Letter pubs.acs.org/NanoLett
Lewis Acid−Base Interactions between Polysulfides and Metal Organic Framework in Lithium Sulfur Batteries Jianming Zheng,†,∥ Jian Tian,†,∥ Dangxin Wu,‡ Meng Gu,§ Wu Xu,† Chongmin Wang,§ Fei Gao,‡ Mark H. Engelhard,§ Ji-Guang Zhang,† Jun Liu,† and Jie Xiao*,† †
Energy and Environmental Directorate, ‡Fundamental and Computational Science Directorate, and §Environmental Molecular Science Laboratory, Pacific Northwest National Laboratory, 902 Battelle Boulevard, Richland, Washington 99352, United States S Supporting Information *
ABSTRACT: Lithium−sulfur (Li−S) battery is one of the most promising energy storage systems because of its high specific capacity of 1675 mAh g−1 based on sulfur. However, the rapid capacity degradation, mainly caused by polysulfide dissolution, remains a significant challenge prior to practical applications. This work demonstrates that a novel Ni-based metal organic framework (Ni-MOF), Ni6(BTB)4(BP)3 (BTB = benzene-1,3,5-tribenzoate and BP = 4,4′-bipyridyl), can remarkably immobilize polysulfides within the cathode structure through physical and chemical interactions at molecular level. The capacity retention achieves up to 89% after 100 cycles at 0.1 C. The excellent performance is attributed to the synergistic effects of the interwoven mesopores (∼2.8 nm) and micropores (∼1.4 nm) of Ni-MOF, which first provide an ideal matrix to confine polysulfides, and the strong interactions between Lewis acidic Ni(II) center and the polysulfide base, which significantly slow down the migration of soluble polysulfides out of the pores, leading to the excellent cycling performance of Ni-MOF/S composite. KEYWORDS: Metal organic framework, sulfur composite, polysulfides confinement, Lewis acidic center, cycle life, lithium sulfur battery
L
different kinds of hosting materials,8,9,12−16 sulfur cathode surface modification,17 electrolyte modification,18−20 and anode protection by employing LiNO3 as the electrolyte additive.5,21−24 While the development of new electrolyte/additive and lithium anode protection remain challenges for a long history in lithium batteries,25,26 more progress was achieved in the sulfur cathode modulation. A variety of carbons, for example, mesoporous carbons,8,27−29 graphene,3,30,31 and hollow carbon nanofiber,5 have been adopted as hosting substrates to provide good conductivity and uniform dispersion of sulfur within the frameworks. Recent work from this group further correlates carbon properties with the real current density and revisits their functions at different current densities.32 Polymers,1,33,34 porous aromatic framework (PAF),35 intercalation compounds,36 and silica4 have also been extensively investigated for Li−S battery system. Another type of high surface area hosts, metal organic framework (MOF), however, has received much less attention probably due to its poorly conducting nature compared to the carbon scaffold.28 Recently, MOF MIL100(Cr) mixed with sulfur has been reported as the composite cathode.12 The confinement of polysulfides in the bimodal pores of MIL100(Cr) improves the cycling performances of sulfur cathode compared with the mesoporous carbon host, but the sulfur content in the whole
ithium-sulfur (Li−S) batteries have attracted increasing attention because of their high theoretical capacity, natural abundance and environmental friendliness.1−7 Li−S battery operates by reduction of sulfur during discharge to form lithium polysulfides with different chain lengths and finally to produce insoluble Li2S2 or Li2S. Assuming complete reduction of sulfur to form Li2S, the theoretical specific capacity and energy from Li−S batteries are 1675 Ah kg−1 and 2650 Wh kg−1, respectively, substantially higher than those of state-of-the-art lithium ion batteries. However, fast capacity degradation remains a significant challenge prior to the practical applications of Li−S batteries. The poor cycle life of Li−S batteries originates from the formation of soluble intermediate discharge products, a series of polysulfides Li2Sx (x > 2), which could easily diffuse to the lithium metal anode and participate in the well described “sulfur shuttle mechanism” reactions.1,8 The undesired shuttle effect leads to the low charge−discharge efficiency and precipitation of insoluble/insulating Li2S2/Li2S on lithium metal anode, causing the loss of energy-bearing materials. Similar precipitation of highly insulating Li2S2/Li2S also occurs on the cathode at deep discharge, leading to a largely increased overpotential, low Coulombic efficiency, as well as irreversible capacity loss.9−11 All those detrimental effects are then reflected by the largely shortened lifespan of Li−S batteries. In recent years, many efforts have been pursued to overcome the hurdles in Li−S battery technology. Various approaches have been proposed, spanning from immobilization of sulfur in © 2014 American Chemical Society
Received: December 19, 2013 Revised: March 19, 2014 Published: April 4, 2014 2345
dx.doi.org/10.1021/nl404721h | Nano Lett. 2014, 14, 2345−2352
Nano Letters
Letter
Figure 1. (a) Crystal structure of Ni-MOF containing two different types of pores represented by dark yellow sphere and blue sphere: mesopore (yellow sphere indicates pore volume; gray, C; red:,O; green, Ni; blue, N); micropore (blue sphere indicates the pore volume). (b) N2 adsorption/ desorption isotherms of pristine MOF and MOF/S composite; inset of panel (b) shows the corresponding pore size distribution. (c) XRD patterns of pure sulfur, pristine MOF and MOF/S composite. (d) HAADF Z-contrast image and carbon and sulfur EELS maps.
model is shown in Figure 1a. The spheres inside the framework are two different types of pores presented in the structure: a smaller pore (blue) with approximately 13.8 Å in diameter and a larger dodecahedral pore (yellow) with a diameter of approximately 27.6 Å, which are derived from the structural data. Each face of the larger dodecahedral pores is connected to a small pore through a microporous pentagonal window with a size of ∼13.8 Å. The pore structure of Ni-MOF is characterized by Type I39 N2 adsorption/desorption isotherms with a very high specific surface area (SA) of 5243 m2 g−1 calculated by BET method (Figure 1b) and a pore size distribution peak located at ∼1.82 nm (combination of micropores and small mesopores) (inset of Figure 1b). The BET surface area of pristine Ni-MOF is among the highest ones that have been reported for porous MOFs, such as NU-100 (SALm = 1502 m2 g−1),40 MIL100(Cr) (SABET = 1485 m2 g−1),12 and MOF-210 (SABET = 6240 m2 g−1).41 The cumulative pore volume of NiMOF is determined to be 2.15 cm3 g−1. This translates to maximum sulfur loading of 82 wt % in this Ni-MOF based on the density of sulfur (2.07 cm3 g−1). The high surface area and unique bimodal porous structure of Ni-MOF are very favorable for the efficient confinement of intermediate soluble polysulfides generated during charge/discharge processes. Figure 1c shows the XRD patterns of pristine sulfur, NiMOF, and Ni-MOF/S composite. The elemental sulfur powder used for synthesis of Ni-MOF/S composite shows XRD pattern that could be assigned to Fddd orthorhombic structure.42 The original Ni-MOF shows several characteristic peaks below 20° in 2θ, which is identical to the simulated pattern from its single crystal structure. XRD pattern of heat-treated Ni-MOF/S composite contains no obvious characteristic peaks of sulfur, which indicates a good dispersion of sulfur into the inner pores
cathode electrode is relatively low which makes it difficult for the effective evaluation of the function of MOF. In addition to the pore confining effects from MOF, the existence of transition metal ions as well as the organic framework may also alter the nature of fundamental interactions between polysulfides and the host framework during the electrochemical process, which is worth to be further explored to fully understand the roles of MOF in Li−S cells. In the present work, we report a novel nickel-based MOF (Ni-MOF) for sulfur impregnation. This new Ni-MOF belongs to a family of highly porous MOFs with framework composition of [M6(BTB)4(BP)3] (M = Co(II), Zn(II), Cu(II), and Ni(II), BTB = benzene-1,3,5-tribenzoate and BP = 4,4′-bipyridyl).37,38 These noninterpenetrated frameworks possess a cross-linked pto-like topology, uniform mesopores, very high surface areas and pore volumes, which are reported to show excellent performance in the adsorption of a variety of gases, such as methane, hydrogen, and n-butane.37 Here, we demonstrate that polysulfides can be effectively harnessed by this novel Ni-MOF, displaying remarkably improved cycling performances. In addition to the structural benefits of Ni-MOF, the fundamental interactions between MOF building blocks and polysulfides are also discussed in detail. Because of the versatile structures and the cation-ligand coordination chemistry in MOFs, this work provides insights to design new open structured cathode materials for Li−S battery chemistry. The Ni-MOF used in this work is a noninterpenetrated porous framework with composition of [Ni6(BTB)4(BP)3]. It is constructed by linking BTB and BP as a three-connected node and a two-connected linker, respectively, with a Ni(II) paddlewheel (Ni2(COO)4) secondary building unit (SBU) as a sixconnected node. The framework of Ni-MOF depicted in “stick” 2346
dx.doi.org/10.1021/nl404721h | Nano Lett. 2014, 14, 2345−2352
Nano Letters
Letter
of Ni-MOF instead of residing on the outer surface. A similar phenomenon was also found when sulfur was infiltrated into porous carbon substrates.29,32 On the other hand, the XRD peaks that could be assigned to Ni-MOF matrix become less pronounced in the MOF/S composite due to the pore filling of MOF by sulfur. BET surface area of Ni-MOF decreases dramatically from 5243 to 514 m2 g−1 (Figure 1b) after sulfur impregnation, accompanied by a significant drop of pore volumes from 2.15 to 0.27 g−1 cm3. Figure 1d shows the highangle annular dark-field (HAADF) Z-contrast image and electron energy loss spectroscopy (EELS) maps of Ni-MOF/ S composite. HAADF Z-contrast image shows that the NiMOF/S composite is composed of aggregated nanoparticles with particle size of about 25 nm. The EELS maps of sulfur suggest a uniform distribution of sulfur among the Ni-MOF structural units. Figure 2a displays the thermogravimetric analysis (TGA) curve for the Ni-MOF/S composite in which TGA curves for pure sulfur and pristine Ni-MOF are also presented for comparison. Pristine activated Ni-MOF shows excellent thermal stability, with negligible weight loss before 420 °C, indicating that Ni-MOF is stable up to 420 °C and not supposed to decompose at 155 °C. Ni-MOF treated at 155 °C for 12 h shows exactly the same XRD pattern with that of pristine Ni-MOF, indicating that the crystal structure of NiMOF is stable at the temperature for sulfur melting and infiltration (Figure S1 in the Supporting Information). The sulfur content of Ni-MOF/S composite is determined to be ∼60 wt %, which is consistent with the predetermined value during preparation. In addition, the weight loss profile of NiMOF/S composite overlaps with that of pure sulfur at temperature lower than 310 °C, strongly suggesting that there was no chemical reaction occurring between Ni-MOF and sulfur when the mixture was heated at 155 °C. To further substantiate this point, the samples of Ni-MOF before and after sulfur impregnation were characterized by Fourier transform infrared spectroscopy (FTIR) and the recorded spectra are shown in Figure 2b,c. The peak positions in IR spectrum of pristine Ni-MOF are in good agreement with those reported by early researcher.38 The existence of absorption peaks at 3420 and 1621 cm−1 could be attributed to the vibration of adsorptive water (Figure 2b). The strong, wide band at 3420 cm−1 is mainly the stretching vibration mode O−H group and the absorption peak at 1621 cm−1 is indexed to the O−H bending band.43 The strong and broad absorption peaks at 1610−1580, 1535, and 1400 cm−1 can be assigned to the skeleton stretching resonance of the aromatic ring (phenyl ring and pyridine ring),44 that is, the stretching vibration of CC, CN bonds of BTB and BP of Ni-MOF (inset of Figure 2b).45 The wide peaks at 1610−1580 and 1400 cm−1 also contain the absorption bands of asymmetric stretching (∼1600 cm−1) and symmetric stretching (∼1400 cm−1) of carboxylate (−COO−) groups of BTB.46,47 The bands at 1180 and 1219 cm−1 are ascribed to the C−C/C−N stretching vibrations,1 while those at 781 and 1016 cm−1 are due to −C−H out-of-phase and inphase bending vibration.48 The weak, broad absorption at 484 cm−1 characterizes the combination of Ni−O stretching vibration and Ni−N stretching vibrations (Figure 2c).49,50 After sulfur impregnation, three peaks at 515, 492, and 470 cm−1 are observed for Ni-MOF/S, which is different from that for pristine Ni-MOF with only one broad peak observed. The weak absorption at 470 cm−1 could be undoubtedly assigned to the S−S vibration mode of elemental sulfur (S8), while the one
Figure 2. (a) TGA of pure sulfur, pristine Ni-MOF, and Ni-MOF/S composite. (b) FTIR spectra of elemental sulfur, pristine Ni-MOF, and Ni-MOF/S composite; inset of panel (b) is the enlarged part of the IR spectra ranging from 900 to 1800 cm−1. (c) Enlarged part of the IR spectra ranging from 350 to 900 cm−1.
at 492 cm−1 and the newly emerged peak at 516 cm−1 could be explained by the fact that the sulfur impregnation/absorption largely reduces the flexibility of the Ni-MOF framework, leading to the separation of Ni−O and Ni−N absorption bands. Except the minor peaks located below 550 cm−1, all the rest of the absorption bands of Ni-MOF/S composite match well with 2347
dx.doi.org/10.1021/nl404721h | Nano Lett. 2014, 14, 2345−2352
Nano Letters
Letter
Figure 3. (a) Charge/discharge profile evolution for Ni-MOF/S composite during cycling at 0.1 C. (b) Cycling performance of Ni-MOF/S composite at 0.1 and 0.2 C rates at a voltage range of 1.5−3.0 V; inset of panel (b) is the corresponding Coulombic efficiency during cycling. (c) EIS spectra of Ni-MOF/S composite in Li−S battery during cycling at 0.2 C; inset of panel (c) shows full impedance spectrum of Ni-MOF/S composite measured before cycling. (d) Rate performance of Ni-MOF/S composite.
those observed for pristine Ni-MOF, revealing that Ni-MOF did not react with sulfur during heating treatment at 155 °C. The possibility of reaction between Ni-MOF and sulfur can be further ruled out by the absence of absorption peaks at 1090 cm−1 (C−S stretching),51 1129 cm−1 (ring-sulfur stretching),52 and the peak related to Ni−S bond absorption (
