Article | Open

Unravelling the hidden link of lithium halides and application in the synthesis of organocuprates

  • Nature Communications 8, Article number: 14794 (2017)
  • doi:10.1038/ncomms14794
  • Download Citation
Published online:


As a versatile metal, copper has demonstrated a wide application in acting as both organometallic reagent and catalyst. Organocuprates are among the most used organometallic reagents in the formation of new carbon–carbon bonds in organic synthesis. Therefore, revealing the real structures of organocuprates in solution is crucial to provide insights into the reactivity of organocuprates. Here we provide several important insights into organocuprate chemistry. The main finding contains the following aspects. The Cu(0) particles were detected via the reduction of CuX by nBuLi or PhLi. The Cu(II) precursors CuX2 (X=Cl, Br) could be used for the preparation of Gilman reagents. In addition, we provide direct evidence for the role and effect of LiX in organocuprate synthesis. Moreover, the EXAFS spectrum provides direct evidence for the exact structure of Li+ CuX2 ate complex in solution. This work not only sheds important light on the role of LiX in the formation of organocuprates but also reports two new routes for organocuprate synthesis.


Since the pioneering work of Gilman et al.,1 organocuprates have been widely employed as organometallic reagents in organic synthesis (including conjugate additions, the opening of epoxides and cross-coupling reactions)2,3,4,5,6,7. In the textbook, organocuprates are usually prepared through transmetalation of lithium, magnesium or zinc organometallics with Cu(I) salts8. Different coordination environments always drastically affect the reactivity or stabilities of organocuprates9. Up to now, a lot of synthetic methodologies involving organocuprate reagents have been developed, while great uncertainty still exists in the related mechanism10. Although several important crystal structures of organocuprates were reported11,12,13,14, it should be noted that solid-state structures often reflect the most thermodynamically stable species and are not necessarily the same as in solution state. Besides, organocuprates can exhibit complex behaviour in solution, often existing as a number of different species in equilibrium, thus further complicating their characterization. For that reason, the structure of organocopper compounds in solution cannot be inferred directly from crystal structures and must be determined independently.

The structures of organocuprate reagents in an ethereal solution have received wide attention, because they are strongly relevant to reactivity in real reaction conditions15,16,17. Nuclear magnetic resonance18,19,20 and electrospray ionization–mass spectrometry21,22 served as powerful tools and have been widely used in determining the structures of organocuprates in solution. The linear bonding geometry of the C–Cu–C moiety in cuprates such as MeCu(CN)Li, Me2CuLi and Me2-CuLi3LiX (X=I, CN) has been well established. In 1996, Knochel and colleagues23,24 first introduced the extended X-ray absorption fine structure (EXAFS) to study the local structure of organocuprates from the reaction between CuCN and nBuLi. EXAFS spectroscopy provides a unique probe of the local structural environment of metal ions in non-crystalline systems25,26,27,28,29,30,31,32,33. The preliminary structure for lithium cyanocuprates based on EXAFS data has been elucidated. However, the role of cyanide and the difference between cyanide and other halide atoms still remain in debate9. Lipshutz et al.21 and Koszinowski and colleagues34 have pointed out the LiX could have a positive effect on the solubility of CuX (X=I, Br, Cl, CN) independently. The electrospray ionization–mass spectrometry was used to study the structure of formed ate complex21,34. However, determination of the exact structure, the role for LiX and application in organocuprates have still been not well-studied up to date. We started our research by investigating the effect of anion on organocuprates preparation. Here we show the anion effect of different Cu(I) precursors on Gilman reagent preparation. The EXAFS reveals that the LiX (X=Br, Cl) serves as the hidden link for organocuprates preparation from unfavoured CuX. A soluble cupric bromide anion intermediate is evidenced by EXAFS when adding LiBr to CuBr in tetrahydrofuran (THF). This CuX2 Li+ ate complex serves as a key intermediate in the generation of Gilman reagent (Fig. 1). In addition, we also shed two other important findings in this work. First, the detection of copper nanoparticles produced after the addition of nBuLi or PhLi to CuX. Second, the Cu(II) precursors CuX2 (X=Cl, Br) can be used for the preparation of Gilman reagents.

Figure 1: Hidden link of lithium halides.
Figure 1

Scheme of the role of LiX (X=Br, Cl) and application in Gilman reagent preparation.


Detection of Cu(0) particles via the reduction of CuX

In organic synthesis, different Cu(I) precursors are always applied in different reaction systems10. Initially, we investigated different cuprous salts with excess nBuLi in THF under −78 °C for organocuprates synthesis. From X-ray absorption near-edge spectroscopy (XANES) spectra (Fig. 2a), we observed the difference of reactions from CuCN and CuX (X=Br, Cl). In Fourier-transformed EXAFS spectra, an obvious copper nanoparticle feature at high shells in CuBr and CuCl complexes appeared (Fig. 2b). However, such peaks at 3.4, 4.1 and 4.8 Å were not detected in the CuCN system, which is accordance with previous literature23 that CuCN is a good precursor to Gilman reagent. In addition, such results also indicate that organocuprates made by CuBr or CuCl are very unstable to decompose into zero valance copper nanoparticle. Meanwhile, it seemed that the smaller the anion atom is, the more Cu(0) particle is formed.

Figure 2: XAFS analysis of CuX reduction.
Figure 2

(a) XANES spectra of various cuprous salts with excess nBuLi in THF under −78 °C. (b) Comparison of FT magnitudes of k2-weighted EXAFS of various Cu(I) species mixed with excess nBuLi in THF under −78 °C. (3.0 Å−1<k<12.4 Å−1).

Then, we also used X-ray absorption spectroscopy (XAS) to study the reaction between CuX (X=Br, Cl) and PhLi. The XANES spectra were shown in Supplementary Fig. 1. From the EXAFS spectra in Fig. 3a, we found that the mixture of Cu (0) and Cu (I) species was formed when mixing CuBr or CuCl with PhLi. The CuCl was easier to be reduced to Cu(0) particle than CuBr by PhLi, which was consistent with the reaction with nBuLi. In addition, we also investigated the solvent effect on this process. We found that the reduction process was even faster when using ethyl ether (Et2O) as the solvent (Supplementary Figs 2 and 3). To further evidence the Cu(0) species and this reduction process, X-ray powder diffraction experiments were performed and the results are shown in Fig. 3b. The figure shows the main existence of Cu in the reaction between CuCl and nBuLi in THF or Et2O, in which the three peaks at 43.3°, 50.4° and 74.1° are corresponding to the (111), (200) and (220) planes of Cu (JCPDS number 04-0836), respectively.

Figure 3: EXAFS and X-ray powder diffraction analysis.
Figure 3

(a) EXAFS spectra of of CuBr and CuCl mixed with excess PhLi in THF under −78 °C. (b) X-ray powder diffraction experiments, blue line: CuCl+5.0 equiv nBuLi in Et2O, black line: CuCl+5.0 equiv nBuLi in THF.

Reduction of Cu(II) precursors

As the nBuLi could serve as a reductant to reduce CuX (X=Br, Cl) to Cu(0) species, we also employed EXAFS to investigate the reactions between Cu(II) salts and nBuLi. We are very excited to discover that instead of using Cu(I) as the starting reagent, more stable and cheaper Cu(II) halide salts could also be good Gilman reagent precursors in the presence of excess organolithium reagent. EXAFS provided us a direct view of these transformations. Compared with traditional Gilman reagent prepared from CuCN, we could see that in the presence of 5.0 equivalent of nBuLi in THF under −78 °C, both CuBr2 and CuCl2 were reduced into Cu(I) with edge energies about 8979.9 and 8980.0 eV, respectively, in the XANES spectra (Fig. 4a). The PhLi could also reduce CuBr2 to corresponding Cu(I) species (Fig. 4b). The fitting result further verified the existence of 2-coordinated [C-Cu(I)-C] short-range structure (Fig. 4c). Owing to the fact that CuX2 is stable and easy to store, this method will be a new route for organocuprates preparation from CuX2. We also applied this method to organic synthesis. When adding the electrophile such as benzyl bromide or (2-bromoethyl)benzene into the reaction system, the desired C–C bond formation was formed, which implied that this method for organocuprates could be used for cross-coupling reactions. The detail application of this method in complicated molecular synthesis is on the way in our lab.

Figure 4: XAFS analysis of CuX2 reduction.
Figure 4

(a) XANES spectra of CuCN, CuBr2 and CuCl2 mixed with excess nBuLi in THF under −78 °C. (b) XANES spectra of CuBr2 mixed with excess PhLi in THF under −78 °C. (c) Fitting result for CuBr2+5.0 equiv nBuLi in THF solution (2.910 Å−1<k<11.472 Å−1 and 1.065 Å<R<2.127 Å).

Role of lithium halides in the synthesis of organocuprates

The lithium salts may have an effect on the formation of organocuprates, which has been awaked by several groups21,34,35,36,37,38,39. Although knowing this phenomenon for a long time, determination of the exact structure, the role for LiX and application in organocuprates have still been not well researched. When one equivalent LiBr was added to the mixture of CuBr and 5 equivalent nBuLi, we obtained a similar XANES spectra of which obtained from CuCN and excess nBuLi under low temperature (Fig. 5a). The edge energy was determined as 8979.7 and 8979.9 eV, respectively. Furthermore, high shell features for copper nanoparticle disappeared in the EXAFS spectrum, which indicates the formation of relatively pure organocopper compound (Fig. 5b). The results of fitting show two carbon atoms at 1.94 Å around the copper atom equally (Fig. 5c). Thus, we developed a new method of preparing Gilman reagents using CuX with the help of LiX, which is very similar to what we get from mostly used cuprous cyanide in traditional synthesis route.

Figure 5: Role of lithium halides in the synthesis of organocuprates.
Figure 5

(a) XANES spectra of CuCN (black), CuBr+1.0 eq LiBr (red) and CuBr2 (blue) with excess nBuLi in THF under −78 °C. (b) EXAFS spectra of CuCN (black), CuBr+1 eq LiBr (red) and CuBr2 (blue) with excess nBuLi in THF under −78 °C. (c) Fitting results of R-space k2-weighted EXAFS spectra of CuBr+1 equiv LiBr in 5.0 equiv nBuLi THF solution (2.957 Å−1<k<11.145 Å−1 and 1.172 Å<R<1.914 Å).

To further elucidate the role of LiX in preparing Gilman reagent, the mixture of LiBr and CuBr in THF was used for demonstration. As shown in Supplementary Fig. 4, the CuBr species alone look polymer-like and very insoluble in THF. In contrast, CuBr can be dissolved in THF with the aid of one equivalent quantity of LiBr. A green solution is quickly formed after adding 1 equivalent LiBr into the system. This result indicated that the CuBr has an interaction with LiBr and a new copper species is possibly formed. The interaction between CuBr and LiBr might be the key to stabilizing the Gilman reagents prepared through this method. Valance alternation in cuprous bromide was invisible in the presence of LiBr from XANES spectra (Fig. 6a, edge energy of 8980.9 eV). In addition, EXAFS fitting results indicate two bromine atoms coordinated to the copper(I) centre (Fig. 6b). Thus, we claim that this type of ate complex was a [Br-Cu-Br] Li+ anion. This [CuBr2] ate complex shows good solubility and serves as a key intermediate in the generation of Gilman reagent.

Figure 6: XAFS analysis of reaction between CuX and LiX.
Figure 6

(a) XANES spectra of CuBr+LiBr in THF species. (b) Fitting results of k2-weighted R-space EXAFS spectra of CuBr+1.0 equiv LiBr (2.890 Å−1<k<12.134 Å−1 and 1.487 Å<R<2.376 Å).

Density functional theory calculations

Density functional theory calculation was also performed to provide support for the EXAFS fitting results. As shown in Fig. 7a, the coordination of THF to [Br-Cu-Br] anion is endergonic by 6.6 kcal mol−1. The optimized structure suggests that the Cu–O distance is 3.62 Å, which means the interaction between Cu and O is very weak. As previous reports, the monomer state of organocuprates was always present in more polar solvent such as THF. Thus, the monomer structure of organocuprate was calculated. Meanwhile, the coordination of THF to [nBu-Cu-nBu] anion is found to be endergonic by 9.9 kcal kcal mol−1 and the corresponding Cu–O distance is determined to be 4.43 Å (Fig. 7b). Thus, the coordination of THF to [Br-Cu-Br] and [nBu-Cu-nBu] anion are both energetically unfavourable19,20,40. Moreover, optimized structures reveal that the bond length of Cu–Br in [Br-Cu-Br] anion is 2.27 Å and the bond length of Cu–C in [nBu-Cu-nBu] anion is 1.97 Å, which are very close to that obtained from EXAFS (2.24 and 1.94 Å ). Consequently, the theoretical study and EXAFS fitting results have reached the same conclusion.

Figure 7: Density functional theory calculation.
Figure 7

Density functional theory calculation for the coordination of THF to (a) [Br-Cu-Br] anion and (b) [nBu-Cu-nBu] anion.


To sum up, we elucidate LiX (X=Br, Cl) as a hidden link in the preparation of organocuprate reagents from CuX (X=Br, Cl) with a key intermediate CuX2 Li+ ate complex evidenced by EXAFS. Meanwhile, this Cu(I) ate complex can serve as a good precursor to prepare Gilman reagents following a tandem process. In addition, we also developed the organocuprate reagents synthesis from Cu(II) precursors. This discovery might help open a new perspective in understanding the organocopper chemistry and mechanisms of copper-catalysed reactions as well.


General information

X-ray absorption measurements were acquired in transmission mode at beamline 17C1 at National Synchrotron Radiation Research Center in Taiwan. A pure Cu foil spectrum (edge energy 8979, eV) was acquired simultaneously with each measurement for energy calibration. Multiple scans were taken to reduce the noise. The Supplementary Tables 1–3 revealed the detailed parameters of the XAFS spectral.

Reaction system

Cu salt (0.5 mmol) was added to the schlenk tube cell in a glovebox beforehand. Then, 5.0 ml of THF was injected into the cell and the solution was stirred under N2 at −78 °C for 20 min. Subsequently, RLi (2.5 mmol) was added into the system and stirred for 30 min. As the last step, the liquid nitrogen was quickly added to reaction system, which would be frozen into solid immediately, and it was transferred into the XAFS cell with the protection of nitrogen gas.

Detection system (beamline)

The detection system was cooled using cooled nitrogen gas. The Supplementary Fig. 5 was the picture of cell holder used in the beamline. This hold connected with a liquid nitrogen cooled gas stream. The Supplementary Fig. 6 showed our idea for low-temperature system. We used a gas stream passing through the liquid nitrogen Dewar to cool the system. The temperature could be controlled by the tuning of the flow rate. The Supplementary Fig. 7 was the whole picture of experimental set-up in beamline.

Data availability

Data supporting the findings of this study are available within this article and its Supplementary Information file and from the corresponding authors on reasonable request.

Additional information

How to cite this article: Yi, H. et al. Unravelling the hidden link of lithium halides and application in the synthesis of organocuprates. Nat. Commun. 8, 14794 doi: 10.1038/ncomms14794 (2017).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    , & The preparation of methylcopper and some observations on the decomposition of organocopper compounds. J. Org. Chem. 17, 1630–1634 (1952).

  2. 2.

    , & The chemistry of carbanions. XII. The role of copper in the conjugate addition of organometallic reagents1. J. Org. Chem. 31, 3128–3141 (1966).

  3. 3.

    , & Enantioselective and enantiospecific transition-metal-catalyzed cross-coupling reactions of organometallic reagents to construct C-C bonds. Chem. Rev. 115, 9587–9652 (2015).

  4. 4.

    & The oxidation of organocuprates-an offbeat strategy for synthesis. Chem. Soc. Rev. 35, 218–225 (2006).

  5. 5.

    , , , & Reaction of lithium dialkyl- and diarylcuprates with organic halides. J. Am. Chem. Soc. 91, 4871 (1969).

  6. 6.

    & Regio- and stereoselective syntheses with organocopper reagents. Angew. Chem. Int. Ed. 36, 186 (1997).

  7. 7.

    & Selective formation of carbon-carbon bonds between unlike groups using organocopper reagents. J. Am. Chem. Soc. 89, 3911 (1967).

  8. 8.

    & Preparation and reactions of polyfunctional organozinc reagents in organic synthesis. Chem. Rev. 93, 2117 (1993).

  9. 9.

    & Elongated Gilman cuprates: the key to different reactivities of cyano- and iodocuprates. J. Am. Chem. Soc. 136, 5765–5772 (2014).

  10. 10.

    & Mechanisms of nucleophilic organocopper(I) reactions. Chem. Rev. 112, 2339–2372 (2012).

  11. 11.

    et al. Electron transfer from trialkyltin radicals to nitrosugars: the synthesis of C-glycosides with tertiary anomeric carbon atoms. J. Am. Chem. Soc. 107, 4332 (1985).

  12. 12.

    & Isolation and first structural characterization of dimethyl sulfide solvates of phenyllithium, phenylcopper, and lower and higher order lithium phenylcuprate reagents. J. Am. Chem. Soc. 112, 8008 (1990).

  13. 13.

    The structures of lithium and magnesium organocuprates and related species. Coordin. Chem. Rev. 255, 1226 (2011).

  14. 14.

    , , & The X-ray crystal structure of a cuprate-carbonyl pi-complex. Angew. Chem. Int. Ed. 52, 10250–10252 (2013).

  15. 15.

    , & An electrospray ionization mass spectrometry study of the aggregation states of organocopper complexes in solution. Organometallics 18, 1571 (1999).

  16. 16.

    New copper chemistry. 20. Lithium cyanocuprates, RCu(CN)Li: first observation of two-bond carbon-13-carbon-13 NMR couplings in organocuprates. J. Am. Chem. Soc. 113, 5470 (1991).

  17. 17.

    , & Tetraalkylcuprates(III): formation, association, and intrinsic reactivity. J. Am. Chem. Soc. 134, 613–622 (2012).

  18. 18.

    , & Higher order, mixed organocuprates. 12. Chemical and spectroscopic observations on the properties and composition of higher order, mixed organocuprates, R2Cu(CN)Li2. J. Org. Chem. 49, 3943 (1984).

  19. 19.

    , , , & Me2CuLi*LiCN in diethyl ether prefers a homodimeric core structure [Me2CuLi]2 and not a heterodimeric one [Me2CuLi*LiCN]: 1H,6Li HOE and1H,1H NOE studies by NMR. J. Am. Chem. Soc. 123, 7299–7304 (2001).

  20. 20.

    Organocuprates and diamagnetic copper complexes: structures and nmr spectroscopic structure elucidation in solution. Chem. Rev. 108, 3029–3053 (2008).

  21. 21.

    , , , & Analyses of anioniccu(I) complexesviaelectrospray mass spectrometry. J. Am. Chem. Soc. 118, 6796 (1996).

  22. 22.

    & Association and dissociation of lithium cyanocuprates in ethereal solvents. Organometallics 30, 4771 (2011).

  23. 23.

    , & Structural characterization of organocopper reagents by EXAFS spectroscopy. J. Am. Chem. Soc. 115, 348 (1993).

  24. 24.

    et al. Structural characterization of organocuprate reagents. EXAFS spectroscopy and ab initio calculations. J. Am. Chem. Soc. 117, 12489 (1995).

  25. 25.

    & X-Ray Absorption: Principles, Applications, Techniques of EXAFS, SEXAFS, and XANES Wiley (1988).

  26. 26.

    & An introduction to X-ray absorption spectroscopy and its in situ application to organometallic compounds and homogeneous catalysts. Catal. Sci. Technol. 2, 461 (2012).

  27. 27.

    et al. Direct observation of reduction of Cu(II) to Cu(I) by terminal alkynes. J. Am. Chem. Soc. 136, 924–926 (2014).

  28. 28.

    et al. Evidence of Cu(I)/Cu(II) redox process by X-ray absorption and EPR spectroscopy: direct synthesis of dihydrofurans from beta-ketocarbonyl derivatives and olefins. Chem. Eur. J. 21, 18925–18929 (2015).

  29. 29.

    et al. Operando X-ray absorption and EPR evidence for a single electron redox process in copper catalysis. Chem. Sci. 6, 4851 (2015).

  30. 30.

    , , & Selective alcohol oxidation by a copper TEMPO catalyst: mechanistic insights by simultaneously coupled operando EPR/UV-Vis/ATR-IR spectroscopy. Angew. Chem. Int. Ed. 54, 11791–11794 (2015).

  31. 31.

    , , & Iron(0) particles: catalytic hydrogenations and spectroscopic studies. ChemCatChem. 4, 1088 (2012).

  32. 32.

    & X-Ray absorption in homogeneous catalysis research: the iron-catalyzed Michael addition reaction by XAS, RIXS and multi-dimensional spectroscopy. Phys. Chem. Chem. Phys. 12, 5575–5584 (2010).

  33. 33.

    , , & X-ray spectroscopic verification of the active species in iron-catalyzed cross-coupling reactions. Chem. Eur. J. 19, 15816–15821 (2013).

  34. 34.

    , & Ionic aggregates of lithium organocuprates. Chem. Eur. J. 19, 10992–10999 (2013).

  35. 35.

    , & Chemical and spectroscopic studies on copper iodide derived organocuprates: new insight into the composition of Gilman’s reagent. J. Am. Chem. Soc. 107, 3197 (1985).

  36. 36.

    , , & Efficient synthesis of functionalized organozinc compounds by the direct insertion of zinc into organic iodides and bromides. Angew. Chem. Int. Ed. 45, 6040–6044 (2006).

  37. 37.

    , , & New copper chemistry. 23. Preparation of ethereal lithium dimethylcuprates (Me2CuLi)2 and Me2CuLi.cntdot.LiI displaying narrow line width 13C NMR resonances. Organometallics 14, 1213 (1995).

  38. 38.

    , , & Role of LiCl in generating soluble organozinc reagents. J. Am. Chem. Soc. 138, 11156–11159 (2016).

  39. 39.

    , & Theoretical studies on SN2-reaction of MeBr with Me2CuLi·LiCl. Solvent and cluster effects on oxidative addition/reductive elimination pathway. J. Am. Chem. Soc. 120, 8273 (1998).

  40. 40.

    et al. The relation between ion pair structures and reactivities of lithium cuprates. Chem. Eur. J. 6, 3060–3068 (2000).

Download references


This work was supported by the 973 Program (2012CB725302), the National Natural Science Foundation of China (21390400, 21520102003, 21272180 and 21302148), the Hubei Province Natural Science Foundation of China (2013CFA081), the Research Fund for the Doctoral Program of Higher Education of China (20120141130002) and the Ministry of Science and Technology of China (2012YQ120060). The Program of Introducing Talents of Discipline to Universities of China (111 Program) is also appreciated. The X-ray absorption spectroscopy studies were carried out at the beamline 17C1 of the National Synchrotron Radiation Research Center (NSRRC) in Taiwan. We thank Ms Z. Alaji and Dr Atul K. Singh for kind discussions.

Author information


  1. College of Chemistry and Molecular Sciences, The Institute for Advanced Studies (IAS), Wuhan University, Wuhan, Hubei 430072, China

    • Hong Yi
    • , Dali Yang
    • , Jie Xin
    • , Yi Deng
    •  & Aiwen Lei
  2. School of Chemistry and Chemical Engineering, Chongqing University, Chongqing 400030, China

    • Xiaotian Qi
    •  & Yu Lan
  3. National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan

    • Chih-Wen Pao
    •  & Jyh-Fu Lee
  4. State Key Laboratory of Organometallic Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai 200032, China

    • Aiwen Lei


  1. Search for Hong Yi in:

  2. Search for Dali Yang in:

  3. Search for Jie Xin in:

  4. Search for Xiaotian Qi in:

  5. Search for Yu Lan in:

  6. Search for Yi Deng in:

  7. Search for Chih-Wen Pao in:

  8. Search for Jyh-Fu Lee in:

  9. Search for Aiwen Lei in:


H.Y. and A.L. conceived the project. H.Y., D.Y., X.J., D.Y., C.-W.P. and J.-F.L. performed the experiments, analysed the data and discussed the results. X.Q. and Y.L. performed the density functional theory calculations. H.Y., D.Y. and A.L. wrote the paper. H.Y. and D.Y. wrote the Supplementary Information and contributed other related materials. H.Y. and D.Y. contributed equally to this work. All the authors checked the manuscript.

Competing interests

The authors declare no competing financial interests.

Corresponding authors

Correspondence to Yi Deng or Aiwen Lei.

Supplementary information

PDF files

  1. 1.

    Supplementary Information

    Supplementary figures, supplementary tables, supplementary methods and supplementary references.


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Creative Commons BYThis work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit