History and Thermodynamics of Electrolysis - PowerPoint Presentation

1. History and Thermodynamics of ElectrolysisMogens B. Mogensen Technical University of Denmark, DTU Risø CampusDK-4000 RoskildeDenmarkmomo@dtu.dk2nd Joint European Summer School on Fuel Cell and Hydrogen Technology,Crete, September 25th, 2012Acknowledgements to colleagues at DTU Energy Conversion

2. OutlineDefinition of electrolysisRelation between electrolysis and fuel cellsTypes of electrolysersHistoryReversible thermodynamicsElectric potentialsIrreversible lossesProblem solving by students in between226 September 2012Add Presentation Title in Footer via ”Insert”; ”Header & Footer”

3. Definition of electrolysisElectrolysis is the use of direct electric current (DC) to drive an otherwise non-spontaneous electrochemical reactionThe words “Electrolysis” is constructed from Greek ἤλεκτρον [ɛ̌ːlektron] "amber" and λύσις [lýsis] "dissolution“Very reactive elements such as alkaline, alkaline earth metals, aluminium and fluorine are made by electrolysisElectrolysis or water into its elements, hydrogen and oxygen – or water splitting – is the kind of electrolysis that is known widestOne compound may be turned into another compound and and element by electrolysis, e.g. CO2  CO + ½O2326 September 2012

4. Electrolysis in this lecture is limited to electrolysis in the context of energy conversion and storage This lecture covers only electrolysis of water/steam and of CO2 or combined electrolysis of H2O + CO2 into H2 + CO Wish to do the electrolysis as efficient as possible and with lowest possible investment Energy storage often implies a need of conversion of H2 + CO back into electricity and H2O + CO2 Therefore, cells that can work as both electrolysis cells and fuel cells may be preferred dependent on costs The solid oxide cell is an example 426 September 2012

5. Principle of reversible Solid Oxide Cell0.8 V1.4 VWorking principle of a reversible Solid Oxide Cell (SOC). The cell can be operated as a SOFC (A) and as a SOEC (B).600 - 900 C EMF ca. 1.1 VFuel cellElectrolyser cell5

6. Ni/YSZ supportNi/YSZ electrodeYSZ electrolyteLSM-YSZ electrode10 mm Acc. voltage: 12 kV SE imageNi-YSZ supported SOC6

7. Electrolysis Cell Types Simple aqueous electrolytes (e.g. KOH or K2CO3), room temperature to ca. 100 C, 0.1 - 3 MPa pressureLow temperature “solid” proton conductor membrane (PEM), 70 – 90 C, and high temperature PEM 120 - 190 C.Immobilized aqueous K2CO3, Na2CO3 etc. in mesoporous structures – pressurized 200 – 300 C, 0.3 – 10 MPa Solid acids, 200 – 250 C, pressurized?Molten carbonate electrolytes, 800 – 950 C, 0.1 MPaHigh temperature solid oxide ion conductor (stabilized zirconia), 650 – 950 C, pressurized 0.1 – 1 (later 5) Mpa Fuel cells of the kinds exists, i.e. all the cells are in principle reversible. But, the same cell cannot in all cases practically work in both directions. Why not?

8. Some electrolysis cells and characteristics826 September 2012Type electrolyteAlkalineAcidPolymerSolid oxideElectrolyteNaOH or KOHH2SO4 or H3PO4 PolymerCeramicCharge carrierOH-H+ H+O2-ReactantH2OH2OH2OH2O and/or CO2ElectrodesNiGraphite with Pt + polymerGraphite with Pt + polymerNi + ceramicsTemperature80 -150 C140 - 180 C60 - 80 C700 - 900 C

9. Electrolyzer statusFew types commercialized but - from an energy conversion and storage point of view - none of them are commercial in today's energy markets.The classical alkaline electrolyser was commercialized during the first half of the 20th century. If significant amounts of synfuel via electrolysis in the very near future (the next 1 -5 years) – only alkaline electrolysers is available on some scale.

10. Early historyHydrogen was discovered by Henry Cavendish in 1766 and oxygen by Carl Wilhelm Scheele in 1771However, they did not know that water was a chemical compound of hydrogen and oxygen.The concept of chemical compounds and molecules was not yet establishedThe electrochemistry had to be invented first1026 September 2012

11. The Volta Pile = first electric battery1126 September 2012was constructed by Alessandro Volta in 1799 and published in 1800This stable electricity supply started the possibility of doing electrochemical experimentsRight after this invention a number of electrochemical discoveries took place

12. Volta’s pilePositive electrode material: silver, cupper or brassElectrolyte: paper of cloth soaked with a sulfuric acid or salt solutionNegative electrode: zinc or tin Thus , there was several options for materials combinations1226 September 2012

13. Early historyVolta’s pile (the battery) was actually know already during 1799 as the news about this amazing device spread fast in the scientific communityWilliam Nicholson (1753 – 1815) together with  Anthony Carlisle (1768 – 1840) discovered that water could be decomposed by electricity, i.e. electrolysis of water in May 1800. The same year of 1800 Johann Wilhelm Ritter (1776 – 1810) discover the same independently Sir Humphry Davy (1778 – 1829) proved in 1806 using electrolysis that the constituents of water is hydrogen and oxygen in the ratio of 2:1By 1838 chemistry of water was reasonably well understood, and the reversible cell could be invented and constructed1326 September 2012

14. Schönbein and Grove1426 September 2012During the year 1835 to 1845 these two friends worked with both electrolysis of water into H2 + O2, and H2 – O2 fuel cell forming water. The discovered that a cell with Pt electrodes in sulfuric acid worked reversible. They finally understood the electrochemical process. Grove invented the fuel cell device with cells in series in 1939. C.F Schönbein (left) was a Swiss professor, and W.R. Grove was a Welsh judge and a scientist

15. Schönbein’s U-tube experiments1838 (published 1839)1526 September 2012Taken from: Ulf Bossel, The Birth of the Fuel Cell, Published by The European Fuel Cell Forum, CH, 2000.

16. Grove’s experiment of 18421626 September 2012Questions:What do you think this Figure illustrates?The is a very basic error in the Figure. What is wrong?Pt wires inside glass tubes in containers with diluted sulfuric acid.W.R. Grove: On a Gaseous Voltaic Battery. Phil. Mag. III, 21 (1842) 417.Taken from: Ulf Bossel, The Birth of the Fuel Cell, Published by The European Fuel Cell Forum, CH, 2000.

17. History of industrial water electrolysis1726 September 2012YearEvent1800Discovery of electrolytic splitting of water1902More than 400 industrial electrolyzers in operation1939First large electrolysis plant with capacity 10,000 m3 H2 h-11948First pressurized electrolyzer by Zdansky/Lonza1966First solid polymer electrolyte system (General Electric)1972Development of solid oxide water electrolysis

18. An electrochemical cell is also called a galvanic cell or just a cell  The relation between the chemical energy, G (Gibbs free energy of reaction) of a cell reaction and the equilibrium (ideal) electrical voltage, also called the electromotive force, Emf, of the cell is given by  -G = n∙F∙Emf n is the number of electrons exchanged in the total reaction, and F is The Faraday constant = 96485 A s mol-1Reversible thermodynamics

19. Important: G and n must refer to the same reaction scheme! Example 1: H2 + O2-  H2O + 2e-½O2 + 2e-  O2- H2 + ½O2  H2O n = 2 and G0298 = - 237 kJ/mol H2Example 2: 2H2 + 2O2-  2H2O + 4e- O2 + 4e-  2O2- 2H2 + O2  2H2O n = 4 and G0298 = - 474 kJ/mol O2 Basics

20. At standard conditions (25 C and 1 atm):Emf = -G0/(nF) =- (- 237 kJ/mol)/(2*96485 As/mol) = - (- 474 kJ/mol)/(4*96485 As/mol) = 1.23 V G = G0 + RTlnK, K is the constant in the law of mass actionThis gives us the Nernst equation:Basics

21. The cell voltage may deviate from the theoretical Nernst voltage. Some possible reasons are 1. The cell is under external electrical load 2. The cell has an internal electronic leak 3. The concentration of reactants are different from the assumed values e.g. due to gas leakage 4. The actual cell temperature is different from the measured5. One (or both) electrodes are not in equilibrium with its reactants - bad electrocatalyst(s) 6. The cell has a thermal gradient, i.e. the Emf + a thermoelectric voltage is measuredBasics

22. Thermodynamics Energy (“volt”) = Energy (kJ/mol)/2F Etn = DH/2FEcell = Etni  Ecell - DG/2F Price  1/i [A/cm2] , DH/DG > 1 ,  = 100 % at E = Etn (no heat loss)22

23. Thermodynamics.23

24. Thermodynamics: CO2 and H2O750ºC – 900ºC24

25. 25ThermodynamicsH2O formation energiesSource: NIST chemistry webbook

26. Pressurized Electrolysis We get pressurized hydrogen with lower electricity input!

27. Question My question to you: Is the cell voltage of an electrolysis cell higher or lower than the cell voltage of a fuel cell in case that the gas composition at the electrodes is the same?

28. Electric potentials in galvanic cellsNaturally, electric potentials are needed in order to describe the behavior of electrochemical cellsHowever, this subject is quite complicated when we go into the details 2826 September 2012

29. Anode and cathode At open circuit voltage, a given electrode in a cell is not having a general name - you have to describe it by its constituents, e.g. "the Pt-H2 -", "the Ni-H2 -" or "the Ag-O2 - electrode" When the cell is working, reactions take place at the electrode and then:The anode is the electrode where the oxidation takes place, e.g. in fuel cell mode: H2 + O2-  H2O + 2 e-The cathode is the electrode where the reduction takes place, e.g. in fuel cell mode: O2 + 4 e-  2 O2- When the direction of current and the sign of the polarisation is changed the electrodes swap names!

30. Potential through a cell with no current – valid for all kinds of cell, yet simplifiedV4 – V1 = EmfEmf = -G/(n∙F) POSITION+-+-v1v4v2=v3v2v3tv4v1ßO--àElectrolyteH2+ O--DH2O + 2e-½O2+ 2e-DO--POTENTIAL, VOLTElectrolyte0

31. Potential through a cell with a current load in fuel cell mode -simplifiedCell voltage smaller than EmfPOSITION+-+-v1v4v2v3v2v30tv4v1ßO--ElectrolyteAnodeCathodeH2+ O--àH2O + 2e-½O2+ 2e-àO--POTENTIAL, VOLTßO--Electrolyte

32. Potential through a cell with a current load in electrolyser cell mode -simplifiedCell voltage larger than Emf

33. Potential concepts - energy and voltage The electrochemical potential of an electron is defined as where e- is the chemical potential of the electron, F is Faradays number, and  is the electrical potential inside the material in which the electron is.  is called the Galvani potential, the inner potential or the electrostatic potential. It is not possible to measure the absolute value of , but we can by measurements determine the difference in Galvani potential of two points (planes) in a material if the material and the current distribution is homogeneous. The electrochemical and the chemical potential are both specific energy quantities, J/mol, whereas the Galvani potential has the unit of voltage, V.

34. Potential concepts - energy and voltage (cont.) What we can measure with a voltmeter is the electromotive or Fermi potential, , defined as This means that the Fermi potential is dependent on both the concentration of electrons and the Galvani potential where is the standard state (or reference) concentration of the electrons Let us look at a picture of the general potential concepts and afterwards see what this may be used for in practise.

35. The electric potentials in more details

36. Questions Which kind of potentials are “ruling” where in the previous cell slides?(go back to slide 30 – 32)Are there any special differences between solid electrolytes and aqueous electrolytes?3626 September 2012

37. The electric potentials in more details (cont.)Sketch of profiles for the electromotive potential, , and the Galvani potential, , in a solid oxide fuel cell. The oxygen electrode is to the right. The absolute positions of the potentials are arbitrarily chosen.The electrolyte is Zr084Y0.16O1,92 (YSZ)

38. The electric potentials in more details (cont.)Sketch of profiles for the electromotive potential, , and the Galvani potential,  in a solid oxide electrolyser cell. The oxygen electrode is to the right. The absolute positions of the potentials are arbitrarily chosen.

39. Examples of a YSZ based cellThe following results are from: T. Jacobsen and M. Mogensen, ECS Transactions, 13 (26) (2008) 259-273Transport and concentration data for 8 mol% YSZ from J. H. Park and R.N. Blumenthal, J. Electrochem. Soc., 136 (1989) 2867 Further data used in calculations given below (not from experiments):Electrode thickness L = 200 μmTemperature T = 1000 COxygen pressure, right pO2 = 0.2 barOxygen pressure, left pO2 = 1.00·10−15 barSOFC current i = 1.00A cm−2SOEC current i = −1.00A cm−2Charge transfer resistances:H2 + O2−  H2O+2e− Rct,H = 0.05 ohm cm2½ O2 + 2e−  2O2− Rct,O = 0.1 ohmcm2Electron transfer Re,H = 0.01 ohm cm2Electron transfer Re,O = 0.01 ohm cm2

40. Electron defect concentration in YSZ0 V1000 C -1 V [h∙][e'] From Park & Blumenthal, JES 136 (1989) 2867

41. YSZpO2 = 1 atmPotential, V0 - NiLSMDistance, m l0l200pO2 = 10-15 atm-1-Potential course, OCV, 1000 C -  Thus, from this equation the local pO2 may be calculated

42. Potential course, SOFC modeElectromotive, , and Galvani, , potentials in a cell operating in SOFC mode compared to open circuit conditions at 1000 C.pO2 = O.2 bar

43. Electromotive, , and Galvani, , potentials in a cell operating in SOFC mode compared to open circuit conditions at 1000 C.Potential course, SOEC mode

44. Conclusion on potentialsSeveral types of concepts of potential existThe driving force for the electrons of very low concentration in a good solid electrolyte, is the Fermi (or electromotive) potential, which is mainly reflecting the concentrations of electrons, n, and holes, h.The driving force for ionic charge carriers - oxide vacancies, interstitial protons of high constant concentration, or ions in water - is the Galvani (also called electrostatic or inner) potential gradient, which is formed by the chemical driving forces of the electrode reactions.There are no free electrons inside an aqueous electrolyte – consequently no Fermi potential is defined inside the electrolyte, but the measured potentials of the electrode is still the Fermi potential.

45. QuestionsIs it possible to measure the absolute potential () of a metal (or other electron conducting) electrode?Is it possible to measure the absolute potential () inside an electrolyte? What can you derive from these concepts practically? Refs.:Jacobsen, T.; Mogensen, M; The course of oxygen partial pressure and electric potentials across an oxide electrolyte cell, ECS Transactions, (2008), 13(26), 259-274, DOI:10.1149/1.3050398 Mogensen, M.; Jacobsen, T., Electromotive potential distribution and electronic leak currents in working YSZ based SOCs, Solid Oxide Fuel Cells 11, ECS Transactions, (2009), 25(2), 1315-1320, DOI:10.1149/1.3205660 4526 September 2012

46. Irreversible losses, including electrode kinetics and efficiencyAs soon as a current is running then some of the driving change in free energy (“driving force”) will be converted into heat with a great risk of the energy being lost to the surroundings, and thereby decrease the efficiency of the electrolysis (and of any electrochemical conversion processHowever, the calculation of a real and generally valid efficiency is close to impossibleWe can calculate an efficiency for a given process in specific cases where all relevant parameter values are available4626 September 2012

47. Definition of two efficiencies The total efficiency, t is in the simplest form given as: where Esup is the total energy supplied and L is the total loss in energy The electrochemical efficiency, el, for an electrolyser is: where EMF is the electromotive force and U is the applied cell voltage 4726 September 2012

48. LossesWhich losses that are relevant depend on the circumstances, e.g.:Electrical efficiency in electrolysis onlyRound trip efficiency, i.e. electricity  H2 + O2 by electrolysis  electricity by fuel cellFirst a look at the thermodynamics again and then a look at the types of electrochemical losses that we call polarization4826 September 2012

49. Thermodynamics Energy (“volt”) = Energy (kJ/mol)/2F Etn = DH/2FEcell = Etni  Ecell - DG/2F Price  1/i [A/cm2] , DH/DG > 1 ,  = 100 % at E = Etn (no heat loss)49HHVLHV

50. Polarisation of a cellWhen current is drawn from a fuel cell the voltage decreases below the Emf, the bigger drop, the higher the current density.This voltage drop is often referred to as cell polarisation.The part of the polarisation that is due to the sluggishness of the electrode reaction is called activation overvoltage (or reaction overpotential) and the overvoltage due to changes in the reactant concentrations is called concentration overvoltage.The voltage drops due to the resistance of the electrolyte and due to non-ideal contacts are not overvoltages but just polarisations.Everything is just with opposite sign in case of electrolysis as compared with fuel cells.

51. Definitions:Voltage drop due to simple ohmic resistance is polarisation (not overvoltage)Overvoltage (= overpotential):  = Eeq - Ei = ( - )eq - ( - )i E is electrode potential, index eq means equilibrium (current density = 0), and index i is indicating "at a current density of size i"

52. Types of polarisation resistanceThe area specific resistance, ASR, may be broken down into five contributing area specific polarisation resistances:ASR = Relyt + Rconnect + Rp;elchem + Rp;diff + Rp;converThe ohmic polarisation, Uelyt, is due to the electrolyte resistance and follows Ohm's law: Uelyt = Relyt∙i [ cm2 ∙A cm-2 = V]

53. Contact resistance Contact resistance may often be equal to constriction resistance (e.g. current collectors in the cell test set-up) , because:Two bodies in contact (without pressure) will touch each other in 3 points.If the bodies are made of hard materials the contacts areas are almost only contact points.Thus the current path is constricted to go through these small contact areas. Literature: R. Holm, Electrical contacts, Theory and Applications, 4. edition, printed in 2000, Springer, Berlin

54. Small electrical contactsα is the radius of a circular contact is a parameter, and the distance along the y-axis is y = sqrt() is the specific electrical conductivityTotal resistance R = 1/(4α)5/6 of R @ 3.7 α From: R. Holm, Electrical contacts

55. Parameters important for constriction resistanceThe contact geometry, in particular roughnessContact load, i.e. mechanical pressureMaterials properties, conductivity, elasticity, ductility (creep and deformation strengths) – temperature is affecting these properties significantly – and current may affect the temperatureSurface layers (dirt, oxidation products, coatings)

56. Concluding remarks about contact resistancesThe matter is in the exact details very complicatedMany example are treated mathematically in Holms bookUsually you have to do own investigation in each single concrete case: electrical measurements, microscopy and modeling

57. Electrode reaction overvoltage or activation overvoltageActivation overvoltage is an unspecific term used when you do not know what you have at hand. There may be many different reasons for electrode reaction rate limitations at an electrode. e.g.:adsorption of reactant molecules at the electrodebond breaking in the reactant moleculesurface diffusion of reaction intermediates from the catalytic sites to the three phase boundary (TPB) linediffusion of ions through the bulk of electrode particles with mixed conductionconduction through or around segregated phases at the surface/at the TPBdesorption of reaction productstransfer of ions across the electrode/electrolyte interfacetransfer of electrons from electrode to molecule

58. The current density in low temperature electrochemistry is often well described by the Butler ­ Volmer equation:    Activation overvoltage = E – E0, the difference between the actual, E =  - , and the equilibrium, E0 (i = 0), electrode potential; a and c are anodic and cathodic symmetry factors, 0 <  < 1

59. At low overvoltage the Butler-Volmer equation becomes linearAt high overvoltage it gets the same form as the Tafel equation:   = a  b x logi  using the absolute value of the current density and the ± sign for anodic and cathodic overpotentials, respectively.Activation overvoltage

60. Activation overvoltageTo my best knowledge there is no experimental evidence that charge transfer as described simply by the Butler - Volmer equation is rate limiting SOC electrode reactions above 700 C!Further, the “bottle neck” theory - i.e. only one rate determining step is present at a given condition – is often taken for granted and is actually a prerequisite for the simple Tafel / Butler - Volmer analysis. This is very seldom seen in the case of SOCs and not common in electrochemistry.

61. Diffusion overvoltage If the diffusion of a reactant or product through a stagnant layer of gas or liquid electrolyte either outside an electrode or inside a thick porous electrode then a diffusion overvoltage appears. The size is given by the difference in the equilibrium potentials determined by the Nernst equation applied to: 1) the OCV situation, and 2) the chemical composition at the electrode surface in the situation with a current load high enough to change the composition at the electrode-electrolyte interface significantly.

62. Gas conversion overvoltageGas conversion overvoltage occurs when the gas concentration (partial pressure) cannot be maintained in the electrode compartment. When this is the case, a contribution to the change in electrode potential, E (=  - ) will appear as given by the difference in the equilibrium potentials determined by the Nernst equation applied to: the OCV situation the situation with a current load, i, high enough to change the gas composition in the electrode compartment. conv = Eocv - Ei

63. Any further questions? You are allowed to e-mail, but I do not promise to answer immediately Further reading especially wrt SOC: Mogensen, M., Hendriksen, P.V. "Testing of Electrodes, Cells and Short Stacks", Chapter 10 in High Temperature Solid Oxide Fuel Cells: Fundamentals, Design and Applications, Eds. S.C. Singhal and K. Kendall, pp. 261 - 290, Elsevier, 2003. The end6326 September 2012