The effect of chromium on the atmospheric corrosion resistance of weathering steels

MISSING IMAGE

Material Information

Title:
The effect of chromium on the atmospheric corrosion resistance of weathering steels
Physical Description:
vi, 175 leaves : ill. ; 29 cm.
Language:
English
Creator:
Wurth, Laura, 1967-
Publication Date:

Subjects

Subjects / Keywords:
Materials Science and Engineering thesis, Ph. D
Dissertations, Academic -- Materials Science and Engineering -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 172-174).
Statement of Responsibility:
by Laura Wurth.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002070309
oclc - 34336941
notis - AKQ8574
System ID:
AA00002057:00001

Full Text












THE EFFECT OF CHROMIUM ON THE ATMOSPHERIC CORROSION
RESISTANCE OF WEATHERING STEELS
/


BY

LAURA WURTH


DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA















ACKNOWLEDGMENTS


I vould like to express my deepest gratitude and appreciation to my research


advisor, Dr. Ellis Verink for his guidance, encouragement, generous support, and


assistance throughout the course of this project.

Dr. Holloway, Dr. DeHoff, Dr. Ambrose, and Dr


My sincere thanks are also extended to

. Winefordner for their participation on


the doctoral committee.

Sincere thanks must also go to Dr. Priya Bendale for her guidance and assistance.

Sincere appreciation is also extended to David Daniels for his assistance in the great

corrosion laboratory clean-up.

I would also like to thank my husband, Jim Plaia, for his encouragement, support,

and infinite patience.

















TABLE OF CONTENTS

ACKNOW LEDGM E NT S .i.i... ............... ...................... 44* ............. ....i*

*AB ST R A C TS ........................................ .. .... ...................... . v


CHAPTERS


* *4 * *4 4~~*~~* *~ *4~4 *4~4~**4** ** ** 4*~**~~* *44~4~~ *4~9*~44~~4 4**** I


Alloy-Based Exposure Studies........................
Characterization Of The Weathering Steel Pat
Corrosion Product Basics................................
Electrochemistry Of Iron And Iron-Chromium
Atmospheric Corrosion Fundamentals.............


.a...
ina ..


44~9*~
ABc


. ..... .. .. .....''. .. '..' ...... 5
. . . . . . . . . 9
... .. ......... ............... 14
)y s............. ......... ......... 17

. .. .. . . .. . . . .2 0


Sam ple Fabrication..............................................
Potentiodynamic Polarization Experiments..........
Long-term Chromium Enrichment Experiments...
Oxygen Uptake Experiments ...............................
X-ray Diffraction And SEM Analysis Of Corrosic


Initial Exposure Experiments...........
One Year Equivalent Exposure.......
Three Year Equivalent Exposures...
Base Metal Electrochemistry...........
Neutron Activation Analysis Studies


1


* 4**** 4~4~4~~** *4~4~4*~ .. .. .. .
* ** *99~~4~4~4 44**~4* 4'~*~~ ~4 44*~* .~41
* *4** ~ 4* *4 ** *~ ..
* 4~*~4~~~ 4*~~ *4* *4~4~t~ '15
fl Products


* *..4....... 4~** *4*4~~~~*4 .4 *4~~~9~4 ~ *.......... .52
* *4 *4* 4~***~ *44~~4 ~ 4~*~4~* *~ . *44*~9 *~ .88
* 4~~~4~~~ . *4 4*~* ~ 4~4*~**~~~ 4*~*~~ *44~~ .99
* *4~~~4~~ . 4*~~**~* *4** *4*4*~~~.. *4*~*~** *~ *~ 126










APPENDICES

A SAMPLE FABRICATION AND VERIFICATION............ 144


Suppliers And Purities Of Pure Materials.........
Average Composition Of Fabricated Samples..
Microstructure Of Samples............................


.... ...... ... ................... . .4 4


... .........14













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE EFFECT OF CHROMIUM ON THE ATMOSPHERIC CORROSION
RESISTANCE OF WEATHERING STEELS

By


Laura Wurth


December, 1995


Chairman:


Dr. Ellis D.


Verink


Major Department:


Materials Science and Engineering


Weathering steels are low-alloy steels which contain less than three percent total of


copper, chromium, nickel, phosphorous, and silicon.


When exposed in atmospheric


environments, these steels display a corrosion resistance two to eight times that of plain-


carbon steel

exposures.


s.


No enhanced corrosion resistance is observed for continuous immersion


The goal of this study was to discover the role of the chromium addition in the


superior corrosion resistance displayed by weathering steels.


During atmospheric corrosion, steels are subjected to alternate periods of wetting


and drying as a result of precipitation and subsequent evaporation.


Under these









thin electrolyte layer present.


A barometric technique was used to study the effect of


small chromium additions on the electrochemistry of iron during wetting and drying.


Iron-


chromium binary alloys with chromium concentrations in the range of zero to five weight


percent were used.


Alloys were studied after various periods ofpre-exposure ranging


from the first four wet/dry cycles to three years equivalent'exposure.


Corrosion product phase form and composition were analyzed with x-ray
C


diffraction.

microscopy.


Corrosion product morphology was analyzed with scanning electron

Chromium distribution in the corrosion product layers of corroded samples


were determined by applying electron microprobe wavelength-dispersive x-ray analysis to


cross-sectioned samples.


Electrochemical characteristics of iron-chromium binary alloys


was determined by potentiostatic polarization in solutions similar to those found on the

surfaces of atmospherically-exposed steels.
















CHAPTER 1
INTRODUCTION


Weathering steels are low-alloy steels


which display enhanced resistance to


atmospheric corrosion when compared with plain carbon steels.


The alloying elements


include copper, chromium, nickel, phosphorous, and silicon, with less than 3w/o total


alloying additions.


Results of both field and laboratory studies have shown that, in an


exposure environment where the steels are periodically wetted and dried, the corrosion


rate is between two and eight times less than that displayed by plain carbon steels.


improved performance is only seen when the steels are exposed in periodic wetting, and,

no improvement is seen for continuous immunersion environments.3


The chromium concentrations in weathering steel are insufficient to cause the

formation of an epitaxial, mixed-metal oxide passive film such as those found on stainless


steels


It is thought that the alloying additions function by promoting the formation of a


dense, adherent layer of iron corrosion products.


Many studies have been done to


characterize this protective corrosion product layer, or, "patina".4"21


From these studies,


the structure, morphology, and chemical properties of the weathering steel patina are fairly
S


well known.


What is not well-understood, however, is the role of individual alloying


C *t~* -. A, n~. A-.. n~.






2


this study is to discover the role of the chromium addition in the enhanced atmospheric

corrosion performance of weathering steels exposed in SO2-contaminated environments.


As atmospheric corrosion is electrochemical in nature, corrosion reactions on the


steel surface will consist of both anodic and cathodic reactions.


reaction is metal dissolution.


In this case, the anodic


Due to the small thickness of electrolyte layers which are


present during atmospheric corrosion, oxygen is readily available, and, therefore, is almost
C


always the predominant cathodic reaction.


Resistance to atmospheric corrosion will be


increased if the rate of either the anodic or the cathodic reactions is decreased.


Since the


chromium levels in weathering steel are too low to promote the formation of a epitaxial,

mixed-metal oxide passive film, it is unlikely that the chromium in weathering steels is


having a significant effect on the metal dissolution reaction.


22-24


It is more likely,


therefore, that chromium is affecting the cathodic reaction.


Since oxygen for the oxygen reduction reaction is supplied by the gas phase (air)

above the electrolyte layer, it must diffuse through the electrolyte until it reaches a surface


which is in electrical contact with the dissolving metal.


Rapid rates of oxygen reduction


are, therefore, dependent on a continuous physical and electrical pathway to the metal


surface.


A tight, nonporous patina of corrosion products will decrease the rate of oxygen


reduction simply by decreasing the rate of oxygen mass-transport.


If the chromium


addition is able to promote the formation of denser, less porous iron corrosion products,






3


would enable higher rates of oxygen reduction than could be supported on the bare metal


surface.


Since many iron corrosion products are conductive, their formation can increase,


rather than decrease the rate of atmospheric corrosion on steels.


If the chromium addition


in the steel can decrease the electrical conductivity of the corrosion product patina, the

rate of oxygen-reduction on the metal surface will decrease, as will the corrosion rate.


Intuitively, limiting the rate of oxygen reduction appeared to be the most likely

method by which the chromium additions could decrease the overall corrosion rate of


weathering steels.


Experiments were conducted which allowed the measurement of the


rate of oxygen reduction on samples as a function of the chromium concentration in the


metal


Chromium concentrations were initially limited to




to the chromium concentrations found in commercial weathering steels.


Since periodic


wetting and drying is essential to the performance of weathering steels, oxygen reduction


experiments were done under conditions of periodic wetting.3


Samples were studied both


in the first few wet/dry cycles and after one, and three years equivalent wetting and drying


exposure using a special dynamic alternate immersion apparatus.


A non-invasive


barometric technique was used to measure the rate of oxygen reduction during different

parts of single wetting and drying cycles on pure iron and various iron-chromium binary

alloys.


Since the density, porosity, and electrical conductivity of the corrosion product






4


Corrosion product morphology and porosity were determined by observation with


scanning electron microscopy (SEM).


Corrosion product phase form was determined


using x-ray diffraction (XRD).


A few previous long-duration field studies have shown that, after extremely long

periods of periodic wetting exposure, many of the weathering steel alloying additions will


enrich at the rust/metal interface. '2'3


It is known from potentiodynamic electrochemical


studies of iron-chromium binary alloys, that, at concentrations as low as 5w/o, chromium


gives promise of facilitating iron passivation.


22-24


If the corrosion reaction were able to


enrich chromium to these levels, it would explain the effect on the metal dissolution


reaction, as well as the oxygen reduction reaction.


iron-chromium binary alloys containing


In order to investigate this possibility,


w/o Cr were tested using the same apparatus


and procedures as were applied to the lower chromium alloys, and the results were


compared.


Of particular importance was the observation that chromium enrichment was


observed, but was not effective until the local chromium concentration at the reaction


interface approached about 5%.


Chromium enrichment was determined by cross-


sectioning samples after testing and estimating chromium levels by x-ray wavelength

dispersive spectroscopy.
















CHAPTER


LITERATURE SURVEY
Discussion of previous weathering steel studies is divided into five different

sections: alloy-based exposure studies, characterization of weathering steel corrosion
4
products, structure and formation of iron oxides and oxyhydroxides, electrochemistry of

iron-chromium alloys, and fundamental atmospheric corrosion studies.


Alloy-Based Exposure Studies
Several investigations have been conducted to find the effect of individual alloying


additions on the corrosion rate.


In all of these studies, the quantity of each alloying


element was systematically varied, and the resulting change in corrosion rate was


measured.


Exposure protocols included both field and laboratory environments.


Field Exposure Studies
In a study by Larrabee and Coburn, 270 different samples were exposed in various


locations for 1


.5 years.


Performance was determined by weight-loss methods.


Table


1 shows the variation of metal loss with chromium concentration for samples exposed in


an SO2 environment.


One sample of each composition was used.


In a study by Taylor, Boden, and Holmes, steel samples with various chromium

concentratinns were ennsdi fnr thrse mnnthc dlirno hnth mmmar an.t unntAr in on cLr






6


In a study by Hudson and Stanner, steel wire and plate samples of various


chromium concentrations were exposed for 1


years.


The metal losses for each


shape sample and exposure duration were averaged and expressed as a corrosion index in


microns per year.


These results are shown in Table 2-3.


In a study by Maxwell, various iron-chromium binary alloys were exposed for 10


years.28S


One sample of each chromium concentration was exposed.


These results are


shown in Table 2-4.


As can be seen in Tables


led to a decrease in corrosion rate.


1 through 2-4, an increase in chromium concentration

The one exception to this trend is the Fe-0.61 w/o Cr


sample in the study by Larrabee and Coburn where the chromium addition resulted in

1.44X increase in corrosion rate.

Accelerated (laboratory) Tests
Although accelerated laboratory tests cannot exactly duplicate the type of

exposure seen in the field, they do provide a more controlled environment for quantifying


the effect of alloying additions.


The results of the two studies described below show how


metal loss varied with chromium concentration in a variety of controlled exposure

environments.










Table


Metal Loss vs Cr Concentration For 15 Years Exposure


Chromium Concentration Metal Loss milss) Percent Improvement Over
(%) Base
<0.1 28.8 base
0.61 41.7 -45
1.20 16.5 43

(Larrabee and Coburn, 1961)25
4



Table 2-2
Metal Loss vs Cr Concentration For Three Month Exposure Durations

Cr composition Summer Metal Percent Winter Metal Percent
(%) Loss (mg/cm2) Improvement Loss (mg/cm2) Improvement
0 15.2 base 11.1 base
1 11.3 26 9.4 15
5 8.4 45 5.6 50
10 2.7 82 3.0 73

(Taylor, Boden, and Holmes, 1971)26




Table 2-3
Annual Metal Loss vs Cr Concentration

Cr Concentration (%) Corrosion Index (pl/yr) Percent Improvement
0 93.5 base
1 64.5 31
1.5 61.2 35











Table


Metal Loss vs Cr Concentration For Ten Years Exposure Duration


(Maxwell)28


Cr Concentration (%) Metal Loss (g/cm2) Percent Improvement
0 0.18 base
3.1 0.04 78
5.7 0.038 79
7.7 0.014 92





9


In a study by Taylor, Boden and Holmes, pure iron and various iron-chromium


binary alloys were subjected to continuous exposure in a fog chamber.

exposed 300 hours in 90% relative humidity (RH), and 0.01lppm SO2.


Samples were


The results of their


study are shown below in Table


In a study by Schwitter and Bohni, steels with chromium contents between


O.Olw/o and 0.75w/o were exposed to cyclic wetting and drying.


Two different


protocols were used: one in which samples were wet 46% of the time, and one where


samples were wet 77% of the time.


In each protocol, the SO2 concentration was 20ppm.


The results of their study are shown below in Table 2-6.


From the results of both field and accelerated exposure studies, it can be seen that

in general, chromium in low concentrations causes some improvement in the atmospheric

corrosion performance of iron and steels.

Characterization Of The Weathering Steel Patina
Heretofore, the great majority of weathering steel studies have been aimed toward


characterization of the protective patina which forms on the steels during exposure.


most of these studies, weathering steels were exposed either in the field, or by accelerated


methods, and the resulting corrosion products were studied by a variety of methods.

review of these studies is broken down by type of characteristics investigated; these

include composition and phase form; morphology, porosity, and surface area; and
1 1 *










Table


Metal Loss vs Chromium Content For Continuous Humid Exposure

Alloy Metal Loss Percent
(mg/cm2) Improvement
pure Fe 3.49 base
Fe-lw/oCr 3.32 5
Fe-5w/oCr 0.85 76


(Taylor,


Boden, and Holmes, 1971)26


Table 2-6
Metal Loss vs Chromium Content In Cyclic Wetting And Drying


(Schwitter and Bohni, 1980)29


Alloy weight loss Percent weight loss Percent
(w/oCr) (46% wet time) Improvement (77% wet time) Improvement
0.01 78 g/m2 base 161 g/m2 base
0.75 61 g/m2 22 122 g/m2 24









Morphology. Porosity. and Surface Area


Protective weathering steel rust


have been found to have a "two-layer" structure.


Observation by optical microscopy has shown the inner layer to be optically inactive, and


the outer layer optically active.4 According

crystallized, and the inner layer amorphous.


,ly, the outer layer was considered to be well-


The inner (amorphous) layer was found to


be more compact and uniform than plain-carbon steel rusts, and was also found to contain


a larger fraction of bound water.6


Separate analysis of inner and outer weathering steel


layers has shown a higher number of pores in outer layer.


The inner (amorphous) layer in


the weathering steel patina is generally thought to be responsible for the enhanced

corrosion resistance displayed by weathering steels.8


Surface area studies using BET analysis have shown that, while the specific surface

area of weathering steel rusts is higher than that of plain carbon steels, the average pore


diameters were smaller.


When the size distribution of these defects was measured as a


function of time, the average size of the pores in the weathering steel rusts decreased with


time, while those of plain carbon steels did not.10


with radioactive


This result is important since studies


S have shown that sulfate ions penetrate rust films only through


macroscopic defects such as cracks and large pores. 10


SEM observation of weathering


steel rusts showed particle sizes between 0.5mm and 4.5mm, layer thicknesses of 0.5mm

to 1.0mm, and four or five layers total."


Elemental analvmic nFhnth filrd Ynnucrp ecmnlc antI arlarattaA tactinn coamhae









extent in and below the corrosion product layer.12


In some studies, enrichment of alloying


elements was only found in the atmospheric corrosion pits.13


The results of these studies indicate that weathering steel rust patinas are denser,

less porous, and, therefore, better able to impede the transfer of oxygen to the metal

surface.

Phase Form and Composition
Weathering steel patinas consist of a xture of different iron oxides and
Weathering steel patinas consist of a mixture of different iron oxides and


hydroxides.


Many studies have been undertaken to determine the phase forms of these


oxides and hydroxides in both the inner and outer layers of the patina.


Misawa, Asami,


Hashimoto, and Shimodaira, using infra-red (IR) spectroscopy, XRD, and SEM found the

inner, amorphous layer of the weathering steel patina to consist of amorphous ferric

oxyhydroxides, whereas Brown and Keiser, using Raman spectroscopy, found it to consist


almost exclusively of 8-FeOOH.14,15


A similar analysis using Mossbauer spectroscopy


showed both inner and outer layers to consist of a mixture of a and y-FeOOH, with the


inner layer having a a-FeOOH fraction almost twice as high as the outer layer.16


Separate


analysis of the inner and outer layers of the weathering steel patina by Meybaum showed


both layers to consist of a mixture of a, y,


and 6-FeOOH, as well as a small quantity of


magnetite.


Mossbauer spectroscopy has been used to study not only the overall phase form,






13


Mossbauer spectroscopy (XMS) were used to study weathering steel rusts from samples


exposed in both rural and marine environments. '


The results showed the rusts from both


sites to consist of a mixture of ferrihydrite, y-FeOOH, and a-FeOOH.


Additionally, they


found that longer wetting durations tended to favor smaller a-FeOOH crystallite sizes.

Additional XMS and CEMS studies by Cook on a 5.4 year marine exposure and a 5.7 year

rural exposure revealed an overall rust composition of ferrihydrite and a-FeOOH plus a


crusty "amorphous oxyhydroxide" that was not possible to characterize.


A study by


Namura using CEMS and XMS on a weathering steel sample exposed 15 years in an

urban environment showed the rust patina to consist of a mixture of a-FeOOH, y-FeOOH,

y-Fe2O3, and magnetite.12


As shown in the results of the studies above, the composition and phase forms of

iron oxides and oxyhydroxides vary widely with both exposure location and exposure


time.


The composition of the poorly-crystallized components of the patina is not clear.


Electrochemical Characteristics


Electrochemical studies of weathering steels have included both potentiostatic


polarizations and galvanostatic reduction. Potentiodynamic polarizations were conducted

mainly to obtain general electrochemical characteristics. Galvanostatic reduction


experiments were often conducted in combination with a spectroscopic technique which

allowed detection of species formed and species reduced during the galvanostatic
c' *n4a






14


Schwitter and Bohni ran anodic polarizations on both plain-carbon and weathering


steels subjected to an "accelerated exposure"


cycles with a wet-time fraction of 46%


consisting of fourteen twelve-hour wet/dry


Both the plain-carbon and weathering steels


showed a large, broad anodic peak between -400mV SCE and 400mV SCE.


The anodic


peak on the weathering steel was forty percent lower than that of the plain carbon steel.

Matsushima and Ueno used potentiodynamic techniques to measure the open-circuit

corrosion current of plain-carbon and weathering steels periodically throughout the course

of a two year exposure.20 The corrosion currents on the weathering steel samples were

consistently between 30% and 80% lower than those on the plain carbon steel samples.


The same investigators examined the cathodic polarization behavior of


month


exposure weathering steel in both aerated and deaerated sodium sulfate solution.


change in limiting current density on changing from aerated to deaerated solution

indicated that oxygen reduction was not the sole reduction reaction occurring on the metal


surface.


From these results it was inferred that trivalent iron oxyhydroxide species were


being reduced on the metal surface.


Suzuki, Masuko, and Hisamatsu performed galvanostatic reduction experiments on

weathering steel and plain carbon steel samples exposed three years in an urban


environment. 21


Reduction was performed in dearated 0.1M sodium sulfate solution.


ray diffraction was used to track the formation of magnetite as a function of current


passed.


The results of the experiment showed that.


while the electrode notentials of









reduction experiment by Okada on plain carbon and


weathering steel samples subjected to


five years of urban exposure showed a larger reduction plateau between -930mV SCE and


-950mV SCE for the weathering steel sample.


Although this is the electrode potential


normally associated with magnetite, XRD showed the weathering steel sample to contain


less magnetite than the plain carbon steel sample.


The reduction plateau was, therefore,


attributed to an "amorphous, spinel-type" iron oxide.

Galvanostatic reduction of the er layer of a weathering steel patina was
Galvanostatic reduction of the inner layer of a weathering steel patina was


performed by Brown and Keiser. The authors separated inner and outer layers by

sanding off the outer layers with abrasive paper. Results of their work showed a reduction


peak between -0.83 and


FeOOH.


-1.00V SCE, corresponding to reduction of a mixture of 8 and y-


IR spectra taken before and after reduction confirmed the initial corrosion


products to be 5 and y-FeOOH, and showed the reduction product to be magnetite.


From the results of the studies described above, it can be seen that exposed

weathering steels exhibit a lower rate of anodic dissolution than a similarly exposed plain


carbon steel


It has also been shown that the corrosion products in the weathering steel


patina can be reduced electrochemically, and that the reduction products are normally

magnetite, or another spinel type oxide.

Corrosion Product Basics
A considerable amount of work has been done in characterizing the structures and






16


metals, their results are useful for predicting the likelihood of various phase

transformations of corrosion products.


The most comprehensive study of formation mechanisms of iron oxides and


oxyhydroxides was done by Misawa, Hashimoto, and Shimodaira. Figure


I shows a


schematic which summarizes the reaction pathways described in their study.


Table


contains a summary of crystallographic information for the iron oxides


and oxyhydroxides which can form on steel exposed in SO2 environments.


From a


corrosion standpoint, some of the most important crystallographic information is the type


of close-packed oxygen lattice that forms the basis for the oxide.


Species both of which


are based on the same type of oxygen close-packed lattices will require less time and


energy to transform.


This becomes important in predicting whether or not a certain


species is likely to form under reducing conditions.


Investigations have been made into the effect of Cu


morphology of colloidally-formed y-FeOOH.


on the structure and


In one, y-FeOOH was synthesized from


an FeSO4 solution with 0 to 50 atomic percent CuSO4 added.


SEM observation of the


resultant y-FeOOH particles showed the particles to be roughly elliptical in shape with the


length decreasing as the Cu


concentration increased.


XRD analysis showed that the


interplanar spacing of the (031) plane increased from 0.247nm to 0.252nm with Cu


doping.


This increase in interplanar spacing was thought to be due to lattice distortion










































































































-- _-


1- I I I I I










Table


Crystallographic Information For Various Iron Corrosion Products


Compound Crystal Dimensions Oxygen layer Structure Notes
System (rnm) sequence
green rust II hexagonal a=0.317 -ABAC- 4 oxygen layers
c=1.09
Fe304 cubic a=0.83963 -ABC- inverse spinel
(magnetite)
y-FeOOH orthorhombic a=0.388 -ABC- based on ccp
(lepidocrocite) b=1.254 oxygen
_________c= 0.307
a-FeOOH orthorhombic a=0.464 -AB- based on hcp
goethitee) b=1.00 oxygen

c=0.303
6-FeOOH hexagonal a=0.2941 -AB- disordered CdI2
c=0.449 structure
y-Fe203 cubic a=0.833 -ABC- spinel


(Fasiska, 1967)3 '





19


important for atmospheric corrosion resistance since a lower conductivity rust layer will


result in a lower oxygen reduction rate, and, therefore, a lower overall corrosion rate.


Electrochemistry Of Iron And Iron-Chromium Allovs


Although, as stated earlier, it is unlikely that the chromium additions in weathering

steels promote the formation of a mixed-metal oxide passive film such as those seen in

stainless steel, it is possible that the chromium additions still have an effect on the anodic


dissolution during atmospheric exposure.


Since any decrease in the rate of either the


anodic or cathodic reactions will result in an improvement in corrosion resistance, it is

important to investigate the known effect of chromium on anodic dissolution of iron.


Diagrams have been constructed which show regions of stability for various iron

and chromium species vs potential and pH.34 These diagrams can be used to predict the

species formed when the metal surface is covered with a thick, reasonably dilute


electrolyte layer.


As the metal dries and the electrolyte layer thins, the rate of the oxygen


reduction reaction increases, as does the


pH and concentration of species in solution.


Unfortunately, due to the extremely small thickness of the electrolyte layer during drying,


measurement of corrosion potential and solution composition is very difficult.


For this


reason, the potential and pH parameters necessary to use these diagrams are not well

known for drying conditions, therefore, the use of these diagrams to predict stable species

under drying conditions is not advisable.









consists of an outer layer resembling y-Fe2O3 with an inner layer of Fe304.


In all cases,


the structure was cubic with the cation concentration decreasing from the inside to the


outer surface..


Thickness varied with exposure time, but, the greatest thickness, measured


after several hours in air, was


2.5nm.


In iron-chromium alloys, this film also contained tri-


valent chromium, which tended to make it more protective.


Several studies have been conducted to determine the effect of low-level chromium


additions on the electrochemical properties of iron in acidic sulfate solutions.


22,23,24.37


Since the SO2-contaminated water layers on weathering steels range from neutral to acidic

sulfate environments, the results of these studies provide useful information on how

chromium weathering steel additions might affect metal dissolution during exposure.


Tables 2-8 through


10, list the results of various anodic polarization studies of


Fe-Cr binary alloys in acidic-sulfate solutions.


In all cases, samples displayed active to


passive transitions with the potential and current required for passivation varying with

chromium concentration.


In all cases, an increase in chromium concentration resulted in a decrease in


the charge necessary to passivate.


most cases, examination of the passive film with x-


ray photoelectron spectroscopy (XPS), and auger electron spectroscopy (AES) showed a

chromium enrichment between three and sixteen times that found in the base metal with

the largest enrichments being seen for alloys with less than one percent chromium in the






21



Table 2-8
Potentiostatic Polarization Results For Fe-Cr Binary Alloys In
Dearated 0.5M sulfate solution, pH=3


Alloy E open circuit E passive I critical to passivate
________(mV vs NHE) (mV vs NHE) (A/cm2)
Fe-5Cr -400 +300 0.10
Fe-7.5Cr -420 +100 0.01
Fe-I 0Cr -400 -100 0.0075
(El-Basiouny and Haruyama, 1976)22

Table 2-9
Potentiodynamic Polarization Of Fe-Cr Binary Alloys In Dearated 0.5M H2SO4,
Sweep Rate 0. Imy/s, Sample Diameter 3mm

Alloy E open circuit E passive I critical to passivate
(mV vs Hg/HgSO4) (mV vs Hg/HgSO4) (A/cm2)
pure Fe -957 -112 2.8
Fe-6Cr -989 -184 2.4
Fe-9Cr -1005 -752 0.80
(Dobbelar, Herman, and DeWit, 1992)23

Table 2-10
Potentiostatic Polarization Of Fe-Cr Binary Alloys In Dearated 1 .ON H2SO4

Alloy E open circuit E passive I critical to passivate
(mV vs SHE) (mV vs SHE) (A/cm2)
pure Fe -300 +625 0.5
Fe-lCr -250 +563 0.5
Fe-6Cr -250 +375 0.25
Fe- 10Cr -250 +219 0.025


(Kirchheim Heine. Fischmeister Hofman Knote and Staiz lO Ro4








current required for passivation dropped by 47.5%


From these results it seems that


chromium enrichment at the metal surface facilitated passivation in the environments


studied


Although both studies showed a correlation between Cr enrichment and ease of


passivation, neither study gave much information on the distribution of Cr relative to the

oxide/metal interface.


Atmospheric Corrosion Fundamentals


In this section a review is made of studies whose goal was a better understanding

of some of the fundamental processes which govern atmospheric corrosion in general.

Although none of these studies dealt specifically with weathering steels, their results can


be directly applied to weathering steels exposed in S02-contamninated environments.


section has been divided into five sections, each dealing with a different fundamental

process in atmospheric corrosion.

Formation of an Electrolyte
Formation of a water layer begins with H20 adsorption onto the oxyhydroxides in


the outer layers of the native oxide.


Adsorptive bonding occurs by bonding of H20


molecules to the OH groups of the oxyhydroxides by formation of H-bridges.39 The water

which attaches in this manner is bound to the native oxide and, therefore, can not serve as


an electrolyte.


If the metal surface is free from salts and other impurities, an extremely


high RH is necessary to start the accumulation of free water for an electrolyte.


hydrosgopic salts are present on the metal surface, free water will begin to accumulate








which coarsen as more and more water is adsorbed onto the surface.


This discontinuous


nature of the water formation is at least partially responsible for the discontinuous sulfate


distribution on exposed metal surfaces.


Once corrosion products have formed, the


quantity of water present on the metal surface will increase due to the capillary action.4143

SO, Interactions
Atmospheric SO2 dissolved in the water layer on a metal surface is hydrolyzed and


oxidized to form sulfate ions b-y a variety of reactions.42-45


varies with RH.


The adsorption rate of SO2


below 60-70%RH, no SO2 adsorption was measurable, but above


70%RH the adsorption increased dramatically.


Presence of corrosion products on the


metal surface will also facilitate SO2 adsorption.42


In the absence of chlorides or other corrosive contaminants, no significant


corrosion will occur without SO2.


46 In addition to increasing the conductivity of the


surface water layer, sulfate ions can participate directly in the corrosion reactions.


Some


investigators have proposed that SO2 acts as a cathodic depolarizer according to the

reaction below:4"


2S02 + 4e


+S04


2- (2-1)


As sulfide compounds have not been observed on atmospherically-exposed steels,


the likelihood of this reaction is considered very small.42


Many more studies have


concerned themselves with the effect of sulfate ion on metal oxidation reactions.41,'42









Fe(II) sulfate salt


The iron in the salt is later converted to an oxyhydroxide, liberating the


sulfate ion for reuse.


The distribution of sulfate on the surface of atmospherically exposed iron and


steels is not homogeneous, but tends to form "sulfate nests"


A cross-section


schematic of a typical sulfate nest is shown in Figure 2-2.

Anodic and Cathodic Reactions in Atmospheric Conditions


The most important distinction between atmospheric electrochemical reactions and
The most important distinction between atmospheric electrochemical reactions and


immersion electrochemical reactions is the thickness of the electrolyte layer.


In the case of


immersion conditions, concentrations of dissolved solids remain fairly low, and the supply


of dissolved gases is limited.


In atmospheric exposure, electrochemical reactions often


occur in extremely thin layers of electrolyte where dissolved gases are much more readily

available, and concentrations of dissolved solids can become very high.

Anodic Reactions
According to Barton, charge transfer is the rate limiting step in metal oxidation. I


this case, the metal oxidation is considered to follow the Butler-Volmer equation.42


Since


conditions are favorable, for corrosion product formation, the dissolution is thought to


occur through a layer of corrosion products.


Rosenfeld conducted an investigation


comparing anodic polarization behavior of iron in 160pm electrolyte films to that in


immersion conditions.41


For current values of 2


iA/cm2, the thin film samples polarized


7SfmV mnre than the immAr'inn camnlc


Fnr anrrent vnine nf


I


O. A/eam2 thp thin film






























-- QO
-"Jwz
-1h-o


I-<


DI <.0
00
00z
_ LL < :




0)
C,'
0
C.
w
I

U
C)
C4
C,'
0
4..






'2
Do

4..
0)
z
4)
4-
z
U,
*<
(4-
0

0
* -
*"~
C.)
C)
tfl
I
In
C,)
0
I-
U









observed even at current values as high as


10 OmA/cm2


Results of AC impedance studies


on thin-film covered iron samples showed the presence of a conductive oxide film.49

Cathodic reactions
Because of the high availability of oxygen in thin films, oxygen reduction is the


major cathodic reaction occurring on atmospherically exposed iron surfaces.


reaction path will vary with solution pH and availability of a catalyst.


The specific


Typical oxygen


reduction reactions are listed:42


02 + 2H20 +


= H202 + 201


(basic)


H202 + 2e


= 20H


(2-3)


02 + 2IT


H202 + 2IW


= H202


+2e


= 2H20


(acidic)


(2-4)


(2-3)


(catalyst)


0~~+2W


+ 2e" = H20


(2-6)


The rate of charge transfer of the oxygen reduction reaction is dependent on the


substrate on which it is occurring.


Potential-current relationships for oxygen reduction on


iron, chromium, and iron oxide are shown in Table


11.42


Because of the low solubility of oxygen in water (2


* 10'"M), oxygen can be


=










Table


Potential-Current Relationships for Oxygen Reduction on Various Substrates


(Barton, 1976)42


Substrate E for E for
_____i=5A/m2 i=10A/m2
Fe 1.00 1.07
Cr 1.15 1.20
Fe203 1.11 1.26









estimation of the limiting oxygen reduction current can be made by applying Fick'


law for semi-infinite linear diffusion:


= (DnFc)/(d)


(2-7)


where: iD = diffusion-limited current


D = diffusion coefficient


= Faraday'


Constant


n = number of electrons passed


c = oxygen concentration


d = diffusion-limited layer thickness


The expression for the diffusion coefficient is given by:


= (RT/N)(6xvr)"1


where: R


T


(2-8)


= gas constant


= temperature (K)


= Avogadro's


number






29

Combination of these two expressions gives an expression for diffusion-limited

oxygen reduction current:


= (RTvFc)/(6dNnr)


(2-9)


This expression is limited to conditions where non convective stirring occurs.

constraint usually limits application to films thinner than 30pm and isothermal

conditions.42


Rosenfeld used cathodic polarization on various thickness electrolyte films to study


the effect of film thickness on the rate of the oxygen reduction reaction4'


Under


conditions of constant film thickness, the rates of oxygen reduction on the thin film


samples were three to four times higher than those of immersion samples.


The same


experiments were performed under vaporizing conditions by varying the RH above the


thin films. The results of cathodic polarization under vaporizing conditions showed two

distinct trends. First, when relative humidities were kept constant, but the initial layer


thickness was decreased, the oxygen reduction currents increased.


Secondly, when the


initial layer thicknesses were kept constant, but relative humidities were lowered, oxygen


reduction currents also increased.


From these results, it was concluded that the oxygen


diffusion length was varying independently of the actual film thickness.


Work by Levich


and Eisher showed that small temperature variations could result in surface tension


rhanoes which nulrd canse "canillarv cnnventinn"


50 Fnr thin electrnlvte nlavers the rate of





30


Although Rosenfeld was able to study fairly thin layers under vaporizing

conditions, he was not able to study the effect of drying to near completion since solution

from the reference electrode could contaminate the sample surface under these conditions.

A technique developed by Stratmann and Streckel allowed the measurement of corrosion


potential to complete drying.


Their technique involved the measurement of sample


surface work function with a non-invasive probe known as a Kelvin probe.


work function could then be related directly to corrosion potential.

not contact the sample, sample contamination was avoided. Using


The surface


Since this probe did


the Kelvin probe as the


reference electrode of a three electrode system, it was possible to conduct cathodic and

anodic polarizations on pure Fe in 1M Na2SO4 with electrolyte layers as thin as 2-3 pm.

Summaries of the corrosion current, corrosion potential, and critical current to passivate


vs film thickness are listed in Tables


12 and


The maximum in corrosion current with decreasing film thickness was due to a

simultaneous increase in the rate of oxygen reduction and decrease in the rate of metal


dissolution.51


As seen from the results of Tables


12 and


film thickness has a


significant effect on electrochemical characteristics.


When metal surfaces are covered with thick electrolyte layers, the slow rate of

oxygen mass-transfer to the reaction interface limits the rate of the overall corrosion


reaction.


As the electrolyte layer thins and becomes self-stirring, however, very high rates










Table


Current To Passivate


vs Film Thickness


(Stratmann and Streckel, 1990)5"


Table


Corrosion Current and Potential vs Film Thickness


(Stratmann and Streckel, 1990)51


film thickness (pm) current to passivate (A/cm2)
285 0.03
10 0.002
2-3 0.0004


film thickness (pm) icorr (pA/cm2) Ecorr (mV SHE)
bulk 48 -311
285 594 -207
10 836 0
2-3 109 293






32


rates of anodic reactions are not high enough to keep up, and the overall corrosion


reaction becomes limited by the metal dissolution reactions.


This switch from cathodic to


anodic control has been demonstrated by Justo and Ferreira who measured the limiting

current for oxygen reduction on thin-film cells.49 The measured value of oxygen reduction


limiting-current exceeded the open circuit corrosion current,


so, it was concluded that the


sample was corroding under. anodic control.

Effect of Wet/dry Cycling
During wet/dry cycling, the thickness of the electrolyte layer on the metal surface


varies from a few millimeters to practically zero.


Since the availability of oxygen varies


with the thickness of the electrolyte layer, the electrochemical reactions which occur on a


rust-covered metal surface will also vary.


Because of the variation of reactions and


reaction rates with electrolyte layer thickness, it is reasonable to expect the rates and types

of reactions on exposed metal surfaces to vary with the progress of single wet/dry cycles.

Mansfeld conducted experiments using a laminated cell consisting of alternating layers of

iron and zinc separated by sheets of polyester film.4 Laminated cells were wet with a


known thickness of electrolyte, then allowed to dry. Galvanic current flowing between the

iron and zinc was measured as a function of drying time. As drying progressed, currents


increased moderately, then increased drastically just before complete drying.


oxygen resulted in little or no galvanic current.


Exclusion of


The results of this experiment demonstrate


the variation of reaction rates during wetting and drying.









Stage One (wetting): During this stage the rust/covered metal surface is covered

with a thick layer of electrolyte. Due to the thickness of the electrolyte layer, mass-


transport of oxygen to the reaction interface near the rust/metal interface is slow, and,


therefore, the corresponding rate of oxygen reduction is also slow.


Because of the low


rate of oxygen reduction, metal dissolution is balanced by reduction of trivalent iron


oxyhydroxide species in the rust layer.


Reaction rates during this stage are fairly low.(See


Figure


Stage Two (drying) During this stage the rust covered metal surface is still


covered by a fairly thick layer of electrolyte, so, the rate of the oxygen reduction reaction


is still very low.


At this point, however, all the reducible species in the rust layer have


been exhausted, so, metal dissolution is now balanced by oxygen reduction.


Reaction


rates in this stage are also fairly low.(See Figure 2-4).


Stage Three (critical wetting) In this stage the electrolyte layer has thinned due to


evaporation.


Now the rate of mass transport of oxygen to the reaction interface is very


rapid, and, consequently, the rate of oxygen reduction is very high.


Now the rate of


oxygen reduction is high enough to support not only a high rate of metal dissolution, but


also to reoxidize the species which were reduced in Stage One. It is during this part of the

wet/dry cycle that the majority of the metal loss occurs.(See Figure 2-5).


This three stage model was confirmed by the results of work by Stratmann,










STAGE


anodic:


- Fe(II) +


cathodic:


8FeOOH + Fe2


+ 2e


3Fe3O4


4H20


Figure 2-3
Stage One Of Wetting And Drying
Because of low rates of oxygen reduction, metal dissolution is balanced by reduction of
trivalent oxyhydroxide species in the rust film.


ONE











anodic:


Fe(II)


cathodic:





O


1/20


20H


Figure,2-4
Stage Two Of Wetting And Drying
Metal dissolution is now balanced by oxygen reduction.


STAGE


TWO










STAGE


anodic:


- Fe(II)


3Fe3O4


cathodic:


+ 3/4 0,


1/20


9/2HO -9FeOOH


p


20H



02
I


Figure 2-5
Stage Three Of Wetting And Drying
The rate of oxygen reduction is high enough to support not only a high rate of metal
dissolution, by also to reoxidize the species which were reduced during Stage One.


THREE









during the course of single wet/dry cycles.


The rate of metal dissolution was measured


using a magnetic technique which allowed detection of quantities of iron metal and iron


spinels


The rate of oxygen reduction was measured using a barometric technique which


relied on the change in oxygen partial pressure from reaction of gaseous molecular


oxygen.


Wet/dry cycling was achieved by flushing the sample to deposit an electrolyte


layer, then heating the sample to dry it.


In the case of pure iron, both oxygen reduction


and metal dissolution currents were between zero and 209iA/cm2 for the first two thirds


of drying period, then jumped to approximately 930pA/cm2 for the last third.


Rates of


both metal dissolution and oxygen uptake dropped to zero as the sample reached complete


drying.


The initial period of low current corresponds to stages one and two, and, the


period of higher current corresponds to stage three.


The same experiments run with an


Fe-0.5Cu alloy showed maximum currents (stage three) of only 500pA/cm2


Continued


cycling of the pure iron samples showed no decrease in maximum currents, whereas the

Fe-0.5Cu samples showed a progressive decrease in maximum (stage three) current with

each cycle.


Additional information on the effect of copper on the cyclic wetting and drying

behavior of iron was provided in a study by Stratmann and Streckel using the Kelvin probe


to monitor corrosion potential during drying.5'


Pure iron samples exhibited a rise in


corrosion potential on drying from -450mV SHE to +200mV SHE.


The copper-bearing


.1 t S a fl n trnrr








corrosion potential would be expected.


Since the copper additions lead to a simultaneous


decrease in both corrosion potential and corrosion current, it was thought that the copper

was acting to inhibit the rate of the oxygen reduction reaction.


From the results of Stratmann and Streckel'


study, it is obvious that an inhibition


of the oxygen reduction reaction should lead to a decrease in the corrosion current


exhibited during the third stage of wetting and drying.


Many studies have shown that the


site of the oxygen reduction reaction on a rust-covered metal surface is not at the

rust/metal interface, but at the oxide/electrolyte interface within electrolyte-filled pores in


the rust layer.56'57"


In order for oxygen reduction to occur on the corrosion product


surfaces, there must be an electrically continuous path to the metal surface. In order to

achieve this path, the corrosion products must be at least partially conductive. At first

glance, the presence of an electrically conductive rust layer seems unlikely. Although

magnetite is somewhat conductive, it is not always present in every rust layer. The other


main constituents of rust, goethite and lepidocrocite, are not considered to be very


conductive.


During reduction in stage one, however, lepidocrocite is partially reduced,


turning it into an n-type semiconductor and greatly increasing its conductivity.

Reduction of Corrosion Products
Since the partial reduction of corrosion products during stage one provides the

conductive path for the high rate of oxygen reduction during stage three, the reduction of


atmospheric corrosion products has been the subject of several studies.


Work by Suzuki,









magnetite, while the a-FeOOH phase remained intact.21


This phase change was ascribed


to a solid state reduction reaction since colorimetric analysis did not detect any Fe


Fe in solution.


Detection of phase changes in this study was done using XRD, which


required that the reduction be halted, and the sample removed from solution.


Other


investigators have used a variety of in-situ techniques to study the phase changes which

occur in atmospheric rust layers during reduction.sS

o
Stratmann, Bohnenkamp, and Engell used an in-situ magnetic technique to monitor

the quantity of magnetite present during potentiostatic reduction of an atmospherically


formed rust film.ss


Magnetite formation was measured as a function of pH, Fe


concentration, and reduction potential.


Their results showed that magnetite formation


began only at potentials below -400mV SHE, and was favored by higher solution pH's


and presence of Fe" in solution.


As in the XRD study by Suzuki, Masuko, and


Hisamatsu, magnetite was found to have formed from y-FeOOH, with no changes in the


a-FeOOH being noted.


Reduction was thought to occur by two different pathways.


First,


in conditions of no solution Fe


and lower pH, a partially reduced intermediate is thought


to form on the surface of the y-FeOOH according to the reaction below:


y-FeOOH + I-


+ e = {Fe*OH*OH}


(2-10)


Since this intermediate is partially reduced, it would be very conductive, and could

serve as a conductive pathway for oxygen reduction to occur in the rust laver during stage









y-FeOOH to magnetite by solid state transformation.


presence of FeC


This is thought to occur in the


according to the reaction below:


2y-FeOOH + Fe"+


= Fe304


+2H


This direct conversion by solid state transformation is thought to be quite likely

due to the similarities of their crystal structures. y-FeOOH having a ccp oxygen lattice


with Fe+ in octahedral sites, and magnetite having a ccp lattice with Fe&


octahedral and tetrahedral sites.


and Fe++ in


Transformation ofy-FeOOH to magnetite would simply


involve movement ofFe+ into the lattice and movement of IH


out of it.


High mobilities


of both species in the y-FeOOH have been previously documented.


Another in-situ study by Stratmann and Hoffinan involved the use of in-situ

Mossbauer spectroscopy to better characterize the reduction intermediate identified in the


study described above.


Based on Mossbauer spectra taken at both 298K and 85K, the


reduction product was thought to be similar in structure to Fe(OH)2, but strained.


From


this information, it was concluded that the reduction intermediate grew on the surface of

the y-FeOOH and the observed strain was due lattice mismatch.


Dunnwald and Otto used in-situ Raman spectroscopy to study the effects of

potentiostatic reduction and reoxidation on the phase forms present in atmospherically


formed rusts.'


The phases present in the rust layer consisted ofy-FeOOH, a-FeOOH,








and appearance of a magnetite peak.


Reduction below -600mV resulted in reduction of


the a-FeOOH to magnetite.


Samples were reoxidized by exposure to air.


The air


exposure did not result in any reoxidation of the magnetite.


The work of this study is


consistent with previous studies involving reduction of'y-FeOOH in atmospheric rust

layers.















CHAPTER 3
EXPERIMENTAL PROCEDURES

Sample Fabrication


Samples were fabricated from pure iron, pure chromium, and pure copper from


various suppliers.

Appendix A. Sar

melting. Larger s


Purities and suppliers for each element are listed i Table A-1 of


mples were fabricated using either arc-melting or vacuum-induction


samples (>50g) were fabricated by vacuum-induction melting, and all


others were fabricated by arc-melting.

Arc-melting procedures
Small pieces of pure materials were weighed out to desired compositions, then


cleaned ultrasonically in ethanol. Clean pi

placed on the water-chilled copper hearth.

avoid oxidation of pure materials. Sample


eces were then transferred to the arc-melter and

Melting was performed in an Ar atmosphere to

;s to be used in electrochemical experiments


were soldered to brass, flat-head screws and machined to a uniform 1cm diameter.

Samples were then mounted in epoxy and ground so that only the front face was exposed.

Samples to be used in long-term exposure, (e.g.chromium-enrichment)studies were

prepared in a similar manner except for the machining to uniform diameter.

Induction-melting procedures









susceptor.


Because of the high melting point of chromium, it was necessary to add


chromium in the form of Fe-20Cr and Fe-30Cr master alloy buttons.


buttons were fabricated by arc-melting.


solidified under Ar.


The master alloy


Samples were melted under vacuum, then


Induction-melted samples to be used in oxygen-uptake experiments


were machined into hollow cylinders.


Schematics of all samples are given in Figure 3-1.


All alloyed samples were checked with WDS on the electron microprobe to ensure
a.


composition and homogeneity


Results of the EPMA checks can be found in Tables A-2


through A-6 of Appendix A.


According to the ternary phase diagram for iron, chromium, and copper, the

microstructure of all samples used should consist of single-phase a-ferrite with chromium


and copper in solid solution.


microscopy.


Microstructures of all samples were confirmed by optical


Representative microstructures of arc-melted and vacuum-induction melted


samples can be found Figures A-1 through A-4 in Appendix 1.


Potentiodynamic Polarization Experiments
In order to better understand how chromium affects metal dissolution, anodic

potentiodynamic scans were performed on pure iron and a variety of iron-chromium binary


alloys.


Polarizations were performed in deareated 0.05M Na2S04.


Deaeration was


performed by purging with nitrogen.


Sample preparation required grinding with SiC paper


to a 600 grit finish, rinsing with deionized water, rinsing with ethanol, then drying in warm










44

























a

C
C
IC
C
0
S.
-
*
o a..
4: 4)
a
C
0'e {1
4- aM
SIC
2w



0.


VI

a.)
0
-n
I

0.0
I-cu



a- e- C,)
S U
o -
U) 4-
Cu
2
4)
a
U
E S
C
-c En
U
o
4- a
C
4) -
a-
w
C
- a
-
o
E x
4/) C

I
a,








used was 0.1 mV/s, and, scans went to 1400mV positive of open-circuit. The platinum

counter electrode was separated from the working electrode by a glass frit. The saturated


calomel reference electrode was used with a luggin capillary positioned just below the


working electrode surface


Polarization was performed using a Princeton Applied


Reaserch model 273 potentiostat interfaced to a 386 computer.

Long-term Chromium Enrichment Experiments
In this series of experiments, various iron-chromium binary alloys were subjected

to wet/dry cycling by alternate immersion and then analyzed to determine the distribution


of chromium in the corrosion products and at the rust/metal interface.


Chromium


distribution in the outer parts of the rust layer was analyzed with neutron activation


analysis


Chromium distribution of the inner layers and the rust/metal interface was


determined by cross-sectioning and analysis with electron microprobe, wavelength

dispersive spectroscopy.


Alternate immersion exposure was performed using a testing apparatus, a


schematic of which is shown in Figure 3


positioned evenly around the rim of the whe


The tester could accommodate six samples

el The tester could be set to rotate


intermittently,


giving each of the six samples a set amount of immersion time.


Since only


the sample min the lowest position could be immersed, the ratio of immersed time to non-


immersed time was always 1:6.


The ratio of wet time to dry time, however, varied with


the rate at which samlDies dried after having been immersed.


In general longer cycle times










































/

a)





47


Neutron Activation Analysis (NAA) samples consisted of large, arc-melted


buttonsmounted in epoxy and ground to expose only one side.


The exposed area of each


samplewas approximately 8cm2


, and, six such samples were exposed at one time.


At the


end of an exposure period, exposed NAA samples were removed from the wheel and


corrosion products were removed for analysis

layers using squares of clear, strapping tape.


Corrosion products were removed in


Since the tape did not contain any elements


which activated upon irradiation, it did not interfere with the chromium analysis.


Squares


of tape were laid over each sample, rubbed with a teflon scraper, and peeled off to remove


a layer of corrosion products.


This procedure was repeated until no more corrosion


products could be removed (normally four times). In this manner, it was possible to

obtain a rough depth profile of chromium concentration. In order to obtain an adequate


signal-to-noise ratio with NAA, it was necessary to have at least 250mg of each sample.


In order to reach this weight,


it was necessary to combine similar layers of all six exposed


samples.


Microprobe samples consisted of small, arc-melted samples mounted in epoxy and


ground to expose only one side.


0.79cm2


The exposed area of each sample was approximately


. At the end of an exposure period, samples were removed from the wheel, and,


the entire sample vacuum-mounted in epoxy. Samples were then cross-sectioned using a

diamond watering blade and polished to a 0.3um finish. A schematic of a cross-sectioned


sample is shown in Figure 3-3.


In order to avoid charging of the corrosion product laver









48















r

I

a
0~
0)

a,
S







cn




0


a-
en0
en4)




-o
03
C
0
m
4-
C)
4)
LO
S
cn
cn
0
I-.
U








=1
L
0~
C
0
a
&





49


Oxygen Uptake Experiments
In order to better understand the effect of chromium on cathodic reactions during

wet/dry cycling, experiments were designed to monitor the rate of oxygen reduction


during wetting and drying


These experiments essentially involved placing the sample to


be studied in an airtight chamber with an oxygen atmosphere, allowing it to wet and dry,


and measuring the resultant pressure drop.


The change in oxygen partial pressure can be


related to an oxygen reduction current through the use of Faraday'


Law.


A schematic of


the cell used for the oxygen uptake experiments is shown in Figure 3-4.


The samples were


hollow and screwed onto a plastic coupler which was attached to the sample holder.

Chilled fluid could be circulated through the hollow sample via the sample holder,


allowing the sample surface temperature to be varied during the experiment.


The cell


contained a temperature control bath which was filled with a saturated solution of CaCl2 .

When the CaCI2 solution was at room temperature in the sealed cell, it established a


relative humidity of 35%.


bath.


Higher relative humidities could be established by heating the


Sample contamination by CaCl2 was avoided by placing a splash guard between the


bath and the sample.


The cell was equipped with a humidity/temperature sensor which


was used to monitor relative humidity throughout the course of each experiment.


Itwas


necessary to monitor relative humidity because, during the course of a wet/dry cycle,

water vapor partial pressure changes were significant and constituted a significant fraction


of the total measured pressure change.


Measured variations in water vapor partial


* .1I I a-. I




































.3 1
-B
I









subtracted from the total pressure before reduction currents were calculated.


Total


pressure was monitored through use of an absolute pressure transducer.


At the start of each experiment, dry samples were attached to the sample holder


and inserted into the cell.


To begin wetting, the sample surface temperature was lowered


by circulating chilled anti-freeze fluid through the hollow sample interior.


The relative


humidity of the chamber was simultaneously raised by heating the solution of saturated
4'


CaCI2.


To ensure even wetting over the entire sample surface, two valves were opened,


and humid oxygen was blown through the cell at a rate of 1


Hour (SCFH).


Standard Cubic Feet Per


After one hour of wetting with flowing, humid oxygen, the outlet valve


was closed and cell pressure was adjusted to


17 ton-rr.


At this point, monitoring of cell


pressure for oxygen reduction rate determination began.


For the first hour of monitoring,


the sample was forced to remain wet by continuing the circulation of chilled fluid through


the sample interior and heating in the CaCI2 bath.


One hour into the monitored portion of


the experiment, the sample chilling and bath heating were shut off and the sample was

allowed to dry.


The quantity of reduced oxygen at any point in the experiment is given by the

formula below:


An(O2)


= (V/RT)[(Pm aed tO-PN20o to)


- (Pmcamued t-PmoH20t)]


(3-1)









= gas constant


= gas phase temperature at time t


Pmeaed to = measured pressure at start of experiment


PH20 to


PHas0rd t


PHmot=


= initial water vapor pressure


= measured pressure at time t


water vapor pressure at time t


The rate of oxygen reduction is given by:


i(02)


= 4F{d[An(O2)/A]/dt}


(3-2)


where: i(O2)=oxygen reduction current


F=Faraday' s


constant


A=sample area


Since it was not possible to introduce corrosive gases directly into the oxygen

uptake measurement cell, all samples received a pre-corrosion exposure before


introduction into the measurement cell.


Long-term exposure samples received the


equivalent of either one or three years of cyclic wetting exposure (assuming one cycle per


eJ a. A n .UL ~L al4a..mnn4n *s4..ina...ne..: nfl Sasans la.tt LZ qi -- - 1 1 L L -





53


X-ray Diffraction And SEM Analysis Of Corrosion Products


After completion of oxygen uptake experiments, corrosion products were removed

from the sample surface with a pointed scalpel and examined using x-ray diffraction, and,


where appropriate, with SEM.


SEM examination was only performed for the longer-


duration exposure samples since the corrosion products from these samples tended to be

thicker and could be removed intact, while those of the initial exposure samples could not.


All SEM samples were coated with Au-Pd to avoid charging during examination.


examination was performed with a JEOL 6400.


SEM


XRD samples were ground to a


particle size and attached to glass-slides with collodian amyl acetate.


Diffraction scans


were run at a rate of 0.01degree per second, with a collection interval of 0.05 degrees.

ray diffraction was performed on a Phillips 3520 powder diffractometer, using Cu Ka


radiation.


Quantitative analysis was performed using matrix-flushing, a variation of the


method of internal standards.


The confidence interval was +/-5%.















CHAPTER 4
EXPERIMENTAL RESULTS


Results are divided into three sections based on exposure time.


The first section


deals with samples exposed for only the first four wet/dry cycles, the second with samples


exposed for 365 wet/dry cycles, or one year'


equivalent exposure; and the third with


samples exposed for 1095 wet/dry cycles, or three year'


equivalent exposure.


Results of


potentiodynamic and NAA experiments are listed separately.

Initial Exposure Experiments
Initial exposure samples were first subjected to a three hour pre-corrosion


exposure in a humid sulfur dioxide/oxygen environment.


After the precorrosion


treatment, samples were transferred to the oxygen uptake measurement chamber where


they were subjected to four measured wet/dry cycles.

pure Fe, Fe-0.5Cr, Fe-0.5Cu, and Fe-0.5Cu-0.5Cr. 1


The first samples to be tested were


[hese sample compositions were used


first since they are most representative of chromium and copper concentrations found in


many commercial weathering steels.


Although the purpose of the study is to determine


the role of chromium in weathering steel performance, the copper-bearing alloys have

been much more thoroughly studied and can, therefore provide a good reference for

comparison.


























a)
C2

W -
XC
0>


C
Cu
a

IC
'Ito)
00
4)).,
3-


o 04


000


4.~
nfl-






















C)-

C>
4J~


C)
C)
>1S~
xc
0>


tflOLflOU'~O&flOtOOW)Ou)or)ouotoo


a
66666666666666 6dddcc


Cu
-



4;
Cu



I 4)
too


O -

I I


C)

C
a
s-tn


















0
a
0~
*0
0>
Q40)


Cu


xc
0>


0 tO 0 '0 0 In 0
o~cJ~Nr-tDcD
6666666


4-
0.


@2
too
@2>l

-


10 10 ^- FO
00000


V
C
0!'





















































I
4)



4)
4)
I
4)
2


O~O~OLflOLnoLnotnoLnoLnou~otnotno


a
e00666666666666666666oo


II






59


to the three-stage model of wetting and drying, the oxygen-uptake vs time curves


shouldcontain four distinct regions corresponding to stage I,


stage II,


stage III, and


complete drying


According to this model,


the rate of oxygen uptake should be low


during stage I,


slightly higher during stage II, significantly higher during stage II, and


approximately zero during complete drying. Regions corresponding to stages I, II, III, and


complete drying are shown on each curve.


As can be seen in Figures 4-1 through 4-4, the


shape of the oxygen uptake vs time curves conforms fairly well to the shape predicted by


the three-stage model.


There are two areas in which the shape of the time-dependent


oxygen uptake curves differs from that predicted by the three-stage model.


First, during


the first 3000-4000s of each experiment, the rate of oxygen uptake becomes negative,


indicating an off-gassing process.


These negative values are attributed to the inaccuracies


of temperature measurement during this part of the experiment.


Due to simultaneous


heating of the internal bath and chilling of the sample, temperature gradients were set up in

the cell which could not be completely accounted for by the in-situ temperature sensors.

After the end of the first hour of stagnant wetting, both the sample and bath temperatures

returned quickly to room temperature, and measured temperatures corresponded to actual


gas-phase temperatures. The second area of discrepancy is in the relative rates of oxygen

uptake during stage I and stage II. According to the three-stage model, the thickness of


the electrolyte layer present during stage II is still fairly thick, and,


therefore, the rate of


oxygen uptake during stage II should not be significantly higher than that of stage I. A








difference in the morphology of the electrolyte layers in each case.


The three-stage model


assumes a single, continuous layer of electrolyte covering the entire sample surface.


Since


the samples in this study were wet by condensation, the electrolyte was present in the form


of droplets with a distribution of sizes. During the wetting phase, the mean droplet size

increased, but a distribution of sizes was still present. As the sample was allowed to dry,


the smaller droplets dried more quickly, and, thus reached the critical thickness for self-


stirring, or, stage III, much sooner than the larger droplets.


If the samples had been


allowed to continue to cycle and a continuous layer of corrosion products had been

allowed to form, the capillary properties of the corrosion products would control the


morphology of the condensing water,


giving a curve shape much more consistent with the


predictions of the three-stage model, as will be seen for longer exposure durations.


Certain aspects of the time dependent oxygen uptake curves can be related directly

to properties of the corrosion product layer and overall atmospheric corrosion resistance.

These aspects include: total drying time, maximum current, and total per-cycle oxygen


uptake.


Total drying time is defined as the time from the start of the experiment to the


point at which complete drying is achieved. Total drying time gives a measure of the

ability of the corrosion product layer to retain moisture. The moisture retention properties


of a corrosion product layer will depend on many things such as thickness, total


continuous pore volume, and average pore diameter.

can only give a measure of the combined effects. Mu


Unfortunately, the total drying time


iximum current can be calculated









limited almost exclusively by the conductivity of the corrosion products themselves.


maximum current can, therefore, be used as a measure of corrosion product conductivity.

Total per-cycle oxygen uptake is defined as the total number of moles of oxygen reduced


per wet/dry cycle.


Since the total number of moles of oxygen reduced and metal dissolved


per cycle are equal, the total per-cycle oxygen uptake gives a direct measure of the

corrosion resistance at any given point in the exposure lifetime.

A plot of total drying time vs number of cycles is shown for all compositions
A plot of total drying time vs number of cycles is shown for all compositions in


Figure 4-5.


As can be seen in Figure 4-5, there is no real trend in total drying time with


either composition or number of cycles.


This lack of trend is more than likely due to the


fact that the corrosion products which form in the first four cycles are both thin and sparse

and are, therefore, unable to retain much moisture.


Plots of maximum current vs number of cycles are shown for all compositions in


Figures 4-6a and 4-6b. As can be seen in these figures, the maximum current does tend to

decrease with increased cycling. This is consistent with the conclusion that, while the


corrosion products are too thin and sparse to contain a significant quantity of moisture,


they are thick enough to impede the diffusion of oxygen to the reaction interface.


Figure


4-7 shows the variation in maximum current for the fourth wet/dry cycle with alloy


composition.


As can be seen in Figure 4-7, the addition of 0.5w/o chromium alone


decreases the maximum current only slightly, whereas the addition of 0.5 w/o copper,























*UL)
U:4 tO

'-II
~0v


I...
CD-c
E
tn
S
4-h

gJ


oo o o o
o 0 0 0 0
o a a oi o
o0 0 0 0 0
00 0 CM 0 10
00000














































m
a,
no
Cs,
I.-.
0
L..
a)
.0


z


\9&L
fl.-0

cc


z
C-


U) C





















































































o U')
CN~~ C LI)
C
S


@2
a)
C.)

C)


In')
I-
tOo
I I
It


a)
.0









65









a-
a)
C-
C-
0
C.)
0

In
0
0

0
-
*
so
-o
0
42
1) -~
cv
t Ca
-
o
4) I-.
-4 -c4)
o A

CV
-.. -
S -C
o
C) C) 0mb
U, (V 'd~ S
-c
inc
o ~
0.
C
V Li
5., tcnz

C~) U)
r..
C
0~
S
S
-I
x

*~ca
cU 43
c~u~
U'-
4)
0 0 0 0 0 0.0

S C0
S S
o 0 0 0 0 0 0 0

02
14
'.9 0
CS)
U
o
E u

x.ic
-





66


effecting a significant decrease in the conductivity of the corrosion products which form in


the first four cycles.


This decrease in conductivity could be due to either a densification of


the products, or a decrease in the electrical conductivity of the products.


A densification


of the corrosion products would result in a decrease in the rate of mass transport of


oxygen to the reaction interface.


A decrease in product electrical conductivity would


prevent oxygen reduction from occurring on the corrosion product surfaces, thereby

reducing the total cathode surface area and forcing the oxygen to diffuse to the rust/metal

interface to react.


Plots of total per-cycle oxygen uptake are shown for all compositions in Figures 4-


8a and 4-8b.


As can be seen from these figures, the total per cycle uptake decreases with


increasing numbers of cycles.


This decrease is attributed to an increased resistance to


atmospheric corrosion, most likely due to a thickening of the corrosion product layer.

Figure 4-9 shows the variation in total per-cycle oxygen uptake with alloy composition.


As can be seen in Figure 4-9, addition of 0


total per-cycle oxygen uptake, while addition of 0.


change.


w/o Cu results in a significant decrease in


w/o Cr alone results in very little


The reasons for this decrease are considered to be similar to those proposed for


the decrease in maximum current.


Photographs of the side of each sample were taken at the end of the fourth cycles


and are shown in Figures 4-10 through 4-13


As can be seen in Figures 4-10 through 4-13,


- a S a a -. -- -






































t



C,
a)

C-)

0
L.
a?
.0
(NE

z


IOOU)O~)0~)OU~0U)OUOW)O
C CD WI in nn CNJ r'4


'0 0


<-0


000


o00oooo


0 0


0 0


*
































I I I I I I I I p


I I I I I I I I I


.1)
0'~
z


CNJC4
oooco666


u~o
























































6 S N
C









































Figure 4-10


Side View Photo Of Pure Fe Initial Exposure Sample










































Figure 4-11


Side View Photo Of Fe-0.5Cr Initial Exposure Sample









































Figure 4-12


Side View Photo Of Fe-0.5Cu Initial Exposure Sample










































Figure 4-13


Side View Photo Of Fe-0.5Cr-0.5Cu Initial Expsoure Sample





74


At the end of the fourth wet/dry cycle the corrosion products were removed from


each sample and analyzed with XRD.


The results of the XRD analysis are shown in Table


Full XRD spectra of these and other XRD scans can be found in Appendix B.


Table 4-1 reveals that the corrosion products consisted of a binary mixture of


lepidocrocite and magnetite.


When the effect of alloying elements are considered, it can


be seen that copper additions favor the formation of lepidocrocite over magnetite, while
chromium additions have ittle or no eect on the relative actions oflepidocrocite and
chromium additions have little or no effect on the relative fractions of lepidocrocite and


magnetite in the corrosion products.


Since magnetite is more conductive than


lepidocrocite, a decrease in magnetite fraction should lead to a decrease in corrosion

product layer conductivity, and, thereby, to an increase in atmospheric corrosion


resistance.


These results are consistent with the results of the oxygen uptake experiments


shown in Figures 4-7 through 4-9.


Since the results of both oxygen uptake and XRD studies have shown that an

addition of 0.5%Cr alone has very little effect on atmospheric corrosion resistance of iron,

similar experiments were run on iron-chromium binary alloys with higher chromium


concentration

and Fe-5Cr.


IS.


The higher chromium concentration samples included Fe-lCr, Fe-2Cr,


Time-dependent oxygen uptake curves for these higher chromium


concentration samples are shown in Figure 4-14, 4-1


and 4-16, respectively.


Just like


the lower-chromium samples, shown min Figures 4-1 and 4-2 these curves display the





75



Table 4-1
Species And Phase Forms Present In Corrosion Products


Alloy %lepidocrocite goethitee %magnetite
pure Fe 66 not detected 34
Fe-0.5%Cr 65 not detected 35
Fe-0.5%Cu 88 not detected 12
Fe-0.5%Cr-0.5%Cu 82 not detected 18









76













C-
C
'S
C>







0>


I
C
0
C
L~.
*
-
*
C,
0
o ow
C
0 w
** -I 1~~)
L.
t
'*0
0
0
o I-
to
LN
U4
4
'0 CU
1 ____
o~ 0.
4 -J
C.)
** *w*g) -~
CN92
4- ~CO
C) C
-4
OW L.X
r-~r
In
C)
S.-


q) *. 0
I-N
'V I '-I
t C C
S 0.
-*-* **** *** 4)




*- *- *
In p
'4 S..
02 I-.
QI
Ott) Ott) Ott) GIn Ott) Oto OW) C toO 'no


4 4 o 4)
OO)O~~r-NWflttflrflCNCNr-O0O~

I I

C)

C

0
a






















































S
4


t
S

I
S
S





a
S
I
S
p
*


S


0
C
C
LO
m

0~

CO


m


OLOOLOOLOOLOOLOOu)OIflOW)OLOOU)OLOO

S S
-dodddddd6ddcSd6ddoooci6d
I I


4)
0>

0
ed

i"o

















4)
0.
0

4.J~

'C)
a)'


0>


K







h



U

to

U)


I I I I I I I I I I I I

t








S


II I I


N
U)
S
'U
.6.)


C)
C
4d~


-ooooooocooddooododaded
0II


C)
M
C
*J -
or'


I






79

Figure 4-17 shows the variation of drying time with number of cycles for the


higher Cr alloys.


Unlike the lower-chromium alloys, the higher-chromium samples all


show a decrease in drying time with continued cycling.


Figure 4-18 shows the variation of


drying time on the fourth wet/dry cycle with chromium concentration.


Figure 4-18


confirms that there is no significant trend in drying time with chromium concentration.


Figure 4-19 shows the variation of maximum current with number of cycles for the
4
higher Cr alloys. As shown in Figure 4-19, maximum currents do not vary significantly

with continued cycling. Figure 4-20 shows the variation in maximum current on the


fourth wet/dry cycle with Cr concentration.


As can be seen in Figure 4-20, chromium in


concentrations of 1 w/o or more produces a significant decrease in the maximum current


sustained during each wet/dry cycle.


A comparison to Figure 4-7 shows that the effect of


1-5w/o Cr additions are similar to those produced by the addition of0.5w/o Cu.


Figure 4-21 shows the variation of total per-cycle oxygen uptake with number of


cycles for the higher-chromium alloys.


As can be seen in Figure 4-21


the total per-cycle


uptake decreases with increased cycling, indicating that the corrosion products are


becoming more protective with continued cycling. Figure 4-22 shows the variation of

total per-cycle oxygen uptake with chromium concentration. Just as in the case of the


maximum current (see Figure 4-20), the total per-cycle oxygen uptake shows a significant

decrease with the addition of one to five w/o Cr.























L)QL)
I I I
0CC)
rz4z4rz~

0.t~


0/



> 00 0




> 0 0 0n 0 0
0 CO 0 0 0 C
tOOOOO0 1 C








81

























C
o 0
-
tO
CU
I-
in -C
-~ t

0
s-I t0

C
2
C) cU~
C ~0
*" 4-.
o
C) 4)0
0 C)
Lfl*J
(Nw
-~

-
a-
0)
'4 '4.
*04 ~ C'
H 2
an
C
* ~"4
in
ci

o cC
*
o
In
o 0 0 0 0 0 016
o 0 0 0 0 0 0 o
o 0 0 0 0 0 0 c
o tO 0 tO 0 tO 0 U,
CN (N -
4)
I-
0)
v
F-
~1%
I-Il'.
CCC
Co
a-C)

0'-








82













La S..
ouu
-cv to
I I I
vww





LI)
C,,
4)
0) -
-~ U
Cdfri U
(6~
0
I-.
'-4 4.)
o .0

V
.0
S n.~ C'-

Z Ls>~
C)
en
C.) O\E~

o 4Z
C) ( 14
4)
.0
z .-G)


z
Qo
S
C
-
S S
S .- U,
x
~
Lu.
z
U
o Ut 0 Ut 0 0 22
c'~i (N 0 0
o 0 0 0 0 0 0


4)
54
5.4
C.)


C)








83






'U-
-o



-C
C)
cv
C)
Of)
C
-
I-

*0
*0
0)
LI) C
In I.

C
o 10
4)
Lw
LV I-
Ut
C.)
Ce.. B
0) .
C) *0
N.
O~C
o S
U 0
00
tfl~ I.
C)
CC
U) L.
0 en
CN~ QO)
0) Cd
CC,
Eo
0) CNO .- C.)
00
eQ LQu>~

E P
-

E I
bcu
Zen
QO)
9 C)
2'
ES
o
(N 0 CD Ut (N CI
r 0 0 0 0 C)
S 5 2
o o o o o 6 6 o
I-
4J 0
0
V
'.4
5-
-
U
0
EN U,
C
0
-
-
C
C.)C)























00
cQto
I I
WV


ci,
w
C.)
U
a'-
0
a)
-o
S

z


a)
C
4-)


C
0)
aD
x
0


C


ow)
001
-o


ddd6666555~,4A


C4)

'-4)






~CTI
4-. I-

-o
C
4)0)

Xc.


u~0
I- -






















































Lc~
*
4-)
00
*1~
CN4~
C
'no)
*c)

0


1.-i Li L3


00


00


Cooco









lower-chromium samples.


The results of the XRD of the higher-Cr samples is show in


Table 4-2.


Refering to Table 4-2, the corrosion products on the higher-chromium-containing


alloys consist of a binary mixture of lepodocrocite and magnetite.


When the relative


fractions oflepidocrocite and magnetite in the samples with chromium concentrations of

one to five w/o are compared to those of the pure Fe and Fe-0.5Cr samples (see Table 4-

1), it is evident that the magnetite fraction is higher in the higher-chromium samples.

Since a higher magnetite fraction results in a more conductive, and therefore less

protective, corrosion product layer, it is unlikely that the superior corrosion resistance

indicated by the results of the oxygen uptake experiments is due to a more protective

corrosion product layer.


Photographs of the sides of each of the three higher-chromium samples were taken


at the end of the fourth wet/dry cycle and are shown in Figures 4-23,


4-24


and 4-25.


Clearly, the coverage by corrosion products is much less for the higher chromium-


containing alloys than for lower-Cr samples.


sample shown in Figure 4-25.


This effect is most obvious for the Fe-5Cr


This lower coverage is attributed to the more protective


oxide which forms on the higher chromium alloys.


Since the naive oxide layer on the


higher-chromium samples is thicker and more protective, there are fewer sites available for

sulfate nest initiation, and hence, fewer sulfate nests after comparable exposure times.






87



Table 4-2
Species And Phase Forms Present In Corrosion Products


Alloy %lepidocrocite goethitee %magnetite
Fe-1Cr 37 1 62
Fe-2Cr 52 1 47
Fe-5Cr 38 not present 62





88




































Figure 4-23 Side View Photo Of Fe-lCr Initial Exposure Sample










































Figure 4-24


Side View Photo Of Fe-2Cr Initial Exposure Sample





90



































Figure 4-25 Side View Photo Of Fe-5Cr Initial Exposure Sample






91

samples correlate with the smaller number of sulfate nests, which, in turn, reflect the more

protective native oxide on these alloys.


From the results of the initial exposure experiments, it is clear that, for short


duration exposures, copper and chromium behave quite differently.


Copper additions


appear to decrease the rate of atmospheric corrosion by favoring the formation of less-

conductive reaction product species in the rust film, while chromium merely decreases the


number of anodic sites on the metal surface.


The presence of the protective, pre-existing


film on the chromium-containing alloys is believed to account for the superior atmospheric

corrosion resistance of these alloys in long duration tests despite almost a complete

coverage by corrosion products.

One Year Equivalent Exposure
Samples were exposed by alternate immersion to give 365 wet/dry cycles, or,


approximately one year's equivalent exposure.


After this exposure, the samples were


subjected to the same oxygen-uptake tests as were the initial exposure samples.


compositions included pure Fe, Fe-0.5Cr, Fe-lCr, Fe-2Cr, and Fe-5Cr.


Alloy


After the "one


year", alternate-immersion pre-exposure, all samples were covered with a continuous layer


of corrosion products.


Figures 4-26, 4-27,4-28,4-29,and 4-30, respectively show the time dependent


oxygen uptake curves for pure Fe, Fe-0.5Cr, F


e-1Cr, Fe-2Cr, and Fe-5Cr, respectively.

















a)
C.
0


~4V





xc


N(DU~tflC~rOQ)~NcD~
r 1~ o a a a


6


-o 0
00


a)
C
.3 r-.
0
*
















































0)
x
CO







0
* -







5 -I
0





(0





o -0t N


*
- --- 0


W~NCDW)tflc'J


00


* 0
00


a)














































































N(DU~


O)~NCDU~tflc'J


- I- 6


0000