Shear induced micromechanical synthesis of Ti₃SiC₂ MAXene nanosheets for functional applications

Abstract

Herein, we report the synthesis of ultrathin 2D Ti₃SiC₂ (MAXene) nanosheets via a facile shear induced micromechanical cleavage strategy. The very high dispersion stability, the UV absorption properties, high electrical conductivity and castability into thin films make the newly derived Ti₃SiC₂ nanosheets an ideal candidate for many functional applications.

---

MAX phase is a classical layered ternary carbide/nitride with the chemical composition of M_n+1AX_n where M is an early transition metal, A represents a group IIIA or IVA element and X is either carbon and/or nitrogen.¹ Its crystal structure consists of alternate MX layers separated with A layer atoms. MAX phase materials are reported to have unusual and unique properties such as solid lubrication, damage tolerance, thermal shock resistance, machinability and low friction. Hence, they are preferred as multifunctional modifiers for ceramics as well as polymer composites.^2–4 The discovery of graphene and its unique nanostructures via mechano-chemical/sonication techniques triggered interest in the processing of similar kinds of 2D nanolayered materials via any novel strategies.⁵ Recently synthesis of 2D nanostructures from metal chalcogenides (MoS₂, WS₂ etc.), BN, and double layered hydroxides have been successfully demonstrated through exfoliation of their 3D counterparts.^6–12 These materials garnered great attention because they showed enhanced optical, thermal, mechanical and electrical properties compared to their bulk counterparts. They have been suggested for the applications in thermal fluids, light emitting diodes, super capacitors, thermoelectricity, Li-ion batteries and also as reinforcements in composites.^13–21

Very recently, Naguib et al. demonstrated an experiment to obtain 2D MX layers by selective etching of Al atoms from the selected MAX phase material.^22,23 This newly derived materials were named as MXenes to spotlight their similarity with graphene. Further, MXenes are explored for the applications in Li-ion batteries, Lead ion absorption and photo catalaysis.^24–27 However, the method proposed by Naguib et al. requires toxic HF as etchant for the selective removal of A atoms and also this method can be success only when A layer is Al atoms. The loss of A elements from the crystal structure can probably alter the intrinsic properties of MAX phases.

A relatively strong inter atomic-bonding between MX and A layers of MAX phase makes the simple exfoliation techniques inappropriate for the delamination of the nanolayers into a single MAX phase nanosheets. Recently, Zhang et al. synthesized ultrathin nanosheets of MAX phases by substitutional solid solution method.²⁸ They fabricated doped phases Ti₃Si_0.75Al_0.25C₂ and Ti₃Al_0.9Si_0.1C₂ by high-temperature self-propagation synthesis and ultrasonically exfoliated these phases. It finally ends up with the same composition as the parent MAX phase i.e. with two types of A atoms. To the best of our knowledge, the preparation of phase pure MAX phase nano sheets is yet to be reported. Herein, we attempted a modified micromechanical cleavage strategy for obtaining single phase 2D MAX phase nanosheets. The newly processed nanosheets were named as MAXenes to highlight their 2D morphology. Micromechanical cleavage is reported to be a successful technique for the exfoliation of MoS₂, BN, NbSe₂, Bi₂Sr₂CaCu₂O_x and graphite.⁷ In these materials, a simple rubbing technique was used to delaminate the layered structures. Unfortunately, adopting such technique directly gives great failures and delamination of MAX phase is not seen. Therefore, we put forward a new idea, shear-induced micromechanical cleavage, wherein a simultaneous application of compressive and rotational shear force along with mechanical grinding in order to induce shearing of MAX phase layers. To the best of our knowledge, this is the first report on such shear-induced micromechanical cleavage for exfoliated MAX phase nanosheets.

Fig. 1 depicts the present strategy wherein we explained the simultaneous application of compressive and rotational forces, one in parallel and another in the perpendicular direction to the basal planes for the exfoliation. Application of shear-induced mechanical grinding within a chemical bath is found to be quite successful to design 2D MAX phase nanostructures. In a dense MAX phase bulk material, mechanical deformation is studied systematically and said to be mediated by anisotropic basal dislocations which resulted in partial delaminations and produce lamellas with a few tens to hundred nanometers thickness.²⁹ However, mechanical delamination in micrometer size MAX phase particles in the presence of prolonged shear force is not documented before. Herein, we have investigated the effect of shear induced micromechanical delamination of micronic Ti₃SiC₂ (hereafter denotes as bulk TSC) MAX phase and examined the evolution and properties of exfoliated nanostructures (hereafter denotes as MAXene nanosheets). The nanolayered ternary transition metal carbides consist of M–X slabs interleave with A atom layers. In Ti₃SiC₂, Ti–C slabs are intercalated with planar packed Si layers. Ti and C atoms are connected through strong covalent bonds, which are mechanically rigid. On the other hand, the Ti–Si bonds are relatively weaker than Ti–C bonds and are prone to compression and stretching under shear deformation.^30,31 These characteristics favours easy basal plane slip and offer low shear deformation resistance. The ab initio calculations reported by Medvedeva et al. also support the weak coupling between Ti₃C₂ and Si layers.³² However, the bond strength between Ti–C and Si layers is not as weak as in the case of graphite and its analogues, where weak van der Waals force holds the layers together.^6,33 A simple ultrasonication can mechanically break the weakly bonded graphite layers and produce freestanding individual graphite sheets. However, such a technique is inadequate to cleave the covalently bonded TiC–Si layers. Application of directionally varied shear force is expected to show advantages in shearing or the cleavage of the stacked nanolayers. Retsch mortar grinder offers provision for applying both pressure and friction simultaneously. In Retsch grinder ceramic mortar is rotating at a speed of 100 rpm and ceramic pestle can induce constant mechanical pressure. The pestle is situated not in the center of the mortar but is offset. Mortar and the pestle rotate in opposite directions (pestle in clockwise and the mortar in anticlockwise directions) caused friction and compressive forces upon the bulk TSC particles dispersed in polar solvents. During grinding, the compressive and frictional forces resulted in mechanical shearing along the TSC basal planes. Each contact point of the pestle and mortar acted as micro-scale cleavage points. Ultimately shearing occurs effectively in bulk TSC nanolayers. The polar medium helps to keep the separated or partially delaminated nanolayers in colloidal range from the micronic parent particle.

Fig. 1 Schematic representation of the synthesis procedure to obtain ultrathin MAXene nanosheets using shear induced micromechanical cleavage technique.

The electron microscopic images of the TSC particle observed before and after micromechanical cleavage clearly show different morphological features and the same are presented in Fig. 2. The SEM image in Fig. 2(a) shows the morphology of bulk TSC material before micromechanical cleavage. The stacking of multiple numbers of TSC nanolayers is apparently seen in the SEM image. Fig. 2(b) gives the representative TEM morphology of exfoliated MAXene nanosheets. This clearly confirms the effective exfoliation of TSC stacks into ultrathin TSC nanosheets. The microstructure of the exfoliated MAXene nanosheets consists few layers. Single layered (dotted white circle) sheets are also clearly identified in the micrograph. The SAED pattern of the corresponding image is given as inset in Fig. 2(b) that confirms the crystalline nature of the exfoliated MAXene nanosheets. Fig. 2(c) represents the corresponding high resolution TEM (HRTEM) images which reveal the (101) plane of single MAXene nanosheets, indicating that nanosheets exposed the XZ plane. It is clear from the TEM images that the exfoliated MAXene nanosheets possess sheet length lower than its bulk counterpart. The dimensional decrease in nanosheets is obvious due to the grinding effect where particle size reduction is expected. Other than this, the formation of micro-scale kink bands may also be expected during mechanical deformation. The kink band is reported to obstruct the propagation of delamination in MAX phase materials and prevent the full-scale delamination of the particle.²⁹ However, the continuous application of shear force for longer hours resulted in micro scale peeling of nanostacks at the kink junctions that finally yielded delaminated nano morphologies. In order to understand the role of milling time on morphology of the MAXene nanosheets, we performed the experiments at 12 h and 30 h also. It was observed that the product obtained after 12 milling consists of partially delaminated large sheets with relatively large number of layers and that of 30 h consists ultrasmall (∼10 nm) sheets resulted by breakdown of nanostructures due to excessive milling. The TEM images of the products obtained after various milling time is depicted in Fig. S1.† The EDX spectrum (Fig. 2(d)) shows the chemical composition of exfoliated MAXene nanosheets obtained by 24 h milling. It confirms that the stoichiometry is identical to that of the bulk TSC. However, the presence of trace amounts of surface oxygen is also evidenced. It is envisaged a microthermal heating during mechanical milling that can cause atomic-level oxidation.

Fig. 2 The morphological features of bulk and exfoliated MAXene nanosheets (a) SEM image of bulk TSC (b) TEM image of exfoliated MAXene nanosheets (c) HRTEM image of exfoliated MAXene nanosheets showing respective lattice spacing and (d) corresponding EDX spectrum.

In this work high boiling point solvents are used for the exfoliation. It was observed that the exfoliation comparatively effective in polar solvents. We also conducted the identical experiments in non-polar solvents such as toluene and o-dichlorobenzene (o-DCB) to understand the role of polarity in exfoliation efficiency. In non-polar medium the particle settling is repeatedly observed, whereas the nanolamellar TSC delaminated effectively in polar solvents and form a stable dispersion. In fact, it is a surprise to see the dispersion is stable even after four months duration. Fig. 3 shows the dispersion stability of MAXene nanosheets processed at different particle concentrations in DMF medium over a period of 100 days. It is noticed the stability of MAXene exfoliates is irrespective of the concentration. It is likely that the extensive milling in polar medium generate chemically identical charged surface functional groups on the nanosheets which in turn give stability via long range attraction or short range repulsive force by the interaction with the polar groups in the solvent. This was confirmed by measuring the zeta potential of the MAXene dispersion using the Malvern Zetasizer. The MAXene exfoliates in DMF is found to have a zeta potential value of ∼−29.2 mV. However, the same for the bulk TSC is only ∼0.08 mV. A surface hydroxylation is expected because mechanical milling in polar solvents was carried out in the open atmosphere. The extended milling time given to the effective dissemination generated a temperature of ∼60 °C and such micro-thermal effect probably the reason for the surface –OH group formation. This surface –OH layer is helping the nanosheets to form a stable dispersion in polar solvents. Probable interaction with the polar groups of solvent with the oxide layer may be the reason for extended stability of the MAXene nanosheets in polar solvents. Fig. 3(b) confirms further the typical Tyndall effect which is observed by allowing the laser beam to pass through the MAXene dispersions prepared with different solid concentrations. The photographs in Fig. 3(b) and (c) taken at an angle perpendicular to the direction of the laser beam is clearly showing a well intense Tyndall effect in both dilute as well as concentrated dispersions. The dilute dispersion is physically transparent that ascertains the MAXene exfoliates have ultrathin dimensions. In order to check the aggregation property of the MAXene dispersion, turbidity values were recorded using Nephelometer for a period of three months at 15 days intervals. The turbidity values are shown in Fig. 3(d). No significant change in the turbidity values confirms that there is no agglomeration occurring in nano dimensional MAXene dispersions and are very stable without any settling of nanosheets.

Fig. 3 MAXene nanosheets prepared in DMF medium via micromechanical cleavage. The photographs of the (a) dispersion of MAXene nanosheets at different concentrations, (b) & (c) Tyndall effect shown by the dispersions and (d) the turbidity values of the dispersions showing the stability in various polar solvents.

Fig. 4 depicts the powder X-ray diffraction patterns of the bulk TSC before and after micromechanical exfoliation. The XRD pattern indicates a change in the crystallinity of the TSC upon exfoliation. The intense, high order peaks correspond to the planes (006), (008), (107), (108) and (109) present in the bulk TSC are disappeared during exfoliation indicate the occurrence of delaminated TSC nanolamellar superstructure. As expected, the peaks are also broadened and down shifted to lower angles with a slight increase in the c lattice parameter. The crystallite size of the delaminated MAXene nanosheets was calculated by the well-known Scherrer equation and it was noticed that the crystallite size is reduced to ∼13 nm upon exfoliation from the bulk crystallite size of ∼60 nm.

Fig. 4 The powder X-ray diffraction patterns of TSC before (a) and after (b) micromechanical cleavage. The broadening of the peaks on exfoliation is shown in the inset.

The surface chemical compositions of the MAXene nanosheets were examined by X-ray photoelectron spectroscopy (XPS) and the spectra are given in Fig. 5. It supports strongly the atomic level oxidation of TSC during milling. The survey spectrum (Fig. 5(a)) revealed the surface of MAXene nanosheets consist of Ti, Si and C atoms along with oxygen atom. The high resolution spectra of Ti, C and Si indicate a deviation from the surface chemical composition reported earlier for the bulk TSC.^34,35 As evidenced from the high resolution XPS spectra of Ti 2p, that the O contribution mainly comes from the TiO_x fractions. The peak at ca. 458.2 eV and 463.6 eV corresponds to the Ti–O bonds.³⁶ The peak at 454 eV represents carbide contribution from Ti–C bonds.³⁵ The C 1s spectra indicate the existence of graphitic carbon (ca. 284.6 eV) and carbide contribution (ca. 280 eV) which is identical to the values seen for the bulk TSC. The presence of SiO_x was also identified by the peak at ca. 101 eV in the Si 2p spectra. The O 1s spectra (Fig. S2†) represents that the oxide contribution from M–O (M = Ti, Si) bonds. The absence of metal oxide peaks corresponds to TiO₂ and SiO₂ in the powder X-ray pattern (Fig. 4) suggests that the functionalization occurs only at the surface and not in the bulk of the TSC nanosheets.

Fig. 5 XPS of exfoliated MAXene nanosheets. The survey spectrum showing all the elements present on the surface (a) and high resolution spectra of (b) Ti 2p, (c) C 1s and (d) Si 2p.

It was noticed that the exfoliated MAXene nanosheets processing UV absorption properties. UV/visible absorption spectra of the as exfoliated MAXene dispersion prepared in DMF medium is given in Fig. 6(a). It is clear from the spectra that the bulk TSC has no absorption in the UV/visible region. Whereas the exfoliated TSC nanosheets exhibit high absorption in the UV region with an onset abortion at 410 nm with a strong band centered at 290 nm. This corresponds to band gap energy of 3.05 eV. It suggests the exfoliated TSC nanosheets are becoming semiconducting in nature. The spectrum is reminiscent to that is reported for graphene oxide.³⁷ The presence of surface TiO_x groups can probably contribute to the UV absorption at λ < 400 nm. It is important to note that these are only preliminary results and further studies are required before understanding this property. More detailed studies are in progress to uncover the mechanism involved in this novel behaviour of MAXene nanosheets.

Fig. 6 (a) UV-Visible spectra of bulk TSC and exfoliated MAXene nanosheets in DMF and (b) photograph of flexible MAXene film prepared by deposition on nylon 6,6 membrane.

Separation of the exfoliated MAXene nanosheets from the DMF medium is a trivial task as in the case of most of the 2D materials, because the separation by simple evaporation results in restacking of layers. In the present work, we have successfully extracted the exfoliated MAXene nanosheets from the solvent by precipitation with chloroform, a less polar volatile solvent (Fig. S3†). On dilution with excess chloroform, the nanosheets settle at the bottom. The decantation of the solvent followed by freeze drying gives the MAXene nanosheets devoid of restacking. It was also noticed that the exfoliated MAXene nanosheets can be easily compressed into definite cylindrical pellets which is difficult in 2D graphene and SWCNTs. The photographs of the green pellets are given as Fig. S4 (ESI†). The exfoliated MAXene possesses a good green strength than bulk TSC.

The resistivity and conductivity of the exfoliated MAXene nanosheets compressed into a cylindrical disc of thickness <1 mm and a diameter of 10 mm are given in Table 1. It is worth to note that the resistivity values mentioned in Table 1 is not exactly the value of a single MAXene layer but a bulk resistivity of the green compact.²³ It is noticed that the MAXene nanosheets possess conductivity six times higher than that of the bulk TSC. Such a comparison has earlier been reported between graphene and bulk graphite. Graphene sheets have increased conductivity than the parent graphite. Similar kind of comparable conductivity is seen with exfoliated MAXene with the bulk TSC.

Table 1 Resistivity, electrical conductivity and green density of cold-pressed MAXene discs

Property	Bulk TSC	Exfoliated TSC (MAXene)
Resistivity (Ω m)	0.00847	0.00158
Conductivity (S m^−1)	1.18 × 10^2	6.30 × 10^2
Density of cold pressed disc (% of theoretical density)	2.46 (55%)	1.45 (32%)

---

The viability of making MAXene membrane using the exfoliated MAXene nanosheets was also demonstrated. For this the dispersion of MAXene nanosheets was injected through a nylon 6,6 membrane having a pore size of ∼0.2 μm. While passing through this nylon membrane, the MAXene nanosheets were deposited and form a crack free thin film membrane module. The newly developed MAXene membrane is flexible as demonstrated in Fig. 6(b). One can fabricate an ultra thin film to thick nanostructured membrane with any thickness by varying the volume MAXene dispersion.

Conclusions

In this work, a modified micromechanical cleavage strategy was successfully demonstrated to produce 2D derivatives of Ti₃SiC₂, the most famous member among the MAX phase ternary carbides. The exfoliation by micromechanical cleavage is more effective in polar solvents which result in 2D Ti₃SiC₂ (MAXene) dispersion stable for several months. TEM images confirmed that the ultrathin MAXene nanosheets possess only a few layer thickness. The presence of surface oxide groups formed during the milling process enhanced the dispersion stability and UV activity make the nanosheets a new semiconducting material. The exfoliated nanosheets compact has the electrical conductivity six times higher than the bulk counterpart. Thus, this work presented a design of nanostructured 2D carbides via a simple, innovative method for the first time and this new class of 2D functional material can be utilized to make a stable conducting ink, fuel cell electrodes and conducting membranes. Also, the method developed in this work is a general technique which can be effective in delamination of all the MAX phases irrespective of the type of ‘A’ atoms.

Acknowledgements

This work was partially supported by the Department of Science and Technology (DST), Govt. of India. The authors are grateful to the Director, CSIR-National Institute for Interdisciplinary Science and Technology for providing the necessary lab facilities. Authors are thankful to Dr K. G. Nishanth for the fruitful discussions and Mr Irshad Ahamed for his help in obtaining XPS. K. V. M. acknowledges CSIR, Govt. of India for Senior Research Fellowship.

Notes and references

1. M. W. Barsoum, Prog. Solid State Chem., 2000, 28, 201281.
2. H. Chunfeng, Z. Yanchun, B. Yiwang and W. Detian, J. Am. Ceram. Soc., 2006, 89, 3456–3461.
3. S. Sui Lin and P. Wei, J. Am. Ceram. Soc., 2007, 90, 3331–3333.
4. K. V. Mahesh, S. Balanand, R. Raimond, A. P. Mohamed and S. Ananthakumar, Mater. Des., 2014, 63, 360–367.
5. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
6. G. Deepesh, D. Dijo and M. S. Manikoth, ACS Nano, 2014, 8, 5297–5303.
7. K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, S. V. Morozov and A. K. Geim, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 10451–10453.
8. P. F. Joensen, R. F. Frindt and S. R. Morrison, Mater. Res. Bull., 1986, 21, 457–461.
9. J.-W. Seo, Y.-W. Jun, S.-W. Park, H. Nah, T. Moon, B. Park, J.-G. Kim, Y. J. Kim and J. Cheon, Angew. Chem., Int. Ed., 2007, 46, 8828–8831.
10. B. K. Miremadi and S. R. Morrison, J. Appl. Phys., 1988, 63, 4970–4974.
11. D. Pacile, J. C. Meyer, C. O. Girit and A. Zettl, Appl. Phys. Lett., 2008, 92, 1–3.
12. Z. Zeng, T. Sun, J. Zhu, X. Huang, Z. Yin, G. Lu, Z. Fan, Q. Yan, H. H. Hng and H. Zhang, Angew. Chem., Int. Ed., 2012, 51, 9052–9056.
13. L. Yang, S. Wang, J. Mao, J. Deng, Q. Gao, Y. Tang and O. G. Schmidt, Adv. Mater., 2013, 25, 1180–1184.
14. J. M. Soon and K. P. Loh, Electrochem. Solid-State Lett., 2007, 10, A250–A254.
15. W.-L. Song, P. Wang, L. Cao, A. Anderson, M. J. Meziani, A. J. Farr and Y.-P. Sun, Angew. Chem., Int. Ed., 2012, 51, 6498–6501.
16. K. J. Reynolds, G. L. Frey and R. H. Friend, Appl. Phys. Lett., 2003, 82, 1123–1125.
17. J. Heising and M. G. Kanatzidis, J. Am. Chem. Soc., 1999, 121, 11720–11732.
18. C. Feng, J. Ma, H. Li, R. Zeng, Z. Guo and H. Liu, Mater. Res. Bull., 2009, 44, 1811–1815.
19. M. S. Dresselhaus, G. Chen, M. Y. Tang, R. Yang, H. Lee, D. Wang, Z. Ren, J.-P. Fleurial and P. Gogna, Adv. Mater., 2007, 19, 1043–1053.
20. Z. Chunyi, X. Yibin, B. Yoshio and G. Dmitri, ACS Nano, 2011, 5, 6571–6577.
21. T.-T. Jaime, N. N. Tharangattu, G. Guanhui, R. Matthew, A. T. Dmitri, P. Matteo and M. A. Pulickel, ACS Nano, 2012, 6, 1214–1220.
22. M. Naguib, M. Kurtoglu, V. Presser, J. Lu, J. Niu, M. Heon, L. Hultman, Y. Gogotsi and M. W. Barsoum, Adv. Mater., 2011, 23, 4248–4253.
23. M. Naguib, O. Mashtalir, J. Carle, V. Presser, J. Lu, L. Hultman, Y. Gogotsi and M. W. Barsoum, ACS Nano, 2012, 6, 1322–1331.
24. X. Xiaohong, C. Siguo, D. Wei, N. Yao and W. Zidong, Chem. Commun., 2013, 49, 10112.
25. P. Qiuming, G. Jianxin, Z. Qingrui, X. Jianyong, L. Baozhong, Z. Aiguo, L. Riping and T. Yongjun, J. Am. Chem. Soc., 2013, 136, 4113–4116.
26. O. Mashtalir, K. M. Cook, V. N. Mochalin, M. Crowe, M. W. Barsoum and Y. Gogotsi, J. Mater. Chem. A, 2014, 2, 14334–14338.
27. M. Naguib, C. Jeremy, D. Boris, P. Volker, T. Pierre-Louis, S. Patrice, W. B. Michel and G. Yury, Electrochem. Commun., 2012, 16, 61–64.
28. X. Zhang, J. Xu, H. Wang, J. Zhang, H. Yan, B. Pan, J. Zhou and Y. Xie, Angew. Chem., Int. Ed., 2013, 52, 4361–4365.
29. M. W. Barsoum and M. Radovic, in Encyclopedia of Materials Science and Technology, ed. K. H. J. Buschow, R. W. Cahn, M. C. Flemings, E. J. Kramer, S. Mahajan and P. Veyssiere, Elsevier Science, Amsterdam, 2004, pp. 1–16.
30. W. Jingyang and Z. Yanchun, Annu. Rev. Mater. Res., 2009, 39, 415–443.
31. M. W. Barsoum and M. Radovic, in Annual Review of Materials Research, ed. D. R. Clarke and P. Fratzl, 2011, vol. 41, pp. 195–227.
32. N. I. Medvedeva, D. L. Novikov, A. L. Ivanovsky, M. V. Kuznetsov and A. J. Freeman, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 58, 16042–16050.
33. C. Zhi, Y. Bando, C. Tang, H. Kuwahara and D. Golberg, Adv. Mater., 2009, 21, 2889–2893.
34. S. Myhra, J. A. A. Crossley and M. W. Barsoum, J. Phys. Chem. Solids, 2001, 62, 811817.
35. E. H. Kisi, J. A. A. Crossley and S. Myhra, J. Phys. Chem. Solids, 1998, 59, 1437–1443.
36. F. Sordello, G. Zeb, K. Hu, P. Calza, C. Minero, T. Szkopek and M. Cerruti, Nanoscale, 2014, 6, 6710–6719.
37. X. Zhang, A. C. Coleman, N. Katsonis, W. R. Browne, B. J. van Wees and B. L. Feringa, Chem. Commun., 2010, 46, 7539–7541.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental section and supplementary Fig. S1–S4. See DOI: 10.1039/c5ra07756g

---