1、Phil. Trans. R. Soc. A (2010) 368, 3227 3241doi:10.1098/rsta.2010.0112REVIEWKey challenges in future Li-battery researchBYJ.-M. TARASCON1,2,*1Laboratoire de Ractivit et Chimie des Solides, Universit dePicardie Jules Verne, CNRS (UMR-6007), 33 rue Saint-Leu,80039 Amiens, France2ALISTORE-ERI European
2、Research InstituteBatteries are a major technological challenge in this new century as they are a keymethod to make more efficient use of energy. Although today s Li-ion technology hasconquered the portable electronic markets and is still improving, it falls short of meetingthe demands dictated by t
3、he powering of both hybrid electric vehicles and electric vehiclesor by the storage of renewable energies (wind, solar). There is room for optimism aslong as we pursue paradigm shifts while keeping in mind the concept of materialssustainability. Some of these concepts, relying on new ways to prepare
4、 electrode materialsvia eco-efficient processes, on the use of organic rather than inorganic materials ornew chemistries will be discussed. Achieving these concepts will require the inputs ofmultiple disciplines.Keywords: cells, batteries; energy1. IntroductionEnergy is the lifeblood of modern socie
5、ty. Global warming, finite fossil-fuelsupplies and city pollution conspire to make the use of renewable energy, togetherwith electric transportation, a worldwide imperative. There is a pressing needto design electrical-energy-storage systems to balance supply with demands,as renewable sources are in
6、termittent, and to power the upcoming plug-inhybrid electric vehicles (PHEVs) or electric vehicles (EVs). Numerous energy-storage solutions enlisting mechanical, magnetic, chemical storage, etc., are beingpresently investigated. Therefore, as we want to store energy in order to restoreit as electric
7、ity, the most attractive path is to convert chemical energy intoelectrical energy, as they both share a common carrier, namely the electron.Fuel cells and batteries are electrochemical storage devices that ensure suchconversions to occur in a single or reversible way, respectively. As fuel cells are
8、frequently referred to as an emerging technology, in spite of 200 years of research,we have focused this paper on batteries since they are the main contendersto better manage the renewable resources of our planet, and to favour the*jean-marie.tarasconsc.u-picardie.frOne contribution of 13 to a Discu
9、ssion Meeting Issue Energy materials to combat climate change .This journal is2010 The Royal Society3227 on March 1, 2015http:/rsta.royalsocietypublishing.org/Downloaded from 3228J.-M. Tarascondeployment of electric cars so as to limit pollution. It should indeed be recalledthat each time we burn 1
10、litre of gasoline, we are liberating 1.5kg (e.g. 750l)of CO2.Amongpresentbatterytechnologies,Li-iontechnologyisthebestperforming one owing to its delivered energy density (210Whkg1; 650Whl1),whichexceedsanycompetingtechnologiesbyatleastafactorof2.5(Tarascon & Armand 2001). With such attractive perfo
11、rmances, coupled with itslong life cycle and rate capability, Li-ion technology has captured the portableelectronic market, invaded the power tool equipment market, previously keptfor Ni MH technology, and is on the verge of penetrating the EV market oncondition that improvements can be achieved in
12、terms of cost and safety. Indeed,long-term stability, high-energy density, safety and low cost seem to be the over-riding factors in high-volume applications. Therefore, it implies assessing presentLi-ion technology so as to define the challenges that lie ahead to ensure its long-lasting success (Ar
13、mand & Tarascon 2008). Attempts to answer such questionswill constitute the core of this paper, which will mainly, but not exclusively,deal with Li-ion technology. Research directions to increase the Li-ion batteryenergy density, lower its cost, improve its safety and make it more sustainableand gre
14、ener will be discussed. New upcoming chemistries will also be broughtto the scene and benchmarked with today s systems.(a) Means to increase energy densityDespite the new lease of life recently brought by the arrival of nanomaterials,present Li-ion technology falls short of meeting all of the requir
15、ements dictatedby the large volume of applications linked to renewable energy and electrictransportation fields. First, we should be aware that a colossal task is awaitingus if we really want to compete with gasoline, as an increase by a factor of 15is needed for the energy delivered by a battery (1
16、80Whkg1) to match the oneof a litre of gasoline (3000Whl1; taking into account corrections from Carnot sprinciple). Knowing that the energy density of batteries has only increased by afactor of five over the last two centuries, our chances to have a 10-fold increaseover the next few years are very s
17、lim, with the exception of unexpected researchbreakthroughs. Fortunately, the automotive industry has set a more realistictarget: doubling the present Li-ion energy density in the next 10 years so thatthe autonomy of EVs approaches 500km. The energy density of a battery isthe product of its capacity
18、 and its potential, and is mainly governed by thecapacity of the positive electrode. Simple calculations show that an increase incell energy density of 57 per cent can be achieved by doubling the capacity ofthe positive electrode, while one needs to increase the capacity of the negativeelectrode by
19、a factor of 10 to get an overall cell energy density increase of 47per cent (Tarascon 2002). So, the chances of drastically improving today s Li-ion cells energy density are mainly rooted in spotting better positive electrodematerials, i.e. materials that could either display greater redox potential
20、s (e.g.highly oxidizing) or larger capacity (materials capable of reversibly inserting morethan one electron per 3d metal).Knowing that the redox potential of an intercalation compound mainly tracksthe iono-covalency of the metal X bonding, fluorine (F) substitution appears tobe quite an attractive
21、direction to increase the material redox voltage (V), as FPhil. Trans. R. Soc. A (2010) on March 1, 2015http:/rsta.royalsocietypublishing.org/Downloaded from Review. Future Li-ion research trends322910 nmactive material:electronic insulator10 nmpores: space forliquid electrolyteelectronicconducting
22、phaseinterphase adhesionvia van der Waal forces,chemical bonding electron-conductingmoleculesactiveparticlesliquidelectrolyteconductivecarbonelectrolyte10 nmcurrent collectormixed conducting matrix conductive film onwalls of pores active porous skeleton withthin walls between poresLi-based electroly
23、tewithin the poresLi+Li+Li+Li+(a)(c)(b)(d)current collectorFigure 1. Schematic of a perfectly wired electrode (a) with variable options to be pursued givenfrom left to right as follows: (b) the molecular bridging, (c) the mixed ion conducting and (d)the reverse-structuring approaches, which each hav
24、e their merits and pitfalls as discussed in thetext (courtesy of G. Amatucci and M. Gaberscek).is very electronegative. Therefore, a drawback brought about by F substitutionconcerns materials having large band gaps (e.g. poor electronic conductivity),which usually display poor ionic conductivity whe
25、n containing Li. As the latter isan intrinsic value, there is a need to synthesize these materials at the nanoscale,as the intercalation time (e.g. the time for Li ions to move from the core tothe particle surface) decreases with the square root of the particle size. Thesenanoparticles must addition
26、ally be carbon coated if one ever wants these F-basedmaterials to be competitive in terms of power-rating capabilities. Apart fromfuelling electrons through the entire electrodes, carbon coatings, which standas a buffer layer between the active material and the electrolyte, also help inminimizing el
27、ectrolyte degradation; the latter is usually quite important, withhighly oxidizing electrode materials functioning at potentials well beyond thethermodynamic potential of the electrolyte (Simon & Tarascon 2009).Another approach towards increasing energy density consists of identifyingmaterials that
28、could reversibly accept more than one electron per 3d metal. Atfirst sight, Li2MnSiO4appears quite attractive, as the presence of Mn+2couldimply the feasibility of removing two lithium atoms (Mn2+Mn+4+ 2e). Sofar, experimental attempts (Gong et al. 2006) have remained unsuccessful inaccordance with
29、theoretical calculations (Gaberscek et al. 2007; Kokalj et al.2007), which indicate the non-thermodynamic stability of the fully delithiatedMnSiO4phase. This further confirms the difficulty in spotting compounds thatPhil. Trans. R. Soc. A (2010) on March 1, 2015http:/rsta.royalsocietypublishing.org/
30、Downloaded from 3230J.-M. Tarasconcould reversibly react with lithium via an insertion/de-insertion process involvingmore than one Li+per 3d metal. This contrasts with conversion reactions that canenlist, according to the overall reaction (MnXy+ .), the reversible uptake of two(CoO) (Poizot et al. 2
31、000) to six Li+(NiP2) (Gillot et al. 2005) per 3d transitionmetal at potentials that depend on the iconicity of the M X bond, with thehighest redox potential for the highest electronegative anion (e.g. F). However,the main drawback in the conversion of positive electrodes (FeF3) lies both in theirhi
32、gh polarization and in the absence of Li in their initial stage, so that they canonly be used in Li rather than Li-ion cells. To eschew the latter, there is always thefeasibility of making (MX LiX) composites via ball-milling, which can be usedas a Li reservoir and be stable in a moisture environmen
33、t if the conversion redoxpotential exceeds 3.2V (Amatucci & Pereira 2007) burning 1 litre. In contrast,chances of lowering the polarization for conversion electrodes remain slim sincewe have shown that this polarization is linked to the iconicity of the M X bondand its redox potential, both increasi
34、ng with compounds showing a high ioniccharacter (fluorides) and decreasing with highly covalent compounds (hydrides)(Oumellal et al. 2008). However, owing to the staggering capacity gain that suchelectrodes provide, they should continue to be a high-topic research, even thoughthe remaining task of i
35、ncreasing their energy efficiency is not perceived as aneasy one.Interestingly, most of the new paths mentioned to increase energy densityresult in the use of materials whose both electronic and ionic properties are verypoor; therefore, the need to grow highly divided materials in order to decreaset
36、he Li-ion transfer path for Li uptake/removal is a must, as well as the use ofconductive coatings for providing electrons throughout the entire electrode (Aricoet al. 2005). Carbon coatings combined with particle-size reduction have beensuccessfully used to trigger the electroactivity of LiFePO4(Rav
37、et et al. 1999).Although quite useful, such an approach cannot be used for highly oxidizingcompounds such as LiCoO2and LiFePO4F, for instance, because of the reducingnature of the carbon treatment, which results in decomposition of the treatedmaterial; hence, the need to find alternatives to improve
38、 electrode kinetics(Gaberscek et al. 2005; Gaberscek & Jamnik 2006). With all the limitationsthat such a comparison could bear, an electrode material can be visualized asour brain, where all the neurones (in our case active nanoparticles) must beelectronically and ionically nicely wired to ensure pr
39、oper functioning. Mindfulof these considerations, several approaches, most of which consist of controllingelectrode architecture at the nanolevel, have been pursued (figure 1). They enlistthe design of electrodes based on molecular bridging via the coating of particleswith electron-conducting polyme
40、rs such as polyethylene di-oxyethylene thiphene(PEDOT; Meng et al. 2003; Shen & Lee 2005), or the grafting of a molecularconductor redox relay (e.g. 3.5V versus Li/Li+) such as triaryl amine substitutedfor a benzyl phosphonate group, as recently demonstrated by Greatzel s groupon LiFePO4(Wang et al.
41、 2007). Parallel approaches consisting of (i) preparinghighly mixed solid conductor composites (e.g. CuF2 MoO3), as demonstratedby Amatucci s group (Badway et al. 2007) and (ii) growing ordered mesoporousmaterials (Bruce et al. 2008), consisting of micrometre-sized particles, withinwhich pores of 2
42、5nm do exist to host the electrolyte, have been successfullymade. For such concepts to fully reach fruition, there are some remainingchallenges. While the approach referred to as either inverse structuring, replica orPhil. Trans. R. Soc. A (2010) on March 1, 2015http:/rsta.royalsocietypublishing.org
43、/Downloaded from 电 极 结 构Review. Future Li-ion research trends3231chemical templating, has been successful in preparing numerous binary phases,it has been solely limited to LiMn2O4for Li-bearing positive electrodes (Bruceet al. 2008). Hence the need to grow ordered mesoporous Li-bearing positiveelect
44、rodes if one wants such a new class of poorly conducting electrodes (silicates,phosphates and fluorophosphates) to be efficiently implemented in Li-ion cells.Additionally, further studies must focus on the molecular-bridging approach, withthe need to (i) design, on demand of molecular layer, relays
45、having a redoxpotential closely matching that of the insertion material, (ii) search for electron-conducting polymers beyond PEDOT, together with better chemical meansto ensure a uniform coating, and (iii) identify mixed ion/electron-conductingpolymers. Although this last task may appear tremendous,
46、 the pay-off couldbe huge, as mastering electrode wiring is keen in the optimum functioning ofelectrochemical storage devices, regardless of what they are (supercapacitors,batteries, fuel cells, etc.).(b) Means of increasing safetyManipulating energy inevitably leads to intrinsic safety risks that h
47、oldregardless of the devices or electrochemical systems used, with the rule ofthumb being that the risks increase with the size of the energy-storing device;hence the requirement for extra safety conditions for large-sized batteries in thecase of PHEV and EV applications. Numerous attractive solutio
48、ns relying oneither the use of chemical additives to the battery electrolyte (solid electrolyteinterface (SEI) modifiers, shut-down and redox shuttles additives, ionic liquids)or improved cell design and electronics have been pursued, as shown in figure 2;therefore, Li-ion technology is not exempt f
49、rom incidents. More conservativeapproaches (Ariyoshi & Ohzuku 2007) consist of coupling a highly oxidizingpositive electrode material (LiMn2O4or others) with a less reducing material(Li4Ti5O12instead of carbon) in order to eliminate the formation of a solidelectrolyte interface at the negative elect
50、rode, which is a source of concern whendealing with safety. Therefore, this approach provides extra safety at the expenseof the battery performance, as LiMn2O4/Li4Ti5O12Li-ion cells display a lowenergy density of 85Whkg1when compared with 140Whkg1for LiMn2O4/Ccells (Amezutsumi et al. 2007). Overall,
