Surface Morphology at the Microscopic Scale, Swelling/Deswelling, and the Magnetic Properties of PNIPAM/CMC and PNIPAM/CMC/Fe3O4 Hydrogels

Poly(N-isopropylacrylamide) (PNIPAM) hydrogels containing carboxymethylcellulose (CMC) and CMC/Fe3O4 nanoparticles were prepared. Free-radical polymerization with BIS as cross-linker was used to synthesize the hydrogels. The morphology at the microscopic scale of these materials was investigated using field emission scanning electron microscopy (FESEM). The images show that CMC in the PNIPAM hydrogels induces the formation of a honeycomb structure. This surface morphology was not observed for pure PNIPAM hydrogels prepared under similar conditions. The equilibrium swelling degree of the PNIPAM/CMC hydrogels (5200%) is much larger than that of the pure PNIPAM hydrogels (2500%). The water retention of PNIPAM/CMC hydrogels above the volume phase transition temperature is strongly reduced compared to that of pure PNIPAM hydrogel. Both PNIPAM/Fe3O4 and PNIPAM/CMC/Fe3O4 hydrogels exhibit a superparamagnetic behavior, but the blocking temperature (104 K) of the former is higher than that of the latter (83 K).


Introduction
Poly(N-isopropylacrylamide) (PNIPAM) hydrogels are the most investigated thermoresponsive hydrogels [1]. Because of their capability to change volume reversibly at a transition temperature, hydrogels of PNIPAM or co-polymer of PNIPAM have applications in several fields ranging from drug delivery to environmental remediation [2,3]. The introduction of magnetic nanoparticles (NPs) into PNIPAM hydrogels allows to heat remotely the hydrogels by applying alternating magnetic fields of suitable frequencies [2]. In most of these applications the rate of the deswelling/swelling processes across the volume phase transition temperature (VPTT) is important [4]. Unfortunately, the response of macroscopic PNIPAM hydrogels to temperature changes is rather slow [5][6][7]. The collapse of the swollen hydrogel (with expulsion of water molecules out of the polymer matrix) has a complex dynamics involving diffusion of polymer segments and conformation changes of the polymer chains [8]. A contribution to slow the thermal response of PNIPAM hydrogels comes from the so-called "skin effect", which is the formation of a dense, vitrous layer which prevents the exchange of water molecules between the hydrogel and the surrounding [9]. Several methods have been introduced to obtain PNIPAM hydrogels featuring a faster response to temperature variations. Some of these methods use the copolymerization of NIPAM with hydrophilic molecules to eliminate the skin effect and form hydrophilic channels for the diffusion of water molecules [9,10]. Other methods are based on promoting the formation of interconnected pores or use PNIPAM microgels embedded in Gels 2016, 2, 30 2 of 8 a macrogel [11,12]. It has been shown that the interpenetrating polymer network (IPN) or semi-IPN hydrogels of NIPAM with a hydrophilic polysaccharide such as sodium alginate have a faster response to temperature variation [13]. In the present work, we prepared PNIPAM hydrogels containing carboxymethylcellulose (CMC) and magnetite (Fe 3 O 4 ) NPs coated with CMC. CMC, a hydrophilic, biocompatible derivative of cellulose, appears to be a suitable candidate to prepare PNIPAM hydrogels exhibiting a faster thermal response. The choice of CMC was also motivated by its ability to stabilize aqueous dispersions of magnetite NPs [14]. One of the aims of this study was to investigate the effect of CMC and CMC/Fe 3 O 4 NPs on the morphology at micrometric scale and on the swelling/deswelling properties of PNIPAM hydrogels. The other aim was to study the effect of coating magnetite NPs with CMC on the magnetic properties of PNIPAM hydrogels containing Fe 3 O 4 NPs. The morphology of the PNIPAM hydrogels was investigated by means of field emission scanning electron microscopy (FESEM). The magnetic properties of PNIPAM/Fe 3 O 4 and PNIPAM/CMC/Fe 3 O 4 hydrogels were studied by measuring magnetization versus magnetic field curves at various temperatures, and zero field cooling (ZFC) and field cooling (FC) curves.

Results and Discussion
The attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) spectrum measured for the PNIPAM hydrogel containing 20 wt % of CMC (PNIPAM/CMC20) can be fitted by summing the spectra of PNIPAM dry hydrogel and CMC. The best fit is obtained by optimizing the weight factors of the intensities of the spectra of PNIPAM hydrogel and CMC ( Figure 1). This result suggests that there are no large interactions between the PNIPAM network and the CMC chains which, otherwise, would produce variations of the peak positions and widths in the spectrum of PNIPAM/CMC hydrogel with respect to those of the pure components. Because of partial overlapping of the main peaks of PNIPAM and CMC and the relatively small contribution of CMC to the intensity of the PNIPAM/CMC20 spectrum, Fourier transform infrared spectroscopy (FTIR) is not able to reveal subtle changes in the spectrum, resulting from possible interactions between PNIPAM and CMC or cross-linking of CMC chains. Hence, attenuated total reflectance (ATR)-FTIR results are consistent with the formation of a semi-IPN, but the cross-linking of CMC chains, leading to a full IPN [15], cannot be ruled out. embedded in a macrogel [11,12]. It has been shown that the interpenetrating polymer network (IPN) or semi-IPN hydrogels of NIPAM with a hydrophilic polysaccharide such as sodium alginate have a faster response to temperature variation [13]. In the present work, we prepared PNIPAM hydrogels containing carboxymethylcellulose (CMC) and magnetite (Fe3O4) NPs coated with CMC. CMC, a hydrophilic, biocompatible derivative of cellulose, appears to be a suitable candidate to prepare PNIPAM hydrogels exhibiting a faster thermal response. The choice of CMC was also motivated by its ability to stabilize aqueous dispersions of magnetite NPs [14]. One of the aims of this study was to investigate the effect of CMC and CMC/Fe3O4 NPs on the morphology at micrometric scale and on the swelling/deswelling properties of PNIPAM hydrogels. The other aim was to study the effect of coating magnetite NPs with CMC on the magnetic properties of PNIPAM hydrogels containing Fe3O4 NPs. The morphology of the PNIPAM hydrogels was investigated by means of field emission scanning electron microscopy (FESEM). The magnetic properties of PNIPAM/Fe3O4 and PNIPAM/CMC/Fe3O4 hydrogels were studied by measuring magnetization versus magnetic field curves at various temperatures, and zero field cooling (ZFC) and field cooling (FC) curves.

Results and Discussion
The attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) spectrum measured for the PNIPAM hydrogel containing 20 wt % of CMC (PNIPAM/CMC20) can be fitted by summing the spectra of PNIPAM dry hydrogel and CMC. The best fit is obtained by optimizing the weight factors of the intensities of the spectra of PNIPAM hydrogel and CMC ( Figure 1). This result suggests that there are no large interactions between the PNIPAM network and the CMC chains which, otherwise, would produce variations of the peak positions and widths in the spectrum of PNIPAM/CMC hydrogel with respect to those of the pure components. Because of partial overlapping of the main peaks of PNIPAM and CMC and the relatively small contribution of CMC to the intensity of the PNIPAM/CMC20 spectrum, Fourier transform infrared spectroscopy (FTIR) is not able to reveal subtle changes in the spectrum, resulting from possible interactions between PNIPAM and CMC or cross-linking of CMC chains. Hence, attenuated total reflectance (ATR)-FTIR results are consistent with the formation of a semi-IPN, but the cross-linking of CMC chains, leading to a full IPN [15], cannot be ruled out.    On the other hand, FESEM images of PNIPAM/CMC are characterized by a honeycomb-like structure (Figures 3a,b), with features smaller than those observed in the images of pure PNIPAM. Images of internal cross sections of the PNIPAM/CMC hydrogels show that a network of interconnected pores exists inside the hydrogels. The observed honeycomb-like structure is probably due to the emergence of micropores at the surface of the PNIPAM/CMC hydrogels as suggested by the images measured for the inner cross-sections of the hydrogel (Figure 3c). The surface morphology of the PNIPAM/CMC hydrogels does not change significantly in the explored range of CMC concentration (that is from 5 wt % to 20 wt %) as far as the presence of pores is concerned. However, the size of the micropores tends to be smaller for the lower CMC concentration (Figure 3d).
The surface morphology of the hydrogels is not significantly affected by the presence of CMC/Fe3O4 NPs as shown by the FESEM images of the PNIPAM/CMC hydrogel containing 4.3 wt % magnetite (PNIPAM/CMC5/Fe3O4) samples (Figure 4a,b).
The attainment of the equilibrium swelling degree (SDeq) was checked by measuring swelling kinetics starting from lyophilized samples of the hydrogels ( Figure 5). These data indicate that the swelling process is slightly faster for the PNIPAM/CMC20 hydrogel than for the pure PNIPAM hydrogel. The SDeq measured at 295 K of the PNIPAM/CMC20 sample (5200 ± 200 %) is much larger than that of the PNIPAM hydrogel (2500 ± 200 %). On the contrary, the addition of 5 wt % of CMC to the PNIPAM hydrogels slightly reduces its SDeq (1900 ± 100 %). Similar SDeq values (ca. 2000%) were measured for the PNIPAM/CMC5/Fe3O4 hydrogel. These results can be explained considering that the interactions of the hydrophilic groups of PNIPAM with those of CMC are more favorable than with water at low CMC concentration; thus, a decrease of the swelling degree is observed [15]. On the other hand, FESEM images of PNIPAM/CMC are characterized by a honeycomb-like structure (Figure 3a,b), with features smaller than those observed in the images of pure PNIPAM. Images of internal cross sections of the PNIPAM/CMC hydrogels show that a network of interconnected pores exists inside the hydrogels. The observed honeycomb-like structure is probably due to the emergence of micropores at the surface of the PNIPAM/CMC hydrogels as suggested by the images measured for the inner cross-sections of the hydrogel (Figure 3c). The surface morphology of the PNIPAM/CMC hydrogels does not change significantly in the explored range of CMC concentration (that is from 5 wt % to 20 wt %) as far as the presence of pores is concerned. However, the size of the micropores tends to be smaller for the lower CMC concentration (Figure 3d).
The surface morphology of the hydrogels is not significantly affected by the presence of CMC/Fe 3 O 4 NPs as shown by the FESEM images of the PNIPAM/CMC hydrogel containing 4.3 wt % magnetite (PNIPAM/CMC5/Fe 3 O 4 ) samples (Figure 4a,b).
The attainment of the equilibrium swelling degree (SD eq ) was checked by measuring swelling kinetics starting from lyophilized samples of the hydrogels ( Figure 5). These data indicate that the swelling process is slightly faster for the PNIPAM/CMC20 hydrogel than for the pure PNIPAM hydrogel. The SD eq measured at 295 K of the PNIPAM/CMC20 sample (5200% ± 200%) is much larger than that of the PNIPAM hydrogel (2500% ± 200%). On the contrary, the addition of 5 wt % of CMC to the PNIPAM hydrogels slightly reduces its SD eq (1900% ± 100%). Similar SD eq values (ca. 2000%) were measured for the PNIPAM/CMC5/Fe 3 O 4 hydrogel. These results can be explained considering that the interactions of the hydrophilic groups of PNIPAM with those of CMC are more favorable than with water at low CMC concentration; thus, a decrease of the swelling degree is observed [15].
The VPTT estimated from the temperature at which the hydrogel samples start to be opalescent is not affected (within an uncertainty of ±2 K) by CMC. This result was confirmed by the swelling degree measurements as a function of temperature of PNIPAM/CMC20 hydrogels which show a drastic change from 303 to 307 K. Moreover, the presence of CMC/Fe 3 O 4 , at the concentrations used in this work, does not influence, within the accuracy of the measurements, the VPTT of the PNIPAM hydrogels.     The VPTT estimated from the temperature at which the hydrogel samples start to be opalescent is not affected (within an uncertainty of ±2 K) by CMC. This result was confirmed by the swelling degree measurements as a function of temperature of PNIPAM/CMC20 hydrogels which show a drastic change from 303 to 307 K. Moreover, the presence of CMC/Fe3O4, at the concentrations used in this work, does not influence, within the accuracy of the measurements, the VPTT of the PNIPAM hydrogels.
The deswelling kinetics of PNIPAM/CMC20 and of pure PNIPAM hydrogel samples, previously swollen in water at 295 K, were measured by monitoring the water retention % as a function of the immersion time in water at 313 K. After 30 min, the water retention of the PNIPAM/CMC20 hydrogel is reduced to about 35%, whereas that of pure the PNIPAM hydrogel reaches a steady-state value of 70% ( Figure 6). Similar results were reported for PNIPAM/sodium alginate semi-IPNs [13]. The higher water retention of the pure PNIPAM hydrogel is due to the formation of a compact, vitreous layer at the surface of the samples above the VPTT, which hampers the release of water. The addition of CMC prevents the formation of this impermeable layer. This interpretation is supported by the different macroscopic morphology of the two kinds of hydrogels in water at 40 °C. The bubbles (full of water) that form at the surface of pure PNIPAM hydrogels in water above the VPTT [7] are not observed for the PNIPAM/CMC samples.  The deswelling kinetics of PNIPAM/CMC20 and of pure PNIPAM hydrogel samples, previously swollen in water at 295 K, were measured by monitoring the water retention % as a function of the immersion time in water at 313 K. After 30 min, the water retention of the PNIPAM/CMC20 hydrogel is reduced to about 35%, whereas that of pure the PNIPAM hydrogel reaches a steady-state value of 70% ( Figure 6). Similar results were reported for PNIPAM/sodium alginate semi-IPNs [13]. The higher water retention of the pure PNIPAM hydrogel is due to the formation of a compact, vitreous layer at the surface of the samples above the VPTT, which hampers the release of water. The addition of CMC prevents the formation of this impermeable layer. This interpretation is supported by the different macroscopic morphology of the two kinds of hydrogels in water at 40 • C. The bubbles (full of water) that form at the surface of pure PNIPAM hydrogels in water above the VPTT [7] are not observed for the PNIPAM/CMC samples. The VPTT estimated from the temperature at which the hydrogel samples start to be opalescent is not affected (within an uncertainty of ±2 K) by CMC. This result was confirmed by the swelling degree measurements as a function of temperature of PNIPAM/CMC20 hydrogels which show a drastic change from 303 to 307 K. Moreover, the presence of CMC/Fe3O4, at the concentrations used in this work, does not influence, within the accuracy of the measurements, the VPTT of the PNIPAM hydrogels.
The deswelling kinetics of PNIPAM/CMC20 and of pure PNIPAM hydrogel samples, previously swollen in water at 295 K, were measured by monitoring the water retention % as a function of the immersion time in water at 313 K. After 30 min, the water retention of the PNIPAM/CMC20 hydrogel is reduced to about 35%, whereas that of pure the PNIPAM hydrogel reaches a steady-state value of 70% ( Figure 6). Similar results were reported for PNIPAM/sodium alginate semi-IPNs [13]. The higher water retention of the pure PNIPAM hydrogel is due to the formation of a compact, vitreous layer at the surface of the samples above the VPTT, which hampers the release of water. The addition of CMC prevents the formation of this impermeable layer. This interpretation is supported by the different macroscopic morphology of the two kinds of hydrogels in water at 40 °C. The bubbles (full of water) that form at the surface of pure PNIPAM hydrogels in water above the VPTT [7] are not observed for the PNIPAM/CMC samples.