Chemical Physics Letters xxx (2016) xxx–xxx Contents lists available at ScienceDirect Chemical Physics Letters journal homepage: www.elsevier.com/locate/cplett Research paper The prospect of sensitizing organic dyes attached to the MoS2 surface: Physical insights from density functional theory investigations Hung M Le a,b,⇑, Viet Q Bui c, Phuong Hoang Tran d, Nguyen-Nguyen Pham-Tran d, Yoshiyuki Kawazoe e, Duc Nguyen-Manh f a Computational Chemistry Research Group, Ton Duc Thang University, Ho Chi Minh City, Viet Nam Faculty of Applied Sciences, Ton Duc Thang University, Ho Chi Minh City, Viet Nam Department of Applied Chemistry and Institute of Natural Sciences, Kyung Hee University, Gyeonggi-do 446-701, South Korea d Faculty of Chemistry, University of Science, Vietnam National University, Ho Chi Minh City, Viet Nam e New Industry Creation Hatchery Center, Tohoku University, Sendai 980-8579, Japan f Theory and Modeling Department, Culham Centre for Fusion Energy, United Kingdom Atomic Energy Authority, Abingdon OX14 3DB, United Kingdom b c a r t i c l e i n f o Article history: Received 26 September 2016 In final form November 2016 Available online xxxx Keywords: Organic dye MoS2 DFT a b s t r a c t In this theoretical study, we employ first-principles calculations to explore the bonding nature of organic dyes on the semiconducting MoS2 surface To achieve good bonding interaction and charge transfer, the ACOOÅ residue needs to form ionic bonds with the defected MoS2 surface In the cases of L0 and a newly synthesized dye named as TN1, we observe the manifestation of an in-gap state at À1 eV from the Fermi level, which might enhance photon trapping capability of the complex Ó 2016 Published by Elsevier B.V Introduction Molybdenum disulfide, a two-dimensional (2D) material with multi-layer stacking of MoS2, has been known to possess applications in several aspects In industry, this material is well-known as a lubricant because of the weak van der Waals’ interactions between layers, which thereby produces a low friction coefficient [1] Interestingly enough, such weak van der Waals interactions have a significant influence on the electronic property of MoS2 and its close derivative in the family, WS2 Adopting angleresolved photoelectron spectroscopy and first-principles calculations, Klein et al [2] showed that in the multi-layer form, the material would possess an indirect band gap of around 1.2 eV, while the standalone single layer establishes a larger direct band gap of $1.8 eV Besides, its novel catalytic capability also guarantees the wide applications of MoS2 as a participating catalyst in hydrodesulfurization for petroleum refinery [3,4] and water splitting for hydrogen production [5] In fact, for a period of time, MoS2 had been considered as an inert material This traditional belief is no longer true until the successful synthesis of highlyreactive anionic [Mo3S13]2À nanoparticles [6] ⇑ Corresponding author E-mail address: leminhhung@tdt.edu.vn (H.M Le) The single layer form of MoS2 finally finds its position in electronics due to the successful synthesis of highly qualitative mono-crystalline layers [7] Not only integrated into functional electronic devices, such a material with a direct band-gap can be employed in phototransistors with high sensitivity and low noise [8] The functionalization of the MoS2 layer have attracted much attention from the research community because of its promising applications in electronics, energy storage, sensing, and catalysis [9] The covalent functionalization of MoS2 was previously discussed by Presolski and Pumera [10] Recently, Chen et al [11] demonstrated a functionalization of exfoliated 2H-MoS2 with cysteine, an organic thiol, and the results showed physisorption rather than covalent attachment Using a first-principles approach, Ataca and Ciraci proposed the attachment of adatom and vacancy creation, which consequently caused MoS2 to gain a net magnetic moment [12] Due to the difficulty of functionalizing pristine MoS2, a new strategy was proposed to attach the acetate group on the surface by employing transition metal bridges [13] In the storyline of photo-sensitivity, there have been two remarkable efforts to tailor the performance of MoS2 in photocatalysis [5] and photodetector [14] In those studies, organic structures, being employed as ‘sensitizing dyes’ and possessing compatible photosensitivity with the heterogeneous layer, are employed to decorate the surface of MoS2, and dedicate an essential role in ‘trapping’ photoexcitations In the content of this study, we demonstrate a http://dx.doi.org/10.1016/j.cplett.2016.11.007 0009-2614/Ó 2016 Published by Elsevier B.V Please cite this article in press as: H.M Le et al., Chem Phys Lett (2016), http://dx.doi.org/10.1016/j.cplett.2016.11.007 H.M Le et al / Chemical Physics Letters xxx (2016) xxx–xxx theoretical investigation of organic dye attached to an MoS2 surface to get more insights of the binding nature and in-gap occupations The organic dyes of interest consist of the well-known L0 structures [15,16] and one in-house-factorized dye However, prior to investigating the interactions between a large organic molecule and MoS2 surface, it is necessary to get deeper understanding how a basic carboxylic residue, i.e formic acid (HCOOH), could establish attachment to MoS2 Computational details First-principles calculations based on density functional theory (DFT) are employed as the main investigating method in this study The Perdew-Bucke-Ernzerhof (PBE) functional [17–19] implemented in the Vienna Ab Initio Simulation Package (VASP) [20– 22] is utilized and the projector-augmented wave method [23] is employed to construct electronic wave-functions for the participating atoms Grimme’s D3 empirical corrections for long-range van der Waals’ interactions are activated for all investigated models [24] For the assumption of lattice circulation in the x and y directions, a (5 Â 5) super-cell of MoS2 consisting of 75 atoms are employed with the c-axis length chosen as 28 Å to guarantee surface isolation For computational feasibility, the constant volume optimization scheme is executed with a force convergence criterion of 10À4 eV The cut-off energy level of 400 eV and a k-point mesh of (3 Â Â 1) are chosen Results and discussion 3.1 Attachment of HCOOH/HCOOÅ on the pure/defected MoS2 surface The basis of binding between a heterogeneous surface and organic dye structures relies on the terminated carboxylate residue, in which oxygen atoms can be attached to the surface [25] Before going into the discussion with dye attachments, we first explore the physics and chemistry understanding of binding origin between an MoS2 surface and the simplest carboxylic residue, HCOOH In the first case, we assume there is neither surface defect nor formic acid reduction, i.e the original structure of formic acid (HCOOH) is in direct contact with MoS2 Because of surface inertness, only van der Waals interaction is formed to keep formic acid quite immobilized By looking at the charge density cloud in Fig 1(a), we observe that the H and O atoms seem to establish weak interactions with those S atoms on the surface Quantitatively, to justify the statement of stability, we examine binding energy using the following equation: Ebinding ẳ Esurface ỵ Eresidue Ecomplex 1ị where Esurface, Eresidue, Ecomplex denote the total energies of the MoS2 surface (with S defect or without S defect depending on case study), organic ligand, and the whole binding complex, respectively In Eq (1), the magnitude of positive Ebinding indicates how strongly the residue is stabilized on the MoS2 surface As in the very first case, the binding energy is only 0.02 eV, which can be regarded as a very weak physisorption In two previous studies [26,27], the physisorption of H2O on the MoS2/WS2 monolayer was shown to be very weak and caused no adjustments on the electronic properties of the 2D layer The eigenstates representing an HCOOH orbital show up as a non-bonding state, and the electronic structure of the thin film layer remains unaltered Upon analyzing charge distribution (Fig 1(a)), we observe insignificant charge transfer between MoS2 and formic acid In the second case, we alternatively consider the attachment of the radical formate residue (HCOOÅ) There is a clear improvement on binding stability (i.e binding energy is elevated up to 0.53 eV) In this case (Fig 1(b)), both O atoms seem to reside on the surface and enhance van der Waals interactions with the most nearby S Fig (a) Charge density distribution of HCOOH interacting with the pure MoS2 monolayer, (b) partial DOS of HCOOÅ absorption on MoS2, and (c) partial DOS of HCOOÅ absorption on MoS2Àe Please cite this article in press as: H.M Le et al., Chem Phys Lett (2016), http://dx.doi.org/10.1016/j.cplett.2016.11.007 H.M Le et al / Chemical Physics Letters xxx (2016) xxx–xxx Fig Molecular structures of L0 and TN1 (2-cyano-3-(N-butyl-3-indolyl) acrylic acid) dyes Fig Partial DOS of L0 radical residue absorption on the defected MoS2 surface In the Bader charge analysis, red contribution corresponds to positive charge, while green contribution depicts negative charge The dye residue is well immobilized with a binding energy of 0.51 eV (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Please cite this article in press as: H.M Le et al., Chem Phys Lett (2016), http://dx.doi.org/10.1016/j.cplett.2016.11.007 H.M Le et al / Chemical Physics Letters xxx (2016) xxx–xxx atoms Another piece of evidence showing a minor interaction of charge density between the O and S anions also implies weak bonding Even though the interacting model in the second case is more improved than that in the first case, we not conceive significant change in the electronic property of MoS2 In the electronic structure, the highest-occupied orbital of HCOOÅ leads to the creation of an additional eigenstate at the Fermi level, which is in conjunction with an energy occupation of the surface S atoms because of the van der Waals interaction In this circumstance, such an electronic state might behave like an ‘‘agent” between the highest occupied (HO) and lowest unoccupied states of MoS2 It should be noted that the electronic structure of MoS2 is very similar to that of the pure MoS2 layer (see Fig S1, Supplementary Material) In several previous studies, the band gap of pure MoS2 was predicted to be in the range of 1.63–1.78 eV by PBE calculations with various types of pseudo-potential sets [28–30] Upon the examination of the two cases above, we observe that it is not easy for formic acid and even the radical formate residue to sneak into MoS2 On the other hand, the formate residue can form strong bonds to the surface if there is a vacancy at the S site In that case, the oxygen atoms may penetrate into the layer and establish direct interactions with Mo During the past few years, progressive steps have been made toward S vacancy creation in the MoS2 layer [31,32] In reality, sulfur vacancies are common and play an essential role in catalysis [10,33] To our awareness, Ma et al [34] even devoted an effort to repair S vacancy by introducing gaseous molecules, such as CO, NO, and NO2 By adopting first-principles calculations, the formation energy of a vacancy by removing one S atom from the MoS2 monolayer was shown to be an endothermic reaction with an activation energy of 2.35 eV [35]; at the same time, the valence and conduction bands were shown to expand to lower energy levels and thereby reduce the band gap [36] In fact, our first-principles calculations demonstrate that this is really the case when the radical O atom connects to two leftbehind Mo atoms due to the absence of S In addition, the van der Waals interactions between the other atoms in the residue and the surrounding S atoms should also be taken into account Analytically, our binding energy calculation using Eq (1) with the total energy of defected MoS2 suggests that the organic residue is magnificently stabilized (Ebinding = 2.47 eV) compared to the previous two cases As can be seen in Fig 1(c), the amount of charge transfer is much more significant from Mo(4d) to O(2p) Adopting Bader charge density analysis with a qualitative isosurface value of 0.001 eV/cell, we really observe a chemisorption behavior In this chemical connection, it is the HCOOÅ radical group that possesses positive charge, while defected MoS2 has negative charge The electronic structure seen from the density of states (DOS) in Fig 1(c) is different from the previous two cases Due to the strong bonding with ÅOOCH, the electronic structure of MoS2Àe changes significantly The highest-occupied state resulted from the Mo(4d)-O (2p) ionic interaction is located around the Fermi level The energy gap between the highest-occupied state at the Fermi level and the next occupied state is 1.2 eV Furthermore, the next occupied state of the MoS2 layer is shifted drastically from the Fermi level In the previous physisorption case shown in Fig 1(b), the HO band is constituted solely by the radical formate group, which might not be meaningful in electronic applications 3.2 Binding L0 and a newly-synthesized dye to defected MoS2 At this point, we have a clear understanding of carboxylate residue interacting with the MoS2 layer When the hydrogen atom from carboxylic acid groups is not removed from the organic dye, the interaction is extremely weak; in addition, neither electronic tuning nor charge transfer can be found Therefore, in the later investigation of organic dye attachments on MoS2, we only consider the binding of ACOOÅ to an MoS2 surface with vacancy defect at the S site It should be kept in mind that the chirality of those large dye molecules makes it harder to stabilize the binding sites between S defects and ACOOÅ groups As a result, the binding energies might be lower than the previous case of HCOOÅ adsorption The structural conformations of two investigated organic dyes are provided in Fig As the first attempt to present a realistic model, we explore the possibility of decorating the MoS2 layer with L0, a well-known dye belonged to the TPA-based class In reality, this dye has been attached to the surface of TiO2 for sensitized solar-cell applications [37], and the electronic structure properties have been verified Fig Partial DOS of bonding Mo(4d) and O(2p) orbitals in (a) MoS2Àe-OOCH, (b) MoS2Àe-L0, and (c) MoS2Àe-TN1 Please cite this article in press as: H.M Le et al., Chem Phys Lett (2016), http://dx.doi.org/10.1016/j.cplett.2016.11.007 H.M Le et al / Chemical Physics Letters xxx (2016) xxx–xxx using DFT calculation methods [38] Apart from the traditional TiO2 surface, we believe there is a prospect of this organic structure to deliver interesting electronic features on MoS2 From the result of our optimizations, L0 is favorably attached to defected MoS2 with a binding energy of 0.51 eV in a bidentate mode More specifically, both O atoms make connections to the two Mo sites sharing a common S defect to establish two chemically equivalent Mo-O linkages With those two bridges, it is quite surprising that the binding energy in this case is lower compared to the attachment of HCOOÅ We believe the hardship of chirality adjustment is due to the clumsy conformation of the organic structure Recall that in the HCOOÅ case, we observe only a monodentate bond from the radical O atom to the layer, but the binding energy is much higher In a previous study concerning dye-sensitized MoS2, it was experimentally demonstrated that the Eosin Y organic dye formed covalent bond with the defected single MoS2 layer [5] More specially, the evidence of ps time-resolved photoluminescence spectroscopy showed significant electronic transfer from Eosin Y to the MoS2 layer We will see later in our DOS analysis that the in-gap states induced by the presence of the dye molecules is responsible for such electronic transfer In terms of covalent bonding, it was also pointed out in another study by scanning tunneling microscopy that the organic thiols established interactions with MoS2 at the vacancy site [39] In the L0-MoS2 structure, the amount of charge exchange revealed in Fig seems to be more significant At the connection bridge between L0 and the MoS2Àe layer, the organic structure has negative charge by perceiving electron density, while the defected MoS2 layer possesses positive charge Looking at the DOS plot, we observe two interesting features The first in-gap band describes electron occupation at the Fermi level, which is a hybridized band of Mo, S, and L0 radical To some physical extent, this state describes a strong bonding nature like the previous case shown in Fig 1(c) Upon the analysis of partial DOS (Fig (a) and (b)), such occupation is originated from the electron exchange of O(2p) and Mo(4d) orbitals We also observe another in-gap band, which is mostly constituted by the molecular orbital of L0 This second peak is located at around À1 eV in Fig 3, and the dominant contribution comes from O(2p) Such interesting in-gap occupation features may allow the dye molecule to absorb photon energy and give up to the MoS2 surface In general, the presence of HCOOÅ, L0, or the later dye causes the HO bands of MoS2 to be drifted away from the Fermi level At this stage, we urge to design a new dye molecule so that binding stability to the defected MoS2 surface can be further improved A new dye molecule is first designed by performing ab initio calculations, then synthesized in our laboratory This new dye molecule is 2-cyano-3-(N-butyl-3-indolyl) acrylic acid Fig Partial DOS of TN1 radical residue absorption on the defected MoS2 surface In the Bader charge analysis, red contribution corresponds to positive charge, while green contribution depicts negative charge The dye residue is well immobilized with a binding energy of 0.71 eV (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Please cite this article in press as: H.M Le et al., Chem Phys Lett (2016), http://dx.doi.org/10.1016/j.cplett.2016.11.007 H.M Le et al / Chemical Physics Letters xxx (2016) xxx–xxx (Fig 2(b)) For simplicity, we denote it as TN1 More detailed information regarding experimental synthesis with FT-IR, GC–MS, H NMR, and 13C NMR spectra can be consulted from the Supplementary Material (Figs S2–S7) Upon the omission of two phenyl rings, the new structure TN1 seems to settle better on the surface of defected MoS2 According to our calculations, the binding energy is reported as 0.79 eV, higher than that in the L0 case The binding conformation of TN1 is more perpendicular to the MoS2 surface, where we observe sorts of tilting behavior caused by the van der Waals’ interactions between one aromatic phenyl group and MoS2 Moreover, only one O atom establishes ionic bonds with the two nearby Mo atoms This bonding behavior is different from that seen in the case of the L0 bidentate attachment Examining the DOS of defected MoS2 (Fig for the TN1 adsorption case), we observe that the electronic behavior of the layer is very similar to that when L0 is attached to MoS2 This observation makes sense in terms of chemical interaction equivalence For both organic structures (L0 and TN1), it is the ACOOÅ radical residue that establishes chemical ionic bonding to two Mo sites nearby the S vacancy, while the AC„N residue also seems to establish weak van der Waals’ interaction to the surrounding S atoms In the partial DOS plot of the MoS2Àe-TN1 complex, we observe there is a polar covalent bond formed as a result of Mo and TN1 orbital interactions (the hybridized peak at the Fermi level), which is dominant by the O(2p) contribution There is also another band (À1 eV) originated from the ligand contribution to the hybridization, which serves as an intermediate in-gap state Such an in-gap occupation resides at a quite lower energy level compared to that of the L0 absorption case This result is not surprising, but implies the fact that the complex with TN1 is more stable because its bonding orbitals tend to reside at lower energy state Summary In summary, we have demonstrated a theoretical investigation of two different organic dye structures on the surface of defected MoS2 In the initial attempt, we perform three testing cases for formic acid/formate residues to be attached to the MoS2 surface When there is no surface defect, the ACOOH/ACOOÅ can only establish weak van der Waals interactions with the layer The removal of an S atom actually prevails The HCOOÅ residue is shown to bind strongly to the Mo atom with a binding energy of 2.47 eV When considering actual large dye molecules such as the L0 and newly-synthesized TN1 structures, we find the binding energies to be lower due to chirality adjustment of the organic ligands Hybridized occupation states and charge transfer clearly indicate strong ionic connections, while there is also one in-gap state showing up at around À1 eV from the Fermi level, which might be supportive in photon trapping Acknowledgments We are grateful for a research fund from Ton Duc Thang University and the supercomputing support from the High Performance Computing Infrastructure Office (project hp150037) and the Institute for Material Research, Tohoku University, Japan Pham-Tran thanks a financial support from Vietnam National University under grant HS-2014-18-01 Appendix A Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cplett.2016.11 007 References [1] T Bartels, W Bock, J Braun, C Busch, W Buss, W Dresel, C Freiler, M Harperscheid, R.-P Heckler, D Hörner, F Kubicki, G Lingg, A Losch, R Luther, T Mang, S Noll, J Omeis, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co KGaA, 2000 [2] A Klein, S Tiefenbacher, V Eyert, C Pettenkofer, W Jaegermann, Phys Rev B 64 (2001) 205416 [3] W Han, P Yuan, Y Fan, G Shi, H Liu, D Bai, X Bao, J Mater Chem 22 (2012) 25340 [4] J.-F Paul, E Payen, J Phys Chem B 107 (2003) 4057 [5] T Jia, M.M.J Li, L Ye, S Wiseman, G Liu, J Qu, K Nakagawa, S.C.E Tsang, Chem Commun 51 (2015) 13496 [6] J Kibsgaard, T.F Jaramillo, F Besenbacher, Nat Chem (2014) 248 [7] J Shi, D Ma, G.-F Han, Y Zhang, Q Ji, T Gao, J Sun, X Song, C Li, Y Zhang, X.-Y Lang, Y Zhang, Z Liu, ACS Nano (2014) 10196 [8] D Lembke, S Bertolazzi, A Kis, Acc Chem Res 48 (2015) 100 [9] X Chen, A.R McDonald, Adv Mater 28 (2016) 5738 [10] S Presolski, M Pumera, Mater Today 19 (2016) 140 [11] X Chen, N.C Berner, C Backes, G.S Duesberg, A.R McDonald, Angew Chem Int Ed 55 (2016) 5803 [12] C Ataca, S Ciraci, J Phys Chem C 115 (2011) 13303 [13] C Backes, N.C Berner, X Chen, P Lafargue, P LaPlace, M Freeley, G.S Duesberg, J.N Coleman, A.R McDonald, Angew Chem Int Ed 54 (2015) 2638 [14] S.H Yu, Y Lee, S.K Jang, J Kang, J Jeon, C Lee, J.Y Lee, H Kim, E Hwang, S Lee, J.H Cho, ACS Nano (2014) 8285 [15] T Kitamura, M Ikeda, K Shigaki, T Inoue, N.A Anderson, X Ai, T Lian, S Yanagida, Chem Mater 16 (2004) 1806 [16] D.P Hagberg, T Marinado, K.M Karlsson, K Nonomura, P Qin, G Boschloo, T Brinck, A Hagfeldt, L Sun, J Org Chem 72 (2007) 9550 [17] J.P Perdew, J.A Chevary, S.H Vosko, K.A Jackson, M.R Pederson, D.J Singh, C Fiolhais, Phys Rev B 46 (1992) 6671 [18] J.P Perdew, J.A Chevary, S.H Vosko, K.A Jackson, M.R Pederson, D.J Singh, C Fiolhais, Phys Rev B 48 (1993) 4978 [19] J.P Perdew, K Burke, M Ernzerhof, Phys Rev Lett 77 (1996) 3865 [20] G Kresse, J Hafner, Phys Rev B 47 (1993) 558 [21] G Kresse, J Hafner, Phys Rev B 49 (1994) 14251 [22] G Kresse, J Furthmüller, Comput Mater Sci (1996) 15 [23] P.E Blöchl, Phys Rev B 50 (1994) 17953 [24] S Grimme, J Antony, S Ehrlich, H Krieg, J Chem Phys 132 (2010) 154104 [25] J Chen, R.E Ruther, Y Tan, L.M Bishop, R.J Hamers, Langmuir 28 (2012) 10437 [26] Q Yue, Z Shao, S Chang, J Li, Nanoscale Res Lett (2013) [27] V.Q Bui, T.T Pham, D.A Le, C.M Thi, H.M Le, J Phys.: Condens Matter 27 (2015) 305005 [28] C Ataca, S Ciraci, Phys Rev B 85 (2012) 195410 [29] T Li, Phys Rev B 85 (2012) 235407 [30] E Scalise, M Houssa, G Pourtois, V Afanas’ev, A Stesmans, Nano Res (2011) 43 [31] H Li, C Tsai, A.L Koh, L Cai, A.W Contryman, A.H Fragapane, J Zhao, H.S Han, H.C Manoharan, F Abild-Pedersen, J.K Norskov, X Zheng, Nat Mater 15 (2016) 48 [32] B.H Kim, M Park, M Lee, S.J Baek, H.Y Jeong, M Choi, S.J Chang, W.G Hong, T K Kim, H.R Moon, Y.W Park, N Park, Y Jun, RSC Adv (2013) 18424 [33] R.R Chianelli, M.H Siadati, M.P De la Rosa, G Berhault, J.P Wilcoxon, R Bearden, B.L Abrams, Catal Rev 48 (2006) [34] D Ma, Q Wang, T Li, C He, B Ma, Y Tang, Z Lu, Z Yang, J Mater Chem C (2016) 7093 [35] D Liu, Y Guo, L Fang, J Robertson, Appl Phys Lett 103 (2013) 183113 [36] L.-P Feng, J Su, S Chen, Z.-T Liu, Mater Chem Phys 148 (2014) [37] M Pastore, E Mosconi, F De Angelis, M Grätzel, J Phys Chem C 114 (2010) 7205 [38] M Pastore, F De Angelis, Phys Chem Chem Phys 14 (2012) 920 [39] M Makarova, Y Okawa, M Aono, J Phys Chem C 116 (2012) 22411 Please cite this article in press as: H.M Le et al., Chem Phys Lett (2016), http://dx.doi.org/10.1016/j.cplett.2016.11.007 ... xxx–xxx theoretical investigation of organic dye attached to an MoS2 surface to get more insights of the binding nature and in-gap occupations The organic dyes of interest consist of the well-known... believe the hardship of chirality adjustment is due to the clumsy conformation of the organic structure Recall that in the HCOOÅ case, we observe only a monodentate bond from the radical O atom to the. .. calculations demonstrate that this is really the case when the radical O atom connects to two leftbehind Mo atoms due to the absence of S In addition, the van der Waals interactions between the other