Introduction

Light emitting diodes (LEDs) have drawn considerable interests in the last several decades due to their high lighting efficiency and lighting wavelength tunability1,2,3,4. In a typical LED device, lighting chip and packaging material are the main components which determine the light emitting efficacy. Concerning the long-term service of LEDs, transmittance, refractive index (RI) and thermal stability are the key criteria required for packaging materials. Silicone resin is endorsed as the first choice over other types of materials (such as epoxy resin, polymethyl methacrylate, etc.) due to its high thermal stability and facile fabrication process5,6,7. However, the RI mismatch between the lighting chip and silicone resin leads to severe light reflection at the interface of LED chip and silicone resin, and thus greatly lowers the light emitting efficacy8. Great efforts have been made to increase the RI of silicone resin by introduction of functionalized atoms and side chains, or bottom-up design of polymer structure9,10,11,12,13,14. Currently, commercial silicone resins with relatively high RI are available15, whereas their very high prices and complicated processing procedures hinder their wide applications as LED packaging materials.

According to the Maxwell–Garnett effective medium theory16, the incorporation of inorganic filler with high RI can increase the RI of the polymer matrix8. TiO2 with a high RI (n = 2.45 and 2.7 for anatase and rutile phase, respectively) and a very low absorption coefficient in the visible range is highly attractive as a filler for the fabrication of high RI packaging materials17, 18. However, the poor dispersion of nanoparticles in the polymer matrix is a great challenge due to their high surface energy. Although various methods have been developed to improve the dispersion of TiO2 nanoparticles in silicone resin, likewise, surface modification of TiO2 nanoparticle in the process of sol-gel19,20,21, modification of silicone resin and etc.22, 23. However, these approaches usually involve toxic agents and complicated processing procedures of synthesis, purification, mixing of TiO2 nanoparticles24, and so on. Therefore, a simple yet environmental-friendly approach is strongly required to fabricate highly transparent silicone nanocomposites with high RI and high thermal stability.

Recently, a facile approach to prepare size-controllable ZnO nanoparticle has been reported by the introduction of coupling agent into precursor24, 25. In this work, a similar strategy has been applied to synthesize surface modified ultra-fine TiO2 (S-TiO2) nanoparticles through the introduction of titanate coupling agent in the process of TiO2 nanoparticle growth. The protective layer formed on the surfaces of TiO2 nanoparticles can effectively prevent the growth of TiO2 nanoparticles and significantly improve the compatibility between TiO2 nanoparticles and the silicone resin. In principle, the introduction of inorganic TiO2 can enhance the thermal stability of polymers. Thus, a highly transparent TiO2/silicone nanocomposite with a high RI and high thermal stability can be obtained by directly mixing the ultra-fine S-TiO2 nanoparticles with the silicone resin. To evaluate their potential as LED packaging material, the transmittance, RI and thermal stability of the as-prepared silicone nanocomposite are examined. As a control, several types of TiO2 nanoparticles without the introduction of coupling agent and the corresponding composites are prepared and compared with the surface modified S-TiO2 case.

Results and Discussions

Synthesis of S-TiO2

Figure 1 shows the general steps to fabricate S-TiO2/silicone nanocomposites. First, the TiO2 nanoparticles synthesized via sol-gel approach were introduced into silicone resin/acetic ether solution under ultrasonic treatment, forming a uniform suspension. Subsequently, the resulting suspension was subjected to a vacuum chamber to remove acetic ether and bubbles. Finally, the transparent nanocomposites were prepared by curing the resin at 100 °C. The only difference of the preparation procedure between S-TiO2/silicone nanocomposite and un-treated TiO2/silicone nanocomposite is the introduction of titanate coupling agent, which endows TiO2 nanoparticles with protective layers on their surface.

Figure 1
figure 1

Illustration of the fabrication of S-TiO2/silicone nanocomposites.

The morphology of the as-prepared TiO2 materials was characterized by TEM. As shown in Fig. 2, the TiO2-1 has a diameter of approximately 5 nm. Increasing the reaction time from 48 h to 72 h leads to an increase of the diameter of TiO2-2 nanoparticles from ca. 10 to ca. 15 nm, accompanying with TiO2 aggregation. TiO2-3 obtained by drying TiO2-2 at 120 °C for 24 h shows a significant increase in the diameter of particles up to the scale of micrometer. By contrast, S-TiO2 nanoparticles with a diameter of 3–4 nm are mono-dispersed in the composite. The afore-mentioned results are supported by the BET measurement. As shown in Table S1, specific surface areas of 35.8, 17.6, 2.84, and 75.6 m2/g are achieved for TiO2-1, TiO2-2, TiO2-3 and S-TiO2, respectively, which agrees well with the result for the diameter of TiO2 nanoparticles. X-ray diffraction patterns (top insets of Fig. 2a–d) indicate that the crystallinity of TiO2 increases with the increases of reaction time and temperature. Additional evidence on the chemical structure of the TiO2 particles was provided by EDS analysis. C originated from titanate coupling agent is observed in S-TiO2 while no trace of C is found in TiO2-1, TiO2-2 and TiO2-3 (Table S1), indicating formation of protective layers on the surfaces of S-TiO2. Due to the formation of protective layers on their surfaces during the growth of TiO2 nanoparticles, the increase of the TiO2 diameter is declined and stopped until the formation of intact layers on their surfaces.

Figure 2
figure 2

TEM images of (a) TiO2-1, (b) TiO2-2, (c) TiO2-3 and (d) S-TiO2; solubility of (e) S-TiO2 and (f) silicone resin in (A) acetone, (B) petroleum ether, (C) cyclohexane and (D) acetic ether.

Dispersibility of TiO2

To achieve a homogenous dispersion of TiO2 in silicone resin, co-solvent is needed to dilute silicone resin due to the poor solubility of silicone resin in ethanol. As shown in Fig. 2e and f, the suspensions of S-TiO2 and silicone resin formed in acetic ether both are transparent, while their suspensions formed in other solvents including acetone, petroleum ether and cyclohexane are turbid. In addition, acetic ether featured by low toxicity and low boiling point is a good candidate for the massive fabrication of TiO2/silicone nanocomposite. The dispersity of TiO2-1, TiO2-2 and TiO2-3 in acetic ether was also investigated via the same strategy. As presented in Table 1, all of the suspensions of TiO2-1, TiO2-2 and TiO2-3 formed in acetic ether are turbid, indicating that the dispersibility of S-TiO2 in acetic ether is much better than that of TiO2 without surface modification. In addition, to reveal the effectiveness of various surface modifiers in improving the dispersibility of TiO2, the transparency of their suspensions in acetic ether was further studied. As presented in Table S2, all of the suspensions of TiO2 modified by oxalic acid, citric acid, KH-550 and KH-560 are turbid, indicating that titanate coupling agent is the most effective in improving the dispersibility of TiO2 in acetic ether.

Table 1 Dispersibility of various TiO2 particles in acetic ether.

Transmittance of TiO2/silicone nanocomposites

The cross sections of silicone nanocomposites of various TiO2 nanoparticles were observed with SEM. As shown in Fig. 3a–c, aggregates sized ca. 150 nm, 30 nm and 10 nm are observed in the nanocomposites incorporated with TiO2-1, TiO2-2, and TiO2-3, respectively, suggesting the poor dispersion of TiO2-1, TiO2-2 and TiO2-3 in the silicone resin. In comparison, no visible aggregates appear in the S-TiO2/silicone nanocomposite. S-TiO2 nanoparticles are mono-dispersed in the silicone resin with a diameter around 3–4 nm (see the top inset of Fig. 3d). The narrow size distribution and excellent dispersion of S-TiO2 in the silicone resin are attributed to the titanate coupling agent layers formed on the surfaces of TiO2 nanoparticles, endorsing the excellent compatibility between TiO2 nanoparticles and silicone resin, and thus eliminating the aggregation of TiO2 nanoparticle in both acetic ether and silicone resin.

Figure 3
figure 3

SEM images of the cross sections of silicone nanocomposites with 1 wt% (a) TiO2-3, (b) TiO2-2, (c) TiO2-1, and (d) S-TiO2 (top inset: TEM image of the corresponding TiO2/silicone nanocomposite).

The following equation gives the light intensity of a spherical particulate composite26, 27:

$$\frac{{\rm{I}}}{{{\rm{I}}}_{0}}=\exp \{[-\frac{32{{\rm{V}}}_{{\rm{P}}}{\rm{x}}{\pi }^{4}{{\rm{r}}}^{3}}{{{\rm{\lambda }}}^{4}}{(\frac{{({{\rm{n}}}_{{\rm{p}}/}{{\rm{n}}}_{{\rm{m}}})}^{2}-1}{{({{\rm{n}}}_{{\rm{p}}/}{{\rm{n}}}_{{\rm{m}}})}^{2}+2})}^{2}]\}$$
(1)

where r, np and Vp are the radius, RI and volume fraction of spherical particles, respectively. nm is the RI of matrix, λ is light wavelength, and x is the thickness of the composite. As revealed in Equation 1, the diameter of nano-filler significantly affects the transmittance of the composite. Therefore, to maintain the high transparency of the composite, the ultra-fine nano-fillers with mono-dispersion are strongly required. Meanwhile, the transmittance of a composite is highly related to the thickness of composites, it is quite difficult to achieve a high transparency for bulk materials than thin films. To our best knowledge, all of the transparent composites with enhanced RI reported previously are thin films with a thickness of 10−8 to 10−4 m, whereas bulk transparent nanocomposites with a thickness of 10−3 to 10−2 m are rarely reported27,28,29,30. The thickness of the resin layer as encapsulating material surely has some effect on the lighting efficiency of LED. In this study, the thickness of the as-prepared silicone composite is ca. 1 mm (namely ca. 10−3 m), close to that of realistic packaging materials for LEDs. This could give us practical evaluation of the current protocol on the performance of the as-prepared silicone composite. The high specific surface area of nanoparticles would lead to irreversible aggregation32, 33. Aggregated nanoparticles usually have much larger sizes than individual nanoparticles, and thus would have an adverse effect on the transparency and lighting efficiency of LED. In this work, S-TiO2 nanoparticles are mono-dispersed in the silicone resin, and thus the resultant composite as encapsulating material has a high transparency (Fig. 4). As a result, the introduction of S-TiO2 would have no adverse effect on the efficiency of LED.

Figure 4
figure 4

(a) Transmittance and (b) photograph of the silicone nanocomposite with 1 wt% TiO2-3, TiO2-2, TiO2-1 and S-TiO2, and pure silicone as the control; (c) transmittance and (d) corresponding photograph of titanate S-TiO2/silicone nanocomposites with varied S-TiO2 mass fractions.

Figure 4a shows the transmittance of the pure silicone resin and silicone nanocomposites with 1 wt% TiO2. Compared with the pure silicone resin, only 2% decrease in transmittance is achieved in the S-TiO2/silicone nanocomposite. As the TiO2 particle size increases, the transparency of silicone nanocomposite decreases. Figure 4b visually demonstrates the transparency of the corresponding silicone nanocomposites. The transparency of TiO2-1/silicone nanocomposite and TiO2-2/silicone nanocomposite is approximately 83% and 75%, respectively. Further increase in the particle size of TiO2 results in an opaque silicone nanocomposite with 20% of transmittance in the visible light range. This change obeys Equation 1.

The effect of S-TiO2 filler content on the transmittance is shown in Fig. 4c. With the addition of 0.5 and 1.0 wt% S-TiO2, the transmittance of silicone nanocomposite is close to that of pure silicone resin, and the average decrease of the transmittance in the visible range of 400–800 nm is less than 2%. As revealed in Fig. 4d, no visible transparency difference is seen with naked eyes between the pure silicone resin and the nanocomposites filled with 0.5 and 1.0 wt% of S-TiO2. Even increasing the content of S-TiO2 up to 5.0 wt%, the transmittance in the visible range still maintains at a high level of approximately 70%, indicating that the surface modification of TiO2 with titanate coupling agent is very effective in fabricating highly transparent TiO2/polymer nanocomposite.

RI of TiO2/silicone nanocomposites

The lighting efficiency of LED can be determined by the difference of refractive index between lighting chips and packaging materials as indicated by Equation 231:

$${{\rm{\eta }}}_{{\rm{encapsulated}}}=(\frac{1-\,\cos (\arcsin \frac{{{\rm{n}}}_{{\rm{pm}}}}{{{\rm{n}}}_{{\rm{chip}}}})}{1-\,\cos (\arcsin \frac{1}{{{\rm{n}}}_{{\rm{chip}}}})}){{\rm{\eta }}}_{{\rm{unencapsulated}}}$$
(2)

where ηencapsulated and ηunencapsulated are the lighting efficiency of encapsulated and un-encapsulated LED chip, respectively; npm and nchip are the refractive index of packaging materials and LED chip, respectively. It is understandable from Equation 2 that as the RI of the packaging material increases, the lighting efficiency first increases quasi-linearly, then sub-linearly, and finally reaches a saturation. Since the RI of conventional lighting chip material is 2.5~3.0, the packaging material with a higher RI will lead to a higher lighting efficiency of LED. In this work, the introduction of fine TiO2 particles with a high RI brings about a high RI of the composite as encapsulating material, aiming at increasing the lighting efficiency of LED.

The RI measured for the silicone nanocomposite with 1 wt% S-TiO2, TiO2-1, TiO2-2, and TiO2-3 is 1.596, 1.596, 1.602, and 1.642, respectively. The RI of silicone nanocomposite filled with TiO2-3 is somewhat higher than other nanocomposites, due to the crystallinity of TiO2-3 formed with longer aging time is better than that of S-TiO2, TiO2-1, and TiO2-2. Figure 5 shows the RI of S-TiO2/silicone nanocomposites as a function of S-TiO2 mass fraction. As the S-TiO2 loading increases from 0 to 0.5 and 1 wt%, the RI of silicone resin is dramatically increased from 1.42 to 1.56 and 1.596, achieving the improvement of 9.8% and 12.3%, respectively. In comparison with other researches, the current strategy shows higher improvements in RI. For instance, the incorporation of 1% of silane modified TiO2 and oleic acid capped TiO2 in to silicone resin increases the RI of silicone composite from 1.51 to 1.56 and 1.575, respectively32, 33. As the content of S-TiO2 increases further, the increasing rate of the RI is declined. Less than 2% of increase in RI is achieved as the content of S-TiO2 increases further to 5 wt%, indicating that the effectiveness of RI enhancement at a higher filler content is limited. In principle, the porosity if any would reduce the RI of the composite since the air has a RI of unity. In the current work, the porosity measured on the surfaces of TiO2 particles was around 3%. On the other hand, silicone resin might penetrate into the porosity. As a result, the effect of the porosity on the RI could be neglected. Therefore, the RI of the composite was determined mainly by TiO2 particles and enhanced by the incorporation of TiO2 particles as confirmed by Fig. 5.

Figure 5
figure 5

RI of S-TiO2/silicone nanocomposites with different mass fractions.

Thermal stability of TiO2/silicone nanocomposite

Figure 6 shows the TGA curves of the pure silicone resin and the silicone nanocomposites with various S-TiO2 loadings. Although the starting decomposition temperature is almost the same for the pure silicone resin and the silicone nanocomposites, the incorporation of TiO2 nanoparticles decreases the decomposition speed of the silicone nanocomposite. Taking the example of the weight loss of 50%, the temperature of the silicone nanocomposites with 0%, 1% and 5% S-TiO2 is 532, 605 and 632 °C, respectively; namely, the temperature corresponding to the weight loss of 50% has been increased by 73 and 100 °C, respectively. The “cross-link density” of the nanocomposites is increased with the incorporation of TiO2 nanoparticles, which leads to the enhancement of their thermal stability with a decreased decomposition rate34.

Figure 6
figure 6

TGA curves of pure silence resin and S-TiO2/silicone nanocomposites with 1 wt% and 5 wt% S-TiO2.

Conclusions

In summary, a facile approach to produce highly transparent TiO2/silicone nanocomposite with a high RI and a high thermal stability has been demonstrated by the incorporation of S-TiO2 nanoparticles treated by titanate coupling agent into diluted silicone resin via simple solvent mixing. The titanate coupling agent layers formed on the surfaces of ultra-fine 3–4 nm TiO2 nanoparticles endorses the excellent dispersion of S-TiO2 nanoparticles in the silicone resin and an excellent compatibility between the TiO2 nanoparticles and the silicone resin. Consequently, with the addition of 1 wt% S-TiO2, the as-prepared silicone nanocomposite shows an enhanced RI from 1.42 to ca. 1.60 (13% enhancement), a high transmittance of 88% in the entire visible light range for 1 mm thick samples and meanwhile an enhanced thermal stability. The current strategy is cost-effective, environmental-friendly, and easy-processing, which has a great potential to be widely used in producing transparent packaging materials with the greatly enhanced RI and meanwhile a high thermal stability.

Materials and Methods

Materials

Tetrabutyl titanate (C16H36TiO4) was purchased from Beijing Xingjin Chemical Reagent Company. Absolute ethanol, acetic ether, acetone, petroleum ether, and cyclohexane were supplied by Beijing Chemistry Reagent Company and used as received. Titanate coupling agent (TM-P) was bought from JiangSu YiZheng TianYang Chemical Plant. Silicon resin (KMT-1252) with two parts A and B was purchased from Beijing KMT Technology Co., Ltd.

Synthesis of S-TiO2 nanoparticles

S-TiO2 nanoparticles were prepared with a typical sol-gel procedure. First, 0.01 M tetrabutyl titanate was added to 50 mL ethanol, then the solution was stirred for 15 min to ensure complete dissolution of tetrabutyl titanate. Afterwards, 0.12 mL TM-P was introduced into the suspension and stirred for another 15 min. Subsequently, 1 mL distilled water was added to the suspension obtained and kept stirring for 48 h at 40 °C.

In parallel, TiO2 nanoparticles with different sizes were synthesized using the similar protocol in the absence of the titanate coupling agent TM-P. The samples react for 48 h and 72 h, denoted as TiO2-1 and TiO2-2, respectively. TiO2-3 was prepared by vacuum drying of TiO2-2 at the temperature of 120 °C for 24 h.

Fabrication of TiO2/silicone nanocomposites

In a typical fabrication procedure, 10 g silicon resin with the mass ratio of 10:1 for Part A and Part B was dissolved in 20 mL acetic ether by ultrasonic treatment until a transparent solution was obtained. Then, 0.11 g TiO2 particles (TiO2-1, TiO2-2, TiO2-3 or S-TiO2) were homogenously mixed with silicone resin/acetic ether solution by ultrasonic treatment. The bubbles and acetic ether were removed by vacuum pump. Finally, the resultant suspension was casted into a stainless steel mould and then transferred into an oven at 100 °C for 2 h. The thickness of as-prepared silicone composites is approximately 1 mm.

Characterizations

The transmittance of the nanocomposite was obtained from a Hitachi U-3900. The RI was measured on an Abbe Refractometer (WAY-2S) in ambient atmosphere. Transmission electron microscopy (TEM) images and electronic diffraction were taken on a JEM-2100F instrument (operated at 200 kV). Scanning electron microscopy (SEM) images and energy dispersive spectroscopy (EDS) were collected on a Hitachi S-4300. The Brunauer–Emmett–Teller (BET) specific surface areas were calculated using adsorption data in P/P0 = 0.05–0.3 (six points collected). The weight loss analysis of the samples was conducted on a Netzsch STA 409 PC/PG at the rate of 10 °C/min under nitrogen atmosphere. Fourier transform infrared (FTIR) spectra were recorded using a Varian 3100 FT-IR spectrometer with 2 cm−1 resolution and accumulation of 24 scans.