within the SAV-treated films are highly oriented in the plane of the substrate. The GISAXS profile for P(S₅₀₀₀₀2VP₁₆₅₀₀) with SVA in trichloromethane vapor displays sharp diffraction rods located at q_‖ ratio 1 : 3^1/2, with the first order rod being at q_‖ = 0.0175 Å⁻¹, whereas these diffraction features with q_‖ ratios 1 : 4^1/2 are discerned for P(S₅₀₀₀₀2VP₁₆₅₀₀) with SVA in acetone vapor (Fig. 2c and f). The 3^1/2 rod associated with the diffraction of the {11} plane can be present only for hexagonally packed nanocylinders.⁴² We thus ascribe the absence of the 3^1/2 rod to the fact that lying P2VP nanocylinders lack hexagonal order within the PS matrix due to the effect of confinement imparted by the film thickness along the substrate normal. In what follows, for brevity we use S, C_⊥, and C_‖ to denote truncated core–shell nanospheres, upstanding and lying nanocylinders, respectively.

Fig. 2 (a and d) AFM, (b and e) 2D GISAXS and (c and f) corresponding 1D in-plane GISAXS profile with intensity distribution versus q_‖ for vertically- (a–c) and parallel-aligned (d–f) nanocylinders within the P(S₅₀₀₀₀2VP₁₆₅₀₀₀) film.

Fig. 3 shows the AFM, SEM and GISAXS results of carbon nanostructures fabricated from pyrolysis of the specimen of Fig. 1 after UVIN treatment of 6 h. Our previous work has demonstrated that in the absence of oxygen molecules, UV irradiation could only bring about cross-linking of polymeric chains.⁴⁹ 6 h of UVIN was found to effectively stabilize the morphology as well as spatial order of nanodomains of monolayered thickness during thermal pyrolysis at 430 °C (see Fig. S2 and S3 of ESI†). As shown in Fig. 3a and b, after UVIN of 6 h followed by 430 °C-pyrolysis, the morphology of spherical nanodomains with short-range order was retained. The thickness of the nanostructured carbon film was approximately 45 nm. As the GISAXS data show, the Bragg diffraction rods and fringes shift to high q_‖ values (Fig. 3c and d). The downward shifts indicate that pyrolysis led to a decrease in inter-domain distance and dimensional shrinkage of carbon nanostructures. In addition, there is additional isotropic scattering, enhancing the intensities of the diffraction rods, fringes and the stripe. The high q regime of the isotropic scattering displays power-law scattering of q^−3.34, which corresponds to the asymptotic scattering behavior of mesopores. The intensity decay with the exponent of D_f = −3.34 suggests an interconnected mesoporous framework formed by the aggregation of carbon nanostructures. The growth of mesoporous structures resulted from shrinkage, aggregation and/or inter-domain fusion of carbon nanostructure.²⁵ Zhong et al. reported CTNCs fabricated from pyrolysis of polyacrylonitril-block-poly(n-butyl acrylate), PAN-b-PBA, for which PAN serves as a nitrogen-enriched nanocarbon precursor and PBA acts as a sacrificial porogen.²⁵ Thus, complete removal of the PBA block would yield mesopores within the PAN-based carbons. Unlike the PAN-b-PBA system, both PS and P2VP blocks were cross-linked by UVIN. None of the two blocks could be completely decomposed during thermal pyrolysis at 430 °C. Instead, thermal pyrolysis caused domain shrinkage, formation of volatile molecules through partial chain scission during thermal pyrolysis, and inter-domain fusion of spherical nanodomains. This process led to an interconnected 3D network structure and brought about the growth of mesoporous channels within the network structure. Thus, the density contrast between micropores and the interconnected network enhanced the scattering intensity.

Fig. 3 (a) A 2 × 2 μm AFM topographic image, (b) SEM image, (c) GISAXS 2D pattern, and (d) 1D in-plane profile of carbon nanostructures fabricated from 430 °C-pyrolysis of the UVIN-treated P(S₄₈₅₀₀2VP₇₀₀₀₀) film with S nanodomains.

Graphitization of carbon upon pyrolysis of S nanodomains can be studied with Raman scattering spectra. Fig. 4 shows Raman spectral profiles of the granular carbon nanostructures. These spectral results reveal two characteristic Raman shifts centered at 1355 and 1587 cm⁻¹. The Raman shift at 1587 cm⁻¹ is called the graphite mode (G band) and can be taken to indicate the formation of graphitic species as this Raman shift corresponds to the E_2g mode. The Raman shift at 1355 cm⁻¹ is called the defect A_1g mode (D band) and is ascribed to the formation of disordered species.^6,17,19 The presence of these two bands provides clear evidence for the growth of granular nanocarbon resulting from pyrolysis of UVIN-treated S nanodomains at 430 °C.

Fig. 4 Raman spectra of granular carbon nanostructures fabricated from a 430 °C-pyrolyzed P(S₄₈₅₀₀2VP₇₀₀₀₀) film.

Fig. 5 shows the AFM, SEM and GISAXS results of carbon nanostructures fabricated from pyrolysis of prolonged UVIN-treated P(S₅₀₀₀₀2VP₁₆₅₀₀) films. Before structural characterization, the P(S₅₀₀₀₀2VP₁₆₅₀₀) specimens containing C_⊥ or C_‖ nanodomains were subjected to UVIN to stabilize the nanodomains and pyrolysis at 430 °C for 1 h. After pyrolysis, the thickness of the films decreased from 90 nm to 30 nm. As Fig. 5a, b, e and f show, the surface morphology of C_⊥ and C_‖ nanodomains within the P(S₅₀₀₀₀2VP₁₆₅₀₀) films with thermal pyrolysis at 430 °C respectively shows pillar-like carbon nanostructures and streak-like carbons. As the corresponding 2D GISAXS patterns and 1D profiles show, the pillar-like nanocarbon displays a feature of hexagonal order with sharp diffraction-truncated rods at q_‖ ratios 1 : 3^1/2 : 4^1/2 with respect to the position of the principal rod while the feature of inter-domain correlation order with q_‖ ratio 1 : 4^1/2 is clearly discerned for the streak-like nanocarbon (Fig. 5c, d, g and h). The presence of D and G features shown in the Raman spectra of Fig. 5i provides direct evidence for the formation of carbon nanostructures with a graphite domain of varied degree.

Fig. 5 (a and e) 1 × 1 μm² AFM topographies, (b and f) SEM images, (c and g) GISAXS 2D patterns, (d and h) 1D in-plane profiles and (i) Raman spectra of carbon nanostructures fabricated from 430 °C-pyrolysis of prolonged UVIN-treated P(S₅₀₀₀₀2VP₁₆₅₀₀) films, each of which contains C_⊥ or C_‖ nanodomains.

Pyrolysis at 430 °C produced a morphological conversion for C_⊥ nanodomains. Before pyrolysis, the position of each P2VP nanodomain displays a shallow concavity with the free surface being lower than the PS matrix (see Fig. 2a). Pyrolysis at 430 °C led to a reversed height contrast between the P2VP and PS nanodomains. Here we discuss possible mechanisms of this morphological conversion. The formation of the pillar-like carbon nanostructures with convex surfaces is due to surface reconstruction. PS (γ = 41.5 mN m⁻²) has been shown to have a smaller surface energy than P2VP (γ = 46.7 mN m⁻²).⁵⁰ We speculate that the carbonaceous species resulting from pyrolysis of the PS chains possesses a surface energy smaller than those from pyrolysis of the P2VP chains. Furthermore, after pyrolysis at 430 °C, the cross-linking density of the P(S₅₀₀₀₀2VP₁₆₅₀₀) film would be decreased as a partial chain scission of polymer chains occurred to form volatile species. As volatile species increasingly developed, the mechanical strength of the cross-linked film decreased. Once the driving force for migration of the PS block onto the P2VP block exceeded the material strength, a morphological conversion occurred via surface reconstruction. Such a conversion can decrease the overall free energy of carbonaceous species, producing pillar-like carbon nanostructures with convex surfaces. For these convex nanostructures, the inner part would be enriched with nitrogen atoms.

To identify whether nitrogen atoms enrich the inner part of the carbon nanostructures, we conducted characterization with X-ray photoelectron spectra to analyze quantitatively the spatial distribution of nitrogen and carbon. Because XPS provide information only about chemical elements in material surfaces, we used an ion gun to sputter the specimen surface for various durations to reveal an elemental depth profile. For a direct comparison, the depth profile of a pristine P(S₅₀₀₀₀2VP₁₆₅₀₀) film was also monitored with XPS. Fig. 6 shows the C1s and N1s spectra of a P(S₅₀₀₀₀2VP₁₆₅₀₀) film with C_⊥ nanodomains before pyrolysis. Before ion bombardment, two lines, at 284.6 and 399.3 eV, are discernible (see Fig. 6a and b). The former line originates from a superimposition of three contributions: one is from hydrocarbons at 285 eV, another is from aromatic carbons at 284 eV and a third is from C–N components at 286.2 eV (see Fig. S4a†).^51–53 Beside the line at 284.6 eV, there is a shoulder at 291.5 eV, which corresponds to the π to π* shake-up line of phenyl groups in PS. The line at 399.3 eV corresponds to the pyridine nitrogen (Fig. S4b†).^54,55 Upon ion bombardment, the line at 284.6 eV shows a slightly increased intensity whereas the intensity of the line at 399.3 eV was depressed. Fig. 6c shows the ratio of N to C as a function of duration of ion bombardment; before bombardment, the fraction of nitrogen is greatest. After ion bombardment for 10 s, the N/C ratio rapidly decreased from 0.067 to 0.026 and then remained stable in a range 0.024–0.034. This result indicates that P2VP chains were preferentially located at the surface of the film. Although the exact mechanism is not entirely clear, we propose a possible mechanism according to early literature reported by Kramer's group who explained the system of PEO-based BCP with solvent annealing under great humidity.⁵⁶ As mentioned above, the surface energy of PS is less than that of P2VP. From a thermodynamic view of point, PS is expected to predominate in the wetting of the free surface; the surface wetting by P2VP is thus a kinetically trapped and metastable state upon drying. We ascribe this state to condensation of water onto block copolymer containing P2VP. After SVA and drying, the rapid removal of solvent vapor results in the cooling of the free surface of a thin film. Once the temperature of the free surface attains a dew point, moisture begins to condense to form a thin layer of dew on the free surface. The water layer attracts hydrophilic P2VP chains, leading to P2VP block domains swollen with water at the surface. As trichloromethane is a neutral solvent, the difference in surface energy between the two blocks becomes decreased in the presence of trichloromethane at the stage of SVA; the free surface could be neutrally wetted with both PS and P2VP, but, at the stage of drying, P2VP becomes dominant over PS to wet the free surface of thin films because water is a selective solvent. Furthermore, as water is much less volatile than trichloromethane, water leaves last from the film, making some chains trapped at the free surface. The indentations within the center of the P2VP domains are ascribed to a mechanism whereby particular P2VP chains are trapped at the free surface (see Fig. 2a).

Fig. 6 XPS (a) carbon 1s, (b) nitrogen 1s spectra and N/C ratio versus ion bombardment time of a pristine P(S₅₀₀₀₀2VP₁₆₅₀₀) film with C_⊥ nanodomains.

Fig. 7 shows C1s and N1s spectra of a P(S₅₀₀₀₀2VP₁₆₅₀₀) film with C_⊥ nanodomains after pyrolysis. A quantitatively different trend, showing an increased fraction of N with ion bombardment, is present in the XPS spectra of the pillar nanocarbon. Without ion bombardment, the N/C ratio was only 0.036; after sputtering over the surface of the pillar nanocarbons, the N/C ratio increased to approximately 0.061. This result provides clear evidence for an interpretation that the inner part of carbon became enriched with nitrogen through surface reconstruction. The formation of convex nanocarbon is a thermodynamically driven morphology. Through graphitization during pyrolysis at 430 °C, the content of trigonally bonded carbons (284 eV) became a major proportion, and tetrahedral carbons and hydrocarbons (at 285 eV) were present as minor proportions (see Fig. S4c†). Furthermore, most pyridine rings were transformed to form nitrogen-containing groups of other types such as pyrrolic nitrogen (400.3 eV) and quaternary nitrogen (401.3 eV) (see Fig. S4d†).

Fig. 7 XPS (a) carbon 1s, (b) nitrogen 1s spectra and N/C ratio versus ion bombardment time of pillar-like nanocarbons fabricated from 430 °C-pyrolysis of a P(S₅₀₀₀₀2VP₁₆₅₀₀) film with C_⊥ nanodomains.

Effect of pyrolysis temperature

Fig. S5† shows AFM height images of the specimens after UVIN and pyrolysis at elevated temperatures – 600, 750 and 900 °C; the materials were almost entirely degraded on pyrolysis at these elevated temperatures. The dependence on temperature of loss of mass for hPS₄₅₀₀₀, hP2VP₁₅₂₀₀₀ and P(S₅₀₀₀₀2VP₁₆₅₀₀) bulk samples acquired by TGA measurements provides a hint (Fig. S6 and S8†). A loss of mass is taken to indicate the formation of volatile carbonaceous species. As Fig. S6† shows, T_d of both hPS₄₅₀₀₀ and hP2VP₁₅₉₀₀₀ specimens are above 400 °C; T_d of P2VP ranged from approximately 380 (denoted as Tⁱ_P2VP,d) to 440 °C (denoted as T^f_P2VP,d), whereas T_d of PS is in a range 420 (Tⁱ_PS,d) to 460 °C (T^f_PS,d). Upon heating, the decomposition of a P(S₅₀₀₀₀2VP₁₆₅₀₀) specimen began at 397 °C and terminated at 492 °C, regardless of a negligible difference in T_d for the two constituting blocks (Fig. 8). The thermal decomposition of PS, P2VP and PS-b-P2VP specimens yielded only small proportions of loss of mass; all were less than 1 percent of their original mass when the temperature rose over T^f_d. Furthermore, prolonged UVIN treatment increased the decomposition temperature of the specimens only slightly, and did not effectively improve the final yield of carbonaceous residue when the temperature exceeded T^f_d. This result indicates that, above T^f_d, all specimens predominantly decomposed to form gaseous hydrocarbons.

Fig. 8 TGA curves of the P(S₅₀₀₀₀2VP₁₆₅₀₀) bulk specimen.

We used Raman spectra to determine the carbonaceous residues left on the substrate after pyrolysis of hPS₄₅₀₀₀ and hP2VP_1520000 bulks, by monitoring the signals of the D and G bands (Fig. S7†). For hPS₄₅₀₀₀ and hP2VP₁₅₂₀₀₀, when pyrolysis was performed at temperatures above Tⁱ_d, the signals of both D and G bands are discernible, whereas no signals corresponding to these two bands appeared when the temperature was less than Tⁱ_d. This result indicates that carbon indeed grew when the temperature rose above Tⁱ_d. The intensities of the D and G bands nevertheless gradually decreased with increasing temperature of pyrolysis, indicating that the yield of carbonaceous residue left on the substrate gradually decreased with increasing temperature of pyrolysis. The degree of crystallinity of carbon nanostructures, determined through the ratio I_D/I_G, also decreased with increasing temperature of pyrolysis (see Fig. S8†).

The AFM, TGA and Raman results indicate that the extent of cross-linking produced with UVIN might not be great enough to improve completely the retention of shape, morphological fidelity and dimensional stability of nanostructured carbon during pyrolysis, in particular when the temperature of pyrolysis is greater than T^f_d of polymeric materials. Thermal degradation is well known to produce gaseous hydrocarbons through scission of polymeric chains, whereas carbonization converts organic substances into carbonaceous species through cross-linking.⁵⁷ When the temperature of pyrolysis is increased, chain scission reactions are dominant to form small molecular hydrocarbons. Most hydrocarbons are volatile products and evaporate readily from the substrate during pyrolysis. If volatile compounds cannot re-adsorb physically or bond chemically on the substrate, the areal yield of pyrolytic products left on the substrate is small after pyrolysis. The same effect explains also the inverse dependence of carbon crystallinity on the temperature of pyrolysis. Such problems of small carbonaceous residue and small degree of graphitization exist even though polymeric materials are intensively cross-linked with greatly prolonged UVIN. The reason is that UVIN imposed on thin films imparts only a trapping barrier to retard kinetically the diffusion of gaseous species that are generated via chain scission during pyrolysis at elevated temperatures. When the extent of chain scission increases, the mechanical strength of the cross-linked thin film decreases. As a result, the gaseous species, once formed, become no longer kinetically trapped within the cross-linked film. These results indicate that the keys for increased carbonaceous residue and improved morphological fidelity involve retarding diffusion and increasing the probability of recombination of gaseous species.

To increase the carbonaceous residue of a solid and to retain the morphological fidelity after pyrolysis at elevated temperatures, we deposited a capping layer of SiO₂, thickness 40 or 100 nm, atop thin films before pyrolysis. The capping layer of SiO₂ exhibits a uniform thickness regardless of the thickness (see Fig. S9a and c†). The nanostructured surface associated with PS-b-P2VP nanodomains was totally absent due to the covering of SiO₂ (Fig. S9b and d†). Such a capping layer has two effects – to prevent morphological conversion and to increase the carbonaceous residue. After pyrolysis, the specimens were dipped in HF to remove the capping layer and then measured with an AFM to show the surface morphology. As SiO₂ is highly soluble in HF but carbonaceous residues are insoluble in HF, removing the capping layer is readily effected without damaging the surface morphology of nanostructured carbons. As Fig. 9a shows, after removal of the capping layer followed by AFM measurements, the surface of the P(S₅₀₀₀₀2VP₁₆₅₀₀) film pyrolyzed at 430 °C initially having C_⊥ nanodomains revealed the same morphology of hexagonally packed C_⊥ nanodomains within the matrix. Nevertheless, for complete preservation of the morphology of nanocarbons fabricated from P(S₅₀₀₀₀2VP₁₆₅₀₀) films pyrolyzed at 900 °C, the capping layer should be thicker than 100 nm (compare Fig. 9b and c). The reason is that pyrolysis at such a high temperature inevitably caused cracks in the surface of the capping layer (Fig. 9d and e). The presence of these cracks would deteriorate a barrier to mass transport because the volatile hydrocarbons become able to diffuse through the cracks to the air above. Increasing the thickness of the capping layer can prevent the depth of cracks induced upon pyrolysis from reaching the thin films. A thick layer is thus still effective to retain the morphological fidelity of nanocarbon.

Fig. 9 AFM topographic images of carbon nanostructures with (a) 430 °C- and (b and c) 900 °C-pyrolysis of C_⊥ nanodomains within the P(S₅₀₀₀₀2VP₁₆₅₀₀₀) thin films with a capping layer of SiO₂, thickness (a and b) 40 or (c) 100 nm. (d and e) OM images of sample-b and -c without removal of the capping layer.

This result demonstrates that capping a protective layer on thin films acts as a superior barrier to mass transport of gaseous hydrocarbons generated through chain scission, and that this capping layer thus increases the probability of recombination reactions of volatile products during pyrolysis. Furthermore, P2VP with pyridine rings has an interaction with the hydrophilic surface of SiO₂ with hydroxyl groups, making P2VP chains reside firmly at the free surface in the presence of the capping layer during pyrolysis. Both factors account for the improved areal yields and morphological fidelity of carbon nanostructures.

Conclusions

In summary, we fabricated various nanocarbon materials with perfect morphological retention and a large carbonaceous residue directly from pyrolysis of S, C_⊥ and C_‖ PS-b-P2VP nanodomains. UVIN cross-linking and capping a SiO₂ layer on top of the cross-linked films proved effective in combination to retain effectively the nanodomains within thin films when pyrolysis was performed at temperatures above the decomposition temperatures of the PS and P2VP blocks. The keys of improved morphological retention are to increase the carbonaceous residue and to tailor preferential interactions at the free surface of thin films. The capping layer has two effects, acting as a barrier for mass transport through diffusion of volatile hydrocarbons generated through chain scission during pyrolysis, and preventing thermodynamically driven morphological conversion.

Acknowledgements

Ministry of Science and Technology provided supports (MOST 103-2221-E-008-084-MY3).

References

1. L. Dai, Acc. Chem. Res., 2013, 46, 31–42.
2. K. Gong, F. Du, Z. Xia, M. Durstock and L. Dai, Science, 2009, 323, 760–764.
3. Y. Li, Y. Zhao, H. Cheng, Y. Hu, G. Shi, L. Dai and L. Qu, J. Am. Chem. Soc., 2012, 134, 15–18.
4. L. T. Qu, Y. Liu, J. B. Baek and L. Dai, ACS Nano, 2010, 4, 1321–1326.
5. S. Valkama, A. Nykanen, H. Kosonen, R. Ramani, F. Tuomisto, P. Engelhardt, G. Brinke, O. Ikkala and J. Ruokolainen, Adv. Funct. Mater., 2007, 17, 183.
6. C. D. Liang, K. L. Hong, G. A. Guiochon, J. W. Mays and S. Dai, Angew. Chem., Int. Ed., 2004, 43, 5785.
7. Y. Wan, Y. Shi and D. Zhao, Chem. Mater., 2008, 20, 932–945.
8. J. Schuster, R. Köhn, M. Döblinger, A. Keilbach, H. Amenitsch and T. J. Bein, J. Am. Chem. Soc., 2012, 134, 11136–11145.
9. J. Xue, G. Singh, Z. Qiang, A. Karim and B. D. Vogt, Nanoscale, 2013, 5, 7928–7935.
10. J. Xue, C. Henry, J. Lee and B. D. Vogt, RSC Adv., 2014, 4, 3675–3683.
11. L. Chuenchom, R. Kraehnert and B. N. Smarsly, Soft Matter, 2012, 8, 10801–10812.
12. Y. Meng, D. Gu, F. Q. Zhang, Y. F. Shi, L. Cheng, D. Feng, Z. X. Wu, Z. X. Chen, Y. Wan, A. Stein and D. Y. Zhao, Chem. Mater., 2006, 18, 4447–4464.
13. S. Tanaka, N. Nishiyama, Y. Egashira and K. Ueyama, Chem. Commun., 2005, 16, 2125–2127.
14. Y. H. Deng, T. Yu, Y. Wan, Y. F. Shi, Y. Meng, D. Gu, L. J. Zhang, Y. Huang, C. Liu, X. J. Wu and D. Y. Zhao, J. Am. Chem. Soc., 2007, 129, 1690–1697.
15. J. Y. Zhang, Y. H. Deng, J. Wei, Z. K. Sunm, D. Gu, H. Bongard, C. Liu, H. H. Wu, B. Tu, F. Schüth and D. Y. Zhao, Chem. Mater., 2009, 21, 3996–4005.
16. Y. H. Deng, J. Liu, C. Liu, D. Gu, Z. K. Sun, J. Wei, J. Y. Zhang, L. J. Zhang, B. Tu and D. Y. Zhao, Chem. Mater., 2008, 20, 7281–7286.
17. C. B. Tang, K. Qi, K. L. Wooley, K. Matyjaszewski and T. Kowalewski, Angew. Chem., Int. Ed., 2004, 43, 2783–2787.
18. C. Tang, A. Tracz, M. Kruk, R. Zhang, D.-M. Smilgies, K. Matyjaszewski and T. Kowalewski, J. Am. Chem. Soc., 2005, 127, 6918–6919.
19. T. Kowalewski, N. V. Tsarevsky and K. Matyjaszewski, J. Am. Chem. Soc., 2002, 124, 10632–10633.
20. J. M. Leiston-Belanger, H. Penelle and T. P. Russell, Macromolecules, 2006, 39, 1766–1770.
21. R. M. Ho, T. C. Wang, C. C. Lin and T. L. Yu, Macromolecules, 2007, 40, 2814–2821.
22. J. Huang, C. Tang, H. Lee, T. Kowalewski and K. Matyjaszewski, Macromol. Chem. Phys., 2007, 208, 2312–2320.
23. Y. Wang, J. Q. Liu, S. Christiansen, D. H. Kim, U. Goesele and M. Steinhart, Nano Lett., 2008, 8, 3993–3997.
24. D. A. Rider, K. Liu, J. C. Eloi, L. Vanderark, L. Yang, J. Y. Wang, D. Grozea, Z. H. Lu, T. P. Russell and I. Manners, ACS Nano, 2008, 2, 263–270.
25. M. Zhong, E. K. Kim, J. P. McGann, S.-E. Chun, J. F. Whitacre, M. Jaroniec, K. Matyjaszewski and T. Kowalewski, J. Am. Chem. Soc., 2012, 134, 14846–14857.
26. J. P. McGann, M. Zhong, E. Y. Kim, S. Natesakhawat, M. Jaroniec and J. F. Whitacre, Macromol. Chem. Phys., 2012, 213, 1078–1090.
27. M. Zhong, S. Jiang, Y. Tang, E. Gottlied, E. K. Kim, A. Star, K. Matyjaszewski and T. Kowalewski, Chem. Sci., 2014, 5, 3315–3319.
28. X. Wang, J. S. Lee, Q. Zhu, J. Liu, Y. Wang and S. Dai, Chem. Mater., 2010, 22, 2178.
29. M. Aizawa and J. M. Buriak, J. Am. Chem. Soc., 2005, 127, 8932–8933.
30. M. Aizawa and J. M. Buriak, Chem. Mater., 2007, 19, 5090–5101.
31. Y. Kim, H. Han, Y. Kim, W. Lee, M. Alexe, S. Baik and J. K. Kim, Nano Lett., 2010, 10, 2141–2146.
32. S.-H. Yun, S. I. Yoo, J. C. Jung, W.-C. Zin and B.-H. Sohn, Chem. Mater., 2006, 18, 5646–5648.
33. H.-G. Boyen, G. Kästle, K. Zürn, T. Herzog, F. Weigl, P. Ziemann, O. Mayer, C. Jerome, M. Möller, J. P. Spatz, M. G. Garnier and P. A. Oelhafen, Adv. Funct. Mater., 2003, 13, 359–364.
34. J. P. Spatz, V. Z.-H. Chan, S. Mömer, F. M. Kamm, A. Plettl, P. Ziemann and M. Möller, Adv. Mater., 2002, 14, 1827–1832.
35. T. Lohmueller, E. Bock and J. P. Spatz, Adv. Mater., 2008, 20, 2297–2302.
36. S. Förster and M. Antonietti, Adv. Mater., 1998, 10, 195–217.
37. S. Förster and T. Plantenberg, Angew. Chem., Int. Ed., 2002, 41, 688–714.
38. X. Gu, I. Gunkel, A. Hexemer, W. Gu and T. P. Russell, Adv. Mater., 2014, 26, 273–281.
39. S. Park, D. H. Lee, J. Xu, B. Kim, S. W. Hong, U. Jeong, T. Xu and T. P. Russell, Science, 2009, 323, 1030–1033.
40. Z. Sun, Z. Chen, W. Zhang, J. Choi, C. Huang, G. Jeong, E. B. Coughlin, Y. Hsu, X. M. Yang, K. Y. Lee, D. S. Kuo, S. Xiao and T. P. Russell, Adv. Mater., 2015, 27, 4364–4370.
41. J. Yin, X. Yao, J. Y. Liou, Z. Wang, Y. S. Sun and Y. Wang, ACS Nano, 2013, 7, 9961–9974.
42. C. Y. Chang, P. J. Wu and Y. S. Sun, Soft Matter, 2011, 7, 9140–9147.
43. C. Y. Chang, Y. C. Lee, P. J. Wu, J. Y. Liou, Y. S. Sun and B. T. Ko, Langmuir, 2011, 27, 14545–14553.
44. C. Stegelmeier, A. Exner, S. Hauschild, V. Filix, J. Perlich, S. V. Roth, V. Abetx and S. Förster, Macromolecules, 2015, 48, 1524–1530.
45. Y. H. Jang, S. T. Kochuveedu, Y. J. Jang, H. Y. Shin, S. Yoon, M. Steinhart and D. H. Kim, Carbon, 2011, 49, 2120–2126.
46. Y. J. Jang, Y. H. Jang, S.-B. Han, D. Khatua, C. Hess, H. Ahn, D. Y. Ryu, K. Shin, K. W. Park, M. Steinhart and D. H. Kim, ACS Nano, 2013, 7, 1573–1582.
47. A. F. M. Barton, Chem. Rev., 1975, 75, 731–753.
48. J. Y. Liou and Y. S. Sun, Macromolecules, 2012, 45, 1963–1971.
49. J. Y. Liou and Y. S. Sun, Soft Matter, 2015, 11, 7119–7129.
50. B. B. Sauer and G. T. Dee, Macromolecules, 2002, 35, 7024–7030.
51. M. M. Browne, G. V. Lubarsky, M. R. Davidson and R. H. Bradley, Surf. Sci., 2004, 553, 155–167.
52. D. O. H. Teare, N. Emmison, C. Ton-That and R. H. Bradley, Langmuir, 2000, 16, 2818–2824.
53. C. Ton-That, D. O. H. Teare, P. A. Cambell and R. H. Bradley, Surf. Sci., 1999, 433–435, 278–282.
54. K. Stańczyk, R. Dziembaj, Z. Piwowarska and S. Witkowski, Carbon, 1995, 33, 1383–1392.
55. J. R. Pels, F. Kapteijn, J. A. Moulijn, Q. Zhu and K. M. Thomas, Carbon, 1995, 33, 1641–1653.
56. C. Tang, J. Bang, G. E. Stein, G. H. Fredrickson, C. J. Hawker, E. J. Kramer, M. Sprung and J. Wang, Macromolecules, 2008, 41, 4328–4339.
57. M. P. Stevens, Polymer Chemistry, Oxford University Press, 2nd edn, 1990, vol. 9, pp. 296–321.

---

Footnote

† Electronic supplementary information (ESI) available: AFM, GISAXS 2D pattern and 1D in-plane profile of an as-spun P(S₅₀₀₀₀2VP₁₆₅₀₀₀) film (Fig. S1); AFM and GISAXS characterizations of stabilization of nanodomains by UVIN for pyrolyzed carbonaceous nanostructures (Fig. S2 and S3); XPS C1s and N1s spectra of a P(S₅₀₀₀₀2VP₁₆₅₀₀) film with C_⊥ nanodomains before and after pyrolysis (Fig. S4); AFM height images of P(S₅₀₀₀₀2VP₁₆₅₀₀) thin specimens after pyrolysis at various elevated temperatures (Fig. S5); TGA profiles of pristine polymer bulk prepared from hPS and hP2VP (Fig. S6); Raman spectra of carbons obtained through pyrolysis of hP2VP₁₅₂₀₀₀ and hPS₄₅₀₀₀ at elevated temperatures in the range of 350–800 °C (Fig. S7); I_D/I_G ratio and in-plane crystal size of hPS₄₅₀₀₀ and hP2VP₁₅₂₀₀₀ thin films at different carbonized temperature (Fig. S8); SEM characterization of the SiO₂ capping layers (Fig. S9). See DOI: 10.1039/c5ra17500c

---