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.
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.
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.
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