Sulfur Composite

Subscriber access provided by SELCUK UNIV

Article

Ionic Liquid–Derived Nitrogen–Enriched Carbon/Sulfur Composite Cathodes with Hierarchical Microstructure – A Step Toward Durable High Energy and High Performance Lithium–Sulfur Batteries Artur Schneider, Christoph Weidmann, Christian Suchomski, Heino Sommer, Jürgen Janek, and Torsten Brezesinski Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/cm504460p • Publication Date (Web): 29 Jan 2015 Downloaded from http://pubs.acs.org on January 31, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Ionic Liquid–Derived Nitrogen–Enriched Carbon/Sulfur Composite Cathodes with Hierarchical Microstructure – A Step Toward Durable High Energy and High Performance Lithium–Sulfur Batteries Artur Schneider,# Christoph Weidmann,# Christian Suchomski,# Heino Sommer,#,† Jürgen Janek#,§,* and Torsten Brezesinski#,* #

Battery and Electrochemistry Laboratory, Institute of Nanotechnology, Karlsruhe Institute of

Technology, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany. †

BASF SE, 67056 Ludwigshafen, Germany.

§

Institute of Physical Chemistry, Justus-Liebig-University Giessen, Heinrich-Buff-Ring 58,

35392 Giessen, Germany.

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 35

Keywords Hard templating, hierarchical porosity, sulfur loading, polysulfide shuttle, in operando X-ray diffraction

Abstract We describe the preparation of turbostratic carbon in monolithic form by silica template method and its use as host matrix to trap polysulfides in Li–S batteries. The synthesized ionic liquid-derived carbon is hierarchically structured with pore size maxima around 6 nm and 750 nm, and it has a room temperature electrical conductivity of 7 to 8 S cm –1, owing to a high content of pyridinic and graphitic nitrogen. Further, this carbon can accommodate up to ~80% sulfur by weight and the resulting nanocomposite demonstrates promising performance as novel electrode material for Li–S batteries. Cathodes with low areal mass loading (1 mgsulfur cm–2) can be cycled at C/5 and 1C over several hundreds of cycles, with low polarization and very little capacity fading (4% between the 5th and 1000th cycles at 1C). Highly loaded cathodes (4 mgsulfur cm–2) also exhibit good cyclability at C/5 with areal capacities of 2.6 mAh cm –2 on average over 200 cycles – yet the high capacities make the cells more prone to electrolyte decomposition, eventually giving rise to low round trip energy efficiencies depending on the electrolyte-to-sulfur mass ratio. Results from in operando X-ray diffraction show that nanocrystalline Li2S forms midway through the second discharge plateau and sulfur recrystallizes in the metastable β-S8 phase, with 30 to 40 nm crystallites, but undergoes amorphization under open circuit conditions. Overall, this work represents a step forward in the development of durable high energy Li–S batteries and provides valuable insights into the operation of such cells.


2

Page 3 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction The development of cost-efficient battery systems that meet future demands (i.e., energy density of >300 Wh kg–1 on cell level, high energy efficiency, stability over thousands of cycles, low self-discharge rate etc.) is of great relevance.1–3 Elemental sulfur (S8) was early considered as a promising candidate and is currently one of the most studied cathode materials for rechargeable lithium batteries of the next generation.4 This is due, in part, to the high theoretical specific capacity of 1675 Ah kgsulfur–1, providing a theoretical energy density of ~2500 Wh kg–1 for the Li–S system at an average voltage of 2.15 V with respect to Li/Li+. Furthermore, sulfur is abundant and of low cost and toxicity, which makes it an attractive alternative to electroactive metal oxides, in particular those containing cobalt. In contrast to the vast majority of practically used intercalation or insertion-type cathode active materials, it is a conversion-type material.5,6 This means that sulfur undergoes major changes upon electrochemical reaction with Li, which involve the formation of higher- and lower-order polysulfides of different solubility.7 Li2S is the end product when the Li–S battery is fully discharged and forms by transfer of 16 electrons according to: S8 + 16Li+ + 16e– → 8Li2S. While the Li–S system looks rather simple on paper, there are severe performance and safety concerns preventing it so far from being used on a commercial scale for e.g., stationary storage or vehicles. Despite significant progress over the past decades, many obstacles still need to be overcome in order to make especially highly loaded Li–S batteries work, including electrical conductivity issues, gas evolution/drying out effects due to continuous electrolyte decomposition and capacity fading/low round trip energy efficiencies due to polysulfide shuttle, to mention only a few.8–13 The general working principle and major challenges of Li–S batteries have been described in excellent reviews elsewhere and thus will not be discussed further here.14–16


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 35

In the present work, we report the preparation of highly conductive, nitrogen-enriched carbon with an interconnected hierarchically porous structure by hard templating using a silica monolith template and an ionic liquid (IL) as the carbon precursor and its use as backbone/host matrix in Li–S batteries. This carbon is an excellent model system for gaining insight into critical issues of the Li–S system, regardless of the fact that ILs are generally expensive. While functional carbons of different shape have been produced from a broad range of precursor materials,17–21 the use of ILs to achieve centimeter-size monoliths having a high nitrogen content and porosity on multiple length scales has, to the best of our knowledge, not yet been reported. As shown by Antonietti, Thomas and others, ILs are very versatile precursors for the synthesis of heteroatom-rich carbonaceous materials with a turbostratic microstructure.22–24 Apart from that they provide the opportunity to greatly reduce the synthesis time of hardtemplated carbons as the impregnation step can be carried out at reduced pressure – the overall process is less tedious. We demonstrate that the monolithic carbon studied here is a promising lightweight and mechanically robust matrix material for application in Li–S batteries. Both low and highly loaded coin cells display excellent cyclability, with little capacity decay over many cycles. However, the purpose of this paper is not to be just another example of a sulfur composite cathode with optimized microstructure, but instead, the goal was to make a step forward towards high energy and high performance Li–S batteries and to understand their functioning. For that reason, cells with a reasonable sulfur content (~60%) and loadings of up to 4 mgsulfur cm–2 as well as electrolyte-to-sulfur mass ratios as low as 10:1 were tested and analyzed in detail using different characterization techniques. We note that the latter ratio is still not in the range needed to achieve specific energy densities much larger than those of current state-of-the-art rechargeable Li-ion batteries, but it is smaller than in the majority of reports on Li–S systems.25


4

Page 5 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


High electrolyte-to-sulfur mass ratios and low sulfur loadings typically result in improved cell characteristics, but strongly reduce the energy on cell level.26–30 This means that, in effect, energy density is traded for cycling stability. Overall, we show in this paper that the IL-derived carbon/sulfur composite cathodes possess a beneficial microstructure for Li–S battery applications – even highly loaded cells exhibit a stable performance, also without the use of commonly used additives such as carbon nanotubes or graphene.

Experimental Section Synthesis of the Nitrogen-Enriched Carbon/Sulfur Monoliths Silica monoliths with a hierarchically porous structure were synthesized according to an established method.31 In a typical procedure, poly(ethylene oxide) (MW = 10.000 g mol–1, 1.2 g) and urea (99%, 0.9 g) were added to an aqueous solution of acetic acid (0.01 M, 10 mL). The mixture was cooled in an ice-bath and stirred for 15 min, followed by addition of tetramethyl orthosilicate (98%, 5.6 mL) and vigorous stirring for 30 min. Thereafter, it was poured into conical tubes and aged at 25 °C for 12 h to induce hydrolysis and condensation. The obtained solid monoliths were then put into an aqueous solution containing urea (9.0 g) and acetic acid (0.01 M, 100 mL) and treated at 80 °C for 15 h. Finally, the siliceous material was thoroughly washed with methanol, followed by calcination at 330 °C for 15 h. Nitrogen-enriched carbon monoliths with a hierarchical pore structure were prepared by hard templating method. Briefly, the monolithic silica described above was immersed in 1-ethyl-3methylimidazolium dicyanamide (EMIM-DCA, ≥98.5%) and the impregnation was carried out at reduced pressure to achieve maximum pore filling (the overall process is finished when the pressure inside the flask reaches 10–3 mbar). After excess EMIM-DCA was wiped off using a ACS Paragon Plus Environment

5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 35

precision wipe (KIMTECH SCIENCE), the semi-transparent monoliths were transferred to a tube furnace and heated to 900 °C under nitrogen using a 2 h ramp, followed by aging for 30 min. The latter steps were repeated twice. Lastly, the silica template was removed by leaching with an aqueous solution of potassium hydroxide (3 M) at 70 °C, followed by hot water rinsing and vacuum drying. Carbon/sulfur composite monoliths were prepared by facile melt-infiltration of sulfur (Aldrich, reagent grade) into the nitrogen-enriched carbon.32,33 Excess sulfur at the outer surface of the monoliths was removed mechanically with a scalpel. Finally, the composite material was ground to a fine powder by ball-milling for 1 h, leading to 5–15 µm particles. Electrode Processing, Cell Assembling and Electrochemical Testing Carbon/sulfur composite powder (72 wt%), Super C65 (Timcal, 9 wt%), Printex XE2 (Orion, 9 wt%) and poly(vinyl alcohol) Selvol 425 (Sekisui, 10 wt%) dissolved in water, isopropanol and 1-methoxy-2-propanol (65:30:5 by weight) were thoroughly mixed to form a homogeneous slurry (composition adapted from patents by Sion Power).34,35 The slurry was then coated onto primed aluminum foil with a doctor blade and dried in vacuum at 60 °C for 12 h. 50–100 µm thick electrodes with sulfur loadings of 1–4 mg cm–2 were obtained. Finally, circular discs 13 mm in diameter were punched out using an EL-Cut electrode cutter (EL-CELL) and transferred to an argon-filled glove box with [O2] < 1 ppm and [H2O] < 1 ppm. Coin-type cells consisting of cathode and lithium foil (China Lithium Ltd., 600 µm), separated by Celgard EK2040 15 mm polyethylene membrane, were assembled inside the glove box. For pouch cells, the assembling was performed in a dry room by stacking lithium foil (Chemetall Foote Corp., 50µm), separator and cathode. The electrolyte used was a solution of lithium bis(trifluoromethanesulfonyl)imide (Aldrich, 99.95%, 8 wt%), lithium nitrate (Merck, 99.995%,


6

Page 7 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4 wt%), 1,2-dimethoxyethane (Alfa Aesar, >99%, 44 wt%) and 1,3-dioxolane (Acros, 99.8%, 44 wt%). The electrolyte-to-sulfur mass ratio was either 15:1 or 10:1. Galvanostatic measurements were performed under stable environmental conditions in a BINDER cooled incubator at 25 °C in the potential range of 2.5–1.7 V vs. Li/Li+ using a MACCOR Series 4000 (Tulsa, Oklahoma) multichannel battery cycler. The cut-off charge voltage was kept constant until a current drop of 90% was achieved before starting the following discharge cycle. All capacity values given in the manuscript were calculated on the basis of the sulfur mass. After 1 cycle at C/50 (with 1C = 1672 mA gsulfur–1) was completed, the cells were charged and discharged at a rate of C/5. Characterization Methods Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images were taken on a LEO 1530 microscope operated at 10 kV and an FEI Tecnai G2 F20 Super-Twin operated at 200 kV, respectively. X-ray diffraction (XRD) measurements were carried out both on an X’Pert PRO diffractometer from PANalytical instruments (λ = 0.15418 nm) using an X’Celerator RTMS detector and at the Synchrotron Light Source ANKA on the PDIFF beamline (λ = 0.0729 nm) using a Pilatus 300k detector (counting time = 30 s). X-ray photoelectron spectroscopy (XPS) data were collected on a VersaProbe PHI 5000 Scanning ESCA Microprobe from Physical Electronics equipped with a monochromatic Al-Kα X-ray source and a hemispherical electron energy analyzer at an electron takeoff angle of 45°. The C1s signal from adventitious hydrocarbon at 284.8 eV was used as energy reference to correct for charging. N2physisorption measurements were carried out at T = 77 K using an Autosorb-6 automated gas adsorption station from Quantachrome Corporation. The Brunauer-Emmett-Teller (BET) method was used to calculate the specific surface area. The total pore volume was determined from the amount adsorbed at a relative pressure of 0.97. The pore size distribution was calculated using a


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 35

nonlocal density functional theory (NLDFT) equilibrium model for carbon. For mercury porosimetry, a setup from Thermo Fisher Scientific (Pascal 140/440) was used. The pressure range was from 0 to 400 MPa. Raman spectra were acquired using a SENTERRA dispersive Raman microscope from Bruker Optics equipped with an objective from Olympus (MPLN 100x, F.N.22, N.A.0.9) and a Nd:YAG laser (λ = 532 nm, P = 0.2 mW). Thermogravimetric analysis (TGA) was performed on a Netzsch TG 209 F1 Libra. Elemental analysis (C/H/N/S mode) was carried out using a vario MICRO cube from Elementar. The electrical conductivity of the nitrogen-enriched carbon monoliths with dimensions of 9 mm × 3 mm × 2 mm was determined by four-probe measurements at room temperature. Commercial silver paste was used for the electrodes.

Results and Discussion The nitrogen-enriched carbon/sulfur composite monoliths employed in this work were prepared by facile hard template method and subsequent melt-infiltration of sulfur into the carbon (see synthesis steps in Scheme 1). One of the biggest advantages of having monolithic carbon over powder is that the impregnation process is straightforward and excess sulfur can be readily removed mechanically. In this way, it is ensured that the sulfur is in fact located in the pores or, in other words, confined in the carbon matrix. Yet, the final composite must be ground to a fine powder to be used as cathode active material in Li–S batteries.


8

Page 9 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 1. Preparation of the nitrogen-enriched carbon/sulfur composite monoliths from hierarchically porous silica, IL as the carbon precursor and molten sulfur. Refer to the experimental section for details.

Both the morphology and structure of the nitrogen-enriched carbon were investigated by means of scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Figure 1 shows a photograph along with representative top view SEM and bright-field TEM images. As is seen, centimeter-size monoliths with a hierarchical pore structure are obtained by the templating approach used in this study (see also photograph and SEM image of the silica monolith template for comparison in Figure S1 of the ESI). While SEM evaluation reveals an interconnected 3-dimensional network of sub-1 μm macropores, TEM indicates the presence of mesopores in the walls.


9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 35

Figure 1. Electron microscopy of the nitrogen-enriched carbon monoliths. (a) Photograph showing that crack-free specimens can be prepared by facile silica template method. Top view SEM images at different magnifications in (b, c) and bright-field TEM image in (d) revealing an interconnected hierarchically porous structure.

The porosity of the monolithic carbon was analyzed in more detail by nitrogen physisorption and mercury porosimetry measurements. Figure 2a shows type IV N2-adsorption/desorption isotherms with a distinct capillary step at p/p0 ≈ 0.6, characteristic of a small-pore mesoporous material. The specific surface area from BET method and the total mesopore volume are 350 m 2 g–1 and 0.5 cm3 g–1, respectively, suggesting that the pores in the interior are accessible. As expected, these values are smaller than those of the silica template (see nitrogen and mercury porosimetry data in Figure S2 of the ESI). The reason for this is the large weight loss of the IL ACS Paragon Plus Environment

10

Page 11 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


during carbonization (approx. 80% at 900 °C) and the resulting incomplete replication of the silica structure.22 The pore size distribution shown in the inset of Figure 2a was determined from the desorption branch using a NLDFT equilibrium model for carbon with slit/cylindrical pores – it is rather narrow with a maximum at 6–7 nm. Figure 2b shows the cumulative and relative pore volumes from mercury intrusion porosimetry and the pore size distribution reveals two distinct maxima at about 7 nm and 750 nm. The former value is in agreement with that from N2physisorption. The pore volume is 1.7 cm3 g–1, thus indicating that the mesopores contribute approx. 30% to the total pore volume. The density of the carbon monoliths (assuming bulk density of graphite) can be estimated to be in the range of 0.5 to 0.6 g cm–3. Overall, the data in Figures 1 and 2 are consistent and demonstrate that the structure of the carbon can be templated by the silica matrix to produce monoliths with porosity on multiple length scales. We are unaware of any similar IL-derived carbon in monolithic form.

Figure 2.

Porous properties

of the nitrogen-enriched carbon

monoliths.

(a) N2-

adsorption/desorption isotherms. Inset: Pore size distribution (the y-axis, dV(d), is not shown). (b) Cumulative and relative pore volumes derived from mercury intrusion porosimetry.


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 35

The chemical composition of the nitrogen-enriched carbon was investigated via X-ray photoelectron spectroscopy (XPS) and elemental analysis. The latter gives a nitrogen content of 12% as well as hydrogen and sulfur contents of 1% and 0.2%, respectively. Similar values have been reported for other EMIM DCA-derived carbonaceous materials although they are strongly depend on calcination temperature and other parameters, such as the inert gas flow rate.22,23 The presence of trace amounts of sulfur is most likely due to impurities in the IL. XPS detail spectra of the N1s and C1s levels are shown in Figures 3a and b. During fitting of the data, the full-width at half-maximum (FWHM) was constrained to be equal for all peaks of a given core level. Both spectra are rich in details because different forms of nitrogen and carbon, with different bonding configurations, are generated during carbonization. The N1s spectrum can be fitted reasonably well assuming five different nitrogen species. The three main peaks at binding energies of (398.2 ± 0.05) eV, (399.8 ± 0.05) eV and (401.1 ± 0.05) eV are assigned to pyridinic, pyrrolic and graphitic nitrogen, respectively.22,23,36 These forms have been shown to increase the electronic conductivity of carbon through the contribution of either one or two electrons to the π band.37 From the peak areas, we conclude that most of the nitrogen atoms are pyridinic and graphitic in nature (almost 1:1). The minor peaks centered at (403.3 ± 0.05) eV and (405.8 ± 0.05) eV indicate the presence of oxidized species, such as pyridine-type nitrogen oxides. The C1s spectrum can also be deconvoluted into five components. However, a clear distinction between different forms of carbon is not possible. Still, the main peak at (284.7 ± 0.05) eV can be assigned to sp2 (graphitic) carbon and the small peaks at higher binding energies [namely, (286.2 ± 0.05) eV, (287.8 ± 0.05) eV, (289.5 ± 0.05) eV and (291.3 ± 0.05) eV] contain contributions from different C–C, C–N and C–O bonding configurations.36,38 Quantitative analysis of the XPS data reveals that the C:N:O atomic ratio is 31.9:4.1:1.0, which corroborates the elemental analysis results described above.


12

Page 13 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Both X-ray diffraction (XRD) and Raman spectroscopy were used to gain insight into the microstructure of the nitrogen-enriched carbon. The XRD pattern in Figure 3c indicates the partially graphitic nature. In the range of 2θ from 10° to 110° several broad peaks can be clearly observed. The one with the largest intensity at 26.0° stems from the interlayer scattering (stacking) of the graphene sheets and the distinct peak at 44.1° from the intralayer scattering. 39 Overall, these data establish that the degree of order is not very high, but is in the range usually observed for EMIM-DCA and related IL-derived carbons calcined at temperatures of ≤900 °C without using graphitizing catalysts.22,38,40 Despite that relatively low crystallinity, the hierarchically porous monoliths have an electrical conductivity of 7–8 S cm–1 at room temperature, which is comparable to the highest reported values for turbostratic carbons in monolithic form.41 Figure 3c further shows XRD data collected after sulfur melt-infiltration. The broad peak at 2θ ≈ 26° can be attributed to the carbon while the others are indicative of orthorhombic sulfur; using the Scherrer equation gives a crystallite size of ~40 nm, suggesting that the majority of the crystalline sulfur is located in the macropores rather than the mesopores. Figure 3d shows a nonpolarized Raman spectrum obtained on the same composite material. It exhibits characteristic features of partially graphitic carbon with strong D (disordered, defectactivated) and G (graphitic, in-plane stretching of sp2 bonds) bands at ~1350 cm–1 and 1590 cm– 1

, respectively, thereby confirming the results from XRD.42 The spectrum also reveals weak

bands at ~153/218 cm–1 and 472 cm–1 due to the presence of orthorhombic sulfur. These can be assigned to bending and stretching modes of the S8 ring, respectively.43 The sulfur content was determined by thermogravimetric analysis (TGA) under inert gas atmosphere. TGA curves for both the hierarchically structured composite monoliths and a Super C65/sulfur physical mixture are shown in Figure S3 of the ESI. From these data, it can be seen that the extrapolated weight loss onset temperature, ϑonset, due to sulfur evaporation is higher for


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 35

the composite monoliths (274 °C vs. 257 °C). The reason for this might be associated with sulfur confinement in the pores. However, interactions with the surface functional groups may play a role as well, as recently demonstrated by Sun et al. and Song et al.33,44,45 They showed that nitrogen-doping of carbon produces a material that has the ability to weakly bind polysulfides through interactions between the sulfur species and oxygen-containing groups. Furthermore, TGA indicates a sulfur content in the range of 80–82% – in good agreement with the theoretical value based on the total pore volume/porosity – thereby providing additional evidence that the sulfur is indeed located in the interior of the monoliths.

Figure 3. XPS detail spectra of the N1s (a) and C1s (b) core excitations in the nitrogen-enriched carbon monoliths. Solid black curves are fits to the data and solid red curves are the sums of the


14

Page 15 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


fits. The main peaks are identified as pyridinic, pyrrolic and graphitic nitrogen and graphitic carbon, respectively. (c) XRD patterns of the bare carbon and carbon/sulfur composite (inset) indicating the partially graphitic nature of the IL-derived material and the presence of crystalline sulfur after melt-infiltration. (d) Raman spectrum of the composite. The D and G bands of carbon are marked.

Electrochemical studies were performed by room temperature galvanostatic cycling of coin cells using the carbon/sulfur composite. The cut-off voltages were 1.7 V (discharge) and 2.5 V (charge) with respect to Li/Li+ and the electrolyte-to-sulfur mass ratio was 15:1, unless stated otherwise. Voltage-capacity curves for the 1st, 50th and 150th cycles of a Li–S coin cell with a loading of 4 mgsulfur cm–2 are presented in Figure 4. The discharge profiles show two characteristic plateaus at about 2.3 V and 2.1 V, corresponding to the reduction of S8 to higher-order lithium polysulfides (Li2Sx with 6 ≤ x ≤ 8) and formation of lower-order lithium polysulfides (Li2Sy with 2 ≤ y ≤ 6) and Li2S, respectively.46 The subtle plateau at about 1.8 V is due to lithium nitrate decomposition on the cathode side.47 The charge profiles reveal plateaus at about 2.2 V and 2.4 V, indicating the reoxidation of the different polysulfide species. No significant increase in overpotential is observed during cycling at C/5, thus suggesting that the mechanical integrity of the cathode is maintained.


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 35

Figure 4. Typical voltage profiles of the Li–S coin cells with a sulfur loading of 4 mg cm–2. After 1 cycle at C/50 was completed, the C-rate was increased to C/5 for the subsequent cycles.

Figures 5a–c show the cycling performance of the Li–S cells with different sulfur loadings. As can be seen from Figure 5a, the carbon/sulfur cathodes with a loading of 1 mgsulfur cm–2 demonstrate a specific capacity close to 1450 mAh g–1 in the first discharge at C/50. Such a low C-rate appears to be crucial to achieving stable cell characteristics on the subsequent cycles. After this low-rate “formation cycle”, the cells were cycled at C/5 – the specific capacities level off quickly at about 600 mAh g–1. Overall, the capacity declines by 7% between the 2nd and 500th cycles, corresponding to a fading of only 0.014% per cycle. Furthermore, it can be seen that the coulombic efficiency stabilizes above 99.8% after a few cycles. However, a closer look reveals that the values decrease gradually, although very slowly. After 500 cycles, the coulombic efficiency is only 99.3%. This indicates that deleterious side reactions, including lithium nitrate consumption, occur during cycling. Lithium nitrate was used as an additive in the electrolyte to suppress the internal redox shuttle (polysulfide shuttle).9


16

Page 17 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Since electrodes with a low areal mass loading are only of moderate interest to battery technology, cathodes with a loading of 4 mgsulfur cm–2 were also tested under identical conditions. Literature reports on sulfur composite cathodes with similar or even higher sulfur loadings are scarce.44,48-50 Nevertheless, we believe that only measurements on electrodes with a reasonably high loading (areal capacity of ≫1.0 mAh cm–2) provide a clear picture of the performance of novel materials for Li–S batteries – high loading is required to achieve competitive areal capacities. To ensure comparability among the cathodes studied here, the sulfur content in the electrodes was kept constant. This means that only the layer thickness was varied to achieve the desired loading. The morphology and microstructure of the sulfur cathodes did not change with increasing thickness (see top view SEM images in Figure S4 of the ESI). Even though the monolithic carbon is conductive and can accommodate up to ~80% sulfur by weight, certain amounts of carbon additive(s) and polymer binder are necessary to produce homogeneous electrode tapes. However, higher sulfur contents are conceivable. In this context, we note that multi-walled carbon nanotubes (MWCNTs) can be completely substituted for the carbon blacks used as electrically conducting additives in this work. The resulting cathodes, with a sulfur loading of 2 mg cm–2, also show stable cyclability over many cycles (see performance data in Figure S5 of the ESI). As is evident from Figure 5b, the highly loaded cells using the carbon/sulfur composite deliver an initial specific capacity of about 1400 mAh g–1. This value is close to that mentioned above. The fact that the theoretical capacity of 1672 mAh g–1 is not achieved in the first discharge at low C-rate might be associated with sulfur confinement in the hierarchically structured carbon matrix and a resulting lower electrochemical addressability. Similar observations with regard to reduced sulfur utilization – also at higher C-rates – have been made for other nanocomposite cathodes with optimized microstructure, including those using microporous carbon as host matrix for the


17


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 35

sulfur.44,51,52 After 10 cycles at C/5, specific capacities of 650 mAh g–1, corresponding to areal capacities of 2.6 mAh cm–2, can be achieved on average over 200 cycles. The latter capacity value is among the largest thus far reported for high performance Li–S batteries, especially when considering the number of cycles and the relatively low electrolyte-to-sulfur mass ratio. However, a direct comparison with other Li–S systems is difficult not only because the cathode composition and electrode processing parameters typically differ across studies, but also because the electrolyte-to-sulfur mass ratio is often not given. Furthermore, it can be seen that the coulombic efficiency shows more scatter and declines more rapidly than that of cells with a loading of only 1 mgsulfur cm–2. As mentioned in the introduction to this article, in particular Li–S batteries with a high sulfur loading often exhibit poor cyclability and show a much higher failure rate than low loaded cells. This is due, in part, to the compressive and tensile stresses arising from the large volume changes of sulfur during electrochemical cycling and the fact that much more lithium is cycled between the positive and negative electrodes. The former may result in electrode fracture and debonding between the carbon/sulfur composite and the polymer binder (i.e., loss of contact between the backbone structure and the sulfur), thus leading to active material loss and increased electrical resistance, while the latter makes the cells more prone to electrolyte decomposition and other unwanted side reactions. However, it should be noted that the effect of loading on the Li anode (dendrite growth, deposition of “mossy” lithium etc.) has not yet been investigated in detail. In general, passivation of the sulfur cathode with highly insulating Li 2S and other insoluble species has been shown to be one of the main degradation mechanisms of Li–S batteries. Therefore, a homogeneous distribution of “small” sulfur particles in the electrode should help to increase the cycle life and keep the overpotential at a minimum.53,54 Moreover, encapsulation strategies seem to be promising to hinder the intermediate lithium polysulfides from leaving the cathode


18

Page 19 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


architecture.26,55–57 But they are not the “ultimate solution” to the problems, because increasing the thickness of an electrode with a poorly conducting active material, like S8, will results in worse performance, regardless of the other causes. From the cycling data in Figure 5, we conclude that the nitrogen-enriched carbon employed in this work is a suitable host matrix for the sulfur to achieve a stable cathode performance in Li–S batteries. It can be readily melt-infiltrated with sulfur to produce homogeneous composites with a reasonably high sulfur content, has sufficient electronic conductivity and seems to be able to effectively accommodate the mechanical forces associated with the volume changes of the active sulfur. Whether the high nitrogen content of ~12% is critical for the functioning of the cells is still under investigation (the effect of nitrogen-doping on the performance of Li–S batteries has not yet verified unambiguously by experiment). According to a recent paper by Li and Sun, the main effects of nitrogen-enriched carbon are the improved conductivity and the immobility of polysulfides.45 Figure 5c shows that the nitrogen-enriched carbon/sulfur cathodes with low sulfur loading can also be cycled at 1C over hundreds of cycles, with virtually no capacity fading. The battery capacity drops by only 4% between the 5th and 1000th cycles, which emphasizes the beneficial microstructure of the composite material. After 1000 cycles, the specific capacity approaches 500 mAh g–1. Voltage-capacity curves for the 300th, 500th, 700th and 900th cycles are shown in Figure S6 of the ESI. The profiles are well defined with plateaus at about 2.34/2.02 V (discharge) and 2.26/2.46 V (charge) with respect to Li/Li+ and barely change in the course of cycling. This finding helps explain the remarkably stable cell performance. In addition, top view SEM images of the cathode after 1000 cycles demonstrate that the electrode integrity is fully maintained – the morphology is virtually unaltered, with very little precipitate at the top surface


19


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 35

in charged state (see Figure S7 of the ESI). The coulombic efficiency is above 99.7% over the first 500 cycles, but declines thereafter in a non-linear way.


20

Page 21 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Capacity retention of the Li–S cells with sulfur loadings of 1 mg cm–2 (red) and 4 mg cm–2 (black.) The specific discharge capacity and coulombic efficiency are shown as a function of cycle number. After the formation cycle at C/50, the cells were cycled at C/5 in (a, b) and 1C in (c).

The rate performance of both the low loaded and highly loaded Li–S cells is shown in Figure 6. Because the discharge is known to be the kinetically rate-limiting process, the performance was measured by increasing the discharge current from C/5 to 5C while keeping the charge rate constant at C/5.58 As expected, the cells with a loading of only 1 mgsulfur cm–2 exhibit a much better rate capability – the discharge capacity declines by only ~40% when increasing the rate from C/5 to 2C as compared to ~90% for the highly loaded cells. This behavior can be explained by shorter pathways for electron and ion transport and is reflected in the less pronounced increase in overpotential (see discharge profiles in Figure S8 of the ESI). However, both the low and highly loaded cells show poor capacity retention at rates as high as 5C, suggesting that they are not suited for high power applications. Nevertheless, the use of other conducting additives might help to improve the rate performance, as recently shown by Xu et al.59 Overall, the data in Figure 5d indicate that the highly loaded cells are capable of delivering specific capacities of >500 mAh g–1 only at rates of C/5 or slightly higher and the kinetic limitations are much less severe for the low loaded cells. The fact that, irrespective of sulfur loading, similar specific capacities are obtained at C/5 is not fully understood, but might be related to a similar amount of polysulfides in the electrolyte that cannot be addressed electrochemically during cycling.60


21


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 35

Figure 6. Rate capability of the Li–S cells with sulfur loadings of 1 mg cm–2 (red) and 4 mg cm–2 (black.) The discharge current was increased from C/5 to 5C while maintaining the charge current rate at C/5. Only the specific discharge capacities are shown for clarity.

Although scarcely discussed, the electrolyte-to-sulfur mass ratio in Li–S batteries plays a very important role with regard to cycling stability and energy efficiency.25,47,61 To achieve competitive energy densities on cell level, a ratio of

Sulfur Composite

Recommend Documents