Gold acetate (purity >99.5%), ferric acetate (purity >99.8%), acetonitrile (purity >99.5%), and 2-ethylhexanoic acid (purity >99.5%) were obtained from Sigma-Aldrich Co., Ltd., and used as received without further purification. Ethylene (C₂H₄, purity >99.9%) and oxygen (O₂, purity >99.9%) were obtained from Nanjing Shangyuan Gas Factory. Other chemicals were of reagent grade from Aldrich.

Multicomponent Au/Fe₃O₄@C NGs were fabricated via enclosed flame spray pyrolysis. Typically, gold acetate with a concentration of 0.2 mol/L and ferric acetate with a concentration of 0.2 mol/L were dissolved into the solvents of acetonitrile and 2-ethylhexanoic acid (1:1 of volume) under constant stirring. This precursor solution of Au/Fe₃O₄ was injected into the reaction chamber with a feed speed of 5 sccm. O₂ gas with a flow of 50 sccm was applied to sheathe this solution. The enclosed flame spray pyrolysis was ignited in a quartz glass tube (40 mm diameter), which was preheated via adding xylene for 5 min before the particle fabrication process. A premixed gas of C₂H₄/O₂ (4.0 L/min with three specific ratios) was utilized as the carbon source. The as-prepared NGs were collected with the assistance of a glass fiber filter. The content of coating carbon in Au/Fe₃O₄@C NGs was regulated to the given values by changing the ratio of C₂H₄ in the C₂H₄/O₂ gas mixture.

Transmission electron microscopy (TEM) was carried out via a Hitachi H-7100 instrument. X-ray diffraction (XRD) was carried out via a D8 Advance X-ray diffractometer (Bruker, Germany). Energy-dispersive X-ray spectroscopic (EDS) line profiles were recorded via a scanning electron microscope (Hitachi SU8700). X-ray photoelectron spectroscopy (XPS) was applied to analyze the surface composition via an ESCALAB 250 device. Raman analysis was performed via a Renishaw inVia microprobe. The magnetic measurements were conducted at 300 K via a commercial Physical Property Measurement System (PPMS-9) (Quantum Design, United States) equipped with a vibrating sample magnetometer (VSM). The T₂-weighted MRI was recorded via an MRI device (T₂) by using a 1.5 T SIGNA Voyager MRI device. The light source at 808 nm was derived from a near-infrared radiation laser device (DBAL-I2). Infrared radiation images with various temperatures were recorded via an infrared radiation camera (FLIR T640-45) at regular time intervals.

The cell culture medium used was DMEM medium, 10% fetal bovine serum albumin, and 1% penicillin–streptomycin double antibody solution. A 20 mg/L Au/Fe₃O₄@C NG suspension was added to the coverslips to incubate esophageal cancer cells for 3 h. Cells were cultured in a CO₂ incubator at 37°C with a CO₂ concentration of 5%. Newly prepared LIVE/DEAD solution obtained from Thermo Fisher Scientific was applied to rinse the cells several times. Then, laser radiation (wavelength 808 nm, power density 0–2.0 W/cm²) was applied to radiate the esophageal cancer cells. The cell images were recorded on a ZEISS Axiocam microscope camera.

A homogeneous alternating current magnetic field with a frequency of 300 kHz and an amplitude of 200–800 Oe was applied for the magnetic hyperthermia of Au/Fe₃O₄@C NGs via a calorimetric approach. In addition, photothermal experiments induced by laser radiation (808 nm of wavelength, 0–2.0 W/cm² of power densities) were carried out in combination with magnetic hyperthermia.

Au/Fe₃O₄@C NGs prepared with a mixed gas of C₂H₄/O₂ (4.0 L/min with three specific ratios including 3.9/0.1 L/min, 3.5/0.5  L/min, and 3.0/1.0 L/min) were termed as Au/Fe₃O₄@C-1 to -3 NGs. Au/Fe₃O₄@C NGs were composed of Au NPs and Fe₃O₄ NPs separately coated with carbon shells (Figures 1A–C). Fe₃O₄ core sizes changed from 51.9 nm to 55.2 nm (Figures 1D, E). The PDI values for Au/Fe₃O₄@C-1 NGs, Au/Fe₃O₄@C-2 NGs, and Au/Fe₃O₄@C-3 NGs were 0.572, 0.593, and 0586, respectively. The thickness of the carbon shell to coat the Fe₃O₄ core decreased from Au/Fe₃O₄@C-1 to Au/Fe₃O₄@C-3. Carbon-encapsulated Au NPs with a shell thickness of 2–4 graphite layers were located and attached with carbon-encapsulated Fe₃O₄ NPs. Au core sizes changed from 4.68 nm to 8.75 nm. Carbon coating shells with 2–4 layers would exert a slight shielding effect toward Au cores. Namely, the plasmonic property of carbon-encapsulated Au NPs should be dominated by Au cores, which may be favorable for the photothermal performance of Au/Fe₃O₄@C NGs.

The formation process of Au/Fe₃O₄@C NGs follows: When the enclosed flame spray pyrolysis started, the precursors, including gold acetate, ferric acetate, and carbon source C₂H_4, would be decomposed by the enclosed flame spray (Li et al., 2017; Mostafa et al., 2021). The species of Au and Fe would coalesce to form NGs as Fe₃O₄/Fe₃C-Au owing to their collision. The carbon species would react with these nanoclusters as a dissolution in Fe NPs and a deposition on the surface of Au NPs (low solubility for carbon). Considering the high temperature of enclosed flame spray, carbon deposited on the surface of Au NPs would form graphite coating shells over Au NPs because of the catalysis of Au for carbon. Graphite coating shells over Fe₃O₄ NPs would also be formed by the re-precipitated carbon from the Fe-based cores. Finally, Au/Fe₃O₄@C NGs would be formed.

EDS spectra (Figure 2A) confirmed that the typical elemental compositions of Au/Fe₃O₄@C NGs were Au, Fe, C, and O without any other impurities. The Au content was 24.4 wt% for Au/Fe₃O₄@C-1, 33.3 wt% for Au/Fe₃O₄@C-2, and 42.8 wt% for Au/Fe₃O₄@C-3, while those of Fe element were 31.3 wt% for Au/Fe₃O₄@C-1, 29.7 wt% for Au/Fe₃O₄@C-2, and 23.9 wt% for Au/Fe₃O₄@C-3. The crystalline structures of Au/Fe₃O₄@C NGs analyzed by XRD are shown in Figure 2B. A diffraction peak located at 26.5° was derived from the (002) crystalline reflection of graphite. Characteristic peaks at 38.5° ((111) reflection) and 44.7° ((200) reflection) demonstrated the existence of Au (JCPDS no. 89-3697) in Au/Fe₃O₄@C NGs (Chang et al., 2021). Meanwhile, peaks ascribed to Fe₃O₄ (JCPDS no.65-3107) were also observed, which verified that the Au/Fe₃O₄@C NGs were composed of three components, including carbon, Au, and Fe₃O₄ (Wang Z. et al., 2021). Given the localized surface plasmon resonance property of nanosized Au in Au/Fe₃O₄@C NGs, characteristic absorption bands were displayed for Au/Fe₃O₄@C-1 at 512 nm, for Au/Fe₃O₄@C-2 at 515 nm, and for Au/Fe₃O₄@C-3 at 519 nm (Figure 2C). With the increasing carbon component from 13.2 wt% to 32.3 wt%, the localized surface plasmon resonance induced by plasmonic coupling was broadened, and the absorption intensity was reduced. These broad optical absorption peaks were located in the visible and infrared regions, which allowed us to assess their applicability in photothermal therapy. As shown by the magnetic hysteresis loops of Au/Fe₃O₄@C NGs displayed in Figure 2D, the values of saturation magnetization of Au/Fe₃O₄@C NGs rose gradually, that is, 33.7 emu/g (58.6 Oe of coercivity), 39.7 emu/g (62.9 Oe of coercivity), and 48.2 emu/g (75.4 Oe of coercivity) for Au/Fe₃O₄@C-1 to -3 NGs, separately. Enclosed flame spray pyrolyzed multicomponent Au/Fe₃O₄@C NGs showed weak ferromagnetism because of the presence of the Fe domain in the NGs (Li et al., 2021; Li et al., 2022). The zeta potentials of Au/Fe₃O₄@C-1 NGs, Au/Fe₃O₄@C-2 NGs, and Au/Fe₃O₄@C-3 NGs at pH 5.0–7.0 were measured, as shown in Supplementary Figure S2 in the Supplementary Material. For instance, the zeta potentials of Au/Fe₃O₄@C-1 NGs, Au/Fe₃O₄@C-2 NGs, and Au/Fe₃O₄@C-3 NGs at pH 7.0 were −0.22, −0.34, and −0.19, respectively. Moreover, the EDS mapping of the nanoformulation image has been provided, as shown in Supplementary Figure S3, to reveal the actual distribution of elements in the nanoformulations.

To analyze stability, 100 mg/L of Au/Fe₃O₄@C-3 NGs was obtained from 100 μg of Au/Fe₃O₄@C-3 NGs resuspended in 1.0 mL of deionized water, and the absorption was further measured at pre-set time points (0 h, 12 h, 24 h, and 36 h). As shown in Supplementary Figure S4A, the characteristic absorption spectra of Au/Fe₃O₄@C-3 NGs at 517 nm displayed a negligible change over this time period. There was little change in the size of Au/Fe₃O₄@C-3 NGs after this time period (Supplementary Figure S4B).

The application of MRI agents based on iodine is restricted because of their relatively low X-ray absorption coefficient. Magnetic materials containing an Fe component have presented superior MRI performance and can be applied as a computerized tomography (CT) contrast agent. Figure 3A showed the effect of Fe concentration in Au/Fe₃O₄@C NGs on the transverse proton relaxation time (1/T₂) of CT signals, where the Fe concentration was measured via an inductively coupled plasma atomic emission spectrometry method. Relaxation time 1/T₂ was linearly correlated with increasing Fe concentration in Au/Fe₃O₄@C NGs. Specifically, as the concentration of Fe increased from 0.05 mM to 0.8 mM, the values of 1/T₂ increased from 7.2 s⁻¹ to 39.7 s⁻¹, 57.4 s⁻¹, and 76.0 s⁻¹ for Au/Fe₃O₄@C-1 to -3 NGs, separately. Figure 3B shows CT images of Au/Fe₃O₄@C NGs as a function of Fe concentration in Au/Fe₃O₄@C NGs. The values of measured relaxivity (r₂) were 43.5 mM⁻¹ s⁻¹, 66.9 mM⁻¹ s⁻¹ and 91.4 mM⁻¹ s⁻¹ for Au/Fe₃O₄@C-1 to -3 NGs separately. The MRIs of Au/Fe₃O₄@C NGs tended to darken as the concentration of Fe increased from 0.05 mM to 0.8 mM. Considering these MRI results, Au/Fe₃O₄@C NGs could be applied as an excellent agent for the potential CT application.

Given the efficient absorbance of Au/Fe₃O₄@C-3 NGs in the region of UV-Vis to NIR, the effect of Au/Fe₃O₄@C-3 concentration from 0 mg/mL to 0.8 mg/mL on the photothermal performance of Au/Fe₃O₄@C-3 solutions was investigated with laser radiation (wavelength 808 nm, power density 1.0 W/cm², and duration 300 s). An obvious concentration-dependent temperature enhancement was observed for the aqueous suspensions of Au/Fe₃O₄@C-3 NGs (Figure 4A). The temperatures of 0.1 mg/mL, 0.2 mg/mL, 0.4 mg/mL, and 0.8 mg/mL of Au/Fe₃O₄@C-3 NGs suspensions reached 34.5°C, 41.9°C, 52.8°C, and 65.4°C, respectively, after only 300 s irradiation. The effect of power density (0.5–2.0 W/cm²) on the photothermal performance of Au/Fe₃O₄@C-3 solutions (0.2 mg/mL) is shown in Figure 4B. Distinct laser-power-dependent photothermal behaviors of Au/Fe₃O₄@C-3 suspensions are displayed; that is, the temperatures of Au/Fe₃O₄@C-3 suspensions reach 34.3°C, 40.7°C, 47.1°C, and 55.6°C with the increased power densities after only 300 s of laser irradiation. The thermal energy converted from the laser was positively related to both the concentration of Au/Fe₃O₄@C-3 NGs and the power densities of the laser. Moreover, the stability of this conversion by Au/Fe₃O₄@C-3 NGs was evaluated for four cycles with a laser radiation duration of 300 s (Figure 4C). The amount of temperature increase of the aqueous suspension changed only slightly, indicating that the stability of photothermal conversion by Au/Fe₃O₄@C-3 NGs was excellent.

Enclosed flame spray pyrolyzed Au/Fe₃O₄@C-3 NGs were applied as a photothermal tumor-ablation agent to a mixture of esophageal cancer cells and Au/Fe₃O₄@C-3 NGs. The conditions for the laser radiation are wavelength 808 nm, power density 2.0 W/cm², duration 5.0 min, and spot size 5 mm. The fluorescent images of cell viability shown in Figure 5 indicate that green was utilized to mark the cell, while red color was used to mark dead cells. The temperature of esophageal cancer cells in the absence of laser radiation was 19.3°C (Figure 5A) as a comparison. The esophageal cancer cells incubated with Au/Fe₃O₄@C-3 NGs (50 mg/L) without any laser irradiation and the esophageal cancer cells alone with laser irradiation are shown in Figures 5B, C; the temperatures reached 20.6°C and 21.4°C, respectively. The viability of esophageal cancer cells remained stable with the addition of Au/Fe₃O₄@C-3 NGs in the absence of laser irradiation, indicating the good biocompatibility and weak toxicity of Au/Fe₃O₄@C-3 NGs. Meanwhile, the temperature of esophageal cancer cells without the addition of Au/Fe₃O₄@C-3 NGs increased lightly due to the laser radiation but presented no obvious cell damage. The above phenomena indicate that such treatments do not compromise cell viability (Cerezo-Navarrete et al., 2022; Chu et al., 2022).

In contrast, the slide temperature in the presence of Au/Fe₃O₄@C-3 NGs rose to 70.3°C (Figure 5D), inducing a significant killing of esophageal cancer cells by the laser. It is demonstrated that the enclosed flame spray pyrolyzed Au/Fe₃O₄@C-3 NGs could be an efficient photothermal agent to produce significant environmental energy from the laser, which could kill the esophageal cancer cells without obvious impact on the healthy tissues. In vitro particle uptake experiments were carried out. Au/Fe₃O₄@C-3 NGs were incubated with esophageal cancer cells for 24 h at an extracellular iron concentration of 100 × 10⁻⁶ mol/L (Table 1). The 24 h incubation at 100 × 10⁻⁶ mol/L of Fe resulted in a mass of iron uptaken by cells of 3.0 ± 0.4 pg of Fe per cell.

To restrain the physical motion of Au/Fe₃O₄@C-3 NGs, Au/Fe₃O₄@C-3 was added into agar with a mass concentration of 2.0 wt%. The temperatures of magnetic hyperthermia with different external alternating current applied magnetic fields were measured as shown in Figure 6A. The heating temperature increases with the applied magnetic field, that is, 33.7°C for 200 Oe, 37.6°C for 300 Oe, 41.5°C for 400 Oe, 46.2°C for 500 Oe, 49.1°C for 600 Oe, 56.9°C for 700 Oe, and 59.3°C for 800 Oe. Because the therapeutic temperature window is located at temperatures ranging from 40°C to 44°C, which enable damaging cancer cells while preserving healthy cells, a magnetic field >400 Oe should be applied in magneto-calorific experiments.

Moreover, the magnetic field dose should be kept as low as possible in the practical application considering its negative impact on normal cells (Mai et al., 2022; Delgado et al., 2023; Myrovali et al., 2023; Wang et al., 2023). Because both the photothermal temperature and magnetic-heating temperature of Au/Fe₃O₄@C NGs were positively related to the laser power density and magnetic field value separately, it is contradictory to improve the efficiencies of photothermal therapy and magnetic hyperthermia without introducing injury to the cell tissues. Given that, we applied the laser radiation and magnetic field simultaneously to investigate the heating performance of Au/Fe₃O₄@C NGs. As the heating curves shown in Figures 6B, C, the temperatures under combined external conditions are obviously higher than those obtained without laser irradiation. For instance, the obtained temperature after 5 min under the magnetic field of 400 Oe increases from 30.6°C to 41.5°C and 47.4°C with the combination of laser exposure (power density: 0.5 W/cm² and 1.0 W/cm² separately). Without applying the laser irradiation, the required value for the magnetic field to enable the target temperature of 40°C was 600 Oe. After adding laser exposure (power density: 0.5 W/cm² and 1.0 W/cm² separately), the required magnetic field value decreases to 400 and 300 Oe, respectively. Thus, it is believed that the usage of these two external stimuli showed a higher heating efficiency than the situation under separate stimuli. The above findings demonstrated that Au/Fe₃O₄@C NGs combined with plasmonic and magnetic components could be promising hyperthermia agents with reduced required laser radiation and magnetic field doses. Furthermore, carbon coating shells over Au/Fe₃O₄ NGs would present a favorable biocompatibility because carbon materials are generally tolerated by organisms.