Towards High‐Safe Lithium Metal Anodes: Suppressing Lithium Dendrites via Tuning Surface Energy

The formation of lithium dendrites induces the notorious safety issue and poor cycling life of energy storage devices, such as lithium–sulfur and lithium–air batteries. We propose a surface energy model to describe the complex interface between the lithium anode and electrolyte. A universal strategy of hindering formation of lithium dendrites via tuning surface energy of the relevant thin film growth is suggested. The merit of the novel motif lies not only fundamentally a perfect correlation between electrochemistry and thin film fields, but also significantly promotes larger‐scale application of lithium–sulfur and lithium–air batteries, as well as other metal batteries (e.g., Zn, Na, K, Cu, Ag, and Sn).


Introduction
The last century has witnessed the soaring development of Li primary batteries (Figure 1). The concept of Li second battery was initialized in 1962. [1] However, the notorious safety issue

Models and Mechanism of Li Dendrites
Several theoretical models have been proposed to describe formation and growth behaviors of Li dendrites, including phase-field model, [9] solid electrolyte interphase (SEI) model, [10] deposition/dissolution model, [11] and charge induced growth model. [12] As shown in Table 2, several aspects can be clarified. These models are catalogued into the following two basic modes (rooted from which phase-field model is a mathematical simulation calculation) from a point view of the dendrite formation. 1) Models in thermodynamics mode, that is, deposition/dissolution model [11] a) Li metal is deposited beneath the SEI film. b) Supplied with an external power, Li ions in the electrolyte transport to Li metal surface through the protective SEI film. The deposition sites on the protective film exhibit a higher Li + conductivity. As a result, crystal defects and grain boundaries in the SEI initiate the continuous deposition of Li. c) The mechanical stress within the Li metal anode induces an asymmetrical deposition of Li, resulting in the formation of Li dendrites.
Supposing that lithium dendrite stems from the mechanical stress, the following Laplace's Equation (1) applies in this model based on the droplet theory of homogeneous nucleation, to be detailed in Section 4: where ΔP is the pressure difference of the surface, γ is the surface tension on a lithium surface, and R 1 and R 2 are any two orthogonal direction radius of the surface curvature. Yamaki et al. set out the dendrites form because of the surface tension effects, and proposed that the surface tension of protective film must be higher than 0.2 Nm -1 by simulation. [11] On the base of this model, a substantial film coating on Li metal can block the Li dendrite obviously.
2) Charge induced growth model [12] (one of the electrochemistry models) Li dendrite growth is considered the adoption of one dimensional nanostructures growth by electrodeposition. [13] Through applying the charge induced growth model, as the Li ions are deposited in combination with the surface charge, the inhomogeneous charge distribution leads to the Li dendrite. This model is on the basis of the Nernst Equation (2): where E Red is a reduction potential ( Red φ E is the standard reduction potential), R is the universal gas constant (8.31 J K -1 mol -1 ), T is the absolute temperature, α is the chemical activity for the relevant species (α Red is for the reductant and α Ox is for the oxidant), F is the Faraday constant (9.65 × 10 4 C mol -1 ), and n is the number of transferred electrons. On the basis of this model, another metal cation (M + ) may have an effective reduced potential lower than that of Li + if M + has a chemical activity α lower than that of Li + . Consequently, it indicates that the electrolyte additive cation should have an effective reduction potential lower than that of Li ion. Although this model does not unravel the dendrite origin, it provides a roadmap to a self-healing electrostatic shield to make the dendrite free.
3) Models in kinetics mode, e.g., SEI model (one of the electrochemistry models) [14] consequence, the lack of Li + layer, coupled with the local space charge layer, is regarded as the main reason for the formation of Li dendrite. Together with many efforts made for this theory based on kinetics [16] aspects of Li dendrites, Chazalviel [15,17] proposed the widely accepted diffusion model (Equation (3)) to correlate the "Sand's time" τ with the transfer nature of Li + ions and electrons empirically as follows, where τ is the initial time of Li dendrites growth, C 0 is the initial concentration of Li salt, D is the diffusion coefficient, e is the electronic charge, and J is the effective electrode current density. μ a and Li µ + are the mobilities of anionic and Li + ,

1910∼1920
The initial study of lithium battery Lewis Li metal as electrode [68] 1962 The rising Li secondary battery Whittingham Li secondary battery in non-aqueous solution [1a] 1970s First commercialization of Li metal battery Sanyo Co. Lithium primary battery [48] 1976 Discovery of Li ions embed into the carbon Besenhard Agarwal Li metal anode neglected gradually in secondary battery [69] 1983 respectively. Hence the SEI region can be also treated as a kinetics-controlling area, on the condition that the SEI film locates within the thermodynamic stability. It is well accepted that a suitable SEI film is obtained to suppress the dendrite by affecting Li µ + in governing the Sand's time (τ). Meanwhile, the effective electrode current density (J) is correlated with the transport of electrons, leading to dendrite-free anodes. This theory model can largely strengthen our understanding of the Li dendrite mechanism, which enables a long-life and safe Li anode through the application of SEI film regulation [10c] or an ultralow current density. [14a] Although the merit of various models have been revealed the formation of Li dendrites, and they are complementary with each other, it is necessary to establish a general model in unraveling the thermodynamic and kinetic aspects of Li dendrites. In fact, the substrate (Li or other materials) and the SEI region (considered from Chazalviel theory) depend partially on the thermodynamics and kinetics factors during the plating, respectively. Herein we describe a film growth model in order to provide fresh insight into the Li dendrite with focusing on the thermodynamics aspect of Li dendrite. Moreover, we demonstrate that it enables an even more thorough understanding of Li dendrites and thereof an effective strategy is established in facilitating Li metal anode protection.

Exploration for Li Metal Protection
The soaring progress of Li metal batteries drives scientific communities to renaissance the Li metal protection. The main technological routes have been established (Figure 3): (1) A "hard film" is prepared to block the Li dendrites. [18] (2) Li is promoted to react with other materials [19] in order to form a targeting "soft film", such as SEI film for the suppression of the Li dendrite growth. (3) Additives are introduced to electrolyte to postpone/retard the dendrite growth [20] or even get a dendrite-free Li anode. [21] (4) The use of nanostructure to modulate the Li deposition behavior through ultralow current density. [14a] Although many efforts have been devoted to reduce the Li dendrites, there is a long way for the ultimate solving towards the protection of the Li metal anode. This is particularly critical for cycling of Li metal anode at very high current density. One representative work of "hard film" coating was proposed by Cui's group. [18a] The thin film with high ionic conductivity and high Young's modulus is introduced to the Li anode in order to block the Li dendrites and prevent the Li metal to immediately contact the electrolyte.
In terms of SEI film, Li salt mixtures offer a thin passivation layer which could suppress the formation of Li dendrites by the Archer's group. [14b,22] However, it is very difficult to form a stable passivation film on Li electrodes, [23] due to the thermodynamically unstable of Li metal in organic solvents. Therefore, the continuing target is to search for a matching SEI film [10a,24] or Li salt film. [20,25] Cheng et al. [26] propose the efficient use of stable solid electrolyte interphase in a high-efficiency battery in order to block the formation of Li dendrites. Alternatively, some additives can be introduced. Zhang and co-workers [27] reported that a dendrite-free lithium deposition was successfully achieved with the self-aligned nanorod structure through adding Cs + . The consumption of the electrolyte suffer from the immediate contacting between Li anode and electrolyte. An interesting work by Kim and co-workers [21b] illustrates that Li dendrites are controlled via a synergistic effect of multilayered graphene coating and an Cs + electrolyte additive. The use of nanostructured framework with high surface area and extraordinary conductivity renders a very low local current together with robust SEI in dual-salt organic electrolyte afford the smooth and continuous Li deposition without dendrite formation. [14a,28]

Film Growth Model
The growth of Li dendrites belong to a typical crystal growth in an organic solution. [29] It is well known that the Volbmer-Weber mode for growth of polycrystalline films, which comprises the island, network, and channel stages by CVD [30] as well as particularly for plasma enhanced chemical vapor deposition (PECVD) [31] towards vertically grown materials under the condition of an electric field. As shown in Figure 4, since PECVD offers an electric field, [32] a vertical film on substrate is achieved from the cooperation of the active species (e.g., the fragment of CH 4 and CH 3 + ) and transition metal catalyst [33] (e.g., Fe, Co, Ni). The resultant products can be the vertical carbon nanotubes (Figure 4), vertical graphene sheet (Figure 4).

www.advancedscience.com
Adv. Sci. 2017, 4, 1600168 www.advancedsciencenews.com Inspired by the PECVD advantages, in analogy to the aforementioned process, the active species (Li + ) and catalyst (the charge accumulation area), under the external electric field, work together, leading to the formation of Li dendrites (Figure 4). The formation and continuous growth of Li dendrites is attributed to the following scenarios: the Li + ions near the anode, promoted under the electric filed, induces Li deposition on the anode; as a consequence, an increase of surface energy occurs. We call it the film growth model and explore the dendrite issue by applying classical theory of film within a battery system later. Our film growth model is to address the Li dendrite issues via the surface energy [30,34] since the latter can be listed as one of the film growth. Based on the well-known capillarity of homogeneous nucleation, a solid nucleates from a prior unstable liquid by establishing a solid-liquid (s-l) interface; in analogue for the droplet theory of homogeneous nucleation (the deposition/dissolution model stem from this theory), a solid nucleates from vapor phase by establishing a solid-vapor (s-v) interface. These two scenarios are depicted as Figure 5A.
According to the Young's equation, cos γ γ γ θ = + sv fs fy (4) where f, s, and v denote the film, the substrate and the vapor, respectively. Then γ sv , γ fs , γ fv represent the interface energy between the two phases, θ is the contact angle. Based on Equation (4), it is well accepted that the film growth depends on two main issues: 1) the surface energy difference between the substrate and the film; 2) the lattice misfit between growth substrate and growing film, as shown in Figure 5B. As shown in Figure 5C, Li deposition is assumed as homogeneous progress of film growth, therefore, the lattice distortion is neglected. With the intermediation of the electric filed, the fluctuation of the surface energy attributed from the dynamics of SEI film induces the Li film growth towards the island type within the framework of Volmer-Weber mode. As shown in Figure 6A, there are m electrons on the anode surface, subjected to an external power supply n (V). Consequently, the electrical potential energy can be expressed as n × m (eV), where n is negative. The surface energy of substrate γ sub consists of two terms, the surface energy of a deposition film γ Li and the electrical potential energy. Since γ sub is less than γ Li , the Li dendrite is induced if the Li metal is electrodeposited. Figure 6B exhibits that the Li substrate surface is stable. The surface is free of any Li dendrite and therefore has a constant surface energy γ sub . After several charge/discharge cycling, the initialization of the Li dendrites leads to a large surface energy. Concurrently, γ den increases. As a result, the substrate becomes rough. The charge/discharge process suffers from the rise of surface energy, which results in the unfavorable low Coulomb efficiency (CE). As expressed by equation (5), sub d en where Q e stems from the partial energy of the electric field. The larger Q e becomes, the lower CE changes.

www.advancedscience.com
Adv. Sci. 2017, 4, 1600168 www.advancedsciencenews.com  Reproduced with permission. [64] 2) vertical graphene sheets. Reproduced with permission. [65] Copyright 2014, Elsevier; 3) Li dendrites. Reproduced with permission. [66] Copyright 2012, Nature Publishing Group. Figure 5. a) Scheme of basic atomistic nucleation on substrate surface during vapor deposition; [67] b) The stability regions of the three-film growth modes in coordination of surface energy differences between growing film and growth substrate (vertical)/the lattice misfit (horizontal); [67] a,b) redrawn after the reference. c) Scheme of correlation between the extent of Li dendrites and the surface energy difference.

Roadmaps to Dendrite-Free Li Anode through our Model
Different from the SEI model mentioned above, we suggest two promising routes to suppress the formation of Li dendrites on the basis of our film growth model derived from the Volmer-Weber theory. [35] The first direction lies to modulate the surface energy through tuning composition and morphology of Li metal. An excellent Li metal anode is expected through 3D substrates with continuous high surface energy. If γ sub increases, the additional barrier is bypassed for the Volmer-Weber mode. As a result, the deposited Li surface becomes smooth, which induces a symmetrical charge/discharge cycling and thereof a higher CE.

www.advancedscience.com
Adv. Sci. 2017, 4, 1600168 www.advancedsciencenews.com  In parallel, more importance has been practically attached to exploration of Li-alloy anode, which are involved with the intercalation and de-intercalation of Li in Li metal battery, [36] LiAl and other Li based alloys. [37] However, for even higher Li content alloy, [38] there is a positive effect in Li protection for robust cycling. The metal mixture in anode results in an increase of surface energy, [39] which promotes a layer growth instead of an island growth. Consequently, the Li dendrite formation is postponed or even completely suppressed.
In order to avoid the occurrence of potential active sites for Li dendrite nucleates, the Li metal surface should be smooth. [40] Recently, both Li powder [41] and Li foam [42] anode are treated as a positive way to suppress Li dendrite formation, as it is common knowledge that the nanostructured counterpart has a larger surface area and surface energy. This can be well explained by our surface energy scenarios. That is, more importance should be attached to controlling and governing of the substrate in Li deposition, e.g., the Cui's latest work. [43] Specifically, Guo's group successfully suppressed Li dendrite [44] by applying a 3D current collector, explained by the electric field effect as anode substrate. The use of 3D porous Cu exhibits a high surface energy, and thereof the mechanism is factually explained by the tuning of surface energy and film growth model. Layer-bylayer growth Li film induces a few dead-Li system, resulting in a long-life anode, as shown in Figure 1. Modulating the anode surface with high surface energy is the very roadmap to obtain a dendrite-free Li metal anode. Very recently, the research work (Figure 7a) of surface-modified three-dimensional (3D) substrate with a "lithiophilic" coating, [43a] presents a novel strategy for the fabrication of metal-scaffold composite. Specifically, the resultant material was employed as anodes in Li metal batteries, exhibiting superior performance compared with bare lithium metal anodes. On the basis of our model, the surface energy (in J cm -2 or eV cm -2 for the surface perpendicular to the electric field during the Li plating) of 3D substrate yields a continuous surface state, as shown in Figure 7b. A long-life and safe Li metal anode can be fabricated.
The second way is reducing the electrical potential energy n × m (eV), which contributes to the decrease of the term γ sub -γ Li . According to our model, the Li dendrites enable being thereof suppressed. This event is achieved by altering the charge mode towards tuning the electrical potential energy, of which the application of pulse charging [45] or charging at a lower voltage would be some of the best choices. The pulse charging mode would lead to a process of relaxation surface charge, and tailoring of the electrical potential energy. The use of gel electrolytes and polarized framework is another route to tune the transport and nucleate behavior of Li ions onto Li metal anode. [10a,46] The use of high current is also verified to protect the robust use of Li metal anode. [47]

Concluding Remarks
The safe use of Li metal is strongly considered for high energy batteries. We introduced the concept of surface energy to understand to the formation and growth of Li dendrites in Li metal batteries. A high charge/discharge current is required if the increased surface energy cannot override the effect of electric field, alternatively a hard film is required to block the Li dendrites. However, the energy relaxation changes to the thermal energy, which is a risk for a working cell. In addition, the proposed model is themed on the effect of thermodynamics, which is the emerging supplement to the kinetics model (SEI model analyzed by Chazalviel theory). The way addressing the kinetics effects focus on the homogenous surface (a robust and homogenous SEI film), whereas the way addressing the thermodynamic effects should gear to the continuous surface (the state of substrate). Consequently, the golden rule lies on both of them for Li protection.
Li anode protection doesn't barely regulate to dendritefree Li metal anodes, the problem of electrolyte consumption should be addressed to avoid an unexpected reaction between Li and electrolyte. A continuous surface possessing high surface energy during the charge/discharge cycling is promising to suppress the lithium dendrites. Such concept may be extended to other metal batteries, such as Zn, Na, K, Cu, Ag and Sn.