Introduction

Hybrid nanostructures (HNs) composed of materials with different physicochemical properties have been widely investigated in recent years since they allow multiple functionalities to be integrated into a single object, a feature that is difficult to achieve by simply mixing the individual constituents1. In particular, special attention has been given to HNs composed of gold (Au) and magnetite (Fe3O4), as they exhibit plasmonic and magnetic properties that make this particular class of heteronanostructure of interest for potential applications in various fields, such as biomedicine2,3,4,5,6, solar energy conversion7,8 and environmental remediation9,10,11. Therefore, several methodologies have been developed to prepare NHs composed of gold and magnetite with narrow shapes and size distributions. The synthesis of this class of HNs is generally performed following a seed-mediated approach whereby secondary nanoparticles (NPs) nucleate and grow on the surface of seeds through heterogeneous nucleation12. According to the specific experimental conditions, core–shell or Janus morphologies were obtained13. The methods for preparing Au-Fe3O4 HNs can be roughly classified according to the solvent employed. Organic-phase routes are generally based on the thermal decomposition of an organometallic iron precursor on the surface of preformed Au NPs, where HNs are formed through epitaxial growth of the iron oxide phase on the Au seeds. The ratio of the Au seed to the iron precursor, the surfactant amount and the reaction temperature are key experimental variables that must be carefully controlled to ensure that the threshold for the homogeneous nucleation of Fe3O4 is not reached throughout the synthesis and therefore to boost the heterogeneous nucleation mechanism. In this regard, Sun and coworkers developed a successful methodology to obtain dumbbell-like bifunctional Au-Fe3O4 HNs, the growth of which was affected by the polarity of the solvent14. Shi et al. reported a general thermal decomposition approach using Fe(acac)3 as the iron precursor for obtaining Au-Fe3O4 HNs with diverse morphologies and narrow size distributions15. Mezni et al. proposed a one-pot approach using triethylene glycol as a solvent and reducing and stabilizing agent that produces mainly Au-Fe3O4 HNs with a core–shell morphology16. Usually, the materials obtained following these organic phase routes are capped with hydrophobic ligands such as oleic acid or oleylamine, so they exhibit low solubility in water, which limits their potential applications17.

The use of less expensive and environmentally friendly solvents has positive implications from both an economic and a sustainability point of view. Therefore, several approaches have been reported for synthesizing Au-Fe3O4 HNs in aqueous media. The synthesis of NHs in an aqueous medium holds promise for replacing organometallic precursors such as iron pentacarbonyl or iron acetylacetonate by aqueous iron salts such as sulfates, chlorides, and nitrates. These salts are non-volatile and less toxic compared to organometallic precursors. Additionally, there are ongoing investigations into replacing organic solvents and stabilizers used in the organic route, such as diphenyl ether, 1,2-hexadecanediol, benzyl ether, triethylene glycol by water. Moreover, these alternative methodologies allow for mild reaction conditions, unlike the thermal decomposition route which necessitates high-boiling organic solvents, temperatures exceeding 200 °C, long reaction times, and sophisticated equipment. This approach generally relies on the heterogeneous nucleation and growth of the Au component on the surface of preformed Fe3O4 NPs that act as seeds, resulting mostly in HNs with Janus or core-satellite morphologies18. For instance, Zeng et al. proposed a general seed-mediated approach to the synthesis of Au-Fe3O4 heterodimers in aqueous media by taking advantage of the reductive nature of Fe3O4 to initiate the reduction of metal cations and deposition of a metal seed on the surface of each Fe3O4 NP before the metal seed is grown to the desired size through a seed-mediated process19. Xing et al. reported a similar two-step method for synthesizing Au-Fe3O4 HNs, which consists of the reduction of HAuCl4 by citrate in the presence of citrate-functionalized Fe3O4 NPs20. However, in contrast with the numerous efforts made to develop methods following organic routes, studies reporting approaches for the synthesis of Au-Fe3O4 HNs in aqueous media using the Au component as a seed are scarce in the literature. Although several methodologies have been established for synthesizing iron oxides in aqueous media, coprecipitation and partial oxidation methods are the most commonly used methods for the preparation of Fe3O4 NPs21,22. According to the coprecipitation method, particles with sizes typically approximately 10 nm are obtained23. On the other hand, following the partial oxidation method, cubic nanometric crystals (30–100) nm or larger spherical particles (0.4–1.1) µm can be obtained depending on the experimental conditions, as O2 or nitrate ions are the oxidizing agents usually employed24,25. In this process, the initial ferrous hydroxide is transformed into magnetite by means of a dissolution–oxidation–precipitation mechanism26. Increasing industrial-scale production of nanostructures could raise concerns about their environmental impact. Comprehensive assessments would be required to address their proper handling, storage and disposal to maximize their benefits and minimize any potential negative impacts, such as the release of nanoparticles into the environment and their potential effects on ecosystems. In this context, the industrial production of nanostructures in aqueous media is highly convenient.

Herein, a partial oxidation-based approach for the synthesis of Au-Fe3O4 HNs in aqueous media is proposed. This method relies on the partial oxidation of the Fe(II) precursor using hydrogen peroxide as the oxidizing agent in the presence of Au NPs dispersed in the reaction medium. The synthesized materials were characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDS) and UV‒Vis spectroscopy. The concentration of the Fe(II) precursor and the value of the parameter R (R = [(OH)−]/[Fe(II)]) were varied to address their effects on the phase, morphology and optical response of the synthesized nanostructures. To achieve deeper insight into the structure‒optical property relationship, extinction spectra were simulated according to the discrete dipole approximation (DDA) method.

Methods

Chemicals

Ferrous sulfate heptahydrate (FeSO4·7H2O, Cicarelli), sodium hydroxide (NaOH, Cicarelli), hydrogen peroxide (H2O2, Anedra), sodium citrate (Na3C6H5O7, Anedra), and tetrachloroauric acid trihydrate (HAuCl4·3H2O, Sigma‒Aldrich), high purity grade reagents, were used as received without further purification. Ultrapure water with a resistivity of 18.2 MΩ cm was used for the preparation of all aqueous solutions.

Synthesis of Fe3O4 NPs

In a typical experiment, the following substances were added sequentially to a three-necked round-bottom flask containing a given amount of FeSO4·7H2O under continuous bubbling of N2(g): 50 mL of a deoxygenated NaOH solution, 50 mL of deoxygenated water, and 50 mL of a deoxygenated H2O2 solution. After all the solutions were added to the reaction vessel, the mixture was left at boiling temperature for an hour under continuous bubbling of N2(g). Prior to mixing the reactants, the dissolved oxygen was removed by bubbling N2(g) through the respective solutions for 30 min. The mass of FeSO4.7H2O was attained at the end of the mixing process at concentrations of 5 × 10–5 M or 1 × 10–4 M. The initial NaOH concentration was adjusted to obtain values of parameter R of 1.3, 2, and 3 after all the substances were added. The H2O2 concentration of the 50 mL of deoxygenated solution was adjusted to 5 × 10–5 M or 1 × 10–4 M depending on whether the final FeSO4 concentration was 5 × 10–5 M or 1 × 10–4 M, respectively. The rate of addition of the solutions was 100 mL/s, unless otherwise specified. A schematic illustration of the experimental setup employed is provided in Supplementary Fig. S1.

Synthesis of Au NPs

Au NPs were synthesized according to Turkevich’s method using sodium citrate as the reducing agent and HAuCl4 as the Au precursor27. In particular, the Au NPs were prepared by adding 4.1 mL of a 0.0254 M HAuCl4 solution and 10.0 mL of a 9.95 × 10–3 M sodium citrate solution to 400 mL of boiling water under magnetic stirring. The heating was turned off when the appearance of the typical ruby-red color was observed after approximately 15 min of reaction. The obtained Au NPs were purified by centrifugation at 1790 RCF for 30 min, after which the pellets were dispersed in ultrapure water. Based on the Lambert–Beer law, the concentration of the purified Au NP suspension was estimated to be 1 × 10–12 M (see the Supplementary Information and Supplementary Fig. S2 for further details).

Synthesis of Au-Fe3O4 HNs

The procedure for the synthesis of Au-Fe3O4 HNs was the same as that described above for Fe3O4 NPs but 50 mL of deoxygenated water was replaced with 50 mL of a deoxygenated 1 × 10–12 M Au NP suspension.

Characterizations

The XRD instrument employed was a PANalytical-X’Pert Pro X-ray diffractometer using Cu Kα radiation (λ = 1.5406 Å). The diffractograms were obtained in the range of 7° − 70° from powders of the particles. To obtain the respective powders, the suspensions containing the reaction products were washed three times with ultrapure water to remove salts formed by counterions such as Na2SO4, and the precipitated particles were obtained by depositing droplets containing the nanostructures over a glass slide and evaporating the solvent in a fume hood at room temperature. The crystallite size was determined by applying the Scherrer equation to the (220), (311), and (400) reflections; instrumental broadening was considered to be equal to 0.02. TEM images of the prepared nanostructures were obtained by using a JEOL 1120 EXII JEM under an accelerating voltage of 80 kV, whereas the HAADF-STEM and HRTEM images were obtained by using a JEOL JEM-2100 plus under an accelerating voltage of 200 kV. The samples were prepared by adding one drop (~ 30 μL) of the colloidal dispersion onto a holey carbon-formvar-coated copper TEM grid (100 mesh). The analysis of the chemical composition of the synthesized nanostructures was performed by EDS (Oxford). The extinction spectra were measured at room temperature using a Shimazdu UV-1700 Pharma Spec spectrophotometer in the range of (300–900) nm with a cuvette with a 1 cm optical path length.

Computational methods

The extinction efficiency spectra of the Au NPs, Fe3O4 NPs and Au-Fe3O4 HNs were simulated using both the discrete dipole approximation (DDA) method and Mie theory. The DDA method is implemented in the DDSCAT 7.3 code28. Further details about this methodology can be found elsewhere29,30. A lattice spacing (interdipole distance) of 1 nm was set in all the simulations to model the Au NPs, Fe3O4 NPs and Au-Fe3O4 HNs. In previous work, we have shown that this interdipole distance is small enough to achieve precise results31. Simulations based on the Mie theory, which is implemented in the BHMIE code32, were performed to further verify that a lattice spacing of 1 nm allows us to reach an acceptable error tolerance. The wavelength-dependent dielectric constants of Fe3O4 and Au were obtained from the literature33,34. The refractive index of the environment was set to 1.33, corresponding to water. To simulate the extinction efficiency spectra, the median sizes of the Fe3O4 NPs and Au-Fe3O4 HNs were obtained from the statistical analysis of the TEM images.

Results and discussion

The crystalline phase of the synthesized nanostructures was characterized through XRD measurements. Figure 1, curve a, shows the XRD pattern of the material obtained for R = 2 and [Fe(II)] = 1 × 10–4 M. For comparison, the reference data of Fe3O4 (JCPDS 01–089-0691) are shown at the bottom as red bars. The diffraction pattern of the obtained material (curve a) shows peaks at 2θ = 18.3, 30.2, 35.6, 37.3, 43.2, 53.7, 57.2, and 62.8, which are assigned to the reflections of the (111), (220), (311), (222), (400), (422), (511), and (440) planes of Fe3O4, respectively. In addition, no diffraction peaks attributed to other iron oxide or oxyhydroxide phases are observed. The average crystallite size D of the synthesized Fe3O4 was calculated by applying the Scherrer equation to three diffraction peaks (2θ = 30.2, 35.6, and 43.2), resulting in a D value of (21.9 ± 0.4) nm. The calculation of the unit cell parameter resulted in a value of a = (8.3832 ± 0.0008) Å, which agrees well with the tabulated value (a = 8.396 Å). The calculated a value not only supports the identification of the iron oxide phase formed but also allows us to estimate the stoichiometry. In general, partially oxidized or nonstoichiometric magnetite can be expressed as Fe3-δO4, where δ = 0 corresponds to stoichiometric magnetite and δ = 1/3 corresponds to maghemite (γ-Fe2O3, tabulated unit cell parameter value a = 8.347 Å)35. In this regard, Yang et al. reported that the unit cell parameter a decreases by 0.20 Å per vacancy δ36. Considering this relationship, a value of δ = 0.064 was estimated for the synthesized material, whose unit formula might be Fe2.936O4. These results indicate that under the abovementioned experimental conditions, the partial oxidation method implemented using H2O2 as the oxidant leads to Fe3O4 NPs with good stoichiometry as the only reaction product.

Figure 1
figure 1

XRD patterns of (a) Fe3O4 NPs and (b) Fe3O4-Au HNs. The reference XRD patterns of Fe3O4 (JCPDS 01-089-0691) and Au (JCPDS 03-065-2870) are shown in the lower parts as red and blue bars, respectively. The patterns have been arbitrarily shifted in the y-axis.

Full size image

Figure 2a shows a representative TEM image of the synthesized material, where it can be appreciated that the Fe3O4 NPs exhibit a rounded quasispherical shape with a rather broad size distribution. The fitting of the number density distribution of the size, q0, of the Fe3O4 NPs to a log-normal distribution revealed that the median size was (30 ± 9) nm (Fig. 2b, red curve). The synthesis of Fe3O4 NPs with a mean size centered at approximately 30 nm has already been reported when using an aging-based approach employing nitrate as an oxidizing37. Interestingly, magnetic particles in the size range of (20–50) nm are of great interest in several fields, such as hyperthermia applications38,39. It has been proposed that the transformation of Fe(OH)2 into magnetite occurs through a mechanism that involves dissolution of Fe(OH)2, oxidation of Fe(II) species in solution and subsequent precipitation of Fe3O440. Some authors state that primary particles of Fe3O4 nucleate heterogeneously on the surface of Fe(OH)241. A comprehensive study published by Sugimoto and Matijevic on the formation of Fe3O4 by means of the partial oxidation of Fe(OH)2 by KNO3 at elevated temperatures revealed that an excess of Fe(II) in the reaction medium led to the formation of relatively large spherical particles (0.4–1.1) µm, whereas an excess of OH- resulted in the formation of smaller cubic particles (30–100) nm42. The different particle sizes and morphologies were attributed to distinct mechanisms of growth, which are determined by the pH of the system in relation to the isoelectric point of magnetite. Under experimental conditions where there is an excess of Fe(II) with respect to the stoichiometric quantity necessary for the precipitation of Fe(OH)2, the pH of the system is maintained close to the isoelectric point of Fe3O4, allowing the growth of particles by aggregation and recrystallization of smaller primary particles. In contrast, an excess of OH- keeps the pH of the system above the isoelectric point, causing the charge of the Fe3O4 surface to prevent its aggregation and producing the growth of the particles due to the addition of species in the solution.

Figure 2
figure 2

(a,c) Representative TEM image of the synthesized (a) Fe3O4 NPs and (c) Au-Fe3O4 HNs. (b,d) Number density distribution of the size, q0, of the (b) Fe3O4 NPs and of the (d) Fe3O4 component in the Au-Fe3O4 HNs.

Full size image

More recently, it has been reported that even in the presence of excess OH-, the crystallization of Fe3O4 occurs through the formation and aggregation of primary particles43. In addition to crystallization via particle attachment, faceted magnetite crystals were found, indicating that coarsening also occurs to minimize the total energy of the developing crystals. The difference found between the average crystallite size D (21.9 ± 0.4) nm and the median Fe3O4 NP size obtained through statistical analysis based on the TEM images (30 ± 9) nm suggested that under the experimental conditions, the crystallization of Fe3O4 might occur through the formation and aggregation of primary particles. Based on these findings, we implemented a strategy for the synthesis of Au-Fe3O4 NHs in an aqueous medium that consists of adding Au NPs to the reaction system, considering that the heterogeneous nucleation of Fe3O4 primary particles might also take place on the Au surface. Figure 1, curve b, shows the diffraction pattern of the obtained material under the same experimental conditions as those already described but with Au NPs dispersed in the reaction medium. In addition to the peaks assigned to reflections of the Fe3O4 phase, peaks at 2θ = 38.2, 44.4, and 64.7 are observed, which are assigned to reflections of the (111), (200), and (220) planes of Au, respectively. For comparison, the reference data for Au (JCPDS 03-065-2870) are shown at the bottom as blue bars. Furthermore, there are also peaks at 2θ = 14.1 and 27.0, which are assigned to reflections of the (020) and (120) planes of lepidocrocite γ-FeOOH (JCPDS 01-074-1877), respectively. This analysis indicated that the Fe3O4 phase was formed along with a detectable amount of a completely ferric phase as a byproduct when Au NPs were present in the reaction medium. Indeed, it has been reported that during the aerial oxidation of Fe(II) sulfate solutions, the precipitation of Fe3O4 occurs via slow oxidation by air, whereas the precipitation of γ-FeOOH occurs via rapid oxidation by air44. The average crystallite size D of the Fe3O4 component was determined following the same procedure as above, resulting in a D value of (26 ± 2) nm. At first glance, this ~ 20% increase in the average D value could indicate that the incorporation of the Au NPs in the system could have an impact on the formation mechanism of the Fe3O4 component. Furthermore, the calculation of the unit cell parameter of the iron oxide component resulted in a value of a = (8.3891 ± 0.0003) Å, indicating the formation of Fe3O4 with good stoichiometry. Figure 2c shows a representative TEM image of the synthesized Au-Fe3O4 composite, where the presence of both Au and Fe3O4 NPs can be observed as a consequence of the substantial differences in the electronic contrast. In particular, Fe3O4 NPs mostly exhibit a rounded quasispherical shape, but crystals with clearer facets and corners that resemble a cubic or octahedral morphology can also be observed in a smaller proportion. Note that according to theoretical predictions, the main morphology of magnetite crystals is octahedral45. Counting more than one hundred NPs, it could be estimated that the amount of rounded (faceted) structures is 70% (30%). Moreover, Altan et al. reported similar results when using NaNO3 as the oxidizing46. The number density distribution of the size of the Fe3O4 NPs (Fig. 2d) shows a broader distribution with respect to that determined for the Fe3O4 NPs synthesized in the absence of Au NPs. Analysis of the size distribution of each morphology revealed that the size distribution of the rounded particles (centered at ~ 30 nm) resembled that of the Fe3O4 NPs synthesized in the absence of Au NPs, whereas the size distribution of the faceted NPs was shifted toward larger values (~ 45 nm). These results are consistent with the larger average D value calculated for the Fe3O4 NPs synthesized in the presence of Au NPs. On the other hand, the Au NPs exhibit a prolate shape with a relatively high degree of sphericity. The fitting of their number density distribution to a normal distribution resulted in a mean size of (57 ± 14) nm (see Supplementary Fig. S3). Figure 2c shows that a significant number of Fe3O4 NPs do not interact with the Au NPs, whose formation might be due to both homogeneous nucleation of primary particles and heterogeneous nucleation on the surface of the Fe(OH)2 platelets that are consumed as the reaction progresses. However, the Au NPs appear to be in close contact with the Fe3O4 component, suggesting that the latter might also nucleate heterogeneously on the Au surface and grow according to the mechanisms described above. Therefore, the structural characteristics of the obtained material might arise as a consequence of different nucleation mechanisms occurring simultaneously during the course of the reaction. Furthermore, the Au-Fe3O4 NHs do not seem to be completely coated by Fe3O4 and present a morphology more related to the Janus type than to the core–shell. Figure 3a shows a representative HAADF-STEM image of the Fe3O4-Au HNs, while its corresponding chemical composition analysis performed using EDS measurements is shown in Fig. 3b–d. The respective EDS element mapping images prove that the synthesized HNs are composed of Au, O, and Fe, respectively, being the elements composing magnetite homogeneously distributed around the Au NPs.

Figure 3
figure 3

(a) HAADF-STEM image and EDS element mapping images for (b) Au, (c) O, and (d) Fe of the synthesized Fe3O4-Au HNs. The scale bar applies for all images.

Full size image

The high-resolution TEM image of the Fe3O4-Au HN, shown in Fig. 4a, illustrates its high crystallinity. Figure 4b,c show the fast Fourier transform (FFT) of the rectangular selected areas delimited by white and blue lines in Fig. 4a, respectively, where a pair of points equidistant from the origin in 1/d can be observed in each case. The interplanar distance was determined by analyzing the respective inverse fast Fourier transform, resulting in a value of 0.20 nm (0.25 nm) for the area delimited by white (blue) lines. Based on the respective JCPDS mentioned above, these values are associated with the family of planes (200) of Au and the family of planes (311) of Fe3O4. Furthermore, Figs. 4d shows the FFT of the rectangular selected area delimited by red lines in Fig. 4a, that is, the interphase between Au and Fe3O4, where the spots corresponding to both phases can be observed.

Figure 4
figure 4

(a) High resolution TEM image of Fe3O4-Au HNs. (b–d) Fast Fourier transform (FFT) of the rectangular selected area delimited by (b) white, (c) blue, and (d) red lines in panel (a), respectively.

Full size image

Colored photos of colloidal dispersions of Au NPs, Fe3O4 NPs, and Fe3O4-Au NHs are provided in Supplementary Fig. S4, where it can be appreciated the visual change in the color of the colloidal dispersion after the formation of Au-Fe3O4 HNs. Figure 5a shows the normalized extinction spectra of the synthesized nanostructures. Note that the concentration of nanostructures is similar in each measured sample. The red curve shows the normalized extinction spectrum of the aqueous dispersion of Fe3O4 NPs, which exhibits a monotonic decrease in wavelength (λ) > 400 nm and resembles typical spectra previously reported for iron oxide nanoparticles suspended in water47,48. The extinction spectra of the Au-Fe3O4 composite is shown in the solid black curve. A broad peak centered at approximately 575 nm attributed to the localized surface plasmon resonance (LSPR) of the Au component is observed, which seems to overlap with the response of the Fe3O4 NPs. For comparison, the extinction spectrum of an aqueous dispersion of the Au NPs added to the reaction system for obtaining the hybrid composite is shown as a solid blue curve and is characterized by an LSPR centered at 535 nm. The redshift observed in the LSPR of the Au NPs after the formation of the Fe3O4 phase is attributed to an increase in the refractive index of the local environment that surrounds the Au NPs, providing evidence of the formation of the Au-Fe3O4 HNs49. These results are in agreement with the morphology of the nanostructures shown in Fig. 2c. To gain deeper insight into the optical properties of the synthesized materials, electrodynamics simulations were performed. Due to the structure–property relationships that are established at the nanoscale, modeling of the extinction spectra provides complementary evidence for the formation of Au-Fe3O4 HNs.

Figure 5
figure 5

(a) Normalized extinction spectra of Fe3O4 NPs (red curve), Au NPs (solid blue curve), and Fe3O4-Au HNs (solid black curve) along with the DDA-simulated normalized extinction efficiency spectra of a 57 nm diameter Au nanosphere (dashed blue curve) and a model Au-Fe3O4 HN schematically depicted in panel (b) (dashed black curve). (b) Scheme of the model Au-Fe3O4 HN, which consists of a 57 nm diameter Au nanosphere in contact with two diametrically opposed 40 nm diameter Fe3O4 spherical caps, with the center of the caps 42 nm displaced with respect to the center of the Au NP.

Full size image

The dashed blue curve in Fig. 5a shows the DDA-simulated extinction efficiency spectrum of a 57 nm diameter Au nanosphere dispersed in water. Note that Supplementary Fig. S5 compares the spectra calculated through the approximated DDA method and the exact Mie solution; the very good agreement between both spectra indicates that a tolerable error is reached when using an interdipole distance of 1 nm to model the synthesized nanostructures. The spectral position of the LSPR peak in the simulated spectrum matches the wavelength peak in the experimental spectrum of the Au NPs, whereas the broadening of the experimental peak with respect to the simulated peak is attributed to the size and shape distributions. On the other hand, the DDA-simulated extinction efficiency spectrum of the Au-Fe3O4 HN, schematically depicted in Fig. 5b, whose size and shape approximately resemble those found for the synthesized material, is shown as a dashed black curve. The model Au-Fe3O4 HN consists of a 57 nm diameter Au nanosphere in contact with two diametrically opposed 40 nm diameter Fe3O4 spherical caps, with the center of the caps 42 nm displaced with respect to the center of the Au NP. As expected, the LSPR peak for the Au-Fe3O4 HN was redshifted in comparison to that for the isolated Au nanosphere, and the magnitude of the redshift was similar to that measured for the synthesized Au-Fe3O4 composite. The experimental peak is notably broader than the simulated peak, which is attributed to the size and shape distributions of the Au-Fe3O4 HNs. Therefore, this modeling performed considering the morphology and size of the synthesized nanostructures allows us to assign changes in the optical response to the formation of Au-Fe3O4 HNs. For the sake of clarity, the modeling of the Fe3O4 NPs is shown in Supplementary Fig. S6, where the accuracy of the DDA results is corroborated by comparison with the Mie theory results. Note that the purpose of performing electrodynamic simulations is to gain insight into the interactions of the components of the of the HN, that is, gold and magnetite. An excellent agreement between the experimental and simulated spectra could be achieved taking into account all the contributions to the extinction of nanostructures of different sizes and shapes present in the sample. However, such a task is beyond the scope of the current work.

To address the effects of the parameter R on the phase, morphology and optical response of the synthesized nanostructures, experiments with R = 1.3 and 3 were performed while keeping the concentration of the Fe(II) precursor, the oxidizing H2O2, and the Au NPs constant. Figure 6a shows the XRD pattern of the material obtained with R = 1.3, where peaks corresponding to reflections of the Fe3O4 and Au phases can be noted. The peak observed at 2θ = 27.0 is assigned to reflections of the (120) plane of γ-FeOOH, whereas the peak at 2θ = 31.7 is attributed to the (112) plane of Na2SO4 (JCPDS 01-079-1553), which is formed as a byproduct. A representative TEM image of the synthesized Au-Fe3O4 composite shown in Fig. 6b reveals the presence of Au NPs attached to the Fe3O4 nanostructures. The iron oxide composite consists of both larger faceted particles than those found for R = 2 and smaller particles showing a granular morphology. According to previous studies, faceted particles might form due to the accretion of disordered primary particles50. The extinction spectrum of the Au-Fe3O4 NHs synthesized with R = 1.3 (Fig. 6c) shows a flat curve, indicating a significant contribution from the scattering component of the relatively large-sized Fe3O4 particles (red curve). However, a shoulder centered at 600 nm can also be attributed to the LSPR of the Au component, which is, in turn, redshifted with respect to the spectral position of the LSPR corresponding to the isolated unmodified Au NPs (black curve).

Figure 6
figure 6

(a) XRD pattern, (b) representative TEM image, and (c) normalized extinction spectrum of Au-Fe3O4 HNs (red curve) synthesized with R = 1.3. (a) The reference XRD patterns of Fe3O4 (JCPDS 01-089-0691) and Au (JCPDS 03-065-2870) are shown in the lower part as red and blue bars, respectively. (c) For comparison, the extinction spectrum of the Au NPs is shown as a black curve.

Full size image

The results obtained when using a value of R = 3 for the synthesis of Au-Fe3O4 NHs are shown in Fig. 7. The peaks in the XRD pattern (Fig. 7a) are assigned to reflections of the phases Au and Fe3O4, whereas the peak at 2θ = 27.0 is assigned to reflections of the (120) planes of γ-FeOOH. A representative TEM image of the synthesized composite (Fig. 7b) indicated the presence of Au NPs that were in close contact with the rounded Fe3O4 nanoparticles. The extinction spectrum of the HNs (Fig. 7c, red curve) presents a broad peak centered at 600 nm corresponding to the excitation of the LSPR of the Au component redshifted with respect to the LSPR in isolated Au NPs (see black curve, Fig. 7c). Therefore, the characterization of the obtained materials indicates that Au-Fe3O4 NHs are also formed when R = 1.3 and 3 are used in the synthesis. For R = 1.3, the size of the Fe3O4 component is relatively large, which impacts the profile of the extinction spectrum, as expected. In all the cases, a substantial amount of Fe3O4 possibly formed following a homogeneous nucleation mechanism, which might be associated with excessively high supersaturation values51,52.

Figure 7
figure 7

(a) XRD pattern, (b) representative TEM image, and (c) normalized extinction spectrum of Au-Fe3O4 HNs (red curve) synthesized with R = 3. (a) The reference XRD patterns of Fe3O4 (JCPDS 01-089-0691) and Au (JCPDS 03-065-2870) are shown in the lower part as red and blue bars, respectively. (c) For comparison, the extinction spectrum of the Au NPs is shown as a black curve.

Full size image

To reduce the amount of Fe3O4 possibly formed by a homogeneous nucleation process, additional synthesis experiments were performed in which the Fe(II) concentration was decreased from 1 × 10–4 M to 5 × 10–5 M using R = 2 and the same Au NP concentration (1 × 10–12 M). The results obtained are shown in Fig. 8. The XRD pattern of the synthesized material presents peaks that correspond to reflections of the Fe3O4 and Au phases, whereas the peak at 2θ = 31.7 is assigned to the reflection of the (112) planes of Na2SO4, which is formed as a byproduct (Fig. 8a). The representative TEM image shown in Fig. 8b suggests that a lower quantity of Fe3O4 is formed in comparison with the results obtained using a Fe(II) concentration of 1 × 10–4 M, as expected. Furthermore, the amount of Fe3O4 particles possibly formed by a homogeneous nucleation process seems to be noticeably reduced. In addition, the extinction spectrum presents a broad peak centered at 570 nm attributed to the excitation of the LSPR in the Au-Fe3O4 NHs (Fig. 8c, red curve). This peak appears to be redshifted with regard to the LSPR peak corresponding to individual Au NPs (see Fig. 8c, black curve) because of the formation of the Au-Fe3O4 NHs. The increase in the extinction coefficient for λ > 700 nm is mainly due to the coupling between the LSPRs that arises as a consequence of the Au NPs being in contact, as shown in Fig. 8b53,54.

Figure 8
figure 8

(a) XRD pattern, (b) representative TEM image, and (c) normalized extinction spectrum of Au-Fe3O4 HNs (red curve) synthesized with R = 2 and a Fe(II) concentration of 5 × 10–5 M. (a) The reference XRD patterns of Fe3O4 (JCPDS 01-089-0691) and Au (JCPDS 03–065-2870) are shown in the lower part as red and blue bars, respectively. (c) For comparison, the extinction spectrum of the Au NPs is shown as a black curve.

Full size image

In all the cases presented thus far, the rate of addition of the H2O2 solution was 100 mL/s. However, quite interestingly, for an Fe(II) concentration of 5 × 10–5 M and in the presence of Au NPs, the iron oxide/oxyhydroxide phase mostly formed can be controlled through the addition rate of the oxidant solution. Figure 9 shows the results obtained when the oxidant addition rate was decreased to 10 mL/s. The XRD pattern of the obtained material presented peaks at 2θ = 38.2, 44.4, and 64.7, which are assigned to reflections of the planes (111), (200), and (220) of Au, respectively, as well as peaks at 2θ = 14.1, 27.0, 36.3, and 46.9 which were assigned to reflections of the (020), (120), (031), and (051) planes of the lepidocrocite γ-FeOOH phase, respectively. Nonetheless, no diffraction peaks attributed to the Fe3O4 phase are observed. This change in the most common phase produced marked changes in the morphology of the synthesized nanostructures. The representative TEM image displayed in Fig. 9b shows Au NPs with ~ 10 nm particles attached on their surface resembling a core-satellite morphology. Importantly, the HNs appear well separated, indicating that the Au NPs do not interact with each other. Note that for an Fe(II) concentration of 5 × 10–5 M in the absence of Au NPs, Fe3O4 NPs are obtained as reaction products irrespective of the addition rate of the oxidant solution (see Supplementary Fig. S7). Owing to the structure–property relationships that occur at the nanoscale, the morphological features of the synthesized HNs are manifested in their optical response. Thus, the extinction spectrum exhibits a defined peak centered at 575 nm (Fig. 9c) due to the LSPR excitation in the Au-γ-FeOOH NHs, whereas the extinction values rapidly decrease for λ > 650 nm as a consequence of the relatively large separation between the Au NPs that comprise the HNs, preventing the occurrence of plasmonic coupling, which results in extinction increases in the NIR region.

Figure 9
figure 9

(a) XRD pattern, (b) representative TEM image, and (c) normalized extinction spectrum of Au-γ-FeOOH HNs (red curve) synthesized with R = 2 and a Fe(II) concentration of 5 × 10–5 M. (a) The reference XRD patterns of γ-FeOOH (JCPDS 01-074-1877) and Au (JCPDS 03-065-2870) are shown in the lower part as green and blue bars, respectively. (c) For comparison, the extinction spectrum of the Au NPs is shown as a black curve.

Full size image

Conclusions

In this work, a simple and environmentally friendly approach for the synthesis of Fe3O4 NPs and Au-Fe3O4 HNs in aqueous media is proposed. This method relies on the partial oxidation of the Fe(II) precursor using H2O2 as the oxidizing agent. In the absence of Au NPs dispersed in the reaction medium, Fe3O4 with a good stoichiometry and a median size of 30 nm was obtained as the only reaction product. On the other hand, in the presence of Au NPs dispersed in the reaction medium, Au-Fe3O4 HNs were obtained as the main reaction product. The morphology of the HNs consisted of approximately 57 nm long quasispherical Au NPs in contact with the Fe3O4 component, although the HNs were not completely coated by the latter, which resembled the Janus type. The structural features of the NHs were employed to perform electrodynamic modeling of their extinction properties. In particular, the redshift observed in the spectral position of the LSPR was successfully accounted for by means of DDA simulations. The joint analysis of both the experimental and theoretical results suggested that the Fe3O4 phase nucleated heterogeneously on the Au surface. The results obtained indicate that the value of the parameter R impacts the size of the Fe3O4 component of the HNs, which in turn modifies their extinction properties. Moreover, a decrease in the concentration of the Fe(II) precursor reduces the amount of Fe3O4 presumably formed by a homogeneous nucleation mechanism. Furthermore, in the presence of Au NPs, the addition rate of the oxidizing agent seems to play a significant role in the formation of the iron oxide/oxyhydroxide phase. The novel strategy implemented for the synthesis of Au-Fe3O4 HNs in aqueous media broadens the possibility of integrating plasmonic and magnetic properties into single nanostructures with potential applications in diverse fields.