Structure and Gas Barrier Properties of Polyimide Containing a Rigid Planar Fluorene Moiety and an Amide Group: Insights from Molecular Simulations.

A novel diamine (FAPDA) bearing rigid planar fluorene and amide groups was successfully synthesized. Using such diamine and pyromellitic dianhydride (PMDA), a high-barrier polyimide (FAPPI) was obtained. FAPPI exhibits an outstanding gas barrier. Its water vapor transmission rate (WVTR) and oxygen transmission rate (OTR) are as low as 0.51 g·m-2·day-1 and 0.43 cm3·m-2·day-1, respectively. Additionally, FAPPI shows excellent thermal stability with a coefficient of thermal expansion (CTE) of 5.8 ppm·K-1 and a glass transition temperature (T g) of 416 °C. Molecular simulations, positron annihilation, and X-ray diffraction were utilized to gain insight on the microstructures for the enhanced barrier properties. Introducing fluorene moieties and amide groups improves the regularity and rigidity of molecular chains and increases interchain interaction of PI, resulting in low free volumes and decreased movement capacity of the chain. The low free volumes of FAPPI restrain the gas diffusivity and solubility. Meanwhile, the decreased chain movement reduces the diffusivity of gases. Consequently, barrier performances of FAPPI are improved. The polyimide possesses widespread application in the microelectronics packaging fields.

Introduction

Active-matrix organic light-emitting diode (AMOLED) displays have garnered great interest owing to their self-emitting properties, wide-angle view, low power consumption, and high-speed video rate.[1,2] The substrate is a vital component in AMOLED displays. In order to achieve wearable electronic devices, polymer films are often chosen as the substrate in flexible AMOLEDs due to their high flexibility, light weight, and outstanding robustness.[3] With widespread application of top-emitting AMOLEDs, transparence is not a necessity for the substrate.[4] Nevertheless, an exceptional barrier to oxygen and moisture, high thermal stability, and good dimensional stability are essential for the substrate. This is because oxygen and moisture from the external environment can crystallize and oxidize the metallic cathode and organic materials in devices, causing dark spots and performance deterioration.[5] In addition, AMOLED assembly, such as deposition of the electrode, coating of the barrier layer, and thin-film transistor (TFT) manufacture, is usually conducted at high temperature.[6] For example, the substrates are required to undergo high temperature (>400 °C) in the fabrication of low-temperature poly-silicon (LTPS) TFTs.[7] Furthermore, excellent dimensional stability of the substrates is helpful for large generation of the LTPS drive backplane.[8] Hence, ideal substrate films suitable for AMOLED displays should achieve the demands of exceptional barrier performances (OTR < 10–5 cm3·m–2·day–1 and WVTR <10–6 g·m–2·day–1), superior thermal properties (Tg ≥ 400 °C), and low CTE (≤ 8 ppm·K–1).[7,9−11] Numerous polymer substrates like polyethylene naphthalate (PEN), polycarbonate (PC), and polyethylene terephthalate (PET) were studied.[12−14] However, their thermostability (Tg < 200 °C), dimensional stability (20–80 ppm·K–1 for CTE), and gas barrier (10–102 cm3·m–2·day–1 of OTR) still need to be enhanced.[15]

Recently, polyimides (PIs) have attracted significant attention as flexible substrates due to their exceptional thermal and mechanical performances.[16] However, most PIs usually show an OTR and WVTR of 101–103 cm3·m–2·day–1 and 101–103 g·m–2·day–1, respectively, which cannot fulfill the application requirement of substrates for display devices.[17,18] To enhance the gas barrier of polyimide substrates, formation of multiple barrier layers on the PI films is usually adopted.[19] However, the coating process requires many steps and specialized equipment, enhancing costs. Obviously, the coating process is based on the intrinsic polyimides. Therefore, improving the barrier performances of intrinsic polyimides is an effective method to reduce the thicknesses of the barrier layer, simplify the coating process, and lower the cost.[20]

Notably, the gas permeation of polymeric materials is highly dependent on their polymer chain structure, aggregation structure, and the interchain interaction.[21] Therefore, barrier performances of PI can be enhanced by altering the chemical structures. In our previous study, a fluorene-containing diamine (FDA, Figure ) was prepared, which was then polycondensed with PMDA to produce PI (FPI).[22] The existence of fluorene in the backbones enhances the chain regularity and rigidity, thus increasing chain stacking and thereby the barrier properties.

Figure 1

Structures of FDA, FAPDA, ODA, and PMDA.

Introducing amide into the polymer chain can raise the interchain force by forming hydrogen bonds. High interchain interaction can promote tight packing of polymer chains and inhibit chain movement, which are advantageous to the enhancement of barrier properties. In light of this, a novel diamine (FAPDA) possessing both a rigid planar fluorene moiety and amide groups was prepared in the study. FAPDA was then reacted with PMDA to yield a new polyimide (FAPPI). Here, the structures of FAPDA and PMDA are presented in Figure . For FAPDA, the fluorene moieties give rise to rigid planar and large rings in the backbone, which is connected to two benzene rings by two amide groups. The fluorene could increase the stacking efficiency of molecular chains. Furthermore, hydrogen bonds could be formed by amide groups to raise the interchain force. Introducing the two groups at the same time was beneficial for the great increase of the gas barrier and thermostable performances of PIs.

In this study, the synthesis and characterizations of FAPPI and its barrier and thermal and mechanical performances were studied. In addition, the barrier properties of FAPPI, Kapton (a typical PI, based on 4,4′-oxydianiline (ODA) and PMDA), and our previously studied analog (FPI, bearing only fluorene in the backbone) were compared. Molecular simulation is a promising means to analyze the structure and gas transport of polymeric materials from the microscopic point of view,[23,24] which are hard to realize by experimental means. To unravel the relationship between polymer structures and barrier properties, molecular simulations, positron annihilation, and X-ray diffraction were employed to investigate the aggregation structure, hydrogen bond interaction, free volume, chain movement capacity, gas diffusion, and solubility of PI, which are beneficial for understanding the barrier mechanism and guiding the design of high-barrier polymers.

Experimental Section

Materials

The materials are given in the Supporting Information.

Instrumentation

The test instruments are displayed in the Supporting Information.

Synthesis of N,N′-(9H-Fluorene-2,7-diyl)bis(4-nitrobenzamide) (FAPDN)

9H-Fluorene-2,7-diamine (FDA) was synthesized through reduction reaction using FDN as the raw material in accordance with our preceding study.[22] FDA (1.963 g, 10 mmol), Py (5.3 mL, 66 mmol), and DMAc (60 mL) were poured into a flask. For removing oxygen, the solution was stirred for 30 min under argon. Afterward, 4-nitrobenzoyl chloride (4.082 g, 22 mmol) was added and stirred for 10 h at room temperature under argon. Subsequently, the solution was heated to 80 °C with stirring for 6 h. After that, the mixture was placed in distilled water. Filtration was performed to collect the precipitate. For removing solvent and impurities, the precipitate was then washed with 600 mL of hot water. The obtained light yellow product was subsequently dried for 24 h at 80 °C. Yield: 86%. IR (KBr, v, cm–1): 1529 (stretching of −NO2), 1644 (stretching of C=O), 1263 (stretching of C–N), 3309 (stretching of N–H), 1100–700 (stretching of Ar–H). 1H NMR (DMSO-d6, 400 MHz, δ, ppm): 10.64 (s, 2H), 8.38 (d, J = 8.8 Hz, 4H), 8.22 (t, J = 11.0 Hz, 4H), 8.08 (s, 2H), 7.94 (DMF solvent peak), 7.84 (d, J = 8.2 Hz, 2H), 7.75 (d, J = 8.3 Hz, 2H), 3.98 (s, 2H). 13C NMR (100 MHz, DMSO-d6, δ, ppm): 163.82, 149.15, 143.68, 140.75, 137.34, 137.10, 129.23, 123.61, 119.76, 119.35, 117.29, 36.72 (DMF solvent peak: 162.33, 35.80, and 30.78). MS (EI, m/z): 495(100) ([M+ + H], calcd for C27H18N4O6, 494.12). Anal. Calcd for C27H18N4O6: C, 65.59; H, 3.67; and N, 11.33; found: C, 65.39; H, 3.58; and N, 11.45.

Synthesis of N,N′-(9H-Fluorene-2,7-diyl)bis(4-aminobenzamide) (FAPDA)

The mixture of FAPDN (2.472 g, 5 mmol) and ethanol (450 mL) was placed into a flask with stirring for 30 min under an argon atmosphere to remove oxygen. After heating to 80 °C, 0.02 g of Pd/C and 5 mL of hydrazine monohydrate were added with stirring for 40 h at 80 °C. After cooling down, filtration was performed to remove Pd/C. Then, diamine (FAPDA) was obtained by crystallization at low temperature. Yield: 93%. IR (KBr, v, cm–1): 3328 (stretching of N–H), 3213 (stretching of −NH2), 1603 (δ N–H), 1651 (stretching of C=O), 1256 (stretching of C–N), 1100–700 (stretching of Ar–H). 1H NMR (DMSO-d6, 400 MHz, δ, ppm): 9.86 (s, 2H), 8.07 (s, 2H), 7.88–7.63 (m, 8H), 6.65 (d, J = 8.6 Hz, 4H), 5.77 (s, 4H), 3.92 (s, 2H). 13C NMR (100 MHz, DMSO-d6, δ, ppm): 165.72, 152.56, 143.74, 138.66, 136.69, 129.83, 121.76, 119.71, 119.43, 117.45, 113.07, 37.12. MS (EI, m/z): 435(100) ([M+ + H], calcd for C27H22N4O2, 434.17). Anal. Calcd for C27H22N4O2: C, 74.64; H, 5.10; and N, 12.89; found: C, 74.51; H, 5.03; and N, 12.93.

Synthesis of PIs

FAPPI was prepared according to the following processes in a clean room. DMF (10 mL) and FAPDA (0.6222 g, 1.432 mmol) were placed into a flask with stirring. Then, PMDA (0.3123 g, 1.432 mmol) was put in, as FAPDA was dissolved. After stirring at 0 °C for 8 h, poly(amic acid) (PAA) was generated, which was defoamed and coated onto glass plates. Then, PAA was thermal-imidized by the following program: 100 °C/100–200 °C/200–300 °C/300–400 °C for 1 h. After that, the FAPPI film was obtained and removed from the substrate. IR (KBr, v, cm–1): 1372 (stretching of C–N), 1775 and 1716 (stretching of C=O), 1100–700 (stretching of Ar–H). Anal. Calcd for C37H20N4O6: C, 72.08; H, 3.27; and N, 9.09; found: C, 71.52; H, 3.52; and N, 9.21. Kapton films were prepared based on ODA and PMDA using a similar method.

Molecular Simulations

Periodic models of PIs containing five polymer chains were built, in which each polymer chain had 25 repeat units. Based on the equilibrated models, the free volumes, radial distribution functions (RDFs), radius of gyration (Rg), movement capacity of chains, gas solubility, and diffusion were investigated. The construction of the model, simulation processes, and descriptions of properties are displayed in the Supporting Information.

Results and Discussion

Synthesis and Characterizations of Monomers and Polyimide

FAPDA was synthesized according to the route in Scheme . FDA was prepared from FDN by reduction reaction firstly. Then, the as-obtained FDA further reacted with 4-nitrobenzoyl chloride to yield FAPDN through the amidation reaction. Eventually, FAPDN was reduced to diamine FAPDA. The structures of the diamine monomer and intermediate were examined using mass spectra, NMR, FTIR, and elemental analysis. The NMR and MS results of the intermediate FAPDN and MS of FAPDA are given in Figures S3 and S4, respectively. The NMR spectrum of FAPDA is illustrated in Figure . The FTIR spectra of FAPDN and FAPDA are presented in Figure S5. Elemental analysis data of FAPDN and FAPDA are given in the Experimental Section. These results agreed with the molecular structures of FAPDN and FAPDA, verifying that the predesigned diamine was successfully synthesized.

Scheme 1

Synthesis Route of FAPDA

Figure 2

(a) 1H NMR, (b) 13C NMR, (c) H–H COSY, (d) C–H QC, and (e) C–H BC spectra of FAPDA in DMSO-d6.

The PI was synthesized according to the two-step method (Scheme ). The two-step procedure included the reaction of PMDA and FAPDA to produce FAPPAA and thermal imidization of FAPPAA to produce FAPPI. The molecular weights of FAPPAA were measured by GPC. Here, the weight-average molecular weight (Mw) was 2.43 × 105, and the polydispersity index (Mw/Mn) was 1.87. The chemical structures of PI were characterized using FTIR as shown in Figure S5. Compared with FAPDA, the peaks at 1603 (δ N–H) and 3200–3500 cm–1 (stretching of N–H) disappeared for FAPPI; meanwhile, the imide peaks at 1372 (C–N stretching) and 1775, 1716 cm–1 (carbonyl stretching) emerged, indicating successful reactions of PMDA and FAPDA and complete imidization of FAPPAA.

Scheme 2

Synthesis Route of Polyimide

Mechanical and Thermal Performances

The thermostability properties of FAPPI are investigated and exhibited in Table . FAPPI had an extremely high Tg of 416 °C (Figure a). In addition, FAPPI demonstrated superior dimensional stability. The CTE was low, 5.8 ppm K–1 (Figure b). Moreover, the TGA analysis showed that FAPPI exhibited good thermostability. The Td5% and Td10% values were 525 and 558 °C, respectively (Figure S6). FAPPI displayed a tensile strength and tensile modulus of 132 MPa and 5.2 GPa, respectively, demonstrating its good mechanical properties (Table ). FAPPI possessed favorable flexibility, as shown in Figure S7. The superior thermal properties and favorable mechanical strength of PI may be due to its rigid structure and high interchain force, which allowed PI to withstand the high processing temperatures in the flexible display industries.

Table 1

Thermal and Mechanical Properties of the FAPPI Film

PI	Tga (°C)	Td5% (°C)	Td10% (°C)	CTEb (ppm·K–1)	tensile strength (MPa)	tensile modulus (GPa)
FAPPI	416	525	558	5.8	132 ± 2.1	5.2 ± 0.2

Measured by DMA.

CTE within the range of 50–200 °C.

Figure 3

(a) DMA and (b) TMA plots of the FAPPI film.

Barrier Performances

For comparison, the gas barriers of FAPPI, Kapton (a typical PI), and our previously reported FPI possessing only fluorene in the backbone are listed in Table . As for Kapton, the diamine ODA possesses ether groups (Figure ). In terms of FAPPI, the WVTR and OTR were as low as 0.51 g·m–2·day–1 and 0.43 cm3·m–2·day–1, respectively, demonstrating its outstanding barrier properties. Furthermore, the WVTR and OTR lowered by 2 and 3 orders of magnitude, respectively, compared with those for Kapton. Especially, FAPPI also displayed superior barrier performances to FPI containing a fluorene moiety in main chains, with 1 order of magnitude reduction in the WVTR and OTR. Apparently, introducing fluorene and amide groups at the same time markedly enhanced the barrier performances of FAPPI. Moreover, the resulting FAPPI presented superior barrier performances to most studied PIs (an WVTR of 101–103 g·m–2·day–1 and OTR of 101–103 cm3·m–2·day–1)[17,18] and other high-barrier polymeric materials (polyvinylidene chloride, PET, ethylene-vinyl alcohol copolymer, and nylon-6).[25−29] The great barrier improvement of PI is of great significance for decreasing the barrier layer thickness on PI substrates and simplifying coating processes.

Table 2

Barrier Properties of the Kapton, FPI, and FAPPI Films

PI	WVP (g·mil·m–2·day–1)	OP (cm3·mil·m–2·day–1)	WVTR (g·m–2·day–1)	OTR (cm3·m–2·day–1)
Kapton	115.63 ± 2.08	331.88 ± 2.50	39.16 ± 0.70	112.40 ± 0.84
FPIa	6.94	2.98	2.35	1.01
FAPPI	1.51 ± 0.10	1.27 ± 0.08	0.51 ± 0.03	0.43 ± 0.03

The barrier property values are obtained from ref (22).

Aggregation Structure Analysis

To understand the mechanism of the improved barrier performances, the aggregation structure and morphologies of the PI were studied using experimental WAXD. As shown in Figure , the FAPPI presented an intense and narrow diffraction peak (2θ = 20.93°). Comparatively, Kapton with a relatively lower intensity at 2θ = 18.53° was observed. The interchain distance (d-spacing) can be determined from the peak in the WAXD curve.[30] The calculated d-spacing for Kapton was 4.78 Å, which was decreased to 4.24 Å for FAPPI (Table ). The chain stackings of FAPPI and Kapton were also studied by RDFs. Figure S8a,b shows the interchain RDFs on the basis of C atoms in the benzene ring and N atoms in the imide ring, respectively. 4000 snapshots of equilibrated models were used, and averaged g(r) was reported. In Figure S8b, each PI exhibited a diffraction peak at ∼4.5 Å, suggesting an average spacing of the interchain in the neighborhood of 4.5 Å, which was comparable to the d-spacing values from experimental WAXD. The g(r) value of FAPPI was larger than that of Kapton, as shown in Figure S8a,b. Normally, a higher g(r) value suggests the larger number of nearest atoms. Herein, the polymer chains of FAPPI were more tightly stacked than those of Kapton, consistent with the WAXD results. FAPPI also exhibited higher density than Kapton (Table ). The low interchain distance and high density of FAPPI revealed its tight chain packing, which was probably a consequence of the chain stacking ability caused by the diamine structures in PIs.

Figure 4

WAXD curves of the FAPPI and Kapton films.

Table 3

Physical Properties of the FAPPI and Kapton Films

PI	density (g·cm–3)	2θ (°)	d-spacing (Å)	Rg (Å)	CED (J·cm–3)
Kapton	1.42	18.53	4.78	41.55	454
FAPPI	1.52	20.93	4.24	77.10	647

Chain Morphologies

The morphologies of FAPPI and Kapton chains were investigated by molecular simulations to understand the chain packing ability. Figure shows the equilibrated conformation of a single chain containing 10 repeat units for FAPPI and Kapton. FAPPI showed relatively stretched and linear chain structures, which favored the compact polymer chain packing and ordered region formation. However, the curved chain structure of Kapton hindered tight chain packing. Additionally, the differences of chain morphologies in the two PIs were also revealed by the radius of gyration (Rg) results. The R vs time curves for the equilibrated Kapton and FAPPI are shown in Figure S9. Table gives the averaged Rg data in the time range of 0–300 ps. FAPPI showed a larger Rg value than Kapton, indicating the relatively expanded chain structure.[31] This was mostly due to the presence of rigid planar fluorene, increasing the regularity and rigidity of molecular chains. Based on the above analysis, it can be inferred that the effective chain packing of FAPPI was mainly due to two factors. One was the regular and rigid polymer backbone of FAPPI resulting from the rigid planar fluorene moieties in main chains. The other was the high hydrogen bonding force in the FAPPI matrix, which will be discussed below.

Figure 5

Stereo view of the lowest energy conformations of the FAPPI and Kapton chains.

Hydrogen Bond Analysis

In addition to chain morphologies, the interchain cohesion has a great influence on chain stacking. RDFs were utilized to analyze the formed hydrogen bonding in PI. Theoretically, two kinds of hydrogen bonding can be generated in the FAPPI matrix: type A (between −NH– and O=C– of amide groups) and type B (between −NH– and O=C– of imide rings); however, hydrogen bonding cannot be formed for Kapton. The RDFs of FAPPI were calculated by simulations based on the H atom of −HN– and O atom of O=C– in amide and imide, as illustrated in Figure . Hydrogen bonding interaction was related to the atom distances of 2.6–3.1 Å.[32] In Figure , two sharp peaks appeared at approximately 2.6 and 3 Å, suggesting the existence of hydrogen bonds in FAPPI. Figure displays the hydrogen bonds in the FAPPI cell, verifying that hydrogen bonds were generated. This agreed with RDFs and theoretical analysis. 100 hydrogen bonds were found in the FAPPI cell through statistical analysis. The simulation cell of Kapton showed no hydrogen bonding. The cohesive energy densities (CEDs) of the two PIs are given in Table . Compared with Kapton, FAPPI showed a greatly higher CED value due to the hydrogen bonds formed by amide groups.

Figure 6

Radial distribution function of FAPPI for the hydrogen atoms of −HN– and oxygen atoms of O=C– in amide groups and imide rings.

Figure 7

H-bonds in the simulation cell of FAPPI: (a) H-bonds between −HN– and O=C– in amide groups; (b) H-bonds between −HN– and O=C– in imide rings.

Free Volume Analysis

Positron Annihilation Technique

The chain morphologies, stacking, and interactions influence free volume in PI matrices, which play a substantially important role in gas permeation properties. Then, for the two PI films of FAPPI and Kapton, the free volumes were investigated by positron annihilation lifetime spectroscopy (PALS). The obtained positron lifetime plots of Kapton and FAPPI are illustrated in Figure . FAPPI showed a higher slope of decay plot than Kapton, suggesting the smaller free volumes in FAPPI. The obtained data were resolved into two decay components through PATFIT programs. The values of positron lifetime are listed in Table . The FAPPI and Kapton films showed two decay components (τ1 and τ2). The intensity (I) corresponding to decay components involves the annihilation number occurring at a particular lifetime.[33] Jean et al. proposed a correlation between τ2 and the mean radius (R) of free volumes in polymers.[34] Based on the correlation, the R values of FAPPI and Kapton are determined and presented in Table . According to our previous work, the size (Vf2) of free volumes and relative fractional free volume (FFV) are also calculated (Table ).[22] The Vf2 of Kapton was 73.58 Å3, which was reduced to 41.03 Å3 in the FAPPI film. Correspondingly, the FFV of 12.82% in Kapton was also decreased to 6.35% in FAPPI.

Figure 8

Positron lifetime spectra measured for the Kapton and FAPPI films.

Table 4

Analyzed Data for the Positron Lifetime and Simulated FFV Values in the Kapton and FAPPI Films

PI	τ1 (ns)	I1 (%)	τ2 (ns)	I2 (%)	R (Å)	Vf2 (Å3)	FFVa (%)	FFVb (O2, %)	FFVb (H2O, %)	FFV0b (%)
Kapton	0.17	1.9	0.38	96.8	2.60	73.58	12.82	10.48	15.80	38.51
FAPPI	0.15	13.8	0.34	86.0	2.14	41.03	6.35	3.08	7.24	33.26

FFV determined by PALS.

Free volume parameters calculated by simulations, FFV(O2), FFV(H2O), and FFV0 based on probe radii of 1.73, 1.325, and 0 Å, respectively.

Free Volume Characteristics by Molecular Simulations

Molecular simulations can offer information about size distributions, connectivity, and arrangement of free volumes in polymers, which are vital elements affecting gas transport performance. The distribution of cavity size in Kapton and FAPPI by molecular simulations is shown in Figure a. It can be observed that the number of voids with a radius larger than 1.2 Å in FAPPI was fewer than that in Kapton. H2O and O2 possessed kinetic radii of 1.325 and 1.73 Å, respectively, which are bigger than 1.2 Å. This meant that the voids available for O2 and H2O to diffuse in FAPPI were fewer than those in Kapton.[35]

Figure 9

Void size distributions (a) and FFVs (b) as a function of probe size in Kapton and FAPPI. The kinetic radii of O2 and H2O are shown by vertical lines.

To analyze the changes of available free volume in relation to the penetrant size, the FFV was determined by a spherical probe with different radii. The relationships between FFV and probe radius for Kapton and FAPPI are given Figure b. The total FFV (FFV0) for Kapton and FAPPI was determined using a probe radius of 0 Å (Table ). FFV decreased with the increase of the probe radius. Evidentially, FAPPI showed lower FFV than Kapton in the studied range. The FFV values accessible for H2O and O2 in FAPPI and Kapton are listed in Table . The H2O and O2 accessible FFV values and FFV0 of FAPPI were much smaller than those in Kapton, in accordance with the PALS results. Moreover, the H2O and O2 accessible volumes for FAPPI and Kapton are illustrated in Figure . It is observed that the voids in Kapton were larger and well-connected while those in FAPPI were smaller and discontinuous. The free volumes available for H2O and O2 were lower in FAPPI. This was a consequence of its effective chain packing due to the regular and rigid polymer chains and high hydrogen bonding force of FAPPI. The discontinuous and lower free volumes in the FAPPI matrix led to low gas permeability. Consequently, the barrier performances of FAPPI were enhanced.

Figure 10

H2O and O2 accessible volumes for (a1,a2) FAPPI and (b1,b2) Kapton (gray: van der Waals surface; blue: Connolly surface; (a1,b1): a probe radius of 1.325 Å; (a2,b2): a probe radius of 1.73 Å).

Polymer Chain Movement

The movement capacity of polymer chains also influences the gas penetration in polymeric materials. To investigate the movement capacity of the PI chain, the mean square displacements (MSDs) of PI chains are shown in Figure . The MSD of FAPPI was lower than that of Kapton, indicating that FAPPI had lower chain mobility than Kapton. This also contributed to the good gas barrier of FAPPI. The chain movement affected the channel formation for gases to diffuse.[36] The fluorene moiety in the FAPPI backbone increased the chain rigidity. Additionally, the hydrogen bonding formed by amide groups enhanced the interchain cohesion. The aforementioned two factors led to poor chain mobility of FAPPI and thus decreased the extent of hole formation for gas permeation, bringing about the high gas barrier.

Figure 11

Polymer chain MSD vs time in FAPPI and Kapton.

Gas Permeation

For revealing the mechanism of the gas barrier, the gas permeation of FAPPI and Kapton was researched, including gas diffusivity and solubility.

Gas Diffusivity

The trajectories of H2O and O2 in Kapton and FAPPI are presented in Figure . Also, the displacements of H2O and O2 in Kapton and FAPPI are given in Figure S10. The plots of FAPPI are shifted up for clear visualization (Figure S10). From Figure S10 and Figure , the transport mode of H2O and O2 in PI matrices involves oscillation in holes and a jump from a hole to the adjacent hole.[37] Obviously, the trajectory lengths and jump frequencies of H2O and O2 molecules in FAPPI were smaller than those in Kapton. In addition, O2 showed a larger jump frequency and moved a larger distance than H2O in a particular PI.

Figure 12

O2 and H2O trajectories in (a) Kapton and (b) FAPPI.

To obtain the diffusion coefficients, the MSD vs time curves of log scale for O2 and H2O in the two PIs are displayed in Figure . Using the Einstein equation, the diffusion coefficient was obtained from the regime of normal diffusions, in which the slope of the MSD plot in log scale was unity.[37]Table displays the O2 and H2O diffusion coefficients in the two PIs. Compared with Kapton, lower H2O and O2 diffusion coefficients of FAPPI were observed. This was because the smaller free volume cavities, poor hole connectivity, and the low chain movement limited O2 and H2O to diffuse in FAPPI. As discussed before, the diffusion coefficient of H2O was smaller than that of O2 for a given PI. Despite the smaller diameter of H2O, the diffusion coefficient of H2O was smaller than that of O2. This is probably attributed to the high interaction of PI and H2O,[38] hampering the diffusion of H2O.

Figure 13

O2 and H2O MSD vs time in Kapton and FAPPI.

Table 5

Simulated permeability, diffusion, and solubility coefficients of H2O and O2 in FAPPI and Kapton

	Da		Sb		Pc
PIs	H2O	O2	H2O	O2	H2O	O2
Kapton	12.8	89.4	3.92	0.056	50.18	5.0
FAPPI	4.3	8.5	0.12	0.009	0.52	0.077

Units of (10–8 cm2·s–1).

Units of (cm3(STP)·cm–3·cm Hg–1).

Units of (10–8 cm2·cm3(STP)·s–1·cm–3·cm Hg–1).

Gas Solubility

The adsorption isotherms of H2O and O2 in FAPPI and Kapton were studied and are illustrated in Figure . The adsorption amounts increased rapidly at low pressures, and the increase trend tended to be flat at high pressures. The adsorption mode at low pressures indicated the presence of microporosities in the polymers.[39] Apparently, the initially rapid uptake of Kapton was higher than that of FAPPI, demonstrating that Kapton had more microvoids. As the micropore adsorption achieved saturation, the gases started to adsorb on the interchain free volumes.[39] The sorption of H2O and O2 fitted well with the dual mode. Table lists the obtained solubility coefficients. FAPPI presented lower solubility coefficients of H2O and O2 than Kapton. This was mainly because the lower free volume of FAPPI brought about fewer sites for gas sorption and therefore the low solubility coefficients. Additionally, the solubility coefficient of H2O was larger than that of O2 as for a particular PI, contrary to the diffusion coefficient trend. The higher affinities between H2O and PI matrix contributed to the adsorption of H2O in the PI matrix.[38] In addition, H2O had higher critical temperature (O2: 155 K, H2O: 647 K), leading to the larger solubility than O2. Moreover, H2O had more adsorption locations because of its small size. The above three factors led to the larger solubility coefficient of H2O.

Figure 14

Adsorption isotherms of H2O and O2 in FAPPI and Kapton.

Gas Permeability

The permeability coefficients (P) was the product of diffusion coefficients with solubility coefficients based on the solution-diffusion principle. The P values of H2O and O2 in FAPPI and Kapton are presented in Table . The simulated permeability displayed the same trend with the experiment values (Table ), proving the viability of simulation in gas penetration studies. In comparison with Kapton, both the solubility and diffusion coefficients of H2O and O2 in FAPPI were reduced by 1 order of magnitude, which in turn led to 2 orders of magnitude reduction for permeabilities. As a consequence, the barrier performances of FAPPI were greatly enhanced. The excellent gas barrier of FAPPI was explainable by the low free volumes and poor chain movement.

Conclusions

A diamine (FAPDA) containing fluorene and amide was successfully synthesized. The polyimide (FAPPI) based on such diamine and PMDA was then prepared. The polyimide presented remarkable barrier performances with WVTR and OTR as low as 0.51 g·m–2·day–1 and 0.43 cm3·m–2·day–1, respectively. Also, the FAPPI presented outstanding dimensional and thermal stability. The exceptional barrier performances associated with polymer structures were investigated using molecular simulation, PALS, and WAXD. Research on aggregation structures showed that FAPPI possessed low interchain distance, indicating its tight chain stacking. Molecular simulations indicated the regular and rigid polymer chains and high interchain hydrogen bonding of FAPPI by introducing fluorene and amide, which led to its tight chain stacking. According to free volume analysis, FAPPI possessed less connected and smaller free volumes, decreasing the solubility and diffusion of H2O and O2 in the polymer matrix. The MSD results of PI chains presented that FAPPI showed poorer chain mobility, which also decreased gas diffusion. The decreases of solubility and diffusion of H2O and O2 in FAPPI brought about the reduction in the permeability coefficient, i.e., improvement of the gas barrier. The PI with excellent overall performance has potential application in the field of microelectronics encapsulation.