Size dependent magnetic hyperthermia of octahedral Fe₃O₄ nanoparticles

Abstract

Magnetic nanoparticle hyperthermia is promising as a cancer therapeutic treatment. Shape and size are two crucial factors for the magnetic hyperthermia performance of nanoparticles. In this work, octahedral Fe₃O₄ nanoparticles with different sizes are successfully synthesized and their magnetic hyperthermia performances are investigated systematically in a gel suspension. The results suggest a wide size range (43–98 nm) for high SAR values (up to 2629 W g⁻¹). The SAR values are verified by hysteresis loss measured in the gel suspension. This study demonstrates that octahedral Fe₃O₄ nanoparticles can serve as an excellent thermal seed for high performance magnetic hyperthermia cancer treatment.

---

Introduction

Magnetic nanoparticle hyperthermia (MNH) is appealing nowadays as a cancer therapy compared with conventional therapeutic methods such as chemotherapy and radiotherapy.^1–4 In MNH, magnetic nanoparticles (MNPs), which are exposed to an external alternating magnetic field, convert electromagnetic energy into thermal energy and induce temperature rises.⁵ When the temperature is above 42 °C, the tumour cell will die because of cell apoptosis or cell necrosis directly.^3,6–8 However, the design of high specific absorption rate (SAR) heating seeds remains a challenge.^3,9–14 Higher SAR is desired because it allows less usage of MNPs and less possible toxicity.¹⁵

It has been reported that SAR is related to MNP hysteresis loss.¹⁶ Higher hysteresis loss results in higher SAR values. In general, hysteresis loss is determined by saturation magnetization (M_s) and coercivity (H_c) of nanoparticle.^6,17–20 Therefore, much attention has been paid to metal nanoparticle (e.g. Fe, FeCo) with high M_s or doping with high anisotropy element (e.g. Co) for high H_c.^6,11,17,21 However, metal nanoparticles are chemically unstable and Co will induce cytotoxicity to cells.²² In comparison, iron oxide nanoparticles (IONPs) are considered as one of the best candidate for MNH because of high M_s, good stability and excellent biocompatibility.²³ Moreover, H_c can be optimized by changing the geometric shape and particle size.^6,24–26 Therefore, IONPs with different shapes (e.g. nanodisk, nanorings and nanocubes) and sizes have been intensively investigated.^27–30 However, to our best knowledge, the magnetic hyperthermia of octahedral IONPs has been rarely studied.³¹ Recently, our group synthesized different octahedral Fe₃O₄ MNPs with different sizes, which inspire us to investigate its hyperthermia performance.³²

In this work, we study the MNH property of octahedral Fe₃O₄ MNPs systematically. Both experimental and theoretical simulation results demonstrate that octahedral Fe₃O₄ MNPs have high SAR values with wide range of particle sizes. Moreover, the SAR values are reproduced by hysteresis loss measured in gel suspension.

Experimental

Materials

Iron(III) acetylacetonate (Fe(acac)₃; 97%), oleic acid (OA; 90%), benzyl ether (99%), cetyltrimethylammonium bromide (CTAB; ≥99%) and chloroform (99.9%) were purchased from Sigma-Aldrich. Hexane (95%) and absolute ethanol were used as received.

Characterization

All transmission electron microscopy (TEM) images were obtained using JEOL 2010 instrument. X-ray diffractometry (XRD) results were acquired using Bruker D8 Advanced Diffractometer System with Cu Kα (1.5418 Å). Magnetic properties of the samples were measured by a LakeShore Model 7407 vibrating sample magnetometer (VSM). Inductively coupled plasma (ICP) results were obtained using iCAP 6000 ICP-OES system. Dynamic light scattering (DLS) measurements were carried out on Malvern Zetasizer Nano-ZS.

Synthesis of octahedral Fe₃O₄ MNPs

The synthesis method is based on a revised thermal decomposition procedure.³² For a typical synthesis of 43 nm octahedral Fe₃O₄ MNPs, 5.5 mmol Fe(acac)₃ and 20 mmol OA were mixed with 25 mL benzyl ether. The mixture was heated to 110 °C for 1 hour under N₂ flow with stirring, and then slowly heated to 165 °C and maintained for 40 minutes. Finally the mixture was quickly heated up to 280 °C with reflux for another 30 minutes. After cooling down, ethanol was added to the mixture and brown precipitate was observed. The precipitate was separated with centrifuge and dissolved in hexane. Washing process was repeated for three times and final powder product was dried for further use.

Magnetic hyperthermia measurement

For magnetic hyperthermia study, as-synthesized octahedral Fe₃O₄ MNPs powder products have to be transferred to aqueous phase with surface modification using CTAB.³³ Typically, 10 mg octahedral Fe₃O₄ MNPs powder dissolved in 1 mL chloroform were mixed with 10 mL 2 mM aqueous CTAB solution. The resulting micro-emulsion is treated with vigorously vortex for 40 minutes, and afterwards, heated at 80 °C for 30 minutes to evaporate chloroform. The resulting solution were washed with ultrapure water and centrifuged at 6000 rpm for 10 minutes twice to remove supernatant. Finally, the resulting precipitate was dispersed in ultrapure water for further use.

After phase transfer, octahedral Fe₃O₄ MNP solution was diluted using ultrapure water with final concentration of 100 μg mL⁻¹ Fe, which determined by ICP results. Agarose gel (20 mg) was mixed with 1 mL Fe₃O₄ solution (100 μg mL⁻¹ Fe) and the mixture was then heated up to 85 °C for 30 minutes. The gel sample was then cooled down for measurement. Afterwards, the gel sample was placed into plastic holder for hyperthermia measurement. The plastic holder was placed into the centre of a water-cooled copper coil driven by an Inductelec A.C. generator (Shenzhen Magtech Company Limited, China, SPG-10AB-II). The sample's temperature was measured by a LuxtronMD600 fibre optical thermometry unit that connected to a computer. After measurement, the specific absorption rates (SAR) of the samples were evaluated using^34,35

where, C (4.18 J g⁻¹ °C⁻¹) is the specific heat of agarose gel (5%). ΔT/Δt donates the initial slope of the time-dependent temperature curve and m_Fe is Fe concentration (100 μg mL⁻¹).

Micromagnetic simulation for SAR

The micromagnetic simulations were carried out using the LLG Micromagnetics Simulator TM package.³⁶ The magnetic parameters of Fe₃O₄ in the simulation are used as follows: saturation magnetization M_s = 500 kA m⁻¹, exchange stiffness constant A = 1.2 × 10⁻¹¹ J m⁻¹, magnetocrystalline anisotropy constant K₁ = −1.35 × 10⁴ J m⁻³, K₂ = −0.44 × 10⁴ J m⁻³.³⁷ The cell size is 2 nm, which is smaller than the exchange length (∼8.6 nm).³⁸ In the simulation, the Gilbert damping coefficient α was used as 0.5. Moreover, we let the nanostructures relax in absence of applied magnetic field from the following either uniform magnetization or vortex state to get the ground states of magnetic nanostructures. After comparing the energy of different resultant states, the ground-state domain structure can be determined. The hysteresis loop simulation was done by applying time-varying magnetic field along certain directions with respective to the Fe₃O₄ octahedron and the magnetizations along these directions are recorded in the meantime. Moreover, the loops were simulated at room temperature (300 K). For the SAR calculation, an average hysteresis loop is calculated to imitate the random orientation of octahedrons during hyperthermia measurement. Then the corresponding SAR is derived as described elsewhere.²⁸

In vitro cell viability assay

MCF-7 mammalian breast cancer cells (ATCC) were cultured in DMEM (Dulbecco's Modified Eagle Medium), 10% FBS, 100 units per mL penicillin, and 100 μg mL⁻¹ streptomycin in 37 °C with 5% CO₂. Thereafter, the cells were seeded in 96-well plates at the density of 5 × 10⁴ viable cells per well and incubated for 24 hours to allow cell attachment. The cells were incubated for another 24 hours with media containing octahedral Fe₃O₄ MNPs at concentration ranging from 6.25 to 100 μg mL⁻¹ Fe of well. Control was defined as wells without octahedral Fe₃O₄ MNPs. After incubation, 10 μL CCK8 was added into each plate well and cells were incubated for another 4 hours. A FluoStar Optima microplate reader was used to measure the absorbance of each well at 450 nm. The cell viability was calculated by:

Results and discussion

Fig. 1 provides the TEM images of the as-synthesized NPs. Octahedral NPs on a copper grid can have different projection shapes when getting TEM images. Therefore, parallelogram, hexagon, square and rectangle shapes can be seen in the TEM images (see ESI Fig. S1†). Throughout this study, the length between two opposite vertices is defined as the particle sizes.³² As shown in Fig. 1a–e, uniform octahedral-shaped NPs are successfully obtained at large scale and the sizes (13 nm, 22 nm, 43 nm, 98 nm, 260 nm) are well tuned by varying the precursor amount and ratio. Fig. 1f provides the high resolution TEM (HRTEM) image of 43 nm nanooctahedron. The 4.8 Å lattice spacing is indexed to {1 1 1} plane for cubic spinel structured Fe₃O₄, suggesting that the eight outer faces are enclosed by {1 1 1} planes. Moreover, the corresponding Fourier transform electron diffraction image (shown as insert) confirms the single crystalline nature of our octahedral NPs.

Fig. 1 TEM images for (a) 13 nm (b) 22 nm, (c) 43 nm, (d) 98 nm, (e) 260 nm octahedral Fe₃O₄ MNPs with size distribution inserted at right up corner. (f) High resolution TEM image for 43 nm MNPs with FFT image inserted at right up corner.

In order to further confirm the crystalline structure of these octahedral NPs, X-ray diffraction (XRD) is performed for all the samples. As shown in Fig. 2a, the XRD patterns clearly demonstrate pure inverse spinel crystalline structure. All the peaks can be well indexed to Fe₃O₄ (JCPDS cards no. 19-0629). No FeO and Fe₂O₃ peaks are found in XRD results, indicating that all these octahedral NPs are of pure Fe₃O₄.

Fig. 2 (a) XRD spectra for different sizes of octahedral Fe₃O₄ MNPs. (b) Magnetization as function of applied field for octahedral Fe₃O₄ MNPs with different sizes (insert: magnetization filed-curves at low magnetic field).

The magneto-static properties of these pure octahedral Fe₃O₄ NPs are investigated using a vibrating sample magnetometer (VSM). Fig. 2b presents the room temperature hysteresis loops. The M_s is 68, 73, 85, 90 and 92 emu g⁻¹ for 13 nm, 22 nm, 43 nm, 98 nm and 260 nm octahedral Fe₃O₄ MNPs, respectively (Table 1). Apparently, M_s decreases with decreasing particles sizes. Such decrease is attributed to the surface spin canting and large surface to volume ratio of small NPs.³⁹ Moreover, the magnetization curves at low applied field are shown as insert. The coercivity (H_c) and remanence (M_r) values are summarized as in Table 1. It can be seen that the 13 nm NPs exhibit almost zero coercivity and remanence due to superparamagnetism (SPM) effect.⁴⁰ For larger octahedral Fe₃O₄ MNPs, the coercivity becomes non-zero and increases with increasing particle size.

Table 1 Saturation magnetization (M_s), coercivity (H_c) and remanence (M_r) values for octahedral Fe₃O₄ MNPs with difference sizes

Size (nm)	Ms (emu g^−1)	Hc (Oe) powder	Mr (emu g^−1) powder
13	68	0	0
22	73	22	7
43	85	37	14
98	90	80	23
260	92	118	13

---

To study their hyperthermia properties, these octahedral Fe₃O₄ MNPs were transferred to aqueous phase by coating with cetyltrimethylammonium bromide (CTAB). When these nanoparticles were freshly prepared, they were capped with oleic acid (OA), which is hydrophobic. After phase transfer, the MNPs were encapsulated with CTAB.⁴¹ CTAB alkane chain would stretch parallel with the oleic acid olefin chain and their ammonia groups would face outward thus making the nanoparticle hydrophilic (Scheme 1). Moreover, we compared CTAB with two other polymers, i.e., Pluronic F-127 and polyvinylpyrrolidone (PVP), which are two typical surfactants for biomedical applications.^42,43 The dynamic light scattering (DLS) results of the 13 nm Fe₃O₄ MNPs coated with different polymers suggest that the CTAB coated Fe₃O₄ MNPs have smaller hydrodynamic size than PVP and F-127 coated MNPs, indicating that CTAB can prevent the MNPs from aggregation more effectively (Fig. 3). Therefore, CTAB is used as surfactant throughout our study. Thereafter, magnetic hyperthermia measurement was carried out in aqueous suspension of the CTAB coated MNPs with Fe concentration of 100 μg mL⁻¹ determined by inductively coupled plasma (ICP) analysis. Additionally, the test was also performed in gel suspension with high viscosity to acquire an overall understanding of hyperthermia performance of these MNPs. It should be emphasized that the maximum product of magnetic field amplitude (H) and frequency (f) in this study is well below the upper limitation to avoid non-selective heating of both cancerous and healthy tissue due to eddy currents.⁴⁴

Scheme 1 Illustration of Fe₃O₄ MNPs surface modification with CTAB.

Fig. 3 Hydrodynamic size for 13 nm Fe₃O₄ MNPs capped with CTAB, PVP and Pluronic F127 as polymers.

Fig. 4 shows the SAR values measured under 358 kHz field with different field strength H (200 Oe, 400 Oe, 600 Oe, 800 Oe). As seen in Fig. 4, the SAR values depend on both H and particle size. Taking the 43 nm octahedral Fe₃O₄ MNPs as example, SAR values increase from 157 W g⁻¹ (200 Oe) to about 2483 W g⁻¹ (800 Oe). This is because SAR values are related to hysteresis loss. Higher magnetic field would result in larger hysteresis loss until it is saturated.¹⁶ More importantly, the SAR value also increases when the particle grows bigger and it reaches a plateau at 43 nm (2483 W g⁻¹) and 98 nm (2629 W g⁻¹). Further increase in particle size to 260 nm leads to the drop of SAR value. It should be noted that such a wide size range (43 nm–98 nm) for high SAR values is beneficial for practical magnetic hyperthermia application. To have a better insight on the shape effect, we have also measured the SAR values for spherical Fe₃O₄ MNPs with almost same volume as 22 nm octahedral Fe₃O₄ MNPs (see ESI Fig. S2†). However, the SAR values for of the spherical MNPs are much lower than octahedral MNPs. It proves that the octahedral Fe₃O₄ MNPs are excellent thermal seeds for magnetic hyperthermia application over the spherical nanoparticles.

Fig. 4 SAR values for octahedral Fe₃O₄ MNPs measured in gel suspension under different magnetic field amplitudes at 358 kHz.

It should be emphasized that for MNPs, the major mechanism for heat dissipation in magnetic hyperthermia is hysteresis loss.¹⁶ Therefore, in order to verify the experimental SAR values in gel, we attempt to reproduce the SAR values using hysteresis loss. For this purpose, the octahedral Fe₃O₄ MNPs were dispersed into agarose gel (5%) with the same Fe concentration as that used in above SAR measurement. In such a way, the MNPs could be well isolated to minimize their magnetic dipolar–dipolar interaction. The full hysteresis loop of the octahedral Fe₃O₄ MNPs measured in gel suspension was shown in ESI (Fig. S3†). The magnetizations of the gel hysteresis loops have been scaled up to the M_s of pure particles as obtained in Fig. 2b. In addition, the detailed magnetizations at low magnetic field are shown in Fig. 5a and the H_c values are presented in Fig. 5b. The H_c values increase with particle sizes and reach a maximum at 98 nm (341 Oe). However, it drops to 132 Oe at 260 nm. This is obviously different from the H_c as revealed by the loops of pure powder in Fig. 2b which shows a monotonic increase. The difference might be attributed to the minimization of magnetic dipolar–dipolar interaction when the nanoparticles are dispersed in gel suspension.

Fig. 5 (a) Magnetization at low magnetic field for octahedral Fe₃O₄ MNPs measured in gel. (b) Coercivity for different sizes of octahedral Fe₃O₄ MNPs. Minor loops at different magnetic fields for (c) 22 nm, (d) 43 nm, (e) 98 nm, (f) 260 nm octahedral Fe₃O₄ MNPs.

Fig. 5c–f provide the minor loops of octahedral Fe₃O₄ MNPs measured in gel suspension under different magnetic fields. Using these minor hysteresis loops, the hysteresis loss and corresponding SAR values can be calculated by²⁸

with

P = μ₀f∮M dH (4)

where ρ is the density of Fe₃O₄ (5.17 × 10³ kg m⁻³), φ is the volume fraction of Fe₃O₄ (equals to 1 for one pure Fe₃O₄ particle), P is the heat dissipation due to hysteresis loss, μ₀ is vacuum permeability, f is the frequency (358 kHz) of the AC field, M is the magnetization.

The calculated SAR values are presented in Fig. 6. For comparison, the SAR values measured in gel suspension are also presented. It is apparent that SAR values calculated from minor loops are quite close to the experimental measured gel suspension values except for the 22 nm MNPs. For 22 nm MNPs, the experimental gel suspension measured SAR values are a lot higher than SAR values calculated by minor loop areas. The discrepancy should be ascribed for the superparamagnetic behaviour of the 22 nm NPs.⁴⁵ For the superparamagnetic nanoparticles, the dynamic hysteresis loops corresponding to the frequency (358 kHz) of hyperthermia measurement should be used to calculate SAR values. The dynamic hysteresis loops are quite different from magnetostatic hysteresis loops at superparamagnetic regime and the hysteresis loss of dynamic hysteresis loop is usually higher than magnetostatic hysteresis loop.¹⁶ Therefore, for 22 nm nanoparticles, the SAR values calculated from magnetostatic hysteresis loop are lower than that of SAR values measured experimentally in gel suspension at all magnetic fields. For the remaining octahedral Fe₃O₄ MNPs sizes ranging from 43 nm to 260 nm, as they fall into ferrimagnetic regime, the SAR values could be calculated by magnetostatic hysteresis loops.²⁸ As a result, SAR values calculated from magnetostatic minor loops are very close to experimental SAR values measured in gel suspension for these octahedral Fe₃O₄ MNPs. The above results prove that SAR values of non-superparamagnetic Fe₃O₄ MNPs in gel suspension could be reproduced from magnetostatic hysteresis loop measured in gel suspension.

Fig. 6 SAR values calculated by minor loop areas and measured by gel suspension experimentally for (a) 22 nm, (b) 43 nm, (c) 98 nm, (d) 260 nm Fe₃O₄ MNPs.

To prove the octahedral Fe₃O₄ MNPs possess high SAR values with wide ranges of nanoparticle sizes, we also did simulation for particle size in between 43 nm and 98 nm MNPs. Fig. 7 shows the simulation results for 60 nm octahedral Fe₃O₄ MNPs. Fig. 7a is the simulated domain structure of 60 nm octahedral Fe₃O₄ MNPs. It reveals a quasi-uniform domain structure pointing along [1 1 1] direction. Based on the domain structure, hysteresis loops are simulated along different directions, as provided in ESI (see Fig. S4†). In fact, all these directions have possibility to occur in hyperthermia measurement. To include the contribution from all these directions, an average loop was calculated as described elsewhere (Fig. 7b).²⁸ With this average loop, the SAR value is calculated to be about 2501 W g⁻¹, which agrees very well with the experimental SAR value found in Fig. 4b. It again confirms the validity of our hyperthermia measurement. Therefore, the octahedral Fe₃O₄ MNPs are shown to have high SAR values with wide nanoparticle size ranges proved by simulation and experiment.

Fig. 7 (a) Simulated ground state domain structure (b) hysteresis loop and corresponding SAR value for 60 nm octahedral Fe₃O₄ MNPs.

Furthermore, the biocompatibility is tested by using standard CCK8 assay on MCF-7 breast cancer cells. The CTAB encapsulated 43 nm Fe₃O₄ MNPs are chosen as it have high SAR value with good suspension (see ESI Fig. S5†). The MNPs are incubated with MCF-7 cells for 24 hours (Fig. 8). As seen from the results, 43 nm Fe₃O₄ MNPs do not induce notable toxicity on MCF-7 cancer cells with final concentration ranging from 6.25 μg mL⁻¹ to 100 μg mL⁻¹ Fe. Therefore, our octahedral Fe₃O₄ MNPs are biocompatible and can be used for further bio application.

Fig. 8 Cell viability for 43 nm CTAB encapsulated octahedral Fe₃O₄ MNPs.

Conclusion

In summary, we investigated hyperthermia performance of octahedral Fe₃O₄ MNPs with different sizes. The results suggest that octahedral Fe₃O₄ MNPs could exhibit high SAR values in wide range of particle sizes. Therefore, the octahedral shaped Fe₃O₄ MNPs offers the possibility for efficient magnetic hyperthermia in biomedical applications. Furthermore, we demonstrate a general method to calculate SAR values from magnetostatic hysteresis loss measured in gel suspension, which is significant for the SAR prediction of magnetic nanoparticle for future research.

Acknowledgements

This work is financially supported by Grants of SERC 1321202068 (R284-000-105-305), NRF-CRP9-2011-01 (R284-000-111-592) and NRF-CRP10-2012-02.

Notes and references

1. C. Grüttner, K. Müller, J. Teller and F. Westphal, Int. J. Hyperthermia, 2013, 29, 777–789.
2. A. Ito, M. Shinkai, H. Honda and T. Kobayashi, J. Biosci. Bioeng., 2005, 100, 1–11.
3. F. Gazeau, M. Lévy and C. Wilhelm, Nanomedicine, 2008, 3, 831–844.
4. P. Moroz, S. K. Jones and B. N. Gray, Int. J. Hyperthermia, 2002, 18, 267–284.
5. R. E. Rosensweig, J. Magn. Magn. Mater., 2002, 252, 370–374.
6. B. Mehdaoui, A. Meffre, J. Carrey, S. Lachaize, L.-M. Lacroix, M. Gougeon, B. Chaudret and M. Respaud, Adv. Funct. Mater., 2011, 21, 4573–4581.
7. J. E. Robinson, M. J. Wizenberg and W. A. McCready, Nature, 1974, 251, 521–522.
8. L. S. Goldstein, M. W. Dewhirst, M. Repacholi and L. Kheifets, Int. J. Hyperthermia, 2003, 19, 373–384.
9. J.-P. Fortin, C. Wilhelm, J. Servais, C. Ménager, J.-C. Bacri and F. Gazeau, J. Am. Chem. Soc., 2007, 129, 2628–2635.
10. A. Figuerola, R. D. Corato, L. Manna and T. Pellegrino, Pharmacol. Res., 2010, 62, 126–143.
11. H. Sohn and R. H. Victora, J. Appl. Phys., 2010, 107, 09B312.
12. D. Ho, X. Sun and S. Sun, Acc. Chem. Res., 2011, 44, 875–882.
13. L. Lartigue, C. Innocenti, T. Kalaivani, A. Awwad, M. d. M. Sanchez Duque, Y. Guari, J. Larionova, C. Guérin, J.-L. G. Montero, V. Barragan-Montero, P. Arosio, A. Lascialfari, D. Gatteschi and C. Sangregorio, J. Am. Chem. Soc., 2011, 133, 10459–10472.
14. A. Jordan, P. Wust, H. Fählin, W. John, A. Hinz and R. Felix, Int. J. Hyperthermia, 1993, 9, 51–68.
15. H. Rudolf, D. Silvio, M. Robert and Z. Matthias, J. Phys.: Condens. Matter, 2006, 18, S2919.
16. J. Carrey, B. Mehdaoui and M. Respaud, J. Appl. Phys., 2011, 109, 083921.
17. J. H. Lee, J. T. Jang, J. S. Choi, S. H. Moon, S. H. Noh, J. W. Kim, J. G. Kim, I. S. Kim, K. I. Park and J. Cheon, Nat. Nanotechnol., 2011, 6, 418–422.
18. P. Guardia, R. D. Corato, L. Lartigue, C. Wilhelm, A. Espinosa, M. Garcia-Hernandez, F. Gazeau, L. Manna and T. Pellegrino, ACS Nano, 2012, 6, 3080–3091.
19. K. H. Bae, M. Park, M. J. Do, N. Lee, J. H. Ryu, G. W. Kim, C. Kim, T. G. Park and T. Hyeon, ACS Nano, 2012, 6, 5266–5273.
20. S. H. Noh, W. Na, J. T. Jang, J. H. Lee, E. J. Lee, S. H. Moon, Y. Lim, J. S. Shin and J. Cheon, Nano Lett., 2012, 12, 3716–3721.
21. X. Hong, M. Li, N. Bao, E. Peng, W. Li, J. Xue and J. Ding, J. Nanopart. Res., 2014, 16, 1–9.
22. J. G. Hengstler, U. Bolm-Audorff, A. Faldum, K. Janssen, M. Reifenrath, W. Götte, D. Jung, O. Mayer-Popken, J. Fuchs, S. Gebhard, H. G. Bienfait, K. Schlink, C. Dietrich, D. Faust, B. Epe and F. Oesch, Carcinogenesis, 2003, 24, 63–73.
23. T. K. Jain, M. K. Reddy, M. A. Morales, D. L. Leslie-Pelecky and V. Labhasetwar, Mol. Pharm., 2008, 5, 316–327.
24. C. Martinez-Boubeta, K. Simeonidis, D. Serantes, I. Conde-Leborán, I. Kazakis, G. Stefanou, L. Peña, R. Galceran, L. Balcells, C. Monty, D. Baldomir, M. Mitrakas and M. Angelakeris, Adv. Funct. Mater., 2012, 22, 3737–3744.
25. P. de la Presa, Y. Luengo, M. Multigner, R. Costo, M. P. Morales, G. Rivero and A. Hernando, J. Phys. Chem. C, 2012, 116, 25602–25610.
26. M. Ma, Y. Wu, J. Zhou, Y. Sun, Y. Zhang and N. Gu, J. Magn. Magn. Mater., 2004, 268, 33–39.
27. D.-H. Kim, E. A. Rozhkova, I. V. Ulasov, S. D. Bader, T. Rajh, M. S. Lesniak and V. Novosad, Nat. Mater., 2010, 9, 165–171.
28. Y. Yang, X. Liu, Y. Lv, T. S. Herng, X. Xu, W. Xia, T. Zhang, J. Fang, W. Xiao and J. Ding, Adv. Funct. Mater., 2015, 25, 812–820.
29. C. Martinez-Boubeta, K. Simeonidis, A. Makridis, M. Angelakeris, O. Iglesias, P. Guardia, A. Cabot, L. Yedra, S. Estrade, F. Peiro, Z. Saghi, P. A. Midgley, I. Conde-Leboran, D. Serantes and D. Baldomir, Sci. Rep., 2013, 3, 1652.
30. X. L. Liu, Y. Yang, C. T. Ng, L. Y. Zhao, Y. Zhang, B. H. Bay, H. M. Fan and J. Ding, Adv. Mater., 2015, 27, 1939–1944.
31. J. Mohapatra, A. Mitra, M. Aslam and D. Bahadur, IEEE Trans. Magn., 2015, 1.
32. L. Li, Y. Yang, J. Ding and J. Xue, Chem. Mater., 2010, 22, 3183–3191.
33. J. Kim, H. S. Kim, N. Lee, T. Kim, H. Kim, T. Yu, I. C. Song, W. K. Moon and T. Hyeon, Angew. Chem., Int. Ed., 2008, 47, 8438–8441.
34. I. Hilger, K. Frühauf, W. Andrä, R. Hiergeist, R. Hergt and W. A. Kaiser, Acad. Radiol., 2002, 9, 198–202.
35. X. L. Liu, E. S. G. Choo, A. S. Ahmed, L. Y. Zhao, Y. Yang, R. V. Ramanujan, J. M. Xue, D. D. Fan, H. M. Fan and J. Ding, J. Mater. Chem. B, 2014, 2, 120–128.
36. M. R. Scheinfein, LLG Micromagnetics Simulator User’s Guide, http://llgmicro.home.mindspring.com/, accessed, June 2015.
37. Y. Yang, X. L. Liu, J. B. Yi, Y. Yang, H. M. Fan and J. Ding, J. Appl. Phys., 2012, 111, 044303.
38. C. A. F. Vaz, C. Athanasiou, J. A. C. Bland and G. Rowlands, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 73, 054411.
39. M. P. Morales, S. Veintemillas-Verdaguer, M. I. Montero, C. J. Serna, A. Roig, L. Casas, B. Martínez and F. Sandiumenge, Chem. Mater., 1999, 11, 3058–3064.
40. M. B. Knickelbein, Phys. Rev. Lett., 2001, 86, 5255–5257.
41. J. Fang, P. Chandrasekharan, X. L. Liu, Y. Yang, Y. B. Lv, C. T. Yang and J. Ding, Biomaterials, 2014, 35, 1636–1642.
42. M. Gonzales and K. M. Krishnan, J. Magn. Magn. Mater., 2007, 311, 59–62.
43. H. L. Liu, S. P. Ko, J. H. Wu, M. H. Jung, J. H. Min, J. H. Lee, B. H. An and Y. K. Kim, J. Magn. Magn. Mater., 2007, 310, e815–e817.
44. R. Hergt and S. Dutz, J. Magn. Magn. Mater., 2007, 311, 187–192.
45. F. X. Redl, C. T. Black, G. C. Papaefthymiou, R. L. Sandstrom, M. Yin, H. Zeng, C. B. Murray and S. P. O'Brien, J. Am. Chem. Soc., 2004, 126, 14583–14599.

---

Footnotes

† Electronic supplementary information (ESI) available: Octahedral nanoparticle projection, gel sample full hysteresis loop, simulation of octahedral sample at all direction, DLS results of sample. See DOI: 10.1039/c5ra12558h

‡ These authors contribute equally.

---