Optical Erasure and Reconfiguration of Surface‐Relief‐Gratings of Azo Polymer and Azo Molecular Glass: A Comparative Study on Soft‐Lithographic Duplicates

Surface‐relief‐gratings (SRGs) of azo polymers and azo molecular glasses have attracted considerable interest for their erasable and reconfigurable characteristics. Herein, the optical erasure of soft‐lithographically duplicated SRGs of an azo polymer (BP‐AZ‐CN) and an azo molecular glass (IAC‐4) is investigated to tackle a long‐standing problem from “memory” effect existing in optically inscribed SRGs. The optical erasure is performed by single laser beam irradiation at wavelength of 488 nm with different polarizations, and the erasing behavior is thoroughly investigated by real‐time diffraction efficiency measurement and atomic force microscopy (AFM) observation. Showing no obvious dependence on the light polarizations, the optical erasure of IAC‐4 grating is a relatively fast process, and the original one can be completely erased and recreated. In contrast, the erasure of BP‐AZ‐CN grating is not only much slower, but also strongly correlated with the light polarizations. Upon the p‐polarized light irradiation, the grating profiles are substantially distorted during the erasure with a significant difference in that of the s‐polarized light irradiation. The gratings cannot be completely erased by the lights with all the polarizations. The erasure behavior is found to be determined by the effects of surface tension and mass transfer along the light polarization direction to different extents.

behavior similar to amorphous polymers and stay in a stable amorphous state below T g s. [33,34] Azo molecular glasses possess several characteristics obviously distinct from their polymeric counterparts, which include well-defined molecular structure, mono-dispersed molecular weight, and no chain entanglement. [11] Compared with azo polymers for inscribing SRGs and other surface structures, azo molecular glasses usually need a much shorter time to complete the processes and the results are relatively easy to be analyzed. [35][36][37] As a unique characteristic, SRGs on azo polymer and azo molecular glass films are erasable, reconfigurable, and highly adaptable. [7,8,11,15] This appealing function has been explored for various important applications, such as wavelength-programmable distributed feedback (DFB) laser, [38] reconfigurable cell-guiding patterns, [39,40] adaptive biological surfaces, [41] and optical Fourier surfaces. [42,43] SRGs can be completely erased by raising the temperature above T g of the amorphous polymers, [20,21] or close to the clear point of the soft liquid crystalline polymers. [44] Moreover, the erasure of SRGs on azo molecular glass films has been accomplished by localized heating with a CO 2 laser. [45] The erasure temperature was found to be strongly correlated with the respective T g s of the materials, which results in the complete and partial erasure of the gratings. Using laser irradiation with interference patterns shifted by half a period, SRGs on supramolecular azo polymer films have been photoerased and regenerated. [46] It has also been reported that the already inscribed gratings can be readily transformed into totally different gratings by exposure to a different interference pattern. [43] Distinct from these reports, the optical erasure caused by irradiation with a single laser beam at ambient temperature is of particular significance. [47][48][49] By this method, the erasing location and size can be precisely controlled without the global heating of the films, which is ideal for the integration of different patterns on a surface through erasure and reconfiguration. Since the erasure caused by the light irradiation is an inverse process of the SRG inscribing, the study can supply a deep understanding of the mass transfer behavior, which is complementary to the investigation of SRG inscribing through the interfering light irradiation. [7,8] However, in contrast to the intensive investigations on the SRG formation, few works have been devoted to the optical erasure of the surface structures through single beam irradiation. [21,[47][48][49] Whether SRG can be completely erased by this method and how the erasure behavior depends on the light polarization state remained controversial for a long time. It has been reported that the grating on an azo polyacrylate (PDR1A) film is not photo-erasable with a single beam of either linearly or circularly polarized light at either absorbing or non-absorbing wavelength. [21] For SRG recorded with p-polarization on an epoxy-based azo polymer (PDO3) film, a linearly polarized erasing beam with the polarization parallel to the grating grooves showed a high efficiency to erase the grating and the amplitude of the grating was reduced from 82 to 14 nm when irradiated for 30 min. [47] However, when the SRG was inscribed with circularly polarized light (CPL), the linearly polarized erasing beam with the polarization perpendicular to the grating grooves can even significantly enhance grating amplitude. On the other hand, based on a separate investigation on an azo polymethacrylate (pDR1M), the erasure efficiency is always in order p-polarized > right circularly polarized (RCP) > s-polarized no matter what polarization configuration is used to inscribe the gratings, where p-polarized, s-polarized, and RCP mean the erasure with linearly polarized light perpendicular or parallel to the grating grooves and right circularly polarized light, respectively. [48] The disagreement has been attributed to the fact that the azo chromophores are oriented by polarized light in the SRG inscribing process and SRG erasure is a process strongly dependent on light polarization. [48] Moreover, the optical erasure of SRGs has also been investigated on polymer-azobenzene complexes. [49] The results showed that the complete optical erasure could be achieved for the low-molecular-weight complexes, which were essentially independent of the polarization state of the erasing beam. But, only partial erasure was observed for the complex with the higher molecular weight.
The aforementioned results show that the optical erasure with a single laser beam is an involved issue including longstanding disagreements over the results. The complexity is mainly caused by the dependence of the optical erasure behavior on the polarization states of both the recording beams and the erasing beam. [47,48] The SRG formation is a process closely related to the polarization states of the writing beams. [23,[50][51][52] The reason is that the directional molecular migration and mass transfer occur along the polarization direction of the laser beam. [24,[53][54][55] The low efficiency for inscribing SRGs with interfering s-polarized beams (intensity modulation) has been rationalized by the fact that the polarization is perpendicular to the vector direction of the interference fringes. [50,51] SRGs can be more efficiently inscribed by using interference of polarization modulation, such as p,p-polarization and opposite circular polarization. [7,50,51] The polarized lights used to inscribe SRG will induce azo chromophores to orientate perpendicularly to the polarization through trans-cis-trans isomerization of the azobenzene moieties. [8,[56][57][58] Therefore, the gratings once formed will hold some memories of the polarization states or show the retroactive effect of the writing conditions, which then affect erasure behavior accordingly. [47,48] As the photoinduced chromophore orientation almost inevitably occurs during the SRG-inscribing process, the chromophores with different orientation states will respond to the polarized lights in different ways to cause the complexity in the investigation of the optical erasure of SRGs. Moreover, during the real-time characterization of SRG erasure by diffraction efficiency measurement, the phase gratings with different phase shifts to SRGs could also cause complications to explain the results. [57,58] Therefore, it is desirable to develop a new approach to study the optical erasing behavior without the complexity related to the chromophore orientation induced during the inscribing of SRGs. Moreover, it is worth noting that although both azo polymer and azo molecular glass films have been adopted in the investigations, there is no relevant research on the optical erasure of SRGs for these two types of materials in a comparative way.
In this study, instead of using SRGs inscribed by the laser interference irradiation, the duplicates of SRGs obtained from the hot embossing soft-lithographic replication were adopted to investigate the optical erasure behavior through a single beam irradiation. This approach can avoid the complexity caused by the chromophore orientation upon the polarized light irradiation during the inscribing of SRGs, and the research can be mainly concentrated on the influences of the material properties and erasing light polarizations on the erasure behavior. An azo polymer (BP-AZ-CN) and an azo molecular glass (IAC-4) were synthesized as the typical azo polymer and azo molecular glass. A systematic study of BP-AZ-CN and IAC-4 with respect to their optical erasure was carried out under comparable conditions. The erasure behavior was investigated by exposing the duplicated gratings to a single laser beam (488 nm) incident perpendicularly. The erasure effect of the beams with different polarizations was investigated by real-time diffraction efficiency measurement and atomic force microscopy (AFM). The nano-mechanical properties of the surfaces before and after light irradiation were also measured by AFM. For comparison, the thermal erasure behavior of the duplicated SRGs was also investigated. Based on the results, the mechanism of optical SRG erasure of the azo polymer and azo molecular glass together with the possible applications are discussed in detail. Figure 1 shows the materials and method used to investigate the optical erasure of SRGs in this study. The optical setup and the chemical structure of the materials are shown in Figure 1a. BP-AZ-CN is an epoxy-based azo polymer, which is a typical type of azo polymer used to inscribe SRGs. [7,59] The number average molecular weight (M n ) and glass transition temperature (T g ) of BP-AZ-CN are 18 500 g mol À1 and 137°C. IAC-4 is an azo molecular glass containing a core of isosorbide moiety, two push-pull-type azo chromophores, and peripheral cinnamate groups. IAC-4 is a low molecular weight compound (MW ¼ 1289.4 g mol À1 ) with T g of 63°C. SRGs can be inscribed on IAC-4 films with high efficiency when irradiated with interfering laser beams. [36,60] The synthesis and characterization details of IAC-4 and BP-AZ-CN have been given in our previous articles, [36,59] which can also be seen in the Supporting Information ( Figure S1ÀS7). Both BP-AZ-CN and IAC-4 show   Instead of using the optically inscribed  SRGs for optical erasure investigation, the SRGs were duplicated  by hot embossing soft-lithography and the duplicates of IAC-4 and BP-AZ-CN were used for the optical erasure investigation. This approach is adopted to avoid the orientations of azo chromophores and adjoining segments in the optically inscribed SRGs as discussed earlier. Thus, the current approach can reveal the erasure behavior without the involved connections with the polarization states of the writing beams. The soft-lithographic method used to duplicate SRGs has been reported in our previous articles. [61,62] To prepare the SRG duplicates, SRGs were first inscribed on IAC-4 films by exposure to the interfering laser beams (488 nm, 100 mW cm À2 ) for 20 min, which were used as the grating masters. Polydimethylsiloxane (PDMS) stamps with negative surface patterns were then obtained by replica molding against the masters. Finally, by hot embossing soft-lithography using the PDMS stamps, the duplicates of SRGs were prepared on BP-AZ-CN films and IAC-4 films, respectively. Figure 1b (Figure 1e). The iridescent parts of the surfaces in the photographs are the areas covered with the diffraction gratings. As proved by the AFM characterization, the grating master, PDMS stamp, and duplicated SRGs all possess a period of 1.18 μm and a grating amplitude of around 300 nm. It can be seen from the figures that the SRGs can be faithfully duplicated by this method. As the replication was highly reproducible, the duplicated gratings with the same amplitudes and periods as the masters were obtained and used for the erasure investigation.

SRGs Replication and Optical Erasure
The duplicated SRGs on IAC-4 and BP-AZ-CN films were optically erased by a single laser beam with three typical polarization conditions and irradiation for different time periods. The beam from a semiconductor laser (λ ¼ 488 nm) was expanded, collimated, and then incident perpendicularly onto the SRG samples. The optical erasure was investigated with the linearly polarized light, both parallel to the grating groove (s-polarized) and perpendicular to the groove (p-polarized), and circularly polarized light. As there was no influence from the different light chirality, the RCP light was adopted in the investigation. The light intensity of the erasing beam was set to be 100 and 200 mW cm À2 . The dynamical processes of the optical erasure were monitored by measuring the diffraction efficiency variation in a real-time manner under laser irradiation. Quantitative characterization of the grating topography and profile was performed by AFM. Moreover, the nanomechanical properties of the erased grating surfaces were measured by AFM PeakForce Quantitative Nanomechanical Mapping (PeakForce QNM or PF-QNM) technique. [63][64][65]

Optical Erasure Monitored by Diffraction Efficiency
The processes of the optical erasure were monitored by measuring the diffraction efficiency variations of the gratings. This method has been typically used to supply information about grating amplitude increase and decrease in the real-time manner. [17,18,[47][48][49] The first-order diffraction efficiency was obtained from the intensities of the zero-order beam (I 0 ) and the first-order diffraction beams (I 1 and I À1 ) by equation (I 1 þ I À1 )/2I 0 . The diffraction efficiencies of the duplicated SRGs before the optical erasure were in the range from 0.35 to 0.40. The diffraction efficiency (DE), defined as the value relative to that before the erasing irradiation, is used to characterize the optical erasure behavior. Figure 2 gives DE variation as a function of irradiation time with the light intensity of 100 and 200 mW cm À2 , respectively. For the azo polymer (BP-AZ-CN), the optical erasure behavior shows an obvious correlation with the light polarization states. As shown in Figure 2a, when exposed to the p-polarized light (100 mW cm À2 ), DE first increases slightly and reaches the maximum (DE ¼ 1.05 À 1.06) at around 5 min. Then, it begins to descend rapidly during 5-60 min, and slowly approaches the steady value (0.04 À 0.06) at 120 min. In contrast, for the s-polarized and RCP erasing beams with the same intensity, DE instantly decreases from the beginning and approaches the steady value after irradiation for 120 min. In the intermediate stage, the decreasing rates of DEs are obviously dependent on the polarization states. The overall decrease rates of DEs show the order, p-polarized > RCP > s-polarized. After the irradiation for 60 min, the DE drops from 1 to 0.12 for p-polarized light and to 0.18 for RCP light, respectively, while DE shows the highest remaining value of 0.31 for the s-polarized light. Although DEs of gratings show different variation trends in the erasure process for the three polarization states, they all decrease substantially and reach almost the same values after the irradiation for 120 min. When the SRGs were exposed to light with a higher intensity (200 mW cm À2 ), DEs are reduced at significantly faster rates in the process of optical erasure (Figure 2b). Comparing Figure 2a with b, it can be seen that the erasing rates are determined by the exposure energy (light intensity Â time), which infers the trivial role of the thermal effect in the erasing processes. The decreasing trends of the DEs for the three polarization states are similar to the cases with the light intensity of 100 mW cm À2 , i.e., the change in light intensity shows no effect on the polarization-dependent characteristics of the optical erasure.
The duplicated SRGs on the azo molecular glass (IAC-4) films show distinctively different erasure behavior under the same single beam irradiation condition. Figure 2c,d shows the changes of DEs for the duplicated SRGs on IAC-4 films upon the polarized light irradiation with intensities of 100 and 200 mW cm À2 , respectively. For the cases upon the light irradiation with the intensity of 100 mW cm À2 and the three different polarization states (Figure 2c), DEs of SRGs all show the monotonic decrease, which is distinct from those of BP-AZ-CN. The erasure rates almost show no difference for the three polarization states, where the erasure with the p-polarized light is slightly slower compared with the others in the intermediate stage. Under the same conditions, the optical erasure of IAC-4 SRGs is much more efficient compared with that of BP-AZ-CN. After the light irradiation for 15 min, the DEs are very close to 0 for all three polarized lights, which implies almost 100% erasure of the SRGs. It is a quite reasonable result as there is no chain entanglement in the low molecular system. Like BP-AZ-CN, the increase of the light intensity to 200 mW cm À2 obviously accelerates the optical erasure of the IAC-4 SRGs (Figure 2d). Similarly, comparing the curves in Figure 2c,d shows that the erasing rates are mainly dependent on the exposure energy, which almost show no dependence on polarization states for these two light intensities.

Topographic Variations during Optical Erasure
Above DE measurements manifest the optical erasure behavior of SRGs of BP-AZ-CN and IAC-4 in a real-time manner, where the clear differences between the azo polymer and azo molecular glass are revealed by their erasing rates and different dependences on the light polarizations. To know the exact structure variations causing the DE changes during the erasing processes, the surface topography was monitored by AFM inspection after exposing the gratings to the single laser beam for a series of time periods. Since the light intensity only shows the effect on the optical erasure rate, the light intensity of 100 mW cm À2 is adopted in this part. The morphological variations during the erasure processes not only confirm the erasing behavior shown in Figure 2, but also reveal some unique characteristics that cannot be seen from the DE measurements. Figure 3a1-a8 gives the AFM topographic images and corresponding cross-sectional profiles for the SRG on BP-AZ-CN film before and after being irradiated with the p-polarized light for different time periods. In the first 5 min (Figure 3a3), the amplitude of the grating remains almost unchanged, but the full width at half maximum (FWHM) increases to some extent ( Figure S8 and S9a, Supporting Information). After being exposed to the light for 20 min (Figure 3a4), the amplitude of the grating shows an obvious decrease. The FWHM of SRG significantly increases ( Figure S9a, Supporting Information), which can also be seen from the profile of the cross-section. At the irradiation time of 40 min (Figure 3a5), the cross-section profile significantly deviates from the original sinusoidal shape. The top of each crest splits into two and the grooves become narrower at the bottom. After the light irradiation for 60 min (Figure 3a6), the bifurcation of the crest can still be seen and the amplitude of the grating continues to decrease. For the further increase of the exposure time (Figure 3a7), the topographic image of SRG shows the obviously decreased amplitude for the flattened ridge structures. The fringes with the small amplitude can still be seen even after the light irradiation for 120 min (Figure 3a8). Similar topographic changes are observed during the erasing process of SRG under RCP light (Figure 3b1-b8). After the light irradiation for 40 min, the tops of the crests become flat and FWHM increases accordingly ( Figure 3b5 and S9a, Supporting Information), and the bifurcation of the crests remains at 60 min ( Figure 3b6). In contrast, the optical erasure with s-polarized light shows the topographic transition obvious different from the cases with the p-polarized and RCP light. The grating profiles nearly keep the sinusoidal shape during the erasure process (Figure 3c1-c8). As the amplitude decreases, the FWHM of the profile remains almost unchanged ( Figure S9a, Supporting Information). After the light irradiation for 120 min (Figure 3c8), the fringes with the small amplitude can also be seen. The results show that the topographic transitions of the gratings in the erasure process depend on the light polarization states. For the optical erasure with the different polarizations, the gratings on BP-AZ-CN films cannot be completely erased even after the light irradiation for 120 min, though the remaining grating amplitudes are small.
The optical erasure behavior of the duplicated SRGs of IAC-4, investigated under the same conditions, displays a significant difference from that of BP-AZ-CN. Figure 4 shows the AFM   www.advancedsciencenews.com www.adpr-journal.com topographic images and cross-sectional profiles after being irradiated with three polarized lights at different times. It can be clearly seen that the topographic changes are rather simple compared with those of BP-AZ-CN. For the three kinds of polarized lights, the profiles of the IAC-4 gratings maintain the sinusoidal shape without the bifurcation of the crests and shape distortion in the optical erasure process. Distinct from the partial erasure of the BP-AZ-CN gratings, the IAC-4 gratings can almost be completely erased in 20 min upon light irradiation with an intensity of 100 mW cm À2 . The surfaces covered with the gratings revert to a smooth plane, which is consistent with the results of the DE measurements. Moreover, the erasure behavior shows very little dependence on the light polarization state, which can only be seen from the FWHM values of the profiles in the middle of the optical erasure process ( Figure S9b, Supporting Information). By comparing the topography of IAC-4 gratings for the irradiation time of 3 min (100 mW cm À2 ), the FWHM values show that order p-polarized > RCP > s-polarized, while the values become almost the same after the irradiation for 12 min. A more quantitative comparison of the erasure behavior can be seen from the amplitude, surface area, and volume of a representative grating ridge in the AFM image. The surface area and volume refer to the values for a single grating ridge with the unit length, which is thus equal to the arc perimeter and the area of its cross-section profile ( Figure S8, Supporting Information). Figure 5a-c shows the variations of the amplitude, surface area, and volume of the BP-AZ-CN SRGs when irradiated with laser beams with three different polarizations. When the grating is irradiated with the p-polarized light, the amplitude, surface area, and volume all show the initial increase, quick decline, and slow decrease before reaching the steady value. The irradiation time corresponding to the maximum value of the amplitude is approximately equal to the time when the DE reaches its maximum ( Figure 2a). The significant decreases in the amplitude and surface area occur in the period from 10 to 50 min (Figure 5a,b). However, as the FWHM of the grating increases with the irradiation time ( Figure S9a, Supporting Information), the ridge volume has its maximum appearance within 10-20 min (Figure 5c). In contrast, for the erasure with s-polarized light, the amplitude, surface area, and volume of the SRG all exhibit a monotonic decrease with the irradiation time, but the decreases are slower in the first 5 min. The slow decreases are also observed after the light irradiation for 40 min. When the grating is exposed to the RCP light, the variations of these values show a manner between those of the p-and s-polarized light. The amplitude variation shows a short activation stage, a rapid decline in the middle, and slower decrease before finally stabilizing (Figure 5a). This variation trend is consistent with the variation of DE (Figure 2a). For RCP light irradiation, the surface area and volume of the ridge slightly increase in the first 10 min, due to the FWHM increases in the same period of time, and then show a monotonic decrease with the irradiation time (Figure 5b,c). But, the declining rates of these three values are slower compared with those under the p-polarized light irradiation. Although the optical erasures of the BP-AZ-CN gratings all include three stages, the variation trends are different for the lights with different polarizations. For the p-polarized light, it appears as the initial increase, quick decline, and slow decrease, while for the s-polarized light, it appears as the slow decrease, quick decline, and steady decrease. For the erasure with RCP light, the three-stage is also observed, which appears as the manner between the two cases of the linearly polarized lights. Due to these differences, the amplitude, surface area, and volume show the different variation curves for the three polarization states. Compared with the BP-AZ-CN grating, the variations of the amplitude, surface area, and volume of the IAC-4 grating show a rather simple manner during the optical erasure processes. As shown in Figure 5d-f, these values decrease quickly in the first 4-8 min and then decline slower to steady values in less than 20 min. This result reflects the quicker erasure of the IAC-4 grating compared with that of BP-AZ-CN, which is fully consistent with the results of the DE measurement (Figure 2c). The change of the polarization states of the erasing light only shows a weak influence on the erasing behavior. An observable difference is that the volume upon the p-polarized light irradiation decreases slightly slower compared with those under irradiation with RCP and s-polarized light (Figure 5f ). This observation is correlated with the increased FWHM during the light irradiation ( Figure S9b, Supporting Information).

Nanomechanical Property and Grating Reconfiguration
After optical erasure with the lights of three different polarizations (100 mW cm À2 ), the nanomechanical properties of the surfaces were probed by the AFM PeakForce QNM method, [63][64][65] where the pristine films of BP-AZ-CN and IAC-4 were used as the control references. The details of the measurement can be seen in the Supporting Information.  (Figure 6a,e). Although BP-AZ-CN has a much higher molecular weight, its surface modulus is in the same range as that of IAC-4. This result is also confirmed by the nanoindentation test ( Figure S10, Supporting Information), the Young's moduli obtained by the measurements are 4.2 AE 0.1 GPa for BP-AZ-CN films and 4.1 AE 0.2 GPa for IAC-4 films, respectively. It proves the reliability of the AFM PeakForce QNM technique in measuring nanomechanical properties. As shown by AFM topographic images in the top row of Figure 6b-d, the duplicated SRGs on the BP-AZ-CN films cannot be completely erased even after the light irradiation for 150 min. For the p-polarized light irradiation, the surface morphology shows no obvious difference compared with that after the light irradiation for 120 min (Figure 3a8). In contrast, for the irradiation with RCP and s-polarized light for 150 min, more complicated surface morphologies are observed in comparison with those after irradiation for 120 min (Figure 3b8,c8), especially for the case with the s-polarized light. It can be seen from the bottom row of Figure 6b  The complete erasure of the SRGs on IAC-4 films supplies a unique prospect to rewrite and reconfigure the grating on the surface. To prove this, after the SRGs were optically erased, the obtained IAC-4 films were irradiated with interfering laser beams to inscribe the new SRGs. Figure 7 shows the AFM images of SRGs inscribed on a pristine IAC-4 film and the IAC-4 films obtained after the optical erasure with the polarized lights, where the new SRGs were all inscribed with the interfering laser beams (150 mW cm À2 ) for 20 min. As can be seen from the figures, the SRGs inscribed on the IAC-4 films obtained from the optical erasures are the same as that inscribed on the pristine film. This result indicates that there is no change in the photoresponsive properties of IAC-4 and the optical erasure causes no damage to the functional groups. As SRGs can be optically erased and rewritten, the surface structures are reconfigurable for different applications expected for this type of material.

Mechanism of Optical Erasure
A necessary condition for optical erasure is to overcome the potential energy barriers between the molecules and activate molecule motion. If this condition can be met, the effect of surface tension and other effects related to the light irradiation can result in erasure. A study of a thermal erasure process can supply insight into the activation of molecule motion by heat and the effect of surface tension. To compare with optical erasure, the duplicated SRGs of these two types of materials were heated at temperatures above their T g s for a period of time. The heating temperatures were 150°C for BP-AZ-CN and 70°C for IAC-4, which were higher than their respective T g s (137°C for BP-AZ-CN and 63°C for IAC-4). The erasures of the gratings are indicated by the decline of DEs (Figure 8), where the DE decrease of IAC-4 grating is obviously faster compared with that of BP-AZ-CN grating. Figures S12 and S13 (Supporting Information) show the variations of AFM topographic images and profiles of the duplicated SRGs on BP-AZ-CN and IAC-4 films in the thermal erasure processes, which show the erasing behavior consistent with the DE results. Moreover, as manifested by the AFM, the profiles of the gratings keep the sinusoidal shape without obvious distortion in the thermal erasure processes. The quantitative results of DE measurements and the amplitudes of the gratings after the thermal erasure are given in Table S1 (Supporting Information), compared with those from the optical erasure. As IAC-4 is an azo molecular glass, for which the   molecule motion is relatively easy and not restricted by the chain entanglement, the thermal erasure is achieved at a much lower temperature than that of BP-AZ-CN. The grating on IAC-4 film is completely erased by heating at 70°C for 40-60 min, while the grating on BP-AZ-CN film cannot be completely erased even after heating at 150°C for 120 min. The variations of the amplitude, surface area, and volume of a typical grating ridge with the unit length for BP-AZ-CN and IAC-4 are plotted versus the heating time ( Figure S14, Supporting Information). The values all show a monotonic decrease from the beginning without the complication observed in the optical erasure with the p-polarized and RCP lights. The FWHM values of these two types of gratings almost keep constant during the thermal erasure process ( Figure S15, Supporting Information). The thermal erasure is attributed to the effect of surface tension after the molecular motion is activated by heating and the free volume is significantly increased above T g . Due to the effect of the surface tension, the surface area quickly declines in the process ( Figure S14b, Supporting Information). For optical erasure, the activation of molecular migration is also necessary for both azo molecular glass and azo polymer. However, the activation mechanism is not the same as that caused by the thermal energy. It has been proved by previous studies that the temperature increase caused by laser irradiation is less than 10°C for thin films. [60,66] This point can also be seen from the aforementioned result that the optical erasure only depends on the exposure energy, while the thermal effect should be more significant for the light irradiation with a higher intensity and shorter exposure time. For the SRG inscribing as an inverse process of erasure, the migration of azo polymers is activated by the repeated trans-cis isomerization cycles of push-pull-type azo chromophores under the light irradiation, [7,8] which causes the free volume to increase as well. [21,22,67] Photoinduced isomerization likewise activates the molecule motion in the optical erasure process. In contrast, distinct from the overriding role of surface tension for thermal erasure, optical erasure can also be caused by the mass transfer along the electric vibration direction of the polarized light, which has been well documented in the previous reports. [23,24,[53][54][55] The virtual importance of these two effects is different for the azo polymer (BP-AZ-CN) and azo molecular glass (IAC-4), which results in the differences in the optical erasure behavior. The optical erasure of the duplicated SRGs on IAC-4 film does not cause obvious profile shape change during the erasure process (Figure 4), and there is almost no difference in erasure behavior upon the irradiation with the three types of polarized lights (Figures 2c,d, and 4). These results indicate that the optical erasure of IAC-4 grating is mainly caused by the surface tension effect to reduce the surface area after the molecular motion is activated by the light irradiation. On the other hand, the optical erasure of the BP-AZ-CN grating not only has a much slower rate (Figure 2a,b), but also appears to be much more sensitive to the change in the light polarization ( Figure 3). It means that besides the surface tension effect, the directional mass transfer in the light polarization direction plays an important role in the erasure process.
The distinctive differences between BP-AZ-CN and IAC-4 are attributed to the relative easiness of the molecular migration under light irradiation, which can be seen from the following two points. First, the glass transition of a polymer occurs at the temperature when the free volume in the system is significantly increased and the thermal energy can overcome the potential energy barriers for the segment rotation and transition. [68] The much higher T g of BP-AZ-CN compared with that of IAC-4 indicates that a larger amount of energy is required for the former to increase the free volume and trigger the segment motion. In other words, the molecular migration of BP-AZ-CN is much more difficult to be activated in comparison with IAC-4. Second, BP-AZ-CN and IAC-4 show the different variation trends of the surface moduli after the optical erasure. As shown by the DMT modulus histogram curves ( Figure S11a, Supporting Information), the overall average modulus of BP-AZ-CN significantly increases after the optical erasure, though the moduli of some areas on the surfaces become lower. In contrast, after erasing the IAC-4 gratings with the polarized lights, the DMT moduli of the surfaces are all lower than that of the pristine film ( Figures S11b, Supporting Information). Although the measurement was performed after the light irradiation, the local structures were partially "frozen" when switching off the radiation as the relaxation was a relatively slow process. Therefore, the results reflect the moduli of the surfaces under the light irradiation, where substantially higher moduli are observed for BP-AZ-CN. These results show that the molecular motion of IAC-4 is relatively easy to be activated by light irradiation. In this case, the optical erasure is mainly caused by the surface tension and thus shows a weak correlation with the light polarizations. In contrast, as the segment migration of BP-AZ-CN is more difficult to be activated and suffers from the restriction of the chain entanglement as well. Therefore, the driving force of the surface tension alone is not enough to erase the grating. In this case, the optical effect to drive the mass transfer along the electric vibration direction of the polarized light becomes much more important. As the local structures after the forced migration are partially "frozen" in the ridge areas of the erased gratings, these regions show distorted shapes and higher DMT moduli after the irradiation with the p-polarized and RCP lights (Figures 3a,b, and 6b,c). The effects of the light polarizations on the optical erasure behavior for BP-AZ-CN and IAC-4 are further discussed in the following section.

Effects of Light Polarization States
For the case with p-polarized light, the optical erasure of BP-AZ-CN grating exhibits three stages. In the first stage (the first 5 min), the DE shows some increase with the irradiation (Figure 2a,b). The amplitude and surface area of the grating ridge also show a slight increase, and the volume shows a more obvious increase in the process (Figure 5a-c). These variations are believed to be caused by the photoinduced orientation of the azo chromophores, which preferentially rotate to the positions perpendicular to the light polarization. [8,15] The building-up stress from the molecule orientation can cause yielding deformation in the azo polymer film. [31] In the second stage, the DE shows a sharp fall (Figure 2a,b), and the amplitude, surface area, and volume of the representative ridge also abruptly decrease (Figure 5a-c). The morphologies of the gratings exhibit a dramatic change from the sinusoidal profiles to those with split crests, broadened ridges, and narrow groove bottoms www.advancedsciencenews.com www.adpr-journal.com (Figure 3a3-3a6). It indicates that the mass transfer along the light polarization direction plays an important role in this stage besides the surface tension effect. It is the mass transfer from ridges to grooves that results in the broadened ridges and doublet crests. The distributions of the DMT modulus curves also give evidence of this effect ( Figure S11a, Supporting Information). As the directional mass transfer causes the rapid decline of the amplitude and volume of the ridges, the p-polarized light shows the most efficient way to erase the grating in this stage.
As the surface area of the grating ridge increases in the first stage, the enhanced effect of the surface tension could also contribute to the acceleration of the erasing process to some degree.
In the third stage, as both the ridge amplitude and the surface area have small values, the effects of mass transfer and surface tension become weak. In this stage, the optical erasure turns out to be slow and even uncompleted. A relatively simple variation mode was observed for the optical erasure of BP-AZ-CN grating with the s-polarized light. The key difference for the s-polarized light is that the electric vibration direction is now parallel to the grating ridges and grooves. As no preferential molecule orientation exists in the duplicated SRGs, the trans-cis photoisomerization rate shows no distinction between the s-and p-polarized lights, and thus the driving force to cause the erasure should be quite similar. In contrast, due to the restriction to the mass transfer in the direction parallel to the grating ridges, it could hardly occur upon the light irradiation. In the first several minutes, the amplitude, surface, and volume of the grating ridge show slow decreases, which can be attributed to the time required to activate the molecule motion and the effect of molecular orientation discussed earlier. After this period, the s-polarized light shows the erasure behavior similar to that of the thermal erasure, which is mainly caused by the surface tension effect. The erasure caused by the surface tension is distinct from that caused by the mass transfer from ridges to grooves, which shows no distortion of the grating profiles. Only when the grating was irradiated with the s-polarized light for a sufficiently long time (150 min), the spontaneous surface buckling occur to show the irregular surface patterns (Figure 6d). The buckling and surface instability are caused by the building-up stress exceeding a threshold value when the surface expansion is confined. [69,70] The surface instability was also observed for the optical erasure of optically inscribed SRGs after the light irradiation for a long time. [48] The surface expansion and building-up stress mainly appear in the direction parallel to the light polarization. For azo polymer film under single beam exposure, the formed wrinkle-like structures are thus perpendicular to the light polarization. [71] In the case of the erasure with s-polarized light reported here, the expansion in the direction parallel to the grating ridges is confined. It causes the buckling structures formed on the ridges due to the building-up stress (Figure 6d). In contrast, for the erasure with p-polarized light, the expansion direction is perpendicular to the stripes, which only causes the ridges to become broader (Figure 6b). In the optical erasure process, the surface tension and directional mass transfer effects are dominant for almost the whole process. Only when the input electromagnetic energy exceeds the threshold value, the surface tension effect to minimize the surface area is overridden by the surface buckling, where the surface irregular structures appear for the irradiation with the s-polarized light.
The optical erasure of BP-AZ-CN grating with the RCP light shows the features between the p-and s-polarized lights, as can be seen from the DE and surface profile variations (Figure 2a,b,  and 5a-c). The crest bifurcation phenomenon is less obvious compared with that observed in the erasure with the p-polarized light (Figure 3b5-b7). Similar to the case of s-polarized light, the buckling appears after the light irradiation for 150 min (Figure 6c), but it is relatively weak. All these observations can be attributed to the fact that RCP is composed of the s-and p-polarized components with π/2 phase retardation.
For SRGs on IAC-4 films, the observed minor difference in the erasure behavior between the s-and p-polarized lights can also be attributed to the mass transfer in the polarization direction, but it is a much weaker effect compared with that observed on BP-AZ-CN grating. For the s-polarized light, the electric vibration direction is parallel to the grating ridges and grooves, for which the mass transfer is restricted and limited. Therefore, light irradiation mainly acts to increase the free volume and activate molecular motion, the erasure is virtually a result of surface tension similar to the thermal erasure. Only a slight increase of FWHM of the grating ridge was observed during the erasure process. In contrast, as the polarization of the p-polarized light is perpendicular to the grating ridges, the mass transfer in this direction causes the FWHM to increase in the first 3 min of the optical erasure ( Figure S9b, Supporting Information). As RCP light is the superposition of the s-and p-polarized lights with the π/2 phase retardation, it shows the erasure behavior between the two cases of the linearly polarized lights (Figure 5d-f, S9b, Supporting Information). When the irradiation time is long enough for activating the molecular migration (such as 12 min), the surface tension becomes the overriding effect and the lights with different polarizations almost show no difference in the optical erasure results, which can be seen from the FWHM curves given in Figure S9b (Supporting Information). After the light irradiation for 30 min, no buckling was observed for the s-polarized and RCP lights, where the surface tension is still the dominant effect for the optical erasure of the IAC-4 grating.

Discussion
The use of the duplicated SRGs supplies a new way to investigate the optical erasure. As no photoinduced chromophore orientation exists in the gratings, which is the main reason to cause the complexity, the results obtained are more feasible to be analyzed by considering the influences of the erasure conditions only. The aforementioned results not only reveal the clear distinction between the azo polymer and azo molecular glass, but also show that the complicated erasure behavior is mainly related to the azo polymer under the p-polarized and RCP light irradiation. The observed increases of the DE and amplitude in the first stage of the optical erasure is consistent with the reported increases of these values for the erasure of SRGs inscribed by interfering laser beams. [47] In the second stage of the optical erasure, the erasure rate and overall efficiency show the order p-polarized > RCP > s-polarized. The most efficient erasure effect observed for the p-polarized light in this stage is in accord with the previous result using optically inscribed SRGs. [48] However, as the gratings used in those reported investigations www.advancedsciencenews.com www.adpr-journal.com were fabricated by the irradiation with the interfering s-, p-, and RCP lights, the "memory" of the grating formation conditions through the chromophore orientation makes the analyses of the results to be much more complicated. In contrast, the observations of the current study can be well rationalized by considering the relative importance of the surface tension and directional mass transfer in the optical erasure, which are found to be some key factors to control the processes. From the perspective of applications, these two types of materials have their own advantages and weaknesses. Although SRGs are extremely stable at a temperature below T g s and unaffected under the irradiation at a wavelength longer than the absorption bands, they are erasable in different degrees upon the light irradiation at 488 nm. The gratings on IAC-4 films can be more easily erased by light irradiation, which shows the advantages of reversible SRG writing and erasing. This property is appealing for optical device fabrication and other applications that need reconfigurable function. In contrast, SRGs on BP-AZ-CN films have much better stability than that of IAC-4. The gratings and other surface structures on BP-AZ-CN films are optically stable for the light irradiation even under the actinic light irradiation for minutes, and not easy to be erased under the conditions, which are suitable for use as DOEs to endure the light irradiation. This property is also useful for the application of SRGs as processing tools such as phase masks for direct pattern transfer with light. [47,48]

Conclusions
In this study, the optical erasure is thoroughly investigated by using SRGs obtained from hot embossing soft-lithography. This new approach aims at addressing the long-standing controversial issue of optical erasure, where the complexity is caused by the "memorized" or "retroactive" effects existing in optically inscribed SRGs, to supply a better understanding of this intriguing function. Through this new approach, the systematic study of duplicated BP-AZ-CN and IAC-4 gratings with respect to their optical erasure with 488 nm lights was carried out under comparative conditions for the first time. As revealed by DE measurement, the optical erasure of SRGs on IAC-4 films demands a much shorter time compared with that of BP-AZ-CN, which shows no obvious dependence on the light polarization states. As revealed by AFM, the profiles of the IAC-4 gratings maintain the sinusoidal shape during the erasing process, while the amplitude, surface area, and volume of the grating ridges decline monotonically. Due to the completely erasable property, new gratings can be inscribed on the IAC-4 films obtained from the optical erasure, which shows no difference from the use of pristine films. On the other hand, the optical erasure of the BP-AZ-CN gratings not only has a much slower rate, but also is closely related to the light polarization states. For the optical erasure with the p-polarized light, the DE and amplitude of the duplicated SRGs show a slight increase in the first 5 min. It is followed by the rapid decline of these values with the irradiation time, and the erasure process becomes slow when approaching the steady state. In the rapid erasure stage, the shapes of the gratings are dramatically altered from the sinusoidal profiles, which show the split crests, broadened ridges, and narrow groove bottoms. In contrast, for the erasure with the s-polarized light, the DE and amplitude of the duplicated SRGs show a slow decrease in the first few minutes. Then, the rapid decline of these values with the irradiation time is observed, which is followed by a steady decrease before reaching a stable state. In the rapid erasure stage, the profiles of the gratings remain the sinusoidal shapes and the erasure is slower compared with that of the p-polarized light. The optical erasure with the RCP light shows behavior between the p-and s-polarization states, as it is the superposition of these two components with π/2 retardation. The aforementioned optical erasure behavior can be rationalized by considering the relative importance of the mass transfer along the electric vibration direction of the polarized light. For IAC-4 grating, this effect is weak and the optical erasure is mainly caused by the surface tension effect similar to the thermal erasure, though the molecular motions are activated by trans-cis photoisomerization cycles instead of thermal energy. In contrast, due to the higher potential energy barriers to activating the molecule motion and the restriction of the chain entanglement of BP-AZ-CN, the surface tension effect itself is not enough to cause the erasure and the directional mass transfer driven by the polarized light plays a much more important role in the process, especially for the erasure with the p-polarized and RCP lights. This understanding can not only clarify some controversial issues, but also supply a solid knowledge base for applications of the materials.

Experimental Section
Materials: Diglycidyl ether of bisphenol-A (BP), aniline (AN), and 4-aminobenzonitrile were purchased from Sigma-Aldrich. Isosorbide, 4-nitrobenzoyl chloride, and N,N-di(hydroxyethyl)aniline were purchased from J&K Scientific Ltd. Cinnamoyl chloride was purchased as a commercial product from Alfa Aesar. These compounds were used for preparing the epoxy-based azo polymer (BP-AZ-CN) and the azo molecular glass (IAC-4). The elastomer base and curing agent (w/w ¼ 10:1) of PDMS were purchased from Dow Corning (Sylgard 184). Deionized water (resistivity > 18.25 MΩ cm) was obtained from a Millipore water purification system and used for all experiments. Glass slides were treated with the H 2 SO 4 /30% H 2 O 2 (7:3) mixture (Caution: this solution is extremely corrosive) for 3 h, rinsed with acetone and washed with plenty of deionized water for several times. Unless otherwise stated, all other chemicals and solvents, such as N,N-dimethylformamide (DMF) and dichloromethane (DCM), were purchased commercially and used as received without further purification. BP-AZ-CN was obtained through post-polymerization azo-coupling reaction between a precursor polymer (BP-AN) and the diazonium salt of 4-aminobenzonitrile, where BP-AN was prepared by step polymerization between BP and AN. [59] The functionalization degree of BP-AZ-CN was about 100% as proved by 1 H NMR spectroscopy ( Figure S1, Supporting Information). The number-average molecular weight of BP-AZ-CN was estimated to be 18 500 with a polydispersity index (PDI) of 1.78 ( Figure S4, Supporting Information). BP-AZ-CN has a strong absorption band in the visible light range with λ max ¼ 467 in DMF ( Figure S5a, Supporting Information). The glass transition temperature (T g ) of BP-AZ-CN was measured to be 137°C ( Figure S6a, Supporting Information). IAC-4 was synthesized by an azo-coupling reaction between isosorbide bis(4-aminobenzoat) and di(hydroxyethyl)aniline biscinnamate as reported in the previous literature. [36] Briefly, isosorbide bis(4-aminobenzoat) was prepared by esterification reaction of isosorbide with 4-nitrobenzoyl chloride and the subsequent reduction reaction of isosorbide bis(4-nitrobenzoat). Di(hydroxyethyl)aniline biscinnamate was obtained through the esterification reaction between N,N-di(hydroxyethyl)aniline and cinnamoyl chloride. IAC-4 is a low-molecular-weight compound with a molecular weight of 1289.4, which has a strong absorption band in the visible light range with λ max ¼ 443 nm in DMF ( Figure S5b, Supporting Information). The glass transition temperature (T g ) of IAC-4 was measured to be 63°C ( Figure S6b, Supporting Information). The synthesis and characterization details of BP-AZ-CN and IAC-4 had been reported in our previous articles, [36,59] which can also be seen in the Supporting Information.
Characterization: 1 H-NMR spectra were recorded at 25°C on a JEOL JNM-ECA600 NMR spectrometer (Japan, 600 MHz for proton) by using dimethyl sulfoxide-d 6 (DMSO-d 6 ) or chloroform-d (CDCl 3 ) as the solvent and tetramethylsilane (TMS) as the internal standard. Fourier transform infrared (FT-IR) spectra were acquired by a Nicolet Magna-IR 560 spectrophotometer (Thermo Fisher, USA) by incorporating samples in KBr powder and then pressing them into thin IR-transparent disks. The number-average molecular weight and PDI of BP-AZ-CN were measured by using a gel permeation chromatograph (GPC, Shimadzu, Japan) with THF as eluent at a flow rate of 1.0 mL min À1 . The apparatus was equipped with a RID-20A refractive index detector, a LC-20AD liquid chromatograph, and a CTO-20AC column oven. The measurement was carried out at 25°C and the data was calibrated with linear polystyrene standards. UV-vis absorption spectra of the samples were obtained by an Agilent Cary 8453 UV-vis spectrophotometer (USA). The thermal analysis, including differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) of the materials, was carried out by utilizing TA Instruments DSC Q2000 and TGA Q5000 systems (USA) with a heating rate of 10°C min À1 in a nitrogen atmosphere. Nanoindentation tests were performed on a Nano Indenter G200 (Keysight) system using a Berkovichtype diamond indenter tip. The maximum indentation depth was 500 nm with a loading rate of 10 nm s À1 . The BP-AZ-CN and IAC-4 films for nanoindentation tests were prepared by dip-coating the DMF solutions (10 wt%) of the materials onto clean glass slides, evaporating solvents at 60°C and drying in a vacuum oven for 48 h. The thicknesses of the films were controlled to be more than 10 times the maximum indentation depth, which ensured the absence of substrate effect.
SRG Inscribing: SRGs were optically inscribed on IAC-4 films by a method reported in the previous articles, [18,51,59] which were used as masters for soft-lithographic replication. Firstly, a suitable amount of IAC-4 was dissolved in DMF to obtain a homogeneous solution with a concentration of 0.2 g mL À1 . Then, the solution was filtered through a 0.45 μm membrane and spin-coated onto clean glass slides to obtain the films with smooth surfaces. By adjusting the spinning speed, the thicknesses of the films were controlled to be about 1.0 μm. All the films were dried in a vacuum oven at 50°C for 24 h to remove the residual solvent before being used. For inscribing SRGs, a semiconductor laser (Genesis CX 488-2000 SLM, Coherent Corporation, λ ¼ 488 nm) was used as the light source. The laser beam was expanded with a spatial filter and collimated by a plano-convex lens to generate a homogeneous beam with an intensity of 100 mW cm À2 . By using the Lloyd mirror setup, one half of the p-polarized beam was reflected onto the IAC-4 film surface and coincident with the other half beam that was directly incident onto the surface to form an interference pattern. The spatial periods of the gratings were controlled by adjusting the incident angle of the Lloyd mirror. The amplitudes of the gratings were adjusted by the irradiation time (20 min typically). The SRGs were inscribed at room temperature under air-ambient conditions.
Soft-Lithographic Replication: The duplicates of the SRGs were obtained by using the hot embossing soft-lithographic method reported before. [62] In the process, the elastomeric PDMS stamps were prepared by replica molding against the grating masters and used to prepare the azo polymer and azo molecular glass gratings through the hot embossing. First, the PDMS stamps were prepared by casting a 10:1 (w/w) mixture of the silicone elastomer base and curing agent on the molds of IAC-4 SRGs. The prepolymer was stationed in a vacuum oven for about 15 min to remove air bubbles and then poured onto the grating masters. After being cured in a 50°C oven overnight, the PDMS stamps were obtained by peeling off the molds. Second, for the hot embossing, the PDMS stamps were placed in a conformal contact manner on the surfaces of BP-AZ-CN and IAC-4 films. Moderate pressure was used to keep the stamps in seamless touch with the sample films. The PDMS stamps and films were heated together on a hot stage to 190°C (for BP-AZ-CN) or 110°C (for IAC-4) and stayed for 30 min. Then, the temperature was gradually lowered to 60°C, while keeping the stamps in conformal touch with the films. After the PDMS stamps were carefully peeled off at this temperature, the surface patterns of the original SRGs were successfully transferred onto the BP-AZ-CN and IAC-4 films.
Optical Erasure: A beam of linearly polarized light from a semiconductor laser (λ ¼ 488 nm, Genesis CX 488-2000 SLM, Coherent Corporation) was used as the light source, which was spatially filtered, expanded, and collimated to generate a homogeneous beam. A polarizer was placed in the light path to get higher linear polarization for the linearly polarized light irradiation. A quarter-wave (λ/4) plate, where its fast ax was oriented by 45°t o the vibration direction of the linearly polarized incoming beam, was applied to achieve circular polarization for circularly polarized light irradiation. The duplicated SRGs were irradiated by the laser beam impinging perpendicularly onto the film surfaces for different time periods. For measuring the diffraction efficiency in real-time, a low-power He-Ne laser beam (λ ¼ 632 nm, JDW-3, Peking University) was incident nearly perpendicularly to the central position of the SRG sample. The light intensity exhibited the Gaussian distribution in the beam cross-section with a diameter of 2 mm, and the total intensity was measured to be 1.2 mW using a silicon photodiode (Thorlabs, 400-1100 nm). The light intensities of the zero-order beam (I 0 ) and first-order diffraction beams (I 1 and I À1 ) were recorded with three silicon photodiodes (Thorlabs, USB Power Meter PM16-120, 400-1100 nm) to obtain data for calculating the diffraction efficiencies. The data were transmitted to a computer and processed with Power Meter Driver Switcher software (Thorlabs) in a real-time manner. All optical erasure experiments were carried out at room temperature under an air-ambient condition.
AFM Measurement: An AFM apparatus (Dimension Icon, Bruker Corporation, USA), equipped with a Nanoscope V controller and NanoScope 8.15 software, was used to characterize the SRG surface variations before and after the laser irradiation. For SRG morphology measurement in the AFM test, the tapping mode was adopted with an amplitude set-point of about 320 mV, while the probe RTESPA-300 had a tip normal radius of 8 nm. The nanomechanical properties of BP-AZ-CN and IAC-4 were determined in the AFM PeakForce QNM mode. During the experiments, a series of force-indentation curves were obtained through PeakForce Tapping technology, and then the force curves were fitted and analyzed using the DMT model, [63] which is described in the Supporting Information ( Figure S16 and S17). Based on Bruker's protocol, [64] the standard cantilever holder and the probe RTESPA-525 were used for operation. The deflection sensitivities of the cantilevers were calibrated using a sapphire standard provided in the PF-QNM kit, and the spring constant k was calculated using the Sader method. [65] The tip radius R of the probe was calibrated by the relative method on a polystyrene test sample (Bruker, PS Film, 2.7 GPa). The spring constant k of the cantilever was measured to be 120-160 N m À1 for the probe RTESPA-525, and the tip radius R of the probe RTESPA-525 was found to be in the range of 20-50 nm. The feedback gain value was dynamically and automatically controlled by the software for good sample tracking. The peak force set-point used in PF-QNM measurements was 300 nN for both BP-AZ-CN and IAC-4. The vertical oscillation frequency of the probe during the measurements was equal to 2 kHz with an amplitude of 75 nm. The lateral scanning rates were 0.5 Hz for images smaller than 5 Â 5 μm 2 and 0.2 Hz for images larger than 5 Â 5 μm 2 at a digital resolution of 256 Â 256 pixels.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.