Influence of monomer conformation on the mechanical and tribological properties of thermosetting polyimides

Abstract

A series of thermosetting polyimide oligomers with different diamines (isomers) and same calculated number-average molecular weights (M_n) were synthesized with BPDA (biphenyl-tetracarboxylic acid dianhydride), 4,4′-ODA (4,4′-oxydianiline), 3,4′-ODA (3,4′-oxydianiline) and 4-PEPA (4-phenylethynylphthalic anhydride). Herein, we investigate the effect of monomer configuration on the mechanical, thermal and tribological properties of thermosetting polyimides. Experimental results demonstrate the relationship between molecular configuration and mechanical, thermal and tribological properties. It is revealed that as the content of 4,4′-ODA increases, T_g, Young's modulus and hardness increases and coefficient of friction (COF) and wear rate (WR) decrease under most conditions. Compared with variations of COFs and WRs of TPIs to velocity or load, it is easy to confirm the major influencing factor under different conditions.

---

1. Introduction

Polyimides play an important role in polymeric composites, and thus many researchers have investigated the properties of polyimide and its composites in many fields such as medical, electronic, automotive, and some other industrial aspects.¹ As is well known, the imide bond can be formed through condensation between a diacid, dianhydride and dihalide and diamine or via other methods. The two components of polyimide have a great impact on the properties of polyimide. Aromatic polyimides with a rigid structure, high strength, thermal stability and excellent anti thermal oxidation are commonly used for applications in engineering and also in tribology and microelectronics industry. Polyimide composites that are used as self-lubrication materials in engineering often receive further attention due to their excellent mechanical and thermal properties and many studies have discussed their tribological properties under different types of conditions.^2–4

Chitsaz-Zadeh,⁵ J. W., Jones,⁶ Cong⁷ and Samyn⁸ have done major work on the chemical structure effect on polyimide tribological performance and they also proposed some conclusions on the relation between chemical structure and tribological properties. J. W. Jones proposed that WR showed the lowest value for polyimide with a flexible linkage. Polyimide with the highest friction coefficient showed the highest density of polar side groups. Moreover, Chitsza-Zadeh found that WR was positively correlated with the flexibility of the molecular chains in polyimides. Cong investigated the tribological difference between thermoplastic and thermosetting polyimide. In addition, Samyn discussed polyimides performance under high temperature sliding and assessed their tribological performance through the reaction occurring at the sliding interface.

All of these references, except for Cong, are mainly about thermoplastic polyimides. For thermoplastic polyimides with uncrosslinked polymer chains, the polymer chains interdigitate through van der Waals forces or electron interactions (for polar groups). When sliding, the polyimide chain at the sliding interface will unloop firstly and rearrange along the sliding direction. Hence, the conclusion about chemical structure effect on polyimide performance was drawn by J. W. Jones, Chitsza-Zadeh and Samyn.

However, for thermosetting polyimides, in addition to van der Waals forces and electron interactions, thermosetting polyimide chains also interdigitate through chemical forces.⁹ Scission occurs firstly at the sliding interface, which causes severe fatigue wear and a higher COF. However, there is little research on the effect of molecular structure on the tribological properties of thermosetting polyimides.

In contrast with thermoplastic polyimides, many factors deeply influence the manifestation of thermosetting polyimides such as molecular weight,^10,11 monomer configuration,^12–14 and end-capping reagent. We investigated the relationship between molecular weight and tribological performance. The clear conclusion was drawn that there is significant intrinsic correlation between molecular weight and the performance of polyimides. Thermosetting polyimides with low molecular weight pre-oligomers present superior mechanical strength from room to high temperature.¹⁰

Moreover, due to the critical role of molecular structure in polymers,^5–8 research on the effect of molecular dynamic movement on polymer performance has been conducted. Many simulations have been established to demonstrate and quantize the relation between macroscopy and microscopy. To adjust the desired property, researchers often synthesize polyimides with two or more types of diamines and dianhydrides.^15–20 Due to the huge impact of synergy of diamine on T_g, solubility, strength and other properties,^21,22 it is interesting to exploit the influence of molecular structure on the tribological properties of co-polyimides. However, the effect of molecule configuration on tribology has not yet been considered. It is noteworthy to exploit the relation between molecular configuration and tribological properties of polymers. To eliminate the influence of molecular weight, three types of thermosetting co-polyimides with the same molecular weight are synthesized with biphenyl-tetracarboxylic acid dianhydride, 4,4′-oxydianiline (4,4′-ODA) and 3,4′-oxydianiline (3,4′-ODA). 3,4′-ODA is one of the isomers of 4,4′-ODA and it has an asymmetric structure. The sole difference between these three types of polyimide lies in the different ratios of the two types of diamines with different conformations. Mechanical strength, micro hardness, and tribological properties and their relation under room temperature are discussed in detail.

2. Experimental

2.1 Materials

Thermosetting polyimide molding powder was synthesized using the early method.¹⁰ The mole ratio of monomers is listed in Table 1. The synthetic process is depicted in Fig. 1 and the 3D model structure of BPDA with 4,4′-ODA and 3,4′-ODA is depicted in Fig. 2.

Table 1 Mole ratio of the different monomers

	BPDA	3,4′-ODA	4,4′-ODA	4-PEPA	η^a/cP
a η: measured in DMF at a concentration of 15 wt% at 30 °C.
TPI-1	3	4	0	2	19.8
TPI-2	15	4	16	10	27.3
TPI-3	3	0	4	2	29.2

---

Fig. 1 Synthesis of polyimide oligomers with different mole ratios of diamine.

Fig. 2 3D model molecular structure of BPDA with 4,4′-ODA and 3,4′-ODA.

BPDA (biphenyl-tetracarboxylic acid dianhydride), 4-phenylethylphthalic anhydride (4-PEPA), 4,4′-ODA (4,4′-oxydianiline) and 3,4′-ODA (3,4′-oxydianiline) were purchased from Changzhou Sunlight Pharmaceutical Co., Ltd. 1-Methyl-2-pyrrolidinone (NMP), which was purchased from Sinopharm Chemical Reagent Co., Ltd., was dried with molecular sieves and distilled under vacuum, then stored in a sealed flask.

2.2 Methods and measurements

Differential scanning calorimetry (DSC) was performed on a NETZSCH DSC-200F3 thermal analyzer under nitrogen. Thermal gravimetric analysis (TGA) was performed on a NETZSCH-STA449F3 series thermal analysis system at a heating rate of 10 °C min⁻¹ under nitrogen at a flow rate of 50 cm³ min⁻¹.

Nano-indentation tests were performed on a TI-950 triboindenter (Hysitron, American) using a Berkovich diamond indenter. After contacting with the surface, the indenter was stuck into the friction specimens with a constant strain rate of 0.05 s⁻¹ until 2000 nm of depth was reached and then withdrawn from the surface with the same rate as loading. At least ten indents were performed on each specimen and the average value was finally adopted. The hardness was calculated using the Oliver and Pharr method.

The bending strength of the specimens was determined on a three-point test machine with a span of 30 mm and crosshead speed of 2 mm min⁻¹. The specimens were 50 mm × 8 mm × 3 mm and the test surface was 50 mm × 8 mm. The specific bending strength (σ_f) of the specimen was calculated from eqn (1).

where P is the maximum load (N), l is the span length (mm), b is the width of specimen (mm), and h is the thickness of the specimen (mm).

The worn surface of the TPI and transfer film was characterized via SEM (JSM-5600LV). The worn surfaces were gold coated using the ion sputtering method.

Before the tribological test, the worn surface of the specimen was polished with metallographic sand paper to make sure that all specimen surfaces were tested under the same roughness concentration (R_a = 0.1 μm).

2.3 Preparation of PI specimens

The composites were fabricated by means of the hot press molding technique. The molding condition was determined through DSC data. Before hot press molding, the composite powders were pulverized. The powders were compressed and heated to 375 °C (the temperature was determined by the DSC characteristics) and held at 20 MPa for 1 h to allow full compression and curing. At the end of each run of compression sintering, the resulting specimens were cooled with a stove in air, and cut into preset sizes for tribological (2 × 2 × 18 mm) and mechanical tests.

2.4 Friction test

The tribological tests were conducted on the ball (GCr15, 3 mm) on a disc wear testing machine (CSEM-THT07-135). The contact schematic of the frictional couple is shown in Fig. 3. The wear rate was calculated using the following equation.¹⁰ Tribological tests were conducted under the loads of 1 N, 5 N, 10 N, and 20 N. Moreover, every specimen was tested under the given loads with three different velocities of 0.05 m s⁻¹, 0.1 m s⁻¹ and 0.5 m s⁻¹. The sliding distance was 300 m, except for the test under 20 N/0.5 m s⁻¹ (1000 m). Wear rate is calculated using eqn (2).
where ΔV: volume wear rate (mm³); R: couple diameter (3.000 mm); b: circle width (μm); d: circle diameter (10 mm); K: wear (mm³ N⁻¹ m⁻¹); P: load (N); L: sliding distance (m).

Fig. 3 Schematic of the contact configuration of the reciprocating friction.

3. Characteristics

3.1 Thermal properties

From the first DSC scans of the TPI oligomers shown in Fig. 4, it is obvious that there is no crystallization produced during the synthetic process. The exothermic peak indicates that the crosslink reactions of the three TPI oligomers started at approximately 350 °C, 358 °C and 366 °C. Moreover, the oligomer with a more rigid diamine (4,4′-ODA) reacted at a higher temperature (more energy to trigger the reaction). This is attributed to the close stack of the polymer chain in TPI-3, which greatly hindered the movement of the polymer molecule. Therefore, the molding temperature was set at 375 °C.

Fig. 4 The first DSC scan of the TPI oligomers under nitrogen.

Fig. 5 shows the infrared spectrum of TPI after hot-pressing. It is obvious that the ethynyl C≡C absorption band at 2213 cm⁻¹ disappears in the TPI specimens, which demonstrates the accomplishment of the thermal curing reaction of 4-PEPA.¹⁵ It can be concluded that all the TPI specimens are all already well cross-linked based on the FT-IR spectrum. Theoretically, it should be noted that under the same molding process parameters, the TPIs with different molecular configurations have slightly different crosslink densities, which can be ignored. The polyimide with a more asymmetric diamine corresponds to a slightly higher crosslink density. Therefore, the difference of mechanical and tribological properties between the TPIs in this study should be considered as a result of the synthesized effects of molecular configuration and crosslink density. However, we propose that the diamines with different configurations dominate the properties in this study.

Fig. 5 Attenuated total reflectance IR spectra of TPI-1,2,3.

Consistent with our speculation, molecular configuration has little influence on the initial decomposition temperature in nitrogen (see Fig. 6). TPI-1, TPI-2 and TPI-3 exhibited excellent thermal stability with onset decomposition temperatures at 540 °C, 544 °C and 549 °C, respectively. In contrast, it has a huge impact on the stacking of the polymer chain, which influences the residual weight retention. As can be seen, TPI-1 with a more asymmetric diamine in its polymer chain has the highest residual weight retention at 1000 °C, which is consistent with the result in the reference.²³

Fig. 6 Weight loss of the three types of polyimides under nitrogen.

Moreover, molecular configuration also affects the glass transition temperature (T_g) substantially (see Fig. 7). The thermosetting polyimide with the more asymmetric 3,4′-ODA presents a lower T_g, and the corresponding T_g sequence is TPI-1 < TPI-2 < TPI-3. Furthermore, the loss tangent (tan σ), which closely relates the molecular structure of a polymer and molecular interaction, also presents a higher value with a higher 3,4′-ODA content in the polymer chain at the glass transition temperature. The more asymmetric units in the molecular chain of the polymer, the stronger the molecular interaction, which results in a higher value of tan σ and different peak shapes at the glass transition temperature.

Fig. 7 DMA of the thermosetting polyimides.

3.2 Mechanical properties

The hardness and Young's modulus of the TPIs show a higher value with a higher content of 4,4′-ODA in the polymer chain (see Fig. 8a). With the same crosslink density, a symmetrical monomer will result in well stacked polymer chains when molding. Thus, hardness increases with a lower 3,4′-ODA content in polyimide. The Young's modulus of the TPIs also presents a similar trend.

Fig. 8 Mechanical properties of TPI-1, TPI-2 and TPI-3, (a) hardness and elasticity modulus (b) bending strength and modulus.

The effect of monomer configuration on mechanical property was investigated through bending strength and modulus testing (see Fig. 8b). It is obvious that the polyimide with more 4,4′-ODA presents a higher bending strength and modulus, which result from its symmetric and rigid structure.

3.3 Friction behavior

The average COFs of the TPIs are depicted in Fig. 9 under different velocities and loads. It is obvious that the COFs of the TPIs decrease with load increase under different velocities. Moreover, as the 3,4′-ODA content in the TPIs increases, the TPIs show a higher COF under most conditions, except for the COF of TPI-1 under 1 N and 0.5 m s⁻¹. The COF of TPI-2, which consists of both the diamines of 3,4′-ODA and 4,4′-ODA, presents the middle COF between TPI-1 and TPI-3, except for the condition mentioned above. The higher COF of TPI-1 results from the entanglement of the polyimide chains due to the greater asymmetric molecular configuration of 3,4′-ODA. Moreover, friction heat plays a more important role on TPI-1 with a lower T_g than the polymers with a higher T_g (TPI-2, 3), since the velocity increases from 0.05 m s⁻¹ to 0.5 m s⁻¹. This can be clearly seen from the performance of TPI-1 under 20 N and 0.5 m s⁻¹ (see Fig. 10). The sharp drop of COF predicts the variation of polymer at the friction interface, which affects the friction process at the interface. This can be proven by the SEM of the worn surface (see Fig. 11c) and the lumpy transfer film (see Fig. 10), which is attributed to the friction heat under such harsh conditions.²⁴

Fig. 9 Coefficient of friction of TPI-1, TPI-2 and TPI-3 under different velocities as a function of load.

Fig. 10 Sliding course of TPI-1, 2 and 3 under 20 N, 0.5 m s⁻¹ and transfer film of TPI-1.

Fig. 11 Worn surface of TPI-1,2,3 under different conditions, TPI-1; a, b, c; 20 N/0.1 m s⁻¹, 10 N/0.5 m s⁻¹, 20 N/0.5 m s⁻¹, TPI-2; d, e, f; 5 N/0.05 m s⁻¹, 5 N/0.1 m s⁻¹, 20 N/0.5 m s⁻¹, TPI-3; g, h, i; 1 N/0.05 m s⁻¹, 5 N/0.1 m s⁻¹, 20 N/0.5 m s⁻¹.

In contrast to the monotonic decrease in COF of TPI-1, the WR of TPI-1 decreases from 1 N to 5 N under different velocities (see Fig. 12) and WR increases with the increase in load from 5 N to 20 N. Under a lower load, the friction heat produced by velocity dominates the test process. The WR of TPI-1 decreases with an increase in velocity and the difference in TPI-1 between 1 N and 5 N becomes smaller, from 3.17 × 10⁻⁴ and 2.6 × 10⁻⁴ to 8.8 × 10⁻⁶, respectively, for 0.05 m s⁻¹, 0.1 m s⁻¹ and 0.5 m s⁻¹. Moreover, the WR of TPI-1 also decreased with an increase in velocity. However, with the load increase, WR escalates rapidly under all velocities, which apparently indicates the huge dependence of TPI-1 on load, especially under high velocity (0.5 m s⁻¹).

Fig. 12 Wear rate of TPI-1, TPI-2 and TPI-3 under different velocities as a function of load.

Different from TPI-1, the WR of TPI-2 decreases with load increase only under 0.0 5 m s⁻¹ and 0.1 m s⁻¹. However, under 0.5 m s⁻¹, the WR of TPI-2 decreases from 1 N to 5 N, then increases from 5 N to 20 N; moreover, WR decreases with velocity increase. As for TPI-3, WR decreases with load or velocity increase under every velocity and load condition. This is attributed to its high T_g, which enables the polymer to maintain its strength under the same conditions.

4. Discussion

4.1 Effect of monomer configuration on COF

To qualitatively evaluate the effect of molecular configuration on the tribological properties of the TPIs, the variations of COFs and WRs of the TPIs to velocity and load are presented in Fig. 13–16. The dependence of material on operating velocity or load is rated as low if the variation is as close to zero as possible or undulations in the curve are minimal or in other words, the curve tends to be a straight line parallel to the X-axis.²⁵ Moreover, if the value is positive, it means that COF decreases with an increase in velocity or load. On the contrary, if the value is negative, it means that COF increases with an increase in velocity or load.

Fig. 13 Difference in value of COF to velocity, (a) Δ_COF(0.05 m s⁻¹–0.1 m s⁻¹); (b) Δ_COF(0.1 m s⁻¹–0.5 m s⁻¹).

Fig. 14 Difference in value of coefficient of friction to load, (a) Δ_COF(1 N–5 N), (b) Δ_COF(5 N–10 N), (c) Δ_COF(10 N–20 N).

Fig. 15 Difference in value of WR to velocity (a) Δ_WR(0.05 m s⁻¹–0.1 m s⁻¹); (b) Δ_WR(0.1 m s⁻¹–0.5 m s⁻¹).

Fig. 16 Difference in value of wear rate to load, (a) Δ_WR(1 N–5 N), (b) Δ_WR(5 N–10 N), (c) Δ_WR(10 N–20 N).

Fig. 13 depicts the variation of COF against load at two speed transitions (transition one: from 0.05 m s⁻¹ to 0.1 m s⁻¹ and transition two: from 0.1 m s⁻¹ to 0.5 m s⁻¹). It is obvious that TPI-1 has a low dependence on load at the first speed transition, which corresponds to the small variation in Fig. 13a. However, at the second speed transition, the COF of TPI-1 decreases vastly (see Fig. 13b). The high value of COF variation of TPI-1 indicates the huge impact of velocity on TPI-1 under different loads. In addition, the relatively straight curves of TPI at the two speed transitions predict that TPI-1 has a low dependence on load.

Different with the high variation in the first speed transition and low variation in the second speed transition of TPI-1, the variation of TPI-2 shows an opposite trend. Compared to the performances of TPI-1 and 2, it can be drawn that monomer configuration has great impact on the friction properties of polymers. For TPI-1, composed with one type of diamine (3,4′-ODA, see Table 1), its polyimide chains intertwined with each other, which hinders the rearrangement and movement of the molecule at the contact surface. Only enough energy can conquer its internal friction or entanglements (physical and chemical crosslink) in its polymer chain such as high velocity and load. Under high load and velocity, more energy input will produce more friction heat, which can affect polymers' mechanical strength at the sliding interface and further friction coefficients. The COF variation of TPI-1 is more stable at the first speed transition than at the second transition. Hence, it is easy to understand the high value of variation of COF for TPI-1 at the second speed transition.

TPI-3 is more stable in the whole process except at 1 N. Under most conditions, the COF variation of TPI-3 always maintains a relatively low value. Moreover, combined with its smooth worn surface in Fig. 11, it is easy to predict that the COF of TPI-3 can easily reach a steady state under low load (corresponding to low energy input), which is attributed to the symmetric molecular structure in this polymer theoretically. Equally, TPI-2, composed of two types of diamines, presents a higher COF variation at the first speed transition and lower COF variation at the second transition. At the first speed transition, the low energy input can conquer the entanglements in the polymer chain of TPI-2. Thus, the low variation value of TPI-2 occurs at the first speed transition, compared to the high variation value of TPI-2 at the second speed transition.

To evaluate the dependence of the TPIs on load and velocity comprehensively, the difference in COFs against load as a function of velocity is plotted in Fig. 14. From Fig. 14a, it can be deduced that the big variations in COF of TPI-2 and TPI-3 indicate the easy shear for TPI-2 and TPI-3 compared to TPI-1. This can be attributed to the different interactions in the polymer chains, which is consistent with the conclusion drawn from the analysis of Fig. 12. Moreover, as the load increases, the variation of TPI-2 and TPI-3 shows a small value compared with that of TPI-1 (see Fig. 14b) and this also can contribute to the strong interaction in the molecular chain of TPI-1 compared to that in TPI-2 or TPI-3. For TPI-2 and TPI-3, the higher variations in the first load transition compared to that in the second load transition also implies the easy shear of TPI-2 and TPI-3 under lower load. As the load increases, in Fig. 14c, the nearly straight line predicts the stability of COF of TPI-3 under higher load as a function of velocity.

4.2 Effect of monomer configuration on WR

In general, the WR of TPI-1, which has more 3,4′-ODA, has great dependence on load and velocity. This can also be demonstrated in its SEM image (see Fig. 11a–c). The apparent cracks on the worn surface of TPI-1 under 20 N and 0.1 m s⁻¹ diminished on the worn surface of TPI-1 under 20 N and 0.5 m s⁻¹ and 10 N and 0.5 m s⁻¹. Traces of melt of polymer can be found in the SEM image of 20 N, 10 N and 0.5 m s⁻¹. Moreover, the sudden drop of friction of TPI-1 under 20 N and 0.5 m s⁻¹ is the most direct evidence of change of physical state of the polymer at the sliding interface.

The WR of TPI-2 shows a different trend under different velocities as a function of load. From the worn surface morphology in Fig. 11d–f, it is easy to understand the variation. Under a lower velocity, the worn surface of TPI-2 presents severe cracks. Under lower velocity and load, the synergy of load and velocity is not enough to support the movement of the molecule at the contact surface and more cracks occur on the worn surface (see Fig. 11d). Under a suitable velocity (5 N/0.1 m s⁻¹, see Fig. 12b, TPI-2), the input energy promotes the movement of the polymer chain at the sliding interface without causing huge cracks in the polymer (see Fig. 11e). However, under higher load and velocity (20 N, 0.5 m s⁻¹), scallops caused by peeling off appears on the worn surface again (see Fig. 11f), which corresponds to the higher wear at 0.5 m s⁻¹.

The WR of TPI-3 has a low dependence on load and velocity under the test conditions in our study. Under low load and velocity or high load and velocity, the worn surfaces of TPI-3 all present a relatively smooth surface morphology (see Fig. 11g–i).

To qualitatively assess the effect of load and velocity on the WR of the TPIs, the difference between velocity and load is plotted as a function of load or velocity, respectively, in Fig. 15 and 16. Fig. 15 depicts the WR variations of the TPIs under two speed transitions from 0.05 m s⁻¹ to 0.1 m s⁻¹ (first transition in Fig. 15a) and 0.1 m s⁻¹ to 0.5 m s⁻¹ (second transition in Fig. 15b) as a function of load. The higher WR variations of TPI-1 and TPI-2 indicate that velocity has great influence on the wear performances of TPI-1 and TPI-2. The WR variation of TPI-1 decreased from 1 N to 5 N firstly and then increased beyond 5 N. This also indicates that load has an important impact on the WR of TPI-1 at the first speed transition. Moreover, this predicts the great impact of load on WR of TPI-1 at the first speed transition. The WR variations of TPI-2 and 3 decreased and stabilized at the first speed transition as load increased, which implies that load has less influence on the WR of TPI-2 and 3, especially beyond 5 N. Compared with TPI-2 and 3, TPI-1 relies greatly on load at the first speed transition, when the load is beyond 5 N. The increase in load led to an increase in contact area, which has a close relation with friction and wear.²⁶ For TPI-3, load and velocity has little effect on their WRs when load is beyond 5 N, in the first speed transition. Moreover, TPI-3 is more stable in the second speed transition.

At the second speed transition, the WR variation of TPI-1 presents a similar trend with that in the first speed transition. Under high load and speed, friction heat has a serious influence on the polymer molecular morphology at the contact surface, which results in softness or melting and this can accelerate the wear of TPI-1 (see Fig. 11c). Combined with the worn morphologies in Fig. 11b and c, it is easy to determine the reason why TPI-1 shows such a variation at 20 N under the higher speed transition. As for TPI-2 and 3, only a small variation occurs at the high speed transition, which indicates the low dependence on the test conditions and stability of TPI-2 and 3. All these differences can be attributed to the different mole ratios of diamine in the polymer chains. TPI-1 with asymmetric 3,4′-ODA resulted in loose chain stacking, which exhibits a lower T_g, modulus and hardness, compared with TPI-2 and 3. The lower T_g results in a sharp drop of COF under high speed and load, which is due to friction heat (see Fig. 10). For the TPIs, their lower modulus, hardness and bending strength resulted from loose stacking in their polyimide chain, which plays a crucial role on tribological performance.

Fig. 16 depicts the variations of WR of the TPIs against load as a function of velocity. At the first load transition, the variations of WR of the TPIs decrease as velocity increases and are all above zero. This means that the WRs decrease as load reaches 5 N. In addition, speed also has a great impact on the WR of TPI-1. This can be attributed to the strong entanglement of the polyimide chain of TPI-1, which hinders the rearrangement of the molecular chain and results in the poor anti-wear of the polymer and higher COF.²³

At the second load transition, the positive WR variation value of TPI-3 predicts that WR decreases as load increases, compared with the negative WR variation value of TPI-1 and TPI-2. This implies the better anti-wear property of TPI-3 under the same test conditions. Significantly, the variation values of TPI-1 and TPI-2 are below zero. Moreover, the big absolute variation value of TPI-1 compared to TPI-2 indicates the worse anti-wear property of TPI-1.

At the third load transition, the variations of WR are more stable for TPI-2 and TPI-3. The relatively straight curves in the second and third load transition predict the stability of WR to velocity. The negative value means that WR increases as load increase. However, for TPI-1, in contrast to the variation in the first load transition, WR increases as load increases, especially under lower velocity (0.05 m s⁻¹ and 0.1 m s⁻¹). Compared with the different performances of TPI-1, TPI-2 and TPI-3 in Fig. 16, it is easy to conclude that variations of WR of the TPIs have a close relation with polymer composition. With a higher content of 4,4′-ODA in the polymer chain, the lower absolute variation value of WR implies the stability of the TPI under the test conditions. Thus, combined with the anti-wear performance in Fig. 12, the conclusion can be drawn that TPI-3 with a higher 4,4′-ODA content in its polyimide composition presents better anti-wear property than TPI-1 and TPI-2.

5. Conclusion

Three types of thermosetting co-polyimides with the same crosslink density and molecular weight were synthesized using BPDA (biphenyl-tetracarboxylic acid dianhydride), 4-PEPA, and different mole ratios of 4,4′-ODA (4,4′-oxydianiline) and 3,4′-ODA (3,4′-oxydianiline). Some mechanical, thermal and tribological properties were examined in our study and we found that there exist some intrinsic relations between molecular configuration and macro performance.

The conclusions drawn are as follows:

(1) With a higher 4,4′-ODA content in the polymer chain, the T_g, modulus, hardness and bending strength of the TPI present a higher value due to the more symmetric structure in the polymer molecule, which results in more close stacking of the polymer chain.

(2) With a higher 4,4′-ODA content in the polymer chain, the COF decreases under three types of velocities at different loads, except for TPI-1 at 1 N, 0.5 m s⁻¹. The high COF of TPI-1 can be ascribed to the asymmetric diamine (3,4′-ODA), which results in serious entanglement and internal friction in its polymer chain, which hinders movement in the friction test and leads to the high value of COF. Equally, under 0.1 m s⁻¹ and 0.5 m s⁻¹, the WR of the TPIs presents a similar trend.

(3) Through variations of COFs and WRs, we can qualitatively conclude that the configuration of the monomer has an important effect on the polymer's performance. For example, the COF of TPI-1 depends greatly on speed (see Fig. 13 and 14, the second speed transition); however, the WR of TPI-1 depends on speed and load simultaneously (see Fig. 15 and 16). In contrast to TPI-1, the tribological properties of TPI-3 appear to be more stable and immune to load and velocity under most conditions.

Acknowledgements

The authors would like to acknowledge the financial support from the National Basic Research Program of China (973 Program, Grant No. 2015CB057502), the National Natural Science Foundation of China No: 51303189 and 51103168 (NSFC No.: 51303189; No.: 51103168) and the National Defense Innovation Fund of Chinese Academy of Sciences (CXJJ-14-M43).

References

1. C. E. Sroog, Polyimides, Prog. Polym. Sci., 1991, 16, 561–694.
2. J. Kim, H. Im and M. H. Cho, Tribological performance of fluorinated polyimide-based nanocomposite coatings reinforced with PMMA-grafted-MWCNT, Wear, 2011, 271, 1029–1038.
3. Q. H. Wang, X. R. Zhang and X. Q. Pei, Study on the synergistic effect of carbon fiber and graphite and nanoparticle on the friction and wear behavior of polyimide composites, Mater. Des., 2010, 31, 3761–3768.
4. P. Samyn, J. Quintelier, G. Schoukens and P. de Baets, The sliding behaviour of sintered and thermoplastic polyimides investigated by thermal and Raman spectroscopic measurements, Wear, 2008, 264, 869–876.
5. M. R. Chitsaz-Zadeh and N. S. Eiss, Friction and wear of polyimide thin films, Wear, 1986, 110, 359–368.
6. J. W. Jone and N. S. Eiss, Effect of chemical structure on the friction and wear of polyimide thin films, ACS Symp. Ser., 1985, 287, 135–148.
7. P. H. Cong, T. S. Li and X. J. Liu, Effect of temperature on the friction and wear properties of a crosslinked and a linear polyimide, Acta Polym. Sin., 1998, 5, 556–561.
8. P. Samyn, G. Schoukens, F. Verpoort and J. Craenenbroeck, Friction and Wear Mechanisms of Sintered and Thermoplastic Polyimides under Adhesive Sliding, Macromol. Mater. Eng., 2007, 292, 523–556.
9. N. Maeda, N. Chen, M. Tirrell and J. N. Israelachvili, Adhesion and friction mechanisms of polymer-on-polymer surfaces, Science, 2002, 297, 379–382.
10. J. G. Smith, J. W. Connell and P. M. Hergenrother, The effect of Phenylethynyl Terminated Imide Oligomer Molecular Weight on the Properties of Composites, J. Compos. Mater., 2000, 34, 614–628.
11. J. W. Connell, J. G. Smith, P. M. Hergenrother and M. L. Rommel, Neat resin adhesive and composite properties of reactive additive/PETI-5 blends, High Perform. Polym., 2000, 12, 323–333.
12. R. Yokota, S. Yamamoto, S. Yano, T. Sawaguchi, M. Hasegawa, H. Yamaguchi, H. Ozawa and R. Sato, Molecular design of heat resistant polyimides having excellent processability and high glass transition temperature, High Perform. Polym., 2001, 13, S61–S72.
13. H. Masatoshi, N. Sensui, Y. Shindo and R. Yokota, Structure and Properties of Novel Asymmetric Biphenyl Type Polyimides. Homo- and Copolymers and Blends, Macromolecules, 1999, 32, 387–396.
14. T. Ogasawara, R. Yokota, W. Toshitaka, T. Aoi and J. Goto, Processing and properties of carbon fiber/Triple-A polyimide composites fabricated from imide oligomer dry prepreg, Composites, Part A, 2007, 38, 1296–1303.
15. Y. M. Wang, T. M. Wang and Q. H. Wang, Effect of molecular weight on tribological properties of thermosetting polyimide under high temperature, Tribol. Int., 2014, 78, 47–59.
16. W. Qiu, K. Zhang, F. S. Li, K. Zhang and W. J. Koros, Gas Separation Performance of Carbon Molecular Sieve Membranes Based on 6-FDA-mPDA/DABA(3 : 2) Polyimide, ChemSusChem, 2014, 7, 1186–1194.
17. E. M. Maya, I. Garcia-Yoldi, A. E. Lozano, J. G. de la Campa and J. de Abajo, Synthesis, Characterization, and Gas Separation Properties of Novel Copolyimides Containing Adamantyl Ester Pendant Groups, Macromolecules, 2014, 44, 2780–2790.
18. R. M. Huertas, A. Tena, A. E. Lozano, J. de Abajo, J. G. de la Campa and E. M. Maya, Thermal degradation of crosslinked copolyimide membranes to obtain productive gas separation membranes, Polym. Degrad. Stab., 2013, 98, 743–750.
19. W. Li, X. Guo and J. Fang, Synthesis and properties of sulfonated polyimide–polybenzimidazole copolymers as proton exchange membranes, J. Mater. Sci., 2014, 49, 2745–2753.
20. C. F. Kins, E. Sengupta, A. Kaltbeitzel, M. Wagner, I. Lieberwirth and H. W. Spiess, et al., Morphological Anisotropy and Proton Conduction in Multiblock Copolyimide Electrolyte Membranes, Macromolecules, 2014, 47, 2645–2658.
21. W. Wang, H. Zhou, D. Chen, M. Wu, Y. Liu and C. Chen, Preparations and Properties of a High Performance Semicrystalline Copolyimide and Composites Reinforced with Glass Fiber, J. Appl. Polym. Sci., 2014, 131, 40410.
22. T. S. Li, J. S. Tian, T. Huang, Z. Y. Huang, H. Y. Wang and R. G. Lu, et al., Tribological Behaviors of Fluorinated Polyimides at Different Temperatures, J. Macromol. Sci., Part B: Phys., 2011, 50, 860–870.
23. T. Huang, P. Liu, R. Lu, Z. Huang, H. Chen and T. Li, Modification of polyetherimide by phenylethynyl terminated agent for improved tribological, macro- and micro-mechanical properties, Wear, 2012, 292–293, 25–32.
24. P. Samyn, P. de Baets and G. Schoukens, Influence of Internal Lubricants (PTFE and Silicon Oil) in Short Carbon Fibre-Reinforced Polyimide Composites on Performance Properties, Tribol. Lett., 2009, 36, 135–146.
25. B. K. Satapathy and J. Bijwe, Composite friction materials based on organic fibres: Sensitivity of friction and wear to operating variables, Composites, Part A, 2006, 37, 1557–1567.
26. T. Aoike, D. Yokoyama, H. Uehara, T. Yamanobe and T. Komoto, Tribology of ultra-high molecular weight polyethylene disks molded at different temperatures, Wear, 2007, 262, 742–748.

---