Open Access

PPMS-based set-up for Raman and luminescence spectroscopy at high magnetic field, high pressure and low temperature

  • Matthias Hudl1, 2Email author,
  • Peter Lazor3,
  • Roland Mathieu4,
  • Alexander G Gavriliuk5, 6 and
  • Viktor V Struzhkin7
Contributed equally
EPJ Techniques and Instrumentation20152:3

https://doi.org/10.1140/epjti/s40485-015-0014-x

Received: 3 October 2014

Accepted: 9 February 2015

Published: 27 February 2015

Abstract

We present an experimental set-up permitting Raman and luminescence spectroscopy studies in a commercial Physical Properties Measurement System (PPMS) from Quantum Design. Using this experimental set-up, gaseous, liquid and solid materials, in bulk or thin film form, may be investigated. The set-up is particularly suitable for the study of the spin-lattice coupling in strongly correlated oxide materials utilizing several different stimuli, e.g. magnetic and electric fields, high pressure and low temperatures. Details for the Raman extension, sample holder assembly and optical design, as well as data acquisition and measurement routine are described. Finally, we present exemplary results collected using the set-up, measured on reference materials, as well as on a correlated transition metal oxide.

Keywords

Raman spectroscopy Quantum Design PPMS High pressure Ruby Hydrogen conversion Strongly correlated materials Langasite

Introduction

Large efforts in the field of material science are spent on the development of novel functional materials. In this context, strongly correlated electron systems are a class of materials that promise a multitude of applications. Several transition metal oxides are being considered for the designing of new devices, owing e.g. to their magnetic, electric and dielectric [1-3] or optical and photovoltaic properties [4,5]. In such materials, the functionalities or physical properties are determined by the localization and delocalization of charge carriers [6]. As a result, magnetic, (di)electric, thermal, and optical properties are correlated with each other [6]. Inelastic light scattering from electronic charge density and phonons in correlated materials is a promising approach to evaluate such correlation properties [7]. This approach is deemed to be particularly fruitful if it can be done under the simultaneous influence of several stimuli that probe these properties such as, e.g. magnetic and/or electric fields, pressure and/or temperature.

Raman spectroscopy for instance is a versatile technique for studying structural properties, lattice distortions, and e.g. the coupling of spin and lattice degrees of freedom. This is due to different types of excitations that can be studied under the influence of external stimuli leading to observation of shifts in vibrational frequencies, changes in peak intensities and polarization states. This was successfully employed to study structure-property relationships in strongly correlated transition metal oxides [8,9]. A particular strength of electronic Raman scattering is the possibility to detect information on the electron dynamics from different regions of the Brillouin zone.

A Raman set-up for in-situ measurements of such properties at low temperature, magnetic field, electrical field and high pressure was constructed as an extension to the Quantum Design, Inc. Physical Properties Measurement System (PPMS) [10] - a standard cryogenic device in many laboratories world wide. An optical probe and an automatized data acquisition system were designed and linked with the already automatized temperature and magnetic field control provided by the PPMS. Several material properties can thus be investigated in the same PPMS set-up, using the standard set of options such as resistivity, magnetization or heat capacity, as well as using extensions permitting e.g. magneto(di)electric measurements [11] or the here proposed Raman spectroscopy set-up. Dedicated custom-made devices for Raman spectroscopy with multiple excitations not using a Quantum Design PPMS were reported in literature previously [12,13].

We present results collected on different reference systems, such as ruby and H 2, as well as a transition metal oxide, Ba 3 NbFe 3 Si 2 O 14, a langasite material with magnetically induced electronic polarization.

Experimental apparatus and methods

PPMS host system and Raman extension

The Raman set-up is based on a commercial Quantum Design, Inc. PPMS [10] which incorporates (i) a cryostat providing a controllable variable temperature in the 2 - 400 K interval and (ii) a superconducting magnet up to 9 T (superconducting magnets up to 16 Tesla are available). The native hardware control of temperature and magnetic field provided by the PPMS was used and a new extension, permitting the collection of Raman spectroscopy data, was designed and fabricated. The Raman extension comprises the following parts: (a) The custom-made Raman top part (including several imaging lenses, beamsplitter, notch filter, white light illumination and CCD camera) and a modified PPMS Raman insert (using multifunction probe temperature sensor) serving as a sample/DAC holder. (b) A miniature nonmagnetic diamond anvil cell (DAC). (c) A 488 nm argon-ion CW laser system. (d) A Shamrock spectrometer with thermoelectrically cooled CCD detector.

The data acquisition, including Raman spectroscopic module (Solis, Andor Technology Ltd.), as well as software control of PPMS temperature and magnetic fields, was implemented using the QD PPMS third party option. A schematic view of the set-up is shown in Figure 1. It comprises a probe, akin to the commercially available multifunction probe (MFP), albeit mechanically more resistant, providing optical and electrical contacts to the sample, and an optical illumination/data collection system. A single sapphire optical window creates an interface between the evacuated interior of the magnet cryostat and the Raman top part.
Figure 1

Schematic view of the PPMS Raman set-up. BS - beamsplitter, F - optical filter, L - lens, NF - notch filter, and OW optical window.

Pictures of the upper optical part of the PPMS Raman extension, sample holder and diamond anvil pressure cell are shown in Figure 2(a)-(c), respectively. Magnetic field may be applied up the (+/-) maximum accessible value of the PPMS system. Electric field can also be applied, as indicated by the arrow in Figure 1, using an external voltage source. The high pressure conditions are created using miniature nonmagnetic diamond anvil cell (DAC), specially designed to hold pressure stable during temperature scans [14]. The DAC is made in a piston-cylinder configuration with a vertical optical axis and a maximum diameter of 10-13 mm. The maximal pressure for diamonds with 0.3 mm culets is 60 GPa. Our Raman extension is suitable for long-term experiments lasting several days, exploiting the typical operational parameter space of the PPMS.
Figure 2

Detailed view of the optical set-up. (a) Picture of the custom-made PPMS Raman top part. (b) Modified PPMS multi-function probe with a miniature diamond anvil cell. (c) Miniature diamond anvil cell with focusing/imaging lens mounted at the top.

Optical design

Raman spectra are recorded with a SR-303I-A Shamrock 303i spectrometer (Andor Technology) coupled by a single mode optical fiber to the PPMS Raman extension. The spectrometer is equipped with 1200l/mm grating and thermoelectrically cooled multichannel CCD detector (iDus, Andor Technology, 1024 ×256 pixels, -50°C). A linearly polarized argon ion laser (488 nm line) is used for the excitation at powers up to 15 mW. The spectral axis is calibrated by the fluorescence lines of a neon lamp. Inside the PPMS, Raman spectra are collected in the back scattering geometry, at a resolution of about 4 cm −1. Elastically backscattered fundamental light (488 nm) is filtered out using a narrow band notch filter (Semrock Inc.). The accuracy of the spectral measurements, resulting from the wavelength calibration procedure and experimental conditions, is estimated to be about 2 cm −1. Typical acquisition times vary between 20 - 120 sec. Experimentally accessible ranges of temperature, pressure, magnetic and electric fields available for spectroscopic acquisitions correspond to 3 - 350 K (higher temperatures with additional heating possible), 0 - 60 GPa, 0 - ±9 Tesla and 0 - ±10 MV/m, respectively. So far, spectral acquisitions involving strong electric fields have been carried out outside the PPMS only. Typically, the experimental arrangement involves application of hundreds of volts across a sample which is few tens of micrometers thick.

Alignment and measurement routine

Measurements are taken in the back scattering geometry. Due to space limitation inside the cryostat and low temperature conditions right angle scattering geometry is difficult to achieve in the described set-up. For sample illumination and focusing, white light is coupled via an optical fiber into the set-up and applied in transmission. Prior the measurements the optical system is pre-aligned and tested outside the cryostat using a room-temperature probe stand.

A standard (high-pressure) experiment is done in the following way. (i) The sample is mounted onto the Raman probe using ambient-pressure sample holder or DAC. For high-pressure experiments specific sample-, ruby- and pressure transmitting medium loading procedures have to be applied as described in Ref. [14]. (ii) The sample is aligned with the optical set-up and inserted into the PPMS cryostat. (iii) Single measurements are performed manually or an automatized measurement sequence is programmed using the PPMS third-party option together with the spectrometer software scripting option.

Exemplary results and discussion

Ruby fluorescence lines

Cr 3+-doped Al 2 O 3, also known as ruby, is a well-established standard material with interesting properties that led to significant technological advances both in academia and industry [15]. For instance, ruby was the material that led to the invention of the first laser by Maiman in 1960 [16]. The characteristic red color of ruby is related to electronic transitions involving the Cr 3+ ions manifested in the two fluorescence lines, R1 and R2 [17]. Ruby is furthermore a well-established pressure gauge in optical high-pressure experiments [18] and hence has been studied extensively. Zeeman splitting in single crystalline ruby has been observed in pulsed magnetic fields up to 60 T and hydrostatic pressure up to 10 GPa [19]. In the presented PPMS Raman set-up static magnetic fields up to 9 T (16 T with different cryostat) and hydrostatic pressure up to 60 GPa can be achieved. In Figure 3 the magnetic field dependence for the ruby R1 line at 5 K and 34 GPa is depicted. Ruby can hence serve as a local (in-situ) probe of the magnetic field (Δ λ/Δ H0.046 nm/T).
Figure 3

Exemplary measurements showing the magnetic field dependence of ruby (Cr 3+ -doped Al 2 O 3 ) fluorescence lines at 5 K and 34 GPa. Inset shows the shift of ruby fluorescence line with applied magnetic field.

Hydrogen conversion

Hydrogen is the most common element in the known universe. It exists in two different spin isomers, ortho hydrogen with two parallel proton spins and para hydrogen with two antiparallel proton spins. At room temperature and ambient pressure the ortho- to para hydrogen ratio is approximately 1:3. This equilibrium ratio is temperature dependent and at low temperatures the equilibrium state is para hydrogen. H 2 conversion, from ortho- to para hydrogen can be traced by measuring H 2 rotational modes (rotons), i.e. the integrated intensities of the rotational Raman peaks. Figure 4(a) shows H 2 conversion data collected using the PPMS Raman extension. The H 2 conversion rate is strongly enhanced under hydrostatic pressure [20]. In the presented PPMS Raman set-up the combined effect of hydrostatic pressure and high magnetic fields can be studied. The evolution of the rotational Raman modes (from 0 h (black line) to 8 h) at 20 K, 9 T magnetic field, and 27 GPa hydrostatic pressure is depicted. Vibrational Raman modes (vibrons) of hydrogen can be observed in a factor of 10 higher frequency range. Figure 4(b) illustrates the hydrogen Q 1(1) vibron at temperatures close to room temperature (300 K) and at low temperature (5 K) for both zero and 9 T magnetic fields. Hydrogen vibron modes are sensitive to structural changes and a clear difference between fluid and solid hydrogen is observed.
Figure 4

Exemplary measurements showing (a) the ortho-to-para conversion of molecular hydrogen at 20 K at applied magnetic field (9 T) and high pressure (27 GPa) and (b) the hydrogen vibron Q 1 (1) at 5 K and 300 K in zero and 9 T magnetic field.

Spin-phonon coupling in a strongly correlated oxide

Langasite materials are well known for their piezoelectric and non-linear optical properties [21,22]. A recent interest in the magnetic and electrical properties of langasite materials is due to the discovery that they could crystalize in a Kagomé-like lattice structure consisting of rare-earth and transition metal cations [23,24]. While a detailed report of the study of the spin-phonon coupling in those materials is beyond the scope of this article, we present here Raman spectroscopy data collected in our set-up as a function of temperature and magnetic fields. Simple qualitative analysis are also included, in order to illustrate the typical studies which may be performed on such materials.

Ba 3 NbFe 3 Si 2 O 14 crystallizes in the non-centrosymmetric trigonal P321 space group with 23 unit cell atoms. Magnetic Fe 3+ ions are arranged in triangle units on a triangular lattice [23]. The structure is non-polar and hence no spontaneous polarization can be expected for point group 32 [25]. It has been reported that Ba 3 NbFe 3 Si 2 O 14 undergoes an antiferromagnetic phase transition at approximately 27 K [23]. A much larger Curie-Weiss temperature of θ≈−175 K [26], reflects the large magnetic frustration in the system [27]. The magnetic transition is accompanied by a symmetry lowering from P321 to C2 [28]. Further details on the magnetic structure and Raman mode assignment can be found in literature [25,29,30].

Unpolarized Raman spectra on a single crystal of Ba 3 NbFe 3 Si 2 O 14 recorded as a function of temperature and magnetic field are shown in Figure 5. All T and H dependent measurements were done with an the incident beam along the [1-10] direction. For better visualization spectra (50 - 300 K) have been shifted vertically with a linear offset. The phonon spectrum of Ba 3 NbFe 3 Si 2 O 14 exhibits a frequency hardening and narrowing of the line-width on reducing temperature.
Figure 5

Exemplary measurements showing the magnetic field and temperature dependence of Raman spectrum of Ba 3 NbFe 3 Si 2 O 14 along the [1-10] direction. The black arrows indicate the 506 cm −1 phonon mode. (a) Ba 3 NbFe 3 Si 2 O 14 Raman spectra as a function of temperature in zero magnetic field. (b)+(c) Detail view showing the influence of magnetic fields to the 5 K Ba 3 NbFe 3 Si 2 O 14 Raman spectrum. (d) Frequency shift of the 506 cm −1 mode around the phase transition temperature T N in zero magnetic field. Dashed line indicates T N, extracted from magnetic and heat capacity measurements (not shown).

In a quasi-harmonic approximation, the frequency shift can be attributed to the volume change only and anharmonic effects due to phonon-phonon and spin-phonon interaction are ignored. Nevertheless, additional changes in the phonon frequencies when approaching T N are evident, as shown in Figure 5(d). For instance, the 506 cm −1 phonon mode reveals softening behavior with an onset above T N and a subsequent hardening below 25 K. Similar behavior is observed for the other modes, suggesting spin-phonon coupling. Figure 5(d) also illustrates precision and stability of the Raman set-up, which allow detection of a magnetic transition involving a change of only a fraction of wavenumber.

The magnetic field dependence of the unpolarized Raman spectra of Ba 3 NbFe 3 Si 2 O 14 at 5 K is shown in Figure 5(b)+(c). In this measurement the magnetic field was set to the highest value (9 T) and the first Raman spectrum was collected. Subsequently, the magnetic field was swept from 9 T to -9 T. At a first glance, a comparably ‘strong’ change in the integrated intensity of the 310, 416 and 506 cm −1 modes and in the position of the 615 cm −1 mode are found. Furthermore, several new modes appear at 250, 322, 760 and 840 cm −1 when the field is changed from 9 T to -9 T. The influence of the magnetic field on phonon modes 506, 573, 675 and 982 cm −1 is weak and only the 416 and 615 cm −1 modes exhibit major changes in the frequency [31].

Recently, the magnetic (and structural) properties of related langasites were found to be greatly affected by application of pressure [32], suggesting the relevance of T,P,H-dependent Raman studies on Ba 3 NbFe 3 Si 2 O 14 and other strongly correlated materials.

Conclusion

A Raman and luminescence spectroscopy system for the study of strongly correlated electron materials is presented. The system is designed as an extension of the Quantum Design, Inc. PPMS and allows for measurements in a large temperature range (3 - 350 K), high magnetic fields (0 - 9 T), high pressure (0 - 60 GPa) and electric field strengths up to 10 MV/m. Gaseous, liquid and solid materials, in amorphous, poly- and single-crystalline bulk form or as a thin film can be studied. Exemplary results for ruby and solid hydrogen at low temperature and high hydrostatic pressure and high magnetic field are shown. Finally results of the study of the spin-phonon coupling in the complex transition metal oxide Ba 3 NbFe 3 Si 2 O 14 are presented and discussed.

Notes

Declarations

Acknowledgements

We are very grateful to Dr. Y. Tokunaga, and Profs Y. Taguchi and Y. Tokura for advices and support in the floating zone growth of Ba 3 NbFe 3 Si 2 O 14 single-crystal. Financial support from the Carl Tryggers Stiftelse för Vetenskaplig Forskning (project CTS 09:217), the Swedish Foundation for International Cooperation in Research and Higher Education (STINT), the Swedish Research Council (VR) and the Göran Gustafsson Foundation is acknowledged. MH thanks the Swedish Research Council (VR, contract 2012-6562) for financial support and Profs M. Fiebig and O. Tjernberg for hosting him during his postdoc. AG thanks the Russian Foundation for Basic Researches (grant 14-02-00483-a), and the program of RAS “Physics of elementary particles, fundamental nuclear physics and nuclear technologies” for financial support.

Authors’ Affiliations

(1)
KTH Royal Institute of Technology, ICT Materials Physics
(2)
Department of Materials, ETH Zürich
(3)
Department of Earth Sciences, Uppsala University
(4)
Department of Engineering Sciences, Uppsala University
(5)
Institute of Crystallography, Russian Academy of Sciences
(6)
Institute for Nuclear Research, Russian Academy of Sciences
(7)
Geophysical Laboratory, Carnegie Institution of Washington

References

  1. Bibes M, Barthelemy A. Oxide Spintronics. IEEE Trans Elec Dev. 2007; 54(5):1003.View ArticleADSGoogle Scholar
  2. Scott JF. Applications of magnetoelectrics. J Mater Chem. 2012; 22:4567.View ArticleGoogle Scholar
  3. Choudhury D, Mandal P, Mathieu R, Hazarika A, Rajan S, Sundaresan A, et al.Near-Room-Temperature Colossal Magnetodielectricity and Multiglass Properties in Partially Disordered La 2 NiMnO 6. Phys Rev Lett. 2012; 108:127201.View ArticleADSGoogle Scholar
  4. Grinberg I, West DV, Torres M, Gou G, Stein DM, Wu L, et al.Perovskite oxides for visible-light-absorbing ferroelectric and photovoltaic materials. Nature. 2013; 503(7477):509–12.View ArticleADSGoogle Scholar
  5. Hosono H. Exploring electro-active functionality of transparent oxide materials. Jpn J Appl Phys. 2013; 52:090001.View ArticleADSGoogle Scholar
  6. Tokura Y, Nagaosa N. Orbital Physics in Transition-Metal Oxides. Science. 2000; 288:462.View ArticleADSGoogle Scholar
  7. Devereaux TP, Hackl R. Inelastic light scattering from correlated electrons. Rev Mod Phys. 2007; 79:175–233.View ArticleADSGoogle Scholar
  8. Guennou M, Bouvier P, Chen GS, Dkhil B, Haumont R, Garbarino G, et al.Multiple high-pressure phase transitions in BiFeO 3. Phys Rev B. 2011; 84:174107.View ArticleADSGoogle Scholar
  9. Iliev MN, Gospodinov MM, Singh MP, Meen J, Truong KD, Fournier P, et al.Growth, magnetic properties, and Raman scattering of La 2 NiMnO 6 single crystals. J Appl Phys. 2009; 106(2):023515.View ArticleADSGoogle Scholar
  10. Quantum Design Inc.Physical Property Measurement System (PPMS). http://www.qdusa.com/products/ppms.html 2015 Accessed 6 Feb 2015.
  11. Hudl M, Mathieu R, Ivanov SA, Weil M, Carolus V, Lottermoser T, et al.Complex magnetism and magnetic-field-driven electrical polarization of Co 3 TeO 6. Phys Rev B. 2011; 84:180404.View ArticleADSGoogle Scholar
  12. Isaacs ED, Heiman D. Fiber optics for Raman scattering at low temperatures and high-magnetic fields. Rev Sci Instrum. 1987; 58(9):1672–4.View ArticleADSGoogle Scholar
  13. Matsuda Y, Kuroda N, Nishina Y. High-field, high-pressure, and low-temperature magneto-optical apparatus using a diamond anvil cell. Rev Sci Instrum. 1992; 63(12):5764–6.View ArticleADSGoogle Scholar
  14. Gavriliuk AG, Mironovich AA, Struzhkin VV. Miniature diamond anvil cell for broad range of high pressure measurements. Rev. Sci. Instrum. 2009; 80:043906.View ArticleADSGoogle Scholar
  15. Syassen K. Ruby under pressure. High Pressure Res. 2008; 28(2):75–126.View ArticleADSGoogle Scholar
  16. Maiman TH. Stimulated Optical Radiation in Ruby. Nature. 1960; 187:493–94.View ArticleADSGoogle Scholar
  17. García-Lastra JM, Barriuso MT, Aramburu JA, Moreno M. Origin of the different color of ruby and emerald. Phys. Rev. B. 2005; 72:113104.View ArticleADSGoogle Scholar
  18. Mao HK, Xu J, Bell PM. Calibration of the Ruby pressure gauge to 800 kBar under quasi-hydrostatic conditions. J Geophys Res-Solid. 1986; 91(B5):4673–6.View ArticleADSGoogle Scholar
  19. Millot M, Broto J-M, Gonzalez J. High-field Zeeman and Paschen-Back effects at high pressure in oriented ruby. Phys Rev B. 2008; 78:155125.View ArticleADSGoogle Scholar
  20. Eggert JH, Karmon E, Hemley RJ, Mao HK, Goncharov AF. Pressure-enhanced ortho-para conversion in solid hydrogen up to 58 GPa. Proc Nat Acad Sci USA. 1999; 96(22):12269–72.View ArticleADSGoogle Scholar
  21. Kaminskii AA, Mill BV, Khodzhabagyan GG, Konstantinova AF, Okorochkov AI, Silvestrova IM. Investigation of trigonal (La 1−x Nd x ) 3 Ga 5 SiO 14 crystals. Phys Stat Sol (A). 1983; 80:387.View ArticleADSGoogle Scholar
  22. Sato J, Takeda H, Morikoshi H, Shimamura K, Rudolph P, Fukuda T. Czochralski growth of RE 3 Ga 5 SiO 14 (RE = La, Pr, Nd) single crystals for the analysis of the influence of rare earth substitution on piezoelectricity. J Cryst Growth. 1998; 191:746.View ArticleADSGoogle Scholar
  23. Marty K, Simonet V, Ressouche E, Ballou R, Lejay P, Bordet P. Single domain magnetic helicity and triangular chirality in structurally enantiopure Ba 3 NbFe 3 Si 2 O 14. Phys Rev Lett. 2008; 101(24):247201.View ArticleADSGoogle Scholar
  24. Zhou HD, Lumata LL, Kuhns PL, Reyes AP, Choi ES, Dalal NS, et al.Ba 3 NbFe 3 Si 2 O 14: A New Multiferroic with a 2D Triangular Fe 3+ Motif. Chem Mater. 2009; 21(1):156–9.View ArticleGoogle Scholar
  25. Birss RR. Symmetry and Magnetism: North-Holland Publishing Company - Amsterdam; 1966.Google Scholar
  26. Marty K, Bordet P, Simonet V, Loire M, Ballou R, Darie C, et al.Magnetic and dielectric properties in the langasite-type compounds: A(3)BFe(3)D(2)O(14) (A = Ba, Sr, Ca; B = Ta, Nb, Sb; D = Ge, Si). Phys Rev B. 2010; 81:054416.View ArticleADSGoogle Scholar
  27. Greedan JE. Geometrically frustrated magnetic materials. J Mater Chem. 2001; 11:37–53.View ArticleGoogle Scholar
  28. Lyubutin IS, Naumov PG, Mill’ BV, Frolov KV, Demikhov EI. Structural and magnetic properties of the iron-containing langasite family A 3 M Fe 3 X 2 O 14 (A= Ba, Sr; M= Sb, Nb, Ta; X= Si, Ge) observed by Mössbauer spectroscopy. Phys Rev B. 2011; 84:214425.View ArticleADSGoogle Scholar
  29. Li Y, Lu G, Yang H, Lan J, Zhang J, Huang W, et al.Lattice vibration of Sr 3 TaGa 3 Si 2 O 14 single crystal. Phys Stat Solidi (b). 2007; 244(2):518–28.View ArticleADSGoogle Scholar
  30. Kroumova E, Aroyo MI, Perez-Mato JM, Kirov A, Capillas C, Ivantchev S, et al.Bilbao crystallographic server: Useful databases and tools for phase-transition studies. Phase Transit. 2003; 76(1-2):155–70.View ArticleGoogle Scholar
  31. Hudl M. Magnetic materials with tunable thermal, electrical, and dynamic properties: An experimental study of magnetocaloric, multiferroic, and spin-glass materials: Uppsala: Uppsala Dissertations from the Faculty of Science and Technology 98; 2012.Google Scholar
  32. Gavriliuk AG, Lyubutin IS, Starchikov SS, Mironovich AA, Ovchinnikov SG, Trojan IA, et al.The magnetic P-T phase diagram of langasite Ba 3 TaFe 3 Si 2 O 14 at high hydrostatic pressures up to 38 GPa.Appl Phys Lett. 2013; 103(16).Google Scholar

Copyright

© Hudl et al.; licensee Springer. 2015

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.