⇦Chinese Physics Letters, 2019, Vol. 36, No. 11, Article code 116401Express Letter Magnetic Coupling Induced Self-Assembly at Atomic Level * Weiyu Xie (解伟誉)†, Yu Zhu (朱瑜)†, Jianpeng Wang (王健鹏), Aihua Cheng (程爱华), Zhigang Wang (王志刚)** Affiliations Institute of Atomic and Molecular Physics, Jilin University, Changchun 130012 Received 5 October 2019, online 15 October 2019 *Supported by the National Natural Science Foundation of China under Grant No 11674123.
†Weiyu Xie and Yu Zhu contributed equally to this work.
**Corresponding author. Email: wangzg@jlu.edu.cn Citation Text: Jie W Y, Zhu Y, Wang J P, Cheng A H and Wang Z G et al 2019 Chin. Phys. Lett. 36 116401    Abstract Developing accurate self-assembly is the key for constructing functional materials from a bottom-up approach. At present, it is mainly hindered by building blocks and driving modes. We design a new self-assembly method based on the magnetic coupling between spin-polarized electrons. First-principles calculations show that spin-polarized electrons from different endohedral metallofullerene (EMF) superatoms can pair each other to ensure a one-dimensional extending morphology. Furthermore, without ligand passivation, the EMF superatoms maintain their electronic structures robustly in self-assembly owing to the core-shell structure and the atomic-like electron arrangement rule. Therefore, it should noted that the magnetic coupling of monomeric electron spin polarization can be an important driving mechanism for high-precision self-assembly. These results represent a new paradigm for self-assembly and offer fresh opportunities for functional material construction at the atomic level. DOI:10.1088/0256-307X/36/11/116401 PACS:64.75.Yz, 82.35.Np, 31.15.-p © 2019 Chinese Physics Society Article Text Self-assembly, a scientific issue that has garnered much interest,[1,2] is a widely applied strategy in chemical and material science for designing artificial structures from simple building blocks.[3–5] Realizing accurate self-assembly at the atomic level is the ultimate goal, however, huge challenges remain. For instance, it is still difficult to control the morphology and the electronic states in a self-assembly process. Further, traditional self-assembly methods have their own limitations, such as often destroying the stability of the building block itself and requiring very complex ligand structures.[6–9] In fact, during the self-assembly process, it is difficult or even contradictory for building blocks to have strong bond strengths without destroying their structures. In addition, the traditional self-assembly methods are usually regulated by thermal diffusion, which often leads to uncertainty. Therefore, it is particularly important to realize atomic-level control rather than relying on thermal motion. Achieving controllable self-assembly of both morphology and electronic state has always been a subject of interest.[10] Its success depends on many factors, such as building block structure and interaction, which provides a forward direction for designing the system at the atomic level. Consequently, we can use more stable building blocks of atomic-level precision, e.g., superatom clusters. Such superatom clusters always show extraordinary stability with electronic closed-shell structures.[11–13] Then, if the driving method can be controlled at the atomic level, such as electronic or magnetic recognition effects, we may break the traditional limitations imposed by thermally regulated self-assembly processes. In this Letter, a new atomic-level self-assembly method, which is driven by magnetic coupling between the building blocks, is presented. Unlike the previous reports, these building blocks can be recognized and combined spontaneously, by means of their own magnetism, and can form stable one-dimensional (1D) chains. Moreover, further analyses indicate that the building blocks can clearly maintain robust structures and have strong bond strengths during the assembly process. Predictably, this approach will lead to the development of atomic-level self-assembly. To achieve this thought, we employed endohedral metallofullerene (EMF) superatoms U@C$_{28}$ (uranium embedded in C$_{28}$ fullerene) as the building blocks[14] U@C$_{28}$ exhibits the following electronic structural properties: the cage adopts the 32-electron principle using the bonding $s$-, $p$-, $d$-, and $f$-type orbitals of uranium. The remaining two spin polarized electrons cannot break the strong U-cage interactions caused by the 32-electron principle.[15] Subsequently, a series of superatom-assembled structures (U@C$_{28})_{n}$ ($n = 2$–7) were studied based on first-principles calculations. Despite the existence of multiple assembly structures, we found that the 1D chain was the most stable for each (U@C$_{28})_{n} $ structure (Fig. S1 and Table S1 in the Supplementary Material). While investigating the reason for the stability of the 1D chains, it was found that a spin-matching phenomenon occurred among the building blocks. Some spin-polarized electrons would flip their spin directions and chemical bonds were formed between the superatoms.[16] Because of the number of spin-polarized electrons in U@C$_{28}$, only two bonds could be formed, enforcing a chained-link configuration. From the spin density analysis, the net spin electrons of these structures were all located at the chain ends. This demonstrates the extensibility of these self-assembly structures and further indicates that the self-assembly process is a local behavior. Thus, the magnetic coupling is a cooperative effect here, including spin flip bonding and spin polarization. The interaction energies between the building blocks (Fig. 1) decrease gradually as the number of building blocks increases, and this trend is approximately linear. The large interaction energies of these structures show the good thermal stability. These illustrate that the electronic structure of the building blocks is not destroyed. In fact, the reason is that the spin-polarized electrons in U@C$_{28}$ only flip their spin directions during this process, which ensures that this self-assembly process is only a local effect. Furthermore, the inset of Fig. 1 exhibits that the interaction energy curve has a deep potential well in the ES position, which shows the strong bond strength of the assembly structure. The interaction energies between the building blocks were also investigated using energy decomposition analysis (EDA). Orbital interaction contributed more than 70% of the attracting interaction, therefore, the building blocks were connected by strong covalent bonds.
cpl-36-11-116401-fig1.png
Fig. 1. Interaction energies between the building blocks of (U@C$_{28})_{n}$ ($n = 2$–7). The black lines represent the trend of the interaction energies from two isolated U@C$_{28}$ to the equilibrium structures (U@C$_{28})_{n}$. The brown line was fitted to show the interaction energy more intuitively. The inset image exhibits the interaction energy during the self-assembly of (U@C$_{28})_{2}$ in detail, and other structures indicate the same change process. The distance represents the bond length between nearest C atoms on different built blocks. ES refers to the equilibrium structure. The isosurface value of the spin density diagram is 0.0005.
cpl-36-11-116401-fig2.png
Fig. 2. Density of states of (U@C$_{28})_{n}$ ($n = 2$–7). The black and red dotted lines indicate the locations of HOMO and LUMO of (U@C$_{28})_{n}$, respectively. The isosurface value of the frontier molecular orbital (MO) diagrams is 0.02. Other frontier MO diagrams can be found in Fig. S2 in the Supplementary Material.
According to the above analyses, the structure of the building blocks is still robust under the action of strong bonding, which breaks the limitation of ligand passivation self-assembly. Therefore, unlike the previous chain assembly scheme, EMF building blocks can be assembled by magnetic coupling, which ensures that the structure is not destroyed. Obviously, it is closer to the self-assembly in nature. And for EMFs, there have been successful preparing methods, such as the electrochemical deposition technique for Sc@C$_{82}$ nanowire and tube,[17] liquid-liquid interfacial precipitation (LLIP) technique for Sc$_{3} $N@C$_{80} $ nanorods and three-dimensional (3D) nanostructures.[18–19] Furthermore, the aforementioned structures do not re-quire ligand passivation, which highlights the feasibility of magnetic coupling assembly. To further understand the collective properties of these chains, the following analyses were conducted. First, the density of states (DOS) was plotted (Fig. 2). The results indicated that the highest occupied molecular orbital-lowest unoccupied molecular orbital (HOMO-LUMO) gap of these chains was 0.99 eV. The trends of these DOSs were identical, which showed that the electronic structures do not change with the increase of the number of building blocks. Overall, these DOS diagrams showed a superimposed property related to the number of building blocks, which revealed that the properties of superatoms can be maintained. The HOMO and LUMO for every chain were located at the chain ends, suggesting that the end building blocks have the highest reactivity. Interestingly, these results revealed that this self-assembly process resembles that of free-radical polymerization.[20] In these chains, there are two electronic states of near-identical energies (Tables S2 and S3 in the Supplementary Material), the low-spin broken-symmetry state and the high-spin triplet state, corresponding to antiferromagnetic coupling and ferromagnetic coupling, respectively.[16] In addition, spin-orbit coupling (SOC) was considered, owing to the existence of low-spin electronic states. The results showed that the frontier molecule orbital energy level splitting was very small (Figs. S3 and S4, Table S4 in the Supplementary Material), demonstrating that the SOC does not significantly impact the properties of the electronic structure.
cpl-36-11-116401-fig3.png
Fig. 3. Redox activity of (U@C$_{28})_{n}$ ($n = 2$–7). The abscissa represents the number of building blocks in these chains. VIP: vertical ionization potential, VEA: vertical electron affinity, AIP: adiabatic ionization potential, AEA: adiabatic electron affinity (Table S5 in the Supplementary Material for details).
To measure redox activity in these chains, the ionization potential and electron affinity were analyzed (Fig. 3). With the increase of the number of building blocks, the ionization potential gradually decreased, while the electron affinities increased. It shows that these systems tend to lose electrons, suggesting an enhancement of reducibility. It is worth noting that when $n \ge 6$ for (U@C$_{28})_{n}$, the electron affinity and ionization potential will not change anymore. Therefore, the redox properties of the chains remain unchanged as the number of building blocks continues to increase. This analysis proves the reliability of this assembly method. A new accurate self-assembly method is reported via the first-principles DFT calculations. During the self-assembly process, the building blocks tend to be chained and the driving method is magnetic coupling. The interaction energy analyzing indicates that the building blocks have strong bond strengths. The DOS and HOMO-LUMO gap analyses demonstrate that the properties of these chains are not dependent on the number of building blocks. Then, when the building blocks are increased to a certain number, the redox properties of the system will not change anymore. All the above results prove the feasibility of this magnetic-coupling-induced self-assembly method and demonstrate that the assembled building blocks is robust. Admittedly, in this study, we employed the selected EMF superatoms, whereas traditional self-assembly building blocks are almost always transition metals. Because of the inherent chemical activity of transition metals, if we want to assemble them, they will often need to be passivated by ligands. However, the method introduced in this work is different, and the reasons are as follows. Firstly, not only do EMF building blocks have thermal stability like fullerenes, but also they can maintain the excellent quantum properties of embedded metals. Secondly, EMF building blocks have stronger bonds compared with those of transition metal building blocks, allowing them to be assembled without relying on ligands. Finally, the driving method is based on magnetic coupling, which implements effective and accurate control at the atomic level. Furthermore, since magnetic fields can be created and controlled, this study shows great potential for future meaningful applications of magnetically controllable self-assemblies. In conclusion, the 1D morphology-controllable self-assembly method presented in this study has great application value in the fields of optics and electricity because of its directional transmission characteristics.[21] Our work provides an effective method for precisely positioning a 1D morphology. Further, this method is not limited to linear chains. If the basic building block has another number of spin-polarized electrons, the potential for assembling variously shaped systems becomes possible. Since the self-assembly extension mechanism is closely dominated by the electron-spin polarization characteristics of the assembly building blocks, this study highlights how the magnetic effect of electron-spin polarization in physics can be an important driving mechanism for high-precision self-assembly. These findings will open up a new field for better functionalization[2,22] and control of building blocks, such as superatoms. Data Availability: The data that support the findings of this study are available from the authors upon request. Competing Interests: The authors declare that there are no competing interests. Author Information: Correspondence and requests for materials should be addressed to Z. Wang (wangzg@jlu.edu.cn). Author Contributions: W. Xie and Y. Zhu contributed equally to this work. Z. Wang conceived this project. Computational details were performed by W. Xie, Y. Zhu and led by Z. Wang. W. Xie, Y. Zhu, J. Wang and Z. Wang acquired and analyzed the data. W. Xie, Y. Zhu, J. Wang, A. Cheng and Z. Wang prepared the manuscript. We warmly thank W. Jiang and Y. Gao for the stimulating discussions. Z. Wang acknowledges the High Performance Computing Center of Jilin University. References How Far Can We Push Chemical Self-Assembly?Charting a course for chemistryChemical Topology: Complex Molecular Knots, Links, and EntanglementsStimuli-Responsive Metal–Ligand AssembliesSelf-assembly of polycyclic supramolecules using linear metal-organic ligandsSynthesis and structural characterization of an AI77 clusterChemically Modified Gold Superatoms and Superatomic MoleculesMagnetization of Pt 13 clusters supported in a NaY zeolite: A XANES and XMCD studyResistance and resilience to changing climate and fire regime depend on plant functional traitsBeyond molecules: Self-assembly of mesoscopic and macroscopic componentsElectronic Shell Structure and Abundances of Sodium ClustersBeyond the Periodic Table of Elements: The Role of SuperatomsCluster-Assembled MaterialsUranium Stabilization of C28: A Tetravalent FullereneU@C 28 : the electronic structure induced by the 32-electron principleBinding for endohedral-metallofullerene superatoms induced by magnetic couplingTemplate Synthesis of Sc@C82(I) Nanowires and Nanotubes at Room TemperaturePreparation of endohedral metallofullerene nanowhiskers and nanosheetsUltrasonication-switched formation of dice- and cubic-shaped fullerene crystals and their applications as catalyst supports for methanol oxidationLiquid bridge induced assembly (LBIA) strategy: Controllable one-dimensional patterning from small molecules to macromolecules and nanomaterialsMulti-step self-guided pathways for shape-changing metamaterials
[1] Service R F 2005 Science 309 95
[2] Aspuru-Guzik A et al 2019 Nat. Chem. 11 286
[3] Forgan R S, Sauvage J P, Stoddart J F 2011 Chem. Rev. 111 5434
[4] McConnell J, Wood C S, Neelakandan P P, Nitschke J R 2015 Chem. Rev. 115 7729
[5] Song, Kandapal S, Gu J, Zhang K, Reese A, Ying Y, Wang L, Wang H, Li Y, Wang M, Lu S, Hao X Q, Li X, Xu B, Li X 2018 Nat. Commun. 9 4575
[6] Ecker A, Weckert E, Schnöckel H 1997 Nature 387 379
[7] Nishigaki J, Koyasu K, Tsukuda T 2014 Chem. Rec. 14 897
[8] García L M, Roduner E, Akdogan Y, Wilhelm F, Rogalev A 2009 Phys. Rev. B 80 14404
[9] Zhou M, Jin R, Sfeir M Y, Chen Y, Song Y, Jin R 2017 Proc. Natl. Acad. Sci. USA 114 4697
[10] Whitesides G M, Boncheva M 2002 Proc. Natl. Acad. Sci. USA 99 4769
[11] Knight W D, Clemenger K, W A de Heer, Saunders W A, Chou M, Cohen M L 1984 Phys. Rev. Lett. 52 2141
[12] Jena P 2013 J. Phys. Chem. Lett. 4 1432
[13] Claridge S A, Castleman A W, Khanna S N, Murray C B, Sen A, Weiss P S 2009 ACS Nano 3 244
[14] Guo T, Diener M D, Chai Y, Alford M J, Haufler R E, McClure S M, Ohno T, Weaver J H, Scuseria G E, Smalley R E 1992 Science 257 1661
[15] Dai X, Gao Y, Jiang W, Lei Y, Wang Z 2015 Phys. Chem. Chem. Phys. 17 23308
[16] Xie W, Jiang W, Gao Y, Wang J, Wang Z 2018 Chem. Commun. 54 13383
[17] Li J, Guo G, Li S, Wang R, Wan J, Bai L 2005 Adv. Mater. 17 71
[18] Wakahara T, Nemoto Y, Xu M, Miyazawa K, Fujita D 2010 Carbon 48 3359
[19] Xu Y, Chen X, Liu F, Chen X, Guo J, Yang S 2014 Mater. Horiz. 1 411
[20]Moad G, Solomon D H 1995 The Chemistry of Free Radical Polymerization (Oxford: Pergamon Press)
[21] Yan C, Su B, Shi Y, Jiang L 2019 Nano Today 25 13
[22] Coulais C, Sabbadini A, Vink F, M van Hecke 2018 Nature 561 512
⇦Chinese Physics Letters, 2019, Vol. 36, No. 11, Article code 116801 Observation of Simplest Water Chains on Surface Stabilized by a Hydroxyl Group at One End * An-Ning Dong (董安宁)1,2,5, Li-Huan Sun (孙丽欢)1,2, Xiang-Qian Tang (唐向前)1,2, Yi-Kun Yao (姚一锟)1,2, Yang An (安旸)1,2, Dong Hao (郝东)1,2, Xin-Yan Shan (单欣岩)1**, Xing-Hua Lu (陆兴华)1,2,3,4** Affiliations 1Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190 2School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100190 3Collaborative Innovation Center of Quantum Matter, Beijing 100190 4Songshan Lake Materials Laboratory, Dongguan 523808 5Engineering Technology Department, Zolix Instruments Co. Ltd, Beijing 101102 Received 9 July 2019, online 21 October 2019 *Supported by the National Natural Science Foundation of China under Grant Nos 11774395 and 91753136, the Beijing Natural Science Foundation under Grant No 4181003, the Strategic Priority Research Program (B) of the Chinese Academy of Sciences under Grant Nos XDB30201000 and XDB28000000.
**Corresponding author. Email: shanxinyan@iphy.ac.cn; xhlu@iphy.ac.cn Citation Text: Dong A N, Sun L H, Tang X Q, Yao Y K and An Y et al 2019 Chin. Phys. Lett. 36 116801    Abstract The key to fully understanding water-solid interfaces relies on the microscopic nature of hydrogen bond networks, including their atomic structures, interfacial interactions, and dynamic behaviors. Here, we report the observation of two types of simplest water chains on Au(111) surface which is expected unstable according to the rules of hydrogen network on noble metal surfaces. A common feature at the end of chain structures is revealed in high resolution scanning tunneling microscopy images. To explain the stability in observed hydrogen bond networks, we propose a structure model of the water chains terminated with a hydroxyl group. The model is consistent with detailed image analysis and molecular manipulation. The observation of simplest water chains suggests a new platform for exploring fundamental physics in hydrogen bond networks on surfaces. DOI:10.1088/0256-307X/36/11/116801 PACS:68.37.Ef, 68.43.Fg, 82.65.+r © 2019 Chinese Physics Society Article Text The behavior of water molecules on solid surfaces at the microscopic scale has attracted significant research interest, due to the ubiquitous water-solid interfaces in our daily life, industrial manufactures, and biosystems.[1–3] Various water structures have been observed and interpreted by scanning tunneling microscopy (STM) combined with density functional theory (DFT), including cyclic water clusters,[4–12] quasi one-dimensional (1D) water chains[13–15] and two-dimensional (2D) water films.[16–19] It is worth noting that all water structures, except very small species, prefer a closed-loop structure on solid surfaces wherever water-substrate interaction is not playing a dominating role. This is because water molecules tend to build as many H-bonds as possible to reach the lowest-energy state, and it has been accepted as one of the rules for sub-monolayer ice growing on noble metal surfaces.[11,20] A water chain is an interesting type of water structure and it is believed to relate to the fundamental living activities in a cell.[21] For example, water molecules transport across the nanochannels on biological membranes in the form of chains.[22] It is thus of great interest to explore the characteristics of such water chains at the molecular scale. Following the rules of water structures on surfaces, the simplest water chains are unstable due to the unsaturated hydrogen bond at one of the chain ends. Water chains as reported in previous STM studies are either under strong constrained conditions or built with complex cyclic water clusters. Examples include water chains along the step edges on Pt surface[23] and crystal structure on Cu(110) surface,[8] and the quasi one-dimensional water structure constituted with face-sharing pentagonal water units on Cu(110) surface.[14] The question is whether it is possible to have the simplest water chains on a solid surface that are not constrained by linear surface reconstructions. Such nearly free water structures are better mimics of water chains in biological activities, and their properties and behaviors are expected to better reflect the nature of hydrogen bonds between water molecules in aqueous solutions. Here, we report the observation of the simplest water chains being stable and unconstrained on a noble metal surface using a low-temperature STM. Two types of water chains are imaged with molecular resolution in which sequential water molecules are linked by hydrogen bonds in zigzag and armchair patterns. To the best of our knowledge, there has not yet been any report of such water structures. Based on the observed topographic features, a structure model has been proposed to explain the stability in these water chains. Molecular manipulations have also been tried to validate the model and to explore the dynamic behavior in such linear hydrogen bond networks. An STM experiment was performed on Au(111) crystal surface in ultrahigh vacuum with pressure lower than 10$^{-10}$ Torr. The Au substrate was cleaned by cycles of sputter with Ar$^{+}$ and anneal around 900 K, then cooled down to about 17 K with liquid helium. Water vapors are then leaked into the vacuum chamber (with a pressure increase by about $0.5\times 10^{-11}$ Torr) and dosed onto the crystal surface. Figure 1(a) presents a typical STM topographic image where a variety of water structures are observed. Besides small water clusters, which have been reported and discussed in a previous report,[12] two new types of water chain structures are observed. We note here that such high-resolution images can be obtained only with carefully functionalized STM tips (see Method in the Supplementary Material).
cpl-36-11-116801-fig1.png
Fig. 1. Water chains on Au(111) surface. (a) Typical high resolution STM image of water structures on surface. Most water molecules constitute closed-loop structures, while water chains are occasionally observed as well. Imaging set point: sample bias $V=-200$ mV, tunneling current $I=50$ pA. [(b), (c)] STM image of water chains with molecular resolution, illustrating armchair and zigzag patterns, respectively. Both chains present a common feature at one of the ends, as indicated by the dotted line. Imaging set point: $V=-100$ mV, $I=50$ pA.
The water molecules in the chains are linked in a zigzag or armchair pattern. The chain length as observed is in the range of 3–16 molecules. The distance between adjacent water molecules in the chains is about 2.95 Å, close to the substrate lattice constant (2.90 Å). Careful examination, however, reveals that the centers of each water molecule are not located at the same position of the substrate atomic lattice. The O–O–O angle between adjacent water molecules is 108$^{\circ}\pm10^{\circ}$ for the armchair and 94$^{\circ}\pm10^{\circ}$ for the zigzag patterns (see Fig. S1 in the Supplementary Material). Some water chains present a mixture of zigzag and armchair patterns. These observations indicate that the hydrogen bonding in these water chains is stronger than or at least comparable to the strength of water-substrate interaction. It is thus a valuable platform in exploring nature of hydrogen bonds in a nearly free two-dimensional environment. The height of water molecules along the water chain is around 0.90 Å, which is close to the height of water molecules in a cyclic hexamer. It is very interesting that each water chain has two distinct ends. At one end of the chains, the first two water molecules (especially the second molecule) are slightly higher than the other molecules in the chain. The last water molecule at the other end, on the other hand, appears lowest along the chain. The difference in heights between the highest one and the lowest one reaches about 0.2 Å (see Fig. S1 in the Supplementary Material). The other distinct feature is the depression in topography near the lowest molecule at the end. The depression appears as an arc around the molecule at a distance of about 5 Å, and it is about 0.06 Å deep refer to the surface nearby, as shown in Fig. 2.
cpl-36-11-116801-fig2.png
Fig. 2. Feature of depression at one end of the water chain. (a) STM image of an armchair water chain. Imaging set point: $V=-100$ mV, $I=50$ pA. (b) Height profile along the line-cut as indicated in (a). The depression is marked by the arrow.
The existence of such a water chain on a flat surface is surprising because according to the rules for ice growing on surfaces,[20] water molecules tend to build closed-loop structures with as many hydrogen bonds as possible to reach the lowest energy state. Based on the observed topographic features, we propose a hydroxyl-terminated structure model to explain the stability of such nearly free water chains on noble metal surfaces. Figure 3 shows the atomic model for the proposed hydroxyl-terminated water chain. Adjacent water molecules are linked through hydrogen bonds, the molecule at one end acts as a pure hydrogen 'donor' and the species at the other end being a hydroxyl group. This hydroxyl group makes the formation of close-loop unnecessary and thus stabilizes the chain structure.
cpl-36-11-116801-fig3.png
Fig. 3. Atomic model of hydroxyl-terminated 1D water chain. (a) Armchair water chain, (b) zigzag water chain. The black dotted line corresponds to the depression position in STM topographs.
The validity of this model is based on following arguments. First, the depression at one end of the chain suggests the presence of a hydroxyl group. The species at the chain end could be a complete water molecule, a hydroxyl group, or an oxygen atom. Obviously, one can exclude the possibility of being a complete water molecule. An oxygen atom on noble metal surface will present as a round depression in STM topography, as reported previously,[24] which is different from the end feature as observed. An hydroxyl group on noble metal surface, on the other hand, appears as a symmetric double depression,[24] the half of which presents the feature at the end of the water chain. The reason for the appearance of single depression instead of a symmetric doublet at water chain end is due to the interaction with linked water molecules, which prohibits the wobbling of the hydrogen atom. The depression is not observed in small water clusters imaged with the same STM tip, which excludes the tip effect as well. Secondly, the depression is not related to any defect on surface. The water chain can be manipulated to a different site, exposing intact flat terrace without any surface defect. The feature can even be switched from one end to the other end of the chain by applying a voltage pulse with tip set at the middle of the chain, as shown in Fig. 4. Lastly, a hydroxyl group at water chain end establishes an energy-favorable configuration. Due to the unsaturated coordination in hydroxyl group, the lone pair orbital in the O atom leads to an enhancement in O-metal interactions and renders a local nadir of potential energy. Consequently, the hydroxyl group bonds to the underlying metal atom and stabilizes the 1D water chains on the surface.
cpl-36-11-116801-fig4.png
Fig. 4. Manipulation of water chain. (a) Configuration change between armchair and zigzag in a water chain consists of six molecules, stimulated by a voltage pulse through STM tip. (b) Stimulated proton transfer along an armchair water chain. The feature of depression at one end disappears and emerges at the other end of the chain after a voltage pulse. Imaging set point: $V=-100$ mV, $I=50$ pA. The marks indicate the positions where the voltage pulses ($-0.5$ V with 200 ms duration) were applied.
The zigzag and armchair configurations can be mixed in a single water chain, indicating that they are close in energy. This can be further confirmed by the fact that a water chain consist of one pattern can be manipulated into the other pattern. As shown in Fig. 4(a), following a voltage pulse stimulation, a water chain composed of six water molecules in armchair configuration is converted to the zigzag configuration. The basic features, including slight slope in height along the chain and the depression at one end of the chain, remain unaltered though. Further manipulation experiments show that the hydroxyl group can transfer from one end of the water chain to the other end, when stimulated by a voltage pulse from STM tip, as demonstrated in Fig. 4(b). One possible method for such transfer process is the actual move of hydroxyl group along the surface. Considering the location of stimulation and the topographic positions of the water chain, the other way is more reasonable in which the transfer of hydroxyl group is achieved by reconfiguring the hydrogen bonds between adjacent water molecules. The hydroxyl group accepts one hydrogen atom from its neighbor and turns into a water molecule, and the last water molecule donates one hydrogen atom and becomes a hydroxyl group. As a result, a hydrogen atom is transferred through the whole molecular chain, which is also referred to as proton transfer process.[25–27] The interesting point is whether the hydrogen atom or the proton is transferred through each molecule sequentially or, all water molecules act simultaneously (co-tunneling). The fact that the stimulation is performed at the middle of the chain suggests that a co-tunneling scenario is more plausible. Further controlled experiments, of course, are needed to confirm such hypothesis. Nevertheless, the manipulation experiments demonstrate that these simplest water chains are of significant advantage in exploring dynamic behaviors in hydrogen bond networks. A hydroxyl-terminated water chain on a noble metal surface has not been predicted by any density function theory calculation. Recently, Peköz and Donadio investigated the partial dissociation of water chains at (221) stepped metallic surface by density functional theory and molecular dynamics simulations.[28] Their results predicted that the favorable site for a hydroxyl group is in the middle of the water chains instead of the ends, which maximizes the water-metal and water-water interactions and leads to a low energy state. The discrepancy, of course, can be due to the difference in substrate activities (active Pt, Ir, Rh, and Pd surfaces, versus nonactive Au surface) and the constraint of step edges. Further attempts with first-principle calculations, based on PBE functional[29] with modified parameters (optB86b-vdW[30] scheme), still cannot explain the experimental observations.[31] The inconsistency between theoretical calculation and experimental observation, again, illustrates the long-standing challenges in exploring fundamental physical principles in nature of water where the subtle balance between various interactions renders exuberant phenomena. In summary, the simplest water chains on noble metal surfaces consisting of armchair and zigzag configurations have been observed with molecular resolution using a low-temperature STM. A hydroxyl-terminated model has been proposed to explain the stability in these water chains, based on common features in high-resolution STM images. The model is further validated by manipulation with STM tip in which both chain reconfiguration and proton transfer are observed. These simplest water chains on noble metal surfaces provides a desired platform in studying hydrogen bond networks on nearly free two-dimensional environment, and the discrepancy between theoretical simulation and experimental observation may stimulate new insights in nature of hydrogen bond networks, as well as their dynamic behaviors. References The interaction of water with solid surfaces: fundamental aspects revisitedWater adsorption and the wetting of metal surfacesWater at InterfacesWater Diffusion and Clustering on Pd(111)Ice nanoclusters at hydrophobic metal surfacesVibrational study of water dimers on Pt(111) using a scanning tunneling microscopeWater Dimer Diffusion on Pd111 Assisted by an H-Bond Donor-Acceptor Tunneling ExchangeWater clusters on Cu(110): Chain versus cyclic structuresReal-space imaging of interfacial water with submolecular resolutionInducing Transient Charge State of a Single Water Cluster on Cu(111) SurfaceWater–Ice Analogues of Polycyclic Aromatic Hydrocarbons: Water Nanoclusters on Cu(111)Identifying Few-Molecule Water Clusters with High Precision on Au(111) SurfaceAnisotropic Water Chain Growth on Cu(110) Observed with Scanning Tunneling MicroscopyA one-dimensional ice structure built from pentagonsUltrahigh-resolution imaging of water networks by atomic force microscopyNovel Water Overlayer Growth on Pd(111) Characterized with Scanning Tunneling Microscopy and Density Functional TheoryPentagons and Heptagons in the First Water Layer on Pt(111)Nucleation-Limited Dewetting of Ice Films on Pt(111) c ( 2 × 2 ) Water-Hydroxyl Layer on Cu(110): A Wetting Layer Stabilized by Bjerrum DefectsWater growth on metals and oxides: binding, dissociation and role of hydroxyl groupsWater adsorption on metal surfaces: A general picture from density functional theory studiesLight-Induced Structural Changes in Photosynthetic Reaction Center: Implications for Mechanism of Electron-Proton TransferAnisotropy in the Adsorption of H 2 O at Low Coordination Sites on Pt(111)Tunneling dynamics of a hydroxyl group adsorbed on Cu(110)H-atom relay reactions in real spaceWater-Mediated Proton Hopping on an Iron Oxide SurfaceDirect visualization of concerted proton tunnelling in a water nanoclusterDissociative Adsorption of Water at (211) Stepped Metallic Surfaces by First-Principles SimulationsGeneralized Gradient Approximation Made SimpleVan der Waals density functionals applied to solids
[1] Henderson M A 2002 Surf. Sci. Rep. 46 1
[2] Hodgson A and Haq S 2009 Surf. Sci. Rep. 64 381
[3] Björneholm O et al 2016 Chem. Rev. 116 7698
[4] Mitsui T, Rose M, Fomin E, Ogletree D F and Salmeron M 2002 Science 297 1850
[5] Michaelides A and Morgenstern K 2007 Nat. Mater. 6 597
[6] Motobayashi K, Matsumoto C, Kim Y and Kawai M 2008 Surf. Sci. 602 3136
[7] Ranea V A et al 2004 Phys. Rev. Lett. 92 136104
[8] Kumagai T et al 2011 J. Chem. Phys. 134 024703
[9] Guo J et al 2014 Nat. Mater. 13 184
[10] Guo Y, Ding Z, Sun L, Li J, Meng S and Lu X 2016 ACS Nano 10 4489
[11] Liriano M L et al 2017 J. Am. Chem. Soc. 139 6403
[12] Dong A, Yan L, Sun L, Yan S, Shan X, Guo Y, Meng S and Lu X 2018 ACS Nano 12 6452
[13] Yamada T, Tamamori S, Okuyama H and Aruga T 2006 Phys. Rev. Lett. 96 036105
[14] Carrasco J et al 2009 Nat. Mater. 8 427
[15] Shiotari A and Sugimoto Y 2017 Nat. Commun. 8 14313
[16] Cerdá J et al 2004 Phys. Rev. Lett. 93 116101
[17] Nie S, Feibelman P J, Bartelt N C and Thürmer K 2010 Phys. Rev. Lett. 105 026102
[18] Thürmer K and Bartelt N C 2008 Phys. Rev. Lett. 100 186101
[19] Forster M, Raval R, Hodgson A, Carrasco J and Michaelides A 2011 Phys. Rev. Lett. 106 046103
[20] Salmeron M et al 2009 Faraday Discuss. 141 221
[21] Meng S, Wang E G and Gao S 2004 Phys. Rev. B 69 195404
[22] Stowell M H B et al 1997 Science 276 812
[23] Morgenstern M, Michely T and Comsa G 1996 Phys. Rev. Lett. 77 703
[24] Kumagai T et al 2009 Phys. Rev. B 79 035423
[25] Kumagai T et al 2012 Nat. Mater. 11 167
[26] Merte L R et al 2012 Science 336 889
[27] Meng X et al 2015 Nat. Phys. 11 235
[28] Peköz R and Donadio D 2017 J. Phys. Chem. C 121 16783
[29] Perdew J P, Burke K and Ernzerhof M 1996 Phys. Rev. Lett. 77 3865
[30] Klimeš J, Bowler D R and Michaelides A 2011 Phys. Rev. B 83 195131
[31]Discussion with Professor Sheng Meng at Institute of Physics, CAS (unpublished results)