Triggered by light and magnetism: smart foam PLLA/HAP/Fe₃O₄ scaffolds for heat-controlled biomedical applications

Abstract

Ternary composite foam materials containing poly-L-lactic acid (PLLA), calcium hydroxyapatite (HAP) (20 nm), and morphologically controlled Fe₃O₄ nanoparticles (80 nm) were fabricated using the thermally induced phase separation (TIPS) technique over a broad concentration range of the magnetic component (1–30 wt%). The foam scaffolds were highly porous (>95%), and lightweight, with a high capacity for soaking in Ringer's solution. The foam density varied with the inorganic component content, ranging from 0.02 to 0.079 g mL⁻¹, while the mean pore size was approximately 330 μm. The magnetic behavior of Fe₃O₄ nanocubes and the foam composites was characterized. The presence of the inorganic filler caused a shift towards a lower decomposition temperature of PLLA. The conversion energy of both dry and Ringer's solution soaked foams was studied in detail demonstrating that the fabricated ternary composites are highly temperature-responsive under the influence of an alternating magnetic field (AMF), near-infrared (NIR) laser radiation (808, 880, and 1122 nm), and the synergistic effect of both external stimuli. This synergy resulted in faster heating and a higher maximum temperature (T_max ≈ 80 °C). Biological characterization and heating ability analysis enabled the selection of the most reliable foam, which contained 15% magnetic filler, based on its appropriate microstructure, sufficient biocompatibility, and ability to reach biologically relevant temperatures under AMF exposure and the combined action of NIR and AMF. The fabricated materials exhibit high potential for biomedical applications as well as other areas requiring temperature-controlled stimulation of various processes.

---

1. Introduction

Composites are considered among the most promising, modern future materials as they allow for the flexible integration of different compounds, and combine the physicochemical properties of their components within the same construct.^1–3 Such materials include hybrids composed of naturally derived polymers (chitosan, collagen, glucan, polylactide, etc.), which upon integration with inorganic substances of the apatite family, i.e., hydroxyapatites, fluorapatites, carbonated apatites, calcium and/or strontium phosphates, etc. enable the development of advanced bone substitute scaffolds for regenerative medicine.^4–10 The ability to use various apatite compounds and their different properties in biological media allows control over new bone tissue formation, which is one of their key advantages.^11–13 Polymeric materials are often used as a matrix that maintains the appropriate microstructure of the bone replacement material, serving as a scaffold for newly colonizing cells. In the case of bone tissue, apatites combined with polymers facilitate achieving the desired properties of implants by stimulation of remodeling processes and integration with bone tissue, as well as affect mechanical properties and degradability.¹ An important aspect of composite materials is their microstructure, which significantly determines whether the scaffold can be applied to compact or cancellous (spongy) bone. Spongy bones are commonly found at the ends of long bones, within flat bones and inside short bones, and are characterized by a highly porous structure with pore sizes ranging from 150 to 800 μm.¹⁴ Polymers and their composites containing inorganic materials with a foam-like microstructure represent an innovative solution for modern implants, whereas their microstructure, porosity, and other critical parameters can be controlled during the fabrication process.^15–17

Poly-L-lactic acid (PLLA) is a well-known biodegradable polymer widely used in tissue engineering due to its favorable mechanical properties and high biocompatibility. PLLA can be processed into various structures, such as scaffolds and films, that support cell growth and tissue formation.¹⁸ Calcium hydroxyapatite (HAP) is a key component of bone tissue, making it an ideal material for bone regeneration and repair. Its biocompatibility, together with its osteoconductive properties, is well-documented. It supports cell adhesion and proliferation in tissue engineering applications.¹⁹ On the other hand, the combination of PLLA with HAP can lead to an increase in the mechanical strength and bioactivity of the resulting composite materials, making them suitable for load-bearing applications in bone regeneration.²⁰ There are several techniques for the fabrication of the hybrid organic–inorganic composite towards highly porous materials, namely, solvent casting and thermally induced phase separation (TIPS), microsphere sintering, selective polymer extraction, gas foaming, 3-D printing, and many more.^21,22

The TIPS is seen as a very effective and versatile technique for the production of polymeric and composite foams with well-controlled porosity and morphology. It relies on the cooling of an earlier prepared homogenous mixture containing the polymer of choice with filler (organic, inorganic), leading to the solidification of a porous structure. This method is widely applied to biodegradable polymers such as poly(lactic acid) (PLA), poly(ε-caprolactone) (PCL), polyglycolic acid (PGA), polyglycerol sebacate (PGS), and polyvinylidene fluoride (PVDF), which are of great importance in specific biomedical and filtration applications. The main advantage of TIPS is its ability to produce foams with a highly porous microstructure with a pore size and distribution that can be tuned through changes of the synthetic parameters (polymer and filler concentration, solvent, quenching temperature, mixing rate, additives, porophore, etc.).^23–25 This makes TIPS particularly attractive for biomedical applications such as tissue engineering, where uniform pore connectivity is essential for cell migration, nutrient diffusion, and vascularization.^26,27

A very interesting idea is to use magnetic compounds as a functional additive to polymeric or inorganic scaffolds, such as ferrites (MFe₂O₄, M–Co²⁺, Mn²⁺, Ni²⁺, Fe²⁺, Zn²⁺)^2,3 as a functional additive to polymeric or inorganic scaffolds. The main advantage of magnetic nanoparticles is their ability to generate heat under the action of external stimuli, i.e., an alternating magnetic field (AMF) or near-infrared light (NIR), which makes them very useful in magneto-photothermal therapies.^2,28–30 The mechanisms of heat generation vary strongly depending on the type of stimulation. In the case of an AMF, three main contributions to heat induction were identified: (1) hysteresis losses, (2) generation of eddy currents, and (3) residual losses that include Néel's internal and Brownian external losses.³¹ Meanwhile, temperature effects associated with NIR light interaction are predominantly caused by non-radiative processes that lead to energy dissipation (lattice vibrations).^30,32 Both forms of magnetic nanoparticle stimulation (AMF and NIR) allow the achievement of relevant temperature ranges that affect biological processes (<43 °C) or can lead to cancer cell elimination (>43 °C).^32–35 The literature reports mainly binary organic–inorganic composites of PLLA/Fe₃O₄,^36–44 HAP/Fe₃O₄^45,46 ceramic materials or HAP embedded iron apatite⁴⁷ for enhancing cellular activity, cell adhesion, biocompatibility and mechanical properties, and only a few studies are devoted to the temperature effects, mainly focused on the interaction with the AMF. In the context of the ternary hybrids PLLA/HAP/Fe₃O₄ of any form, there is no data on their responsiveness to AMF and NIR. Some results can be found for other types of polymer, like collagen/HAP/Fe₃O₄⁴⁸ (AMF only), chitosan/HAP/Fe₃O₄⁴⁹ (no AMF and NIR), PLA/ENR/Fe₃O₄⁴¹ (ENR – epoxidized rubber) shape memory triggered by AMF and NIR, PLLA/PGA/Fe₃O₄⁴³ (PGA – polyglycolic acid, AMF only), polycaprolactone/HAP/HAP:Fe⁴⁷ (AMF only), and cellulose/HAP/Fe₃O₄⁵⁰ (no AMF and NIR). Therefore, the engineering of new, multifunctional composite materials is necessary to open attractive directions for possible practical applications. Such stimuli-responsive composite materials can serve as tools supporting regenerative medicine by moderate hyperthermic effects. It was noted that temperatures of 39 °C can differentially affect specific human stem cells and their differentiated derivatives. For example, the elevated temperature evokes the upregulation of genes engaged in the regulation of ossification in human embryonic stem cells, and upregulation of genes involved in the regulation of angiogenesis and vascular development in human vascular stem cells.⁵¹ In another study, it was shown that exposure of hFOB cultures to 39 °C resulted in spontaneous osteogenic differentiation to mineralized nodules.⁵² Also, another group confirmed that hFOB cells undergo spontaneous differentiation to mature osteoblasts when cultured at 39.5 °C.⁵³ Therefore, moderate AMF and NIR light stimulation of selected composites can be used in regenerative medicine to speed up the healing processes.

This work proposes an innovative ternary foam composed of PLLA/HAP/Fe₃O₄ fabricated via the TIPS technique. The novelty relies on the development of a highly porous composite that exhibits dual responsiveness to both an alternating magnetic field (AMF) and near-infrared irradiation (NIR_808nm, NIR_880nm, NIR_1122nm) due to the incorporation of morphologically controlled cubic nanomagnetite particles. In comparison to the previous reports, which mainly focus on binary systems or single-mode stimulation, this study demonstrated the possibility of temperature control above and below the therapeutic threshold of 43 °C, with controllable heating under different stimulation modes. Our design strategy offers a versatile, smart heating platform with significant potential for advanced tissue engineering and regenerative medicine applications.

2. Experimental

2.1. Synthesis procedure of magnetic foams

The synthesis of magnetic foams with incorporated magnetite cubic nanoparticles, calcium hydroxyapatite, and polylactide polymer was carried out in three distinct steps: (I) preparation of morphology-controlled Fe₃O₄ cubic nanoparticles, (II) synthesis of nano-sized Ca₁₀(PO₄)₆(OH)₂, and (III) integration of the inorganic components into the polylactide matrix to fabricate magnetic and light-responsive functional foam hybrids (see Fig. 1).

Fig. 1 Scheme of the PLLA/HAP/Fe₃O₄ foam fabrication.

(I) Synthesis of magneto and light temperature-responsive Fe₃O₄ nanocubes. For the preparation of magnetite cubic particles, the well-known thermal decomposition method reported by Kim et al.⁵⁴ was adopted with minor modifications. In a typical procedure, 2 mmol of the iron(III) complex (iron(III) acetylacetonate, Fe(acac)₃, 99.7%, Thermo Fisher Scientific, Germany) was mixed with 10 mL of dibenzyl ether (BE, 98%, Sigma Aldrich, Poland) in a three-neck glass flask until the substrate was completely dissolved. Subsequently, 4 mmol (1.4 mL) of oleic acid (OA, 90%, Sigma Aldrich, Poland) was added to the reaction mixture. All operations involving the iron source and reaction solution were performed under an inert N₂ atmosphere (99.999%, Linde, Poland) using a glove box (GS Glove Box Systemtechnik GmbH P10R250T2, Germany) equipped with an automatic gas control unit. The flask containing all necessary chemicals was then assembled into a reaction setup comprising a mechanical stirrer, a temperature controller with a Pt-100 sensor (LTR 2500, Juchheim, Germany), a heating mantle, an Allihn's condenser column, and a N₂ support gas line. In addition, the N₂ was passed through laboratory scrubbers containing mineral oil and molecular sieves (separately). After that, a solution was degassed with N₂ for 60 min at room temperature. Subsequently, the reaction mixture temperature was increased to 285 °C, and left for an additional 30 min under constant stirring. The final mother liquor with a visible black precipitate was cooled down to room temperature and further separated by centrifugation and washing cycles with the solvent mixture containing hexane/acetone/ethanol (1 : 1 : 1 v/v ratio, all delivered by Chempur, Poland, pure for analysis). The resulting magnetic particles were resuspended in ethanol and stored in a laboratory refrigerator. Nanoparticle concentration in a suspension was determined using a microbalance technique with a Radwag MYA 5.4Y scale by filling up an aluminum foil crucible with 50 μL of particle dispersion and solvent evaporation until dry mass. The average mass value was calculated after three repetitions.

(II) Preparation of calcium nanohydroxyapatite polymer filler. For the fabrication of Ca₁₀(PO₄)₆(OH)₂ inorganic nanoparticles, a well-known precipitation technique proposed by Rodriguez-Lorenzo et al.⁵⁵ was employed, with modified process parameters to obtain smaller particles. Briefly, to prepare approximately 100 g of hydroxyapatite, 1 mol (236.15 g) of Ca(NO₃)₂·4H₂O (99%, Sigma Aldrich, Poland) and 0.6 mol (79.23 g) of (NH₄)₂HPO₄ (99%, Sigma Aldrich, Poland) were used, maintaining a Ca : P ratio of 1.66. Both salts were separately dissolved in deionized water in glass beakers. The phosphate-containing solution was then slowly added to the calcium nitrate solution, leading to the rapid precipitation of a white powder. Vigorous stirring, preferably mechanical for large batches, was essential to prevent powder adhesion and dense material formation. During mixing, a 25% NH₄OH solution (99%, Avantor, Poland) was added in 20 mL portions to keep a strongly basic pH above 9. This step was crucial for transforming the initially precipitated apatite into a hydroxyapatite structure. Maintaining the pH above 9 was particularly important, as a neutral pH could lead to the formation of a TCP/HAP (TCP – tricalcium phosphate) mixture.⁵⁶ The presence of biphasic apatite can be desirable in certain cases to adjust phosphate resorption rates for bone integration processes in apatite-based implants.⁵⁷ The whole mixing stage lasted approximately 2 h, with continuous pH monitoring. The product was then separated via centrifugation, and washed with Milli-Q water to purify it and remove the ammonium nitrate by-product, which accounted for nearly half of the total mixture. Purification was critical for safety reasons, preventing the risk of explosion and NO_x contamination during thermal treatment. However, the separated NH₄NO₃ could be collected and reused as a key ingredient in plant fertilizers. Finally, the resulting white powder was dried at 60 °C for 24 h and annealed at 500 °C for 3 h. Due to significant particle aggregation caused by sintering, a final milling step was performed to enhance sample uniformity. The primary advantage of this protocol lies in scalability, enabling large-scale hydroxyapatite production.

(III) Fabrication of functional foam hybrid materials. The integration of inorganic components, magneto-light-responsive Fe₃O₄ cubic nanoparticles, and bioresorbable HAP nanoparticles into the PLLA matrix was carried out by using the TIPS technique supported with SLP (salt-leaching process). Further details on this approach can be found elsewhere.⁵⁸ However, for this study, the process parameters were significantly modified. Specifically, PLLA (Resomer L210s, Evonik, Germany), along with the previously prepared HAP and cubic Fe₃O₄ nanoparticles, served as substrates. In the standard procedure, PLLA and HAP in a 50 : 50 weight ratio were suspended in 1,4-dioxane (99%, Sigma Aldrich, Poland) at 75 °C under mechanical stirring overnight, ensuring the formation of a homogeneous mixture. Subsequently, different amounts of Fe₃O₄ in dioxane were added to achieve magnetic particle concentrations ranging from 1 to 30 wt%. The final solution was mixed vigorously with a stirrer and placed on a shaker to eliminate air bubbles. Sodium chloride (NaCl, Stanlab, Poland) with an average particle size of 500 μm was used as a porogen. After mixing, the blend was dispensed into a 24-well plate (0.5 mL per well) and frozen at −40 °C for 24 h. This was followed by freeze-drying at −50 °C under a vacuum of 10 Pa for an additional 24 h to form porous foams. In the final step, foams, including the reference sample without Fe₃O₄, were subjected to SLP, which was repeated multiple times with fresh water exchange to ensure NaCl removal. Washing effectiveness was confirmed by the reaction of the AgNO₃ solution with the supernatant to verify the absence of Cl⁻ anions. Final foam composites were dried at 40 °C for 24 h and used for physicochemical and biological characterization.

2.2. Characterization of the physicochemical properties of the fabricated materials

The X-ray powder diffraction technique (XRD) was employed to study the structural properties of the stock Fe₃O₄ magnetite and HAP particles as well as fabricated foams. All diffraction patterns were recorded using a Bruker D8 Advanced diffractometer with an X-ray source copper lamp (K_α1 1.54060 Å) within a 2θ range of 15–65°. In addition, a Ni filter was utilized to filter out K_α2 associated reflections. In the case of highly X-ray-absorbing Fe₃O₄ particles and foams with incorporated magnetite, the data curation relied on background correction (Diffrac.Eva software (V.2)), normalization, and signal smoothing (Origin Pro 9, Origin Lab). The final results were compared with the reference cards from the ICDD database (International Centre for Diffraction Data). Sample preparation involved the evaporation of part of the ethanol suspension of Fe₃O₄ to a dry powder. No special treatment for HAP and foam measurements was used.

Magnetite and calcium hydroxyapatite particle size, morphology, and distribution were determined by means of transmission electron microscopy (TEM) with a Tecnai Osiris X-FEG HRTEM microscope (FEI Company, USA) operating at 200 kV. Sample preparation required placing a droplet of the ethanol-based nanoparticle suspensions of all inorganic materials (0.25 mg mL⁻¹) on a carbon-coated 200 mesh copper grid (EM Resolutions, United Kingdom) and slow drying overnight at room temperature under dust protection.

In the case of the foam hybrids, microstructure imaging was carried out using a Tescan Vega 3 (Tescan Group, Czech Republic) scanning electron microscope. The sample was deposited onto carbon tape, mounted on an aluminum stage, and placed in the microscope chamber. Due to the risk of polymer melting, a low-voltage regime was used (max 5 kV). The size distribution analysis of Fe₃O₄, HAP particles, and pore diameter was performed using freeware ImageJ software (ver. 1.46r). The density and porosity of the foams were measured using a hydrostatic balance Shimadzu AX 220 (Shimadzu Ltd, Japan). The porosity was calculated with the following formula:

where Φ_p is porosity, ρ_sc represents scaffold density, and ρ_b is the bulk density of the polymer. Both ρ_sc and ρ_b (1.24 g mL⁻¹) were measured by the buoyancy method.

Fourier transform infrared spectroscopy (FTIR) was employed to analyze the structural features of the fabricated foams. A Thermo Scientific Nicolet iZ10 spectrometer equipped with an attenuated total reflection (ATR) accessory was used and spectra were recorded over a range of 4000–500 cm⁻¹. Before measurement, a standard calibration procedure was performed to correct for air humidity. No special sample treatment was required. Samples were directly placed on the diamond crystal and pressed against the hot-spot using a bolt.

Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were performed using a TGA/DSC1 Mettler Toledo system under a N₂ atmosphere. For each foam sample, 5 mg of material was taken and heated from 50 °C to 550 °C at a heating rate of 20 °C min⁻¹. Data from the TGA characterization were processed with STARe software and plotted in Origin 9.0. Composite glass transition and crystallization temperatures were determined utilizing a DSC apparatus coupled with a TC 100 intercooler under an inert N₂ atmosphere (30 mL min⁻¹). For this purpose, approximately 6 mg of foam was placed inside an aluminum pan and heated from 0 °C to 200 °C at a rate of 5 °C min⁻¹. Afterward, the sample was cooled down to 0 °C, and a reheating scan was performed to eliminate the foam's thermal history and evaluate thermal transitions.

Mechanical testing, in particular compressive strength measurements, was performed using a universal testing machine, the INSTRON 5960 (Norwood, MA, USA), equipped with a 10 kN load cell. The compression rate was set to 1 mm min⁻¹. The specimens had a cylindrical shape, measuring 4 mm in height and 15 mm in diameter. The tests were conducted until a maximum strain of 30% was reached. For each sample type, three replicates were tested, and the results were averaged.

Magnetic measurements were performed using a Quantum Design MPMS7XL superconducting quantum interference device (SQUID) magnetometer. Zero-field-cooled (ZFC) and field-cooled (FC) magnetization curves were recorded during warming from 3 K to 350 K under an applied magnetic field of 10 mT. Isothermal magnetization measurements were conducted at 3 K and 300 K with magnetic fields ranging from −5 T to +5 T. To prevent physical rotation of the magnetic nanoparticles during measurements, the samples were immobilized within gelatin capsules using a small amount of adhesive. The organic content, originating from surfactant residues and the PLLA matrix, was quantified by thermogravimetric analysis for each sample, and the magnetization data were subsequently corrected based on the inorganic (Fe₃O₄) nanoparticle mass.

⁵⁷Mössbauer spectroscopy using a ⁵⁷Co source diffused into a Rh matrix with constant acceleration movement was performed at room temperature in transmission geometry with a Wissel spectrometer equipped with a proportional counter. Measurements were conducted in zero external magnetic field. An α-Fe foil was used as the calibration standard. Spectra were analyzed using the NORMOS fitting software.

The heat generation ability of the magnetite-containing foams, as well as the reference PLLA/HAP sample, was measured using two external stimuli, namely alternating magnetic field and near-infrared laser radiation (NIR₈₀₈, NIR₈₈₀, and NIR₁₁₂₂ nm) under separate and synergy mode. A G2 D5 Series Multimode 1500 W driver (nanoScale Biomagnetics, Spain) with a thermally insulated S32 coil was used (polystyrene box). In the case of laser stimulation, 808 nm, 880 nm, and 1122 nm continuous laser modules (power stability no worse than 1%) equipped with 400 μm optical fibers (CNI, China) were utilized. Lasers were calibrated with an Ophir StarLite power meter using a beam track thermal sensor 10 A-PPS (Ophir, Israel) with a measurable laser power range of 20 mW–10 W, with a laser beam size detection accuracy of 5%. The magnetic field frequency and intensity were chosen between 145–396 kHz and 22–32 kA m⁻¹, while the laser output power for all three wavelengths was between 100–600 mW. All measurements were done on dry and Ringer's solution soaked foams (2 h at room temperature), mimicking conditions at possible implantation sites. Temperature effects were recorded with a FLIR T660 scientific thermovision camera (FLIR, USA) and analyzed with dedicated ResearchIR software (FLIR, USA). The final data presentation was performed in Origin 9.0 software.

2.3. Cell culture experiments

The normal human fetal osteoblast cell line (hFOB 1.19, American Type Culture Collection (ATCC), USA) was used to assess cytotoxicity through both indirect and direct methods, as well as to evaluate cell proliferation. Cells were cultured in a 1 : 1 mixture of DMEM/Ham's F12 medium without phenol red (Gibco Thermo Fisher Scientific, USA), supplemented with 10% fetal bovine serum (Pan-Biotech GmbH, Germany), 100 U mL⁻¹ penicillin, 100 μg mL⁻¹ streptomycin, and 300 μg mL⁻¹ G418 (Sigma-Aldrich, Poland). The cells were maintained at 34 °C in a 5% CO₂ atmosphere.

2.3.1. Evaluation of cytotoxicity by the indirect method. Cytotoxicity was assessed using the MTT assay, in accordance with ISO 10993-5, using extracts of biomaterials prepared according to ISO 10993-12. In summary, extracts were obtained by incubating 25 mg of the composites in 1 mL of culture medium, which was dictated by the high capacity of the materials to absorb fluids, at 37 °C for 24 h. Concurrently, a negative control for cytotoxicity was prepared. Cells were seeded in 96-well plates at a density of 1 × 10⁵ cells per mL in a volume of 100 μL per well and cultured at 34 °C for 24 h. Subsequently, the medium was replaced with 100 μL of the respective foam extracts, and the cells were incubated for an additional 24 h. After incubation, an MTT colorimetric assay (Sigma-Aldrich, Poland) was performed to evaluate cell viability and metabolic activity. Briefly, 25 μL of MTT solution (5 mg MTT/1 mL PBS) was added to each well and incubated for 3 h at 34 °C to allow mitochondrial dehydrogenase enzymes in viable cells to reduce MTT to formazan. As a result of this reduction, formazan accumulated as crystals within the living cells. After the specified incubation period, 100 μL of 10% SDS solution in 0.01 M HCl was added to each well to dissolve the formazan crystals. After 12 h, the absorbance of the wells referring to the solubilized formazan concentration was measured using a plate reader at a wavelength of 570 nm. The amount of formazan is proportional to the number of viable cells. The results of the MTT assay were expressed as a percentage of the optical density (OD) value obtained relative to the negative control for cytotoxicity. Cytotoxicity was also assessed qualitatively with the live/dead double staining kit (Sigma-Aldrich, Poland) following the manufacturer's protocol. The hFOB 1.19 cells treated with biomaterial extracts were stained with fluorescent dyes and observed utilizing a confocal laser scanning microscope (CLSM, Olympus Fluoview equipped with FV1000, Japan)

2.3.2. Evaluation of cytotoxicity by the direct method. In order to assess cytotoxicity using the direct method, biomaterials weighing 25 mg were placed in a 24-well plate and pre-soaked in complete culture medium. Subsequently, hFOB 1.19 cells were seeded directly onto the materials at a density of 5 × 10⁴ cells per mL. After 48 hours of culture, the cells were stained with the live/dead double staining kit (Sigma-Aldrich, Poland), following the manufacturer's protocol. Subsequently, the samples were visualized with a confocal laser scanning microscope (CLSM, Olympus Fluoview equipped with FV1000, Poland).

2.3.3. Evaluation of cell proliferation. The impact of biomaterial extracts on cell proliferation was assessed both qualitatively and quantitatively using the following procedure. Cell line at a density of 5 × 10⁴ cells per mL was seeded in a volume of 100 μL into 96-well plates and incubated in a culture medium with 10% FBS for 24 h at 34 °C. On the following day, the culture medium was replaced with biomaterial extracts at a volume of 100 μL (prepared at a ratio of 25 mg of composite per 1 mL of 10% culture medium). The experiment was conducted over 5 days with two time points for the evaluation of cell proliferation. Specifically, on days 3 and 5, the cell number was determined employing a cell counting kit-8 (WST-8 assay, Sigma-Aldrich, Poland) according to the manufacturer's instructions. Simultaneously, at the same time points, the cells were fixed in a 3.7% formaldehyde solution and permeabilized with 0.2% Triton X-100 solution. They were then stained with the fluorescent dyes Hoechst 33342 and AlexaFluor™ 635 Phalloidin to visualize actin cytoskeleton fibers and cell nuclei. The samples were observed using confocal laser scanning microscopy (CLSM, Olympus Fluoview equipped with FV1000, Japan) for a qualitative evaluation of cell growth over time.

3. Results and discussion

3.1. Physicochemical characterization of stock nanoparticles and foams

The structural phase identification of the stock Fe₃O₄ and HAP nanoparticles, as well as the representative foam, was carried out using the XRD technique (Fig. 2). The diffraction patterns for Fe₃O₄ and HAP were compared with reference diffraction standard cards no. 19-0629 and 09-0432 from the ICDD international database, respectively. Both compounds exhibited a very good agreement with characteristic crystalline phase reflections, confirming their structures as magnetite and hydroxyapatite, respectively. In the case of the PLLA/HAP/Fe₃O₄ foam, the diffraction pattern reveals a combination of the primary characteristic peaks of all its constituents. The PLLA is classified as a semicrystalline polymer, and it can exist in several crystalline forms (α, α′, β, γ). The inset in Fig. 2 highlights typical signals observed within the 15–25° 2θ range. These features, especially at around 17° and 19° 2θ are indicative of the α PLLA crystalline phase^59,60 and are clearly visible in the composite material as well. The obtained result supported evidence of the successful integration of all compounds in the hybrid material.

Fig. 2 X-ray powder diffraction patterns of the stock Fe₃O₄, HAP particles, and PLLA/HAP/Fe₃O₄. The inset shows a pattern of the neat PLLA polymer.

The primary particle size, distribution, and morphological characteristics of stock Fe₃O₄ and HAP particles were analyzed using TEM (Fig. 3). The magnetite nanoparticles predominantly showed a cubic morphology with an average size of approximately 80 nm. In contrast, the hydroxyapatite particles formed a finer-grained product with a mean size of 23 nm. Unlike Fe₃O₄, HAP crystallites demonstrated a tendency to form elongated shapes, highlighting a distinct difference in their growth behavior.

Fig. 3 TEM images of the Fe₃O₄ and HAP nanoparticles and their size distribution histograms.

Fig. 4 presents a representative photo of the actual PLLA/HAP/Fe₃O₄ 5% composite sample, showing that the foam is magnetically responsive. The sample floated on the water surface, which is an indication of its low density. The final composite foams were prepared with a broad concentration of Fe₃O₄ ranging from 1 up to 30% to study the effect of magnetic components on microstructure (together with HAP), heat generation ability, and biological activity.

Fig. 4 Photograph of the representative PLLA/HAP/Fe₃O₄ foam with 5% of the magnetic component.

Foam microstructure was analyzed using the SEM technique. A representative image of the PLLA/HAP/Fe₃O₄ composite is shown in Fig. 5. As one can see, the hybrid material exhibited a significant porosity, with an average pore size of 329 ± 53 μm, which corresponded to the size of the NaCl porogen used during the fabrication process. In addition, smaller pores (below 50 μm) were observed, particularly within the foam wall areas. Their formation is likely attributed to solvent (dioxane) evaporation during the foam processing. The signs of the microstructure collapse were detected upon a significant increase in the inorganic content (above 70% – HAP and Fe₃O₄). In addition, the microstructure of those two samples was measured after 24 months of aging showing no signs of any changes (see Fig. S1, ESI†).

Fig. 5 SEM images of the PLLA/HAP/Fe₃O₄ 5% and PLLA/HAP/Fe₃O₄ 20% composite foam materials.

The density of foams with varying Fe₃O₄ and HAP nanoparticle content was measured using a hydrostatic balance. Subsequently, porosity was calculated, and the results are summarized in Table 1. The PLLA foam shows the lowest scaffold density (0.0208 g mL⁻¹), while the incorporation of HAP (3.16 g mL⁻¹ bulk) slightly increases the skeletal density (0.0267 g mL⁻¹). A progressive increase in the Fe₃O₄ (5.17 g mL⁻¹) concentration led to higher density. Notably, a significant change occurs at Fe₃O₄ concentrations exceeding 10%, where the scaffold density nearly quadruples (from 0.0208 to 0.0790 g mL⁻¹) due to the increase of the high density magnetic component. Foam porosity remains relatively unchanged, with only an approximate 3% variation when comparing a PLLA matrix containing 20% Fe₃O₄. However, the highest difference is observed at 30% of Fe₃O₄ (80% of total HAP and Fe₃O₄), where porosity reaches 93% and is caused by ongoing microstructure collapse. These findings are consistent with Szustakiewicz et al.⁵⁸ on PLLA/HAP scaffolds fabricated using the same approach.

Table 1 The density and porosity of the PLLA/HAP/Fe₃O₄ foams with different concentrations of magnetite

Sample	Porosity Φp (%)	Density ρsc (g mL^−1)
PLLA	98.32	0.021 ± 0.012
PLLA/HAP	97.85	0.027 ± 0.002
PLLA/HAP/Fe3O4 1%	97.77	0.028 ± 0.006
PLLA/HAP/Fe3O4 5%	97.35	0.033 ± 0.011
PLLA/HAP/Fe3O4 10%	96.55	0.043 ± 0.009
PLLA/HAP/Fe3O4 15%	95.99	0.050 ± 0.004
PLLA/HAP/Fe3O4 20%	95.44	0.057 ± 0.007
PLLA/HAP/Fe3O4 30%	93.62	0.080 ± 0.011

---

The PLLA/HAP/Fe₃O₄ foams with a concentration of magnetite above 20% are notably stiffer, showing the first signs of microstructural degradation. The observed changes in porosity and density suggest that the incorporation of HAP with Fe₃O₄ into PLLA composites leads to denser structures with reduced porosity when the total inorganic content exceeds 80%. Highly concentrated foams are brittle and prone to fracture (above 70% of Fe₃O₄ and HAP). However, foams below 20% of magnetite will show potential for cell colonization and proliferation due to the high porosity and preferential pore size. An attempt was also made to determine the BET specific surface area (SSA) (see Fig. S3, S4 and Table S1 in the ESI†). However, the low SSA values suggest a predominantly macroporous character, as supported by SEM imaging.

FTIR-ATR spectroscopy was employed to study the structural features as well as surface properties of the Fe₃O₄, HAP, and PLLA/HAP/Fe₃O₄ foams (Fig. 6). For the stock cubic Fe₃O₄ nanoparticles, a typical broad and high-intensity vibration mode was recorded within the range of 680–500 cm⁻¹ associated with the Fe–O characteristic for tetrahedrally coordinated iron cations.⁶¹ The spectra of the HAP nanoparticles prepared by the precipitation technique show characteristic bands ascribed to the PO₄³⁻ and OH⁻ group vibrations.⁶² Therefore, the modes with maxima at around 1091, 1023, and 964 cm⁻¹ are attributed to the stretching vibrations, while those at 601, 564 and 470 cm⁻¹ correspond with bending modes associated with the ν₃, ν₁, and ν₄ vibrations, respectively. The prominent band at around 630 cm⁻¹ is characteristic of the hydroxyl unit vibrations in the HAP structure. In the case of the PLLA/HAP/Fe₃O₄ foams, there are several features. The first one was associated with the occurrence of bands related to the organic polymer matrix PLLA. The following vibration band attribution was made, namely 2990 cm⁻¹ (CH₃), 2940 cm⁻¹ (C–H), 1756 cm⁻¹ (C=O), 1455 cm⁻¹ (C–H), 1381 and 1359 cm⁻¹ (C–H), 1181 cm⁻¹ (C–O), 1129 cm⁻¹ (C–O), 870 and 754 cm⁻¹ (CH₃). The spectral features of PLLA are in good agreement with the literature data.^59,60,63 Integration of inorganic compounds into the PLLA matrix resulted mainly in a visible shift of the PO₄³⁻ modes, being evidence of interaction with the organic material. We observed that the contribution of the Fe₃O₄ nanoparticles results in peak broadening between 630–500 cm⁻¹ for elevated concentrations. Moreover, bands at 630 and 600 cm⁻¹ associated with HAP modes diminished or were covered by the intense Fe–O, and strong ν₄ (564 cm⁻¹) overlapping vibrations. In the case of the foams with 20 and 30% of Fe₃O₄, the FTIR-ATR spectra can be found in the ESI† (Fig. S2).

Fig. 6 The FTIR-ATR spectra of the Fe₃O₄, HAP stock nanoparticles, and PLLA/HAP/Fe₃O₄ foams with different magnetite concentrations.

The thermal properties of the foam scaffolds were investigated using TGA-DTA and DSC measurements as a function of Fe₃O₄ content, for the PLLA/HAP binary composite and PLLA reference foam (Fig. 7 and Fig. 8). The greatest weight loss was observed for the PLLA/HAP/Fe₃O₄ 1% foam, with a subsequent decrease in weight loss. This was attributed to the increase of inorganic content and the PLLA polymer fraction reduction. Generally, PLLA undergoes complete decomposition between 328 °C and 410 °C, with a peak centered at 372 °C. The addition of HAP to the polymer did not significantly alter the decomposition temperature. However, the presence of Fe₃O₄ nanoparticles caused a notable shift of approximately 30 °C toward a lower decomposition temperature (around 340 °C). Similar behavior was observed by Yang et al. in PLLA/Fe₃O₄⁴⁴ binary composites as well as upon ZnO nanoparticle addition to the PLLA matrix.⁶⁴ Both authors attributed the temperature shift to the expansion of the interface space and the possible catalytic activity of the nanoparticles, which decreased the polymer thermal stability. It was found that the initial decomposition temperature of the foams changed significantly. It is worth noting that the decomposition of the PLLA/HAP/Fe₃O₄ scaffolds was not only faster but also had a narrower temperature range. This suggests that magnetite nanoparticles may act as catalysts or oxidants, accelerating PLLA depolymerization at lower temperatures.

Fig. 7 TGA (left panel) and DTA curves (right panel) of the chosen PLLA/HAP/Fe₃O₄ foams as a function of magnetite concentration.

Fig. 8 DSC analysis of the PLLA/HAP and PLLA/HAP/Fe₃O₄ foams (a) 1^st heating DSC curves, (b) the cooling stage and (c) 2^nd heating DSC curves.

Analysis of the DSC results (Fig. 8) shows three distinct thermal effects, namely glass transition (T_g) within a temperature range of 61–69 °C, cold crystallization around 100 °C, and a melting temperature (T_m) between 178–185 °C. During the 1st heating, a slight decrease in glass transition temperature was found, followed by an increase in magnetite concentration. A similar effect was noticed in the melting process, whose temperature peak for binary PLLA/HAP foam was at 185 °C and decreased to 181 °C for the composite with 20% Fe₃O₄. Upon integrating the magnetic particles into the PLLA/HAP, a noticeable promotion of nucleation in the crystallization process was observed. This phenomenon is particularly evident during heating and cooling cycles. It manifests as an increase in the endothermic and exothermic peaks. During the cooling, one can note that the exotherm peak is monomodal with a maximum of around 100 °C for all foams. We did not observe a bimodal shape of the exothermic peak (two T_c temperatures) and their shift towards higher temperatures (up to 119 °C) upon HAP amount increased as reported by Szustakiewicz.⁵⁸ In this study, the T_c remained stable. We believe that this is probably due to the different character of the interaction of magnetite particles. They did not significantly alter the crystallization of the PLLA polymer, like HAP. A similar observation was reported by Laraba et al.⁶⁰ with graphene oxide (GO) as a filler of the PLLA matrix. However, GO shifted the melting temperature toward higher values. Interestingly, in the plot representing the 2nd heating, one can distinguish an exothermic peak with a maximum of around 161 °C. In accordance with the literature data, this behavior is associated with the phase transition of α′ to α form of PLLA, meaning that composite foams consist of a mixture of both PLLA structures.⁶⁵ The former one was not detected by XRD, probably due to its low content and the instrument's sensitivity. Moreover, the rise of concentration of magnetic particles led to an increase in the exothermic peak surface area, indicating the effect of Fe₃O₄ on the formation of a greater amount of the α crystalline phase.

The data showing results of the PLLA/HAP/Fe₃O₄ foams’ mechanical properties including compression modulus and compressive strain at 10% and 30% are gathered in Table S2 (ESI†). It was observed that the incorporation of the HAP and Fe₃O₄ nanoparticles into the PLLA polymer matrix significantly affects the mechanical behavior of the composite foams. The compression modulus of the neat PLLA 47 kPa increases up to 106 kPa at 80% (50% HAP and 30% of Fe₃O₄) with inorganic filler loading, which indicates a significant increase in foam stiffness that becomes brittle. The composite material up to 15% of the Fe₃O₄ content (65% in total both HAP and Fe₃O₄) showed a consistent enhancement of compression modulus and compressible stress at 10% strain that results in foam reinforcement. It is particularly evident for the PLLA/HAP/Fe₃O₄ 15% hybrid, which showed a compression modulus of 93 kPa and a compressive stress around 11 kPa at 10% strain (double the stress value of neat PLLA). However, the mechanical properties at 30% strain applied revealed a complex behavior, i.e., neat PLLA showed the highest stress value of 151 kPa, while the addition of the Fe₃O₄ filler for the concentration range between 1–15% resulted in reduced stress that could be a reflection of structure softening under compression. What is more interesting, the 20% foam showed a sharp increase of compressive stress at 30% strain (185 kPa), which can indicate pore wall collapse and localized densification at higher loadings. These results indicated the need for precise control over the inorganic filler content and the porous foam structure to assure a proper balance between stiffness and deformation resistance in multifunctional composite scaffolds.

ZFC and FC magnetization curves (Fig. 9) exhibit a bifurcation above 350 K, indicating that all samples remain in a magnetically blocked state throughout the entire measured temperature range.⁶⁶ This behavior is consistent with the critical size for single-domain Fe₃O₄ nanoparticles (approximately 91 nm).^67,68 The ZFC curves display a pronounced drop in magnetization around 100 K, which is attributed to the Verwey transition⁶⁹ – a phenomenon characteristic of magnetite and absent in maghemite. No significant shift in the Verwey transition temperature was detected with increasing Fe₃O₄ nanocube concentration in the PLLA matrix, suggesting that interparticle interactions remain relatively weak and do not significantly influence the transition behavior. Isothermal magnetization measurements at 3 K reveal moderate coercivity (μ₀H_c ≈ 0.04 T, or 400 Oe)⁵⁴ across all samples (Fig. 9). The saturation magnetization values (83, 97, 87, and 94 Am² kg⁻¹ for 1%, 5%, 10%, and 15% Fe₃O₄ loading, respectively) remain close to that of bulk magnetite (Ms ≈ 90 Am² kg⁻¹), indicating the high crystallinity and preservation of the magnetic properties of the Fe₃O₄ nanocubes. Only slight variation in remanent magnetization was observed across the PLLA composites, with values of 26, 23, 25, and 23 Am² kg⁻¹ for 1%, 5%, 10%, and 15% Fe₃O₄ loadings, respectively. At room temperature, the isothermal magnetization curves exhibit very low coercivity, further supporting the presence of a blocked magnetic state for the Fe₃O₄ nanocubes under ambient conditions. Across the measured temperature range, the magnetic response of the samples remains effectively unchanged, indicating that the Fe₃O₄ nanocube loading has minimal influence on the overall magnetic behavior in this regime.

Fig. 9 ZFC/FC magnetization curves (a) of Fe₃O₄ nanocubes and PLLA/HAP/Fe₃O₄ foams (dashed line indicates the Verwey transition, T_v) and their isothermal magnetization (b) at 300 K (inset: zoomed in 3 K isothermal magnetization).

To unequivocally confirm the presence of the magnetite phase, room-temperature ⁵⁷Fe Mössbauer spectroscopy was performed on the Fe₃O₄ nanoparticles (Fig. 10). The spectrum was best fitted using three sextets: two corresponding to the magnetite phase and one to maghemite. The magnetite components exhibit characteristic isomer shifts of 0.272 and 0.67 mm s⁻¹ and hyperfine fields of 48.58 T and 46.1 T, assigned to Fe³⁺ in tetrahedral (A) sites and Fe²⁺ in octahedral (B) sites, respectively.^70,71 The A : B spectral area ratio of 1 : 2 strongly supports the presence of the inverse spinel structure typical of magnetite. The third sextet, with an isomer shift of 0.27 mm s⁻¹ and a hyperfine field of 49.8 T, corresponds to high-spin Fe³⁺ and is attributed to a minor maghemite fraction. Quantitative analysis reveals that the nanoparticles consist predominantly of magnetite (70%) with a smaller contribution from surface-oxidized maghemite or nonstoichiometric magnetite (30%). These Mössbauer results provide direct and definitive evidence for the dominance of the magnetite phase in the synthesized Fe₃O₄ nanoparticles.

Fig. 10 ⁵⁷Fe Mössbauer spectrum of the synthesized NPs. Red dots, black line, and blue line represent experimental data, overall fit, and residual, respectively. Different fitted iron sub spectra are shown as colored surfaces.

3.2. Heating ability of the ternary PLLA/HAP/Fe₃O₄ foam composites under AMF and NIR stimulation

The effectiveness of PLLA/HAP/Fe₃O₄ foam in converting energy into heat under AMF and NIR light was evaluated for both separate and synergic action (dual mode) of external stimuli. All measurements were conducted as a function of Fe₃O₄ concentration (1–30%) and optimized for the magnetic field frequency (149–496 kHz), intensity (22–32 kA m⁻¹, corresponding to 27.6–40.2 mT), laser wavelengths (808, 880, 1122 nm), and laser power (100–600 mW) for dry and Ringer's solution soaked materials. A binary PLLA/HAP hybrid foam was used as the reference material. Optimization of the external stimulation parameters was performed on a PLLA/HAP/Fe₃O₄ sample containing 8% Fe₃O₄, which resulted in obtaining temperatures exceeding 43 °C crucial for potential biological applications (Fig. 11). The complete characterization of stock Fe₃O₄ nanoparticles and their heat-generation capabilities was the subject of our previous study on magneto-plasmonic heterostructures.⁷²

Fig. 11 Heating curves of the PLLA/HAP/Fe₃O₄ 8% dry foam measured for (a) AMF stimulation, (b) 808 nm laser, (c) 880 nm laser, and (d) 1122 nm laser as a function of the AMF parameters and laser power (separate modes).

The Fe₃O₄ concentration dependence in PLLA/HAP/Fe₃O₄ was measured for Ringer's solution soaked foams by using the following AMF parameters 336 kHz, 27 kA m⁻¹ (Fig. S5, ESI†). The choice of this particular hybrid foam was dictated by their ability to respond to the AMF field, while samples with the Fe₃O₄ amount below 8% did not show a pronounced temperature effect upon Ringer's solution soaking. For dual mode (synergy), heating experiments were conducted at 336 kHZ, 27 kA m⁻¹, and 600 mW for three wavelengths 808, 880, and 1122 nm covering the Ist and the IInd optical biological window (Fig. 13). In all cases, the cooling stage was recorded for both dry and soaked composites. The optimization of the AMF parameters and laser power was performed on a dry PLLA/HAP/Fe₃O₄ foam containing 8% magnetite cubic particles to choose the most reliable conditions for contactless heat generation. As shown in Fig. 11a, the best performance was achieved for the 336 kHz and 27 kA m⁻¹ (33.9 mT) with a maximum temperature of 54.1 °C. The 145 kHz and 32 kA m⁻¹ (40.2 mT) also gave a high temperature of around 50 °C for the dry composite. In the case of the laser light, three wavelengths were tested 808, 880, and 1122 nm (see Fig. 11b–d) for the 100, 200, and 300 mW laser power, respectively. A higher power was not applied since the T_max of 93 °C was achieved. Therefore, further increase of this parameter might lead to polymer melting and microstructure degradation, which is seen as a detrimental effect.

The variation of the T_max (47–98 °C) observed for the different lasers can be attributed to the absorption of the Fe₃O₄, which decreases up to 900 nm and increases again above it,⁷³ as well as differences in laser spot areas. In the case of the 880 nm line, the laser light has the highest area (2.01 cm²), while the 808 nm and 1122 nm lasers' spot sizes were 0.95 and 1.33 cm², respectively. These differences directly influence the laser optical density (LOD), which ranged from 0.1–0.63 W cm⁻² depending on the laser power (100–600 mW) and wavelength. All used wavelengths have a capacity for significant heat induction due to the Fe₃O₄ efficient absorption that covers a broad spectral range.⁷⁴ To estimate the heat induction effectiveness, the specific absorption rate (SAR)⁷⁵ was calculated using the following formula:

where C_p is specific heat capacity (J g⁻¹ °C⁻¹), and dT/dt defines temperature increase over time (°C s⁻¹), a product of linear fitting of the heating curve for the first tens of seconds of exposure to the given stimulation. The C_p (1.23 J g⁻¹ °C⁻¹) for the PLLA/HAP/Fe₃O₄ 8% foam was estimated as the weighted average (see Table 2)⁷⁶ by taking specific heat capacities of PLLA (1.8 J g⁻¹ °C⁻¹), HAP (0.75 J g⁻¹ °C⁻¹), and Fe₃O₄ (0.67 J g⁻¹ °C⁻¹).

Table 2 AMF and NIR conversion parameters, i.e., heating speed dT/dt and maximum temperature (T_max) for the PLLA/HAP/Fe₃O₄ 8% dry foam

Sample	Parameters (AMF and NIR)	dT/dt (°C s^−1)	T max (°C)	SAR (W g^−1)
AMF
PLLA/HAP/Fe3O4 8%	145 kHz, 32 kA m^−1	1.91	50.3	29.3
PLLA/HAP/Fe3O4 8%	336 kHz, 27 kA m^−1	2.49	54.1	38.3
PLLA/HAP/Fe3O4 8%	496 kHz, 22 kA m^−1	1.59	43.5	24.4
Laser 808 nm
PLLA/HAP/Fe3O4 8%	100 mW	6.05	50.5	93.0
PLLA/HAP/Fe3O4 8%	200 mW	18.17	76.5	279.4
PLLA/HAP/Fe3O4 8%	300 mW	25.27	97.6	388.5
Laser 880 nm
PLLA/HAP/Fe3O4 8%	100 mW	4.32	47.7	66.4
PLLA/HAP/Fe3O4 8%	200 mW	9.58	60.2	147.3
PLLA/HAP/Fe3O4 8%	300 mW	15.43	78.3	237.2
Laser 1122 nm
PLLA/HAP/Fe3O4 8%	100 mW	5.10	50.8	78.4
PLLA/HAP/Fe3O4 8%	200 mW	9.17	73.8	140.9
PLLA/HAP/Fe3O4 8%	300 mW	16.32	93.4	250.9

---

A comparison of the heating speed (Table 2) revealed a difference in the effectiveness of the heat generation mechanisms over one order of magnitude. For the AMF stimulation, the dT/dt is between 1.59–2.49 °C s⁻¹, while laser stimulation led to values within the range of 5.1–25.2 °C s⁻¹, showing its superior role over AMF heat induction. In the case of the SAR, heat induction under AMF led to values between 24–38 W g⁻¹, depending on the field frequency and intensity, with a maximum for the foam measured with 336 kHz and 27 kA m⁻¹ field parameters. Further increase of the field frequency reduces the conversion ability since at certain frequencies, reorientation of the magnetic moments will no longer follow with fast changes of the magnetic field. Typically, magnetic particle dispersion with good heating ability shows SAR above 100 W g⁻¹.²⁸ The literature describes three main mechanisms associated with the heat generation on magnetic particles, namely (I) hysteresis losses due to the coercivity field, (II) eddy currents (depend on the type of material, metal–semiconductor–dielectric, AMF frequency and field intensity), and (III) residual losses such as Néel and Brown relaxations (internal and external relaxivity).⁷⁷ It is important to note that magnetic material heating ability will be critically dependent on the particle size, chemical composition, shape, agglomeration degree, surface to volume ratio, surface functionalization, hydrodynamic size, colloidal stability, liquid viscosity, magnetic anisotropy, magnetic saturation, coercivity, and other factors as well.^28,78–84 In the case of the composite foam, magnetic particles are embedded within the polymer matrix. Thus, physical movements of the particles upon AMF are not possible. Therefore, the Brownian relaxation losses responsible for heating are no longer available, leading to an SAR decrease.

In the case of the NIR laser light induced heating, the calculated SAR values for the 8% Fe₃O₄ sample were an order of magnitude higher than those for AMF action, and depending on laser wavelength and power, are within the range of 66 to 388 W g⁻¹. It needs to be mentioned that the heat generation mechanism is completely different, and relies on the light absorption as well as energy dissipation mainly through non-radiative processes towards net vibrations (phonons) that cause a temperature increase.³⁰ Therefore, the use of NIR in the case of the composite foams is far more effective than AMF stimulation. However, it has to be stressed that the use of laser light in biomedical applications will be problematic due to tissue penetration limitations, while AMF overcomes this drawback. For that reason, laser light might be limited to on-skin and shallow treatments. For deeper-tissue procedures, more invasive techniques such as laparoscopy must be used.^85,86 Although foams are primarily designed for bone tissue replacement, other potential applications, such as wound-healing composites, remain viable.

It is of great importance to measure the ability of heat induction in an environment that resembles the implantation site. Therefore, the PLLA/HAP/Fe₃O₄ foams were soaked for 2 h in Ringer's solution (Table 3), and further characterization of energy conversion was performed (Fig. 12 and Fig. S6, ESI†). The interesting feature of all foams is their ability to achieve water-based solution loading of ≈440–500%, except for PLLA/HAP/Fe₃O₄, above 20%, when the water loading drops below 400% due to the microstructure degradation. Fig. S5 (ESI†) presents the heating curves recorded for the different concentrations of magnetic component (1–30%) under stimulation with the optimal AMF (336 kHz, 27 kA m⁻¹) parameters.

Table 3 Mass change of the PLLA/HAP/Fe₃O₄ foams before and after soaking with Ringer's solution

Sample	Foam mass (g)	Ringer's solution soaked foam (g)	Water (g)	Water loading (%)
PLLA/HAP/Fe3O4 1%	0.0318	0.1612	0.1294	506.9
PLLA/HAP/Fe3O4 5%	0.0229	0.1230	0.1001	537.1
PLLA/HAP/Fe3O4 10%	0.0382	0.1694	0.1311	443.2
PLLA/HAP/Fe3O4 15%	0.0373	0.1988	0.1615	533.0
PLLA/HAP/Fe3O4 20%	0.0465	0.1902	0.1437	409.2
PLLA/HAP/Fe3O4 30%	0.0592	0.1803	0.1211	304.6

---

Fig. 12 Heating curves of the PLLA/HAP/Fe₃O₄ 8% soaked foam measured for (a) 808 nm, (b) 880 nm, and (c) 1122 nm laser modules as a function of the laser power 300–600 mW.

It is clear that the presence of a high water loading (Ringer's solution) dramatically affects the heating ability due to the high specific heat capacity of water (4.184 J g⁻¹ °C⁻¹). In general, the values of the heating speed dT/dt (Table 4) decreased strongly, if directly compared to the dT/dt obtained for the PLLA/HAP/Fe₃O₄ 8% composite (dry foam – 2.49 °C s⁻¹, soaked foam – 0.07 °C s⁻¹). It also caused a two-fold decrease of the T_max from 54.1 °C to 27.8 °C. Since the AMF is a key stimulant for deep-tissue penetration, the bottom limit for samples that will allow achieving a biologically relevant temperature was a PLLA/HAP/Fe₃O₄ with 15% of magnetite (43.8 °C with dT/dt of 0.52 °C s⁻¹). In the case of the higher-doped foams, one can see a progressive increase in the T_max (up to 74 °C) and heating speed (0.52 to 1.39 °C s⁻¹). Again, for the evaluation of the efficacy of the heating, SAR was calculated for the Ringer's solution soaked foams (see Tables 4 and 5). However, since the composite contains a distinct amount of water (Table 3), the specific heat capacity (3.5 J g⁻¹ °C⁻¹) was calculated as a weight average, including water loading. Based on Table 3, the average content of water infiltrating the foams was taken to simplify calculations, since water loading is comparable for each foam type. The estimated SAR for the Ringer's solution loaded PLLA/HAP/Fe₃O₄ foams showed a clear dependence on the magnetic component content under the action of the AMF. At low Fe₃O₄ concentrations (1–5%) SAR was low (below 2 W g⁻¹) and thus magnetic heating is insufficient due to the too low magnetite content within the polymer matrix. The polymer starts to respond to the AMF at 8% of Fe₃O₄ (3.6 W g⁻¹), and it further shows a progressive SAR increase up to 22.4 W g⁻¹ for 20% and 16.2 W g⁻¹ for 30% Fe₃O₄ content. The observed increase of SAR with magnetite amount and further decrease above 20% arises from interparticle magnetic interactions, and changes in collective magnetic response within the composite.⁸⁷ However, as mentioned earlier, the degradation of the foam microstructure was already noticed above 20% of the Fe₃O₄. We believe that for the decline of SAR at 30% of magnetite nanoparticles in the polymer matrix, magnetic coupling-induced blocking can be responsible since close-packed particles may form clusters with magnetic moments no longer able to reorganize independently, suppressing Néel relaxation. Another factor contributing to such behavior can be associated with the reduced anisotropy, as densely packed domains can experience increased magnetic stiffness, thereby limiting their ability to respond dynamically to the AMF. It cannot be excluded that heat generated by packed particles may dissipate less efficiently, especially in the presence of water within the pores of the foam, which reduces the SAR. Taking all that into account, only the hybrid material with a maximum of 15% of magnetite was considered for further biological evaluation. The SAR for that foam composite was 12.1 W g⁻¹, which is enough to achieve a relevant biological temperature range.

Table 4 AMF conversion parameters, i.e., heating speed dT/dt and maximum temperature (T_max) for the PLLA/HAP/Fe₃O₄ Ringer's solution soaked foams – concentration dependence

Sample	dT/dt (°C s^−1)	T max (°C)	SAR (W g^−1)
AMF 336 kHz. 27 kA m^−1
PLLA/HAP/Fe3O4 1%	0.01	22.2	3.5
PLLA/HAP/Fe3O4 2%	0.01	22.9	1.75
PLLA/HAP/Fe3O4 5%	0.025	23.7	1.75
PLLA/HAP/Fe3O4 8%	0.07	27.8	3.6
PLLA/HAP/Fe3O4 10%	0.22	35.5	7.7
PLLA/HAP/Fe3O4 15%	0.52	43.8	12.1
PLLA/HAP/Fe3O4 20%	1.28	58.8	22.4
PLLA/HAP/Fe3O4 30%	1.39	74.7	16.2

---

Table 5 NIR conversion parameters, i.e., heating speed dT/dt and maximum temperature (T_max) for the PLLA/HAP/Fe₃O₄ 8% soaked foam

Sample	Parameters (NIR)	dT/dt (°C s^−1)	T max (°C)	SAR (W g^−1)
Laser 808 nm
PLLA/HAP/Fe3O4 8%	300 mW	0.64	34.5	28.0
PLLA/HAP/Fe3O4 8%	450 mW	1.59	40.8	68.6
PLLA/HAP/Fe3O4 8%	600 mW	2.32	45.2	101.5
Laser 880 nm
PLLA/HAP/Fe3O4 8%	300 mW	0.46	33.2	20.1
PLLA/HAP/Fe3O4 8%	450 mW	0.42	34.7	18.4
PLLA/HAP/Fe3O4 8%	600 mW	1.04	40.2	45.5
Laser 1122 nm
PLLA/HAP/Fe3O4 8%	300 mW	0.36	33.7	15.6
PLLA/HAP/Fe3O4 8%	450 mW	0.39	38.6	17.1
PLLA/HAP/Fe3O4 8%	600 mW	1.35	43.1	59.1

---

To evaluate the effect of laser light stimulation, a measurement of the laser power dependence within the 300–600 mW range was necessary. It was anticipated that the presence of water molecules in the soaked foams would significantly affect heating efficiency across all three wavelengths (Fig. 12). To ensure a proper comparison, heat induction was analysed using PLLA/HAP/Fe₃O₄ foams containing 8% magnetite. The final data are summarized in Table 5. It is evident that the high heat capacity of the solvent strongly influences energy conversion. This phenomenon was reflected in a drastic change of dT/dt values, which ranged from 0.36 to 2.32 °C s⁻¹ for the soaked composites (dry 4.32–25.27 °C s⁻¹, T_max 43–97 °C), along with a reduced T_max of 33–45 °C at 8% Fe₃O₄ doping.

The calculated SAR values for the hydrated PLLA/HAP/Fe₃O₄ 8% composite foams were strongly dependent on both laser power and wavelength. The highest SAR was recorded under 808 nm irradiation, reaching 101.5 W g⁻¹ Fe₃O₄ at 600 mW, which corresponds to a laser optical density (LOD) of 0.63 W cm⁻². This value is nearly three times lower than that of the same foam in its dry state, a difference primarily attributed to the presence of water (high specific capacity, 4.184 J g⁻¹ °C⁻¹) filling the foam's pores, increasing the system's overall heat capacity. In addition, the enhanced photothermal efficiency observed at 808 nm is consistent with the relatively strong absorption of Fe₃O₄. In contrast, the other two NIR wavelengths (880 nm and 1122 nm) yielded lower SAR values of approximately 45 and 59 W g⁻¹, respectively. This SAR decrease can be attributed to a combination of factors, including larger laser spot area (resulting in lower LODs) and wavelength-dependent optical properties. Although the 1122 nm wavelength lies within the spectral range where Fe₃O₄ extinction coefficient begins to rise, the contribution of optical scattering becomes more prominent at longer wavelengths, which may limit photothermal conversion efficiency compared to the <900 nm range, where absorption dominates. Overall, the results confirm the wavelength-dependent nature of photothermal conversion and support the use of 808 nm laser light as an optimal NIR source for non-invasive heating applications in such composites.

Notably, NIR stimulation is significantly more effective than AMF in terms of both heating speed and maximum temperature. This feature is crucial to avoid the induction of cell thermotolerance in hyperthermia applications.⁸⁸ The direct comparison of the dry and soaked foams during the action of the NIR light (808, 880, and 1122 nm) is shown in Fig. 13. The concentration dependence of cubic magnetite nanoparticles in PLLA/HAP/Fe₃O₄ foams was measured under synergy mode. It involved, referred to here as dual mode, a combined action of AMF and NIR laser light (AMF: 336 kHz, 27 kA m⁻¹; NIR: 600 mW at 808, 880, and 1122 nm) for soaked ternary composites (Fig. S6, ESI†). For the PLLA/HAP/Fe₃O₄ sample with 8% magnetite, the maximum temperature (T_max) reached was 49 °C (AMF: 27.8 °C, NIR₈₀₈: 45 °C, NIR₈₈₀: 40 °C, NIR₁₁₂₂: 43 °C), demonstrating that the synergistic interaction of both stimuli leads to a further temperature increase.

Fig. 13 Comparison of the dT/dt and T_max values for the dry, and Ringer's solution soaked foams with 8% of Fe₃O₄ under NIR stimulation (open symbols – dry, full-colored – soaked foams).

As shown, the T_max obtained for all foams under dual mode was significantly higher than that achieved with individual stimulation. The highest recorded T_max was observed for PLLA/HAP/Fe₃O₄ foams containing 30% magnetite, reaching 84 °C (NIR₈₀₈), 80 °C (NIR₈₈₀), and 82 °C (NIR₁₁₂₂). Although the composites containing 20 and 30 wt% of Fe₃O₄ showed the most efficient heating response, the collapse of the foam microstructure excludes them from potential biomedical applications. Nevertheless, the addition of the high-loading samples (20% and 30%) was essential to comprehensively understand the effect of the magnetite content on the overall foam energy conversion performance. These samples allowed us to assess how increased magnetic nanoparticle content affects not only the heating ability but also the density, porosity, mechanical integrity, and microstructure stability of the composite foams. This systematic approach enabled the identification of an optimal filler concentration that balances magnetic responsiveness with structural requirements for future biomedical use. Fig. 14 shows a comparison of different stimulation modes for PLLA/HAP/Fe₃O₄ with 15% magnetite, highlighting striking differences between the various energy conversion mechanisms. Notably, this sample could reach biologically relevant temperature ranges using AMF, NIR, and their synergistic combination.

Fig. 14 Comparison of the heating generation using different stimuli AMF, and NIR separately and under synergy mode (dual) for different laser wavelengths (808, 880, 1122 nm) on soaked PLLA/HAP/Fe₃O₄ 15% foam.

3.3. Assessment of biological interactions with engineered biomaterials

Our studies have shown that extracts from biomaterials composed solely of HAP and PLLA exhibited some cytotoxicity towards hFOB 1.19 cells (Fig. 15). According to ISO 10993-5, biomedical materials are considered cytotoxic when the percentage of viable cells falls below 70%.⁸⁹ MTT assay results after 24 h of exposure to extracts from these materials indicated that only extracts containing 15% of Fe₃O₄ nanoparticles did not exhibit cytotoxicity achieving a cell viability of 76.6% ± 2.38. After 48 h of exposure, it was observed that biomaterials containing 10% of Fe₃O₄ also showed no cytotoxicity with a cell viability of 80.2% ± 1.2 while those with a 15% addition achieved an even higher viability level of 91% ± 3.2 (Fig. 15a and b). Therefore, Fe₃O₄ nanoparticle integration into the PLLA/HAP composites enhanced their biological safety in a concentration-dependent manner. We believe that the PLLA/HAP cytotoxicity might be due to the high affinity of metal cations towards the porous material surface. This phenomenon was observed by Fijoł et al.,⁹⁰ showing enhanced cation chemisorption on the 3D PLLA/HAP (15% of HAP) scaffolds after incorporation of small apatite nanoparticles with extended surface area.

Fig. 15 Cytotoxicity of the extracts derived from the biomaterials as well as from polystyrene (PS), which served as the control extract: (a) and (b) after 24 and 48 h of incubation, hFOB cell viability (MTT assay). Statistically significant differences were observed compared to the control extract (two-way ANOVA test followed by a Bonferroni comparison test P < 0.05); (c) and (d) images obtained using a confocal laser scanning microscope presenting live/dead staining of cells after exposure to extracts from the tested biomaterials (green fluorescence – live cells, red fluorescence – dead cells). Magnification 100×, scale bar 100 μm.

Similar behavior was reported by Klimek et al.⁹¹ on high specific area HAP ceramics that caused significant false-toxicity due to the depletion of the medium from the important ions through a strong adsorption to the surface of HAP (35 nm particles). They proposed a modified protocol for the estimation of biomaterial cytotoxicity. However, this approach was not used by us, since the most valuable composite, from the biological point of view, was PLLA/HAP/Fe₃O₄ 15% foam that showed no effect on hFOB cells.

The live/dead staining confirmed the results of the MTT assay, clearly demonstrating a significantly higher number of viable cells after exposure to extracts from the tested materials containing 10% and 15% of Fe₃O₄ nanoparticles (Fig. 15c and d). Moreover, the cells showed normal morphology. Both the MTT assay and live/dead staining indicated that biomaterials with these concentrations of Fe₃O₄ exhibited an increased biocompatibility, which was manifested as reduced cytotoxicity. A similar trend was observed in the test of direct biomaterial–cell contact. After 48 h of cell incubation on the biomaterials, microscopic observations revealed that hFOB cells had populated the surfaces of the biomaterials containing the highest concentrations of Fe₃O₄ (10% and 15% of Fe₃O₄), confirming the non-cytotoxicity of the materials (Fig. 16). The hFOB cells were alive (green fluorescence) and exhibited a flattened morphology, which suggested good cell adhesion to the biomaterials. It is worth noting that in the other material variants, the cells were also alive (green fluorescence), and no dead cells were observed (absence of red fluorescence).

Fig. 16 Viability of human osteoblasts (hFOB 1.19 cells) cultured on the biomaterials. Live cells – green fluorescence; dead cells – red fluorescence. Magnification: 100×, scale bar: 150 μm.

However, they displayed a round shape and abnormal morphology. These observations indicated that the Fe₃O₄ addition may positively modify the surfaces of the biomaterials, promoting cell adhesion. The conducted cell proliferation assay demonstrated that extracts from the developed biomaterials did not negatively affect the proliferation of hFOB cells. Both on days 3 and 5, the number of cells incubated with the extracts was comparable to the control cells (those incubated in 10% culture medium) (Fig. 17a). The CLSM staining revealed that with an increase in Fe₃O₄ concentration, the number of cells also increased (Fig. 17b).

Fig. 17 Quantitative (a) and qualitative (b) representation of the cell proliferation assay results. CLSM images, 200× magnification, scale bar 100 μm.

The in vitro biocompatibility studies conducted with human osteoblasts revealed that the PLLA/HAP/Fe₃O₄ 15% biomaterial most effectively supported both cell viability and cell proliferation, suggesting its potential for medical applications. Notably, the results indicated that cells adapted to the conditions created by the biomaterials over time. This is supported by the viability assay, which showed reduced cytotoxicity after 48 h compared to 24 h (in biomaterials containing 5%, 10%, and 15% of Fe₃O₄), and the proliferation assay, which demonstrated no negative impact of the biomaterials on cell growth. Furthermore, it is worth emphasizing that regardless of the in vitro biocompatibility test performed, all cells observed via CLSM were viable. While the PLLA/HAP/Fe₃O₄ 15% foam undoubtedly appeared to be the most promising, the potential utility of the other biomaterials cannot be definitively excluded.

Although our study demonstrated the physicochemical properties and heating behaviour of PLLA/HAP/Fe₃O₄ foams, biological validation remains a necessary next step. Future work will include in vivo evaluation of selected foam composites under NIR and AMF stimulation to assess their effectiveness in bone regeneration and tissue integration. These studies will help translate the proof-of-concept demonstrated here into functional biomedical applications.

4. Conclusions

Foam PLLA/HAP/Fe₃O₄ composites with a wide range of magnetic component concentrations (1–30%) were successfully fabricated using the TIPS technique, supplemented by the salt leaching process (SLP). Calcium hydroxyapatite and Fe₃O₄ nanoparticles, with primary sizes of 20 nm and 80 nm, were synthesized through wet chemical methods, specifically co-precipitation (HAP), and thermal decomposition (Fe₃O₄), and subsequently incorporated into the PLLA matrix. The density of the hybrid materials ranged from 0.02 to 0.079 g mL⁻¹, depending on the inorganic content, while their porosity varied between 93% and 98% with a pore size of 330 μm. A decrease in the PLLA polymer decomposition temperature was observed due to the presence of Fe₃O₄. Structural collapse of the porous microstructure occurred upon doping with 20% Fe₃O₄ (70% of fillers) as evidenced by the SEM and mechanical properties evaluation. Upon material characterization, the PLLA/HAP/Fe₃O₄ with 15% of the magnetic component was chosen as the most reliable sample in terms of its microstructure and achievable T_max (AMF – 43.8 °C, laser 45 °C, dual mode – 58 °C for soaked hybrids), as well as sufficient biocompatibility and cell adhesion toward the hFOB line. Although higher heating effects were observed for the 20% and 30% Fe₃O₄ content in PLLA/HAP/Fe₃O₄ foams, their microstructure collapse excluded them from consideration as a potential bone replacement material.

Energy conversion under alternating magnetic field (AMF) and near-infrared (NIR) irradiation (808, 880, and 1122 nm) was thoroughly investigated in both separate and synergistic modes. For AMF stimulation, the optimized parameters for the most efficient heat generation were identified as 336 kHz and 27 kA m⁻¹. The specific absorption rate (SAR) values obtained for the dry composites were approximately three times higher than those in Ringer's solution soaked foams, due to the high specific heat capacity of water. Importantly, safety limitations must be considered when applying AMF and NIR. For AMF, it is generally recommended that the product of field intensity and frequency (f·H) does not exceed 4.85 × 10⁸ A ms⁻¹ (Atkinson–Brezovich limit) or 5 × 10⁹ A ms⁻¹ (Hergt limit).^33,92 In this study, the f·H value slightly exceeded the Hergt limit (calculated as 9.072 × 10⁹ A ms⁻¹). While a lower frequency might have been preferable, our equipment did not allow fine-tuning of the field frequency. Our results demonstrated a high responsiveness of the foams, evidenced by rapid heating and the ability to reach biologically relevant temperatures in both dry and Ringer's solution soaked composite foams. In bio-related applications, the generally accepted safe limit for near-infrared (NIR) laser irradiation is between 0.3 and 0.5 W cm⁻². However, higher optical power densities may be tolerated for shorter exposure durations, depending on the biological system involved.³⁰ In our experiments, considering the applied laser power range (100–600 mW) and the spot area, the effective optical density ranged from 0.1–0.63 W cm⁻², depending on the specific laser wavelength and power. In the literature there is still ongoing debate regarding the estimation of adverse biological effects under dual stimulation (NIR and AMF), as these effects depend on multiple variables such as: exposure time (short, long), irradiated area (internal, external), cell type (different thermotolerance), and biological context of application (diathermia, hyperthermia, laser ablation, on-skin therapies, stimulation of regenerative processes etc.). Therefore, additional in vitro and in vivo studies are necessary to clarify safety margins. Our data showed that under certain stimulation parameters, the foam temperature can exceed 43 °C, which suggests that there is space to reduce NIR and AMF parameters to better conform to established safety guidelines. Moreover, as we underlined, the intended application of the foams is for implantable, stimuli-responsive materials in regenerative medicine (not for cancer treatment). Therefore, maintaining precise hyperthermia thresholds (e.g., 43 °C) may not always be critical to therapeutic success.

Author contributions

E. Z. preparation of foams, density, porosity, TGA, DSC, manuscript writing and editing. A. T. synthesis of Fe₃O₄ and HAP nanoparticles, XRD and FTIR measurements, data analysis, and manuscript editing. M. K.-G. TEM imaging, energy conversion measurements, and manuscript editing. P. K. SEM characterization, data analysis, and manuscript editing. J.-M. N. data analysis, manuscript writing, and editing. D. Z. and Š. H. magnetic and Mössbauer measurements, data analysis, manuscript writing and editing, A. B.-R. and A. N. biological characterization of foams, data analysis, manuscript writing and editing. R. P. idea, conceptualization, data analysis, energy conversion analysis, manuscript writing, and editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

All data are accessible from the corresponding authors upon request.

Acknowledgements

Financial support of PCI (Podkarpackie Centrum Innowacji) within project N3_032 ‘Multifunctional foams for temperature stimulated regenerative process of bone tissue’ is gratefully acknowledged. R. P. was a holder of the SSHN scholarship funded by the French Government. Financial assistance was partially provided by Ministry of Science and Higher Education in Poland within the DS6 project of Medical University of Lublin. D. Z. and Š. H. acknowledge the assistance provided by Advanced Multiscale Materials for Key Enabling Technologies project, supported by the Ministry of Education, Youth, and Sports of the Czech Republic Project No. CZ.02.01.01/00/22_008/0004558, co-funded by the European Union. D. Z. is also financially supported by Charles University Research Centre program no. UNCE/24/SCI/010.

References

1. Z. Zhou, W. Feng, B. K. Moghadas, N. Baneshi, B. Noshadi, S. Baghaei and D. A. Dehkordi, Tissue Cell, 2024, 88, 102390.
2. R. Sun, H. Chen, L. Sutrisno, N. Kawazoe and G. Chen, Sci. Technol. Adv. Mater, 2021, 22, 404–428.
3. H. Sadeghzadeh, H. Dianat-Moghadam, A. R. Del Bakhshayesh, D. Mohammadnejad and A. Mehdipour, Stem Cell Res. Ther., 2023, 14, 194.
4. A. Kirillova, T. R. Yeazel, D. Asheghali, S. R. Petersen, S. Dort, K. Gall and M. L. Becker, Chem. Rev., 2021, 121, 11238–11304.
5. X. Sheng, C. Li, Z. Wang, Y. Xu, Y. Sun, W. Zhang, H. Liu and J. Wang, Mater. Today Bio, 2023, 20, 100636.
6. L. Borkowski, T. Lübek, M. Jojczuk, A. Nogalski, A. Belcarz, K. Palka, M. Hajnos and G. Ginalska, J. Biomed. Mater. Res., Part B, 2018, 106, 2653–2664.
7. D. Iannazzo, A. Pistone, M. Salamò and S. Galvagno, in Hybrid Polymer Composite Materials, ed. V. K. Thakur, M. J. Thakur and R. K. Gupta, Elsevier B.V, 2017, pp. 125–155.
8. A. Pistone, D. Iannazzo, C. Celesti, E. Piperopoulos, D. Ashok, A. Cembran, A. Tricoli and D. Nisbet, Materials, 2019, 12, 2321.
9. K. Yamaguchi, T. Hirano, G. Yoshida and K. Iwasaki, Biomaterials, 1995, 16, 983–985.
10. M. Ebrahimi, M. G. Botelho and S. V. Dorozhkin, Mater. Sci. Eng., C, 2017, 71, 1293–1312.
11. H. Oonishi Jr., H. Oonishi, H. Ohashi, I. Kawahara, Y. Hanaoka and L. L. Hench, in Advances in Calcium Phosphate Biomaterials, ed. B. Ben-Nissan, Springer, Verlag Berlin Heidelberg, 2014, pp. 19–49.
12. M. Ansari, Prog. Biomater., 2019, 8, 223–237.
13. Y. Sugiura, F. Ono, M. Nohara, M. Funabiki, K. Kutara, T. Kanda, E. Yamada and M. Horie, Regener. Ther., 2024, 26, 760–766.
14. E. Cunningham, N. Dunne, G. Walker, C. Maggs, R. Wilcox and F. Buchanan, J. Mater. Sci.: Mater. Med., 2010, 21, 2255–2261.
15. S. Tan, X. Cai, W. Li, W. Liu, Q. Liu, Y. Liu, F. Zhang, S. Guo, Q. Wang and C. Li, ACS Appl. Polym. Mater., 2024, 6, 2232–2242.
16. A. Bansiddhi and D. C. Dunand, in Bone Substitute Biomaterials, ed. K. Mallick, Woodhead Publishing, 2014, pp. 142–179.
17. Q. Z. Chen, Bubble Sci., Eng., Technol., 2011, 3, 34–47.
18. E. M. Materón, C. M. Miyazaki, O. Carr, N. Joshi, P. H. S. Picciani, C. J. Dalmaschio, F. Davis and F. M. Shimizu, Appl. Surf. Sci. Adv., 2021, 6, 100163.
19. L. Gao and X. Yan, in Advances in Experimental Medicine and Biology, ed. S. Perret, A. K. Buell and T. P. J. Knowles, Springer, Singapore, 2019, vol. 1174, pp. 291–312.
20. S. Mondal, T. P. Nguyen, V. H. Pham, G. Hoang, P. Manivasagan, M. H. Kim, S. Y. Nam and J. Oh, Ceram. Int., 2020, 46, 3443–3455.
21. C. Schugens, V. Maquet, C. Grandfils, R. Jerome and P. Teyssie, Polymer, 1996, 37, 1027–1038.
22. A. Salerno and P. A. Nettis, Biomedical Foams for Tissue Engineering Applications, Elsevier Ltd, 2014, pp. 71–100.
23. R. Banerjee, A. Gebrekrstos, J. T. Orasugh and S. S. Ray, Ind. Eng. Chem. Res., 2023, 62, 6807–6842.
24. M. Ghorbandoust, M. Fasihi and R. Norouzbeigi, Sci. Rep., 2024, 14, 12395.
25. J. F. Kim, J. H. Kim, Y. M. Lee and E. Drioli, AIChE J., 2016, 62, 461–490.
26. F. Carfì Pavia, V. La Carrubba, S. Piccarolo and V. Brucato, J. Biomed. Mater. Res., Part A, 2008, 86, 459–466.
27. M. S. Mohammadi, E. Rezabeigi, J. Bertram, B. Marelli, R. Gendron, S. N. Nazhat and M. N. Bureau, Polymers, 2020, 12, 231.
28. H. Gavilán, S. K. Avugadda, T. Fernández-Cabada, N. Soni, M. Cassani, B. T. Mai, R. Chantrell and T. Pellegrino, Chem. Soc. Rev., 2021, 50, 11614–11667.
29. A. Espinosa, J. Kolosnjaj-Tabi, A. Abou-Hassan, A. Plan Sangnier, A. Curcio, A. K. A. Silva, R. Di Corato, S. Neveu, T. Pellegrino, L. M. Liz-Marzán and C. Wilhelm, Adv. Funct. Mater., 2018, 28, 1–16.
30. D. Jaque, L. Martínez Maestro, B. del Rosal, P. Haro-Gonzalez, A. Benayas, J. L. Plaza, E. Martín Rodríguez and J. García Solé, Nanoscale, 2014, 6, 9494.
31. L. Zhang, Q. Li, J. Liu, Z. Deng, X. Zhang, N. Alifu, X. Zhang, Z. Yu, Y. Liu, Z. Lan, T. Wen and K. Sun, Colloids Surf., B, 2024, 234, 113754.
32. D. H. Ortgies, F. J. Teran, U. Rocha, L. de la Cueva, G. Salas, D. Cabrera, A. S. Vanetsev, M. Rähn, V. Sammelselg, Y. V. Orlovskii and D. Jaque, Adv. Funct. Mater., 2018, 28, 1–11.
33. R. Hergt and S. Dutz, J. Magn. Magn. Mater., 2007, 311, 187–192.
34. A. Espinosa, M. Bugnet, G. Radtke, S. Neveu, G. A. Botton, C. Wilhelm and A. Abou-Hassan, Nanoscale, 2015, 7, 18872–18877.
35. D. Chang, M. Lim, J. A. C. M. Goos, R. Qiao, Y. Y. Ng, F. M. Mansfeld, M. Jackson, T. P. Davis and M. Kavallaris, Front. Pharmacol., 2018, 9, 831.
36. E. Díaz, M. B. Valle, S. Ribeiro, S. Lanceros-Mendez and J. M. Barandiarán, Int. J. Mol. Sci., 2019, 20, 4664.
37. S. Bin, A. Wang, W. Guo, L. Yu and P. Feng, Polymers, 2020, 12, 2045.
38. W. Y. Lai, S. W. Feng, Y. H. Chan, W. J. Chang, H. T. Wang and H. M. Huang, Polymers, 2018, 10, 804.
39. W. Zhao, Z. Huang, L. Liu, W. Wang, J. Leng and Y. Liu, Compos. Sci. Technol., 2021, 203, 108563.
40. Y. H. Pan, H. T. Wang, T. L. Wu, K. H. Fan, H. M. Huang and W. J. Chang, Polym. Compos., 2017, 38, 2881–2888.
41. J. Huang, L. Cao, D. Yuan and Y. Chen, ACS Sustainable Chem. Eng., 2019, 7, 2304–2315.
42. D. Shan, Y. Shi, S. Duan, Y. Wei, Q. Cai and X. Yang, Mater. Sci. Eng., C, 2013, 33, 3498–3505.
43. C. Shuai, W. Yang, C. He, S. Peng, C. Gao, Y. Yang, F. Qi and P. Feng, Mater. Des., 2020, 185, 108275.
44. W. Yang, Y. Zhong, P. Feng, C. Gao, S. Peng, Z. Zhao and C. Shuai, Polym. Test., 2019, 76, 33–42.
45. S. Sprio, S. Panseri, A. Adamiano, M. Sandri, M. Uhlarz, T. Herrmannsdorfer, E. Landi, Y. Pineiro-Remondo and A. Tampieri, Front. Nanosci. Nanotechnol, 2017, 3, 1–9.
46. T. Iwasaki, R. Nakatsuka, K. Murase, H. Takata, H. Nakamura and S. Watano, Int. J. Mol. Sci., 2013, 14, 9365–9378.
47. A. Gloria, T. Russo, U. D’amora, S. Zeppetelli, T. D’Alessandro, M. Sandri, M. Bañobre-López, Y. Piñeiro-Redondo, M. Uhlarz, A. Tampieri, J. Rivas, T. Herrmannsdörfer, V. A. Dediu, L. Ambrosio and R. De Santis, J. R.Soc., Interface, 2013, 10, 20120833.
48. E. Andronescu, M. Ficai, G. Voicu, D. Ficai, M. Maganu and A. Ficai, J. Mater. Sci.: Mater. Med., 2010, 21, 2237–2242.
49. A. V. Chaves, R. M. Freire, V. P. Feitosa, N. M. P. S. Ricardo, J. C. Denardin, D. M. A. Neto and P. B. A. Fechine, J. Compos. Sci., 2021, 5, 37.
50. S. Torgbo and P. Sukyai, Mater. Chem. Phys., 2019, 237, 121868.
51. S. Wang, F. Cheng, Q. Ji, M. Song, Z. Wu, Y. Zhang, Z. Ji, H. Feng, J. C. I. Belmonte, Q. Zhou, J. Qu, W. Li, G. H. Liu and W. Zhang, Protein Cell, 2022, 13, 615–622.
52. S. Marozin, B. Simon-Nobbe, S. Irausek, L. W. K. Chung and G. Lepperdinger, Sci. Rep., 2021, 11, 10933.
53. M. Yen, C.-C. Chien, I. Chiu, H.-I. Huang, Y.-C. Chen, H.-I. Hu and B. L. Yen, Stem Cells, 2007, 25, 125–131.
54. D. Kim, N. Lee, M. Park, B. H. Kim, K. An and T. Hyeon, J. Am. Chem. Soc., 2009, 131, 454–455.
55. L. M. Rodriguez-Lorenzo, M. Vallet-Regi and J. M. F. Ferreira, Biomaterials, 2001, 22, 583–588.
56. R. J. Wiglusz, B. Pozniak, K. Zawisza and R. Pazik, RSC Adv., 2015, 5, 27610–27622.
57. I. Salma, G. Salms, M. Pilmane, D. Loca, J. Locs, N. Romanchikova, A. Skagers and L. Berzina Cimdina, J. Oral Surg., 2011, 40, 1216.
58. K. Szustakiewicz, M. Gazińska, B. Kryszak, M. Grzymajło, J. Pigłowski, R. J. Wiglusz and M. Okamoto, Eur. Polym. J., 2019, 113, 313–320.
59. L. Su, J. Zou, S. Dong, N. Hao and H. Xu, RSC Adv., 2017, 7, 55364–55370.
60. S. R. Laraba, N. Ullah, A. Bouamer, A. Ullah, T. Aziz, W. Luo, W. Djerir, Q. A. Zahra, A. Rezzoug, J. Wei and Y. Li, Molecules, 2023, 28, 6442.
61. Y. Liang, L. Jiang, S. Xu, W. Ju, Z. Tao, Y. Yang, X. Peng and G. Wei, J. Mater. Eng. Perform., 2024, 33, 6804–6815.
62. N. K. Nguyen, M. Leoni, D. Maniglio and C. Migliaresi, J. Biomater. Appl., 2013, 28, 49–61.
63. B. Xue, L. Xie and J. Zhang, RSC Adv., 2017, 7, 47017–47028.
64. J. Yao, F. Luo, J. Mao, Y. Li, Y. Liu and X. Sun, Polym. Test., 2021, 100, 107244.
65. R. Androsch, C. Schick and M. L. Di Lorenzo, Macromol. Chem. Phys., 2014, 215, 1134–1139.
66. J. Muro-Cruces, A. G. Roca, A. López-Ortega, E. Fantechi, D. Del-Pozo-Bueno, S. Estradé, F. Peiró, B. Sepúlveda, F. Pineider, C. Sangregorio and J. Nogues, ACS Nano, 2019, 13, 7716–7728.
67. H. F. Schmidts and H. Kronmüller, J. Magn. Magn. Mater., 1994, 130, 329–341.
68. L. Rebbouh, R. P. Hermann, F. Grandjean, T. Hyeon, K. An, A. Amato and G. J. Long, Phys. Rev. B:Condens. Matter Mater. Phys., 2007, 76, 174422.
69. F. Walz, J. Phys.: Condens. Matter, 2002, 14, 285–340.
70. J. Winsett, A. Moilanen, K. Paudel, S. Kamali, K. Ding, W. Cribb, D. Seifu and S. Neupane, SN Appl. Sci., 2019, 1, 1636.
71. G. K. Rozenberg, M. P. Pasternak, W. M. Xu, Y. Amiel, M. Hanfland, M. Amboage, R. D. Taylor and R. Jeanloz, Phys. Rev. Lett., 2006, 96, 045705.
72. M. Kulpa-Greszta, A. Tomaszewska, A. Michalicha, D. Sikora, A. Dziedzic, R. Wojnarowska-Nowak, A. Belcarz and R. Pązik, RSC Adv., 2022, 12, 27396–27410.
73. J. Tang, M. Myers, K. A. Bosnick and L. E. Brus, J. Phys. Chem. B, 2003, 107, 7501–7506.
74. S. M. Devi, A. Nivetha and I. Prabha, J. Supercond. Novel Magn., 2020, 33, 3893–3901.
75. C. K. Chou, Int. J. Hyperthermia, 1990, 6, 367–370.
76. I. Baker, Q. Zeng, W. Li and C. R. Sullivan, J. Appl. Phys., 2006, 99, 97–100.
77. M. Vasilakaki, C. Binns and K. N. Trohidou, Nanoscale, 2015, 7, 7753–7762.
78. E. Lima, E. De Biasi, M. V. Mansilla, M. E. Saleta, M. Granada, H. E. Troiani, F. B. Effenberger, L. M. Rossi, H. R. Rechenberg and R. D. Zysler, J. Phys. D: Appl. Phys., 2013, 46, 045002.
79. C. Blanco-Andujar, A. Walter, G. Cotin, C. Bordeianu, D. Mertz, D. Felder-Flesch and S. Begin-Colin, Nanomedicine, 2016, 11, 1889–1910.
80. A. Sathya, P. Guardia, R. Brescia, N. Silvestri, G. Pugliese, S. Nitti, L. Manna and T. Pellegrino, Chem. Mater., 2016, 28, 1769–1780.
81. H. Kakwere, M. P. Leal, M. E. Materia, A. Curcio, P. Guardia, D. Niculaes, R. Marotta, A. Falqui and T. Pellegrino, ACS Appl. Mater. Interfaces, 2015, 7, 10132–10145.
82. D. Cabrera, A. Lak, T. Yoshida, M. E. Materia, D. Ortega, F. Ludwig, P. Guardia, A. Sathya, T. Pellegrino and F. J. Teran, Nanoscale, 2018, 9, 5094–5101.
83. M. V. Zyuzin, M. Cassani, M. J. Barthel, H. Gavilan, N. Silvestri, A. Escudero, A. Scarpellini, F. Lucchesi, F. J. Teran, W. J. Parak and T. Pellegrino, ACS Appl. Mater. Interfaces, 2019, 11, 41957–41971.
84. M. Levy, A. Quarta, A. Espinosa, A. Figuerola, C. Wilhelm, M. García-Hernández, A. Genovese, A. Falqui, D. Alloyeau, R. Buonsanti, P. D. Cozzoli, M. A. García, F. Gazeau and T. Pellegrino, Chem. Mater., 2011, 23, 4170–4180.
85. P. Southern and Q. A. Pankhurst, Int. J. Hyperthermia, 2018, 34, 671–686.
86. Q. A. Pankhurst, J. Connolly, S. K. Jones and J. Dobson, J. Phys. D: Appl. Phys., 2003, 36, R167–R181.
87. R. Fu, Y. Yan, C. Roberts, Z. Liu and Y. Chen, Sci. Rep., 2018, 8, 4704.
88. J. Landry, P. Chretien, D. Bernier, L. S. Nicole, N. Marceau and R. M. Tanguay, Int. J. Radiat. Oncol. Biol. Phys., 1982, 8, 59–62.
89. ISO 10993-5:2009, Biological evaluation of medical devices— Part 5: Tests for in vitro cytotoxicity, International Standard: Geneva, Switzerland.
90. N. Fijoł, H. N. Abdelhamid, B. Pillai, S. A. Hall, N. Thomas and A. P. Mathew, RSC Adv., 2021, 11, 32408–32418.
91. K. Klimek, A. Belcarz, R. Pazik, P. Sobierajska, T. Han, R. J. Wiglusz and G. Ginalska, Mater. Sci. Eng., C, 2016, 65, 70–79.
92. S. Dutz and R. Hergt, Int. J. Hyperthermia, 2013, 29, 790–800.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5tb00998g

---