Novel scanning magnetic microscopy method for the characterization of magnetic nanoparticles

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm

Research articles

Novel scanning magnetic microscopy method for the characterization of magnetic nanoparticles

T

⁎

Jefferson F.D.F. Araujoa, , Tahira, Soudabeh Arsalanib, Fernando L. Freire Jr.a, Gino Mariottoc, Marco Cremonaa, Leonardo A.F. Mendozad, Cleanio Luz-Limae, Quaid Zamana, Tommaso Del Rossoa, Oswaldo Baffab, Antonio C. Brunoa a

Department of Physics, Pontifical Catholic University of Rio de Janeiro, 22451-900 Rio de Janeiro, Brazil Department of Physics, FFCLRP, University of São Paulo, 14040-901 Ribeirão Preto, SP, Brazil c Department of Informatics, Università di Verona, Strada le Grazie 15, I-37134 Verona, Italy d Department of Electrical Engineering, State University of Rio de Janeiro, 20550-900 Rio de Janeiro, Brazil e Department of Physics, Federal University of Piauí, 64.049-550, Teresina, PI, Brazil b

A R T I C LE I N FO

A B S T R A C T

Keywords: Iron oxide magnetic nanoparticles Co-precipitation Pulsed laser ablation in liquid Magnetic scanning microscope

In this paper, a new method is presented for the magnetic characterization of nanoparticles that is especially suitable for samples with a low mass, on the order of tens of micrograms. We investigated the magnetic and morphological properties of the colloidal dispersions of iron oxide magnetic nanoparticles that were synthesized by two methods: chemical precipitation (co-precipitation) and pulsed laser ablation in liquid (PLA). We measured the stray field generated above the samples by scanning magnetic microscopy (SMM) and used a nonstandard model to obtain the magnetization of the nanoparticles. We assessed the performance of the method by comparing the magnetization curves with measurements obtained using commercial magnetometers. The errors in the saturation and remanent magnetization were found to be approximately ± 0.18 Am2/kg and ± 0.6 Am2/ kg, respectively. As the samples exhibited a superparamagnetic state, we also used the magnetization curves to estimate the average size of the synthesized nanoparticles, which were found to be consistent with the results obtained using other techniques.

1. Introduction The magnetic properties of materials are a research topic in physics and engineering [1–5]. The applications of magnetic materials are determined by important parameters, such as the Curie temperature, the saturation magnetization, the remanent magnetization, coercive fields and the magnetic anisotropy. Magnetic nanoparticles (MNPs) have demonstrated considerable potential for medical applications, including magnetic resonance imaging (MRI), drug delivery, magnetic particle imaging (MPI) and hyperthermia treatment [6–7]. Frequently used MNPs in biomedical applications are magnetite (Fe3O4), maghemite (Fe2O3) and their mixtures, which have low toxicity and high susceptibility and biocompatibility compared with other MNPs, such as CoFe2O4, ZnFe2O4 and MnFe2O4. The physical properties of MNPs, including size, composition, shape and surface chemistry, vary widely and influence the biological properties and clinical applications of these materials. Many physical and chemical methods are used for MNP synthesis. The most commonly used methods are co-precipitation and

⁎

pulsed laser ablation in liquid (PLA) [8–11]. Our present configuration for producing MNPs by PLA does not allow us to fabricate samples of more than a few tens of micrograms (µg). Such small amounts can be used for in vitro studies [12]. However, the magnetic properties of such low-mass samples cannot be characterized using some magnetometry techniques [13–15]. Accordingly, we develop a method for characterizing low-mass magnetic samples from magnetic field maps obtained by a scanning magnetic microscope (SMM) [16–18]. We compare the magnetization results with measurements obtained by standard highsensitivity commercial magnetometers, yielding errors of approximately ± 0.18 Am2/kg (0.37%) at saturation and below ± 0.60 Am2/ kg (1.2%) for the remanent magnetization. We also use the magnetic measurements to estimate the average sizes of the MNPs, which are consistent with the results obtained using traditional techniques, such as transmission electron microscopy (TEM). The custom-built SMM used in this study can apply magnetic fields of up to 550 mT to a sample. The SMM has a scanning range of 150 to 150 mm with µm resolution. In its current configuration, the

Corresponding author. E-mail address: jferraz@fis.puc-rio.br (J.F.D.F. Araujo).

https://doi.org/10.1016/j.jmmm.2019.166300 Received 5 July 2019; Received in revised form 28 November 2019; Accepted 12 December 2019 Available online 17 December 2019 0304-8853/ © 2019 Elsevier B.V. All rights reserved.

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

Fig. 1. (a) Schematic of SMM, the sensors (Sensor 1 and Sensor 2) are soldered to a printed circuit board (PCB). (b) Photo of sample holder, with printed text “PUCRio”. (c) Experimental map made by SMM under a 20 mT field.

microscope is equipped with a pair of commercial Hall effect sensors operating as an axial gradiometer. The output noise measured at 6.0 Hz is approximately 250 nTrms/√Hz in an unshielded environment, and the magnetic moment sensitivity is 4.20 ×10−11 Am2.

had a 3.0 mm diameter and a depth of 3.0 mm. Considering the sourceto-sensor distance and the cavity geometry, we used a cylindrical current sheet model (see Eq. (1)) to obtain the reference sample magnetization [19]:

2. Magnetic microscope

mz =

The SMM can be used to characterize magnetic samples with masses on the order of µg. We used double-sided adhesive tape to fix a sample upside down onto a sample holder, which was placed between the poles of an electromagnet that could generate uniform magnetic fields of up to 550 mT between 40 mm diameter pole tips in a 17 mm gap under a 4.0 DC current. The pole axis of the electromagnet was oriented in the vertical direction, as shown in Fig. 1(a). The response of the sample to the applied field was detected by two Hall effect sensors, hereafter denoted as Sensor 1 and Sensor 2, wherein a GaAs element was incorporate in a surface-mount technology package [16]. The Hall element was 200 μm in diameter and was placed at a nominal distance of 130 μm from the sample surface. The capability of the equipment is illustrated in Fig. 1(c) in terms of the magnetic field response of black toner particles under a 20 mT field for a text printed on regular paper, which is shown in Fig. 1(b). The small sensor-to-sample distance and the spatial resolution of the SMM results in a magnetic image that is very similar to the physical printed image. The toner particles that are missing in the small region at the top of the letter C in Fig. 1(b) can be seen in the magnetic image in Fig. 1(c). The axial gradiometer was primarily designed to attenuate the applied field, thereby increasing the dynamic range of the instrument and enabling its operation under strong applied fields. An applied field attenuation of approximately 10-3 was attained by adjusting the polarization current of Sensor 1 [16,19]. We observed that the gradiometer also attenuated the ambient magnetic noise.

4π 2a2L −L 0

μ0 ∫

∫0

2π (x − x 0 ) acos (ϕ) r3

dϕdx

Bz (x , y, z ) (1)

where r = (x − x 0)2 + (y − asin (ϕ))2 + (z − acos (ϕ))2 , µ0 is the permeability of free space, and a is the radius of a uniformly magnetized cylinder of length L. We determined the actual distance z between the top of the sample and the sensor by performing a linear scan over the sample to measure Bz . We analyzed only the spatial dependence of Bz for a = 1.5 mm and L = 3.0 mm. We used a least-squares fit to obtain a distance of 138 μm between the Hall sensor element and the top of the sample. We used this distance and a bulk Fe3O4 density of 5.19 × 10+3 kg/m3 for to obtain a magnetization of 75.8 Am2/kg under 500 mT, which is approximately 0.16% greater and 0.21% lower than the values obtained from a vibrating sample magnetometer (VSM) and a Hall magnetometer, respectively [20–21]. 3.2. Choice of sample holder We tested several materials available in the laboratory for fabrication into a sample holder with a spatial uniform response and low susceptibility. Fig. 2(a)-(c) show the magnetic field maps of sample holders made of glass, quartz, and acrylic, respectively. The maps were obtained under 500 mT and 100 mT applied in the z-direction, which was the axis of the Hall element sensitivity. Fig. 2(a) shows two maps of the z-component of the magnetic field, which were generated by the glass sample after an acid bath: the first map shows an induced field of approximately 3.0 mT under 500 mT, and the second map shows an induced field of approximately 0.8 mT under 100 mT. Fig. 2(b) shows two magnetic maps of the quartz sample after an alcohol bath: the first and second maps show induced fields of approximately 1.5 mT and 0.4 mT, respectively. Fig. 2(c) shows two scans of the acrylic sample after an alcohol bath: the first and second maps show induced fields of approximately 1.0 mT and 0.25 mT, respectively. All of the maps show a rather uniform spatial response relative to the sample size. To quantify these results in terms of the magnetization and aid in the selection of a sample holder, we considered different theoretical models to obtain the magnetization curve (the magnetic moment

3. Sample holder selection and SMM calibration 3.1. Microscope calibration A calibration procedure was performed to obtain the magnetic moment of the samples from the SMM output voltage. We used 99.9% magnetite microparticles (Merck KGaA) as the calibration source. We inserted 18.2 mg of these microparticles into a cylindrical cavity, which 2

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

(a) 500 mT

(b)

100 mT

500 mT

(c) 500 mT

100 mT

(d) 100 mT

Fig. 2. Magnetic maps of samples under 500 mT and 100 mT for: (a) 2.5 mm × 1.3 mm rectangular glass sample; (b) 2.5 mm diameter quartz disk and (c) 1.7 mm × 0.9 mm rectangular acrylic sample and (d) magnetization results obtained from modeling for an external magnetic field step of 5 mT.

4. Experimental methods

divided by the material mass) of each sample holder. A uniformly magnetized rectangular prism model was chosen because of its similarity to the sample geometries shown in Fig. 2(a) and (c) (glass and acrylic, respectively). For a rectangular prism with a uniform volume of magnetization mz, the z-component of the magnetic field generated at any point in the Bz(x, y, z) space can be written as [16]

4.1. Synthesis of colloidal dispersions of iron oxide MNPs The iron oxide MNPs were synthesized by both PLA (MNPs-I) in water and co-precipitation (MNPs-II) of ferric and ferrous salts. A pure iron disk (Kurt & Lesker, USA, 99.99% purity) was used as a target material for the synthesis of iron MNPs-I in water. The disk was first gently rubbed with sandpaper to remove surface oxides and then ultrasonically cleaned twice in Milli-Q water for 10 min. Afterwards, the cleaned iron disk was rinsed in 2 mL of ultrapure deionized water and irradiated through the air–liquid interface with a nanosecond laser beam (lambda = 1064 nm, 10 Hz pulse repetition rate), where the experimental setup has been described elsewhere [22]. The concentration of the colloidal dispersion of iron oxide MNPs was controlled by adjusting the distance between the lens (focal length: 13 cm) and the target to optimize the fluence of the laser beam at the sample plane on the order of 0.2–0.4 J/cm2. The PLA process lasted for two hours, and a bar magnet was used to collect only the precipitated MNPs from the colloidal solution. Finally, the resulting MNPs were washed several times with deionized water. In each ablation cycle, the total mass/ concentration of MNPs was measured by inductively coupled plasma mass spectrometry (ICPMS) to be less than 65 µg/32 ± 2 µg/mL. The co-precipitation procedure is described in [9]. Briefly, we first dissolved 2.25 g of FeCl3·6H2O in 8.5 mL of distilled water. Then, we dissolved 1.32 g of FeCl2·4H2O in 3.5 mL of aqueous hydrochloric acid (5.45 M). We added a mixture of FeCl3(4 mL) and FeCl2(1 mL) to a basic NH4OH solution (1.28 M) with vigorous stirring in a water bath at 90 °C. The solution immediately turned black, indicating the formation of iron oxide (Fe3O4) MNPs. We precipitated the MNPs by magnetic separation and washed the MNPs several times with Milli-Q water until the solution reached a neutral pH. Finally, we dried the Fe3O4 at room

Bz (x , y, z ) μ mz = 0 [F (−x , y, z ) + F (−x , y, −z ) + F (−x , −y, z ) 4π + F (−x , −y, −z ) + F (x , y, z ) + F (x , y, −z ) + F (x , −y, z ) + F (x , −y, −z )]

(2)

where the F function is defined as

F (x , y, z ) = arctan

(x + a)(y + b) (z + c ) (x + a)2 + (x + b)2 + (x + c )2

and a, b and c are the half-width, the half-length and the halfthickness of the rectangular prism, respectively. We used a cylindrical current sheet to model the cylindrical quartz sample holder shown in Fig. 2(b). This model was the same as that used to calibrate the microscope (see Eq. (1)). Fig. 2(d) shows the magnetization results obtained from the modeling process. All of the samples exhibited paramagnetic behavior up to 500 mT: the quartz and acrylic samples exhibited the lowest slope and thus the lowest magnetic response under an applied magnetic field. The acrylic sample also exhibited the lowest susceptibility and was therefore selected to characterize the magnetic samples.

3

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

30

4.2. Raman spectroscopy

27

Intensity (arb. units)

temperature [9].

To prepare the samples for the Raman analysis, we deposited the nanomaterials that were synthesized by the chemical and physical routes over a gold thin film and quenched the luminescence of the nanomaterials following the experimental procedures in [22]. We used a Horiba Jobin-Yvon (model LabRAM HR800) MicroRaman spectrometer to carry out a careful vibrational spectroscopic analysis of the precipitated nanoparticles, which were in the form of granular aggregates. The spectrometer was equipped with a He-Ne laser as an excitation source (λexc = 632.8 nm) and a notch filter for Rayleigh line rejection. We focused the laser beam onto the sample surface through a 100× objective lens with a numerical aperture of 0.9 to produce a spot size of approximately 1 µm. The scattered radiation was dispersed by a diffraction grating of 600 lines/mm and detected at the spectrograph output by a multichannel detector, a CCD with a resolution of 1024 × 256 pixels, which was cooled by liquid nitrogen, and exhibited a maximum efficiency in the visible red regime. The average spectral resolution was approximately 1 cm−1/pixel over the spectral range of interest, and the notch filter produced a low wavenumber limit of approximately 200 cm−1. We collected Raman spectra for the backscattering geometry over the Stokes-shifted region, maintaining the irradiation power at the surface of the sample below 100 µW to prevent damage to the sample. Under these excitation conditions, very long integration times (typically 1800s) were required to collect spectra with an optimal signal-to-noise ratio. MicroRaman measurements were carried out on different aggregates of both types of MNPs under the experimental conditions described above, and the recorded spectra exhibited very good reproducibility. Moreover, repeated measurements over the same sample grain produced overlapping spectra, indicating that there was no structural damage to the sample surface from the laser beam.

24 21

668 712

500 465

MNPs - I MNPs - II

382 317

548

18 15 12 9 6 300

600

900

1200

1500

1800

-1

Raman shift (cm ) Fig. 3. Raman spectra of iron oxide MNPs synthesized by PLA (MNPs-I, red line) and co-precipitation (MNPs-II, black line). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

which according to group theory, gives rise to five Raman modes: three T2g, one Eg and one A1g [24]. Three of these modes are clearly present in the MNPs-I spectrum: the strongest peak at approximately 670 cm−1 is identified as the A1g mode and corresponds to the stretching vibrations of oxygen atoms along the Fe-O bonds, whereas the weaker bands at 548 and 317 are associated with the T1g and E2g modes, respectively [28]. Finally, the relatively low intensity and very broad bands above 1000 cm−1 probably result from second-order scattering processes. The MNPs-II spectrum has a strong scattering component at approximately 712 cm−1, which turns out as a shoulder of the main peak at approximately 668 cm−1. This spectral component, together with other broad features resulting from band overlap between 350 and 550 cm−1 (i.e., bands at 382, 465 and 500 cm−1), are characteristic of the iron oxide maghemite [γ-Fe2O3], which has an inverse spinel structure and can be considered as an iron-deficient form of magnetite [24,29]. In conclusion, the chemically precipitated powders have a more complex structure (maghemite mixed with magnetite) than that of the MNPs-I, in which only the iron oxide magnetite phase is detected by Raman spectroscopy.

4.3. Dynamic light scattering and transmission electron microscopy We initially determined the size distributions of the synthesized nanomaterials by both dynamic light scattering (DLS) and TEM. We performed DLS using a Nano SZ-100 HORIBA Scientific nanoparticle analyzer, Japan, with an incident beam at a wavelength of 532 nm. For the TEM measurements, we used a Tecnai G2 Spirit TWIN FEI microscope, USA, operating at 120 kV with a LaB6 (lanthanum hexaboride) filament. We prepared the sample for TEM analysis by diluting the nanomaterial colloidal solution to a concentration of 8 µg/mL and depositing a 20 µL drop of the solution over a copper grid covered with a conductive polymer. The grid was then allowed to air dry overnight.

5.2. Magnetization curve Fig. 4(a) shows the fabricated acrylic 17 × 17 × 4 mm sample holder. A cylindrical cavity (400 µm diameter, 400 µm depth) lies approximately at center of the holder and is filled with 50 µg of the MNPsII produced by co-precipitation. Fig. 4(b) shows the magnetic map of the z-component of the field response obtained over a 2 mm × 2 mm scan under an applied 420 mT field in the same direction. The measured response is the stray field of the MNPs and the contribution of the sample holder. We used this map to obtain the maximum intensity response line over a longer scan of 20 mm, as shown in Fig. 4(c) and 4(d). The S1 curve represents this line, where we can clearly observe the edge of the sample holder at approximately x = 6 mm. The S2 line is the measured contribution of the sample holder before being filled with the MNPs. The line labeled S1-S2 is the difference between the two curves. In this way, we determined the field induced by the MNPs alone (see inset). We repeated this process to obtain maps under different applied fields that are shown in Fig. 5. We used the cylindrical current sheet model (Eq. (1)) to obtain the magnetic moments of the sample for each applied field. Fig. 6(a) corresponds to selected cross-sections of the field maps (Fig. 5) that pass

5. Results and discussion 5.1. Raman spectroscopy Fig. 3 shows two typical Raman spectra of the colloidal aggregates of the iron oxide MNPs synthesized by co-precipitation (MNPs-II, black) and PLA (MNPs-I, red). In general, the observed main Raman bands of iron oxide were imposed on a lower-intensity, continuous, nearly flat luminescent background. The spectra of the two precipitated MNPs clearly show certain common features that appear to be related to the same iron oxide phase present in both types of MNPs. However, additional bands in the MNPs-II spectrum suggest a more complex structural arrangement in this system, as is also reflected by the broad bandwidths of the MNPs-II peaks. The MNPs-I spectrum is simpler than that of the MNPs-II and consists of three main spectral features centered at approximately 317, 548 and 668 cm−1, which are typical signatures of magnetite (Fe3O4) [23–27]. Magnetite has a spinel structure belonging to space group Oh7, 4

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

(b)

(a)

(c) Sample Sample holder

(d)

z Sensor 1 Y

Sensor 2

x

Sample

Fig. 4. (a) Acrylic square-shaped sample holder with 1.7 cm side length used to characterize magnetic MNPs, which are placed in the center of the 400 × 400 µm cylindrical cavity. (b) Magnetic map of ≈50 µg magnetic Fe3O4 MNPs-II placed in cylindrical cavity of sample holder. (c) Results obtained along one axis under a 420 mT applied field; S1 path passes through the center of the cavity, which contains 50 µg of MNPs, and S2 path passes through the center of the empty cavity; S1-S2 is the difference between two paths. (d) Spatial position of the magnetic reading system and the magnetic map shown in (a).

peak temperature of the ZFC curve is identified as the blocking temperature (TB). Fig. 6(e) shows that the magnetization curves at ambient temperature are above the blocking temperature [30]. We also obtained the magnetic moment directly from the experimental maps (Fig. 7(a)). We constructed a theoretical current cylinder model (see Eq. (1)) for the entire mapping region (see Fig. 7(b)) using all the points in the 2.0 × 2.0 mm2 map, instead of only the curve passing through the point with the maximum field intensity, as in Fig. 6(a). Note that the maximum intensity region occurs in different regions in the maps obtained by SMM and the theoretical model (see Fig. 7(a) and 7(b)). To verify if this theoretical technique is appropriate, we subtracted the map obtained directly from the SMM of the theoretical model to produce the result shown in Fig. 7(c). The figure shows that the points with the maximum field intensity do not lie in the same region. Thus, this method is less precise than the previous modeling method (Fig. 6(a)-(c)). To verify the model, we compare the magnetic moment values obtained by this technique with measurements made with the SQUID magnetometer. The magnetic characterization curve in

through the maximum in each map. A 200 µm step was used for each point measurement. The blue and red curves were obtained under positive and negative magnetic fields, respectively. Fig. 6(b) demonstrates an example (at 50 mT) of fitting the theoretical model adjusting the magnetization value (solid black curve) and the values obtained for the induced field of the sample (blue circles). Fig. 6(c) compares the magnetization results obtained using our technique for a 50 µg sample (blue circles) with the values obtained using a MPMS SQUID magnetometer (Quantum Design Inc.) for 1.7 mg of the same MNPs (solid red curve). The absolute magnetization errors are shown in the inset of Fig. 6(c). The errors are ± 0.18 Am2/kg near saturation and approximately ± 0.6 Am2/kg for field values close to zero. The magnetization curves of the MNPs-I and MNPs-II obtained using our method are compared in Fig. 6(d). Both curves exhibit superparamagnetic behavior and do not exhibit remanence or coercivity. Fig. 6(e) shows the zero-field-cooled (ZFC) and field-cooled (FC) curves of the MNPs produced by PLA (MNPs-I) and co-precipitation (MNPs-II). In the ZFC curves, magnetization increases with the temperature. The 5

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

400 mT

300 mT

230 mT

140 mT

70 mT

0 mT

-70 mT

-140 mT

-230 mT

-260 mT

-300 mT

-400 mT

-500 mT

500 mT

260 mT

Fig. 5. Maps of 2.1 mm × 2.1 mm area around a cylindrical cavity containing a few tens of µg of MNPs under applied fields from 500 mT to 500 mT.

particles. The particle diameter determined by DLS is the hydrodynamic size of the particle and therefore provides information on the size distribution of the aggregates rather than the sizes of individual particles. Both methods provide different types of information and very often correspond with the magnetic diameter of a particle. Thus, the accuracy of the characterization method was assessed by comparing the values of the obtained magnetic diameter with those obtained using TEM and DLS. As the samples were at room temperature and in a superparamagnetic state, our measurements could be used to estimate the average magnetic diameter of the MNPs from the magnetization curves. Using Eq. (3) [19,33] yields

Fig. 7(d) shows that the error in the magnetic moment is approximately ± 9 Am2/kg in the 0.5 T region. This result could be attributed to the difficulty in spatially adjust the maps generated by the model and experimental data and also to the inclusion in the fitting routine of data with low signal-to-noise ratio at positions far from the center of the map.

5.3. Average diameter In addition to characterizing the MNPs behavior at room temperature, we estimated the average diameter of the MNPs from the magnetization curves (see Fig. 6(d)) for the samples that exhibited superparamagnetic behavior at room temperature, as shown in Fig. 6(e) [31–32]. The average diameter of MNPs is usually obtained using techniques such as TEM and DLS. The average diameter of MNPs determined by TEM (see Fig. 8(a) and (c)) gives the diameter of the

Dmag =

(18kB Tχ0 )1/3 (μ0 πMS2)1/3

(3)

where Dmag is the average diameter to be calculated, T is room temperature, and χo is the initial susceptibility. As the two samples were 6

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

(a)

Fig. 6. (a) Several curves of induced magnetic field of sample: blue and red curves were obtained under positive and negative magnetic fields, respectively. (b) Comparison between theoretical results (solid black curve) and values obtained from the induced field of sample (blue circles). (c) Comparison between magnetization results obtained for the same nanoparticle sample using our technique (blue circles) and a commercial magnetometer (MPMS SQUID, Quantum Design Inc.) (solid red curve). (d) Magnetization curves obtained using our method for samples produced by PLA (MNPs-I) and co-precipitation (MNPs-II) and (e) Low-temperature measurements of MNPs. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(b)

(c)

(d) MNPs – I MNPs – II

(e)

(a)

(b)

Experimental map (H a)

Model

Maximum intensity

Maximum intensity

(c) H a - Model

(d)

7

Fig. 7. (a) Experimental map under 20 mT. (b) Map obtained using theoretical current cylinder model. (c) Subtraction of experimental map from theoretical model. (d) Comparison between magnetization results obtained for the same nanoparticle sample using our technique (blue circles) and a commercial magnetometer (MPMS SQUID, Quantum Design Inc.) (solid red curve). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

25

(a)

(b)

MNPs - I LogNormal - Size (0 25)

Count

20 15 10 5 0

0

5

10

15

20

25

Size (nm) 20

(c)

(d)

MNPs - II LogNorrmal - Size (0 20)

Count

15

10

5 100 nm

0

0

5

10

Size (nm)

15

20

Fig. 8. (a) TEM image of MNPs produced by co-precipitation. (b) Histogram obtained from TEM images of MNPs produced by co-precipitation. (c) TEM image of MNPs produced by PLA. (d) Histogram obtained from TEM images of MNPs produced by PLA.

100

primarily composed of magnetite (see Fig. 3), we used a value of ρ = 5.20 × 103 kg/m3 [34]. We used the curves in Fig. 6(d) for fields near zero to estimate the value of χo. Hence, Ms was estimated by extrapolating the magnetization curve as a function of the inverse field (1/H) to 1/H = 0 [15,19]. In this way, we obtained Ms = 88 and 91 Am2/kg for MNPs produced by co-precipitation and PLA in water, respectively. Therefore, the diameters for the MNPs obtained from the experimental magnetic data are 9 and 4 nm, respectively. These values are consistent with those calculated from the TEM images (11.6 nm and 4.4 nm, respectively; see Fig. 8(b) and (d)). A sample of approximately 200 MNPs for each image was used to build the statistical size distributions, and the corresponding mathematical fit was performed using a log-normal distribution following the same procedure as reported in [35]. The results indicate that mean sizes of the MNPs synthesized by the two different techniques are quite different. The mean sizes and standard deviation of the radius distributions are 〈r〉cp = 11.6 nm and 〈δ〉cp = 3.4 nm, respectively, for the chemically synthesized MNPs and 〈r〉PLA = 4.4 nm and 〈δ〉PLA = 1.9 nm, respectively, for the nanomaterial produced by PLA in water. The same trend is observed for the DLS measurements, as shown in Fig. 9. The mean diameters of the MNPs synthesized chemically and by PLA are approximately 106 nm and 86 nm, respectively. As expected, the MNPs sizes obtained by DLS are greater than those obtained by TEM, because DLS technique measures the hydrodynamic radius of the MNPs in a suspension, which is always larger than the metal nucleus size [22]. Table 1 presents the MNPs diameter and standard deviation (σ) determined using different techniques. We attribute the differences to the intrinsic characteristics of each analysis method. Thus, we have demonstrated the estimation of the average nanoparticle size, which is a quantity of practical interest. In addition, this estimation method is less expensive than TEM.

MNPs - I MNPs - II

Count

80 60 40 20 0 100

200

300 400

Hydrodynamic Diameter (nm) Fig. 9. Mean radius distributions of magnetic MNPs-I (red line) and MNPs-II in water (black line) obtained by the DLS technique. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1 Comparison of particle diameters obtained by various techniques. Technique

Co-precipitation Laser ablation

8

TEM

Magnetic

DLS

Diameter

σ

Diameter

σ

Diameter

σ

11.6 nm 4.4 nm

3.4 nm 1.9 nm

9 nm 4 nm

2.5 nm 1.8 nm

106 nm 86 nm

31.4 nm 16.3 nm

Journal of Magnetism and Magnetic Materials 499 (2020) 166300

J.F.D.F. Araujo, et al.

6. Conclusion

[12] N.L. Martínez-Rodriguez, S. Tavárez, Z.I. González-Sánchez, In vitrotoxicity assessment of zinc and nickel ferrite nanoparticles in humanerythrocytes and peripheral blood mononuclear cell, Toxicol. In Vitro 57 (2019) 54–61, https://doi.org/ 10.1016/j.tiv.2019.02.011. [13] J.F.D.F. Araujo, J.M.B. Pereira, A.C. Bruno, Assembling a magnetometer for measuring the magnetic properties of iron oxide microparticles in the classroom laboratory, Am. J. Phys. 87 (2019) 471–475, https://doi.org/10.1119/1.5100944. [14] J.F.D.F. Araujo, M.C. Costa, S.R.W. Louro, A.C. Bruno, A portable Hall magnetometer probe for characterization of magnetic iron oxide nanoparticles, J. Magn. Magn. Mater. 426 (2017) 159–162, https://doi.org/10.1016/j.jmmm.2016.11.083. [15] J.F.D.F. Araujo, A.C. Bruno, S.R.W. Louro, Versatile magnetometer assembly for characterizing magnetic properties of nanoparticles, Rev. Sci. Instrum. 85 (2015) 105103–105107, https://doi.org/10.1063/1.4931989. [16] J.M.B. Pereira, C.J. Pacheco, M.P. Arenas, J.F.D.F. Araujo, G.R. Pereira, A.C. Bruno, Novel scanning dc-susceptometer for characterization of heat-resistant steels with different states of aging, J. Magn. Magn. Mater. 442 (2017) 311–318, https://doi. org/10.1016/j.jmmm.2017.07.004. [17] O. Kazakova, R. Puttock, C. Barton, H. Corte-León, M. Jaafar, V. Neu, A. Asenjo, Frontiers of magnetic force microscopy, J. Appl. Phys. 125 (2019) 060901–60929, https://doi.org/10.1063/1.5050712. [18] A. Oral, S.J. Bending, Scanning Hall probe microscopy of superconductors and magnetic materials, J. Vac. Sci. Technol. B 14 (2) (1996) 1202–1204, https://doi. org/10.1116/1.588514. [19] J.F.D.F. Araujo, D.R.P. Vieira, F. Osorio, W.E. Pöttker, F.A. Porta, P. Presa, G. Perez, A.C. Bruno, Versatile Hall magnetometer with variable sensitivity assembly for characterization of the magnetic properties of nanoparticles, 165431-165431, J. Magn. Magn. Mater. 489 (2019), https://doi.org/10.1016/j.jmmm.2019.165431. [20] J.F.D.F. Araujo, A.C. Bruno, H.R. Carvalho, Characterization of magnetic nanoparticles by a modular Hall magnetometer, J. Magn. Magn. Mater. 322 (2010) 2806–2809, https://doi.org/10.1016/j.jmmm.2010.04.034. [21] J.F.D.F. Araujo, J.M.B. Pereira, A practical and automated Hall magnetometer for characterization of magnetic materials, Mod. Inst. 4 (2010) 43–53, https://doi.org/ 10.4236/mi.2015.44005. [22] T. Del Rosso, S.R.W. Louro, F.L. Deepak, E.C. Romani, Q. Zaman, Pandoli, M. Cremona, F.L. Freire Junior, P. De Buele, De St, T. Pierre, R.Q. Aucelio, G. Mariotto, S. Gemini-Piperni, A.R. Ribeiro, S.M. Landi, A. Magalhães, Biocompatible Au@Carbynoid/Pluronic-F127 nanocomposites synthesized by pulsed laser ablation assisted CO 2 recycling, Appl. Surf. Sci. 41 (2018) 347–355, https://doi.org/10.1016/j.apsusc.2018.02.007. [23] R.J. Thibeau, C.W. Brown, R.H. Heidersbach, Raman spectra of possible corrosion products of iron, Appl. Spectrosc. 32 (1978) 532–535, https://doi.org/10.1366/ 000370278774330739. [24] D.L.A. Faria, S.V. Silva, M.T. Oliveira, Raman microspectroscopy of some iron oxides and oxyhydroxides, J. Raman Spectrosc 28 (1997) 873–878, https://doi.org/ 10.1002/(SICI)1097-4555(199711)28:11<873::AID-JRS177>3.0.CO;2-B. [25] S.J. Oh, D.C. Cook, H.E. Townsend, Characterization of iron-oxides commonly formed as corrosion products on steel, Hyperfine Interact. 112 (1998) 59–65, https://doi.org/10.1023/A:10110763085. [26] M.A. Legodi, D. de Waal, The preparation of magnetite, goethite, hematite and maghemite of pigment quality from mill scale iron waste, Dyes Pigments 74 (2007) 161–168, https://doi.org/10.1016/j.dyepig.2006.01.038. [27] M. Hanesch, Raman spectroscopy of iron oxides and (oxy) hydroxides at low power laser and possible applications in environmental magnetic studies, Geophys. J. Int. 177 (2009) 941–948, https://doi.org/10.1111/j.1365-246X.2009.04122.x. [28] O.N. Shebanova, P. Lazor, Raman spectroscopic study of magnetite (FeFe2O4): a new assignment for the vibrational spectrum, J. Solid State Chem. 174 (2003) 424–430, https://doi.org/10.1016/S0022-4596(03)00294-9. [29] T. Ohtsuka, K. Kubo, N. Sato, Raman spectroscopy of thin corrosion films on iron at 100 to 150 C in air, Corrosion 42 (1986) 476–481, https://doi.org/10.5006/1. 3583054. [30] M. Knobel, W.C. Nunes, L.M. Socolovsky, E. De Biasi, J.M. Vargas, J.C. Denardin, J. Nanosci. Nanotechnol. 8 (2008) 2836–2857, https://doi.org/10.1166/jnn.2008. 15348. [31] Q. Li, C.W. Kartikowati, S. Horie, T. Ogi, T. Iwaki, K. Okuyama, Correlation between particle size/ domain structure and magnetic properties of highly crystalline Fe3O4 nanoparticles, Sci. Rep. 7 (2017) 9894–9895, https://doi.org/10.1038/ s41598-017-09897-5. [32] Pacakova, B.; Kubickova, S.; Salas, G.; Mantlikova, A. R.; Marciello, M.; Morales, M. P.; Niznansky D.; Vejpravova, J. The internal structure of magnetic nanoparticles determines the magnetic response, 2017, 9, 5129-12. doi: 10.1039/C6NR07262C. [33] B.D. Cullity, C.D. Graham, Introduction to Magnetic Materials, IEEE Press, NJ, 2009, p. 391. [34] D.J. Dunlop, Ö. Özdemir, Rock Magnetism, Cambridge University Press, UK, 1997, p. 51. [35] Q. Zaman, J. Souza, O. Pandoli, K. Costa, V. Dmitriev, D. Fulvio, M. Cremona, R. Aucelio, G. Fontes, T. Del Rosso, Two-color surface plasmon resonance nanosizer for gold nanoparticles, Opt. Express 27 (2019) 3200–3216, https://doi.org/10. 1364/OE.27.003200.

Magnetic measurements were made for MNPs using an SMM, and a technique was developed for obtaining magnetization curves. The results were consistent and statistically comparable with those magnetometers obtained using other instruments, such as a commercial magnetometer (SQUID). The errors obtained for the saturation and remanent magnetization were approximately ± 0.18 Am2/kg and ± 0.6 Am2/kg, respectively. Our results demonstrate that the proposed method is an excellent alternative to conventional magnetic characterization techniques for microgram samples of MNPs. These analyses also demonstrate that the characterization based on magnetic curves obtained by this new technique can be used to reliably estimate the average particle size. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This work was supported in part by the Brazilian agencies CNPq, CAPES, FAPERJ and FAPESP. We thank Fredy Osorio and Amanda Farias dos Santos for providing Fig. 1(a) and Fig. 4(d). References [1] J. Dou, Q. Zhang, M. Ma, Gu, Fast fabrication of epoxy-functionalized magnetic polymer core-shell microspheres using glycidyl methacrylate as monomer via photo-initiated miniemulsion polymerization, J. J. Magn. Magn. Mater. 324 (2012) 3078–3082, https://doi.org/10.1016/j.jmmm.2012.05.005. [2] M.S. Boon, W.P.S. Saw, M. Mariatti, Magnetic, dielectric and thermal stability of Ni–Zn ferrite-epoxy composite thin films for electronic applications, J. Magn. Magn. Mater. 324 (2012) 755–776, https://doi.org/10.1016/j.jmmm.2011.09.009. [3] R. Wang, S. Nie, J. Zhao, Y. Ji, Measuring magnetic anisotropy with a rotatable ac electromagnet, Measument 79 (2016) 15–19, https://doi.org/10.1016/j. measurement.2015.10.032. [4] Y. Yeh, J. Jin, C. Li, J.T. Lue, The electric and magnetic properties of Co and Fe films percept from the coexistence of ferromagnetic and microstrip resonance for a T-type microstrip, Measument 42 (2009) 290–297, https://doi.org/10.1016/j. measurement.2008.06.012. [5] S.H. Noh, W. Na, J.T. Jang, J.H. Lee, E.J. Lee, S.H. Moon, Y. Lim, J.S. Shin, J. Cheon, Nanoscale magnetism control via surface and exchange anisotropy for optimized ferrimagnetic hysteresis, Nano Lett. 12 (2012) 3716–3721, https://doi. org/10.1021/nl301499u. [6] M. Wu, S. Huang, Magnetic nanoparticles in cancer diagnosis, drug delivery and treatment, Mol. Clin. Oncol. 7 (2017) 738–746, https://doi.org/10.3892/mco. 2017.1399. [7] E.Y. Yu, M. Bishop, B. Zheng, R.M. Ferguson, A.P. Khandhar, S.J. Kemp, K.M. Krishnan, P.W. Goodwill, S.M. Conolly, Magnetic Particle Imaging: A Novel in Vivo Imaging Platform for Cancer Detection, Nano Lett. 17 (2017) 1648–1654, https://doi.org/10.1021/acs.nanolett.6b04865. [8] S. Arsalani, E.J. Guidelli, M.A. Silveira, C.E.G. Salmon, J.F.D.F. Araujo, A.C. Bruno, O. Baffa, Magnetic Fe3O4 nanoparticles coated by natural rubber latex as MRI contrast agent, J. Magn. Magn. Mater. 475 (2019) 458–464, https://doi.org/10. 1016/j.jmmm.2018.11.132. [9] S. Arsalani, E.J. Guidelli, M.A. Silveira, J.F.D.F. Araujo, A.C. Bruno, O. Baffa, Green synthesis and surface modification of iron oxide nanoparticles with enhanced magnetization using natural rubber latex, ACS Sustainable Chem. Eng. 11 (2018) 13756–13765, https://doi.org/10.1021/acssuschemeng.8b01689. [10] W.E. Pottker, R. Ono, M.A. Cobos, A. Hernando, J.F.D.F. Araujo, A.C. Bruno, S.A. Lourenço, E. Longo, F.A. La Porta, Influence of order-disorder effects on the magnetic and optical properties of NiFe2O4 nanoparticles, Ceram. Int. 44 (2018) 17290–17297, https://doi.org/10.1016/j.ceramint.2018.06.190. [11] L. Courtney-Davies, Z. Zhu, C.L. Ciobanu, B.P. Wade, N.J. Cook, K. Ehrig, A.R. Cabral, A. Kennedy, Matrix-matched iron-oxide laser ablation ICP-MS U-Pb geochronology using mixed solution standards, Minerals 6 (2016) 1–17, https:// doi.org/10.3390/min6030085.

9