Carbon dioxide in the cage: manganese metal–organic frameworks for high performance CO2 electrodes in Li–CO2 batteries

Siwu Li, Yu Dong, Junwen Zhou*, Yuan Liu, Jiaming Wang, Xing Gao, Yuzhen Han, Pengfei Qi and Bo Wang*
Beijing Key Laboratory of Photoelectronic/Electrophotonic Conversion Materials, Key Laboratory of Cluster Science, Ministry of Education, School of Chemistry and Chemical Engineering, Beijing Institute of Technology, Beijing 100081, P. R. China. E-mail:;

Received 7th February 2018 , Accepted 12th March 2018

First published on 13th March 2018

Improving the reversibility and energy efficiency of Li–CO2 electrochemistry would help in developing practical Li–air batteries capable of providing stable power supply in the presence of CO2. However, it is hard for most existing electrodes to convert CO2 effectively (high discharge capacity) and efficiently (low charge potential). Herein, we have, for the first time, identified the potential of metal–organic frameworks (MOFs) as porous catalysts in CO2 electrodes, taking advantage of their high capability for CO2 capture and monodispersed active metal sites for Li2CO3 decomposition. In particular, eight porous MOFs (Mn2(dobdc), Co2(dobdc), Ni2(dobdc), Mn(bdc), Fe(bdc), Cu(bdc), Mn(C2H2N3)2, and Mn(HCOO)2) and two nonporous materials (MnCO3 and MnO) were investigated. Among them, Mn2(dobdc) achieves a remarkable discharge capacity of 18[thin space (1/6-em)]022 mA h g−1 at 50 mA g−1, while Mn(HCOO)2 retains a low charge potential of ∼4.0 V even at 200 mA g−1 for over 50 cycles. The Li–CO2 electrochemistry on MOF electrodes was investigated with an arsenal of characterization techniques, including X-ray diffraction, scanning electron microscopy, electrochemical impedance spectroscopy, Raman spectroscopy and in situ differential electrochemical mass spectrometry. The findings provide useful design principles for improving the reversibility and energy efficiency of Li–CO2 electrochemistry and will herald the advent of practical technologies that enable energy-efficient utilization of CO2 for electrochemical energy storage and supply.

Broader context

Li–CO2 batteries, an emerging battery technology, hold promise in coupling CO2 utilization with electrical energy storage. They also share fundamental understandings and technical solutions with Li–air batteries in the decomposition of Li2CO3, an electrochemically inactive product that undermines the reversibility of the batteries. However, great challenges still remain in the development of Li–CO2 batteries—especially high charge overpotentials, limited capacity and poor cycling stability, requiring high performance CO2 electrodes to be created. A well-designed CO2 electrode must be multifunctional: in the discharge process, it can enrich CO2 and control the deposition of Li2CO3, while in the charge process, it should be capable of activating Li2CO3. Metal–organic frameworks (MOFs) are a rapidly growing family of multifunctional porous materials with great structural versatility and functional tunability. In this work, we have, for the first time, demonstrated the potential of MOFs in Li–CO2 batteries. MOFs with Mn(II) centers significantly reduce the charge overpotential, while high porosity with exceptional CO2 uptake tends to expand their discharge capacity. The deposition and decomposition of Li2CO3 are sensitively affected by a MOF's porosity and isosteric heat of CO2 adsorption. Our findings provide useful insight into and guidelines for designing novel electrode materials for high performance Li–CO2 batteries.

1. Introduction

Electrochemical energy storage (EES) technologies play a key role in powering the modern society.1–6 Among various EES technologies, metal–air batteries hold promise in terms of high volumetric and gravimetric energy density.7–10 In confined spaces, such as aircraft, space shuttles, submarines and so on, portable batteries with high energy density are urgently demanded; metal–air, especially Li–air, batteries are extremely promising in such circumstances. Despite exciting progress in recent years, formidable challenges still remain in the Li–air system, including low energy efficiency, a short cycle life and decomposition of electrolyte. Things may become even worse when other reactive components (H2O and CO2) are present in the air.11–13 In the confined spaces mentioned above, the utilization of pure O2 as an active material for Li–air batteries will be costly because it competes against human beings in O2 consumption; meanwhile, a high concentration of CO2 (up to ∼1000 ppm, three times the normal level) can be accumulated as a result of human activity, which is hazardous to human health.14,15 CO2 in the air would complicate the Li–O2 electrochemistry, and its discharge product Li2CO3 often causes severe passivation of O2 electrodes.16 Therefore, understanding the electrochemistry of CO2 would be greatly helpful for developing practical Li–air batteries that can provide stable power supply in the presence of CO2.17,18 In some circumstances, CO2 is even suitable to be directly utilized as an active material. Over the past few years, there has been growing research interest in the electrochemistry of Li–CO2 batteries. In 2011, Takechi et al. pioneered the use of CO2 as an active material in a lithium battery, but the working gas was indeed an O2/CO2 mixture.19 Li et al. have, for the first time, demonstrated the reversible cycling of a Li–CO2 battery in pure CO2.20 Thereafter, the research interest in the Li–CO2 system is ever growing, as it not only shares fundamental understandings and technical solutions in the decomposition reaction of Li2CO3 with Li–air batteries, but also provides options for power supply in versatile applications.16,18

The CO2 electrode is the core component that determines the performance of a Li–CO2 battery, in terms of capacity, working potentials, and cycling stability. The major challenge in building high-performance CO2 electrodes stems from the electrochemical inactivity of the discharge product Li2CO3, making it difficult to convert CO2 effectively (high discharge capacity) and efficiently (low charge potential). So far, several materials have been explored in CO2 electrodes to promote the electrochemistry of Li2CO3, including various carbonaceous materials and nanoparticles of noble metals.20–26 Zhou et al., for example, reported CO2 electrodes based on carbon nanotubes (CNTs) and graphene, which operate for <30 cycles at 50 and 100 mA g−1 with a capacity limit of 1000 mA h g−1, but the charge potentials are high (up to ∼4.5 V), leading to a large voltage hysteresis of >1.5 V.22,23 Zhou et al. introduced Ru nanoparticles into CO2 electrodes to lower the charge potential, which can be reduced to ∼4.0 V at 100 mA g−1 with good cycling stability (∼70 cycles).25 However, the high cost of Ru metal would be disadvantageous in practical applications.

From our perspective, a well-designed CO2 electrode must bear the following functions: (1) capturing and enriching CO2, (2) restricting the deposition of Li2CO3, and (3) catalysing the decomposition of Li2CO3. (1) and (2) together ensure a high practical discharge capacity of the CO2 electrode (high energy density), while (2) and (3) warrant a low overpotential for recharging the battery (high energy efficiency); (1)–(3) contribute to an efficient and reversible conversion between CO2 and Li2CO3. Thus it is of central importance to design a multifunctional porous cathode material with regard to the above requirements. As a rapidly developing family of porous materials with great structural versatility and functional tunability, metal–organic frameworks (MOFs) have been extensively investigated in CO2 capture, separation and catalytic conversion into useful compounds.27–30 Identically, it would be a fascinating strategy to relieve the environmental issues of carbon emissions if the CO2 captured by the pores of MOFs can be utilized for power supply.31,32 In pursuing a multifunctional porous material for high-performance CO2 electrodes, we targeted the MOF family for the following considerations: (1) MOFs have strong capability in CO2 capture; the affinity with and the uptake amount of CO2 can be easily tuned by adjusting the pore size and functionality of MOFs. (2) MOFs possess regularly arranged micropores, providing separated and strictly confined nanospaces that can dictate the nucleation and early-stage deposition of Li2CO3, and induce its growth towards a uniform nanostructured morphology. (3) Catalytic metal centers can be monodispersed in the highly porous framework, facilitating a complete and efficient decomposition of the nanostructured Li2CO3. (4) The metal centers and pore structures of MOFs can be systematically varied for drawing out the structure–performance relationships, making rational design and optimization possible (Fig. 1).

image file: c8ee00415c-f1.tif
Fig. 1 Scheme of the advantages of a Li–CO2 battery equipped with a MOF-based CO2 electrode.

Herein, we have, for the first time, identified the potential of MOFs as porous catalysts in CO2 electrodes. In particular, eight porous MOFs (Mn2(dobdc), Co2(dobdc), Ni2(dobdc), Mn(bdc), Fe(bdc), Cu(bdc), Mn(C2H2N3)2, and Mn(HCOO)2) and two nonporous materials (MnCO3 and MnO) were investigated. Among them, Mn2(dobdc) achieves a remarkable discharge capacity of 18[thin space (1/6-em)]022 mA h g−1 at 50 mA g−1, and Mn(HCOO)2 retains a low charge potential of ∼4.0 V even at 200 mA g−1 for over 50 cycles. The Li–CO2 electrochemistry on MOF electrodes was investigated with an arsenal of characterization techniques, including X-ray diffraction (XRD), scanning electron microscopy (SEM), electrochemical impedance spectroscopy (EIS), Raman spectroscopy and in situ differential electrochemical mass spectrometry (DEMS). The results show that (1) the selection of Mn(II) as metal centers is essential in alleviating the charge overpotential, (2) the upper limit of discharge capacity exhibits an increasing trend with increasing porosity and CO2 uptake of MOFs, (3) the isosteric heat of CO2 adsorption of MOFs has an impact on charge polarization in the way of, presumably, dictating the energy penalty for CO2 evolution, and (4) the deposition of Li2CO3 upon discharge is sensitively affected by MOFs, with which it tends to grow into a nanostructured morphology with a uniform dispersion in the whole electrode. The last effect is pivotal in activating the inert Li2CO3, which mitigates the growth of the electrode's charge transfer resistance and guarantees a facile and complete conversion of Li2CO3 back to CO2 in the charge process. The unique advantages of MOFs in the adsorption and enrichment of CO2 and the deposition and decomposition of Li2CO3 have endowed them with great potential as multifunctional catalysts for high performance Li–CO2 batteries in the future.

2. Experimental

2.1 Preparation of MOF materials

Ni2(dobdc), Co2(dobdc), Mn2(dobdc), Mn(bdc), Fe(bdc), Cu(bdc) and Mn(C2H2N3)2 were synthesized according to previously reported methods respectively.33–38 Mn(HCOO)2 was synthesized as follows: MnCl2·4H2O (1.04 g) and formic acid (1.5 mL) were dissolved in 70 mL of methanol to form a solution. 2.68 g of triethylamine was then rapidly added under vigorous stirring at room temperature for 30 min to form a white precipitate. The solid was collected by centrifugation and washed with methanol several times. After being dried, the powders were further activated under vacuum at 100 °C for 12 h.

2.2 Material characterization

XRD measurements were performed on a MiniFlex 600 diffractometer with a Cu-Kα X-ray radiation source (λ = 0.154056 nm). SEM observations were carried out on a JEOL model JSM-7500F under an accelerating voltage of 5.0 kV. N2 sorption isotherms were measured at 77 K on a Quantachrome Instruments ASiQMVH002-5 after pre-treatment by activating the samples under vacuum at 150 °C for 12 h. CO2 sorption isotherms were measured at 273, 283 and 298 K with the same instrument after pre-treatment. The discharged or recharged electrodes for XRD and SEM experiments were washed with tetraethylene glycol dimethyl ether (TEGDME) thoroughly to remove any residual lithium salt from the electrolyte prior to tests. Raman characterization was performed on a Horiba LabRAM HR Evolution Raman spectrometer with a 633 nm laser.

2.3 Electrode preparation and cell assembly

CO2 electrodes were prepared in a glove box filled with pure argon. Typically, 45 wt% of degassed MOF crystals, 45 wt% of carbon nanotubes (CNTs) and 10 wt% of polyvinylidene fluoride (PVDF) were blended to form a uniform slurry. The slurry was coated on a carbon paper (TGP-H-060 carbon paper, Toray) and dried in a vacuum oven at 80 °C for 12 h. The dried carbon paper was punched into 1.12 cm2 circular discs as working electrodes, with 0.3–0.5 mg of active material (the total mass of MOF material and CNTs) on each. The Mn2(dobdc)@Ni electrode for Raman investigation was prepared by growing Mn2(dobdc) on a Ni foam under the solvothermal conditions of Mn2(dobdc). Before the Raman tests, the Ni foam and Mn2(dobdc)@Ni electrodes were coated with Au (∼0.2 mg) by magnetron sputtering using a Au target. Li–CO2 cells were assembled in a glove box filled with high-purity argon, using CR2032-type coin cells with holes on the cathode side. Li foil was used as an anode and a piece of glass fiber (16 mm in diameter) soaked with 50 μL of electrolyte (1 mol L−1 LiTFSI dissolved in TEGDME) was used as a separator.

2.4 Electrochemical tests

The cells were tested in a glovebox filled with CO2 using a LAND-CT2001A tester. Galvanostatic tests were carried out subject to either a capacity limit or a cutoff voltage. All the capacity and current density values were normalized by the total mass of catalyst and CNTs. EIS experiments were conducted on an electrochemical workstation (CHI 760E: CH Instruments Inc.) with an amplitude of 5 mV in the frequency range of 1 MHz to 1 mHz. In situ DEMS was performed using a commercial quadrupole mass spectrometer (Pfeiffer Vacuum, ThermoStar) and a customized Swagelok-type DEMS cell. The DEMS system was purged with 1.0 sccm pure Ar for 10 h before charge.

3. Results and discussion

3.1 Properties of MOFs based on different transition metals in CO2 electrodes

In attempt to screen out a metal center that has activity in recharging the Li–CO2 system, we initiated our study with six MOFs: three members of the M-MOF-74 series (M2(dobdc) (M = Mn, Ni, Co; dobdc = 2,5-dioxido-1,4-benzenedicarboxylate)) and Mn(bdc), Fe(bdc) and Cu(bdc) (bdc = 1,4-benzenedicarboxylate). The M-MOF-74 series is a famous class of MOFs that have 1D hexagonal channels of 11–12 Å in diameter lined with open metal sites (Fig. 2a–c, left side),33 while Mn(bdc), Fe(bdc) and Cu(bdc) are built up by the same linker with permanent open channels.35–37 The five 3d transition metals Mn, Co, Ni, Fe and Cu have proved to be active centers in a variety of electrocatalytic systems, and thus are selected in our study.34,39–42 CO2 electrodes were prepared by blending degassed MOF powders with CNTs, a commercially available conductive agent that can help in providing robust electrode integrity and facilitating rapid mass transport. The electrodes were assembled into Li–CO2 cells with Li anodes and cycled galvanostatically at 50 mA g−1 with a capacity limit of 1000 mA h g−1. Fig. 2 shows the discharge–charge curves of the MOF electrodes. Among the six electrodes, Ni2(dobdc), Co2(dobdc), Fe(bdc) and Cu(bdc) are charged at 4.59, 4.53, 4.53 and 4.46 V, respectively, much higher than that of the CNT electrode (4.46 V, Fig. S1a, ESI). Mn2(dobdc), impressively, shows a charge potential of 3.96 V—a 0.50 V decrease compared with the CNT electrode; Mn(bdc) also shows a much reduced charge potential of 4.19 V. Preliminarily, we speculate that Mn(II) centers may play a positive role in activating Li2CO3 upon charge.
image file: c8ee00415c-f2.tif
Fig. 2 Crystal structures presenting 1D porous channels (left) of (a) Mn2(dobdc), (b) Ni2(dobdc), (c) Co2(dobdc), (d) Mn(bdc), (e) Fe(bdc) and (f) Cu(bdc) and their corresponding discharge–charge voltage curves at 50 mA g−1 with a capacity limit of 1000 mA h g−1 (right).

3.2 Properties of different Mn(II) materials in CO2 electrodes

To confirm the electrocatalytic activity of Mn(II), we tested other Mn(II)-based materials including two porous Mn(II)–MOFs, Mn(C2H2N3)2 (C2H2N3 = 1H-1,2,3-triazole) and Mn(HCOO)2, together with non-porous MnCO3 and MnO. Mn(C2H2N3)2, known as MET-2, is formed via Mn2+ ions octahedrally coordinated with N atoms in triazolate (Fig. S2a, ESI);38 it shows a diamond-type structure with a pore diameter of 6.1 Å. Mn(HCOO)2 possesses narrower zigzag 1D channels of ∼5.5 Å diameter (Fig. S2b (ESI) and Fig. 4a). Compared with Mn(dobdc) and Mn(bdc), these two Mn(II)–MOFs have smaller pore sizes and different types of organic linkers, with varied chemical environments around Mn(II) centers. As shown in Fig. 3, all the Mn(II)-based electrodes exhibit reduced charge potentials (<4.40 V) at the same current density; Mn(II)–MOF electrodes, except Mn(C2H2N3)2, show lower charge potentials (<4.20 V) than those of non-porous MnCO3 and MnO (>4.25 V). It can be deduced that all studied Mn(II)–MOFs with carboxylate linkers are likely to achieve high catalytic activity owing to their easily accessible metal sites, while the catalytic activities of densely packed crystals like MnCO3 and MnO are restricted by their limited surface areas. The relatively low catalytic activity of Mn(C2H2N3)2 (Fig. S2a, ESI) compared with other Mn(II)–MOFs suggests that the triazole linker or the coordination mode may pose negative effects on the catalytic activity of Mn(II) centers: all the Mn(II) centers in the studied Mn(II)–MOFs are in octahedral coordination, but those in Mn(C2H2N3)2 are closely surrounded by bulky triazolate penta-heterocycles, resulting in large steric hindrances around them.
image file: c8ee00415c-f3.tif
Fig. 3 Average discharge (red, right side) and charge (blue, left side) potentials of different electrodes upon operation at 50 mA g−1 with a capacity limit of 1000 mA h g−1.

To further investigate how the porosity of MOFs affects the activity of Mn(II) centers and their electrochemical behaviours, we specifically selected Mn2(dobdc), Mn(HCOO)2 and MnCO3 as representatives for comparison. Mn(HCOO)2 is a Mn(II)–MOF linked by short formate ligands. Albeit with smaller pore size, it is also an active material for CO2 adsorption.43 MnCO3 is a typical non-porous Mn(II) carboxylate almost incapable of adsorbing CO2; its octahedral Mn(II) centers are densely packed in a hexagonal–rhombohedral structure. We first investigated their CO2 adsorption behaviours by measuring the adsorption isotherms at three different temperatures (273, 283 and 298 K). As shown in Fig. 4b, the CO2 uptakes of Mn2(dobdc) and Mn(HCOO)2 at 1 atm and 298 K are 143 and 31 cm3 g−1, respectively, whereas MnCO3 and CNTs show almost no CO2 adsorption under identical conditions. The trend is similar in the isotherms at other temperatures (Fig. S6a and b, ESI). The Brunauer–Emmett–Teller (BET) surface areas for MnCO3, Mn(HCOO)2 and Mn2(dobdc) are 7, 180 and 1019 m2 g−1, respectively, illustrating their substantial difference in porosity (Fig. 4a).

image file: c8ee00415c-f4.tif
Fig. 4 (a) Crystal structures of MnCO3, Mn(HCOO)2 and Mn2(dobdc) along certain directions. (b) CO2 adsorption isotherms at 298 K and (c) discharge voltage curves of Mn2(dobdc), Mn(HCOO)2, MnCO3 and CNTs at 50 mA g−1 with a cut-off voltage of 2.0 V.

In voltage-limited battery tests, the different porosities of the three Mn(II) materials clearly translate into varied discharge capacities of the corresponding CO2 electrodes. Fig. 4c shows the discharge curves at 50 mA g−1 with a cutoff voltage of 2.0 V. All the electrodes start with an open circuit voltage (OCV) of ∼2.7 V, in line with previous reports.22,23,25 The discharge capacities of MnCO3, Mn(HCOO)2 and Mn2(dobdc) are 11[thin space (1/6-em)]110, 15[thin space (1/6-em)]510 and 18[thin space (1/6-em)]022 mA h g−1, respectively, whereas the CNT electrode only achieves 8396 mA h g−1. Remarkably, the value of Mn2(dobdc) (18[thin space (1/6-em)]022 mA h g−1) is more than twice that of the CNT electrode (8396 mA h g−1), exceeding those of all the reported CO2 electrodes (Table S1, ESI). Importantly, both Mn(HCOO)2 and Mn2(dobdc) release the majority of capacity above 2.6 V, showing the promise of the Li–CO2 system in high-energy power supply with the two MOFs. Clearly, the porous nature of MOFs and their capability in CO2 capture are favourable for boosting the discharge capacity, as suggested by the trend of increasing discharge capacity with increasing porosity and CO2 uptake among the three Mn(II) compounds.

We examined the cycling performance of these CO2 electrodes at different current densities (50, 100 and 200 mA g−1) subject to a capacity limit of 1000 mA h g−1 (Fig. S1–S5, ESI). The average charge potentials of the electrodes are summarized in Fig. 5a. At 50 mA g−1, Mn2(dobdc) shows the lowest charge potential of ∼3.96 V; as the current density increases to 100 and 200 mA g−1, the charge potential increases slightly to 3.98 V and then to 4.21 V, suggesting an aggravated polarization upon high power operation. Notably, Mn(HCOO)2 keeps low charge potentials of 4.00 and 4.02 V at 100 and 200 mA g−1, respectively, indicative of excellent catalytic activity with faster electrochemical kinetics. As expected, the charge potentials of MnCO3 are above 4.20 V at all current densities—though still below those of CNTs—substantially inferior to the two porous MOF counterparts. This can be ascribed to the poor accessibility of CO2 molecules to and the low utilization rate of the Mn(II) active sites in the bulk phase of non-porous MnCO3.

image file: c8ee00415c-f5.tif
Fig. 5 (a) Average charge potentials of different CO2 electrodes at various current densities. (b) Discharge–charge cycling performance of a Mn(HCOO)2 electrode at 200 mA g−1. (c) Isosteric heats of adsorption (Qst) for Mn2(dobdc) and Mn(HCOO)2 as a function of CO2 uptake (derived from the CO2 adsorption isotherms collected at 273, 283 and 298 K).

3.3 Effects of MOFs on Li–CO2 electrochemistry

To rationalize the difference in electrochemical performance, isosteric heats of adsorption (Qst) as a function of CO2 uptake were determined for Mn2(dobdc) and Mn(HCOO)2 (Fig. 5c) from the adsorption isotherms collected at 273, 283 and 298 K. Mn2(dobdc) shows a high initial Qst of 35.5 kJ mol−1, indicative of a strong affinity with CO2; the initial Qst of Mn(HCOO)2 is relatively modest (23.9 kJ mol−1). Considering the fact that CO2 molecules evolve during the charge process according to the following equation: 2Li2CO3 + C = 3CO2 + 4Li, the weaker the affinity of CO2 molecules with the porous catalyst, the easier the reaction.20 In other words, although the strong affinity of Mn2(dobdc) with CO2 is favourable for CO2 capture that boosts the discharge capacity, it also causes energy penalty for CO2 evolution upon the decomposition of Li2CO3. As the current density increases, this negative effect becomes increasingly pronounced, rendering a marked growth in the charge overpotential of the Mn2(dobdc) electrode. In contrast, the CO2 generated in the charge process can be more easily released from Mn(HCOO)2, given the lower Qst, leading to smaller polarizations at high current densities. Notably, the Mn(HCOO)2 electrode shows good reversibility, keeping the charge potential below 4.05 V after being discharged and charged at 200 mA g−1 for over 50 cycles (Fig. 5b).

To examine the structural stability of the cathode materials and the electrochemical products of the reactions, the discharged and recharged electrodes were disassembled from the Li–CO2 cells for PXRD measurements. Fig. S7 (ESI) shows the diffraction patterns of Mn(HCOO)2 electrodes at different states. The peaks of Mn(HCOO)2 can be clearly found in the discharged and recharged electrodes, revealing good retention of the MOF structure in the whole electrochemical process. Other materials, including the three M2(dobdc) (M = Mn, Ni, Co) and MnCO3, also retain their crystal structures in the battery tests (Fig. S8–S11, ESI). As illustrated in the enlarged picture in Fig. S7 (ESI), Li2CO3 (PDF#22-1141) is the major product in the discharged Mn(HCOO)2 electrode, and its diffraction peaks completely vanish after a full recharge, verifying a highly reversible CO2 conversion cycle on the MOF electrodes.

Raman spectroscopy was employed to further identify the discharge products. In particular, to rule out the interference of carbonaceous materials (carbon cloth and CNTs) on the detection of carbon, one of the discharge products, we prepared a carbon-free MOF electrode by solvothermally growing Mn2(dobdc) on a Ni foam; Au was sputtered on the electrode to enhance the conductivity. As shown in Fig. 6, the characteristic Li2CO3 peak at 1087 cm−1 can be identified on both bare Ni and Mn2(dobdc)@Ni electrodes after discharge. The peaks emerging around 1343 cm−1 and 1581 cm−1 in the spectrum of the discharged Ni foam (Fig. 6a) can be ascribed to the D band and G band of carbon, respectively; the appearance of similar signals can also be identified in that of the discharged Mn2(dobdc)@Ni electrode (Fig. 6b), which are around 1285 cm−1 and 1601 cm−1, respectively.25 These results confirm that Li2CO3 and C are formed as the discharge products. We note that the relatively weak strengths of the D and G signals in Raman spectra are in line with previous Li–CO2 studies.25,44–46 This perhaps is due to the insufficient concentration of discharge products on the carbon-free electrode. Owing to the lack of a conductive agent (carbon black or CNTs), the activity of the electrode is rather limited, resulting in low discharge capacity and thus insufficient concentration of discharge products. With regard to the other peaks in Fig. 6b, those at 1197, 1396, 1485, 1554 and 1571 cm−1 can be identified in the electrode before discharge and are thus ascribed to Mn2(dobdc), whereas the one at 1455 cm−1 is ascribed to the electrolyte solvent TEGDME adsorbed in the pores of MOF.47

image file: c8ee00415c-f6.tif
Fig. 6 Raman spectra of (a) Ni foam and (b) Mn2(dobdc)@Ni before and after discharge.

In situ DEMS was employed to investigate the gas emission of a Mn2(dobdc) electrode upon recharge. The electrode was first discharged to 1000 mA h g−1 at 50 mA g−1 in a coin cell. Afterwards, the electrode was transferred into a customized Swagelok cell and recharged at 200 mA g−1, during which the gas product was monitored by in situ MS. The DEMS data explicitly show that the cell releases CO2 upon recharge without any trace of O2 (Fig. 7), excluding the possibility that the charge reaction proceeds as 2Li2CO3 = 2CO2 + O2 + 4Li.25,48 Besides, a Mn2(dobdc) electrode tested in pure Ar shows negligible capacity (Fig. S19, ESI). These pieces of evidence together support the reversible reaction path of 3CO2 + 4Li = 2Li2CO3 + C on the MOF electrode.

image file: c8ee00415c-f7.tif
Fig. 7 Gas emission (bottom) of a Mn2(dobdc) electrode upon charge at 200 mA g−1 (top).

EIS was applied to track the evolution of the charge transfer kinetics of different CO2 electrodes. Data were collected for each electrode at a series of discharge and charge states. As shown in Fig. 8a and Fig. S12, Table S2 (ESI), the CNT electrode experiences a dramatic increase in charge transfer resistance (Rct) upon discharge, while the growth of Rct in electrodes containing Mn(II)-based materials is largely suppressed; the values only reach 1/3 of that of the CNT electrode when the cells are discharged to 1000 mA h g−1. Upon recharge, all the cells show a downward trend in Rct; after being fully recharged, the Rct values of the Mn(II)-based electrodes almost return to their original levels, but an obvious net increase in Rct is witnessed for the CNT electrode. Since the growing Rct reflects the deterioration of the electrode's charge transfer kinetics and is directly linked with the passivation of the electrode's surface caused by the inactive solid product (Li2CO3), the retarded growth of Rct brought about by the Mn(II)-based materials inspired us to compare the deposition and decomposition of Li2CO3 on different electrodes by SEM. As shown in Fig. S13–S16 (ESI), the morphologies of the discharge products show tremendous differences. The CNT electrode is covered by aggregates of plate-like products, several hundred nanometers in size (Fig. S13e, ESI). The morphology of Li2CO3 on the discharged MnCO3 electrode (Fig. S16e, ESI) shares some similarity, but being more uniform, smaller in aggregates’ size and thinner in plates’ thickness. In the discharged MOF electrodes (Fig. S14e and S15e, ESI), only a thin layer of Li2CO3, composed of tiny nanoparticles, can be observed, which conformally covers the surfaces of the Mn(HCOO)2 and Mn2(dobdc) crystals. After scrutinizing the morphology evolution in the MOF electrodes during the discharge process (Fig. S14b–d and S15b–d, ESI), we speculate that the nucleation of Li2CO3 at the early stage initiates in the regularly arranged micropores of the MOF; the separated and strictly confined nanospaces direct the growth of Li2CO3 and finally result in uniform nanostructures. A similar phenomenon was reported by Chen et al.,49 who observed that at the very beginning of the discharge of a Li–O2 battery with a MOF-based O2 electrode, the solid product Li2O2 initially emerges in the pores of the MOF and further grows into nanofibers. A nanostructured morphology is highly favourable for promoting the electrochemical activity of inert discharge products, ensuring a facile and complete conversion of the products in the following recharge process (Fig. S14f and S15f, ESI). Large aggregates of inactive products would otherwise result in a difficult and incomplete decomposition, as can be observed in the recharged CNT electrode (Fig. S13f, ESI) and indicated by the EIS results (as mentioned above).

image file: c8ee00415c-f8.tif
Fig. 8 (a) Evolution of the charge transfer resistance (Rct) of different electrodes at varied discharge and charge states during operation at 50 mA g−1. (b) Typical discharge–charge voltage curves with plots showing the capacity at which an EIS experiment was conducted.

4. Conclusions

In summary, we have demonstrated the promise of MOFs that significantly activate Li–CO2 electrochemistry. Useful guidelines for the selection of MOFs are provided: (1) Mn(II) metal centers play an essential role in alleviating the charge overpotential; (2) rich porosity with high CO2 uptake tends to raise the upper limit of discharge capacity; (3) a moderate isosteric heat of CO2 adsorption is more favourable in achieving smaller charge polarization at high current density. Combining the advantages of high porosity with monodispersed Mn(II) centers, Mn(II)–MOFs have enabled efficient (low charge potential) and reversible (good cycling stability) CO2 electrodes. Mn2(dobdc) achieves a record high discharge capacity of 18[thin space (1/6-em)]022 mA h g−1 with a charge potential as low as 3.96 V at 50 mA g−1. Mn(HCOO)2, another Mn(II)–MOF with smaller pores and moderate isosteric heat of CO2 adsorption relative to Mn2(dobdc), keeps a low charge potential of 4.02 V even at a high current density of 200 mA g−1 and is able to retain for over 50 cycles with a capacity limit of 1000 mA h g−1. The deposition of Li2CO3 upon discharge is found to be sensitively affected by the MOF, with which it tends to grow into a nanostructured morphology with a uniform dispersion in the whole electrode. This effect is of critical importance in activating the inert Li2CO3, which mitigates the growth of the electrode's charge transfer resistance and guarantees a facile and complete conversion of Li2CO3 back to CO2 in the charge process. The enhancement of the conductivity of MOF catalysts is the future direction to further improve their performance. In this line, using conductive MOFs or growing MOFs on conductive materials like CNTs or graphene are possible strategies. In addition, as revealed by studies of MOFs in Li–S50 and Li–O251 systems, the morphology of a MOF (especially particle size) can also make contributions to electrochemical performance—most probably, capacity. Accordingly, systematic investigation on the effects of the pore size, framework geometry, side group functionality and morphology of MOFs is on-going in our lab within a broad scope of Mn(II)–MOFs; the behaviours at varying current densities and capacity limits will be studied comprehensively. Exploration of batteries working in mixed CO2/O2 gas that simulates real-world scenarios (such as confined spaces) is underway. While the systems presented here are preliminary, they have provided useful design principles for improving the reversibility and energy efficiency of Li–CO2 electrochemistry. The knowledge will herald the advent of practical technologies that enable energy-efficient utilization of CO2 in EES.

Conflicts of interest

There are no conflicts to declare.


This work was financially supported by the 973 Program (2013CB834702), the National Natural Science Foundation of China (Grant No. 21625102, 21471018, 21404010, 21490570, 21674012, 21701012) and the 1000 Plan (Youth).


  1. V. Augustyn, P. Simon and B. Dunn, Energy Environ. Sci., 2014, 7, 1597–1614 CAS.
  2. D. Kundu, E. Talaie, V. Duffort and L. F. Nazar, Angew. Chem., Int. Ed., 2015, 54, 3431–3448 CrossRef CAS PubMed.
  3. W. Xia, A. Mahmood, R. Q. Zou and Q. Xu, Energy Environ. Sci., 2015, 8, 1837–1866 CAS.
  4. M. S. Balogun, Y. C. Huang, W. T. Qiu, H. Yang, H. B. Ji and Y. X. Tong, Mater. Today, 2017, 20, 425–451 CrossRef CAS.
  5. B. Long, M. S. Balogun, L. Luo, W. T. Qiu, Y. Luo, S. Q. Song and Y. X. Tong, Adv. Energy Mater., 2018, 8, 1701681 CrossRef.
  6. Z. Q. Liu, H. Cheng, N. Li, T. Y. Ma and Y. Z. Su, Adv. Mater., 2016, 28, 3777–3784 CrossRef CAS PubMed.
  7. T. Ogasawara, A. Debart, M. Holzapfel, P. Novak and P. G. Bruce, J. Am. Chem. Soc., 2006, 128, 1390–1393 CrossRef CAS PubMed.
  8. A. C. Luntz and B. D. McCloskey, Chem. Rev., 2014, 114, 11721–11750 CrossRef CAS PubMed.
  9. H. G. Jung, J. Hassoun, J. B. Park, Y. K. Sun and B. Scrosati, Nat. Chem., 2012, 4, 579–585 CrossRef CAS PubMed.
  10. Q. Sun, Y. Yang and Z.-W. Fu, Electrochem. Commun., 2012, 16, 22–25 CrossRef CAS.
  11. M. M. Ottakam Thotiyl, S. A. Freunberger, Z. Peng, Y. Chen, Z. Liu and P. G. Bruce, Nat. Mater., 2013, 12, 1050–1056 CrossRef CAS PubMed.
  12. Z.-W. Chang, J.-J. Xu, Q.-C. Liu, L. Li and X.-B. Zhang, Adv. Energy Mater., 2015, 5, 1500633 CrossRef.
  13. L. Grande, E. Paillard, J. Hassoun, J. B. Park, Y. J. Lee, Y. K. Sun, S. Passerini and B. Scrosati, Adv. Mater., 2015, 27, 784–800 CrossRef CAS PubMed.
  14. U. Satish, M. J. Mendell, K. Shekhar, T. Hotchi, D. Sullivan, S. Streufert and W. J. Fisk, Environ. Health Perspect., 2012, 120, 1671–1677 CAS.
  15. J. G. Allen, P. MacNaughton, U. Satish, S. Santanam, J. Vallarino and J. D. Spengler, Environ. Health Perspect., 2016, 124, 805–812 CrossRef PubMed.
  16. S. R. Gowda, A. Brunet, G. M. Wallraff and B. D. McCloskey, J. Phys. Chem. Lett., 2013, 4, 276–279 CrossRef CAS PubMed.
  17. X. Li, S. Yang, N. Feng, P. He and H. Zhou, Chin. J. Catal., 2016, 37, 1016–1024 CrossRef CAS.
  18. Z. Xie, X. Zhang, Z. Zhang and Z. Zhou, Adv. Mater., 2017, 29, 1605891 CrossRef PubMed.
  19. K. Takechi, T. Shiga and T. Asaoka, Chem. Commun., 2011, 47, 3463–3465 RSC.
  20. Y. Liu, R. Wang, Y. Lyu, H. Li and L. Chen, Energy Environ. Sci., 2014, 7, 677–681 CAS.
  21. S. M. Xu, S. K. Das and L. A. Archer, RSC Adv., 2013, 3, 6656–6660 RSC.
  22. Z. Zhang, Q. Zhang, Y. Chen, J. Bao, X. Zhou, Z. Xie, J. Wei and Z. Zhou, Angew. Chem., Int. Ed., 2015, 54, 6550–6553 CrossRef CAS PubMed.
  23. X. Zhang, Q. Zhang, Z. Zhang, Y. Chen, Z. Xie, J. Wei and Z. Zhou, Chem. Commun., 2015, 51, 14636–14639 RSC.
  24. L. Qie, Y. Lin, J. W. Connell, J. Xu and L. Dai, Angew. Chem., Int. Ed., 2017, 56, 1–6 CrossRef PubMed.
  25. S. Yang, Y. Qiao, P. He, Y. Liu, Z. Cheng, J.-J. Zhu and H. Zhou, Energy Environ. Sci., 2017, 10, 972–978 CAS.
  26. Y. Hou, J. Wang, L. Liu, Y. Liu, S. Chou, D. Shi, H. Liu, Y. Wu, W. Zhang and J. Chen, Adv. Funct. Mater., 2017, 1700564,  DOI:10.1002/adfm.201700564.
  27. H. Deng, S. Grunder, K. E. Cordova, C. Valente, H. Furukawa, M. Hmadeh, F. Gandara, A. C. Whalley, Z. Liu, S. Asahina, H. Kazumori, M. O'Keeffe, O. Terasaki, J. F. Stoddart and O. M. Yaghi, Science, 2012, 336, 1018–1023 CrossRef CAS PubMed.
  28. J. R. Li, R. J. Kuppler and H. C. Zhou, Chem. Soc. Rev., 2009, 38, 1477–1504 RSC.
  29. H. C. Zhou, J. R. Long and O. M. Yaghi, Chem. Rev., 2012, 112, 673–674 CrossRef CAS PubMed.
  30. E. M. Miner, T. Fukushima, D. Sheberla, L. Sun, Y. Surendranath and M. Dinca, Nat. Commun., 2016, 7, 10942 CrossRef CAS PubMed.
  31. K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, E. D. Bloch, Z. R. Herm, T. H. Bae and J. R. Long, Chem. Rev., 2012, 112, 724–781 CrossRef CAS PubMed.
  32. B. Seoane, J. Coronas, I. Gascon, M. Etxeberria Benavides, O. Karvan, J. Caro, F. Kapteijn and J. Gascon, Chem. Soc. Rev., 2015, 44, 2421–2454 RSC.
  33. S. R. Caskey, A. G. Wong-Foy and A. J. Matzger, J. Am. Chem. Soc., 2008, 130, 10870–10871 CrossRef CAS PubMed.
  34. D. Wu, Z. Guo, X. Yin, Q. Pang, B. Tu, L. Zhang, Y. G. Wang and Q. Li, Adv. Mater., 2014, 26, 3258–3262 CrossRef CAS PubMed.
  35. F. Luo, Y.-X. Che and J.-M. Zheng, Inorg. Chem. Commun., 2008, 11, 358–362 CrossRef CAS.
  36. P. Horcajada, F. Salles, S. Wuttke, T. Devic, D. Heurtaux, G. Maurin, A. Vimont, M. Daturi, O. David, E. Magnier, N. Stock, Y. Filinchuk, D. Popov, C. Riekel, G. Ferey and C. Serre, J. Am. Chem. Soc., 2011, 133, 17839–17847 CrossRef CAS PubMed.
  37. C. G. Carson, K. Hardcastle, J. Schwartz, X. Liu, C. Hoffmann, R. A. Gerhardt and R. Tannenbaum, Eur. J. Inorg. Chem., 2009, 2338–2343 CrossRef CAS.
  38. F. Gandara, F. J. Uribe-Romo, D. K. Britt, H. Furukawa, L. Lei, R. Cheng, X. Duan, M. O'Keeffe and O. M. Yaghi, Chem. – Eur. J., 2012, 18, 10595–10601 CrossRef CAS PubMed.
  39. Z. H. Cui and X. X. Guo, J. Power Sources, 2014, 267, 20–25 CrossRef CAS.
  40. R. Wang, X. Yu, J. Bai, H. Li, X. Huang, L. Chen and X. Yang, J. Power Sources, 2012, 218, 113–118 CrossRef CAS.
  41. Y. Yang, Q. Sun, Y. S. Li, H. Li and Z. W. Fu, J. Power Sources, 2013, 223, 312–318 CrossRef CAS.
  42. L. Pan, J. Tang and F. Wang, Cent. Eur. J. Chem., 2013, 11, 763–773 CAS.
  43. D. N. Dybtsev, H. Chun, S. H. Yoon, D. Kim and K. Kim, J. Am. Chem. Soc., 2004, 126, 32–33 CrossRef CAS PubMed.
  44. X. Hu, Z. Li and J. Chen, Angew. Chem., Int. Ed., 2017, 56, 5785–5789 CrossRef CAS PubMed.
  45. C. Li, Z. Guo, B. Yang, Y. Liu, Y. Wang and Y. Xia, Angew. Chem., Int. Ed., 2017, 56, 9126–9130 CrossRef CAS PubMed.
  46. Y. Qiao, J. Yi, S. Wu, Y. Liu, S. Yang, P. He and H. Zhou, Joule, 2017, 1, 359–370 CrossRef.
  47. V. de Zea Bermudez, G. Lucazeau, L. Abello and C. Poinsignon, J. Mol. Struct., 1993, 301, 7–19 CrossRef CAS.
  48. S. Song, W. Xu, J. Zheng, L. Luo, M. H. Engelhard, M. E. Bowden, B. Liu, C. M. Wang and J. G. Zhang, Nano Lett., 2017, 17, 1417–1424 CrossRef CAS PubMed.
  49. X. Hu, Z. Zhu, F. Cheng, Z. Tao and J. Chen, Nanoscale, 2015, 7, 11833–11840 RSC.
  50. J. W. Zhou, X. S. Yu, X. X. Fan, X. J. Wang, H. W. Li, Y. Y. Zhang, W. Li, J. Zheng, B. Wang and X. G. Li, J. Mater. Chem. A, 2015, 3, 8272–8275 CAS.
  51. W. Yan, Z. Guo, H. Xu, Y. Lou, J. Chen and Q. Li, Mater. Chem. Front., 2017, 1, 1324–1330 RSC.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c8ee00415c

This journal is © The Royal Society of Chemistry 2018