Ultrasmall and monodisperse colloidal amorphous Nd–Fe–B–Na magnetic nanoparticles with high T_C

Abstract

Ultrasmall and monodisperse air-stable colloidal Nd–Fe–B–Na nanoparticles (NPs) have been synthesized by the solution phase colloidal method and characterized by TEM, EXAFS, Mössbauer spectroscopy, and SQUID. Their size can be controlled in the sub-5 nm regime. The critical temperature (T_C) of Nd–Fe–B–Na nanoparticles with 3 nm diameter is surprisingly high, higher than 650 K. EXAFS gives the chemical environmental information of the Fe center, and Mössbauer spectroscopy confirms that the 3 nm Nd–Fe–B–Na nanoparticles contain the Fe nanocluster. A possible magnetic coupling mechanism is proposed.

---

Colloidal magnetic nanoparticles (NPs) have attracted intensive research attention in the past few decades from synthesis to various applications, ranging from ultrahigh-density magnetic recording,¹ single-electron tunneling magnetic devices,² magnetic energy storage,³ biomedical applications,^4–6 and magnetic separable catalysis⁷ to magnetically responsive photonic crystals.⁸ As a potential candidate to realize ultrahigh-density recording at the single particle level, an essential prerequisite for practical one particle one bit ultrahigh-density magnetic recording is to synthesize ultrasmall colloidal magnetic nanoparticles with enough high transition temperature. However, the blocking temperatures (T_B) of colloidally synthesized freestanding monodisperse magnetic nanoparticles and surfactant-assisted ball milled Nd₂Fe₁₄B nanoparticles with about 5 nm diameter based on long-range ferromagnetic interactions are far below room temperature, such as 20–30 K for 4 nm chemically disordered fcc FePt,¹ 80 K for 10 nm ε-Co,^2b 160 K for 5.2 nm CoFe₂O₄,^4a 25 K for 4 nm γ-Fe₂O₃,^4b 20 K for 1.8 nm CoPt₃,^4c 40 K for 5 nm Fe₃O₄,^5a and 65 K for 2.7 nm Nd₂Fe₁₄B.^9a At the same time, amorphous phase RE–TM and RE–TM–B alloys based on short-range ferromagnetic interactions (RE = rare-earth metal, TM = transition metals) with a high magnetic anisotropy constant (K_u usually larger than 10⁷ erg cm⁻³) have drawn much attention for their commercial applications in a magnetooptic sensor, a high-density magnetooptic disk and high T_C.^9b,c,10–12 The short-range ordered structure and random magnetic anisotropy (RMA) of the above mentioned amorphous ferromagnetic alloys suggest that the undesirable effects of decreasing size on T_B do not behave like those of the crystalline phase. However, controlled synthesis of colloidal magnetic nanoparticles of amorphous phase metal alloys remains a challenge. It is necessary to develop a colloidal method to synthesize amorphous magnetic nanoparticles and investigate their magnetic properties at the nanometer scale.

In this communication, we report the controllable colloidal synthesis and magnetic properties of air-stable monodisperse amorphous Nd–Fe–B–Na nanoparticles, one of them with 3 nm diameter and T_C higher than 650 K.

To increase the ratio of magnetic centers, we try to eliminate sodium from the final product by selecting tetrabutylammonium borohydride (TBAB) and triethylamine borane complex. For the former, the reaction system produces a large amount of bubbles when the reducing agent is added in at 100 °C and no product is separable by centrifugation; for the latter, triethylamine borane is sublimated on the cool part of glassware and no product can be obtained. We also tested KBH₄ and LiBH₄: no reaction has been found for KBH₄, but the reaction is very fast for LiBH₄ and no product is separable by centrifugation. The hydride, with low melting point (LiBH₄, 275 °C) or compatible with the organic phase (TBAB), reacts too fast to form stable nuclei before the dissociation of Nd(acac)₃ and Fe(acac)₃. The triethylamine borane complex is stable at the reflux temperature of the solvent and surfactants for its three B–C covalent chemical bonds. This could be the reason why superhydride (LiBEt₃H) has been selected to synthesize colloidal ε-Co nanoparticles: to form a homogeneous solution in the organic phase (dioctylether) and to avoid possible incorporation of B atoms into ε-Co. In the series of LiBH₄, NaBH₄ (m.p., 400 °C), and KBH₄ (m.p, 500 °C), only NaBH₄ can react steadily with the precursors and surfactants (OA and OM) in organic solution near 300 °C to form stable nuclei and its solid state ensures the following growth process. Thus, the suitable chemical reactivity of NaBH₄ and its solid state at the selected aging temperature realize a reasonable separation between nucleation and growth. When the temperature increases to 275 °C, there are small bubbles in the reaction, which are the bubbling effects taking place through absorbing local latent heat released from the exothermic reactions involved in the nucleation and growth of magnetic nanoparticles. The bubbling effects facilitate the formation of a huge number of nuclei.¹³ The reactions in the system are rather complex. At least, the amidation reaction between OA and OM, the dissociation of Nd(acac)₃ and Fe(acac)₃ (facilitated by surfactants, such as OA and OM, and/or by hydrides), and the reduction of Fe^(III) by borohydride are involved.¹⁴ Because of the very high positive redox potential of the metal Nd, NaBH₄ cannot reduce Nd(III) to Nd, supported by the XPS peak of Nd 3d_5/2 at 982 eV, being the same value as that of Nd₂O₃ (see ESI, Fig. S3†). Even if this reduction reaction occurs, the water molecules from the above-mentioned amidation reaction and OA can oxidize it to Nd(III). Considering the intrinsic oxophilic properties of Nd(III) and the presence of water and other oxygen-containing molecular species, the Nd–Fe–B–Na nanoparticles could contain some oxygen. These factors can contribute to the chemical stability to air, evaluated by the magnetic properties in the following part. These chemical reactions are different from the sputter deposition and melt spinning techniques on chemical composition control. We speculate that the reasons for formation of amorphous colloidal nanoparticles instead of crystalline ones are as follows. Firstly, the chemical potentials of the monomers (reactive atomic or molecular species) are usually high, and according to Ostwald's rule, the kinetically favoured metastable amorphous phase forms first.¹⁵ Secondly, the presence of Nd^(III), sodium, some oxygen, and high content of boron could suppress the formation of the crystalline phase. Thirdly, the temperature allowing us to synthesize ultrasmall and monodisperse nanoparticles (300–320 °C) is not high enough to rearrange and anneal the constituent atoms during growth. This rearrangement is usually associated with significantly high thermal barriers.¹⁶ It is worth noting that it is a likely trend that ultrasmall nanoparticles adopt disordered structure compared to their bulk crystalline phases, such as chemically disordered face-centered cubic (fcc) for FePt colloidal nanocrystals,¹ (γ-Fe₂O₃)_0.8(Fe₃O₄)_0.2 for Fe₃O₄ nanocrystals with a diameter of 5 nm,¹⁰ and non-periodic and lacking long-range atomic ordered structure for the Co and FeCo nanoparticles.¹⁷

The magnetic properties of Nd–Fe–B–Na colloidal nanoparticles with size being 1 nm, 2 nm, and 3 nm corresponding to Fig. 1 are measured with a SQUID–VSM system (MPMS, Quantum) in the temperature range lower than 400 K. The M–H magnetic hysteresis curves are depicted in Fig. 2. The linear paramagnetic characterization is for the colloidal nanoparticle with 1 nm diameter. When the diameter increases to 2 nm, the ferromagnetic magnetic hysteresis loop is observed, with the coercive field (H_c) being 120 Oe at 300 K (Fig. 2a). The coercive fields (H_c) of nanoparticles with 3 nm diameter are 262 Oe at 300 K and 245 Oe at 400 K (Fig. 2b). These H_c values are higher than that of bulk amorphous alloy measured at 300 K, about 10 Oe for Nd₈Fe₇₂B₂₀.^12d The saturation magnetization (M_s) at 6 T is 14 emu g⁻¹, comparable with that of Fe₃O₄ nanoparticles with the same size,^5c which can exclude the possibility of the ferromagnetic properties originating from defect magnetism. In Fig. 2, the time for the M–H curve measured at 400 K is twenty days later than the M–H curve measured at 300 K. From the M–H curve at 300 K in Fig. 2b inset, we can see that the magnetic nanoparticles with 3 nm diameter cannot be spontaneously magnetized or induced by the magnetic bar during colloidal synthesis. However, after the magnetic measurement at 300 K, the nanoparticle maintains its magnetic remanence at least for twenty days, as shown in Fig. 2b inset for the M–H curve at 400 K. The stability in air of the Nd–Fe–B–Na colloidal nanoparticles is interrogated by the change of the magnetic properties after being exposed to air for two months. The coercive field (H_c) and magnetic remanence (M_r) decrease only by 0.5 percent, while the saturation magnetization (M_s) at 6 T decreases by 2.5 percent (see ESI, Fig. S4†). These indicate that the 3 nm nanoparticles are stable in air. The magnetic properties decreasing after exposing to air for several days is due to the oxidation of the trace surface unstable Fe (e.g., Fe^(I) and Fe^(II)) or Fe⁽⁰⁾ nanocluster.

Fig. 1 TEM characterizations of colloidal Nd–Fe–B–Na nanoparticles with different sizes. (a) 1 nm; (b) 2 nm; (c) 3 nm; (d) 5 nm.

Fig. 2 Magnetic properties of colloidal Nd–Fe–B–Na nanoparticles with different sizes. (a) Magnetic hysteresis of nanoparticles with 1 nm diameter (black curve) and 2 nm (red curve). (b) Magnetic hysteresis of nanoparticles with 3 nm diameter at 300 K (black curve) and 400 K (red curve).

To shed more light on the structure, the magnetic properties of the 3 nm nanoparticles under temperatures higher than 400 K are determined using a LakeShore 7407 vibrating sample magnetometer (VSM) equipped with a Hall probe. The M–H magnetic hysteresis curves and the magnetic parameters are given in Fig. 3. As the temperature of the sample increases to 650 K, the saturation magnetization and magnetic remanence decrease. However, the coercive field (H_c) decreases first and then increases a little, from 308 Oe to 270 Oe and then to 293 Oe (see ESI, Fig. S5†). We do not measure the magnetic properties at temperatures higher than 650 K for the following considerations: higher temperature would decompose the organic surfactants (OA and OM) and result in the aggregation by annealing and sintering. The thermal analysis of the colloidal magnetic nanoparticles under a nitrogen atmosphere is conducted (see ESI, Fig. S6†). There are an endothermic peak of the DSC curve and a weight loss peak of the TGA curve near 723 K, and these peaks can be tentatively assigned to the corresponding process of decomposition of the organic surfactants of OA and OM and the fusion and aggregation of the nanoparticles. The magnetic properties at temperatures higher than 650 K would arise from the fusion and aggregation of the ‘naked’ nanoparticles, not at the single nanoparticle level. When the size increases to 5 nm (Fig. 1d), the magnetic nanoparticles can be spontaneously magnetized or induced by the magnetic bar during colloidal synthesis, shown in Fig. S7.† The coercive field (H_c) and magnetic remanence (M_r) are 612 Oe and 11.2 emu g⁻¹ measured by a LakeShore 7407 VSM at 300 K, while the saturation magnetization (M_s) at 1.8 T is 29.6 emu g⁻¹.

Fig. 3 Magnetic properties of colloidal Nd–Fe–B–Na nanoparticles with 3 nm diameter at high temperatures between 400 K and 650 K.

To study the magnetic properties of the 3 nm Nd–Fe–B–Na nanoparticles, we performed extended X-ray absorption fine structure (EXAFS) and Mössbauer spectroscopy characterization. The radical distribution functions (RDF) of the 3 nm Nd–Fe–B–Na nanoparticle and Fe foil, obtained from their k³χ(k) by Fourier transform in Fig. S8,† are depicted in Fig. 4. There are several distinct peaks of the sample. Compared with the peaks of Fe foil, the peaks at 2.58 Å, 3.9 Å, and 4.6 Å of the sample may partially result from the Fe⁰ nanocluster. The peak at 1.86 Å can be assigned to the amorphous phase. Structural parameters obtained by fitting the Fe K-edge data shown in Fig. 4 are listed in Table S1.† There are about 3.5 boron atoms and 1.2 Fe atoms around each Fe center. The nearest Fe–B and Fe–Fe distances are 1.99 Å and 2.91 Å. Although Nd₂Fe₁₄B nanomagnets and nanocomposites as crystlline sample are characterized by EXAFS, the data on the fitted results for chemical bonds of Fe–B and Fe–Fe are still lacking. It is difficult to directly compare the crystalline phase with the amorphous phase. The distance of Fe–B is shorter while the distance of Fe–Fe is larger than those of the amorphous Fe–Si–B alloy, 2.27 Å and 2.54 Å, respectively.^11b The first peak in the RDF of the 3 nm nanoparticle is in a slightly higher position than those of standard Fe₂O₃ and Fe₃O₄ in Fig. S9,† indicating the Fe–B bonds being slightly larger than the Fe–O chemical bonds.

Fig. 4 RDFs of the 3 nm Nd–Fe–B–Na nanoparticles and the standard Fe foil.

⁵⁷Fe Mössbauer spectroscopy of the sample measured at 300 K in Fig. 5 and Table S2† confirms that there is a sextet subspectrum of subarea of 24% (subspectrum 2) with the same isomer shift (δ) as that of Fe foil and a hyperfine field of 33.3 kOe attributed to unoxidized Fe⁰ nanoparticles in a blocked state, assigned to Fe⁰ nanocluster(s).¹⁸ To exclude the possibility of the sextet subspectrum resulting from very large Fe nanoparticles not observed by TEM, we characterize the dynamic light scattering (DLS) of the 3 nm Nd–Fe–B–Na nanoparticles dispersed in hexane; there is no sign of particles with hydrodynamic diameter larger than 7 nm (Fig. S10†). For freestanding colloidal Fe nanoparticles, the blocking temperature of unoxidized 4.5 nm Fe nanoparticles is just 110 K.¹⁹ We speculate that the sextet subspectrum of the Fe⁰ nanocluster(s) could originate from their exchange-coupling with the amorphous parts surrounding them inside the 3 nm Nd–Fe–B–Na nanoparticles. This structural feature is in accordance with the RDF of the nanoparticles in Fig. 4. Although the temperature to measure the Mössbauer effect is far below T_C, there is still a doublet peak with 40% subarea. This phenomenon is different from the magnetic behavior of CoFe₂O₄ nanocrystals and Fe nanocube assembly with only one sextet peak under the blocking temperature, probably due to structural disorder and surface effect.^4a,18 Because of the surface effect and complex chemical composition, the Mössbauer spectra are also different from those of bulk amorphous Fe–B–Si, the bulk without any doublet peak.^11b However, there is no sign of Fe⁰ nanoclusters in the 5 nm nanoparticles shown in Fig. S11 and Table S3,† probably due to higher synthetic temperature being able to destroy the Fe nanocluster by forming Fe–B chemical bonds. As the size increases to 5 nm, the subarea of the doublet peak from Fe nuclei with unblocked state decreases slightly to 37%. For the nanoparticles with 1 nm or 2 nm diameter, there are only two doublet peaks, although the H_c is 120 Oe for the 2 nm one (Fig. S11 and Table S3†).

Fig. 5 ⁵⁷Fe Mössbauer spectrum of the 3 nm Nd–Fe–B–Na nanoparticles measured at 300 K.

For the present colloidal Nd–Fe–B–Na nanoparticles, when the size decreases to 1 nm, the zero-field-cooling (ZFC) magnetization (Fig. S12†) under 1000 Oe magnetic field indicates that there is no peak magnetization between 2 K and 305 K, unlike those of ε-Co, CoFe₂O₄, and Fe₃O₄ magnetic nanoparticles.^1,4a,5a These results probably arise from finite size effects suppressing the magnetic ordering. Considering the ferromagnetic properties of magnetic nanoparticles larger than 2 nm (Fig. 2a, red curve), the size is a crucial factor to determine the magnetic exchange model. The critical size of the present colloidal Nd–Fe–B–Na nanoparticles from the experimental results should be between 1 nm and 2 nm, which is close to the estimated grain size of Gd–Co amorphous films (smaller than 25 Å) and the diameter of Fe-enriched clusters simulated in the inhomogeneous Fe₆₇Y₃₃ alloys (about 15–20 Å).^10a,11a

Conclusions

In summary, the controlled colloidal synthesis of amorphous Nd–Fe–B–Na nanoparticles and the size dependence of the magnetic properties have been demonstrated in the sub-5 nm regime. The magnetic nanoparticles with 3 nm diameter exhibit ferromagnetic magnetic properties even at temperatures as high as 650 K. The structure of the 3 nm Nd–Fe–B–Na nanoparticles has been characterized by EXAFS and Mössbauer spectroscopy. These characteristics clearly demonstrate that such Nd–Fe–B–Na nanoparticles offer opportunities for fundamental studies of nanomagnetics, as well as for potential ultrahigh-density magnetic recording, permanent magnetic nanocomposites, catalysis, and biological applications. The present colloidal method can also be used to grow colloidal nanoparticles of other amorphous metal alloys with nanometer size, with potential importance for magnetic, magnetooptic, and electronic applications.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (91122003) and the Developing Science Funds of Tongji University.

Notes and references

1. S. H. Sun, C. B. Murray, D. Weller, L. Folks and A. Moser, Science, 2000, 287, 1989.
2. (a) S. H. Sun and C. B. Murray, J. Appl. Phys., 1999, 85, 4325; (b) C. T. Black, C. B. Murray, R. L. Sandstrom and S. H. Sun, Science, 2000, 290, 1131; (c) V. F. Puntes, K. M. Krishnan and A. P. Alivisatos, Science, 2001, 291, 2115.
3. H. Zeng, J. Li, J. P. Liu, Z. L. Wang and S. H. Sun, Nature, 2002, 420, 395.
4. (a) C. Liu, B. S. Zou, A. J. Rondinone and J. Zhang, J. Am. Chem. Soc., 2000, 122, 6263; (b) T. Hyeon, S. S. Lee, J. Park, Y. Chung and H. Bin Na, J. Am. Chem. Soc., 2001, 123, 12798; (c) J. I. Park and J. Cheon, J. Am. Chem. Soc., 2001, 123, 5743; (d) S. H. Sun and H. Zeng, J. Am. Chem. Soc., 2002, 124, 8204.
5. (a) J. Park, K. J. An, Y. S. Hwang, J. G. Park, H. J. Noh, J. Y. Kim, J. H. Park, N. M. Hwang and T. Hyeon, Nat. Mater., 2004, 3, 891; (b) J. Park, E. Lee, N. M. Hwang, M. S. Kang, S. C. Kim, Y. Hwang, J. G. Park, H. J. Noh, J. Y. Kini, J. H. Park and T. Hyeon, Angew. Chem., Int. Ed., 2005, 44, 2872; (c) B. H. Kim, N. Lee, K. An, Y. I. Park, Y. Choi, K. Shin, Y. Lee, S. G. Kwon, H. B. Na, J. G. Park, T. Y. Ahn, Y. W. Kim, W. K. Moon, S. H. Choi and T. Hyeon, J. Am. Chem. Soc., 2011, 133, 12624; (d) J. Dobson, Nat. Nanotechnol., 2008, 3, 139.
6. (a) C. Tassa, S. Y. Shaw and R. Weissleder, Acc. Chem. Res., 2011, 44, 842; (b) D. Yoo, J. H. Lee, T. H. Shin and J. Cheon, Acc. Chem. Res., 2011, 44, 863; (c) D. Ho, X. L. Sun and S. H. Sun, Acc. Chem. Res., 2011, 44, 875; (d) 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.
7. (a) A. H. Lu, W. Schmidt, N. Matoussevitch, H. Bonnemann, B. Spliethoff, B. Tesche, E. Bill, W. Kiefer and F. Schuth, Angew. Chem., Int. Ed., 2004, 43, 4303; (b) M. Shokouhimehr, Y. Z. Piao, J. Kim, Y. J. Jang and T. Hyeon, Angew. Chem., Int. Ed., 2007, 46, 7039; (c) V. Mazumder, M. F. Chi, K. L. More and S. H. Sun, J. Am. Chem. Soc., 2010, 132, 7848; (d) Y. M. Lee, M. A. Garcia, N. A. F. Huls and S. H. Sun, Angew. Chem., Int. Ed., 2010, 49, 1271; (e) S. J. Guo, S. Zhang, X. L. Sun and S. H. Sun, J. Am. Chem. Soc., 2011, 133, 15354.
8. (a) J. P. Ge, Y. X. Hu and Y. D. Yin, Angew. Chem., Int. Ed., 2007, 46, 7428; (b) J. P. Ge, L. He, J. Goebl and Y. D. Yin, J. Am. Chem. Soc., 2009, 131, 3484; (c) J. P. Ge and Y. D. Yin, Angew. Chem., Int. Ed., 2011, 50, 1492; (d) Y. X. Hu, L. He and Y. D. Yin, Angew. Chem., Int. Ed., 2011, 50, 3747.
9. (a) N. G. Akdogan, W. F. Li and G. C. Hadjipanayis, J. Appl. Phys., 2011, 109, 07A759; (b) C. W. Kim, Y. H. Kim, U. Pal and Y. S. Kang, J. Mater. Chem. C, 2013, 1, 275; (c) A. Hussain, A. P. Jadhav, Y. K. Baek, H. J. Choi, J. Lee and Y. S. Kang, J. Nanosci. Nanotechnol., 2013, 13, 7717.
10. (a) P. Chaudhari, J. J. Cuomo and R. J. Gambino, Appl. Phys. Lett., 1973, 22, 337; (b) Y. Togami, Appl. Phys. Lett., 1978, 32, 673; (c) R. Harris, M. Plischke and M. J. Zuckermann, Phys. Rev. Lett., 1973, 31, 160.
11. (a) D. Spisak, J. Hafner and R. Lorenz, J. Magn. Magn. Mater., 1997, 166, 303; (b) Y. Takahara and N. Narita, Mater. Sci. Eng., A, 2001, 315, 153; (c) N. Hassanain, H. Lassri, R. Krishnan and A. Berrada, J. Magn. Magn. Mater., 1995, 146, 37; (d) N. Hassanain, H. Lassri, R. Krishnan and A. Berrada, J. Magn. Magn. Mater., 1995, 146, 315.
12. (a) M. Slimani, M. Hamedoun, H. Lassri, S. Sayouri and R. Krishnan, J. Magn. Magn. Mater., 1996, 153, 132; (b) M. Slimani, M. Hamdoun, M. Tlemcani, H. Arhchoui and S. Sayouri, Physica B, 1997, 240, 372; (c) E. H. Sayouty, F. Annouar, H. Lassri, N. Randrianantoandro and J. M. Greneche, J. Alloys Compd., 2005, 397, 47; (d) N. Hassanain, H. Lassri, A. Berrada and R. Krishnan, Phys. Status Solidi C, 2006, 3, 3239.
13. J. Lynch, J. Q. Zhuang, T. Wang, D. LaMontagne, H. M. Wu and Y. C. Cao, J. Am. Chem. Soc., 2011, 133, 12664.
14. (a) N. R. Jana, Y. F. Chen and X. G. Peng, Chem. Mater., 2004, 16, 3931; (b) J. Joo, S. G. Kwon, J. H. Yu and T. Hyeon, Adv. Mater., 2005, 17, 1873; (c) H. M. Wu, Y. G. Yang and Y. C. Cao, J. Am. Chem. Soc., 2006, 128, 16522; (d) H. T. Zhang, D. H. Ha, R. Hovden, L. F. Kourkoutis and R. D. Robinson, Nano Lett., 2011, 11, 188; (e) Q. Zhang, B. Yan, F. Lei and H. H. Chen, Nanoscale, 2012, 4, 7646.
15. W. Z. Ostwald, Phys. Chem., 1897, 22, 289.
16. L. E. M. Howard, H. L. Nguyen, S. R. Giblin, B. K. Tanner, I. Terry, A. K. Hughes and J. S. O. Evans, J. Am. Chem. Soc., 2005, 127, 10140.
17. (a) D. P. Dinega and M. G. Bawendi, Angew. Chem., Int. Ed., 1999, 38, 1788; (b) B. Chaudret, C. Desvaux, C. Amiens, P. Fejes, P. Renaud, M. Respaud, P. Lecante and E. Snoeck, Nat. Mater., 2005, 4, 750.
18. F. Dumestre, B. Chaudret, C. Amiens, P. Renaud and P. Fejes, Science, 2004, 303, 821.
19. T. C. Monson, E. L. Venturini, V. Petkov, Y. Ren, J. M. Lavin and D. L. Huber, J. Magn. Magn. Mater., 2013, 331, 156.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4qi00049h

---