Electrochemical Impedance Spectroscopy

Transcription

1 Electrochemical Impedance Spectroscopy May 2012 Designing the Solution for Electrochemistry Potentiostat/Galvanostat І Battery Cycler І Fuel Cell Test Station І І І

2 Nomenclature : EIS Electrochemical? In electrochemistry, everything of interest takes place at the interface between electrode & electrolyte! Controlling REDOX by Potentiostat/galvanostat Impedance? AC circuit theory describes the response of a circuit to an alternating current or voltage as a function of frequency Impedance is a totally complex resistance encountered when a current flows through a circuit made of resistors, capacitors, or inductors, or any combination of these Ohm s Law, V = R I V = I (complex number ) Spectroscopy? No Quantum Process Small Perturbation Response

3 Excitations used in E chem Techniques 1. DC Potentiostatic CA, CC, CP 2. Sweep LSV, Tafel, PD, LPR CV Cyclic Polarization 3. Pulse DPV NPV RNPV SWV 4. Sine PEIS, GEIS

4 Electrochemical Interface and Electrochemical Process

5 Electrochemical Interface Everything happens at the interface Charge Transfer R ct R ct ~ 1/ i 0 Butler Volmer Equation Diffusion Layer W Bulk Electrolyte R ser, R Ω Double Layer C dl Non Faradaic Process

6 Randles Circuit

7 Process of Energy Storage in Electrochemical System Common Steps Ionic charge conduction through electrolyte in pores of active layer Electronic charge conduction through conductive part of active layer Electrochemical reaction on the interface of active material particles including electron transfer Diffusion of ions or neutral species into or out of electrochemical reaction zone. P (SOC,SOH,T) Ex) R s, R ct, C dl, W Study of Mechanism Evaluation & Diagnosis CO poisoning Water flooding in FC Performance Simulation Arbitrary Load DC/AC/Transient Power/Energy Discharge Process Polarization Curve CV EIS (SOC,SOH,T;ω)

8 Impedance Spectra of a Li ion battery Impedance Spectra upon cycling Nyquist Plot vs. level of discharge CF)Discharge curve upon cycling Effect of temperature

9 PEMFC A. PEMFC Under Dead End ( H 2 O outlet in cathode closed) B. PEMFC Under CO Poisoning (H ppm CO as fuel gas) Ref. ahner elektrik Kronach Impedance Day 2012

10 Circuit Elements (1)

11 Basic Circuit Elements Resistor Inductor Capacitor E RI Idt C C Q E 1 dt di E L R C j C j 1 1 L j t j e I t I 0 ) ( t j e I t I 0 ) ( t j e I t I 0 ) ( I E I E I E

12 AC Current, Voltage, and Impedance " ' ) sin (cos /, ) ( ) ( ) ( ) ( Impedance ) cos( ) ( Current 2 1 & where, ) cos( ) ( Voltage ) ( ) ( j j I E where e I E e I t I I f j e E t E E o o o o j o t j o o t j o o Modulus & Phase (Bode Plot) Real & Imaginary part (Nyquist Plot) Ohm s Law

13 Presentation of Impedance Spectrum Nyquist Plot Vectors of length Individual charge transfer processes are resolvable. Frequency is not shown. Small can be hidden by large. Bode Plot C may be determined graphically. Small s in presence of large s are usually easy to identify. Imaginary, ʺ 0 log Real, ʹ log

14 Basic Circuit Elements Resistor R Inductor jl Capacitor 1 jc 1 j C

15 Combinations of Elements Serial Combination Parallel Combination

16 Combinations of Circuit Elements R C R 1 jc R C 1 1 R jc R R C R S 1 R 1 jc

17 R s R C " ' j C R RC R j C R R Rs C j R R S min max max max max phase, " " ' 4. 2 " ) 2 ( ' 1 ", 1 ' 3., 2. 0, 1. RC R R C RC R R R R R R R C RC R R C R R R R R s S S s s

18 R s R C Phase ϕ ω max = 2f max = 1 / RC ω R ω 0 R s R S +R logω

19 Coating Capacitance Ideal Coating Imperfect Coating C coat A d

20 Uniqueness of Models There is not a unique equivalent circuit that describes a spectrum. Measuring is simple and easy, but analyzing it is difficult. Physically relevant model is important. It can be tested by altering physical parameters. Be cautious in handling empirical models even if you get a good looking fit. Use the fewest elements Test it by T test Same Impedance Spectrum

21 Disadvantages of EIS Ambiguities in interpretation All cells have intrinsically distributed properties Ideal circuit elements may be inadequate to describe real electrical response Use of distributed elements (e.g. CPE) There is not a unique equivalent circuit describes measured impedance spectrum

22 Advantages of EIS Relatively simple electrical measurement But analysis of complex material variables: mass transport, rates of chemical reactions, corrosion. Predictable aspects of the performance of chemical sensors and fuel cells Providing empirical quality control procedure Theory Physically Relevant Model Mathematical Model theory(ω) Electrochemical System EIS Experiment e(ω) Curve Fitting System Characterization Empirical Circuit em(ω) Modified figure shown in page 10, IS(2 nd ed.)

23 Circuit Elements and Electrochemical Meanings

24 Physical Electrochemistry & Equivalent Circuit Elements Electrolyte Resistance 3 electrode: between WE and RE 2 electrode: all series R in the cell are measured incld. R of contacts, electrodes, solution, and battery separators Depends on ionic concentration, type of ions, temperature, and geometry

25 Physical Electrochemistry & Equivalent Circuit Elements Charge Transfer Resistance Echem charge transfer reactions are generally modeled as resistances. When an EIS spectrum is measured on a corrosion cell at E corr, the resistance at low frequency is identical to the polarization resistance. For a one step, multi electron process, O + ne R small overpotential is given by RT nf C (0, t) C O * O CR (0, t) C * R i i o R ct E i RT nfi 0 C O (0, t), C R (0, t)

26 Physical Electrochemistry & Equivalent Circuit Elements Double Layer Capacitance A electrical double layer forms as ions from the solution stick on the electrode. There is an Å wide separation between charge in the electrode and ionic charges in the solution. Charges separated by an insulator form a capacity. On a bare metal, estimate 20 to 40 μfof C for every cm 2 of electrode area. Depends on electrode potential, temperature, ionic concentrations, types of ions, oxide layers, electrode roughness, impurity adsorption, etc From A. J. Bard & L. R. Faulkner, Electrochemical Methods

27 Physical Electrochemistry & Equivalent Circuit Elements Constant Phase Element (CPE) The CPE is basically an imperfect capacitor. It s phase shift is less than 90. CPE Unlike C, a CPE has 2 parameters α is generally between 0.9 and 1.0 A is similar to C Possible Explanations 1 A( j) Surface roughness Fractal Dimension, D=1+1/α Distribution of reaction rates on a surface Varying thickness or composition of a coating

28 Physical Electrochemistry & Equivalent Circuit Elements Diffusion Diffusion processes can create an impedance, which is small at high frequency and increases as frequency decreases. Warburg Impedance Warburg looks like a special CPE with A=1/s and α=1/2. However, remember that Warburg is derived from electrochemical kinetics. Parameters you obtain with Warburg have physical meanings. It is only partly true for CPE. You can get a good fit, but how to interpret the resulting W (1 parameters? j) e j 4 ( j) 1/ 2 For a one step, multi electron process RT / 2 * 1/ 2 * n F A D C D C 2 O O R R

29 Physical Electrochemistry & Equivalent Circuit Elements Diffusion Nernstian & Finite Diffusion Impedance (1 j)tanh( j D ) Homogeneous reaction (Gerischer) Spherical Diffusion (1 j)coth( j D ) 1 A B j 1 1 A B j

30 Physical Electrochemistry & Equivalent Circuit Elements Diffusion Transmission Line Model Warburg ( 1 j) W Element Nernstian Impedance: Diffusion by Constant Concentration (1 Finite Diffusion Impedance: Diffusion by Phase Boundary (1 j)tanh( j D ) O Element j)coth( j D ) T Element

31 Nernstian Impedance R: 100Ω C: 0.001F

32 Nernstian Impedance R: 100Ω C: 0.001F

33 Finite Diffusion Impedance R: 100Ω C: 0.001F

34 Finite Diffusion Impedance R: 100Ω C: 0.001F

35 Validation of Impedance Data Kramers Kronig Relation

36 Validation of Impedance Data Ideal impedance data must fulfill: Causality: The output must be exclusively a result of the input Linearity: The response must be a linear fn. of the perturbation Stability: The system must not be changing during measurement a serious problem for corroding systems Finite Valued: Impedance must be finite value at any frequency Kramers Kronig Relation: Validation Test Low Frequency Extrapolation The integration range includes the frequencies zero and infinity Note pure capacitor cannot be calculated a. " ' 2 '( ) '( ) b. ' " 2 "( ) 0 0 x"( x) "( ) dx 2 2 x '( x) '( ) dx 2 2 x

37 Electrochemistry: A Linear System? Circuit theory is simplified when the system is linear. in a linear system is independent of excitation amplitude. The response of a linear system is always at the excitation frequency (no harmonics are generated). Look at a small enough region of a current versus voltage curve and it becomes linear. If the excitation is too big, harmonics are generated and EIS modeling does not work.

38 E chem: A Stationary System? Measuring EIS spectrum takes time (often many hours). The sample can change during the time the spectrum is recorded. If this happens, modeling results may be wildly inaccurate. To shorten the measuring time of impedance spectrum, use FFT EIS method. Non Stationary Conditions result in non stationary spectra!

39 K K Relation Ref. Scribner Associates Inc. ViewK K Transform Tutorial

40 Validation of Impedance Data HIT

41 Limitation of K K Relation The integration range includes the frequencies zero and infinity and Phase are measured independently with different accuracy and sensitivity, but in theory, they are correlated with each other.

42 Heating NTC, PTC LOG Impedance / 4,0 3,9 3,8 3,7 3,6 3,5 KTY 10k (NTC) Phase / LOG Impedance / 3,15 3,10 3,05 Pt 1000 (PTC) Phase / 3, LOG frequency / Hz LOG frequency / Hz 0 Phase is more stable than! Ref. ahner elektrik Kronach Impedance Day 2012

43 HIT Approximation ln ( ) 0 const S ( ) d ln d( O dln ) Local relationship between impedance and phase => Not affected by the limited bandwidth problem => Reliable detection of artifacts and instationarities (drift) => Reconstruction (!!) of causal spectra => Reliable interpretation of spectra Ref. Kronach Impedance Day 2012

44 Ex. Spectrum of a fuel cell under load / m phase / o HIT / m phase / o FIT m 100m K 10K 100K frequency / Hz m 100m K 10K 100K frequency / Hz Ref. Kronach Impedance Day 2012

45 Ex. Fuel cell under CO poisoning Raw data Refined data 0,1 0.1 Hz DATA FIT 0,1 0.1 Hz DATA FIT imaginary part / 0,0-0,1 10K 2 Hz 3 Hz imaginary part / 0,0-0,1 10 KHz 2 Hz 3 Hz -0,2-0,2 0,0 0,1 0,2 0,3 0,4 0,5 real part / 0,0 0,1 0,2 0,3 0,4 0,5 real part / Ref. ahner elektrik Kronach Impedance Day 2012

46 Other Methods to Measure EIS

47 Multi Sine Wave Method Time Single Sine Method In Galvanostatic EIS, Freq V : Linear Region? Pseudo Galvanostatic EIS Multi Sine Method

48 imag. part / Ohm im ag. part / Ohm imag. part / Ohm impedance Time Resolved EIS drift affected data interpolated data time course of the impedance at single frequencies 1. Real Time Drift Compensation time -10m 2. Time Course Interpolation frequency 0-10m 3. HIT Refinement 10m 10m 50m real part / Ohm 0 10m 10m 50m real part / Ohm An EIS recorded at the PEMFC with flooded cathode, after 4450s of dead end operation at 2A. After pp , IS (2nd ed.) & Schiller -10m 0 10m 10m 50m real part / Ohm

49 Artifacts

50 Inductive Loop at High Frequency The effects of inductances are often seen at the high frequencies The value of inductor is very small, however, this can be important if the electrode impedance is low. Possible Causes Actual physical inductance of loop or coil of wire between electrode and potentiostat Self inductance of the electrode itself: even a straight piece of rod has some self inductance ~ several nh Some cylinder type batteries also shows this effect ~ uh Instrumental artifacts, notably capacitance associated with the current measuring resistor, however. potentiostat manufacturers may have already made corrections for this effect di E L dt ʺ L jl

51 Mutual Induction Twisting Cables A B C D Arbitrary CE/RE & WE/WS CE/WS & WE/RE CE/WE & RE/WS Cell: 20 mω planar resistor Ref. ahner elektrik Kronach Impedance Day 2012

52 Others

53 Galvanostatic EIS is Better for Low Potentiostatic Mode Vacis 1 mv Minimum! 1 mv rms = A rms X min = 707 uω These are Absolute Minimum Values! Limitation is APPLIED E Measured E is still Accurate! Galvanostatic Mode Can Measure Smaller E Values! ~Microvolts CMR of electrometer may limit the absolute minimum Values! > 5 uω Refer to Shorted Lead Test

54 How to Extract Model Parameters Building equivalent circuit model Physically relevant model Each component is postulated to come from a physical process in the EChem cell based on knowledge of the cell s physical characteristics. Empirical model Complex Nonlinear Least Square (CNLS) Fitting Algorithm is used to find the model parameters that cause the best agreement between a model s impedance spectrum and a measured spectrum. starts with initial estimates of model parameters. Iterations continue until the goodness of fit exceeds an acceptance criterion, or until the number of iterations reaches a limit. Please check the change of χ 2 after each iteration. Sometimes, CNLS algorithm cannot converge on a useful fit because of An incorrect model Poor estimates for the initial values Noise and etc. Don t care if the fit looks poor over a small section of the spectrum.