Constraining the magnetic properties of ultrafine- and fine-grained biogenic magnetite

Abstract

Four samples containing ultrafine- and fine-grained magnetite of magnetoferritins and magnetotactic bacteria cells were magnetically characterized at both room and low temperatures. Transmission electron microscopy analysis showed that the biometrically synthesized magnetoferritins (M-HFn) have magnetite cores with a mean size of 5.3 ± 1.2 nm inside protein shells, while Magnetospirillum gryphiswaldense MSR-1 cell produced intracellular magnetosome magnetites have a mean size of 29.6 ± 7.6 nm, arranged in a single chain. A pure M-HFn sample (M₁), MSR-1 whole cell sample (M₄) and two samples (M₂, M₃) mixing M-HFn with MSR-1 whole cells in different weight percentages were measured, including hysteresis, temperature dependency of magnetization and remanence and frequency dependence of AC susceptibility at low temperature. At room temperature, the ultrafine-grained magnetite core of M-HFn of M₁ sample has a typical superparamagnetic (SP) behavior. The chain-arranged magnetosome magnetite of MSR-1 cells of M₄ sample shows a stable single-domain (SD) state. At low temperature, the M₂ sample with ~ 16 wt% SD magnetosome magnetite and the M₃ sample with ~ 43 wt% SD magnetosome magnetite behave somewhat similar to the M₁ (pure M-HFn), due to the SP component from M-HFn magnetite. With the dominance of SP magnetite in samples M₁, M₂, and M₃, the coercivity and saturation remanence decrease significantly as temperature increasing from 5 to 20 K. Of note, the magnetization and frequency dependence of AC susceptibility at low temperature are sensitive to SP magnetites in measured samples. The magnetosome magnetite produced by MSR-1 has a Verwey transition temperature at around 100 K, which is consistent with previous observations on magnetotactic bacteria. This study provides useful clues for identification of SP and SD magnetite in sediments, as well as related potential biomedical and biomagnetic applications.

Introduction

Superparamagnetic (SP) magnetite has been found in numerous geological samples, e.g., soils, pelagic sediments, tuffs and ice sheets, usually with some distinct magnetic properties; therefore, they are of great interests in rock magnetism, environmental magnetism and paleomagnetism. While progress has been made in identifying the SP magnetite in nature samples (Bedanta and Kleemann 2009; Creer 1961; Dunlop 1973; Lanci and Kent 2006; Liu et al. 2010; Maher 2016; Oldfield et al. 1981; Smirnov and Tarduno 2001; Tarduno 1995; Tauxe and Wu 1990; van de Moortele et al. 2007; Worm and Jackson 1999), some ambiguities in measurement interpretation remain, partially because of uncertainties in size distribution and particle magnetostatic interaction.

Worm and Jackson (1999) studied the magnetite in the Yucca Mountain tuff samples through measurement of hysteresis loops, isothermal remanent magnetization acquisition, thermal demagnetization and frequency and temperature dependence of susceptibility. They noted deviations exist between modeled and measured susceptibility, which was associated with size-dependent anisotropy, non-uniform magnetization and also uncertainties in the pre-exponential time (Worm and Jackson 1999). The pre-exponential time of magnetite was thought to be size-dependent and sensitive to particle interaction (Cao et al. 2010; Moskowitz et al. 1997).

Although SP magnetite grains are unable to retain a remanence at room temperature, numerical simulations and experimental measurements indicate that SP magnetite in sediments is abundant evidenced by their contribution to hysteresis loops (Tauxe et al. 1996). Their presence is indicative of magnetite reduction diagenesis at the Fe-redox boundary in pelagic sediments (Roberts et al. 2013; Tarduno 1995) and can reflect climate-associated pedogenesis in loess and soils (Maher 2016).

Stoichiometric SP magnetite with very narrow size distribution and good biocompatibility is also of great interest in magnetic nanomaterial production and related medical applications, e.g., contrast agents of magnetic resonance imaging (Bonnemain 1998; Hergt et al. 1998; Roch et al. 1999; Thorat et al. 2016; Tromsdorf et al. 2007; Wang et al. 2001), hyperthermia treatments of tumors (Hergt et al. 2008; Jordan et al. 1999), biomedical application (Fan et al. 2012; Gao et al. 2017; Schaefer et al. 2007; Thorek et al. 2006) and nanometric biomaterials (Amstad et al. 2011; Cao et al. 2014; Zhang et al. 2017).

Ferritin is a widely existing iron-storage protein in many living organisms throughout animals, plants and bacterias. It is a cage-like protein with an external diameter of 12 nm and an inner diameter of 8 nm, an ideal versatile platform for synthesis of size-controllable nanometer-scale ferrimagnetic particles. The structure of mature mammalian ferritin consists of a 24-subunit protein, composed of heavy subunits (H) and light subunits (L). Cao et al. (2010) used the recombinant human H-chain ferritin (HFn) and successfully synthesized mono-dispersed, non-interacting ferrimagnetic magnetoferritin (M-HFn) nanoparticles, which have magnetite cores with average diameters of a few nanometers (Cai et al. 2015; Cao et al. 2010). These biomimetic synthesized ferrimagnetic cores have extremely narrow size distribution and high crystallinity, are superparamagnetic at ambient temperature and due to their intact protein shell separation, have nearly no magnetostatic interactions (Cao et al. 2010; Walls et al. 2013).

Magnetotactic bacteria (MTB) intracellularly produce nanosized single-domain (SD) magnetite magnetosomes (30–120 nm), usually arranged in chains, allowing the microbes orientate in the ambient magnetic field (Blakemore 1975; Bazylinski and Frankel 2004), which is a model microorganism for biogeomagnetism study. MTB play important roles in sedimentary magnetism and iron cycling and have application in paleoenvironmental studies (Lin et al. 2014). Over the past decades, MTB have been found in a diverse range of aquatic environments, such as freshwater lakes, rivers, ponds, estuaries, lagoons, mangrove swamps, intertidal zones, deep-sea sediments, marine oxygen minimum zones, saline–alkaline lakes and hot springs, and appear to be important in the geochemical cycling of Fe, S, N and C and so forth (Bazylinski and Frankel 2004; Faivre and Schuler 2008; Kirschvink 1980; Kopp and Kirschvink 2008; Lin et al. 2014; Pan et al. 2005a; Schuler and Frankel 1999; Simmons and Edwards 2007; Zhou et al. 2012). Fossil magnetosomes (magnetofossils) have been widely identified in numerous of sediments, and they are important magnetic carriers (Chang et al. 2014; Channell et al. 2013a, b; Kopp et al. 2007; Li et al. 2013; Liu et al. 2015; Mao et al. 2014; Moskowitz et al. 1993, 2008; Pan et al. 2005b; Petersen et al. 1986; Yamazaki and Shimono 2012; Zhao et al. 2016).

In this study, we characterized the ultrafine-grained magnetite of M-HFn and fine-grained SD magnetosome magnetite by both room- and low- temperature magnetic measurements. The magnetoferritin, Magnetospirillum gryphiswaldense MSR-1 whole cells and two mixed samples with different concentrations of magnetoferritins were analyzed. The objective of this study is to examine magnetic properties of these two types of biogenic magnetites, in particular the traits of ultrafine-grained superparamagnetic magnetite. Applications of ultrafine- and fine-grained magnetite in rock magnetism, biomagnetic and biomedicine are also discussed.

Materials and methods

Preparation of samples

The M-HFn nanoparticles were synthesized by the recombinant ferritin cage using the method of Cao et al. (2010) with minor modifications. Fe(II) (25 mM (NH₄)₂Fe(SO₄)₂·6H₂O) was added in a rate of 80 Fe/protein/minute. Simultaneously, freshly prepared H₂O₂ (8.33 mM) was added as an oxidant in accordance with stoichiometric equivalents (1:3, H₂O₂/Fe²⁺). Then a theoretical 7000 atoms of Fe per protein cage was added to the reaction vessel and allowed to react for another 10 min. Then, 200 μl of 300 mM sodium citrate was added to each sample to chelate any free metal ions. Finally, purification was performed through size exclusion chromatography (Sepharose 6B, GE Healthcare) after centrifugation at 10,000 rpm for 30 min at 4 °C.

Magnetospirillum gryphiswaldense strain MSR-1 was cultured in a sodium lactate medium (SLM) at 30 °C while spinning at 100 rpm (Ding et al. 2010; Jogler and Schuler 2009); sterile ferric citrate was added as iron source. Fresh whole cells were collected by centrifugation at 8000 rpm for 10 min at 4 ^◦C after culturing for 24 h.

In this study, four samples (M₁, M₂, M₃ and M₄) were prepared and used for magnetic measurements, namely M₁, pure M-HFn nanoparticles, dry weight 3.1 mg; M₂, a mixture of M-HFn nanoparticles (39.4 wt%) and MSR-1 whole cells (60.6 wt%), dry weight 15.5 mg; M₃, a mixture of M-HFn (14.0 wt%) and MSR-1 whole cells (86.0 wt%), dry weight 22.5 mg; and M₄, pure MSR-1 whole cells, dry weight 12.8 mg. Note that the magnetization and remanence values of this paper were calculated using the sample’s dry weight. Samples were transferred into non-magnetic capsules and stored in anoxic chamber before magnetic measurements.

Transmission electronic microscopy analysis

Magnetite cores within M-HFn nanoparticles and magnetosome magnetite in MSR-1 whole cells were examined by transmission electronic microscopy (TEM, JEOL JEM2100) operating at 200 kV. The sizes of magnetoferritin and magnetosome were analyzed using standard analytical software. The major and minor axes of magnetite were used as the length (L) and width (W) of the crystal, respectively. The grain size was defined as (L + W)/2.

Room-temperature hysteresis, first-order reversal curve (FORC) and saturation isothermal remanent magnetization (SIRM) measurements

Room-temperature hysteresis loops, FORCs and SIRMs were measured using a VSM3900 magnetometer (Princeton Measurements Corporation, USA, sensitivity 5.0 × 10⁻¹⁰ Am²). A total 120 curves were measured in FORCs using an increasing field step of 0.721 mT with an averaging time of 500 ms. The FORC diagrams were processed using FORCinel version 1.18 software (Harrison and Feinberg 2008) with a smooth factor of 3.

Low-temperature magnetic measurements

Low-temperature magnetic measurements were taken with a Quantum Design MPMS XP-5 SQUID magnetometer (sensitivity 5.0 × 10⁻¹⁰ Am²). Hysteresis loops were measured at temperatures 5 K, 10 K and 20 K. Low-field magnetization curves were measured between 5 and 300 K in a field of 1 mT field after the sample was cooled from 300 to 5 K in zero-field (zero-field cooling, ZFC) and 1 mT field (field cooling, FC), respectively. The thermal decay of saturation isothermal remanence (acquired in 2.5T at 5 K) was measured between 5 and 300 K after the sample was cooled in ZFC and FC (in a 2.5-T field) from 300 to 5 K, respectively. The AC susceptibility measurements were taken between 5 and 300 K at frequencies of 1, 5, 10, 50, 100, 200, 500 and 1000 Hz in a peak AC field of 0.4 mT.

Results

Grain size and mass of magnetites in samples

Figure 1a–c shows that the magnetite cores of the synthesized M-HFn nanoparticles were mono-dispersed, well crystalline and with a mean size of 5.3 ± 1.2 nm (N = 509). Each MSR-1 cell contains a single chain that consists of 10–15 octahedral magnetite magnetosomes, which have a mean size of 29.6 ± 7.6 nm (N = 311) (Fig. 1d–g).

Fig. 1

TEM micrographs of M-HFn nanoparticles and MSR-1 cells. a M-HFn nanoparticles; b an enlarged M-HFn core; c size distribution histogram of magnetite cores in M-HFn nanoparticles. d MSR-1 whole cells; e magnetosomes chain; f a magnetite magnetosome; g size distribution histogram of magnetosomes in MSR-1 cells. h–j Mixed M-HFn nanoparticles and MSR-1 cell of sample M₂. j An electron diffraction pattern of the M-HFn nanoparticles of (i)

Figure 1h–i shows the M-HFn nanoparticles adhered to MSR-1 cell surface in the mixed sample M₂. Magnetite nanoparticles are separated by cell membrane and protein shell. Figure 1j shows the X-ray diffraction of the magnetite of M-HFn nanoparticles of Fig. 1i, confirming the crystalline ultrafine-grained magnetite in the M-HFn nanoparticles.

The proportion of magnetite in magnetoferritin and magnetosome in MSR-1 cells is 32.4% (Table 1) and 4% (Faivre et al. 2007), respectively. The mass of magnetite in each sample M₁–M₄ was calculated.

Table 1 Hysteresis parameters obtained from room-temperature measurements

FORCs diagram and magnetic interaction analysis

FORCs diagrams are presented in Fig. 2. Of note, magnetostatic interactions between magnetite particles in measured samples are very weak (H_{b, 1/2} = 2.1 mT), due to protein and cell separations. Samples M₂, M₃ and M₄, containing different portions of SD magnetosome magnetite of MSR-1 whole cells, have nearly the same feature with a coercive value of 13.8 mT, and a negative region at the lower left region, a indicative of SD magnetite. No signal of sample M₁ was detectable under the used measurement condition.

Fig. 2

FORCs diagrams of samples M₂, M₃ and M₄ measured at room temperature

Hysteresis

Room-temperature hysteresis loops of four samples are shown in Fig. 3. Hysteresis parameters are summarized in Table 1. It is noted that the sample M₁ containing ultrafine-grained magnetite of M-HFn is a typical superparamagnetic, which has high saturation magnetization (M_s) but no detectable coercivity (H_c), consistent with previous studies (Cao et al. 2010; Tauxe et al. 1996). Sample M₄ containing chain-arranged SD magnetite magnetosomes of MSR-1 cells has a potbelly shape loop, a uniaxial single-domain behavior. Samples M₂ and M₃ containing both M-HFn and MSR-1 have H_c values of 1.8 mT and 4.5 mT and M_s values of 4.1 Am² kg⁻¹ and 1.6 Am² kg⁻¹, respectively, varying with the portions of SP and SD components.

Fig. 3

Hysteresis loops of four samples measured at room temperature

The hysteresis loops measured at low temperature (Fig. 4 and Table 2) showed that all samples have the capacity to carry remanence at T ≤ 20 K. Their M_s and saturation remanence (M_rs) are 23.7, 10.9, 3.8 and 0.4 Am² kg⁻¹ and 7.4, 3.4, 1.2 and 0.2 Am² kg⁻¹ at 5 K, respectively, for M₁, M₂, M₃ and M₄. From 5 to 20 K, M_s of sample M₁, M₂ and M₃ decreases slightly but the M_rs drops significantly due to thermal unblocking of contained ultrafine-grained magnetite; for sample M₄, the M_s and M_rs remain nearly constant, while the H_c drops from 70 to 50 mT. Interestingly, although the H_c values of each sample dramatically decrease with temperature increment, it is noted that the samples M₁, M₂ and M₃, in spite of difference concentration of M-HFn, have very similar coercivity values at 5 K, 10 K and 20 K. This can be explained by the dominance of the ultrafine-grained magnetite in M-HFn.

Fig. 4

Hysteresis loops of four samples measured at 5 k

Table 2 Hysteresis parameters obtained from low-temperature measurements

Low-temperature magnetic variation, blocking temperature and Verwey transition

Figure 5 shows the temperature dependence of low-field (1.0 mT) magnetization from 5 to 300 K after the ZFC and FC treatments. The samples M₁, M₂ and M₃ display similar behaviors but different blocking temperatures (T_b): 35 K, 34.1 K and 27.1 K, respectively. Of note, the ZFC and FC curves of M₁ sample merges at about 60 K, while the ZFC and FC curves of M₂ and M₃ are diverged and join at much high temperatures, indicating the presence of magnetosome magnetite. On the contrary, sample M₄ containing pure MTB displays totally different trend in the ZFC and FC curves. On the warming curve, the magnetization of M₄ after either ZFC or FC treatment increases rapidly at temperature between 90 and 130 K, indicative of the Verwey transition temperature (T_V) of magnetite around 100 K.

Fig. 5

Low-field (1.0 mT) magnetization as a function of temperature (5–300 K) measured after zero-field-cooling (ZFC) and field-cooling (FC) treatments

Thermal demagnetization curves of SIRMs acquired in a 2.5 T field at 5 K for samples M₁, M₂ and M₃ showed a rapid drop from 5 to 20 K (Fig. 6). The 2.5 T ZFC and FC curves of M₁ are nearly superimposed. The rapid drop of SIRM is due to unblocking (Brown 1963). The median unblocking temperature in the FC curve is T_bm = 10.5 K. The remanence of ZFC and FC decreased to zero near at 25 K, which indicate all the particles are unblocked. As expected, the decay curve of the sample M₄ containing pure MTB sample shows a Verwey transition behavior around 100 K; the calculated δ-ratio is 1.37. The Verwey transition in samples M₂ and M₃ is nearly completely suppressed, although these two samples contain significant amount of magnetosome magnetite.

Fig. 6

Decay curves of saturation isothermal remanent magnetization acquired in a 2.5 T field after zero-field-cooling (ZFC) and field-cooling (FC) treatments

Frequency dependence of AC susceptibility

AC magnetic susceptibility measurements yield both a real component (χ′) and an imaginary component (χ′′) (Fig. 7). It is clear that M₁, M₂ and M₃ containing the ultrafine-grained M-HFn magnetite display generally similar variation with temperature and frequency: on warming, χ′ increased rapidly up to 34–46 K and decreased up to 300 K, while χ′′ increased rapidly up to 25–34 K and decreasing again up to 80 K. Above this temperature, χ′′ approaches zero. The peak values of χ′ decrease with increasing frequency, and blocking/unblocking temperature identified by peak values of χ′ increases with increasing frequency (see inset in Fig. 7). The peak values of χ′ and χ′′ decreased with decreasing proportions of M-HFn magnetite in the samples. A slight concave appears in the χ′′-T curve around 80–100 K in M₃ sample probably due to the effect of SD magnetosome magnetite.

Fig. 7

Curves of real and imaginary components, χ′ and χ″, of the AC susceptibility versus temperature. Samples were measured at frequencies of 1 Hz, 5 Hz, 10 Hz, 50 Hz, 100 Hz, 200 Hz, 500 Hz, 1000 Hz in a peak AC field of 0.4 mT

Sample M₄ containing only MTB cells has a totally different behavior in the χ′ -T and χ′′-T curves. Both χ′ and χ′′ increased slightly up to 85 K and then increased much more rapidly between 85 and 135 K, corresponding to the Verwey transition of magnetite. Above 130 K, χ′ continuously increased up to 300 K, while χ′′ decreased slightly. M₄’s χ′ shows frequency dependence, and the values of χ′ decrease with increasing frequencies.

Discussion and conclusions

Magnetic signature of SP magnetite

Ferritin-based ultrafine-grained magnetites can be taken as ideal sample for superparamagnetism study, because of (1) controllable and uniform grain size, (2) mono-dispersed, (3) lack of magnetic interaction, (4) good availability and (5) “ideal” SP behaviors at room temperature. Hysteresis loop confirmed that the ultrafine-grained magnetites in sample M₁ with a mean size of 5.3 ± 1.2 nm are superparamagnetic at room temperature (Fig. 3), which corresponds to a T_b of 35 K (Fig. 5); at T < T_b, they can carry remanence (Fig. 6). Their frequency dependences of AC magnetic susceptibility are significant, e.g., peak temperature and value of AC susceptibilities decrease with frequency from 1 Hz to 1 kHz (Fig. 7). Cao et al. (2010) previously determined the value of pre-exponential factor f₀ in the Néel–Arrhenius equation from AC susceptibility data of M-HFn nanoparticles with a mean grain size of 3.9 ± 1.2 nm: (9.2 ± 7.9) × 10¹⁰ Hz and the extrapolated value of M_rs/M_s = 0.5 and B_cr/B_c = 1.12 at 0 K, which suggests the ferrimagnetic M-HFn is dominated by uniaxial anisotropy.

Grain size effect of SP magnetite is an interesting subject to probe. In the combination of data of this study and available magnetic results of magnetoferritins (Cai et al. 2015; Yang et al. 2017; Zhang et al. 2017), it is noted that M_s, T_b, H_c and peroxidase-like activity enhance with grain sizes. Recently, Cai et al. found that the synthesized M-HFn containing Fe₂O₃ cores with a mean grain size of 2.2 nm has a high longitudinal relaxivity value of 0.94 mM⁻¹ s⁻¹ and they proposed it as a potential positive contrast agent for magnetic resonance angiography (Cai et al. 2018). Efficiency of hyperthermia using ultrafine-grained iron oxides is also related to grain size of magnetic minerals (Banobre-Lopez et al. 2013; Deatsch and Evans 2014). Therefore, grain size effects of SP magnetites on mineral magnetism and related applications for biomedical detection and therapy need investigations.

Magnetic signature of bacterial SD magnetite

Over the past several decades, there are few studies on their magnetic properties of MTB samples (Ding et al. 2010; Li et al. 2009, 2010; Moskowitz et al. 1993; Prozorov et al. 2007; Weiss et al. 2004; Wang et al. 2015). The cultivated MSR-1 cells used in current study (M₄ sample) contain in average 10–15 magnetite magnetosomes with a mean size of 29.6 ± 7.6 nm, arranged in single chains (Fig. 1d–g); the size of magnetosome magnetite in MSR-1 is slightly smaller than 38 nm of the wild-type MSR-1 (Ding et al. 2010), 44 nm of Magnetospirillum strain XM-1 (Wang et al. 2015) and 42 nm of Magnetospirillum magneticum AMB-1 (Li et al. 2009), probably, due to the strain degradation in cultivation. Strain degradation of MSR-1 may be caused by genetic instability (Ullrich et al. 2005), which is a common cause of strain degradation in industrial production (Gravius et al. 1993). The FORC diagram is nicely characterized by a rather narrow distribution around H_c,FORC ~ 13.8 mT along the horizontal axis, the so-called central ridge distribution (Egli et al. 2010) and a negative area in lower left region (Fig. 3). The T_v of 100 K is comparable to other cultivated MTB and uncultivated MTB (Li et al. 2010; Ding et al. 2010; Pan et al. 2005b; Moskowitz et al. 1993; Wang et al. 2015), confirming the lower T_v as a good indicative of MTB-produced magnetites.

Analysis of SP contribution in mixed samples

In this study two mixed samples with different portions of SP and SD magnetite were magnetically characterized. As mentioned previously, SP magnetite and SD magnetite in measured samples are well separated by protein and membranes; it led to magnetic interaction ignorable. With the known composition and grain size distribution, results of samples M₂ (containing 84 wt% SP magnetite and 16 wt% SD magnetite) and M₃ (containing 57 wt% SP magnetite and 43 wt% SD magnetite) provide us unambiguous constraints on contributions of either SP or SD component.

The room-temperature FORC diagrams for M₂ and M₃ are similar to that of M₄ (Fig. 3), which is clearly signal of chain-arranged SD magnetosome magnetites. In contrast, the bulk magnetic properties, e.g., H_c, M_s and M_rs, χ′ and χ′′, measured at low temperatures of sample M₂ and M₃ (Table 2; Figs. 4, 5, 6, 7) are controlled by the SP magnetite. Comparing samples M₁ to M₃, at T ≤ 20 K, M_rs values decay rapidly with both temperature and SP concentrations; the M_s decay with SP concentrations but not with temperatures (Table 2); it suggests that the decay rates of M_rs and M_s at T < 20 K may be useful in estimating SP component in samples.

At room temperature, as expected, M_rs/M_s of samples M₄ and M₂ decrease from 0.37 to 0.02 while their H_rs/H_c increase from 1.58 to 8.72, respectively, reflecting significant influence of SP component to the hysteresis (Tauxe et al. 1996; Dunlop 2002). However, it should be aware that the measured mixed samples contained only non-interacting SP and SD magnetite. More measurements on samples with different concentrations and compositions are needed.

Identification of fossil magnetite in sediments

Although it has been found that present-day magnetotactic bacteria are ubiquitous in a diverse range of aquatic environments, identification of fossil magnetite from sediments as well as deciphering their paleoenvironmental and paleomagnetic records is somehow not straightway, because of relatively low concentrations of magnetofossils in sediments and mixtures with abundant detrital origin single- and multi-domain magnetites. This study confirms the magnetosome magnetite in MSR-1 does have a T_v around 100 K and central ridge behavior in FORCs. Through extensive investigation on both uncultivated and cultivated MTBs over past decades, it has been well demonstrated that magnetofossils can be identified through measurement of combination of the delta ratio (Moskowitz et al. 1993), FORCs (Egli et al. 2010), identification of lower T_v temperatures (Chang et al. 2014; Pan et al. 2005b), ferromagnetic resonance (Weiss et al. 2004), unique chemical signatures (Amor et al. 2015), Fe isotope signature (Amor et al. 2016), as well as TEM examinations. Recently, Lin et al. found through genetic and genomic analyses that magnetotactic bacteria may appear in Archean time, indicative of an Archean geomagnetic field and a small amount of oxygen in oceanic habitats (Lin et al. 2017). To uncover potential paleoenvironmental and paleomagnetic records carried by magnetofossils in sediments over geological time is desired.