The Optical Properties of Metal-Free Polymer Films with Self-Assembled Nanoparticles

In this paper, it is reported that a metal-free and non-conjugated polymer, MA-PEG 8000-BADGE (MP8B), exhibits an antireflective property and substrate-dependent photoluminescence (SDP). MP8B was constructed from maleic anhydride, poly(ethylene glycol) and bisphenol-A diglycidyl ether. Self-assembled nanoparticles are found in MP8B and can prospectively act as scattering centers to improve light trapping and extraction. MP8B films prepared from MP8B solutions have been characterized by photoluminescence (PL), atomic force microscopy (AFM), tunnelling electron microscope (TEM), reflectance, transmittance, and UV-Vis absorption spectrum. MP8B films can suppress light reflection and enhance light transmission. The PL spectrum of MP8B film on ITO peaks at approximately 538 nm, spanning from 450 to 660 nm at a concentration of 25 mM. Meanwhile, the effects of concentration and substrate on the PL of MP8B films are also investigated in this study. Surface roughness becomes larger with concentration. A red shift of the PL spectrum is observed as solution concentration increases. Meanwhile, aggregation-caused quenching (ACQ) is insignificant. Moreover, the PL spectra of MP8B films show a substrate-dependent phenomenon due to dielectric screening. The optical band-gap energy of MP8B is approximately 4.05 eV. It is concluded that MP8B is a promising candidate for a host material, and its film can be utilized as a multifunctional layer (i.e., antireflective and light-scattering functions) for optoelectronic applications.


Thin-Film Preparation
MP8B polymer was dissolved in tetrahydrofuran (THF). MP8B films prepared from MP8B solution were spin-coated for 40 s with 2000 rpm on ITO-coated glass, Si, and TiO2-coated Si substrates, and then annealed at 80 °C in N2 ambient for 1 h. Moreover, to confirm the influence of ACQ on the light emission of MP8B films, the samples prepared from using various MP8B solutions (1,5,10, and 25 mM) were spin-coated for 40 s at 2000 rpm on ITO-coated glass substrates, and then annealed at 80 °C in N2 ambient for 1 h.

Characterization
The thicknesses of MP8B films prepared from different solutions (1, 5, 10, and 25 mM) were measured by surface profiler (Bruker Dektak XT, Billerica, MA, USA) and were 63 nm, 65 nm, 81 nm, and 84 nm, respectively. The morphology of the polymer films were observed by atomic force microscope (Bruker, Dimension ICON, Tucson, AZ, USA). For the analysis of self-assembled nanoparticles, TEM images were captured (JEOL JEM-2100 TEM, Tokyo, Japan). PL properties were measured by a FluoroMax-4PL spectrometer (Horiba Jobin Yvon, Kyoto, Japan) and recorded in the range of 400-700 nm at 325 nm excitation wavelength at room temperature. UV-Vis absorption, reflectance, and transmittance spectra were measured using an UV/VIS/NIR spectrophotometer (UV-3150, Shimadzu Corporation, Tokyo, Japan). UV-Vis absorption spectra were also employed to estimate the optical band-gap energy (Eg). Figure 1 shows the chemical structure of the MP8B polymer. The chemical structure of the MP8B polymer was confirmed by 1 H nuclear magnetic resonance (NMR) and IR spectra, as given in the Supplementary Materials, Figures S1 and S2, respectively. The elemental composition of MP8B film was investigated through XPS measurement. Figure S3 displays the XPS spectrum of MP8B film deposited on an ITO-coated glass. C 1s and O 1s peaks are obvious in the XPS spectrum. This shows that MP8B film is metal-free. Furthermore, an N 1s peak is also visible, which most likely originates from contamination.

Results and Discussion
To observe the optical behaviors of MP8B film, it was spin-coated onto an ITO-coated glass substrate (i.e., MP8B/ITO-coated glass). Figure 2a,b shows the reflectance and transmittance spectra of MP8B/ITO-coated glass, respectively. The reflectance of MP8B/ITO-coated glass decreases compared to that of the ITO-coated glass over a wide wavelength range of 400-800 nm, as shown in Figure 2a. This shows that MP8B The elemental composition of MP8B film was investigated through XPS measurement. Figure S3 displays the XPS spectrum of MP8B film deposited on an ITO-coated glass. C 1s and O 1s peaks are obvious in the XPS spectrum. This shows that MP8B film is metal-free. Furthermore, an N 1s peak is also visible, which most likely originates from contamination.
To observe the optical behaviors of MP8B film, it was spin-coated onto an ITO-coated glass substrate (i.e., MP8B/ITO-coated glass). Figure 2a,b shows the reflectance and transmittance spectra of MP8B/ITO-coated glass, respectively. The reflectance of MP8B/ITOcoated glass decreases compared to that of the ITO-coated glass over a wide wavelength range of 400-800 nm, as shown in Figure 2a. This shows that MP8B film possesses antireflection properties. Meanwhile, the transmittance of MP8B/ITO-coated glass increases compared to that of the ITO-coated glass, as illustrated in Figure 2b. Notably, very little light is absorbed in the MP8B/ITO-coated glass structure (transmittance > 90%), indicating excellent optical quality. Moreover, the transmittance spectrum is relatively flat (93-95%) in the visible range, which means the optical property is insignificantly varied when it is used for lighting applications. These are attributed to the step gradient refractive-index distribution in constituent materials, i.e., air (n = 1)/ MP8B (n~1.61)/ITO (n~1.82)/glass (n~1.52). In the case of a single-layer ARC, one single coating is applied on the surface of a device, such that the reflection from the air-ARC and ARC-device interfaces experience destructive interference. To achieve this and zero reflection, the optical thickness of a single-layer ARC should be equal to an odd number multiple of a quarter of the wavelength of the selected light, and its refractive index should equal the square root of the refractive index of the surface of the device. In Figure 2, the thickness of MP8B film is approximately 152 nm and is not optimized, so that the reflectance is not minimized. the step gradient refractive-index distribution in constituent materials, i.e., air (n = 1)/ MP8B (n ~ 1.61)/ITO (n ~ 1.82)/glass (n ~ 1.52). In the case of a single-layer ARC, one single coating is applied on the surface of a device, such that the reflection from the air-ARC and ARC-device interfaces experience destructive interference. To achieve this and zero reflection, the optical thickness of a single-layer ARC should be equal to an odd number multiple of a quarter of the wavelength of the selected light, and its refractive index should equal the square root of the refractive index of the surface of the device. In Figure  2, the thickness of MP8B film is approximately 152 nm and is not optimized, so that the reflectance is not minimized. Figure 3a demonstrates the TEM image of self-assembled nanoparticles. Obviously, self-assembled nanoparticles are highly dispersive, and the size of the self-assembled nanoparticle is not uniform, from 154 to 500 nm. The size distribution fitted by a Gaussian curve is also shown in Figure 3b. It shows that the mean of the size is approximately 365.7 ± 127.9 nm. Generally, self-assembly processes of organic molecules to form nanostructures are developed in solution. Therefore, it has been suggested that the nanoparticle is produced in a self-assembly process with bisphenol-A aggregates and poly(ethylene glycol) moieties serving as core and shell compartments, respectively [44]. On the other hand, the nanoparticles in the film are able to scatter the incident light [46]. Therefore, scattering caused by nanoparticles is more efficient than absorption for light trapping in photovoltaic devices [49]. In addition, the larger the nanoparticle size, the more scattering [46]. It has also been published in the literature that the size of the self-assembled nanoparticle increases with the concentration of the polymer solution [45]. Similarly, it is expected that the size of the self-assembled nanoparticle can be tuned by the concentration of MP8B solution. However, the introduction of larger nanoparticles in the antireflective (AR) film leads to a further increase in the thickness of the AR film so that the absorption of the AR film increases correspondingly. Thus, the control of the diameter of the nanoparticle is an important issue. So far, many nanoparticles have been developed, which can be categorized into metallic and dielectric nanoparticles. Alt- Figure 3a demonstrates the TEM image of self-assembled nanoparticles. Obviously, self-assembled nanoparticles are highly dispersive, and the size of the self-assembled nanoparticle is not uniform, from 154 to 500 nm. The size distribution fitted by a Gaussian curve is also shown in Figure 3b. It shows that the mean of the size is approximately 365.7 ± 127.9 nm. Generally, self-assembly processes of organic molecules to form nanostructures are developed in solution. Therefore, it has been suggested that the nanoparticle is produced in a self-assembly process with bisphenol-A aggregates and poly(ethylene glycol) moieties serving as core and shell compartments, respectively [44]. On the other hand, the nanoparticles in the film are able to scatter the incident light [46]. Therefore, scattering caused by nanoparticles is more efficient than absorption for light trapping in photovoltaic devices [49]. In addition, the larger the nanoparticle size, the more scattering [46]. It has also been published in the literature that the size of the self-assembled nanoparticle increases with the concentration of the polymer solution [45]. Similarly, it is expected that the size of the self-assembled nanoparticle can be tuned by the concentration of MP8B solution. However, the introduction of larger nanoparticles in the antireflective (AR) film leads to a further increase in the thickness of the AR film so that the absorption of the AR film increases correspondingly. Thus, the control of the diameter of the nanoparticle is an important issue. So far, many nanoparticles have been developed, which can be categorized into metallic and dielectric nanoparticles. Although metallic nanoparticles are strong light scatterers at wavelengths near their resonant frequency, their intrinsic losses are significant. Furthermore, for metallic nanoparticles, the resonant scattering is only dominated by the electric-type resonances, whereas dielectric nanoparticles have both electric and magnetic dipole resonances simultaneously excited inside the same particle by light incidence [50]. Thus, dielectric nanoparticles are more suitable for light scattering for photovoltaic applications.
hough metallic nanoparticles are strong light scatterers at wavelengths near their resonant frequency, their intrinsic losses are significant. Furthermore, for metallic nanoparticles, the resonant scattering is only dominated by the electric-type resonances, whereas dielectric nanoparticles have both electric and magnetic dipole resonances simultaneously excited inside the same particle by light incidence [50]. Thus, dielectric nanoparticles are more suitable for light scattering for photovoltaic applications.  Figure 4 depicts PL spectra of MP8B films spin-coated on ITO glass, Si, and TiO2-coated Si substrates at an excitation wavelength of 325 nm. The PL peaks of MP8B films on ITO, Si, and TiO2 are 538, 522, and 488 nm, respectively. The PL spectra of MP8B films spin-coated on Si and TiO2 exhibit blue shifts (16 and 50 nm, respectively) compared with that on ITO. It has been shown that the PL of nanostructures is affected by doping [51][52][53][54][55], strain [54,[56][57][58][59][60], and dielectric screening [61]. MP8B films are annealed at 80 °C after spin-coating so that the doping effect on the PL of MP8B films is not significant in this study. Moreover, the lattice constants of ITO, Si, and TiO2 are 10.12 Å, 5.43 Å, and 4.59 Å, respectively. Thus, the induced strain of MP8B film on ITO is less than on Si and TiO2, which arises from a relatively small lattice mismatch between MP8B film and ITO. Therefore, if PL is dominated by lattice mismatch induced strain, a red shift in PL spectra should be observed with increasing strain [56][57][58][59]. However, this is contradictory to our observations in Figure 4. Furthermore, it has been shown that the PL may be intensively influenced by the screening of the dielectric environment [61]. In this study, it is observed that the wavelength of the peak decreases (blue shift) with the increase in the relative dielectric constant of the underlying film, from 3.4 (ITO) to 63.7 (TiO2). This result may be ascribed to the dielectric screening of the Coulomb interactions, which can affect the binding energies of excitons [61]. Therefore, the larger the relative dielectric constant of the underlying layer or substrate, the smaller the Coulombic interaction between electrons and holes, which thus decreases the exciton binding energy [62,63], so that the wavelength of the luminescence decreases (blue shift). Hence, if the effect of dielectric screening dominates PL, a blue shift is expected with the relative dielectric constant of the surrounding dielectrics because the PL peak energy can be estimated by subtracting the exciton binding energy from the band-gap energy, which is in good agreement with our observations. Consequently, it is suggested that the blue shift in PL results from dielectric screening in this study.  Figure 4 depicts PL spectra of MP8B films spin-coated on ITO glass, Si, and TiO 2coated Si substrates at an excitation wavelength of 325 nm. The PL peaks of MP8B films on ITO, Si, and TiO 2 are 538, 522, and 488 nm, respectively. The PL spectra of MP8B films spin-coated on Si and TiO 2 exhibit blue shifts (16 and 50 nm, respectively) compared with that on ITO. It has been shown that the PL of nanostructures is affected by doping [51][52][53][54][55], strain [54,[56][57][58][59][60], and dielectric screening [61]. MP8B films are annealed at 80 • C after spin-coating so that the doping effect on the PL of MP8B films is not significant in this study. Moreover, the lattice constants of ITO, Si, and TiO 2 are 10.12 Å, 5.43 Å, and 4.59 Å, respectively. Thus, the induced strain of MP8B film on ITO is less than on Si and TiO 2 , which arises from a relatively small lattice mismatch between MP8B film and ITO. Therefore, if PL is dominated by lattice mismatch induced strain, a red shift in PL spectra should be observed with increasing strain [56][57][58][59]. However, this is contradictory to our observations in Figure 4. Furthermore, it has been shown that the PL may be intensively influenced by the screening of the dielectric environment [61]. In this study, it is observed that the wavelength of the peak decreases (blue shift) with the increase in the relative dielectric constant of the underlying film, from 3.4 (ITO) to 63.7 (TiO 2 ). This result may be ascribed to the dielectric screening of the Coulomb interactions, which can affect the binding energies of excitons [61]. Therefore, the larger the relative dielectric constant of the underlying layer or substrate, the smaller the Coulombic interaction between electrons and holes, which thus decreases the exciton binding energy [62,63], so that the wavelength of the luminescence decreases (blue shift). Hence, if the effect of dielectric screening dominates PL, a blue shift is expected with the relative dielectric constant of the surrounding dielectrics because the PL peak energy can be estimated by subtracting the exciton binding energy from the band-gap energy, which is in good agreement with our observations. Consequently, it is suggested that the blue shift in PL results from dielectric screening in this study.
The AFM images in Figure S4 show that the roughness of MP8B film changes significantly as the concentration of MP8B solution is varied. The root mean square (RMS) roughness was estimated on an area of 30 × 30 µm 2 . Figure 5 illustrates RMS roughness values vary at different concentrations. From the AFM measurements, we find the RMS roughness values, which are 0.95, 1.84, 3.00, and 4.48 nm for MP8B films prepared from 1, 5, 10, and 25 mM of MP8B solutions, respectively, as shown in Figure 5. It is proposed that the diameter of the self-assembled nanoparticle increases with the increase in the concentration of MP8B solution. This is due to the formation of the self-assembled nanoparticle from bisphenol-A aggregates and poly(ethylene glycol) moieties [45]. Therefore, the higher the solution concentration, the larger the size of the self-assembled nanoparticle, so that MP8B film is rougher at higher concentrations. The AFM images in Figure S4 show that the roughness of MP8B film change nificantly as the concentration of MP8B solution is varied. The root mean square (R roughness was estimated on an area of 30 × 30 μm 2 . Figure 5 illustrates RMS rough values vary at different concentrations. From the AFM measurements, we find the roughness values, which are 0.95, 1.84, 3.00, and 4.48 nm for MP8B films prepared fro 5, 10, and 25 mM of MP8B solutions, respectively, as shown in Figure 5. It is prop that the diameter of the self-assembled nanoparticle increases with the increase in concentration of MP8B solution. This is due to the formation of the self-assembled noparticle from bisphenol-A aggregates and poly(ethylene glycol) moieties [45]. T fore, the higher the solution concentration, the larger the size of the self-assemble noparticle, so that MP8B film is rougher at higher concentrations.  The AFM images in Figure S4 show that the roughness of MP8B film changes nificantly as the concentration of MP8B solution is varied. The root mean square (R roughness was estimated on an area of 30 × 30 μm 2 . Figure 5 illustrates RMS rough values vary at different concentrations. From the AFM measurements, we find the R roughness values, which are 0.95, 1.84, 3.00, and 4.48 nm for MP8B films prepared fro 5, 10, and 25 mM of MP8B solutions, respectively, as shown in Figure 5. It is propo that the diameter of the self-assembled nanoparticle increases with the increase in concentration of MP8B solution. This is due to the formation of the self-assembled noparticle from bisphenol-A aggregates and poly(ethylene glycol) moieties [45]. Th fore, the higher the solution concentration, the larger the size of the self-assembled noparticle, so that MP8B film is rougher at higher concentrations.   Figure 6a illustrates the PL spectra of MP8B films prepared using various solutions (1 mM, 5 mM, 10 mM, and 25 mM) and annealed at 80 • C. The PL peak of MP8B film on ITO-coated glass is approximately 480 nm at a concentration of 1 mM. The PL peak is gradually changed from 480 to 538 nm when solution concentration rises from 1 to 25 mM. In other words, the PL peak shifts toward the longer wavelength (a red shift) with solution concentration. It was generally suggested that the PL peak exhibits a blue shift when lowering the size of the nanostructure due to the quantum size effect [64]. Hence, it is proposed that the diameter of the self-assembled nanoparticle decreases with a decrease in solution concentration in this study. That is, the higher the concentration of the MP8B solution, the larger the size of the self-assembled nanoparticle. This can be attributed to the larger amount of MP8B (higher concentration), which leads to more formation of bisphenol-A aggregates and poly(ethylene glycol) moieties so that the size of the self-assembled nanoparticle increases. This is consistent with the result of Figure 5. solution concentration. It was generally suggested that the PL peak exhibits a blue shift when lowering the size of the nanostructure due to the quantum size effect [64]. Hence, it is proposed that the diameter of the self-assembled nanoparticle decreases with a decrease in solution concentration in this study. That is, the higher the concentration of the MP8B solution, the larger the size of the self-assembled nanoparticle. This can be attributed to the larger amount of MP8B (higher concentration), which leads to more formation of bisphenol-A aggregates and poly(ethylene glycol) moieties so that the size of the self-assembled nanoparticle increases. This is consistent with the result of Figure 5. On the other hand, it also indicates that the PL peak intensity of MP8B film is concentration-dependent. As in the cases of lower concentrations (≤ 10 mM), PL peak intensity increases with solution concentration. However, it slightly decreases at a concentration of 25 mM, as shown in Figure 6a. Although traditional fluorescent conjugated materials generally exhibit intensive emission in diluted solution, severe luminescence quenching of the conjugated materials is usually observed in highly concentrated solutions or in the solid state. This phenomenon is referred to as aggregation-caused quenching (ACQ) [48].
In general, it is known that the PL intensity depends on the sample thickness. To study the concentration dependence of MP8B luminescence efficiency in more detail, the PL intensity is normalized to film thickness to exclude the contribution of thickness shown in Figure 6b. In this way, the normalized PL spectra are not dependent on the thickness. No concentration quenching effects have been observed up to 5 mM. Unfortunately, there is a slightly downward trend in PL peak intensity at higher concentrations (≥ 10 mM), as illustrated in Figure 6b The UV-Vis absorption spectra of MP8B films prepared from different solutions are shown in Figure 7. It is observed that the spectra of MP8B films are red-shifted with solution concentration. This implies that more molecular aggregation and stronger intermolecular interactions are present in the polymer film prepared from the higher concentrated solutions [65]. That is, the higher the concentration, the more molecular aggrega- On the other hand, it also indicates that the PL peak intensity of MP8B film is concentration-dependent. As in the cases of lower concentrations (≤10 mM), PL peak intensity increases with solution concentration. However, it slightly decreases at a concentration of 25 mM, as shown in Figure 6a. Although traditional fluorescent conjugated materials generally exhibit intensive emission in diluted solution, severe luminescence quenching of the conjugated materials is usually observed in highly concentrated solutions or in the solid state. This phenomenon is referred to as aggregation-caused quenching (ACQ) [48].
In general, it is known that the PL intensity depends on the sample thickness. To study the concentration dependence of MP8B luminescence efficiency in more detail, the PL intensity is normalized to film thickness to exclude the contribution of thickness shown in Figure 6b. In this way, the normalized PL spectra are not dependent on the thickness. No concentration quenching effects have been observed up to 5 mM. Unfortunately, there is a slightly downward trend in PL peak intensity at higher concentrations (≥10 mM), as illustrated in Figure 6b.
The UV-Vis absorption spectra of MP8B films prepared from different solutions are shown in Figure 7. It is observed that the spectra of MP8B films are red-shifted with solution concentration. This implies that more molecular aggregation and stronger intermolecular interactions are present in the polymer film prepared from the higher concentrated solutions [65]. That is, the higher the concentration, the more molecular aggregation. This is in agreement with the results shown in Figures 5 and 6. On the other hand, the estimate of the optical band-gap energy (E g ) is related to optical absorption edge. An exciton is generated as a photon is absorbed. The optical E g is, therefore, derived from the onset of the optical absorption. The absorption onset is determined by linear extrapolation of the low energy edge of the UV-Vis absorption spectrum. Consequently, as solution concentrations are 1, 5, 10, and 25 mM, the onset wavelengths (λ onset ) are 305.2 nm, 306.3 nm, 306.5 nm, and 306.8 nm, respectively. The optical E g is evaluated by [66,67]: where λ onset is the onset wavelength. Thus, the optical E g of MP8B film decreases from 4.06 to 4.04 eV as solution concentration is increased from 1 to 25 mM. This result indicates that the optical E g of MP8B film slightly decreases with the increase in solution concentration. Therefore, this also implies that MP8B with a wide E g (around 4.05 eV) can be a promising candidate for a host material for polymer light-emitting diode (PLED) applications.

Conclusions
MP8B is a metal-free and non-conjugated polymer. MP8B film prepared from M solution can lower reflectance and improve transmittance. The PL peaks of MP8B f on Si and TiO2 show blue shifts (16 and 50 nm, respectively) compared to that on The SDP is probably due to the dielectric screening in this study. The TEM imag MP8B film has shown the existence of a size distribution of self-assembled nanopart ranging between 154 and 500 nm. Furthermore, the RMS roughness of MP8B film comes larger with solution concentration. It is suggested that the increase in the siz the nanoparticle with increasing solution concentration can be ascribed to the m self-assembled formation of bisphenol-A aggregates and poly(ethylene glycol) moi in the higher concentrated solution so that the RMS roughness of MP8B film incre correspondingly. Accordingly, the control of the diameter and density of self-assem nanoparticles is an important issue for future work. On the other hand, the PL peak hibits a blue shift (~54 nm) with decreasing concentration, from 25 to 1 mM, owing to

Conclusions
MP8B is a metal-free and non-conjugated polymer. MP8B film prepared from MP8B solution can lower reflectance and improve transmittance. The PL peaks of MP8B films on Si and TiO 2 show blue shifts (16 and 50 nm, respectively) compared to that on ITO. The SDP is probably due to the dielectric screening in this study. The TEM image of MP8B film has shown the existence of a size distribution of self-assembled nanoparticles ranging between 154 and 500 nm. Furthermore, the RMS roughness of MP8B film becomes larger with solution concentration. It is suggested that the increase in the size of the nanoparticle with increasing solution concentration can be ascribed to the more self-assembled formation of bisphenol-A aggregates and poly(ethylene glycol) moieties in the higher concentrated solution so that the RMS roughness of MP8B film increases correspondingly. Accordingly, the control of the diameter and density of self-assembled nanoparticles is an important issue for future work. On the other hand, the PL peak exhibits a blue shift (~54 nm) with decreasing concentration, from 25 to 1 mM, owing to the quantum size effect. Therefore, it is deduced that the diameter of the self-assembly nanoparticle decreases with a decrease in solution concentration so that the PL peak shifts toward the shorter wavelengths (a blue shift) with decreasing solution concentration. This is consistent with the results of the RMS roughness. In addition, PL peak intensity is concentration-dependent. No concentration quenching effect has been observed up to 5 mM, excluding the contribution of film thickness. Nevertheless, as in the cases of higher concentrations (≥10 mM), there is a slightly downward trend in PL peak intensity. A red shift in UV-Vis absorption spectra is also observed with the increase in solution concentration. This indicates that more molecular aggregation and stronger intermolecular interactions exist in MP8B films prepared at higher concentrations. The optical band-gap energy is around 4.05 eV. These results illustrate that MP8B is a promising candidate for a host material, and its film can be utilized as a multifunctional layer (i.e., antireflective and light-scattering functions) for optoelectronic applications.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.