Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Application of electrochemical methods in corrosion and battery research
(USC Thesis Other)
Application of electrochemical methods in corrosion and battery research
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. ProQuest information and Learning 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA 800-521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPLICATION OF ELECTROCHEMICAL METHODS IN CORROSION AND BATTERY RESEARCH Copyright 2001 by Zhaoli Sun A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Materials Science) December 2001 Zhaoli Sun Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number 3065856 U M I* U M I Microform 3065856 Copyright 2002 by ProQuest Information and Learning Company. A ll rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor. M l 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA T he G raduate School U niversity Park LOS ANGELES, CALIFORNIA 90089-1695 This dissertation , w ritten b y Zfiaok Sun U nder th e direction o f h S s.. D issertation C om m ittee, an d approved b y a ll its m em bers, has been p resen ted to an d accepted b y The G raduate School in p a rtia l fu lfillm en t o f requirem ents fo r th e degree o f DOCTOR OF PHILOSOPHY Dean o f Graduate Studies D ate December 17. 2001_____ DISSERTATION COMMITTEE ___________________________m t i . Chalrperstm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGMENTS I would like to express my heartiest appreciation to my research advisor Prof! Florian B. Mansfeld. for his valuable guidance, encouragement and generous support through all these years o f my Ph.D study, without which this work would not have been possible. I would also like to thank Dr. Mark E. Thompson, Dr. Edward Goo, Dr. Steven R, Nutt and Dr. Bruce E. Koel for their academic guidance and for being on my qualifying and dissertation committees. Many thanks go to Jack Worrall and John Curulli for their help in surface analysis at CEMM A USC. I am very grateful to Advanced Mechanical Technology, Inc. (AMTT) for funding the three and half-year project o f corrosion protection o f absorption heat pumps, which is part I o f this thesis. Special thanks go to Mr. Chuck Hannon. AMTI for providing me with useful references and suggestions. Dr. Tracy Piao and Mrs. Susan Phillips, Quallion, LLC are thanked for their help and collaboration during the battery project which is part II in this thesis. I am also very grateful to the former and present CEEL members, especially to Dr. Caiyun Chen, Dr. Chu-Cheng Lee. Dr. Gang Zhang and Mr. Stan Hsu for their continuous collaboration and help in many aspects over the last five years. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. iii Finally. I would like to thank my whole family back in China for their love, sacrifices, understanding and support without which this thesis would be still in the air. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IV TABLE OF CONTENTS ACKNOWLEDGMENTS______________________________________________ ii LIST OF TABLES_____________________________________________________x LIST OF FIGURES___________________________________________________xii ABSTRACT — ___......._.........__..._ _ ....__........_.......____..._____ ..._____xviii PART I: DEVELOPMENT OF CORROSION PROTECTION METHODS FOR AMMONIA/WATER ABSORPTION SYSTEMS____________ I 1. INTRODUCTION___________________________________________________2 2. LITERATURE REVIEW____________________________________________ 8 2 .1 Corrosion o f carbon steel in A-W absorption system s...........................................8 2.2 Corrosion inhibition o f carbon steel........................................................................10 2.2.1 Traditional inhibitor - chromate........................................................................1 1 2.2.2 Chromate-ffee inhibitors.....................................................................................12 2.3 Corrosion protection by REM Ss..............................................................................15 2.3 . 1 REMSs as inhibitors........................................................................................ 16 2.3.2 Surface modification........................................................................................... 17 3. EXPERIMENTAL TECHNIQUES___________________________________ 21 3.1 Electrochemical techniques...................................................................................... 21 3.1.1 Polarization techniques...................................................................................... 22 3 .1. 1.1 Potentiodynamic technique.......................................................................22 3 . 1.1.2 Potentiostatic technique.............................................................................. 24 3.1.2 Galvanic Current Measurement....................................................................... 24 3.1.3 Electrochemical impedance spectroscopy (EIS)............................................26 3.2 Surface analysis techniques......................................................................................33 3 .3 Factorial design o f experiments and optimization o f process parameters 33 4. EXPERIMENTAL APPROACH_____________________________________ 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 .1 Materials and pretreatments.....................................................................................35 4.2 Electrochemical measurements............................................................................... 36 4.2.1 Electrochemical cells......................................................................................... 36 4.2.2 Electrochemical testing methods...................................................................... 37 4.2.2.1 EIS................................................................................................................. 37 4.2.2.2 DC polarization technique......................................................................... 40 4.3 Surface analysis......................................................................................................... 41 4.4 Evaluation o f cerium salts........................................................................................ 41 4.4.1 Tests in 0 .1 N N aC l............................................................................................42 4.4.1.1 Cerium salts as inhibitors............................................................................42 4.4.1.2 Chemical conversion coatings....................................................................42 4.4.1.3 Electrodeposition......................................................................................... 43 4.4 1.4 Cerating process.................................................................................... 43 4.4.2 Tests in the baseline solution at RT................................................................. 43 4.4.3 Tests in the baseline solution at I00°C............................................................43 4.5 Optimization o f the cerating process......................................................................45 4.5 .1 Optimization o f process parameters through factorial design.....................45 4.5.2 Investigation o f pretreatment processes...........................................................47 4.5.3 Investigation o f post-treatment processes (sealing)......................................47 4.6 Evaluation o f Yttrium Salts......................................................................................48 4.7 Evaluation o f Organic Inhibitors.............................................................................49 5. EXPERIMENTAL RESULTS______________________________________.50 5 .1 Evaluation o f cerium salts.......................................................................................50 5.1.1 Evaluation in 0 .1 N NaCI.................................................................................50 5.1.1.1 Cerium salts as inhibitors............................................................................50 5.1.1.2 Cerium salts used in chemical conversion coatings............................... 54 5.1.1.3 Cerium salts used in electrodeposition.....................................................56 5.1.1.4 Cerium salts used in cerating.....................................................................60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.1.2 Evaluation in the baseline solution at R T....................................................... 70 5 1.3 Evaluation in the baseline solution at 100°C..................................................74 5 .2 Optimization o f the cerating process...................................................................... 82 5.2.1 Factorial design......................... 84 5 .2.2 Effects o f pretreatments, post treatments and additives..............................102 5.2.2.1 EIS results....................................................................................................105 5.2.2.2 SEM results................................................................................................. 112 5.2.3 Effect o f cerating solution pH ......................................................................... 115 5.2.4 Effect o f "'aging" of cerating solution.............................................................119 5 .2.5 Effect o f sealing in silicate solution...............................................................119 5 .3 Evaluation of Yttrium Salts...................................................................................119 5.3.1 Fast chemical conversion ("Yttrating").......................................................... 122 5.3.1.1 SEM results................................................................................................. 122 5.3.1.2 EIS results in 0 .1 N NaCl at RT...............................................................125 5.3.1.3 EIS results in the baseline solution at RT...............................................125 5.3.1.4 EIS results in the baseline solution at 100 °C ........................................ 128 5.3 .2 Y;(S04>3 as Inhibitor..................................................................................... 130 5.4 Evaluation o f organic inhibitors...........................................................................130 5.4.1 EIS results.......................................................................................................... 132 5.4.2 DC polarization results.....................................................................................137 6. DISCUSSION____________________________________________________ 141 6 .1 Evaluation o f cerium salts..................................................................................... 141 6 .1.1 Cerium salts as inhibitors................................................................................. 141 6.1.1.1 Corrosive medium dependence................................................................ 142 6 .1.1.2 pH dependence........................................................................................... 142 6 .1.1.3 Cerium salt anion dependence................................................................. 143 6 .1.1.4 Time dependence........................................................................................ 144 6.1.2 Chemical conversion coatings......................................................................... 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vu 6.1.3 Electrodeposition...............................................................................................147 6.1.4 Cerating...............................................................................................................148 6.1.4.1 0.1 N N aC latR T ........................................................................................149 6 .1.4.2 Baseline solution at RT..............................................................................150 6.1.4.3 Baseline solution at 100°C........................................................................ 152 6.2 Optimization o f the cerating process.................................................................... 152 6.2.1 Concentration o f solution constituents.......................................................... 153 6.2.2 Cerating tim e......................................................................................................155 6.2.3 Cerating solution pH......................................................................................... 156 6.2.4 Additives.............................................................................................................156 6.2.5 Pretreatments......................................................................................................157 6.2.6 Post treatm ents.................................................................................................. 159 6.2.7 Aging o f cerating solution...................................................... 162 6.3 Evaluation o f yttrium salts......................................................................................164 6.4 Evaluation o f organic inhibitors.............................................................................166 7. CONCLUSIONS__________________________________________________ 170 7.1 Evaluation o f cerium salts.......................................................................................170 7.2 Optimization o f the cerating process.....................................................................171 7.3 Evaluation o f yttrium salts......................................................................................173 7.4 Evaluation o f organic inhibitors.............................................................................173 8. SUGGESTIONS FOR FUTURE WORK_____________________________ 175 8 .1 Optimization o f cerating process............................................................................175 8.2 Mechanistic studies on sealing...............................................................................175 8.3 Recycling o f the cerating solution......................................................................... 176 8.4 Dual protection strategy for mild steel in A-W systems.....................................176 8.5 Application o f the cerating process and REMSs in LiBr system s...................177 9. REFERENCES___________________________________________________ 178 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. v i i i PART O: STABILITY EVALUATION OF METALLIC MATERLALS IN LI-ION BATTERY ELECTROLYTE_________________________ 188 1. INTRODUCTION_________________________________________________189 2. LITERATURE REVIEW___________________________________________193 2 .1 Stability o f the cathode current collector..............................................................193 2 .1.1 AI composition and passive film integrity.................................................... 193 2 .1.2 Electrolytes.........................................................................................................194 2.1.3 Contaminants......................................................................................................196 2 .1.4 Effect o f temperature........................................................................................196 2.2 Stability o f anode current collector.......................................................................197 3. EXPERIMENTAL TECHNIQUES__________________________________ 198 4. EXPERIMENTAL APPROACH____________________________________ 199 4 .1 Materials and Pretreatments................................................................................... 199 4.2 Electrochemical experiment configuration..........................................................200 4.3 Electrochemical techniques.................................................................................... 201 4.3.1 Electrochemical impedance spectroscopy (EIS)..........................................201 4.3.2 Potentiodynamic polarization..........................................................................202 4.3.3 Potentiostatic polarization................................................................................203 4.3.4 Galvanic corrosion experim ent...................................................................... 203 4.4 Surface Analysis...................................................................................................... 204 5. EXPERIMENTAL RESULTS AND DISCUSSION____________________ 20S 5.1 Potentiodynamic polarization.................................................................................205 5 .1.1 Polarization curves and SEM results.............................................................205 5.1.2 Fitting o f polarization curves..........................................................................214 5.1.3 Stable potential window...................................................................................217 5.2 Potentiostatic polarization...................................................................................... 219 5.2.1 Polarization at + - 4.5V vs. Li/Li*....................................................................219 5.2.2 Polarization at + 3.5 V vs. L i/L F ...................................................................221 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ix 5.3 Galvanic corrosion tests..........................................................................................226 5.4 EIS measurements................................................................................................... 231 5.4.1 EIS results..........................................................................................................235 5.4.2 Analysis o f EIS results..................................................................................... 235 5.4.3 Corrosion rate time law and stability prediction..........................................242 5.4.4 Surface analysis.................................................................................................248 6. CONCLUSIONS__________________________________________________ 251 7. SUGGESTIONS FOR FUTURE WORK_____________________________ 253 7 1 The effect o f Alloying element AI on the stability o f Ti-AI alloys..................253 7.2 Corrosion protection o f Cu...................................................................................253 8. REFERENCES___________________________________________________ 255 BIBLIOGRAPHY___________________________________________________ 257 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES Table Page Part I 4.1 23 factorial design for optimization o f the cerating treatment.............................46 5.1 Rp. C. r^ofr and inhibition efficiency values for 1018 steel in NaCl for 2 hours............................................................................................................................ 53 5.2 Rp, C, T o o n - and inhibition efficiency values for 1018 steel after different chemical conversion treatment....................................................... 58 5.3 Rp, C and inhibition efficiency values for 1020 steel after 2 days exposure to different ammonia solutions at I00°C..............................................77 5.4 Rp. C. and r0 )rT values for 1020 steel exposed to 5 wt% N ’H? - 0.2 vvt% NaOH + 5 mM CefNO^h at 100°C for 2 days....................................................80 5.5 Fit parameters Rp and C values for the 23 factorial design experiments............86 5.6 The main and interactive effects and corresponding percentage o f factors FI (concentration ofCeClj). F2 (concentration o f HtOt) and F3 (cerating time) and the average values o f the responses......................................................87 5.7 Factors and responses for the 2 ' factorial design (2:-I)....................................... 89 5.8 Main and interactive effects o f factors F2 and F3 for 22 -I..................................90 5.9 Factors and responses for the 22 factorial design (22-IT)..................................... 92 5.10 Main and interactive effects o f factors F2 and F3 for 22 -tI................................ 93 5.11 Factors and responses for the 32 -I factorial design...............................................95 5.12 Factors and responses for the 32-tl factorial design............................................. 99 5.13 Experimental arrangements for the investigation o f pretreatments, post treatments and additives......................................................................................... 104 5.14 Corrosion rate r^ o n - in pm/year for experiments over one week in the baseline solution at RT........................................................................................... 109 5.15 Fit parameter C (F/cm2) values tor experiments over one week in the baseline solution at RT........................................................................................... 110 with permission of the copyright owner. Further reproduction prohibited without permission. XI 5.16 Fit parameters A* and Br, and correlation coefficient (R2), calculated thickness loss At (pm) and volume o f evolved hydrogen (at STP) for I m ' in one-year exposure......................................................................................I l l 5.17 Fit parameters -V .., Bc and correlation coefficient (R*) for C in Table 5.15... 113 5.18 Fit parameters aR, bR and the correlation coefficient (R: ) for 1018 steel in different solutions for two weeks at RT.............................................................. 134 5.19 Fitting results from potentiodynamic polarization curves for 1018 steel in different solutions at RT.........................................................................................140 Part II 5.1 Fit results from potentiodynamic polarization curves...................................... 216 5.2 Stable potential windows determined from anodic polarization curves........ 218 5.3 Charge calculations for applied potential (4.5 V vs. Li ’ / Li) experiments (for 24 hours)...........................................................................................................224 5.4 Charge calculations for applied potential (3.5 V vs. Li ’ / Li) experiments (for 24 hours)...........................................................................................................227 5.5 Charge calculations for galvanic corrosion experiments (for 24 hours)........230 5.6 Parameters a, b and correlation coefficient (R: ) from fitting Rp to the equation o f log ( 1/Rp) = a log (t) +b and predictions o f thickness loss after 10 years for Cu and Pt-Ir.............................................................................. 245 5.7 Conservative thickness toss calculations for Ti materials................................247 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. \ 1I LIST OF FIGURES Figure Page Part I I. I A schematic drawing o f an ammonia/water absorption heat system [4].............3 2.1 Pourbaix diagram for iron [11].................................................................................. 9 3.1 Potentiostatic tests for the study o f materials passivity and stability.................25 3 .2 Schematic illustration o f galvanic corrosion potential Eg and current...L.........27 3.3 (a) One-time-constant (OCT) model, (b) Bode plot and (c) complex plane plot for R* = 20 Q cm2. Rp = 10J Q-cm2 and C = I O '4 F/cm2...................29 3.4 (a) A two-time-constant (TTC) model, (b) Bode plot and (c) complex plane for R, = 20 Q cm ". Rp = 103 Qcm ". Y0 (CPE I) = I O '4 S/cm2 and Y„ (CPE2) = 10‘5 S/cm2........................................................................................... 32 4 .1 Cathodic polarization curve (a) and galvanostatic curve at the cathodic current density o f 3 mA/cm2 (b) for mild steel in 0 .1 M Ce(NQj)3 solution at RT............................................................................................................ 44 5 .1 Bode plots o f 10 18 steel in blank and inhibited 0 .1 N NaCl for 2 hours at RT; (a) impedance, (b) phase angle........................................................................52 5.2 Bode plots for 1018 steel exposed to 0.1 N NaCl + 20 mM CefNCh);*. pH 7.0 at RT for 2 days................................................................................................ 55 5.3 Bode plots o f 1018 steel exposed to 0 .1 N NaCl for 2 days after different chemical conversion treatments. For comparison, data are also shown for untreated samples exposed to NaCl with and without 20 mM Na2Cr2C >7 57 5.4 Bode plots o f 10 18 steel in 0 .1 N NaCl for 3 days after chemical conversion treatment in 20 mM Ce(NO})3 for 2 hours........................................59 5.5 Bode plots o f 1018 steel in 0. IN NaCl for 2 hours at RT after different cathodic polarization processes in 0 .1 M Ce(N0 3 ) 3 at RT.................................61 5.6 1018 steel exposed to 0 .1 N NaCl for 24 hours at RT after cerating in 10 g/L CeCI? + 5 wt% H2O2 aqueous solution at RT for 8 minutes; (a) Bode plots, (b) simulated data vs. experimental data 24-hour data.............................62 5.7 Micrographs o f cerated layer on 1018 steel formed in 10 g/L CeCh + - 5 wt% H2O2 aqueous solution at RT for 8 minutes.................................................64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. x i i i 5 .8 EDS results o f the cerated layers on 1018 steel; (a) mounts, (b) flat cracked areas..............................................................................................................65 5.9 Bode plots for 1018 steel in 0 .1 N NaCl after cerating and sealing treatment by cathodic polarization in 20 mM CefNChh at 3 mA/cm2 for 2 minutes.....................................................................................................................67 5 10 SEM images o f the cerated layer on 1018 steel after cathodic polarization in 20 mM CefNOsh at 3 mA/cm2 for 2 minutes............................6 8 5 .11 EDS analysis results o f the cerated layer on 1018 steel after cathodic polarization in 2 0 mM CefNCbb at 3 mA/cm' for 2 minutes; (a) mounds. (b) flat area................................................................................................................. 69 5.12 Bode plots o f 1018 steel samples exposed to different solutions at RT at the 7th day after different treatments.......................................................................71 5.13 log (1/Rp) vs. log (t) (a) and log (l/C ) vs. log (t) (b) plots for 1018 steel in different solutions at RT after different treatments. Dotted lines are the plots according to the fit equations.........................................................................72 5.14 Bode plots for 1020 steel after 2 days exposure to different ammonia solutions at 100°C; (a) impedance, (b) phase angle.............................................75 5 .15 Comparison of experimental and fitted data for 1020 steel after 2 days exposure to 5 wt% NH3 solution at 100°C; (a) equivalent circuit, (b) Bode plots...................................................................................................................76 5.16 Bode plots for 1020 steel exposed to 5 wt% NHj - f - 0.2 wt% NaOH - 5 mM CefNChb solution at 100°C for 2 days..........................................................79 5.17 (a) Bode plots and (b) fit parameters Rp and C vs. time for cerated 1020 steel exposed to 5 wt% NHj - 0.2 wt% NaOH solution at 100°C for 5 days..............................................................................................................................81 5.18 SEM images of cerated 1018 steel before (a) and after (b) immersion in 5 wt% NH3 + 0.2 wt% NaOH solution at 100°C for 24 hours..............................83 5.19 Bode plots for the cerated 1020 steel for the 23 factorial design experiments exposed to 5 wt% NH3 + - 0.2 wt% NaOH solution at 100°C after 48 hours; (a) impedance, (b) phase angle.....................................................85 5.20 Response surface o f Rp in 3D mesh (a) and 2D contour (b) for the 3 '-l factorial design o f CeCI3 and H2O2 concentrations..............................................96 5.21 Response surface o f C in 3D mesh (a) and 2D contour (b) for the 32-! factorial design o f CeCIs and H2O2 concentrations..............................................97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XIV 5.22 Response surface o f Rp in 3D mesh (a) and 2D contour (b) for the 3MI factorial design o f CeCL and H2O2 concentrations............................................1 0 0 5.23 Response surface o f C in 3D mesh (a) and 2D contour (b) for the 32-LI factorial design of CeCL and H2O2 concentrations............................................101 5.24 Bode plots for 1020 steel exposed in 5 wt% NH3 - 0.2% NaOH solution at I00°C for 48 hours after cerating at the optimization condition o f 2.3 g/L CeCL and 4.4 wt% H2O2 for 2 0 minutes and 5 minutes............................ 103 5 25 Bode plots for EIS tests carried out in 5 wt% NH3 - 0.2 wt% NaOH (baseline) solution at RT for one week; (a) Micro I, (b) Actane3. (c) Actane6 and (d) Actane7......................... 107 5.26 SEM images for mild steel, cerated in 3.0 g/L CeCL + 2.5 wt % H2O2 at room temperature for 5 minutes (A ctanel)......................................................... 114 5.27 SEM images for mild steel, cerated in 3.0 g/L CeCL + - 2.5 wt % H2O2 at room temperature for 5 minutes, then sealed in 10 % N ^SiO t • 9H20 at 50°C for 30 minutes................................................................................................ 116 5.28 SEM images for mild steel, cerated in 3 .0 g/L CeCL - 2.5 wt % H2O2 - 2 g/L NaN0 2 + 4ppm Pb(AC>2 + 3ppm Triton X - 1 0 0 at room temperature for 5 minutes (Actane2 )...................................................................117 5.29 1020 steel exposed to the baseline solution at 100 °C for 48 hours after cerating in cerating solutions o f 2.3 g/L CeCL ~ 4.4 wt% H2O2 - 2 g/L NaN0 2 + 4 ppm Pb(AC)2 + 3 ppm Triton X -100 with different pH at RT for 20 minutes: (a) Bode plots, (b) fit parameters Rp and C vs. pH.......... 118 5.30 Bode plots for 1 0 2 0 steel exposed to baseline solution at 1 0 0 °C for 48 hours after cerating treatments with and without aging o f cerating solutions o f 2.3 g/L CeCL + 4.4 wt% H2O2 + 2 g/L NaNC>2 + 4 ppm Pb(CH3COO>2 ~ 3 ppm Triton X -100 at RT for 20 minutes. The pH o f cerating solution was 1.9........................................................................................ 120 5.31 Bode plots for 1020 steel exposed to baseline solution at 100 = C for 48 hours with and without sealing in 1 0 % sodium (meta) silicate at 50 °C for 30 minutes after cerating treatment (without aging) as described in Fig. 5.30................................ 7 .................................... 7...7...................................... 121 5.32 SEM images o f mild steel vttrated in 12.5 g/L YCL ~ 2.5 wt % H2O; solution for 20 minutes at room temperature...................................................... 123 5.33 SEM images o f mild steel vttrated in 12.5 g/L YCL ~ 2.5 wt % H2O2 solution for 20 minutes at RT, and immersed in 5 wt % NH3 + 0.2 % NaOH solution for 48 hours at I00°C..................................................................124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XV 5.34 1018 steel, yttrated in 12.5 g/L YClj - 2.5wt %HzOz for 20 minutes at RT, and then exposed to 0 .1 N NaCl at RT for one week; (a) Bode plots. (b) fit parameters Rp and C vs. time.................................................................. 126 5.35 1018 steel, yttrated in 12.5 g/L YCI3 + 2 5wt %H2C >2 for 20 minutes at RT. and then exposed to 5wt% NH3 - 0.2% NaOH at RT for one week; (a) Bode plots, (b) fit parameters Rp and C vs. time......................................... 127 5.36 1020 steel, yttrated in 12.5 g/L YCI3 - 2.5wt %H~02 for 20 minutes at RT, and then exposed to 5 wt% NH3 + 0.2% NaOH at 100 °C for one week; (a) Bode plots, (b) fit parameters Rp and C vs. time..............................129 5.37 1020 steel in 5wt% NH3 + 0.2 % NaOH + ImM Y2(S04>3 for one week at 100 °C. (a) Bode plots, (b) fit parameters Rp and C vs. time....................... 131 5.38 Time dependence o f fit parameters Rp (a) and C (b) for 1018 steel in different solutions for two weeks at RT.............................................................. 133 5.39 Bode plots for 1018 steel in 5wt%NH3 +0.2% NaOH with 10 mM L-aspartic acid (AA) monosodium salt for 2 weeks at RT................................ 135 5.40 (a) TTC model [159,160], (b) experimental and fit data for Id. 5d and I4d spectra in Fig. 5.30. and (c) fit parameters R and C vs. time....................136 5.41 1 0 18 steel in different solutions after 2-hour immersion at RT; (a) cyclic anodic polarization curves, (b) Cathodic polarization curves...........................139 6.1 Cathodic polarization curves for 1018 steel with different treatments in 0 .1 N NaCl at RT after 2-hour immersion...........................................................151 6.2 Dependence o f cerating solution pH on aging time as a function o f (a) concentration of NaNOi additive, (b) concentration of CeCL and (c) concentration o f H2O2............................................................................................ 158 6.3 Bode plots for mild steel, (a) in 0 .1 N NaCl at RT for 24 hours after cerating and yttrating, (b) in baseline solution at RT for 7 days after cerating and yttrating, and (c) in the hot baseline solution inhibited by yttrium sulfate and cerium nitrate after 48 hours............................................... 165 6.4 Comparison o f 2-hour impedance spectra for mild steel exposed to RT baseline solutions without or with organic inhibitors........................................168 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XVI Part II I . I Schematic diagram o f a Li-ion cell on discharge [6].......................................... 190 5.1 Potentiodynamic polarization curves for different metallic materials in LiPF6 + EC + DEC at 37°C (after 2 hours immersion).....................................206 5.2 SEM images for Cu foil after anodic potentiodynamic polarization test; (a) exposed area, (b) unexposed area................................................................... 208 5.3 SEM images for A 1 foil after anodic potentiodynamic test; (a) exposed area, (b) unexposed area........................................................................................209 5.4 Potentiodynamic polarization curves for Ti and Ti alloys in LiPF* - EC + DEC at 37°C (after 2 hours immersion)...........................................................210 5.5 SEM images for Ticp sheet after anodic potentiodynamic test; (a) exposed area, (b) unexposed area......................................................................... 211 5.6 SEM images for Ti-6Al-4V sheet after anodic potentiodynamic test; (a) exposed area, (b) unexposed area......................................................................... 212 5.7 SEM images for Ti-3A1-2.5V sheet after anodic potentiodynamic test; (a) exposed area, (b) unexposed area......................................................................... 213 5 .8 SEM images for Pt-lr foil after anodic potentiodynamic test; (a) exposed area, (b) unexposed area........................................................................................215 5 .9 Current vs. time at an applied potential o f (a) 4.5 V and (b) 3 .5 V' (b) vs. Lf/Li in LiPF6 + EC + DEC at 37°C (Sample area = 1.98 cm2).................... 220 5.10 SEM images for Ti-6A1-4V sheet after applied potential test (4.5 V vs. Lf/Li); (a) exposed area, (b) unexposed area.................................................... 222 5 .11 SEM images for Ti-3AI-2.5V sheet after applied potential test (4.5 V vs. Li7Li); (a) exposed area, (b) unexposed area.................................................... 223 5.12 SEM images for Ticp foil after applied potential test (3.5 V vs. Lf/Li); (a) exposed area, (b) unexposed area..................................................................225 5.13 Galvanic corrosion test for A 1 / Ti-6Al-4V couple in LiPF6 + EC + - DEC at 37°C (Sample area = 1.98 cm"); (a) OCP vs. time for AI and Ti-6AI- 4V before test, (b) Eg vs. time and (c) Ig vs. tim e............................................. 228 5.14 Galvanic corrosion test for AI / Ti-3A1-2.5V couple in LiPF6 + - EC + - DEC at 37°C (Sample area = 1.98 cm2); (a) OCP vs. time for AI and Ti- 3A1-2.5V before test, (b) Eg vs. time and (c) Ig vs. tim e ..................................229 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XVII 5.15 Galvanic corrosion test for Ticp/Pt-Ir couple in LiPF6 - EC + - DEC at 37°C (Sample area = 1.98 cm2); (a) OCP vs. time for Ticp and Pt-lr before test, (b) Eg vs. time and (c) Ig vs. tim e....................................................232 5.16 Galvanic corrosion test for Cu/Ticp couple in LiPF6 - EC - DEC at 37°C (Sample area = 1.98 cm2); (a) OCP vs. time for Cu and Ticp before test, (b) Eg vs. time and (c) Ig vs. time.................................................................233 5.17 SEM images for Cu after Ticp -C u galvanic corrosion test; (a) exposed area, (b) unexposed area....................................................................................... 234 5.18 Bode plots for AI foil in LiPF6 - EC DEC at 37°C........................................236 5.19 Bode plots for Pt-lr foil in LiPF6 + • EC + DEC at 37°C.................................... 237 5.20 Bode plots for Ti-6A1-4V sheet in LiPF6 - EC - DEC at 37°C...................... 238 5.21 Bode plots for Ti-3 AI-2.5V sheet in LiPF6 + EC + DEC at 37°C....................239 5.22 Bode plots for Ticp sheet in LiPF* - EC - DEC at 37°C................................. 240 5.23 Bode plots for Cu foil in LiPF6 + EC + DEC at 37°C....................................... 241 5.24 (a) log ( I/Rp) vs. log (t) and (b) log ( l/C) vs. log (t)......................................... 243 5.25 SEM images after 30-day immersion EIS test; (a) Cu foil, (b) Ticp foil. (c) Ti-6AI-4V sheet and (d) Ti-3A1-2.5V sheet............................................... 250 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT In part I, rare earth metal salts (REMSs) and organic additives were extensively evaluated for their use in the corrosion protection o f mild steel in the A-W baseline solution o f 5 wt% NHj + 0.5 wt% NaOH. The REMSs. specifically cerium salts and yttrium salts, were systematically investigated with electrochemical methods (EIS, DC polarization) and surface analysis techniques (SEM. EDS). Cerium nitrate as inhibitor, despite its low solubility in the 100°C A-W baseline solution, provided high corrosion protection that was at least comparable to that provided by dichromate. The cerating process which was the fastest and most convenient method to form CedV) (hydr)oxide film on mild steel is a multi-factor dependent process. Through factorial design and single-variable experimental methods, optimum process conditions were determined to obtain the cerated layer with the best corrosion protection for mild steel in the A- W baseline solution. Fores and cracks in cerated layer were sealed in the hot A-W baseline solution, and the corrosion protection provided by the cerated layer increased with prolonged exposure time. The optimum cerating process for mild steel includes pickling o f mild steel in I . I volume ratio of hydrochloric acid and Dl water at room temperature (RT) for 60 seconds, cerating in solution o f 2.3 g/L CeCl3 + 4.4 wt% HjO; - appropriate additives at pH 2.2 at RT for 20 minutes with 30 minutes solution aging prior to cerating, and sealing o f the cerated layer in 10 % sodium silicate or molybdate solution at 50°C for 30 minutes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XIX The uniqueness o f the working environment in A-W systems makes it verv challenging to select appropriate organic inhibitors. Four promising organic inhibitors were screened in terms of their long-term inhibition capabilities in the A- W baseline solution at room temperature. Glycerophosphate (GPH) was found to be the most promising inhibitor, but confirmation of its high temperature stability is still needed in the hot A-W baseline solution. In part II. the stability o f six metallic materials was evaluated with different electrochemical methods and surface analysis techniques. All electrochemical evaluation work was conducted in IM lithium hexafluorophosphate (LiPF6) dissolved in a 1:1 volume mixture of ethylene carbonate (EC. CH2CH2CO3) and diethyl carbonate (DEC. (CH?CH2hCCM at 37°C in a drv-box. AI foil and 90Pt/I0Ir sheets demonstrated high stability in the test electrolyte. Virtually no sign o f corrosion could be observed on both materials with SEM after every electrochemical evaluation test, e.g. potentiodynamic and potentiostatic polarizations, galvanic corrosion and 30-day EIS experiments. However. 90Pt/10Ir could cause chemical decomposition o f battery electrolyte resulting in low stable potential window. Cu foil can be easily attacked in the form o f pitting corrosion even under small anodic polarization; therefore, it has practically no stable potential window. Coupling o f Cu with other metallic materials having higher or similar OCPs could Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. XX introduce localized galvanic corrosion to Cu. While at its OCP. no significant corrosion was observed with SEM for Cu after the 30-day EIS test. The three Ti materials, i.e. commercial pure Ti foil. Ti6A14V and Ti3AI2.5V sheets showed passive behavior in the test electrolyte. Higher concentrations o f alloying elements AI and/or V in Ti alloys resulted in a larger stable potential window. After the 30-day EIS test. localized corrosion was noticed with SEM on commercial pure Ti foil, while the two Ti alloys were covered with corrosion products. AES results suggested that the corrosion products consisted mainly o f oxides and fluorides o f lithium and titanium. The battery electrolyte exhibited high chemical stability at 37°C in contact with AI and the three Ti materials up to 6.5 V vs. Li/Li'. However, the electrolyte would decompose at applied potentials in the presence o f 90Pt/1 OIr due to its catalytic activity. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I PART I: DEVELOPMENT OF CORROSION PROTECTION METHODS FOR AMMONIA/WATER ABSORPTION SYSTEMS Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1. INTRODUCTION Ammonia/water (A-W) absorption heat systems, such as heat pumps or chillers have regained their popularities in the industry o f heating, ventilation and air conditioning (HVAC) since the late 1980s, because ammonia is a low-cost, environmentally safe refrigerant with excellent thermodynamic properties, while the existing refrigerants, such as chloroflurocarbons (CFCs) and hydrochloroflurocarbons (HCFCs) are believed to have ozone depleting potentials with effects on global warming. They are therefore under strict attack and in a phase-out schedule [1.2]. Another stimulus to the rejuvenation o f the A-W absorption system has been its flexibility on almost any energy source, especially the low-cost natural gas or industrial waste heat sources. In the United States. A-W absorption systems have been used mostly in large capacity applications, however, because o f environmental and energy concerns, they are under intensive development to be accommodated in smaller capacity systems, such as gas-fired residential air-conditioning systems, for example the Generator Absorber Heat Exchange C'GAX") cycle system [3], One A-W absorption system along with its main components is schematically shown in Fig. I. I [4], The basic operating principle is based on the fact that ammonia is very soluble in cold water, but not in hot water. Water only acts as the transport medium in the system. Detailed descriptions o f the cooling and heating mechanisms o f an A-W absorption system can be found in related references [4-7]. It is important to notice that different components in the system experience different environments. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A I II I I I I I Evaporator In v I Ammonia Vapor | [ Hydrogen Absorber Water Dissolved ammonia . Separtor Heat Fig. t.I A schematic drawing o f an ammonia/water absorption heat system [4]. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 Generator, separator and absorber are exposed to an A-W mix solution with the highest temperature at the generator and separator, around 200°C, and the lowest at the absorber, around 35°C. The condenser transfers heat from ammonia vapor to its surroundings, and is maintained at about 35°C, while the evaporator withdraws heat from its surroundings to the low pressure hydrogen / liquid ammonia mix inside, and is kept at about -20°C. To be competitive in the HVAC market. A-W absorption systems have to be constructed with inexpensive materials and have a long equipment lifetime, e.g. 20 years with free or low maintenance [8], Carbon steel is the material of choice since it is readily available, economical, and easily formed and welded. However it is susceptible to corrosion in A-W mix solutions at high working temperatures. Moreover, one of the corrosion products is hydrogen gas. The build-up of hydrogen gas at the condensation and absorption surfaces can greatly reduce the efficiency o f the absorption system [9], Conventionally, chromate or dichromate is added as inhibitor in the A-W solution. Through adsorption, chromate or dichromate passivates mild steel and forms a very- thin and protective ferric and chromic oxide mixture, which inhibits corrosion to tolerable levels. However, since chromate and dichromate are known carcinogens and environmentally hazardous, their usage has been restricted in many localities, even with zero tolerance [8-10]. Moreover, chromate is called a '^dangerous inhibitor” since it is an anodic inhibitor and is consumed gradually during operation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Periodical replenishment is required to maintain the concentration o f chromate above a safe level. Therefore successful development o f chromate-free. maintenance-free and environmentally friendly inhibitors or alternative corrosion protection methods for carbon steel in A-W absorption systems have become very urgent and critical, and are the aim and theme o f this thesis. REMSs and organic additives are environmentally friendly, efficient and effective, i.e. so-called E3 corrosion inhibitors. REMSs have been predominantly studied for the localized corrosion prevention o f A 1 alloys in halide containing media. Corrosion protection results from the formation o f an insulating REM (hydr)oxide layer preferably on the cathodic sites o f the alloy surface due to local pH increase as the result o f oxygen reduction. The REM (hydr)oxide layer can be formed through various methods, such as addition of REMSs in corrosive environments as inhibitors, chemical conversion coating formed by immersion in aqueous REMSs solutions, electrodeposition in REMSs solutions, and formation of fast chemical conversion coating with the help of strong oxidizing agents. However these methods can produce REM (hydr)oxide layers with different properties and compositions leading to different corrosion protection capabilities. Organic inhibitors have also been receiving more attention mainly due to their non toxic and biodegradable nature. Organic additives are widely used in the corrosion inhibition o f mild steel in acidic or neutral solutions, however less frequently in alkaline media, since in high pH solution, a protective magnetite film can easily form Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on mild steel [11. In A-W absorption systems, alkali metal hydroxides and buffers are usually added to the working fluid to maintain an appropriate pH range [9], Furthermore, the A-W system usually operates at high temperatures (up to 200°C). which could cause decomposition o f organic additives. Thus, selection o f appropriate organic inhibitors for A-W absorption systems is a challenging task In this thesis, REMSs. specifically cerium salts and yttrium salts, and various rare earth metal (REM) (hvdr)oxide formation methods were evaluated extensively in terms o f corrosion protection o f mild steel. The evaluation was first performed in corrosive 0 .1 N NaCI at room temperature (RT) with a comparison o f dichromate as inhibitor, then the promising REMSs and REM (hydr)oxide formation methods were further investigated in the A-W baseline solution of 5 wt% NTh - 0 2 wt°"0 NaOH. simulating the working fluid o f A-W absorption systems at RT as well as high temperature ( t00°C). The cerating process is a multi-variable dependent process. The optimization o f the cerating process was thoroughly pursued through factorial design and regular single-variable methods in order to obtain a cerated layer with the best corrosion protection for mild steel in the A-W baseline solution. Some organic inhibitors for mild steel reportedly effective in alkaiine solutions were also assessed in terms o f their corrosion inhibition efficiency in the A-W baseline solution at RT in an attempt to understand their inhibition mechanisms and to select promising inhibitors for use in the high temperature baseline solution. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 During the evaluation o f REMSs and organic inhibitors, electrochemical impedance spectroscopy (EIS) and DC polarization methods were generally employed in the determination o f time dependence of corrosion rate and in mechanistic studies of corrosion. Surface analysis techniques, such as SEM and EDS were selectively applied to REM (hydr)oxide films formed on mild steel to obtain morphological and surface chemistry information to aid in a better understanding and interpretation of corrosion protection provided by REMSs. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 2. LITERATURE REVIEW 2 .1 Corrosion o f carbon steel in A-W absorption systems The corrosion susceptibility o f iron in aqueous solution at different pH can be described by a Pourbaix or potential-pH diagram, as shown in Fig. 2.1 [II], At relatively high pH and room temperature, iron is in the passivation region due to the formation o f a magnetite (Fe3 0 4) passive film. However, the corrosion protection o f the passive film decreases with the increase o f solution temperature [12-21], It was reported that there exists a corrosion region in which soluble Fe(Il) species may form, between the immunity and passivity’ regions in the revised Pourbaix diagram in the temperature range o f 100 to 300°C [21 ]. Components such as generator, separator and absorber in an absorption system are exposed to weak alkaline A-W solution with temperatures up to 200°C, which is corrosive to carbon steel [8.9.22], In the de-aerated A-W solution at elevated temperatures, iron corrodes to form a layer o f magnetite and release hydrogen gas through the formation o f ferrous hydroxide (Fe(OH)-) as an intermediate. This process was first proposed by Schikorr [23-24] and is called the Schikorr reaction which can be written as: 3 Fe + 4 H20 -► Fe3 0 4 + 4 H2 (g) 2-1 During the corrosion process, the magnetite layer can get thick enough to crack and loss integrity, and finally peel off and expose the fresh substrate. The peeled-off Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 1.8 H i = 1.0 C O d O 0.6 c « 0 o o. -0.2 - 0.6 Fe (Immunity) 1.0 HFeO. 1.4 1 0 1 3 5 7 9 11 13 15 PH Fig. 2.1 Pourbaix diagram for iron [11]. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 magnetite forms sludge, whose accumulation with time inside an absorption pump can block the circulation o f the ammonia/water mix solution, and decease the efficiency o f the heat transfer [8.9. 25-27]. The other corrosion product is hydrogen gas, which is considered more harmful to the absorption system [28], Since it is light and non-condensable. the hydrogen gas can be accumulated and trapped at the condenser and absorber surfaces [9.22]. Therefore, the absorption system must be purged of the hydrogen gas frequently, with a resulting loss o f ammonia. In the worst case, the presence o f the non- condensable gas causes the absorption system to operate at higher temperatures and pressures, thereby causing greater leakage, more corrosion and more danger from possible explosions [28]. The best way known so far to protect the absorption system is the addition o f inhibitors to the A-W working fluid to suppress the corrosion of carbon steel according to Eq. 2-1. 2.2 Corrosion inhibition o f carbon steel The corrosive environment in an A-W absorption system is unique, i.e. oxygen-free, high temperature, dynamic, and weak alkaline aqueous solution with varying concentrations o f ammonia in a closed system. This places a strict restriction on the selection o f appropriate inhibitors. The traditional inhibitors for the A-W absorption system are chromate salts, found effective about 80 years ago [28]. The chromate salts can provide acceptable corrosion inhibition with low volume generation o f hydrogen gas and trouble-free operation for the absorption system [8,9.28]. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I I However, due to the toxic and carcinogenic nature o f the chromate salts, extensive attempts have been made to develop chromate-ffee. environmentally friendly, efficient and cost effective, so-called EJ (ecology, efficiency and economy) inhibitors for the absorption system as well as water-cooling systems [8.9.29-49], The REMSs, especially the cerium salts, exhibit a promising tendency to replace chromates. They can also provide corrosion protection with formation o f chemical conversion coatings for mild steel and other metallic materials in aqueous solutions [50-99], 2.2.1 Traditional inhibitor - chromate Chromate is a strong oxidizer or passivator It is believed that the inhibition mechanism o f chromate is a combination o f adsorption and oxide formation on the steel surface [100-104], Through adsorption, chromate polarizes the anode to the passivity region and forms a very thin, but protective ferric and chromic oxide mixture. Therefore, chromate is an anodic inhibitor. The concentration o f chromate has to be kept above a critical level to repair the thin oxide film; otherwise, chromate can accelerate the corrosion o f steel through localized attack [103.104], Chromate is effective in the corrosion inhibition o f carbon steel in aqueous solution in a wide temperature range, up to 200°C [105] and a wide pH range, at least 6-11 [106], with or without the presence o f oxygen in solution. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 In the hot A-W solution, chromate has shown many drawbacks as inhibitor. In addition to its toxic and carcinogenic properties, it was found that a relatively high concentration o f chromate. about 0 .2 wt% must be maintained in order to effectively suppress the evolution o f hydrogen gas in A-W absorption systems [28]. However, it was found that chromate can react with ammonia at high temperatures resulting in the formation o f non-condensable nitrogen gas [25]: NHi - H20 * Na2C r04 -> Cr(OH)3 - 2NaOH - ' > \ 2 (g) 2-2 This reaction causes additional consumption o f chromate and increases the chances o f localized corrosion. Chromate is effective only up to about 200°C for carbon steel, which is near the peak solution temperature in the GAX cycle [ 105], However, the temperature at the surface of the heat exchanger in GAX cycle system is at least 10°C higher [107], Chromate may no longer be the proper choice o f inhibitor, since at such high temperatures the adsorption o f chromate to steel surface is greatly decreased [104], and the rate o f scale (magnetite) formation and chromate breakdown increases with increasing temperature [26]. 2.2.2 Chromate-free inhibitors The search for chromate-ffee E3 inhibitors has been launched several decades ago with a trend from toxic inorganic to low toxic or non-toxic inorganic to organic compounds [30,32]. For carbon steel in aqueous solution, the research has been mainly focused on acidic media (pH 2-5) [108-112] and neutral media (pH 5-8) [29- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 36.38-46], Only little work has been done on absorption systems with lithium halide/water working fluids [37,47.48] or A-W working fluids [8,9,22.49] and other alkaline media (pH 8-12) [113-116] due to the fact that the corrosion rate of carbon steel in alkaline solution is relatively low at room temperature (Fig. 2.1) [117], The uniqueness of A-W working fluid makes the selection of appropriate inhibitors difficult. The presence o f ammonia with varying concentration can prevent many inhibitors from working well in aqueous systems [8], The anaerobic environment may invalidate the anodic non-oxidizing inhibitors, such as molybdates [118] and vanadates [119], which are thought to be effective, low toxic and promising chromate substitutes in otherwise oxygen-rich conditions for carbon steel [30], The alkaline nature of the A-W solution can reduce the efficiencies o f many scale- forming inhibitors, which only work in the neutral environments, such as polyphosphates which are effective only in the pH range of 5-7 for iron and steel [106], The high working temperature o f the A-W fluid may rule out the use of organic inhibitors, since most organic compounds are unstable above 400F (204°C) [93], which is the working temperature o f absorption systems. Battelle [8.22] examined sodium zincate and sodium silicate as alternatives to chromate-based inhibitors in A-W chillers. Results showed that zincate was not as effective as chromate-based inhibitors with a higher amount o f hydrogen gas evolution and early stage failure from cracking near a weld at the absorber. Silicate- inhibited (0.17 wt% NaOH • + • I wt% sodium metasilicate) chillers would generally Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 4 generate less amount o f hydrogen gas. However, silicate caused excessive temperature failure at the generator resulting in early shutdown due to the low solubility o f silicates in A-W solutions, which clogged the circulation o f the working fluid. Silicate could also be responsible for the stress cracking corrosion (SCC) failure o f chillers pretreated and inhibited with silicates. Silicate-accelerated SCC o f steel was also reported in alkaline boiler water above 208°C (407F) [120]. However, other studies [121.122] found that silicates were effective inhibitors for SCC of stainless steel. Phillips et al. [9] found that the hydroxyl ion (OH~) concentration played a critical role in the corrosion inhibition o f carbon steel in aqueous ammonia solution at temperature up to 2I8°C (425F). The hydrogen generation rate was much reduced when the OH" concentration was maintained in the range of 0 015 - 0.2 N. adjusted by alkali bases, with or without the addition of soluble, non-toxic buffers, such as alkali borates, molvbdates and acetates. Similarly, the alkali bases were also used as inhibitors with or without molybdate buffers in lithium halide / water absorption systems [29.49], A thorough literature search has been conducted in an attempt to find potential chromate-free non-toxic inhibitors for the A-W solutions. However, only a few papers dealing with inhibitors for carbon steel in alkaline aqueous solutions were found [113-116.123.124], Two papers described inorganic inhibitors containing the toxic oxidizers dichromate and nitrite [123,124], while others examined organic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 inhibitors [113-116]. 3-[TrimethoxysilyI]-l-propanethiol (TMSPT) [115] could provide corrosion inhibition for carbon steel to some extent at pH 12, however, the inhibition efficiency decreased with increasing exposure time. In addition, TMSPT contains silane groups, which are considered highly toxic [125], Di-sodium 3- glycerophosphate (GPH) [113] was examined as a non-toxic inhibitor for steel in concrete. It was found that GPH inhibits both the anodic and cathodic reactions and affords good inhibition efficiency. Aspartic acid [ 114], an amino acid o f low molecular weight, was found to protect steel at pH > 10, and polyaspartic acid had a similar inhibition effect and pH dependence. Polygiutamic acid was found to effectively inhibit the corrosion o f mild steel in neutral artificial seawater at a low concentration o f 20 ppm [126]. It is an amino acid with high molecular weight, is non-toxic and biodegradable. 2.3 Corrosion protection by REMSs Since their introduction as cathodic inhibitors in the early 1980s, REMSs. especially cerium salts, have been studied extensively in order to replace chromate in many corrosion protection applications. The rare earth elements include the lanthanide elements (s7La - ?iLu), 2iSc and »Y [127]. Their oxides or hydroxides are insoluble in neutral or alkaline aqueous solutions, and hence the rare earth cations are potentially effective cathodic inhibitors. Moreover, the toxicity o f REMSs is very low, considered similar to that o f sodium chloride [128]. Some o f the rare earth elements, like cerium (66 ppm in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 crustal rocks, ranked 26), are by no means rare, but relatively abundant on the earth, similar to Cu (68 ppm in crustal rocks, ranked 25) [129]. Therefore, REMSs are promising E3 alternatives to chromate. So far, REMSs have been used as inhibitors [50-72] and in surface modification agents [72-99] on many metals and alloys including steel [50-54], but mostly on A 1 alloys. The surface modification methods include immersion in REMS aqueous solutions at room temperature or elevated temperatures [73-89], cathodic polarization [90-93], cerating [94-97] and surface cleaning or desmutting [94.98.99], 2.3.1 REMSs as inhibitors REMSs have been examined as inhibitors on various metals and alloys, such as carbon steel [50-54], zinc or zinc galvanized steel [54-57], bronze [58], nickel [59], stainless steel [60-64] and A 1 alloys [54.65-70], The inhibition mechanisms have also been investigated [69-72], It is commonly agreed that REM cations are cathodic inhibitors operating through precipitation o f REM (hvdr)oxide film due to the local pH increase at or near the substrate as a result o f oxygen reduction at cathodic sites. The REM (hydr)oxide film acts as a barrier to the supply o f oxygen and electrons to stifle the oxygen reduction reaction. As a result, the corrosion current density drops and the OCP moves in the negative direction. For mild steel in soft aerated tap water with an addition o f lOOppm CeCb, potentiodynamic polarization data demonstrated that the corrosion Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 current density decreased about an order of magnitude and OCP shifted about 230 mV in the negative direction [54]. Auger electron spectroscopy (AES) studies by Hinton et al. [54.55.130] showed that the composition o f the protective film depends on the substrates and specific REM cations used. For steel and zinc exposed to solutions containing lOOOppm CeCb-7H;0. the films consisted o f cerium and substrate metal oxides, while the film on AI7075 under the same condition consisted almost entirely o f cerium oxide with virtually no aluminum oxide. 2.3.2 Surface modification In order to replace chromate-based surface modification methods. REMSs-based surface modification methods have been examined widely in order to torm protective REM (hydr)oxide films on the substrate surface. They include immersion in REMSs aqueous solutions at room temperature or elevated temperatures with or without follow-up procedures [73-89], cathodic polarization [90-93], cerating [94-97] and surface cleaning or desmutting [94.98.99], Open-circuit immersion in REMSs aqueous solutions at room temperature could produce the protective (hydr)oxide film, but this process takes a long time, at least 100 hours [52]. Cathodic polarization could form the (hydr)oxide film in less than 0.5 hour, but the film was non-uniform, full o f cracks and lacking of durability in test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 solutions [52, 92,93], Furthermore, the film consisted mainly o f cerous (III) (hydr)oxide, which is not as protective as eerie (IV) oxide [71]. Hinton et al. [94,95] developed a rapid and convenient film deposition method, called “cerating”, by immersion o f metals or alloys in an aqueous solution mixture o f cerium chloride and hydrogen peroxide (H2O2). In the presence o f H2O2. the cerium (III) cation exists in the form o f a peroxo complex through the following reaction at pH < 2.5 [97]: Ce3 ‘ (aq) + H2O2 -> Ce(H20 2)3 ‘ (aq) 2-3 For pH > 2.5. the cerium cation can be oxidized progressively from the III to the IV valence state with a solution color change from colorless to golden yellow [97] due to the reaction [90]: 2Ce3' * 20H~ + H2O2 -> 2Ce(HO)2: ' 2-4 For a metallic alloy in contact with the cerating solution, reduction of H2O2 occurs and pH increases at cathodic sites: H2O2 + 2e -> 20FT 2-5 which causes precipitation o f insoluble CeC> 2 at cathodic sites[90]: Ce(HO)2: ~ + 20HT -► C e02 + H20 2-6 For the cerium peroxo complex species, the following precipitation reactions may occur [97]: Ce(H2 02)3* -> CefHO*)2* + H~ 2-7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2Ce(H02)2' ~ H2 0 z -> 2Ce(0-): " ' 2H2 0 2-8 CefO?)2 - 2e — > CeO; 2-9 During the cerating process, hydrogen gas may be produced at cathodic sites due to the low solution pH. Generation o f oxygen gas can occur due to the oxidation o f H2O2 at anodic sites [97]: or due to the catalytic decomposition of H2O2 on the metal or alloy surface: The dissolution o f the metallic alloy occurs at anodic sites providing the electrons for the cathodic reactions, which may cause corrosion of the substrate to an extent that depends on the formation rate o f the cerium oxide film. Hughes and co-workers [96] analyzed the surface chemistry with X-ray photoelectron spectroscopy (XPS) and observed the morphology with SEM o f a cerated film, which was formed on AA2024 by immersion in 10 g/L CeCl? and 0.3 % (v/v) H2O2 solution with pH 1.92 for 10 minutes at 43°C. They found that the cerated film was hydrated cerium oxide, that the composition o f the film depended on pretreatment and cerated time, and that nearly all cerium existed in the IV valence state in the cerated film. The morphology o f the cerated film was non-uniform and full o f cracks. Therefore, a sealing treatment was necessary after cerating. H2O2 -► 0 2 + 2H' + 2e 2-10 2 H2O2 -► 2H;0 - 0 ; 2-11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 The color o f the cerated film was golden yellow. Its corrosion protection quality depended strongly on the pretreatment and sealing treatments, aging o f the cerating solution and the cerating process parameters, such as concentrations o f the solution components. pH. temperature and cerating time and additives as well as the nature o f the metallic substrate [94.95.97], Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 3. EXPERIMENTAL TECHNIQUES The experimental techniques employed in this thesis include electrochemical techniques, surface analysis techniques and factorial design. The electrochemical techniques include polarization curves, electrochemical impedance spectroscopy (EIS) and galvanic corrosion measurements. The surface analysis techniques are scanning electron microscopy (SEM). energy dispersion X-ray spectroscopy (EDS) and Auger electron spectroscopy (AES). Factorial design or design o f experiment (DOE) is a powerful and efficient method to identify the important parameters and their optimum ranges. 3.1 Electrochemical techniques Mixed potential theory is the corner stone for the electrochemical techniques for corrosion study [131.132], For an electrochemical corrosion system at equilibrium or steady state, the anodic reaction rate, or anodic current density i a equals the cathodic reaction rate, or cathodic current density /c which defines the corrosion current density i COr r o f the system. The OCP o f the system is the compromise or mixed potential between the equilibrium potentials o f the anodic and cathodic reactions. When the corrosion system is under activation polarization, the net current density i is the sum o f the anodic and cathodic current densities, and can be expressed as the Butler-Volmer equation [133,134]: / = /corr[exp(2.303AE/ba) - exp(- 2.303AE/bc)] 3-1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. where AE = E - E»n- is called overpotential and E is the applied potential. ba and bc are the anodic and cathodic Tafel slopes, determined by the electrode material, reaction mechanism and temperature. 3.1.1 Polarization techniques Polarization techniques are commonly used to obtain the polarization curves o f the studied corrosion system through external application o f potentials or currents away from Ecotr or Un-. They can be classified as potentiodynamic technique if the potential is changed continuously at a constant rate or potentiostatic technique if the potential is fixed at a constant value. Details concerning the measurements with these techniques are provided in ASTM standard G5-94 [135], 3 .1. 1.1 Potentiodynamic technique The polarization curves obtained by this technique are usually drawn in the format o f potential vs. logarithm o f current density according to Eq. 3-1. They can provide the characteristics o f the corrosion system in a wide potential range, such as reaction mechanisms, passivity or stability o f the electrode material [136], Fitting o f the experimental data to Eq. 3-1 can be used to obtain the electrochemical parameters ba, bc and w in the non-Tafel region close to E**,. o f the polarization curve, usually E ^ ± 30 mV [137,138]. The value o f i ^ can also be determined from Rp. which is defined as the slope o f potential vs. current density at E ot- o f the polarization curve, i. e., Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. through the Stem-Geary equation [139]: _ b u b B 2 . 3 0 3 - A )~ Rp where B = ba-bc/2.303(ba - + - bc) [140], Eq. 3-3 is very useful in corrosion study. Since B is usually a constant for a corrosion system, w can be determined from Rp. Eq. 3- l and 3-3 and potentiodynamic technique are the basis o f the polarization resistance technique. Pros and cons of this technique have been discussed by Mansfeld [ 140- 142], ico rr can be related to corrosion rate r^n- through Faraday's Law if uniform corrosion occurs and the nature o f the anodic reaction is known [ 143]: i -W = 3.27-*”- - 3-4 // • D . orr where is given in pm/vear. i ^ in jjiA/cm2. W is the atomic weight, in g/mole and D is the density o f electrode in g/cm3. The potentiodynamic technique has been widely used in inhibitor selection and studies o f inhibition mechanisms. The inhibition efficiency E is defined as: E = (rcn.” - r ^ / r j = I - r ^ / r ^ 3 3-5 where r^ o n -0 and rC D n - in h are the corrosion rate in the absence and presence o f inhibitor, respectively. If B in Eq. 3-3 is the same for a corrosion system with and without inhibitor, then the inhibition efficiency E can be expressed in terms o f Rp: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 E = I- Rp'7Rpin h 3-6 3.1.1.2 Potentiostatic technique The potentiostatic technique, also called constant potential test or bulk electrolysis test, allows a more detailed and accurate study o f a corrosion system than the faster potentiodynamic technique. In a potentiostatic test, a constant potential is applied, and the corresponding current or current density is recorded versus time. Potentiostatic technique can be used in the measurement o f steady state current densities, estimation o f passivity and evaluation o f materials long-term stability at potentials o f interest. Fig. 3.1 illustrates the current density vs. time curves at different applied potentials for a passive system [144.145]. If the applied potential is below the pitting potential Ep rt, the current density drops fast to values comparable to the passive current densities and no pits will initiate. If applied potential is above Epa, the current density first decreases and then increases after some pit initiation time. The higher the applied potential above Ep*, the shorter is the initiation time and the larger is the current density due to pit initiation and propagation. 3.1.2 Galvanic Current Measurement When two dissimilar metals immersed in an electrolyte are connected electrically, galvanic corrosion occurs. In a galvanic corrosion system, the electronegative or active metal becomes the anode (A) and corrodes preferentially, while the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 E » E p ,t Rapid initiation and propagation of new pits E > EpU New pits initiate and propagate co Z uU E E pt t No initiation o f new pits a TIME Fig. 3 .1 Potentiostatic tests for the study of materials passivity and stability. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. electropositive or noble metal becomes the cathode (C) and is protected. According to mixed-potential theory, when the galvanic corrosion system reaches steady state, the galvanic corrosion potential Es falls between the uncoupled anode corrosion potential E ^ . A and the uncoupled cathode corrosion potential E ^ .c , as shown in Fig. 3.2. The current lg flowing between the galvanic couple measures the galvanic attack rate o f the anode. To correctly measure the Eg and lg o f a galvanic corrosion system, a zero resistance ammeter (ZRA) is usually connected between the anode and the cathode. In a modem electrochemical-testing instrument, the ZRA can be realized with a potentiostat by applying zero potential between the dissimilar metals allowing the Eg and Ig to be recorded simultaneously versus time. 3.1.3 Electrochemical impedance spectroscopy (EIS) Compared to previous electrochemical techniques, electrochemical impedance spectroscopy (EIS) is an alternating current (AC), non-destructive technique. EIS allows the long-term study o f a corrosion system with regards to its corrosion behavior or stability, determination o f corrosion rate time laws and prediction o f system lifetimes. In an EIS test, an alternating, sinusoidal potential or current signal is applied to a corrosion system, and the output signals, such as impedance and phase angle are recorded in a wide frequency range, usually between 100 kHz to 10 mHz or I mHz, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 ■corr. C •corr. A loe I w Fig. 3 2 Schematic illustration of galvanic corrosion potential Eg and current Ig. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 8 depending on the system and information sought. The selection of the applied signal amplitude depends on the system impedance. The higher the system impedance, the larger the applied signal amplitude, provided that the system remains linear. The output signals can be displayed in a Bode format in which the logarithm o f impedance modulus |Z| and phase angle < t> are plotted versus the logarithm o f frequency f, i.e. log |Z| vs. log f and < D vs. log f; or in a complex plane format in which the negative imaginary part o f the impedance is plotted against the real part. Equivalent circuits (EC) are used to interpret impedance spectra. In an EC. various electrical components, such as resistance, capacitance. Warburg, constant phase element (CPE) are combined to describe the electrochemical phenomena occurring at the electrode / electrolyte interface. The simplest EC. also called one-time-constant (OTC) model, is shown in Fig. 3 .3(a), in which R, is the electrolyte solution resistance. Rp accounts for the polarization resistance and C represents the capacitance o f the electrode / electrolyte interface. This EC can be used to describe corrosion systems under charge transfer control. The impedance Z of this EC is a function o f the frequency/of the applied signal: A, Z = R + p - 3-7 I + jeoCRp with co = 2 ;/. and j = V -T The Bode plot and complex plane for this EC are schematically shown in Fig. 3.3(b) and (c), respectively. The Bode plot shows clearly the characteristics o f the EC. such as R* in the high-frequency region. Rp in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 9 R p 1 = 3 (a) * •t I a 0.9 -90 -80 -70 -60 -50 -40 -30 -20 -10 ; ■ ;• 0 0 1 2 log(f(Hz)) (b) -500 ^ -400 i -300 * 3- -200 -100 0 t 0 200 400 600 800 1000 1200 T (Q * cm2 ) (C) Fig. 3.3 (a) One-time-constant (OCT) model, (b) Bode plot and (c) complex plane plot for E C = 20 Q -cm \ Rp = 103 Q-cm2 and C = 10-4 F/cm2. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase A n g le (Degree) 30 the low-frequency region and capacitive intermediate frequency region with a slope o f-I. In many systems, deviations from the ideal plot frequently exist, with the absolute value o f the slope less than I and a smaller maximum absolute value o f the accommodate these deviations, an exponent a was introduced to the impedance Eq. 3-7 by Mansfeld and coworkers [146-149], with - I < a < 0. Fitting of the impedance spectra with software, such as AN ALE IS [150] allows determination of the electrical components. Rp is the polarization resistance, and capacitance C can be related to electrode / electrolyte interface parameters through: while s and d are the dielectric constant and thickness o f an oxide film on the electrode or the double layer at the electrode / electrolyte interface, respectively. The constant phase element (CPE), a general diffusion related element, has also been used in ECs to account for the non-ideal impedance spectra. The admittance representation o f a CPE is given by: phase angle in the Bode plot, and a depressed semicircle in the complex plane. To 3-8 g „-e-A d 3-9 where &, = 8.85x 10'1 4 F/cm is the permittivity o f free space. A is the electrode area. Y (©) = Y0 (jo)" 3-10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 1 which is a very generalized formula with an exponent 0 < n < 1. CPE can be related to other EC elements for different values o f n. When n = 0, CPE is a resistance, with R . = Y0' 1 , when n = I, a capacitance with C = Y0 and when = 0.5, a Warburg impedance. For interpretation o f EIS data. CPE usually represents capacitive elements when n is close to I. and diffusion-related. Warburg-like elements when n is close to 0.5. .An EC with two CPEs and its theoretical Bode plot and complex plane are shown in Fig. 3.4. In the EC. CPE1 is a capacitive element (e.g. nl = 1). while CPE2 is a diffusion-related element (e.g. n2 = 0.6). When CPE I is a capacitance (nl = 1). and CPE2 is a Warburg impedance (n2 = 0.5). the EC becomes the Randles Circuit. Fining o f impedance spectra with this EC gives R *. Rp and two parameters Y0 and n for each CPE. For a capacitive CPE, the true capacitance C can be determined from the Y( ) and n values through the following equation first reported by Von Westing [151]: sin(«/r/ 2) at ©m, the maximal absolute phase angle appears. However. Hsu and Mansfeld [152] demonstrated that Eq. 3-11 is incorrect and had derived an equation between C. n and Yo. C = 3-12 where © m is the frequency at which the imaginary impedance Z" has a maximum. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A CPE1 — a n — Rp CPE2 C Z J (a) -90 --n.= 0 9. n.=0 6 n,= 1.rt,= 0 5 -80 4 9 r > % -70 ® - 5 ' ~ cb E ^ -60 ® » * « • Q ~ 3 " * - - 50 • ^ « • - -40 ? ~ - - O - «r o ^ _ - 2 -j - ~ 2 o ' d -20 v _ -r— « s > - - y - •? 1 , T i> < ?■ < »■ -» • > ■ 0 -3 -2 -1 0 1 2 3 4 5 log(f(H2)) -30 ® -20 -10 -8000 £ -4000 o * = -2000 N (b) C O ' 0 2000 4000 6000 Z (Q * cnr) (C) Fig. 3.4 (a) A two-time-constant (TTC) model, (b) Bode plot and (c) complex plane for R« = 20 Q-cm2 . Rp = 10J Q-cm2 . Y0 (CPEl) = 10-1 S/cm2 and Y0 (CPE2) = 10'3 S/cm2. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 3 3.2 Surface analysis techniques SEM. EDS and AES are the most frequently used surface analysis techniques in corrosion studies. These techniques allow detailed surface investigation o f materials after interaction with electrolytes to aid in the interpretation o f electrochemical data and to facilitate the understanding o f corrosion processes and mechanisms. SEM and EDS are usually conducted together for surface morphology observation and chemical composition determination in near surface depth. AES is suitable to determine surface chemistry at very thin depth, about one monolayer into the surface layer, and to obtain concentration profiles versus depth with Ar ion milling [82], 3 .3 Factorial design o f experiments and optimization of process parameters In a factorial design of experiments, the variables whose influence on a particular quantity is being investigated are referred to as factors. The values of these factors that are set for each experiment are called levels. For a study o f the influence o f n factors at m levels, an mn factorial design is employed, where m" indicates the number o f experiments required. The two-level (m = 2) factorial design is commonly used for systems with a small number o f variables. Although these designs only explore a small region o f the response surface, they can determine a promising direction for further exploration by analysis o f the main and interaction effects o f treatments. Main effects measure the influence that a particular variable has on the response, while interaction effects measure the influence that a particular combination o f variables has on the response. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 4 To explore a large response surface, a three or higher-level factorial design is needed. The response equation, which is a function o f variables, can be obtained via least-square methods. Detailed descriptions o f the factorial design o f experiments can be found in many references [153-155]. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 5 4. EXPERIMENTAL APPROACH 4.1 Materials and pretreatments Materials employed in this study were two construction mild steels, i.e. AISI-SAE 1018 steel hot-rolled sheet with a thickness o f I mm and AISI-SAE 1020 steel hot- rolled rods with a diameter o f 6.35 mm (1/4 inches). The 1018 steel sheets were cut into 7 cm < 7 cm or 3 .5 cm * 7 cm panel specimens for tests conducted at room temperature (RT). The 1020 steel rods were cut into specimens o f the length o f 30 cm for tests conducted at 100°C. Before each experiment or treatment, specimens were pretreated according to following procedures: • Degreasing in an Alconox detergent solution followed by rinsing with tap water. • Polishing with SiC sand papers to at least grit 400 followed by rinsing with tap water. • Brushing in an Alconox detergent solution followed by thorough rinsing with DI water. • Pickling in a 1:1 volume ratio solution o f hydrochloric acid (36.5 - 38.0%) and DI water for 30 seconds, followed by thorough rinsing with DI water. • Immersion in dehydrated alcohol, then drying with compressed air at RT. • Storing in a desiccator for further tests, if not used immediately. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.2 Electrochemical measurements EIS and DC polarization methods were carried out with three-electrode and two- electrode cell configurations. EIS and DC polarization data were analyzed with software developed at CEEL/USC [138,150]. 4.2.1 Electrochemical cells For EIS tests, the three-electrode cell was employed for the evaluation o f REMSs in neutral aerated 0.1 N NaCl at RT The working electrode, i.e.. the 1018 steel panel specimen of the size 7 cm < 7 cm. was placed horizontally with an exposed area o f 20 cm". The counter electrode was a 304 stainless steel panel, and the reference electrode was a saturated calomel electrode (SCE). The two-electrode cell consisting o f two identical, vertically placed specimens was used for evaluation o f REMSs in the baseline solution at RT and at 100C C. The specimens for RT tests were 1018 steel panel specimens o f the size 3 .5 cm < 7 cm with an exposed area o f 4.5 cm" for each specimen and an effective exposure area of 2.25 cm*. For the I00°C tests, the specimens were 1020 steel rods with an immersed area o f 6 cm" for each specimen and an effective area o f 3 cm". The temperature o f the test solution was controlled within 100 ± 2°C with a temperature controller (Model CN-A8005M from Thermolvne Co ). A PYREX brand Graham condenser with a spiral inner condensing tube that has a high condensing surface Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 7 area per unit length o f jacket was used to prevent the ammonia gas inside the cell from escaping during the experiments. For DC polarization tests, a three-electrode cell configuration was employed in which the working electrode, i.e.. the mild steel specimen with the size o f 3.5 cm x 7 cm. was vertically placed in the cell. The counter electrode was a 304 stainless steel panel and the reference electrode was the saturated calomel electrode (SCE) placed about 3 mm from the working electrode via a Luggin capillary containing saturated KCl solution to reduce IR drop. The exposed areas o f the working and counter electrodes were 4.5 cm" and 20 cm2, respectively. 4.2.2 Electrochemical testing methods EIS tests were carried out to determine the time dependence o f Rp and C. DC polarization tests were used to obtain electrochemical kinetic parameters for mechanistic investigations. 4.2.2.1 EIS The impedance measurements were conducted either with a Zahner universal electrochemical interface driven by the Thales IM6 impedance measurement program or a Gamry Instruments electrochemical measurement system driven by the software EIS300. The amplitude o f the applied sinusoidal AC signal was 5 mV in a frequency range from 5 KHz or I MHz to 5 mHz. All impedance data were displayed in Bode format and analyzed with the software ANALEIS developed at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 8 CEEL / USC according to equivalent circuits containing one time constant or two time constants to obtain Rp, which can be converted into the corrosion rate according to Eq. 3-3 and 3-4. For uniform corrosion o f iron or steel with the anodic corrosion reaction: Fe — > Fe: - 2e 4-1 the Stem-Geary constant B is about 20 mV [140], and the equivalent weight EW equals 27.9. The density p equals 7.8 g/cmJ for mild steel. The relationship between the corrosion rate ro^- (in jim/year) and the Rp (in Q-cm") for mild steel is: l/v5 B « EW 20 . 27 9 2.34 « 10* r , = j.27< 10 <----------- = j .27 <-------------= -------------- 4-2 p<Rp l%<Rp Rp The dependence o f r ^ (in pm/year) on exposure time usually can be fitted to the time law: r„„ = A * t a' 4-3 or in a bi-logarithmic form: loS(r^ ) = l°g (Ar) + Br Iog(/) 4-4 in which Ar is the corrosion rate o f mild steel at t = I, i.e. after one hour, day or year depending on the time units used for the analysis. The fit parameter Br describes the time dependence o f corrosion rates and is related to the prevailing corrosion mechanism. Eq. 4-2 can also be written as: lo g (r_ ) = 5 .3 7 -lo g (-^ -) 4-5 K p Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 therefore, combining Eq. 4-4 and 4-5. we have: Iog(l/R p ) = log(4r )-5 .3 7 + Br Iog(/) 4-6 = a R +bR log(/) in which aR = log(.\-) - 5 .37 and bR = Br. aR is the logarithm o f the inverse o f Rp when t is unity. bR has the same significance as Br. The smaller the values o f aR and bR are. the lower is the corrosion rate. Similar time laws as in Eq. 4-3 to 4-6 were observed by many researchers [156. 157], They suggest that transport o f reactants through a growing corrosion product layer or a protective film determines the rate o f the corrosion process, if the corrosion product layer or the protective film is porous. In these cases. Br and bR equal or approach - 0 .5. If the corrosion product layer or the protective film is compact and difficult for the reactants to penetrate. Br and bR will be close to - I But removal o f film from the surface due to dissolution or bad adhesion will lead to larger Br and bR toward 0. These time laws have been used extensively in studies o f atmospheric corrosion with the purpose o f extrapolating results from short term tests (i.e. a few years) to very long times (i.e. 20 years or more) [158]. This approach can be used for lifetime prediction o f structural materials exposed to different corrosive atmospheres. Corrosion rate time law can be obtained through equations 4-3 to 4-6. The time dependence o f the C o f the corrosion product layer or the protective film on a steel surface can also be described with equation similar to 4-3: C = Ac x t s‘ 4-7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 0 According to Eq. 3-9, C depends on the area A. dielectric constant e and thickness d o f the corrosion product layer or the protective film. In cases where the area A and the dielectric constant s remain relatively unchanged. l/C is proportional to d. The time dependence can be described with the following equation: log( I / O = a ,. ~ hr x log(/) 4-8 However, in most cases, A and e are not constant during corrosion process due to changes in surface roughness and changes in structure and composition o f corrosion product or protective film. Eq. 4-7 does not simply reflect the time dependence o f d. but rather the time dependence of d. A and e svnergisticallv. 4.2.2.2 DC polarization technique The DC polarization experiments were conducted with a Zahner universal electrochemical interface controlled via l/E software or a Gamry Instruments electrochemical measurement system controlled by the software DC 105 The scan rate for both the anodic potentiodynamic or cyclic polarization and cathodic polarization tests was 0.17 mV/second (0.6 V/hour) according to the ASTM standard G5 [135]. The anodic scan was started 30 mV negative o f the open-circuit potential E con- and reversed when a current density o f 2 mA/cm* was reached according to ASTM G 61 [143], The scan was continued in the negative direction and stopped at a potential a few hundred mV negative o f the new value o f Ec^r. The cathodic scan was started at E ^ and was stopped after the limiting current density region had been reached. The polarization data were analyzed with POLFIT developed at CEELI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 1 USC [138], The program outputs include the electrochemical parameters Rp. b3 , bc , icons ico n -, E co n - and B along with their corresponding error terms as well as statistical evidence o f the fit quality. 4.3 Surface analysis SEM and EDS techniques were employed to obtain surface morphology and chemistry information on mild steel after surface modifications, which included cerating, yttrating, and different sealing processes. SEM and EDS were performed with a Cambridge Model Stereoscan 360 and a Link Analytical Model 1000 Analyzer, respectively, at the Center for Electron Microscopy and Microanalvsis (CEMMA) o f USC. The system was operated at electron beam energy of 10 to 15 Kev with a beam current o f about 200 pA for SEM and 2 nA for EDS. The working distance was about 10 mm for SEM. and around 20 mm for EDS. 4.4 Evaluation o f cerium salts In order to evaluate cerium salts as chromate alternatives for the corrosion protection o f mild steel, different cerium salts were investigated in four different applications, i.e.. as inhibitors, as chemical conversion agents, as electrodeposition agents and as cerating agents. Their effects were first studied with EIS in aerated 0.1 N NaCl at RT in an attempt to choose promising cerium salts and applications providing corrosion protection comparable or better than that provided by chromate. The promising Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 2 cerium salts and applications were further investigated and compared with chromate in the baseline solution at RT and at 100 °C. 4.4.1 Tests in 0 .1 N NaCl Aerated NaCl is a corrosive medium to mild steel. The corrosion behavior o f mild steel in NaCl is well known, therefore it was used here as the screening solution for REMSs. 4.4.1 .1 Cerium salts as inhibitors Considering the high corrosivity o f Cl- for mild steel at high temperatures and the low solubility of Ce^SO-tb in aqueous solution, only CefNCbb and Ce(CH?COO)? at a concentration o f 20 mM were tested as inhibitors in 0 .1 N NaCl. Due to the acidifying effect o f CefNCbb. EIS measurements were conducted in 0. IN NaCl with the natural pH = 2.5 and neutral pH = 7.0 adjusted with 1% sodium borate fNazB-tOT-lOHiO) solution. 4.4.1.2 Chemical conversion coatings CefCHjCOOh. CeClj and CefNCbb were used to produce chemical conversion coatings by immersing 1018 steel specimens in 20 mM cerium salts solutions for 2 hours at RT. Ce(CH3 COOb did not dissolve completely and the color of the solution was slightly white. The pH values o f CefCHjCOOb- CeCl? and Ce(NCbb solutions were 6.7. 5.0 and 2.5. respectively. The conversion coatings were evaluated in 0 .1 N NaCl with EIS at RT. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 3 4.4.1.3 Electrodeposition Electrodeposition o f a cerium containing coating was carried out in 0 .1 M Ce(N0 3 ) 3 at 3 mA/cnT for 300 seconds and 900 seconds. The cathodic polarization curve of mild steel in 0 .1 M CefNCbb is shown in Fig. 4. la. The change o f the applied cathodic potential with time during galvanostatic polarization is given in Fig. 4. lb. After the treatment, a light yellow layer was observed on the 1018 steel samples. 4.4.1.4 Cerating process The cerating process for the 1018 and 1020 steels consisted o f the following steps: a. Surface pretreatment as described in section 4.1.1 b. Immersion in 10 g/L CeCl3-7H20 ■ + • 5 wt% H2O2 at RT for 20 minutes. c. Rinsing thoroughly with DI water, then drying with compressed air at RT. 4.4.2 Tests in the baseline solution at RT Based on the results in 0 .1 N NaCl. EIS measurements were performed for untreated and treated 1018 steel specimens for 7 days in the baseline solution. For comparison, EIS tests were also conducted on untreated 1018 steel specimen for 7 days in the baseline solution inhibited with 2.4wt% Na2Cr2< > 7. 4.4.3 Tests in the baseline solution at 100°C Cerium nitrate was used as inhibitor at a concentration o f 5 mM and cerium chloride was used in the cerating treatment in these tests. Their protection efficiencies for 1020 steel specimens were evaluated with EIS in the baseline solution at I00oC. For Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 4 ■ 0 S C 0 70 0 80 ^ 0 90 LU O ,0 0 e f t I 10 > o C O \ I 2 0 — \ * . - Catnccic polarization curve ter 1C2C steel in 0 I M Ce(C03 solution : -ic - \ i I so — ' I 60 — ------------------------------------------------------ -------------------- log(i CA/cm^)) (a) 2 5C I SC I 70 U J oo _ ' Jo < i 1 30 - OOGc-a 1 90 Q 30 60 90 120 Deposition time (second) (b) Fig. 4.1 Cathodic polarization curve (a) and galvanostatic curve at the cathodic current density o f 3 raA/cm2 (b) for mild steel in 0 .1 M CefNChh solution at RT. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 5 comparison, impedance spectra were also collected at 100°C for two days in 5wt% NHj, baseline solution, and baseline solutions inhibited with 2.4 wt% Na^C^O? and 1% sodium (meta) silicate (Na2Si0 3 -9 Hi0 ). respectively. 4.5 Optimization o f the cerating process Cerating is a multi-factor dependent process. The optimum conditions o f the different factors were determined through factorial design or single-variable experimental methods. 4.5.1 Optimization o f process parameters through factorial design In the cerating treatment, three process parameters were found to be important, i.e. the concentrations o f cerium chloride and hydrogen peroxide and the cerating time. In order to optimize the cerating process to give high Rp and low C. i.e.. the responses for the cerated layer, a factorial design o f experiments was conducted which employed three factors (concentration o f cerium chloride, concentration o f hydrogen peroxide and cerating time, labeled as factors FI. F2 and F3, respectively) at two levels (2Jdesign). The 8 experiments for the 2J factorial design o f cerating treatment are displayed in Table 4 .1. The responses of Rp and C were obtained through EIS tests in the baseline solution at I00°C for 48 hours. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 Table 4. 1 23 factorial design for optimization of the cerating treatment. Experiment U r r Factors CeCl3 (g/L) (FI) H2O2 (wt%) (F2) Time (minutes) (F3) 7.5 12.5 2.5 7.5 10 20 23-l < < < 23-2 X X X 23-3 X X 23-4 < < < 23-5 < .< < 23-6 < < 23-7 < < X 23-8 < < Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 7 4.5 .2 Investigation o f pretreatment processes Before the cerating treatment, the mild steel samples were pretreated to degrease the surfaces, to remove surface oxides and to produce active surfaces. The effects o f the following three pretreatment processes on the corrosion characteristics o f cerated mild steel were investigated: a. Immersion in a 2 volume % Micro solution (International Products Corporation) at 66°C for 10 minutes. The Micro solution consisting o f mainly organic acids was phosphate-ffee and biodegradable. This process was labeled as Pre-Treat 1 . b. Immersion in I: I volume ratio o f hydrochloric acid (36.5-38%) and DI water for 60 seconds at room temperature. This process was labeled as Pre-Treat2. c. Immersion in I20g/L ACTANE* 345 solution (Enthone-OMI Inc.) for 90 seconds at R.T. ACTANE* 345contains fluoride and surface-active agents and is effective in removing tenacious oxides, smut and scale. This process was labeled as Pre-Treat3. 4.5 .3 Investigation o f post-treatment processes (sealing) Since the cerium conversion coating has a cracked mud-like crack structure, a post- cerating sealing treatment is necessary. Cathodic polarization in cerium salts, such as CefN0 3 > 3 after the cerating treatment can deposit a cerium oxide/hvdroxide layer (Ce(OH)3) on the cerated mild steel, but this layer is not compact [52]. Therefore, four additional post treatment processes were studied. a. Immersion in 10% sodium silicate (meta) (Na2SiC>3 * 9 H2O) solution at 50°C for 30 minutes. This process was labeled as Post- Treat I. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 8 b. Immersion in 10% (3-glycerophosphate disodium salt (C3H7O6 PNa2‘ 9 H2O) at 50°C for 30 minutes. This process was labeled as Post-Treat2. c. Immersion in 1% sodium borate (Na2B4 0 7 - 10H2O) + 1% cerium nitrate (CeCNChh' 6 H2O) at 100°C for 60 minutes. This process was labeled as Post-Treat3. d. Immersion in 10% sodium molybdate (Na2Mo0 4 * 2 H2O) solution at 50°C for 30 minutes. This process was labeled as Post- Treat^ These sealing solutions contained inorganic compounds such as sodium silicate, sodium molybdate. sodium borate, glycerophosphate sodium salts and cerium nitrate that are effective as inhibitors for mild steel. These inorganic compounds, especially sodium silicate and cerium nitrate have limited solubility in ammonia-water solution. The purpose o f sealing treatments with these compounds was not only to seal the cracks, but also to incorporate the inorganic compounds into the cerated film in order to make the film more compact as well as to form a reservoir o f inorganic inhibitors. 4.6 Evaluation o f Yttrium Salts Yttrium chloride (YCI3) was employed in the surface modification o f mild steel ("vttrating") at room temperature. The vttrating procedures were similar to those used in cerating. The yttrating parameters were 12.5 g/L YCI3,2.5 wt % H2O2 with an immersion time o f 2 0 minutes at room temperature. The corrosion resistance o f yttrated layer on mild steel was assessed in 0. IN NaCl at room temperature (RT). in the baseline solution at RT and at 100°C by the EIS technique for one week. The morphology o f the yttrated layer was studied with SEM. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 9 Yttrium sulfate Y^SO-iri was evaluated as a corrosion inhibitor o f mild steel at a concentration of ImM in the baseline solution at 100°C with EIS for one week. 4.7 Evaluation o f Organic Inhibitors Four organic inhibitors were evaluated in the baseline solution at RT: L-aspartic acid (AA) sodium salt with monohydrate (CtHfiN'O^Na-HiO) at a concentration o f 10 mM, 3-glycerophosphate acid (GPH) di-sodium salt with four and one half hydrate (CjHTOfePNai^.SHsO) at a concentration o f 50 mM. poly-i-aspartic acid (PAA) sodium salt (molecular weight 5000 to 15000) at a concentration o f 20 ppm and the poly- t.-glutamic acid (PGA) sodium salt (molecular weight about 1000) at a concentration o f 20 ppm. All organic chemicals were purchased from the Sigma Chemical Co. The inhibition efficiency of the organic additives was determined with EIS and the DC polarization techniques. EIS was used to monitor the time dependence o f the inhibition efficiency for two weeks. The DC polarization technique, which included anodic and cathodic polarization, was used to determine the inhibition mechanisms o f the organic additives. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 0 5. EXPERIMENTAL RESULTS 5. 1 Evaluation o f cerium salts The corrosion protection capabilities o f cerium salts for mild steel were first evaluated in aerated 0.1 N NaCl at RT using different cerium oxide coating methods. Promising cerium salts and coating methods were further investigated in the baseline solution at RT and 100°C. 5.1.1 Evaluation in 0 . 1 N NaCl Cerium (hydr)oxide coatings can be formed on mild steel through addition o f cerium salts as inhibitors in 0 I N NaCl. immersion (chemical conversion), electrodeposition and cerating (fast chemical conversion) in different cerium salt solutions. The corrosion protection efficiency o f cerium (hydr)oxide coatings on mild steel in 0 .1 N NaCl (open to air) at RT was extensively and systematically studied with EIS. polarization curves. SEM and EDS 5.1.1. 1 Cerium salts as inhibitors According to the literature [66.67], o f all the studied REM cations, the Ce (m ) cation provided the best degree o f protection against uniform corrosion for A17075 in NaCl. The ranking o f the REM cations in degree o f protection was: Ce > P r> Nd > La > Y 5-1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 O f the studied cerium (III) salts, cerous acetate gave the highest inhibition efficiency for A17075 in NaCl [54], The ranking of the REM salt anions in inhibition efficiency was: c h 3c o o ' > c i o t > s o 4 :' > c r 5 - 2 NO3' was not in the list, perhaps due to its strong acidifying effect. Ce(NCbb and CefCHjCOOb were tested as inhibitors at a concentration of 2 0 mM in 0 . 1 N NaCl at RT. Due to the strong acidifying ability of Ce(NCbb. the effect of pH on the inhibition efficiency o f Ce(NChb was studied at the natural pH 2.5 and pH adjusted to 7.0 with 1% sodium borate fNa2B4O7* 1 0 H2O) solution. For comparison. 1018 steel was also tested in 0 .1 N NaCl without and with 2 0 mM Na2Cr2<>7 Fig. 5.1 shows the impedance spectra for 1018 steel specimens exposed to five different solutions for 2 hours at RT Table 5. 1 lists the fit parameters Rp and C values as well as the corrosion rates r ^ and inhibition efficiencies calculated according to (4-2) and (3-6) respectively. Due to low solution pFI2.5. Ce(NChb provided practically no corrosion protection for 1018 mild steel with inhibition efficiency- only 7%. But when the pH was neutral. Ce(NCbb brought about a high inhibition efficiency o f 93%. higher than that o f Na2Cr2 0 r (84%). Though pH was near neutral (6.7). Ce(CHjCOOb did not protect mild steel satisfactorily with an inhibition efficiency o f only 41%. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 ♦ G 3 iiL 00 o Blank 20 mM Na2Cr207 0 - - 20 mM Ce(N0 3 )3, PH2.5 - 0 — 20 mM Ce(N0 3 )3 , PH7.0 A — 20 mM Ce(CH3COO)3 ' ^ ^ eo eeeo seeso a eceeecc C J 2 = o u < 2 a -90 -75 -60 2 (b) -45 -15 V 0 log(f (Hz)) Fig. 5 .1 Bode plots o f 1018 steel in blank and inhibited 0.1 N NaCl for 2 hours at RT; (a) impedance, (b) phase angle. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5 . 1 Rp , C, rco rr and inhibition efficiency values for 1018 steel in NaCl for hours. ! j Inhibitor ! , ! R p (H-cnr) j ! ■ C (F/cm:) Ic o r r (pm/year) Inhibition efficiency (%) | Blank l.32x l03 ! 1.42x 1 O '3 177 N/A 1 20 mM Na^Cr^O? ! 8.34* 103 1 5.64 * I O '4 28 85 ! 20 mM Ce{N03)3. pH 2.5 i 1.42 < 103 6.76* 10" * 165 7 1 20 mM Ce(N03)3. pH 7.0 . 1.77* 10 ! 9 42 * iO -4 13 93 1 20 mM Ce(CH3COO)3, |. pH 6.7 ■ 7 n v in 3 “ l° .1 2.72 * I O '4 105 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 4 After 2 hours, the specimen exposed to blank NaCl was covered with black corrosion products, especially at the deep immersion part where the oxygen concentration was low. The exposed area in NaCl with 20 mM CefNChh at pH 2.5 was coved with a blue-colored film, however uniform corrosion and pits were observed at the deep immersion part. When the solution pH was adjusted to neutral, a fairly uniform blue- colored film was formed on the exposed area. The blue-colored interference film was also observed on mild steel by Hinton et a/. [54] in CeCIj THsO inhibited soft tap water. The blue-colored film deteriorated at the deep immersion area with prolonged exposure time resulting in a decrease o f Rp, from 1.77 < l04 Q-cm: after 2 hours to 8.87 • < 103 Q-cm" after 24 hours as illustrated in Fig. 5.2. After 2 days, a second time constant appeared at the intermediate frequency range, which was believed to be associated with the breakdown o f the film at the deep immersion area. Visual observation o f the film after 7 days showed that the upper part was covered with a thick yellow layer, while the lower part was covered with a thick brown layer of corrosion products. 5 .1.1.2 Cerium salts used in chemical conversion coatings Applications o f cerium salts in chemical conversion coating were mainly investigated in the localized corrosion prevention o f AI alloys [73-89], Studies o f chemical conversion treatment on mild steel have not been reported in the literature. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 5 4 . 5 - 9 0 4.0 ~ 3.5 E o a 3.0 61 e» 2 2.5 2.0 1.5 2h -75 2d -60 -45 o. -30 -15 -2 1 0 2 1 3 5 4 logf (Hz) Fig. 5.2 Bode plots for 1018 steel exposed to 0 .1 N NaCl - 20 mM Ce(NCh)?. pH 7 0 at RT For 2 days. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase A n g le (degree) 5 6 Fig. 5.3 presents the 2-day impedance spectra o f 1018 steel exposed to 0.1 N NaCl after 2 -hour immersion in 2 0 mM cerium salt solutions o f CefCHjCOOb, CeCl3 and CefNCbb at R.T, respectively. Also shown in Fig. 5.3 for comparison are the impedance spectra o f 1018 steel in blank NaCl and NaCl inhibited with 20 mM Na2Cr:0 . Table 5.2 lists the fit Rp, C values for the spectra and the calculated corrosion rates r ^ and inhibition efficiencies. From Fig. 5 3 and Table 5 .2. it can be seen that after chemical conversion treatments in different cerium salt solutions, corrosion rates decreased to different extents. The inhibition efficiencies of layers formed in Ce(CH3COO )3 and CeCl3 were low. only 43% and 28%. respectively. Immersion in Ce(N03)3 solution yielded the highest inhibition efficiency, which was higher than that for the untreated sample in dichromate-inhibited NaCl. However, the conversion coating did not provide lasting protection in uninhibited 0 . 1 N NaCl as demonstrated in Fig. 5.4. where after an increase between the first and second day, Rp dropped to 4.92x 103 Q-cm2 (48 urn/year) after three days. 5. 1 . 1.3 Cerium salts used in electrodeposition The cathodic polarization curve of mild steel in 0 .1 VI CefNChb at natural pH = 2.5. is shown in Fig. 4. Ia, in which two Tafel regions can be clearly recognized. The first one exists between -0.70 - -0.83 V vs. SCE. corresponding to the evolution of hydrogen gas. The second region occurs in the region o f -0.95 — 1 . 10 V vs. SCE. where reduction o f oxygen and production o f hydroxyl anions take place. In the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 7 * a N 'oi O — Immersion in Ce(N0 3 > 3 -£ •— Immersion in (CH3 COO)3 Ce FT Immersion in Ce<CI) 3 O — Untreated, without chromate -90 -75 Untreated, with chromate -60 -45 -30 -15 < 0 £ o > < D -® o > at C O (0 tf -3 -2 -1 0 1 2 3 4 5 log(f(Hz)) Fig. 5.3 Bode plots o f 1018 steel exposed to 0 I N NaCl for 2 days after different chemical conversion treatments. For comparison, data are also shown for untreated samples exposed to NaCl with and without 20 mM NaiCriO- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 Table 5. 2 Rp, C, rco rr and Inhibition efficiency values for 1018 steel after different chemical conversion treatment. Chemical conversion solution Test solution ^ 2 (Q-cm2 ) C (F/cm2 ) Tcon- (pm/year) Inhibition efficiency (%) 20 mM CeCNOsk 0.1 N NaCl 2.74xl04 3.89X10-4 9 88 20 mM Ce(CH3COO)3 0.1 NNaCI 5.98xl03 9.37X10-4 39 43 20 mM Ce(CI)3, 0.1 N NaCl 4.72XI03 8.48XI0-4 50 28 Untreated 0.1 NNaCI 3.40xl03 2.56x1c4 69 N/A Untreated 0.1 NNaCI + 20 mM Na2Cr207 8.81xl03 3.51 xlO "4 27 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 -90 E o * g O ) o Id O ' 2d - - 3d -75 -60 -45 □ < 5 -30 - - •> -15 2 5 1 3 4 log(f(Hz)) Fig. 5.4 Bode plots of 1018 steel in 0 .1 N NaCl for 3 days after chemical conversion treatment in 2 0 m iV I Ce(NOjh for 2 hours. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase A ngle (degree) 6 0 beginning o f electrodeposition, the cathodic current density was chosen at 30 mA/cm2. However, after 30 seconds deposition at this current density, the cerium hydroxide coating did not adhere to the substrate but cracked into pieces. Therefore, a much smaller current density o f 3 mA/cm' was employed. Fig. 5.5 shows the EIS spectra for 1018 steel samples exposed to 0.1 N NaCl for 24 hours after cathodic polarization in 0 .1 M Ce(NOih at 3 mA/cm2 for 300 and 900 seconds, respectively. The cerium hydroxide coating formed after 300s deposition was not compact. Pitting corrosion was observed on the steel substrate after 2 hours exposure. After 24 hours exposure, the coating lost its integrity and became sediments loosely covering the substrate. The coating formed during 900s deposition had a very high Rp o f 2.4 < 104 Q cm* after 2 hours exposure, but also lost integrity after 24 hours with the Rp dropping to 3 .7 < 103 Q c m l similar to that o f the coating after 300s deposition. 5.1.1.4 Cerium salts used in cerating 1018 steel was cerated in 10 g/L CeClj - 5 wt% H^Ot aqueous solution at RT for 8 minutes, then tested in 0.1 N NaCl at RT for 24 hours. The impedance spectra at 2 hours and 24 hours are shown in Fig. 5 .6 a. The capacitive parts o f the spectra had shifted to lower frequency region suggesting that the capacitance value o f the cerated steel electrode had increased greatly. The spectra cannot be fitted with conventional fitting methods, but can be simulated under the assumption that the phase angle spectra are symmetric along the axis o f the logarithm of frequency. The simulated Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 1 ‘cow 3 * 0 CM 3 - , E - j - 2 . u - ■ - — -90 ----- 300 seconds at 3mA/cm2 ° ----- 900 seconds at 3mA/cm2 . 7 5 -60 • o 5 . 3 ' 0 -45 n T tQ, * O OOO w O ni -30 * - - 2 _ -15 ‘ ‘?o_ -3 -2 - 1 0 1 2 3 4 5 log(f(Hz)) Fig. 5.5 Bode plots o f 1018 steel in 0. IN NaCl for 2 hours at RT after different cathodic polarization processes in 0 .1 M Ce(N0 3 ) 3 at RT. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase A ngle (degree) 6 2 -9 0 E u * SL 3 51 cn o 2 h 24h -75 -60 -45 © © 0 3 © O ) ® 0 1 < D -15 -2 0 1 2 log(f(Hz)) [ n [ij — 0 3 4 5 (a) E o • 3. 3 o > o ♦ V Expenmentat data Slnulanon data: Rs“ 260 U-cm2 Rp= 4000C*cm2 C » 1,8-ii. -F/cnt? •x* 0.87 2 ♦ % -2 -1 0 1 2 log(f(Hz)) -90 -75 -60 -45 -30 -15 0 ® 5s ® 2 . ® O l c < ® m ta (b) Fig. 5.6 1018 steel exposed to 0 .1 N NaCl for 24 hours at RT after cerating in 10 g/L CeCh + 5 wt% H2O2 aqueous solution at RT for 8 minutes: (a) Bode plots, (b) simulated data vs. experimental data 24-hour data. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 3 values showed that cerated steel had low Rp o f 1.5 < 103 Q-cm2 and a very high capacitance of 3 .65 < 10'' F/cm2 after 2 hours exposure, but after 24 hours exposure, Rp had increased to 4.0* 103 Q-cm2 and C had decreased to 1.8 «I O ' 2 F/cm2 The simulated data together with the experimental data for 24h exposure are shown in Fig. 5.6b. After the experiment, the 1018 sample was visually inspected and it was found that some exposed areas were covered with brown corrosion products, while some other areas were still covered with cerated layers with good adhesion to the substrate. A similar phenomenon was observed on one cerated sample after exposure to humid air for 24 hours. The surface morphology and chemical composition o f the cerated layer on 1018 steel were investigated with SEM and EDS. The SEM micrographs (Fig. 5.7) demonstrate that the cerated layer was relatively thick and had a cracked mud-like network. A large number o f globular and cracked mounds or pores with varying size up to 2 0 were distributed randomly in the cerated layer. EDS results in Fig. 5.8 indicate that the mounds (Fig. 5.8a) and flat areas (Fig. 5 8 b) had similar chemical composition. Fe was also found in a large amount which could be due to the formation o f a mixed cerium / iron (hvdr)oxides cerated layer and/or due to the signal from the steel substrate. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 (a) (b) Fig. 5.7 Micrographs o f cerated layer on 1018 steel formed in 10 g/L CeCb + 5 wt% H2O2 aqueous solution at RT for 8 minutes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 9 5.220 keU F 5 .2 2 0 keU 10.3 > Fig. 5.8 EDS results o f the cerated layers on 1018 steel; (a) mounts, (b) flat cracked areas. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 6 Since the cerated layer was full o f cracks and pores, several sealing methods were evaluated in an attempt to improve the performance o f the cerated steel. Sealing by dipping in 100 g/L Na2Cr2C>7 aqueous solution at RT for 5 minutes provided no improvement as evidenced by the resulting impedance spectra that were similar to those in Fig. 5.6a. Sealing by immersion in 0 .1 M CefNChh at RT impaired the cerated layer due to its low solution pH. Sealing by cathodic polarization seemed to improve the corrosion protection provided by the cerated layer on mild steel. Fig. 5.9 shows the impedance spectra for 1018 steel exposed to 0 .1 N NaCl after cerated in 10 g/L CeClj * 5 wt% H2O2 solution at RT for 8 minutes and then sealed by cathodic polarization in 20 mM CefNChb at 3 mA/cm" for 2 minutes at RT. The spectra were quite stable during the 3-day test with relatively high Rp values and low C values (Rp = 3.4*10J Q-cm", C = 5.54 * I O '4 F/cm" at 2 hours, and Rp = 4.2* 103 Q -cm \ C = 5.61 * I O '4 F/cm* at the 3rd day). SEM images (Fig. 5.10) indicate that after cathodic polarization sealing, the cerium (hydr)oxide film had much less, but wider cracks and looked totally different from that o f cerated layer (Fig. 5 .7). This film could have been formed by cathodic polarization on top o f the porous and cracked cerated layer. By comparing Fig. 5.10(b) and Fig 5.7(b), it can be seen that the cracked mounts formed during cerating were sealed by cathodic polarization. EDS analysis, as shown in Fig. 5.11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 7 E o * a FT o> o 2 — -3 2 h - 0 — td - g — 2d 0 — 3d -90 -75 -60 -45 -30 -15 2 0 1 2 5 1 3 4 Fig. 5.9 Bode plots for 1018 steel in 0 .1 N NaCl after cerating and sealing treatment by cathodic polarization in 20 mM CelNChb at 3 mA/cm: for 2 minutes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase A ngle (degree) 68 (b) Fig. S.10 SEM images o f the cerated layer on 1018 steel after cathodic polarization in 20 mM Ce(NC>3)3 at 3 mA/cm2 for 2 minutes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 69 (a) C t F C N eea < . 1 5.220 keU 1 0 .3 > (b) F i < .1 5.220 keU 1 0 .3 > Fig. 5.11 EDS analysis results o f the cerated layer on 1018 steel after cathodic polarization in 20 mM Ce(N03 > 3 at 3 mA/cm2 for 2 minutes; (a) mounds, (b) flat area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 0 demonstrates that the flat areas had a larger Ce/Fe ratio (Fig. 5 .11(b)) than that of cerated layer (Fig. 5.8(b)). The mounts seemed to consist entirely o f cerium (hydr)oxide (Fig. 5 .11(a)). 5 .1.2 Evaluation in the baseline solution at RT The baseline solution, i.e. 5 wt% NHj - 0.2 vvt0 /o NaOH. has a pH value o f 13.3 at RT. Cerium salts are not soluble in this solution at RT. since the cerium cations form insoluble cerium (III) hydroxide, whose solubility product is 1.6«I O'2 0 [82], On the other hand, in the baseline solution magnetite can readily form on the iron surface and provide corrosion protection (Fig. 2 .1). The protective properties o f cerated layers on 1018 steel specimen were evaluated in the baseline solution for one week. For comparison, untreated 1018 steel samples were also tested in the baseline solution and baseline solution inhibited with 2.4 wt% NazCriO- for one week. The 7 days impedance spectra all had OTC characteristics, but were more capacitive than in 0.1 N NaCl. The untreated sample in the blank baseline solution had the highest Rp and lowest C values after one-day exposure, which was more evident after 7-day exposure (Fig. 5.12). The time dependence of the fit parameters Rp and C values for the above three cases is shown in bilogarithmic plots o f log ( l/Rp) vs. log (t) (Fig. 5 13a) and log ( l/C) vs. log (t) (Fig. 5 .13b). Rp and C can be fitted according to Eq. 4-7 and 4-9. respectively. For untreated 1018 steel in the blank baseline solution, the fit equations are: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 1 E o .90 Untreated, in 5WF&NH3 - 0.2 wt% NaOH ~ ^ Untreated, in 5**tHNH3- ^ ^ 0.2 wt'*NaOH-2.4%Na2Cr20 7 ' -75 £ 0 r> . cerated. in 5wt%NH3- >i “ ' w 0.2 wt% NaOH 5 O ^ o _co c - ' 3 o g , > c Q < * V - S « ~ ~ 0 o * o - N J _ O O ^X ■5 0 < 5 ^ O 0 3 - OO _ — -30 ry 5 ~ “ ' >s--------------------------- ----- o - o o - e - 0 - 0 o o o o o o 2 “ 4 * 15 * 5 1 Sk_ - © 1 ^-^-.5 . jr ' ^C 0 3 0 — 0 -3 -2 -1 0 1 2 3 4 5 log(f(Hz)) Fig. 5 .12 Bode plots of 1018 steel samples exposed to different solutions at RT at the 7th day after different treatments. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase A n g le (degree) 7 2 -4.5 CM o -5.5 Q. 0 1 ^ - 6 — ______ Untreated, in 5 wt% NH3 2 * 0 .2 wt% NaOH Untreated, in 5 wt% NH 3 - -6.5 0 2 wt% NaOH + 2.4 vK % Na2Cr2P 7 Cerated. in 5 wt% NH, ^ 0.2 wt% NaOH 0 4 0 8 1 2 1 6 2 2.4 Log(t(hrs)) (a) 3.7 —---------------------- — ------------------------------------------------------------------------------- . 0 0.4 0.8 12 16 2 2.4 Log (t (hrs)) (b) Fig. 5.13 log ( 1/Rp) vs. log (t) (a) and log ( l/C) vs. log (t) (b) plots for 1018 steel in different solutions at RT after different treatments. Dotted lines are the plots according to the fit equations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 3 log ( l/Rp) = - 4.73 - 0.93 log (t), R2 = 0.98 5-3a log ( l/C) = 4.03 - 0.05 log (t). R2 = 0.93 5-3b where the R" is the correlation coefficient indicating the fitting quality for each equation. For untreated 1018 steel in the baseline solution inhibited with 2.4 wt% Na2Cr2 0 7 , the fit equations are: log (l/Rp) = - 4.91 - 0.50 log (t). R2 = 0.96 5-4a log ( l/C) = 3.94 + 0.03 log (t), R2 = 0.65 5-4b For cerated 1018 in blank baseline solution, the fit equations are: log ( l/Rp) = - 4.67 - 0.25 log (t), R2 = 0.87 5-5a log ( l/C) = 3.65 + 0 .12log (t), R2 = 0.93 5-5b For mild steel, untreated or cerated. in the baseline solution, inhibited or not inhibited with dichromate. the relative corrosion rates ( l/Rp) have similar trends, i.e.. they decrease with increasing exposure time. However, they differ in the rate of decrease, as evidenced by the different slopes o f the double-logarithmic plots (Fig. 5.13a). The smallest value o f the slope. - 0 93 for the untreated sample in the blank baseline solution indicates that the magnetite film provides the best long-term protection for mild steel in the baseline solution at RT From the log ( l/C) vs. log (t) equations, it can be seen that for the rougher and thicker cerated layer caused a smaller ac (3.65) and a larger be (0.12) in Eq. 5-5b than those in Eq. 5-3b and 5-4b. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 4 5.1.3 Evaluation in the baseline solution at 100°C Considering that the solubility o f cerium salts increases in the baseline solution at high temperature ( 100°C). Ce(NChh as inhibitor and cerating process were evaluated in the hot baseline solution. 5.1.3.1 Cerium salts as inhibitors 5 mM Ce(N0 3 )3 was added to the hot baseline solution as inhibitor. For comparison, 2.4 wt% NajCr^O— 2H ;0 and I wt% sodium (meta) silicate were also examined as inhibitors in the baseline solution. In order to determine the effect o f the 0 . 2 wt% NaOH in the baseline solution, a blank 5 wt0 /o NH^ solution was also tested. The 2-day impedance spectra for 1020 steel exposed to five different solutions at 100°C are shown in Fig. 5 14. The plateaus at the highest frequencies (Fig.5.14a) indicate different solution resistance due to the addition o f different inhibitors. All spectra except those obtained in the blank 5 wt% NH? solution can be fitted with a OTC model. The spectra obtained in blank 5 wt% NH3 solution can be fitted with a modified TTC model, whose equivalent circuit is depicted in Fig. 5.15a. The fit data vs. experimental data are compared in Fig. 5 .15b. The CPE in series with R * represented the incompact magnetite film formed on mild steel. The fit parameters Rp, C and the calculated inhibition efficiency with respect to the blank 5 wt% NH3 solution for each test at the 2nd day are listed in Table 5.3. It can be seen that with the addition o f the inhibitors, including the hydroxyl anions. Rp Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 5 7 ---------------------- : -----------------;------------------------------------- -- -----5%NH3+0.2%NaOH --0 — - 5%NH3+0 2%NaOH+2.4%Na2Cr2O7 6 t_i ~ 5%NH3*0.2%NaOH+5mM Ce(N03)3 § * - O - 5%NH3 * ☆ 5%NH3+Q.2%NaOH+1 %Na2Si03(mela) i 5 ~ ------------------------------------------------- rsl 4 o > ° • a _l 3 2 1 °o& (a) I * iy - % ^ * “ -3 - 2 - 1 0 1 Logf(Hz) -90 -00 — O '0 0 .vt-ifcl-p. v .- © -7 0 o> -60 * ■ ' C x i © ^ •o _© a > j= a. * * -50 « 2 -40 ° S 1 -3 0 - " s - 2 0 - (b) - 0 ' . S *1° ~ ° c „ n - - % o ----------------- - . - . - -1 0 1 2 3 4 °5 Logf(Hz) Fig. 5.14 Bode plots for 1020 steel after 2 days exposure to different ammonia solutions at 100°C; (a) impedance, (b) phase angle. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 6 c CPE R c E o OS o (a) 7 —- — 6 * • | 5 — O Experimental data # - Fitted data • • • • S*. 4 3 v \ V -3 -2 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 « ra as ■ o _ 0 ) G O £ 0 ) c o as 0 1 Logf(Hz) (b) Fig. 5 .15 Comparison o f experimental and fitted data for 1020 steel after 2 days exposure to 5 wt% NHj solution at 100°C. (a) equivalent circuit, (b) Bode plots. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 7 Table 5. 3 Rp , C and inhibition efficiency values for 1020 steel after 2 days exposure to different ammonia solutions at 100°C. EIS test solution Rp (Q-cm2) | C (F/cm2) Inhibition efficiency (%) 5 wt% NH3 | 6 .6 0 x 105 1 | 3.05-10"* N/A 5 wt% NH3 -t- 0.2 wt% NaOH 1.18x10° j 1.09x10"* 4 4 5 wt% NH3 + 0.2 wt% NaOH + 5 mM Ce(N 03)3 1.51x10° i i 9 .3 3 - 1 O'5 56 5 wt% NH3 - 0.2 wt% NaOH +2.4 wt% Na2Cr2 07-2H20 1 ! 1 1 .3 4 x 1 0 ° I 7.75x 10*5 51 5 wt% NH3 -r 0.2 wt% NaOH -r l wt0 /o Na2Si03-9H20 ! 2 .3 4 x 1 0 ° 8 .0 3 - 1 0 '5 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 8 increased to different extent. 1 wt% NaiSiQj-OHiO had the highest corrosion inhibition efficiency, while 5 mM CefNOj^ was the second best, slightly better than 2.4 wt% NaiCriOT-IH^O Addition o f 0 2 wt% NaOH to the 5 wt% NHj solution increased the solution pH at RT from 12.1 to 13.3, and gave an inhibition efficiency o f 44% at 100°C. In the presence o f inhibitors, the capacitance also decreased to different extent indicating that more protective and thicker films were formed on the steel surface or inhibitor anions were adsorbed on the steel surface. 5 mM CefNOsb was not totally soluble in the baseline solution even at 100°C and made the solution unclear. Some grav-colored deposit was found on the bottom of the test cell. However, the impedance spectra o f 1020 steel exposed to the solution during 2 days, as shown in Fig. 5.16. indicate strong interaction between the solution and 1020 steel surface resulting in a continuous increase o f Rp and decrease o f C (Table 5 .4). which implies the formation o f a protective film with the incorporation o f a small amount o f dissolved cerium cations. 5.1.3.2 Cerium salts used in cerating 1020 steel was cerated in 10 g/L CeCh + 5 wt% H2O2 at RT for 8 minutes, then tested in the baseline solution at I00°C for 5 days. The impedance spectra are given in Fig. 5 .17a. The inverses o f fit parameters Rp and C values are drawn vs. exposure time in a double-logarithmic graph (Fig.5.17b). During the five days exposure, Rp continuously decreased and can be fitted to the following time law: log ( t/Rp) = -3.20 - 0.73 log (t), R2 = 0.97 5-6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 7 ---------------------------------------------------------------------------------------------------------------------------- • -90 :nours log(f(Hz)) Fig. 5.16 Bode plots for 1020 steel exposed to 5 wt% NHj + 0.2 wt% NaOH CeiN O jh solution at 100°C for 2 days Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 mM Phase A n g le (degree) 8 0 Table 5 .4 Rp, C, and rco rr values for 1020 steel exposed to S wt% N H3 + 0.2 wt% NaOH + 5 mM CefNQjfc at 100°C for 2 days. j Immersion I time (hours) ! R p (Qcm 2) | C (F/cm2) rcorr (pm/year) 1 2 1 l.67x IO 5 | l.53x 10-4 1.2 I 24 1 4.30x105 | 1.29x1 O '4 0.5 0 0 •t 1.51x10° | 9.33x1 O '5 . ..O' _ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. S I C M E o * g C T O 5 — -3 2 -3.6 -3 -2 0 1 2 log(f(Hz)) (a) - 9 0 -75 -60 © £ o > © ■ o © -45 o) c < © C O C O -30 -15 C M i . g 'a . <r -4.4 a t o C M E o O a> o -4.8 -5.2 0.4 0.8 1.2 1.6 log(t(hrs)) 2.4 (b) Fig. 5.17 (a) Bode plots and (b) fit parameters Rp and C vs. time for cerated 1020 steel exposed to 5 wt% NtL$ — 0.2 wt% NaOH solution at 100°C for 5 days. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 2 The C value had a drastic drop during the first day from 1.54* I O'1 F/cm2 at 2 hours to l.54x I O'3 F/cm' at 24 hours, but after the first day it remained relatively unchanged. SEM micrographs (Fig. 5.18) o f 1018 steel, cerated in 12.5 g/L CeCl? - + • 2.5 wt% H2O2 at RT for 20 minutes show that the as-cerated surface (Fig.5.18a) had a fully developed microcrack network, but all microcracks were essentially closed after 24 hours immersion in the hot baseline solution (Fig. 5 .18b). The phenomenon is termed with ^self-sealing”. One cerated 1020 steel sample was first sealed via cathodic polarization in 50 mM Ce(Cl)3 at RT for 3 minutes with a cathodic current density o f 10 mA/cm2. and then exposed to the hot baseline solution for 5 days. Its impedance spectra had almost identical features with the spectra in Fig. 5.17a. which means that cathodic deposition sealing had no obvious effect in improving the performance of the cerated layer in the hot ammonia solution or that the improvement due to cathodic polarization was overshadowed by the self-sealing effect occurring in the baseline solution. 5.2 Optimization o f the cerating process Cerating is a multi-variable or multi-factor dependent process. The factors investigated here include the cerating solution composition (concentrations o f CeClj and H2O2), cerating time, pretreatments and post treatments, additives to the cerating solution, pH and aging o f cerating solution. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 (a) (b) Fig. 5.18 SEM images o f cerated 1018 steel before (a) and after (b) immersion in 5 wt% NH3 + 0.2 wt% NaOH solution at I00°C for 24 hours. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.2.1 Factorial design The cerating solution composition (concentrations o f CeClj and H2O2, labeled as factors FI and F2) and cerating time (labeled as factor F3) were first optimized through a three-factor and two-level (2J) factorial design, as arranged in Table 4 .1. A total o f 8 experiments were carried out. The protective properties o f the cerated layer on 1020 steel were evaluated with EIS in the baseline solution at 100°C at 2. 24 and 48 hours for each experiment. The impedance spectra o f the 8 experiments at 48 hours are shown in Fig. 5.19. The fit values o f Rp and C. which are the responses for the 2J factorial design, are listed in Table 5.5. There were large quantitative differences in these impedance spectra as well as in the responses. The lowest Rp and highest C occurred in experiment #8. which employed the higher levels o f all three factors. The highest Rp was obtained from experiment #2. in which the lower level o f CeCh and H2O2 concentrations and the higher level o f cerating time were used. The main and interactive effects (Ei) o f the three factors on the responses Rp and C in Table 5 5 are listed in Table 5 .6 together with the average values o f the responses (XA v. i). The percentage for each effect is defined as the value of the effect divided by the average value o f the responses. It indicates the percentage change o f the average value when one factor changes from its lower level to its higher level. Table 5 .6 shows that the concentration o f H2O2 and the cerating time have a larger influence on Rp than the concentration o f CeCb; while all three factors have strong effects on C. Increasing the cerating time and decreasing the concentration o f H2O2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 5 Experiment # i Experiment *2 Ex pen m em *3 Experiment Expen ment *5 Expenment#6 Exoenment »7 Expenment #8 -3 2 -1 0 1 2 3 4 5 log(f(Hz)) . 9 0 -------------------------------------------------------------------------- ----------- (b) -75 < u 3 2 1 0 1 2 3 4 5 logCfCHz)) Fig. 5.19 Bode plots for the cerated 1020 steel for the 23 factorial design experiments exposed to 5 wt% NH3 + 0.2 wt% NaOH solution at 100°C after 48 hours; (a) impedance, (b) phase angle. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 86 Table 5 .5 Fit parameters Rp and C values for the 23 factorial design experiments. Experiment # Rp (ohm-cm2) C (F/cm2 ) 23-l 9.00* 104 6.75*10“* 23-2 1.80* 105 2.03*10“* 23-3 1.06* IO5 1.79*10“* 23-4 1.92* 10; 3.20*10“* 23-5 9.91* IO4 4.70* 10"* 23-6 9.6* IO4 2.78* IO'3 23-7 9.14* IO4 2.76* 10“* 23-8 4.86* 104 7.61 *10"* Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 7 Table 5. 6 The main and interactive effects and corresponding percentage of factors FI (concentration of CeCh), F2 (concentration of H2O2 ) and F3 (cerating time) and the average values of the responses. Rp (x IO4) (O-cm*) C(xl0"*) (F/cm2) X A V 11.290 15.641 E fi -0.678 10.643 (-6%) (68%) E f2 -5.823 24.398 (-52%) (156%) E f3 3.253 23.283 (29%) (149%) E fIF2 -2.078 12.538 (-18%) (80%) E f iF3 -1.090 14.093 (-9 7%) (90%) E f2F3 -5.550 24.938 (-49%) (159%) E fIF2F3 -1.035 17.338 (-9%) (111%) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. could lead to a higher Rp. However, during cerating. it was observed that a long cerating time could result in a dull brown color instead o f a golden color film and cause decomposition o f H2O2 as evidenced by the continuous bubbling after the removal o f samples from the cerating solution. Therefore, the beneficial effect o f longer cerating time is traded off by the short service life of the cerating solution due to the decomposition o f H2O2. Based on these results, a 22 factorial design. labeled as 22-l. was constructed, in which the CeCU concentration was fixed at the higher level (12.5 g/L) and the H2O2 concentration was decreased and had levels of I wt% and 2.5 wt%. The cerating time remained at the same two levels as in 2s. Thus, only two new experiments were needed for the 2'-t design. The arrangement o f the 2M factorial design and the responses o f Rp and C o f the cerated samples after 48 hours immersion in the baseline solution at 100°C are tabulated in Table 5.7. Table 5.8 lists the main and interactive effects o f H2O2 concentration (factor F2) and cerating time (factor F3), the corresponding percentage, and the average values o f the two responses. The differences among the Rp and C values, as shown in Table 5.7 were much reduced from those shown in Table 5.5 for the 23 factorial design. Higher average responses (Rp and C) were obtained for 2"-l (Table 5 8). It can be seen from Table 5.8 that increasing in the concentration o f H2O2 resulted in a slight increase in Rp (7%) and a large decrease in C (-61%). Longer cerating time caused a sharp increase in Rp (87%), but only a relatively small increase in C (43%). The maximum Rp Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 89 Table 5.7 Factors and responses for the 22 factorial design (22-l). Experiment Factors Responses H2O2 (wt%) Factor F2 Cerating time (Minutes) Factor F3 Rp ^ (Q-cm*) C (F/cm2) 2M-1 I 10 5. 58 a IO4 3.83 a 10"1 2M-2 (23-3) 2.5 10 1.06 x 105 L.79x L0"1 2--r-3 I 20 2.21 xIO5 5.53 x 10"1 22-I-4 (2j-4) 2.5 20 I.92x 105 •t 0 X 0 ri Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 90 Table 5. 8 Main and interactive effects of factors F2 and F3 for 22 -l. R p(*l05 ) (Q cm :) Cp (x iq-4) (F/cm*) X av 1.437 3.588 Ef2 0.106 -2.185 7% -61% E f3 1.256 1.555 87% 43% E f2F3 -0.396 -0.145 -28% -4% Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 (2.21 < IO5 fl-cm ') among the experiments of both 23 and 22 -I factorial designs was obtained with the cerating condition o f 12.5 g/L CeCU and I wt% H2O2 for 20 minutes. Considering the fact that the high CeCh concentration resulted in a cerated layer with a dull brown color, and decomposition of cerating solution, and that the lower CeCh concentration had a small beneficial effects on both Rp and C (Table 5.6), another 2* factorial design (2MI) was performed. The arrangement of 2MI is listed in Table 5.9. in which the cerating time was fixed at the high level (20 minutes), while the concentration ofCeCU was decreased and had levels o f 2.5 g/L and 7.5 g/L. Three more experiments were needed for 2:-lI and the responses of Rp and C are shown in Table 5.9. The main and interactive effects o f CeCU concentration (F I) and H2O2 concentration (F2), the corresponding percentage, and the average values o f the responses are listed in Table 5 10 Compared with the results in the 22-l factorial design (Table 5.8). comparable average Rp. but larger average C values were obtained in the 22-II factorial design (Table 5.10). From Table 5.8, it can be seen that the higher H2O2 concentration had beneficial effects on both Rp and C. and that the high level CeCU concentration resulted in a slightly smaller Rp (-9.7%), but much smaller C (-69%) than the low level CeClj concentration. With low concentrations o f CeCU and H2O2, a golden yellow cerating film on mild steel was obtained. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 92 Table 5.9 Factors and responses for the 22 factorial design (22-II). Experiment U rr Factors Responses CeCb (g/L) H ;0 ; (\V t% ) Factor F2 R p (Q cm 2) C (F/cm2) 22-II-I 2.5 I 9.43* 104 7.41 * I O '4 N 1 1 7.5 I 7.31 x 104 6.24x10"* 2MI-3 2.5 2.5 1.85 < IO5 9.57«10"* 22-ll-4 (23-2) 7.5 2.5 1.80- IO5 2.03 -10"* Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table S. 10 Main and interactive effects of factors F2 and F3 for 22- R p ( x l 0 5) (Q -cm * ) C p fx lO * 4) (F /c m 2) X AV 1.33 6 .3 1 E fi -0 .1 3 -4 .3 5 - 9 .7 % -6 9 % E f2 0 .9 9 - 1 .0 2 7 4 % -1 6 % E f iF2 0 .0 8 - 3 .1 8 4 6 .2 % -5 0 % Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 94 By employing some results from the 23, 22 -I and 22 -lI factorial designs and performing one more experiment with the cerating condition of 2.5 g/L CeClj. 7.5 wt% H20 2 and the cerating time o f 20 minutes, a two-factor three-level (32) factorial design, labeled as 32-I was constructed with cerating time fixed at 20 minutes, three levels ofC eC h concentration (factor Fi), i.e. 2.5. 7.5 and 12.5 g/L, and three levels of H2 0 2 concentration (factor F2) o f 1. 2.5 and 7.5 wt%. Its arrangement and the responses o f Rp and C are listed in Table 5 .11. Through the least-square method, the prediction equations for Rp and C in 3:-I were obtained: Rp (Q-cm: ) = 2.45x 104 -4 .9 0 * IO3 (F,) + 8.88* I04(F2) + 1.13* 103(F i) 2 - 8.69* 103(F2) 2 - 2.79* IO3 F,F2 5-7 C (F/cm2 ) = 3.78* 10'3 - 6.24< 10"* (F,) - 1.37-10'3 (F2) * 2.64« 10'5 ( F t ) * ^ l.I7 * I0 ‘ 4 (F: ) 2- 115* 10“* F i F 2 5-8 These equations are valid only in the range o f CeCh: 2.5 - 12.5 g/L and H20 2: I - 7.5 wt% for 20 minutes cerating. The response surfaces in 3D and 2D formats for the fit Rp and C values are presented in Fig. 5.20 and Fig. 5.21. respectively. No maximum or minimum value was obtained for either Rp or C. The response surface o f Rp had a saddle point at intermediate ranges o f both CeCU and H20 2. However, it showed two trends o f Rp to increase. One was in the direction o f decreasing the concentration of CeCL and increasing the concentration of H2 0 2. The other was in the direction o f decreasing the concentration o f H2 0 2 and increasing the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 Table 5. 11 Factors and responses for the 3Z -I factorial design. Experiment # Factors Responses CeCIj (g/L) H2O2 ( wt%) Rp (H em2) C (F/cm2) 32 -I-l (22 -II-l) 2.5 1 9.43 x 104 7.41 x IO "4 32 -I-2 (22 -ll-3) 2.5 2.5 l.85x IO5 9.57x10"* U V N 1 1 2.5 75 1.27vl05 9.92x10"* 32-I-4 (22-1I-2) 7.5 t 7.31 «104 6.24x 10"* 3:-I-5 (23-2) 7.5 2.5 1.80-IO3 2.03-10"* 3:-I-6 (23-6) 7.5 7.5 9 .6 0 -IO4 2.78-10'3 3--I-7 (2:-I-3) 12.5 I 0 X r i 5.53x10"* 3:-[-8 (2:-I-4) 12.5 2.5 I.92xI05 3.20x10"* 32-I-9 (23-8) 12.5 7.5 4.86x 104 7.61 x I O '3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (b) 7.0 - 1.4e+5 1.6e+5 1.8e+5 " 5.5 2.0e+5 S P 1 C M o C M X 4.0 1 Oe+5 1.2e+5 . 1.4e+5 1.6e+5 1.8e+5 a n p ^ i 6.06+4 8.0e+4— — 8.0e+4---- 1.0e+5-------- l.0e+5- 1 2e*5~------• , 2e+5— ~ " 1.4e+5-------- 1.4e+5" _ _ 16e+5' -1.6e+5 1.8e+5— 1.8e+5 1.8e+5 2.0e+5 2.06+5 2.5 - 1.8e+5' ,_1.6e+5 -1.4e+5- 1.2e+ ~1.8e+5.^ 1.6e+5 -1.4e+5. 2. Oe+5 2.2e+5 1.8e+5 1.0 1 f0e+l 2.5 '1.2e+5 1.0e+5 r- " ; « _ 2.06+5 ^ 1 8e+5 1.4e+5 ^ 7.5 CeCl3 (g/L) 10.0 12.5 Fig. 5.20 Response surface o f R p in 3D mesh (a) and 2D contour (b) for the 32 -I factorial design o f CeCh and H2O2 concentrations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 7.2e-3 6.4e-3 5.6e-3 4.8e-3 « 4.0e-3 u. ^ 3.2e-3 ° 2.4e-3 (a) 1.6e-3 8.0e-4 Wj 0 2 ^ ° /o) (b) I ts O 7 6 5 4 3 4.7e-4 2 4 7e-4 1 3e-3 1 2.5 5.0 7.5 10.0 12.5 CeCl3 (g/L) Fig. 5.21 Response surface o f C in 3D mesh (a) and 2D contour (b) for the 32 -I factorial design o f CeCl* and H2O2 concentrations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 concentration o f CeCb The response surface o f C indicated that C decreased in the direction o f decreasing the concentration o f CeCb and increasing the concentration of H2 0 2. similar to the first trend tor Rp. Since a cerated layer with high Rp and low C was desired, the trend showing an increase in Rp. but a decrease in C was further investigated by implementing another two-factor, three-level 3* factorial design. labeled as 3 '-II. By employing six experimental results in 32 -l. the construction o f 32 -II needed only three more experiments at a lower CeCb concentration, i.e. 1 g/L. The arrangement o f the 32 -II factorial design together with its responses o f Rp and C are listed in Table 5 .12. The prediction equations for Rp and C in 32 -II were: Rp (Q-cm2) = 2.79* IO4 + 4.58x IO 3 (F,) + 7.92x 104(F2) - 7.38 a 10:(F ,): - 8.85a 103(F2 ) : - 2.77* IO2 F,F2 5-9 C (F/cm2) = 1.27' IO'3 - 2.15* IO'5 (Fi) - 4.62«lO ^F;) - 1.36* 10'5 (F,) 2 + 4.32* IO'5 (F:) 2 + 5.95* I O'5 F,F2 5-10 These equations were valid in the range of CeCb: 1 - 7 .5 g/L and H20 2: I — 7 .5 wt% for 20 minutes cerating. The response surfaces in 3D and 2D formats for Rp and C are presented in Fig. 5.22 and Fig. 5.23. respectively. A maximum value of Rp existed at 2.3 g/L CeCb and 4.4 wt% H2 0 2 in the response surface o f Rp, while no minimum or maximum point existed in the response surface o f C. According to Eq. o f 5-9 and 5-10. at the point o f 2.3 g/L CeCb and 4.4 wt% H20 2. Rp was 2.09* IO5 Q-cm2 . and C was 5.96x IO "4 F/cm2. To verify the prediction Eq. o f 5-9 and 5-10 and confirm the cerating time effect, two EIS experiments were performed for 1020 steel Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 Table 5. 12 Factors and responses for the 32 -II factorial design. Experiment # Factors Responses CeCl3 (g/L) H2O2 (wt%) Rp (Q-cm*) C (F/cm2) 3MI-1 1 I 1.28xl05 7.38x10"* 32 -ll-2 I 2.5 1.43* 105 5.29x10"* 32-II-3 t 7.5 1.31 *105 7.51*10"* 32 -I-l (22 -II-l) 2.5 1 9.43.104 7 41*10"* 3M-2 (22-H-3) 2.5 2.5 1.85 - IO5 9.57*10"* 1 4 1 1 0 4 2.5 7.5 1.27* IO5 9 92 * 10"* 3*.[_t (2MI-2) 7.5 1 7.31 * IO4 6.24* 10"* 3“-I-5 (2*-2) 7.5 » n ri 1.80*10-' 2.03x10"* 32-I-6 (23-6) 7.5 7.5 9.60x104 2.78x IO * 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (b) C M O C M X 1 4e+5 1.6e+ 1.6e+5- 1.8e+5 1.8e+ 2.0e+J 2.0e+5- 1.8e+5- 2. Oe+5.. -2.0e+5' -l.8e+5 ~1.4e+5—. 1.6e+5_ 1.8e+5 _ _ _ 2.0e+5 2.0e+5 -1.8e+5" '1 2e+5m ~ ~ " 1.4e+5 " 1.6e+5 1.8e+5 1.8e+5 1.6e+5 — 1.6e+5 1.6e+5 l.6e+ 1.4e+5 i.4e+5 t.4e+5 t.2e+5 1.0e+ 4.0 5.5 CeCl3 (g/L) Fig. 5.22 Response surface o f Rp in 3D mesh (a) and 2D contour (b) for the 32 -II factorial design o f CeCb and H2O2 concentrations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (b) 6.7e-4 8.2e-4 * 5 o'* k7e^5.2e-4 N A - o * ^ C M X 3 - 2 - 1 _6.7e-4 8 .2 e -4 -_ 1.3e-3 ' 1.1e-3_ ' 9.7e-4-__ " 8 2e-4 6.7e-4 6.7e-4 5.2e-4 T 1 --------- 1 --------- r 1 2 3 4 5 CeCl3 (g/L) T 7 Fig. 5.23 Response surface of C in 3D mesh (a) and 2D contour (b) for the 32 -II factorial design o f CeCIj and H 2O2 concentrations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 2 in the baseline solution at 100°C for 48 hours after cerating at the optimum condition o f 2.3 g/L CeCU and 4.4 wt0 /o H2O2 for 20 minutes as well as 5 minutes. The 48-hour impedance spectra of the two experiments are shown in Fig. 5 .24. The fit results were Rp = 2.56* 105 Q-cm" and C = 6.41 * 10 * 4 F/cm* for 20 minutes cerating and Rp = 1.24* 10s Q-cm2 and C = 4.03 * I O '4 F/cm* for 5 minutes cerating. It can be seen that the experimental results of Rp and C were close to the predicted values. Cerating time had a pronounced effect on both Rp and C. With longer cerating time, a higher Rp. but also a higher C was obtained. 5 .2.2 Effects of pretreatments, post treatments and additives The effects o f three pretreatment (pickling) processes, four post treatment (sealing) processes and three additives on the corrosion protection efficiencies o f cerated layers on mild steel were evaluated in the baseline solution at RT for 7 days. The three chemicals were added to the cerating solution in order to obtain more uniform and compact cerated layers on mild steel. Triton X-100 with a concentration o f 3 ppm was used as a surfactant in the cerating solution for uniform formation o f the cerated layers. Lead acetate ((C^HjO^h Pb * 3 H2O) with a concentration o f 4 ppm was used as stabilizer for the cerating solution. Sodium nitrite was added at a concentration o f 2g/L to scavenge oxygen gas evolved at anodic sites (Eq. 2-12). The arrangement of experiments for investigation o f the effects o f the pretreatments, post treatments and additives are listed in Table 5.13. According to the pretreatment applied all experiments were classified into 3 groups, i.e. the Micro, HCl and Actane Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 3 -------------------- ,g 0 Cerating for 20 minutes. Cerating for 5 minutes. 75 4 - " - rrr-- - 3 2 - 1 0 1 2 3 4 5 log(f<H») Fig. 5.24 Bode plots for 1020 steel exposed in 5 wt° o NH* - 0.2° b NaOH solution at 100°C for 48 hours after cerating at the optimization condition o f 2.3 g/L CeCh and 4.4 wt% H ;0 ; for 20 minutes and 5 minutes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 Table 5.13 Experimental arrangements for the investigation of pretreatments, post treatments and additives. Experiment ID Pretreatment Cerating ! Post treatment Micro 1 Pre-Treat I Ceratingl 1 NO HCll Pre-Treat2 Ceratingl NO HC12 Pre-Treat2 j Ceratingl | NO HC15 Pre-Treat2 Cerating3 | NO HC16 Pre-Treat2 Cerating4 , NO Actanel Pre-Treat3 Ceratingl ; n o Actane2 Pre-Treat3 Ceratingl additives i n o Actane3 Pre-Treat3 Ceratingl 1 Post-Treat I ActaneS Pre-Treat3 Ceratingl ( Post-Treat2 Actane6 Pre-Treat3 Ceratingl 1 Post-Treat3 I ActaneT Pre-Treat3 Ceratingl ' Post-Treat4 Pretreaunents. Pre-Treat I: Immersion in 2v/% Micro solution (International Products Corporation) at 66°C for 10 minutes. Pre-Treat2: Immersion in 1:1 volume ratio o f HC1 (36.5-38%) and DI water for 60 seconds at room temperature. Prc-TrcaG: Immersion in I20g/L ACTANE* 345 solution (Enthone-OMI Inc.) for 90 seconds at RT. Cerating processes: Ceratingl: In 12.5 g/L CeCW ~ 2.5 wt% H;0; at RT for 20 minutes. Cerating2: In 3 g/L CeCI-, - 2.5 wt% H;0 ; at RT for 5 minutes. Cerating3: In 2.3 g/L CeCI? + 4.4 wt% H;Q; at RT for 20 minutes. Cerating4: In 2.3 g/L CeCI-. * 4.4 wt% H;C> at RT for 5 minutes. Additives: 2/L NaNO; + ■ 4 ppm (CHjCOO^Pb + ■ 3 ppm Triton X-100. Post treatments: Post-Treat I: Immersion in 10% sodium silicate (meta) (Na;SiCK • 9H-0) solution at 50°C for 30 minutes. Post-Treat2: Immersion in 10% ^-glycerophosphate disodium salt (C?HO* PNa;- 9H;0 ) at 50°C for 30 minutes. Post-TreaG: Immersion in 1% sodium borate (Na;B4 0 -- lOlTO) + 1% cerium nitrate (Ce/NO?)?- 6H;0) at I00°C for 60 minutes. Post~Treat4: Immersion in 10% sodium molybdate (Na^MoOi * 2H;0 ) solution at 50°C for 30 minutes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 groups. In these experiments, different cerating processes were employed to further study the influences o f cerating parameters on the corrosion protection provided by the cerated layers in the baseline solution at RT. Experiments Micro I. HCll. HC12 and Actanel were used to compare the three pretreatment processes. Experiments HCll. HC12. HC15 and HC16 were designed to compare the four cerating processes. Experiments Actanel and Actane2 were employed to compare the effects of additives, while experiments ActaneS. ActaneS. Actane6 and Actane7 were intended to compare the four post treatment processes. The effects o f different cerating processes, additives and post treatments on the morphology of the cerated layers were investigated with SEM. 5.2.2.1 EIS results The impedance spectra for Micro I are presented in Fig. 5.25a. The spectra did not show pure one-time constant characteristics. At low frequencies as seen in the phase angle plots, there seemed to be a second time constant possibly due to an oxide formed during the Micro solution pretreatment (Pre-TreatI. Table 5 13). There were obvious changes from the 2-hour spectra to the I-day spectra, i.e. an impedance increase at low frequencies accompanied by an increase in the absolute value o f the phase angle. After Id immersion, the spectra remained relatively unchanged. The impedance spectra for the HCl group and Actane group experiments all had simple one-time constant features, but differentiated in impedance at low frequencies and in their time dependence. For experiment Actane3. which was sealed with sodium Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 6 5 - N J > s 3 — ------- 2 h — z--- Id S 2d < 3 3d 5d 7 a 3 3 (a) £k^jT *1 < O g C f(H :rT ! (b) - 75 50 4 5 30 15 0 90 75 50 45 X 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phflio Angle (Ucgioc) p |,a M Angle (Degiec) 1 0 7 90 r v £ >2 3 2 h 75 3d 5 0 60 X 3 3 m O g (f(H z )) (C) 90 Id £ A * 3 o • ■ -ft* * 3 - 3 0 £ 5d * * 0 X 0 0 ap < fo a (f(M ry > (d) Fig. 5.25 Bode plots for EIS tests carried out in 5 wt% NH3 - 0 2 wt% NaOH (baseline) solution at RT for one week; (a) Micro I. (b) ActaneS, (c) Actane6 and (d) Actane7. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 8 silicate solution (Post-TreatI), the spectra (Fig. 5.25b) were capacitive and became more capacitive with prolonged test time as evidenced by the change o f the phase angles at low frequencies. With sealing in a mixed solution o f sodium borate and cerium nitrate (Post-Treat3), the spectra o f experiment Actane6 (Fig. 5 .25c), shifted toward lower frequencies indicating the formation o f a high-capacitance cerium oxide layer. After sealing in sodium molybdate solution (Post-Treat4), the spectra o f experiment Actane7 showed a strong time dependence (Fig. 5.25d). i.e. the impedance increasing rapidly with the increase o f exposure time. The fit Rp values for all experiments were converted into corrosion rates according to Eq. 4-2 and are tabulated in Table 5.14. The fit C values for all experiments are listed in Table 5 15. The dependence o f corrosion rates on exposure time for each experiment was determined by fitting to Eq. 4-3 or 4-4 The fit parameters Ar. Br and the correlation coefficient (R2) for each experiment are listed in Table 5.16 together with the calculated thickness loss At o f mild steel and the volume o f hydrogen gas Vh2 released after one-year exposure o f I m2 assuming uniform corrosion. V'h; was calculated for hydrogen gas under standard temperature and pressure (STP) condition, i.e. 0°C and one atmosphere pressure from the Schickor reaction (2-1). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 109 Table 5. 14 Corrosion rate rcorr in pm/year for experiments over one week in the baseline solution at RT. r t Experiment Immersion time (hour) ID I I 2 ! 1 24 j 48 i 1 72 ' I ; 120 168 1 Micro I j 1 17.33 I 4.98 1 ! 4.23 I t-- 00 2.87 2.79 i HCll ! 21.66 ( 6.80 i j 4.00 i 3-56 I 3.03 " ! 2.84 1 i HC12 i 15.19 | 9.72 j 8.81 I 776 1 5.53 1 4.71 HC15 j 15.40 | 8.67 | 7.12 | 5.78 j 4.41 3.75 i HC16 | 6.71 | 4.30 | 3.10 i 2.56 ! 1.89 1.48 Actanel ! 14.72 ! 11.94 1 N/A I 6.89 ! 6.42 1 5.98 Actane2 > i 1 6.89 5.20 ! N/A ! 4.54 ! 3.44 3.69 ' A t - ! Actanej 0.64 0.52 ' N/A i i 1 0.34 ' 0.24 0.20 i ActaneS ! 2.96 1.30 N/A 1 ! i 1.30 i i 1 30 1.12 1 I ! Actane6 , i I 19.83 2.00 . N/A 1 i I-46 i 1.40 1.37 1 ! i Actane7 ; 5.68 2.25 j 1.46 : 1.03 i i ! 0.58 0.46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 0 Table 5. 15 Fit parameter C (F/cm2) values for experiments over one week in the baseline solution at RT. 1 i Experiment : Immersion time (hour) i id r i 2 24 48 1 n 1 I i 120 1 168 Micro I | 1.70E-03 _ 6.49E-04 1 ; 5.51E-04 1 1 5.87E-04 1 502E-04 ; 6.22E-04 j HCll 4.49E-04 3.04E-04 i ! 2.67E-04 1 2.43E-04 ■ 2.26E-04 : 2.1 IE4)4 1 ! j HC12 j 1.56E-04 1.27E-04 j 1.14E-04 i ' I.10E-04 j I.04E-04 \ 9.96E-05 j HCI5 | 3.76E-04 2. L8E-04 | 1.93E-04 1.74E-04 I 1 I.43E-04 1.32E-04 | HC16 | 1.14E-04 9.2 IE-05 j S.S6E-05 8.67E-05 | 8.52E-05 8.02E-05 ! Actanel ! 1.22E-04 I.OOE-04 I N/A 9.87E-05 | 9.64E-05 9.64E-05 ! Actane2 I 1.44E-04 1.15E-04 ! N/A L09E-04 i 9.60E-05 9.60E-05 I i Actane3 1 I.25E-04 I.I8E-04 ! N/A ! 1.I2E-04 1 1.12E-04 L09E-04 > ActaneS 1 ! 4.44E-04 3.28E-04 1 N/A i 1 2.88E-04 1 ! 2.77E-04 2.73E-04 j ! Actane6 I i 2.80E-02 1.52E-02 i N/A i 1 1 1.34E-02 j ! I.I2E-02 1.02E-02 1 i j Actane7 , 1 i 2.04E-03 , _ . 7.75E-04 | 7.06E-04 1 6.55E-04 | 5.80E-04 5.17E-04 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill Table 5. 16 Fit parameters Ar and B„ and correlation coefficient (R2), calculated thickness loss At (pm) and volume of evolved hydrogen Ve q (at STP) for I m2 in one-year exposure. | Exp. ID j Ar (pm/year) j B r | | ^ j R* ' At (pm) ' Vhi (L) 1 Micro I j 1 I 1 21.88 ! i i - 0.42 ' 0.98 1 1 0.8 ! 1 - 1 j J J 1 I | HCll ! 1 i 2951 i | - 0.48 ; i 1 0.99 1 0.7 I 2.9 ! 1 ; HC12 1 | 1 9 - 9 5 | -0.25 ; 0.90 | 2 7 ! i 11.3 ! I I HCI5 1 20.94 i 1 i 1 O 0.96 | i 1.8 j 7.5 | i | HC16 I 9.71 j -0.33 ! 1 0.91 | 0.7 j 2.9 1 l 1 Actanel ! 18.62 1 -0.21 I 0.89 | 3.5 ! 14.6 | ! Actane2 ! 7.94 1 - 0 15 1 0.92 ' 2.4 1 10.0 I ! ActaneS I 1 0.89 1 - 0.25 ' 0.84 1 i 0.1 i j 0 4 1 l | Actane5 ■ 3.09 ■ f ! - 0.20 ^ | 0.88 0.6 * ! i 2.5 j Actane6 I 1 i 23 44 I ! i - 0.62 i 0.91 ! ] 0.2 0.8 | [ Actane7 i i 1 0 , 2 1 i -0.56 ! 0.94 i 0.1 ° - 4 ! Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. L12 The fit parameter A, equals the corrosion rate of cerated mild steel at t = I hour. It reflects the protective properties o f cerated layers for different pretreatments, cerating and sealing treatments. The fit parameter Br describes the (long term) time dependence o f corrosion rates and is related to the prevailing corrosion mechanism, i.e. the interaction between the corrosive media and the cerated layer or the substrate mild steel. Unusually high C values were observed for some experiments without sealing, such as Micro I and HCll. or with sealing, such as ActaneS and Actane6. as shown in Table 5.15. The C values for all experiments have been fitted to Eq. 4-7 in order to establish the corresponding time laws. The fit parameters Ac. Be and R: are given in Table 5 17 .A c is the capacitance after exposure for I hour in the baseline solution at RT. while Be describes the time dependence o f cerated layer capacitance in the baseline solution. 5.2.2.2 SEM results Fig. 5.26 shows the SEM images for 1018 steel cerated in 3.0 g/L CeClj 2.5 wt % H ;0 ; at RT for 5 minutes (experiment Actanel). By comparing Fig. 5.26b and Fig.5.18a. which shows the SEM image o f 1018 steel cerated in 12.5 g/L CeCL 2.5 wt % H2O2 at RT for 20 minutes, it can be noticed that after decreasing the concentration o f CeCb and cerating time, the crack width and the crack density were decreased. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 3 Table 5. 17 Fit param eters A „ Bc and correlation coefficient (R2) for C in Table 5.15. 1 Experiment ID ; Ac(pF/cm2) ■ I Bc R2 i i j Micro I 1 ! 1778.28 i -0.26 1 0.84 i HCll | i 512.86 | 1 -0.17 I 1.00 1 I ; HC12 I I 169.82 | 1 -0.10 0.99 j 1 HC15 | 451.86 | -0.23 l 0.99 i I HCk> j 119.40 | - 0.08 | j 0.98 | J 1 Actanel j 123.03 | - 0.05 | 0.93 : 1 ! Actane2 1 154.88 ! -0.09 i 0.97 | i i 1 Actane 3 ! 128.82 1 - 0.03 1 0.97 | I 1 Actanel 478.63 * -on 1 0 99 i 1 I Actane6 32359.37 ! 1 -0.22 0.99 1 I Actane7 i ...1 T 2360 47 1 - 0 30 ; 0 97 | Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 114 (b) Fig. 5.26 SEM images for mild steel, cerated in 3.0 g/L CeCb + 2.5 wt % H2O2 at room temperature for 5 minutes (Actanel). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 5 Fig. 5.27 presents the SEM images o f 1018 steel with the same cerating treatment as for Actanel (Fig. 5.26). but followed by sealing in 10 % Na2Si(>j • 9H30 at 50C C for 30 minutes. It can be seen that after the sealing treatment, most microcracks and micropores formed during cerating were sealed and the cerated layer became much smoother. The circles in Fig. 5.41 are believed to be due to bubbles clinging onto the cerated layer during the cerating process. Fig. 5.28 gives the SEM images for 1018 steel cerated in 3 0 g/L CeCI? - 2.5 wt % H2O2 + 2 g/L NaN02 + ■ 4 ppm (CFLCOOEPb + 3 ppm Triton X -100 at RT for 5 minutes (experiment Actane2). Fig.5.28 and Fig. 5.26 demonstrate that with the additives in the cerating solution, the cerated layer had fewer cracks and smaller micropores. 5.2.3 Effect o f cerating solution pH The effect o f the cerating solution pH on the performance o f the cerated layer was evaluated in the range o f 1.6 to 2.5. Cerating was carried out using the optimum conditions, i.e. 2.3 g/L CeCI? - 4.4 wt% H2O2 with additives o f 2 g/L NaNOi + - 4 ppm (CHiCOOEPb + 3 ppm Triton X -100 at RT for 20 minutes. The pH o f cerating solution was adjusted with dilute HC1 solution. The impedance spectra for 1020 steel exposed to the hot baseline solution for 48 hours after cerating in different pH solutions are shown in Fig. 5 .29a. The dependence o f the fit Rp and C values on cerating solution pH is presented in Fig. 5.29b. The highest Rp and lowest C values were obtained when the pH o f the cerating solution was 2.2. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 (b) Fig. 5.27 SEM images for mild steel, cerated in 3.0 g/L CeCl3 + • 2.5 wt % HzOz at room temperature for 5 minutes, then sealed in 10 % Na2SiC> 3 - 9 H2O at 50°C for 30 minutes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 117 (b) Fig. 5.28 SEM images for mild steel, cerated in 3.0 g/L CeCL + 2.5 wt % H2O2 + 2 g/L NaN02 + 4ppm Pb(AC)2 + 3ppm Triton X — 100 at room temperature for 5 minutes (Actane2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. log(f(Hz» (a) 2.4x10s — — ------ ---------- 2 8x 1 0 ° Q 2.0x10s 2.4x1(7* 1 6x10s - <— 2.0x10T s c r" 12 x10s ^ 16 x1 or3 8.0x10* — c 1 2x10Ts - 4.0x10* —---------- * ---------- 8 .0 x 1 0 ° 1 5 17 1.9 2.1 2.3 2.5 pH (b) Fig. 5.29 1020 steel exposed to the baseline solution at 100 °C for 48 hours after cerating in cerating solutions o f 2.3 g/L CeCI* 4.4 wt% H2O 2 ~ - g/L NaNOi ~ 4 ppm Pb(AC)2 + 3 ppm Triton X -100 with different pH at RT for 20 minutes; (a) Bode plots, (b) fit parameters Rp and C vs. pH. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C(F/cm2) 119 5.2.4 Effect o f ‘‘aging” o f cerating solution The cerating solution was aged for 30 minutes by simply keeping it unused at RT. The cerating conditions were the same as those in 5.2.3 except that the solution pH was 1.9. The impedance spectra for mild steel exposed to hot baseline solution for 48 hours after cerating with and without aging o f cerating solution are shown in Fig. 5.30. After aging, Rp increased to 1.73 < 105 Q-cm2 from 9.53 * 104 Q-cm2. while C decreased to 9 .03 x I O '4 F/cm2 from 1.53 * 10° F/cm2. 5 .2.5 Effect o f sealing in silicate solution The effect o f sealing o f the cerated layer in 10 % sodium (meta) silicate at 50 °C for 30 minutes was evaluated in the hot baseline solution for 48 hours. The cerating treatment was the same as that employed in 5 .2.4. but without aging o f the cerating solution. The impedance spectra for 1020 steel exposed to hot baseline solution after 48 hours with and without sealing treatment are given in Fig. 5.31. The sealing treatment improved the performance o f cerated layer in the hot baseline solution by increasing Rp to 1.39 x 10s Q-cm2 from 9.53 < 104 Q-cm2 . and decreasing C to 7.71 < I O'4 F/cm2 from 1.53 * 10*3 F/cm2. 5.3 Evaluation o f Yttrium Salts Yttrium salts were investigated in a similar way as employed for the evaluation o f cerium salts. The protective properties o f yttrium salts result from the formation of yttrium (hydr)oxide Sims on mild steel since yttrium cation is also a cathodic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 2 0 Without aging O - At e ! ag in g 9 0 -75 3 ' 6 0 - 3 - ' ♦ - „ § % 3 . ' ® ~ f i 3 - v ‘ » - -« I ^ * " * jf a » A * 2 — ^ - - 3 0 * ® * « - - S - « = 2 r. -S * 4 * 1 . 4 . 4 . - * a £ -15 O - O ; t ------------ " ^ . ^ s r s ^ w ,. J ^ + — ~ r .... » — — _ 3 -2 1 0 t 2 3 4 5 (og(f(Hz)> Fig. 5.30 Bode plots for 1020 steel exposed to baseline solution at 100 °C for 48 hours after cerating treatments with and without aging o f cerating solutions o f 2.3 g/L CeCI? + 4.4 wt0 /o FLO; - 2 g/L N'aNO; - 4 ppm Pb(CH}COOh * 3 ppm Triton X -100 at RT for 20 minutes. The pH o f cerating solution was 1.9. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. togt«Hz» Fig. 5.31 Bode plots for 1020 steel exposed to baseline solution at 100 °C for 48 hours with and without sealing in 10% sodium (meta) silicate at 50 °C for 30 minutes after cerating treatment (without aging) as described in Fig. 5.30. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 2 2 inhibitor [67], The yttrium (hydr)oxide film can be formed on mild steel through fast chemical conversion (“yttrating”) in oxidized yttrium salt solution and through addition o f yttrium salts as inhibitors. 5.3.1 Fast chemical conversion ("Yttrating") Yttrating was carried out by immersion o f mild steel in solution 12.5 g/L YClj - 2.5 wt % H2O2 for 20 minutes at RT. The morphology o f the yttrated layer was observed with SEM. and its protective efficiency was evaluated in 0.1 N NaCl at RT and in the baseline solution at RT and 100 °C. 5.3.1.1 SEM results The yttrated layer formed on mild steel had a similar appearance as the cerated layer, but it had wider cracks than the latter (Fig. 5.32). The morphology o f the yttrated layer after immersion in the baseline solution at 100°C for 48 hours is shown Fig. C ** The adhesion o f the yttrated layer to the mild steel substrate was poor. Some pieces o f the layer had peeled off during the experiment and had exposed the mild steel substrate (Fig. 5.32a). After 48 hours immersion in the hot baseline solution the cracks had shrunk to some extend, but had not closed completely (compare Fig. 5.33b with Fig. 5.33b). After immersion for 48 hours in the hot baseline solution some brown scales were observed on the bottom o f the test celL which were assumed to be yttrium and iron (hydr)oxides detached from the mild steel surface. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 123 (b) Fig. 5.32 SEM images o f mild steel yttrated in 12.5 g/L YC13 + - 2.5 wt % H2O2 solution for 2 0 minutes at room temperature. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 (b) Fig. 5.33 SEM images o f mild steel yttrated in 12.5 g/L YCI3 + 2.5 wt % H2O2 solution for 20 minutes at RT, and immersed in 5 wt % NH3 + 0.2 % NaOH solution for 48 hours at 100°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.3.1.2 EIS results in 0 .1 N NaCl at RT Immediately after yttrating, 1018 steel samples were immersed in 0.1 N NaCl at RT and EIS experiments were performed for one week. The impedance spectra over one week are presented in Fig. 5.34a. It can be seen that the spectra were shifted to the lower frequencies and that the impedance did not reach a DC limit in the investigated frequency range due to the high capacitance of the yttrated layer. The fit values o f Rp and C are plotted vs. time in bi-logarithmic graphs, as shown in Fig. 5.34b. Rp increased with time, while C decreases with time. The fit equations for Rp and C are: 5.3.1.3 EIS results in the baseline solution at RT The one-week impedance spectra o f 1018 steel exposed to the baseline solution at RT after yttrating are shown in Fig. 5.35a. The time dependences o f the Rp and C are plotted in Fig. 5.35b. Compared with the results in 0 .1 N NaCl at RT. much higher Rp and smaller C values were obtained due to the lower corrosivity o f the baseline solution at RT. The fit equations for Rp and C in Fig. 5 .35b are. log (I/Rp) = - 2.51 - 0.12 log (t). R2 = 0.98 log ( l/C) = 0.76 + 0 .18 log (t), R2 = 0.86 5-1 lb 5-1 la log ( l/Rp) = - 3.52 - 0.22 log (t), R2 = 0.95 log (l/C ) = 3.84 + 0.03 log (t). R2 = 0.79 5-12b 5-12a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 2 6 3 ; : 3 : ; a i 0 g(f(nz)) (a) -2 5 - .....____ . .... - 1 5 R p -2.8 — 0.5 0.0 0.5 1.0 1.5 2.0 2.5 log(t(hrs» (b) Fig. 5.34 1018 steel, yttrated in 12.5 g/L YCI3 + 2.5wt ^oHzCh for 20 minutes at RT. and then exposed to 0 .1 N NaCl at RT for one week; (a) Bode plots, (b) fit parameters Rp and C vs. time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 2 7 9 0 4 — (a) -3.5 3 95 5 o * S 4 0 Q. o o .d -" o 3 90 ^ ' " * ■ (j - ^ u_ £ < * * ' * 3.85 O ■ — i ------------ Rp ^2 on 00 J j -O’ ----- capacitance C 3 80 ~ -4.5 0 1 2 3 log(t(hrs)> (b) Fig. 5.35 1018 steel, yttrated in 12.5 g/L YCI3 + 2.5wt %H2 0 i for 20 minutes at RT. and then exposed to 5wt% NH3 - 0.2% NaOH at RT for one week; (a) Bode plots, (b) fit parameters Rp and C vs. time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 2 8 5.3.1.4 EIS results in the baseline solution at 100 °C After yttrating, 1020 steel samples were immersed in 100 °C baseline solution for one week. The impedance spectra are presented in Fig. 5.36a. The impedance at low frequencies increased drastically during the first day immersion (compare 2-hour and 1-day spectra), and after that it increased gradually with immersion time. The time dependences o f Rp and C are plotted in Fig. 5 .36b. The fit equations for Rp and C are: log ( I/Rp) = - 3.50 - 0.60 log (t). R2 = 0.94 5-l3a log ( I/C) = 1.94 + 0.44 log (t), R2 = 0.94 5 -13b It can be seen from the aR and bR values in the Eq. 5-1 la - 5-13a that with the yttrating layer the corrosion rate o f mild steel at one hour in the baseline solution decreased by about one order o f magnitude compared to that o f the untreated mild steel, and the corrosion rate decreased with time more rapidly due to the smaller bR values. The high temperature o f the baseline solution facilitated the decrease o f corrosion rate as indicated by the smaller bR value (- 0.60) in Eq. 5 -13a than that in Eq. 5-12a. Th ac values in equations 5-1 lb - 5-13b imply that a smaller capacitance at one hour was obtained after yttrating, which could be due to a thicker vttrated layer. The largest be in Eq. 5-13b indicates that the capacitance o f the vttrated layer in the hot baseline solution decreased most rapidly, perhaps due to a decrease of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 129 2 h = 3 3 2 — 60 » 3 0 3 Q ■ 8 a i < o S t ■ ■ n £ egtfC H sy, (a) -3.5 o 4 0 * a — ■ Q. Q C -4.5 Rp — O capacitance C o 00 o 3.2 2 8 E o 2.4 P 00 2.0 — -5.0 1 2 log(t(hrs)) 1.6 (b) Fig. 5.36 1020 steeL yttrated in 12.5 g/L YCI3 + 2.5wt Q / oH;0:> for 2 0 minutes at RT. and then exposed to 5 wt% NH3 + - 0.2% NaOH at 100 °C for one week; (a) Bode plots, (b) fit parameters Rp and C vs. time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 dielectric constant e and/or decrease o f surface area due to the self-sealing o f the yttrated layer in the hot baseline solution. At RT in the baseline solution, the yttrating layer could not be sealed; therefore the capacitance remained relatively unchanged as proved by be = 0.03 in Eq. 5 -12b. 5.3 .2 Y;( SCUta as Inhibitor Y2(S04>3 as inhibitor with the concentration o f 1 mM was tested only in the hot baseline solution. The one-week impedance spectra o f 1020 steel are shown in Fig. 5.37a. The spectra had a remarkable change during the first 24-hour immersion, and after that they remained relatively stable. The dependences o f Rp and C on immersion time are presented in Fig. 5 .37b. The fit equations for Rp and C are: log (l/Rp) = -4 .7 2 -0 .3 7 log (t). R: = 0.97 5-I4a log (I/C) = 3.14 + 0.25 log (t), R2 = 0.94 5-I4b The small aR (- 4.72) and large ac (3 .14) values indicate the formation o f a relatively thick film o f magnetite. However. bR (- 0 .37) is still large, implying a porous film and possibly ineffective incorporation o f yttrium into the magnetite film. 5 .4 Evaluation o f organic inhibitors Four organic inhibitors^ i.e. L-aspartic acid (AA) at a concentration o f 10 mM. (3- glycerophosphate acid (GPH) at a concentration of 50 mM. polv- L-aspartic acid (PAA) at a concentration o f 20 ppm and poly- L-glutamic acid (PGA) at a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 3 1 2a 5 60 M 30 3 c o 0 0 4 5 a :2 *0g(f(H2)) (a) I S -4.5 3 8 o -5.0 * ■ a al a : 00 o -5.5 -6.0 o . ✓ Rp capacitance C 3 6 3.4 E o O 00 3.2 o 3.0 0 1 2 3 log(t(hrs)) (b) Fig. 5.37 1020 steel in 5wt% NH3 + 0.2 % NaOH + - ImM Y2(SO.»)3 for one week at 100 °C; (a) Bode plots, (b) fit parameters Rp and C vs. time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 concentration of 20 ppm were evaluated to determine their inhibition efficiency for the mild steel 1018 in the baseline solution at RT using EIS and DC polarization electrochemical methods. 5.4.1 EIS results The two-week impedance spectra for 1018 steel in the blank baseline solution and the baseline solutions with the four organic additives all had simple OTC features and can be fitted to the OCT model (Fig. 3.3a). The dependences o f the fit values o f Rp and C on exposure time are given in Fig. 5.38. The time dependence o f Rp in Fig. 5 38 was fitted according to Eq. 4-7. and the fit parameters aR, bR as well as R: for each fitting are listed in Table 5 18 The impedance spectra obtained in the baseline solution + 10 mM AA (Fig. 5.39) had two time constants. The second time constant became more and more obvious after 24-hour exposure as indicated by the changes at the low frequencies in the phase angle spectra. These spectra could be fitted to a TTC model, usually called coating model [159.160] (Fig. 5.40a). The comparisons between experimental and fit data for the 1-day. 5-day and 14-dav spectra are given in Fig. 5.40b. The fit values for the two time constants are plotted vs. exposure time in Fig. 5.40c. The time dependence o f the resistances in Fig. 5.40c was fitted according to Eq. 4-7. The fit parameters aR, bR and R2 are listed in Table 5.18. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 133 Baseline + 50mM GPH Baseline + 20ppm PGA log(t(hrsV) 4 0 --- -------------------------------------------------------------------------------------------- - 0 1 2 3 log(tChrs)) Fig. 5.38 Time dependence o f fit parameters Rp (a) and C (b) for 1018 steel in different solutions for two weeks at RT. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 3 4 Table 5. 18 Fit parameters aR, bn and the correlation coefficient (R2) for 1018 steel in different solutions for two weeks at RT. ! Solution ! aR bR j R2 ! Baseline (BL) ' i i -4.60 -0.86 ! i 0.97 i t ! | BL - 50 mM GPH j -4.13 1 - 0.99 j 0.89 j BL -r 20 ppm P A A j -4.29 -0.97 , l 0.96 | BL + 20 ppm PGA j -4.45 -0.75 j 0.96 I BL * 1 0 ! RI ! -4.24 - 0.32 ! 1 0.99 1 m M A A j i 1 ! R2 1 0 4 O o -1.27 i 0 9 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 •90 2 h 2a 7a © — 14a 75 60 45 3 0 :s '0 g(f(hz)) Fig. 5.39 Bode plots for 1018 steel in 5wt°oNHi -0 2O /o NaOH with 10 mM L-aspartic acid (AA) monosodium salt for 2 weeks at RT. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase Angle (Degree) 136 Rs R1 C P E l -I I CPE2 R2 (a) 4 .Q 0 0 o 4. 5 - 5 0 - 5 5 - 60 - 6 5 — iafit 3 • 5 1 J !o^f(Hz)) (b) \ _ X. N V - R i O c i -0 R 2 — C2 -v -v * x G V '- V 90 75 60 u v § - 1 5 I < 30 1 a . 15 - : 4xio4 ; 2xio^ u_ : oxio* o "5 _o 3 OxiG - 6 0x10' log(t(hrs)) (c) Fig. 5.40 (a) TTC model [159.160], (b) experimental and fit data for Id, 5d and 14d spectra in Fig. 5.30. and (c) fit parameters R . and C vs. time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.4.2 DC polarization results The DC polarization experiments consisted o f cyclic anodic polarization and cathodic polarization conducted on 1018 steel in the baseline solutions with and without AA, GPA. PAA or PGA. The anodic cyclic polarization curves are shown in Fig. 5.41a. while the cathodic polarization curves are summarized in Fig. 5.4lb. As seen from the anodic scan, addition o f inhibitors to the baseline solution moved E ^ -o rr in the positive direction to different extent. The passive current density ip a® remained relatively unchanged at about 2 [xA/cm* for solutions with PAA and PGA and was reduced to about 1 pA/cm2 for solutions with AA and GPH. On the return scan ip a ss was much lower in all solutions due to passive film thickening during the anodic scan. The lowest values were observed in the presence o f GPH (Fig. 5.41 a). The data within the range o f E ^ ± 30 mV in the anodic scan were fitted with the software POLFIT developed at CEEL/USC [138], The fit results o f Eom -. ba. bc, B. ico n - and the calculated E are listed in Table 5.19. AA has the highest E (59 %). while PAA. increased the corrosion rate o f 1018 steel in the baseline solution. The cathodic polarization curves in Fig. 5.41b reached a broad region at almost the same polarization potential (about - 600 mV vs. SCE). where the current density it was limited by mass transport o f oxygen. With the addition o f the four inhibitors, iL was reduced from about 25 jiA/cm* to 20 pA/cm2 for GPH and to about 15 pA/cm2 for AA. PAA and PGA. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 The simple shape of the anodic (Fig. 5.41a) and cathodic (Fig. 5.4lb) polarization curves suggests that the additives to the baseline solution did not undergo oxidation or reduction reactions on the mild steel surface during polarization in a wide range of potentials. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 3 500 0 8 0 0 0 700 0 600 0 500 0400 3 300 0 200 0 100 0000 0 100 0 200 0 300 — Baseline — C * — Baseline - 5GmM GPH Baseline * lOmM A A — Baseline * 20 com P A A Baseline 20 ppm PG A i £ 012 :e o i: 1E010 IE 009 IE 008 1E0C7 1E306 1E005 1E004 1E0C3 1E002 i (A/'cnO (a) 0 100 0200 0300 _ 0 400 — ij U 3 -n 0500 • > U i 0 600 ■ 0 700 — 0 S C O - ■ 0 90C IE 010 IE 009 leooa IE 007 IE 006 IE 005 IE 004 (b) Fig. 5.41 1018 steel in different solutions after 2-hour immersion at RT; (a) cyclic anodic polarization curves, (b) Cathodic polarization curves. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 4 0 Table 5. 19 Fitting results from potentiodynamic polarization curves for 1018 steel in different solutions at RT. | Solution ! j i Baseline (B L i | BL - 10 mM ,\A l BL - 50 mM G PH BL - 20 j ppm PAA t 1 BL - 20 j ppm PGA | I | L<»r (m V vs. ■ j SC E) | -249 1 i *214 1 -124 -241 -232 1 j b . | im V decade) j 40 r I " 106 62 i | be (m V decade) j 20 ! 1 5 1 5 57 46 | B im M j 6 j 5 5 16 n I w ( .V c m : ) | 5.9 • 104 ! 2 .4 . I04 3. 9 . 10 * . I 4 . I0‘ 5 1 . 104 1 E l % ) ! \ A ! ! 59 54 V \ 13 Note. The inhibition efficiency E = ( I - ieo rr'h b V Wrb a ' sclm c) « 100%. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 1 6. DISCUSSION 6.1 Evaluation o f cerium salts Cerium salts were used in different ways to form protective films on mild steel. These films were different in coverage, porosity, thickness and adhesion and therefore provided different extents o f corrosion protection. The properties o f these films were evaluated in 0 .1 N NaCI and the baseline solution at RT and 100 °C. The formation methods and corrosion protection properties o f the cerium (hydr)oxide films will be discussed in the following. 6 .1. 1 Cerium salts as inhibitors The corrosion inhibition efficiency o f cerium salts for mild steel depends on the corrosive medium. pH o f the corrosive medium, anions o f cerium salts and exposure time. The inhibition mechanism o f REMSs. especially cerium salts, has been investigated with various techniques [69-72], It is generally agreed that REM cations are cathodic inhibitors functioning through the precipitation of REM (hydr)oxide films on the substrate due to local pH increase at or near the substrate as the result o f oxygen reduction at cathodic sites. The REM (hydr)oxide film behaves as a barrier to the transport o f oxygen to stifle the oxygen reduction reaction. .\s a result, the corrosion current density drops accompanied by a shift o f OCP in the negative direction. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 6 .1.1.1 Corrosive medium dependence Cerium nitrate was effective in inhibiting the corrosion o f mild steel in 0 .1 N NaCl (pH adjusted to neutral after addition o f inhibitor) and baseline solution at 100 °C. It performed even better than dichromate as illustrated in Tables 5 .1 and 5.3. The inhibition efficiency was higher in the former solution (E = 93%) than in the latter solution (E = 56%). The higher corrosion inhibition efficiency is due to the fact that the solubility o f cerium cation (III) is much higher in N'aCl than in the alkaline baseline solution, which makes the formation o f cerium (hydr)oxide film easier on mild steel in the former solution. It was observed that after 2 days exposure, a brown-yellow film was formed on mild steel in 0 .1 N NaCl. while no visible film was found on mild steel in the hot baseline solution. 6 .1.1.2 pH dependence The pH o f a corrosive medium plays a very important role in the effectiveness o f cerium salts as inhibitors, as proved in Table 5.1. When pH o f 0.1 N NaCl was 2.5. cerium nitrate was practically ineffective with an inhibition efficiency o f only 7 %. But when pH was adjusted to 7.0 with borate buffer solution, inhibition efficiency increased drastically to 93 %. The borate anions might have some contribution to the efficiency increase, since borate is also a known inhibitor for mild steel. But the major contribution would be ascribed to the uniformity and integrity o f the blue- colored cerium (hydr)oxide film formed on mild steel. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 3 Hinton er al. [54] also observed the formation o f a blue-colored film on mild steel immersed in CeCb-7H20 (> 50 ppm) inhibited soft tap water. In the literature [50.51], pH values above 6 were suggested for solutions inhibited with REMSs. The explanation would be that the anodic dissolution rate o f mild steel was smaller at the neutral pH than at pH = 2.5. therefore, a film with majority cerium (hvdr)oxide could be more easily precipitated on mild steel. 6.1.1.3 Cerium salt anion dependence It was reported that o f all studied REMS cations, cerium cation (III) was most effective in corrosion protection o f A17075 in NaCl [66.67], while o f all the cerium salts studied. CH?COO‘ anion was ranked the best in the corrosion inhibition for A17075 in NaCl [54], The efficiencies o f both NO3' and CH3COO' of cerium salts for corrosion inhibition of mild steel in 0 .1 N NaCl are listed in Table 5.1. It can be seen that at the same concentration and comparable solution pH values (7.0 for 20 mM CefNChh, and 6.7 for 20 mM (CHsCOObCe). NO3' (E = 93 %) was much more effective than CH3COO' (E = 41 %). If Eq. 5-2 holds for mild steel in NaCL then the ranking o f cerium salt anions based on inhibition efficiency at the same concentration and solution pH would be: NO3* > CHjCOO > CIO4 > SO42 ' > Cl* 6 -1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 The high efficiency o f NO3’ could be due to the following aspects. First, the solubility o f CefNChb is much higher than that o f (CFfjCOOjjCe in 0 .1 N NaCl. Secondly, the oxidizing nature o f NO3' facilitates the oxidation of cerium cation from the HI to the IV valence state, and speeds up the formation o f a cerium (IV) (hydr)oxide film, which is more stable than a cerium (III) (hydr)oxide film [69], NO3' also helps the formation o f the more compact oxide film on steel that has been widely used in industry as a pretreatment procedure for rust-proofing [22], The compact oxide film inhibits the etching o f the steel substrate, i.e. the anodic dissolution o f iron, and assists the formation o f a more uniform cerium (hvdr)oxide or a mixture o f cerium / iron (hydr)oxide film. On the other hand. CFHCOCT is non oxidizing and has an ability o f forming complex compounds with most metals [82], which could limit the availability o f cerium cations for formation o f oxide films. REM nitrate salts were also highly recommended when the REMSs were used as corrosion inhibitors in aqueous systems [50]. 6.1.1.4 Time dependence The corrosion inhibition efficiency o f cerium salts for mild steel in 0.1 N NaCl is strongly dependent on the exposure time, as shown in Fig. 5.2. With prolonged exposure time. Rp decreased from I.77x I04n-cm 2 after 2 hours to 8.87* 103 Q-cm2 after 24 hours and to about 3 5* 103 Q-cm2 after 48 hours. The second time constant observed at intermediate frequencies as illustrated in Fig. 5.2 after 48 hours is believed to be the result o f the deterioration o f the cerium (hydr)oxide film. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 On the contrary, the efficiency’ o f cerium nitrate in the hot baseline solution increased with immersion time, as demonstrated in Fig. 5.16 and Table 5 .4. The corrosion rate dropped from 1.2 pm/year at 2 hours to 0.1 pm/vear at 48 hours. The 2-hour impedance spectra were unstable showing the strong interaction between the cerium nitrate inhibited solution and mild steel. The higher Rp value obtained in the cerium nitrate inhibited baseline solution than that obtained in the blank baseline solution (Table 5.3) suggests effective corrosion inhibition by the small amount of dissolved cerium nitrate. In the hot alkaline baseline solution (pH 13.3), a compact magnetite film can be readily formed on mild steel (Fig. 2 .1). The incorporation of cerium cations into the magnetite film modifies its properties to more effectively impede the penetration of hydroxyl ions across the film, thus resulting in a higher polarization resistance with prolonged immersion time (Fig. 5.16). This protection mechanism could be similar to that provided by the 1.24 at% cerium ion-implanted on stainless steel surface [161], However, the concentration o f cerium cations incorporated into the oxide film on mild steel in the hot baseline solution are presumably very low due to the low solubility o f cerium cation in the alkaline solution. For mild steel in NaCl inhibited by cerium nitrate at RT. C f can easily penetrate the cerium (hvdr)oxide film and attack the substrate preferably at anodic sites and form the yellow-brown cerium iron (hydr)oxide. With longer exposure time, the film losses its integrity and causes the inhibition efficiency decrease o f cerium salts. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 4 6 6.1.2 Chemical conversion coatings A cerium (hydr)oxide film was formed on mild steel through immersion in cerium salt solutions at RT. The properties of the film depend on the cerium salt anions used (Table 5.2). NCb* gives the best protection efficiency (E = 88%) after 2 days immersion in 0 .1 N NaCl. followed by CH3COO* (E = 43%) and Cf (E = 28%). Interestingly, this sequence also agrees with the ranking in Eq. 6-1. Comparing Table 5.2 and Table 5.1, it can be noticed that without the presence of Cl* the cerium (hydr)oxide film formed in cerium salt solution can provide better and longer protection for mild steel. That implies that a more uniform and compact cerium (hydr)oxide film was formed without the involvement of Cl*. Fig. 5.4 illustrates the time dependence of the protection efficiency of cerium (hydr)oxide film formed by conversion coating. The reason for the efficiency decrease with time would be the film deterioration caused by Cl* penetration, similar to that for the efficiency decrease for the film formed in NaCl inhibited by cerium nitrate, as discussed before. Similar phenomena were observed by Rudd et al. [89] in the study o f corrosion protection on pure magnesium and a magnesium alloy by rare earth (cerium, lanthanum and praseodymium) conversion coatings tested in weak alkaline borate buffer solution. The non-lasting corrosion protection for mild steel and magnesium by REMSs chemical conversion coatings could also be attributed to the porous nature o f the oxide films formed during treatments. The oxide film is actually a (hydr)oxide Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. L 4 7 mixture o f REM and substrate metal since some amount of corrosion o f the substrate metal is needed during the chemical conversion process [54], The corrosion o f the substrate is in the form o f uniform corrosion, not like the case in the REMSs chemical conversion on A 1 and alloys, in which localized corrosion dominates and the conversion film is formed in a long period o f time. e.g. 100 hours, leading to the formation o f compact REM (hydr)oxide film [54,77], 6.1.3 Electrodeposition Cerium nitrate was used in the formation o f the cerium (hydr)oxide film via electrodeposition. The properties o f the film depend on cathodic current density and deposition time. High current density resulted in a fragile film. Short deposition time would form a porous film as evidenced in Fig. 5.5. which shows the impedance spectra with a diffusion tail at low frequencies for the 300-second deposition film. With longer deposition time at an appropriate current density, the film became more insulating, therefore more negative external potentials were needed to maintain the galvanostatic deposition condition (Fig. 4 lb). For a film deposited on mild steel in 900 seconds, the Rp value was 2.4* 104 Q-cm2 after 2 hours exposure in 0 .1 N NaCl. much higher than that for untreated mild steel ( 1.32* 103 Q-cm2 ). and even the dichromate (8.34x I03 Q-cm2) (Table 5.1). However, this film could not withstand the CF attack, and Rp value dropped to 3.7x 103 Q-cm2 after 24-hour exposure. The color o f the cerium (hydr)oxide film formed by electrodeposition was light yellow indicating a mixture o f (hydr)oxide with a small amount of cerium (IV) and a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 148 majority of cerium (III). Within a 24-hour exposure to 0 .1 N NaCl. the film lost its integrity and became a powder or slurry like substance loosely covering the mild steel substrate. Wilson et al. [52] reported that cerium (hydr)oxide coatings formed on Al alloys via cathodic treatment in cerium chloride or nitrate solution also suffered from a lack o f durability due to coating blistering or cracking. These drawbacks together with the special equipment and process requirements in electrodeposition have limited it from practical applications [52], 6.1.4 Cerating Cerating is a fast chemical conversion process for the formation o f cerium (hydr)oxide films, first developed, named and patented for the protection of Al alloys by Hinton and Wilson [95]. The uniqueness o f this method benefits from its fast speed, convenient application and the resulting protective golden yellow cerium (IV*) (hydr)oxide film, referred to as cerated layer or film in differentiation from the cerium (hydr)oxide films formed by the other methods discussed above. The cerating solution consists mainly o f cerium salts and strong oxidizing agents, most preferably cerium chloride and hydrogen peroxide (H ;0;). During the cerating process, gas evolution is inevitable due to the side reaction at cathodic sites (H2) and anodic sites (O2) (Eq. 2-10) or catalytic decomposition o f H2O2 at the metal surface (Eq. 2-11), which introduces pores in the cerated layer. A crack network is formed when the thickness o f a cerated layer is greater than 0 .8 pm [96] possibly due to the brittle nature and dehydration o f the cerated layer. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fig. 5.7 clearly shows the pores or mounts formed due to gas evolution and a fully developed crack network in a thick cerated layer formed on mild steel. EDS results (Fig. 5.8) indicate that the cerated layer is a mixture o f cerium and iron (hydr)oxide, contrary to that obtained on Al alloys, which is mainly ceria (CeC>2) [94], The corrosion protection efficiency o f the cerated layer on mild steel depends on many factors o f the cerating process as well as on the corrosive medium. The optimization o f the cerating process to obtain the best corrosion protection will be discussed in section 6.2. The corrosion protection behavior and mechanisms o f cerated layers in different corrosive media, i.e. 0 .1 N NaCl at RT. baseline solution at RT and 100 C C will be discussed in the following. 6.1.4.1 0.1 N NaCl at RT In aerated 0 .1 N NaCl. the cerated layer cannot provide any protection for the mild steel substrate as demonstrated in Fig. 5 6a. The simulated results for the 2-hour experimental data indicate that Rp ( l.5x 103 Q-cm2) was comparable to the value for untreated mild steel in the same solution ( I ,32x 103 Q-cm2 . Table 5 .1). while the C value was 3 .65 x I O'2 F/cm2 . more than 25 times larger than the capacitance for untreated mild steel ( I 42x 10‘ J F/cm". Table 5.1). The large capacitance value could be due to the roughness or porosity and high dielectric constant of the cracked cerated layer. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 5 0 To further investigate the electrochemical behavior o f the cerated layer in 0 .1 N NaCl, DC cathodic polarization curves were measured in 0.1 N NaCl for mild steel untreated and cerated for 20 minutes in solution of 12.5 g/L CeClj and 2.5 wt% HjOi. The polarization curves are given in Fig. 6 .1. After cerating treatments, the OCP was shifted slightly towards positive (noble) direction: wide Tafel regions were observed, and the limiting current densities were increased. One reason for these changes could be peroxo ions (O') from the H;Oi incorporated into the cerated layer during cerating treatment and/or from the unreduced cerium peroxo complex species, such as C e(0:)‘‘ (Eq. 2-9). However, peroxo ions are very active, and can be easily reduced to oxygen ions (0'~) during drying after cerating treatment. Therefore, high peroxo concentration (leading to high limiting current density) would not be expected. Another possible reason could be the roughness o f the cerated layer that increases the actual surface area and causes a higher limiting current density, higher film capacitance C. but smaller Rp if the nominal surface area is used in a calculation. 6 .1.4.2 Baseline solution at RT In the baseline solution at RT. the differences in impedance behavior between cerated and untreated mild steel were not as pronounced as those in 0 I N NaCl. But overall, cerated mild steel had lower Rp and higher C, as illustrated in Fig. 5.13. which could be due to the strong passivation ability o f hydroxyl ions in the alkaline (pH 13.3) baseline solution at RT resulting in the formation o f a protective magnetite Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 151 0 50 - O SO _ B a rr = -0 6 1 ! V E ccrr= 0 6 2 7 V '* i ^ a * - A ' " ‘ ■v *> —r \ - , UJ O > -J — ■ — CJ Z A c n 3 a 2 ! * > a a c . a \ ji a _____________ ~ A a a ± - - U n tre a te d d » \ A - A 1 30 0 90 4. w O A Cerateo IZ 5 «t% nZCZ) A A : o e oc? : o e coe : c e c o s : c e c c a : c e 0 0 3 Current density (A/crm ) Fig. 6 .1 Cathodic polarization curves for 10IS steel with different treatments in 0 .1 N NaCl at RT after 2-hour immersion. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 152 film on mild steel (Fig. 2 .1). The slope of - 0.93 indicates the formation o f a compact and protective film, which could be magnetite, while the slope of - 0.25 implies a porous and less protective film, which is the cerated layer. In the presence o f dichromale inhibitor, the slope was - 0.5 (Eq. 5-4a), suggesting the formation of a porous film that acts as a diffusion barrier. 6.1.4.3 Baseline solution at 100°C In the hot baseline solution, cerated mild steel experienced dramatic changes within the first 24-hour exposure as depicted in the impedance spectra (Fig. 5.17a) and the Rp and C vs. time plots (Fig. 5 .17b). The slope o f the log ( 1/Rp) vs. log (t) plot was - 0 73 (Eq. 5-6). indicating a change of the cerated layer to a much more compact and protective film in the hot baseline solution from the porous film in the baseline solution at RT (slope o f - 0.25). The two orders of magnitude decrease in cerating film capacitance C (Fig. 5.17b) during the first day of exposure indicates a drastic decrease in surface area or decrease in dielectric constant. The SEM images before and after 24-hour exposure in the hot baseline solution (Fig. 5.18) suggest that the decrease o f the real surface area would be more reasonable, since almost all pores and cracks were sealed within the first 24-hour exposure in the hot baseline solution. 6.2 Optimization o f the cerating process The cerating process is a multi-variable or multi-factor dependent process. The studied variables included concentration o f solution constituents (CeClj. H2O2), solution pH, additives, cerating time, pretreatments, post treatments (sealing) and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. solution aging. These variables were investigated in the hot baseline solution or first screened in the RT baseline solution and then evaluated in the hot baseline solution through factorial designs or single variable experiments. The ultimate goal was to obtain a protective cerated layer on mild steel in the hot baseline solution with as high as possible Rp and as low as possible C. The effects of these variables on the properties of the cerated layers will be addressed and discussed in the following. 6.2.1 Concentration o f solution constituents The effects of the concentration o f CeCH and H2O2 were first studied with a 2s factorial design (Tables 4.1. 5.5 and 5.6). followed by a 2Z design (Tables 5.9 and 5.10) and finally two 32 designs (Tables 5 .11 and 5.12. Figs. 5.20 - 5.23). The cerated layer properties (responses) were evaluated in the hot baseline solution. Through all the factorial designs, it was found that Rp had a weak dependence on CeClj. but a strong dependence on H2O2 concentration. This conclusion was proven by the small main effect o f CeCL (E fi) and large main effect o f H2O2 (E f 2) on Rp in Table 5.6 and Table 5.10 as well as the 3D profile and 2D contours for Rp (Fig. 5.20 and Fig. 5.22). The weak dependence on CeClj concentration, but strong dependence on H2O2 concentration for Rp could be due to the fact that the cerating process was mainly controlled by the strong oxidizing agent H2O2. Low H2O2 concentration would not completely oxidize cerium cations from III to IV valence therefore the cerating process would be slow. High H2O2 concentration could leave too much H2O2 in the cerating solution and cause vigorous gas evolution resulting in a heavily Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 154 porous and cracked cerated layer. The capacitance C was found to depend on both CeCh and H2O2 concentrations, as evidenced by the relatively large Efi and Ef2 values for C (Table 5.6 and 5.10) and the 3D profile and 2D contours for C (Fig. 5.21 and Fig. 5.23). This observation could be attributed to the fact that higher concentrations o f CeClj or H2O2 lead to the formation o f a rougher and more porous cerated layer. The influences o f CeCb and H2O2 concentrations could also come from their acidifying effect. Higher concentration of CeClj or H2O2 causes lower solution pH, which has a strong influence on cerating. Through factorial design 32 -I. two trends were observed to obtain higher Rp (Fig. 5.20): one was to decrease the concentration o f H2O2- but increase the concentration of CeCli. the other was to decrease the concentration o f CeCh, but increase the concentration o f H2O2 The differences between the two trends are that the first one leads to a higher Rp, but also a higher C. therefore the second trend leading to a higher Rp and a smaller C was preferred. After another factorial design (32 -U), the optimum conditions were identified, i.e. 2.3 g/L CeClj and 4.4 wt% H2O2, under which a maximum Rp and lower, though not minimum C can be obtained. This result was justified by the good agreement between the prediction values (Rp = 2.09x 105 Q-cm2 . and C = 5.96x I O '4 F/cm2) and the experimental data (Rp = 2.56x I05 Q -cm \ C = 6.41 x IC T 1 F/cm2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6.2.2 Cerating time The effect o f cerating time on the cerated layer properties was evaluated in the hot baseline solution with 23 (Table 5.5) and 2"-I (Table 5.7) factorial designs. It was observed that a longer cerating time had a beneficial effect on both Rp and C. as substantiated by the high and positive main effect (E r) on Rp and C (Tables 5.6 and 5.8) for mild steel in the hot baseline solution. This finding could be due to the fact that longer cerating time results in a thicker cerated layer. A thicker cerated layer had a more heavily cracked surface, however the crack network can be sealed during hot baseline exposure as discussed earlier. This result was further confirmed by the 48- hour impedance spectra (Fig. 5.24) collected for mild steel samples exposed to hot baseline solution after cerating treatments in 2.3 g/L CeCW and 4.4 wt% H ;0; for 20 minutes and 5 minutes. The cerated layer obtained with the 20 minutes treatment had a higher Rp (2.56 < 105 Q cm ')but also a higher C (6 41 < I O ’ 4 F/cm: ) due to the rougher surface than the cerated layer obtained with the 5 minutes treatment (Rp = 1.24* 105 Q-cm% C = 4.03 < I O '4 F/cm:). The effect o f cerating time on the cerated layer properties evaluated in RT baseline solution was different from the above observation in the hot baseline solution. It was found that a shorter cerating time produced cerated layers with higher Rp (Table 5.14) and lower C (Table 5.15), as demonstrated by the experiments HCI5 and HC16. which were different only in cerating time (Table 5.13). The explanation could be that shorter cerating time results in thinner cerated layers with fewer cracks or pores. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 5 6 Since the pores and cracks cannot be sealed at RT in the baseline solution, the cerated layer with fewer pores and fewer cracks contributed to the higher Rp and lower C. 6.2.3 Cerating solution pH The effect o f solution pH was evaluated in the range o f 1.6 - 2.5 for cerated layer in the hot baseline solution. It was found that the highest Rp and lowest C were obtained at pH 2.2 as shown in the impedance spectra (Fig. 5.29a) and fit results vs. pH plots (Fig. 5.29b). The reason for these results could be that at low pH the dissolution rate o f the mild steel substrate is high, which on the one hand introduces more H> gas during the cerating process, and on the other hand causes too much iron hydroxide to be incorporated into the cerated layer thereby degrading its quality. At high solution pH. cerium (hydr)oxide could start to precipitate even before the cerating treatment which could give rise to the formation o f a loose and non-uniform cerated layer. At pH 2.2. the two pH effects compromise and produce a cerated layer with highest Rp and lowest C. 6.2.4 Additives The additives used in the cerating solution include NaNOj, Triton X-IOO and (CHjCOOhPb. The original purpose o f adding NaNCh was to scavenge oxygen produced during the cerating treatment at anodic sites (Eq. 2-10). but later it was found that NaNO- could also be used to adjust and stabilize the pH o f cerating solution. Fig. 6.2a demonstrates the relationship between the concentration o f Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 5 7 NaNO? and solution pH. The solution pH did not change with aging time indicating good solution stability after addition o f NaNOi. From Fig. 6.2a it can be noticed that solution pH can be easily adjusted to 2.2 with an addition o f I g/L NaNO;. Triton X- 100 is an effective surfactant, and was used at a concentration o f 4 ppm to prevent clinging o f gas bubbles to the mild steel surface during cerating. (CHjCOOhPb is a widely used solution stabilizer in the electroless plating industry, here it was employed at a concentration o f 3 ppm to suppress the decomposition o f cerating solution. Beneficial effects o f additives on the properties o f cerated layers were observed from SEM results collected from experiments Actanel and Actane2 that were designed to study the effects o f the additives (Table 5.13). With the additives in the cerating solution (experiment Actane2). the cerated layer became smoother with fewer cracks and smaller pores than that for Actanel (compare Fig. 5.28 with 5.26), which resulted in higher Rp values and lower corrosion rates for Actane2 in the baseline solution (Tables 5.14 and 5.16). 6.2.5 Pretreatments The purpose o f pretreatment was to produce a clean, active surface to facilitate the formation and adhesion o f cerated layers on mild steel. Experiments Micro I and HCIl were designed to compare Pre-Treat I and Pre-Treat2 (Table 5.13). The phase angle spectrum at low frequencies for Micro I (Fig. 5.25a) suggests a possibly porous oxide film was left after Pre-Treat 1. The porous oxide film could benefit Micro I in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 158 3.0 ____________________________________ ____________ E f f e c t o f N a N 02 c o n c e n t r a t i o n . 2 8 - ' 025 3 / L 0.5 g/L —B — 1 g /L 2.6 • O 2 g / L (a) 1 2.4 2.2 2.0 10 20 30 Aging time (Minutes) 3.5 3.3 3.1 (b) ^ 2.5 wt% H 2O2 ♦ Ced3 with different concentration. ------------2 g /L V" 3 g /L E r 4 g A - o . c 2.9 2.7 2.5 — - v +• i t — 10 2 0 3 0 4 0 5 0 6 0 Aging time (Minutes) (C) r a. 3 5 3.3 - 3.1 • •J 2.9 2.7 - 2.5 — 3 g/L CeCl3 > H 2O2 wdh different concentration. 1.5 wt% -> 2.5 wt% --g -- 3.5 w t% n * ■i- 10 2 0 3 0 4 0 5 0 6 0 Aging time (Minutes) Fig. 6.2 Dependence o f cerating solution pH on aging time as a function o f (a) concentration o f NaNO^ additive, (b) concentration o f CeCb and (c) concentration o f H2O-. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 159 the first week o f exposure in the baseline solution with a slightly higher Rp. therefore lower corrosion rate (Table 5.14) and smaller time law parameter Ar (Table 5.16). But in the long run. HCll could perform better than Micro I due to its lower time law parameter Br (- 0.48 vs. - 0.41) and better cerated layer quality as indicated by its low capacitance C (Table 5.15). Experiments HC12 and Actanel differed only in pretreatments (Pre-Treat2 vs. Pre- Treat3) (Table 5.13). However HC12 outperformed Actanel in every aspect, i.e. a higher Rp leading to smaller corrosion rates (Table 5.14). lower C (Table 5.15). Ar and Br (- 0.25 vs. - 0.21) and therefore smaller At and VH ; (Table 5.16). In summary. Pre-Treat2 was the best pretreatment processe studied, which could be due to the strong etching ability o f the HCl solution. Micro and Actane solutions contain some inorganic or organic substances that could become residues on the mild steel surface in the form o f insoluble oxides or adhesives after treatment and thereby prevent the formation o f a uniform and sound cerated layer. 6.2.6 Post treatments Post pretreatment (sealing), the last, but maybe the most important step in the whole cerating treatment on mild steel was aimed to seal the pores and cracks in the cerated layer. In the hot baseline solution, the cerated layer sealed itself within the first 24- hour exposure resulting in a remarkable increase in performance, as stated and discussed in section 6.1.4.3. This process is termed with self-sealing. In order to find Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 0 additional means for improvement o f the cerated layer by sealing treatment, five methods were investigated, i.e. cathodic polarization in cerium salt solution as well as Post-Treat 1 to Post-Treat4. as described in chapter 4 and in Table 5.13. Cathodic polarization in cerium salt solution did not provide an improvement over self-sealing. Considering its difficulty in practical applications, this method is not recommended. Post-Treat I to Post-Treat4. as represented by experiments ActaneS, Actane5. Actane6 and Actane7. were first screened against Actanel (without post treatment) and ranked in the RT baseline solution. Promising methods were further evaluated in the hot baseline solution. Based on the corrosion rate data in Table 5.14. it is clear that after these post treatments, the performance o f the cerated layer was significantly improved after one-week exposure, though Actane6 (Post-Treat3) had a larger capacitance and a lower 2-hour Rp than A ctanel. From the time law parameters (Ar and Br) and the predicted long-term thickness loss At and volume o f gas evolution Vh2 (Table 5.16), the ranking in corrosion protection efficiency from high to low for the tour post treatments are: Post-Treat I - Post-Treat4 > Post-Treat3 > Post-T reatl 6-2 Post-Treat I and Post-Treat4 were the most efficient methods in sealing the cerated layer. However it can be noticed that the fit time law parameters (Ar and Br) were Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 161 very different for these two treatments (Table 5.16). Post-Treat 1 (Actane3) had a smaller Ar (0.89 jim/year) and a larger Br (- 0.25), while Post-Treat4 (Actane7) had a large Ar (10.12 jim/year) and a smaller Br (- 0.56). The differences in the time law parameters would imply different sealing mechanisms for Post-Treat 1 and Post- Treat4. which employed silicate and molybdate. respectively. Silicate is an effective film-forming compound and is widely used as a precipitation inhibitor for steel [104], During Post-Treat I (Na;SiOi * 9HiO solution), silicate could be incorporated into the cerated layer to form a layer o f cerium, iron and silicon oxides, such as CeOc*Fe?0 4 -SiOt. which effectively sealed the pores and cracks and resulted in a small Ar. The SEM images in Fig. 5.27 in comparison with Fig. 5.26. clearly demonstrate the high sealing effectiveness of Post-Treat I. However, since silicate is a non-oxidizing anion, it could not further react with the mild steel during later exposure in the baseline solution. Therefore, the time dependence o f the sealed cerated layer was weak, as characterized by a larger Br value (small absolute value). Molybdate is an oxidizing compound, and an effective anodic inhibitor for many metals and alloys including steel [104], During Post-Treat4 (Na2MoC>4 * 2 H2O). due to its lack o f film forming ability, molybdate did not seal the cerated layer with a consequent large Ar, but it could be effectively incorporated into the cerated layer in the high concentration (10%) post treatment solution. The oxidizing molybdate might further interact with mild steel or even the cerated layer to form a mixture o f Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ce, Fe and Mo oxides during exposure to the baseline solution resulting in the observed stronger time dependence o f the corrosion rate and a smaller Br. The sealing efficiency o f Post-Treat I was further investigated in the hot baseline solution. EIS results (Fig. 5.31) indicated that after 48-hour exposure, Post-Treat I increased Rp to 1.39 * 105 Q-cm* from 9.53 < 104 Q-cm*. providing a 46 % improvement compared to self-sealing. 6 .2.7 Aging o f cerating solution Hughes et al. [97] first reported the importance o f cerating solution aging in their study o f (H2O2) accelerated cerium-based conversion coatings on AI alloys and suggested that a 30-minute solution aging is necessary prior to use in order to achieve the optimum coating activity The beneficial effect o f cerating solution aging on improvement o f the cerated layer properties was also observed for mild steel in the hot baseline solution, as shown in Fig. 5.30. With 30 minutes aging before cerating, the cerated mild steel sample had a higher Rp o f 1.73 < 105 Q-cm* than that for a cerated steel, but without aging (9.53 < 104 Q-cm*), showing 82 % improvement after aging. The reason for the aging phenomenon is not clear. Hughes et al. [97] thought that aging is related to the formation o f a cerium (III) peroxo complex, which may be a necessary intermediate in the cerating process through progressive substitution o f water by H2O2, as described in Eq. 2-5. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 3 During the study o f the cerating process, it was observed that aging o f the cerating solution was accompanied by a pH decrease, especially in the first 30 minutes, as illustrated in Fig. 6.2b and Fig. 6.2c. if concentrated H20 2 was first mixed with CeCl* solution, and then diluted to the final volume. In such a case, the cerating solution was golden yellow, since the cerium (III) was oxidized to cerium (IV): 2CeJ* + 2HzO + H20 2 -> 2Ce(HO)22' + 2FT 6-3 Therefore, aging was associated with the oxidization process o f Ce (III) to Ce (IV) accompanied by the decrease o f solution pH. However if H20 2 was first diluted and adjusted to low pH (< 2.0). and then mixed with CeCU solution, the obtained solution was colorless with practically no pH change during the following 30 minutes aging. In this case. Ce (III) was not oxidized to Ce (IV) and was present as a Ce (III) peroxo complex. It can be concluded that in this case the aging could be dominated by the reaction in Eq. 2-5, as suggested by Hughes el al. [ 104], The latter case could be more preferable, since during cerating only the Ce (III) cations near the cathodic sites are oxidized to Ce (IV). and then precipitate as cerium (IV) (hydr)oxide due to the local pH increase. This could result in a more stable cerating solution and produce a more compact and uniform cerated layer due to the controlled availability o f Ce (IV). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6.3 Evaluation o f yttrium salts Yttrium salts were evaluated in a similar way as in the study o f cerium salts. Evaluation results indicated that yttrium salts were less effective than cerium salts in the test solutions. The impedance spectra for yttrated and cerated mild steel in 0.1 N NaCl at RT for 24 hours and in the baseline solution at RT for 7 days are compared in Fig. 6.3a and Fig. 6.3b. respectively. The impedance spectra for yttrium sulfate and cerium nitrate as inhibitors in the hot baseline solution are compared in Fig. 6.3 c. These spectra demonstrate that higher Rp and lower C values were obtained in the presence o f cerium than yttrium. Yttrated layers had wide cracks, and did not adhere well to the mild steel substrate (Fig. 5.32). therefore providing practically no corrosion protection for mild steel in NaCl and the baseline solution at RT. as indicated by the low Rp values (Fig. 5.34b and Fig.5.35b) and large (small absolute) slopes o f plots o f log ( 1/Rp) vs. log (t) plots in Eq. 5-1 la and 5-12a. After 48-hour exposure to the hot baseline solution, the yttrated layer could not be as effectively sealed as in the case o f the cerated layers. Corrosion products were observed on the yttrated layer (Fig. 5.33 a) as well as in the hot baseline solution after the experiment. The difference in the corrosion protection efficiency between yttrium salts and cerium salts could be associated with the fact that yttrium cations exist only in the [II valence state, therefore can only act as a precipitation inhibitor, while cerium cations Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 5 (b) -> Ceratea v ttra te c ■60 * ieeeeees> •1 0 1 2 log((f (Hz)) 9066- 0 4 5 (C) -fcj— I mM Y 2(S04.)3 o 5 r r ’M CefN0 3 ) 3 -7 5 ^ ------------- o 0 ) 60 ® 0 1 2 log(f(Hz)) Fig. 6.3 Bode plots for mild steeL. (a) in 0 .1 N NaCl at RT for 24 hours after cerating and yttrating, (b) in baseline solution at RT for 7 days after cerating and yttrating, and (c) in the hot baseline solution inhibited by yttrium sulfate and cerium nitrate after 48 hours. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 6 can exist in Eli valence as well as the strongly oxidizing IV valence state which is able to form the more protective IV valence ceria (CeO?) film than the III valence cerium oxide (CezOj). 6 .4 Evaluation o f organic inhibitors Four organic inhibitors for mild steel in alkaline environments were evaluated in the baseline solution at RT. GPH consists o f CH2OH (methylol) and PO4 (phosphate) groups, while A A PAA and PGA consist o f COOH (carboxyl) and NH; (amino) groups. Their inhibiting ability mainly comes from the adsorption o f PO4 or NH2 groups on the steel surface. The concentrations for GPA and .AA were determined according to literature results [113.114] to obtain the best inhibition efficiency. For PAA and PG A a low concentration o f 20 ppm was employed due to cost considerations [126], Due to the high passivation ability o f hydroxyl ions in the baseline solution at RT. a passive film o f magnetite can readily form on mild steel. However the film formation process can be affected by the addition o f organic inhibitors, even at low concentration o f 20 ppm for PAA and PGA due to competitive adsorption onto steel surface between the organic inhibitor and hydroxyl anions resulting in the formation o f an inhibitor-adsorption film. The inhibitor-adsorption film could be less protective than the passive magnetite film, if it does not fully cover the mild steel surface at an early exposure period. The 2h-impedance spectra for all systems are summarized in Fig. 6.4, which clearly demonstrate the decrease o f the impedance at low frequencies Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 7 due to the addition o f organic inhibitors. The fit parameter aR (Table 5.18) further supported this observation. With the presence o f the organic inhibitors in the baseline solution, aR increased to different extents, which indicates the increase in the corrosion rate in the early exposure period. However, the inhibitor-adsorption film could become more or less protective than the magnetite film with increasing exposure time as indicated by the fit parameter bR. Addition o f GPH or PAA resulted in a similarly more protective passive film (bR = - 0.99 for GPA. - 0.97 for PAA vs. - 0.86 for BL) (Table 5.18). while addition o f PGA produced a less protective passive film (bR = -0 75 for PGA). Addition of AA yielded a less protective porous film, as illustrated in Fig. 5.39 with two-time- constant characteristics. The abnormal bR value (- 1.27) for R2. which represented the charge transfer resistance (in units o f ohms, not area-normalized) at the pores, could be related to the change of the total pore area during the two-week exposure. According to bR, the ranking for the inhibitors in terms o f long-term corrosion protection from high to low would be: GPH > PAA > OFT > PGA > AA 6-4 The DC polarization results listed in Table 5.19 present a different picture from that given by the EIS results. According to the inhibition efficiencies calculated from the corrosion current densities, the ranking would be: AA > GPH > PGA > OH" > PAA 6-5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 8 t t \ o ✓ 8 —■ ------ Saseune iBL) O 3L * 10 mM AA • — r ~ ~ BL - 5C mM GPH O SL - 20 pom PAA £ r ='- - 20 oom PGA 9C — 75 * C a _ * 1 - c0 •15 30 15 .ag(KHz)) Fig. 6 4 Comparison of 2-hour impedance spectra for mild steel exposed to RT baseline solutions without or with organic inhibitors. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Phase Angle (D egiee) 1 6 9 The discrepancies between these two rankings could be due to the fact that ranking in Eq. 6-5 was based on the short-term (2 hours) DC polarization results. After 2 hours, the studied systems had not reached steady states, nor had the inhibitor- adsorption films been fully developed. Therefore, the ranking based on the long-term (two weeks) data should be more reliable. GPH had a better performance than OH*, as indicated in Eq. 6-4. More importantly, the lowest ip a .s s on the return scan was observed in the presence o f GPH as seen from the cyclic polarization curves (Fig. 5.4 la), which suggests that incorporation and/or adsorption o f GPH into the mild steel surface would produce a more protective film than the magnetite film. Further studies focusing on the inhibition mechanism o f GPH and further evaluation in the hot baseline solution on its thermo stability are necessary. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 170 7. CONCLUSIONS 7.1 Evaluation o f cerium salts Corrosion protection o f mild steel by cerium salts can be realized by formation o f cerium (hydr)oxide layers functioning as a barrier separating the mild steel substrate from the corrosive environment. The cerium (hydr)oxide layer on mild steel can be produced by four methods: addition o f cerium salts as inhibitors in corrosive environments, immersion in cerium salt solutions, electrodeposition in cerium salt solutions and cerating. Different methods yield cerium (hydr)oxide layers with different properties and corrosion protection efficiencies. The efficiency o f a cerium salt as an inhibitor depends on the cerium salt anion, the corrosive environment, the pH o f the corrosive environment and the exposure time. In neutral 0.1 N NaCl. cerium nitrate had the best inhibition efficiency among all cerium salts studied, even better than dichromate, but it didn’t provide lasting protection due to the formation o f a porous mixture o f cerium (IV) and iron oxides. In the hot alkaline baseline solution, the small amount o f dissolved cerium salt provided better protection than dichromate. and the inhibition efficiency increased with time due to the incorporation o f cerium (III) into the magnetite film. Chemical conversion coatings can be formed on mild steel by immersion in cerium salt solutions at RT. Immersion in a cerium nitrate solution produced a conversion coating with the best corrosion protection, even better than that provided by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 71 dichromate as inhibitor in 0.1 N NaCl. However, the protection efficiency o f the conversion coatings decreased quickly with prolonged exposure time in NaCl. A cerium (hydr)oxide film with a small amount of cerium (IV). but a majority cerium (LH) can be formed on mild steel by electrodeposition. Compact and protective cerium (hydr)oxide deposits can be obtained with appropriate choice of the current density and deposition time. However, after 24 hours in NaCl, the deposit lost its integrity and did not provide corrosion protection for the mild steel substrate any longer. A cerium (IV) (hydr)oxide cerated layer can be obtained on mild steel as a chemical conversion coating cerating formed in an aqueous solution o f CeCl? and H ;0 ; at RT. In NaCl and the A-W baseline solution at RT. the cerated layer provided no corrosion protection for mild steel due to its porous and heavily cracked nature. However, in the hot baseline solution, the pores and cracks were effectively sealed within 24 hours and the cerated layer was found to provide increasing protection with increasing exposure time. 7.2 Optimization o f the cerating process The cerating process is a multi-variable process. The properties o f the cerated layer depend on the concentrations o f cerating solution constituents (CeCU, H^O^), solution pH, additives, cerating time, pretreatments, post treatments (sealing) and solution aging. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 7 2 Through factorial design, the optimum conditions for obtaining the cerated layer with the highest Rp in the hot baseline solution were found to be 2.3 g/L CeCh, 4.4 wt% H2O2 and 20 minutes cerating time. However, a high Rp can also be obtained in a cerating solution containing a high concentration o f CeCh and a low concentration o f H2O2, for example. 12.5 g/L CeCL, I wt% H2O2 with 20 minutes cerating time. In cerating solutions with pH ranging from 1.6 to 2.5. the highest Rp and lowest C can be obtained at pH 2.2. This result could be a compromise o f two pH effects on the cerating process, i.e.. at lower pH the cerating reaction is vigorous leading to the formation o f a porous and heavily cracked cerated layer, while at higher pH the precipitation o f cerium hydroxide before cerating could give rise to the formation o f a loose and non-uniform cerated layer. Addition o f sodium nitrite as an oxygen scavenger and pH adjuster at an appropriate concentration. Triton X-100 as a surfactant at 4 ppm and lead acetate as a solution stabilizer at 3 ppm made the cerated layer smoother with less pores and cracks. Such a cerated layer exhibited higher Rp and lower corrosion rates in the A-W baseline solution. Pretreatment and post treatment (sealing) are very important steps in the whole cerating process. Pretreatment of mild steel in hydrochloric acid solution produced a clean and active mild steel surface and facilitated the formation o f a compact and more protective cerated layer. Sealing in a silicate or molybdate solution remarkably Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 7 3 improved the performance o f the cerated layer in the A-W baseline solution, though the sealing mechanisms in these solutions could be different. Aging o f the cerating solution prior to use had a pronounced beneficial effect on the performance o f the cerated layer in the baseline solution. Although the mechanism o f aging is unclear, it is believed that aging facilitates the formation of more compact and uniform cerated layer. 7 .3 Evaluation o f yttrium salts Yttrium salts can be used in a similar way as cerium salts in the corrosion protection o f mild steel in the baseline solution, i.e. as inhibitors, chemical conversion coatings through immersion and fast chemical conversion coating (“ytt rating’’). However, they were outperformed by their cerium counterparts in almost every aspect, perhaps due to the difference between the yttrium cation, which exists only in III valence, and the cerium cation, which has the III and the oxidizing IV valences. 7 .4 Evaluation o f organic inhibitors Four organic inhibitors (GPH. AA. PAA and PGA) were evaluated in the baseline solution with EIS and DC techniques. Competitive adsorption between inhibitor and hydroxyl anions could have interfered with the formation o f a uniform magnetite film on mild steel and resulted in a less protective inhibitor-adsorption film in the early exposure period. However, within two weeks, the inhibitor-adsorption film would become fully developed over the mild steel surface. The inhibitor-adsorption Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 174- film o f GPH or PAA exhibited superior corrosion protection compared to the magnetite film, while that o f AA or PGA provided less protection with an obviously porous film formed by AA. From the perspective o f long-term corrosion protection, the ranking o f the inhibitors in inhibition efficiency from high to low would be: GPH > P.AA > OFT > PGA> AA where OFT represents the sodium hydroxide in the baseline solution. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 7 5 8. SUGGESTIONS FOR FUTURE WORK 8 .1 Optimization o f cerating process Through factorial designs a trend was found to exist, i.e. higher concentration of CeCh and low concentration o f H2O2 leading to the formation of cerated layers with higher Rp, but not lower C (Fig. 5.31). One or more 32 factorial designs would be necessary to locate another possible maximum Rp condition. The initial two-factor three-level factorial (3") design is preferably CeClj at 10. 12.5 and 15 g/L. and H;0 ; at 0.5, 2 and 3.5 wt%. Based on the responses o f the above 32 factorial design, further designs can be performed on a larger or smaller response surface. 8.2 Mechanistic studies on sealing Sealing o f the cerated layer in either silicate or molybdate solution yielded similar improvements in performance in the baseline solution with possible different sealing mechanisms. Further mechanistic studies would be necessary to obtain a better understanding of the role played by silicate or molybdate anions during sealing and subsequent exposure to the baseline solution. These studies would include electrochemical investigation by EIS and DC polarization curves, surface analysis by SEM (morphology). EDS (Si. Ce and Mo distribution in the cerated layer). AES or XPS (elemental concentration profile through the thickness o f the cerated layer for Si, Ce and Mo) as well as X-ray Diffraction (XRD) for structural analysis o f the cerated layer. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sealing in silicate and molybdate mixture solution or sealing in silicate solution and then in molybdate solution is also worth o f further investigation to obtain a cerated layer with better performance due to the synergetic effects o f Ce. Mo and Si. 8 .3 Recycling o f the cerating solution For economical and practical applications o f the cerating process in industry, the cerating solution has to be recycled and rejuvenated as many times as possible before final discharge. After a cerating process, changes in the cerating solution include the concentrations o f the constituents (CeClj, H2 0 2, and additives), solution pH. build up o f reaction products and formation o f insoluble precipitates such as Ce(III) or Ce(IV) hydroxide. Extensive and systematic investigation is required to recover the used creating solution producing quality cerated layers on mild steel. 8.4 Dual protection strategy for mild steel in A-W systems A dual protection strategy has been proposed and a US patent application has been filed [162] for mild steel in A-W heat absorption systems, in which addition of REMSs. preferably cerium and yttrium salts, as inhibitors to the A-W working fluid and the cerating treatment on mild steel are employed. Experimental results in the test rig designed to simulate the working conditions o f a real A-W heat absorption system showed that better corrosion protection was achieved by this strategy with lower corrosion rates and smaller volumes o f evolved non-condensable gases than with chromate as inhibitor [163]. However, considering the low solubility o f REMSs Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 7 7 in the alkaline A-W working fluid, the possibility o f adding more soluble inhibitors such as silicate, molybdate or the promising organic inhibitor GPH to the A-W working fluid should be considered. On the one hand, they act as inhibitors, on the other hand, they can repair defects such as pores or cracks in the cerated layer. 8.5 Application o f the cerating process and REMSs in LiBr systems LiBr heat absorption systems share a much larger HVAC market than A-W systems, however, they encounter more corrosion threats than the latter due to the corrosive nature o f B r. Application o f the cerating treatment and the use o f REMSs as inhibitors in LiBr systems could be an effective alternative to the toxic chromate or dichromate. A preliminary study at CEEL/USC showed that cerating o f mild steel provided comparable protection with dichromate as inhibitor in the LiBr system working fluid o f 64 wt% LiBr ■ + • 0.2 wt% LiOH at 100°C. Further work is needed to investigate the applicability o f the newly developed cerating process and REMSs as inhibitors or the dual protection strategy discussed above in LiBr systems. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 9. REFERENCES 1. J. Philips, Control and Pollution O ptions fo r Am monia Emission, p. 47, U. S. Environmental Protection Agency. NC (1995). 2. P. A. Domanski. J. Res. Natl. Inst. Stand. Technol., 103, 529 (1998). 3. DOE Energy Resources R&D Portfolio: FY 1999-2000. p. 244. DOE (2000). 4. http://www.polarpowerinc.com/products/refrigerator/ref-tech-overview.htm 5. D. K. Miller. Handbook o f Applied Thermal Design. edited bv E. C. Guver and D. C. Brownell. McGraw-Hill Inc. (1989). 6. G. F. Hewitt. G. L. Shires and T. R. Bott. Process H eat Transfer. CRC Press Inc. (1994). 7. Heat Pump Systems: A Technology Review. p. 17. Organization for Economic Co-operation and Development. Paris (1982) 8. US Patent 5,342,578. 9. US Patent 5.811.026. 10. R. J. Lewis. Sr.. Hazardous Chemicals Desk Reference. 4th ed.. p. 292. Van Nostrand Reinhold. New York (1997). 11. H. H. Uhlig and R. W. Revie, Corrosion and Corrosion Control. 3rd ed.. p.414. John Wiley & Sons. New York (1985). 12. H. E. Townsend. Corr. Sci. 10. 343 (1970). 13. V. Ashworth and P. J. Boden. Corr. Sci. 10. 709 (1970). 14. P. A. Brook, Corr. Sci. 11. 389 (1971). 15. P. B. Linkson. P. B. Phillips and C D. Rowles. Corr. Sci. 19. 613 (1979). 16. D. D. Macdonald. B.C. Syrett and S. S. Wing, Corrosion 35, I (1979). 17. S. J. Lennon and F. P. A. Robinson. Corr. Sci. 26, 995 (1986). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 179 18. D. Cubicciotti. J. Nucl. Mater. 167. 241 (1989). 19. P. Paine. Corrosion 46. 19 (1990). 20. H. H. Le and G. Ghali. J. Appl. Electrochem. 23. 72 (1993). 21. B. Beverskog and I. Puigdomenech. Corr. Sd. 38. 2121 (1996). 22. B. Hindin. Evaluation o f Alternative Corrosion Inhibitors for Gas-Fired Ammonia/Water Refrigeration Machines. Final report, p. 1, GRI (1995). 23. G. Schikorr. Z. Elektrochem. 25. 65 (1929). 24. G. Schikorr. Z. Anorg. Allg. Chem. 212. 33 (1933). 25. F Gruber and K. Mall. Linde Berichte aus Technik und Wissenschaft. 61 (1988). 26. H. G. Mosle. Metalloberflaeche, 23, 358 (1969). 27. http ://www. angelfire. com/mt/chemicalcleaning/magnetite. html 28. E. C. McKeivy. and A. Isaacs, Technologic Papers o f the Bureau of Standards. No. 180. Washington. Government Printing Office (1920). 29. US patent 2.775.170. 30. E. Kalman. Corrosion Inhibitors. Euro. Fed. o f Corr. Pub. No. 11. p. 12. The Institute o f Materials (1994). 31. G. E. Ibbotson and A. Marshall. Chemical Inhibitors fo r Corrosion Control. Ed. B. G. Clubley. p. 89. Royal Sodetv o f Chemistry (1990). 32. Y. I. Kuznetsov. Organic Inhibitors o f Corrosion o f M etals, p. 225. Plenum Press. New York (1996). 33. S. A. Bradford. Corrosion Control. p. 241. Van Nostrand Reinhold. New York (1993). 34. G. B. Hatch, Corrosion Inhibitors, Ed. C. C. Nathan, p. 126, NACE (1973). 35. V. Kain. G. E. Prasad and H. S. Gadiyar, India J. of. Tech.. 30. 341 (1992). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ISO 36. I. B. Singh. G. Venkatachari and K. Balakrishnan. J. of App. Electrochem. 24. 179(1994). 37. S. Iyoki and T. Uemura. Reffegeration. 55, 353 (1980). 38. A. Bahadur. Mater. Trans. JIM. 37. 605 (1996). 39. D. G. Kolman and S. R. Taylor, Corrosion. 49. 635 (1993). 40. US patent 4.961.878. 41. US patent 4,915.872. 42. US patent 4.098.720. 43. US patent 4,724,125. 44. US patent 4,382.008. 45. US patent 2.972.581. 46. US patent 5.871.691. 47. US patent 6.004.475. 48. M. -H. Ahn and K. -K. Baek. Corrosion 99. Paper No. 97. NACE. Houston (1999). 49. US patent 5,725,749. 50. US patent 4.749.550. 51. US patent 5,130.052. 52. International Patent WO 88/06639. 53. 53. H. S. Isaacs. A. J. Davenport and A. Shipley, J. Electrochem. Soc.. 138, 390(1991). 54. B. R. W. Hinton. Journal o f Alloys and Compounds, 180, 15 (1992). 55. B. R. W. Hinton and L. Wilson. Corrosion Science. 29. 967 (1989). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 181 56. S. VI. Powell. H. N. McMurrav and D. A. Worsley. Corrosion. 55. 1040. (1999). 57 S. L. Li. Mat. Prot.. 26. 6 (1993). 58. N. Verma. W. R. Singh. S. K. Tiwari. and R. N. Singh. Brit. Corr. Jour.. 25. 131 (1990). 59. R. N. Singh, S. K. Tiwari and W. R. Singh. J. Appl. Electrochem.. 22. 1175 (1992). 60. T. Suwa. Corr. Eng., 37. 73 (1988). 61. S. K. Mitra. S. K. Roy and S. K .. Bose. Oxidation o f Metals. 39. 221 (1993). 62. S. Seal, S. K. Bose and S. K. Roy, Oxidation o f Metals. 4 t. 139 (1994). 63. Y. C. Lu and M. B. Ives. Corr. Sci.. 37. 145 (1995). 64. S. Bernal. F. J. Botana. J. J. Calvino and M. A. Cauqui. J. Allovs Comp.. 225. 638.(1995). 65. B. R. W. Hinton. D. R. Amott and N. E. Ryan. Metals Forum. 7. 211 (1984). 66. B. R. W. Hinton. D. R. Amott and N '. E. Ryan. Vlater. Forum. 9. 162 (1986). 67. D. R. Amott. B. R. W. Hinton and N. E. Rvan. Materials Performance. 2 6 .42 (1987). 68. F. Mansfeld. C. Chen and C. B. Breslin. J. Electrochem. Soc.. 145. 2792 (1998). 69. A. J. Davenport. H. S. Isaacs and M. W. Kendig, Corr. Sci.. 3 2 .653 (1991). 70. A. J. Aldykewicz. Jr.. H. S. Isaacs and A. J. Davenport. J. Electrochem. Soc.. 142. 3442(1995). 71. S. Bohm. R. Greef H. N. McMurray, S. M. Powell and D. A. Worsley, J. Electrochem. Soc.. 147. 3286 (2000). 72. M. Bethencourt. F. J. Botana and M. A. Marcos. J. Alloys Comp.. 250,455 (1997). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 8 2 73. F. Mansfeld. S. Lin. S. Kim and H. Shih. Proc. EuroCorr’87. p. 713. Karlsruhe (1987). 74. F. Mansfeld, S. Lin. S. Kim and H. Shih. Corr. Sci.. 27. 997 (1987). 75. F. Mansfeld. S. Lin. S Kim and H. Shih. Electrochim. Acta. 34. 1123 (1989). 76. F. Mansfeld. S. Lin. S. Kim and H. Shih. J. Electrochem. Soc.. 137. 78 (1990). 77. F. Mansfeld. Y. Wang and H. Shih. J. Electrochem. Soc.. 138. L74 (1991). 78. S. Lin. H. Shih and F. Mansfeld. Corr, Sci., 33. 1331 (1992). 79. F. Mansfeld and Y. Wang, Brit. Corr. Jour.. 29. 194 (1994). 80. US patent 5,582.654. 81. US patent 5.635.084. 82. Y. Wang. “Corrosion Protection o f A 1 Alloys by Surface Modification Using Chromate-free Approaches". Ph. D. thesis. University of Southern California. Dec. 1994. 83. C. B. Breslin. C. Chen and F. Mansfeld. Corr. Sci.. 39. 1061 (1997). 84. C. Chen and F Mansfeld. Corr. Sci.. 39. 1075 (1997). 85. A. E. Hughes. J. D. Gorman and P. J. K. Paterson. Corr. Sci.. 38. 1957 (1996). 86. J D. Gorman. S. T. Johnson. P N. Johnston. P. J. K. Paterson and A. E. Hughes. Corr. Sci., 38, 1977 (1996). 87. X. W. Yu. C. N. Cao and Z. M. Yao. J. Mater. Sci. Letters. 19. 1907 (2000). 88. S. Powell, Surf Eng. 16. 169 (2000). 89. A. L. Rudd. C. B. Breslin and F. Mansfeld. Corr. Sci.. 42. 275 (2000). 90. A. J. Aldykewicz, Jr., A. J. Davenport and H. S. Isaacs. J. Electrochem. Soc.. 143, 147(1996). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 8 3 91. F. Lin. R. C. Newman and G. E. Thompson. Electrochim. Acta. 42. 2455 (1997). 92. M. Geary and C. B. Breslin, Corr. Sci., 39, 1341 (1997). 93. C. B. Breslin and M. Geary. Corrosion. 54. 964. (1998). 94. International Patent WO 95/08008. 95. Australian Patent. P I0649. 96. A. E. Hughes, R. J. Taylor. B. R. W. Hinton and L. Wilson. Surf. Inter. Anal.. 23. 540 (1995). 97. A. E. Hughes. S. G. Hardin. R. M. Peter and W. W. Klaus. Paper no. 111. NACE 2000. Orlando. Florida (2000). 98. A. E. Hughes, R. J. Taylor. K. J. H. Nelson. B. R. W Hinton and L. Wilson. Mater. Sci. & Tech.. 12. 928 (1996). 99. A. E. Hughes. K. J. H. Nelson and P. R. Miller. Mater. Sci. & Tech.. 15. 1124 (1999) 100. J. E. O. Mayneand M. J. Pryor, J. Chem. Soc.. 1831 (1949). 101. VI. Cohen. J. Phvs. Chem.. 56. 451 (1952). 102. M. Cohen. Corrosion. 3 2 .4 6 1 (1979). 103. S. Tougoose. Chemical Inhibitors fo r Corrosion Control. Ed. B. G. Clubley. p. 73. Royal Society of Chemistry (1990). 104. N. Hackerman and E. S. Snavely. NACE Basic Corrosion Course, p. 9-11. NACE (1970). 105. C. L. Hannon. J. Gerstmann. F. Mansfeld and Z. Sun. in Symp. o f Heat Pump and Refrigeration System Design. Analysis and Applications, p. 35. ASME. Orlando (2000). 106. A. Bahadur, Mater. Trans.. JTM. 34. 1191 (1993). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 184 107. J. Gerstmann and C. L. Hannon, Development o f Corrosion Inhibitors fo r Absorption H eat Pumps, p. I, AMTI (Advanced Mechanical Technology, Inc.), Watertown. MA(1999). 108. G. Trabanelli, Corrosion, 47, 410 (1991). 109. Y. I. Kuznetsov and T I. Bardasheva. Protection o f Metals. 28. 450 (1992). 110. M. AjmaL, A. S. Mideen and M. A. Quraishi, Corro. Sci., 36. 79 (1994). 111. B. Heeg. T. VIoros and D. Klenerman. Corro. Sci.. 40. 1313(1998). 112. B. Heeg, T. Moros and D. Klenerman, Corro. Sci., 40. 1303(1998). 113. C. Monticelli. A. Frignani. G. Brunoro. G. Trabanelli. F. Zucchi and M. Tassinari, Corro. Sci . 35, 1483 (1993). 114. D H. Kalota and D. C Silverman. Corrosion. 50. 138 (1994). 115. A. M. Beccaria. M. Ghiazza and G. Poggi, Corro. Sci., 36. 1387 (1994). 116. E. Mueller. C. S. Sikes and B. J. Little. Corrosion. 49. 1 (1993). 117. H. H. Uhlig and R. W. Revie. Corrosion and Corrosion Control. 3rd ed.. J. Wiley, 1985. 118. C. M. Hwa, Mater. Perform.. 10. 249 (1971). 119. J R. Ambrose. Corrosion. 34. 27 (1978). 120. F. M. Kharafi and F. H. Hajjar. British Corrosion Journal. 25. 209 (1990). 121. W Tsai. A. Moccari and D. D. Macdonald. Corrosion. 39. I (1983). 122. B. Bavarian. A. Moccari and D. D. Macdonald, Corrosion, 38. 104 (1982). 123. F. H. Hajjar and W. T. Riad. British Corrosion Journal. 25. 119 (1990). 124. V. Kain. G. E . Prasad and H. S. Gadiyar, Indian Journal o f Technology. 30, 341 (1992). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 185 125. J. L. Zaldes, J. W. Mitchell and J. A. Mosovsky, Proceedings o f 27th International SAMPE Technical Conference, p. 315, Ed. R. J. Martinez, Albuquerque, NM, USA (1995). 126. F. Mansfeld and T. Wood, personal communication between. Aug.. 1999. 127. N. N. Greenwood and A. Eamshaw. Chem istry o f the Elements. 2nd Ed., p. 1227, Butterworth-Heinemann, Oxford, GB (1997). 128. T. J. Haley. J Pharm. Sci.. 54. 633 (1965). 129. N. N. Greenwood and A. Eamshaw. Chem istry o f the Elements. 2n d Ed., p. 1294, Butterworth-Heinemann. Oxford, GB ( 1997). 130. B. R. W. Hinton and R. R. Amott. Microstructural Sci.. 17. 31 1 (1989). 131. C. Wagner and W. Traud, Z. Electrochem.. 44. 391 (1938). 132. J. V Petrocelli. J. Electrochem. Soc.. 97. 10 (1950). 133. J. A. V. Butler. Transactions o f the Faraday Society. 19. 729 (1924). 134. T Erdev-Gruz and VI. Volmer. Zeitschrift fur Phvsikund Chemie. 150A. 203 (1930)." 135. A nnual Book o f A SV vf Standards. Vol. 03.02. G5-94. Standard Reference Test VIethod for Making Potentiostatic and Potentiodvnamic Anodic Polarization Measurements. ASTM (1997). 136. D. A. Jones, Principles and Prevention o f Corrosion, 2n d Ed., p. 119. Prentice- Hall. Inc.. NJ (1996). 137. F. Mansfeld. Corrosion, 29. 397 (1973). 138. H. Shih and F. Mansfeld. in Computer M odeling in Corrosion. ASTM STP 1154, R. S. Munn. Ed., p. 174, ASTM. Philadelphia (1992). 139. M. Stem and A. Geary. J. Electrochem. Soc.. 104. 56 (1957). 140. F. Mansfeld. in Advances in Corrosion Science an d Technology, Vol. 6, p. 163. Plenum Press (1976). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 8 6 141. F. Mansfeld, in Electrochem ical Techniques fo r Corrosion Engineering, R. Baboian. Ed., p. 67. NACE. Houston (1986). 142. F. Mansfeld and K. B. Oldham, Corr. Sci.. 11. 787 (1971). 143. Annual Book o f A STM Standards, Vol. 03.02. GI02-89. Standard Guide for Calculation o f Corrosion Rates and Related Information from Electrochemical Measurements, ASTM (1997). 144. Z. Szlarska-Smialowska. Pitting Corrosion o f Metals. Chap. 3. NACE. Houston (1986). 145. J. Aromaa and A Klarin. M aterials. Corrosion Prevention a n d M aintenance. p. 99. TAPPI press. Helsinki (1999). 146. F. Mansfeld and M. W. Kendig, Werkst. Korros.. 34, 397 (1983). 147. F. Mansfeld. M. W. Kendig, A. F Allen and W J. Lorenz. Proc. 9th Int. Conf. Metallic Corrosion. Toronto. Canada. Vol. I. 368 (1984). 148. F. Mansfeld, M. W Kendig and W. J. Lorenz. J. Electrochem. Soc.. 132. 190 (1985). 149. F. Mansfeld and M. W. Kendig. Proc. 6th Euro. Symp. Corros. Inhibitors. Ferrara. Italy (1985). 150. F. Mansfeld. C. H. Tsai and H. Shih. in Computer M odeling in Corrosion, ASTM STP 1154. R. S. Munn. Ed., p. 186. ASTM. Philadelphia (1992). 151. E. V. Westing, Determination o f Coating Performance with Impedance Measurements. Ph.D. Dissertation, p. 104. TNO Centre for Coating Research. Netherlands (1992). 152. C. Hsu and F. Mansfeld. Technical Note: Concerning the Conversion o f the Constant Phase Element Parameter Y0 into a Capacitance. Corrosion (in press). 153. E. Heitz and G. Kreysa. Principles o f Electrochem ical Engineering, p. 169. VCH Publishers. New York (1986). 154. A 1. Khuri and J. A Cornell, Response Surfaces: Designs and analyses, Marcel Dekker, Inc.. New York (1987). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 8 7 155. 156. 157. 158. 159. 160. 161. 162. 163 K. Hinkelmann and O. Kempthome. Design and Analysis o f Experiments. VI. Introduction to Experimental Design. John Wiley & Sons, Inc., New York (1994). M. Pourbaix. in Pro. o f the 7th Int. Congress on Metallic Corrosion, Rio de Janeiro, 1978. V. Kucera and E. Mattsson. in Corrosion M echanisms, F. Mansfeld Ed, p. 211. Marcel Dekker. Inc., New York (1987). Annual Book o f A STbl Standards. Vol. 03.02. G101-94. Standard Guide for Estimating the Atmospheric Corrosion Resistance of Low-Allov Steels. ASTM (1997). F. Mansfeld. M. W. Kendig and S. Tsai. Corrosion. 38. 478 (1982). F. Mansfeld. J. Appl. Electrochem. 25. 187 (1995) Y. C. Lu and M. B. Ives. Corrosion Science, 34. 1773 (1993). F Mansfeld and Z. Sun. US patent application 09/774.540. Corrosion Protection o f Steel in Ammonia/Water Heat Pumps. Unpublished experimental results from AMTI. Watertown. VIA. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 188 PART II: STABILITY' EVALUATION OF METALLIC MATERIALS IN LI-ION BATTERY ELECTROLYTE Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1. INTRODUCTION The last decade has seen a spectacular growth o f rechargeable Li-ion battery systems since their introduction by Sony in 1991 [ I ]. because o f their outstanding properties at ambient temperature, such as higher energy density, lighter weight, fewer safety concerns and more environmental friendliness than their conventional counterparts, such as lead-acid and nickel/cadmium systems [1-3]. In recent years, Li-ion batteries have become one o f the best choices as implantable and/or external power sources for implantable biomedical electronic devices, such as neurological stimulators for treatment o f patients suffering from stroke, Parkinson's disease, limbs paralysis, etc. [4,5]. However due to its vital role in a biomedical device package, a rechargeable Li-ion battery system has to meet extreme stability and safety requirements in order to guarantee long term, for example 10 years, service in the human body. The stability o f metallic materials used in a rechargeable Li-ion cell is one o f the major safety concerns. These metallic materials include anode and cathode current collectors and casing materials. Copper and aluminum are the most commonly used current collectors for anode and cathode electrodes, respectively, due to their high conductivity and good workability. They also function as supporting substrates for lithiated composite anode and cathode electrodes. Fig. I I [6] illustrates schematically the positions o f the anode current collector (Cu) and cathode current collector (Al) in a Li-ion cell. Stainless steels are usually used as casing materials, but for implantable batteries. Ti materials, which are the most popular implantable Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 190 (n- - -0 - Cu L ► • i t i t L f t t t 8 8 8 » • Anode current collector UX nOm Anode electrode (source) Li* Li* Li' conducting electrolyte ► t t i 888 • • I •♦ •♦ •••I AI ► e -► e • • 8888 A ,B . Cathode electrode (sink). Cathode current collector Fig. I. I Schematic diagram o f a Li-ion cell on discharge [6]. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 191 metallic materials [7,8], are preferred mainly due to their lightweight as well as high passivity. The metallic materials are required to be stable and compatible with the electrolyte in a battery under normal operating conditions and even under abusive conditions, such as overdischarge or overcharge during which the metallic materials, especially the current collectors may experience high potential polarization. In a real battery system, there usually exist joints o f metallic materials, which could introduce galvanic corrosion and accelerate corrosion o f the more active material. In electrolytes with lithium hexafluorophosphate (LiPF6) as the solute, the presence o f water, even in a trace amount, could cause severe corrosion o f metallic materials, since the LiPF6 can react with water to produce HF according to the following reaction [9]: LiPF6 + H20 = 2HF - LiF + POF} I - 1 In this study, the stability of various metallic materials in a LiPF6 electrolyte at a specified temperature was extensively evaluated with different electrochemical methods and surface analysis techniques. All electrochemical experiments were carried out in a glove box with moisture and oxygen concentrations maintained below 0.5 ppm and 100 ppm. respectively. Potentiodvnamic polarization was employed to determine the stable potential window o f each test material as well as the test electrolyte. Potentiostatic polarization was used to estimate the corrosion rate o f each test material under a specified anodic potential simulating typical battery operating voltages. Galvanic corrosion tests were carried out to examine the extent o f Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 192 galvanic corrosion between dissimilar metals. The time dependence o f the corrosion rate for each test material was obtained through a 30-day EIS test, and then used to predict the material’s long-term, i.e. 10 years stability. The SEM was used to observe morphology changes o f each sample after an experiment. For a selected number o f samples covered with corrosion products. EDS and/or AES were performed to determine the surface chemistry o f the corrosion products. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2. LITERA TU RE REV IEW The stability o f current collectors, especially A 1 as the cathode current collector in rechargeable Li-ion batteries has become an intense research subject in recent years. The preliminary results [10-161 have indicated that current collectors are susceptible to degradation in the organic nonaqueous Li-ion battery electrolytes. Up to now no studies of the corrosion behavior o f casing materials has been reported in the literature. 2 .1 Stability o f the cathode current collector Al foil is the most commonly used cathode current collector in commercial rechargeable Li-ion batteries. Results obtained so far on the corrosion behavior o f Al in Li-ion battery electrolytes have shown that Al is more susceptible to localized corrosion than to general corrosion [10-14], At the OCP, Al is very stable. The general corrosion rate is very small and can be neglected [11]. The corrosion behavior o f Al depends on its composition and passive film integrity, electrolyte components, contaminants and temperature. 2.1.1 Al composition and passive film integrity Al foil used in Li-ion battery is o f commercial purity, but it still suffers from pitting corrosion in battery electrolytes due to the small amount o f impurities. Brahhwaite and coworkers showed that commercial pure aluminum AAl 145 (99.45% minimum) may have superior pitting resistance than AAl 100 (99.0% minimum and 0.12% Cu) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 194 in an electrolyte o f 1M LiPF6 dissolved in I: I (v/v) o f propylene carbonate (PC) and diethyl carbonate (DEC) under simulated electric cycling conditions [14], The initiation o f pits was believed to be due to intermetallic compounds, such as CuAl?, which is in consistent with the pitting initiation mechanisms of Al and its alloys in aqueous solutions [17], The intermetallic compounds on the Al surface break down the integrity o f the passive film and act as the cathode producing OH", which accelerates the dissolution of the surrounding aluminum. It was reported that commercial aluminum with purity higher than 99.5% suffers very little pitting in the Li-ion battery electrolyte [18], With such high purity Al. the density o f the intermetallic compounds on Al surface has been much reduced, and therefore, the passive film is more uniform and protective Yang and colleagues [13] found that pitting corrosion was exacerbated to different extent when the passive film was removed from Al specimen in a glove box filled with argon (water and oxygen concentrations less than 5 ppm) and then tested in different Li-ion battery electrolytes. 2.1.2 Electrolytes Lithium salts dissolved in aprotic organic solvent mixtures or blends are the most popular electrolytes in rechargeable Li-ion batteries thanks to their excellent solubility in aprotic organic solvents, high conductivity and high kinetic stability towards oxidation [19]. The lithium salts most commonly used have a formula o f LiXFn , where n = 4 if X is boron and n = 6 if X is phosphorous or arsenic. However. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 195 this type o f lithium salts suffers from chemical instability in the presence o f water and thermal instability [3,9.20-22]. For example, LiPF6 can easily react with water to produce HF [9], and also decompose at a temperature around 30°C in the solid state and 130°C in solution [3], Consequently, many alternative lithium salts have been studied, of which the lithium imides. such as lithium bis(trifixoromethanesulphonyl) imide LiNfCFjSChh are the most promising ones [23, 24], However, this imide class o f lithium salts has one drawback preventing them from applications in the commercial batteries, which is the severe corrosion o f the Al cathode current collector in their presence [3. 24], The susceptibility o f Al to corrosion, especially localized corrosion depends on the solvent mixture, but more on the lithium salts used in the electrolytes. Peter and Arai [12] studied the anodic dissolution o f Al in various solvent and lithium salt combinations. Based on their results, they classified the electrolytes into three categories o f noncorrosive, moderately corrosive, and very corrosive. Electrolytes with lithium salts LiXF„ were noncorrosive to Al. since they were apparently excellent passivating agents. LiXF„ can easily release fluoride ions, which contributes to the formation o f the passive film protecting Al at low anodic potentials [12,24], This is in good agreement with other findings [10,13] that LiPF6 and LiBF4 were effective inhibiting additives at appropriate concentrations (such as 0.5M) to corrosive electrolytes containing LiNfCFjSOzh- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 196 2.1.3 Contaminants Due to their high hygroscopic nature o f LiXF„ lithium salts, LiXF„ lithium salts are unstable and can produce HF resulting in dissolution o f cathode materials and current collector in the presence of even very small amounts o f water [20-22], But surprisingly, Braithwaite et al. [14] observed that addition o f 20 ppm water to IM LiPF6 dissolved in I: I volume ratio o f PC: DEC actually appeared to have improved the corrosion resistance o f aluminum after more than 100 simulated discharge/charge cycles. They attributed this phenomenon to the stabilizing effect o f water on the passive layer [25], but suggested that a more detailed analysis would be needed before reaching to a definitive conclusion. 2.1.4 Effect o f temperature Rechargeable Li-ion batteries operate ideally at ambient temperature. At elevated temperatures, though ionic conductivity o f electrolyte is increased, the electrolyte and composite electrodes could become thermally and chemically unstable, and that in turn could affect the stability o f the aluminum current collector. Unfortunately, there has not been much information thus far on this topic. Braithwaite et al. [14] tried, but failed to determine the effect o f temperature (at 35°C and 50°C) on the corrosion behavior o f Al due to the significant corrosion o f Li reference electrode, which totally blackened the electrolytes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.2 Stability o f anode current collector Compared to the cathode current collector AL the anode current collector Cu has received less attention in corrosion studies. Since copper usually has a higher OCP than that of the lithiated carbonaceous anode electrode, it is cathodically protected most o f the time by the anode. But in case the battery is abusively used, such as being overdischarged and then remaining unrecharged. Cu would be exposed to the electrolyte at its OCP or at anodic polarization potentials. Zhao and coworkers [15.16] studied the oxidation-reduction ofbattery-grade copper in three LiPF6 based electrolytes containing different mixtures of organocarbonate solvents at ambient temperature. They found that the stability of copper depended on the blends o f the solvents. The corrosion products o f copper were soluble in the studied electrolytes. The introduction o f a small amount o f water (500ppm) or HF ( lOOOppm) greatly accelerated the copper oxidation. A graphite coat provided the copper substrate some protection against oxidation, but tended to peel off in some electrolytes. Braithwaite et al. [14] investigated the stability of commercial copper foil under simulated discharge/charge conditions in IM LiPF6 dissolved in 1:1 volume ratio of PC: DEC and EC: DMC. respectively. Since the OCP o f Cu was nobler than the cycling potential range, the Cu foil samples were actually cathodically protected. No sign o f uniform or localized attack o f copper was observed under various conditions such as water contaminants, carbon coating and elevated temperatures. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 198 3. EXPERIMENTAL TECHNIQUES The experimental techniques employed in this study include different electrochemical methods and surface analysis techniques. The electrochemical methods were EIS. potentiodvnamic polarization, potentiostatic polarization and galvanic corrosion measurement. Surface analysis techniques were SEM. EDS and AES. A detailed description o f these techniques is given chapter 3 o f Part I o f this thesis. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 199 4. EXPERIMENTAL APPROACH 4 .1 Materials and Pretreatments The test materials were Ti-6AI-4V (Al 6%, V 4%. balance Ti) sheet with a thickness of 0.4 ram. Ti-3A1-2.5V (Al 3%. V 2.5 %. balance Ti) sheet with a thickness o f 0.4 mm, commercial pure titanium (Ticp) foil with a thickness o f 0.01mm, pure copper foil with a thickness of 0.01 mm. pure aluminum foil with a thickness o f 0.02mm and Pt90/IrI0 (Pt 90%, Ir 10%) foil with a thickness of0.05mm. All materials were provided by Quallion. LLC. Svlmar. CA. All test samples were cut into 1.5 by 1.5 inch (38 by 38mm) square pieces. Before each experiment, the sheet samples were cleaned according to the following procedure: 1) Polishing with SiC sandpaper from grit #200 to #2400. 2) Ultrasonic cleaning in acetone for 10 minutes. 3) Rinsing with Dl water. 4) Cool air blow-drying. 5) Cleaning in vacuum oven for 24 hours at room temperature (RT) at vacuum o f 30mmHg. The foil samples were cleaned according to the following procedure: 1) Ultrasonic cleaning in acetone for 10 minutes. 2) Rinsing with DI water. 3) Cool air blow-drying. 4) Cleaning in vacuum oven for 24 hours at RT at 30mmHg. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 0 After pretreatments, the samples were stored in a glove box with moisture and oxygen concentrations maintained below 0.5 ppm and 100 ppm. respectively. After the experiment, the samples were ultrasonicallv cleaned in acetone for 10 minutes, then rinsed with DI water, and finally blow-dried with compressed cool air at RT and stored in a dessicator. 4.2 Electrochemical experiment configuration The electrolyte used in all tests was provided by Quallion. It consisted o f 1VI lithium hexafiuorophosphate (LiPFb) dissolved in a 1:1 volume mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) with a water concentration o f less than 10 ppm. Since EC and DEC are organic solvents and LiPFt, is prone to a hydrolytic reaction to produce hydrofluoric (HF) acid, it was required that the test cell be stable during experiments to both the organic solvents and hydrofluoric (HF) add. Therefore, special cells made from Teflon flats with an exposed sample area of 1.98cm2 were designed at CEEL. USC and fabricated at Quallion. Because o f the poor resilient nature o f Teflon, spedal O-rings were chosen for the cells, which were Teflon-encapsulated Silicone or Viton O-rings with excellent chemical inertness provided by the Teflon and outstanding resilience given by Silicone or Viton. The counter and reference electrodes were made from a pure lithium foil with a thickness o f 0.25 mm, which was provided by Quallion, and stored in a glass jar in the glove box. The reference electrode was made by wrapping a I cm x I cm lithium Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 1 foil on a pure nickel wire with a diameter o f 0.5 mm. The reason o f choosing nickel is that nickel is stable when exposed to hydrofluoric acid. The counter electrode was made by putting a 38 mm x 38 mm lithium foil on a Teflon tape-wrapped stainless steel sheet with a size o f 38 mm x 38 mm. The lithium foil and its stainless steel support were covered with a piece o f pure nickel foil with a 25 mm diameter hole in the center to expose the lithium foil with an area o f 4.9 cm2 to the electrolyte. All experiments were conducted in a glove box installed in a Class 1000 cleanroom at Quallion. During each experiment, the cell was put into an incubator in the glove box to maintain the electrolyte temperature at 37 ± 2 °C. 4.3 Electrochemical techniques The electrochemical techniques employed in this study included electrochemical impedance spectroscopy (EIS). potentiodynamic polarization, potentiostatic polarization (“constant potential test”), and galvanic corrosion measurements. All techniques were performed with the Electrochemical Measurement System (Model PC4/300™ DCH1) with an electrochemical multiplexer (Model ECM8) from Gamrv Instruments, Inc. 4.3.1 Electrochemical impedance spectroscopy (EIS) All EIS measurements for the six test materials were controlled with the EIS 300 software version 3.20 from Gamry Instruments. A two-electrode cell configuration was employed for all impedance measurements in which two identical samples with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 2 the same exposed area were used and the reference electrode was eliminated. The sample’s exposed area was 1.98 cm2; therefore, the effective exposed area was 0.99 cm2. The amplitude o f the applied sinusoidal ac signal was 10 mV rms in a frequency range from 100 KHz to I mHz. An eight-channel electrochemical multiplexer (Model ECM8) from Gamry Instruments was used to connect the six cells containing the six test materials placed in two incubators. Commercial software was employed to remotely control the Electrochemical Measurement System in the clean room o f Quallion from the computer at CEEL, (JSC via commercial telephone line. A total o f 20 measurement data was collected for each test material (test cell) at an immersion time in the electrolyte spanning from 2 hours to 30 days. All EIS data were analyzed with the ANALEIS software developed at CEEL. USC [26], 4.3 .2 Potentiodynamic polarization The potentiodynamic polarization experiments for the six test materials were conducted with the software DC 105 version 3.20 from Gamry Instruments. A three- electrode cell configuration was employed in which the working electrode was the test material, while the counter electrode and the reference electrode were a pure lithium foil. The exposed area for the working electrode was 1.98 cm ' and for the counter electrode it was 4.91 cm2. The potential was scanned from -30 mV vs. the OCP to -i- 6.5 V vs. the reference electrode or to a maximum anodic current density o f 100 pA/cm:. whichever comes first. The scan rate was 600 mV/hour according to ASTM G5-94. All potentiodynamic curves were fitted with the POLFIT software Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 203 developed at CEEL, USC [27] in order to determine kinetic parameters and corrosion rates. 4.3.3 Potentiostatic polarization The constant potential tests were conducted for AI foil, Ti-6A1-4V sheet and Ti-3 AI- 2.5V sheet at - 4.5 V vs. Li'/Li, and for Ticp foil and Pt/Ir foil at - 3.5 V vs. Lf/Li. All the experiments were controlled with the software ESA400 version 1.20 from Gamry Instruments. The three-electrode cell configuration was employed. One hour after the cell was set up. the OCP o f the test material vs. reference electrode was recorded for one hour. Then the constant potential was applied between the test material and reference electrode and the current flowing between the test material and the counter electrode was recorded for 24 hours with a sampling rate of 2 points per second. 4.3.4 Galvanic corrosion experiment The galvanic corrosion experiments for four couples, i.e. Al foil / Ti-6A1-4V sheet. Al foil / Ti-3 AI-2.5V sheet. Ticp foil / Cu foil and Ticp foil / Pt-Ir foil, were conducted with the software ES A400 version 1.20 from Gamry Instruments. The three-electrode cell configuration was employed with the reference electrode made from a Li foil, while the working and counter electrodes were the two dissimilar materials. One hour after the cell was setup, the OCP o f the two test materials vs. the reference electrode was recorded for one hour each. Then a zero potential difference was applied with a zero resistance ammeter (ZRA) between the two Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 204 electrodes and the galvanic current and the potential o f the coupled electrodes were recorded for 24 hours with a sampling rate o f 2 points per second. 4.4 Surface Analysis The SEM was employed to observe surface changes between the unexposed and exposed areas after each electrochemical test. For a few selected samples with obvious morphological changes, e.g. coverage with corrosion products, EDS and/or AES was performed to identify surface chemistry o f chemical elements. The SEM and EDS analysis were carried out with a Cambridge Model Stereoscan 360 and a Link Analytical Model 1000 Analyzer, respectively, at the Center for Electron Microscopy and Microanalysis (CEMMA) o f USC. The system was operated at the electron beam energy of 10 to 15 KeV with a beam current of about 200 pA for SEM and 2 nA for EDS. The working distance was about 10 mm for SEM. and around 20 mm for EDS. AES was conducted with a Perkin-EImer AES PHI660 at CEMMA. During an AES measurement, samples were placed in a chamber at a vacuum o f 4.3 < 10‘9 torr. Ar sputtering was applied at the beam energy o f 2 KeV and an emission current o f 30 mA for a total o f about 30 minutes on a sample area of 2 mm by 2 mm. The surface chemistry was analyzed with an electron probe at beam energy o f 3 KeV. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 205 5. EXPERIMENTAL RESULTS AND DISCUSSION The stability o f six metallic materials in Li-ion battery electrolyte at 37°C was studied with electrochemical methods and surface analysis techniques in a glove box. Potentiodynamic polarization was employed to determine the stable potential window of each test material. Potentiostatic polarization was used to measure the dissolution rate o f each test material under applied potential simulating the operating voltage o f a battery. The galvanic corrosion test was intended to examine the galvanic corrosion between dissimilar metals. EIS measurements were carried out to determine the time dependence o f corrosion rate for each test material and to predict the material’s long-term stability. SEM, EDS or AES was performed to detect morphology and surface chemistry changes after a given test. 5 .1 Potentiodynamic polarization Potentiodynamic polarization tests were carried out to determine the stable potential windows o f six metallic materials as well as the electrolyte. The surface morphological changes after polarization for each material were observed with SEM. 5.1.1 Polarization curves and SEM results The anodic polarization curves for all materials are presented in Fig. 5 .1. For Cu active behavior was observed, i.e. the current density increased sharply to 100 pA/cm2 with applied potential, while for Al passive behavior was found. i.e. the current density remained constant at about 0.3 pA/cm2 after a small activation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 6 7 00 > > 6 0 0 5 0 0 — 3 0 0 — 3 0 0 Ti64 Ti3 2 5 - Q Al — O ' T'CP - A Cu Pt I r 0 c r O ■ > . 3 . I : « * os * 3 3 3 3 3 3 C t , s > * v i* - \ t © t A « 3 s • ' $ * jSs' 3 r ® 3 J - * r ^ r 5-< S3 ♦ t* s2 ■ Q ’ J^ siT . e 200 IEOI2 1E01I IE 010 IE 009 1E00S IE007 IE006 lEGOS IEC03 1E003 Current Density (A/ cm2) Fig. 5 .1 Potentiodynamic polarization curves For different metallic materials in LiPF6 + EC + DEC at 37°C (after 2 hours immersion). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 7 region. SEM observation on the exposed and unexposed areas revealed that Cu was severely pitted (Fig. 5.2), while Al remained virtually intact (Fig. 5.3). Passivity of Al in the test electrolyte suggests that the elevated temperature (37°C) and the contaminants o f water and oxygen at the controlled low levels did not have detrimental effects on the stability o f Al. The polarization curves for all three Ti materials (Fig. 5.4) have a common current density peak at about + 5.0 V suggesting that this peak is associated with dissolution and passivation o f Ti. The two Ti alloys have a more negative Eorr than Ticp and the polarization curves show a peak at about - 3.0 V. which could be due to preferential dissolution and passivation o f the alloy constituents Al or V. SEM results (Fig. 5 .5 - 5.7) indicate that the three Ti materials experienced corrosive attack during anodic polarization. However, the attack varied in extents in accordance with the highest anodic current densities for the three Ti materials (Fig. 5.4). The attack on Ti-6A1-4V was the least severe and occurred mainly in the vicinity o f inclusion particles, while other regions were covered with a thin film o f corrosion products (Fig. 5.6). The polish scratches were still visible in the exposed area (Fig. 5.6a). The attack on Ti- 3 AI-2.5V was the most severe with a rough and porous surface observed after the test (Fig. 5.7a). The attack on Ticp was moderate and localized on surface mechanical defects formed during manufacturing (Fig. 5.5). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 8 (b) Fig. 5.2 SEM images for Cu foil after anodic potentiodynamic polarization test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 209 (b) Fig. 5.3 SEM images for Al foil after anodic potentiodynamic test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 0 ■ n » > 7 00 6 0 0 5 0 0 4 00 3 0 0 2 0 0 1E0I1 IE010 IE 009 1E008 1E007 1E0G6 IE005 IE004 Current Density (A/cm*) Fig. 5.4 Potentiodynamic polarization curves for Ti and Ti alloys in LiPF6 + EC + - DEC at 37°C (after 2 hours immersion). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fig. 5.5 SEM images for Ticp sheet after anodic potentiodynamic test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fig. 5.6 SEM images for TI-6AI-4V sheet after anodic potentiodynamic test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 213 (b) Fig. 5.7 SEM images for Ti-3AI-2.5V sheet after anodic potentiodynamic test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 214 There was no obvious passive region in the polarization curve o f Pt-Ir. Instead, two Tafel regions can be observed in the potential range o f 3.5 - 4.0 and 5.0 - 6.5 V, respectively. Considering the known stability and catalytic activity o f Pt. the active behavior of Pt-Ir could be ascribed to the decomposition o f electrolyte or selective dissolution of Ir. However. SEM observations o f the Pt-lr sample after polarization found virtually no difference between the exposed area and unexposed area (Fig. 5 .8). It is possible that some decomposition o f the electrolyte could have occurred during polarization. Chemical analysis o f the test electrolyte is necessary to reach more definitive conclusions. 5.1.2 Fitting o f polarization curves Fitting of the polarization curves within E ^ r 30 mV results in the anodic and cathodic Tafel slopes ba and bc, constant B, corrosion current density Lwr and polarization resistance Rp (Table 5 .1). The values o f ic o n - were very small ranging from about IxlO'1 0 A/cm2 for Al to 2.4x10 A/cm" for Cu. Quantitative corrosion rates can be calculated from the experimental values o f io o rr using Faraday’s law. A value o f IxlO*7 A/cm" corresponds to a corrosion rate o f about I pjn/year. Low corrosion rates for all test materials indicate high stability in the electrolyte at the OCP. Based on L on-, the stability o f the materials at OCP from high to low can be ranked as: Al > Ti-6AI-4V > Ti-3AI-2.5V > Ticp > Pt-Ir > Cu 5-1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 5 (b) Fig. 5.8 SEM images for Pt-Ir foil after anodic potentiodynamic test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 6 Table 5.1 Fit results from potentiodynamic polarization curves. b>(mV/decade) bc (mV/decade) B (mV) icon (A/cm-) j Rp (ohm-cnr) j Ti-6A1- 4V 43.0 IS. 1 5.5 3.2* 10'9 | 1.7* to6 1 T1-3AI- 2.5V 55.6 12.9 « 5 .O x lO'9 I 1 7.9* 10 -' 1 Ticp 54.8 19.4 6.2 6.0* 10'9 1 j 1.2 <10° * 45.6 68.9 11.9 2.4* ltr" 1 | 5.O x 105 PM, 62.7 27.5 8.3 8.5* I0'9 j 1.0x10° Al 86.9 20.8 7.3 l.OxlO'1 ' * | 7.0* 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 7 5.1.3 Stable potential window The stable potential window is defined here as the potential range bounded by two limits. The lower limit is the OCP. while the upper limit is the anodic potential corresponding to the anodic current density o f 1 p.A/cm2 The stable potential window for each test material, tabulated in Table 5.2 was determined from its anodic polarization curve (Fig. 5.1). Based on the stable potential window, the stability of the test materials from high to low can be ranked as. Al > Ti-6AI-4V > Ti-3 A1-2.5V > Pt-lr > Ticp > Cu 5-2 From Table 5.2 and Eq. 5-2. it can be concluded that Cu had practically no stable potential window, while Al was the most stable material. For the three Ti materials, the lower limit (OCP) decreased and upper limit increased with increasing concentration o f alloying elements Al and/or V. resulting in a larger stable potential range. The effects o f the alloying elements Al and V on the stable window are worth o f further investigations. Eq. 5-1 and 5-2 are consistent with each other except for the Pt-Ir alloy, which could induce decomposition o f the electrolyte. The consistency between Eq. 5-1 and 5-2 suggests that the stable potential window o f a metallic material is qualitatively reflected by w or Rp. i.e.. the tower the Wr or the higher the Rp, the larger is the stable potential window. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 8 Table 5.2 Stable potential windows determined from anodic polarization curves. Material Lower limit I (OCP) (V vs. LiTLi) | Upper limit (V vs. LiTLi) | Potential range (V) Ticp 3.3 | 4.1 | 0.8 Ti-3AI-2.5V 3.7 | 4.2 1 15 Ti-6AI-4V 3 , | 4.7 i 2'2 Al 3.3 j > 6.5 , >4.2 Cu 3.7 j 2.8 1 - Pt-Ir 3.3 j 4.4 i . . *2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 9 5.2 Potentiostatic polarization Potentiostatic polarization was carried out to determine the dissolution rate o f a test material under an applied constant anodic potential, which simulates the battery operating voltage. The time dependences o f the polarization currents are presented in bilogarithmic plots for Ah Ti-6AI-4V and Ti-3AI-2.5V at an applied potential of + - 4.5 V in Fig. 5.9a. and for Ticp and Pt-Ir at + - 3.5 V in Fig. 5.9b. 5.2.1 Polarization at - 4.5V vs. Li/Li' The bilogarithmic plot o f polarization current vs. time in Fig. 5.9a for Al has a slope o f - 0.97 after about 30 seconds indicating fast passive film formation. SEM observation found no morphological differences between polarized and non polarized areas, which implies that the passive film is very thin and compact. For the two Ti alloys (Fig. 5.9a), a linear region existed in the time span o f 30 to 1000 seconds for Ti-3 A1-2.5V. and 30 to 2000 seconds for Ti-6A1-4V. Following the linear region, current fluctuations occurred for both Ti alloys. The linear region has a slope o f - 0.57 for Ti-3A1-2.5V and - 0.60 for Ti-6AI-4V. which might indicate the formation o f a porous film on Ti-3 AI-2.5V as well as on Ti-6A1-4V. The current fluctuation regions may be caused by the dissolution and passivation o f Ti. It has to be noted that the applied potential (+ 4.5V) was within the stable windows for Al and Ti-6A1-4V. but higher than the upper limit o f the stable window for Ti- 3AI-2.5V (Table 5.2). After 24-hour polarization, the current density was about 10'5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 0 < 3 3 o I Q E 0 0 4 I OE-005 I QE-006 1 OE007 — I 0E 008 (a) A l foil Ti3-2 5 sneet 7i6A sheet I 0E 009 IE-000 IE-001 IE-002 IE-003 IE-CCA rim e(s) 1 O E O O A 10EC05 — 10EC06 - 1 0EC07 1 0E0C8 (b) P t-I r ton r ic d foil L OEC09 I E-C O O IE-001 IE-002 IE-003 T im e (s ) IE-004. Fig. 5.9 Current vs. time at an applied potential of (a) 4.5 V and (b) 3 .5 V (b) vs. LFTLi in LIPF6 + EC + DEC at 37°C (Sample area = 1.98 cm2). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A/cm2 for Ti-3A1-2.5V, 10"° A/cm4 for Ti-6A1-4V, and 10** A/cm2 for Al, which means that Ti-3A1-2.5V had the highest dissolution rate. A layer of corrosion products can be observed with SEM on Ti-6AI-4V (Fig. 5 .10a), but a porous layer on Ti-3A1-2.5V (Fig. 5.11a). The amount o f charge passed in the 24-h tests was determined by integration of the current-time curves. After normalization by the electrode area the equivalent thickness loss was calculated according to Faraday’s law (Table 5.3). The thickness loss for the 24-h tests was 0.13 ptm for Ti-3 A1-2.5V. and 0.02 urn for Ti-6Al-4V. 5.2.2 Polarization at + 3.5 V vs. Li/Li' The applied potential (3.5 V) was within the stable potential windows for both Ticp and Pt-lr (Table 5.2). A linear region can be observed in the bilogarithmic plots (Fig. 5.9b) in the time span o f 30 to 3000 seconds with a slope o f - 0.90 for Pt-lr and - 0.97 for Ticp indicative passive behavior for both materials. Since there were no surface changes observed with SEM, it is assumed that the current vs. time behavior o f Pt-lr was due to the changes in the electrolyte. For Ticp, a current plateau was reached at current density of about I O '8 A/cm2, which suggests that no additional passivation was obtained. SEM images o f Ticp are presented in the exposed area (Fig. 5.12a) and unexposed area (Fig. 5.12b). Only slight corrosion attack can be discerned, especially at mechanical defects (“white spots” in Fig. 5.12a). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 2 (b) Fog. 5.10 SEM images for T1-6A1-4V sheet after applied potential test (4.5 V vs. LiTLi); (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 223 0> ) Fig. 5.11 SEM images for TI-3AI-2.5V sheet after applied potential test (4.5 V vs. LiTLi); (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 224 Table 5.3 Charge calculations for applied potential (4.5 V vs. L i / Li) experiments (for 24 hours). E xperime nts Charge (C) Normalized Charge (C/cm2) Equivalent Thickness Loss d (pm) Ti-6A1-4V 8.58x1 O'2 4.33 xlO*2 2.21 xlO'2 Ti-3Ai-2.5V 4.69x1O '* 2.37x 10'1 1.25x10'* Al 5.30x 10‘3 2.68 x 10'3 9.25* IQ -4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 5 (a) (b) Fig. 5.12 SEM images for Ticp fofl after applied potential test (3.5 V vs. Lf/Li); (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 6 The amount o f charge passed during a test was also calculated for Ticp and Pt-Ir. Since the charge could be associated with the decomposition of electrolyte for Pt-Ir, it was only converted to thickness loss for Ticp (Table 5.4). At 3.5 V, dissolution o f Ticp can be neglected assuming uniform corrosion. 5.3 Galvanic corrosion tests In these tests both the galvanic potential Eg and the galvanic current Ig o f the coupled materials were recorded over a 24-hour period. Before each experiment, the OCP o f each material was monitored for one hour. For AI/Ti-6Al-4V, the OCP vs. time curves before coupling are shown in Fig. 5.13a, which indicates that the OCP difference is quite large (about 0 8 V). with Al being the anode after coupling. The time dependence o f Ig is presented in Fig. 5.13c. For Ig, a linear region was observed after 100 seconds in the bilogarithmic plot indicating o f the formation o f a passive film on Al due to polarization by Ti-6A1-4V. Similar results were observed for AI/Ti-3 AI-2.5V with an OCP difference o f about 0.7 V (Fig. 5.14a) and changes o f Ig with time (Fig. 5 .14c). SEM analysis found no sign o f corrosive attack on both Al and the Ti alloys. The amount o f charge converted to a thickness loss o f the anode Al was about I O '4 pm for each galvanic test (Table 5.5). It can be concluded that there was no additional corrosion loss caused by galvanic coupling Al and Ti alloys in the test electrolyte. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 2 7 Table 5.4 Charge calculations for applied potential (3.5 V vs. Li / Li) experiments (for 24 hours). Experiments Charge (C) J Normalized Charge | (C/cm:) Equivalent Thickness Loss d (jam) Ticp 2.49x1 O'3 | l.26x!0*3 6.93 x 10"* Pt-lr 1.87x10 2 i 9.44x 10‘3 N/A N/A: does not apply. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 228 3 60C (a) 2 800 Eoc vs time n electrolyte uPP5 - E C - C E C at 37 °C ♦ — A i after one hour immersion - ♦ — T i6* t after 2 ncur immersion 2*00 2 000 2 1 C C C 3 C*C O 3 020 3 3000 -t (b) * 2 980 > - " 2960 " — - - - 2990 - 3 20000 t 0E OO 4 I 0E005 1 (c) < ^ I0E 006 ,, s > 5 1QE007 <J : QEOoa — ----- ------ — — I O E 009 " ..... ........... .... ;3 C C 7 lin e iS, *000 * 0 0 0 0 Time is) 600CO 3 0 0 0 0 1 0 100 1000 1 0 0 0 0 'irne (s> Fig. 5.13 Galvanic corrosion test for AI / Ti-6AI-4V couple in LtPF6 + EC ~ DFG at 37°C (Sample area = 1.98 cm"); (a) OCP vs. time for AI and Ti-6A1-4V before test, (b) Eg vs. time and (c) Ig vs. time Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 229 3 € 0 0 - (a) 1 2 0 0 - C Esc v$ rime .n electrolyte U?F6 h - E C * O E C at 370c 2 300 - — A t after ane ncur immersion >• — ♦ ---- E i3 2 5 after 2 nour immersion 2 AOO 20C C ".me i SI 3 200 2 300 2 -W O 2000 i 600 : 200 0 8 0 0 1 O E 004 T I O E O O S < " ~Z : o e 006 t C : < o = : O E 007 I O E 008 I 0E 009 (b) 20000 -wocc T .m e (s i 60C00 30000 ( C ) :oo :oco :oooo Tim e (s) Fig. 5.14 Galvanic corrosion test for AI / Ti-3 AI-2.5V couple in LiPF( 1 - EC - DEC at 37°C (Sample area = 1.98 cm‘): (a) OCP vs. time for AI and Ti-3A1-2.5V before test, (b) Eg vs. time and (c) Is vs. time Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 230 Table 5.5 Charge calculations for galvanic corrosion experiments (for 24 hours). I 1 j Experiments j | 1 Charge (C) I I ■ Normalized Charge j j (C/cm: ) j i i Equivalent Thickness Loss d (pm) o f the Anode i AI 1 Ti-6Al-4 V , 7.26x1 O ’ 4 1 3.67< 1 O '4 | 1.27x10"* j AI / Ti-3A1-2.5V 5.38 * I O '4 2.72. 1 O '4 j 9.38« 10'5 ' Cu / Ticp 6.82. 10*J < ; ! 3.44« 10" i I I 28. 1 O'5 1 Tic|> / Pt-lr 1 III. 10‘3 ! 5.58. 1 0 * 4 1 3.08* IC T 4 Note: the underlined material is the anode o f the galvanic couple. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. For Ticp/Pt-Ir, Ticp was the anode, and the OCP difference was about 0.05 V before coupling (Fig. 5 .15a). Ig was about 3 * lO'8 A at the start (Fig. 5 .15c), reached a plateau after about 3,000 seconds, and then dropped sharply after 60.000 seconds. Calculated charges and thickness loss o f Ticp are presented in Table 5 .5 No definite changes could be observed with SEM for both materials indicating that galvanic corrosion between Ticp and Pt-lr was negligible. For Cu/Ticp. the OCP difference between Cu and Ticp was only 0.02 V. with Ticp being the anode in the galvanic couple (Fig. 5.16a). However, their positions switched after about 20. 000 seconds with Cu becoming the anode, as indicated by the sign change o f Ig in Fig. 5.16c. The spikes in Fig. 5.16b and 5.16c were introduced by the operation o f the glove box. lg was low. within ±1 x 1 0 * * A (Fig. 16c), but its absolute value increased quickly at the end of experiment. After the 24-h experiment, a slight extent o f localized corrosion of Cu can be seen in the SEM picture (Fig. 5 .17a) compared to the unexposed area (Fig. 5 .17b). Although the charge and the calculated equivalent thickness loss were very small (Table 5.5). localized corrosion o f Cu implies a potential risk for Cu/Ticp in the test electrolyte. 5.4 EIS measurements A total o f 20 impedance spectra were recorded for each material in 30 days. The spectra were analyzed to determine the time dependence o f Rp and C in order to determine the time law o f the corrosion rate for each material. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 2 3 500 - (a) 3 400 - 200 Eoc us time in electrolyte L 1PF6 - EC * • DEC at 3T C ► — t icd ron after one nour immersion ► Pt ir foil after 2 nour immersion 2000 7 ,m e (S ) 3 23C ; 3 240 ‘ 3 2 0 0 \ 3 1 6 0 3 120 I OE 0 07 (b) 20000 40000 7im e(S) 60000 30000 < § 1 QE008 (C) L 0E 009 1 0 ICO 1000 10000 Time (s) Fig. 5.15 Galvanic corrosion test for Ticp/Pt-Ir couple in LiPF6 - EC - DEC at 37°C (Sample area = 1.98 cm2); (a) OCP vs. time for Ticp and Pt-Ir before test, (b) Eg vs. time and (c) Ig vs. time Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 ? (a) 3 360 h 3 3 2 0 - > Eoc vs time in electrolyte uPf6 • E C - G £C at 37'C 3 IS O — ♦ a f t e r o r e n o u r i m m e r s io n Ticp after 2 hour immersion 3 2 0 0 - i 32G 3 310 0 LGQG 2 0 0 0 3000 * m e < S> ; > 3 300 3290 : C E O C S 6 C E 009 - 20E009 Jj_ 20E009 — 6 O E 0C9 I O E 008 ------- (b) A 0 20000 A G O C O 60000 60000 " m e (si (c) 2 0 0 0 0 40000 60000 80000 nme (sl Fig. 5.16 Galvanic corrosion test for Cu/Ticp couple in LiPF6 - EC - DEC at 37°C (Sample area = 1.98 cm2); (a) OCP vs. time for Cu and Ticp before test, (b) Eg vs. time and (c) Ig vs. time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fig. 5.17 SEM images for Cu after Ticp -C u galvanic corrosion test; (a) exposed area, (b) unexposed area. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.4.1 EIS results A selected number o f impedance spectra for the six materials are shown in Fig. 5.18 to 5.23. The spectra for AI (Fig. 5.18) indicate that Ai behaved like an ideal capacitor since the slope o f the log |Z| - log f plot was - I and the phase angle had a value o f - 90° over a wide frequency range. The spectra for Pt-Ir (Fig. 5.19) were similar to those for AI (Fig. 5.18). however some deviation from ideal capacitive behavior could be identified at low f For the Ti materials (Fig. 5.20 to 5.22). similar spectra were obtained. There were some changes in the spectra during the first 10 days, indicating a continuous increase o f the capacitance. The changes for Ti-3 A1-2.5V and Ticp were more rapid than those for Ti-6A1-4V After the first 10 days, the spectra reached a stable state. For Cu (Fig. 5 23), pronounced changes were observed in the first 10 days. The frequency dependence o f the spectra at low f suggests that a mass transport controlled mechanism was involved in the corrosion process. 5.4.2 Analysis of EIS results Since the impedance spectra for AI (Fig. 5.18) were capacitive even at the lowest test frequency o f I mHz. Rp. i.e. the DC limit o f the impedance cannot be obtained through fitting. However, the capacitance C can be estimated from the modulus of the impedance according to the following equation: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 6 5 o * a v_/ 5i ' Bo o 7 6 V, t * ■'! B i i i a 1 <D < D \_ Q O < D T 3 _o 00 < D tf) c a 90 75 - 2h <> Id r r 5d O 10d A 15d ☆ ----- 21d 3K— 26d 53 30d 60 -45 30 -15 0 2 1 0 1 2 3 4 5 lo g (f(H z)) Fig. 5.18 Bode plots for AI foil in LiPF6 - EC - DEC at 37°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 7 E o * a '55 o 2h 7 5 d lO d 17d 2 1 d 2 6 d 3 0 d 6 5 4 3 2 3 •90 •75 a > < D 00 a > T3 N --/ 60 < D O b -45 c < c ® - s to -30 ta -C Q _ •15 3 2 1 0 1 2 3 5 4 io g ( f ( H z ) ) Fig. 5.19 Bode plots for Pt-lr foil in LiPF6 - EC - DEC at 37°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 238 E C J * a v— / 00 o 8 7 2h 6 r r — 5d 0 lOd A 15d - 6 - 2 Id 3K 26 d g f 3 0 d 5 4 3 ♦ ■ m m 2 -3 2 90 © a > oo a > X3 0 ) o o c < a > tn « a 0 1 2 • o g (f(H z )) 75 -60 “ •45 -30 -15 3 2 1 0 2 1 3 5 4 lo g ( f ( H z ) ) Fig. 5.20 Bode plots for TI-6A1-4V sheet in LiPF6 - EC - DEC at 37°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 239 £ a * a fsT 00 o 8 7 6 5 4 3 2 1 -3 o ■*sr ☆ & 2 h I d 5 d lO d 1 5 d 2 1 d 2 6 d 3 0 d * ■ m m a m * ■ m ■ 0 1 2 lo g ( f( H z )) •90 •75 a > < D w 00 < D T3 •60 < D oo '45 c < S > -30 .c Q _ •15 • 3 - 2 -1 0 1 2 3 4 5 lo g ( f (Hz)) Fig. 5.21 Bode plots for T1-3A1-2.5V sheet in LiPF6 - EC - DEC at 37°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 240 8 7 2 h E o * a n _/ 00 o 5d lOd 15d 21d 26d 30d 6 a 5 4 3 2 -1 0 1 2 lo g (f(H z )) •9 0 ---- •75 — < v 2 > §? -60 A O M -45 oo C -n < ® 30 — r C O -15 0 V -3 -1 2 1 3 4 5 l° g ( f ( H z » Fig. 5.22 Bode plots for Ticp sheet in LiPFf t - EC - DEC at 37°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 241 E o * a 0 0 o 8 7 6 2h Id 5d O lOd i£sr— 15d ☆------ 21d ^ — 26d ^ ---- 30d 5 4 3 2 10 12 lo g ( f ( H z ) ) 90 y - - - - « < D aj ab < D T3 •60 03 00 c <c < D tn c o 45 s z Q. ■15 • 3 - 2 - 1 0 1 2 3 4 5 lo g C f(H z )) Fig. 5.23 Bode plots for Cu foil in LiPF6 - EC - DEC at 37°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 242 where jZj is the modulus o f the impedance at the frequency f. For accuracy and consistency, C values were calculated and averaged in the frequency range o f 0.1 to 10 Hz for all spectra. The impedance spectra for Pt-Ir and the three Ti materials had simple OTC characteristics and were fitted to the OTC model (Fig. 3.3 of Part I). For Cu. the impedance spectra were fitted to a TTC model, as illustrated in Fig. 3 4 o f Part I. The inverse values o f fit Rp and C vs. time for all the test materials are presented in Fig. 5.24a and Fig. 5.24b, respectively. For Pt-Ir. linear relations exist for both Rp and C. For Cu. a linear relation was observed only for Rp (Fig. 5.24a). For the three Ti materials, no linear relation could be obtained for either Rp or C The capacitance o f AI remained almost unchanged during the 30-day test period (Fig. 5.24b). 5.4.3 Corrosion rate time law and stability prediction The corrosion rate r ^ is inversely proportional to polarization resistance R p: B * F.W I /* = - > 27 «-------------« — 3-4 1 D R p where rw is given in pm/year, EW is the equivalent weight o f the test material in g. and D is the density o f electrode in g/cm3 B values can be found in Table 5 1 for each test material. l/Rp is usually called relative corrosion rate and is used here in the determination o f corrosion rate time laws. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 243 -40 ao - : o 0 5 oo 05 : o i s LogCt(day)) (a) 6 00 -------------- - * ® a -a a - a a- aaaa— m 5 60 •20 g> 480 a - 4 40 4 00 — ■ T i o 4 i > ~ - T'3-2 5 1 t .c 0 'O C u & 4 1 * Pt I r i f •4 a 150 1 00 0 50 0 00 0 50 _oe(t(cayV. : oo 1 50 (b) Fig. 5.24 (a) log (1/Rp) vs. log (t) and (b) log (I/C) vs. log (t). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 244 The linear relation between log ( 1/Rp) and log (t) for Pt-Ir and Cu can be fitted to the following equation: log ( l/Rp) = a log (t) + b 5-5 The fit parameters a, b and correlation coefficient (R: ) for each fit are presented in Table 5.6. The estimated thickness loss over 10-vear period in the test electrolyte can be predicted by first combining Eq. 5-4 and 5-5 to obtain r ^ , then integrating r ^ over 10 years. The predicted thickness losses for Cu and Pt-Ir are shown in Table 5.6. A Cu foil current collector with a thickness o f 10 pm is commonly used in commercial Li-ion batteries. Assuming uniform corrosion from both sides, after 10 years the thickness o f the Cu foil would be 6.4pm. still thick enough to act as current collector and support the anode electrode. However, since Cu is prone to localized corrosion, as discussed earlier, pitting penetration on Cu foil within 10 years is possible even under small anodic polarization, which would lead to failure in collecting current from the anode electrode. Due to the high stability of Pt-Ir. the calculated thickness loss is not due to corrosion o f Pt-Ir; rather it could be due to the decomposition of the test electrolyte. Since the thickness loss is very small (only 0.05 pm over 10 years), there should be no real concern from the perspective o f practical application o f whether corrosion o f Pt-Ir or decomposition o f the electrolyte would occur. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 245 Table 5.6 Parameters a, b and correlation coefficient (R2) from fitting Rp to the equation of log (1/Rp) = a log (t) +b and predictions of thickness loss after 10 years for Cu and Pt-Ir. j j ' | “ j Thickness Loss I 1 ------------------------ Cu j ... _ -0.208 1 - 5 , 5 7 ! t i — i 0.81 ' i u Vtuu) 1.777 | Pt-Ir i -0.369 1 i -7.251 i 0.94 ; 1 0.005 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 6 For the Ti materials, no linear relation existed between log ( I/Rp) and log (t) (Fig. 5.24a), and therefore no time law could be obtained. However, after 30 days l/Rp showed a decreasing trend for all the three Ti materials (Fig. 5.24b). Therefore, a conservative strategy for stability prediction was employed by assuming that the corrosion rate determined at the end o f the 30-dav test period would remain constant for the 10-year test period. The calculated thickness loss for the 30-day test period, the corrosion rate at the 30th day and the thickness loss from the 30th day to 10 years as well as the total thickness loss in 10 years for the Ti materials are listed in Table 5.7. Ticp had the largest thickness loss (about 0.15 pm), while Ti-6AI-4V had the smallest thickness loss o f about 0 .07 pm. The Ti materials are used as case materials, usually with thickness o f 400 pm (0.4 mm); therefore, the thickness loss resulting from uniform corrosion during 10 years can be ignored. The capacitance values for oxide covered materials such as AI and Ti alloys can be expressed as: C - esoA/d 5-6 where e is the dielectric constant of the oxide. s„ the dielectric constant o f vacuum. A is the electrode area and d is the thickness o f the oxide layer. Fig. 5 .24b illustrates the time dependence of the relative oxide thickness, expressed as l/C. for all the test materials assuming that e was constant. It can be seen that the thickness o f the oxide layer did not change for AI and Pt-Ir over the 30-day test period, indicating the high stability o f these two materials. Assuming e - 10 for aluminum oxide, a thickness o f Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 7 Table 5.7 Conservative thickness loss calculations for Ti materials. ! ! 1 Thickness loss in the first 30 days (pm) ! Corrosion ! rate at 30th | day (pm/year) Thickness loss from 1 30th day to 10 years , (pm) calculated from ! 30th day corrosion j rate Total thickness loss d (pm) Ti-6A1- 4V 0.001 I , 6.605 < I O'3 ! I 0.066 I 0.067 Ti-3A1- 2.5V 0.003 ! | 7.909x 10'3 i 0.073 1 0.081 i Ticp 0.004 l.499x I O'2 0.149 | 1 0.153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 248 about 4 nm was determined, which agrees with the thickness o f aluminum oxide naturally formed on AI at ambient temperature [28]. For the three Ti materials the oxide layer thickness seemed to decrease continuously with exposure time (Fig. 5 .24b), most likely due to dissolution o f the titanium oxide in the electrolyte. 5.4.4 Surface analysis The test materials were examined with SEM after the 30-day EIS tests (Fig. 5.25). It was found that the surfaces of the .A I and Pt-Ir samples were virtually unattacked. No significant localized or uniform corrosion damage was observed on the surface o f Cu (Fig. 5.25a). as a result o f the low corrosion rate of Cu in the test electrolyte at OCP. Localized attack seemed to have occurred on Ticp with a porous layer evident after the 30-dav experiment (Fig. 5 25b). The surfaces o f Ti-6AI-4V (Fig. 5 25c) and Ti- 3 AI-2.5V (Fig. 5.25d) were covered with a layer o f corrosion products. EDS analyses were performed in an attempt to identify the chemical compositions o f the corrosion products on the Ti materials. No obvious differences in chemical composition between exposed and unexposed areas were detected for the three Ti materials. Therefore. AES was employed for surface chemistry analyses on the EIS samples o f Ti-6A1-4V and Ti-3 A1-2.5V. The results were similar for the two Ti alloys. The corrosion products consisted mainly o f oxide and fluoride o f lithium and titanium. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 249 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 0 (C) (< D Fig. 5.25 SEM images after 30-day immersion EIS test; (a) Cu foil, (b) Ticp foil, (c) Ti-6AI-4V sheet and (d) Tt-3A1-2.5V sheet. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 251 6. CONCLUSIONS 1. Electrochemical techniques have been applied successfully and effectively to the evaluation o f the stability o f metallic materials in non-aqueous organic Li-ion battery electrolyte at elevated temperature. These techniques include EIS, potentiodynamic and potentiostatic polarization curves, and galvanic corrosion measurements. 2. The stable potential window for a metallic material in the battery electrolyte can be determined using potentiodynamic polarization curves. The extent of the stable potential window for each test material from high to low can be ranked as: AI > Ti-6AI-4V > Ti-3A1-2.5V > Pt-Ir > Ticp > Cu Cu has practically no stable potential window. It can be easily activated under anodic polarization and severely attacked by localized corrosion. The stability of the Ti material increased with increasing concentration o f alloying element AI or V. AI and Pt-Ir are remarkably immune to corrosion in the non-aqueous organic electrolyte. Pt- Ir could cause decomposition o f electrolyte resulting in a low stable potential window. 3. Due to the high degree o f passivity o f AI. Pt-Ir and the three Ti materials, coupling o f any two o f these materials would cause only negligible galvanic corrosion. Since Cu is highly sensitive to anodic potentiaL coupling between Cu and one o f the other materials with higher OCP could cause localized corrosion on Cu. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 252 4. At OCP. no significant corrosion will occur on AI or Pt-lr in the battery electrolyte even for a long-term period, e.g. 10 years. Negligible corrosion can occur for the Ti materials with the main corrosion products probably being lithium and titanium oxides and fluorides. Among the test materials, Cu has the highest corrosion rate, which is controlled by a mass transfer process. However, corrosion o f Cu could not be a big concern if it occurs as uniform corrosion. 5 The battery electrolyte demonstrated high chemical stability in the drv-box at 37°C. No chemical decomposition had been observed or detected when the electrolyte was in contact with Al and the three Ti materials polarized up to 6.5 V vs. L i/L f. The 90Pt/10Ir would decrease the stability o f the electrolyte due to its well- known stability and high catalytic activity. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 3 7. SUGGESTIONS FOR FUTURE WORK 7.1 The effect o f Alloying element A 1 on the stability o f Ti-Al alloys It was found in the study o f the stable potential window o f Ti materials that a higher concentration o f the alloying elements A 1 and V in Ti alloys produced a lower OCP and a higher upper limit leading to a larger stable potential window (Table 5.2). Since the A 1 foil had the lowest OCP and highest stable window. Al could be the major alloying element that contributes to the stability o f the Ti alloys. To prove this hypothesis, potentiodynamic polarization tests are needed to determine the stable potential windows o f a Ti-Al alloy and a Ti-V alloy with same alloying element concentration. If the Ti-AI alloy has higher stable window, cheaper Ti-AI alloys can be used in Li-ion batteries in place of the current Ti-AI-V alloys. 7.2 Corrosion protection o f Cu Battery-grade pure Cu foil is the most common anode current collector in Li-ion batteries. However, it was found in this study that Cu was susceptible to localized corrosion even under small anodic polarization. In Li-ion batteries, over-discharge can increase the OCP o f anode electrode above the OCP o f the Cu current collector [15], and therefore impose a polarization risk on Cu. Because o f safety concerns in a long-term period, e.g. 10 years, proper corrosion protection o f Cu is necessary. It was reported that graphite coatings could not provide lasting protection o f Cu due to adhesion problems [16]. In aqueous environments. Cu and alloys can be effectively Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 254 protected from corrosion by the addition o f inhibitors [29], The inhibition efficiency o f candidate soluble inhibitors in battery electrolytes can be determined in the Li-ion battery electrolyte with polarization curves or EIS. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 255 8. R EFER EN CES 1. R. J. Brodd. Interface. 8. 20 (1999). 2. S. Hossain. in Handbook o f Batteries. 2n d Ed.. D. Linden. Editor, p. 36.1, MacGraw-Hill Inc., New York (1995). 3. C. A. Vincent and B. Scrosati. in M odern Batteries- An Introduction to Electrochem ical Power Sources. 2nd Ed., p. 199. John Welev & Sons Inc.. New York (1997). 4 C F Holmes. Interface. 8. 32 (1999). 5. http://www. nist.gov/public_afFairs/atp2000/00004050. htm. 6. C. A. Vincent and B. Scrosati. in M odern Batteries- An Introduction to Electrochem ical Power Sources. 2nd Ed., p. 204. John Welev & Sons Inc.. New York (1997). 7. A. L. Bement. Biomaterials: Bioengineering Applied to Materials for Hard and Soft Tissue Replacement. University o f Washington Press. Seattle and London (1971). 8. N. D. Green et al. Journal o f Materials. I. 2 (1966). 9 C G. Barlow. Electrochem. and Solid-State Letters. 2. 362 (1999). 10. W. K. Behl and E. J. Plichta. J. Power Sources. 72, 132 (1998). 11. Y Chen. T. M. Devine. J.E. Evans. O. R. VIonteiro. and I. G. Brown. J. Electrochem. Soc.. 146, 1310 (1999). 12. L. Peter and J. Arai. J. Applied Electrochem.. 29. 1053 (1999). 13. H. Yang, K. Kwon, TM. M. Devine, and J. E. Evans, J. Electrochem. Soc.. 147. 4399 (2000). 14. J. W. Braithwaite. A. Gonzales. G. Nagasubramanian. S. J. Lucero, D. E. Peebles. J. A. Ohlhausen and W. R. Cieslak. J. Electrochem. Soc.. 146. 248 (1999). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 6 15. M. Zhao, S. Kariuki. H. D. Dewald. F. R. Lemke, R. J. Staniewicz. E. J. Plichta and R. A. Marsh, J. Electrochem. Soc., 147, 2874 (2000). 16. M. Zhao, H. D. Dewald, F. R. Lemke and R. J. Staniewicz. J. Electrochem. Soc., 147, 3983 (2000). 17. E. H. Dix, Corrosion o f Light M etals, p. 131, American Society for Metals, 1964. 18. J. W. Brahhwaite. G. Nagasubramanian. A. Gonzales, S. J. Lucero and W. R. Cieslak. The Electrochem ical Society Proceedings Series, p 44. PV96-17, Pennington, NJ, 1996. 19 B. Scrosati. Electrochimica Acta 45. 2461 (2000). 20. D. FI. Jang and S. M. Oh. J. Electrochem. Soc., 144. 3342 (1997). 21 Y Xia. Y. Zhou and VI. Yoshio. J Electrochem. Soc.. 144. 2593 (1997). 22. G. G. Amatucci. J. M. Tarascon and L. C. Klein. Solid States Ionics. 83. 167 (1996). 23. A. Webber. J. Electrochem. Soc.. 138. 2586 (1991). 24. L. J. Krause. W Lamanna. J. Summerfield. M. Engle. G. Korba. R. Loch and R. Atanasoski. J. Power Sources. 68. 320 (1997). 25. D. A. Shifler. P J. Moran and J. Kruger. Electrochim. Acta. 40.897 (1995). 26. F. Mansfeld, C. H. Tsai and H. Shih. Computer M odeling in Corrosion. ASTM STP 1154. R. S. Munn. Ed., p. 186. ASTM. Philadelphia (1992). 27. H. Shih and F. Mansfeld. Computer M odeling in Corrosion. ASTM STP 1154. R. S. Munn. Ed., p. 174. ASTM. Philadelphia (1992). 28. H. Kaesche, M etallic Corrosion, p. 269, NACE (1985). 29. Y. I. Kuznetsov. Organic Inhibitors o f Corrosion o f M etals, p. 49. Plenum Press, New York (1996). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 7 BIBLIOGRAPHY A M. -H. Ahn and K. -K. Baek. Corrosion 99, Paper No. 97, NACE, Houston (1999). M. AjmaL, A S. Mideen and M. A Quraishi, Corro. Sd., 36, 79 (1994). A. J. Aldykewicz. Jr., A. J. Davenport and H. S. Isaacs, J. Electrochem. Soc., 143, 147(1996). A J. Aldykewicz, Jr., H. S. Isaacs and A J. Davenport, J. Electrochem. Soc., 142, 3442(1995). G. G. Amatucci. J. M. Tarascon and L. C. Klein, Solid States Ionics, 83, 167 (1996). Unpublished experimental results from AMTI, Watertown, MA. J. R. Ambrose. Corrosion. 34.27 (1978). A nnual Book o f ASTM Standards, Vol. 03.02, G102-89. Standard Guide for Calculation o f Corrosion Rates and Related Information from Electrochemical Measurements, ASTM (1997). Annual Book o f ASTM Standards, Vol. 03.02, G10I-94, Standard Guide for Estimating the Atmospheric Corrosion Resistance o f Low-Alloy Steels, ASTM (1997). Annual Book o f ASTM Standards, Vol. 03.02, G5-94, Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements, ASTM (1997). D. R Amott, B. R W. Hinton and N. E. Ryan. Materials Performance, 26,42 (1987). J. Aromaa and A Klarin, M aterials, Corrosion Prevention an d M aintenance, p. 99, TAPPI press, Helsinki (1999). V. Ashworth and P. J. Boden, Corr. Sd. 10, 709 (1970). Australian Patent, P 10649. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 258 B A. Bahadur, Mater. Trans., JIM, 34, 1191 (1993). A. Bahadur, Mater. Trans. JIM, 3 7 .605 (1996). C. G. Barlow, Electrochem. and Solid-State Letters, 2, 362 (1999). B. Bavarian, A. Moccari and D. D. Macdonald, Corrosion, 38, 104 (1982). A. M. Beccaria, M. Ghiazza and G. Poggi, Corro. Sci., 36, 1387 (1994). W. K. Behl and E. J Plichta, J. Power Sources, 72, 132 (1998). A. L. Bement, Biomaterials: Bioengineering Applied to Materials for Hard and Soft Tissue Replacement, University of Washington Press, Seattle and London (1971). S. Bernal, F. J. Botana, J. J. Calvino and M. A. Cauqui, J. Alloys Comp., 225,638, (1995). M. Bethencourt, F. J. Botana and M. A. Marcos, J. Alloys Comp., 250,455 (1997). B. Beverskog and I. Puigdomenech, Corr. Sci. 38, 2121 (1996). S. Bohm, R. G reet H. N. McMurray, S. M. Powell and D. A. Worsiey, J. Electrochem. Soc., 147, 3286 (2000). S. A. Bradford, Corrosion Control, p. 241, Van Nostrand Reinhold, New York (1993). J. W. Braithwaite, A. Gonzales, G. Nagasubramanian. S. J. Lucero, D. E. Peebles, J. A. Ohlhausen and W R. Cieslak, J. Electrochem. Soc., 146, 248 (1999). J. W. Braithwaite, G. Nagasubramanian, A. Gonzales, S. J. Lucero and W. R. Cieslak, The Electrochem ical Society Proceedings Series, p. 44, PV96-17. Pennington, NJ. 1996. C. B. Breslin, C. Chen and F. Mansfeld, Corr. Sci., 39, 1061 (1997). C. B. Breslin and M. Geary, Corrosion, 5 4 ,964, (1998). R. J. Brodd, Interface, 8, 20 (1999). P. A. Brook, Corr. Sci. 11,389 (1971). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 259 J. A _ V. Butler, Transactions o f the Faraday Society, 19, 729 (1924). C Y. Chen, T. M. Devine, J.E. Evans, O. R. Monteiro, and I. G. Brown, J. Electrochem. Soc., 146, 1310(1999). C. Chen and F. Mansfeld, Corr. Sci., 39, 1075 (1997). M. Cohen, Corrosion, 32,461 (1979). M. Cohen, J. Phys. Chem., 5 6 ,4 5 1 (1952). D. Cubicciotti, J. Nucl. Mater. 167, 241 (1989). D A. J. Davenport, H. S. Isaacs and M. W. Kendig, Corr. Sci., 3 2 ,653 (1991). E. H. Dix, Corrosion o f Light M etals, p. 131, American Society for Metals, 1964. DOE Energy Resources R&D Portfolio: FY 1999-2000, p. 244, DOE (2000). P. A. Domanski, J. Res. Natl. Inst. Stand. Technol.. 103, 529 (1998). E T. Erdey-Gmz and M. Volmer, Zeitschrift fur Physikund Chemie, 150A, 203 (1930). G M. Geary and C. B. Breslin, Corr. Sci., 39, 1341 (1997). J. Gerstmann and C. L. Hannon. Development o f Corrosion Inhibitors fo r Absorption H eat Pum ps, p. I, AMTI (Advanced Mechanical Technology, Inc.), Watertown. MA (1999). J. D. Gorman, S. T. Johnson, P. N. Johnston, P. J. K. Paterson and A. E. Hughes, Corr. Sci., 38, 1977 (1996). N. D. Green et al. Journal o f Materials, 1 ,2 (1966). N. N. Greenwood and A. Eamshaw, Chem istry o f the Elem ents, 2nd Ed., p. 1227, Butterworth-Heinemann, Oxford, GB (1997). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 0 N. N. Greenwood and A. Earnshaw, Chem istry o f the Elem ents, 2nd Ed., p. 1294, Butterworth-Heinemann. Oxford, GB (1997). F. Gruber and K. Mall, Linde Berichte aus Technik und Wissenschaft, 61 (1988). H N. Hackerman and E. S. Snavely, N A C E B asic Corrosion Course, p. 9-11, NACE (1970). F. H. Hajjar and W. T. Riad, British Corrosion Journal, 25, 119 (1990). T. J. Haley, J. Pharm. Sci.. 54, 633 (1965). C. L. Hannon. J. Gerstmann, F. Mansfeld and Z. Sun, in Symp. o f Heat Pump and Refrigeration System Design, Analysis and Applications, p. 35, ASME, Orlando ( 2000). G. B. Hatch, Corrosion Inhibitors, Ed. C. C Nathan, p. 126, NACE (1973). Heat Pump Systems: A Technology Review, p. 17, Organization for Economic Co operation and Development, Paris (1982). B. Heeg, T. Moros and D. Klenerman, Corro. Sci., 40, 1313(1998). B. Heeg, T. Moros and D. Klenerman. Corro. Sci.. 40, 1303(1998). E. Heitz and G. Kreysa, Principles o f Electrochem ical Engineering, p. 169, VCH Publishers, New York (1986). G. F. Hewitt, G. L. Shires and T. R. Bott, Process H eat Transfer, CRC Press Inc. (1994). B. Hindin, Evaluation o f Alternative Corrosion Inhibitors for Gas-Fired Ammonia/Water Refrigeration Machines, Final report, p. I, GRI (1995). K. Hinkelmann and O. Kempthome, Design a n d Analysis o f Experiments. VI, Introduction to Experim ental Design, John Wiley & Sons, Inc., New York (1994). B. R. W. Hinton, Journal o f Alloys and Compounds, 180, 15 (1992). B. R. W. Hinton and R. R. Amott, Microstructural Sci., 17, 3 11 (1989). B. R. W. Hinton, D. R. Amott and N. E. Ryan, Metals Forum, 7 ,2 1 1 (1984). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 1 B. R. W. Hinton, D. R _ Amott and N. E. Ryan, Mater. Forum, 9, 162 (1986). B. R. W. Hinton and L. Wilson, Corrosion Science, 29,967 (1989). C. F. Holmes, Interface, 8, 32 (1999). S. Hossain, in Handbook o f Batteries, 2nd Ed., D. Linden, Editor, p. 36.1, MacGraw- Hill Inc.. New York (1995). C. Hsu and F. Mansfeld, Technical Note: Concerning the Conversion o f the Constant Phase Element Parameter Yo into a Capacitance, Corrosion (in press). http://www.angelfire.com/mt/chemicalcleaning/magnetite.html http://www. nist. gov/public_afFairs/atp2000/00004050. htm. http://www.polarpowerinc.com/products/refrigerator/ref-tech-overview.htm A. E. Hughes, J. D. Gorman and P J. K. Paterson, Corr. Sci., 38, 1957 (1996). A. E. Hughes, S. G. Hardin, R. M. Peter and W. W. Klaus, Paper no. III. NACE 2000, Orlando. Florida (2000). A. E. Hughes, K. J. H. Nelson and P. R. Miller. Mater. Sci. & Tech.. 15, 1124 (1999). A. E. Hughes, R. J. Taylor, B. R. W. Hinton and L. Wilson. Surf Inter. Anal.. 23. 540(1995). A. E. Hughes, R. J. Taylor, K. J. H. Nelson, B. R W. Hinton and L. Wilson, Mater. Sci. & Tech., 12, 928 (1996). C. M. Hwa. Mater. Perform.. 10. 249 (1971). I G. E. Ibbotson and A. Marshall, Chem ical Inhibitors fo r Corrosion Control, Ed. B. G. Clubley, p. 89, Royal Society o f Chemistry (1990). International Patent WO 88/06639. International Patent WO 95/08008. H. S. Isaacs, A. J. Davenport and A. Shipley, J. Electrochem. Soc., 138, 390 (1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 2 S. Iyoki and T. Uemura, Reffegeration, 55, 353 (1980). J D. H. Jang and S. M. Oh, J. Electrochem. Soc., 144, 3342 (1997). D. A. Jones, Principles an d Prevention o f Corrosion, 2nd Ed., p. 119, Prentice-Hall, Inc., NJ (1996). K H. Kaesche, M etallic Corrosion, p. 269, NACE (1985). V. Kain, G. E. Prasad and H. S. Gadiyar, India J. o f Tech., 30, 341 (1992) E. Kalman, Corrosion Inhibitors, Euro. Fed. o f Corr. Pub. No. 11, p. 12, The Institute o f Materials (1994). D. H, Kalota and D. C. Silverman, Corrosion, 50, 138 (1994). F. M. Kharafi and F. H. Hajjar. British Corrosion Journal, 25, 209 (1990). A. I. Khuri and J. A. Cornell, Response Surfaces: Designs and analyses, Marcel Dekker, Inc., New York (1987). D. G. Kolman and S. R. Taylor, Corrosion, 49, 635 (1993). L. J. Krause, W. Lamanna, J. Summerfield, M. Engle, G. Korba, R. Loch and R. Atanasoski, J. Power Sources, 68, 320 (1997). V. Kucera and E. Mattsson, in Corrosion M echanism s, F. Mansfeld Ed, p. 211. Marcel Dekker, Inc., New York (1987). Y. I. Kuznetsov, Organic Inhibitors o f Corrosion o f M etals, p. 225, Plenum Press. New York (1996). Y. I. Kuznetsov, O rganic Inhibitors o f Corrosion o f M etals, p. 49, Plenum Press, New York (1996). Y. I. Kuznetsov and T. I. Bardasheva, Protection o f Metals, 2 8 ,450 (1992). L H. H. Le and G. Ghali, J. Appl. Electrochem. 23, 72 (1993). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 3 S. J. Lennon and F. P. A. Robinson, Corr. Sci. 26, 995 (1986). R. J. Lewis, Sr., Hazardous Chemicals Desk Reference, 4th ed., p. 292, Van Nostrand Reinhold, New York (1997). S. L. Li, Mat. Prot., 26, 6 (1993). F. Lin, R. C. Newman and G. E. Thompson, Electrochim. Acta, 42,2455 (1997). S. Lin, H. Shih and F. Mansfeld, Corr, Sci., 33, 1331 (1992). P. B. Linkson, P. B. Phillips and C. D. Rowles, Corr. Sci. 19,613 (1979). Y. C. Lu and M. B. Ives, Corrosion Science, 34, 1773 (1993). M D. D. Macdonald. B.C. Syrett and S. S. Wing, Corrosion 35, I (1979). F. Mansfeld, in Advances in Corrosion Science an d Technology, Vol. 6, p. 163. Plenum Press (1976). F. Mansfeld, J. Appl. Electrochem. 25. 187 (1995). F. Mansfeld, Corrosion. 29, 397 (1973). F. Mansfeld, in Electrochem ical Techniques fo r Corrosion Engineering, R. Baboian. Ed., p. 67, NACE, Houston (1986). F. Mansfeld, C. Chen and C. B. Breslin. J. Electrochem. Soc., 145, 2792 (1998). F. Mansfeld. S. Lin. S. Kim and H. Shih, Corr. Sci., 27, 997 (1987). F. Mansfeld. S. Lin, S. Kim and H. Shih, Electrochim. Acta, 34, 1123 (1989). F. Mansfeld, S. Lin, S. Kim and H. Shih, J. Electrochem. Soc.. 137, 78 (1990). F. Mansfeld, S. Lin. S. Kim and H. Shih. Proc. EuroCorr’87, p. 713. Karlsruhe (1987). F. Mansfeld and M. W. Kendig, Proc. 6th Euro. Symp. Corros. Inhibitors, Ferrara. Italy (1985). F. Mansfeld and M. W. Kendig, Werkst. Korros., 34, 397 (1983). F. Mansfeld, M. W. Kendig, A. F. Allen and W. J. Lorenz, Proc. 9th Int. Conf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 4 F. Mansfeld, M W. Kendig and W. J. Lorenz, J. Electrochem. Soc., 132, 190 (1985). F. Mansfeld, M. W. Kendig and S. Tsai, Corrosion, 38,478 (1982). F. Mansfeld and K. B. Oldham, Corr. Sci., 11, 787 (1971). F. Mansfeld and Z. Sun, US patent application 09/774,540, Corrosion Protection o f Steel in Ammonia/Water Heat Pumps. F. Mansfeld, C. H. Tsai and H. Shih, in Computer M odeling in Corrosion, ASTM STP 1154, R. S. Munn, Ed., p. 186, ASTM, Philadelphia (1992). F. Mansfeld, Y. Wang and H. Shih, J. Electrochem. Soc., 138, L74 (1991). F. Mansfeld and Y. Wang, Brit. Corr. Jour., 29, 194 (1994). F. Mansfeld and T. Wood, Personal communication between, Aug., 1999. J. E. O. Mayne and M. J. Pryor, J. Chem. Soc., 1831 (1949). Metallic Corrosion, Toronto, Canada, Vol. 1, 368 (1984). E. C. McKelvy, and A. Isaacs, Technologic Papers o f the Bureau o f Standards, No. 180, Washington, Government Printing Office (1920). D. K. Miller, Handbook o f Applied Thermal Design, edited by E. C. Guyer and D. C. Brownell, McGraw-Hill Inc. (1989). S. K. Mitra, S. K. Roy and S. K. Bose, Oxidation o f Metals, 39, 221 (1993). C. Monticelli, A. Frignani, G. Brunoro, G. Trabanelli, F. Zucchi and M. Tassinari, Corro. Sci., 35, 1483 (1993). H. G. Mosle, Metalloberflaeche, 23, 358 (1969). E. Mueller, C. S. Sikes and B. J. Little, Corrosion, 49, ? (1993). P P. Paine, Corrosion 46, 19 (1990). L. Peter and J. Arai, J. Applied Electrochem., 29, 1053 (1999). J. V. Petrocelli, J. Electrochem. Soc., 97, 10 (1950). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 5 J. Philips, C ontrol a n d Pollution O ptions fo r Am monia Em ission, p. 47, U. S. Environmental Protection Agency, NC (1995). M. Pourbaix, in Pro. o f the 7th Int. Congress on Metallic Corrosion, Rio de Janeiro, (1978). S. Powell, Surf Eng. 16, 169 (2000). S. M. Powell, H. N. McMurray and D. A. Worsley, Corrosion, 55, 1040, (1999). R A. L. Rudd, C. B. Breslin and F. Mansfeld, Corr. Sci., 42, 275 (2000). S G. Schikorr, Z. Anorg. Allg. Chem. 212, 33 (1933). G. Schikorr, Z. Elektrochem. 25, 65 (1929). B. Scrosati, Electrochimica Acta 45, 2461 (2000). S. Seal, S. E C Bose and S. K. Roy, Oxidation o f Metals, 41, 139 (1994). D. A. Shifler, P. J. Moran and J. Kruger, Electrochim. Acta, 40,897 (1995). H. Shih and F. Mansfeld, in Com puter M odeling in Corrosion, ASTM STP 1154, E L S. Munn, Ed., p. 174, ASTM, Philadelphia (1992). E L N. Singh, S. E C . T'rwari and W. R. Singh, J. Appl. Electrochem., 22, 1175 (1992). I. B. Singh, G. Venkatachari and E C Balakrishnan, J. o f App. Electrochem. 24, 179 (1994). M. Stem and A. Geary, J. Electrochem. Soc., 104, 56 (1957). T. Suwa, Corr. Eng., 37, 73 (1988). Z. Szlarska-Smialowska, Pitting Corrosion o f Metals, Chap. 3, NACE, Houston (1986). T S. Tougoose, C hem ical Inhibitors fo r Corrosion Control, Ed. B. G. Clubley, p. 73, Royal Society o f Chemistry (1990). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 6 H. E. Townsend, Corr. Sci. 10, 343 (1970). G. Trabanelli, Corrosion, 47,410 (1991). W. Tsai, A. Moccari and D. D. Macdonald, Corrosion, 39, 1 (1983). U H. H. Uhlig and R. W. Revie, Corrosion and Corrosion Control, 3rd ed., p.414, John Wiley & Sons, New York (1985). US patent 2,775,170. US Patent 5,342,578. US Patent 5,811,026. US patent 4,961,878. US patent 4,915,872. US patent 4.098,720. US patent 4,724,125. US patent 4.382.008. US patent 2,972,581. US patent 5,871,691. US patent 6,004,475. US patent 5,725,749. US patent 4,749,550. US patent 5,130,052. US patent 5,582,654. US patent 5,635,084. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 7 V N. Verma, W. R. Singh, S. K. Tiwari, and R. N. Singh, Brit. Corr. Jour., 25, 131 (1990). C. A. Vincent and B. Scrosati, in M odem B atteries-A n Introduction to Electrochem ical Power Sources, 2n d Ed., p. 199, John Weley & Sons Inc., New York (1997). C. A. Vincent and B. Scrosati, in M odem B atteries-A n Introduction to Electrochem ical Power Sources, 2nd Ed., p. 204, John Weley & Sons Inc., New York (1997). W C. Wagner and W. Traud, Z. Electrochem., 44, 391 (1938). Y. Wang, “Corrosion Protection o f A 1 Alloys by Surface Modification Using Chromate-free Approaches”, Ph. D. thesis, University o f Southern California, Dec. 1994. A. Webber, J. Electrochem. Soc., 138, 2586 (1991). E. V. Westing, D eterm ination o f Coating Performance with Impedance M easurements, Ph.D. Dissertation, p. 104, TNO Centre for Coating Research, Netherlands (1992). X Y. Xia, Y. Zhou and M. Yoshio, J. Electrochem. Soc., 144, 2593 (1997). Y H. Yang, K. Kwon, TM. M. Devine, and J. E. Evans, J. Electrochem. Soc., 147,4399 (2000). X. W. Yu, C. N. Cao and Z. M. Yao, J. Mater. Sci. Letters, 19, 1907 (2000). Z J. L. Zaldes, J. W. Mitchell and J. A. Mosovsky, Proceedings o f 27* International SAMPE Technical Conference, p. 315, Ed. R. J. Martinez, Albuquerque, NM, USA (1995). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 8 M. Zhao, H. D. Dewald, F. R. Lemke and R. J. Staniewicz, J. Electrochem. Soc 147, 3983 (2000). M _ Zhao, S. Kariuki, H. D. Dewald, F. R. Lemke, R. J. Staniewicz, E. J. Plichta and R. A. Marsh, J. Electrochem. Soc., 147, 2874 (2000). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Electrochemical test methods for advanced battery and semiconductor technology
PDF
EFAB: A novel, high-aspect-ratio true three-dimensional microfabrication process for rapid, low-cost desktop micromachining of MEMS
PDF
Development of steel foam processing methods and characterization of metal foam
PDF
Evaluation of the superplastic potential in commercial aluminum alloys through equal -channel angular pressing
PDF
Deformation of bulk nanocrystalline aluminum alloys
PDF
Creep deformation in superplastic yttria -stabilized zirconia ceramics and yttria-stabilized zirconia /alumina composites
PDF
Corrosion protection of anodized aluminum alloys and aluminum metal matrix composites and stainless steels
PDF
Developments of chromium-free conversion coatings on galvanized steels
PDF
Chemisorption on the (111) and (100) faces of platinum-tin bimetallic surfaces
PDF
Investigation of sacrificial layer and building block for layered nanofabrication (LNF)
PDF
Dispersive properties in photonic crystals applications to beam-steering and wavelength demultiplexing
PDF
Design and applications of MPEG video markup language (MPML)
PDF
Generation, filtering, and application of subcarriers in optical communication systems
PDF
Combining compile -time and run -time parallelization
PDF
Design considerations in the application of membrane reactors for use in steam-carbon dioxide reforming environments
PDF
Chromatic and polarization mode dispersion monitoring for equalization in optical fiber communication systems
PDF
Chemisorption studies of hydrocarbons on clean and hydrogen precovered platinum(111) and tin/platinum(111) surface alloys
PDF
An implicit-based haptic rendering technique
PDF
Ferroelectric-Paraelectric Phase Transitions In The Lead(X) Calcium(1-X) Titanium Trioxide And The Lead(X) Barium(Y) Strontium(1-X-Y) Titanium Oxide Perovskite Systems
PDF
Development of ceramic-to-metal package for BION microstimulator
Asset Metadata
Creator
Sun, Zhaoli
(author)
Core Title
Application of electrochemical methods in corrosion and battery research
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Materials Science
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
engineering, materials science,OAI-PMH Harvest
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Mansfeld, Florian (
committee chair
), [illegible] (
committee member
), Goo, Edward (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-213080
Unique identifier
UC11334693
Identifier
3065856.pdf (filename),usctheses-c16-213080 (legacy record id)
Legacy Identifier
3065856.pdf
Dmrecord
213080
Document Type
Dissertation
Rights
Sun, Zhaoli
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
engineering, materials science