This journal is©The Royal Society of Chemistry 2014 Chem. Commun., 2014, 50, 7011--7013 | 7011

Cite this:Chem. Commun., 2014,

50, 7011

TPPi as a flame retardant for rechargeable lithiumbatteries with sulfur composite cathodes†

Hao Jia,a Jiulin Wang,*a Fengjiao Lin,a Charles W. Monroe,b Jun Yanga andYanna NuLia

Triphenyl phosphite (TPPi) is adopted as a flame retardant to improve

the safety of rechargeable lithium batteries with sulfur composite

cathodes. The thermal stability of the electrolyte is greatly enhanced

after the addition of TPPi, which also has a positive impact on the

electrochemical performance of the Li–S batteries. TPPi facilitates

the formation of SEI, resulting in a smaller interfacial impedance and

a better rate performance. The addition of about 5 wt% TPPi greatly

reduces the polarization voltage, stabilizing the cycle performance

of the battery. This indicates that an optimized addition of TPPi is a

favorable additive in conventional liquid electrolytes for recharge-

able Li–S batteries with high performances and good safety.

Lithium ion batteries demonstrate promising power sources forelectrical vehicles and energy storage devices for smart grid, solarand wind stations. Because of the active electrode materials andthe flammable organic electrolyte, the safety of lithium ionbatteries is a very important issue for their practical applications.Recently secondary lithium–sulfur (Li–S) batteries, one of mostpromising types of battery, have been investigated intensively dueto their high theoretical capacity, low cost and non-toxicity.1

However, Li–S batteries suffer more serious safety issues comparedto lithium ion batteries. Three main factors influencing the safetyperformances of Li–S batteries are as follows: the formation oflithium dendrites in batteries using the inevitable lithium anode,which are the causes of battery dysfunction and heat emission;2,3

liquid electrolytes used in batteries are highly volatile and inflam-mable, and at high temperatures the electrolytes tend to react withboth the anode and cathode, generating even more heat in thebattery and causing thermo runaways;4 and insulator sulfurwith conductive agents (usually acetylene black) are combustible,bringing hazards to the system.5

In our previous work, the safety of sulfur cathodes has beensignificantly improved by combining elemental sulfur with non-flammable polyacrylonitrile (PAN).6 However, traditional electrolyteswith carbonates or ethers are highly flammable and might causesafety hazards when the cells are misused. In traditional lithium ionbatteries, the thermal stabilities and electrochemical performancesof conventional electrolytes have been enhanced by flame retar-dants. The most explored flame retardants used in batteries can becategorized into the following groups: alkyl phosphates,7–9 fluori-nated alkyl phosphates,10 ionic liquids,11 and phosphazenes.12 Mostof these flame retardants were found to greatly enhance the safety oflithium ion batteries. However, flame retardants also have a negativeimpact to some extent on lithium ion batteries’ electrochemicalperformances, for example they can destroy the cycling stability ordeteriorate the power rate properties. Ideal flame retardants whichcompletely resolve the safety issues of batteries and at the same timehave no negative impact on the electrochemical performances arestill under investigations.

Triphenyl phosphite (TPPi) is a chemical compound with theformula P(OC6H5)3, as shown in Fig. S1 (ESI†). This colorlessliquid with a viscosity of 17.7 mm2 s�1 at 20 1C is the ester ofphosphorous acid and phenol, and generates phosphates andretards combustion when heated. In this work, TPPi was adoptedas a flame retardant additive into the traditional electrolyte 1 MLiPF6/EC + DMC (1 : 1, v/v) which not only enhanced the carbo-nates’ thermal properties but also improved the electrochemicalperformances of rechargeable Li–S batteries.

The thermal properties of the blank electrolyte and thefunctional electrolytes with different doses of TPPi were char-acterized via DSC as shown in Fig. 1(a). For the blank electrolyte,the thermolysis temperature was about 266 1C, which shifted toabout 275 1C after 10 wt% TPPi addition. The thermal stability ofthe electrolytes coupled with the cathode as a whole was alsoinvestigated as shown in Fig. 1(b). The thermolysis temperaturewas pushed 20 1C higher, demonstrating an enhanced thermalstability. DSC studies carried out in a sealed pan withoutpinholes showed similar results (Fig. S2, ESI†). Since such asmall dose of TPPi is incapable of bringing significant changes

a School of Chemistry and Chemical Engineering, Shanghai Jiao Tong University,

Shanghai 200240, P. R. China. E-mail: wangjiulin@sjtu.edu.cn;

Fax: +86-21-54747667; Tel: +86-21-54745887b Department of Chemical Engineering, University of Michigan, Ann Arbor,

MI 48109, USA

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4cc01151a

Received 13th February 2014,Accepted 15th April 2014

DOI: 10.1039/c4cc01151a

www.rsc.org/chemcomm

ChemComm

COMMUNICATION

Publ

ishe

d on

15

Apr

il 20

14. D

ownl

oade

d by

Mic

higa

n T

echn

olog

ical

Uni

vers

ity o

n 21

/10/

2014

22:

21:3

2.

View Article OnlineView Journal | View Issue

7012 | Chem. Commun., 2014, 50, 7011--7013 This journal is©The Royal Society of Chemistry 2014

to the heat capacity of the system, the change in the heat flowindicates that the pyrolysis reaction is efficiently suppressedby TPPi. When heated above the pyrolysis temperature, TPPigenerates free radicals (such as PO�), which are able to activelycapture other free radicals emitted by the burning electrolyte(usually H� and �OH) to retard the reaction.12 Additionally,metaphosphoric acid and polyvinylidene acid, along with otherphosphorous compounds produced in this process form anoxygen-proof layer that helps distinguish the flame on thesurface of the combustible materials.13

The influence of TPPi on the conductivity is shown inFig. 2(a). The conductivity of the electrolytes decreased as theconcentration of TPPi increased due to its high viscosity. Withthree equivalent phenyl groups in the molecule, the polarity ofTPPi is weak. This hinders the dissociation of LiPF6, resulting ina lower conductivity. However, electrolytes with TPPi demon-strated lower interfacial impedances. After being fully charged inthe 2nd cycle at 0.1 C, the electrochemical impedances of thesebatteries were tested by AC impedance. As is shown in Fig. 2(b),the impedances of the electrolytes diminished after adding TPPi.This is possibly attributed to the electron-rich phenyl structure ofTPPi, which could facilitate the formation of SEI,14 enabling afaster lithium ion transportation.15 The mechanism concerningthe formation of SEI is still not thoroughly elucidated as it variesin different systems. It is commonly accepted that SEI formationin all these systems involves a reductive decomposition of theelectrolyte components (solvents, salts, additives).14 TPPi consistsof three reductive phenoxy groups, which probably promote theformation of SEI and induce a lower interfacial impedance.The addition of TPPi does not affect the electrochemical windows

of the electrolyte, and the benefits to the plating/stripping of theLi metal on the anode.

Fig. 3(a) presents a comparison of batteries using electrolyteswith different TPPi contents. It should be noted that the sulfurcontent in the composite materials was ca. 45 wt%, as determinedby elemental analysis. The specific capacities in this paper werecalculated based on the whole weight of the composite materialsincluding the pyrolyzed PAN matrix and the pure sulfur, and thisis noted in the units of (mA h g�1)(composite) and (mA h g�1)(sulfur) inthe figures, respectively. After 45 cycles, the capacity retentionrates of these batteries were 90.4% for the normal electrolyte,94.5% for the functional electrolyte containing 5 wt% TPPi and94.2% for the functional electrolyte containing 10 wt% TPPi. Thebatteries using electrolytes with 20 wt% TPPi demonstratedrelatively lower capacities. When the concentration of TPPireached 20 wt%, the high viscosity of TPPi decreased the ionicconductivity of the electrolyte, and even influenced the intimatecontact between the electrolyte and the cathode material. As aresult, the Li+ transfer at the interface deteriorated. Fig. 3(b) showsthe polarization voltage profiles after the TPPi additions. Thebattery using an electrolyte doped with 5 wt% TPPi manifestedthe smallest polarization voltage. It can be inferred that when theconcentration of TPPi is constrained to around 5 wt%, thecontradictory effects brought by TPPi in the electrolyte conduc-tivity and interface impedance reach an optimal level.

The rate performance of the S–PAN cathode also benefitsfrom the presence of TPPi. The batteries with different TPPicontents were recharged and discharged at different ratesvarying from 0.2 C to 5 C. The battery with blank electrolytewas more sensitive to rate variations. As shown in Fig. 4(a), inlow rate zone (o1 C), the differences among three electrolytesare negligible. Significant discrepancies emerged in the high

Fig. 1 (a) DSC profiles of the electrolytes, in which the black line indicatesthe blank electrolyte 1 M LiPF6/EC + DMC (1 : 1, v/v) and the red line indicatesthe blank electrolyte with 10 wt% TPPi additive; (b) DSC profiles of theelectrolytes with fully charged sulfur cathodes, in which the black lineindicates the blank electrolyte and the red line indicates the blank electrolytewith 10 wt% TPPi additive. The electrolytes were sealed in an aluminum panin the glove box. A pinhole was punched on the pan prior to the DSC testing.

Fig. 2 (a) Effects of TPPi addition on SET and ionic conductivity;(b) impedance spectra (at the fully charged state) of the cells withelectrolytes containing 0, 5 and 10 wt% TPPi.

Communication ChemComm

Publ

ishe

d on

15

Apr

il 20

14. D

ownl

oade

d by

Mic

higa

n T

echn

olog

ical

Uni

vers

ity o

n 21

/10/

2014

22:

21:3

2.

View Article Online

This journal is©The Royal Society of Chemistry 2014 Chem. Commun., 2014, 50, 7011--7013 | 7013

rate zone, where the capacity of the battery with a blankelectrolyte faded to 160 mA h g�1 (5 C), while the 5 wt% TPPidoped battery maintained a capacity of 400 mA h g�1 (5 C). The5% TPPi doped battery illustrates the best rate performance, afact which coincides with the recharge–discharge pattern whichshows that the 5 wt% TPPi battery reduces the polarization

voltage most significantly. Similar to the battery with the blankelectrolyte, the battery with 10 wt% TPPi additive also suffers arapid capacity loss at a high rate. As the current increases, therate performance of the batteries is primarily determined by thetransportation of the lithium ions in the electrode, interface,and electrolyte. The electrolyte with 10 wt% TPPi possesses ahigh viscosity, which hinders fast lithium ion transportation inthe electrolyte and therefore deteriorates the rate performance.All batteries with TPPi addition recovered well from the highrate test, indicating that the reversibility of the electrodes iswell preserved by TPPi. The effect of 5 wt% TPPi addition on thedischarge voltage at different rates is shown in Fig. 4(b). Underthe same rate, the PAN–S cathode demonstrates a higherdischarge voltage and a larger capacity when doping with TPPi.The initial voltage of the regular battery decreased from 2.1 V(0.1 C) to 1.7 V (2 C) (Fig. S2, ESI†), whereas the voltage values ofthe batteries with additives were sustained at 1.85 V, as shownin Fig. 4(b), since they benefited from the formation of SEI withless impedance. Since the interface is enhanced, lithium is ableto more readily enter the internal structure of the electrodes,resulting in a low polarization voltage and insensitivity to ratevariation.

In summary, safety is a very important issue for rechargeablebatteries and in this work TPPi was successfully adopted as aflame retardant in rechargeable lithium batteries with a PAN–Scomposite cathode. Despite its disadvantages such as highviscosity and weak polarity, TPPi is applicable as an additivein the traditional electrolyte to enhance its thermal stabilityand improve the electrochemical performance of Li–S batteries.The optimal content of TPPi is around 5 wt%, with which thebatteries show the small polarization voltage, a good rateperformance and stable cycle performances.

This work is financially supported by the National NaturalScience Foundation of China (51272156, 21333007) and theSJTU-UM joint research project.

Notes and references1 P. G. Bruce, S. A. Freunberger, L. J. Hardwick and J.-M. Tarascon,

Nat. Mater., 2011, 11, 19.2 O. Crowther and A. C. West, J. Electrochem. Soc., 2008, 155, A806.3 C. Monroe and J. Newman, J. Electrochem. Soc., 2003, 150, A1377.4 T. Ohsaki, T. Kishi, T. Kuboki, N. Takami, N. Shimura, Y. Sato,

M. Sekino and A. Satoh, J. Power Sources, 2005, 146, 97.5 J. Wang, J. Chen, K. Konstantinov, L. Zhao, S. Ng, G. Wang, Z. Guo

and H. Liu, Electrochim. Acta, 2006, 51, 4634.6 F. J. Lin, J. L. Wang, H. Jia, C. W. Monroe, J. Yang and Y. N. NuLi,

J. Power Sources, 2013, 223, 18.7 X. Xia, P. Ping and J. R. Dahn, J. Electrochem. Soc., 2012, 159, A1834.8 R. Shibutani and H. Tsutsumi, J. Power Sources, 2012, 202, 369.9 Y. Shigematsu, M. Ue and J. Yamaki, J. Electrochem. Soc., 2009, 156, A176.

10 G. Nagasubramanian and C. J. Orendorff, J. Power Sources, 2011,196, 8604.

11 J. Choi, Y.-K. Sun, E.-G. Shim, B. Scrosati and D.-W. Kim, Electrochim.Acta, 2011, 56, 10179.

12 T. Tsujikawa, K. Yabuta, T. Matsushita, T. Matsushima, K. Hayashiand M. Arakawa, J. Power Sources, 2009, 189, 429.

13 E. P. Roth and C. J. Orendorff, Electrochem. Soc. Interface, 2012, 45,summer.

14 M. Inaba, Y. Kawatate, A. Funabiki, S.-K. Jeong, T. Abe andZ. Ogumi, Electrochim. Acta, 1999, 45, 99.

15 M. Winter and R. J. Brodd, Chem. Rev., 2004, 104, 4245.

Fig. 3 (a) Cycle performances at 0.5 C, except for the cell with 20 wt%TPPi which was initially activated at 0.06 C for 3 cycles, then 0.5 C for thefollowing cycles; and (b) charge–discharge profiles for the 40th cycle ofthe cells with electrolytes containing 0, 5 and 10 wt% TPPi at 0.1 C.

Fig. 4 (a) Rate performances of the cells with and without the TPPiadditive; (b) discharge profiles for the PAN–S composite electrodes atdifferent rates with the blank electrolyte and the electrolyte with 5 wt% TPPi.

ChemComm Communication

Publ

ishe

d on

15

Apr

il 20

14. D

ownl

oade

d by

Mic

higa

n T

echn

olog

ical

Uni

vers

ity o

n 21

/10/

2014

22:

21:3

2.

View Article Online