Introduction

The search for a comprehensive understanding of the physisorption and chemisorption behaviors of small molecules, oligomers and polymers at various interfaces (including liquid–solid, liquid–gas and solid–gas) is one of the most important issues in surface and materials science.1, 2, 3, 4 However, real-time visualization with direct detection of physisorption and chemisorption behaviors by the adsorbate at these interfaces remains a substantial challenge since the pioneering work in this area began in the 1910s.5, 6, 7, 8

In general, chemisorption indicates irreversible immobilization of the adsorbate on the surface by covalent connections with strong chemical bonds.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22 Indeed, the most common chemisorption processes rely on (i) a grafting-to approach of organic compounds and polymers with the aid of functional groups (e.g., Si–X 9, 10, 11, 12 (X=Cl, OR, and H) and S-H20, 21, 22) and (ii) a grafting-from approach using a catalyst-loaded surface by applying conventional13 and controlled radical polymerization (CRP) techniques14, 15, 16, 17, 18 (i.e., ATRP (atom transfer radical polymerization), RAFT (reversible addition-fragmentation chain transfer polymerization) and NMP (nitroxide-mediated polymerization)).

However, physisorption typically indicates the reversible adsorption and desorption behavior of the adsorbate at the surface that results from non-covalent weak forces.23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33 The van der Waals (vdW) force, which is one of the weakest but ubiquitous forces, is responsible for physisorption and the initial process of chemisorption,2, 3, 4 as well as template polymerization for the production of it- and st-PMMA stereocomplexes.34 The vdW force is recognized as the sum of the attractive and repulsive forces between forms of matter. The attractive forces originate from the instantaneous fluctuation (London dispersion) of electron clouds within matter, and the repulsive forces occur at a short distance caused by the overlap between wavefunctions of electrons owing to the Pauli exclusion principle.32, 33, 34, 35

According to several textbooks by Israelachvili4 and Butt, Graf, and Kappl,3 the attractive vdW force (F) between two particles with an interdistance r is proportional to r–7 for two atoms, r–6 for two molecules, r–3−r–4 for metallic plates (Casimir force) and r–2 between μm-size solid spheres, as well as between random-coil macromolecules. However, the r dependence between solid spheres and macromolecules has not been fully elucidated regardless of the flexible-, stiff- and rigid-rod chains. The r–2 dependence of the vdW force is identical to that of Coulombic and gravitation forces.4 Indeed, the Coulombic force is widely applied to layer-by-layer (LbL) deposition using anionic and cationic polymeric electrolytes.23, 24, 25, 26, 27, 28, 29 This LbL method facilitates irreversible solution-phase physisorption to build-up polymeric adsorbates from the surface.

Unanswered questions remain, such as (i) whether the invisible attractive vdW force exists in heterogeneous fluids between neutral sphere-like soft matter (i.e., a random coiled polymer in a fluidic solution) and a neutral hard sphere (i.e., a ceramic solid particle) and (ii) whether this phenomenon is detectable with a visible method. Other unanswered questions include (iii) whether physisorption is influenced by the nature of the solvent molecule in suspension, (iv) whether the dominant driving factor is an entropic or enthalpic effect in the physisorption of chain-like polymers, (v) determination of the major differences between polymers and small molecules, and more importantly, (vi) the demarcation of reversible physisorption and irreversible chemisorption.

In this study, we assume the existence of an attractive vdW force between non-charged (neutral) highly polarizable polymers and neutral polarizable inorganic spheres at the liquid–solid interface. The vdW force between the small molecules and the surface might be on the order of 1–2 kcal mol–1. However, when a chain-like polymer is employed, this weak vdW force should be compatible with irreversible covalent chemical bonds (e.g., the bond dissociation energy of a C–C linkage is ≈90 kcal mol–1 if a number of repeating units are added together). Indeed, in a series of hydrocarbons and alkanethiols on a clean gold surface, the physisorption enthalpy linearly increases with increasing methylene numbers. However, the chemisorption enthalpy is a constant and is independent of the methylene numbers.20 A recent experiment has confirmed that in random-coil polymeric electrolytes, the entropy effect due to release of solvated molecules and paired ionic species is responsible for the LbL association.31

Herein, we designed a hybridized photoluminescent core-shell particle by applying solution-phase physisorption based on the attractive vdW hypothesis. The particle is composed of a spherical silica (the core as a model of a neutral hard sphere) and a stiff π-conjugative poly(di-n-octylfluorene; PF8, a shelled adsorbate as a model of a neutral soft sphere). The PF8 is a polymeric probe that exhibits an intense chromophore in the UV-visible region and a highly efficient fluorophore in the visible region with a quantum yield of ~0.90.35 The spherical silica particles have a large specific surface area without detectable Si-OH groups and the ability to detect the adsorbate compared with a solid with a flat surface.

In this study, we determined that PF8 spontaneously physisorbs onto μm-sized spherical silica in suspension in toluene within an hour at room temperature. This result was confirmed by photoluminescence (PL)/PL excitation/UV-visible spectroscopy, as well as observation by the naked eye and in photographs. The level-off adsorption amount of PF8 was controlled by the choices of dispersion solvent, initial concentration of PF8 and physisorption time. Similarly, two non-charged photoluminescent materials (i.e., poly[(9,9-dioctylfluorenyl-2,7-diyl)-alt-bithiophene], PF8T2) and poly[2-methoxy-5-(2-ethyl-hexyloxy)-1,4-phenylenevinylene], MEH-PPV; Figure 1)) spontaneously physisorbed onto the surfaces of several μm-sized ceramic particles (SiO2, Al2O3, ZnO and TiO2 (anatase and rutile)) in chloroform within several hours. These results were confirmed by PL spectra, as well as observation by the naked eye and in photographic images. Our experimental results indicate the feasibility of irreversible and spontaneous physisorption of a wide range of non-charged π-conjugated polymers onto several inorganic solid surfaces in suspension conditions with catalyst-free coating processing at room temperature.

Figure 1
figure 1

Chemical structures of PF8 (Mw=7.5 × 104, Mn=1.7 × 104, Mw/Mn=4.4), PF8T2 (Mw=2.3 × 104, Mn=9.0 × 103, Mw/Mn=2.6) and MEH-PPV (Mw=3.1 × 105, Mn=6.8 × 104, Mw/Mn=4.5).

Full size image

Materials and methods

Materials

PF8 (Mw=7.5 × 104, Mn=1.7 × 104, Mw/Mn=4.4), PF8T2 (Mw=2.3 × 104, Mn=9.0 × 103, Mw/Mn=2.6) and MEH-PPV (Mw=3.1 × 105, Mn=6.8 × 104, Mw/Mn=4.5) were purchased from Sigma-Aldrich (St Louis, MO, USA) and used as received (Figure 1). The spherical silica (0.5±0.1 μm in diameter (Supplementary Information (SI) Figure S1a), specific surface area=9.5±3.5 m2 g–1, no detectable silanol according to fourier transform infrared spectroscopy) was supplied by Toa-Gosei (Tokyo, Japan). Spherical particles (5-μm in diameter) of ZnO and TiO2 (anatase and rutile) were purchased from Wako Pure Chemicals (Osaka, Japan). Spherical Al2O3 particles (MA800 10-μm in diameter) were donated by the company Micron (Himeji, Hyogo, Japan), a Nippon Steel and Sumikin Materials Group (Tokyo, Japan). The specific surface areas of TiO2 (anatase and rutile) were 50 and 7 m2 g–1, respectively. However, the specific surface areas of Al2O3 and ZnO were not disclosed.

Method

The schematic outline of the solution-phase physisorption of PF8 followed by the work-up process is shown in Figure 2. The quantity of the solvent for the physisorption experiment was fixed at 5 ml, and the [PF8]0 ranged from 1.0 × 10–2 m to 1.0 × 10–5 m. The quantity of 0.5-μm silica particles was fixed at 250 mg, corresponding to a surface area of 2.38 m2. The weighed 0.5-μm spherical silica was placed in a solution of PF8 followed by gentle stirring with a magnetic stir bar. A homogeneous solution of the non-physisorbed PF8 and the physisorbed PF8 on silica was separated by centrifugation (Kubota, Tokyo, Japan), 3000 r.p.m. for 20 min). After the initial centrifugation, the physisorbed PF8 particles in solvent were repeatedly washed by decantation and centrifugation (Hitachi Koki, Tokyo, Japan), 15 000 r.p.m. for 20 min). Finally, a transparent solution of non-physisorbed PF8 was obtained by filtration with a polytetrafluorethylene syringe filter (Whatmann, GE healthcare, Tokyo, Japan, 0.45-μm pore size). This material was subjected to UV-visible measurements and analysis. The UV-visible spectra of the mixed solutions were recorded simultaneously at 25 °C on JASCO (Hachioji, Tokyo, Japan) V-550 and V-570 spectrophotometers equipped with a Peltier-controlled housing to control the solution temperature. As shown in Figures 2a–d, the changes in absorbance at 415 nm were detectable but subtle. A synthetic quartz (SQ)-grade cuvette was used with path lengths of 1 mm and/or 5 mm and a bandwidth of 2 nm, and the measurement wavelength was set to 415 nm. The PL spectra were recorded at 25 °C on a JASCO FP-6500 spectrofluorometer without a Peltier-controlled housing and with an SQ-grade cuvette. A scanning rate of 100 nm min–1, an excitation bandwidth of 3 nm, a monitoring bandwidth of 3 nm, a response time of 1 s and a normal-mode photomultiplier gain were applied. A field-emission scanning electron microscope ((JEOL, Hachioji, Tokyo, Japan), JSM-7400F, acceleration voltage with 200 kV) was used to view the 0.5-μm spherical silica.

Figure 2
figure 2

Workup processes of the solution-phase physisorption experiment for PF8. Changes in UV-visible absorption spectra at two typical physisorption times (t=0 h (dotted line) and t=24 h (solid lines)) of (a) PF8 (1.0 × 10–2 m) with 0.5-μm silica (250 mg) in 3.0 ml of toluene, (b) their magnified changes in the visible absorption spectra, (c) PF8 (5.0 × 10–5 m) and 0.5-μm silica (250 mg) in 3.0 ml of toluene and (d) their magnified changes in the visible absorption spectra. The physisorption behavior was monitored by the change in absorbance at 415 nm owing to absorption tail of the π−π* transition35 with path lengths of 5.0 and 1.0 mm when [PF8]0 ranges from 5.0 × 10–5 m to 1.0 × 10–2 m.

Full size image

Results and Discussion

Physisorption behavior

The physisorption of PF8 on 0.5-μm spherical silica is discussed. In evaluating the physisorbed PF8 on silica, the Beer–Lambert law was applied to a solution of non-adsorbed PF8. The decrease in π−π* absorption of PF8 at 415 nm35 in the solution indirectly measures the amount of physisorbed PF8 because the physisorbed PF8 on silica does not contribute to the π−π* absorption at 415 nm, as shown in Figures 4a–d. The amount of physisorbed PF8 can be evaluated as follows:

where [Abs]0 and [Abs]t are the initial absorbance of PF8 in solution and the absorbance of PF8 in solution at physisorption time t, respectively. Next,

where [PF8]t and [PF8]0 are the PF8 dissolved in solution because of the physisorbed PF8.

By considering the total volume (5.0 ml) of the suspension solvent, the weight of physisorbed PF8 per surface area of silica particles was evaluated in units of mg m–2.

Figure 3 shows the amount of physisorbed PF8 on 0.5-μm silica in toluene as a function of physisorption time (t) at room temperature (23–25 °C) under gentle stirring with a magnetic stir bar. In this example, [PF8]0=1.0 × 10–2 m (based on the repeating unit of PF8), toluene=5 ml, and 0.5-μm silica=250 mg (2.38 m2 of surface area) were the chosen parameters. The level-off physisorption time under this condition is 3 h, leading to saturated adsorption of PF8 on the silica. This observation suggests the existence of an equilibrium between the adsorption and desorption processes of PF8 at the toluene-silica interface. In addition, the amount of adsorbed PF8 depends on the initial concentration of PF8 in toluene (i.e., if [PF8]0 is <3.0 × 10–3 m), as shown in Figure 4a.

Figure 3
figure 3

(a) Amount of PF8 adsorbed onto 0.5-μm silica in toluene (n=3 as open circles and averaged value as filled square), [PF8]0=1.0 × 10–2 m, monitoring wavelength=415 nm. (b) Fractional surface coverage θ (relative to the level-off amount value at 12 h) as a function of PF8 on the silica surface. The data were taken from a.

Full size image
Figure 4
figure 4

(a) Amount of PF8 adsorbed onto 0.5 μm-silica in toluene (n=3 as open circles and averaged value as filled square) as a function of initial concentration of PF8 [PF8]0 and (b) the log–log plots. The monitoring wavelengths of PF8 were 415 nm. (c) Reciprocal plots of physisorbed PF8 as a function of the initial concentration of PF8 in toluene. The data were taken from the average values in a.

Full size image

Figure 4a shows the amount of physisorbed PF8 on 0.5-μm silica in toluene at room temperature (23–25 °C) as a function of [PF8]0, which ranges from 2 × 10–5 m and 1 × 10–2 m, and the physisorption time was fixed at 24 h. The amount of physisorbed PF8 monotonically increased as [PF8]0 increased from 2 × 10–5 m to 3 × 10–3 m. Figure 4b plots the amount of physisorbed [PF8] and [PF8]0 on log–log scales. More specifically, the amount of physisorbed PF8 nearly linearly increases when [PF8]0 is in the 2.0 × 10–5 m to 3.0 × 10–3 m range.

On the basis of the Langmuir-type isotherm equation, we attempted to analyze the physisorption behaviors of PF8 even though the isotherm denotes a non-specific deposition of adsorbate at the surface. Figure 4c plots the reciprocal value of the physisorbed PF8 on the silica (mg m–2) as a function of [PF8]0–1 (in m–1), which is in line with the Langmuir isotherm because of the adsorption occurring in solution.3, 4 In this plot, W is the amount of physisorbed PF8 on the silica surface, C is [PF8]0 in toluene, Ws is the saturated value of physisorbed PF8, and a is the adsorption equilibrium constant.

The saturated value (Ws) of physisorbed PF8 and the adsorption–desorption equilibrium constant (a) cannot be obtained from a plot of 1/W as a function of 1/C (Figure 4c) owing to the highly non-linear relationship between 1/W and 1/C. In particular, a high deviation from the expected linear plot can be observed for the dilute condition of [PF8]0=2.0 × 10–5 m. When [PF8]0–1 <20 000 m–1, the Langmuir adsorption constant (a) is ~0.25 from the linear region (dotted line). In addition, although the level-off amount of physisorbed PF8 is ≅0.35 mg m–2 (Figure 4a), this value is nearly half of the 0.68 mg m–2 value evaluated from the linear slope of Figure 4c. Therefore, these inconsistencies indicate that the physisorption behavior of PF8 does not simply obey the conventional Langmuir isotherm that was formulated based on the non-specific physisorption hypothesis.

A possible explanation for this observation is discussed below. Because PF8 is a stiff polymer with a shorter persistence length (q) of 8.5–9.5 nm (or Kuhn segment length (lK) 17–19 nm)36, 37, it may adopt a semi-coiled conformation in a fluid solution due to PF8 possessing a Mw=7.5 × 104 (averaged repeating number ≅192, averaged segment number ≅13). The intramolecular excluded-volume effect of a stiff chain is known to be rather minimal when the segment numbers are less than ≅50.36 The nature of the stiff conformation may be crucial for spontaneous production of the β-phase.

Indeed, when the PF8 chains are nonspecifically adsorbed on the silica surface, significant changes in the conformation of PF8 at the liquid–solid interface might occur. The existence of interchain interactions between PF8 chains and intrachain interactions within the PF8 chains may be responsible for the appearance of the crystalline β-phase where the highest content (~53%) of PF8 was observed at the silica surface 35, 38, 39, 40, 41 based on the PL excitation spectrum (see Results and Discussion, later section). The packing entropy of stiff PF8 chains at the liquid–solid interface, which is similar to smectic liquid crystals, might be one of the driving forces of spontaneous physisorption followed by the formation of the highly ordered β-phase.

In an attempt to examine the solvent effect on the physisorption behavior, we chose four common solvents (i.e., chloroform, benzene, toluene and carbon tetrachloride) (physisorption time=24 h, physisorption temperature=23–25 °C, [PF8]0=1.0 × 10–2 m). These solvents have inherent Hansen’s solubility parameters (δ)42, 43 and relative permittivities (ɛ).43 Figures 5a and b plot the level-off amounts of physisorbed PF8 as functions of δ and ɛ of the four solvents.

Figure 5
figure 5

Level-off amount of physisorbed PF8 on 0.5 μm-silica in four organic solvents (n=3 as open circles and averaged value as filled square) as a function of (a) Hansen's solubility parameter and (b) relative permittivity. Physisorption time=24 h, physisorption temperature=23–25 °C, [PF8]0=1.0 × 10–2 m.

Full size image

The amount of physisorbed PF8 in the solvent was interpreted as the difference in δ values between PF8 and the solvent even though the δ value of PF8 has not yet been determined. From Figure 5a, among the four solvents, chloroform imparted poor adsorption ability but carbon tetrachloride imparted the highest adsorption ability. Because chloroform acts as a good solvent, PF8 may display many segmental sequences with a shorter persistence length compared with its behavior in a poor solvent owing to the excluded-volume effect. This effect might lead to a decrease in the adsorption ability of PF8 on the silica surface in the good solvent. Among these solvents, chloroform acts as a good solvent, and carbon tetrachloride acts as a poor solvent. In addition, the suitability of benzene and toluene fall between that of chloroform and carbon tetrachloride. The δ value of PF8, δPF8, was extrapolated to be 9.21–9.25. If a pure solvent with a higher δ value than PF8 is employed, this hypothesis might be verified because the difference between δ for the solvent and δPF8 (i.e., Δ δ) is a critical factor for the degree of chain coiling in solution.

Another explanation may be related to the ɛ value of the solvents. The results in Figure 5b suggest that the amount of physisorbed PF8 tends to increase as the ɛ value decreases. If attractive vdW forces are responsible for the spontaneous physisorption behavior, the physisorbed amount is independent of the ɛ value. However, when the ɛ value of apolar and less-polar solvents partially contributes to the attractive vdW forces between apolar and less-polar PF8 and the polar oxide surface of silica particles, a lower ɛ value might enhance the attractive force that obey ∝–1/ɛ.

Photoluminescent spectra

The PL and PL excitation spectra of the physisorbed PF8 on 0.5-μm silica (referred to as the silica-core/PF8-shell particle) aids in characterizing the conformational and higher-order structures of PF8 at the silica surface. The existence of physisorbed PF8 in silica-core/PF8-shell particles is evident from the bluish PL in photographs and naked-eye observation of these particles placed in Petri dishes and excited by 375 nm UV-LED light sources (Nichia, Tokushima, Japan; Supplementary Figure S1b). Similarly, the bluish PL in photographs of the particles prepared in chloroform ([PF8]0=1.0 × 10–3 m, physisorption time=1 h) can be observed by the naked eye (Supplementary Figure S1c).

Plural PL π−π* transitions at 440.5 nm, 466.3 nm, 498.3 nm and 535.9 nm, which were assigned to the 0–0’, 0–1’, 0–2’ and 0–3’ vibronic bands, respectively, and are characteristic of β-phase PF8, were observed. However, a weak PL band was observed at ~420 nm because the α-phase with a minimal intensity35, 38, 39, 40, 41 when carbon tetrachloride was used as the solvent (Figure 6a). From the PL excitation spectra, the amount of β-phase decreases in the order 56% (carbon tetrachloride), 52% (toluene), 51% (benzene) and 51% (chloroform). This trend may be related to the amount of physisorbed PF8 on silica, as shown in Figure 5. However, these amounts of β-phase (i.e., >51%) are superior to those previously reported for β-phase PF8.35, 38, 39, 40, 41 In particular, the 56% β-phase (in carbon tetrachloride) is a notably high value. In addition, it is important to note that the Stokes' shift of the β-phase at the silica in four solvents is only 4 nm or ≅589 cm–1 (Figure 6b). This small Stokes' shift indicates that the highly rigid β-phase PF8 structure is spontaneously formed on the silica surface. It is possible that a poorer solvent and/or lower relative permittivity provides the rather densely packed β-phase at the silica surface from existing as free PF8 chains rather than the α-phase in the bulk solution. Therefore, the high quantity of β-phase with the small Stokes’ shift was deposited spontaneously at the silica surface without any thermal treatment.

Figure 6
figure 6

(a) PL (dotted curve) spectrum excited at 360 nm and PL excitation (PLE; solid curve) spectrum monitored at 460 nm, (b) magnified PL and PLE spectra of physisorbed PF8 on 0.5-μm silica. [PF8]0=1.0 × 10–2 m in carbon tetrachloride, physisorption time=24 h. For comparison, PL and PLE spectra physisorbed in (c) toluene, (d) benzene and (e) chloroform under the same conditions as those used for carbon tetrachloride.

Full size image

Next, we discuss the reason for the β-phase with the high fractions. For amorphous polystyrene (PS), the main chain mobility of PS at the air-polymer interface is much higher than that of bulk PS. The glass transition temperature (Tg) at the surface substantially decreases at ≅0 °C relative to that of the bulk (i.e., Tg≅100 °C).44, 45 In contrast, a more recent study revealed that for PS deposited on a silicon wafer, the Tg value at the PS ultrathin layer significantly increases at ≅140 °C.46 In addition, the segmental motions of PSt carrying luminophores as pendants on three inorganic solids with a high Tg are very restricted, as confirmed by PL dynamics of the luminophores.47 If we assume that an inorganic solid substrate with a notably high Tg is crucial for this discrepancy, a silica solid with a high-melting temperature of ≅1500 °C could prevent main chain mobility of PF8 and enhance production of the β-phase at the air/PF8/silica interfaces.

Open questions remain as to whether PF8 at the silica surface deposits in monolayers or multilayers and whether PF8 adsorbs in face-on, edge-on or end-on states,48 as schematically outlined in Figure 7. Under the physisorption condition of [PF8]0=1.0 × 10–2 m in toluene for 24 h, the estimated cross-sections of di-n-octylfluorene with face-on, edge-on and end-on orientations at the surface are 1.55 nm2, 1.53 nm2, and 1.33 nm2, respectively (Supplementary Table S1). The surface coverages for the three orientations are assumed to be 0.96, 0.95 and 0.82, respectively.

Figure 7
figure 7

Schematic models of (left) face-on, (center) edge-on, and (right) end-on states of the β-phase PF8 layer at the silica surface (rectangular shape and solid line denote fluorene ring and rotatable C–C bonds, respectively, and the n-octyl chains are omitted for clarity).

Full size image

Although there are no definitive answers to these questions, regardless of these orientations, these evaluations suggest that PF8 prefers to deposit as a monolayer on the silica surface. Indeed, the preliminary molecular weight dependence of the physisorption behavior implies that the level-off amount of PF8 is nearly proportional to the Mw value of PF8 when the Mw is in the 38 × 103–300 × 103 range. Therefore, PF8 may be able to adopt face-on, edge-on and end-on orientations at the toluene-silica interface. The edge of the fluorene ring and/or end terminus of the PF8 chains contact the silica surface (Figure 7 and Supplementary Figure S15), leading to the densely packed β-phase of PF8. Another preliminary measurement and analysis using QCM-D (quartz crystal microbalance with dissipation monitoring system, Meiwafosis, Tokyo, Japan) indicated monolayer physisorption of PF8 onto a SiO2 substrate in toluene. Further physisorption behaviors of the molecular weight dependence at different physisorption temperatures (10–35 °C) and the n-alkyl side chain-length dependence of polyfluorene derivatives are currently in progress.

Other π-conjugated polymers and inorganic particles

With the existence of attractive vdW forces between PF8 and the silica particle, spontaneous solution-phase physisorption of PF8 on 0.5-μm silica was confirmed by several measurements and analyzes. To validate whether this physisorption is general or specific, we further tested two additional non-charged π-conjugated polymers (i.e., PF8T2 and MEH-PPV) and other 5-μm-sized ZnO and TiO2 (anatase and rutile) and 10-μm-sized Al2O3. Chloroform was employed as a common good solvent for these polymers. The PL spectra and photographs after excitation with a UV light source offer additional direct confirmation of these physisorption events, and the results are summarized in the Supplementary Figures S2–S14. In this study, the initial concentration of the polymers was fixed at ≅10–3 m with a physisorption time of 24 h using ≅250 mg of ceramic with 5 ml of chloroform at room temperature (23–25 °C).

First, it is evident that PF8 can spontaneously physisorb onto the surfaces of 5-μm ZnO (Supplementary Figures S4a and S4b) and rutile 5-μm TiO2 (Supplementary Figures S5a and S5b) according to the PL spectra and photographs, as well as onto the 10-μm Al2O3 particles (Supplementary Figure S12) according to the PL spectrum. Similarly, PF8T2 spontaneously physisorbs onto the surfaces of 0.5-μm SiO2 (Supplementary Figures S2a and S2b), 5-μm ZnO (Supplementary Figures S6a and S6b), 5-μm anatase TiO2 (Supplementary Figures S7a and S7b) and 5-μm rutile TiO2 (Supplementary Figures S8a and S8b) according to the PL spectra and photographs, as well as onto the 10-μm Al2O3 particles (Supplementary Figure S13) according to the PL spectrum. In addition, the MEH-PPV spontaneously physisorbs onto the surfaces of 0.5-μm SiO2 (Supplementary Figures S3a and S3b), 5-μm ZnO (Supplementary Figures S9a and S9b), 5-μm anatase TiO2 (Supplementary Figures S10a and S10b) and 5-μm rutile TiO2 (Supplementary Figures S11a and S11b) according to the PL spectra and photographs, as well as onto the 10-μm Al2O3 (Supplementary Figure S14) according to the PL spectrum.

Therefore, it is possible to design and construct various photoluminous and colored core-shell hybridized structures from various non-charged π-conjugated polymers and neutral insulating and semiconducting ceramics in the appropriate organic solvents without the use of any specific chemical reagents or any charged π-conjugated polymers. The photoluminescent and colored semiconducting core and semiconducting shell structures offer potential for use as photoconversion device materials in the future.

Summary and outlook

This solution-phase physisorption method allowed us to design various photoluminescent core-shell structures consisting of non-charged photoluminescent π-conjugated polymers and neutral ceramics in common organic solvents without the use of any specific chemical reagents over a short physisorption time at ambient temperature. Using the attractive vdW forces between these neutral materials, it was possible to design various luminous and colored core-shell structures and construct them from various non-charged π-conjugated polymers and neutral insulating and semiconducting ceramics in appropriate organic solvents without the use of specific chemical reagents or charged π-conjugated polymers. Our experimental results provide an alternative approach that enables hybridization of various π-conjugative polymers in a highly ordered structure onto ceramics in addition to the conventional chemisorption and physisorption approaches (i.e., (i) grafting-from by covalently linked polymerization with catalyst-modified surfaces, (ii) grafting-to with organic and polymeric adsorbates with functional Si–X and S–H groups, and (iii) LbL deposition of anionic and cationic polymer electrolytes driven by Coulombic forces.