"" Art is I; science is we." i i i " ‐‐ Claude Bernard (1813 – 1878), French medic and physiologist
IBA 2013, Barcelona, Spain; March 11 ‐15, 2013
AN ELECTROLYTE MENU IN SEVERAL COURSES R. Schmitz, R. Schmitz, R. Wagner, M. Amereller, J. Kasnanscheew, E. Krämer, T. Schedlbauer, H.J. Gores, S. Passerini, and M. Winter
MEET Battery Research Center, Institute of Physical Chemistry, Univ. of Muenster, GER, , , E‐mail: martin.winter@uni‐muenster.de
Specials of Today El t l t th “E bl ” f B tt i Electrolytes, the “Enabler” of Batteries Non‐Aqueous Liquid Electrolyte Economics MTFMP as Co‐Solvent with Propylene Carbonate The SEI as Indicator for Li Plating on Graphite Al Current Collector “Corrosion”: It is Different than Reported The Zeppelin or the Gasoline (ICE) Car?
1) Volta‐Pile (1791): Zn/Cu; NaClaq as Electrolyte: Actually a Metal Air Battery Actually a Metal‐Air Battery Volta-Cell (open; to O2 from air) O2 reacts with Cu forming CuO at the surface. Salt water (slightly acidic) medium @ 1.1 V: (Anode) (A d ) (Cathode)
Zn Zn Z Z 2+ + 2 eCuO + 2H+ + 2e- Cu + H2O Cu react with O2, regen. CuO
Closed @ 0.76V (Anode) (Cathode)
Zn Zn2+ + 2e2 H+ + 2e- H2 on inert Cu
Failure mechanism of the Volta-Pile: Drying out, because of H2O evaporation Technology progress: Pile Electrolyte reservoirs or “crown of cups”
Electrolytes, The “Enabler” of Batteries
2) Non‐Aqueous Liquid Electrolyte Economics • Standard non‐aqueous electrolytes are blends of linear and cyclic carbonates q y y
High permittivity cyclic carbonate
Low viscosity linear carbonates
• Blends are formulated to achieve specific physical & chemical electrolyte properties. • Typical range of Li salts, such as LiPF6 is 0.8 to 1.2 molar (10 ‐ 15% by weight) • Electrolyte contributes ca. 6 to 10 % of the overall lithium ion battery material costs. • With a mass fraction of <15%, the salt costs are up to 90% of the electrolyte costs depending on the salt concentration, purity, and supplier and use of electrolyte additives. • Bulk solvents add only 5 to 15% of the total cost. Additives (0.1 – 5%) increase costs. • Moderate Moderate changes in solvent and additive chemistry are tolerable in view of costs. changes in solvent and additive chemistry are tolerable in view of costs Lower salt content and lower purity requirements can reduce costs.
3) New Salts and Solvents Developed 3) New Salts and Solvents Developed in Muenster* Salts Salts
Difluoro(oxalato)borate
Li
O3S
C F2
F2 C
O
F2 C
CF3
Lithium 1,1,2,2‐tetrafluoro‐2‐ (perfluoroethoxy)ethanesulfonate
Lithium‐1,1,2,2,‐tetrafluoro‐ 2‐methoxy‐ethanesulfonate
Additives Additives
Solvents
Adiponitiile 1‐Fluoropropane‐2‐one
N PF5 N (1,3‐dimethyl‐4,5‐dihydro‐1H‐ imidazol‐3‐ium‐2‐yl) pentafluorophosphate(V)
1,3,2‐Dioxathiolane 2,2‐dioxide
Methyl 2,3,3,3‐tetrafluoro‐ 2‐methoxypropionate yp p
3‐Fluorodihydrofuran‐2(3H)‐one
*Partially Partially in cooperation with Gert‐Volker Röschenthaler group from Jacobs‐University Bremen in cooperation with Gert‐Volker Röschenthaler group from Jacobs‐University Bremen
Methyl 2,3,3,3‐tetrafluoro‐2‐methoxy propionate: y , , , yp p 2‐Step Synthesis of MTFMP
Hexa fluoro propene (HFP)
Perfluoro propylene oxide
R = Me MTFMP
Graphite is Compatible with a PC:MTFMP (9:1) Electrolyte at a High Coulombic Efficiency 2.0
•
•
•
NMC/graphite full cells with NMC / counter and Li reference electrodes MTFMP‐decomposition at >1.0 V vs. Li/Li+ MTFMP prevents graphite exfoliation Similar 1. cycle efficiencies compared to EC:DEC (3:7), when low surface area graphite is used: ca. 85% efficiency Distinctive graphite intercalation stages indicate reasonable kinetics
1.8
+ Pote ential vs (Li/L Li ) / V
•
MAGD-graphite 1M LiPF6 in PC:MTFMP (9:1 wt)
1.6 1.4
1M LiPF6 in EC:DEC (3:7 wt)
1.2 1.0 0.8 0.6 0.4 0.2 0.0 0
50
100
150
200
250
300
350
400
-1
Capacity / mAh g
•
MTFMP enables the use of PC in Li MTFMP enables the use of PC in Li‐ion ion batteries
Schmitz, R. et al.; Journal of Power Sources, 2012, 205, 408‐413
450
Graphite Shows Competitive Discharge Rates and Cycling Stability in PC:MTFMP (9:1) • NMC/graphite full cells /g p
• Good cycling stability at 1 C y g y
• Capacity of graphite is given.
• 1% capacity loss after 300 cycles.
• High discharge capacity even at 5 C (discharge rate = 5D)
• High average efficiency of >99.9 %
MTFMP forms a stable, low resistive SEI at the graphite anode when used in PC MTFMP forms a stable low resistive SEI at the graphite anode when used in PC 400
500 -1
1 M LiPF6 in EC:DEC (3:7)
300
1 M LiPF6 in PC:MTFMP (9:1) 250
200
D/5
100
400 80
300
200
100 D/3
D/2
D
2D
Discharge rate
3D
5D
Discharge capacity Efficiency
60
Efficiency / % E
350
Discharge capacity / m mAh g
Deintercala ation capacity / mAh g
-1
120
40
20
0
50
100
150
200
250
300
Cycle no.
Schmitz, R.; Schmitz, R.W..; Müller, R.; Kazakova, O.; Kallinovich, N.; Röschenthaler, G-V.; Winter, M.; Passerini, S.; Lex-Balducci, A., Journal of Power Sources, 2012, 205, 408-413
4) How the SEI Tells Us, that Li Metal Deposition ) p Takes Place on Graphite* Graphite is the widely used anode material because of a beneficial combination of high discharge/charge efficiency, long cycle life and low lithium intercalation potential, at relatively low costs. The low intercalation potential of Li into graphite may result in Li metal deposition. deposition In general, (localized) deposition of Li on graphite may take place during cycling of graphite, in particular, during formation, at lower temperatures and when the cells are aged. Deposited Li is local and finely distributed and is often not easy to detect. How can the SEI help to detect the deposited Li? How can the SEI help to detect the deposited Li? *Schmitz, R. et al.; Journal of Power Sources, 2013, 10.1016/j.jpowsour.2013.01.105 Schmitz, R.; et al. Journal of Power Sources, 2012, 217, 98‐101.
In situ Raman on Copper Electrodes In situ Raman on Copper Electrodes 3.5
•
Constant Current with ‐0.1 mA cm‐2
•
>0 V vs. Li/Li+: no changes in Raman spectrum
Poten ntial / V vs Li/Li
• <0 V vs. Li/Li+: formation of Li2C2; Lithium‐ Carbide: Increase of signal at 1844 cm‐1
+
3.0 2.5 2.0 1.5 1.0 0.5
Li2C2
0.0
Formation of Li2C2 starts with the Li metal deposition process the Li metal deposition process.
1 M LiPF 1 M LiPF6 in EC:DEC (3:7 by in EC:DEC (3:7 by wt.) Li: CE ,RE; Cu as WE 0.1 mA cm‐2 current density
0.0
0.2
0.4
0.6
0.8
1.0
time / h
-0.1 V
Raman inttensity arb.unit
Similar results, when Li is deposited on a graphite electrode (see next).
-0.5
0.0 V 0.5 V 1.0 V 1.5 V OCP
500
1000
1500
Wavenumber / cm
-1
2000
Li2C2 as Indicator for Li‐Deposition on Graphite: Raman Mapping of Graphite at <0V vs. Li/Li+
• IInhomogeneous distribution of Li h di ib i f Li2C2 • Li2C2 forms only below <0V vs. Li/Li+, when Li metal deposition takes place
Li2C2 35 Rel. integral 0.0 500
30
0.1
y-Position / µm m
• Ex situ measurement Ex situ measurement • 400 spectra taken in a designated area Integration of corresponding signals • Red: High concentration; Red: High concentration; Blue: Low concentration
25
0.2 28000 0.3
20
0.5 55500
0.4
0.6 0.7 83000 0.8
15
09 0.9
10
110500 1 111200
5
1844 cm‐1
0 0
5
10
15
20
25
30
35
x-Position / µm
Simpler detection of Li2C2 with mass spectrometry, after hydrolysis of (partially) overcharged graphite after hydrolysis of (partially) overcharged graphite electrode: acytelene (H2C2) gas formation.
SEI formation on Li in the Presence of IL: SEI formation on Li in the Presence of IL: Li2C2 • Li electroplated on Cu • Investigation of the distribution of SEI components by Raman mapping SEI components by Raman mapping • Almost homogeneous distribution of Li2C2 • Lower amounts at the position of the scratch y-position \ µm
Li2C2 is part of the SEI formed by 0.3 M LiTFSI in PYR14TFSI
50
0 0.1 0.2 0.3 0.4
40
0.5 0.6 0.7 0.8 0.9 1.0
30 20 10 0
10
20
30
40
x-position / µm
50
0 120 240 360 480 600 720 840 960 108 120
SEI formation on Li in the Presence of IL: SEI formation on Li in the Presence of IL: LiCN • Li deposition on Cu • Investigation of the LiCN distribution by Raman mapping Raman mapping • Almost homogeneous distribution of p LiCN on deposited lithium • Lower amounts at the position of the scratch
The origin of LiCN Th i i f LiCN is the PYR i th PYR14 cation ti
0 4800 0.1
y-position \ µm
LiCN is part of the SEI formed by 0.3 M LiTFSI in PYR14TFSI
50 9600 0.2 14400 0.3
40
19200 0.4 24000 0.5 28800 0.6
30
33600 0.7 38400 0.8 43200 0.9
20
47600 1.0
10 0
10
20
30
40
x-position / µm
50
5) Current Collectors: Requirements for LIB* Excellent electronic conductivity: Ag, Cu, Au, Al,… g, , X , XAg, Cu, Au, Al
Low cost:
Electrochemically stable within the electrode operation potentials: He
Metals that alloy with Li Fe Co
Ni
B
C
N
O
F
Ne
Al
Si
P
S
Cl
Ar
Cu Zn Ga Ge As Se
Br
Kr
I
Xe
At
Rn
Ru Rh Pd Ag Cd
In
Sn Sb Te
Os
Tl
Pb
Ir
Pt
Au Hg
Bi
Po
Al alloys with Li at carbon anode potentials Al alloys with Li at carbon anode potentials Cu is oxidized at >3.6V vs. Li/Li+ (= cathode potentials); surface impurities even at lower potentials tl t ti l Cu anode, Al (!) cathode (LiPF6 is necessary!)
Processing to thin foils (in the 10‐20 µm range) possible Rel. light weight Chemically and thermally stable/inert
*Considerations are valid for lithium ion cells with carbon anode and 4‐V cathode materials!
Al Current Collector: General Definition of Corrosion
“Corrosion is the destruction of materials by a chemical or electrochemical reaction with their environment.” (DIN 50900)
no applied external current or potential “electrolytic electrolytic dissolution dissolution“
Al: Current Collector Corrosion or Anodic Oxidation and Dissolution; Anodic Oxidation and Dissolution; Depending on the Electrolyte Salt LiClO4, LiCF LiCF3SO3, LiN(SO LiN(SO2CF3)2*, LiTFSI, etc. LiTFSI etc No significant chemical reaction with Al No effective Al passivation by a chemical reaction! a chemical reaction! Al dissolves in the electrolyte by anodic oxidation (= oxidation by an external force) via a complex Al‐TFSI‐anion via a complex Al TFSI anion formation formation*
LiPF6,, LiBF4,, etc. Chemically and thermally unstable. In the presence of H+ (H2O impurities after assembly, solvent oxidation after charge) y, g ) the salts tend to form HF. HF corrodes the current collector forming AlFx**and/or AlzOyFx films. This passivates Al against anodic oxidation.
1M Salt in EC:DMC (1:1) Dahbi et al., J. Power Sources, 2011, 196, 9743 * Wang et. al.: Electrochimica Acta, 2000, 45, 2677 **Kanamura et al., J. Power Sources, 1995, 57, 119
Al: Current Collector Corrosion or Anodic Oxidation and Dissolution: LiPF6 vs. LiTFSI (LiN(SO2CF3)2)* LiPF6
< 4.3 V (Corrosion & Passivation)
< 4.3 V (No Corrosion) LiTFSI
More HF due to protons created by electrolyte solvent oxidation
> 4.3 V (Further Reaction & More Passivation)
> 4.3 V (Anodic dissolution & no passivation) *with organic carbonate based solvents
Al: Current Collector Corrosion or Anodic Oxidation and Dissolution: Anodic Oxidation and Dissolution: LiPF6 vs. LiTFSI (LiN(SO2CF3)2)* LiPF6
< 4.3 V (Corrosion & Passivation)
< 3.5 V (No Corrosion) LiTFSI
More HF due to protons created by electrolyte solvent oxidation
> 4.3 V (Further Reaction & more Passivation)
> 3.5 V (Anodic Dissolution & no Effective Passivation) *with organic carbonate based solvents
Al: Current Collector Corrosion or Anodic Oxidation and Dissolution Depending on the Electrolyte Solvent* •Keep the electrolyte salt fix: 1M LiTFSI
cyclic
linear
• Vary the electrolyte solvents y y • Various org. carbonate mixtures • Lactone • Nitrile (adiponitrile) Nitrile (adiponitrile) • Examine the stability vs. anodic dissolution.
Different functional groups *E. Kraemer et al.; ECS Electrochemistry Letters 1, 2012 , C9‐C11.
Anodic Dissolution Behavior of Different Electrolyte Solvents with 1M LiTFSI
Experiment: 1 mV/s from OCP to 5 V vs. Li/Li+; then a 24 h holding step: Accumulated electric charge: Influence of the electrolyte solvents on anodic oxidn of Al Accumulated electric charge: Influence of the electrolyte solvents on anodic oxidn. of Al. LiTFSI/GBL shows the largest, LiTFSI/ADN the smallest charge values. Different mixtures of carbonate show different oxidation behavior.
((Quasi) Online Al‐Ion Detection ) in the Electrolyte
Flow cell developed in house: Every hour an electrolyte sample is taken out from the electrochemical cell and the Al‐ion amount is detected by ICP‐OES*. The electrolyte in the cell is replaced by new electrolyte from the reservoir The electrolyte in the cell is replaced by new electrolyte from the reservoir. Electrochemical experiment: 1 mV/s from OCV to 5 V; 1 mV/s from OCV to 5 V; then a 24h holding step. Electrolyte is taken out every hour. Electrolyte is taken out every hour. *ICP‐OES: Inductively Coupled Plasma – Optical Emission Spectroscopy p p py E. Krämer et al., Journal of The Electrochemical Society 160, 2013, A356‐A360.
Quasi‐On‐line Detection of Al: Electrochemical Oxidation Reactions and Al Dissolution Do Not Occur in Parallel* 0.50
225
• 1M LiTFSI in EC:DEC (3:7)
150
• Complex data. Grey: Electrochemical data. Black: Al3+ concentration.
0.45 0.40 0.35 0.30 0 25 0.25
75
0.20 Analytical Al Detection
0.15 0.10
Al-Con ncentratio on / ppm
Curren nt density / mA cm
--2
0.55
Electrochemical experiment: 1 mV/s from OCV to 5 V; then a 24h holding step
Electrochemical Oxidation
0
6
12
18
24
0
Time / h
• Steep increase in current density for the electrochemical process from the beginning. • Little Al3+ in the electrolyte for the first 3 ‐ 5 hours, but increasing current density. What is happening in the “First First Hours Hours”:: Electrolyte decomposition: Both solvents and TFSI. Electrolyte decomposition: Both solvents and TFSI • What is happening in the • In this time period, countermeasures, such as protective film formation on Al, should apply to prevent Al oxidation.
Role of the Al Surface Oxide Layer? y Preparation of Boehmite (AlOOH)‐Coated Al 2‐Step Preparation Process : A) 2 min pre‐treatment of the Al surface with diluted NaOH and B) 60 min treatment with water at 100°C
(Partial) removal of the natural y Aluminum oxide layer
Growth of a thick aluminum oxide/hydroxide / y (boehmite) layer
After treatment the surface layer thickness increases to ca. 400 nm.
Anodic Dissolution of Al Before and After Boehmite Treatment
14
Electrochemical experiment: 1 mV/s from OCV to 5 V; then a 24h holding step at constant potential Little accumulated electric charge with the boehmite surface layer
Electrric charg ge / C
1M LiTFSI in EC:DEC (3:7)
Aluminium foil Boehmite foil
12 10 8 6 4 2 0 0
5
10
15
Time / h
20
25
A Model for Al Dissolution, Surface Layer Removal and Electrolyte Decomposition Removal, and Electrolyte Decomposition
Al Al dissolution is preceded by electrolyte decomposition and Al di l ti i d db l t l t d iti d Al2O3 surface layer removal. f l l Protons (H+) created during solvent oxidation seem to be one of the instrumental species. This depends on the type of electrolyte solvent. The TFSI‐ anion also decomposes. When the protective Al When the protective Al2O3 layer is removed, Al dissolution takes place layer is removed Al dissolution takes place and pits are formed in the Al foil. (Similar) models may be valid to other electrolyte salt/solvent combinations.
6) The Future of the Electric Vehicle: ) Gasoline Car or Zeppelin?
or
Zeppelin: pp High Hopes, but… Zeppelin pp was pioneered by the German Count Ferdinand von Zeppelin. His plans p y pp p from 1874 and 1983 were patented in the United States on 14 March 1899. ZZeppelins were operated by the Deutsche Luftschiffahrts‐AG (DELAG), the first li t d b th D t h L ft hiff h t AG (DELAG) th fi t commercial airline. In World War I, the German military made extensive use of Zeppelins as bombers and scouts. Civilian zeppelins became popular in the 1920s again. Their heyday was during the 1930s when the airships LZ 127 Graf Zeppelin and LZ 129 Hindenburg operated regular transatlantic flights from Germany to North America and Brazil. The Hindenburg disaster in 1937, along with political and economic issues, ended the era of the zeppelin. Today, there only a few zeppelins in operation.
The Invention of the Automobile with Gas Engine: 127 years ago ith G E i 127 1886: 1886: 1888:
Carl Benz filed a patent on a gas engine powered vehicle (“tricyle”). M Max. speed 18 km/h d 18 k /h Gottlieb Daimler (GER) constructed an engine‐powered coach. First long distance (>100 km) trip was made by Carl Benz’ wife Bertha Carl Benz himself Bertha. Carl Benz himself was not allowed to drive.
From the beginning: Sceptiscm about the “ “Coach without horse”. h ih h ” Carl Benz was warning to use a car beyond a speed of 50 l i b d d f km/h (ca. 31 mph), because high speed would cause irreversible brain damage. 1901: “The 1901: The world‐wide demand for cars will not exceed one million – world wide demand for cars will not exceed one million because there will be not enough chauffeurs available. “, G. Daimler (GER) (from wikipedia) Attention! Gottlieb Daimler (1834 – 1900) (from wikipedia)
th Century: “Motorphobia” in the late 19 p y Scepticism, Criticism and Hostility
Medical doctors were afraid that traveling with high speed will lead to brain g g p damage and suffocation. Economics scientists feared for the mass unemployment of coachmans, saddler and coachmakers The automobile will be always too expensive for a mass and coachmakers. The automobile will be always too expensive for a mass market. g, g Architects were fearing, that the vibration of the engine will destabilize buildings along the road. g Farmers were afraid that the loud noise of the engine will kill their animals. To cover the bad smell of the exhaust gas, it was suggested to blend gasoline with perfume. Very few believed that gasoline powered car will become a success story. Obviously society was not ready for ICE powered car at that time Obviously, society was not ready for ICE powered car at that time.
Acknowledgments Federal Ministry of Economics and Technology (BMWi) Federal Ministry for the Environment, Nature Conservation & Nuclear Safety (BMU) Federal Ministry of Education and Research (BMBF) North‐Rhine‐Westphalia (NRW) University of Muenster (WWU)
" An expert is a person who has made all the mistakes that can be made in a very narrow field. An expert is a person who has made all the mistakes that can be made in a very narrow field " ‐‐ Niels Henrik David Bohr (1885 ‐ 1962); Danish footballer, physicist, and philosopher.