3.1. CPI resin synthesis and film preparation
One pure CPI (ABTFCPI-0) and seven CPI nanocomposite films (ABTFCPI-5~ABTFCPI-35) were prepared, respectively based on the ABTFCPI-0 matrix resin, as shown in
Figure 1. The resin was soluble in the polycondensation system despite the existence of the rigid-rod benzanilide units in the diamine moiety. The non-conjugated molecular structure feature in the HPMDA dianhydride moiety endowed the derived resin good solubility in the reaction media. The resin was soluble in both of polar aprotic solvents, such as N,N–dimethylformamide (DMF), DMAc, NMP, dimethyl sulfoxide (DMSO) and common polar solvents, such as cyclopentanone, chloroform, and so on at the solid content of 10 wt%.
The good solubility of the ABTFCPI-0 resin in organic solvents made it possible to confirm the structure by
1H-NMR measurements, as illustrated in
Figure 2. In the figure, the characteristic absorptions of hydrogen protons in the amide (–CONH–) units, in the aromatic rings (H
1~H
5), and in the alicyclic rings (H
a, H
b, and H
b’) could be clearly assigned. As expected, the amide proton exhibited the absorption at the chemical shift of 10.77 ppm, which was at the farthest downfield in the spectra. This is due to the strong electron-withdrawing carbonyl group. Similarly, the absorption of H
3 appeared at the second farthest downfield in the spectra due to the ortho-substituted –CF
3 groups with the high electronegativity. Contrarily, the absorptions of the cyclohexane protons were present at the upfield in the spectra due to the electron-donating features of the alicyclic rings. This is consistent with the structural features of the ABTFCPI-0 resin.
The pure ABTFCPI-0 film and the ABTFCPI-SiO
2 nanocomposite films (ABTFCPI-5~ABTFCPI-35) were prepared according to the procedure shown in
Figure 3. All the films showed flexible and tough characters except ABTFCPI-35, which was a bit brittle and couldn’t be folded like the other counterparts. The good flexibility of the nanocomposite films was on one hand due to the merits of the ABTFCPI-0 matrix and on the other hand owing to the good dispersion and distribution of the SiO
2 nanoparticles. The ABTFCPI-0 resin showed the numerical average molecular weight (M
n) of 1.11×10
5 g/mol, the weight average molecular weight (M
w) of 2.08×10
5 g/mol, and the polydispersity index (PDI) of 1.87. The high molecular weights of the resin could endow the derived films good strength and toughness. As for the colloidal silica nanoparticles, they could be uniformly dispersed into the pristine ABTFCPI-0 matrix due to the specific physical and chemical features of the fillers. Few aggregations were observed for the nanoparticles in the composite films, as could be evidenced by the clear and transparent appearance of the composite solutions and films shown in
Figure 3. Even the loading amounts of the SiO
2 fillers reached 35 wt%, the composite films still maintain good optical transparency.
The successful dispersion of the SiO
2 nano-fillers in the composite films could further be proven by the XRD and XPS measurements. As could be seen from the XRD plots of the polymer films shown in
Figure 4, either the pristine ABTFCPI-0 film or the nanocomposite films exhibited amorphous nature although the SiO
2 nanoparticles showed somewhat crystalline characters. This is mainly ascribed to the non-conjugated feature of the cyclohexane rings in the dianhydride moiety, which efficiently prohibited the formation of crystalline regions in the polymers. In addition, it could be deduced from the full width at half maxima (FWHM) values of the samples labeled in the figure that the FWHM values increased with the increasing SiO
2 contents in the films. This indicates that the partially ordered molecular packing structures in the films were gradually destroyed with the increasing contents of the SiO
2. On the other hand, it also indicates that with the increase of the SiO
2 fillers, some interactions between the SiO
2 particles and the ABTFCPI-0 matrix occurred, causing a lower crystallinity in the films. The loose molecular packing in the nanocomposite films was beneficial for the penetration of visible light and endowed the films good optical transparency.
Figure 5 depicts the XPS plots of the ABTFCPI films. All the samples showed clear absorptions of common elements, including the F1s at the binding energy of 688 eV, O1s at 532 eV, N1s at 400 eV, and C1s at 284 eV. However, only the composite films showed the absorptions of Si2p in the range of 100~153 eV, as could be seen from
Figure 5a. In addition, from the expanded binding energy ranged from 99 to 107 eV (
Figure 5b), one can observe that the absorption of Si2p increased with the increasing contents of the SiO
2 fillers in the polymers. This is in good agreement with the composition characters of the nanocomposite films.
After confirming the successful incorporation of SiO
2 nanoparticles into the ABTFCPI-0 matrix, the chemical structures of the polymers were identified by the FTIR measurements.
Figure 6 shows the FTIR spectra of the ABTFCPI films. The spectra revealed the characteristic absorptions peaks of the imide rings for all the polymers, including the asymmetrical and symmetrical carbonyl stretching vibrations at 1785 cm
-1 and 1710 cm
-1 and the C–N at 1381 cm
-1. In addition, the phenyl C=C stretching vibrations at 1504 cm
-1 and the amide (–CONH–) carbonyl stretching vibrations at 1670 cm
-1 were also observed. The Si–O stretching vibrations at 1049 cm
-1 were only detected for the composite films, indicating the incorporation of the silica fillers. At last, the C–H stretching vibrations in cyclohexane units at 3000~2900 cm
-1 were also detected.
3.2. Optical properties
Deterioration of the optical properties of the nanocomposite films due to the aggregation of the incorporated nanoparticles has been becoming one of the most concerned issues in the development of high-performance organic/inorganic composite optical films. After all, maintaining the intrinsic optical transparency and low color parameters of the pristine optical films is always the prerequisite in practical applications. Nano-sized silica fillers have been one of the most important functional fillers to compensate the defects of the common polymeric films, such as the relatively poor dimensional stability at elevated temperatures, the low modulus, the low flame retardancy, and the so on. However, the common silica nanoparticle usually aggregated in the optical film matrixes, especially at the high loading amounts. Thus, special surface treatments and additional dispersion treatments are usually required during the fabrication of the nanocomposite films. Even by these methodologies, the loading proportions of the nano-fillers in the final composite films were usually controlled at a relatively low level (<10 wt%). By contrast, colloidal silica (cSiO2) or silica sols nanoparticles dispersed in water or organic solvents exhibited good dispersion and distribution in the polymer films. Usually, the functionalization of the SiO2 surface introduces organic groups that make the colloidal SiO2 particles surface-active and sterically stable. The particles are amorphous, but the particle surface is composed of silanol groups, which is the hydroxyl group connected to the silicon atom. The increased surface activity of SiO2 particles enables them to act as an emulsifier. In general, silica sols are an aqueous suspension of colloidal SiO2 particles. In order to endow the silica sols good compatibility with the ABTFCPI-0 matrix, the commercially available aqueous cSiO2 dispersion was modified by solvent replacement in the current work, in which the water was replaced by DMAc.
Figure 7 and
Figure 8 show the UV-Vis spectra and the 3D CIE Lab optical parameters of the ABTFCPI films and the detailed data are summarized in
Table 2. As expected, both of the pristine ABTFCPI-0 and the nanocomposite films showed good optical transparency and low yellow indices. Incorporation of the cSiO
2 nanoparticles didn’t deteriorate the optical properties of the derived films even at the high loading contents of 35 wt%. As shown in the figures and
Table 2, the optical transmittance of the ABTFCPI films at the wavelength of 400 nm (T
400) showed a trend of increasing first and then decreasing with the incorporation of silica nanoparticles. For example, the T
400 value of the pristine ABTFCPI-0 film was 81.8%, while the ABTFCPI-5 film containing the 5 wt% of silica fillers showed the value of 82.0%. With the further increase of the nanoparticles, the T
400 values of the nanocomposite film gradually decreased, and the T
400 value of the ABTFCPI-35 film was as low as 77.2%. This shows that although the introduction of silica nanofillers maintained the intrinsict optical transparency of the nanocomposite films, a certain degree of aggregation still occurred at the high loading contents of the nanoparticles, making the optical transmittance of the films slightly deteriorated. As can be seen from the CIE Lab parameters shown in
Figure 8, the brightness (L
*) values of the nanocomposite films slightly increased with the increasing of the silica fillers, but the yellow indices (b
*) also increased, showing a trend of yellowing for the films. The haze values of the nanocomposite films remained at a relatively stable level. It can be concluded seen that the nanocomposite films could maintain the inherent optical characteristics of the pristine ABTFCPI-0 film when the silica contents were below 30 wt%.
3.3. Thermal properties
The thermal properties of the ABTFCPI films, including the 5% weight loss temperatures (T
5%), the residual weight ratios at 750
oC (R
w750) the glass transition temperatures (T
g), and the linear coefficients of thermal expansion (CTE) were investigated by TGA, DMA, and TMA measurements, respectively. The thermal properties data are tabulated in
Table 2.
First, the TGA and the derivative TG (DTG) plots of the samples are illustrated in
Figure 9. All the ABTFCPI films showed good thermal stability before 450
oC. The films began to thermally decompose around 500
oC and revealed the similar T
5% values in the range of 488~494
oC. According to the DTG plots, the most rapid thermal decomposition occurred in the range of 500~525
oC. At 750
oC, the samples maintained 47.8~64.7 % of their original weights. From the above thermal data, it can be deduced that incorporation of the cSiO
2 nanoparticles had little effects on the initial thermal decomposition behaviors; however could efficiently increase the R
w750 values of the composite films. This is mainly ascribed to the high thermal stability of the inorganic cSiO
2 fillers at elevated temperatures.
Secondly, the T
g values of the ABTFCPI films were determined by the DMA measurements, as shown in
Figure 10. It can be seen that, the ABTFCPI films maintained most of the initial storage and loss modulus up to 350
oC, after which the modulus dramatically decreased. As for the tanδ plots, the inflection point temperatures of the samples were not detected before 400
oC, which is the upper limitation of the DMA equipment in the current work. The peak temperatures of the loss modulus plots were recognized as the T
g values of the films. According to this definition, the ABTFCPI-0 film showed the T
g of 390.3
oC and the nanocomposite films exhibited similar T
g values, indicating the little effects of the cSiO
2 nanoparticles on the glass transition behaviors of the films. This also verified the thermal stability of inorganic silica fillers at elevated temperatures. The high-T
g features of the current ABTFCPI films were mainly attributed to the rigid-rod benzanilide and biphenylene units in the polymers. Strong chemical bonds, such as hydrogen bonds might form among the ABTFCPI molecular chains, which efficiently prohibited the free motions of the molecular segments in the polymers at elevated temperatures.
At last, the high-temperature dimensional stabilities of the ABTFCPI films were evaluated by the TMA measurements. Figure 11 depicts the TMA curves of the ABTFCPI films together with the expanded image in the temperature range of 360~420
oC. The pristine ABTFCPI-0 film showed a CTE value of 32.7×10
-6/K in the range of 50~250
oC (
Table 2). This value ranked lower in all of the semi-alicyclic PI films, which are also the effects of the rigid-rod benzanilide and biphenylene units in the polymers. Incorporation of the rigid amide or benzanilide linkages have been widely used to develop CPI films with the low-CTE features. When incorporating with the cSiO
2 nanoparticles, the CTE values of the CPI films further decreased. For example, ABTFCPI-30 film showed a CTE value of 25.4×10
-6/K, which was 22.3% lower than that of the pristine ABTFCPI film. In addition, it can be seen from the inserted picture that the ABTFCPI-25 and ABTFCPI-30 films with the higher loading contents of cSiO
2 fillers showed much lower dimensional expansion at elevated temperatures. Undoubtedly, the inorganic cSiO
2 fillers efficiently increased the dimensional stability of the nanocomposite films. On one hand, this might be contributed to the good dispersion and distribution of the cSiO
2 nanoparticles. On the other hand, the possible strong interactions between the hydroxyl (–OH) groups on the surface of the silica fillers and the amide bond (–CONH–) in the molecular structure of the ABTFCPI matrix film may also contribute to the reduction of CTE.
Figure 13.
TMA curves of PI films.
Figure 13.
TMA curves of PI films.