MSU

LIBRARIES
-—

‘—

 

 

RETURNING MATERIALS:
P1ace in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped be10w.

 

 

 

 

By
Patricia L. Lang

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF SCIENCE

Department of Metallurgy, Mechanics and Materials Science

1988

ABSTRACT

SEVERAL PUBLISHED ANALYTICAL FUNCTIONS FOR MODELLING
ATMOSPHERIC CORROSION OF ZINC AND GALVANIZED STEEL ARE
REVIEWED. SOME ARE MODIFIED TO A STANDARD FORMAT IN ORDER TO
COMPARE THEM WITH ONE ANOTHER. A DAMAGE FUNCTION HAS BEEN
PROPOSED RECENTLY BY HAYNIE AND SPENCE (EPA) WHICH ACCOUNTS
FOR ACID PRECIPITATION AND OTHER ENVIRONMENTAL FACTORS. THIS
NEW FUNCTION AND SEVERAL OLDER ONES ARE EVALUATED CRITICALLY
WITH RESPECT TO THEORETICAL AND EXPERIMENTAL CONSIDERATIONS.
THIS PROVIDES THE FOUNDATION FOR COMPARING OLDER ANALYTICAL
FUNCTIONS WITH MORE RECENT EXPERIMENTAL DATA. RIGOROUSLY
MEASURED RESULTS FROM THE BUREAU OF MINES ARE COMPARED WITH
EPA MODEL . THE PROBLEM OF ACCURATELY MODELLING ZINC CORROSION
IN TERMS OF ENVIRONMENTAL FACTORS REMAINS FORMIDABLE.

ACKNOWLEDGEMENTS

I would like to thank my advisor, Professor R. Summit, for his
friendship and guidance that made this thesis possible. I would like to
thank my husband, Frederick W. Lang, Jr., for supporting me with love and
encouragement throughout the years of schooling, leading up to, and
including this Master's thesis.

iii

1.

1.1

2.

2.1
2.1.1
2.1.2
2.1.2.1
2.1.2.2
2.1.2.3
2.1.3
2.1.3.1
2.1.3.2
2.1.3.3
2.1.3.4
2.2
2.2.1
2.2.2
3.

3.1
3.1.1
3.1.2
3.1.3
3.2
3.2.1
3.2.1.1

3.2.2
3.2.2.1
3.2.3

TABLE OF CONTENTS

LIST OF FIGURES
LIST OF TABLES
INTRODUCTION
THE EPA/MSU COOPERATIVE AGREEMENT PROGRAM
BACKGROUND
FACTORS AFFECTING CORROSION
ACID RAIN
ENVIRONMENTAL FACTORS
RELATIVE HUMIDITY
WIND VELOCITY/DEPOSITION VELOCITY
TEMPERATURE
POLLUTANTS
SULFUR DIOXIDE
CHLORIDES
PARTICULATES
NITROGEN OXIDES
REVIEW OF DAMAGE FUNCTIONS
CONVERSIONS
EQUATIONS
EVALUATION OF THE HAYNIE-SPENCE DAMAGE FUNCTION
BUREAU OF MINES
RUNOFF ANALYSIS
ACID RAIN
FILM CHEMISTRY
OUTLINE OF THE DAMAGE FUNCTION
COMPARISON WITH EDNEY/STILES MODEL
FORMULATION OF EDNEY/STILES DAMAGE
FUNCTION
DAMAGE FUNCTION TERMS AND THEIR BASIS
LABORATORY EXPERIMENTS
CONVERSIONS FOR COMPARING WITH EXPERIMENTAL

iv

DATA
DATA ANALYSIS
RESULTS/CONCLUSIONS
DISCUSSION/SUGGESTIONS
REFEREMBES
APPENDICES
APPENDIX A
PAST DAMAGE FUNCTIONS GRAPHED INDIVIDUALLY
APPENDIX B
COMPUTER PROGRAM USED TO SUM WEEKLY AND DAILY
ENVIRONMENTAL FACTORS

10.

11.

12.

LIST OF FIGURES

Current U. S. Levels of Sulfur Dioxide Emission expressed as Median
Sulfate Concentrations, peg/l.

Current U. 8. Levels of Sulfur Dioxide Emission expressed as Median
Sulfate Deposition, kg/ha-yr.

Median pH levels for the United States in 1980.

Potential-pH diagram, Zn-coz-Hzo; 25°C, 10"1 M Zn, 10‘5 M H2C03.

Stability Domains of Zinc Carbonate and Basic Zinc Carbonate (a) and
(b), Basic Zinc Sulfate (c).

Corrosion rate vs. Sulfur Dioxide for ten damage functions.

Corrosion rate vs. Time-of-Wetness for ten damage functions.
Regression Analysis of Hudson and Stanners data- Corrosion as a

Linear Function of Sulfur Dioxide.

Zinc Flux as a Function of SOx Flux in Condensation.

Zinc Flux as a Function of Incident Hydrogen Ion Flux for $02

Concentrations of 3, 30, 45 ppb.
Zinc Concentration After 1000 Seconds as a Function of Hydrogen
Ion Cencentration.

Zinc Concentration as a Function of Residence Time for Deionized

vi

13.
14.
15.
16.
17. _
18.
19.
20.
21.

23.

24.

.88

28.

30.
31.

Water.

Probable Time-of-Wetness vs. Relative Humidity (12).
Possible Fitted Solutions to Guttman's data (a,b,c).

Average Relative Humidity, %, for 5 NAPAP Sites for 1983-86.
Average Sulfur Dioxide, ug/ma, for 5 NAPAP Sites for 1933-33.
Dew point, °C, for 5 NAPAP Sites for 1983-1986.

Average Temperature, °C, for 5 NAPAP Sites for 1983-1986.
Average Precipitation, mm, for 5 NAPAP Sites for 1983-1986.
Average Wind Speed, m/s, for 5 NAPAP Sites for 1983-1986.
Experimental vs. Calcuated Weight Loss, mchm2 x 10's, for Zinc
at Research Triangle Park, NC, 1983-84 .

Corrosion, mol/cm2 x 10'6, vs. Time data range for Research
Triangle Park, NC, 1982-85, combined with calculated data points
for 1983-84.

Calcuated Weight Loss, mol/cm2 x 10’s, for Zinc at Research
Triangle Park, NC, 1983-84 showing each term of the Damage
Function.

Calcuated Weight Loss, mol/cm2 x 10‘s, for Zinc at Research
Triangle Park, NC, 1983-84 indicating contributions from Clean
Rain, Acid Rain and Dry Deposition terms.

Corrosion Rate vs. Time-of-Wetness for Equation (2.7).

Corrosion Rate vs. Sulfur Dioxide for Equation (2.7).

Corrosion Rate vs. Time-of-Wetness for Equation (2.9).

Corrosion Rate vs. Sulfur Dioxide for Equation (2.9).

Corrosion Rate vs. Time-of-Wetness for Equation (2.11).
Corrosion Rate vs. Sulfur Dioxide for Equation (2.11).

Corrosion Rate vs. Time-of-Wetness for Equation (2.13).

vii

Corrosion Rate vs. Sulfur Dioxide for Equation (2.13).
Corrosion Rate vs. Time-of-Wetness for Equation (2.15).
Corrosion Rate vs. Sulfur Dioxide for Equation (2.15).
Corrosion Rate vs. Time-of-Wetness for Equation (2.18).

Corrosion Rate vs. Sulfur Dioxide for Equation (2.18).

viii

LIST OF TABLES

labia

1. Environmental Parameters Measured at Materials Exposure Sites.
2. Organizations Conducting Environmental Monitoring Activities.

3. Review of Damage Functions

4. Additive Format of Ten Damage Functions

5. Sum of mean daily Kg, $02. tw, and products compared with sum of

mean weekly Kg, $02, tw, and products.

ix

1 . INTRODUCTION

Galvanized zinc coatings protect underlying steel from corrosion
primarily as a corrosion-resistant barrier, but also provide some additional
galvanic protection as a sacrificial electrode when the steel has been
exposed. The barrier protection function is effective because of a relatively
insoluble complex zinc carbonate surface coating which forms from the
reaction of zinc with atmospheric water, carbon dioxide, and oxygen. When
sulfur dioxide from the atmosphere reacts with water (rain, dew) the
surface moisture is acidified and the solubility of the coating is modified
leading to more rapid corrosion, hence the effects of environmental
pollution are reflected in increased damage costs to galvanized steel
structures. In order to reach political decisions related to pollution, it is
desirable to estimate these costs monetarily.

Zinc has been a pepular testing material from the beginning of
corrosion testing, and there exists a large literature of experimental data
for zinc. It also has been fashionable to model such data in terms of the
ambient experimental conditions because it has long been known that
environments are not equally corrosive. Environment-based corrosion
models vary widely in sophistication, from simple ”rural, urban, industrial”
to complex mathematical expressions using relative humidity (RH), sulfur
dioxide concentrations, and other variables. Obviously, if an accurate
model were available, it could be used, together with the national

distribution of, say galvanized steel, and accurately monitored

2

environmental conditions, to calculate the overall annual corrosion costs of
”acid deposition.”

The U. S. Congress in 1980 resolved to face the political question of
whether cleanup costs will be larger or smaller than the resulting benefits
by passing the National Acid Precipitation Act (18). Recognizing that the
questions are complex and 'off-the-shelf' answers were not available, the
Act established a ten-year program to develop the required database.
Specifically needed details, with respect to galvanized steel, are:

- A national inventory of the metal, detailing amounts, age, location,
structure type,

- aerometric data recorded with sufficient frequency to reflect
average environmental conditions as well as more-damaging
excursions, and

- a damage function that uses values from the aerometric data base

to compute metal corrosion from the known corrosive factors.

In addition to these requirements, an experimental data base, obtained under
precisely controlled conditions, is needed to demonstrate the validity and
reliability of the damage function's calculations.

Without question, this is a tall order, even for a single metal. In fact
information is needed for many metals, building materials, and cultural
structures. The Environmental Protection Agency, in cooperation with
several other agencies, notably the Bureau of Mines, Department of the
Interior, has undertaken this program aiming to meet the
Congress-mandated 1990 deadline. Data collection efforts have moved

‘ forward more rapidly for zinc and galvanized steel than other materials,

3
hence these offer the best opportunity to confirm the proposition that costs
of acid precipitation damage to materials can be assessed with the required
accuracy using this approach.

Michigan State University has entered into a Cooperative Agreement
with the EPA where, in close cooperation with the EPA and the BOM, MSU
will evaluate:

- The literature of damage functions regarding zinc and galvanized
steel, particularly a function recently proposed by the EPA, and revise them,
in so far as possible, to bring them into conformity with one another;

- aerometric data provided by EPA, with respect to accuracy and
statistical validity, then integrating them as appropriate for comparison
with experimental results;

- experimental data provided by the BOM and from the literature to
determine goodness-of—fit between damage functions as well as accuracy,
reliability, and precision. The overall objective of the MSU effort is to
provide scientific coordination and review of the various efforts related to
zinc and galvanized steel. This thesis specifically addresses:

- A literature review of the possible variables affecting zinc

corrosion,

- Comparison between previous damage functions for zinc,

- Comparison of the recent EPA proposed damage function with BOM

experimental data for Research Triangle Park, NC.

1.1 W

In the overall program relating to zinc; the evaluation of the cost for

4
galvanized steel and zinc due to acid rain, the activities of each research

group may be outlined as follows:

Environmental Protection Agency:
Haynie, F. H., and Spence, J. W. (EPA)
Edney, E. 0., and Stiles, D. C. (Northrop Services)

- Development of damage functions and models based on
controlled laboratory experiments and theoretical background
in cooperation with Northrop Services, Inc.

- Integration of the damage function with a cost analysis model

- Analysis of aerometric data

Bureau of Mines, Department of Interior:
Cramer, S. D., Flinn, D. R., and Carter, J. P.

- Refinement of atmospheric test data at NAPAP sites: NC, NJ,
NY, OH and DC

- Statistical evaluation of atmospheric tests

Michigan State University:
Summitt, R., Lang, P. L., and Li, Z.
- Statistical analysis of environemental data (provided by the EPA)

- Evaluation of damage functions for zinc and galvanized steel

5
Comparison of predicted values with experimental results (provided
by the BOM)

2. BACKGROUND

Zinc is used as a protective coating for ferrous products because of
its resistance to atmospheric corrosion. Numerous atmospheric tests have
been performed to determine what factors affect zinc corrosion at various
sites. Almost from the beginning, analytical function modeling has been
used to correlate experimental corrosion rates with environmental data.
Such functions often differ considerably from one another depending on
what environmental factors are measured in experiments.

Early on, Vernon (2) performed controlled laboratory experiments

using a bell jar system to control relative humidity, RH, sulfur dioxide, 802,

carbon dioxide, C02, and other environmental factors, and found that

atmospheric corrosion rates for zinc increase rapidly when the RH exceeded
a certain critical value. Vernon also performed ”open-air exposure testing”
now known as outdoor testing, in which precipitation runoff was collected

and specimen weight loss/gain were measured. Many techniques used today
to evaluate corrosion of metals are not substantially differerent from those
used by Vernon. Ellis (20) later linked the initial corrosion of zinc to
corrosion rate for one year exposures, correlating 28-day weight losses

with hours of rainfall and the time during which 100% RH was achieved
during the first five days of exposure. Others (7,12,15, 26, 36) have found

that $02, moisture and rain, temperature, wind, particulates and other

measurable quantities affect the corrosion rate of zinc.

7

Two types of testing typically performed are outdoor exposure and
laboratory chamber studies. Early studies may not have included
environmental measurements but today's tests often are quite sophisticated
with controls aimed at separating variables. Studies have included weight
loss, pollutants, and meteorological measurements at outdoor exposure
sites, as well as laboratory chamber studies where known pollutants are
injected into chambers which simulate dew and the diurnal illumination and
temperature cycles.

One objective of such studies is to provide a basis for predicting
material corrosion in service environments. Today there is considerable
interest in using such models to compute corrosion costs caused by ”acid
rain” and dry pollutants. Pollutants such as sulfur dioxide and nitrates are
being investigated for the purpose of providing a cost/benefit analysis in
regards to environmental "cleanup”, i.e. to deterimine whether emission
restrictions will be cost effective, if new replacement materials offer an
acceptable alternative, or if present pollution levels are tolerable. Since
galvanized steel materials are used widely as roofing, siding, fencing etc.,
they are exposed to the atmosphere and are affected by pollutants and
environmental changes.

U. S. Congress has mandated studies of corrosion rates caused by
acid precipitation via the National Acid Precipitation Assessment
Program, NAPAP (62). Task Group VII of NAPAP is evaluating effects on
Materials and Cultural Resources. The goal of the task group is to develop a
damage function for construction materials affected by wet and dry acidic
deposition. A damage function mathematically relates the rate of material

deterioration to environmental factors. Congress requires a NAPAP report

8

in 1990 that will give an assessment of materials damage caused by acid
deposition. As the research continues, it becomes evident that the cost
”acid" rain has had on materials will not be an easy problem to evaluate.

Current levels of sulfur dioxide emission are shown in Figure 1 as
median sulfate concentrations in peq/L and in Figure 2 as median sulfate
deposition in kg/ha—yr (38). The deposition is a derived quantity calculated
by multiplying the observed concentration by the amount of precipitation
associated with the sample, thereby obtaining a value of the mass deposited
per unit area. The deposition pattern gives the appearance of mostly
short-range transport from the primary source region, since the heaviest
deposition occurs almost directly over the highest sulfur dioxide emission
area. The median pH values for 1980 are shown in Figure 3 (62). Values
below 5.0 can be seen for most of the eastern US. These values, although
much lower than ”normal” rain, may not be significantly different from
those of the 1950's (38).

For the United States, corrosion costs have been estimated to be $5.5
billion in 1947, $15 billion in 1965 , and $9.7 billion in 1975 (39). Haynie

estimates the cost of materials damage from $02 accelerated corrosion in

1980 as, “ $1.50 1: 0.50 for each ug/m3 of $02 plus $0.10 for substitution

costs and $0.50 for prior years of accumulated damage. The costs when

related to coal-fired utility power production can range from 0.0049
mills/kWh to 0.25 millls/kWh depending on the level of 802 contributed by
the power plants (3).“ For 1985 the cost of acid deposition damage to zinc

gutters/downspouts, wire-mesh fencing, and chain-link fencing was

estimated at $3.69 per ft2 area of material(40). These figures and others

r '
.5 .. J

 

' Median Sulfate
Concentration (ueq L")
(Samples through 7/84)

 

Figure 1
Current: U. S. levels of Sulfur Dioxide Emissions expressed as Medsim
Sulfate Concentrations, peq/l (38).

III

 

Figure 2
Oman: U. 8. levels of Sulfur Dioxide Buissions acpreesed as
Median Sulfate Deposition, kg/ha-yr (38),

II

 

 

 

 

 

 

 

 

Figure 3
Annual Average pH 1980 (62)-

 

1 2
give a wide range of possible estimation schemes to evaluate the cost of
materials damage due to pollutants. Most of the figures are the costs over

and above the corrosion costs under pristine conditions which are usually

taken to be 802 at .5 ug/m3 and pH at 5.2 (40). It can be seen from these

figures that the controllable costs resulting from pollutants is substantial.

2.1 W

Zinc corrosion rates vary widely. Weight losses, for example, have
been found to vary from 0.280 gm/4x6 panel to 0.035 gm/4x6 panel for
two-year exposure tests (5). Other tests exhibit seasonal variations for
weight loss (5). These variations can be explained in part from variation in
environmental factors. Since, for zinc, atmospheric factors control the
initial corrosion rate and also have been reported to control its corrosion
for longer term exposures, it is critical that atmospheric factors be
closely studied (4,9).

Early investigators set up standard testing procedures to make the
basis of knowledge involving outdoor corrosion testing as generic as
possible. Test sites were categorized early as industrial, rural, urban, or
marine. A semi-rural site has been established on the Michigan State
University campus roughly one quarter mile distant from an interstate
highway. The rack is in a fenced in area near two small inland lakes, one
being only about 30 feet away, the other roughly 250 feet away. In a
preliminary trial, samples of zinc were exposed in August 1987. These had
been cleaned in 23% ammonium hydroxide solution for 2 minutes, and then

polished using 600 grit paper. Preliminary results from the first set of five

1 3
samples set out showed almost no corrosion one week later since they
experienced two hot, low humidity days in which to form a thin film as
compared to the five samples set out two days later on which rain fell the
night after exposure. The galvanized steel panels set out in August 1987

exhibited no rusting or coating corrosion after six months of exposure.

2.1-1 AQIQBAIN

Normal rain is considered to have a pH of approximately 5.6, as a

result of equilibrium with 002 and carbonic acid, hence acid rain is defined

to be wet deposition (snow, sleet, rain or dew) having a pH below 5.6. Zinc
forms a protective film which usually is stable in a higher pH range than
that of "acid“ rain (Figures 4 and 5) (19). A reaction will occur between the
rain and the protective film if the pH of the rain is too low. This film
consists of corrosion products, mainly zinc oxide, zinc hydroxide, and basic

zinc carbonate (1). ln moist conditions zinc oxidizes to form zinc hydroxide:

The zinc hydroxide then reacts with atmospheric consitituents such as 002,

SOX, Cl', etc. to form basic zinc salts at the hydroxide/air boundary,

provided the pH of the surface is sufficiently high. The zinc hyroxide and
salts formed protect the film from further attack. If the pH of the surface
moisture is low, as for acid rain, no protective layer will form and films
formed early may be dissolved, hence, the corrosion rate increase. Although
neutral rain does dissolve the film, the solubility of the film will be greater

at a lower pH.

PotentIaI,V(SI-IE)

1.2 .

1.0
0.8
0.8
0.4
0.2

- 0.2
- 0.4
-'0.6
" 0.8
" 1. 0
- 1. 2
- 1.4

14

 

 

 

Zn

Figure 4
Potential-pH diagram,.Zn-CO

 

2'32

8 10 12

o; 25 c, 10'

1

M Zn, 10"

5

 

1
pH

M‘HZCO

4

3 (19).

I5

 

 

 

 

 

 

 

 

 

 

 

 

o
’24“ 2
V .2
'5 -2- g
E u
v-3.
§ §
3‘ 3‘.
35-5.- .33.
a .0,
2 -6 "9' ,
o
(a)
“o 1
zn(cn{%ucsch)mfi
-14
A
{'2‘
2 .
O -3- an
TE! C )
' - - e-Zn OH
333 4 2
3).-5-
a
2 '6 l I 1 '

 

02468101214
pH

Figure 5
Stability Domains of Zinc Carbonate and Basic Zinc Carbonate (a) and (b),

Basic Zinc Sulfate, (c) (19).

16
2.1.2 ENMIBQNMENIALEACIQBS

2.1.2.1 RELATIVE HUMIDITY

The direct dependence of corrosion rates on the relative humidity (RH)
later evolving into the term time-of-wetness, was first demonstrated by
Vernon (2). Sereda developed a device for measuring the time that a
corrosion specimen was wet, which is, in effect, the time that Vernon's
critical humidity is exceeded (5). Time-of-wetness is defined as RH
conditions resulting in an adequate film of water forming on a metal's
surface. This condition is influenced by surface contaminants, i.e., soluble
ions which lower equilibrium vapor pressure,and the nature of products
forming on the surface (63). The critical relative humidity varies from one
metal to another. Guttman and Sereda used a Dew Detector to study
corrosion of rolled zinc and found that the measured time- of-wetness
corresponded to the time the RH exceeded 86.5% (5). The critical RH value
for zinc has been known to vary from 80 to 100 percent depending on
conditions and site locations as well as pollutants (63). Relative humidity
has been expressed as a function of temperature T, and dew point DP, by the
relationship (42):

RH - 100 exp [-0722 + 0.00025 (T 4- DP) [T - DP] ] (2.2)

where T and DP are in °C.

1 7
2.1.2.2 WIND VELOCITY

The wind velocity is important in relating ambient pollutant
concentrations with the amount of pollutant expected to be deposited on the
samples, i.e., higher speed wind will deposit proportionally more airborne
reactant on a surface. Thus, deposition velocity (cm/sec) multiplied by
pollutant concentrations (pg/m3) provides flux data. Deposition velocities
are a function of wind speed and for gaseous pollutants are on the order of 1

chsec (30), and can be calculated on an hourly basis from windspeed data.

2.1 .2.3 TEMPERATURE

Temperature plays an important role in the corrosion process. The
observed temperature delay of metals behind changing ambient values
resulting from metal's heat capacity, along with the normal increase in
corrosion activity, which can theoretically double for each ten degree
Celcius increase in temperature, are two temperature dependent factors
affecting material behavior in the atmosphere. Some zinc studies in natural
environments have Shown a temperature dependence in the analysis of
corrosion rate data (2,12), but usually it is not specifically considered
because it is intimately related to other factors, e.g., relative humidity,
time-of-wetness and dew point, and its effects are contained within these
factors.

13
2.1.3 EQLLLIIIANIS

Pollutants are undesirable aerosols and gases added to the

atmosphere by human activity. Although anthropogenic pollutants

contribute to corrosion damage, many pollutants i.e., SOZ, chlorides, 002,

also come from natural sources and are not amenable to control.
2.1.3.1 SULFUR DIOXIDE

Sulfur dioxide is a major pollutant and is known to be linked with the
atmospheric corrosion of materials. Sulfur dioxide emissions from

metallurgical and other industrial acvitivy, e.g., burning fossil fuels, causes

much of the 802 pollution. Estimated U. S. emissions of S02 in ppm (one
part 802 to 1 million parts air), were 2.2 in 1965 and .8 in 1969 (63). In

1970 33.9 million tons (2) of 802 were emitted and in 1980 it decreased to
27 million tons (62).

Zinc surfaces are sensitive to 802 because it reacts with the
protective zinc carbonate film and forms a soluble sulfate (1). In addition,
802 slowly reacts with oxygen to become 803 which in turn rapidly

combines with atmospheric moisture to form sulfuric acid aerosol. These
reactions involving sulfur are estimated to cause deposits which are about
50% wet and 50% dry (62).

A number of investigations have attempted to quantify $02

acceleration of atmospheric corrosion (2, 26, 63, 64). The dependence of

19

zinc corrosion on $02 was shown by Shikorr, who found in 1964, that

monthly average corrosion rates followed the annual cyclic variation in $02
concentration (33). Others have established that zinc corrodes more rapidly
in highly industrialized areas than in rural (26). Another study investigated
the corrosion of zinc exposed to environmments containing 802 or H2804
aerosol at varying humidities. Corrosion occured when the RH was high

enough to allow the formation of an electrolyte layer on the zinc surface.

The amount of further corrosion that proceeded was directly linked to the

amount of $02 or H2S04 aerosol on the zinc surface. Under these conditions

zinc corrosion is by chemical mechanisms (2).
2.1.3.2 CHLORIDE ION

Salt content appears to have a direct relationship on the corrosion
rates of materials (5). Although the amount of chlorides in the atmosphere
decreases dramatically as the distance from the seashore increases, large
amounts of chlorides have been found in corrosion products on specimens
exposed at some inland sites (5, 63). Nevertheless, chlorides usually are

monitored only at marine sites.
2.1 .3.3. PARTICULATES

Some studies have indicated particulate matter may affect the

corrosion rate of zinc by increasing the time-of-wetness and by adding

2 0
surface area for the absorption of 802 during periods of dryness (65, 56).

With the exception of chloride salts, particulates are widely different in
corrosiveness, depending on their chemical composition and source.
Accordingly, it is not possible to deal with them effectively in a general

way.
2.1 .3.4 NITROGEN OXIDES

Nitrogen pollutants have been identified as possible corrosion

enhancers. Most nitrogen oxides emitted from combustion processes as

nitrogen oxide, NO, react with the atmosphere to form N02. Nitrogen

dioxide, which is usually considered the main nitrogen pollutant near an

emission source, then slowly oxidizes to HN03 . Because the reaction rate is
slow, the area near the emission source usually shows low HN03 and
nitrate levels. Results from chamber studies of the effects of N02 are

mixed. One such study revealed that 0.05 and 0.5 ppm N02 had no significant
effect on zinc's corrosion rate (54), yet another found some increase in
corrosion when N02 was added to an 802 containing atmosphere (54). So
far, no outdoor studies relating N02 to corrosion rates appear to have been

reported.

21
2.2 W

Numerous experimental studies of zinc and galvanized steel have been
published throughout the 20th century. Frequently the results have been
used to formulate equations modelling corrosion behavior, and a‘ number of
such equations relating corrosion to environmental factors are reviewed in
this study. These equations are useful for understanding the basis for
correlations between environment and damage. Ten equations having defined
coefficients are reviewed in detail. Uniform units for corrosion rate,
urn/yr, were established so that comparisons can be made between the
equations.

These equations are based on a variety of studies performed outdoors
and in laboratories worldwide. The equations are listed in before- and
after-conversion form in Table 3. An attempt to begin correlating

parameters used in most equations into the mathematical format of the

form C - ax1 + bx2 + c, has been done as shown in Table 4. To compare the

functions to one another, equations one through ten are plotted together,
Figures 6 (corrosion rate vs. 802) and 7 (corrosion vs. time-of- wetness),
showing the Similarities between them. An exposure time of 1 year was
used if the initial equation was not given in terms of rate. For Figure 6,

time-of- wetness was set as 2920 hours (1/3 year). For Figure 7, sulfur

dioxide was set at 50 pg/m3. As can be seen in Figure 6, two equations (5

and 6) are not as dependent on 802 as the others, this is because they are

also temperature dependent.

These graphs illustrate how closely many of the equations are to one

22

IABLEJ W

10.

C - .001028 (RH-48.8)802 - Haynie & Upham
C - .01028 (404/[4.04-ln(tw/8760)]-48.8) 302

c - .005431tw-3152(Soz+.02339) - Guttman
c - .005431tW-8152 (802/2600+.02889)*14.98+ yrs. exposed

c - .0049tw'91(SOz+.050) - Hakkarainen 3 Ylasaari
c - .0049tw'91(802/72.8+.050)+yrs. exposed

C - .00076tw'5SOZ'718 - Barton, et. al.
c - (woman/335)-5(Sozx.723)-713}*51.19

c - 04511802 4- .031 pNo2 + 9(17-1334-23750
e(-5.4(100-RH)/RH) .. Haynie

C - .045ttso2 + .031 pNo2 + e<17-133‘4-237E)}
e(-5.4{100-[404/(4.04-ln(f“/8760))]}/ 404/[4.04-ln(tw/8760)])

C - (.0187302+9(41'85-E/RT))*1W -Haynie, Spence, & Upham
c - (01373029441~35‘E’RT))*(tw/3760) + yrs. exposed

C - .22 802 + 6.0 - Atteraas & Haagenrud
C - (.22 SOZ+6.0) * .140 -

C - .037802 + .649 - Hudson & Stanners (data)
C - .18 $02 + 3.7 - Haagenrud, Kucera, & Atteraas
C - (.18 (802/728) 4- 3.7) * .140+ yrs. exposed

C - .17 $02 + 3.6 - Knotknova, Gullman, Holler,
C =- (.17 (802L728) + 3.6) * .140 and Kucera

Corrosion -

2. In

3. In

4. In
5. In

Inp.
6. In

10.

$02 +

$02

In
In(SOz+.02889)
In(SOZ+.05)

.71 8 In

In
[N02 1611“]

In

.16
.18
.17

23

tw + constant
tW or RH B
|n(RH-48.8) -6.88
.8152 In -5.21

.91 In -5.32
.91 In -7.18

-5.4(100-RH)/RH -3.1
{ -5.4(100-RH)/RH + 2.22)

In -3.98
{In 4- 2.60}
6.0
6.32
3.7
3.6

(26)
(1 2)
(48)
(44)
(7)

(15)

(45)
(43)
(46)
(47)

24

no“ izaabiia .
F4 mmmzzfionmoaa a
E 535$ n

E 932:. e

74 so a. 2253 n
E .o 3. (>95sz e
H2. 3 a. aamzuoia A
We oamzueiimméwts u
7.; .o a. 923.: a
:4 23.3» 4 5233.55 2

1111111111

28.333 owns—.3 con. new 03.63 .335 .2, 35m aowuouuoo

8233 5on «:58
P. .me .3 .e .3 ..

 

 

 

.62» F n. 2:: 2:3an

Coo» n\5 8:0: owmu I 3...

Now .m> whosm ZOEOwEOO

 

.20“

(WIND) Elva NOISouaoo

o wuaqu

25

.0 do ZOE—(m

23.—n5 on mozwmm .m_z><I .

25¢: a. m_z><1

1111 it

acowuouzm own—53 don. you acocuoziucioaau. .n> 33— defimouuou

A235 mmmzm; .8 m2:

8% . 82 . o
- b n D I - n I L D . 00°

 

 

I
O.
f‘

 

ro..v —

$52 8.3 a» _ ”l2: mmamoaxm
.23: on i Sm

mmszBllemZH ..m> m._.<m szommoo

n ouawwm

(ax/no) 31w Nousoaaoo

26

other even for a variety of test sites. The equations are illustrated

individually in Appendix A as corrosion rate vs. 802 at three different

values of time-of-wetness (500 hrs, 1000 hrs and 8760 hrs) and are

graphed as corrosion rate vs. tw at three values of S02 (0 ug/m3,

so pg/m3 , and 100 pg/m3).
22-1 SDNMEBSIQNS

Corrosion rates were converted to units of urn/yr for several reasons:
Metric units are more universal than English units; standard practice in the
field is to measure corroson in terms of penetration/time, i.e., miVyr; and
ASTM (G1) outlines a standard for calculating the corrosion rate which is of
the form:

corrosion rate - (K x W)/(A x T x D), (2.3)
where K - a constant,

W - mass loss in g, to nearst1 mg,

T - time of exposure in hours to the nearest 0.01 hr,

A - area in cm2 to the nearest 1 mg, and
D - density in g/cm3.

This formula results in the corrosion rate being equal to (K * cm/hr) which
would indicate units of penetration/unit time. Standard units used are:

pm/yr , for corrosion rate,

ug/m3, for $02, and

hrs, for time-of-wetness.

Zinc weight loss corrosion rate conversions are based on the density

27

of zinc, D - 7.13 g/cm3, and the metric base-10 multiples.
1 g/(m2 yr) [1/0] [1rn/100cm]2 [10, 000 urn/1 cm] - .140 um/yr.
1 g/m2 4- exposure time in years - .140 urn/yrs exposed.
1 g/(m2 day) [.140 pm/yrylg/(m2 yr)] (365 day/yr) - 51.19 pm/yr.
1 g/3 x 5 in2 panel [1/0] [ 1 in2/6.45 cm2] [10, 000 urn/1 cm] +
exposure time in years - 14.5 pm/yrs exposed.
1 mil/yr (25.4 um/mil) - 25.4 pm/yr.

Sulfur dioxide concentrations are converted from part per million,
ppm, to jig/m3 as in ASTM standard (D1914) for room temperature (25°C)

and for standard pressure (760 torr). The conversion is: 1 ppm (molecular
weight)/24,450 - X mg/l, so for S02 this becomes 1ppm [64.0/24,450] -
.00262 mg/l.

The relation 1 mg SO3/dm2/day - 0.028 ppm 802 for lead peroxide
candles by Cominco and 1 mg SOaldmzlday - 0.024 ppm 802 for candles
prepared by National Research Council was established by Foran et. al. at

Birchbank, British Columbia (12). We used 1 mg SO3/dm2/day - 0.023 ppm

802. Conversions from these relations yielded the following conversions:

1 ppm [(.0026 mg/l)/1 ppm] [1000 ug/mg] [1000 I/m3] - 2600 ug/m3.

1 mg/(dm2 day) [0.028 ppm/(1 mg/dmzldayfl [(.0026 mg/l)/1 ppm]
[1000 ug/mg] [1000 l/m3] - 72.3 ug/m3.

1 "is/(m2 day) [728 liglma/mg/(dm2 day)] [1 W10 dm]2 - .723 pg/m3.

1 g/(m2 day) [.728 pig/m3/mg/(m2 day)] [1000 mg/g] - 728 pig/ma.

Percent Relative humidity is converted to time-of-wetness using (15):

28

f_ e(4.04 - 404/RH)’ or (24)
RH - 404/[4.04 - In tw], (2.5)
where f “' t - tw,

f - fractional time-of-wetness,
t - time, hrs, and
tw - time-of-wetness, hrs.

22-2 EQUAIIQNS

Haynie and Upham (26) developed a linear function for zinc corrosion
rates from outdoor testing at eight urban locations. Special high-grade
commercial zinc (99.9% pure) panels used were 4 x 6 inch and .06 inches
thick. Rural sites, expected to have lower pollutant levels but generally
the same meteorological conditions, were used as control sites. The
function for the urban sites suggests that there will be no corrosion if the

mean relative humidity is less than 48.8%. Their equation is:

K . 0.001023 (RH - 43.3) $02, (2.6)

where K - corrosion rate, pm/yr,

$02 - concentration, (lg/m3, and
RH - relative humidity, %.

In uniform format where S02 is expressed as jig/m3 and tW is in hours, this
is:

K - 0.001023 (404/[404 - In (tw/8760)] - 43.3) 302. (2.7)

Guttman (12) found that special high-grade zinc exposed at

29

Birchbank, British Columbia, a semi-rural site, on different dates for a
specific period of time corroded at different rates. An empirical equation
was given which accounted for most variation in corrosion losses and was
valid for exposure periods of up to almost five years duration (December,

1958- December1963). Parameters measured were time-of-wetness, dew

-point, ambient and panel temperature, wind direction and velocity, S02, and

solar radiation, and the equation was:
K - 0.005461 (A) 03152 x (B + 0.02339), (2.3)

where K - corrosion loss, mg/3 x 5 in. panel,
A - time of wetness, hr, and
B - atmospheric $02 content when panels were wet, ppm.

In uniform format where 802 is expressed as jig/m:3 and tW is in hours, this
is:
K - [0.005461 (A) 03152 x ( (B/2600) + 002339)]
*14.5 + yrs exposed. (2.9)

Hakkarainen and Ylasaari (48) developed an equation relating time-of-
wetness and $02 for the pure zinc panels exposed in southern Finland.
Corrosion rates of the order of 1 urn/yr were observed. The equation,
based on four years of data from three of five sites is:

K - .0049 two-91 ( S02 + 0.050), (2.10)

where K - amount of corrosion, pm,
$02 - mean content, mg/dmzlday, and

tw- ti me-of-wetness, hrs.

In uniform format where 802 is expressed as )tg/m3 and tW is in hours, this

30

is:

K - [.0049 two-91 ( {502/723} + 0.050) 1 + yrs exposed. . (2.11)

Barton, et. al. formed an equation of the form (44):

K - 0.00076 two-5 $020713, (2.12)

where K - corrosion rate, g/m2 day,
tW - time-of-wetness (time RH > 80%, temp. > 0°C), hr/day, and

802 - deposition, mg/m2 day.

In uniform format where 802 is expressed as )tg/m3 and tW is in hours, this
is:

K - [0.00076 (V365)°°5 (sog.723)°-713] * 51.19. (2.13)

Haynie (7) used data from test sites in St. Louis, MO to show that
relative humidity, temperature, wind speed, and levels of total sulfur gases
and oxides of nitrogen were found to be statistically significant variables
and introduced two equations based on regression analysis: One for the best
estimate of the corrosion rate and another for the instantaneous corrosion
rate. An equation also was developed to relate site-to-site RH differences
to temperature differences and from this another equation allows

time-of-wetness estimation.

K - [0.045u $02 + 0.0314u . No2 + e(‘7-133'4-237E)] x
e,(-5.4(100- RH)/RH)’ (2.14)

31

where
K - corrosion rate in pm/yr,
E - 1000/ (T + 273.16),
E - temperature parameter,
T - hourly averge temperature in °C (ave. value given - 11°C),
RH - 100 exp(-[6.32(T-D)/(80 +T)],
RH - relative humidity, %,
T - temperature °C,
D - dew point, °C,
802 - concentration, (lg/ma,

N02 - concentration, ug/m3 (ave. value given - 61.44), and
u - deposition velocity, cm/s (ave value given - .54).

In uniform format where 802 is expressed as pg/m3 and Iw is in hours, this
is:
K - [0.045u 502 + 0.03140 . N02 + 41743342373] x
e(-5.4(100- {404/[4.04 - In tw]} )/ 404/[4.04 - In tw]. (2.15)

For instantaneous corrosion:

K - [0.045u 502 + 0.0314p No2 + 41239742373] x
e(-5.4(100-RH)/RH). (2.16)

Haynie, Spence, and Upham (15) statistically designed a laboratory
study to look at the effects of gaseous air pollutants on materials.

Commericial grade 18-gage galvanized steel with an approximate 25 pm

coating was exposed to sixteen polluted and four clean air conditions. 802,

N02, ozone, and RH were used at high and low levels. Temperature was held

constant at 35°C and weight loss was measured after 250, 500 and 1000

32

hours. Corrosion was observed to be linear with time for the 7.6 x 12.7 cm
panels. The empirical equation resembling part of the Arrhenius equation

was arrived at for describing the corrosion rate:

K - (0.0137 302 + exp(41.35 - E/RT)) tw, (2.17)

where K - loss, pm,
$02 - concentration, pg/m3,

R - gas constant, 1.9872 can mol K,

E - activization energy, 23,240,

T - temperature when wet, °K (mean temperature given - 8.1 °C), and
tW - time of wetness, yrs.

In uniform format where 802 is expressed as p.g/m3 and tw is in hours, this
is:
K - (0.0187 802 + exp(41.85 - 23,240/RT))
[fW /8760] + yrs exposed. (2.18)

Atteraas and Haagenrud (45) derived an equation based on one and
two year data from 22 sites in Norway. Regression analysis included the

parameters: 802, rain amount, its pH, sulfate, chloride,and temperature.

The results from the 10 best linear regression equations having a
correlation coefficient R - .76. had the form:

K - 0.22 so2 + 6.0, (2.19)
and K - 0.27 CI + 0.22 S02 + 4.5, (2.20)

where K - corrosion rate, g/m2 yr,
802 - concentration, ug/ma, and

3 3
Cl - chloride deposition, g/m2 yr.

In uniform format where 802 is expressed as (lg/m3 and tw is in hours, this

is:
K - [0.22 802 + 6.0] "' .140. (2.21)

Hudson and Stanners (43) presented data for zinc, 4 x 2 inch
specimens exposed vertically at sixteen sites throughout England for one
year. Sulfur dioxide was measured using the lead peroxide method and their

data were used to give the equation referenced in several articles (45, 47):

K - 0.16 so2 + 6.32, (2.22)

where K - corrosion rate g/(m2 yr), and
802 - concentration, ug/m .

In uniform format where 802 is expressed as ttg/m3 and tW is in hours, this

is:

K - [0.16 802 + 6.32] * .140, (2.23)
or in urn/yr,

K - .0224802 + .8848. (2.24)

An analysis of the data was done using simple regression with 802 vs.

corrosion and the sixteen data points given in Hudson and Stanners article
yielded the equation in urn/yr with R - 0.96 (Figure 8):
K - .037 $02 + .649. (2.25)

Haagenrud, Kucera, and Atteraas (46) presented another form of the

34

Figure 8

Regression Analysis of Hudson and Stanners Data - Corrosion as a Linear Function of S02

 

 

 

 

 

14
12 ~
A I
E 10
5' .
a
2 3L
2
m .
8 6 -
I:
o f
o 4)
’ y . 0.649 + 0.037x n a 0.96
2
o a l x L a I L
0 100 200 300 400

SC2 (UG/M3)

35

linear equation relating corrosion rate to 802:

K - 0.13 so2 + 3.7, (2.26)

where K - g/mz, and
802 I mg/m2 day.

In uniform format where 802 is expressed as jig/m:3 and tW is in hours, this
is:

K -[ 0.13 (802/.728) + 3.7] * .140 + yrs exposed. (2.27)

Knotkova, Gullman, Holler, and Kucera, (47) showed yet another linear
equation from results of a four year field study in Sweden (6 sites) and
Czechoslavakia (5 sites) on zinc (98.5%). The 10 x 15 cm panels were

exposed at 45° to the horizontal beginning in June 1978. The following

relation was found for corrosion and 802:

K - 0.17 so2 + 3.6, (2.28)

where K - g/m2 yr, and
$02 - deposition, mg/m2 day.

In uniform format where 802 is expressed as pg/m3 and tW is in hours, this
is:

K - [0.17 (502/ .723) + 3.6] * .140. (2.29)

Guttman and Sereda (9) used data from seven North American sites

over a four year period to show conclusively that atmospheric factors

36

control the initial rate of corrosion of zinc for at least one month. Good
correlation exists for a corrosion loss equation for one month exposure
periods. Longer term exposures show atmospheric factors control zinc
corrosion in all but one site, and equations developed permit prediction of
corrosion losses, and account for variations in corrosion losses for panels

exposed at different times of the year. For 1 month exposure losses:

K-b+bA+bB+bC+bA2+bB2+bCZ+bAB
+ b AC + b BC, (2.30)
where K - corrosion loss of Zn in g/panel,
b - regression coefficients,
A - (time of wetness, days-14.3)l5.0,

B - (temperature, °F - 5.25)/ 15.0, and
C - (S02, ppm -0.0202)/0.024.

For 3, 6, 9, & 12 month exposure losses:

K-b+bA+bB+bC+bA2+sz+bC2+bAB+b
AC+bBC*(+bBD+bCD), (2.31)

and K - corrosion loss, 9,
A - (time of wetness, days-75)/70,

B - (panel temperature, °F -45)/12, and
C - ($02, ppm-.02)/0.02.

For the marine environment Variable D is added
D - (Cl ,mg-18)/10.

Another equation was developed to correlate data for 3, 6, 9, & 12

month exposure losses:

K - aAn (B + c), (2.32)

37
where K - corrosion loss, 9,
A - time of wetness, days,
inland sites-
B - 802, ppm , B - temperature, °F.

marine sites-
B - chloride count, mg Cl lmzld., B - temperature, °F.

Barton (51) used a very general series expansion to describe the
corrosion rate of materials. The total corrosion over a long period of time

is a sum of corrosion effects during shorter periods.

K II 2 8i tWI' (2.33)

where K - total corrosion loss,
a - a series of coefficients for sulphur dioxide, and
tw - time of wetness (or time RH280% when T20° C), hr/day.

Barton (51) also expressed the corrosion rate in another manner in
which he tried to conbine the theoretical and the empirical approach to
equations.

vk - M tn 8'“, (2.34)

where Vk - corrosion rate, urn/yr,

t - time of wetness (or time RH280% when T209 C), hr/day,
S - $02, mg/mzday, and

M,n and m - constants where M involves the specific corrosion
kinetics of the metal concerned.

Legault and Pearson's (6) equation describes the behavior of
galvanized steel in industrial and marine enviromnents, East Chicago, IN,
and Kurie Beach, NC, respectively, where K and N can be used to separate the

tendency for a corrosion product to form from the effect of that corrosion

38

product on the subsequent reaction. The equation can predict reliably

long-term atmospheric corrosion behavior.

AW - Kt", (2.35)

where AW - wt loss in gms/mz,
t - exposure time in years, and

K, N - constants.
Feliu and Morcillo (52) introduced a different way of describing “1
corrosion for mild steel, zinc, copper, and aluminum based on data from ten 1.

testing stations located throughout Spain. Assuming other secondary
factors such as surface orientation, temperature, etc. to be treated as unity
the equation is:

K - M tfg, (2.36)

where K - corrosion rate,
M - function of material (corrosion module per 1,000 hr of exposure in
a pollutant-free atmosphere with RH 2 85% and T > 0°C),
t - time throughout the year M is met,
f - inhibition due to corrosion product in terms of t, and
g - corrosion stimulation coefficient caused by air pollution,
1 + 3802 + bC.

Another Feliu and Morcillo equation took the form (50):

K - a exp (b 802 - c/RH), (2.37)

where K - corrosion rate, and
a,b,c - coefficients.

39

Mikhailovaskii and Sokolov (25) compared expected corrosion values
with those from COMECON countries corrosion stations and found good
agreement between calculated and observed corrosion (relative error 25%).

They used the folowing expression for the mean annual corrosion:
M - a (0302)b tads 103, (2.33)

where M - mean annual corrosion rate, g/m2 yr,
tads - time relative humidity is 280% and the temperature is between

-15°C and 30°C, and
a,b - regression coefficients.

Another of Haynie's equations (56) involved looking at the lifetime of

zinc:
t - 25 um/ ((2.4 + 0.02 802) f), (2.39)

where t - lifetime in years, and
f - fraction of time when wet.

Lipfert (31) proposed a more recent damage function for zinc by

combining data from eight test programs comprising 72 different test sites

in many countries. Some data was estimated but 802 data was available for

all. The equation is consistent with previous field and chamber results and a
separate effect from wet deposition of H+ has been identified. The effect is

larger than would be indicated by the stoichiometric removal of zinc by H+

and also apparently represents the removal of ZnCOa. The H"' effect appears

to accelerate over time. The function, resulting from deposited 802 based on

surface wetting at 85% RH or above, is as follows:

40'
C - [t0'779 + 0.0456 In (H"’)] x [4.244 + (0.547 :I: 0.023»;35 x so2 .-
(0.029 i 0.006) x Cl + (0.0293 : 0.0053) x 14+), (2.40)

C the corrosion rate (either in pm or g),
802 atmospheric concentration, pg/m ,

H+ hydrogen ion concentration, meg/m2 yr,
Cl concentration, mg/m2, and
f35 the fraction time of wetness when the relative humidity exceeds 85%.

These corrosion equations for zinc and/or galvanized steel span a
wide variety of both form and parameters that have been considered. Most

formulas were based on outdoor exposure studies, but a few involved

laboratory experiments. The most frequent parameters are S02 and

time-of-wetness, tw, but other parameters, e.g., temperature, particulate

matter, and nitrogen dioxide, have been used in some cases. These equations
indicate the historical development of equations and lead to the most recent
formulations based on outdoor exposure, laboratory experiments and

theoretical considerations.

3. EVALUATION OF THE HAYNIE-SPENCE DAMAGE FUNCTION

3.1 BUBEALLQEMINES

As part of the NAPAP the Bureau of Mines, US. Department of the f
Interior, is conducting outdoor atmospheric corrosion testing of zinc, l».
galvanized steel, and other materials to evaluate the effects of acid
deposition (32). These studies were begun in 1982 at four sites in the East
representing a range in environmental factors of interest in atmospheric
corrosion processes. These sites are located at: Washington DC, Newcomb,

NY, Chester, NJ, and Research Triangle Park (RTP), NC. A fifth site was
established at Steubenville, OH, in 1985.
Four types of atmospheric corrosion tests being used are:

1. Boldly exposed, two-sided, (all sites).

2. Boldly exposed, single-sided, (NY and DC).

3. Sheltered two-Sided under a transparent fixed polycarbonate

cover, (RTP and OH).

4. Sheltered two-sided where, during periods of rainfall, one set is
sprayed with pH 5.6 water in an amount equivalent to the
precipitation, and the second set remains unsprayed (OH and RTP).

Only the boldly exposed, two-sided tests for zinc and galvanized steel were
used to compare experimental data with calculated damage functions in this
thesis. The code 191 zinc alloy used was hot rolled, but not tempered, and
consisted of .81% Cu, .003% Fe & Pb, and the balance zinc. The hot-dipped

41

42
galvanized steel had a coating weight (for both sides) of .27 kg/m2 and the
coating was 19 pm thick. The 10 x 15 cm zinc panels were 1.6 mm thick;
the galvanized steel was 26 gage. The panels were exposed at 30° to the
horizonal according to ASTM Standard 650-76. Environmental parameters
measured at the exposure sites are shown in Table 1 (13). Table 2 lists the
organizations involved in on site monitoring of these variables (13). From
these test sites comprehensive corrosion and environmental data have been
gathered and are being used by the the Bureau of Mines, US. Department of
the Interior, and the U. S. Environmental Protection Agency (EPA) to
determine mechanisms of corrosion damage and to establish damage
functions which describe the effects of the environment, including

acid deposition, on materials performance.

3.1.1 BUNQEEANALXSIS

For materials with a fast response to pollutant deposition, an
alternative to a site-to-site cross-sectional analysis of data is
correlating the time series of responses with that of deposits. This method
avoids site-to-site collinearity problems, such as problems with pollutants
which tend to have similar spatial patterns where it is difficult to separate
their effects. The easiest way of doing a time series of materials
degradation responses is through the analysis of precipitation run-off
Chemistry, compared with incident precipitation and run-off from an inert
surface (blank). The difficulty of this approach is defining an adequate
surrogate yet inert surface for the blank; stainless steel has been used in
assessing zinc run-off. The run-off is taken from a single large zinc panel

which is boldly exposed (33). The mass balance on the runoff panel

43

 

 

Parameters Reporting format '
Air. Quality:
Continuous: 80,, NO, NO,“ NO” 0, Hourly average
Passive: SO, . Monthly total
Particulate chemistry: mass, size,
composition ' Weekly analysis
Meteorology: .
Weather: wind speed (W8) and wind
direction (WD), temperature ('1‘),
relative humidity (RH) Hourly average
precipitation, solar radiation Daily total .
Sun-taco wetness: time, events ' Monthly total
Rain Chemistry:
pH . Monthly analysis and by event
Cations (11’, Ca”, Mg“, K‘, Na’, NH” Monthly analysis and by event
Anion (1510;, Cl", 803‘, ro:-, 1100;) Monthly analysis and by event '

 

Table 1; '-

Environnentsl Parmters 'ueasured 3:: Materials Exposure Sites (13) -

44

. 3: 0035300: 553.50: 550505.55 5:305:00 0:05:05:qu

 

 

N 0309
>256 1010—000 50...:m 10130.0 5055 10120.0 5255 1010—000 5.55 10102000 505
5.695 550.15 . .
:031095 8552 3:355
.052 «0 :55 20¢ 503.2 3:059:55 .053 ~0 55 .1519 50.0.5 2.05.500
"$350.5 £05
.032 50 50:5 .058 20 50:5 358 «0 59:5 .058 go 5055 .052 20 50.55 305.3 00.....am
m . :030305
3:059:55
.o .8859: 42 .3 828 no .335
85.03 :02 3.09.3. 1.33.2.
9.65.5. . p.313... . oozes-383
a... 31> .550 3:38: :58. 2 3.855.... o... 33......
030 .3002 50.59.34. 15. 9:55 85.00 503.2 .1529 4580.5 .0533
$530.68.}.
50:02 55.03. 5.3.
:030895 :030395 50:03. :030305 5: :030035 :030895
15.555 15.5555 15.5555 .. 3:059:55 15.5555 580:.
33:03.35
.032 «0 59:5 .032 «0 59:5 333 «0 5.55 53 «0 :55 .053 «0 55 108 .3805
50350.1...— 3:05
05...: 1.559555
5.55:4 :03500 :030805
a: 55> 1505055 3:059:55 .35—00 33.5
030 5:02 20 5053.5 #2 20 0:05:05 22 «0 .0135 01:02.5. A0505 3053.30
. . ..b..~§0 5..
.352.
20 #2 .2 05 02 5:03:02

 

45

directly accounts for all the interactions occuring on the panel (33):

Di-Ri-f-Ai-Wi (3.1)

where Wi - wet deposition,
Ai - corrosion film accumulation,
Ri
Di - dry deposition.

- runoff, and

Corrosion damage is determined from weight-loss measurements, analysis
of corrosion film chemistry and from precipitation runoff chemistry.
Covariance between rain chemistry variables and air quality is probably low
enough to avoid problems in developing linear regression models of

corrosion damage (33).

3.1.2 AQIQBALN

The BOM has found the loss of corrosion product from zinc to depend
linearly on both hydrogen ion loading, which is the product of ion
concentration and rain volume divided by the cross-sectional area of the
collection bucket (13), and corrosion product solubility (32) and although
neutral rain does dissolve the film, the solubility will be greater at lower
pH.

The narrow almost neutral range of 6.58 to 6.74 pH values for the
runoff (see Figure 14) indicates that the corrosion film reacted with nearly
all the acidic species present in the rain (34). This leads one to expect that
the rain dissolution of the film is probably not a function of residence time
for the acid rain component. Loss of corrosion product from zinc in runoff

was found to be a function of both film dissolution by rain water (55%) and

4 6
neutralization by hydrogen ion loading (45%) (34).

3.1-3 W

The corrosion film for zinc has been studied by the BOM using TGA
and RGA analysis. The film was found to be composed of ZnCOa-nZn(OH)2
(33). For TGA analysis, assuming the partial pressure of 002 and H20 are
proportional to the molar production of their gases from the film, then the
ratio of 1.5 for H20/002 corresponds with a film composition of
22n003-32n(OH)2 (33).

All galvanized steel panels as well as zinc panels at the DC site
appear to have large amounts of ZnO in the corrosion film, as well ZnCOs and
Zn(OH)2. For galvanized steel the presence of ZnO is probably because of

the stabilization of an initial passive ZnO film created by Cr3"' which

persists over parts of the surface in exposures up to 3 years because of the

chromate treatment. For Zn at the DC site, the higher NOx levels may have

contributed to the presence of ZnO in the zinc film there (34).

3.2 W

Various damage functions have been proposed by Haynie, Spence, and
coworkers over a number of years. Their most recent, intended to account
for wet and dry deposition on zinc and galvanized steel, was developed by
Haynie and Spence in October 1987 (30). What distinguishes this equation

from most others are factors which account for rain and dew run-off

4 7
analysis of solutions and surface film components. Experimental results
that contributed to the formation of the damage function are that rain
acidity reacts stoichiometrically with the zinc coating, the corrosion film

of basic zinc carbonate is soluble in clean rain, and the confirmation (30)

that $02 reacts stoichiometrically with zinc during periods of wetness.
These findings have been incorporated in the damage function:

C(t) - 1(0) + K9302] fw + R {s + 0.5[H"‘]}, . (3.2)

where
C(t) total amount of corrosion at time t (moles/cmz),
T(O) amount of zinc in the passive film (moles/cmz),
Kg deposition velocity for $02 (cm/sec),

average atmospheric concentration of $02 (moles/cm3),

. SO2
R total amount of rain (cm),
tw time of wetness (sec),

8

solubility of the zinc carbonate (moles/cm3), and
H+ incident precipitation hydrogen ion concentration (moles/cm3).

An analysis of the equation was done to test the applicability of the
damage function and to evaluate the theoretical and experimental basis for

the equation. The equation then was compared with experimental data.

3.2.1 W

Haynie's use of the term T(O) and other factors may be understood by
review of the Edney-Stiles model (36).
Edney defines total corrosion as :

C(t) - T(t) + Bt, (3.3)

where C(t) - the corrosion at exposure time t. The corrosion model is

48
obtained by adding the insoluble corrosion product layer at time t, T(t), to
the contribution of the soluble corrosion products which is given as Bt. The

term T(t) is given as:

T(t) - To {1-exp[-(T(t) + [so/T01}, (3.4)
where To :- f/B, (3.5)

f - fractional time of wetness,
- diffusion coefficient,
A - a constant that converts mol/cm2 to cm,
u- moles of zinc corroded per mole of x deposited,
[x] - gas phase concentration in mol/cm3 for the ith ion,
Vd - dry deposition velocity in cm/s for the ith ion,

l - the amount of rain per unit time,

H+ - the rain concentration on hydrogen ions, and
S - solubility of ZnCOa.

Thus Edney-Stiles write three terms:

Terms: (1) (2) (3)

C(t) - T(t) + [A {f 2 pivdi [xi] + I(.5H+ + S)}] t. (3.7)
compared with Haynie's model,
Terms: (1) (2) (3)

C(t) - T(O) + Kg[802] tw + R {s + 0.50-m} (3.8)

Note that terms one and three are nearly the same with (l)x(t) - R, and

term two is roughly the same with {f 2 “ini [Xi] - Kg[802] tw . In other
words, Haynie's model is the same as Edney's if only 802 pollutants are

considered when “i - 1, as it would for 802 reacting with Zn since Vdi - Kg

49

ft - tw, and xi - $02. The same assumptions should be used for Haynie's

model as for Edney-Stiles: 1) clean dew is ignored and 2) terms of wet
deposition are additive. The first assumption is reasonable since the
contribution would be small for clean dew. There is no clear basis,
however, for the second. It may be noted that Edney's model does not

suggest addition of a particulate term as Haynie's did.
3.2.1.1 FORMULATION OF EDNEY-STILES DAMAGE FUNCTION

Edney and Stiles began with an equation stating that the change in
thickness of the insoluble corrosion product over time is equal to some
diffusion coefficient divided by the film thickness multiplied by the
fractional time of wetness, assuming corrosion only occurs when the panels
are wet and that clean air conditions are present.

dT/dt - /T f (cm/s), (3.9)
T - thickness of film,

f - fractional time of wetness,

- diffusion coefficient, and
t - total exposure time.

I TdT -I fdt,
Integrating equation (5) yields: T(t) - (2 ft)-5. (3.10)

Experimental results agreed with a model in which acidic
components in rain and dew react with insoluble corrosion products (56).
Assuming that all deposited compounds react stoichiometrically with the
film compound, 8 is introduced into Equation 9.

dT/dt - IT (f) -B, (3.11)
Solving Equation 11 yields

5 O
T(t) - To {1 -exp[-(T(t) + Bt)/To]}, (3.12)

where To - f/B, the steady state thickness of the film. T(t) is not defined

but is assumed to be the thickness of the insoluble corrosion product film at
time t. The final corrosion model is obtained by adding the insoluble
corrosion product layer at time t, T(t), to the contribution of the soluble
corrosion products which is given as Bt.

C(t) - T(t) + fit. (as shown above). (3.13)
Assumptions are as stated above.

3.2.2 WW

It has been shown that the Haynie-Spence equation is similar to the
Edney-Stiles model. Haynie also supports his equation with thermodynamic
arguments and laboratory chamber studies, which allow for the separation
of the effects of different pollutants. The chamber's simulation of
turbulence and precipitation effects, however, is difficult to correlate
with natural environments.

C(t) - T(O) + Kg[SO2] M + R {S + 0.5[H+]} (3.14)

1) T(O) is from Edney's model and from the relation :
Total Zn corrosion - amount in the film [T(O)] + amount in the runoff
T(O), the zinc ions in the film, are found through atomic absorption analysis

of the film upon its removal.

2) Kg[SOZ] tw is from lab studies 1 and 2 and from many outdoor

exposure studies (12, 15, 26, 36, 44) with the addition of K9, the deposition

velocity, which is a function of wind velocity, and provides some statistical

51
improvement to $02 dose (7). K9 is from Haynie's outdoor studies (7) and is
from stoichiometry, since 1 mole of 802 should dissolve 1 mole of Zn'H'.

Dry deposition is acceptable if the conversion equation from wind velocity
to deposition velocity is accurate and if the conversion from RH to f,
fractional time-of-wetness, is acceptable.

3) RS, the clean rain term, is from lab studies 2, 3A, and 38. S, the

solubility of ZnCOa, should instead be the solubility of the corrosion film,
experimentally found to be ZZnCO3 * 32n(OH)2.

4) 0.5RH" is from lab experiments 2 and 3A and from stoichiometry

since 1 mole of Zn“ should dissolve for 2 moles of H+.
3.2.2.1 LABORATORY EXPERIMENTS

LAB EXPERIMENT 1 (57)

802 was increased from 3 to 650 ppb. Zn” in the dew was analyzed.
Finding: 1. Zn‘H' has a 1 :1 correlation with $02 flux. A slope of 1.06

was found if all deposited $02 appeared as $03" and $04“ ions in the

condensate. See Figure 9.

LAB EXPERIMENT 2 (57)
At 3, 30 and 45 ppb 802, samples were sprayed with NH4HSO4 at

various pH values. Rain collections were analyzed.
Finding: 1. pH effects the Zn“ flux. Zinc flux represents the number
of nmols of Zn dissolved into the sprayed solution per cm2 of panel per cm

of spray, giving nmol/ml units. For 30 ppb, a factor of 2 increase in Zn”

52

 

 

110-
100'-l
.90‘
so-
an ' 1.06 Fax. -0.“
R2 ' 0.995
A 70"
E ‘5-
"E
c
s o
u. .
50"
40-4 -
30‘
20-r
IO-r C)
C)
‘ I I I I fir I I T
10 20 30 4O 50 60 70 80
F . rune!)
w" (cm’ -h
Figure 9

Zinc Flux as a Function of 80x Flux in Condensation (57).

53
from 5.6 to 4.0 pH was exhibited and a factor of 12 increase from 5.6 to 3.0

pH. Figure 10 shows the Zn flux vs. H+ flux for 3, 30, and 45 ppb 802. The

slope and intercept were observed to decrease as the $02 exposure of the

panels increased. This decline is most probably due to the increased

interaction of [H"'] with $02 decreasing [H+] interaction with [Zn"’"'].

LAB EXPERIMENT 3A (58)
After 10 weeks boldly exposed panels were removed. Films of
sulfuric, nitric, and formic acid at pH levels of 3,4, and 5 were sprayed on L
the surfaces. Runoff was analyzed.
Findings: 1. Zn“ correlates with H"' after 1000 sec. residence time
fitting the relation:
Zn” - .484H"‘ + 58.7 uM/l (3.15)
with R - .858 at the 98% confidence interval. See Figure 11. This supports
the chemical model that two hydrogen ions react with one zinc molecule in
solution.
2. Clean rain causes some Zn” in the runoff due to the intercept of
58.7 uM/l.
3. Residence time may not appear to be important for acid rain
components. The corrosion film reacted with practically all the acidic
species. The mean value for the final pH was 6.73 +/- .20. at the 98%
confidence interval. The possibility that residence time is not a factor for
acid rain is confirmed by the narrow range of pH values 6.58 to 6.74
observed by the Bureau of Mines for rain runoff from zinc panels at the DC.

site (33). Lab study assumptions: 1. 1000 seconds is a reasonable residence

time. 2. ZnCO3 is the film. Question: Why didn't the anions correlate?

120-

110-

100-

40-

30"

54

[5021' 3 ppb
Pg,- 0.097 53. + 19.5
R’ - 0.978

[5021' 30 ppb
Ff" - 0.059 F3. + 4.03
a: - 0.999

(so,1- 45 ppb
52,, - 0.021 F:++ 3.95
a? - 0.975

 

 

 

fl 1 f i T 1 1 r
200 400 600 800 1000 1200 1400 1600
FR. nmol )
'1 cmlocrnr

Figure 10
Zinc Flux as a Function of Incident Hydrogen Ion Flux for SO

Concentrations of 3, 30, and 45 ppb (57). 2

[Zn1 ’] Equilibrium (pM)

55

 

 

500-
O
400 '
" ' o
300-
0
~ 0
200- . .
Regressron Analysrs
-[Zn2’] = 0.454 [H‘] + 58.7
l’=0.858
100-
O
O
.0.
a: O
1 1 1 1 1 1 1 ~ 1 1
100 200 300 400 500 600 700 800 900
[H‘llncident (pM) ' .
Figure 11

Zinc Concentration After 1000 Seconds as a Function of Hydrogen Ion
Concentration (58) .

56

LAB EXPERIMENT 3B (58)
As in 3A preexposed panels were used and sprayed with 5.6 pH H20

(deionized water). [Zn‘H'] was analyzed vs. residence time of H20 on the

samples.

1. Findings: An equilibrium value of [Zn"""] with 5.6 pH water was
found to be 70 (M.

2. Zn""" ‘ runoff is a function of residence time for clean rain fitting
the equation which is sometimes used to represent an approach to
equilibrium: [2n++] - a [1 43‘4””) , (3.16)
[Zn'H'] is the concentration in mM/l at time t (sec),

a is the solubility coefficient, and b is a time coefficient. b - 206 +/- 92

sec. and a - 70 +/- 18 uM/l at the 98% confidence interval with R - .878. The

value for a is consistent with 80 uM/I for ZnCOa solubility in clean water

(59). See Figure 12. Assumptions: The film is ZnCO3.

3.2.3 Ox : On ‘0: I. .AA.A: 0L A I. ." :u kl‘ I; A

SULFUR DIOXIDE
To convert the air quality data from the measured ppm values to
moles/cm2 the following conversions were used:
1 ppm - 2600 ug/m3,
802 has 64 gm/mol S - 32, O - 16,

[2.6 x 10'9 gm/cm3 ] / I 64 gin/mol] - 4.0525 x 10'11 mol/cm3,
1 ppm - 4.0625 x 10'11 mol/cm3.

[Zn2 *1 (11M)

57

 

 

 

 

IOU-l
90-
80—
C)
40
<3
° (3
Regression Fit .
[an ‘] = 59.9 (1— @1006)
r = 0.878
(a I I I I I
100 500 1000 1500 I800
' Residence Time (s)
Figure 12

Zinc Concentration as a Function of Residence Time for Deionized
Water.(58),

5 8
(Other conversions can readily be reached for other 802 units; see Section

2.2.1)

RH TO TlME-OF-WETNESS

To convert from RH to tW the same conversions were used as for the

review of past functions; see Section 2.2.1.

Haynie found the relation between fractional time-of-wetness and RH
from Guttman's data (12). Guttman plotted the probable time-of-wetness
as a function of relative humidity for both groundward and skyward surfaces
of exposed metal panels. Assuming the average probability for the two
surfaces represents the fraction of the total time panels are wet, Haynie fit
the data to the relationship

f_ e (4.04 -[404/RI-I]). (317)

Guttman's data is for rolled zinc at Birchbank, 8.0. from 1959-62
(Figure 13). To evaluate the equation, points along the Guttman curve were
found to fit several formulas with good correlation. An equation close to
Haynie's formula was found when In (y) vs. 1/(x) were plotted, where y -

probable time of wetness and x - relative humidity. The formula arrived at

was:

In f - 8.649 - 404.764/RH, (3.18)
with R-.97. This converts to

f_ e(8.65—404.76/RH), (3.19)

which is close to Haynie's formula. The equation f - -167.609 + 2.629 (RH)
also fit with correlation coefficient R - 0.99 and the equation f- 0.220
"100'027 (RH) fit with R - 0.98. These three equations are shown in Figure

14 a—c. Although Haynie's equation for converting RH to tw can not be used

59

 

 

 

539

 

 

 

u
0

GROUNDWARO SURFACE / /

 

 

5
O

u
0

 

v.74»

\_

 

' 3

 

 

‘PROBABLE TIME OF wefutssm
8 3

_//7

 

 

 

 

 

 

 

 

f // . \
SKY'NARD SURFACE

 

 

 

io --- - . .
. ‘ . ~-I ... . /... ‘ d .
o ' 411‘ . '
4o ,_ so so _ 70 no 90 too
RELATIVE HUMIDITY (%I
Figure 13.

Probable Time-of—Wetness vs . Relative Hum‘dity, Z. (12) .

Figure 111 Possible Fitted Solutions to Guttman's Data (a,b,c).

 

 

a) 5 y =- 8.649 - 404.764x R =- 0.97
In (tw) vs. 1IRH
2 4 '
E
3 .-
I

 

 

2
0.009 0.011 0.019 0.015
mm 0

Guttman's data, simple fit
.bi1 y - - 167.609 + 2.629x n . 0.99

 

88

E

 

 

I n I I n I

 

60
4o-
20.
0 .‘ .
60 70 BORHQO 100110

Guttman's data, exponential fit
C) 11117 0.220 * 10*(0.027x) R . 0.98

 

Tw
3388

 

 

 

I n _I n I n I n

O 70 80 90 100110
RI'I

 

05°

6 1
with great confidence it is the only relation that will consider the relative
humidity at all times. The relation [RH > 86.5%] has been used to

approximate time-of-wetness on the panel, but it would not consider times

when the 802 may be high, yet the RH is below 86.5%. Relative humidity

weekly average values for all sites were between 60 and 75% RH, so this

relation would yield very low tW and eliminate many weeks and months of

802 accumulation. Unfortunately time—of—wetness was not measured and so .‘1

some relation between RH and time-of-wetness must be used.

The Kg Factor

Deposition velocities can be calculated on an hourly basis based on

windspeed data. The equation used for deposition velocity is calculated
from boundary layer theory beginning with the relation (42):
V+ - 8.5 + 2.5 ln(Z/e), (3.20)

where V"’ - dimensionless velocity,
2 - measuring height, nominally that of test rack, and
e - ground roughness height.

The test rack height, 2, has been reported to be about 3 m from the
base (42). This height does not seem reasonable since racks on the ground
usually are less than 1 m from the ground, nevertheless we will use this
value. The meteorological towers are either 10 or 30 m from the ground.
Assuming an average ground roughness height, e, of 0.1 m, the rack height
windspeed then would be 0.75 the windspeed at 30 meters or 0.85 the
windspeed at 10 meters. e.g. 8.5 + 2.5ln(3/.1) - 17.0, 8.5 + 2.5ln(10/.1) -
20.0 hence 17/20 -.85.

6 2
The deposition velocity, u, then is (42):
u - v’2N , (3.21)
where v‘ - friction velocity which - v x (i/2)1’2,
V - average windspeed, and
f - friction factor, which from boundary layer theory for

smooth flat plates is given by the equation
1. 0.03/(REL)1’7, for which REL -vav,

where L- length of surface over which the air flows, assumed to be the
geometric mean of the panel dimensions (- 12.45 cm) and v- kinematic f.
velocity of air (- 0.15 cm2/s). 1
Thus this information yeilds (42):

u - 0.35 V10'86 or u - 0.31 V3096, (3.22)

where u is the deposition velocity (cm/s),
v10 is for towers at 10 m (m/s), and

v30 is for towers at 30 m (rer).

The S Factor
The solubility of ZnCO3, 80 umol/l (59), converted to mol/cm3 is
8 x 10‘8 mol/cm3.

Hydrogen ion concentration, H"'
Hydrogen ion concentration data were reported by the BOM in ug/l;

converting to mol/cm3 these units are 10'9 mol/cm3.

The Term T(O)

Film chemistry data were reported as ug/cmz; converting yields :
[1 gm/1 x 106 pg] [1 /65.4 gm/mol(Zn)] - 1.529 x 10'8 mchmZ.

6 3
Corrosion, C
Corrosion data were reported as mg/dmz; expressed as mol/cmz;

these are 1mg/dm2 [1dr'rl/10cm]2 [1/65.4 gm/mol(Zn)] .. 1.529 x 10'7
mol/cm2.

3.3.4 DAIAANALISJS

Hourly environmental data were recieved from the EPA. Where ever
values were missing, the value 99.99 or 9.99 was listed. Negative values
frequently wee listed, where apparently the instrument "zero' had drifted.
Both kinds of discrepant data were omitted in the present analysis. The EPA
has attempted to correct or 'infill' these discrepancies using a variety of
techniques. We also have done some 'infilling' using different methods in an
effort to estimate the significance of such data gaps. Results are discussed
later.

The parameters provided by the EPA needed to evaluate equation (3.2)
are wind speed, relative humidity, sulfur dioxide concentration, and rain
accumulation. The hourly values were summed to weekly and daily totals
using Fortran programs (Appendix B). The four parameters of interest,
together with dew point and temperature, which are directly linked to RH
(see 2.1.2.1) were plotted vs. time for each of the five NAPAP sites. Trends
for relative humidity, sulfur dioxide concentration, dew point and
temperature are shown in Figures 15-18. Rain and wind speed data (Figures

19 and 20) show a more random pattern.

The second term, (thwSOZ), of equation (3.2) which takes into

54

8.5 o<c<z m mo... on .85 B P 2% 28... $32

n_._.m

.Nm .ma. ..ch .01 .mm .Nm .mN ..vN .ON .2. .NP .m cs .0
_t_.e._.c._.c.__.c._.._

 

.owlmomH «on copy» m<m<z w

.LPLL .o
T

1.0?

1'

IdN

r.

1.0m.

...oe

T

1.0m

To

... .. . ...8

.l.o\.
r
”tom
1.0m
when:
id:
T
1.09

1.2.4

 

Eon .u gnu—van 253.33 ouch—22
ma ensues

986I-7286I 80:! MIGIV‘IRH BAIiV'IZ-IH ESVEEIAV

55

 

6x55 “.3:

 

 

cod

 

..
. is, ...
...... .... m . a x .. a s ‘ 1ND O
1.. «04. a; hit m . fia(~‘ .‘s cot-u ‘ . m ‘M. .... I. . N04; 7. ~ \I
...... r ..N «c ...... ..s. ,s .2
Z .. ...... ... I
...... m .... u .u. "r
. . .. .5. n
on ..T» ... m m
..z nlm «a a” .u
10 «to m m
>2 Ole .u .
02 I : «

 

 

 

[v0.0
Bum o<c<z m
MQXOE leij

$3183 33m $.22 m no.0. .EE .3 H58 09.03. 3 drama

(Wald) zos

.Nm .m.v :11 .O.v .mm .NM .mN...vN .ON 6.. .N— ..m ..v. .0
_ r _ . L . r b _ P L P — . L L F . _ . L _

mmbm n_<n_<z n mombm .95 I F .23.. 201m mxmma

 

 

 

.L . L 3W3 .ONII
. T
I.O_.I
1.0
Id?
33”.. fl
4%- I.ON
. o 1
$5 .6 cl. ...om
77‘ arlLu .. 1
>2 BIB I.O.V
on ole . r .
nth. I
rdm

...eearsz 68533 2.22 _ ... use .3 .633 see
5H ounwflm

986L-‘ZB6IV‘IOBJ (0) .LNIOcI M30

67

mmtm Q<n_<z m mom on .000 I F .23. 20mm mxmmz,

.Nm .ch :14 .0.v .00 .Nm .mN ..VN .0N .0? .NF .0 cu .0
Fc_t_._._._.cuc._ch_Tp+Ft.mFl

bl

 

1.2..

. ...ml
y-........o
. Ex...
....2
...?
...om
rem
1.0...
the
1.9.
{he

 

 

986L~9286l W083 (0) 21290075351131 '3Av

637.33 you «33. as n you .0 0 63309.89 omuuo><
mH enough

mmEm c<c<z 0 mo...» on .80 .. . .23. :9... QB:

 

 

 

 

 

 

 

 

 

 

 

.Nrm 0.04 £100.94 0.0 .Nn 0..N. ..v_N .0_N ..0— N.« ..0 .w. ....0.
are!) n’P .. n’ n .O
.....ldm
z.
Av . j.O.V
. .. , ._ 1.8
. . ‘ r
c _ . 1.0m
1.00?
. f
A ... .. 1..nVNn—
..z elm , , F
>2 ole .. .
. nrrm. mYliu . Ll “was?

 

63782 new .3.3 $526 com .5. 5633233.. $6.33
ma vacuum

(WW) 986I-ZBSI woes NOIiViIdIOI—léicl ’EIAV

.N0 .00 cats .00. .00 .Nm. .0N .0N .0N .0— .NF .0 cs .0
F...........F.L.. .L

mme n_<n_<z 0 m0... on .000 I F .23.. 201... 0x003

F_ r

_.R. .O

 

YE
T:
1.3
TM;
T:
in.

:2

00
>2
nth.

IIIII

 

.eoenlnoed now cause m<m<z n uou .e\a .eouem eons omeuu><
cu ensues

(s/w) BBBI—SIBBL woes CIEIEIdS CININI 'EIAV

70

account dry deposition, was considered initially, since adequate data were
available for it's evaluation. Weekly and daily sums for the years that data

were recieved from the five NAPAP sites are compared in Table 5. The

weekly sums for each term Kg,Tw, and 802 then were multiplied together

to obtain the term KgTwSOZ. For the daily values a program summed the

hourly values, filled in the average for the day where there were missing

data, and then multiplied the three terms Kg,Tw, and $02 together for each

day. The values show a relatively small error between the two summations
even when the mean daily value was inserted for missing points for daily
totals; missing points were not filled in for the weekly values.

Rain chemistry and film chemistry data were available for the RTP
site from the BOM, and the total equation was evaluated. The term T(O), the
amount of zinc in the film, already is a sum for the exposure period, i.e.,
one, three, or twelve months. It was necessary to sum the term .5RH"'
monthly since hydrogen ion concentration data were reported only on a
monme basis. The rain term therefore was totaled monthly and the two
terms then were multiplied to give the monthly totals. The term RS was
calculated using the amount of rain per month multiplied by the solubility of
zinc carbonate, 80 umon3 (59) integrated over monthly values.

The amount of data available at this time limited calculations; hence,
seventeen values were determined. The five one-month values were for
January 1983, March 1983, June 1983, September 1983 and January 1984.
After that time no further one-month values for T(0) were reported.

Eight three-month values were determined for the months of initial

exposure including the five one-month dates as well as for March 1984, June

71

IABLEfi

Sum of mean daily Kg, tw, and $02, and products compared with sum

of mean weekly means Kg, tw, and $02, and products. Data for years where
data were available for all five NAPAP sites.

TEST ' DAILY SUM WEEKLY SUM RELATIVE DIFFERENCE

SITE .. x 1.0‘5- mol/cm? x 10"5 mchm2 . (Weekly-Dailyyweekly
M .3032 .3049 0.56%

NJ 1.600 1.3632 14.60%
00 1.245 1.1888 4.53%

O-l .6370 .7622 6.94%
m .1576 .1314 16.62%

72

1984, and September 1984. Four twelve-month values were calculated for
exposures beginning the same as the one-month dates excluding January
1984. The rain chemistry data recieved was for 1983 and 1984, hence the

limitation for the three- and twelve-month values.

4. RESULTS/CONCLUSIONS

Michigan State University's role in the evaluation of materials
degradation due to acid deposition has been to evaluate damage functions
for zinc and galvanized steel and to compare predicted values with

experimental results.

Data analysis showed yearly trend cycles for $02, temperature, dew

point and relative humidity; none were observed for rain or wind. This
could add to a much greater error where average accumulated rain and wind

values are inserted for missing values. Comparison of weekly and daily

totals for the second term , thwSOZ, suggests that different methods used

to evaluate the data will give variation in the results. In most cases, these
will be small differences as is indicated by the comparison of weekly and
daily data. The values in Table 5 suggest that weekly values may be
adequate for estimating daily values, and the trends suggest that short data
gaps can be reasonably approximated. Where large gaps occur, e.g., seasonal
gaps, summer/holiday vacations, trends recorded in prior (or subsequent)
years possibly can be used to interpolate.

Preliminary results of the evaluation of the equation proposed by Haynie
and Spence have indicated problems with the damage function.

Results for RTP, NC indicate that calculated values anticipate higher
corrosion damage than realized in field studies (Figures 21 and 22).

For one-month exposures, three of the five calculated points are larger

73

74

«medians. .oz .sues oflwcehua couscous ue

dawN HOm .0loH x Nao\Hoa .nmoq uswuuz vouaasoamo .m> Aconcagumcxa LN mpawau

+10.+mm+E~ommx+Sru€o

 

00.243045
om o. o
_ s — p o
«:3 Ezo: . .
Ea Ezoz n c _ _
(Ea Ezo: a. o 4 o
‘ T
. 4
4 3
X
. id
as
. 4 o. m.
. m
N
o .v.
. .II
0
.oE: ho uotoa. L.
. ow

 

Show 05 .2 03098 mac—on ...2 bo> mount team

Glue. x maniacs oz ...E
mmS :60; 8.233346 .m> ...fizmzimoxm loz_N

75
ZINC CORROSION

 

 

 

 

 

 

O
H MEASURED
”o CALCULATED.
F 15+
E
O
>
O
33,
to “r- 0
'T'
ulJ 10-
O
X. .4-
Z
9.
8 54
8 RESEARCH TRIANGLE
PARK, NC
0 "" l I
0 ~ 5 10
TIME (MONTHS)
Figure 22

Corrosim, cpl/c132 vs. Time data range for Research Triangle Park, NC,
1982-1985. ombined with calculated data points for 1983-1983.

76

than experimental weight loss values resulting from the T(O) term alone
(Figure 23). Apparently there is error either in treatment of the film
analysis or in the weight loss measurements. This error would not be
reduced by changing the equation.

The clean rain term is significantly higher than the acid rain term and/or
the dry deposition term (Figure 24). The damage function, if correct, would
therefore not support the major pollutant effect on materials. Others have
shown pollutants do increase the corrosion rates for zinc, hence one must
conclude that the equation does not accurately model the environmental
damage. Reasons for this could be that each rain event's significance is

overestimated and that the clean rain residence time is unaccounted for.

Another probable reason is that the solubility of Zn003 used in the equation

is for equilibrium conditions not probable rain conditions.
Conclusions reached are:

1. The exhibited trends in sulfur dioxide, and relative humidity along with
comparison of daily vs. weekly values suggest that gaps might be filled
by daily or weekly averages.

2. Preliminary results show the damage function is not in good

agreement with experimental data.

77

ON

8.3336

 

B$44
B

X

.coauodsm «mason concomiowahum on» um
oawcwwua noumomom uo onfiN now

 

 

o
0..P o
m new a; a. a
mm
,8 .mm m K *
B d
4
‘ 3%
x44 4
Wm
....
-m
4
cm
. Emommx e -
. +55. 4
mm m
at 4
.59 x
r3
E3 53 e333. $3-82 .oz :38 2 3:»:

e

o” x «Boxaoa .oooq uawaoa vouudooaoo

TVLNBWIHBdXH

78

83:38
o _.

. .oauou noauuuomon mun and .dwnm

vau4 .afiqm cacao any“ oaoauaaauuaoo unauaoaeaa camalnmma coz .xuam

oawoowua nouoouom no onus you

w

-3 x «8:3 .33 330: 33338

 

 

rm”
mm a
.38me >
+57... 3
[five
«N gunmen

'lVlNEIWIHHdXEl

5. DISCUSSION/SUGGESTIONS

The equation should be evaluated against experimental data from all test
sites. The monthly, weekly, and daily calculations should be carried out
where possible, i.e. ion concentration is given only monthly.

The evaluation of time-of-wetness based on relative humidity should be
tested further by researchers at test sites to compare the actual
time—of-wetness with relative humidity. There may be no easy correlation
or perhaps RH 2 86.5% must be used instead of equation (3.17).

Equation (3.2) has no practical use to corrosion researchers in the field
as to the possibilities of an estimate of possible damage based on local
environmental factors because the term T(O), the amount of zinc in the film,
is difficult to evaluate. Although the use of the equation is to approximate
a cost estimate of materials damage, it would be helpful if the equation
could become of common utility. The equation could then also be tested
more thoroughly and updated based on field data.

The equations suggested by the EPA have significant problems because
each term involves significant estimation and error, e.g. deposition velocity
from wind speed, time-of-wetness from relative humidity, missing data
fill-ins. As more and more terms are introduced further error results. It

may turn out that the best evaluation is the simplest, i.e. PACER LIME

program (30), or an equation based solely on variables such as pH and $02,

and that these methods may approximate the corrosion weight loss

79

satisfactorily.

LIST OF REFERENCES

'L_

81

LIST OF REFERENCES

Corrosion, Metal/Environment Reactions, Vol. 1, Shreir,L. L., Ed.,
Butterworths, London, pp. 150-162.

Brown, P. W., and Masters, L. W., "Factors Affecting the Corrosion of
Metals in the Atmosphere," Atmospheric Corrosion, Ailor, W. H., Ed., The
Electrochemical Society, John Wiley and Sons, New York, 1982,

pp. 31 -49.

Haynie, F.H., "Economic Assessment of Pollution-Related Corrosion
Damage," Atmospheric Corrosion, Ailor, W. H., Ed., The Electrochemical
Society, John Wiley and Sons. N.Y., 1982, pp. 3-17.

Mikhailovsky, Y. N., “Theoretical and Engineering Principles of
Atmospheric Corrosion of Metals,” Atmospheric Corrosion, Ailor, W. H.,
Ed., The Electrochemical Society, John Wiley and Sons, N.Y., 1982,
pp. 85-1 04.

Guttman, H, “Atmospheric and Weather Factors in Corrosion Testing,"
Atmospheric Corrosion, Ailor, W. H., Ed., The Electrochemical Society,
John Wiley and Sons, N.Y., 1982, pp. 51-68.

. Legault, RA. and Pearson, V.P., ”Kinetics of the Atmospheric Corrosion

of Galvanized Steel,” Atmospheric Factors Affecting the
Corrosion of Engineering Metals, ASTM STP 646, Coburn, S. K.,
Ed., American Society for Testing and Materials, 1978, pp. 83-96.

Haynie, F.H., "Evaluation of the Effects of Microclimate Differences on
Corrosion,” Atmospheric Corrosion of Metals, ASTM STP 767, Dean, Jr.,
8. W., and Rhea, E. C., Eds., American Society for Testing and Materials,
1982, pp. 286-308.

Haynie, F.H. and Upham, J.B., ”Correlation Between Corrosion Behavior of

82

Steel and Atmospheric Pollution Data,” Corrosion in Natural
Environments, ASTM STP 558, American Society for Testing and
Materials, 1974, pp. 33-51.

9. Guttman, H., and Sereda, P.J., “Measurement of Atmospheric
Factors Affecting the Corrosion of Metals,“ Metal Corrosion in the
Atmosphere, ASTM STP 435, American Society for Testing and
Materials, 1968, pp. 326-359.

10. Dunbar, S. R., “Effect of 1 Percent Copper Addition on the Atmospheric
Corrosion of Rolled Zinc," Metal Corrosion in the Atmosphere, ASTM STP
435, American Society for Testing and Materials, 1968, pp. 308-325.

11. Baker, E. A. and Lee, T. S. ”Calibration of Atmospheric Corrosion Test
Sites,‘ Atmospheric Corrosion of Metals, ASTM STP 767, Dean, Jr., S. W.,
and’Rhea, E. C., Eds., American Society for Testing and Materials,

1982, pp. 250-266.

12. Guttman, H., “Effects of Atmospheric Factors on the Corrosion of
Rolled ch,‘ Metal Corrosion in the Atmosphere, ASTM STP 435,
American Society for Testing and Materials, 1968, pp. 223-239.

13. Flinn, D.R. , Cramer, S. D. , Carter, J. P., and Spence, J. W., “Field
Exposure Study for determining the Effects of Acid Deposition on the
Corrosion and Deterioration of Materials— Description of program and

preliminary results,” W, 1985, pp. 147-175.

14. Dunbar, S. R., and Showak, W., “Effect of 1 Percent Copper Addition on
Atmospheric Corrosion of Rolled Zinc after 20 Years' Exposure,"
Atmospheric Corrosion of Metals, ASTM STP 767, Dean, Jr., S. W.,
and Rhea, E. C., Eds., American Society for Testing and Materials, 1982,
pp. 135-1 62.

15. Haynie, F.H., Spence, J.W., and Upham, J.B., ”Effects of Air Pollutants on
Weathering Steel and Galvanized Steel: A Chamber Study,“ Atmospheric
Factors Affecting the Corrosion of Engineering Metals, ASTM STP 646,
Koburn, S. K., Ed., American Society for Testing and Materials, 1978,
pp. 30-47.

16. Haynie, F.H., and Upham, J.B., ”Effects of Atmospheric Pollutants on

8 3
Corrosion Behavior of Steels,“ W, 10, 1971, pp. 18-21.

17. Mansfeld, F., Ed., Corrosion Mechanisms, Marcel Dekker, New York, 1987.
pp. 21 1-265.

18. "Acid Precipitation Act of 1980," United States Statutes at Large , 94,
pp. 770-775.

19. Mattsson, E., “The Atmospheric Corrosion Properties of Some Common
Structural Metals - A Comparative Study," W, 7, 1982,
pp. 9-1 9.

20. Ellis, 0. B., "Effect of Weather on the Initial Corrosion Rate of Sheet
Zinc,“ Wings. 49. pp. 152-171-

21. Dunbar, S. R. and Showak, W., Atmospheric Corrosion of Zinc and Its
Alloys, Atmospheric Con-osion, Ailor, W. H., Ed., The Electrochemical
Society, John Wiley and Sons, N.Y., 1982, pp. 529-552.

22. Stroud, E. G., “The Quantitiative Removal of Corrosion Products from
Zinc,“ J._AppL_Qh_em._1_, 1951, pp. 93-95.

23. Copson, H. R., ”Factors of Importance in the Atmospheric Corrosion
Testing of Low-Alloy Steels,“ Wm, 48, pp. 591 -606.

24. Anderson, E. A., "The Atmospheric Corrosion of Rolled Zinc,” ASIM
W. PP-127'134-

25. Milhailovski, Y. N. and Sokolov, N. A., “Modelling Atmosphere Corrosion of
Metals in an Atmospheric Testing Unit and Artificial-Climate
Chambers,” W 5, 1982, pp. 531-536.

26. Haynie, F.H. and Upham, J.B., "Effects of Atmospheric Sulfur Dioxide on
the Corrosion of Zinc,” W 1970. Pp. 35-40.

27. Dean, S. W., ”Atmospheric corrosion after 80 years of study,”
WM 987. pp- 9'11-

28. Martin, H. 0., "Acid Rain: Impacts on the Natural and Human
Environment,” W 1, 1982, pp. 36-39.

3O

84

. Baedecker, P., Chairman of Task Group Vll, 1987 NAPAP Annual Report

Draft, unpublished.

. Summitt, R., and Fink, F. T., "PACER LIME: An Environmental Corrosion

Severity Classification System," AFWAL-TR-80-4102 Part I, USAF
Wright Aeronautical Laboritories, Wright-Patterson AFB, OH, 1980.

31. Lipfert, F. W., ”Effects of Acidic Deposition on the Atmospheric

Deterioration of Materials,” W 7,1987, pp. 12-19.

32. Carter, J. P., Linstrom, P. J., Flinn, D. R., and Cramer, S. D., "The Effects

33

34.

35.

36

37.

of Sheltering and Orientation on the Atmospheric Corrosion of
Structural Metals,“ W, 7,1987, pp. 25-32.

. Cramer, S. D., Linstrom, P. J., Carter, J. P., Flinn, D. R., “Environmental

Effects in the Atmosphere Corrosion of Zinc,” The Degradation of Metals
in the Atmosphere, ASTM STP 965, Dean, S. and Lee, L. S., Eds., American
Society for Testing and Materials, Philadelphia, 1987, pp. 229-247.

Flinn, D. R., Cramer, S. 0., Carter, J. P., Hunrvitz, D.M., and Linstrom, P.
J., ”Environmental Effects on Metallic Corrosion Products formed in
Short-Tenn Atmospheric Exposures,“ Materials Degradation Caused
by Acid Rain, Baboian, R., Ed., American Chemical Society,
Washington DC, 1985, pp. 119-151.

Spence, J. W., Haynie, F.H., Edney, E. O., and Stiles, D. C., "A Field
Experiment to Partition the Effects of Dry and Wet Deposition on
Metallic Materials," Materials Degradation Caused by Acid Rain,
Baboian, R., Ed., American Chemical Society, Washington DC, 1985,
pp. 194-199.

. Edney, E.O., Stiles, D.C., Spence, J.W., Haynie, F.H., and Wilson, W.E, ”A

Laboratory Study to Evaluate the Impact of NOx, 80),, and Oxidants on

Atmospheric Corrosion opf Galvanized Steel,” Materials Degradation
Caused by Acid Rain, Baboian, R., Ed., American Chemical
Society, Washington DC, 1985, pp. 172-193.

Spence, J. W., and Haynie, F. H., "Damage Function for the Effects of Acid
Deposition on Galvanized Steel,” 1987, Mass,

85

38. Semonin, R. 6., “Wet Deposition Chemistry,” Materials Degradation
Caused by Acid Rain, Baboian, R., Ed., American Chemical
Society, 1985,Washington D.C., pp. 23-42.

39. Passaglia, E., ”Economic Effects of Materials Degradation,” Materials
Degradation Caused by Acid Rain, Baboian, R., Ed ., American
Chemical Society, Washington DO, 1985, pp. 384-397.

40. Lareau, T. J., et. al, ”Model for Ecomonic Assessment of Acid Damage to
Building Materials,“ Materials Degradation Caused by Acid Rain, Baboian,
R., Ed., American Chemical Society, Washington DC, 1985,pp. 397-411.

41. Kucera, V., ”Influence of Acid Deposition on Atmospheric Corrosion of
Metals: A Review,“ Materials Degradation Caused by Acid Rain,
Baboian, R., Ed., American Chemical Society, Washington DC, 1985,
pp. 104-119.

42. Haynie, F. H., ”Environmental Factors Affecting Corrosion of Weathering
Steel,“ Materials Degradation Caused by Acid Rain, Baboian, R., Ed.,
American Chemical Society, Washington DC, 1985, pp. 163-172.

43. Hudson, J. C. and Stanners, J. F., LAW, 86, 1953.

44. Barton, K. et. al., W, 387, 1980.

45. Atteraas, L., et. al., ”Atmospheric Corrosion Testing in Norway,”
Annospheric Corrosion, Ailor, W. H., Ed., The Electrochemical Society,
John Wiley and Sons, N.Y., 1982, pp. 873- 892.

46. Haagenrud, S. E., et. al., Proceeding of the 9th Scandinavian Corrosion
Congress, Copenhagen, 1983, p. 257.

47. Knotkova, D., Gullman, J., Holler, P., and Kucera, V., 9th International
Congress on Metallic Corrosion, Toronto,1984, pp. 198—205.

48. Hakkarainen, T, and Ylasaari, S. "Atmospheric Corrosion Testing in
Finland,“ Atmospheric Corrosion, Ailor, W. H., Ed., The Electrochemical
Society, John Wiley and Sons, N.Y., 1982, pp. 787-797.

49.

50.

51.

52.

53.

55.

56.

57.

58.

59.

86

Upham, J. B., “Atmospheric Corrosion Studies in Two Metropolitan
Areas.” MW. 6. 1967- p- 398-

Feliu, S., and Morcillo, MuMJQMQBW. 4, 1975.

Barton, K., Protecfi'on Against Atmospheric Corrosion, John Wiley and
Sons, N.Y., 1976. '

Feliu, S. and Morcillo," Atmospheric Corrosion Testing in Spain,“
Atmospheric Corrosion, Ailor, W. H., Ed., The Electrochemical Society,
John Wiley and Sons, N.Y., 1982, pp. 913-922.

Haagenrud, S. E., et. al. 154th Electrocheimical Society Meeting, PA,
1980.

. Kucera, V., "Influence of Acid Deposition on Atmospheric Corrosion of

Metals: A Review," Materials Degradation Caused by Acid Rain, Baboian,
R., Ed., American Chemical Society, Washington DC, 1985, pp. 104-119.

Mikhailovaski, Y, N., etal,Emt,_MeL_1§, 4, 1980.

Haynie, F. H., “Theoretical Air Pollution and Climate Effects on Materials
Confirmed by Zinc Corrosion Data,“ Durability of Building Materials,

ASTM STP 691, Sereda, P. J., and Litvan, G. G., Eds., American Society for
Testing and Materials,1980, pp. 157-168.

Edney, E. O., Stiles, D. C., Corse, E. W., Wheeler, M. L., Spence, J. W.,
Haynie, F. H., and Wilson, Jr., W. E., “Laboratory Investigations of the
Impact of Dry Deposition on $02 and Wet Deposition of Acidic Species on

the Atmospheric Corrosion of Galvanized Steel,” W 1986,
pp. 541 -548.

Stiles, D. C., and Edney, E. O., ”Dissolution of Zinc into Thin Aqueous
Films as a Function of Residence Time Acidic Species and pH,”
Submitted to J. of Air Poll. Cont. Assoc, 1987.

Weast, R. C., et al., Handbook of Chemistry and Physics, CRC Press,
Cleveland, OH, 1979, B-142.

87

61. Lipfert, F. W., and Wyzga, R. E., "Application of a Theory for Economic
Assesment of Corrosion Damage,” Materials Degradation Caused by Acid
Rain, Baboian, R., Ed., American Chemical Society, Washington DC, 1985,
pp. 41 1-430.

62. Gibbons, John H. (Director), Acid Rain and Transported Air Pollutants,
US. Congress Office of Technology Assessment, Washington DC, 1984.

63. Sereda, P. J., ”Weather Factors Affecting Corrosion of Metals,”
Corrosion in Natural Environments, ASTM STP 558, American Society
for Testing and Materials, 1974, pp. 7-21.

64. Stanners, J. F., "Selecting Testing Conditions Representative of the
Atmospheric Environment,” Corrosion in Natural Environments, ASTM
STP 558, American Society for Testing and Materials, 1974, pp. 23-31.

65. Haynie, F. H., “Environmental Factors Affecting the Corrosion of
Galvanized Steel," The Degradation of Metals in the Atmosphere, ASTM
STP 965, Dean, S. and Lee, L. S., Eds., American Society for Testing and
Materials, Philadelphia,1987, pp. 282-289.

APPENDICES

7.1

25.

CORROSION RATE (UM/YR)

CORROSION RATE (UM/YR)

 

 

88
APPENDIX A

Corrosion Rate vs. Time-of-Wetness for Equation (2.7).

.001028*(RH—48.8)*802

HAYNIE AND UPHAM
EXPOSURE TIME 1 YR (8760 hours)

a.

iii

 

III
m

 

 

TIME OF VIETNESS (hours)

Corrosion Rate vs. Sulfur Dioxide for Equation (2.7).

c-o.001_028(RH-48.a)osoz -
HAYNIE a UPI-IAN
an - 404/[404—Inah/0760n

 

I'o. ab. 35. 4b. 58. ab. 76. a5. 97). 160.
SULFUR DIOXIDE (us/M3)

.0 -5

ml

27. Corrosion Rate vs. Time-of-Wetness for Equation (2.9).

.005461 -(Tw-°,‘”)-(Soz /2600+.02889)*1 4.98

GUTTMAN
EXPOSURE TIME 1 YR (8760 hours)

 

 

70.01
3
z
D
V
E
E 5.0-
2
9 .
tn
0
n:
I: ‘l
O
O
H m- !” (03M
H soa- 80 (no/0')
' H 802- “! (nu/I.)
0'0 T V T ‘ F ' ‘ ' ‘ I
' 4300 8700

mm or mass (hours)

28. Corrosion Rate vs. Sulfur Dioxide for Equation (2.9).

GUTTMAN
C-0.005461 (Tw)".8152¢(502+0.02889)

.4. Tu - 870 HRS
9.- H TI - 2820 HRS (1/3 year)
HTI-8760HRSU year)

    

 

A

0:

< 80 '-

2

D 70-

V

E 8. '1

§ 5.-

5 4.4

8 3.4

o: .

g 2.

0 1 .
O.

 

07 1'0. 20. 35 4b. 50. 00. 7'0. ab. 90. 130.
SULFUR DIOXIDE (US/MS)

CORROSION RATE (UM /YR)

29. Corrosion Rate vs. Time-of-Wetness for Equation (2.11).

.0049*(TW"')*(SOZ/72.8+.05)

HAKKARAINEN AND YLASAARI
EXPOSURE TIME 1 YR (8760 hours)

 

 

27.0-‘
I
I
18.0-I
.
I
9.0-
I
I
2 - H 502- 100 (ug/M’)
* , e " H 502- 50 (tag/u!)
o 0 " i H 502- 10 (ug/M’)
. . f f - , . r . . ,
0 4380 8760

TIME OF WETNESS (hours)

30. Corrosion Rate vs. Sulfur Dioxide for Equation (2.11).

91

31 . Corrosion Rate vs. Time-of-Wetness for Equation (2.13).

[.OOO76-(tw/365)‘ *(802/.728)"~'° ]-51 .19

BARTON at. oL
EXPOSURE TIME 1 YR (8760 hours)

 

 

7.5-]
I
:50-1
3
E
&.
§
In
E u:
8
HIE-INhI/I’)
Hm- ”(W/“'7
0 - . . . 1 . Hm- ”(04M
0 ' ' 4300 duo

.TIME OF WETNESS (beam)

32. Corrosion Rate vs. Sulfur Dioxide for Equation (2.13).

c-0.00075cru‘.5).(sozt.71a)
BARTON at.ol.
1 your mum tine I

 

 

10.]
' E O.‘
2
a - '
. 5....
I3
é 1
z
9. 4.-
U)
0
m
m
0
° 2.-
H T'IO7OOI’RSU ”0!)
H II- - 2920 IRS (1/3 year)
H TI - O70 H6
0-4 F 1' I I I I I I I
CI. 10. 20. 30. 40. 50. 00. 70. BO. 90. 700.

SULFUR DIOXIDE (us/M3)

33.

CORROSION RATE (UM/YR)

CORROSION RATE (UM/YR)

34.

'92

Corrosion Rate vs. Time-of-Wetness for Equation (2.15).

15.01

10.0‘

C-[.O450502+.0314UN02+EXP(111355—423378].
EXP[—.4(100-RH)/RH]
HAYNIE
RH-4D4/[4.04-ln(Tw/8760)] E-iOOO/(T+273.16)
u-.54crn/a. N02-61.44. T-1 1' C

 

 

50' 5'
- . ,3 a 2858 ;
a H 000- :0 (cg/In

°". ' ' ' ‘ do " ' ' * ciao

- . THE OF WETNESS (hm)

C-W-l-QJMIMZ-i-uptf7433-4m]. ’
upf—SAOOO-RHVRH)
m

15.1 m - «MW/3100)]. E - Iona/04.21345)

u-Mw/gf-Itqm-NA409/m3

‘2'. //

 

 

 

 

’0‘

I

60‘

3.- _ _ _ _ _ _ _ !.

nth-mummy...)

‘ g _ #Hh-MWUISm)
o. - '2 . . L I - . H“"’“"“

0. - ‘ 50. 160.

 

SULFUR DIOXIDE (UC/MS)

Corrosion Rate vs. Sulfur Dioxide for Equation (2.15).

_ 35.

CORROSION RATE (UM /YR)

CORROSION RATE (UM/YR)

4.0'1

 

 

 

93

Corrosion Rate vs. Time-of-Wetness for Equation (2.18).

(.0 I 87*SOZ,+1 .461 )*(Tw/8760)

HAYNIE. SPENCE AND UPHAM
EXPOSURE TIME 1 YR (8760 hours)
T= 8.1'C

H 001- I” (W)

 

T ' f T4300 ' ' ' ' ciao
nIIEOI-‘IIrErN-s(hom)

COrrosion Rate vs. Sulfur Dioxide for Equation (2.18).

C-[0.0187tSOZ-l-exp(41 .85-23240/RT)]¢Tw
HAYNIE. SPENCE. AND UPHAM
T - 8.1 C

l-ITI-O'IOI'RS /

‘ HN-ZOZOIRSU/Jyaor)

H TI - 8760 HRS (1 your)

 

£‘ .__. :~:-Ifi:#l——I——V
0. ,

I
O.

I I I
10. 20. 30. 40. 50. 50. 70. ab. 35.-"'00

SULFUR DIOXIDE (US/MS)

742 APPENDIX B Itfififitfififittiitiifitfifilflfittfiflfiitifltti.....ttittttifittttfittfiw

pnoonnu cant
”......CfifitfiitifittifiOtflflfitifittififitfifitiflitfiIQCOOOOOIOOQOQOOflitttitttfifitttfittitt
c Pnoonnu l0 cancunama DAIL! annoutrzac Dara (HEAR, anxznun,
c uzlzuul AID armlnann Dnvxnrtou)
CO...Cfifiitlfififiifltttti......fifitiififltitfitfiC......flfiiiflflttflititittiiitfititfiitfiflitt
c

PAIAHITII (l-200)

*caanncrsatao our!

cannacrxntaz zurtnl,oorrtnl

tuwcazn c1,c2.c3.c4,cs.c6,L1,Lz,L3,L4,Ls.ns,31,u1,xaan.0ama

_- IIAL soz.uaa,rp,op.nn.pn, _

-aaoz,m.srp,snp,sna,sn, ma.m,mr, MJGIR, my,

-|III8,KIII,III'2,IIIRD.KIIR.KIIP,DBOZ,DNBA,DTP,DDP, one,

-DVIRJ,DVIR2.DV183,DVIR4,DVIIS,8Vll1,BVIR2,SVIRJ,BVIRA,8VARS.

-Dszpt,DatDz,Dstpa,osrnc.osros

DIIIIIIOI 003(24).IIA(24),ID(24),DD(24).nnIzl).nn(24).DA:I(24)

c

c.....m'ona um! um m AID m m mmxm m I113...
c .

CALL mgnutmmmoflnm
W (10.nu-znm.srma-ronnr)

c 0". (30.7W.M'm')
0" (30,W,m'mm')

100' sane-0

95

3:30 (10.'(00A)',IID-999) sort

I! (lurt(7xa).lg.'00') tall

' lorr(7:0)-' '

Ill!!! (lDtt(1sI).lq.'0')rall
3037(737) - ' '

llnxr

IIID (IUI'.120)!IAI,DASI(I), 002(1),IBA(I).IP(I),DP(I).RH(I),PR(I)
01-01-01
[LII I! (AII(O0R(I)-99.999) .LI. 1.000) Till
' 01-01+1 '
ILII 1’ (002(1) .08. 0.000) Ell.
soon-aooz+aoc(z)
1.1-1.144
It (000(1) .0:. lass) tall
' Ilsa-000(1)
“I,
. Hill-000(1)
. -1!
m,
0000-0000
IIDII
ca-ca-n
no
'IIIDFSIIANIIR(I)
3.2-n+1
1r (Ill(1) .oc. lax!) cull
IIIIbIII(I)

I! (ma) .u'. I!!!) all
Inn-mu)
In"
Ill!!!
' It (AII(!!(I)-999.9) .LE. 1.0) tall
ca-ca-u
It.”
mu)
1.34.344
I! (IN!) .02. In!) mu
Inn-nu)
-1!
It (”(1) .m'. I!!!) I'm
Inn-nu)
Ill!!!
'1!
1r.(aas(np(x)-909.9) .02. 1.0) tutu
c4-c4+1
nu
IDP-lDP+DD(I)
W1

1! (00(1) .02. MAID) Till
lIlD-DP(I)
ll!!!
10 (00(1) .01. Iran) was:
. IIID-DP(I)
I001:
3:01!
It (ana(na(x)-999.9) .L1. 1.0) ran:
.cs-cs+1
ILII
can-033+na(1)

110
120-

130

Inn-nan)
3'01!
1! (“(1) .02. II!!!) mu
urn-mu)
RID"
RID"
-:r (aaa(pn(1)-9999) .03. 1.0) tuna
06-091
I!“
‘ woman)
1.6-1.6+].
I? (”(1) .013. M) I'll!
IIZP-PR(I)
IIDII'
I, ("(1) .LI'. III?) I'll.
Imus-03(1)
It

"I!

00:21:03
m (fl,12,13,108,u.3,4(33,”.1)p31,”.0)
I! (01 .ll. 0) run:
0000-0002/(01)
Ins:
bloc-9.099
. II88-9.999
- urns-9.999
000:!
tr (:2 .13. 0) rain
0000pcuaa/(Lz)
Ins:
DIBAP999.9
Ilsa-999.9
Ian-999.9
El!!!
1! (03 .II. 0) rain
DrD-OIP/(La)

1: (La .03. 0) tall
Don-sop/(L4)

Ins:

DOD-999.9
MAID-999.9
NERD-999.9

1001:.

1: (L5 .II. 0) run:
annulus/(La)

Ina:

DIE-999.9
anal-999.9
u:up-999.9

3:01:

It (00 .c9. 0) runs
Hazy-9999
IIlP-9999

0001!

0033140

Dvanz-o

ovals-0

140

97

wan-o
cmz-o
suns-0
ovum-o
Inns-0
08101-0
08202-0
00103-0
08104-0
08208-0
01-0

no m 21.24.
11' (u .n. 0) 1m
81-81-14
m I, (”mama-99.090) .02. 1.000) rm
III-81+].
In: 11 (002(1) .03. 0.000) m
. m1-0m1+(000(1)-0000)n2
no
Ivan-0m}.
0.01,
3.013
11 (1.2 .II. 0) 1m
.avm-sm+(m(1)-M) “2
0.01,
-1,
u (1.3 .II. 0) ml
MW(T’(I)-0fl)*'2
1.011
non -
I! (M .II. 0) III.
" W(0’(1)-fl0)**2
1:011
3001'
I? (1.5 .03. 0) 1m
WW(u(1)-mp-z
111011
-1!

12 (1.1 .00. 0) 1m
Dam-9.999

11.01
Dun-mum)
Dam-mam)

mu

1! (1.2 .00. 0) 1m
Dana-999.9

31.01
Dvm-svm/(u)
ma-aguwmz)

mu

11 (1.3 .09. 0) 1m
Dana-999.9

11.03
Dma-avm/(La)
Duos-sonmpma)

C)¢)¢)¢I¢ICIGMO

1‘0

170.

190

200
210

220
230

0

Dunc-999.9

01.00
(Imam/(1.4)
Dome-comma”)

m1!

1: (1.5. .00. 0) 1m
Dans-099.9

01.00
ems-mum)
ems-agarwmn

00012

01-24-(01)-(c1)

m. rampant).0002.11,c1,01.Dm1,00101.0m.00102.1.2
1.110 (20,150)

.110 (20.100)Daoa.m,m.oop,om.sn

n11: (20.110)m,m,m,m.m,m

.110 (20,180)m.m.m.nm,nn,mr

00110 (20.100)Dm1,0mz,0m3,0m4,0m5

' 1.110 (20.200)!.1.L2.L3,IA.LS.M

.110 (20.310)c1,cz,c3,04,cs.cs
Inn-0 (20.220)01
00110 (30.230)Dan(1).0000,m,m.001,000.0n

roamt/l/l/l/uum M10 0010 or m Dmv/l
-201,'002',0!,'Ill',0!.'TP',03,'00'.03.'Ifl',OX,'PR')
“(mum 10001.4
41pm: 001w,
-3‘0'10‘0“"6020‘x'"02"""02":p’6020‘x9’601’)
max. 'mm',7x.n.4.4(41,u.2)41.16.34)
IUIII! (‘1.'IIIIIUI'.7!,'7.6.C(CX.'0.2).4X,’0.1/)
100m (u,'nmm-/cx,rnmmow.41,
0'0.0.28,'0.2,2I.’0.3.23.'0.2.23,'0.RI)
100m (41pm err/41.'000n100'.01.13.s(71.13)/)
100m (41mm or]
ouumcm Dm',41.13.s(11.13)/)
100m (annual: or'/
41.010001“: communal/Ill)
loam (41.13.41,n.4.4(41.n.z).4x,re.1)
- 0000 100

rum. 01101“
0100“»)
1:00:00)
mac)

0100

up

c0.........O...‘.0...O..000...Oififififlfififltiiflfiifitt0.0.0.0....fittiitttfifitttiiiflfifltr

000000-1100 00'1_r11.00(1m1.0.00'1111.0)

'QOOOQIOOQOIIOO0.00.0.00...0.0.00.0..000000.ififiitiltiflifitiifiDflttiifiitittfltfifitfifiti

mmmmxmm.mmnmmnm,

c..00...00..00.0000.0.00...Q...000......0000......OIQIQCQOOQOOQOOO......tfitfiitfii

0

man-v32 Immortal,”
1010000 In

00 1° 1 - 1'32

W(III) - ' '
W(III) - ' '

00110(t,20)
20 000001(' ',//,21,'00100 10001;p010 0110 0000']
+, r (32 0000 0011000, 0010011 011000100 -.001)s',///)
c - .
0000(-.'(0)') 100110

1 - 1
001-0
00 00110 ((100110(131).00.' r).000.(011.09.0).000.(1.10.32I)
.10 (100110(1:1). .'.')1000
c.........10001 0110 0000 10010000 00 011000100, 000 0000 00 10,
« 011-1 .

00000 i 100110(2:1-1)
. 10 (001.00.0) 1000
c..........10001 0110 0000 001 1001000 00 011000100...............
c0000....000m mm, W!" (.m’.OOOOOOOOOOOOOOOOOO00.0....
100110 (131+3) - '.001'

00010
c -“
0010110 - 00000
0010110(1:1+4) - v.0011'
c .

100

c000...000.00.000.0000000000000000000000000.000000000000000...Otiittittttittw.’

0 0000000 081.04
c0000..0.000000000000000000000000000000000.0000000000000000...000000000000000at
0 0000000 1'0 W 00001! MIN EMA (MIDI, MAXIM,

0 0101000 All) 01000000 WIN)

c.0000.00......0000000000000000.000.*00*0OlfliOOOQ00......0fitfifitiiflfifiiflttitI.“OOV
000000000 (0-200)
CHIIACTIR*80 0000
Ilfiflll 01.02.03.04.CS.¢$.LI,L3,L3.L‘.L5,L5,SL.I1
'8‘; 002.'IA.TP.DP"30P.0
60002.8'Il,81P,80P,818,8P8. IIIB.KIXfl.HBIT. IBID,IIIR, HAXP.
-IIII.HIII.IIIT.IIID.IINR,IIIP.W002.WWBA. “RP, “DP, "In.
-'VIIJ,'VIRQ'WVIR3.UVIR§,'VIIS.8VIRJ.8V332,3VIR3'8VIR‘,8VARE,
-'I!Dl,fll!bz.IBTD3.IITD‘.II!DS

DIIIIIIOI 000(160).III(160).!P(168).DP(168),Ifl(160).PR(168)

0". (lg'ILl-"ILIIQ.OU!'.SIflEUII'OLD'b ’
00" (2.!ILII'IJI04.002',IIIIUBO'III')

100 0000-0

5502:52222.0003500;

IIIP-IO0.0

00 110, I-l,160
IIID'(1,'(BOA)',IlD-999) 00"
0000 (0000.120) 000(1),000¢1),00(1,,00(1).00(1),00(1)
I? (A38(002(I)-9.999) .02. 1.000) THIN
01-01+1
'L“ I, (h30(802(I)-99.999) 0‘30 10°00. Tun'
01-01+1
IBIS 1’ (002(1) .08. 0.000) Til!

.mflM‘DAM‘! o v \

IO!

10 (000(1) .00. 0000) 0000
HIS-002(1)
Ill!!!
1' (002(1) .Ll'. III.) 0000
HID-002(1)
0.017
0100
0000-0002
‘01?
10 (000(000(1)-999.9) .10. 1.0) 1000
02-0244 "
0100
W(I)
2.2-1.2+).
1' (088(1) .0‘1'. M) 1000
”081(1)
"I!
10 (000(1) .10. 0100) 0000
M(I)
"I!
IIDU
(BS-03+).
0100
' W(I)
3340-01
1’ (”(1) .02. M)
NIH-”(1)
"If
10 (00(1) .10. 0100) 0000
M(I)
“I!
‘1'
04-04-01

0100
000-000-000“)
W1

10 (00(1) .00. 0000) 0000
0000-00(1)
00010
10 (00(1) .10. 0100) 0000
1000-00(1)
0310
00010
10 (mama-999.9) .10. 1.0) 0000
00-091
0100
menu)
10-1s+1
10 (00(1) .00. 0000) 0000
0000-0101)
00010
10 (00(1) .10. 0100) 030
nun-00(1)
00010
00010
10 (000(00(1)-9999) .11. 1.0) 1000
06-13-01

0100
000-000+00(1)
1.0-191
10 (00(1) .01. 0000) m0
Inn-00(1)
00010

110
120

no

102

IIIP-Pfl(1)
I“!!!
IIDI'

00001000
000000 (000,05.3,4(31,0s.1),30.0s.0)
10 (11 .00. 0) 0000
0000-0002/(11)
0100
0000-9.999
0000-9.999
0100-9.999
00010
10 (10 .00. 0) 0000
0000-0000/(10)
0100
0000-999.0
0000-909.9
0100-990.0
00010
10 (13 .00. 0) 0000
000-000/(13)
0100
000-999.0
0000-900.0
0100-999.9
00010
10 (10 .00. 0) 0000
000-000/(10)
0100
000-999.9
0000-099.9
0100-999.9
00010
10 (1s .00. 0) 0000
000-000/(10)
0100

Ina-090.9
0000-990.9
0100-999.9

00010

10 (15 .00. 0) 0000
0000-9999
lIlI-9999

00010

DO 140. 1.1.168

103

81-8144
0100 10 (000(002(1)-99.999) .10. 1.000) 0000
81-81+1
0100 10 (002(1) .00. 0.000) 0000
00001-0v001+(002(1)-0002)n2
0100
mm
“I,
“D!"
I? (1.2 .II. 0) 0000
00002-0v000+(000(1)-0n0)n2
III)!"
”I?
I, (1.3 .II. 0) 0000
WW(2P(I)-M)'*2
-1!
00010
I? (M .II. 0)
I, (W(DP(I)-999.9) .02. 1.0) 0000
W(DP(I)-DD)'*2
IIDII'
-12
10 (10 -00. 0) 0m
00000-00000+(00(1)-n0)«0
IIDII
-1,

10 (11 .00. 0) 0000
Ital-9.999

0100
mum/(11)
00001-0900mmn

00010

10 (1a .09. 0) 0000
00000-9903

0100
mama/(12)
0090-0900(00002)

00010

10 (13 .00. 0) 1000
nus-099.9

01.00
Imam/(13)
madman”)

00010

10 (14 .00. 0) 0000
00004-999.9

01.00
mam/(14)
00004-0000(0mc)

00010

10 (13 .09. 0) 0000
Isms-999.9

0100
mums/(15)
00000-0900NV000)

_ -10

01-160-(11)-(01)

150

160

170
190

200
210

104

101100 (2.150)

“I!" (2.160)m2.m0.000.m0,000,s00
101100 (2,170)0000,0000,0000,0000,0000,0000
101100 (2.180)HIS,HIUU.IIII2,HIND.HINR,KINP
101100 (2,190)00001.00002.00003.00000,00105
00100 (2.200)1.1.1.2.L3.M,L5,L6

'12] (2,010)c1,02.03.04.00.06

101100 (2.220)I1

000md/l/l/l41u000 me 0000 00 0010 0000: VI
-2”. '802'933. 'm'.”. '"'p“, '3’. .33. 'IB' 0"; 'PR')
0000011600. «00010 000014

«0.4100010 0000',
-31.".4g‘x.'6.2,‘X.".2.u.”.2,4x, l6.2.4X.'6.1/)
000000(40.'n01m0',10.07.0.4(40,06.2).40.06.'./)
000000 (40.'01011010'.70,07.4.4(40,00.2),40.06.1/)
000000 (‘8, '00000000'I40.'00v100100',40.
-".4.2!.n.2.21.".2.2X.n.2.28,38.2])

000m (40.00000 02"“, 'WM'.OX.I3,5(7X.I3)/)
m (‘3. '000000 0"]

908. 00100100 0000*,40.13,0(70,13)/)

000m (40. '000000 0"]

0‘8. '00000100 M'.38.13/////)

”1'0 100 '

00.100» '01010000'
(0.000(2)

«000(1)

0000

000

5 6037

mill
Am
Ruflflfl
mm
Lflflflu
Y"
"I
Hflflia
m0
m3
flflfluO

W

1293

 

'IHIWIHIIIIIHI