1. Reliable Electrochemical Energy Storage for Alternative Energy Craig B. Arnold Department of Mechanical and Aerospace Engineering Princeton Institute for Science and Technology of Materials Princeton University 2500  m

6. Case Study: Wind Power P. Denholm, G. L. Kulcinski, and T. Holloway, "Emissions and energy efficiency assessment of baseload wind energy systems," Environmental Science and Technology , vol. 39, pp. 1903-1911, 2005. Fluctuations occur over many different time periods

8. Ragone plot in Hz

9. Charging versus Discharging We all like to think about batteries like cups of coffee. We simply fill it up until it is filled and empty it out until it is empty But in reality, batteries are not that simple  the rate matters

10. Charging is different than discharging

17. Mechanical Degradation Much work has been done looking at mechanical degradation to electrodes after electrochemically cycling  cracking, delamination, etc. However, these types of studies typically ignore effects to the rest of the system Physical constraints can lead to elevated stress states throughout the battery Stress source Electrode Stress State Separator Stress state Electrode Expansion Tensile Compressive SEI Formation Compressive/Neutral Compressive External Compression Compressive Compressive External Bending Tensile/compressive Tensile/compressive

18. Separator Compression Regardless of most internal or external sources, the separator will experience compressive stress But how large is the stress ? Also notice that during rest step, the stress begins to decline J. Cannarella and CBA, in preparation Li + pouch cell (LiCoO 2 cathode) Cycle electrochemically while directly measuring the stress of the system Max stress is 1.3 MPa Average pinch stress for adult is ~1 MPa Mathiowetz et al., Arch Phys Med Rehabil , 66, p.69 (1985)

20. SEM Images of Compressed Battery Unstressed Stressed Electrodes are unchanged, most deformation occurs in separator 30 MPa for 3 hours This is expected as the polymer separator is significantly weaker than the electrode material

23. Creep Strain C. Peabody and CBA, in preparation As applied stress is increased the total permanent strain continues to increases

25. Mechanism of Relaxation 0 MPa 10 MPa 30 MPa Viscoelastic creep causes pore closure in the polymer separator A 30MPa compressive stress for 3 hours decreases pore volume by almost 60% C. Peabody and C.B. Arnold, J. Power Sources 196, 8147 (2011) Pore volume decrease appears linear with respect to permanent strain

26. EIS Characterization of Separators Pore closure impedes ion transport through the separator C. Peabody and C.B. Arnold, J. Power Sources 196, 8147 (2011) Permanent strains as small as 10% can significantly reduce the capacity by a factor 2-3

27. Other Separators Higher porosity corresponds to larger creep strains at same stress state Similar effects occur for other separator materials Composition Porosity (%) Avg Pore Diam ( μ m) Thickness ( μ m) 2320 PP/PE/PP 39 0.027 20 2340 PP/PE/PP 45 0.035 40 3501 PP 55 0.064 25

29. Electrochemical Performance The electrochemical effects of stress induced pore closure in the separator manifest as capacity fade C. Peabody and C.B. Arnold, J. Power Sources 196, 8147 (2011) Cycle separator in t-cell using commercial cathode and Li metal anode Charge/Discharge current constant 0.5 mA

30. Capacity Fade As a function of pore volume, we see a linear decrease in capacity The discharge capacity decreases with permanent strain As permanent strain reaches ~15%, the capacity of the cell has decreased by a factor of 2

31. Relevance Are these values relevant or way out of the realm of a real system ? In a typical Li + ion battery the electrodes may expand up to 10% Since the electrodes are typically 2-3 times thicker than the separator, such expansion if totally accommodated by the separator, could lead to strain values as high as 40-50 % They are very relevant Consider: Mechanical stress on the separator can be an important factor in battery design e.g. if T electrodes = 80, T sep = 20  T electrodes = 8 Then  sep = 40% !

34. History stuff

36. But in the 1600’s, scientists did not really know much about electricity or how to use it. The spark generators were mostly used by scientists to study the nature of the sparks In 1745, scientists (Musschenbroek and Cunaeus) noticed that one could “charge” up a glass filled with water and get a shock by touching a metal nail Shortly thereafter, this was simplified to just metal foil wrapped around the inside and outside of a jar with a chain connecting the inner layer.  Leyden Jar We know these devices as capacitors, but they work by storing charge ELECTROSTATICALLY

37. Although they still didn’t know all that much about electricity, they now had methods of storing and generating electricity, but it was still a research tool (and a parlor trick) In fact, this enabled many important experiments of the time 1746: Nollet assembled a line of 200 monks each holding the end of a wire to test if electricity can travel faster than human communication. Without warning he connected a Leyden Jar to the ends … But Cavendish did not publish all that much and these discoveries were rediscovered years later by Faraday, Ohm, Coulomb, Maxwell 1747-1753 Cavendish used Leyden Jars to discover many of the fundamental physics laws of electricity  Inverse square law for force, electric potential, capacitance, resistance

38. 1752: Ben Franklin and his famous kite experiment Showed that lightening is the same as electricity stored in Leyden Jar Franklin’s other main contributions to the field include the concept of current as the flow of positive charges, and the term battery We later found out he was very wrong, but unfortunately it was too late. This is why current goes in the opposite direction of electron flow Two disadvantages of the Leyden Jar are that it doesn’t store charge all that long (This is true in general for electrostatic storage) and it doesn’t store all that much energy

39. 1786: Galvani’s famous experiments on frog legs He took two dissimilar metals (Zn, Cu) and touched them to the ends of a dead frog’s leg Surprisingly, the leg moved and Galvani attributed this to bioelectricity

40. But Volta did not believe that the electricity came from the frog. He believed the electricity came from the metals In 1799, he showed that by combining different metals that are separated by a salt or acidic solution it was possible to generate electricity  VOLTA PILE First commercially available battery An enabling techno-logy for the telegraph Side note: Galvani died one year earlier and never knowing the answer

42. Separate reactions into half-cells Salt bridge: allows ions to move between cells Oxidation occurs at anode Material gives up electrons i.e. Zn  Zn +2 + 2 e - Reduction occurs at cathode Material takes in electrons i.e. Cu +2 + 2e -  Cu Voltaic or Galvanic Cells

43. M  M n+ + ne - Oxidation Reduction M n+ + e -  M (n-1)+ Oxidation is loss of electrons Reduction gains electrons Electrochemistry: chemical reactions require charge transfer This occurs through redox reactions Oxidation occurs at the anode  anodic reaction Reduction occurs at the cathode  cathodic reaction Both need to happen… electron is generated at the anode and must be consumed at the cathode so net charge is conserved in overall process Ions dissolve into solution Ions deposited from solution These are called Half-Cell reactions We need a salt bridge to complete the circuit so that the charge remains balanced  otherwise, the charge would build up and the battery would stop working

44. In the process of oxidation and reduction, energy is converted from chemical into electrical i.e. Electrons are free to run through the circuit and do work Voltage of the cell is determined by the oxidized and reduced species and related to the change in free energy But they must go through the external wires Can think of this as analogous to water flowing downhill where the voltage is the height of the hill Zn can lower its energy by giving up electrons and dissolving into solution Cu + can lower its energy by capturing electron and ‘plating’ out on electrode In previous example, Zn  Zn +2 + 2 e - Cu +2 + 2e -  Cu +0.763 V +0.337 V 1.10 V

45. The important thing here is that every material has a slightly different potential

46. Just because a reaction is energetically favorable, that doesn’t mean we know how fast it will occur. The rate depends on kinetics It’s got to depend on: Temperature Voltage Concentration or pH Let’s consider a single electrode of Zn We know that oxidation and reduction can occur according to Zn  Zn 2+ + 2 e - Furthermore, we know that if this is in equilibrium, then the forward and back reactions must be occurring at the same rate Lets define a current density i = I/A

47. Then, i f = i b = i o Where i o is called the exchange current density , sometimes denoted i e If we apply a voltage to this electrode, we will shift the reaction one direction or the other i f – i b = i  0 This condition can be described by the Butler-Volmer equation Where  a and  c are called the anodic and cathodic transfer coefficents  s is called the surface overpotential or polarization  s = V-V 0 * This expression for the B-V equation assumes a one electron process. Typically it is safe to use this equation since most reactions involving multiple electrons also involve multiple intermediary steps which involve a single electron and for current, we only have to worry about the slowest reaction. The details of this is way beyond the scope of this introductory course. When we don’t know, we usually assume  a and  c ~ 0.5 The exchange current tells us how fast a reaction can occur There are other prefactors that I have omitted

48. We notice the similarity of this equation to that for diffusion Also, note the sign convention that the forward current direction denotes an anodic reaction i.e. i > 0 oxidation is occurring (corrosion) i < 0, reduction is occurring (deposition)  a =  c = 0.5  a >  c

49. Overpotential is a term that tells us how far from equilibrium we are Activation Overpotential Tafel (1905) observed:  =a+b log i If we do some math on the B-V equation, we can see where this comes from Consider larger overpotentials so that one term dominates the B-V equation (lets choose anodic term) It’s a driving force for transport (just like chemical potential) So we can interpret the B-V equation as telling us the amount of current we get from a reaction given a certain electrode potential

50. So, we can plot this in a different way to find  , i o , and V 0 Rearranging, we find or If we extrapolate the linear regime for both anodic and cathodic, the intersection will be at the equilibrium voltage and exchange current density Tafel Plot Voltage (V) Ln | i | (mA/cm 2 ) V 0 I o anodic cathodic RT/  a F RT/  c F

51. Other Overpotentials In reality, the Tafel equations are only valid in a small regime of potential At potentials close to the equilibrium potential, the slope decreases as reverse reactions become important At potentials much higher, the current becomes limited as the transport properties in the electrolyte limit the reaction  Concentration overpotential The limiting current depends on the concentration in solution and could be increased significantly by mixing electrolyte

52. Fig 17.5W from Callister So depending on the overpotential, we may need to take concentration into account Also, don’t forget about resistive losses in the system as well

53. Electrochemical Energy Storage Batteries are a compact method of converting chemical energy into electrical energy Other methods, fuel cell, photovoltaic cell, electrochemical capacitors etc. All work the same, but the details are different C-rate  charging/discharging rate, 1C is current needed to discharge in 1 hour Anode (Oxidation): Zn + 2 OH -  Zn(OH) 2 + 2e - E = 1.25 V Ag 2 O + H 2 O + 2e -  2 Ag + 2 OH - E = 0.34 V Cathode (Reduction): e - e - e - e - e - e - Anode Cathode Electrolyte/Separator Current Collectors Primary : Non-rechargeable Secondary : rechargeable Voltage  Potential difference between anode and cathode. Related to energy of reactions Capacity  amount of charge stored (usually given per unit mass or volume)

54. So this raises a potential problem with the Volta Pile If Zn is the anode and H is the cathode, there is an evolution of hydrogen gas that could passivate the electrode or generate a large resistance to flow Solution is to use a wet cell that has the proper ions around (1836) Daniel Cell: Cu/Zn in sulfate solutions (1859) Plante: Pb/PbSO 4 cell Wet cells have an obvious disadvantage for transport, but they do work well

55. But actually, a wet cell battery may have existed well before Volta Archeologists found this clay pot in the Baghdad area in the 1930’s Carbon dating places it ~250 BC How does this work?

56. Leclanche in 1866 developed another kind of wet cell that had a better shelf life and was less reactive with the environment This battery was constructed with Zn as the anode and MnO 2 + C as the cathode The cathode was mixed into a paste and placed in a porous pot The Zn anode was immersed in a chloride electrolyte Very popular with the new Telegraph !!

58. Modern Batteries Modern Button Cell

59. Battery Types Battery chemistry has not changed much since 1800’s Leclanché Alkaline Silver-Oxide Rechargeable Alkaline 1866, 1888 1949 1950 1978 Zn-MnO 2 Zn-MnO 2 Zn-Ag 2 O Zn-MnO 2 Lead-acid 1859 Pb-PbSO 4 Ni- based NiCad Metal Hydride 1899 1990 Cd-NiOOH MH-NiOOH Li- based Lithium-Iodine Lithium ion Plastic/Polymer 1968 1991 1995 Li-I 2 Li-LiCoO 2 Li-LiCoO 2

60. How does a Li-ion battery work? Li ions intercalate into the crystal structure of the electrode materials But Li is very reactive (high V) which causes side reactions and passivation layers (solid electrolyte interphase)  Critical to the proper functioning of these cells More advanced topic

62. Impedance Matching So faced with all these choices, how do we choose an appropriate power source? The optimal power transfer to the load will occur when the impedance (resistance) is equal to the internal resistance Can derive this from Ohm’s laws However, we would not typically run a battery at this current as it would heat up too much

63. Ragone relation Energy density  Energy per unit area/volume Specific Energy  Energy per unit mass Other issues such as total weight, size, voltage, environmental concerns can limit our selection process In general, electrochemical systems always show the characteristically downward curved plots High power  Low energy High energy  Low power One might think that it is best to just push the limits of energy or power http://www.powerstream.comz/ragone.gif

64. Rate effects Another major factor in optimizing energy storage is the rate at which the energy is needed As with all batteries, there is a decrease in capacity (efficiency) as the power is increased But many application for batteries do not require extended periods of high power draw Energy demands differ from application to application Main limiting factor is rate of ion movement across electrolyte Duracell ‘D’ cell

65. Although standard specs can be reported, there are a variety of important issues that ultimately affect the lifetime and performance of a battery system for a desired application Current Drain: Different batteries respond differently to current In general, as current is increased, the available voltage and capacity decrease Peukert’s equation: I n x t = C I is current in A t is time in hr C is rated capacity Modified Peukert’s law But since C depends on rate, one must correct for this Where H is the hours for the rated capacity Handbook of Batteries 3e, Eds Linden and Reddy Rate effects

66. Different chemistries will have different voltages and different characteristic discharge curves  Appropriate choice will depend on application limitations For instance, Silver Oxide batteries have a very flat discharge at 1.5 volts compared to Li-ion batteries with a sloping curve around 3 V Cut-off voltage of the device is an important parameter as it will determine the actual capacity that can be used Handbook of Batteries 3e, Eds Linden and Reddy Discharge Characteristics

67. Batteries operated in pulsed applications will last longer than constant discharge at same current During rest time, battery recovers voltage  more of the theoretical capacity can be used Factors such as maximum current and duty cycle will have a profound effect on this issue Pulsing between high and low current (e.g. transmit/receive operations) will have a similar effect  voltage will oscillate Voltage response to pulse will vary with chemistry Alkaline Zn-C High current pulsed can lead to catastrophic failure Handbook of Batteries 3e, Eds Linden and Reddy Pulsed Discharges

69. An ultracapacitor has properties of both battery and capacitor It has a high power density and can be cycled like a capacitor But it also has a significant energy density like a battery Different Flavors: carbon, transition metal oxides Ionic conduction in liquid or solid-state electrolyte Electronically conductive electrodes Electrochemical capacitors Supercapacitors pseudocapacitors First ultracapacitor was patented in 1957 using porous carbon electrodes Require high surface area electrodes to achieve large capacitance Supercapacitors Double Layer: Charge stored in double layer at interface between electrolyte and electrode Pseudocapacitance: Faradaic charge storage at electrode surface ultracapacitors Capacitors Batteries Electrode Electrode Electrolyte