Depletable Peroxidase-like Activity of Fe3O4 Nanozymes Accompanied with Phase Transformation Triggered by Separate Migration of Electron and Iron Ion

As the pioneering Fe 3 O 4 nanozymes, their explicit peroxidase (POD)-like catalytic mechanism remains elusive. Although many studies have proposed the surface Fe 2+ -induced Fenton-like reactions accounting for their POD-like activity, few focus on the internal atomic changes and their contribution to the catalytic reaction. Here we report that Fe 2+ within Fe 3 O 4 transfers electrons to the surface via the Fe 2+ -O-Fe 3+ chain, regenerating the surface Fe 2+ and enabling a sustained POD-like catalytic reaction. This process occurs with the outward migration of excess oxidized Fe 3+ from the lattice, which is a rate-limiting step. After prolonged catalysis, Fe 3 O 4 nanozymes suffer the phase transformation to γ -Fe 2 O 3 with a depletable POD-like activity. This self-depleting characteristic of nanozymes with internal atoms involved in electrons transfer and ion migration is well-validated on lithium iron phosphate nanoparticles. We reveal a key yet ever ignored issue concerning the necessity of considering both surface and internal atoms when designing, modulating, and applying nanozymes.


Introduction
2][13][14][15] Given the intricate structure-activity relationships and restricted characterization techniques, however, fewer breakthroughs have been made in understanding the explicit mechanism of most nanozymes. 3,45][16][17][18][19] To date, it is generally accepted that high-reactive hydroxyl radicals (•OH) generated by Fenton-like reactions (Equation 1-2) involving the surface Fe 2+ under acid conditions contributes to the POD-like activity of Fe3O4 NPs. 18,19Similar to natural horseradish peroxidase (HRP), Fe3O4 nanozymes follow the ping-pong mechanism and Michaelis-Menten kinetics. 101][22][23] Other individual studies have investigated the absorption, activation, and desorption processes of substrates (e.g.H2O2 and TMB) on the surface of Fe3O4 at the atomic level based on density functional theory and developed some descriptors to predict their POD-like activity. 14,15 2+ +  2  2 →  3+ +•  +  −  1 = 76 (/) −1  −1 (1) The above mechanistic studies share a theoretical premise: only the surface-active sites play a decisive role in the enzymatic-like property of nanozymes since catalysis occurs mainly on the particle surface or interface.This view is now widely recognized and works for most types of nanozymes. 1,2,4,11,23,24For example, in a recent controversial question regarding how to define "nanozyme concentration", Liu et al. argued that considering the whole particle or all atomic units within a particle as an enzyme unit would overestimate and underestimate the catalytic activity of nanozymes, respectively, because it is the surface atoms that are really the catalytic active sites. 24However, in the Fenton-like reactions triggered by Fe3O4 nanozymes, we noticed that the reaction rate constant of Equation ( 1) is much higher than that of Equation ( 2), which implies that the surface-active Fe 2+ is hardly recovered after being oxidized.This irreversible oxidation of surface Fe 2+ prompts us to ponder if only the surface atoms of the nanozymes, particularly for metal oxide nanozymes, act in enzymatic-like catalysis, would these active sites be exhausted after long-term catalysis, rendering the nanozymes inactive?So far, nevertheless, no relevant studies can conclusively answer this crucial question.
Here we propose a key yet ever ignored issue regarding the POD-like mechanism of nanozymes by characterizing the chemical composition and catalytic activity of the recycled Fe3O4 NPs participating in cyclic POD-like catalysis.Both surface and interior Fe 2+ were found to impart POD-like property to Fe3O4 nanozymes.Namely, Fe 2+ inside the particle transfers its electron to the surface layer, regenerating the surface Fe 2+ and sustaining the catalytic reaction.This process is coupled with the outward migration of excess oxidized Fe 3+ , which is a rate-limiting step.As the catalysis continues, Fe3O4 is slowly oxidized into γ-Fe2O3 accompanying the depleted enzyme-like activity, similar to the conventional low-temperature oxidation of magnetite, only with different electron receptors.This self-depleting characteristic of nanozymes with internal atoms involved in electrons transfer and ion migration is further demonstrated by a typical model material, LiFePO4, which contains the redox-active metal sites and mobile lithium ions (Li + ) encapsulated in a rigid phosphate network.This paper reveals that internal atoms may also contribute to the nanozyme-catalyzed reactions even though these reactions occur on the surface of NPs, which is thought-provoking when designing, regulating, and applying nanozymes.

Results
Synthesis and Characterization of IONPs.Near-spherical magnetite nanoparticles (Fe3O4 NPs) with an average diameter of 10.16 ± 0.12 nm (Supplementary Fig. 1a) were synthesized using the chemical co-precipitation method. 18Maghemite (γ-Fe2O3) and hematite (α-Fe2O3) NPs were derived by calcining the Fe3O4 NPs powder at 200 °C and 650 °C for 2 hours, respectively (Fig. 1a).XRD and Raman spectra (Supplementary Fig. 1b-c) show the successful synthesis of these three iron oxide NPs (IONPs).These IONPs were uniformly dispersed in an aqueous solution at pH of 3 by ultrasonication (Supplementary Fig. 1d).To avoid affecting the enzymatic-like activity, all particles were free of the surface coating.Their POD-like activities were assessed using different colorimetric substrates, including TMB, ABTS, and OPD, in the presence of H2O2.The results show that their catalytic activity followed the order of Fe3O4 NPs >> γ-Fe2O3 NPs > α-Fe2O3 NPs (Supplementary Fig. 2).To better quantify their POD-like activity, we calculated their specific activity (anano) according to the specified method, 26,27 which were 1.79, 0.44, and 0.03 U•mg -1 , respectively (Fig. 1b).As previously reported, 10,18 the higher catalytic ability of Fe3O4 NPs originates from the Fenton-like reaction triggered by the surface Fe 2+ (Supplementary Fig. 3).The negligible anano of α-Fe2O3 NPs compared with γ-Fe2O3 NPs is ascribed to the change of the inverse spinel structure due to the higher calcination temperature.Cyclic POD-like catalysis of Fe3O4 NPs.To investigate whether the surface Fe 2+ of Fe3O4 NPs is depleted after participating in prolonged catalysis, we continuously increased the amount of substrate TMB under sufficient H2O2 with as-synthesized three IONPs as "continuous catalysts", and monitored the absorbance changes of TMB oxidation products at 650 nm within 12 h.From Supplementary Fig. 4, even though the TMB was increased from 0.087 mM to 0.52 mM, the Fe3O4 NPs were still able to continuously and rapidly engage in the catalytic reaction for a long duration (≥ 12 h) without showing signs of depletion.We speculated two reasons: 1) the amount of substrate is still too low to completely consume the surface-active Fe 2+ ; 2) The Fe 2+ within Fe3O4 NPs provides the impetus for the continuous catalysis.
Cyclic POD-like catalytic assays (Fig. 1c) were carried out as validation, which could provide sufficient substrates for Fe3O4 NPs to keep exerting their POD-like capacity.We evaluated the anano of the recycled Fe3O4 NPs within five days.The results show that the catalytic ability of Fe3O4 NPs decreased to a level comparable to that of γ-Fe2O3 NPs after five days of cyclic catalysis, while the changes of γ-Fe2O3 NPs were negligible (Fig. 1d and Supplementary Fig. 5).It pushed us to wonder how the surface-active Fe 2+ of Fe3O4 NPs alone could sustain the TMB oxidation up to 100 hours?Conceivably, if only the surface-active sites are responsible for the enzyme-like performance, nanozymes will deactivate when the surface-active sites are exhausted.
To reveal the potential reasons for the sustained catalytic capacity of Fe3O4 NPs, we characterized the physicochemical properties of the recycled Fe3O4 NPs using different methodologies.The chemical states of Fe atoms in Fe3O4 NPs recycled from catalysis at days 0, 1, 3, and 5 were first analyzed by XPS technology.The X-ray penetration depth of the analyzed sample ranges from 2 to10 nm.Since the average diameter of as-synthesized Fe3O4 NPs is around 10 nm, the Fe valence state obtained from the Fe2p fitting analysis can be approximated as the oxidation state of individual Fe3O4 NPs.As shown in Fig. 1e, the Fe 2+ in Fe3O4 NPs decreased from the original 30.9% to 0% with the increase of cyclic catalytic days, indicating that the interior Fe 2+ was also oxidized to Fe 3+ in the successive POD-like reactions.
Furthermore, in the Raman spectra of the recycled Fe3O4 NPs, the feature of the A1g mode band shifted from 660 cm -1 to 700 cm -1 , corresponding to a transition from magnetite to maghemite (Fig. 1f). 28,29Besides, this phase transformation was also confirmed by the NEXAFS spectroscopy.Figure 1g shows the Fe L-edge NEXAFS spectra of the control Fe3O4 NPs and the recycled Fe3O4 NPs after 5 days of catalysis, in comparison with two reference spectra of FeSO4 and Fe2O3.1][32] Additionally, TEM images (Fig. 1h) and XRD pattern (Supplementary Fig. 6) show that the influence of this transformation on the particle morphology, size, and lattice structure is negligible.Based on these characterization results, we conclude that both surface and internal Fe 2+ can be oxidized into Fe 3+ accompanied by a gradual phase transformation to γ-Fe2O3 during Fe3O4 nanozymes exerting their POD-like activity.
Aeration oxidation kinetics of Fe3O4 NPs.We assume that the oxidation of Fe3O4 nanozymes induced by POD-like catalysis is comparable to the traditional low-temperature (< 200 °C) air oxidation of magnetite since the crystal structure of both remains unchanged during the oxidation process. 33Both magnetite and maghemite contain 32 O atoms per unit cell.The difference is that the former contains 24 Fe atoms (16 Fe 3+ and 8 Fe 2+ ), while the latter has only 21. 33 Fe atoms (all Fe 3+ ).Namely, once 8 Fe 2+ in magnetite are oxidized to 8 Fe 3+ releasing 8 electrons, a charge imbalance will occur (Equation 3).To maintain electroneutrality, 2.67 Fe 3+ have to migrate to the crystal surface, leaving the cation vacancy (Equation 4). 34The outward moving Fe 3+ will coordinate with the surface absorbed O 2-which is ionized by the electrons generated by the oxidation of Fe 2+ .Therefore, the phase transformation of Fe3O4 to γ-Fe2O3 is a single-phase topological reaction accompanied with the separate migration of electrons and excess Fe 3+ . 34ttice defects have been reported to facilitate the outward migration of excess iron ions, thereby accelerating the oxidation process of magnetite. 35As verification, we compared the aeration oxidation kinetics of Fe3O4 NPs synthesized by two methods with different levels of lattice defects.One was prepared by the chemical co-precipitation method as described above (Fig. 1a), which is considered to possess more lattice defects (named cc-Fe3O4 NPs).The other was prepared by the thermal decomposition method (Supplementary Fig. 7) with a relatively complete lattice structure (named TD-Fe3O4 NPs). 36Both Fe3O4 NPs have a similar average particle size (~10 nm) without surface coating.Their aqueous solutions were stirred under the same aeration rate (with air) for 12 h at 120 °C.For a better comparison, the oxidation system of cc-Fe3O4 NPs (total 170 mL, 3.6 mg Fe/mL) was much larger than that of TD-Fe3O4 NPs (total 30 mL, 0.45 mg Fe/mL).This implies that individual TD-Fe3O4 could gain more oxygen than cc-Fe3O4 to keep it oxidized.From Fig. 2a-b, both Fe3O4 NPs exhibited electronic transitions in the visible and NIR region due to intervalence charge transfer between Fe 2+ and Fe 3+ , 37 which decreased gradually with oxidation time.At the end of aeration oxidation, little absorption beyond 700 nm was observed, indicating a phase transformation from Fe3O4 NPs to γ-Fe2O3 NPs. 37Besides, the color of both suspensions gradually changed from dark-brown to reddish-brown.Notably, despite the less oxygen exposure for individual cc-Fe3O4 NP, its NIR absorption decreased faster than that of TD-Fe3O4 NP, especially during the initial oxidation phase (within three hours).These results confirm that more lattice defects favor the oxidation reaction of Fe3O4 NPs due to the faster electron and ion transfer.
Analogous to aerated oxidation, the rapid electron and ion migration also facilitates the POD-like catalysis of Fe3O4 NPs, with the only difference that the electron receptor changed from O2 in aerated oxidation reaction to H2O2 in POD-like reaction.To prove this, the POD-like activity of cc-Fe3O4 NPs and TD-Fe3O4 NPs as well as their variation with aerated oxidation time were investigated.As seen in Supplementary Fig. 8, the POD-like activity of cc-Fe3O4 NPs was higher (2.8 folds) than that of TD-Fe3O4 NPs, despite TD-Fe3O4 NPs having a smaller hydrodynamic diameter and negative surface potential contributing to a strong affinity with TMB.Aeration oxidation kinetic studies show that the POD-like activity of both Fe3O4 NPs decreased with oxidation time (Fig. 2c), along with slight fluctuations in hydrodynamic size and surface potential (Supplementary Fig. 9).However, the decline rate of cc-Fe3O4 NPs was faster than TD-Fe3O4 NPs, particularly in the initial oxidation stage.This phenomenon is consistent with the changes of NIR spectra shown in Fig. 2a-b.These results further confirm that the more lattice defects of Fe3O4 NPs, the easier the migration of excess Fe ions, and thus the higher the POD-like activity.It also means that Fe3O4 NPs with more defect sites are easier to be depleted when involved in a POD-like reaction due to their excellent catalytic capability.LiFePO4 NPs as an ideal verification model.9][40][41] LiFePO4 undergoes redox reactions along with the lithium insertion/extraction during the charge-discharge process (Equation 5-6) without changing its ordered-olivine structure (Fig. 4a). 38We speculate the charging process of LiFePO4 is resembling the oxidation process of Fe3O4, both of which involve the oxidation of Fe 2+ and the migration of internal ions, which motivated us to focus on whether LiFePO4 NPs also have the POD-like catalytic ability.
Rod-like LiFePO4 NPs with an average length of 321.9 nm and width of 172.2 nm (Fig. 4b) were successfully synthesized using the solvothermal method 38 and characterized via various methodologies (Supplementary Fig. 10 and Table S1-2).As expected, the POD-like activity of LiFePO4 NPs was demonstrated with different chromogenic substrates including TMB, ABTS, and OPD (Fig. 4c and Supplementary Fig. 11).Also, they follow pH, temperature as well as NPs concentration dependence and the Michaelis-Menten kinetics (Supplementary Fig. 12-13).The optimal pH is about 4.0.The ESR spectra show that •OH was produced from the decomposition of H2O2 catalyzed by LiFePO4 NPs in a time-dependent manner (Fig. 4d), which is similar to Fe3O4 NPs.We then compared the POD-like activity of LiFePO4 NPs and cc-Fe3O4 NPs using two oppositely charged substrates (TMB and ABTS) at pH 3.6.The results consistently show that LiFePO4 NPs had higher catalytic ability than cc-Fe3O4 NPs (Supplementary Fig. 14), and the anano of LiFePO4 NPs was approximately four times that of cc-Fe3O4 NPs, despite their larger particle size (Fig. 4e).
These results imply that LiFePO4 NPs may share a similar POD-like catalytic mechanism with Fe3O4 NPs, differing in that the rapid Li + migration in the lattice of LiFePO4 NPs confers them a superior POD-like catalytic activity (Fig. 4f).peaks of the recycled LiFePO4 NPs were shifted toward the higher binding energy (Fig. 5a), indicating the oxidation of Fe 2+ within the NPs.In the XRD pattern (Fig. 5b), the residual LiFePO4 phase (marked with "о" in the yellow pattern) in the recycled NPs was negligible, proving that almost all LiFePO4 were delithiated and oxidized into FePO4 (marked with "+") after cyclic POD-like catalysis.This result was further confirmed by ICP analysis that the Li element content in recycled NPs was almost 0 (Table 1).Moreover, the electrochemical property of the recycled NPs was examined using cyclic voltammetry (CV) 43 at various scan rates in the voltage range of 0.2 to 0.5 V (Supplementary Fig. 15).The redox peak currents of the recycled NPs were dramatically reduced due to the absence of Li + in their lattice (Fig. 5c).
This phase transformation, as expected, severely impaired the POD-like activity of the recycled LiFePO4 NPs (Fig. 5d), in agreement with the self-depleting characteristic of the Fe3O4 NPs described above.Mobile Li-ions as the limiting factor for LiFePO4 NPs-catalyzed POD-like reaction.In the field of sodium (Na)-ion batteries, the charge transfer resistances and lattice volume change upon Na + migration are larger for NaFePO4 electrodes, compared with their Li equivalents due to the larger ionic radius of Na (1.02 Å) than Li (0.76 Å). 44 Inspired by this, We partially replaced Li with Na in the lattice of LiFePO4 NPs to explore the potential effect of Na-doping on their POD-like activity.Concretely, three NaLiFePO4 NPs with similar physicochemical properties but different Na-doping amounts were successfully synthesized (Supplementary Fig. 16 and Table S3).We then compared their POD-like activities under the same reaction conditions and found that the more Na doping, the lower the POD-like activity (Fig. 5e), indicating that the large Na + radius hinders the free migration of Na + and Li + in the crystal, thereby impairing the electron transfer rate.We attempted to use K-doped LiFePO4 NPs as further proof, however, the large ionic radius of K (1.38 Å) makes it difficult to embed into the electrode materials (Supplementary Fig. 17 and Table S3), which is a common issue in K-ion batteries. 45 further prove the decisive role of mobile Li + , we measured the POD-like activity of commercially available LiFePO4, Fe3(PO4)2, and FePO4 materials with similar hydrodynamic dimensions and surface negative potentials (Supplementary Fig. 18).The results show that their POD-like activity follows LiFePO4 >> Fe3(PO4)2 > FePO4 (Fig. 5f), directly confirming that the presence of Fe 2+ alone in Fe3(PO4)2 cannot ensure the superior catalytic performance, but the transportable Li + contributes to the outstanding POD-like activity of LiFePO4.

Conclusion
In summary, a detailed mechanism of the POD-like activity of Fe3O4 nanozymes is elucidated by characterizing the chemical composition and catalytic activity of the Fe3O4 NPs recycled from the long-term POD-like catalysis.These studies demonstrate that all Fe 2+ in Fe3O4 nanozymes contribute to their POD-like activity.The Fe 2+ inside the particle transfers electrons to the surface, regenerating the surface Fe 2+ that is directly involved in the sustained catalytic reaction.This process is accompanied by the outward migration of excess oxidized Fe 3+ from the interior of the crystal, which is a rate-limiting step.Analogous to the low-temperature oxidation of magnetite, Fe3O4 NPs participated in the POD-like reaction are eventually oxidized to γ-Fe2O3 NPs with a reduced POD-like capacity.Furthermore, this mechanism is well-validated on LiFePO4 NPs.This work reveals the depletable characteristic of Fe3O4 nanozymes that differ from natural enzymes and highlights the potential contribution of internal metal atoms in nanozymes-catalyzed reactions.Meanwhile, these findings bring new thoughts for the mechanistic study and rational design of nanozymes.

Fig. 1
Fig. 1 The synthesis of IONPs and cyclic POD-like catalysis.(a) Illustration of the synthesis process of IONPs.(b) The specific activity (anano) of these three IONPs with TMB as colorimetric substrates. (c) Diagram of the cyclic catalysis assay. (d) Kinetic study of anano values of Fe3O4 NPs with the days of cyclic catalytic reaction.(e) The fitted Fe2p XPS spectra and (f) Raman spectra of Fe3O4 NPs recycled after catalysis on days 0, 1, 3, and 5. (g) The Fe L-edge

Fig. 2
Fig. 2 The aeration oxidization kinetics of Fe3O4 NPs.Variation of UV-vis-NIR absorption of (a) cc-Fe3O4 NPs and (b) TD-Fe3O4 NPs with aeration oxidation time.Insets are photos of the suspensions corresponding to oxidation times at 0, 0.5, 1, 3, 5, 8, 10, and 12 h.All spectra and photos were obtained at the same Fe concentration.(c) Changes in anano of the oxidized cc-Fe3O4 NPs and TD-Fe3O4 NPs during the aeration oxidation.

Fig. 3
Fig. 3 Schematic diagram of the catalytic mechanism of the POD-like activity for Fe3O4 NPs.

𝐿𝐿𝐿𝐿𝐹𝐹𝐹𝐹𝐿𝐿𝑂𝑂 4 − 5 )𝐹𝐹𝐹𝐹𝐿𝐿𝑂𝑂 4 + 6 )Fig. 4 LiFePO4 2 M
Fig. 4 LiFePO4 NPs as verification materials and their POD-like activity.(a) The crystal structure of LiFePO4 and FePO4 viewed along the a, b, c-axis.The olivine structure is maintained during Li-ions insertion and extraction.(b) The SEM image of as-synthesized LiFePO4 NPs.Inset is a photo of LiFePO4 NPs aqueous solution.(c) The POD-like activity of LiFePO4 NPs (6.25 ug Fe/mL) with TMB (1.7 mM) as colorimetric substrates under the presence of H2O2 (0.8 M) in 0.2 M acetate buffer (pH = 3.6).(d) ESR spectra of spin adducts DMPO/•OH produced by LiFePO4 NPs (10 ug/mL) in the presence or absence of H2O2 (0.165 M ) in 0.2 M acetate buffer (pH = 3.6).(e) Comparison of the anano of as-synthesized LiFePO4 NPs and cc-Fe3O4 NPs.(f) Diagram of the POD-like catalytic reaction process of LiFePO4 NPs and Fe3O4 NPs.