2.1. Structural Analysis

FTIR is an effective tool to characterize the secondary structure and functional group of silk protein materials [16,17] (Figure 2). For Fe₃O₄-SF samples, there is a sharp peak at 1622 cm⁻¹, suggesting the β-sheet crystal structures are dominant. However, the peak shoulder at 1646 cm⁻¹ (random coils structures) increased with Fe₃O₄ content, indicating that relative fraction of β-sheet crystals decreased. All BaM-SF samples also showed a sharp peak at 1620 cm⁻¹, indicating a predominant β-sheet secondary structure due to the wet-pressing method. The peak at 1650 cm⁻¹ slightly increased when more nanoparticles were present, suggesting that the BaM particles in the silk matrix can also slightly enhance the formation of α-helix or random coils structures. However, this structural change is not as significant for the Fe₃O₄-SF samples. Compared to the FTIR patterns of BaM-SF and Fe₃O₄-SF, Co-SF samples showed much sharper peaks at 1622 cm⁻¹, suggesting it has the highest β-sheet crystal content among the three types of magnetic inclusions. In addition, the shoulder at 1651 cm⁻¹ decreased with the increase in Co particle content. In the amide II region, all three types of composite films showed peaks around 1515 cm⁻¹, which suggests Tyr side chains structure.

Figure 2. FTIR spectra of (a) Fe₃O₄-silk fibroin (SF), (c) BaM-SF and (e) Co-SF composite films. Secondary structure contents of (b) Fe₃O₄-SF, (d) BaM-SF and (f) Co-SF calculated from a Fourier self-deconvolution curve fitting method.

A quantitative analysis of the secondary structure contents was performed with a Fourier self-deconvolution (FSD) curve fitting method (Figure 2b,d,f) [16]. It shows that β-sheet content of SF sample is around 39% composed of mainly inter-molecular β-sheets [10]. The β-sheet content of the Fe₃O₄-SF samples decreases with Fe₃O₄ content, reaching 33% at 30 wt%, while the random coils content increased slightly by 29–32%. Similar behavior was found in BaM-SF samples, where the β-sheet content decreased with an increase in BaM weight fraction, while the random coils content nominally increased. In contrast, for the Co-SF samples, the secondary structure remained almost unchanged with a slight increase in β-sheet content.

2.2. Morphology Analysis

Surface and cross-section morphology of SF and magnetic silk composite films are shown in Figure 3. Lamellar patterns evenly spread out across the cross section of SF and 20% BaM-SF films (Figure 3a,g). Cross section of 20% Fe₃O₄ (Figure 3d) showed a rougher morphology with wrinkles and densely distributed particles. Cross section of 20% Co-SF (Figure 3j) showed evenly distributed holes with connected wrinkles. When comparing the surface samples to the cross-sectional, the surface seems to be much more homogenous, containing fewer aggregates and wrinkles. SF film shows a smooth and uniform surface (Figure 3b). A 20% BaM-SF sample (Figure 3h) showed a relatively rough surface, and BaM particles distribute homogenously instead of forming big aggregates. Compared to the surface morphology of 20% BaM-SF film, 20% Fe₃O₄ (Figure 3e) and 20% Co-SF (Figure 3k) films showed smooth and uniform surface morphology with shallow pits. EDS spectra and analyses of the SF, 20% Fe₃O₄-SF, 20% BaM-SF and 20% Co-SF films are shown in Figure 3c,f,i,l, respectively. As mentioned above, most CaCl₂ residue has been washed out of the pure SF film after the water annealing process (Figure 3c). The EDS spectra of various elements in the composites also showed evidence of Fe (6.403 keV and 0.705 keV, Figure 3f), Ba (4.465 keV and 0.779 keV, Figure 3i) and Co (6.929 keV and 0.776 keV, Figure 3l) elements in their respective composite films.

Figure 3. (a,d,g,j) The cross section of SF film, 20% Fe₃O₄-SF, 20% BaM-SF and 20% Co-SF composite, respectively; (b,e,h,k) are the surface morphology of SF film, 20% Fe₃O₄-SF, 20% BaM-SF and 20% Co-SF composite, respectively; (c,f,i,l) display the EDS spectra for the SF, 20% Fe₃O₄-SF, 20% BaM-SF and 20% Co-SF films, respectively.

2.3. Thermal Analysis

Thermal stability of the silk magnetic composite films was characterized by TGA (Figure 4, Table 1). All three types of magnetic particles are thermally stable with no degradation for all of them up to 800 °C. All Fe₃O₄-SF composite films showed a small degradation in the range of 209~226 °C (T_d1), a major degradation between 296 and 303 °C (T_d2) and a third degradation around 650 °C (T_d3). The residual weight of Fe₃O₄-SF samples at 800 °C was between 26.6 and 48.2%, which generally increased with Fe₃O₄ content. All BaM-SF composite films showed a small degradation between 238 and 251 °C (T_d1) and a major degradation between 301 and 309 °C (T_d2). When the BaM concentration was 10% or above, a third degradation (T_d3) was found around 700 °C (Figure 4c,d). The residual weight of BaM-SF samples at 800 °C was between 27.1 and 46.1%, and the residual weight increased with BaM content. All Co-SF composite films showed a small degradation at 214~233 °C (T_d1) and a major degradation at 300~307 °C (T_d2). The residual weight of Co-SF samples at 800 °C is between 27.1 and 55.1%, and increased with Co content. However, no third degradation peak (T_d3) was observed around 600~700 °C for any of Co-SF samples. The first small degradation is mainly from the unstable part of silk proteins [32,33,34]. BaM-SF samples showed a higher T_d1 than that of the other two types of composites, suggesting that BaM particles were able to best protect silk materials. The third degradation for Fe₃O₄-SF and BaM-SF samples is probably from a stable phase that formed when the Fe combined with the silk protein.

Figure 4. Thermogravimetric curves of (a) Fe₃O₄-SF, (c) BaM-SF and (e) Co-SF composite films. The 1st derivative TG (DTG) curves of (b) Fe₃O₄-SF, (d) BaM-SF and (f) Co-SF composite films.

Heat flow and reversing heat capacity of magnetic silk composite films measured from DSC are shown in Figure 5. Heat flow analysis shows that all three type composite films have a major degradation about 260 °C, which is from the decomposition of silk proteins. The amorphous part of the polymer has a greater mobility with increasing temperature, which is defined as the glass transition. The glass transition is gradual and reversible, and the heat capacity of the polymer changes dramatically during this transition [35]. All Fe₃O₄-SF samples showed a similar glass transition temperature around 172 °C. The glass transition temperature of BaM-SF composite films increased from 172 °C for SF, to 192 °C for 30% BaM-SF, suggesting that the mobility of the amorphous structure in BaM-SF composites can be tuned by the BaM particles. All Co-SF composite films showed a glass transition temperature around 180 °C.

Figure 5. Heat flow of (a) Fe₃O₄-SF, (c) BaM-SF and (e) Co-SF composite films. Reversing heat capacity of (b) Fe₃O₄-SF, (d) BaM-SF and (f) Co-SF.

2.4. Magnetization Analysis

Magnetization of the three types of composite films showed typical ferromagnetic behavior with coercive field of about 128 Oe (Fe₃O₄), 3660 Oe (BaM) and 165 Oe (Co), respectively (Figure 6a–c). The dependence of the magnetization on weight fraction of magnetic particles are shown in Figure 6d. The saturation magnetization M_s of Fe₃O₄, BaM and Co nanoparticles is about 61 emu/g, 68 emu/g and 156 emu/g, respectively. Since silk protein matrix has magnetic susceptibility near zero, it can be assumed that the moment of the composite films depends only on the net weight of the magnetic particles. Therefore, one would anticipate that the M_s value of the composite should be about xM_s, where x is the weight fraction of magnetic particles. All samples display a linear dependence of M_s on weight fraction; however, note that neither line intercepts the origin. Presumably, all three types of magnetic particles partially dissolved in the formic acid solution, saturating the solution at about 3.7 wt%, 3.4 wt% and 1.1 wt% (horizontal intercept of the graph) for Fe₃O₄, BaM and Co particles, respectively. These results are consistent with the study of the secondary structure of SF matrix. In any case, at sufficient loading, of three types of composite films maintain a sizable magnetization for potential use in MRI imaging or targeted drug delivery.

Figure 6. Magnetization and hysteresis loops of (a) Fe₃O₄-SF, (b) BaM-SF and (c) Co-SF composite films at room temperature. (d) Saturation moment of Fe₃O₄-SF, BaM-SF and Co-SF composite films as a function of the magnetic particle content.

2.5. Self-Assembly Mechanism

With the experimental evidence and analysis provided above, we can confirm that Fe₃O₄ and BaM particles can slightly prevent the β-sheet crystal formation, suggesting that the Fe₃O₄-SF and BaM-SF composite films have more noncrystalline structures (Figure 7). This is probably caused by the strong coordination bonding between Fe³⁺ ions and carboxylate ions on silk fibroin chains [36,37,38] due to the dissolved Fe found from the magnetization studies. Most β-sheet crystals usually formed during the water annealing and wet pressing process. However, when the Fe₃O₄ and BaM particles were present, the strong coordination bonding limited the mobility of silk fibroin chain and further prohibited the β-sheet crystal formation (Figure 7). The unique high temperature degradation (T_d3) of Fe₃O₄-SF and BaM-SF samples found in TG analysis may be another indicator of this stable phase, which is formed when Fe is combined with silk protein. On the other hand, the amount of dissolved Co is minimal as compared to that of the Fe, so the addition of Co nanoparticles has less effect on the secondary structure. Since a large amount of Co particles will occupy more space in the Co-SF film, the silk protein chains are able to be closer to each other, resulting in a slight increase in β-sheet crystals (Figure 7).

Figure 7. The effect of iron-based (Fe₃O₄ or BaM) and cobalt-based (Co) magnetic particles on the secondary structures of silk fibroin material.