Soft solution growth of magnetite-maghemite nanocrystals in crosslinked chitosan templates and their superparamagnetic properties

Abstract Crosslinked chitosan/iron oxide nanocomposites (CC/IO) were synthesized at low temperatures using aqueous systems, i.e. hydrothermal and refluxing methods. The CC templates derived from various concentrations of tripolyphosphate crosslinker were used as host materials. The Fe2+ and Fe3+ ions with 1:2 molar ratio were adsorbed into the CC templates by the swelling, allowing to form CC/Fe2+Fe3+ precursors. The CC/IO nanocomposites were created by treating the precursors in NaOH solution using hydrothermal and refluxing methods. The CC/IO nanocomposites contained magnetite-maghemite nanocrystals with quadrilateral shape of 10 − 14 nm embedded in the CC templates. Superparamagnetism was obtained in the CC/IO nanocomposites, which had maximum magnetization (Mmax) values ranging from 8.6 to 15.2 emu/g and coercivity and magnetic remanence values close to zero. The cell viability of CC/IO nanocomposites ranged from 80 to 89%, demonstrating high safety for mammal. The CC/IO nanocomposites were considered to be potential superparamagenetic candidates for alternative medical applications. Graphical Abstract


Introduction
Magnetic nanoparticles, particularly magnetite (Fe 3 O 4 ) and maghemite (c-Fe 2 O 3 ), are materials with distinct magnetic properties. In bulk, magnetite and maghemite exhibit ferrimagnetism. When their particles are in a single domain and their sizes are less than the superparamagnetic critical radius, their magnetic characteristics change to superparamagnetism. The superparamagnetic behavior is advantageous for operating a variety of modern medical devices for alternative medical therapies, such as magnetic resonance imaging (MRI) contrast agents, drug carriers, biosensors, and heating agents in hyperthermia applications, etc [1][2][3][4][5]. The superparamagnetic materials can improve the specificity of medical treatment to the target organ and immediately demagnetize after treatment termination, resulting in no unwanted magnetism effects. Magnetic nanoparticles, on the other hand, agglomerate due to magnetic and Van der Waals forces. In addition, magnetite and maghemite nanoparticles undergo rapid oxidation, which results in phase transformation. Surface modification and coating are crucial procedures to regulate the shape, aggregation, and particle size of magnetic nanoparticles while preventing oxidation [6][7][8]. Organic compounds like fatty acids and synthetic polymers were widely used as surfactants and coating agents for magnetite and maghemite nanoparticles [9][10][11]. However, these coating agents want to employ organic solvents in the syntheses, and some synthetic polymers can trigger allergic reactions in some people.
To provide new alternatives for allergy sufferers and move more toward an environmentally friendly approach, biopolymers and water-based synthesis play a crucial role for various applications, particularly biomedical applications. Chitosan is a nontoxic, antibacterial, biocompatible, and biodegradable natural linear copolymer of glucosamine and N-acetyl glucosamine units. It is widely used as a biocompatible material in medical applications such as drug delivery, wound healing, hemostatic agent, and fat binder [12][13][14][15]. The chitosan can be crosslinked to generate a three-dimensional hydrogel network that can be used as reaction templates. In our previous research, the chitosan was in-situ ionically crosslinked chitosan with incorporated of iron oxide precursors, resulting in the immersion of iron oxide precursors in the chitosan network. After heat treatment, iron oxide nanoparticles with 3.9-4.3 nm were successfully formed, resulting in chitosan/iron oxide nanocomposites with an iron oxide nanoparticle proportion of 70-75% [16]. It was, however, the wide particle size distribution of iron oxide nanoparticles found in synthetic nanocomposites was thought to have a significant impact on the projected magnetic characteristics.
In order to narrow the particle size distribution of iron oxide and improve the homogeneity of the particle dispersion in chitosan matrix in this study, the preformed 3 D-network of the sodium tripolyphosphate (TPP) crosslinked chitosan were utilized as the host material, creating a uniform confine area for formation of iron oxide nanoparticles. The nucleation and growth of iron oxide nanoparticles embedded in crosslinked chitosan templates were considered to minimize oxidation and increase dispersion, hence improving the superparamagnetic characteristics of the resultant CC/IO nanocomposites. These alternate synthesis techniques presented in this work were carried out in an aqueous-based system employing a low-temperature refluxing and hydrothermal methods. Due to the lack of organic solvent and the laborious process of surface modification and coating of iron oxide nanoparticles, these proposed techniques are cost-effective, easy-to-scaleup, and environmentally acceptable. Furthermore, the resulting CC/IO nanocomposites from this study contained superparamagnetic nanoparticles and a high proportion of chitosan matrix. For alternative medical applications, a large proportion of the chitosan matrix could be loaded or functionalized by diagnostic or bioactive agents, such as medicinal substances, DNA, RNA, enzymes, and so on [17][18]. As a result, the CC/IO nanocomposites could be used to create hybrid materials with specificity to the target organ and modification for specific medical purposes.

Materials
Low average molecular weight chitosan synthesized from squid pen and snow crab shell with deacetylation degree (DD) of ca. >98% was purchased from Bio21 Co. Ltd., Thailand.

Preparation of crosslinked chitosan templates (CC)
The CC templates were prepared by dropwise adding the as-prepared chitosan solution into the TPP solutions having various concentrations, i.e. 0.3, 0.4, 0.5 and 0.6%w/v, with vigorous stirring for 30 min. Finally, the resultant CC templates were filtered, washed with 100 mL distilled water and freeze-dried at À100 C for 48 hr.

Preparation of crosslinked chitosan/iron oxide nanocomposites (CC/IO)
The wet precursors of CC templates, prepared using 0.3, 0.4 and 0.5%w/v of TPP crosslinker as described in section 2.3, were added into the 40 mL of Fe(II):Fe(III) mixture. The CC templates were soaked with magnetic stirring in the Fe(II):Fe(III) mixture for 30 min, resulting in the adsorption of Fe 2þ and Fe 3þ ions in the network structure of CC templates as the CC/Fe 2þ Fe 3þ precursors. The CC/Fe 2þ Fe 3þ precursors were filtered and then added into 100 mL of 1.5 M NaOH solution. These mixtures were treated under two different conditions, i.e. refluxing and hydrothermal methods. In the refluxing method, the CC/Fe 2þ Fe 3þ precursors were facilely treated in the refluxing system for 90 min to obtain the CC/IO nanocomposites, designated as CCx/IO-R nanocomposites where x is TPP concentration. While, the Fe 2þ Fe 3þ precursors were hydrothermally treated in the NaOH solution at 100 C for 90 min in the closed glass bottles, resulting in the CCx/IO-HT nanocomposites. The synthesized CC/IO nanocomposites were filtered, washed with 100 mL distilled water and freeze-dried using freeze dryer at À100 C for 48 hr. Finally, the CC/ IO-R and CC/IO-HT nanocomposites were obtained as shown in Fig. 1.

Characterization and testing
X-ray powder diffraction (XRD) was used to characterize the crystalline phases of iron oxide formed in the CC/IO nanocomposites. The XRD analysis was scanned from 2h ¼ 20 to 70 at 0.02 /step and 1 sec/step using D8 ADVANCE X-ray diffractometer, Bruker.
Differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) were used to determine thermal properties of CC/IO nanocomposites.
The DSC analysis was carried out under the N 2 atmosphere with the flow rate of N 2 ¼ 20 mL/min using DSC 204 F1, NETZSCH. The heat flow was monitored by heating up samples from 25 to 400 C at a heating rate of 10 C/min. The thermal decomposition of CC/IO nanocomposites was analyzed in the temperature range of 50 À 800 C (heating rate of 10 C/min) using PYRIS 1 TGA, Perkin Elmer. The mass loss was detected under the O 2 atmospheres, while, the decomposition temperatures were detected under the N 2 atmospheres.
The morphological aspects of CC/IO nanocomposites were investigated by transmission electron microscopy (TEM, FEI, Tecnai G2 20 S-Twin) operating at an accelerating voltage of 200 kV (Bright-field mode). The CC/IO nanocomposites were dispersed in distilled water, then the droplet of the diluted suspension of CC/IO nanocomposites was deposited on a carbon-coated copper grid and dried overnight at room temperature. The average sizes of IO nanocrystals in the nanocomposites and their size distributions were determined by measurement of 50 randomly selected particles in one image from the TEM images at 500,000x magnification using ImageJ software.
Magnetic properties of CC/IO nanocomposites were investigated by an in-house developed vibrating sample magnetometer (VSM) operating at room temperature and 5 s per measurement sample using a À10,000 Oe to þ10,000 Oe magnetic field. The VSM was calibrated using a 3 mm-diameter Ni sphere (Lakeshore 730908, USA).

Cell viability assays
Cell viability of the CC/IO nanocomposites was analysized by using African green monkey fibroblast (Vero) cells for a preliminary evaluation of in vitro cytotoxicity. The CC/IO nanocomposites (0.4 g/sample) were sterilized at 121 C for 15 min, then the mixed solution of Dulbecco's Modified Eagle's Medium (DMEM) and 5% Fetal Bovine Serum (FBS) (10 mL) was added into each sample. The obtained mixtures were incubated at 37 C for 24 h. After incubation, the mixtures were filtered through a 0.45 mm filter membrane and diluted to the concentration of 1,000 mg/mL for the preparation of CC/IO stock solutions. The Vero cells were seeded into a 96-well plate by adding the cell suspension (5 Â 10 4 cells/mL, 100 mL) into each well, then the cell cultures were incubated at 37 C for 24 h in 5% CO 2 atmosphere. After incubation, the suspension was discarded and then the CC/IO stock solution (100 mL) was added to each well (3 wells/sample). The as-prepared sample was incubated at 37 C for 24 h. Then, 10 mL of 5 mg/mL MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution was added to each well. The sample was incubated at 37 C for 4 h with 5% CO 2 atmosphere. After that, the MTT solution was discarded and then the solution mixture of 100% dimethyl sulfoxide (DMSO) and 10% sodium dodecyl sulfate (SDS), with the volume ratio of 9:1-DMSO:SDS, was added to the well (150 mL/well). The sample was mechanically shaken for 5 min, then the optical absorbance was measured at 570 nm with a microplate reader (Thermo Fisher Scientific, Varioskan). The percentage of cell viability (% Cell viability) was calculated as follows: where A is the absorbance of the reference cells in the cell culture media, B is the absorbance of the sample solution, A and B must be corrected for the absorbance of a well, that contained only 100% DMSO:10% SDS.

Effect of TPP concentration on CC templates
The effect of TPP concentrations on crosslinked structures of CC templates were analyzed from DSC thermograms. Figure 2 shows the DSC thermograms of CC0.3, CC0.4, CC0.5 and CC0.6 templates obtained by using TPP concentrations of 0.3, 0.4, 0.5 and 0.6%w/ v, respectively. The melting temperatures (T m ) and of the CC templates were in the range of 158.3 À 178.5 C, corresponded to the crosslinked chitosan. The melting enthalpy (DH m ) values of the CC templates were considered to be related to the crosslinking density of chitosan. It can be seen that the T m values of CC0.3, CC0.4 and CC0.5 templates were almost similar, however, their DH m values significantly increased with the increasing of TPP concentration used in the preparation of CC templates from 0.3 to 0.5%w/v. The higher TPP concentration induced the higher dissociation of phosphoric and hydroxide ions, therefore, these ions speedily interacted with the -NH 3 þ groups of chitosan for crosslinking, resulting in the higher crosslinking densities. In the opposite direction, the T m and DH m values of CC0.6 template drastically decreased when the TPP concentration was increased to 0.6%w/v. The drastically decreased of T m and DH m values of CC0.6 template were because the rapid crosslinking immediately occurred when the droplets of chitosan solution were added to very high TPP concentration, resulting in the difference in crosslinking density between the outer surface layer and inner bulk of the chitosan. Therefore, the CC0.6 template formed as non-uniform crosslinked chitosan, in which it can be evidentally shown as the broad endothermic peak around 158 C. In addition, all samples showed the exothermic change at the temperature ranging from 226.4 to 232.1 C, in which it was related to the decomposition of free amine units in chitosan [19]. It was, therefore, that decomposition enthalpy (DH d ) corresponded to the number of free amine groups of chitosan. These DH d values fluctuated by increasing the TPP concentrations. The DH d value of CC0.6 templates was higher than those of CC0.3, CC0.4 and CC0.5 templates. This was considered to be because of higher quanitity of free amine groups of chitosan and lower crosslink densities were obtained in the CC0.6 template. These results suggested that the usage of high TPP concentration created a gradient crosslinking between the outer surface and inner bulk of the chitosan, resulting in the non-uniform crosslinked density and high quantity of free amine groups of CC templates as observed in the CC0.6 sample [20]. It was, therefore, that CC templates prepared by using TPP concentrations of 0.3, 0.4 and 0.5%w/v would be further used as the precursors for the preparation of CC/IO nanocomposites.  (2). However, the magnetite nanocrystals could be partially oxidized further in the presence of oxygen to maghemite nanocrystals as shown in equation (3) [21,22].

Characterization
The crystallite sizes (D XRD ) of magnetite-maghemite nanocrystals in the CC/IO nanocomposites were calculated from the (311) peak (2h ¼ 35.6 ) using the Scherrer equation with K ¼ 0.9. It was found that the D (311) values were in the range of 11 À 14 nm as summarized in Table 1. The magnetite-maghemite nanocrystals grew and embedded in the restricted free volume among the crosslinked structure of CC templates, leading to the finite particle sizes of magnetitemaghemite nanocrystals. It was, therefore, the CC templates positively impacted on the growth of magnetite-maghemite nanocrystals with the small particle sizes and the narrow size distribution [23].
The TEM images of CC0.3/IO-R, CC0.4/IO-R and CC0.5/IO-R nanocomposites are shown in Fig.  4. The microstructures of all nanocomposites were observed as the quadrilateral nanocrystals of magnetite-maghemite embedded in the CC templates. These nanocrystals were associated with the cubic inverse spinel structure of magnetite and the cubic tetragonal structure of maghemite, inferred from the XRD patterns. The average particle sizes of magnetite-maghemite nanocrystals measured from the TEM images were in the range of 10 À 14 nm (Table 1), in which they were in agreement with their average crystallite sizes from XRD data.
The particle size distributions of magnetitemaghemite nanocrystals in these nanocomposites analyzed from the TEM images are shown in Fig. 5. The average sizes of magnetite-maghemite nanocrystals in the CC0.3/IO-R, CC0.4/IO-R and CC0.5/IO-R nanocomposites were found to be primarily in the range of 8 À 12, 8 À 16 and 8 À 20 nm, respectively. It can be seen that the CC0.3/IO-R nanocomposite possessed the smallest magnetite-maghemite nanocrystals with the narrowest particle size distribution. When the CC0.4 and CC0.5 templates were used as the host materials, the particle size tended to grow larger and with a wider distribution, corresponding to the presence of non-uniform crosslinked structures and a larger free volume in the CC templates obtained when using higher TPP concentrations, in which they were further discussion in section 3.2.2.

3.2.2.
Thermal properties of CC/IO-R nanocomposites TGA curves of CC/IO nanocomposites synthesized by refluxing method and using the CC templates crosslinked with different TPP concentrations, i.e. 0.3%w/v (CC0.3/IO-R), 0.4%w/v (CC0.4/IO-R) and 0.5%w/v (CC0.5/IO-R), as shown in Fig. 6. The TGA analysis under N 2 atmosphere was used to determine the decomposition temperatures (T d ) of CC template as shown in Fig. 6 (N 2 ), while the analysis under O 2 atmosphere was used to quantify the relative crosslinked chitosan and iron oxide in the nanocomposites as shown in Fig. 6(O 2 ). From Fig. 6 (N 2 ), all nanocomposites showed similar thermal events, that is two major stages of degradations at T d1 and T d2 related to the degradation of CC templates. The T d1 value of the first degradation stage in all nanocomposites corresponded to the breaking of glucoside linkage [24] was observed in the range of 250.6 À 256.2 Csee Table 2. The second degradation stage was observed at the T d2 value of 409.2 À 431.6 C. This degradation temperature related to the interfacial interaction between the embedded magnetite-maghemite  nanocrystals and the CC templates. It can be seen that the T d1 values of all nanocomposites were quite similar, while the T d2 values tended to significantly decrease when the nanocomposites were prepared from the CC templates derived with higher TPP concentrations. The higher the TPP concentration, the higher the free volume and non-uniformity in the CC templates, resulting in the formation of larger magnetite-maghemite nanocrystals. Because the surface area of the magnetite-maghemite nanocrystals decreased as they grew larger, the interfacial interaction between the nanocrystals and the CC templates decreased, and thus the T d2 values decreased.
The fractions of CC templates and iron oxide in the CC/IO nanocomposites were determined using weight losses and remaining in the TGA thermograms under O 2 atmosphere. All nanocomposites were composed of water and organic solvent, evaporating at around 53.0 À 55.1 C with the weight loss in the range of 10 À 15%. The large proportion calculated from the sum of weight losses at T d1 and T d2 was attributed to the chitosan quantity of 53 À 57%. Residuals ranged from 28 to 35%, representing the relative amount of iron oxide nanocrystals and a trace of phosphorus from the TPP crosslinker in all nanocomposites.
The DSC thermograms of CC0.3/IO-R, CC0.4/ IO-R and CC0.5/IO-R nanocomposites are shown in Fig. 7. The melting points (T m ) of CC templates in these nanocomposites were in the range of 183.1 À 187.7 C as summarized in Table 2. These T m values were higher than the T m of the as-   prepared CC templates shown in section 3.1, insisting the occurrance of interfacial interaction between magnetite-maghemite nanocrystals and CC templates. The CC0.3/IO-R had a higher DH m value than the CC0.4/IO-R and CC0.5/IO-R nanocomposites, which were attributed to the effect of interfacial interaction of nanocrystals and CC template. As previously mentioned, the CC0.3/IO-R nanocomposite had smaller magnetite-maghemite nanocrystals, which tended to provide a greater interfacial interaction with the CC0.3 template, resulting in a higher DH m value than the others. Furthermore, a large exothermic change was observed in all nanocomposites at temperatures ranging from 249.4 to 251.2 C, which occurred at a higher temperature than the precursor CC templates and corresponded to T d1 values from the TGA data. These results were achieved because the lone pair electrons of free amine groups could interact with the Fe 2þ Fe 3þ precursors and then react with the basic solution, forming the magnetite-maghemite nanocrystals. As a result, nanocomposites containing small magnetitemaghemite nanocrystals have higher DH d values than those containing large nanocrystals.

Magnetic properties
The magnetic properties of CC/IO nanocomposites were measured as a function of the magnetic field at room temperature. The magnetization curves of CC0.3/IO-R, CC0.4/IO-R and CC0.5/IO-R  nanocomposites are shown in Fig. 8 and summarized in Table 2. The absence of hysteresis loops was found in all nanocomposites because their coercivity (H c ) and magnetic remanence values (M r ) were close to zero, which is superparamagnetism. The CC0.3/IO-R nanocomposite showed the lowest value of maximum magnetization (M max ) of $8.6 emu/g. The M max values significantly increased to $15 emu/ g in the CC0.4/IO-R and CC0.5/IO-R nanocomposites. The lowest M max value in the CC0.3/IO-R nanocomposite can be explained by the smallest size of magnetite-maghemite nanocrystals with the narrowest particle size distribution as discussed in section 3.2.1. The magnetic moment of magnetitemaghemite nanocrystals dropped as their size decreased due to volume magnetic anisotropy. Furthermore, the increased surface layer of magnetite-maghemite nanocrystals could interact with the CC templates, resulting in magnetic disorder and magnetic frustration [25,26]. Since the particle sizes of the synthesized magnetite-maghemite nanocrystals in all nanocomposites were less than the superparamagnetic critical radius, they were considered to be single magnetic domain structures. It can be concluded that the proposed synthesis method in this study was able to control the formation of magnetite-maghemite nanocrystals with appropriate particle size and narrow size distribution, superparamagnetic properties and sufficient M max values for medical applications. Hence, the CC/IO-R nanocomposites are promising candidates for targeted delivery for alternative medical treatments due to their superparamagnetism behavior. When an external magnetic field was applied during treatment, the CC/IO-R nanocomposites were thought to respond very quickly to the magnetic field, all magnetisms of the nanocomposites immediately disappearing with the absence of the magnetic field when the treatments were completed (Demagnetization).

Effect of heat treatment method
3.3.1. Crystal structure and morphology In Fig. 9(a), the crystal structure of a CC/IO nanocomposite synthesized from the CC0.4 template and treated by hydrothermal method (CC0.4/IO-HT) is shown in comparison to that treated by refluxing (CC0.4/IO-R). Similar to the CC0.4/IO-R nanocomposite, the XRD pattern of the CC0.4/IO-HT nanocomposite was dominated by typical peaks of magnetite and maghemite phases. The D (311) crystallite size of CC0.4/IO-HT nanocomposite was $11.7 nm, which was comparable to that of CC0.4/ IO-R nanocomposite as summarized in Table 1. The magnetite-maghemite nanocrystals were observed of quadrilateral shape embedded in the CC template as shown in Fig. 9(b). The average particle size of magnetite-maghemite nanocrystals in the CC0.4/IO-HT nanocomposite measured from the TEM images was $10.7 nm, which is consistent with the D (311) value. As demonstrated in Fig. 9(c), the particle size distribution of magnetite-maghemite nanocrystals in the CC0.4/IO-HT nanocomposite was 4-20 nm, which was wider and smaller than that of the CC0.4/IO-R nanocomposite. We assumed that under the pressurized hydrothermal system, the magnetite-maghemite nanocrystals could not grow freely, limiting the particle size of the nanocrystals to a specific range. On the other hand, the refluxing system had no pressure, allowing the magnetite-maghemite nanocrystals to grow to larger particle size.

Thermal properties
The TGA thermogram of CC0.4/IO-HT nanocomposite is shown and compared with that of CC0.4/ IO-R in Fig. 10  which was lower than 35% in the CC0.4/IO-R nanocomposite. These findings were attributed to the smaller size of the nanocrystals, which promoted higher interfacial interactions between nanocrystals and CC templates, resulting in higher T d2 values. However, the proportion of iron oxide nanocrystals in the CC0.4/IO-HT nanocomposite was lower, which could be due to the turbulent movement of materials in the system caused by the pressurized hydrothermal reaction system. Therefore, the iron oxide precursors might partially migrate from the CC0.4 template and hydrothermally reacted in the solution media. Figure 11 compares the DSC thermograms of CC0.4/IO-HT and CC0.4/IO-R. The T m and DH m values of crossinked chitosan template in the CC0.4/IO-HT were 185.4 C and 104.1 J/g, respectively, as summarized in Table 2. Because these nanocomposites were prepared from the CC0.4 template using the same TPP concentration, the T m and DH m values of CC0.4/IO-HT and CC0.4/IO-R would be comparable. In addition, the large exothermic change in both nanocomposites, caused by the decomposition of CC templates, corresponded to the T d1 values of the TGA data. The T d and DH d values in the CC0.4/IO-HT nanocomposite were $258 C and 118.3 J/g, respectively, in which they were higher than those of the CC0.4/IO-R nanocomposite because of the smaller magnetite-maghemite nanocrystals in agreement with the TGA data.    Figure 12 shows the magnetization curves of CC0.4/ IO-HT in comparison with the CC0.4/IO-R. The M max value of CC0.4/IO-HT nanocomposite was $12.7 emu/g, which was slightly lower than that of CC0.4/IO-R nanocomposite as shown in Table 2. The M max value increased with particle size and quantity of magnetite-maghemite nanocrystals in the nanocomposites. However, the shape of the hysteresis curves of both nanocomposites insisted on superparamagnetic behavior. Because the particle sizes of magnetite-maghemite nanocrystals in the CC0.4/IO-HT nanocomposite were less than the superparamagnetic critical radius of both magnetite and maghemite, they were assumed to be a single magnetic domain structure.

Cell viability of CC/IO nanocomposites
To assess the toxic behavior of the CC/IO nanocomposites to mammalian continuous cells, a preliminary study on the viability of African green monkey fibroblast (Vero) cells was performed. As shown in Table 2, all nanocomposites have cell viability in the range of 80-89%, indicating that they have a high cell viability. All nanocomposites exhibit non toxic behavior. These findings supported the safety of CC/IO nanocomposites in mammalian continuous cells. Some researchers [27,28] reported that mild toxicity of magnetite and maghemite nanoparticles was found at high concentrations because they could induce reactive oxygen species (ROS) in human body tissues, leading to tissue damage. The generated CC/IO nanocomposites, on the other hand, contained magnetite-maghemite nanocrystals embedded in high biocompatible CC templates, displaying negligible cytotoxicity. The high biocompatible CC/IO nanocomposites therefore have the possiblity to be applied in medical applications.

Conclusions
The chitosan templates crosslinked with TPP crosslinker were utilized as the host materials for the nucleation and growth of iron oxide nanocrystals in hydrothermal and refluxing systems, resulting in the CC/IO nanocomposites. In the resultant nanocomposites, the hydrothermal and refluxing forms of iron oxide nanocrystals were detected as quadrilateral shaped magnetite-maghemite nanocrystals embedded in a CC matrix. Increasing in TPP concentration resulted in more non-uniform crosslinked CC templates with higher free volume, leading to the growth of larger magnetite-maghemite nanocrystals in the CC/IO nanocomposites. The average crystallite sizes of magnetite-maghemite nanocrystals in the CC/IO nanocomposites were in the range of 10 À 14 nm, according to the XRD and TEM data. Furthermore, magnetite-maghemite nanocrystals hydrothermally grown in the CC/IO nanocomposite had a smaller crystallite size but a wider size distribution than those obtained by refluxing. The resulting CC/IO nanocomposites exhibited superparamagnetic behavior with M max in the range of 8.6 À 15.2 emu/g. Moreover, all nanocomposites demonstrated high cell viability, ranging from 80 to 89%, indicating that they were safe for mammals. The synthesis methodologies proposed in this study were facile, soft solution eco-friendly approaches that are simple to scale up. The crosslinked chitosan templates could reduce agglomeration and oxidation of magnetite-maghemite nanocrystals, in addition, they could improve the narrow size distribution and distribution homogeneity. The CC/IO nanocomposites would demonstrate the synergistic effect of iron oxide nanocrystals and crosslinked CC matrix. The advantages of superparamagnetic nanoparticles were specificity to target organ and demagnetism after medical treatments, while the high proportion of chitosan improved biocompatibility in the CC/IO nanocomposites. Furthermore, the chitosan matrix could be modified for alternative medical treatments such as drug delivery, bioseparation, cell labeling, and hyperthermia. Because of their hybrid functions, the synthesized CC/IO nanocomposites are a promising candidate for improving medical therapies.
Royal Golden Jubilee (RGJ) Ph.D. Programme; materials chacterization had partially been supported by Scientific Instruments Center, School of Science, King Mongkut's Institute of Technology Ladkrabang.

Disclosure statement
No potential conflict of interest was reported by the author(s). Her research group has focused on the chemical synthesis and characterization of biocompatible materials and nanostructured materials, and optimisation of these materials for medical, environmental, agricultural and textile applications. She has published over 50 journal and conference papers. In 2012, she has been awarded the L'Or eal-UNESCO For Women in Science Fellowships from her research finding on "Nanotechnology for Ecofriendly Development in Textile Finishing". Her work on chitosan-clay nanocomposites has been widely recognized and cited over 220 times, and it has been expanded for a wide range of applications. She has international collaboration with some research groups in France, Austria and USA. Furthermore, she has also been applying her expertise to industrial partners.