carbon nanotube hybrid films as interlayer for lithium-sulfur batteries

carbon nanotube hybrid films as interlayer for lithium-sulfur batteries

Accepted Manuscript Porous graphene oxide/carbon nanotube hybrid films as interlayer for lithium-sulfur batteries Jian-Qiu Huang, Zheng-Long Xu, Sara ...

9MB Sizes 0 Downloads 19 Views

Accepted Manuscript Porous graphene oxide/carbon nanotube hybrid films as interlayer for lithium-sulfur batteries Jian-Qiu Huang, Zheng-Long Xu, Sara Abouali, Mohammad Akbari Garakani, JangKyo Kim PII:

S0008-6223(15)30547-9

DOI:

10.1016/j.carbon.2015.12.081

Reference:

CARBON 10607

To appear in:

Carbon

Received Date: 30 September 2015 Revised Date:

8 December 2015

Accepted Date: 25 December 2015

Please cite this article as: J.-Q. Huang, Z.-L. Xu, S. Abouali, M.A. Garakani, J.-K. Kim, Porous graphene oxide/carbon nanotube hybrid films as interlayer for lithium-sulfur batteries, Carbon (2016), doi: 10.1016/ j.carbon.2015.12.081. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

1

ACCEPTED MANUSCRIPT

Porous graphene oxide/carbon nanotube hybrid films as interlayer for lithium-sulfur batteries

RI PT

Jian-Qiu Huang, Zheng-Long Xu, Sara Abouali, Mohammad Akbari Garakani, Jang-Kyo Kim* Department of Mechanical and Aerospace Engineering, The Hong Kong University of

M AN U

SC

Science and Technology, Clear Water Bay, Kowloon, Hong Kong, PR China. *Corresponding author. E-mail: [email protected] (Jang-Kyo Kim)

Abstract

Highly porous, conductive graphene oxide (GO)/carbon nanotube (CNT) composite films are synthesized via facile vacuum filtration of hybrid dispersion. The

TE D

flexible film is used as interlayer between separator and sulfur cathode to entrap active materials and prevent polysulfide shuttle. The lithium-sulfur (Li-S) battery

EP

furnished with an optimal GO/CNT interlayer delivers an excellent reversible capacity

AC C

of 671 mAh/g after 300 cycles with a low degradation rate of 0.33 mAh/g or 0.043% per cycle at 0.2 C. The encouraging outcome arises from synergistic effects of interlayer characteristics: namely, (i) the porous structure facilitates easy ion transport and electrolyte penetration; (ii) the GO layers with oxygenated functional groups entrap active materials, preventing polysulfide shuttle and enhancing their recycling; and (iii) the highly conductive CNTs offer fast pathways for electron/ion transfer. 1

ACCEPTED MANUSCRIPT 1. Introduction Rechargeable lithium-sulfur (Li-S) batteries possess an excellent theoretical

RI PT

capacity of 1675 mAh/g and a large theoretical energy density of 2567 Wh/kg. Other advantages of sulfur include natural abundance, low cost and environmental benignity, making it an attractive choice as cathode material for next-generation energy storage

SC

devices [1-3]. Nevertheless, the commercialization of Li-S batteries has been severely

10-30 S/cm at 25

M AN U

limited by several issues. The inherently low electrical conductivity of sulfur (5× ) leads to low utilization of active materials. There is volume

expansion of ~80% during conversion from sulfur to Li2S, causing large mechanical stresses on the cathode. Another critical challenge is the dissolution of lithium

S8 + 16Li+ + 16e-

TE D

polysulfides, Li2Sx (x = 4~8), as the intermediate products from the reversible reaction, 8Li2S, during the charge/discharge process [2-5]. Polysulfides

EP

are soluble in the electrolyte so that they tend to migrate between the cathode and the

AC C

anode, and are reduced to insoluble solid precipitates, Li2S/Li2S2, depositing on the surface of electrodes and subsequently leading to a loss of active materials and fast capacity fading [6].

Extensive studies have been carried out to alleviate the above issues, focusing mainly on developing novel cathode materials [7, 8]. The modification of electrolytes [9-10] and the protection of lithium anodes [11-13] have also been explored to 2

ACCEPTED MANUSCRIPT improve the electrochemical performance of Li-S batteries with varied successes. Another promising solution is to place an interlayer between the separator and the

RI PT

sulfur cathode. Porous films made from carbon nanotubes [14], microporous carbon [15], carbon nanofibers [16] and polypyrrole [17, 18] have been synthesized as the interlayers. The interlayer can effectively mitigate the shuttling effect of polysulfides

SC

by immobilizing them and recycling the active materials, so as to improve the

M AN U

electrochemical performance of Li-S batteries [14, 15]. Carbon nanotubes (CNTs) − one-dimensional carbon materials possessing excellent electrical conductivities, very high aspect ratios and good mechanical durability − have been widely studied for electrochemical energy storage applications [19]. A freestanding multi-walled CNT

TE D

interlayer inserted between the separator and the cathode improved the electrochemical performance of the battery by offering abundant second redox

EP

reaction sites for active materials [14]. Nevertheless, the capacity degradation

AC C

remained unresolved due to the prevalent polysulfide migration. Graphene oxide (GO) is a two-dimensional monolayer of hexagonal carbon atoms with massive oxygen-containing functional groups on its basal plane and edges. Once hybridized with CNTs that are easily agglomerated due to van der Waals interactions, the amphiphilic nature of GO sheets allows adsorption of CNTs onto their surfaces in water [20]. More importantly, GO sheets are known to immobilize sulfur and lithium 3

ACCEPTED MANUSCRIPT polysulfides through the reactive functional groups, such as hydroxyl and epoxide groups [21, 22]. It follows then that with the unique advantages of oxygenated

RI PT

functional groups, GO sheets can make a promising functional addition to CNT films to form GO/CNT hybrid interlayers.

Taking advantage of the above benefits, this study is dedicated to synthesize

SC

conductive and porous GO/CNT hybrid films as the interlayer for Li-S batteries by

M AN U

facile vacuum filtration. CNTs endow the films with a high electrical conductivity and a porous architecture, while GO sheets are essential to promoting uniform dispersion of CNTs and immobilization of active materials by oxygenated functional groups. These multi-functional synergies offered by hybrid interlayers favor fast ion/electron

TE D

transfer and electrolyte permeation as well as entrapment of polysulfides, with much

EP

improved electrochemical performance of the Li-S batteries.

AC C

2. Experimental

2.1. Materials and preparation of interlayers Preparation of GO/CNT films: GO dispersion was prepared using the modified

Hummers method as described in our previous work [23, 24]. The multi-walled CNTs with a diameter of 10-20 nm and a length of 10-50 µm were synthesized by a chemical vapor deposition method (supplied by Iljin Nanotech). CNTs were purified 4

ACCEPTED MANUSCRIPT to remove catalysts and other impurities by nitric acid treatment. Hybrid dispersions with three different GO/CNT ratios of 1/5, 1/2 and 1/1, or equivalent 16.7, 33.3 and

RI PT

50 wt.% GO were prepared, as shown in Figure 1a. Typically, 50 mg CNTs were added into 90 ml distilled water and ultrasonicated using a probe sonicator for 30 min. 10 ml GO dispersion (1mg/ml) was then added into the CNT dispersion, followed by

SC

ultrasonication in a bath for 2 h. GO/CNT films were obtained by vacuum filtration of

diameter). After drying at 60

M AN U

hybrid dispersions through a membrane filter (Milipore, 220 nm pore size, 100 mm in for 12 h, flexible films were peeled off. These films

were denoted as 1GO/5CNT, 1GO/2CNT and 1GO/1CNT interlayers. Films consisting of 100% CNTs or GO sheets were also prepared following the same

TE D

procedure.

Preparation of Li2S6 solution: To evaluate the polysulfide absorption by the

EP

GO/CNT interlayer, Li2S6 was employed as the representative polysulfide

AC C

intermediate. Elemental S and Li2S with a molar ratio of 5:1 were added in DME and DOL (1:1 in volume) to obtain 1mM Li2S6. The mixture was stirring at 90℃ in Ar atmosphere overnight.

2.2. Characterization The morphologies were examined under scanning electron microscopes (SEM, 6390 and 6700F) and a transmission electron microscope (TEM, JEOL 2010). The 5

ACCEPTED MANUSCRIPT elemental mapping of interlayers after cycles was carried out on an energy dispersive spectrometer (EDS) linked to SEM 6390. The surface areas of GO/CNT interlayers

RI PT

were measured using the Brunauer-Emmett-Teller (BET) equation. Nitrogen adsorption/desorption isotherms were obtained at 77 K with an automated adsorption apparatus (Micrometritics, ASAP 2020). Their electrical conductivities were

SC

measured on a four-probe resistivity/Hall system (HK5500PC, Bio-Rad). Raman

M AN U

spectra were obtained on a Reinshaw MicroRaman/Photoluminescence System. X-ray photoelectron spectroscopy (XPS, PHI5600 by Physical Electronics, Inc.) was conducted using monochromatic Al Kα X-ray at 14 kV. 2.3. Electrochemical tests

TE D

The sulfur slurry was prepared by mixing 70 wt.% commercial sulfur powders (purum p.a., ≥99.5% Sigma-Aldrich), 20 wt.% carbon black (super P) and 10 wt.%

EP

polyvinylidene fluoride (PVDF) binder in N-methy1-2-pyrrolidone (NMP) solvent.

AC C

The slurry was magnetically stirred overnight and was applied onto an aluminum foil to form a thin film. After drying at 60 °C in an air-circulated oven, the cathode electrodes were cut into discs of 14 mm in diameter with a sulfur content of ∼1.0 mg/cm2. The GO/CNT hybrid films were cut into interlayer discs of 15 mm in diameter, ~1.1 mg/cm2 in weight and ~30 µm in thickness, while the film thicknesses were ~20 and ~50 µm for the neat GO and neat CNTs, respectively. 1.0 M lithium 6

ACCEPTED MANUSCRIPT bis-trifluoromethane

sulfonylimide

(LiTFSI)

in

1,3-dioxolane

(DOL)

and

1,2-dimethoxyethane (DME) at a volume ratio of 1:1 with 1 wt.% LiNO3 additive was

RI PT

used as electrolyte. CR2032 coin cells were assembled in an argon-filled glove box using the Li foil as the anode, the prepared sulfur cathode, the electrolyte of 20 µl for each electrode and the interlayer which was placed between the polyethylene

SC

separator (Celgard 2400) and the cathode, as schematically shown in Figure 1b. The

M AN U

coin cells were charge/discharge cycled between 1.5 and 2.8 V on a LAND 2001 CT battery tester at room temperature. The electrochemical impedance spectra (EIS) were obtained on a CHI660c electrochemical workstation in the frequency range between

AC C

EP

voltage.

TE D

0.1 and 100k Hz at a constant perturbation amplitude of 5 mV using an open-circuit

7

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 1 Schematics of (a) synthesis of GO/CNT hybrid films and (b) cell configuration of the Li-S battery with an interlayer.

8

ACCEPTED MANUSCRIPT 3. Results and discussion 3.1 Morphologies and properties of interlayer films

RI PT

The morphologies and structures of interlayers with different GO contents were characterized by SEM and TEM, as shown in Figure 2 and 3. CNTs and GO sheets were seen uniformly distributed layer by layer across the thickness of the hybrid of

SC

films. CNTs imparted the film with a highly porous structure, facilitating easy

M AN U

penetration of electrolyte and fast ion/electron transfer. The interconnected network of individual CNTs intercalated between GO sheets meant high electrical conductivities of the films in both the plane and thickness directions. The SEM image taken from the top surface of 1GO/2CNT is shown in Figure 2d. CNTs were uniformly distributed

TE D

on GO layers with many surface pores which may allow the electrolyte and polysulfides to get access to the interlayer interior. TEM images of three interlayers

EP

are presented in Figure 3, indicating that CNTs were uniformly dispersed and

AC C

attached to the surface of GO sheets due to the adsorption of CNTs on the amphiphilic GO surface [20]. Raman spectra of three different hybrid interlayers and neat films are shown in Figures 3d and S1, respectively. The two peaks at 1350 and 1595 cm-1 correspond to D- and G-bands, representing the disordered sp3 structure and the sp2 hybridized graphitic structure, respectively. It is noted that the G-bands of hybrid films were located between 1587 cm-1 (for neat CNT) and 1599 cm-1 (for neat GO). 9

ACCEPTED MANUSCRIPT There was an up-shift of G-band from 1GO/5CNT to 1GO/1CNT, indicative of

TE D

M AN U

SC

RI PT

increased GO contents in the hybrid films.

EP

Figure 2 SEM images of cross-sections of (a) 1GO/5CNT, (b) 1GO/2CNT and (c)

AC C

1GO/1CNT hybrid films; and (d) top surface of 1GO/2CNT film.

10

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

EP

Figure 3 TEM images of interlayers of (a) 1GO/5CNT, (b) 1GO/2CNT and (c)

AC C

1GO/1CNT hybrid films; (d) Raman spectra of interlayers with different GO/CNT ratios.

Figure 4 depicts the nitrogen adsorption/desorption isotherm curves of the three hybrid films along with their pore size distributions in inset. The corresponding results for the neat CNT and GO films are shown in Figure S2, while the BET surface 11

ACCEPTED MANUSCRIPT area and pore volume data are given in Table 1. It is noted that the BET surface area increased with increasing GO content and peaked at 287 m2/g for the 1GO/2CNT film. The theoretical specific surface area of a completely exfoliated and isolated graphene

RI PT

sheet is ~2620 m2/g [25], an order of magnitude larger than that of MWCNTs [20]. Thus, the surface area of the hybrid film was expected to consistently increase with

SC

the addition of GO sheets. However, the surface area showed a peak at an optimal

M AN U

combination before it decreased rapidly with further increase in GO content. This happened because the GO sheets tended to restack during vacuum filtration although CNTs helped physically separate the GO layers to create a well-defined porous sandwich structure [26]. The pore distribution of the neat CNT film was mainly in the

TE D

range of 20-40 nm, consistent with a previous study [27]. Upon addition of GO sheets, significant amounts of mesopores smaller than 5 nm appeared, especially for the

EP

1GO/2CNT and 1GO/1CNT films. As expected, the total pore volume consistently

AC C

decreased with increasing GO content, or decreasing CNT content (Table 1). Similarly, the electrical conductivity also steadily decreased as the GO content increased because the GO sheets were not reduced in this study to maintain sufficient oxygenated functional groups on them.

12

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 4 Nitrogen adsorption/desorption isotherm curves and pore size distributions

TE D

of GO/CNT hybrid films.

AC C

interlayers.

EP

Table 1 Specific surface areas, pore volumes and conductivities in different

Materials

BET surface area

Pore volume

Electrical

(m2g-1)

(cm3g-1)

conductivity (S cm-1)

Neat CNT

223

1.51

26.78

1GO/5CNT

253

1.31

9.49

1GO/2CNT

287

0.88

5.64

13

ACCEPTED MANUSCRIPT 192

0.57

2.99

Neat GO

160

0.23

~10-4

RI PT

1GO/1CNT

The XPS analysis was employed to investigate the chemical compositions and the presence of functional groups in different interlayers, as shown in Figure 5. Two

SC

prominent peaks centered at binding energies 284.7 and 532.4 eV, are assigned to C

M AN U

1s and O 1s, respectively, and the relative intensity of O increased as GO content increased. The deconvoluted C 1s spectra for the three hybrid interlayers (Figure 5b) show two prominent peaks located at ∼284.4 and ∼285.0 eV which are attributed to

TE D

C=C and C-C, respectively. Four other peaks at ∼286.0, ∼287.0, ∼287.8 and ∼288.9 eV arise from the functional groups, C-OH, C-O-C, C=O and O=C-OH, respectively [28-30]. The corresponding elemental compositions and the functional groups are

EP

summarized in Tables S1 and S2, respectively. A trace amount of N was detected,

AC C

arising from the synthesis process of GO. The O/C ratio increased with increasing GO content, a reflection of surge in oxygen-containing groups as confirmed by the deconvoluted data shown in Table S2. The increases in hydroxyl (C-OH) and epoxide (C-O-C) groups were particularly notable, which may become very important in immobilizing polysulfides, as discussed below.

14

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 5 (a) XPS general spectra; and (b) deconvoluted C 1s spectra of hybrid films.

3.2 Electrochemical performance of interlayers Figures 6 and S3 present the cyclic performance of the cells with interlayers of different GO contents measured at a galvanostatic charge/discharge current of 0.2 C 15

ACCEPTED MANUSCRIPT between 1.5 and 2.8 V, and the results are summarized in Table 2. The cell with a 1GO/5CNT interlayer delivered a remarkable initial discharge capacity of 1600

RI PT

mAh/g, 96% of the theoretical value. After 100 cycles, however, the residual capacity was 670 mAh/g due to fast capacity degradation of 0.6% per cycle. In contrast, the cell with a 1GO/1CNT interlayer showed a relatively low capacity decay of 0.38% per

SC

cycle, but both the initial and residual capacities after 100 cycles were disappointingly

M AN U

low, 784 and 489 mAh/g, respectively. The interlayer made from the 1GO/2CNT film presented the optimal electrochemical performance: the battery displayed an excellent initial capacity of 1370 mAh/g and a stable residual capacity of 787 mAh/g after 100 cycles with a balanced degradation of 0.42% per cycle. The Coulombic efficiencies

TE D

(CEs) of both the cells with 1GO/2CNT and 1GO/1CNT interlayers sustained above 98% after 100 cycles, whereas the CE for the 1GO/5CNT interlayer consistently

EP

degreased, reaching 91% after 100 cycles, indicating a gradual loss of active materials.

AC C

The corresponding cyclic performance of the cells with a neat CNT or GO film interlayer was also studied. While the neat CNT interlayer showed marginally lower capacities and CEs than the 1GO/5CNT interlayer counterpart, the neat GO interlayer presented especially poor initial/residual capacities and CEs. It is suspected that the electrolyte could not easily penetrate into the tightly-packed GO interlayer and the ion/electron transfer was hindered by the almost insulating GO film (Table 1). In light 16

ACCEPTED MANUSCRIPT of the above findings, detailed studies were made of electrochemical performance of

AC C

EP

TE D

M AN U

SC

RI PT

the cell with a 1GO/2CNT interlayer in the following.

Figure 6 Cyclic performance of Li-S batteries with different GO/CNT hybrid interlayers

17

ACCEPTED MANUSCRIPT Table 2 Comparison of cyclic performance of cells with different interlayers measured before and after 100 cycles at 0.2 C.

Degradation rate

(mAh/g)

(mAh/g)

per cycle

609

0.62%

Neat CNT

0

1592

1GO/5CNT

17

1600

1GO/2CNT

33

1370

1GO/1CNT

50

Neat GO

100

RI PT

GO content (wt.%) in interlayers

Residual reversible capacity

SC

Materials

Initial discharge capacity

0.60%

787

0.42%

784

489

0.38%

43

4

0.91%

TE D

M AN U

670

Figure 7a presents the cross-sectional SEM image of the interlayer taken after

EP

100 cycles at a fully charged state. It looked different from the pristine state of the

AC C

same interlayer material (Figure 2b) with no clear signs of CNTs because a significant number of reaction products have been accumulated on its surface after cycles. The corresponding elemental mapping images (Figures 7b-7d) show that sulfur was uniformly distributed on the carbon matrix as the evidence of polysulfides captured by the interlayer. The elemental analysis by EDS gave ∼27.8 wt.% of sulfur (Figure S4). 18

ACCEPTED MANUSCRIPT Li2S6 was employed as the representative polysulfide intermediate to examine the polysulfide absorption by the GO/CNT interlayer. A fresh 1GO/2CNT interlayer was

RI PT

immersed into 1 mM Li2S6 solution and the color of the solution was changed from yellow-green to almost colorless after 2 h (see inset of Figure 7e), suggesting polysulfides were entrapped in the interlayer. The interaction between polysulfides

SC

and the interlayer was further verified by the deconvoluted XPS S 2p spectrum, as

M AN U

shown in Figure 7e. The sample was well protected in argon atmosphere before the analysis. The presence of long-chain polysulfides is reflected by the peaks at 161.8 and 163.0 eV for terminal S, and 163.1 and 164.9 eV for bridging S. The broader peaks appeared in the range of 167-170 eV are a reflection of S-O bonds, confirming

AC C

EP

TE D

the chemical interaction between sulfur and the oxygenated functional groups [31].

19

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 7 (a) SEM image of cross-section of the 1GO/2CNT after 100 cycles at a full charged state, and the corresponding EDS elemental maps of (b) carbon, (c) oxygen 20

ACCEPTED MANUSCRIPT and (d) sulfur; and (e) deconvoluted S 2p XPS spectra of the 1GO/2CNT interlayer after immersion in 1mM Li2S6 solution with a color change of the solution after 2h in

RI PT

inset.

Figure 8a shows the charge/discharge curves of the batteries with an interlayer at

SC

different stages of cycles. Two classical reduction plateaus I and II appeared at

M AN U

2.26-2.36V and 1.97-2.07V, respectively, in the first discharge curve, which correspond to the reaction of elemental sulfur, S8, to form lithium polysulfides, Li2Sx (x=4-8), and the conversion to lower-order solid Li2S2. Subsequently, Li2S was formed from Li2S2 at stage III [3]. A small plateau IV appeared at the end of the 1st

TE D

discharge process is due to the formation of solid electrolyte interface (SEI) films on the anode and the irreversible reduction reaction of LiNO3 on the cathode [32-34].

EP

The plateaus V and VI in the charge curves are consistent with the transition from

AC C

Li2S/Li2S2 to low-order lithium polysulfides and the formation of elemental sulfur [35]. There were no obvious differences in the charge/discharge profiles of the following 300 cycles, suggesting a very stable electrochemical behavior of the cell. The high rate capability of the cells with an interlayer is shown in Figure 8b. The first discharge capacity at 0.1 C reached as high as 1616 mA h/g, equivalent to 96.5% of the theoretical value. The high initial capacity and the rapid capacity decay in the 21

ACCEPTED MANUSCRIPT following cycles are attributed to the low cutting off voltage of 1.5 V where irreversible reactions of LiNO3 on electrodes occurred only in the first

RI PT

charge/discharge cycle [32-34]. When the rate was gradually increased from 0.1 to 0.2, 0.5, 1 and 2 C after every 10 cycles, the cell exhibited high and stable discharge capacities of 1026, 825, 715, 616 and 469 mAh/g, respectively. When the rate was

SC

reverted to 0.1 C, the capacity was recovered to 959 mAh/g, demonstrating excellent

M AN U

stability and reversibility of the Li-S battery with a 1GO/2CNT interlayer. The corresponding cyclic performance of the cell measured at 0.2 and 1 C is displayed in Figure 8c. The cells were tested at a lower rate of 0.1 C in the first two cycles to activate the electrodes and the interlayer. When tested at 0.2 C, the cell

TE D

sustained a high capacity of 671 mAh/g after 300 cycles from the initial discharge capacity of 1370 mAh/g, showing a very low average capacity degradation of 0.043%

EP

per cycle. At a higher current rate of 1 C, the cell delivered 441 mAh/g with an

AC C

average capacity reduction of 0.073% per cycle. The excellent cyclic stability of the Li-S batteries at both low and high rates signifies that the GO/CNT interlayer played an important role to entrap polysulfides and effectively re-utilize the active material.

22

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Figure 8 Electrochemical test results of Li-S batteries with a 1GO/2CNT interlayer: (a) charge/discharge voltage profiles determined at 0.2 C; (b) rate capacities; and (c)

AC C

efficiencies.

EP

cyclic performance measured at 0.2 C and 1 C, and the corresponding Coulombic

The electrochemical impedance spectroscopy (EIS) was conducted to ex situ

monitor the internal resistance of the battery at different cycles (Figure 9a). The characteristic resistance values calculated according to the equivalent circuit shown in inset [36, 37] are presented in Figure 9b and Table S3. Re is the resistance of electrolyte; Rst//CPEst represents the interphase contact resistance; Rct//CPEdl is the 23

ACCEPTED MANUSCRIPT charge-transfer resistance; and W0 represents the diffusion resistance. Two semicircles at high and medium frequency regions correspond to Rst//CPEst and Rct//CPEdl,

RI PT

respectively. It is noted that Re surged after the first cycle because of the formation of soluble polysulfides during the reaction, resulting in increases in viscosity and resistance of the electrolyte [36]. In the following cycles, it gradually decreased due to

SC

the entrapment of polysulfides by the interlayer via the interactions between the

M AN U

polysulfide intermediates and the oxygenated function groups, especially hydroxyl and epoxide groups [21, 38]. The rapid reduction of Rst in the first few cycles indicates close contacts among the electrolyte, active material and interlayer through the rearrangement of active materials at more electrochemical sites in the interlayer

TE D

[39]. Upon the first charge, Rct abruptly increased due to the insulating agglomerates of active materials on the surface of cathodes and with further cycles it gradually

EP

decreased owing to the re-deposition of polysulfides by entrapment on the GO/CNT

AC C

interlayer. It is remarkable that the generally low resistance values after 300 cycles, though mild increases in Rst and Rct from the 50th to 300th cycle, present an extremely stable electrochemical reaction environment in the Li-S battery with the 1GO/2CNT hybrid interlayer [39-41].

24

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 9 (a) Electrochemical impedance spectra (EIS) and (b) corresponding variations of Re, Rst and Rct of the Li-S cell with a 1GO/2CNT interlayer as a function of cycle number.

To identify the standing of this work among published results, a comparison is 25

ACCEPTED MANUSCRIPT made of electrochemical performance between the current work and the cells with carbon-based and other interlayers, as shown in Table 3. Among many different

RI PT

interlayer systems, the performance of the current cells with optimal GO/CNT interlayers was superior or comparable to the other interlayers, such as nickel foam [41], graphene oxide/carbon black [42], carbon fiber monolith [43] and

SC

cellulose/graphene oxide [44]. Although the interlayers made from MWCNTs [14]

M AN U

and microporous carbon [15] showed better performance for short cycles than the GO/CNT interlayer, the hybrid interlayer presents a much longer life of 300 cycles with a slower degradation rate of 0.17% per cycle, offering an alternative choice for

AC C

EP

TE D

applications requiring long-term cyclic stability.

26

ACCEPTED MANUSCRIPT Table 3 Comparison of electrochemical performance of Li-S batteries with different interlayers.

Cycle number

336

819

80

Fe3C/CNF

200

1177

100

Graphene oxide/carbon black

168

1260

336

1120

Cellulose/ graphene oxide

300

2736

Acetylene black

168

Carbon paper

336

MWCNT

GO/CNT

0.33%

893

Reference

[41]

0.24%

[16]

0.29%

[42]

100

630

0.44%

[43]

200

474

0.42%

[44]

1491

50

1062

0.58%

[45]

1500

50

810

0.92%

[46]

TE D

894

838

1446

100

855

0.41%

[14]

1675

1176

100

1000

0.15%

[15]

336

1370

300

671

0.17%

Current study

AC C

Microporous carbon

Degradation rate per cycle

100

EP

Carbon fiber monolith

604

M AN U

Nickel foam

Residual reversible capacity (mAh/g)

RI PT

Initial discharge capacity (mAh/g)

SC

Interlayer materials

Discharge current rate (mA/g)

4. Conclusions Using a facile, low-cost vacuum filtration technique, we successfully synthesized freestanding GO/CNT hybrid films with optimized structures as the interlayer for Li-S 27

ACCEPTED MANUSCRIPT batteries. Among different compositions, the 1GO/2CNT interlayer containing 33.3 wt.% GO showed the most optimal electrochemical performance. The battery

RI PT

displayed an excellent initial capacity of 1370 mAh/g and a stable residual capacity of 787 mAh/g with a Coulombic efficiency over 98% after 100 cycles. It also delivered an excellent capacity of 671 mAh/g after 300 cycles with a very low degradation rate

SC

of 0.33 mAh/g or 0.043% per cycle at a rate of 0.2 C. With CNTs intercalated between

M AN U

the GO sheets, the flexible GO/CNT hybrid interlayer possessed many useful characteristics with synergistic effects that were responsible for the enhanced capacities and cyclic stability. The high electrical conductivity and the high porosity of the interlayer were beneficial to fast ion/electron transfer and electrolyte

TE D

penetration. The oxygenated functional groups present on the GO layers played a crucial role in immobilizing active materials, preventing polysulfide shuttle, and thus

AC C

EP

facilitating their recycling.

Acknowledgments

This project was financially supported by the Research Grants Council (GRF Projects: 613612 and 16212814) and the Innovation and Technology Commission (ITF Project code: ITS/318/14) of Hong Kong SAR. The authors also appreciate the technical assistance from the Materials Characterization and Preparation Facilities (MCPF) of 28

ACCEPTED MANUSCRIPT HKUST.

This document file contains supplementary information.

SC

References

RI PT

Appendix A. Supplementary data

M AN U

[1] Bruce PG, Freunberger SA, Hardwick LJ, Tarascon JM. Li–O2 and Li–S batteries with high energy storage. Nat Mater 2012;11(1):19–29.

[2] Barghamadi M, Kapoor A, Wen C. A Review on Li-S Batteries as a High Efficiency Rechargeable Lithium Battery. J Electrochem Soc 2013;160(8):1256-63.

TE D

[3] Ji X, Nazar LF. Advances in Li–S batteries. J Mater Chem 2010;20(44):9821-6. [4] Seh ZW, Li W, Cha JJ, Zheng G, Yang Y, McDowell MT, et al. Sulphur–TiO2

EP

yolk–shell nanoarchitecture with internal void space for long-cycle lithium–sulphur

AC C

batteries. Nat Commun 2013;4:1331-6. [5] Jung DS, Hwang TH, Lee JH, Koo HY, Shakoor RA, Kahraman R, et al. Hierarchical porous carbon by ultrasonic spray pyrolysis yields stable cycling in lithium–sulfur battery. Nano Lett 2014;14(8):4418-25. [6] Demir-Cakan R, Morcrette M, Gangulibabu, Guéguen A, Dedryvère R, Tarascon JM. Li–S batteries: simple approaches for superior performance. Energy Environ Sci 29

ACCEPTED MANUSCRIPT 2013;6(1):176-82. [7] Manthiram A, Fu F, Su YS. Challenges and prospects of lithium-sulfur batteries.

RI PT

Acc Chem Res 2013;46(5):1125-34. [8] Wang J, He YS, Yang J. Sulfur‐based composite cathode materials for high‐energy rechargeable lithium batteries. Adv Mater 2015;27(3): 569-75.

SC

[9] Zhang SS. Liquid electrolyte lithium/sulfur battery: Fundamental chemistry,

M AN U

problem, and solution. J. Power Sources 2013;231:153-62.

[10] Wang Q, Jin J, Wu X, Ma G, Yang J, Wen Z. A shuttle effect free lithium sulfur battery based on a hybrid electrolyte. Phys Chem Chem Phys 2014;16(39):21225-9. [11] Lee YM, Choi NS, Park JH, Park JK. Electrochemical performance of

TE D

lithium/sulfur batteries with protected Li anodes. J Power Sources 2003;119:964-72. [12] Ma G, Wen Z, Wu M, Shen C, Wang Q, Jin J, et al. A lithium anode protection

AC C

21225-9.

EP

guided highly-stable lithium–sulfur battery. Phys Chem Chem Phys 2014;16(39):

[13] Ma G, Wen Z, Wang Q, Shen C, Jin J, Wu X. A lithium anode protection guided highly-stable lithium–sulfur battery. Chem Commun. 2014;50:14209-12. [14] Su YS, Manthiram A. A new approach to improve cycle performance of rechargeable lithium–sulfur batteries by inserting a free-standing MWCNT interlayer. Chem Commun 2012;48(70):8817-9. 30

ACCEPTED MANUSCRIPT [15] Su YS, Manthiram A. Lithium–sulphur batteries with a microporous carbon paper as a bifunctional interlayer. Nat Commun 2012;3:1166.

RI PT

[16] Huang JQ, Zhang B, Xu ZL, Abouali S, Akbari Garakani M, Huang J, et al. Novel interlayer made from Fe3C/carbon nanofiber webs for high performance lithium-sulfur batteries. J Power Sources 2015;285:43-50.

SC

[17] Ma G, Wen Z, Jin J, Wu M, Wu X, Zhang J. Enhanced cycle performance of Li-S

M AN U

battery with a polypyrrole functional interlayer. J Power Sources 2014;267:542-6. [18] Ma G, Wen Z, Wang Q, Shen C, Peng P, Jin J, et al. Enhanced performance of lithium sulfur battery with self-assembly polypyrrole nanotube film as the functional interlayer. J Power Sources 2015;273(1):511-6.

TE D

[19] Liu XM, Huang ZD, Oh SW, Zhang B, Ma PC, Yuen MMF, et al. Carbon nanotube (CNT)-based composites as electrode material for rechargeable Li-ion

EP

batteries: A review. Compos Sci Technol 2012;72(2):121-44.

AC C

[20] Huang ZD, Zhang B, Oh SW, Zheng Q, Lin X, Yousefi N, et al. Self-assembled reduced graphene oxide/carbon nanotube thin films as electrodes for supercapacitors. J Mater Chem 2012;22(8):3591-9. [21] Ji L, Rao M, Zheng H, Zhang H, Li Y, Duan W, et al. Graphene oxide as a sulfur immobilizer in high performance lithium/sulfur cells. J Am Chem Soc 2011;133(46):18522-5. 31

ACCEPTED MANUSCRIPT [22] Zhang L, Ji L, Glans PA, Zhang Y, Zhu J, Guo J. Electronic structure and chemical bonding of a graphene oxide–sulfur nanocomposite for use in superior

RI PT

performance lithium–sulfur cells. Phys Chem Chem Phys 2012;14(39):13670-5. [23] Geng Y, Wang SJ, Kim JK. Preparation of graphite nanoplatelets and graphene sheets. J Colloid Interf Sci 2009;336(2):592-8.

SC

[24] Zheng Q, Li Z, Yang JH, Kim JK. Graphene oxide-based transparent conductive

M AN U

films. Prog Mater Sci 2014;64:200-47.

[25] Stankovicha S, Dikin DA, Dommett GA, Kohlhaas KM, Jimney EJ, Stach EA, et al. Graphene-based composite materials. Nature 2006;442(7100):282–6. [26] Huang ZD, Zhang B, Liang R, Zheng QB, Oh SW, Lin XY, et al. Effects of

TE D

reduction process and carbon nanotube content on the supercapacitive performance of flexible graphene oxide papers. Carbon 2012;50(11):4239-51.

EP

[27] Jiang Q, Qu MZ, Zhou GM, Zhang BL, Yu ZL. A study of activated carbon

AC C

nanotubes as electrochemical super capacitors electrode materials. Mater Lett 2002;57(4):988-91.

[28] Kim J, Cote LJ, Kim F, Yuan W, Shull KR, Huang J. Graphene Oxide Sheets at Interfaces. J Am Chem Soc 2010;132(23):8180–6. [29] Yang SY, Chang KH, Tien HW, Lee YF, Li SM, Wang YS, et al. Design and tailoring of a hierarchical graphene-carbon nanotube architecture for supercapacitors. 32

ACCEPTED MANUSCRIPT J Mater Chem 2011;21(7):2374-80. [30] Zheng Q, Zhang B, Lin X, Shen X, Yousefi N, Huang ZD, et al. Highly

hybrid

films

produced

by

Langmuir–Blodgett

RI PT

transparent and conducting ultralarge graphene oxide/single-walled carbon nanotube assembly.

2012;22(7):25072–82.

J

Mater

Chem

SC

[31] Liang X, Hart C, Pang Q, Garsuch A, Weiss T, Nazar LF. A high efficient

M AN U

polysulfide mediator for lithium-sulfur batteries. Nat Commun 2015:6;5682 [32] Zhang SS. Role of LiNO3 in rechargeable lithium/sulfur battery. Electrochim Acta 2012;70:344-8.

[33] Zhang SS. Liquid electrolyte lithium/sulfur battery: Fundamental chemistry,

TE D

problem, and solution. J Power Sources 2013;231:153-62. [34] Rosenman A, Elazari R, Salitra G, Markevich E, Aurbach D, Garsuch A. J

EP

Electrochem Soc. 2015;162(3):A470-3.

AC C

[35] Wang J, He YS, Yang J. Sulfur‐based composite cathode materials for high‐ energy rechargeable lithium batteries. Adv Mater 2015;27(3):569-75. [36] Deng Z, Zhang Z, Lai Y, Liu J, Li J, Liu Y. Electrochemical impedance spectroscopy study of a lithium/sulfur battery: modeling and analysis of capacity fading. J Electrochem Soc 2013;160(4):553-8. [37] Zhang B, Liu Y, Huang Z, Oh S, YuY, Mai YW, et al. Urchin-like Li4Ti5O12– 33

ACCEPTED MANUSCRIPT carbon nanofiber composites for high rate performance anodes in Li-ion batteries. J Mater Chem 2012;22(24):12133–40.

RI PT

[38] Zhou G, Yin LC, Wang DW, Li L, Pei S, Gentle IR, et al. Fibrous hybrid of graphene and sulfur nanocrystals for high-performance lithium–sulfur batteries. Acs Nano 2013;7(6):5367-75.

SC

[39] Chung SH and Manthiram A. Lithium–sulfur batteries with superior cycle

M AN U

stability by employing porous current collectors. Electrochim Acta 2013;107:569-76. [40] Kolosnitsyn VS, Kuzmina EV, Karaseva EV, Mochalov SE. A study of the electrochemical processes in lithium–sulphur cells by impedance spectroscopy. J Power Sources 2011;196(3):1478-82.

TE D

[41] Zhang K, Qin F, Fang J, Li Q, Jia M, Lai Y, et al. Nickel foam as interlayer to improve the performance of lithium–sulfur battery. J Solid State Electrochem

EP

2014;18(4):1025-29.

AC C

[42] Wang X, Wang Z, Chen L. Reduced graphene oxide film as a shuttle-inhibiting interlayer in a lithium–sulfur battery. J Power Sources 2013;242:65-9. [43] Zhang L, Wang Y, Peng B, Yu W, Wang H, Wang T, et al. Preparation of a macroscopic, robust carbon-fiber monolith from filamentous fungi and its application in Li–S batteries. Green Chem 2014;16(8):3926-34. [44] Zeng Q, Leng X, Wu KH, Gentle IR, Wang DW. Electroactive 34

ACCEPTED MANUSCRIPT cellulose-supported

graphene oxide interlayers

for

Li–S

batteries.

Carbon

2015;93:611-9.

RI PT

[45] Jeonga TG, Moona YH, Chun HH, Kim HS, Cho BW, YT Kim. Free standing acetylene black mesh to capture dissolved polysulfide in lithium sulfur batteries. Chem Commun 2013;49(94):11107-9.

SC

[46] Zhang K, Li Q, Zhang L, Fang J, Li J, Qin F, et al. From filter paper to carbon

AC C

EP

TE D

M AN U

paper and toward Li–S battery interlayer. Mater Lett 2014;121:198–201.

35