The liquid crystal molecules were synthesized according to the reported methods (Rodle et al., 2016; Tan S. et al., 2020). 1-phenyloctane (99%) used as the solvent was purchased from J&K Chemical Ltd. (Beijing, China). All the chemicals above were used without any further purification.

The LC-amide-1, LC-amide-2, and LC-ester are dissolved in 1-phenyloctane at the corresponding concentrations. Then the self-assembly samples were prepared by depositing a droplet of the solutions (0.1 μL) onto the freshly cleaved HOPG (grade ZYA, NTMDT, Russia) surface, respectively. All the solutions and assembly samples were prepared under ambient conditions.

The self-assembly structures of liquid crystals were characterized using a Nanoscope IIIA system (Bruker, Germany) under ambient conditions. And the theoretical calculations were performed using DFT-D scheme provided by DMol3 code. Further characterization and calculation details can be seen in the Supplementary Material.

The microscopic friction forces of different liquid crystal systems were measured using an MFP-3D AFM (Asylum Research, America) at room temperature. To eliminate the solvent effect on the friction forces, the measurements were conducted at the gas-solid interface after the solvent was evaporated. The silicon CSG10 probe with a rectangular cantilever (nominal normal spring constant 0.3 N/m) was used. All the friction pairs are silicon probe (CSG10)—sample (assembly on HOPG). The selection of probe and loads were optimized after several attempts of friction experiments to get stable and repeatable data. It should be mentioned that the genuine normal and lateral factors need to be calibrated before each measurement. The normal photodetector sensitivity was calculated from the slope of the force curve obtained on a hard substrate. The normal spring constant was estimated from the power spectral density of thermal noise fluctuations under ambient conditions (Hutter and Bechhoefer, 1993). By scanning a commercially available TGF11 silicon grating (MikroMasch, Estonia), the lateral factor was calculated based on an improved wedge calibration method so that the voltage results can be transformed into the force values (Ogletree et al., 1996; Varenberg et al., 2003). The microscopic friction forces were measured by scanning in the perpendicular direction to the cantilever in the contact mode. The scan rate was 1 Hz, and the scan size was 160 nm × 160 nm, which is similar with the scan size of self-assembly structures in STM characterization. During the measurements, the feedback gains should be appropriate to make sure that the noise of the deflection signal was minimal while the height signal was tracked well. At least five positions were measured for each sample. For the data processing, friction values were calculated as half of the difference between the trace and retrace signals, such that the topography effect on friction can be excluded.

Our previous research has revealed the assembly structure of LC-amide-1 at 10⁻⁴ M, in which the dimers formed by two closed-packed molecules are aligned in parallel (Tan S. et al., 2020). However, after further investigation, we found that the assembly structure could change with the increase of concentration. As shown in Figure 1, two absolutely different assembly structures are existing simultaneously at higher concentration. Moreover, when the concentration is increased to 10⁻³ M, the uniform structure formed by new arrangement of LC-amide-1 molecules can be observed.

Figure 2A presents the large scale STM image of LC-amide-1 assembly structure at high concentration (10⁻³ M), in which the linear arrangement of bright rods could be observed. However, the arranging patterns of two adjacent rows are quite different, in which the bright rods are loose-packed in one row and close-packed in the neighbor row. Further details of assembly structure can be observed in the high resolution STM image, as shown in Figure 2B. The loose-packed rows are lined up with dimers formed by two bright rods. The average length of bright rods is measured to be 3.7 ± 0.1 nm, which shows good agreement with the theoretical size of backbone of LC-amide-1 molecules. Therefore, it can be inferred that the bright rods correspond to the LC-amide-1 molecules. And the alkyl chains of LC-amide-1 molecules correspond to the lines distributed in the shade, which cannot be explicitly characterized due to the much lower density of electric states. Each dimer is consisted of two facing LC-amide-1 molecules, as marked with blue rods in Figure 2B. As for the close-packed rows, the LC-amide-1 molecules are closely aligned in parallel, the orientations of which are at an angle to the direction of the row, as marked with blue rods in Figure 2B likewise. The motif of uniform and ordered assembly derives from the π-π stacking interaction between corresponding benzene rings, the N–H···O hydrogen bonding between neighboring amide groups and the van der Waals interactions between the alkyl chains. The parameters of the unit cells overlaid in Figure 2B are measured as follows: a = 7.9 ± 0.1 nm, b = 5.8 ± 0.1 nm, and α = 105 ± 1°.

The unit cell parameters and interaction energies of LC-amide-1 at high concentration are calculated with the DFT method based on the STM observation to further analyze the self-assembly structure. The calculated unit cell parameters (as shown in Figure 2C) are listed in Table 1, which are in accordance with the experimental results. The interaction energies between molecules are also calculated, as listed in Table 2. For comparison, the interaction energies of LC-amide-1 at low concentration, LC-amide-2, and LC-ester presented in our previous research (Tan S. et al., 2020) are also listed in Table 2. Herein, the lower interaction energy indicates the stronger interaction. Compared with three self-assembly structures in our previous research (Tan S. et al., 2020), the arrangement of LC-amide-1 molecules at high concentration is much more compact. Therefore, it can be observed that the intermolecular interaction energy of LC-amide-1 at high concentration (−535.52 kcal/mol) is the lowest. Other than the intermolecular interactions, the interactions between molecules, and the substrate also play an important role in the surface supramolecular assembly process. As shown in the second column in Table 2, the interaction energy between assembled molecules and the substrate of LC-amide-1 at high concentration (−1,704.309 kcal/mol) is much lower than all the others. It can be concluded that the π-π stacking and van der Waals interaction between LC-amide-1 molecules and HOPG at high concentration is strongest due to the much tighter alignment of molecules. Moreover, the total energy per unit area is also calculated and listed in the fourth column in Table 2, which can evaluate the interaction strength and thermodynamic stability of the assembly systems with different unit cells. It can be clearly revealed that the total energy per unit area of LC-amide-1 at high concentration (−0.479 kcal mol⁻¹ Å⁻²) is lowest, which means that the interaction strength is the strongest among four assembly systems.

At present, the tribological research on liquid crystals is mainly conducted at macroscale. To investigate the nanotribological performance, the friction forces of liquid crystals, including self-assembly mentioned above and three other self-assemblies in our previous research (Tan S. et al., 2020), are measured under ambient conditions by using AFM. As displayed in Figure 3, friction forces of LC-amide-1 at high concentration (10⁻³ M) and low concentration (10⁻⁴ M), LC-amide-2, LC-ester and HOPG are presented as a function of the applied load. For these five systems, friction forces increased linearly with the load. In comparison, the friction forces of LC-amide-1 at low concentration, LC-amide-2 and LC-ester are similar, while the friction forces of LC-amide-1 at high concentration are approximately twice larger. The friction-load data are linearly fitted to acquire friction coefficient, which can be defined as the slope of the linear fitting curve, as labeled in Figure 3. It is revealed that the friction coefficient of LC-amide-1 at high concentration is much higher than those of the other self-assembly systems, which is similar as the friction force result. As for the three other liquid crystal systems, their friction coefficients are close to each other and all below 0.01, indicating the achievement of microscale superlubricity. Specifically, the friction coefficient of LC-amide-1 at low concentration is slightly higher than that of LC-ester, while the friction coefficient of LC-amide-2 is the lowest in three systems.

As mentioned above, the interaction strength of LC-amide-1 at high concentration is the strongest among all the self-assemblies. Considering that the friction coefficient of LC-amide-1 at high concentration is also the highest, it seems that the nanotribological properties are related to the interaction strength. Therefore, to better understand such correlation, the experimental friction coefficient and theoretical total energy per unit area of four liquid crystal self-assembly systems are summarized in Figure 4. It is clearly revealed that the friction coefficient is positively correlated with the absolute value of total energy per unit area. In other words, there is a positive correlation between friction coefficient and interaction strength, which agrees well with the conclusion in our previous researches (Shi et al., 2018; Tan S. C. et al., 2020). The interaction herein includes the interaction between assembled molecules and the interaction between molecules and HOPG substrate. It can be hypothesized that during the rubbing process between the AFM probe and the self-assembly, these physical interactions are ruptured by the AFM probe and regenerated rapidly, achieving a dynamic equilibrium state. The friction process can also be considered as the continuous energy barrier overcoming process of AFM probe at the interface. Therefore, the assembly with stronger interaction would display a higher friction coefficient. The microscale superlubricity of liquid crystals can only be achieved in the self-assemblies with weak interaction, while the friction coefficient of LC-amide-1 self-assembly at high concentration is higher than 0.01 because of the tighter arrangement of molecules and stronger interaction.