Influence of diamine structure on the properties of colorless and transparent polyimides

2.1 Materials

Butane-1,4-diyl bis(1,3-dioxo-1,3-dihydroisobenzofuran-5-carboxylate) (BU-DA) was obtained from Lum. Tech. Co. (Taipei, Taiwan). (The nine diamine monomers – m-XDA, p-XDA), 4,4′-ODA, m-APS, m-BAPS, TFB, m-TFAB, BAPP, and 6FBAPP – were purchased from TCI (Tokyo, Japan) and used without further purification. N,N′-Dimethylacetamide (DMAc), purchased from Sigma-Aldrich (Yongin, Korea), was dried thoroughly to remove moisture before use.

2.2 Synthesis of CPI Films

Scheme 1 illustrates the synthesis route of CPI films via a two-step process involving PAA formation followed by thermal imidization. As the procedures for all nine diamine monomers were essentially identical, the synthesis using m-XDA is described here as a representative example: BU-DA (5.698 g, 1.3 × 10⁻² mol) was completely dissolved in 50 mL of dried DMAc under stirring for 30 min. Then, m-XDA (1.770 g, 1.3 × 10⁻² mol) was added to the BU-DA solution, and the mixture was stirred at 0 °C under a nitrogen atmosphere for 1 h. The solution was then polymerized at room temperature for 14 h to form the PAA precursor.

Scheme 1:

Synthetic routes for CPIs based on BU-DA monomer.

The resulting PAA solution was cast uniformly onto a clean glass plate, followed by vacuum stabilization at 50 °C for 1 h. The solvent was then gradually removed under vacuum at 80 °C for an additional hour. The PAA film was subsequently thermally imidized at various temperatures under a nitrogen atmosphere, with specific thermal treatment conditions summarized in Table 1. After thermal imidization, the resulting CPI film was immersed in a 5 wt% aqueous hydrofluoric acid (HF) solution to detach it from the glass substrate. The final freestanding CPI films were obtained with dimensions up to 10 × 10 cm².

2.3 Characterization

Fourier-transform infrared (FT-IR) spectroscopy (PerkinElmer, L-300, London, UK) was used to confirm the imidization of CPI by identifying characteristic functional groups in the range of 4,000–1,000 cm⁻¹.

The ¹H and ¹³C magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectra were recorded at 850 MHz and 125.75 MHz, respectively, on a Bruker spectrometer (Germany) at the Laboratory of NMR, NCIRF, Seoul National University. The ¹H and ¹³C chemical shifts were referenced to tetramethylsilane (TMS), and the ¹H NMR experiment was performed using CDCl₃ as the solvent.

Differential scanning calorimetry (DSC, NETZSCH 200F3, Berlin, Germany), thermogravimetric analysis (TGA), and derivative thermogravimetry (DTG) (TA Instruments Q-500, New Castle, USA) were performed under a nitrogen atmosphere with heating and cooling rates of 10 °C min⁻¹. The coefficient of thermal expansion (CTE) was measured using a thermomechanical analyzer (TMA, TA Instruments TMA2940, Seiko, Tokyo, Japan) under a constant load of 0.1 N and a heating rate of 10 °C min⁻¹.

Tensile properties were evaluated using a universal testing machine (UTM, Shimadzu AG-50KNX, Japan) with a crosshead speed of 5.00 mm min⁻¹. Each film was tested at least ten times. Outliers were excluded from the dataset, and the average of the remaining values was reported. Experimental error margins were maintained within ±1 MPa for tensile strength and ±0.05 GPa for initial modulus.

The yellowness index (YI) was measured using a spectrophotometer (KONICA MINOLTA CM-3600d, Tokyo, Japan). Ultraviolet–visible (UV–vis) spectroscopy (SHIMADZU UV-3600, Tokyo, Japan) was used to determine the cut-off wavelength (λ ₀) and optical transmittance at 500 nm (500 nm^trans). To ensure consistency, all CPI films were prepared with a uniform thickness of 42–46 μm.

3.1 FT-IR, ¹³C MAS NMR, and ¹H NMR, analysis

CPIs were synthesized from a single dianhydride, BU-DA, and nine different diamine monomers, as illustrated in Scheme 1. FT-IR spectra of the resulting CPI films are presented in Figure 1. A separate inset was incorporated into Figure 1 to provide a more detailed view of the spectral range between 2000 and 1,300 cm⁻¹. The absorption bands observed at 3,050–2,874 cm⁻¹ correspond to the C–H stretching vibrations originating from aliphatic and aromatic moieties, reflecting the structural characteristics of the aliphatic and aromatic rings within the main chain. In addition, the strong absorption peaks appearing at 1783–1780 cm⁻¹ and 1717–1704 cm⁻¹ are assigned to the characteristic C=O stretching vibrations of the imide ring, corresponding to the asymmetric and symmetric modes, respectively. The peaks in the range of 1,500–1,420 cm⁻¹ are attributed to the C=C stretching vibrations of the aromatic rings, indicating that the synthesized polymers retain the aromatic backbone structure. Finally, the absorptions observed at 1,385–1,364 cm⁻¹ are assigned to the C–N–C stretching vibrations within the imide ring, which represent a typical characteristic band confirming the formation of polyimide. These results clearly demonstrate that the synthesized CPI films possess the intended chemical structure and that successful cyclization reactions were achieved in all compositions [30].

Figure 1:

FT-IR spectra of various CPIs based on BU-DA monomer.

Further confirmation of successful CPI formation was obtained by solid-state ¹³C MAS NMR analysis. Representative spectra of CPI structures I, IV, and VII are shown in Figure 2. In structure I, both the ester carbon (a) and imide carbon (b) were observed at 166.03 ppm. Aromatic ring carbons (c, d, e) appeared at 135.68, 129.65, and 124.56 ppm, respectively. The carbon of the alkyl group adjacent to the ester oxygen (f) was found at 66.45 ppm, while the central aliphatic methylene group (h) appeared at 26.77 ppm. The methylene group adjacent to the aromatic ring (g) gave a signal at 42.18 ppm. These results, along with the FT-IR data, confirm the successful synthesis of the CPI structures [31]. The full NMR spectra of structures IV and VII are provided in the Supplementary Figure I.

Figure 2:

¹³C NMR spectrum of structure I based on BU-DA monomer.

Among the nine synthesized CPIs, the ¹H NMR spectrum of Structure I (Figure 3) shows proton (¹H) signals consistent with the proposed CPI structure, while the spectra of the remaining CPIs are provided in Supplementary Figure II. The aromatic protons adjacent to the imide ring (a) appeared at 7.1–8.5 ppm due to the strong deshielding effect of the carbonyl groups. The methylene protons bonded to the imide nitrogen (b) were observed at 4.7–4.9 ppm, reflecting the influence of the electronegative nitrogen and neighboring carbonyls. The protons adjacent to oxygen or carbonyl groups (c) resonated at 4.3–4.6 ppm. Finally, the aliphatic protons (d) appeared as broad multiplets at 1.5–2.2 ppm, consistent with shielding in the aliphatic environment. These assignments confirm that the proton environments are in good agreement with the designed CPI structure [31].

Figure 3:

¹H NMR spectrum of structure I based on BU-DA monomer.

3.2 Thermal properties

Since CPI is generally amorphous, it does not exhibit a melting point (T _m ) detectable by DSC [32]. Instead, the primary thermal characteristic of interest is the glass transition temperature (T _g ), which is highly sensitive to monomer structure, chain flexibility, bulky substituents, free volume, and inter- and intramolecular interactions such as hydrogen bonding [33].

The measured T _g values of the nine CPI films are summarized in Table 2, and their proposed 3D polymer structures are illustrated in Figure 4. CPI structures I and II showed the lowest T _g values (112 °C and 130 °C, respectively), attributed to the presence of flexible alkyl groups and the bent m-substitution in structure I. Structure III, containing an ether linkage allowing free rotation, exhibited a slightly higher T _g of 146 °C. The relatively low T _g values of structures I–III may also be due to their lower aromatic content, resulting in reduced rigidity.

Figure 4:

Comparison of the CPI three-dimensional chemical structures based on BU-DA monomer.

Structures IV and V, which incorporate the bulky and electron-withdrawing –SO₂– group, showed T _g values of 149 °C and 157 °C, respectively. The presence of sulfone groups introduces steric hindrance that restricts segmental motion. The higher T _g of structure V compared to structure IV is attributed to its increased aromatic content. Structures VI, VII, and IX exhibited relatively high T _g values (160–171 °C) due to reduced segmental mobility from the bulky –CF₃ substituents [34], 35]. Structure VII showed the highest T _g (171 °C), likely due to the presence of amide groups capable of forming intermolecular hydrogen bonds, which further restrict chain motion and increase the thermal energy required for transition [36], 37]. In contrast, structure VIII displayed a comparatively lower T _g (154 °C), despite its bulky substituents, because of the flexibility imparted by alkyl groups [38]. In the case of CPIs, the T _g is generally lower compared to that of conventional aromatic PIs. This tendency can be attributed to the specific structural modifications introduced to suppress CTC formation, which is essential for achieving optical transparency. For instance, the incorporation of flexible linkages such as ether bonds (–O–) or bulky substituents including –CF₃ and –C(CF₃)₂– not only reduces intermolecular interactions but also decreases chain rigidity. These structural features increase the free volume and enhance segmental mobility, thereby lowering the thermal energy required for the glass transition. Consequently, while CPIs exhibit excellent optical transparency by minimizing intermolecular electronic interactions, this design strategy inevitably compromises chain stiffness, leading to a reduced T _g relative to traditional aromatic polyimides [39], 40]. Figure 5 displays the DSC thermograms of the CPI films.

Figure 5:

DSC thermograms of various CPIs based on BU-DA monomer.

Thermal stability was further assessed using thermogravimetric analysis (TGA), and the initial decomposition temperatures (T _D ⁱ ) and residual weights at 600 °C (wt _R ⁶⁰⁰ ) are summarized in Table 2 and shown in Figure 6. T _D ⁱ values ranged from 346 °C to 382 °C. Structure I exhibited the lowest T _D ⁱ (346 °C), likely due to the thermally unstable m-linked alkyl chains [41]. Structures II to IV, with para-(p-) alkyl, ether, and m-sulfone linkages, respectively, showed moderate stability (T _D ⁱ  = 365–368 °C), attributed to bent configurations that limit close chain packing and intermolecular interactions.

Figure 6:

TGA thermograms of various CPIs based on BU-DA monomer.

Structures V–IX, which feature thermally robust benzene rings and fluorinated substituents, exhibited higher T _D ⁱ values (370–382 °C), indicating improved thermal stability [42]. Their linear configurations promote effective molecular packing and reinforce interchain interactions, enhancing thermal endurance [43]. The reason why the T _D ⁱ of CPIs is lower than that of conventional aromatic polyimides is primarily related to the structural modifications introduced to suppress CTC formation. While these modifications improve optical transparency, they inevitably weaken the intermolecular interactions that typically enhance thermal stability. Moreover, the incorporation of flexible ether linkages, alkyl substituents, and bulky moieties such as –CF₃ reduces the overall bond dissociation energy of the polymer backbone, thereby creating thermally weaker sites that promote earlier thermal degradation. In addition, the increased free volume generated by bulky substituents decreases chain packing density, rendering the polymer matrix more susceptible to the onset of decomposition. Consequently, CPIs exhibit a lower T _D ⁱ compared to traditional polyimides, despite maintaining sufficient thermal stability for practical applications. Residual weights at 600 °C (wt _R ⁶⁰⁰ ) for these structures generally ranged between 50 % and 64 %, whereas structures I and II, which included thermally labile alkyl groups, showed significantly lower residue values (36–37 %).

DTG analysis provides the rate of weight loss as a function of temperature, thereby offering more detailed insight into the thermal degradation behavior than TGA alone. In particular, for CPIs exhibiting multi-step degradation, the DTG curves clearly reveal the maximum degradation rate temperature (T _max) for each stage, which can be correlated with the decomposition of specific structural units such as imide, ether, or sulfone moieties. Furthermore, DTG enables precise identification of the onset (T _onset) and termination (Tend) of degradation, allowing a direct comparison of how additives, fillers, or copolymer compositions affect the decomposition process. Figure 7 shows the DTG curves of nine CPI films prepared from different diamines. All samples exhibited major degradation peaks in the range of 450–550 °C, which are associated with the thermal decomposition of the imide rings and aromatic backbones. However, the position and intensity of the T _max varied depending on the diamine structure. CPIs containing rigid aromatic units or strongly electron-withdrawing substituents displayed higher T _max values, reflecting enhanced thermal stability. In contrast, CPIs incorporating more flexible chain segments exhibited relatively lower T _max, indicating reduced resistance to thermal decomposition. These results demonstrate that DTG analysis provides more detailed information on the stepwise degradation behavior compared with TGA alone, enabling clear identification of structural effects on the thermal stability of CPI films. The TGA/DTG results of the CPI films prepared with various diamines were provided in Supplementary Figure III.

Figure 7:

DTG thermograms of various CPIs based on BU-DA monomer.

Thermal dimensional stability was evaluated using thermomechanical analysis (TMA). The coefficient of thermal expansion (CTE), which reflects the polymer’s tendency to expand upon heating, is sensitive to chain rigidity and aromatic content [44], 45]. Figure 8 and Table 2 show the TMA results. Structures I and II exhibited high CTE values (55.7 and 57.7 ppm °C⁻¹, respectively), due to flexible alkyl content and bent backbones. In contrast, structures IV and V, despite their non-linear chains, showed lower CTE values (48.6 and 49.0 ppm °C⁻¹), attributed to the restricted rotation imparted by the –SO₂– moieties. Among the fluorinated structures (VI–IX), structure VII displayed a relatively low CTE (52.1 ppm °C⁻¹), due to the potential formation of hydrogen bonds via the –NH–CO– groups. Structure IX exhibited the highest CTE (58.8 ppm °C⁻¹), which can be explained by the flexible hexafluoropropyl (–C(CF₃)₂–) substituents, allowing greater chain movement.

Figure 8:

TMA thermograms of various CPIs based on BU-DA monomer.

3.3 Mechanical properties

The mechanical properties of the CPI films – including ultimate tensile strength, initial modulus, and elongation at break (EB) – were evaluated using a UTM. The results are summarized in Table 3.

Structures I–III exhibited moderate tensile strengths ranging from 54 to 63 MPa. Structure IV, which incorporates a bulky –SO₂– group and a m-substitution, displayed the lowest tensile strength (22 MPa) among all samples. This significant reduction in strength is attributed to its highly bent conformation, which limits effective chain alignment (Figure 4). Conversely, although Structure V also contains a –SO₂– moiety, its more linear geometry, facilitated by ether linkages and a higher benzene content, led to a considerably higher tensile strength of 57 MPa. Structures VI–IX demonstrated superior tensile strength values (92–136 MPa), which can be attributed to their relatively linear chain configurations and higher aromatic content. In particular, Structure VI, derived from a simple p-substituted diamine, exhibited the highest tensile strength (136 MPa), attributed to its highly regular and linear 3D molecular conformation that enables effective chain packing.

The initial modulus, reflecting the stiffness of the polymer chain, was strongly influenced by structural rigidity and the presence of rigid aromatic units. As shown in Figure 4, Structure I exhibited the highest modulus (6.81 GPa), likely due to its relatively short and linear backbone. Structures VI, VIII, and IX also showed high modulus values (4.27–5.96 GPa), consistent with their more linear architectures. Although Structure VII exhibited a relatively high tensile strength (92 MPa), its modulus was comparatively lower (2.87 GPa), likely due to its overall bent chain configuration, which reduces stiffness despite the presence of hydrogen bonding interactions via amide linkages.

Overall, Structures VI–IX displayed enhanced mechanical performance, with high tensile strength and stiffness, which can be attributed to tightly packed, linear chains, short repeating units, and intermolecular attractions such as hydrogen bonding [46], 47].

The elongation at break (EB) values for all CPI films ranged consistently from approximately 4 %–10 %, suggesting comparable flexibility across the different structures. The reason why CPIs generally exhibit lower EB compared to conventional aromatic PIs can be attributed to their inherent structural characteristics. To suppress CTC formation and achieve optical transparency, CPIs are often designed with bulky substituents such as –CF₃, –C(CF₃)₂–, or –SO₂– groups, which increase chain rigidity and restrict molecular flexibility. This enhanced rigidity limits the ability of polymer chains to undergo plastic deformation during stretching, leading to brittle fracture and a reduced EB. Furthermore, the incorporation of bulky moieties increases free volume and disrupts uniform chain packing, which facilitates localized stress concentration rather than homogeneous chain extension under load. As a consequence, while CPIs maintain excellent thermal stability and optical transparency, they inevitably suffer from lower ductility, as reflected in their reduced EB [40], 48].

3.4 Optical transparency

To evaluate the optical transparency of the CPI films, three key parameters were measured: the λ ₀, the 500 nm^trans, and the YI. These values reflect the UV absorption onset, visible light transmittance, and visual coloration of the films, respectively [49]. All films were fabricated with a uniform thickness of 42–46 μm to ensure consistent optical comparison.

UV–vis spectra of the CPI films are shown in Figure 9, and the corresponding data are presented in Table 4. The λ ₀ values ranged from 328 to 392 nm depending on the diamine monomer used. Structures III, VIII, and IX exhibited relatively high λ ₀ values of 381, 392, and 374 nm, respectively. These results suggest increased CTC formation due to their more linear chain conformations. As a consequence, their 500 nm^trans values were slightly lower (81–83 %) compared to the other films. Nevertheless, all films demonstrated visible light transmittance greater than 80 % at 500 nm and transmitted light starting below 400 nm, indicating excellent overall optical transparency.

Figure 9:

UV–vis. transmittance of various CPIs based on BU-DA monomer.

Table 4 summarizes the YI values, which were calculated using the ASTM E313-96 and DIN 6167 standards with the equation:

where a and b are constants specific to each method, and X, Y, and Z are the tristimulus values.

Structures I and II exhibited the lowest YI values (both < 1), due to their bent, flexible chain structures, which hinder close chain stacking and CTC formation. Structures IV–VII also showed good optical clarity (YI = 2.2–4.2), attributed to structural distortion and m-substitution effects.

Notably, although Structure VI contains a linear p-substitution, its incorporation of the strongly electron-withdrawing –CF₃ group disrupts π-electron delocalization, thereby reducing CTC and lowering YI. In contrast, Structure III retained a relatively linear structure and displayed a significantly higher YI of 19.3. Structures VIII and IX, due to their linear conformations and lack of significant structural distortion, exhibited the highest YI values (27.5 and 15.7, respectively), consistent with increased coloration resulting from CTC formation.

To visually assess transparency, Figure 10 shows images of a printed logo viewed through each CPI film. Despite color variations due to different monomer structures, all films maintained sufficient clarity for easy logo recognition. These differences are consistent with the measured YI values.

Figure 10:

Photographs of various CPIs based on BU-DA monomer.

3.5 Solubility

Although CPI materials are known for their excellent thermal and chemical stability, their rigid aromatic backbones often lead to poor solubility in conventional solvents [50], 51]. This limitation restricts their processability and application as high-performance polymer films. Therefore, structural modifications that improve solubility without compromising CPI’s desirable properties are of great interest.

In this study, solubility tests were conducted using 12 common solvents, and the results are compiled in Table 5. All CPI films were completely insoluble in polar protic solvents such as ethanol (EtOH) and methanol (MeOH), and only exhibited limited solubility in acetone, dimethyl sulfoxide (DMSO), and toluene. However, higher solubility was observed in halogenated solvents such as chloroform (CHCl₃) and methylene chloride (CH₂Cl₂), and in polar aprotic solvents including N,N′-dimethylacetamide (DMAc), N,N′-dimethylformamide (DMF), N-methyl-2-pyrrolidone (NMP), and tetrahydrofuran (THF).

Structures I–III showed the poorest solubility due to their simple aromatic frameworks and absence of solubilizing substituents. In particular, structures I and II incorporate –CH₂– linkages within the main chain. While such alkyl segments may increase chain flexibility, they essentially function as hydrophobic moieties with negligible polarity. This structural feature is unfavorable when dissolution in polar solvents such as DMAc, DMF, or NMP is required. The presence of alkyl groups diminishes the extent of specific intermolecular interactions between the polymer chains and solvent molecules, including hydrogen bonding and dipole–dipole interactions, which are critical to achieving efficient solvation. Consequently, instead of enhancing solubility, the incorporation of –CH₂– units suppresses polymer–solvent affinity, thereby leading to a marked reduction in overall solubility [52]. In contrast, Structures IV–IX demonstrated significantly improved solubility. This enhancement is attributed to the presence of bulky groups such as –SO₂–, –CF₃, and alkyl substituents, which increase free volume and reduce chain packing density, thereby facilitating solvent penetration [53], 54]. Specifically, Structures VIII and IX exhibited excellent solubility across multiple solvents, owing to the incorporation of voluminous –C(CH₃)₂– (Structure VIII) and –C(CF₃)₂– (Structure IX) substituents. These groups improve solvent–polymer interactions and enhance accessibility to solvent molecules.