|
Issues Impacting Bridge Painting: an Overview
Chapter 3. Task B - Worker Protection/Paint
Removal
Table of Contents
INTRODUCTION
The objective of task B was to review current worker health
and safety practices pertaining to the removal of lead-based
paint from existing structures. This included reviewing containment
concepts and methods of testing the relative efficiency and
safety of different types of paint-removal methods. Development
of a test chamber to assess the relative merits of various
removal techniques also was suggested in the initial plan.
The objectives were achieved by: (1) assessing information
obtained from state and federal DOT personnel, the general
literature (including Steel Structure Painting Council (SSPC)
publications), and materials and equipment suppliers, (2)
conducting and participating in field paint removal and recoating
trials, (3) observation and assessment of contract bridge
repainting operations done in cooperation with State departments
of transportation, especially the Illinois Department of
Transportation (IDOT), and (4), performing laboratory tests
and analysis of samples from the various field operations
in addition to samples prepared in the laboratory.
The program plans were reviewed with an advisory
committee consisting of experts in the field, including State
DOT representatives
from Oregon, Illinois, and Kentucky; the Army Corps of
Engineers; paint and thermal spray companies; and the FHWA
program monitor.
The general consensus of the committee was that it would
be difficult to develop and qualify a universal test chamber
to accurately assess the relative merits of paint-removal
techniques within the time and funds allocated to this
task in this program. It was agreed, however, that a controlled-environment
chamber could be valuable for certain removal-method evaluations
as well as for paint-application control studies. Thus,
the
majority of the paint-removal/worker-protection analyses
were based on field operations, and a lesser number of
analyses were based on chamber tests. Valuable information
also was
provided by the Iowa, Kentucky, Louisiana, North Carolina,
and Wisconsin DOTs.
DISCUSSION
Paint- Removal
Methods
There are many methods of paint removal, including
abrasive or grit blasting, water jets, hand tools, power
tools, chemical
stripping, and heat. Each of these general methods consists
of a number of subcategories as shown in TABLE
2.
An excellent comparison of many of the attributes
of these methods is given in the Industrial
Lead-Paint Removal Handbook by K. A. Trimber (see
chapter 5, table 4). Additional comparisons based on our
work in this program
are provided
later.
In the past, the most common paint-removal
method has been abrasive or grit blasting. It is fast, and
it removes paint,
corrosion, and millscale to produce a clean surface with
a good profile for recoating. Its main disadvantage is
that it produces large volumes of dust. The dust contains
very
fine particulates from which the environment and the workers
must be protected. The cost of this protection is high.
Today, continuous blasting without containment will generate
particulates
in excess of the U.S. Environmental Protection (EPA) regulations.
An alternative is vacuum blasting without confinement,
but the removal rates are greatly reduced (approximately
80 percent).
The added weight of the vacuum attachment tires the operator
and the added bulk reduces mobility. In case of containment,
building and moving the confinement structures are both
costly and time consuming. In both cases, the used blast
media and
the paint debris need to be separated so the media can
be recycled and the volume of hazardous waste can be reduced.
This significantly reduces the cost of treatment and disposal
of the waste.
Steel grit often is selected because it can
be recycled repeatedly because of a low breakdown rate. It
gives a good
surface profile, and provides temporary stabilization against
leaching of lead from the debris as measured by the toxicity
characteristic leaching procedure (TCLP) test. Steel-grit
blasting produces large numbers of paint debris particles
with diameters of less than 1.0 µm. As discussed in
greater detail later, size analysis of the particles collected
on a filter within the blast confinement area showed that
more than 50 percent of the particles had diameters less
than 1.0 µm. Elemental analysis of these fine particles
shows that they contain lead. It is difficult to determine
how many particles with diameters less than 0.3 µm are
produced by the blasting operation because such small particles
pass through most filters and common particle size analysis
techniques are not applicable below 0.3 µm. Particulates
that contain lead/chromium and have diameters less than 0.3µm
would present a hazard to the worker because: (1) they would
not be totally removed by standard high efficiency particle
absolute (HEPA) filters which are rated 99.97 percent efficient
down to 0.3-µm size particles, (2) they would be difficult
to contain within the confinement structure, and (3) they
would produce contamination at greater distances from the
work site.
Despite problems associated with grit blasting,
its high removal rate outweighs the additional costs associated
with
worker and environmental protection when compared with
other paint-removal options. In general, blasting within
confinement
is the method of choice when complete paint removal is
specified, while vacuum blasting may be selected for spot
preparation
requirement. In this process, it is difficult to recover
all dust and debris with the vacuum system. Very few additives
are used in grit blasting. One exception is the addition
of a material such as BlastoxTM which has been
shown to help stabilize the lead in the paint debris against
leaching in the TCLP test. A negative aspect of BlastoxTM is
the need to keep it dry; when wet it clogs the blasting
system.
Water jets can be used to remove the loose
paint and rust, as well as salt, dirt, and soils. Advantages
of this method
are: (1) the degree of loose paint and rust removal can
be altered by adjusting the water pressure, (2) the paint
usually
comes off in chips so that these chips can be collected
on a porous tarp, while allowing the water to pass through,
(3) the water temperature can be adjusted, and (4), it
is
relatively inexpensive. Disadvantages are: (1) the method,
by itself, is not suitable for total paint, corrosion,
and millscale removal; (2) it does not produce any surface
profile;
and (3) care must be taken to be sure all hazardous debris
is contained and properly disposed. Paint removal by hand
tools gives a minimally prepared surface. This may be satisfactory
provided good surface-tolerant paints are available and
limited coating lifetimes are acceptable. Handtool removal
is probably
the most operator sensitive of the removal methods. It
has the advantage of producing the least debris and the debris
that is produced can be easily collected on tarps.
Chemical stripping is used in certain cases,
depending on bridge type and location. It is used most often
to remove
paint from tanks. This process involves messy, dangerous
acidic or alkaline materials, often as gels, which are
hazardous and require thorough rinsing and proper disposal.
Paint removal by laser was examined in the
laboratory using a Coherent EFA51, 1500-W, continuous-wave
CO2 laser.
The substrates (test samples) were 102-mm (4-in) square
or 51-mm by 102-mm (2-in by 4-in) rectangular sections
of painted
steel cut from bridge girders that had been taken out
of service and stored by the Illinois Department of Transportation.
The existing paint was a three-layer system consisting
of a red-lead/alkyd primer, a leafing aluminum/phenolic
matrix
intermediate coating, and a green-pigmented, non-leafing
aluminum/phenolic topcoat. The test samples were placed
within
a glass confinement system with a zinc selenide (ZnSe)
window so that all byproducts from the removal process
were contained.
The containment system consisted of a 2-L glass filter
flask with side tubulation and the bottom cut off. The
test sample
was placed on a steel-base plate to which the flask was
sealed by vacuum-bagging tacky tape. The ZnSe window was
sealed
to the top of the flask and the side arm led through
a TenaxTM adsorbent
tube to a TedlarTM gas bag. The apparatus
is shown in FIGURE 12.
The laser beam covered a 6.4-mm (1/4-in) diameter
area of the sample surface. It was found that approximately
1
s of laser operation was sufficient to remove the paint
without damage to the steel. Thus, the sample was moved to
expose
a different spot on the surface after each 1-s irradiation
until sufficient byproduct could be collected for analysis.
Each time the laser beam impinged upon the
painted surface, the paint burned with a yellow, sooty flame
that rose 76
to 102 mm (3-4 in) above the painted surface. Visibility
through the flask was reduced with each application of
the laser. This was due to the fine soot that formed and
was
deposited on the sides of the flask. It scattered light
within the flask and was deposited on the inner walls of
the flask
as shown in FIGURE 13. Larger,
heavy soot particles fell to the bottom of the system,
depositing on the base plate as well as on the painted
sample itself.
Samples of the soot were collected from various
parts of the apparatus and were analyzed by scanning electron
microscopy
(SEM) and energy dispersive x-ray (EDX). SEM micrographs
of soot from the walls of the flask, from the base plate,
and from the area where the byproduct gases escaped from
the flask are shown in FIGURES 14, 15, and 16, respectively.
By comparing the particle sizes with the 10-µm bar on
the photos, it is evident that particles with diameters much
less than 1 µm are prevalent. In fact, the larger particles
appear to be agglomerates of finer particles.
The condition of the painted surface after
being subjected to the laser beam for 1 s is shown in FIGURES 17a and 17b.
Several different areas are apparent in the x 10 magnification
photograph. The raised black area is where the charred
paint has blistered up away from the steel. A portion of
the char
was broken off, revealing that there is still some residual
red lead left on the steel as seen in the upper right-hand
corner of the photograph. Some large pieces of soot deposited
on untreated green paint are shown at the left and the
bottom of the picture.
EDX analysis of soot samples shown in FIGURES
14, 15, and 16 demonstrated
that the soot contains lead. The other elements detected
by EDX were silicon and aluminum.
FIGURE 18 is
a photograph of the surface after laser treatment. The photograph
is at
x 7 magnification and shading was produced by the light
shining at an angle from the top of the photograph. Several
interesting
features are apparent. Over most of the treated area, the
topcoats were removed leaving red-lead primer. Complete
paint removal down to bare metal occurred only over a small
area
as seen near the top of the treated area. Residual aluminum
can be seen along the periphery of the treated area.
The laser tests lead to the following conclusions:
- Paint can be removed rapidly by laser treatment.
- Care must be taken to insure complete paint removal.
- Large volumes of soot are produced during paint removal
using the laser.
- The particle size of much of the soot produced by laser
treatment is very fine (less than 1 µm).
- The laser-produced soot contains lead and aluminum.
Since the use of heat or light sources to remove the paint
would most likely result in similar conclusions to those above,
they were not extensively studied.
IDOT PAINT-REMOVAL DEMONSTRATION
- TESTING PROTOCOL
The Illinois Department of Transportation
(IDOT) conducted a paint-removal demonstration program at
I-55 between Wood
and Damen Streets in Chicago, Illinois on April 22 and 23,
1993. The project was monitored in this study. The IDOT objective
for the program was to familiarize contractors with various
aspects of paint removal, including removal methods, Illinois
DOT requirements, EPA and Occupational Safety and Health
Administration (OSHA) requirements, full-containment structure
construction, paint-condition determination, and analysis
of paint/blast debris. The objectives of this study were
to monitor the removal methods, assess and compare their
advantages and disadvantages, obtain samples for TCLP and
other analysis, and demonstrate thermal-wave imaging (task
F) and electrochemical impedance spectroscopy (task E) as
methods of evaluating existing conditions.
Work within this study was divided into three
categories: tests before paint removal, tests during the
removal processes,
and tests after the paint removal. All of these tests,
which are listed below, provided information directly related
to
the objectives of task B of this study. In addition, several
of the tests, such as 1, 2, 3, and 4, provided information
generally required in order to make an economic analysis
to fulfill our task A objectives. Several of the tests
(4, 5, 12, and 13) provide the type of data needed to address
issues related to task C (Analysis and Disposal of Paint
Debris), such as assessment of the degree of hazard and
handling
procedures for the waste debris. Paint debris samples were
taken for TCLP analysis in order to verify results reported
by others concerning the ability of steel blast media to
stabilize the lead from leaching in the TCLP test. Other
tests such as 2, 3, and 5, provided information relating
to task D (Advanced Coatings).
Initial Tests
- Videotape the initial bridge steel surface condition.
- Measure the paint film thickness using dry film thickness
gauge.
- Conduct paint adhesion tests using an Elcometer.
- Collect samples of existing paint for metals and chloride
analysis.
- Measure lead (Pb) levels in existing paint using an x-ray
fluorescence (XRF) paint analyzer.
- Collect water samples from web sections of girder to
be washed.
- Document the initial paint surface condition of a section
of bridge girder, similar to the one from which paint is
to be removed, by thermal-wave imaging.
Tests Conducted During the
Paint Removal
- Collect air samples as near as possible to the blast
or tool surface during the paint-removal operations.
- Collect airborne particulates as near as possible to
the blast or tool surface during paint-removal operations
using filters and pumps.
- Measure temperature and humidity.
Tests Conducted After Paint
Removal
- Measure lead (Pb) levels remaining on the cleaned surfaces
using XRF.
- Collect samples of paint debris from each removal method
and submit for total Pb analysis and toxicity characteristic
leaching procedure (TCLP) testing.
- Laboratory tests:
- a. Analyze air samples taken during paint-removal operations
using gas
- chromatography/mass spectrometry (GC/MS) for volatile
organic compounds (VOC) determination.
- b. Analyze particulates collected on filters for metals
content and size distribution.
Double containment was used in the demonstration in order
to minimize the possible escape of dust from the grit-blast
area. The containment structure was recorded on videotape
for future reference and is part of the IDOT records but
is not discussed further in this report. Details are available
from IDOT. Entrance and exit from the containment area was
via an anteroom or change room. The grit blasting was done
using G-40 steel grit. The original paint film thickness
was measured as 0.15-0.20 µm (6 to 8 mils). It was a
three-layer system consisting of a red-lead/alkyd primer,
a leafing aluminum/phenolic intermediate coating, and a gray
non-leafing aluminum/phenolic topcoat. Adhesion varied significantly
from area to area on this structure. Elcometer tests gave
adhesion values of approximately 2067 kPa (300 lb/in2),
but since the paint was brittle, it spalled off while crosshatching
during the ASTM D3359-90 test (Measuring Adhesion by Tape
Test, Method B - Cross-Cut Tape Test) producing a 0B rating
(extremely poor adhesion). Weather conditions were mild throughout
the 2-day removal demonstration. Temperatures ranged between
12.8 and 15.6°C (55°F and 60°F), and relative
humidities were 35 to 45 percent. Details of the tests relating
to task B are discussed below.
Analysis of Paint and Steel
for Lead by X-Ray Fluorescence
Both the paint and the base steel were analyzed
for lead content using a Princeton Gamma Tech (PGT) x-ray
fluorescence
(XRF) lead-paint analyzer. This work was done by George Cardis
of Loyalty Environmental, Inc. of Skokie, Illinois. The XRF
method gives results per unit area -- it analyzes approximately
1 cm2 of surface. For lead, it is accurate to ±0.5
mg/cm2. Paints with lead levels greater than 1 mg/cm2 are
considered hazardous. Analysis of the base steel was needed
to obtain a background lead level. The paint contained such
large amounts of lead that all paint readings were off scale,
i.e., greater than 10 mg Pb/cm2. After the various
paint-removal methods were demonstrated, the cleaned-surface
lead levels were again measured. It was found by visual inspection
that grit blasting did drive a small amount of lead from
the paint into the cleaned surface. This amount was below
0.5 mg/cm2 and thus was no longer considered as
hazardous. Some concern has been expressed as to the possibility
of the presence of sulfur resulting in false lead-concentration
measurements by XRF. While it is true that the lead M series
line measured at 2.3 keV is interfered by the sulfur K series
line, the XRF device used during the IDOT test measured lead
concentrations using the lead K series line at 74.9 keV with
no interference by sulfur.
Although originally designed for use in homes,
schools, and other buildings, the method is finding more
and more
use with steel structures. This is because the measurements
are easy to make, they require only a few minutes to obtain,
and the results are known immediately on site. If the owner
does not want to purchase the equipment, consultants who
will provide the service are readily available.
Water-Jet Cleaning Results
Water-jet cleaning of bridge girders was demonstrated
under various conditions of water pressure and temperature.
The
chloride levels on the surfaces of the I-beams before and
after water-jet cleaning were measured using two different
sampling methods. In the first method, a plastic pipe elbow
with an o-ring that could be sealed to the painted surface
by hand pressure was used. The elbow was filled with 60 mL
of deionized water which was allowed to remain in contact
with approximately 15 cm2 of the surface for 30
s and then was poured into a plastic container by rotating
the fixture. This method was used to sample the surface of
the vertical portions (webs) of the I-beams before and after
water-jet cleaning.
In the second method, runoff water from the
water-jet cleaning operation was collected at the start of
cleaning and at the
end of the cleaning. The initial sample contained both soluble
and insoluble contaminants including loose rust, loose paint,
dirt, and salts coming from the horizontal flange areas as
well as from the vertical web area. The final sample contained
only soluble materials. The pH of the initial water runoff
sample taken under conditions of 17914 kfa (2600 lb/in2psi
water pressure and a starting water temperature of 99.9°C
(210°F) was 5.0 while that of the final run-off sample
at the end of the cleaning was 5.5. The water-sample analyses
are given in TABLE 3.
The chloride content of the starting water
was subtracted from each sample analysis and the results
were reported in TABLE
3 as milligrams of chloride per square foot of girder
surface. The results indicate that the plastic-elbow device
gives reasonably reproducible results since the two samples
taken before cleaning (1A and 1B) had essentially the same
chloride content as did the two samples (2A and 2B) taken
after cleaning. Since these samples were[B taken from a
vertical portion of the beam, there was little or no accumulation
of salt and dirt on the area sampled. Although the chloride
levels were rather low to start with, washing gave approximately
a 70 percent reduction in chloride.
The water-jet runoff samples contained much
more chloride, probably due to an accumulation of salt on
the I-beam flanges.
The chloride level in the initial sample was 12953 mg/m2 (1205
mg/ft2). Washing gave a 94 percent reduction
in chloride down to a final level of 120 mg/m2 (11.2
mg/ft2). The initial sample also was slightly
more acidic than the final sample. In other water-jet cleaning
trials conducted in this program and by IDOT, the use of
unheated water gave results very similar to those obtained
with hot water. As a result, IDOT has standardized on cleaning
with unheated water.
Air-Sampling Tests for VOCs
Generated During Grit Blasting Within Confinement
No volatile organic compounds (VOCs) were
found in a vacuum-bottle test of the atmosphere in the confinement
area during grit
blasting. In addition, no VOCs were collected on a TenaxTM adsorbent
tube through which the confinement area atmosphere was continuously
cycled during the blasting operation. This is the expected
result since there were no organic solvents involved and
no heat to cause thermal decomposition of the paint.
Three different air-sampling tests were conducted
within the confinement area during the steel-grit blasting
procedure.
The first test was a vacuum-bottle test in which the valve
on an evacuated sampling bottle was opened within the confinement
area during blasting to collect an air sample. The objective
was to determine whether or not the process produces any
significant quantity of VOCs. Analysis of this sample by
gas chromatography/mass spectroscopy (GCMS) did not indicate
the presence of any VOCs resulting from the blasting.
The second test also was a test to determine
whether or not VOCs are produced during the blasting process.
In this
test, a small pump was used to continuously draw air through
an adsorbent (TenaxTM) tube at 20 cm3/min.
Organic contaminants in the airstream are adsorbed and
later in the laboratory are driven from the adsorbent into
a GCMS
for analysis. This test is better suited to the determination
of small concentrations of VOCs than is the vacuum- bottle
test since a much larger volume of air is used in this
test and the VOCs are concentrated by adsorption. Nevertheless,
no VOCs were detected.
Airborne Grit-Blast Debris
Particulate Sampling and Analysis
The objective of this air-sampling test was to determine
the size distribution and composition of particulates in
the air within the confinement area during and after blasting.
Samples were collected by drawing ambient air through an
0.8-µm mixed cellulose ester (MCE) in-line filter at
a rate of 400 cm3/min for a total period of 90
min of which only the first 10 min was actual blasting time.
The remaining 80 min was time during which blast debris was
being vacuumed out of the confinement area. From the weight
of the collected residue and the flow of 400 cm3/min
over the full 90-min period, an average airborne particulate
concentration of 280 mg/m3 is calculated. The
residue was analyzed for particle-size distribution by the
Coulter technique and by an ElzoneTM particle-size
analyzer. Particulate size, shape, and composition were determined
by scanning electron microscopy (SEM) and energy-dispersive
x-ray (EDX) microanalysis. Particle-size analysis results
indicate the presence of many particles having diameters
less than the nominal 0.8-µm filter porosity are present
in the filter cake. This apparently is the result of partial
clogging of the filter porosity as the residue built up during
the test. The filter cake was thicker in the center of the
in-line filter due to the high airflow rate.
Coulter particle size analysis: Filter-cake
particulates were dispersed in a saline solution using an
anionic surfactant
to prevent the particles from agglomerating. The Coulter
Counter particle-sizing system allows for a measurement of
a size distribution between certain minimum and maximum particle
size diameters depending of the orifice sizes chosen. A range
of 2 to 60 µm was chosen for analysis because preliminary
SEM analysis indicated that the smaller particles tended
to be more rich in lead. FIGURES
19a and 19b depict the particle-size distribution obtained
for two samples taken from different areas on a single filter
cake. Sample #1 was taken from an area near the outer edge
of the filter while sample #2 was taken from an area near
the center of the filter. The plots in FIGURES
19a and 19b are most easily
understood by imagining the particles passing through a series
of sieves with mesh sizes indicated on the horizontal axis.
The height of each vertical line (or bar) represents the
number of particles of that size. Neither sample contained
many particles with diameters greater than 10 µm. The
plot in FIGURE 19a indicates that 50 percent of the particles
are smaller than 2.5 µm in size. The plot in FIGURE 19b indicates that of all the particles measured
in the 2 µm to 60 µm range, the largest number
occurred at the 2-µm cutoff. Since this strongly suggests
that a large portion of the particulates have diameters less
than 2-µm, samples were sent out for analysis by the
ElzoneTM system, which was capable of measuring
sizes down to well below 1 µm. These samples were representative
of the remainder of the filter cake after having removed
the two small samples analyzed in FIGURES
19a and 19b.
ElzoneTM particle size analysis:
The ElzoneTM technique
is similar to the Coulter technique. The difference was that
in the ElzoneTM case, a smaller orifice (capable
of detecting particle diameters down to 0.6 µm) was
used. Two samples were analyzed and the results are shown
as number percentages in FIGURES
20 and 21 and volume percents in FIGURES
22 and 23. For the sample
corresponding to FIGURE 20,
more than 50 percent of the particles detected had diameters
less than 0.8 µm and the highest number count occurred
at 0.65 µm. For the sample shown in FIGURE
21, more than 50 percent of the particles had diameters
less than 1.0 µm and the highest number counts occurred
between 0.65 µm and 0.73 µm. Although neither sample
contained a large number of particles with diameters greater
than 6 µm, the larger particles make up a significant
volume of the total particulate volume as shown in FIGURES
22 and 23. Note that the
sample in FIGURE 22 contained
a 10-µm latex marker.
The increased residence time of the smaller
particles aerosol degrades visibility and air quality within
the containment
zone and raises safety concerns as the particle size approaches
the filtering efficiency of different classes of respirators.
Neither the Coulter CounterTM nor the ElzoneTM
system were capable of determining whether or not particles
with diameters less than 0.3 µm (the porosity of HEPA
filters) were present.
SEM and EDX microanalysis: The air filter was
heavily loaded with particulates. Sample particles were prepared
for analysis
in the SEM using two different techniques. The first technique
entailed scooping particulates from the filter and dispersing
them in semiconductor-grade acetone. The suspension was
then transferred via pipette to a spectroscopically pure
carbon
plate. The second technique relied on electrically conductive
adhesive carbon tape that was gently pressed against the
surface of the filter and then mounted on a carbon plate.
Both techniques produced satisfactory dispersions as indicated
in FIGURES 24a and 24b that
show the acetone dispersion at x 200 and x 500 magnifications
and FIGURES 25a and 25b that
show the carbon tape dispersion at x 200 and x 250. The
SEM photomicrographs indicate that a wide variation is
particle
size is present in the airborne grit-blast debris.
FIGURES 26a and 26b show
the top surface of the filter cake together with its energy-dispersive
x-ray spectrum (EDX) and composition. The primary constituents
are lead and calcium from the paint, iron from the steel
grit and rust, and minor amounts of aluminum and magnesium
that may come from the paint or from residual dust. The detection
of silicon separately and in combination with aluminum is
nearly always due to ambient dust. Note that this x-ray spectrum
and the spectra that follow were acquired with an x-ray detector
that uses a beryllium window to insulate the x-ray detecting
crystal from the vacuum in the SEM. Although the beryllium
window ensures proper operation of the detecting crystal,
it completely absorbs low-energy x-rays from elements lighter
than sodium. Therefore, elements such as oxygen, nitrogen,
carbon, and hydrogen are not detectable. Three large fragments
of steel grit found in the filter cartridge are shown in FIGURES
27a and 27b. Note that no lead was detected on the steel grit
within the detection limits of the system. Smaller iron-containing
particles were also detected, typically in the size range
down to approximately 10 µm. Two such particles are
shown in FIGURES 28a, 28b, 29a,
and 29b. The iron-rich particle
shown in FIGURES 29a and 29b also contains lead and silicon. These two particles
may contain significant amounts of rust since they are morphologically
different compared with the steel grit in FIGURES 27a and 27b.
The majority of particles analyzed contained
lead, although some did not. Representative samples of particles
are shown
in FIGURES 30a, 30b, 31a, 31b, 32a, 32b, 33a, 33b, 34a, 34b, 35a,
and 35b. The particle in FIGURES
34a and 34b contains
significant amounts of lead and silicon together with iron,
magnesium,
calcium, and zinc. It may have come from the topcoat that
had dust blown on it before it had cured. FIGURES
35a and 35b show a particle
that is primarily zinc and probably originated from an
intermediate layer of paint between the primer and the
topcoat.
The proportion of particles that contain lead
as a major constituent appeared to increase with decreasing
size. Smaller
particles, down to approximately 1 µm, are shown in FIGURES
36a, 36b, 37a,37b, 38a, 38b, 39a,
and 39b. The angular morphology,
together with the cleavage step appearance of these smaller
particles, suggest they originate from a brittle layer, possibly
the primer that is fractured by impingement of steel grit. TABLE
4 gives the composition of various sizes of particles
collected on the filter within the grit-blast confinement
area. The figure number of the corresponding SEM and EDX
analyses also is listed in TABLE
4.
Effectiveness of Paint-Removal
Procedures
Three types of paint-removal methods were
compared in the demonstration. They were vacuum steel-grit
blasting, steel
grit blasting in full containment, and vacuum power-tool
cleaning. All blasting was done using G-40 steel grit. Two
vacuum power-tool system suppliers demonstrated the use of
their equipment. Needle guns, roto-peening, and abrasive-disc
cleaning tools were demonstrated by both companies. The debris
collected for analysis from each power-tool company's cleaning
work was a combination of debris from all three types of
tools.
A qualitative comparison of these methods is
presented in TABLE 5. It is
based solely on observations made during the demonstration.
Each method
has advantages and disadvantages and the method of choice
depends on the size, location, and geometry of the bridge
being refurbished, as well as the size of area to be coated,
degree of surface preparation required, etc.
JULIEN DUBUQUE BRIDGE AT DUBUQUE,
IOWA
Paint debris samples were obtained from a
paint-removal and recoating project on the Julien Dubuque
Bridge over the
Mississippi River between Dubuque, Iowa, and East Dubuque,
Illinois, done by the Iowa Department of Transportation.
The paint-removal method used was steel-grit blasting with
full containment. This required bringing a preconstructed
scaffolding under the bridge on a barge, raising the scaffolding
up under the bridge as shown in FIGURES 40a and 40b,
and putting the confinement structure in place. The grit-recovery
system consisted of vibratory grit separation, air washing,
and cyclone separation, but did not include a magnetic separation
step.
This bridge was originally painted in 1942
using a lead pigment. In 1976, the paint was removed from
all the easily
accessible areas by grit blasting and the bridge was recoated
using a zinc chromate pigment. Paint was removed from the
areas not blasted in 1976 using vacuum-blast equipment.
This paint debris is a combination of the original lead-based
paint and the 1976 zinc chromate system.
The following debris samples were collected:
(1) Wheelabrator GL40 steel grit, (2) used GL40 grit prior
to cleaning, (3)
paint waste separated from the used grit by the separation/recycling
system, (4) debris obtained by magnetic separation from
steel grit initially treated by the separation/recycling
system,
and (5), a sample of the vacuum blast debris described
above. No analysis was done on the first two samples. The
other
three samples were examined by SEM as shown in FIGURES
41a, 41b, and 41c and by
optical microscopy as shown in FIGURES
42a, 42b and 42c. They also
were analyzed by EDX and these results are shown in FIGURES 43a, 43b, and 43c.
The material magnetically separated from the
once-treated grit (see FIGURE 43a)
is very high in silicon and, in fact, appears to contain
glass or
sand particles. The other main component identified is iron
which came either from the steel blast media or from the
structure itself. The sample was extremely nonuniform in
composition and contained some fibrous materials (perhaps
wood). It appeared to contain only small amounts of actual
paint components such as red lead and zinc chromate. In fact,
several other elements such as aluminum, magnesium, calcium,
and titanium, were found in similar small concentrations.
Sieve analysis of this sample showed that despite its very
dusty brown appearance, less than 6 percent of the sample
by weight was smaller than 45 µm in diameter (passed
through a 325-mesh screen sieve).
The paint waste sample (FIGURE
43b) is very high in iron. This sample also contains
silicon, although at a much reduced level compared to the
magnetically separated sample described above. Lead, zinc,
and chromium as well as magnesium, aluminum, calcium, and
titanium were also found. The relatively low levels of
lead in the paint waste are consistent with the fact that
most of the lead had been previously removed. This sample
contained the largest amount of fines. More than 20 percent
of the sample passed through a 45 µm screen.
The vacuum blast debris sample (FIGURE
43c) contains many elements including lead, chromium,
iron, and zinc, that are present in relatively large amounts.
This analysis is consistent with the paint records that
show that both red-lead and zinc chromate pigments were
used in the paints. Again, the iron most likely comes from
steel or rust. This is the only paint-waste sample from
the Dubuque site that actually looks like it contains paint.
Some of these paint particles clearly show four layers
of paint under optical microscopic examination. Approximately
16 percent of this sample was finer than 325 mesh (45 µ m).
The fines (the portions that passed through
the 325-mesh sieve screen) from each of the above three samples
were also
examined by SEM. FIGURES 44a, 44b,
and 44c show the <45-µm
particles in the magnetically separated material at magnifications
of x 500, x 5,000, and x 50,000. In the x 500 photomicrograph,
particles ranging in size from less than 1 µm up to
about 30 µm are clearly discernable. The x 5,000 photo
shows a less than 0.5-µm particle nest to an approximately
10-µm-diameter particle. The x 50,000 photo shows only
the 0.4-µm-wide particle.
FIGURES 45a, 45b,
and 45c show a similar set
of photomicrographs for the paint-waste particles that passed
through the 325-mesh screen. FIGURES
46a, 46b, and 46c present
the results for the vacuum blast fines. In each of the three
sets of figures, the particle featured at x 50,000 can be
identified in the x 5,000 and x 500 photos upon close examination.
Particles smaller than 0.3 µm in diameter appear to
exist in each of the samples, but since this size is near
the limit of resolution of the instrument, further study
is needed to establish whether or not such small particles
are present, and if so, how many.
Two particles, approximately 0.5 µm in diameter, from
each of the above three samples were analyzed by EDX. The
results are presented in FIGURES
47a, 47b, 48a, 48b, 49a, and 49b. For the
paint-waste samples separated from the grit by air and magnetically,
one particle contained lead and the other did not. For the
vacuum blast debris, both particles contained lead. Again,
these results demonstrated that the fine particles produced
during grit-blasting operations contained lead. |
