Control of the macroscopic shape of assemblies formed from microparticles based on host–guest interaction

Biological macroscopic assemblies have inspired researchers to utilize molecular recognition to develop smart materials in these decades. Recently, macroscopic self-assemblies based on molecular recognition have been realized using millimeter-scale hydrogel pieces possessing molecular recognition moieties. During the study on macroscopic self-assembly based on molecular recognition, we noticed that the shape of assemblies might be dependent on the host–guest pair. In this study, we were thus motivated to study the macroscopic shape of assemblies formed through host–guest interaction. We modied crosslinked poly(sodium acrylate) microparticles, i.e., superabsorbent polymer (SAP) microparticles, with β-cyclodextrin (βCD) and adamantyl (Ad) residues (βCD(x)-SAP and Ad(y)-SAP microparticles, respectively, where x and y denote the mol % contents of βCD and Ad residues). Then, we studied the self-assembly behavior of βCD(x)-SAP and Ad(y)-SAP microparticles through the complexation of βCD with Ad residues. There was a threshold of the βCD content in βCD(x)-SAP microparticles for assembly formation between x = 22.3 and 26.7. On the other hand, the shape of assemblies was dependent on the Ad content, y; More elongated assemblies were formed at a higher y. This may be because, at a higher y, small clusters formed in an early stage can stick together even upon collisions at a single contact point to form elongated aggregates, whereas, at a smaller y, small clusters stick together only upon collisions at multiple contact points to give rather circular assemblies. On the basis of these observations, the shape of assembly formed from microparticles can be controlled by varying y.


Introduction
Biological systems utilize macroscopic self-assemblies based on molecular recognition. In these decades, biological macroscopic assemblies have inspired researchers to utilize molecular recognition to develop smart materials, e.g., soft actuators 1-6 and self-healing materials [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] . Recently, macroscopic selfassemblies, which allow ones to see molecular recognition by naked eyes, have been realized using millimeter-scale hydrogel pieces possessing molecular recognition moieties [23][24][25][26][27] . Even if the difference in binding constants of molecular recognition moieties is small, the systems of mixture of gel pieces can exhibit perfect delity by controlling the contents of molecular recognition moieties [28][29][30] . During the study on macroscopic self-assembly based on molecular recognition, we noticed that the shape of assemblies might be dependent on the host-guest pair; For example, the pair of β-cyclodextrin (βCD) and adamantyl (Ad), which shows one of the highest binding constants, led to the formation of linear alternating assemblies, whereas βCD and t-butyl, which possesses a lower binding constant, might form planar checkered assemblies 23 . We were thus motivated to study the macroscopic shape of assemblies formed through host-guest interaction. We have chosen crosslinked poly(sodium acrylate) microparticles, i.e., superabsorbent polymer (SAP) microparticles, because of the availability and ease of modi cation. We have employed a combination of βCD and Ad as host and guest residues, respectively, because this combination forms stable inclusion complexes. In the present article, we describe preparation of SAP microparticles modi ed with βCD and Ad residues (βCD(x)-SAP and Ad(y)-SAP microparticles, respectively, where x and y denote the mol % contents of βCD and Ad residues), and the self-assembly behavior of βCD(x)-SAP and Ad(y)-SAP microparticles through the complexation of βCD with Ad residues, and discuss the effect of the contents of residues, i.e., x and y, on the self-assembly behavior.

Preparation of Superabsorbent Polymer Microparticles Modi ed with β-Cyclodextrin and Adamantyl
Residues. SAP microparticles were modi ed with βCD and adamantyl (Ad) residues (βCD(x)-SAP and Ad(y)-SAP microparticles, respectively, where x and y denote the mol % contents of βCD and Ad residues) by coupling of mono-(6-amino-6-deoxy)-β-CD (βCD-NH 2 ) and 1-adamantanamine hydrochloride (Ad-NH 2 •HCl) with carboxylic acid residues in SAP microparticles using 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4methylmorpholinium chloride n-hydrate (DMT-MM) as a coupling agent (Scheme 1). The βCD(x)-SAP and Ad(y)-SAP microparticles were puri ed by washing repeatedly with water for a few days. The βCD(x)-SAP and Ad(y)-SAP microparticles prepared were characterized by solid-state 1 H eld gradient magic angle spinning (FGMAS) NMR and attenuated total re ectance (ATR) FT/IR spectroscopy (Figures S1 and S2 in Supporting Information, respectively). As can be seen in Figure S1, the 1 H FGMAS NMR spectra exhibited signals ascribable to βCD and Ad residues for βCD(x)-SAP and Ad(y)-SAP microparticles at 5.32 and 1.64 ppm, respectively. As shown in Figure S2, the ATR FT/IR spectra contain absorption bands ascribable to the stretching vibration of C=O of the amide bond at ca. 1650 cm -1 . These spectra are indicative of successful modi cation of SAP microparticles.
The contents of βCD and Ad residues in βCD(x)-SAP and Ad(y)-SAP microparticles, x and y, were roughly estimated by the weights of remaining βCD-NH 2 and Ad-NH 2 •HCl recovered from the reaction mixtures, as listed in Tables 1 and 2. The βCD and Ad contents in βCD(x)-SAP and Ad(y)-SAP microparticles are markedly lower than those in feed. It should be noted here that the differences in contents are much larger for the Ad(y)-SAP microparticles than those for βCD(x)-SAP microparticles. This may be caused by the difference in reactivities of βCD-NH 2 and Ad-NH 2 •HCl. βCD-NH 2 is a primary amine, whereas Ad-NH 2 •HCl is a tertiary ammonium. The hydrochloride, Ad-NH 2 •HCl, was used because of the low solubility of 1-adamantanamine in water. The pH values of reaction media were thus different for modi cation with βCD and Ad residues; pH = 8.1 and 6.3 for βCD and Ad residues, respectively. Since, in amide coupling reactions using DMT-MM, the 4,6-dimethoxy-1,3,5-triazine residue is attacked by a carboxylate ion, the coupling proceeds less e ciently at lower pH 31 . Therefore, the lower pH upon modi cation with Ad residues may be partly responsible for the reduced Ad contents in Ad(y)-SAP microparticles. It should be also noted that SAP microparticles exhibited different absorption properties for βCD-NH 2 and Ad-NH 2 •HCl. The a nities of SAP microparticles for βCD-NH 2 and Ad-NH 2 •HCl were compared by adsorption experiments. SAP microparticles fully swelled with water (2.0 g) were added to an aqueous solution of βCD-NH 2 or Ad-NH 2 •HCl (5.0 mM, 10 mL). After equilibration, the concentration of solute, i.e., βCD-NH 2 or AdNH 2 •HCl, and the volumes of aqueous solution and SAP microparticles were determined. Using these data, the amounts of βCD-NH 2 and Ad-NH 2 •HCl adsorbed were evaluated to be ca. 52 and 2.3 mg, respectively, per 1 g of SAP microparticles, indicating that βCD-NH 2 is adsorbed into SAP microparticles more strongly than Ad-NH 2 •HCl. It is thus likely that βCD residues were introduced e ciently not only to carboxylate residues on the surface of SAP microparticles but also to those inside of SAP microparticles, whereas Ad residues were introduced dominantly to carboxylate residues on the outer layer of SAP microparticles. The average diameters (D av ) for a hundred of βCD(x)-SAP and Ad(y)-SAP microparticles were determined by observing on an optical microscope. Values of D av were plotted in Figure 1 against the βCD or Ad content, i.e., x or y. The D av values (ca. 100 μm) for βCD(x)-SAP microparticles were signi cantly smaller than that for unmodi ed SAP microparticles, whereas the D av values (ca. 190 μm) for Ad(y)-SAP microparticles were almost the same as that for SAP microparticles independent of y. The smaller D av values for βCD(x)-SAP microparticles indicate that βCD residues in βCD(x)-SAP microparticles interact with each other or with carboxylate residues presumably through hydrogen bonding formation, resulting in a reduced swelling ratio of microparticles. These observations may support the different fashions of modi cation for βCD and Ad residues.
For visual discrimination, dyeing of Ad(y)-SAP microparticles was tested using several dyes. The food pigments, which we had used for poly(acrylamide)-based gels in our previous studies 23,24,28-30 , were not useful in this study because the pigment molecules quickly came out of Ad(y)-SAP microparticles. The results of dyeing test indicated that pararosaniline was a useful dye presumably because of electrostatic interaction between the ammonium in pararosaniline and carboxylate residues in Ad(y)-SAP microparticles ( Figure S3 in Supporting Information). Under the dyeing conditions in this study, the size distributions of Ad(y)-SAP microparticles were almost the same before and after dyeing (data not shown). It is thus likely that pararosaniline has no effect on the properties of Ad(y)-SAP microparticles and on the interaction of βCD(x)-SAP and Ad(y)-SAP microparticles. In the following, Ad(y)-SAP microparticles dyed with pararosaniline were used; Colorless and red particles are βCD(x)-SAP and Ad(y)-SAP microparticles, respectively.
Interaction of βCD(x)-SAP and Ad(y)-SAP Microparticles. The interaction of βCD(x)-SAP and Ad(y)-SAP microparticles was investigated by observing on an optical microscope. A suspension containing ca. 50 βCD(x)-SAP microparticles in water (1 μL) and a suspension containing ca. 50 Ad(y)-SAP microparticles in water (1 μL) were mixed on a clean slide glass. Since the SAP microparticles used in this study did not undergo practically Brownian motion because of their larger size, the mixed suspension was agitated at 500 rpm using a mixer to investigate the interaction. Typical examples of assemblies formed are shown in Figure 2. The results of interaction are summarized in Table S1 in Supporting Information. In this table, "-" denotes no assembly formation from βCD(x)-SAP and Ad(y)-SAP microparticles, and "+" denotes the formation of assemblies from microparticles. βCD(x)-SAP microparticles of x = 16. As can be seen in Figure 2, the shape of assemblies seems to be dependent on the Ad content, y. The assembly formed at y = 15.1 (Figure 2b) shows a more elongated shape than does that formed at y = 5.2 ( Figure 2a). Thus, the assemblies formed by agitating βCD(26.7)-SAP and Ad(y)-SAP microparticles of different y at 500 rpm were analyzed by ellipse t using an ImageJ software 33 to evaluate the aspect ratio (a/b, where a and b are the longer and shorter axes, respectively). The same experiments were repeated three times, and the average value of a/b was calculated. Figure 4 demonstrates the aspect ratio, a/b, as a function of the Ad content, y. The aspect ratio, a/b, increases from ca. 1.5 to ca. 2.8 with increasing y from 5.2 to 15.1. These observations indicate that assemblies of a larger aspect ratio are formed at a higher y, i.e., upon stronger interaction.

Discussion
As can be seen in Figure 4, there is a remarkable correlation between the Ad content (y) in Ad(y)-SAP microparticles and the aspect ratio (a/b) of assemblies formed from βCD(26.7)-SAP and Ad(y)-SAP microparticles. In the assemblies, the interaction between βCD(26.7)-SAP and Ad(y)-SAP microparticles is stronger at higher y because of the larger number of inclusion complexes of βCD and Ad residues in a contact region. The correlation between y and a/b can be roughly explained as follows. Since the attractive interaction between βCD(26.7)-SAP and Ad(y)-SAP microparticles of a larger y is strong enough, small clusters formed in an early stage can stick together even upon collisions at a single contact point, resulting in the formation of elongated aggregates. On the other hand, at a smaller y, small clusters stick together only upon collisions at multiple contact points, leading to rather circular assemblies.
Although the βCD(x)-SAP and Ad(y)-SAP microparticles are non-Brownian particles, the formation of assemblies from the microparticles is similar to aggregation of colloidal particles which undergo Brownian motion. Colloidal particle aggregation has been studied in detail for a century or longer [34][35][36][37][38][39] .
The aggregates of colloidal particles often take fractal structures which obey where M and R are the mass and radius of colloidal aggregate, respectively, and D f denotes the fractal dimension [40][41][42][43] 44,45 . In the case of aggregates formed from a number of colloidal particles, the density of aggregates is dependent on D f but the shape of aggregates is practically independent on D f . Figure S4 in Since in the present system of βCD(x)-SAP and Ad(y)-SAP microparticles, assemblies are formed through the formation of inclusion complexes of βCD and Ad residues, we should also discuss binary colloidal aggregation through attractive interactions, e.g., electrostatic interaction of oppositely-charged colloidal particles [48][49][50] . In the case of binary colloidal aggregation, it is known that the structure of aggregates is also dependent on the ratio of radii and the ratio of the numbers of colloidal particles of different types [51][52][53][54][55][56] . In the present study, the ratio of average diameters, D av , was approximately 2 and the ratio of the number of particles was set to almost unity. It should be noted here that the randomness of collision is also important. In the system of βCD(x)-SAP and Ad(y)-SAP microparticles, the collisions of microparticles, i.e., non-Brownian particles, may not be random because the mixtures of microparticles were agitated with a mixer. On the basis of these considerations, we can conclude that the assembly of βCD(x)-SAP and Ad(y)-SAP microparticles exhibited a strong correlation between y and a/b because of the limited conditions, e.g., the limited number of particles, the ratios of radii and the number of microparticles, and the mixing procedure. It may be possible to control the structure of assemblies of βCD(x)-SAP and Ad(y)-SAP microparticles in a different manner under other conditions.

Conclusions
In this study, the self-assembly behavior of βCD(x)-SAP and Ad(y)-SAP microparticles was investigated. SAP microparticles were modi ed with βCD and Ad residues by coupling of βCD-NH 2 and Ad-NH 2 •HCl with carboxylic acid residues in SAP microparticles using DMT-MM as a coupling agent. The contents of βCD and Ad residues in βCD(x)-SAP and Ad(y)-SAP microparticles, x and y, were roughly estimated by the weights of remaining βCD-NH 2 and Ad-NH 2 •HCl recovered from the reaction mixtures. The adsorption experiments indicated that βCD residues were introduced e ciently not only to carboxylate residues on the surface of SAP microparticles but also to those inside of SAP microparticles, whereas Ad residues were introduced dominantly to carboxylate residues on the outer layer of SAP microparticles. For visual discrimination, Ad(y)-SAP microparticles were dyed with pararosaniline. The interaction of βCD(x)-SAP and Ad(y)-SAP microparticles was monitored with an optical microscope using aqueous suspension containing ca. 50 microparticles in each on a clean slide glass. Since the SAP microparticles used in this study did not undergo practically Brownian motion because of their larger size, the mixed suspension was agitated at 500 rpm using a mixer to investigate the interaction. βCD(x)-SAP microparticles of x = 16.2 or 22.3 did not adhere to any of the Ad(y)-SAP microparticles used in this study, whereas βCD(26.7)-SAP microparticles interacted with all the Ad(y)-SAP microparticles examined to form assemblies. These observations indicate that there is a threshold of the βCD content in βCD(x)-SAP microparticles for assembly formation between x = 22.3 and 26.7. Competitive experiments using βCD and AdC as competitors indicated that the formation of assemblies of βCD(26.7)-SAP and Ad(y)-SAP microparticles was based on the inclusion complex formation between βCD and Ad residues on the surface of SAP microparticles, and the interaction of βCD(26.7)-SAP and Ad(y)-SAP microparticles was stronger at higher y. The observation with an optical microscope indicated that the shape of assemblies was dependent on the Ad content, y; the aspect ratio (a/b) increased from ca. (Osaka, Japan). βCD was purchased from Junsei Chemical Co., Ltd. (Tokyo, Japan). Pararosaniline was purchased from Nacalai Tesque, Inc. (Kyoto, Japan). βCD-NH 2 was prepared by previously reported procedure with slight modi cation. 57 Other reagents were used without further puri cation.
Measurements. 1  kHz. ATR FT/IR spectra were measured on a JASCO FT/IR-6100 spectrometer equipped with a diamond ATR accessory. The FT/IR spectrometer was constantly purged with N 2 gas.
Preparation of βCD(x)-SAP and Ad(y)-SAP Microparticles. βCD(x)-SAP microparticles and Ad(y)-SAP microparticles were prepared by amide coupling of carboxylate residues in the SAP microparticles with βCD-NH 2 and Ad-NH 2 •HCl, respectively, using DMT-MM as a coupling agent in water. Predetermined amounts of SAP microparticles, βCD-NH 2 (or Ad-NH 2 •HCl) and water were placed in a reaction vessel.
After cooling the mixture at 5 °C, an aqueous solution (10 mL) of a predetermined amount of DMT-MM was added dropwise to the mixture. After agitating for 18 h at room temperature, the SAP microparticles were recovered by ltration with a glass lter, and the SAP microparticles were thoroughly washed with water to remove impurities soluble in water, i.e., unreacted DMT-MM and βCD-NH 2 (or Ad-NH 2 •HCl) and by-products from DMT-MM. Portions of the βCD(x)-SAP and Ad(y)-SAP microparticles obtained were lyophilized for 1 H FGMAS NMR and ATR FT/IR spectroscopy.
For visual discrimination, the Ad(y)-SAP microparticles obtained were dyed by immersing in an aqueous solution of pararosaniline (0.05 M) for 1 h.
Adsorption Tests. SAP microparticles fully swollen with water (2.0 g) were added to an aqueous solution of βCD-NH 2 or Ad-NH 2 •HCl (5.0 mM, 20 mL) and the mixture was stirred at room temperature (ca. 25 °C) for 18 h. After removing the SAP microparticles by ltration, the concentration of solution of βCD-NH 2 or Ad-NH 2 •HCl was determined by an Anton-Paar DMA5000 density meter using the standard curve prepared separately. Using the volume and concentration of solution, the amount of βCD-NH 2 or Ad-NH 2 •HCl adsorbed per gram of SAP microparticles swollen was estimated.
Interaction of βCD(x)-SAP and Ad(y)-SAP Microparticles. Suspensions of βCD(x)SAP and Ad(y)SAP microparticles (1 μL each) were placed and mixed on a glass plate. The mixture was agitated at ca. 500 rpm using an EYELA CM-1000 cute mixer. The formation of assemblies was observed on an EVOS optical microscope.
Declarations Figure 1 The average diameter (Dav) as a function of x or y for the unmodi ed SAP (black circle), βCD(x)-SAP (black square) and Ad(y)-SAP microparticles (red circle).

Figure 2
Page 16/17 A typical example of optical micrograms for assemblies formed from βCD (27)