Introduction

The synthesis of small (diameter < 30 nm) iron oxide (Fe3O4 or γ-Fe2O3) nanoparticles (NPs) with narrow size distributions (< 5%) and pronounced saturation magnetisation (Ms) is an important prerequisite for their potential technological applications [1, 2]. Magnetite (Fe3O4) crystallises in a cubic inverse spinel configuration where Fe3+ ions are evenly distributed among the tetrahedral (A) and octahedral (B) sites and Fe2+ ions occupy the remaining octahedral sites, yielding the empirical formula [Fe3+]A[Fe3+Fe2+]BO4 [3, 4]. Typically, Fe3O4 NPs exhibit lower saturation magnetisation values (50–60 emu/g) than bulk Fe3O4 (85–100 emu/g) at 300 K, induced by a large spin disorder on their surface [5]. In recent years, several strategies have been considered to improve their Ms by tuning their geometry, size, size distribution and composition. One potential strategy to enhance the Ms value of magnetite is to substitute Fe2+ with M2+ cations (M = Mn, Co, Ni) [3, 6, 7]. This has been achieved through several synthetic protocols such as sol–gel pyrolysis [8], reverse micelle emulsion [9], polyol [10], solvothermal [11], and co-precipitation [12] techniques. Nevertheless, technological success in the synthesis of single-domain crystalline nanoferrites of low dispersity has been limited.

The thermal decomposition method has been proposed for the synthesis of monodisperse nanoferrites with high crystallinity. This method involves the decomposition of metal precursors, such as metal carbonyls M(CO)5 [13], metal acetylacetonates M(acac)3 [14] and iron oleate [2], in high-boiling point organic solvents at very high temperatures (~ 320 °C) and in the presence of surfactants. It encourages the synthesis of monodisperse Fe3O4 and analogous MxFe3−xO4 nanocrystals of great size uniformity, crystallinity and well-shaped configurations. Consequently, various nanoferrites such as Fe3O4 [15, 16], Li0.3Zn0.3Co0.1Fe2.3O4 [17], ZnxFe3–xO4 [18], MgxMn1−xFe2O4 [19], Mn0.5Zn0.5Fe2O4[20], MgFe2O4 [21], Co0.03Mn0.28Fe2.7O4 [22], MnxZn1–xFe2O4 [23] and CoxFe3–xO4 nanoferrites [24] have been synthesised through this method for various applications. Despite the progress in the synthesis of monodisperse crystalline nanoferrites with high precision using this method, there are still some obstacles for its practical realisation. These include the complexity of the chemical synthesis, the need for an inert (N2) atmosphere during the reaction, the use of flammable organic solvents at higher temperatures and the difficult separation of solvents from the product. Another disadvantage is that the synthesised nanocrystals are dispersible in organic solvents, which requires additional steps to transfer them to the aqueous phase specially for biomedical applications. To address this, exchange agents such as PEG-phospholipid (DSPE-PEG2000) [25], methoxy-polyethyleneglycol-silane—500 Da (PEG) [5], 2,3-dimercaptosuccinic acid (DMSA) [26], Acrypol 934 [27], citric acid [28] and polyvinylpyrrolidone (PVP) [29] have been utilised to help with this.

Solvothermal methods (or hydrothermal methods when water is used as the solvent) are alternative synthetic methods which encourage the synthesis of nanocrystals with definite sizes, geometries and narrow size distributions. Generally, these are conducted using an autoclave under mild conditions [30, 31]. For example, MFe2O4 nanoferrites (M = Co, Mn, Ni, Zn) [32], MFe2O4 (M = Cu, Ni) [33], MFe2O4 (M = Mn, Fe, Co, Ni) [34], CoFe2O4 [35], Mg1–xZnxFe2O4 (x = 0.4–0.7) [36], MnFe2O4 and MFe2O4–Ag2O (M = Zn, Co and Ni) [37] have all been synthesised using hydrothermal decomposition of metal precursors; however, the prepared nanocrystals were aggregated even in the presence of surfactants. To synthesise nanoferrites with high monodispersity, organic phase (solvothermal) decomposition of metal precursors has been examined. For example, Tian et al. synthesised ultrasmall monodispersed magnetite NPs of 4–6 nm as potential MRI contrast agents by the decomposition of Fe(acac)3 in n-octanol. Additionally, Dendrinou-Samara et al. have utilised oleylamine (OAm) as both the solvent and surface-functionalising agent to synthesise NiFe2O4 NPs (9–11.7 nm) with Ms values of 32.0–53.5 emu/g, CoFe2O4 NPs (9–11 nm) with Ms values of 84.7–87.5 emu/g and 9-nm-sized MnFe2O4 with Ms values of 65.7 emu/g for biomedical applications. However, some aggregation was observed in the prepared nanocrystals [38,39,40].

In our previous work, we addressed this aggregation issue by the careful control of reaction parameters leading to highly uniform MxFe3−xO4 (M = Fe, Mg, Zn) nanoferrites using a mild solvothermal route. However, we found that our particles exhibited decreased magnetisation after substituting Fe3O4 with diamagnetic Mg2+ ions [41]. Additionally, they were only dispersible in an organic solvent (hexane). In the present work, we have attempted to increase the magnetisation by the substitution of paramagnetic Mn2+ ions into the Fe3O4 structure and subsequently explored the effect of this substitution on the crystallinity and magnetisation values of bare Fe3O4 NPs. In addition, considering that PVP is a water-soluble stabilising agent, we replaced oleic acid (OA) with PVP in an attempt to directly synthesise water-dispersible nanoferrites without additional phase exchange treatments.

Experimental Section

Materials

Iron (III) acetylacetonate (Fe(acac)3 ≥ 99.9% trace metals basis), oleylamine (≥ 70%), polyvinylpyrrolidone (Mw ≈ 25,000), tri-n-octylphosphine oxide (TOPO, 99%), octadecene (ODC, 90%), AR grade 1,5-pentanediol, tetramethylammonium hydroxide (20% w/w) and manganese(II) chloride (MnCl2) were purchased from Sigma–Aldrich. Zinc chloride (ZnCl2) was purchased from Ajax Finechem. All other chemicals were of analytical grade and used as received from commercial sources without further purification.

Synthesis of MxFe3−xO4 (M = Fe, Mn, Zn) spinel nanoferrites

Similar to our previous work [41], we used a solvothermal route with some modifications to develop monodisperse MxFe3−xO4 (M = Fe, Mn, Zn) spinel nanoferrites. For a typical synthesis of spinel MxFe3−xO4 (M = Fe, Mn, Zn) NPs, the defined amounts of metal (the Zn, Fe and Mn precursors), OAm and TOPO (Table 1) were mixed in octadecene (20 mL) under stirring (500 rpm) at 100 °C for 60 min. Polyvinylpyrrolidone (0.20 g) was dissolved in 1,5-pentanediol (5 mL) and heated to 150 °C for 60 min. The reason for using 1,5-pentanediol is to improve the solubility of PVP in octadecene.

Table 1 Synthesis conditions for MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites
Full size table

The two solutions were then mixed and deoxygenated with Ar (2 min) and then transferred into a 100-mL polytetrafluoroethylene (PTFE)-lined autoclave. The autoclave was sealed and maintained at 200 °C for 30 min and then heated to 240 °C for 2 h. After this time, the autoclave was left to cool to RT. The MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites were precipitated upon the addition of ethanol (10 mL), washed with an ethanol/hexane (1:2) solution mixture several times and then dispersed in hexane for further use.

Characterisation

The structure, chemical composition and magnetic features of the synthesised nanoferrites were probed with transmission electron microscopy (TEM; Tecnai G2 Spirit Bio-TWIN), powder X-ray diffraction (XRD; Rigaku Spider X-ray diffractometer), energy-dispersive X-ray-connected scanning electron microscopy (SEM–EDX; FE-SEM FEI Quanta), atomic absorption spectroscopy (AAS; AAS-9000 spectrometer, Shimadzu), thermogravimetric analysis (TGA; TA Instruments Q500), X-ray photoelectron spectroscopy (XPS; Kratos Axis UltraDLD) and vibrating sample magnetometry (VSM; Quantum Design P935A USA, physical property measurement system (PPMS)). Further details regarding characterisation are described in the supporting information.

Results and discussion

Structural and compositional studies

In our previous research, we investigated the effect of various experimental conditions and reaction parameters to successfully synthesise monodisperse MxFe3−xO4 (M = Fe, Mg, Zn) spinel nanoferrites [41]. The optimal conditions for the synthesis of a range of monodisperse NPs with high uniformity were found to be a 1:4 OA to OAm molar ratio, 0.5 mmol TOPO, 120-min reaction time and a temperature of 240 °C [41]. The resulting nanoparticles were not dispersible in water. Therefore, phase exchange treatment was performed using tetramethylammonium hydroxide to transfer hexane-dispersible nanoferrites to the aqueous phase.

In an attempt to directly synthesise water-dispersible nanoferrites without an additional ligand exchange process, we replaced OA with PVP. The rational is that polyvinylpyrrolidone (PVP) is a non-charged amphiphilic polymer which can be dissolved in either water or organic solvents [42]. Huang et al. synthesised water-dispersible stable colloidal Fe3O4 nanocrystals through thermal decomposition of Fe (Co)5 in the presence of PVP as the sole stabiliser for MRI applications [25]. It was expected that PVP molecules adsorbed on the particle surface would help to improve the dispersibility of the resultant nanocrystals in water. However, the obtained nanoferrites were not water dispersible due to the presence of OAm and TOPO in the reaction medium.

The TEM images of Fe1 and Fe2 (Fig. 1a, b) show the arrangement of NPs in a close-packed assembly. The Fe1 and Fe2 NPs exhibit predominantly spherical shapes with mean particle sizes of 7.9 ± 1.2 nm and 8.5 ± 2.2 nm, respectively (Fig. 1c, d). Substitution of Mn2+ ions into Fe3O4 results in deformed NPs with larger sizes than that of Fe1–2 as shown in the corresponding TEM micrographs of MnFe1 (Fig. 2a) and MnFe2 (Fig. 2b). The average particle sizes were determined to be 8.9 ± 1.9 nm (Fig. 2c) and 10.6 ± 2.9 nm (Fig. 2d) for MnFe1 and MnFe2, respectively. The TEM images of ZnFe1 (Fig. 3a) and ZnFe2 (Fig. 3b) show some well-separated spherical particles of good size uniformity. The ZnFe1 and ZnFe2 NPs have average particle sizes of 9.4 ± 2.5 nm (Fig. 3c) and 10.3 ± 2.1 nm (Fig. 3d), respectively. Increasing the concentration of OAm has resulted in both increased [43] and decreased size effects in the past [41]. In terms of size and morphology, an increase in the OAm amount in the present study does not appear to have had an influence (within experimental error).

Figure 1
figure 1

TEM images of the synthesised Fe1 (a) and Fe2 (b) NPs at different magnifications. Insets: Size distribution of Fe1 (c) and Fe2 (d) with mean size (Dh) and standard deviation (σ)

Full size image
Figure 2
figure 2

TEM images of the synthesised MnFe1 (a) and MnFe2 (b) NPs at different magnifications Insets: Size distribution of MnFe1 (c) and MnFe2 (d) with mean size (Dh) and standard deviation (σ)

Full size image
Figure 3
figure 3

TEM images of the synthesised ZnFe1 (a) and ZnFe2 (b) NPs at different magnifications. Insets: Size distribution of ZnFe1 (c) and ZnFe2 (d) with mean size (Dh) and standard deviation (σ)

Full size image

The crystalline phase and purity of MxFe3−xO4 (M = Fe, Mg, Zn) nanoferrites were studied by powder X-ray diffraction (PXRD) (Figs. 4 and S1). The XRD patterns of Fe3O4 NPs (Fe1 and Fe2) at 2θ = 18.50°, 30.1°, 35.6°, 43.1°, 53.2°, 57.2° and 63° can be indexed to the crystallographic planes of (111), (220), (311), (400), (422), (511) and (440) [44, 45]. Importantly, no other impurities corresponding to other iron oxide crystal phases such as γ-Fe2O3 (additional peaks at 2θ = 23.77° (210) and 26.10° (211) (JCPDS no. 39-1346)) [46, 47], α-Fe2O3 (2θ = 24.1° (012), 33.1° (104), 40.8° (113), 49.4° (024), 53.9° (116), 57.4° (018), 62.3° (214) and 63.9° (300) (JCPDS no. 24-0072)) [48] or FeO (2θ = 36.0° (113), 41.8° (200) and 60.7° (220) (JCPDS no. 06-0615))[49] were detected in the XRD patterns, confirming the purity of the synthesised Fe3O4 crystals.

Figure 4
figure 4

a Powder X-ray diffraction patterns and b the highlighted (311) diffraction peaks of MxFe3−xO4 (M = Fe, Mg, Zn) nanoferrites

Full size image

The average crystallite sizes of Fe1 and Fe2 calculated from the Debye–Scherrer equation were 7.2 ± 0.89 nm and 6.9 ± 0.38 nm, respectively. Considering the strong Bragg reflection peak (Miller index 3 1 1), the lattice spacing (d) and lattice constant (a) were determined for all MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites (Table 2). The lattice constants were both 8.3481 Å for Fe1 and Fe2 NPs, consistent with a = 8.35 Å reported for Fe3O4 NPs [50].

Table 2 Calculated values of size and lattice parameters for MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites
Full size table

The XRD patterns of Mn-doped Fe3O4 are well matched with the single-phase spinel cell geometry. The Bragg diffraction peaks at 2θ values of 18.5°, 30.1°, 35.5°, 42.80°, 52.8°, 57.1° and 62.5° corresponded to the crystal reflection planes of (111), (220), (311), (400), (422), (511) and (440) (JCPDS no. 74-2403) [51]. There is no evidence of impurities related to manganese oxide secondary phases such as α-MnO2 (2θ = 12.7° (110), 18.1° (200), 28.8° (310), 37.5° (211), 42.1° (301), 49.9° (411), 56.2° (600) and 60.3° (521) (JCPDS no. 44-0141)) [52] or MnO2 (2θ = 37.12° (100) and 66.75° (110) (JCPDS no. 30-0820)) [53]. This confirms that Mn2+ ions substituted Fe2+/Fe3+ in the Fe3O4 crystal unit rather than being distributed on the surface of Fe3O4 as a manganese oxide. The (311) crystal plane of Mn-doped Fe3O4 shifted from 2θ = 35.6° for Fe1 to 35.56° for MnFe1. Moreover, a increased from 8.3481 Å to 8.3800 Å and d increased from 2.5170 Å to 2.5266 Å [53]. The diffraction peaks of ZnFe2O4 representing the (111), (220), (311), (400), (422), (511) and (440) crystal planes agree well with the standard cubic spinel ZnFe2O4 [54].

Importantly, sharper peaks were observed in the XRD patterns of Zn-doped Fe3O4, which suggested greater crystallinity compared to undoped Fe3O4. No secondary phases of zinc oxides (2θ = 31.7° (100), 34.4° (002), 36.2° (101), 47.5° (105) and 56.5° (110)) (JCPDS no. 36-1451) were detected, indicating the purity of the crystals [55, 56]. Nevertheless, in line with Mn-doped Fe3O4, the (311) crystal plane of Zn-doped Fe3O4 slightly shifted from 2θ = 35.64° for Fe1 to 35.52° for ZnFe1 (Fig. 4b, labelled with a #). Furthermore, the corresponding lattice constant and lattice spacing increased from 8.3481 to 8.3754 Å and 2.5170 to 2.5253 Å. The increase in lattice parameters (expansion of the unit cell volume) of Fe3O4 after doping is due to the replacement of Fe3+ (rionic = 0.64 Å) [57] and/or Fe2+ ions (rionic = 0.76 Å) [58] with larger Mn2+ (rionic = 0.80 Å) [58] or Zn2+ ions (rionic = 0.74 Å) [59]. Lattice expansion of Fe3O4 by Zn2+ [58, 60] or Mn2+ [61] has been previously reported.

The physical sizes obtained from TEM are larger for the ZnFe nanoferrites than that of the crystallite sizes measured by XRD, with the remaining nanoferrites all within experimental error. This increase in size for the ZnFe nanoferrites could be ascribed to the fact that the reaction performance does not always lead to single crystal particles.

Thermogravimetric analysis (TGA) under an N2 flow with a heating rate of 10 °C/min from room temperature to 800 °C was used to quantitatively determine the organic fraction of nanoferrites (Fig. 5). The fraction of inorganic cores inside the nanoferrites is represented by the mass percentage of the residue; hence, the weight fraction of the surfactants in the nanoferrites can be established.

Figure 5
figure 5

TGA curves of MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites

Full size image

The first loss of mass below 200 °C in all samples refers to the removal of water or organic solvents from the nanoferrite surface [62]. The mass loss between 250 and 350 °C in the thermograms of all samples can be attributed to the decomposition of the free surfactants adsorbed on the surface of the NPs [63, 64]. The mass loss between 450 and 620 °C corresponds to the decomposition of directly attached surfactants [65]. The final loss of weight between 600 and 750 °C for all nanoferrites can be assigned to the reduction of the inorganic metal core under a N2 atmosphere. For Fe1, the loss of weight was a two-stage process between 600 and 700 °C and 700 and 750 °C. Ayyappan et al. and Mameli et al. have reported the same behaviour for CoFe2O4 [66] and ZnxCo1−xFe2O4 (0 < x < 0.6) nanoferrites [18].

Notably, there are discontinuities in the thermograms of MnFe2 (just above 400 °C and at approximately 500 °C), Fe1 (just past 600 °C) and MnFe1 (at approximately 300 and 500 °C). This may be the result of the instrument being slightly bumped during the measurements.

The fraction of stabilising surfactants around the nanoferrites tends to reduce with increased magnetic core size (Table 2). Larger nanocores possess lower surface-to-volume ratios, which results in less active sites available for binding surfactants [67, 68].

EDS spectra of nanoferrites were obtained through line scans of selected sites at various magnifications to analyse the chemical compositions in a semi-quantitative manner. The EDS results for all nanoferrites collected from different positions exhibited similar atomic percentages (At%), indicating the composition homogeneity of the nanoferrites [69]. For instance, the At% of Mn atoms were 4.81% and 4.03% for MnFe1 and MnFe2 NPs, respectively. A similar trend could also be seen in EDS results of Fe1–2 and Zn1–2 as depicted in Figures S2–S4.

Peaks corresponding to C atoms, due to surfactants and the carbon coating prior to analysis, were observed in the EDS spectra of all nanoferrites. The characteristic peaks for Fe and O were observed in the EDS spectra of Fe1–2 (Figure S2). The EDS spectra of MnFe1–2 contained Fe, Mn and O peaks (Figure S3), and the EDS spectra of ZnFe1–2 featured Fe, Zn and O peaks (Figure S4). Notably, in the case of Mn-doped Fe3O4 NPs, the signals of Mn and Fe overlap at 6.5 keV. This has also been observed for other Fe–Mn nanoferrites [53, 70] and is due to the close energy levels of Mn Kβ and Fe Kα, which makes it difficult to distinguish them on EDS spectra. Based on EDS analysis, the Fe/O, Mn/Fe/O and Zn/Fe/O atomic ratios differed from the expected stoichiometries.

Atomic absorption spectroscopy (AAS) was used to determine the experimental chemical formula of MxFe3–xO4 nanoferrites (Table 3). The calibration curve equations and R2 for each analyte along with the experimental absorption values are presented in Figures S5–S7 for the nanoferrites. Consistent with the EDS results, the molar ratios and chemical formula derived from AAS results were not in agreement with the corresponding theoretical values. Our group, Hu et al. and Oberdick et al. have also reported this deviation from theoretical stoichiometry in MxFe3–xO4 CoxFe3–xO4 and core/shell Fe3O4/MnxFe3−xO4 NPs [71, 72]. This might be due to the different decomposition temperatures of Fe (acac)3, MnCl2 and ZnCl2 salts [73]. In addition, the large magnitude of surfactants used in the synthetic process may alter the decomposition temperatures of Fe(acac)3, MnCl2 and ZnCl2 salts and influence the growth mechanism [71, 74].

Table 3 Theoretical and experimental (AAS) molar ratios and chemical formula of MxFe3−xO4 (M = Mn, Zn) nanoferrites
Full size table

The chemical compositions and oxidation states of MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites were investigated via X-ray photoelectron spectroscopy (XPS). The C 1s signal at 284.6 eV (adventitious carbon) was utilised as an energy reference in order to calibrate the binding energy (B.E) scale ranging from 0 to 1300 eV. Wide-scan surveys confirmed the peaks of carbon (C 1s), oxygen (O 1s), iron (Fe 2p), zinc (Zn 2p) and manganese (Mn 2p) photoelectron lines recorded at B.E values of 280–300 eV, 520–550 eV, 702–750 eV, 630–665 eV and 1015–1055 eV, respectively (Figs. 6 and S8–S12). A sharp peak at 285 eV was observed in the high-resolution C 1s XPS spectra of all nanoferrites, which corresponded to the C–C bond of surfactants and adventitious carbon [75, 76]. For all nanoferrites, the O 1s core-level spectrum showed a low-intensity signal centred at lower B.E value of ~ 530 eV, attributed to the metal–oxygen bonds of MxFe3−xO4 [77, 78], and a larger peak at ~ 532 eV assigned to carboxylate groups [76].

Figure 6
figure 6

XPS spectra of the Fe1 nanoferrite: a survey scan, b C1s, c O1s and d Fe2p regional scans

Full size image

For the Fe1 NPs, the Fe 2p core-level spectrum indicates the absence of Fe ions on the surface (Fig. 6). For the Fe2 NPs, the doublet peaks centred at B.E values of 710.6 and 723.8 eV are ascribed to Fe 2p3/2 and Fe 2p1/2 of Fe3+ in Fe3O4 (Figure S8) [79]. The spin–orbit splitting energy difference was 13.2 eV which is in accordance with earlier reports for Fe3O4.[80] Furthermore, the presence of both the + 2 and + 3 oxidation states of Fe and the formation of Fe3O4 rather than γ-Fe2O3 are thought to cause the broadness of the Fe 2p peaks [81].

In the case of Mn-doped Fe3O4 (MnFe1 and MnFe2), the presence of Mn 2p3/2 and Mn 2p1/2 spin–orbit peaks at B.E values of ~ 641.5 and 653.5 eV (MnFe1) and 641.7 and 653.8 eV (MnFe2) indicates the oxidation state of Mn2+ (Figures S9–S10) [3, 82, 83].

For the Zn-doped Fe3O4 nanoferrites, the presence of Fe in the + 3 oxidation state is indicated by two major peaks, located at 711.8 and 724.8 eV for ZnFe1 and at 711 and 724.2 eV for ZnFe2, consistent with Fe 2p3/2 and Fe 2p1/2, respectively. Furthermore, two peaks in the Zn 2p core-level XPS spectrum indicate Zn in the + 2 oxidation state, with B.E values of 1021.6 and 1044.8 eV (ZnFe1) and 1022.2 and 1045.4 eV (ZnFe2) and attributed to Zn 2p3/2 and Zn 2p1/2 (Figures S11–S12). The obtained values are in accordance with values reported in the literature for Zn in the + 2 oxidation state, which confirms the formation of zinc ferrite ZnFe2O4 [84, 85].

Magnetometry

Magnetisation (emu/g) as a function of the applied magnetic field (Oe) (M–H loops) was recorded using VSM at room temperature (~ 300 K) for all nanoferrites to obtain Ms, remnant magnetisation (Mr), coercivity (Hc) and normalised remanence (Mr/Ms) values (Fig. 7 and Table 3). Individual VSM plots and sample masses of nanoferrites are provided in Figure S13. The Hc and Mr values at low fields were almost negligible for all nanoferrites, indicating their superparamagnetic characteristics. This reveals that the superparamagnetic behaviour of Fe3O4 (Fe1 and Fe2) does not change after doping Zn2+ and Mn2+ ions into its spinel structure [86].

Figure 7
figure 7

(a) Magnetic hysteresis loops of MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites at room temperature and (b) the magnetic response of MnFe2 to an external magnetic field

Full size image

The relationship between particle size and the magnetic properties of Fe3O4 NPs has been examined. If the particle sizes measured with XRD and TEM are between 10 and 80 nm, particles are in single-domain state; however, particles with diameters > 80 nm present multi-domain structure. Additionally, if the measured size is < 30 nm, particles present superparamagnetic behaviour [87]. The sizes of our particles measured with TEM and XRD are < 30 nm, which indicates they are single-domain superparamagnetic feature as evidenced by VSM measurements.

Magnetisation values decreased with increased OAm content, so that Ms (Fe1) > Ms (Fe2) and Ms (MnFe1) > Ms (MnFe2). Since Ms is described on a per gram basis (emu/g), a non-magnetic coating OAm layer will diminish its value which equates a decrease in the magnetisability [88]. However, in the case of ZnFe nanoferrites, the Ms value of ZnFe2 > ZnFe1. The magnetisation difference between these particles is very small and so is most likely related to the size effect [89].

The concentration and occupation sites of doped ions are also important factors. In the crystal structure of Fe3O4, all Fe2+ ions (magnetic moment = 4 μB) occupy B sites, while the Fe3+ ions (magnetic moment = 5 μB) are distributed equally between A and B sites. Therefore, the net magnetisation of Fe3O4 is decided by Fe2+ ions at B sites, as Fe3+ ions at A and B sites align opposite to each other and their magnetic moments cancel. When doped with Mn2+ or Zn2+ ions, the A and B sites of Fe3O4 can be occupied by these ions and the ferromagnetic and/or antiferromagnetic coupling interactions between Fe2+ and Fe3+ ions are modulated by their concentrations and distribution of the doped ions at the A and B sites.

For Mn-doped Fe3O4, the substitution of Fe3+ ions (5 μB) by Mn2+ ions (5 μB) does not change the net magnetisation of the Fe3O4 unit. However, if the Mn2+ ions (5 μB) substitute the octahedral Fe2+ ions (4 μB), the saturation magnetisation is expected to increase. Doping Mn2+ ions (0.2 mmol) into Fe3O4 increased the Ms from 30.2 to 33.9 emu/g for MnFe1 and from 24.8 to 31.2 emu/g for MnFe2, suggesting that Mn2+ substituted Fe2+ ions. Additionally, MnFe1 and MnFe2 nanoferrites exhibit larger sizes compared to Fe1 and Fe2, resulting in a smaller spin canting effect and higher magnetisation [67].

Doping Zn2+ ions (0.4 mmol) results in a reduced Ms of 14.2 emu/g for ZnFe1 and 12.7 emu/g for ZnFe2. For Zn-doped Fe3O4, the decrease in magnetisation is ascribed to the substitution of diamagnetic Zn2+ ions (0 μB) in the tetrahedral site with the Fe3+ ions occupying the octahedral sites, decreasing the strength of antiferromagnetic coupling interactions (the A–B superexchange interaction) of Fe3+ atoms and subsequent reduced magnetisability [36, 90].

Magnetisation of nanoferrites as a function of the temperature (T) (M–T curves) was investigated in the field-cooled (FC) and zero-field-cooled (ZFC) regimes under a constant magnetic field of 10 Oe to extract magnetic crystalline anisotropy energy (Keff) and blocking temperature (TB) values (Fig. 8 and Table 4). The magnetic anisotropy constant (K) was calculated for all nanoferrites using the equation, K = 25kBTB/V, where TB is the blocking temperature, kB is the Boltzmann’s constant (1.3807 × 10 −23 J K−1) and V is the volume of a single nanocrystal (m3) [2].

Figure 8
figure 8

FC/ZFC curves of the MxFe3−xO4 (M = Fe, Mn, Zn) nanoferrites recorded at a constant magnetic field of 10 Oe

Full size image
Table 4 Parameters of ZFC/FC and M–H hysteresis loops for MxFe3−xO4 (M = Fe, Mg, Zn) nanoferrites
Full size table

A marked increase in TB was observed for Zn- and Mn-doped Fe3O4 NPs. This shift towards higher temperatures could be attributed to increased magneto-crystalline anisotropy of the system (Table 4). The strong coupling strength between orbital angular momentum and electron spin (L–S) can also contribute to higher blocking temperatures [91, 92]. The Mn-doped nanoferrites presented with broader ZFC curves. This is ascribed to the larger size distribution of NPs and increased interparticle interactions as shown by TEM results [93].

Conclusion

The synthesis of a series of low-dispersity MxFe3−xO4 (M = Fe, Mn, Zn) spinel nanoferrites was achieved through the solvothermal route. The substitution of Fe2+ ions with paramagnetic Mn2+ ions resulted in improved magnetisation compared to the previous study which incorporated Mg2+ ions [41]. Zn2+ doping was effective for the formation of well-shaped nanoferrites of improved crystallinity; however, this substitution decreased the magnetisation. XPS provided evidence for the formation of MxFe3−xO4 (M = Fe, Mn, Zn) spinel nanoferrites, but revealed a high organic content on the surface of the NPs. Considering the AAS results, all doped nanoferrites exhibited compositional deficiency. Our rationale to directly synthesise water-dispersible nanoferrites without subsequent treatments proved to be unsuccessful utilising PVP; however, in the future we aim to study the ratio between PVP, OAm and TOPO through which the possibility of synthesis of direct in situ water-dispersible low-dispersity particles might be achieved.