Synergistic photothermal and magnetothermal heat generation with Sc-doped iron oxide nanoflakes: a platform for potential bimodal tumor therapy

Abstract

Enhancement of specific absorption rate (SAR) of iron oxide (Fe₃O₄) is crucial for ensuring selectivity of hyperthermia tumor therapy, yet both magnetothermal and photothermal approaches endure shortcomings i.e., high dosage and laser power densities, that compromise therapeutic efficacy. This work reports the Scandium (Sc) doped Fe₃O₄ nanoflakes synthesized by sol-gel route with superior heat generation properties enabling bimodal tumor therapy. The novel nanoflakes superstructure exhibits pronounced optical extinction and a high saturation magnetization of 90.33 emu/g, arising from lattice expansion and enhanced magnetic exchange coupling. Photothermal conversion efficiency reached 66.84%, while SAR value under magnetic field (32 kA/m, 100 kHz) peaked at 1311.24 W/g (intrinsic loss power (ILP): 11.34 nHm²/kg). Under 808 nm laser exposure (2 W/cm²), photothermal SAR reached to 3480.84 W/g and with combined bimodal (LASER + AMF) heating the SAR synergically boosted to 11857.53 W/g due to temperature modulated magnetic and optical relaxation supporting each other. The material exhibits excellent biocompatibility and in vitro tumor ablation with A375 cell line using small (1 mg/mL) dosage under minimal magnetic field and laser parameters.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s40580-026-00558-w.

Introduction

Cancer persists to be a serious health hazard due to its high mortality and demands novel therapeutic modalities capable of improving treatment specificity thereby mitigating the side effects [1, 2]. Nanoparticles (NPs) can serve as heating nanocarriers for the localized hyperthermia (42–46 °C) by means of an external stimulus such as, X-rays, light, magnetic field etc. [3–5]. Currently, Hyperthermia tumor therapy is being utilized either by activation through alternating magnetic field (AMF), due to its deep tissues penetrability [6] or by Infrared (IR) laser, where transparency window (700–1000 nm) of biological tissues lies [7]. Fe₃O₄ NPs has been widely studied as hyperthermia agents both for magnetic hyperthermia therapy (MHT) and photothermal therapy (PTT), due to good biocompatibility, magnetic, and optical properties [8–10]. However, both magnetic and photothermal approaches endure several limitations, that restrict their extensive clinical adaption.

To avoid long-term health risks, the dosage of NPs should be kept minimal for MHT [11], although current MHT still requires high concentrations (2–10 mg/mL) of NPs [12, 13]. Therefore, the substantial enhancement in heating efficiency, measures in terms of specific absorption rate (SAR) under clinical safety limit (H×f < 9.59 × 10⁹ Am⁻¹s⁻¹) [14] is necessary. For PTT, although the NPs dosage is significantly reduced but requires the higher laser power densities (2–5 W/cm²) [15], far exceeding the biological safe limit (i.e., 0.33 W/cm², 808 nm) [16], due to lower extinction and photothermal conversion efficiency and also some NPs (e.g., carbon-based and metallic) could be bio-accumulative and toxic [17].

The potential of using bimodal heating by employing MHT and PTT, can bring cumulative enhancement in heating efficiency for hyperthermia tumor therapy. However, combined MHT and PTT is generally explored in conventional plasmonic-magneto hybrid structures e.g., core-shell magnetic NPs and semiconductor quantum dots (Fe₃O₄–PbS/CdS) [18], noble metals and magnetic NPs hybrids (Au–Fe₃O₄, Ag/MnFe₂O₄ & Ag/Fe₃O₄ ) [19–21], carbon based magnetic hybrids (MWCNTs/Zn_0.5Mn_0.5Fe₂O₄) [22], and magnetic-polymer nanocomposites (CoFe₂O₄@MnFe₂O₄/polypyrrole, Fe₃O₄/polyelectrolyte-Au) [23, 24]. For such hybrids, the complex interactions between integrated optical and magnetic components can impact their optimal performances, viz., optical absorption edge can be shifted due to proximity of magnetic component or heating efficiency of MHT has been seen to be reduced due to change in magnetic anisotropy and saturation magnetization [25]. Consequently, the overall heating performance is average or sum of both modes, limiting SAR and preventing substantial dosage reduction [26]. Whereas, single nanomaterial is expected to take leverage from these limits and provide synergic enhancement due to mutually reinforcing optical and magnetic relaxation’s pathways via phonon mediated lattice dynamics and surpassing the drawbacks i.e., thermal boundary resistance and energy losses for hybrid systems. The photothermal energy can modulate the effective magnetic relaxation time [27, 28] and vice versa, mediated by phonons [29, 30], under bimodal stimulation and ultimately providing optimal performance for tumor therapy.

The efficiency of Fe₃O₄ NPs for MHT can be enhanced by doping with various transition or rare earth metals, but such dopant (owing large ionic radii differences and unquenched orbital momentum) may also add to the toxicity and lattice distortion, impacting magnetic ordering and anisotropy unpredictably [31, 32]. Zn, Co and Mn doped Fe₃O₄ NPs exhibit high SAR due to improved value of saturation magnetization but generally require high fields i.e., intrinsic loss power (ILP) is particularly low [33–35]. Scandium (Sc) being first d-block element (also classified as rare-earth) has good biocompatibility [36, 37] and much stable valance shell configuration as compared to Fe³⁺ so it is expected that Sc substitution will be modifying exchange couplings between magnetic ions as well as stabilize the lattice by changing defects concentration, alteration of electronic states and mobility [38]. The studies reported for Sc doping on various binary ferrites such as Mn-Zn ferries [39], Ni–Zn ferrites [40] also presented the increase in saturation magnetization up to some extent and for Ni–Co ferrites [41, 42], Mn−Mg ferrites [43], the substitution of Sc ions show the weakening effect in super exchange interactions between two sublattice sites (A–B), (preferential occupation for Sc is octahedral/B sites) and decrease the hyperfine magnetic fields resulting to the changes in its magnetic and electronic properties.

Additionally, the nonoptimal photothermal performance of Fe₃O₄ can be addressed by making superstructures such as in Fe₃O₄ nanoclusters or self-assemble nanocrystals [44, 45] by enhancement of magnetic coupling and likelihood of ligand field (d-d) transitions [46]. So, the synthesis of self-assembled nanoflakes like morphology is promising as efficacy of hyperthermia with Fe₃O₄ NPs is constrained by superparamagnetic limit i.e., the inherent tradeoff between reducing the size of NPs reduces the magnetic energy barrier. Self-assembled nanoflakes like morphology has been observed to took liberty from this limit found in conventional spherical Fe₃O₄ NPs [47, 48] giving enhanced PTT performance due to magnetic coupling [49] and allowing certain indirect band gap transitions that are forbidden in single grain NPs [50, 51]. Further, the Sc doping can also supplement the PT efficiency by additional electron density of states and enhancing electron traps by alteration of localized defects [52, 53].

This work highlights the Sc doped Fe₃O₄ nanoflakes providing improved SAR for MHT and PTT individually as well as synergically enhanced SAR of 11857.53 W/g for bimodal hyperthermia tumor therapy when subjected to AMF (32 kA/m, 100 kHz, ) and laser (808 nm, 2 W/cm²). The obtained magnetothermal SAR is 1311.24 W/g (32 kA/m, 100 kHz) and ILP of 11.34 nHm²/kg is unprecedently high, while photothermal conversion efficiency (66.84%) and SAR is 3480.84 W/g (808 nm, 2 W/cm²). Good biocompatibility and showcased heating performance are auspicious for ablating the tumor using minimal dosage under mild external conditions.

Materials and methods

Materials

Scandium (III) nitrate hydrate (Sc(NO₃)₃. nH₂O, 99.9% trace metals basis), ethylene glycol (C₂H₆O₂, AR, 98%) and ammonia solution (NH₄OH, AR, 25–28%) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd., China. Iron (III) nitrate nonahydrate (Fe(NO₃)₃. 9H₂O, AR, 98.5%) was obtained from Shanghai Macklin Biochemical, China. Citric acid (C₆H₈O₇. H₂O, AR) was purchased from Sinopharm Chemical Reagent Co., Ltd., China. During all experiments, deionized (DI) water was used. All reagents were used as received, without further purification.

Human skin epidermal keratinocyte (HaCaT) and human malignant melanoma (A375) cell lines were obtained from Procell Life Science and Technology Co., Ltd., China. Dulbecco’s Modified Eagle medium (DMEM), supplemented with 10% fetal bovine serum (FBS) and Penicillin-streptomycin (PS) (1000 U/mL) was provided by Gibco Thermo Fisher Scientific Inc, USA. Phosphate-buffered saline (PBS) was supplied by Hyclone, USA. Ethanol disinfectant (70–75%) was purchased from Panjin Tainyuan Pharmaceutical Co., Ltd., China. Cell Counting Kit (CCK-8) was supplied by APExBIO Technology LLC, USA. Calcein acetoxymethyl ester (Calcein-AM) and propidium iodide (PI) were obtained from Jiangsu Keygen Biotech Co., Ltd., China.

Synthesis methodology and procedure

Nanocrystalline Sc-doped i.e., (x = 0, 0.03, 0.05 and 0.07) were synthesized using the citrate sol−gel auto combustion method(Pechini method) [54]. Stoichiometric amounts of Fe(NO₃)₃. 9H₂O and Sc(NO₃)₃. nH₂O (0.03 M) were dissolved in 80 mL of DI water along with citric acid (CA) at 1: 1 molar ratio. The pH of solution was adjusted at 7 using ammonia solution (25–28%). Ethylene glycol (EG) was added to attain a CA: EG molar ratio of 1: 2, followed by heating the solution to 80 °C under continuous stirring for 2.5 h until the formation of a dark brown gel. The resulting gel was dried at 100 °C for 12 h, followed by self-propagating auto combustion at 250 °C for a few minutes, forming a loose orangish brown nanocrystalline powder. The combustion temperature reached approximately 1000 °C, resulting in a highly crystalline mixed phase nanocrystalline fine powder. The powder was calcined at 500 °C (5 °C/min) for 3 h under argon gas flow to obtain final pure nanocrystalline magnetite (Fe₃O₄) phases. The resulting phases was washed twice with a water/ethanol (3:1) mixture and once with DI water to remove impurities and ensure pH neutrality.

Characterizations

The phase structure and purity of Sc doped Fe₃O₄ were analyzed using X-ray diffraction (XRD) (Smartlab 9KW, Rigaku corporation, Japan), with Cu Kα radiations (λ = 1.5405 Å) in the 2θ range of 10 to 80^o, with a step size of 0.01^o and a time step of 10 s. The structural and morphological analysis were conducted with atomic force microscope (AFM) (JPK NanoWizard 4 XP Bioscience, Bruker, Germany) and transmission electron microscope (TEM) (FEI Tecnai TF30, Thermo Fisher Scientific, USA), equipped with energy dispersive X-ray (EDX) spectrometer (X-max 80, Oxford instruments, UK) and field emission scanning electron microscopy (SEM) (Hitachi SU8600 Fe-SEM, Hitachi, LTD., Japan) (X-ray photoelectron spectroscopy (XPS) (Thermo ESCALAB 250XI, Thermo Fisher Scientific, USA) survey was conducted for detailed compositional studies by chemical oxidation states of elements and their proportions. (The occupancy of cations on lattice sites and vibrational modes of functional groups were studied using Fourier transform Infrared (FTIR) spectrometer (Nicolet is50, Thermo Fisher Scientific, USA), by obtaining attenuated total reflectance (ATR) FTIR spectra in spectral range of 100 to 4000 cm^-1 with spectral resolution of 0.09 cm⁻¹. The magnetization measurements were conducted in an external field of ± 20 kOe at room temperature, using a vibrating sample magnetometer (VSM) (Lake Shore 7404-S, Lake Shore Cryotronics, USA). The Curie temperature was determined by thermogravimetric analysis (TGA), using Thermogravimetric analyzer (Mettler Toledo AG-TGA/SDTA851e, Mettler Toledo AG, Switzerland), in temperature range from 25 to 650 °C (5 °C/min) under an Argon gas atmosphere and static magnetic field. Briefly, a 20 mg sample was placed in an alumina pan with a strong N35 magnet positioned in the microbalance chamber to track mass changes. The temperature at which the sample lost its magnetic properties, indicated by an abrupt mass loss, was recorded by differentiating the TGA curve. The colloidal stability of suspensions was analyzed using zeta potential and dynamic light scattering (DLS) measurement (Zetasizer Nao ZS90 laser particle size analyzer, Malvern Instruments, UK). The optical properties were measured by ultraviolet-visible (UV-Vis) spectrophotometer (Shimadzu UV-1800, Shimadzu corporation, Japan).

Magnetothermal performance evaluation

Suspensions of nanocrystalline Sc-doped i.e., (x = 0, 0.03, 0.05 and 0.07) were prepared at concentrations of 1 mg/mL, 2.5 mg/mL, and 5 mg/mL in water. The heating efficiency of each suspension was measured under an alternating magnetic field (AMF) with an amplitude of 32 kA/m and a frequency (f) of 100 kHz. The temperature change was monitored by using fluorescent optical fiber (Opsens Solutions Inc., Canada), over the time interval of 180 s. The shorter exposure duration ensures quasi-adiabatic conditions, enabling more accurate SAR evaluation using the Box-Lucas method [55], that is an essential consideration for practical implementation [56].The specific absorption rate (SAR) and intrinsic loss power (ILP) was extracted from heating curves by using Box-Lucas method and its fitting equation can be written as:

1

where a represents the maximum saturation temperature reaches in T vs. t curve and b represents the heat transfer rate from sample to surrounding or the rate at which the temperature approaches to a. Parameters a and b were obtained by fitting the above Eq. 1 over the heating curves and SAR and ILP are calculated by following equations:

2

3

where C represents the specific heat capacity of water, m_s represents the mass of suspension and m_n represents the mass of nanocrystalline sample.

Photothermal performance evaluation

Photothermal performances of (x = 0, 0.03, 0.05 and 0.07) under 808 nm laser (1 W/cm²) were assessed using 1 mg/mL suspension (avoiding excessive light scattering) of each composition as determined from extinction measurements. Under laser exposure, the temperature change was recorded by Infrared (IR) thermal imaging camera (Ti32, Fluke corporation, USA) at each 30 s interval. The composition exhibiting the maximum extinction and temperature rise over the time was further chosen for calculating the photothermal conversion efficiency (), temperature increase as function of laser power density and photothermal stability. The photothermal conversion efficiency () was determined using the formula:

4

Here, h represents the heat transfer coefficient, S denotes the surface area of tube, T_max refers to the maximum (equilibrium) temperature attained during laser irradiation, T_sur corresponds to the ambient surrounding temperature, is the heat dissipated by only solvent and tube under laser irradiation, I is the laser input power and A₈₀₈ indicates the absorption or extinction of sample at 808 nm. From these, hS can be calculated from cooling period (no heat production) of curve, when laser in turned off, using following equations:

5

6

7

where m_i represents the mass and C_i represents specific heat capacity of sample and solvent, respectively. From Eq. 5, the characteristic time constant of cooling () can be derived using Eq. 6, by taking slope of linear fit of vs. plot. For the quantification of temperature driving force, a dimensionless parameter was defined in Eq. 7, where, T_t is instantaneous temperature at time t, T_sur is ambient surrounding temperature and T_max is maximum steady state temperature. The parameter in Eq. 4 was measure independently using tube only containing DI water i.e., 0.004861 W, to exclude background heat contribution.

For the photothermal measurements of composition, which exhibits the best conversion efficiency was further assessed by taking specific absorption rate (SAR) values using Box-Lucas method under laser irradiation of various power densities (0.1–2 W/cm²), and monitoring the associated temperature changes over time.

Bimodal (photo and magneto-) thermal performance evaluation

For bimodal heating experiments (LASER + AMF), the 1 mg/mL suspension of (x = 0.05) was simultaneously exposed to different laser power densities (0–2 W/cm²) and AMF (32 kA/m, 100 kHz). The temperature was recorded by optical fiber and calibrated using IR thermal camera. The fiber-probe tip was positioned to minimize coupling with incident laser irradiance in order to suppress the background temperature rise. For all three heating modes (a) AMF (b) laser (c) bimodal i.e., simultaneous laser and AMF, the SAR was calculated using similar protocol as described previously and using Eq. 2 of Box-Lucas method.

Cell culture and cytotoxicity

The nanocrystalline (x = 0.05) was thoroughly immersed in 70−75% ethanol for 12 h, washed with PBS solution followed by UV light exposure for 12 h. The HaCaT cells were cultured in 96-well plates (n = 4) for each group, with cell culture medium having 89% DMEM, 10% FBS and 1% PS and incubated at 37 °C with 5% CO₂ humidification in an incubator (Heracell 150i, Thermo Fisher Scientific, USA). The cells were seeded at 2 × 10⁵ cells/well (100 µL cell culture medium) and incubated for 4 h, aiming for proper adhesion to the walls. The pretreated suspensions with various concentrations (0, 0.25, 0.50, 0.75 and 1 mg/mL), in 100 µL culture medium went through the replacement of medium in well plates and again put in the incubator. After different co-culturing periods (24, 48 and 72 h), the 120 µL CCK-8 essay mixtures (CCK-8: medium = 1: 10) were used to replace the culturing medium in each well and continued the incubation for further 3 h. After that, the cell viability was assessed by measuring the absorbance at 450 nm with microplate reader (SpectraMax M2e, Molecular Devices LLC, USA), taking 100 µL of supernatant solution, to quantify the ratio of dead to live cells in colorimetric way. Further, the medium in cell cultures of above adherent cells was then replaced with 100 µL of fluorescent staining mixtures (Calcein-AM: PI: PBS = 5: 3: 1000) and incubated for 30 min. The live or dead states of cells were observed by imaging using fluorescent microscope (Olympus BX71, Olympus corporation, Japan).

In vitro hyperthermia studies

A375 cells (5 × 10⁵ cells/well, 100 µL cell culture medium) were cultured in 96-well plates for 12 h followed by replacement of sterilized and washed suspension (100 µL) of nanocrystalline (x = 0.05) in culture medium. After this, these were divided in to two group classes; (1) Control and (2) Nanoflakes. Each class was randomly subdivided in into 4 treatment groups (n = 4 per group): (i) Without LASER/AMF, (ii) With LASER, (iii) With AMF and (iv) With LASER + AMF. For LASER group, 808 nm NIR laser of 0.1 W/cm² power density was used for continuous irradiating the cultured medium for duration of 8 min. For AMF group, the cultured medium was subjected to AMF (100 kHz, 32 kA/m) for similar duration and the temperature during each exposure was continuously monitored by IR thermal imaging camera. Finally, for LASER + AMF group, it was subjected to both NIR laser (808 nm, 0.1 W/cm²) and AMF (100 kHz, 32 kA/m) simultaneously following similar procedure. After 4 h incubation, CCK-8 testing and fluorescent staining microscopic imaging were performed following the same protocols as mentioned above in details. Further, for LASER + AMF group, the cell viability was assessed for intermediate duration of 4 min and finally 8 min to observe the instantaneous cell killing effects over the time period. For similar experimental group (Control, 0 min, 4 min and 8 min), after heating treatment (LASER + AMF), all cells were washed twice and stained using an Annexin V-FITC/PI apoptosis detection kit. The stained cells were subsequently analyzed by flow cytometry (BD FACSymphony™ A1 cell analyzer, BD Biosciences, USA).

Statistical assessment

All data were presented as the mean ± standard deviation (SD) for various groups (n = 4). To assess the statistical significance among experimental groups, one-way analysis of variance (ANOVA) was applied and significance level was selected at 0.05 (p-value). The significances are indicated symbolically with (*) for p < 0.05, (**) for p < 0.01, (***) for p < 0.001 and (n.s.) for no significance statistically, over the asterisk brackets.

Results and discussion

Material design and structural studies

Nanocrystalline Sc doped Fe₃O₄ was successfully synthesized by Sol-gel auto combustion method as briefly presented and illustrated in Figure S1 in Supporting Information. Conventional low temperature synthesis routes e.g., co-precipitation, suffer from several drawbacks such as limited and non-uniform dopant incorporation and poor crystallinity [57–59]. Here in, during the sol-gel synthesis, the EG (as complexing or chelating agent) is incorporated that endures esterification reaction with metal citrate complex under heat to form metal ions entrapped poly-esterified covalent network that ensures homogeneous distribution and stabilization of metal ions followed by high temperature auto combustion to get highly crystalline nanostructure. As the incorporation of rare earth (RE) ions into ferrite lattice require substantially greater energy than Fe³⁺ due to stronger RE³⁺–O bond; therefore, high temperature synthesis is required to achieve uniform nucleation and grain growth, and resulting ferrites exhibit enhanced thermal stability.

TEM images shown in Fig. 1A(i-ii)-D(i-ii) present the morphological features of all compositions of (x = 0, 0.03, 0.05 and 0.07), respectively. The results indicate that nanocrystalline structures weather doped or undoped have two-dimensional (2D) nano-plates or nano-flakes like morphology sintered together by thermal treatments during combustion and calcination, forming micrometric clusters indicating potential aggregations. This is due to the Pechini auto-combustion method which involves polymeric resin formation with homogenous distribution of metallic ions and rapid combustion, which can create non-equilibrium conditions favoring anisotropic growth [48, 60]. The temperature profile and rapid auto-combustion in the Pechini method can limit particle growth time, favoring 2D growth due to kinetic control rather than thermodynamic equilibrium (which favors spheres). This promotes growth along specific crystallographic planes, and self-assembly resulting to 2D nanoflakes [61, 62] without employing any template. The average size of each flake determined to be approximately 1 μm and high-resolution TEM (HRTEM) images in Fig. 1A(iii)-D(iii) indicate the obvious lattice streaks with characteristic d-spacings of respective atomic planes (311 and 400) of cubic spinel structure. The selected area electron diffraction (SAED) patterns in Fig. 1A(iv)-D(iv) exhibit several in line diffraction spots with d-spacings of as above atomic planes hence confirming the highly crystalline phase and single crystalline nature of small grains. It referred to the indication that each entity is single crystalline but the powder as a whole is polycrystalline in nature. The atomic force microscopy (AFM) analysis depicted in Figs. 1E and 2H also confirms the flakes like shape of nanocrystalline structure having average lateral dimension of approximately 1 μm, with sharp to round edge morphology, an average individual thickness of about 11 nm and relatively rough surface. The height profile confirms the approximate thickness, incorporating multilayer stacking also perceived in TEM studies.

Fig. 1.

Morphological analysis Sc-doped Fe₃O₄ nanoflakes: TEM at low magnification and high magnification, HRTEM and SAED images of for: A(i−iv) x = 0; B(i−iv) x = 0.03; C(i−iv) x = 0.05; D(i−iv) x = 0.07. The inset shows the characteristic lattice planes and d−spacing. AFM images for: E x = 0; F x = 0.03; (G); x = 0.05 and (H) x = 0.07 depicting thickness and lateral dimensions

Fig. 2.

Structural, compositional and spectral analysis of (x = 0, 0.03, 0.05, 0.07): A XRD patterns with subsequent peak shifting by enlarged view of (311) lattice plane. B FTIR spectra and zoom in ~ 300–650 cm⁻¹ region showing the characteristic absorption bands for metal complexes at octahedral and tetrahedral sites. C Dopant dependent variations in lattice parameter, crystallite size and macrostrain from XRD analysis. D(i−iv) EDX spectra with insets images showing scanning TEM results along with elemental mapping and distribution of Fe, O and Sc (left to right) E XPS full spectra of with (F−I) high resolution spectrum of elements Fe 2p, Sc 2p, O 1s and C 1s, respectively

X-ray diffraction (XRD) patterns of all synthesized (x = 0, 0.03, 0.05 and 0.07) nanoflakes are shown in Fig. 2A. The clear bragg diffraction peaks indicate the single-phase cubic spinel structure with Fdm space group and well matched with JCPD reference card no. 85-1436. The as obtained peaks are referred to (111), (220), (311), (222), (400), (422), (511) and (440) crystal planes and no extra peak corresponding to some secondary phase is detected. From x = 0 to x = 0.05, peak shifting toward lower 2 values are observed due to difference between ionic radii of Sc³⁺ (0.073 nm) and Fe³⁺(0.064 nm). However, after x = 0.05, it goes toward higher 2 indicating the non-uniform strain in crystal lattice, as Sc ions reaches their solubility limit and no more replacing Fe³⁺ ions. (Fig. 2A, enlarged view). The observed XRD peaks shifting and broadening is associated with both crystallite size and lattice strain effects and can be appropriately understood by the Williamson–Hall equation:

8

where ε, D, λ, β, k, and θ represent micro strain, crystallite size, x-ray’s wavelength i.e., 0.15405 nm, full width half maximum (FWHM), shape coefficient (chosen k = 0.8 for platelet shaped nanoparticles) and Bragg’s angle, respectively. The lattice parameter (a) average crystallite size (D) and associated dislocation density () was calculated using Debye-Scherrer formula:

9

10

For x = 0.07, the dopant Sc ions are expected to occupy interstitial sites or may reside at grain boundaries, resulting opposite peak shift than earlier. These ions owing bigger ionic radii can produce strain at grain boundaries hence hindering the lattice expansion and causing reduction in both lattice parameter (a) and average crystallite size (D).

The Fig. 2B illustrates the effect to variation in lattice parameter (a), average crystallite size (D) and strain (ε) with dopant content. The detailed calculated structural parameters from XRD data are presented in Table 1, that shows increase in lattice parameter, crystallite size and cell volume up to x = 0.05 due to Sc incorporation in to doping sites causing lattice expansion. On the other hand, strain and dislocation density decreases on same time due to Sc doping stabilizing the lattice but after x = 0.05, opposite phenomenon occurs for all parameters due to Sc ions being occupied at interstitial sites or residing at grain boundaries causing increase in strain and dislocation density. The general disagreement between TEM results and the average crystallite size changes in Table 1 derived from XRD, suggest that nanoflakes have polycrystalline nature composed of crystalline domains of size about 40 ~ 44 nm. The lattice expansion should result in increasing the crystallite sizes but effect on overall nanoflakes sizes by TEM results are non-distinguish due to aggregation and clustering effect.

Table 1.

Structural data of all compositions obtained from XRD measurements

Sc doping content (x)	0.00	0.03	0.05	0.07
Cationic distribution	[Fe3+]A [Fe2+Fe3+]B O4	[Fe3+]A [Fe2+ Fe3+0.97Sc3+0.03]B O4	[Fe3+]A [Fe2+ Fe3+0.95Sc3+0.05]B O4	[Fe3+]A [Fe2+ Fe3+0.93Sc3+0.07]B O4
Avg. crystallite size D (nm)	40.429	42.861	44.458	38.786
Lattice constant a (Å)	8.389	8.396	8.399	8.394
Cell Volume (Å3)	590.37	591.85	592.49	591.43
Strain (ε) ×10− 3	2.297	2.118	2.092	2.471
Dislocation density () ×10− 3	0.611	0.596	0.507	0.751

The ionic distribution and lattice vibrational modes of the (x = 0, 0.03, 0.05 and 0.07) were investigated through IR absorption spectra, as shown in Fig. 2C. The spectra exhibit absorption bands between 1900 and 2100 cm⁻¹, corresponding to the stretching vibrations of carbonaceous residues (C–O) adsorbed on the surface of nanoflakes, rather that intrinsic absorption band of Fe₃O₄ ferrites ^[63]. This is attributed to the sol gel auto combustion synthesis method, which involves organic precursors that leave behind such surface species. The typical ferrite absorption bands are observed in the range of 300–700 cm⁻¹ with three prominent peaks, denoted as v₁, v₂ and v₃, corresponding to the metal complexes at the octahedral (B-site) and tetrahedral (A-site) positions. These bands reflect the differences in bond lengths (Fe²⁺–O/ Fe³⁺O) and are summarized in Table 2. The minor shifts in the absorption band positions are influenced by the synthesis method, porosity and grain size. However, the v₁ sharp absorption band around 580 cm⁻¹ is associated to the stretching vibrations of tetrahedral A site metal complexes (Fe³⁺−O), remains unaffected by any splitting or shifting across all samples. This observation confirms that Sc ions do not occupy A site. There is obvious shoulder splitting in the absorption band v₂, associated to the trivalent metal octahedral complexes. This splitting can be explained by Jahn Teller distortion produced by non-cubic crystal field potential of Fe²⁺, resulting from local deformation induced by doping. Additionally, the peak shift toward lower wavenumber indicates that the Sc bigger ionic radii have substituted the Fe³⁺ ions producing Sc³⁺–O and at higher doping content producing lattice deformation as well. The similar results have been reported for other Sc doped ferrites [39, 40]. The v₃ band at low frequency presents the presence of divalent metal octahedral bond of metal complex and some vibrations involving the displacement of ions residing at octahedral sites. It is also observed that the intensities of bands are also composition dependent, that tends to rise with increasing doping content, implying the increase the extent of interaction between functional groups and hence leading to increased absorption intensity.

Table 2.

Characteristic band positions with compositions from FTIR spectra

Doping content	Infrared (FTIR) bands position
	v1 (cm−1)	v2 (cm−1)	v3 (cm−1)
x = 0.00	580	418	348
x = 0.03	577	411	347
x = 0.05	579	408	344
x = 0.07	583	419	345

The presumed compositions of (x = 0, 0.03, 0.05 and 0.07) were confirmed by energy dispersive X-ray (EDX) spectrum and elemental mapping in Fig. 2D(i)-(iv). The observations show that the undoped composition have right atomic weight percentages of Fe and O, when x = 0 as well as the Sc dopant in three compositions (x = 0.03, 0.05, 0.07) of nanoflakes matrix is present in nominal quantities i.e., 0.43, 0.69, 0.96% respectively, as suggested by atomic weight percentages in inset images. Moreover, the Fe, O and Sc elements are turned out to be homogeneously distributed in total volume of nanoflakes as well as no other impurity elements are detected.

The element valance states of (x = 0.05) was characterized by XPS and the full survey spectra is presented in Fig. 2E along with high-resolution spectrum containing Fe 2p, Sc 2p, O 1s and C1s in Fig. 2F-I, respectively. For Fe 2p high-resolution spectra, two deconvoluted spectral bands at 725.5 and 711.9 eV were corresponds to 2p_1/2 and 2p_3/2 of Fe³⁺ species, respectively. Similarly, two bands at binding energies of 723.3 and 710.7 eV could be assigned to 2p_1/2 and 2p_3/2 of Fe²⁺ species. The other two weak peaks at 719.3 and 733.1 eV are satellite peaks confirming the formation of Fe₃O₄ phase and its purity. As, Fe 2p have spin orbital peak splitting so the ratio of Fe²⁺/Fe³⁺ can be obtained by taking only Fe 2p_3/2 in to account and using relative area of two peaks, the value of Fe²⁺/Fe³⁺ is 2.11 due to possible surface oxidation and γ-Fe₂O₃ phase. The high resolution spectra of Sc 2p in Fig. 2G exhibits three peaks at binding energies of 398.9, 402.02 and 406.6 eV assigned to Sc(0) or sub-ox species due to some incomplete oxidation, Sc 2p_3/2 and Sc 2p_1/2 of Sc³⁺ in oxide matrix, respectively. The high-resolution O 1s and C 1s spectrum are shown in Fig. 2H and I. The O 1s spectral band consisted of two peaks at binding energies of 530.2 and 531.8 eV, that are attributed to lattice oxygen in metal oxide (Fe–O) bond or oxygen anion (O₂⁻) and hydroxide (OH) on the surface of metal oxide, respectively. For C 1s, the three spectral bands at binding energies of 284.7, 286.6 and 288.1 eV are assigned to C=C/C–C (sp₂ carbon), C–O species (e.g., OH) and C=O carbonyl or the low end of COOH related carbon groups, respectively. It validates the FTIR results to indicate the presence of some residual carbonaceous functional groups on the surface.

The Fig. 3 A−E presents the SEM images of taken at various magnifications. These results also suggest predominant nanoflakes like morphological structure that is consistent with TEM and AFM observations (Fig. 1), along with some crystalline grains with well-defined sharp edges. These grains likely arise from localized anisotropic growth or incomplete exfoliation, reflecting a competition between isotropic crystalline domains and kinetically favored 2D nanoflakes during synthesis. The coexisting sharp‑edged grains and nanoflakes suggest simultaneous thermodynamic and kinetic growth mechanisms. The edge width of these grains is observed to be ~ 63 nm, marked in Fig. 3E. The large populations of plate-like structures are randomly oriented and partially overlapping, forming a porous and interconnected network due to potent aggregation in powdered form. The size distribution plot in Fig. 3F shows the average size of 0.87 and the lateral dimensions of the flakes fall within the nanoscale range, with relatively smooth surfaces and sharp edges providing high surface-to-volume ratio. These images confirm that the morphology, size distribution, and overall structure are consistent across the synthesized samples, indicating good reproducibility and macroscopic uniformity of the sol‑gel process.

Fig. 3.

(A-F) Low resolution SEM images showing morphology of large population of nanoflakes along with size distribution plot. G UV-Vis absorption spectra (concentration: 1 mg/mL) H Extinction coefficient () measurement at 808 nm wavelength I Tauc’s plots for calculation of band gap energies () for various compositions. J Respective changes in band gap energies and extinction coefficient at 808 nm wavelength with changing the doping content. (K−L) Zeta potential and PDI of in DMEM supplemented with 10% FBS over various time intervals of 0, 24, 48 and 72 h. The inset image presents magnetic suspension in cell culture medium (DMEM with 10% FBS)

Optical properties studies

UV−Vis spectral studies characterized the effect of Sc doping on extinction ability supported by band gap measurements of as synthesized nanoflakes. The absorption spectra of (x = 0, 0.03, 0.05 and 0.07) taken at constant concentration of 1 mg/mL, presented in Fig. 3G, which show broad extinction of all suspensions in NIR region. The absorption value of samples when x = 0, 0.03, 0.05 and 0.07 are 0.49572, 1.11676, 1.43346 and 0.3919, respectively. The values suggest the increase in light absorption with Sc doping until x = 0.05 possibly due to substitutional doping and the decrease at x = 0.07 due to lattice strain caused by doping at interstitial sites. According to Beer−lambert law:

11

The extinction coefficient () is calculated by taking linear filling of extinction values (A) at 808 nm for different concentrations (c) of suspensions, over the optical path length (). From Fig. 3H, it can be observed that the extinction coefficient first increases almost linearly up to x=0.05 with maximum value of 1.4304 Lg⁻¹cm⁻¹ and then decreases. The optical absorption property that corresponds to the electronic transitions from valance band to conduction band, can be used to determine band gap energy using Tauc’s relation written as:

12

Here, and represent the photon energy and band gap energy respectively. B is proportionality constant (band tailing or Tauc’s parameter) and exponent n is related to electronic transition type (indirect or direct), typically either ½ for direct or 2 for indirect transitions. The parameter depicts the absorption coefficient relation to Eq. 11 as:

13

where A is absorption or extinction and l is optical path length. The Tauc plots of all composition presented in Fig. 3I report the direct band transitions, by plotting versus photon energy and extrapolating the straight portion up to x-axis gives the band gap energy .

Band gap calculations using Tauc plots reveal a decrease in the energy gap from 3.34 to 2.58 eV with increasing concentrations of Sc³⁺ ions up to x = 0.05, indicating changes in the electronic band structure. This reduction is attributed to the introduction of localized shallow impurity states by Sc dopant, which appear near the valance band maximum (VBM) or conduction band minimum (CBM), leading to the merging of these sub-states into continuous bands [64, 65]. Since the 3d subshell of Sc³⁺ ions is empty, the weaker electrostatic interactions between Sc³⁺ and O²⁺ compared to those between Fe³⁺/Fe²⁺ and O²⁺, resulting to the longer Sc–O bond length, inducing lattice distortion and altering the electronic density of states near the band edges, which consequently reduces the band gap [66]. This reduction in the energy gap has also been reported to be inversely related to increasing crystallite size [67, 68], a trend that is supported by the XRD results presented here. For x = 0.07, the band gap increases, and the reduced extinction observed in Fig. 3J suggests the non-substitutional doping at this level, with further deterioration in properties at higher doping concentrations. The observed reduction in the band gap is promising for enhancing photothermal effects, as the additional impurity states facilitate more non-radiative recombination, thereby improving infrared (IR) extinction values [64].

The colloidal stability of nominal composition was systematically evaluated in DMEM supplemented with 10% FBS over 0 h to 72 h using dynamic light scattering (DLS) and zeta potential measurements. The zeta potential results in Fig. 3K show slight shift from −15.2 to −18.4 mV over 72 h. This modest negative charge increase indicates progressive protein adsorption and formation of stabilizing corona, rather than aggregation. The value remain well with in this range (−10 to −20 mV) are typically associated with stable colloids in biological media. This negative zeta potential caused due to deprotonated carboxyl groups from carbonaceous residuals prevent cell adhesion and cytotoxicity for extracellular long term retention. by electrostatic repulsion from anionic cell membrane The DLS results presented in Figure S2 of Supporting Information indicated the good colloidal stability over the 72 h time period as the hydrodynamic size (nm) essentially remain unchanged (~ 101 nm) with a negligible decrease (0.5%) at 72 h. Although, the nanoflakes have average size of about 0.87 μm, but Z-average size of ~ 100 nm is intensity weighted diffusion-averaged size that is hydrodynamic equivalent spherical diameter of anisotropic nanoflakes. These nanoflakes diffuse 3 to 5 times faster than equivalent volume spheres due to lower hydrodynamic fraction and edge-on translational diffusion so the Z-average size is heavily weighted toward smaller dimension. The values of polydispersity index (PDI) (Fig. 3L) over the 72 h indicate monodisperse or narrow size distribution as PDI < 0.2, the threshold acceptable for monodispersity. The value remains stable up to 48 h and slightly decrease by 0.071 at 72 h indicate improved dispersion homogeneity, hence, ruling out the aggregation tendency due to denatured on hard protein corona formation.

Magnetic properties studies

The room temperature hysteresis loops (M-H curves) of the as-synthesized Sc doped Fe₃O₄ nanoflakes are presented in Fig. 4A. The composition (x = 0.05) exhibits the highest saturation magnetization of 90.33 emu/g, which is very close to M_s value of bulk magnetite (i.e., 92 emu/g) [69]. The observed M−H loops of all compositions of nanoflakes display very narrow shaped hysteresis, indicating the soft magnetic behavior and a nearly superparamagnetic nature at room temperature. According to Neel’s two sublattice model, the net magnetic moment per unit formula results from antiferromagnetic ordering between A and B sites i.e., n(µ_B)=M_B − M_A. Here the increase in saturation magnetization from x = 0 to x = 0.05 can be explained by the lattice expansion, as evidenced by the increased lattice parameters and crystallite size from XRD data in Table 1. Substitution of Sc ions causes lattice expansion, increasing the distance between magnetic ions, which reduces orbital overlap and narrows the d-electron band. This spin splitting effect contributes to the observed increase in M_s. The M_s value also shows a clear dependence on the surface to volume ratio and the structure of the magnetic domains. The doping reduces the structural defects and minimizes spin canting of the surface atoms, leading to a higher surface to volume ratio, which stabilizes the magnetic order and contributes to the increased Ms. Additionally, at low doping concentrations, it is possible that Sc ions may occupy the A sites, causing a shift of some Fe³⁺ ions to the B sites, which would enhance the magnetic moment at the B site and increase the net magnetic moment per formula unit. Although the preferential occupation site of Sc ions is at the B site is more experimentally plausible, the observed reduction in crystallite size at x = 0.07, due to lattice strain, leads to a decrease in saturation magnetization. Saturation magnetization (M_s) value decreases from 90.33 emu/g for x = 0.05 to 79.82 emu/g for x = 0.07, to a net reduction of 8% at room temperature. A rather modest decrease in M_s value with x content indicates that the substitution of Sc³⁺ is most likely favored at octahedral sites with the spin-down configuration as also endorsed by Pauling’s rule taking account to the ionic radii ratio of cation to anion (r_C/r_A) i.e., 0.564 nm, that justify coordination number 6 for Sc³⁺ (octahedral/B site) [70]. The reduction in saturation magnetization upon Sc doping can be attributed to several factors: (1) the magnetic dilution effect arising from the decrease in the net magnetic moment per formula unit caused by replacing Fe³⁺ (5 µB) with a non-magnetic Sc³⁺ ion, (2) the lattice distortion, which could weaken Fe³⁺–O²⁻–Fe³⁺ super exchange interactions and (3) decrease in the number of super exchange pairs upon substitution for Fe³⁺ which seems to be more favorably towards Sc³⁺–O²⁻–Fe³⁺ rather than Fe³⁺–O²⁻–Fe³⁺. Further, the values of Coercivity (H_c), Remanence magnetization (M_r) and squareness ration (M_r/M_s), summarized in Table 3 show the opposite trend than that of M_s, due to Sc doping stabilizing the magnetic domain structure and reducing then spin canting of surface atoms up to x = 0.05 and afterword these parameters increase possibly due to lattice strain and structural defects. The effective magnetocrystalline anisotropy constant (k) was calculated using formula:

14

Fig. 4.

Magnetic and magnetothermal analysis: A Magnetic hysteresis loops of (x = 0, 0.03, 0.05 and 0.07) at room temperature. B TGA of nanoflakes for: (i) x = 0, (ii) x = 0.03, (iii) x = 0.05, and (iv) x = 0.07. The inset images exhibit the variation of first derivatives of normalized mass versus temperature curves and the Curie temperature (T_c) maxima. Heating curves of (x = 0, 0.03, 0.05 and 0.07) suspensions with concentration of: C 1 mg/mL D 2.5 mg/mL E 5 mg/mL under magnetic field of 32 kA/m, 100 kHz. F Corresponding evolution of calculated SAR with doping content x, ILP is taken from average SAR at each composition and error bars showing deviation from mean

Table 3.

Saturation magnetization (M_s), coercivity (H_c), remanence magnetization (M_r), squareness ration (M_r/M_s) and effective magnetocrystalline anisotropy constant (k) of Sc_x [Fe²⁺] O₄ against Sc³⁺ doping content from M−H loop

Sc doping content (x)	Saturation magnetization (Ms) emu/g	Coercivity (Hc) Oe	Remanence magnetization (Mr) emu/g	Squareness ratio (Mr/ Ms)	Effective magnetocrystalline anisotropy constant (k) erg/g
0.00	83.56	193	27	0.32	16799.04
0.03	86.47	164	24	0.27	14706.33
0.05	90.33	156	22	0.24	14678.62
0.07	79.82	209	25	0.31	17377.47

Here, M_s and H_c represents the saturation magnetization and coercivity respectively. The variation of magnetocrystalline anisotropy constant (k) with doping content is presented in Table 3. The reduction in the anisotropy constant (k) up to x = 0.05 is attributed to the substitution of Sc³⁺ (no magnetic anisotropy) at B sites, eventually leading to a weakening of the A–B super exchange interactions. Beyond the solubility limit of doping, the anisotropy constant increases due to lattice strain that modulates the exchange interactions. This change in anisotropy is closely linked to the Curie temperature (Tc), a very important parameter for realization of self-regulating magnetic hyperthermia.

Figure 4B(i)−(iv) presents the thermogravimetric analysis (TGA) curves of (x = 0, 0.03, 0.05 and 0.07) nanoflakes, with inset images showing the derivative of curves and the maximum value points to the Curie temperature (T_c). The inset of Fig. 4B(i) shows that the T_c value of undoped nanoflakes (x = 0) is 567.8 °C, which is consistent with several reported values in the literature [71–73], depending on synthesis and morphological changes. The T_c is the critical temperature above which ferro- or superparamagnetic nanomaterial undergo a phase transition to a paramagnetic state with disordered spins. This transition is generally driven by super exchange interaction between the A and B sites. For x = 0.03 and x = 0.05, the T_c values decrease to 498.6 and 442.6 °C, respectively (insets of Fig. 4B(ii) and (iii)). This reduction is attributed to the replacement of Fe³⁺ (5 µB) with a non-magnetic Sc³⁺ (0 µB) at B sites, which induces lattice distortion and increases the distance between magnetic ions at A and B sites. Consequently, the Fe³⁺–O²⁻−Fe³⁺ super exchange interactions are weakened, and the number of super exchange pairs decreases, favoring Sc³⁺–O²⁻–Fe³⁺ interactions instead. It can be seen in inset of Fig. 4B(iv), that the Curie temperature then increases to 484.7 °C, after x = 0.05 to x = 0.07 by enhancing doping content of Sc³⁺ ions, possibly due to strengthening of A–B super exchange interaction and reduction of distance between magnetic ions due to induced strain by interstitial sites occupation, leading to an increase in the T_c. The results suggest that Sc³⁺ substitution could be an effective strategy for developing doped ferrites nanostructures with low Curie temperature for realization of self-regulating hyperthermia treatment.

Magnetothermal studies

The heat generation ability of magnetic nanoparticles is typically quantified in term of the specific absorption rate (SAR) i.e., which measures the heat generated per unit mass (W/g). However, for more standardized comparisons, the intrinsic loss power (ILP) is a more relevant parameter as it is independent of external field parameters and experimental conditions [74, 75]. Figure 4C and E display the temperature rise curves of (x = 0, 0.03, 0.05 and 0.07) magnetic suspensions with concentrations of 5 mg/mL, 2.5 mg/mL and 1 mg/mL under an applied magnetic field of 32 kA/m at 100 kHz, measured over a constant time interval (180 s) in water and Fig. 4F presents the variation trend of calculated SAR/ILP values at all dopant content levels, over the various concentrations.

For 1 mg/mL suspensions, the maximum temperature change of 13.2 °C was observed for x = 0.05 and the temperature difference among the other compositions is not significant, possibly due to limited magnetic couplings and content of magnetic domains per unit volume. For 2.5 mg/mL suspension, the maximum temperature change of 55.7 °C was observed for x = 0.05 and the temperature difference among the other compositions is quite significant and distinguish, which is attributed to the enhancement of magnetic couplings and content of magnetic domains per unit volume. For 5 mg/mL suspensions, the temperature change is also highest i.e., 59.5 °C in 50 s, for x = 0.05, with most significant rise initially and then stabilizes to some extant due to negligible heat exchange with the surrounding medium. The higher concentration (5 mg/mL) suspensions exhibit more rapid temperature increase (rate of change of temperature) employing the increase in SAR values, which is attributed to the increased density of magnetic domains per unit mass and the reduced distance between individual gains, which collectively promote long range ferromagnetic ordering. Stronger dipole-dipole interactions and magnetic coupling further enhance the heating performance. The structural and magnetic studies indicate that the nanoflakes are of the blocked single domain type, exhibiting very narrow hysteresis. Consequently, the primary mechanisms contributing to heat generation are likely hysteresis loss (due to magnetization reversal) and Brownian relaxation (physical rotation of the particles). The contribution of Neel relaxation to the heating mechanism in these large-single domain structures is minimal due to the significant magnetic anisotropy energy barrier [76]. The increase in SAR with higher concentrations further supports the dominance of hysteresis loss and Brownian relaxation as the primary heating mechanisms. At higher concentrations, the stronger inter-grain dipole interactions raise the effective energy barrier, thereby reducing the contribution of Neel relaxation to the overall heating process [77, 78]. Additionally, Brownian relaxation becomes more sensitive to changes in viscosity and hydrodynamic size, both of which increase with concentration.

Referring to Figs. 4F, the experimental result also gives indication that SAR variation trend is consistent with that of M_s (Table 3), so x = 0.05 composition exhibits maximum SAR value of 1311.24 W/g and x = 0.07 exhibit minimum SAR value of 497.73 W/g at 5 mg/mL concentration. Theoretically, SAR can be obtained by calculating the volumetric power dissipation (P) per unit mass of nanomaterial, So:

15

where, f and H represents the frequency and amplitude of AMF respectively, is magnetic permeability of free space and is imaginary part of magnetic susceptibility, given by:

16

With is angular frequency of AMF, is total relaxation time and the static magnetic susceptibility () is defined as:

17

where is Boltzmann constant, V is the magnetic volume, M_s is saturation magnetization and T is the absolute temperature. It can be justified from above Eqs. 15–17, that SAR directly corelates with M_s, which conclude the high SAR value of (x = 0.05). Also, the SAR have significant dependence on size (hydrodynamic and magnetic volume), which may also explain this observation because sample with x = 0.05 has obviously larger crystallite or grain sizes (Table 1). The increase in size may also modulate the effective relaxation time majorly the Brownian relaxation, making closer to maximum theoretical value causing the increment in SAR [79, 80]. The Fig. 4F briefly summarizes the effect of SAR and ILP by changing the doping content at various concentrations. The sample with x = 0.05 shows maximum SAR at each concentration, mainly because of the size and M_s increment resulting from Sc doping. For x = 0.07, the properties become ambiguously deteriorated due to structure distortion effects. The ILP is calculated by taking average of SAR at three concentrations and error bars presents the deviation caused by concentration effect. The as obtained ILP value for x = 0.05 sample is 11.34571.6519 nHm²/kg is unprecedently high, which can realize the hyperthermia using minimal dosage and exposure time under safety limit of magnetic field. Further, the Figure S3A reveals that the temperature rises and SAR of (x = 0.05) remains almost identical across the 4 cycles (180s AMF ON and 180s AMF OFF), indicating the excellent thermal stability and reproducibility of magnetothermal performance. The temperature drop during cooling period is relatively low, which is attributed to the insulated sample holder (for ideal adiabatic system) minimizing heat loss to surroundings. The structural stability is also confirmed by XRD taken before and after the heating-cooling cycles is presented in S3(B).

Photothermal studies

In order to study the photothermal performance of (x=0, 0.03, 0.05, and 0.07), the 1 mg/mL suspension of each composition was irradiated with 808 nm laser at power density of 1 W/cm² and temperature rise was monitored using IR thermal camera. The Fig. 5A shows the sample with x = 0.05 shows the maximum temperature rise up to a stable temperature of 44.4 °C at 10 min compared with x = 0, x = 0.03 and x = 0.07 that shows an obvious temperature change of 31.7, 35.4 and 29.3 °C, respectively. It has been discussed in subsection 3.2 (Fig. 1E and F), that x = 0.05 composition exhibit highest extinction enabling greater energy conversion in to heat and indicating the direct covariation of extinction with photothermal heating. While the lower extinction of yield smaller temperature elevation due to lower energy absorption and conversion to heat. Based on this, the (x = 0.05) i.e., can be used for calculating the photothermal conversion efficiency, using the heating and cooling curve after the laser was turned off in Fig. 5B and the characteristic time constant ( of system by slope value of best fitted line in Fig. 5C. Using the Eq. 4, photothermal conversion efficiency was obtained to be 66.84%, which should be maximum and consistent with the variation fashion of both extinction coefficient and the maximum temperature rise. The as obtained conversion efficiency is also very high as compared to previously reported superparamagnetic Fe₃O₄ nanoparticles and superstructures [44, 81], possibly due to doping effect as well as unique self-assembled morphology with enhanced magnetic coupling. To study the influence of laser density on maximum temperature rise, the Fig. 5D and E present the IR temperature distribution images and corresponding heating curves at the laser power densities of 0.1 W/cm², 0.33 W/cm² and 1 W/cm², respectively. The results enlighten that there is no significant temperature rise when laser power density is minimal i.e., 0.1 W/cm², while it can get to the therapeutic level i.e., ~ 46 °C from the room temperature (25.4 °C), using laser density of 0.33 W/cm² alone and also the rapid heating effect (0–3 min) for all power densities, with an approximate stable temperature (7–10 min). The four successive heating and cooling cycles of curve in Fig. 5F, exhibit good stability for photothermal experimentation, that is an essential requirement for in vivo or in vitro applications.

Fig. 5.

Photothermal performance analysis of (x=0, 0.03, 0.05, and 0.07): A Heating curves under 808 nm laser irradiation (Concentration: 1 mg/mL, Power density: 1 W/cm²). B Calculation of Photothermal conversion efficiency of (Concentration: 1 mg/mL, Power density: 1 W/cm²), C respective Time versus -ln plot from cooling period. D Laser power density dependent IR temperature field distribution images of (Concentration: 1 mg/mL). E Temperature elevation curves from room temperature for under laser power densities of 0.1 W/cm², 0.33 W/cm² and 1 W/cm² (Concentration: 1 mg/mL, n = 3). F Successive cyclic heating curves of (Concentration: 1 mg/mL, Power density: 1 W/cm²). G Heating curves of nanoflakes suspension (Concentration: 1 mg/mL) under 808 nm laser irradiation at different power densities. H Corresponding calculated SAR as function of laser power densities (n = 3), the error bars represent deviation from mean values

The specific absorption rate (SAR) of nanoflakes was also calculated by subjecting its 1 mg/mL suspension to 808 nm laser irradiation with power densities ranging from 0.1 to 2 W/cm². As shown in Fig. 5G, the temperature change is directly proportional to the laser power, with maximum temperature increase, reaching to 55.3 °C at high power density of 2 W/cm² and minimum temperature increase of 5.6 °C at ultra-low power density of 0.1 W/cm². This is due to the fact that the population of electronic excitations is more under higher absorbed power, ultimately undergoing non-radiative relaxation to yield heat and also vice versa. The largest temperature change implies the elevated SAR values with increasing laser power densities, affirming the positive correlation of photothermal conversion and thermal output. The calculated SAR values are expressed in terms of the nanomaterial mass (W/g), rather than the metal content, providing a more accurate representation of the heating performance relevant to practical dosing. The Fig. 5H shows the calculated SAR values at various laser power densities. At an ultra-low power density of 0.1 W/cm², the SAR is 681.3 W/g, which steadily increases to 3480.84 W/g at 2 W/cm². While, the 2 W/cm² power density is significantly high (20 times the solar irradiance), this high laser density was used to facilitate comparison with existing photothermal studies in the literature, which commonly use power densities of 2–3 W/cm². Due to such impressive photothermal SAR, this work suggests that efficient tumor hyperthermia can be achieved under the safe power density limit of 0.33 W/cm² for 808 nm laser irradiation.

Bimodal (photo and magneto-) thermal studies

From studies of individual magnetothermal and photothermal performances of nanoflake, it is expected that these nanoflakes would give much enhanced SAR under bimodal (magnetothermal and photothermal) heating experimentation. To study this, the suspension (1 mg/mL) was simultaneously exposed to AMF of 100 kHz, 32 kA/m and IR laser (808 nm) with various laser power densities and temperature was monitored by fiber probe and corrected using IR thermal camera for accuracy. The schematic setup used for this purpose is illustrated in Fig. 6A. The Fig. 6B and C exhibit the amplified heating under bimodal stimulation and the calculated SAR values with changing laser power densities (0–2 W/cm²). It can be seen from Fig. 6B, under AMF alone and when there is no laser, the temperature change is 13.7 °C, that is persistent to Fig. 4C. When 0.1 W/cm² laser power density is used along with AFM, the temperature change reach to 28.4 °C. Compared with the use of AFM alone (ΔT = 13.7 °C) and laser alone (ΔT = 5.6 °C), it is clearly not a straightforward superposition effect. It can also be emphasized that the single heating mode, either AMF or laser (0.1 W/cm²) individually cannot be utilized for therapeutic treatment but the temperature change governed by bimodal experimentation is very enough to be used for therapeutic treatment. It further signifies that bimodal heating enables attaining the required thermal effects under lower dose and reduced laser power density. Also, the calculated SAR results in Fig. 6C validate that under AMF and no laser it is 807.5 W/g, (also referred to Fig. 3F), while for laser alone it is 681.3 W/g (referred to Fig. 4H). Under bimodal stimulation, the SAR values get 1836.9 W/g using same experimental parameters together and it is not simply additive but leveraged by complementary heating mechanism. The SAR values successively get to the highest value of 11857.53 W/g under AMF and increasing laser power density to 2 W/cm², particularly due to atomic level synergism of two heating effects. The calculation detail of amplified heating, leading to such high value of SAR under bimodal stimulation is illustrated in Figure S4 of Supporting Information, using Box-Lucas fitting curves. Compared with photothermal experimentation results in Figs. 4E and 5D, using 0.1 W/cm² alone cannot get to the required therapeutic temperature. Here this synergistic effect aids the temperature to rapidly get to the therapeutic temperature (48 °C) within 90 s from the room temperature (25 °C) under very low laser density of 0.1 W/cm² and AMF (100 kHz, 32 kA/m) that is already under the safety limit. Further, the SAR measurements of suspension (1 mg/ML) in DMEM supplemented with 10% FBS, over the time, up to 72 h for three heating modes reveals the excellent heating performance retention in biological medium as presented in Figure S2B of Supporting Information.

Fig. 6.

A Schematic setup used for evaluating the heating performance under bimodal stimulation. B Heating curves of nanoflakes suspension (Concentration: 1 mg/mL) under bimodal stimulations, magnetic field (100 kHz, 32 kA/m) and laser (808 nm) with different power densities (0–2 W/cm²). C Corresponding SAR variations with laser with laser power densities (n = 3), the red error bars showing deviation from mean values and inset image showing experimental scheme for bimodal heating evaluation. D Cytotoxicity analysis of nanoflakes: Cell viability of HaCAT cell co-cultured and exposed to nanoflakes under different concentrations (0, 0.25, 0.5, 0.75 and 1 mg/mL) for 24 h, 48 h and 72 h E Associated representative fluorescent microscopy images of co cultured cells. Data is presented as mean values with ± SD (n = 4) by error bars and statistical analysis comparison test for significant differences: (*) p < 0.05, (**) p < 0.01, (***) p < 0.001 and n.s. for no significance

The observed SAR values under bimodal stimulation, for most of the composite nanomaterials is just sum or average of SAR values of individual magnetothermal and photothermal effect due to indirect coupling of these two effects at nanoscale, leading to mere summation. The synergic effect seen here is attributed to direct nanoscale level coupling of these two effects under bimodal stimulation, it which photothermal temperature rise is expected to modulate the magnetic relaxation especially Brownian relaxation [9, 28, 82]. Conversely, at low power density when energy is not enough for frequent excitations, the energy from magnetic losses can increase electron transitions due to phonon couplings [83, 84]. The similar kind of synergism has been observed in some previously reported nanostructures but here both the unique morphology and Sc doping plays important role in achieving synergically enhanced performance, making bimodal modality more convincing. The performance evaluation of these nanoflake is done by conducting a brief comparison with individual and multimodal performance of previously reported nanomaterial, as summarized in Table S1 of Supporting Information. The enhancement of thermal energy dissipation under bimodal stimulation observed for such nanocrystalline with flakes morphology is because of the optical absorption and magnetic responses that are associated with the same atoms in lattice framework. Electronic orbitals responsible for light absorption interact with magnetic moments, leading to coupled energy states and effective photo-thermal-magnetic coupling [85, 86]. The optical and magnetic energy conversion enhances as energy conversion pathways are interlinked and mutually supporting each other, creating ultra-fast synergic heating effect [87]. The result is amplified magnetic relaxation rates enhanced by photothermal heat and conversely the magnetic heating induces the lattice vibrations thereby increasing phonon-assisted electronic transitions and heat conduction by improved lattice dynamics [83, 88]. The both energy conversion path ways are linked and support each other directly through phonons as explained by Debye–Callen–type model (in Supporting Information), that is unachievable in hybrid structures due to obvious interface thermal boundary resistance and energy losses [89, 90]. So, the bimodal heating is synergically enhanced as compared to individual single modality, thereby overcoming the limitations of each and providing full potential for in vitro and in vivo applications. As observed from above results, the nanoflakes exhibit the best performance, supported by high SAR for both magnetothermal and photothermal effects and are promising for high temperature hyperthermia or thermal ablation of tumors, So, in vitro cell experimentation is conducted using these to study the cytotoxic and hyperthermia effect under standardized culture conditions.

In vitro cytotoxicity studies

To ensure the biocompatibility is crucial for both in vitro or in vivo applications, so HaCaT cell line was employed to evaluate the baseline cytotoxicity with human keratinocytes (non-cancerous). For this purpose, HaCaT cells were co-cultures with nanoflakes under varying concentrations of 0 mg/mL (control), 0.25 mg/mL, 0.5 mg/mL, 0.75 mg/mL and 1 mg/mL. The cell viability measured by CCK-8 essay, assessed after 24, 48 and 72 h and the results are illustrated in Fig. 6D. The results approve that the cell viability is more than 90% after 72 h, even at maximum concentration of 1 mg/mL, signifying the good biocompatibility and negligible toxicity. It can be seen in Fig. 6E that most of the cells remain viable in representative images and only a few shows the symptom of death. These results are persistent with the CCK-8 essay’s results, conferring the intrinsic low cytotoxicity of nanoflakes. Although Fe-based nanomaterial may induce toxicity (ferroptosis), but these nanoflakes are non-toxic, probability due to surface morphology and chemistry (residual functional groups) that could help to alleviate the oxidative stresses associates with such ferrite nanostructures [91], as well as Sc doping has not induced any toxicity at all.

In vitro hyperthermia studies

The in vitro tumor killing effect of nanoflakes was evaluated with A375 (human melanoma) cell line using 1 mg/mL concentration of nanoflakes with cultured cells. This minimal constant extracellular concentration (1 mg/mL) was chosen based on previously established heating profiles under different stimulation modes. The cultured cells were subjected to: (i) controlled conditions (Control without laser/AMF), (ii) 808 nm laser at power density of 0.1 W/cm² (LASER), (iii) AMF of 100 kHz and 32 kA/m (AFM) and (iv) laser and AMF simultaneously at the same conditions (LASER + AMF). The heating curves tested with in the cell culture medium for all three treatment modes are presented in Fig. 7A. For second group (LASER only), the temperature elevation of 7.4 °C was observed compared to control group, reaching to 32.4 °C from room temperature (~ 25 °C) after 480 s (8 min). The lower temperature rise at 0.1 W/cm² power density could be attributed to the fact, that due to less energies of photons the electronic transitions are not significant in numbers so heating effect is not sufficient. For third group (AMF only), the maximum stable temperature reaches to 40.3 °C from room temperature, but still this temperature could not enter the therapeutic temperature window for mild hyperthermia (42–46 °C). It is because of the AMF parameters (100 kHz, 32 kA/m) that are much less than safety limit (H×f < 9.59 × 10⁹ Am⁻¹s−¹) and nanoflakes concentration could be not enough for large scale magnetic couplings and relaxations. Finally, for fourth group, under simultaneous LASER + AMF subjection, very fast amplified heating effect was observed reacting the temperature of 53.4 °C after 240 s and 55.5 °C after 480 s. The IR temperature field distribution images show the temperature of controlled cultured cells relative to the temperature of other cell groups exposed under three different modes after 480 s in Fig. 7B. Cell viability measured from CCK-8 essays for various heating treatment groups are presented in Fig. 7C. The group without laser and AMF shows no effect or toxicity with control group, accomplishing that this concentration is very safe. For other groups, one with laser and another with AMF, the effect is very similar i.e., the cell viabilities are > 95%, as the temperature does not reach to the minimal hyperthermia temperature (~ 42 °C), in order to start the cell deaths by apoptosis. Flow cytometry analysis using Annexin V-FITC/PI staining was performed to evaluate the cell death mode combined AMF + LASER treatment for control, 0, 4, and 8 min respectively. The flow cytometry data presented in Fig. 7D(i-iv) revealed the dominant apoptotic phenotype (Annexin V⁺/PI⁺ population of 88.6%), consistent with stress induced apoptosis following severe thermal shock (55 °C) for the combined AMF + LASER treatment group. It has acknowledged that the temperature of 55 °C exceeds the threshold for immediate membrane permeabilization, and literature reports necrosis as the primary mode at such temperatures [92, 93]. The observed apoptosis results represent a delayed execution pathway in cells that did not undergo immediate lysis, while the immunogenic response can be driven by early necrotic/necroptotic cells that released damage-associated molecular patterns (DAMPs, ATP, calreticulin) during the initial thermal shock. At 55 °C for 8 min, the thermal dose is sub-lethal for immediate necrosis but supra-threshold for triggering the mitochondrial apoptotic cascade [94]. The cells do not die during heating but they die after heating via a delayed, programmed execution. Thus, the death mechanism is bimodal: rapid necrosis in a subpopulation indicated by 4 min results (Annexin V⁻/PI⁺ population of 14.7%) and delayed apoptosis in surviving cells. Hyperthermia at 55 °C for 8 min can induce a mixed death response, with flow cytometry indicating late apoptotic execution (dominant form of cell death) in the majority of cells, but early necrotic cell death mode can drive the immunogenic anti-tumor effects. So, the 8 min group exhibits a higher apoptosis than the 4 min group. These results correlates with fluorescent staining images presented in Fig. 7E, as for first three groups (referring to Fig. 7C), there are no death signs observed in cells respective to control group, while for fourth group (LASER + AMF), the maximum dead cell can be observed reducing the cell viability to minimal level [95]. The cell viability starts decreasing as after 4 min, it reduces to 63% and finally 7% after 8 min as illustrated in Fig. 7F. These results are also supported by the fluorescent staining image in Fig. 7G showing half of cell deaths after 4 min compared to control group (0 min) and maximum cell deaths (red cells) after 8 min. These results suggest that the bimodal stimulation (LASER + AMF) can ablate the tumor rapidly using smallest dosage and without surpassing the safety limits of each mode individually, that is crucial for avoiding long term health concerns for clinical applications.

Fig. 7.

In vitro hyperthermia studies of nanoflakes for killing of tumor cells: A Heating curves under LASER (808 nm, 0.1 W/cm²), AMF (100 kHz, 32 kA/m) and bimodal (LASER + AMF) of co-cultures A375 cells in medium (Concentration: 1 mg/mL, Exposure time: 8 min). B Corresponding temperature field distribution images of well plates containing various groups of co-cultured cells at the end exposure duration. C Cell viability assessed for each heat treatment group after end of exposure. D Flow cytometry result to determine cell death mode (i) control group (ii-iv) under LASER + AMF treatment duration of 0 min, 4 min and 8 min, respectively. E Representative live-dead fluorescent images of each heat treatment group after end of exposure. F Cell viability over the time under bimodal stimulation. G Corresponding live-dead fluorescent images for 0 min, 4 and 8 min exposure. Data is presented as mean values with ± SD (n = 4) by error bars and statistical analysis comparison test for significant differences: (*) p < 0.05, (**) p < 0.01, (***) p < 0.001 and n.s. for no significance

Conclusion

To sum up, we synthesized the Sc doped Fe₃O₄ nanoflakes with high saturation magnetization and photothermal conversion efficiency that is promising for enhancement of SAR under both magnetothermal and photothermal effects. The proposed nanoflakes under bimodal stimulation demonstrate the synergically and unprecedently enhanced SAR of 11857.53 W/g, when subjected to AMF (32 kA/m, 100 kHz) and laser (808 nm, 2 W/cm²). This highest SAR enables to conceive tumor ablation with minimal concentration/dosage under very mild conditions of AMF and laser with in the biological safety limits. The nanoflakes also exhibit very good biocompatibility with HaCAT cell line and the in vitro studies showed the successful tumor ablation of A375 cell line under high temperature hyperthermia with bimodal stimulation (808 nm, 0.1 W/cm² and 100 kHz, 32 kA/m), which is not achievable with using either modality individually, reducing the cell viability 7% over shorter duration and minimal concentration (1 mg/mL). These observations stress the potential of bimodal tumor therapy compared to classical individual tumor therapies for safer clinical applicability of hyperthermia.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Ihtisham Ahmad Butt: Conceptualization, Writing—original draft, Data curation, Methodology, Experimentation, Investigation and Validation. Naihan Chen: Experimental validation, Formal analysis, Writing—review and editing. Wei Zhang: Investigation, Funding acquisition, Writing—review and editing, Project administration, Supervision. All authors contributed to manuscript preparation and discussion of the results.

Funding

This work was financially supported by the National Key Research and Development Project of China (2022YFE0115400, 2018YFA0704103, 2018YFA0704104) and National Health Commission of China (HDLH2514).

Data availability

Data will be available from the authors upon request.

Declarations

Competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.