Phosphoryl-rich flame-retardant ions (FRIONs): Towards safer lithium-ion batteries

Article abstract of DOI:10.1002/anie.201310867

The functionalized catecholate, tetraethyl (2,3-dihydroxy-1,4-phenylene) bis(phosphonate) (H₂-DPC), has been used to prepare a series of lithium salts Li[B(DPC)(oxalato)], Li[B(DPC)₂], Li[B(DPC)F ₂], and Li[P(DPC)₃

Full text of DOI:10.1002/anie.201310867

Angewandte
Chemie
DOI: 10.1002/anie.201310867
Functional Materials
Phosphoryl-Rich Flame-Retardant Ions (FRIONs): Towards Safer
Lithium-Ion Batteries**
Michael F. Rectenwald, Joshua R. Gaffen, Arnold L. Rheingold, Alexander B. Morgan, and
John D. Protasiewicz*
Abstract: The functionalized catecholate, tetraethyl (2,3-dihy-
droxy-1,4-phenylene)bis(phosphonate) (H2-DPC), has been
used to prepare a series of lithium salts Li[B(DPC)(oxalato)],
Li[B(DPC)2], Li[B(DPC)F2], and Li[P(DPC)3]. The phos-
phoryl-rich character of these anions was designed to impart
flame-retardant properties for their use as potential flame-
retardant ions (FRIONs), additives, or replacements for other
lithium salts for safer lithium-ion batteries. The new materials
were fully characterized, and the single-crystal structures of
Li[B(DPC)(oxalato)] and Li[P(DPC)3] have been deter-
mined. Thermogravimetric analysis of the four lithium salts
show that they are thermally stable up to around 2008C.
Pyrolysis combustion flow calorimetry reveals that these salts
produce high char yields upon combustion.
novel salt lithium bis(oxalato)borate (LiBOB),[4] which has
been shown to improve certain aspects of the solid electrolyte
interface (SEI) formed at the LIB battery anode.[5]
One means to reduce the flammability of the electrolyte
solution is to add flame-retardant compounds to the mixture.
A number of studies have shown that addition of organo-
phosphorus(V) compounds, such as DMMP,[6] result in
desirable protection to systems.[7] However, such additives
may have deleterious effects on the operation of the battery,
especially if they are susceptible to reduction at the low
potentials of the anode, and could thus lead to capacity fading.
Additives can also disrupt formation of a robust SEI. Finally,
incorporation of additives to a closed system requires
displacement of vital LIB components, and hence also lead
to loss of capacity.
Our particular strategy involves the synthesis of flame-
retardant ions (FRIONs) as bifunctional electrolytic salts.
The ground-up design of FRIONs offers one the opportunity
to replace some or all of a conventional lithium salt, such as
LiPF6, without jeopardizing loss of the LIB capacity as well as
risk incompatible chemistries. Our first foray into FRIONs
resulted in the identification of FRION 1.[8] The particular
FRION combines attributes of LiBOB[9] and the flame-
retardant qualities introduced by the presence of phosphinate
groups. Unfortunately, FRION 1 suffered from low yields and
limited solubility. Herein we communicate the successful
design of a new class of lithium salts having attributes suitable
for applications of anions where flame-retardant qualities are
desired.
L
ithium-ion batteries (LIBs) have become the battery of
choice for portable electronics and mobile communication
devices because of their high energy density and their lack of
a memory effect. As their size scales towards transportation
applications there have been concerns surrounding their
safety in light of an increasing number of well-publicized
battery fires. High energy density LIBs present unique
challenges, for a fully charged battery contains reactive and
hazardous materials in the presence of flammable organic
solvents. Catastrophic failures of a lithium-ion battery often
include thermal runaway combined with fire and toxic fumes.
It is thus unsurprising that a number of strategies are actively
being pursued to manage the risks associated with LIBs.[1]
One of the primary components of LIBs is the electrolyte,
which in most commercial LIBs is a solution of LiPF6 in
mixtures of organic carbonates, including ethylene carbonate,
dimethyl carbonate, and diethyl carbonate.[2,3] Of note is the
[*] M. F. Rectenwald, J. R. Gaffen, Prof. J. D. Protasiewicz
Department of Chemistry, Case Western Reserve University
10900 Euclid Ave., Cleveland, OH 44106 (USA)
E-mail: protasiewicz@case.edu
Prof. A. L. Rheingold
Department of Chemistry and Biochemistry, University of California
at San Diego, La Jolla, CA 92093-0358 (USA)
To enhance the thermal stability of FRIONs, we sought to
utilize a chelating organophosphorus entity to replace the two
phosphinato groups in FRION 1. We thus examined the use
of tetraethyl (2,3-dihydroxy-1,4-phenylene)bis(phosphonate)
[H2-DPC, Eq. (1)]. This particular material was previously
Dr. A. B. Morgan
Energy Technology and Materials Division, University of Dayton
Research Institute, Dayton, OH 45469 (USA)
[**] We thank the Assistant Secretary for Energy Efficiency and Renew-
able Energy, Office of Vehicle Technologies, of the U.S. Department
of Energy under Contract DE-AC02-05CH11231 and Subcontract
6878940, for support, and Prof. D. A. Scherson (CWRU) for helpful
comments.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201310867.
Angew. Chem. Int. Ed. 2014, 53, 4173 –4176
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4173

.
Angewandte
Communications
prepared in an impure fashion by the anionic phospho-Fries
rearrangement of 1,2-phenylene tetraethyl bis(phosphate).[10]
Changing the aqueous ammonium chloride quench to an
aqueous hydrochloric acid (4m) quench leads to a significantly
purer product which is suitable for subsequent use.
Reaction of two equivalents of H2-DPC with boric acid
and LiOtBu in refluxing toluene leads to Li[B(DPC)2], in
73% yield, as an analytically pure white solid after drying
(Scheme 1). Likewise, reaction of H2-DPC with oxalic acid,
Li[B(DPC)(oxalato)], and the results of a crystallographic
analysis are presented in Figure 1.[11] The 1:1 ratio of DPC and
oxalato ligands is rigorously established, and geometries
about the boron and lithium centers are tetrahedral. The
lithium ion is coordinated by a phosphoryl oxygen atom
(O10) and a catecholate oxygen atom (O2) of one [B(DPC)-
(oxalato)] anion, as well as by a phosphoryl oxygen atom (O7)
and an oxalato oxygen atom (O5) of two other [B(DPC)-
(oxalato)] anions, and gives rise to a coordination polymer in
the solid state. The Li-O distances span the range from
1.867(1) to 1.989(2) ꢀ, and are within the range of Li-O
distances reported for other lithium oxalatoborate and
lithium catecholatoborate salts.[12]
Figure 1. Molecular structure of Li[B(DPC)(oxalato)] in the solid state
(hydrogen atoms omitted for clarity).
Scheme 1. Conversion of a 1,2-diphosphinatocatechol into lithium
salts.
Having successfully prepared boron-centered anions we
sought to extend the syntheses to a phosphorus-centered
anion. Reaction of a slight excess (3.3 equiv) of H2-DPC with
PCl5 in toluene led to the evolution of HCl gas. Upon the end
of HCl evolution, the solvent was removed to yield an off-
white solid. 31P{1H} NMR spectroscopic analysis of the crude
material suggested that H[P(DPC)3] ([D6]DMSO: d = À81.3,
19.4 ppm) was produced, along with some amount of [P(H-
DPC)(DPC)2] (À35.6, 15.0). Analogous mixtures were gen-
erated during the synthesis of H[P(catecholate)3].[13] Washing
the solid with diethyl ether and drying for 24 hours under
vacuum yielded H[P(DPC)3] as a white solid in 97.1%.
Treatment of H[P(DPC)3] with LitOBu in THF (18 h) led to
a white precipitate which upon filtration and drying gave
Li[P(DPC)3] in 96.4% yield. This compound displays reso-
nances at d = 10.7 and À85.7 ppm in [D6]DMSO for the
phosphoryl and the central phosphorus atoms, respectively. In
[D6]acetone Li[P(catecholate)3] shows a single phosphorus
resonance at d = À82.0 ppm.[13]
boric acid, and LiOtBu in refluxing toluene leads to the
asymmetric salt Li[B(DPC)(oxalato)] in 86% yield as a white
solid after drying. Reaction of H2-DPC with Me3SiCl, and
then with LiBF4, allows isolation of Li[B(DPC)F2] in 54%
yield as an analytically pure white solid after filtration and
drying under vacuum. These three new materials are readily
identified by multinuclear NMR spectroscopy. The
31P{1H} NMR spectra ([D6]DMSO) for Li[B(DPC)2], Li[B-
(DPC)(oxalato)], and Li[B(DPC)F2] present resonances at
d = 16.4, 15.2, and 17.7 ppm, respectively. Notably, a mixture
of Li[B(DPC)2] and Li[B(DPC)(oxalato)] in [D6]DMSO
shows distinct 31P NMR signals indicating that any ligand
exchange or disproportionation processes are slow on the
NMR timescale. Sharp 11B NMR resonances are also
observed at d = 11.8, 11.8, and 11.5 ppm for Li[B(DPC)2],
Li[B(DPC)(oxalato)], and Li[B(DPC)F2], respectively. These
values are shifted slightly upfield compared to that reported
for LiBOB (d = 12.2 ppm).[4a]
The results of an X-ray diffraction study upon a single
crystal of Li[P(DPC)3] (grown by vapor diffusion of EtOH
into DMF solution of Li[P(DPC)3]) are shown in Figure 2.[11]
As expected, the geometry about the central phosphorus
X-ray-quality crystals of Li[B(DPC)(oxalato)] were
grown by vapor diffusion of hexanes into a THF solution of
4174
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 4173 –4176

Angewandte
Chemie
Organophosphorus-based flame-retardant compounds
are well known.[15] One proposed mechanism for their efficacy
is based on the fact that when many of these materials
undergo combustion, a layer of char (polyphosphoric acids) is
produced that acts as a thermal and gas barrier.[15,16] Further,
boron compounds can form a variety of boron oxide
enhanced chars, depending upon other neighboring species,
during thermal decomposition/pyrolysis.[17] The data for
pyrolysis combustion flow calorimetry (PCFC)[18] obtained
for three of the new salts, as well as corresponding data for
FRION 1 and LiBOB, are shown in Table 1. From the data in
Table 1: Microcombustion calorimetry results for FRIONs and LiBOB.[a]
Compound Avg. char Avg. HRR
Avg. HRR
Avg. total
yield[b]
[%]
Peak
T [8C]
peak value
heat release
[WgÀ1
]
[kJgÀ1
]
FRION 1
23.6(4)
158(1), 178(1), 49(1), 82(3),
300(0), 561(1) 4(2), 225(8)
20.5(5)
Li[B(DPC)2] 42.2(4)
312(2), 446(8) 200(17), 35(3) 12.2(3)
Figure 2. Molecular structure of Li[P(DPC)3] in the solid state (hydro-
gen atoms omitted for clarity).
Li[B(DPC) 49.60(5) 304(1), 327(1), 102(1), 111(1), 8.7(1)
(Oxalato)]
534(1)
239(2), 299(2) 20(1), 488(14) 10.4(1)
360(1), 492(1) 69(2), 9(0) 3.0(1)
42(1)
atom is octahedral. As found for Li[B(DPC)(oxalato)], the
lithium ion is tetrahedrally coordinated. Each lithium ion is
surrounded by four phosphoryl oxygen atoms from four
different [P(DPC)3] anions, and gives rise to a network-type
structure in the solid state. The Li-O distances span the range
from 1.906(7) to 1.917(7) ꢀ. The P-O distances for the
octahedral phosphorus atom range from 1.700(2) to
1.719(2) ꢀ. The structure of Li[P(3,5,-di-tert-butyl-catecho-
late)3] serves as a useful comparison and has for the central
phosphorus atom P-O distances of 1.690(3)–1.724(3 ꢀ.[14]
A desired feature of these anions is their thermal stability.
Each of the new FRIONs were thus examined by thermal
gravimetric analysis (TGA). The results are presented in
Figure 3. Each material is stable to up to 2008C, but the
Li[B(DPC)F2] shows additional stability up to 2808C.
Li[P(DPC)3] 40.2(7)
LiBOB
19(5)
[a] Standard deviation shown within parentheses. Samples were tested at
a 18C/sec heating rate under nitrogen from 75 to 8008C using method A
of ASTM D7309. [b] Average of three runs (full data in the Supporting
Information).
Table 1, we can infer that the lithium boron complexes as well
as the lithium boron phosphorus species are forming high
levels of char and showing reduced heat release. As higher
levels of char are formed, less of the electrolyte is available
for burning in the event of a battery fire, and further,
depending upon how the battery is penetrated/damaged in
a fire, the chars may help form seals/caps between layers to
slow fire growth in the battery. The LiBOB material shows the
lowest total heat release and some of the lowest heat release
rate (HRR) values, thus indicating it is the least flammable
material tested, yet it has a low char yield. Given that LiBOB
has very little carbon present in its structure, we hypothesize
that most of the carbon which is present decomposes in the
form of carbon dioxide, and would thus give off no additional
heat. Evolved gas analysis would be needed to confirm this
hypothesis, however. For the other compounds we can see
higher char yields and a range of total heat release. The higher
char yields are responsible for lower total heat release as
more of the structure is bound up as char, which is preferable
to being part of a substance that can be pyrolyzed and
combusted. These PCFC results show that these new ions
have reduced flammability and should have great potential as
flame-retardant ions in LIBs.
The utility of anions in LIBs depends on their ability to be
stable under a range of oxidation potentials. Studies have now
shown linear correlations between the experimental Eox and
Figure 3. Thermal gravimetric analysis of lithium salts.
Angew. Chem. Int. Ed. 2014, 53, 4173 –4176
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4175

.
Angewandte
Communications
Amine, J. Power Sources 2003, 119 – 121, 383 – 387; c) K. Xu,
M. S. Ding, S. Zhang, J. L. Allen, T. R. Jow, J. Electrochem. Soc.
2003, 150, A161 – A169; d) K. Xu, S. Zhang, J. L. Allen, T. R.
Jow, J. Electrochem. Soc. 2003, 150, A170 – A175; e) S.
Izquierdo-Gonzales, W. Li, B. L. Lucht, J. Power Sources 2004,
135, 291 – 296; f) Q. Wang, J. Sun, X. Yao, C. Chen, Electrochem.
Solid-State Lett. 2005, 8, A467 – A470; g) M. Otsuki, T. Ogino, K.
Amine, ECS Trans. 2006, 1, 13 – 19.
computed EHOMO of anions. Following the computational
procedures of Xue and co-workers,[19] we first minimized the
structure (DFT B3LYP 6-311 ++ G(2df,2p)) of a model anion
[B(DPC’)(oxalato)]À where in the DPC’ group the ethyl
groups of DPC were replaced by methyl groups for computa-
tional simplicity. This structure was then subjected to a single
point calculation (HF 6-311 ++ G(3df,3dp)), thus yielding
[8] A. R. Shaffer, N. Deligonul, D. A. Scherson, J. D. Protasiewicz,
Inorg. Chem. 2010, 49, 10756 – 10758.
[9] A. Xiao, L. Yang, B. L. Lucht, Electrochem. Solid-State Lett.
2007, 10, A241 – A244.
E
HOMO = À5.69. This orbital is comprised of the catechol core
p orbitals (see the Supporting Information). The presence of
the two O = P(OMe)2 units drives the energy of the HOMO
orbital down by 0.83 eV relative to that of [B(catechol)-
(oxalato)]À,[19b] and thus allows one to predict greater
oxidative stability (in the absence of kinetic effects) for
catechol borate anions bearing O = P(OR)2 residues. Electro-
chemical measurements are now underway to examine the
behavior and performance of the new FRIONs in LIBs.
[10] B. Dhawan, D. Redmore, J. Org. Chem. 1984, 49, 4018 – 4021.
[11] CCDC 969990 and 969991 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[12] a) F. Bockstahl, E. Graf, M. W. Hosseini, D. Suhr, A. De Cian, J.
Fischer, Tetrahedron Lett. 1997, 38, 7539 – 7542; b) A. Downard,
M. Nieuwenhuyzen, K. R. Seddon, J.-A. van den Berg, M. A.
Schmidt, J. F. S. Vaughan, U. Welz-Biermann, Cryst. Growth Des.
2002, 2, 111 – 119; c) P. Y. Zavalij, S. Yang, M. S. Whittingham,
Acta Crystallogr. Sect. B 2004, 60, 716 – 724; d) M. Albrecht, M.
Fiege, M. Baumert, M. de Groot, R. Froehlich, L. Russo, K.
Rissanen, Eur. J. Inorg. Chem. 2007, 609 – 616; e) J. L. Allen,
P. D. Boyle, W. A. Henderson, Acta Crystallogr. Sect. E 2011, 67,
m533; f) J. L. Allen, S.-D. Han, P. D. Boyle, W. A. Henderson, J.
Power Sources 2011, 196, 9737 – 9742.
Received: December 15, 2013
Revised: February 1, 2014
Published online: March 11, 2014
Keywords: flame retardant · lithium-ion battery · phosphorus ·
.
weakly coordinating anions
[13] M. Handa, M. Suzuki, J. Suzuki, H. Kanematsu, Y. Sasaki,
Electrochem. Solid-State Lett. 1999, 2, 60 – 62.
[1] G. Nagasubramanian, K. Fenton, Electrochim. Acta 2013, 101,
3 – 10.
[14] D. S. Yufit, Y. T. Struchkov, E. I. Matrosov, D. N. Lobanov, N. V.
Matrosova, M. I. Kabachnik, Zh. Strukt. Khim. 1988, 29, 131 –
137.
[2] D. Aurbach, Y. Talyosef, B. Markovsky, E. Markevich, E.
Zinigrad, L. Asraf, J. S. Gnanaraj, H.-J. Kim, Electrochim. Acta
2004, 50, 247 – 254.
[15] a) B. Schartel, Materials 2010, 3, 4710 – 4745; b) I. van der Veen,
J. de Boer, Chemosphere 2012, 88, 1119 – 1153; c) S. V. Levchik,
E. D. Weil, Adv. Fire Retard. Mater. 2008, 41 – 66.
[16] M. Lewin, E. D. Weil, Fire Retard. Mater. 2001, 31 – 68.
[17] K. K. Shen, S. Kochesfahani, F. Jouffret, Polym. Adv. Technol.
2008, 19, 469 – 474.
[18] R. E. Lyon, R. N. Walters, J. Anal. Appl. Pyrolysis 2004, 71, 27 –
46.
[19] a) Z.-M. Xue, W. Zhou, B.-B. Sun, C.-H. Chen, Electrochim.
Acta 2011, 56, 8770 – 8775; b) Z.-M. Xue, J. Ding, W. Zhou, C.-H.
Chen, Electrochim. Acta 2010, 55, 3838 – 3844; c) Z.-M. Xue, C.-
H. Chen, Electrochim. Acta 2004, 49, 5167 – 5175.
[3] K. Xu, Chem. Rev. 2004, 104, 4303 – 4417.
[4] a) W. Xu, C. A. Angell, Electrochem. Solid-State Lett. 2001, 4,
E1 – E4; b) K. Xu, S. S. Zhang, T. R. Jow, W. Xu, C. A. Angell,
Electrochem. Solid-State Lett. 2002, 5, A26 – A29.
[5] a) K. Xu, S. S. Zhang, U. Lee, J. L. Allen, T. R. Jow, J. Power
Sources 2005, 146, 79 – 85; b) J. C. Panitz, U. Wietelmann, M.
Wachtler, S. Strobele, M. Wohlfahrt-Mehrens, J. Power Sources
2006, 153, 396 – 401; c) L. Larush-Asraf, A. Biton, H. Teller, E.
Zinigrad, D. Aurbach, J. Power Sources 2007, 174, 400 – 407.
[6] H. F. Xiang, H. Y. Xu, Z. Z. Wang, C. H. Chen, J. Power Sources
2007, 173, 562 – 564.
[7] a) C. W. Lee, R. Venkatachalapathy, J. Prakash, Electrochem.
Solid-State Lett. 2000, 3, 63 – 65; b) Y. E. Hyung, D. R. Vissers, K.
4176
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 4173 –4176