University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Colorado School of Mines
|
the solid phase and less will be available in solution. The decreasing concentration of
ferric iron will limit the dissociation of pyrite.
Figure 2.7: pe vs pH diagram for iron(III) hydroxides. This represents the
precipitation of Fe(OH)3 (s) with an increase in pH
This use of lime or limestone in a closed system, like in the containment or
subsurface flow examples, is a treatment solution with a combination of remediation
benefits. The neutralization of mine tailings is considered temporary when a system is
open, however, in the case of using a liner and cap, this creates a closed system, which
can be thought of as a more permanent solution. This is because additional applications
are sometimes not necessary.
Oxidation may also me necessary if the iron is found in the dissolved form, Fe2+.
The soluble iron must be oxidized so that it loses an electron and forms Fe3+. As seen in
Figure 2.7 this requires raising the electro potential, which is commonly done by aerating
18
|
Colorado School of Mines
|
O through the contaminated water. This treatment technology commonly is used in
2
conjunction to remove manganese. However, this conventional treatment technology is
very costly and is also labor intensive if large mechanical devices are used. However, this
technology could be applicable if natural aeration, such as waterfalls, is sufficient enough
to oxidize most of the soluble iron or manganese.
The final chemical treatment technology is the use of ion exchange. This
technology commonly used resin beads that are covered in sodium ions. The iron, in this
example, exchange places with sodium. The beads are either recharged or discarded.
2.4.4 Biological Water Treatment
Reasons why microorganisms are intriguing when treating MIW is because of the
potential reduction in manual labor compared to conventional treatment technologies.
The use of wetlands, with the purpose to increase the hydraulic detention time, is one
technology has holds value for the ability to treatment MIW passively. Wetlands have
often shown considerable promise at pilot scale experiments, however, when
implemented at full scale sizes, they often fail. The second treatment technology is to
employ biochemical reactors to provide sulfate reducing bacteria. These reactors are
filled with substrates such as woodchips, hay, and straw, the more simple the substrate
the more bioavailable the nutrients. This is also a passive system that has high promise
when it comes to cheap and efficient strategies. Again, there are problems that occur,
debris and precipitates from the formation of iron hydroxides clog the pipes and conduits
in these systems. Therefore, to solve these issues, more unforeseen labor is often needed
to keep the reactor operating. These passive technologies still have great promise in the
future; however, a greater understanding of the microbial communities and interactions
with substrates is needed.
When considering all these strategies, this project focuses on the stabilization and
solidification of mine waste in cementitious material. The chemistry behind the setting of
cement paste and the leachability of the hardened material are the factors that are
investigated. The compressive strength and the concentrations of heavy metals are
indications of the level of degradation that may occur when mine tailings are substituted
for a portion of the fine aggregate.
19
|
Colorado School of Mines
|
2.5 Hydration Process
The reaction between cement, such as OPC, and water creates a hardened material
that can be used to bind pieces of aggregate together to make concrete. There is a
transition zone between the cemented paste and the aggregate. This layer that is between
10-15 µm in thickness is a collection of crystals and hydrated calcium oxide. This layer is
weaker compared to the cemented paste and the aggregate, therefore it is often studied
because it is the weakest link in this construction material.
The formation of cemented paste, the silicates and the aluminates from the cement
react with water in two ways. The first is a true hydration process where the addition of
water molecules directly reacts with cementitious material and creates theoretical
componds: tricalcium silicate (C S), dicalcium silicate (C S), tricalcium aluminate (C A),
3 2 3
and tetracalcium aluminoferrite (C AF), (Neville 1996). These reactions are exothermic.
4
The second is a process called hydrolysis.
The hydrated calcium silicate structure can encapsulate heavy metals, Figure 2.8.
During the setting process the cemented paste becomes less porous and less permeable.
This is one reason why this material can be used as a form of encapsulation of unwanted
wastes. Another reason why the solidification and stabilization process is intriguing is
because of the alkaline environment, usually found to be around pH 12 of higher, is an
environment that microbes have a very difficult time surviving.
In the case of cemented paste backfill, the production of dissolved heavy metals
due to MIW has been thoroughly studied previously by other researchers. Different
cementitous material can be used, such as OPC, cement kiln dust, and fly ash, (Nehdi and
Tariq 2008). It has been found that cement kiln dust and fly ash are more effective at
reducing the leaching of heavy metals from hardened material, (Park 2000), while also
increasing the setting time. However, in this study the use of OPC is the material of
choice.
The generation of sulfuric acid will degrade the solidified material over time and
cause corrosion and eventual failure. The amount of pyrite found in the samples will help
determine the potential to produce sulfuric acid. Depending on the mill processing
methods, often times lime or calcium carbonate can be introduced for the purposes of
limiting the production of MIW.
20
|
Colorado School of Mines
|
The corrosion of concrete may not occur in the long run if the mine tailings
potential to generate acid is overcome by the neutralizing capacity from the alkali
material found in the cementitious material. When considering the addition of aggregate
for the purposes of structural concrete, there is another interface at which the cement
paste may interact. Therefore, the transition zone maybe an important area of
encapsulation, not only the bulk cemented paste.
Figure 2.8: The encapsulation of metals in portland cement, (Conner 1990)
2.6 Leachability
The process at which water travels through a solid matrix and dissolves analytes is
considered leaching. A common method of visualizing the flow of contaminants in a
solid matrix is with the use of a conceptual model. The molecular diffusion due to
random Brownian motion can be used to create an analytical model. A bulk diffusion
model and shrinking unreacted core leaching procedure are sources of future study. These
models can be found in more detail in the recommended Future Research chapter.
The first most agreed upon explanation for the immobilization of hazardous
wastes is the chemical bonding within the solid matrix. This type of fixation denotes the
waste has gone under sometime type of alteration. The chemical changes can undergo
oxidation/reduction reactions, acid/base reactions, and sorption. Chemisorption is
stronger than true adsorption at the surface due to the lack of a boundary between the
21
|
Colorado School of Mines
|
waste and the surface, (Conner 1990). The most often used sorption material is activated
carbon, sodium silicates, and gypsum, (Wiles 1987). The process of ion exchange is
another factor which dictates chemical fixation. Ion exchange can sometimes be
incorporated in the sorption process, (Conner 1990). The second most agreed upon
explanation for the stabilizing and solidifying of hazardous wastes is encapsulation. This
physical mechanism can be separated into three different processes: micro-encapsulation,
macro-encapsulation, and embeddment, (Conner 1990). For wastes to leach into solution
there must be a solubilization (dissolution is probablily a better term to use) process to
occur. The most dominate immobilization mechanism will be the one retaining the waste
in the solid matrix. One way to predict the leachability is to create a mechanistic model.
Figure 2.9: The release of iron(III) hydroxides into solution from an OPC and blast
furnace slag cement pastefill, (Benzaazoua, et al. 2004)
It is important to understand if the metals from the finished product will leach out
into solution, thus creating MIW under oxidizing conditions. A common way of
preparing a leaching experiment is through a kinetic test using a soxhlet “reactor”,
(Hornberger and Brady 1998). This device is what Benzaazoua and his collogues used in
their experiments for estimating the leaching of metals from paste backfills with varying
cementitious material (Benzaazoua, et al. 2004).
22
|
Colorado School of Mines
|
Shake flask extractions are also common with conducting leaching tests, Figure 2.9
shows concentrations of iron found in solution during the leaching process for both OPC
and slag in paste backfill. It is important to note leaching was conducted from the
concrete, not from the cement. Even though this is not a substantial finding, due to the
low number of data points, it is an indicator OPC might not be the best cementitious
material.
There is a difference between the leaching occurring in the real world and the
leaching occurring in the laboratory. The real world conditions are not 100%
reproducible because the physical and chemical conditions are never exactly the same.
There are three main different types of kinetic leaching tests which mimic the real world.
First, the extraction test, such as the extraction procedure (EP) and the toxicity
characteristic leachate procedure (TCLP), overestimate the leachability of heavy metals
from such substrates as concrete. The EP test was the predecessor to the TCLP test,
where the TCLP test was designed to handle heavy metals, pesticides, semi-volatile
compounds, and volatile organic compounds (VOCs), (Shively and Crawford 1989).
However, both tests predict similar leaching rates when implemented properly, (Chang, et
al. 2001). The physical properties in the real world are not identical to the EP or the
TCLP test because the sample is crushed and grinded. The surface area of the sample is
usually larger and the particle size distribution is different, (Conner 1990). Close to the
total amount of heavy metals in the sample can be calculated, however, this is not an
accurate prediction for the actual amount contained in the leachate in the real world.
The second testing method, semi-dynamic batch leaching, is often used because the
system replaces the leachate with fresh extraction fluid periodically. The method also
mimics the real world better than batch extraction testing because the sample is not
ground into smaller pieces. The laboratory tests such as the NEN 7345, (Malviya and
Chaudhary 2006), and ARNI 16.1, (Baker and Bishop 1997), are both semi-dynamic
batch tests. The overall time for the semi-dynamic batch leach procedures is much longer
than a TCLP test because the surface area/volume ratio is much lower. The main factor
responsible for the rate of leaching is the permeability of the solid matrix, which has a
range from 10-5 – 10-8 cm/s, (Conner 1990).
23
|
Colorado School of Mines
|
The third type of test is the column leaching test. This type of test implements a
continuous flow of leachant through crushed solidified wastes. The soxhlet reactor, which
was mentioned before, conducts a column test. The device mimics how water runs over
mine tailings pile during a rain storm. The benefit when using the column leach test is the
compatibility for organic wastes such as trichloroethylene (TCE) and phenol, (Kolvites
and Bishop 1989). In this study a bulk diffusion model was used to predict the
leachability of the organic wastes.
The waste which leaches out of the solid matrix is due to a process called
decalcification, (Bonen and Sarkar 1995). Ground and surface water runs over the
stabilized/solidified solid, breaking down the silica/calcium-oxide shell structure,
encapsulating and chemically fixates the wastes. Calcium leaches out into solution as the
concentration of heavy metals increase, Figure 2.10, because the solid matrix becomes a
silica skeleton that does not retain its structural integrity. Therefore, the capacity
decreases for the pour structure to encapsulate and chemically fixate wastes.
Decalcification, Figure 2.10, takes place due to the formation of bicarbonate in the
pore water. Bicarbonate comes from carbon dioxide and calcium carbonate dissolving in
the water. The three phase system will create the following reaction, (Bonen and Sarkar
1995), where C-S-H, the solid matrix of the cement paste, is represented as
CaO·SiO ·H O:
2 2
→ (2.6)
( )
The decalcification process can be thought of as the reciprocal of the process of
cement hardening, otherwise known as hydration. The pH of an inorganic cement
polymer is between 13 and 14, (Jackman and Powell 1991). This makes a strong case for
the lack of microbial activity if mine tailings are used in the cement. But, more
importantly the high pH is lowered through the process of decalcification making the
waste susceptible to microbial activity and thus why the creation of MIW becomes a
concern.
24
|
Colorado School of Mines
|
Figure 2.10: The process of decalcification, simulating the leaching of heavy
metals, (Lin and Huang 1994)
In the case of creating MIW from mine tailings encapsulated in concrete, the main
contaminants of concern are inorganic constituents such as: arsenic, cadmium, copper,
lead, mercury, nickel, selenium, silver, thallium, and zinc.
2.7 Regulatory Issues
The USEPA regulates the maximum contaminant limits (MCL) for inorganics, in
this case the interest lies in heavy metals, (USEPA 2012). The national regulations for
drinking water are often more stringent than the nationally recommended water quality
criteria for human and aquatic life, Table 2.1. However, the fresh water concentrations
for these recommended standards are sometimes higher than drinking water standards.
Taking this into account, the EPA notes that these criteria are not enforceable. The
criterion fall under two categories, the criterion maximum concentration (CMC) and
criterion continuous concentration (CCC), (USEPA 2009). Please note that the MCLs for
silver and zinc are secondary drinking water regulations, these are not enforceable.
The concentrations for the water quality criteria are posted as dissolved metals
concentrations; they were adjusted with a conversion factor. The conversion factor for
cadmium and lead are dependent on the hardness of the water. Therefore, the
concentration of calcium and magnesium will dictate these concentrations. However,
natural organic matter (NOM) will also react with dissolved metals, (Schwarzenbach,
Gschwend and Imboden 2005).
25
|
Colorado School of Mines
|
The junction between government, industry, and the world of academia are
important for the re-use of mine waste for a construction product because of the potential
to be used worldwide. Making sure the waste stream is regulated and implemented
properly is an issue just as important as insuring concrete is designed properly. In regards
to the United States, the American Concrete Institute (ACI) would need new regulations
for concrete with portions of tailings are substituted for fine aggregate. There have been
studies that have focused on various fly-ash geopolymers and looking at properties such
as: static elastic modulus, Poisson’s ratio, compressive strength, and flexural strength
(Diaz-Loya, Allouche and Vaidya 2011). These properties need to be measured for
various mine tailings concrete specimens so that they can related to values found in
normal concrete mixes. Therefore, there can be established relationships between the
amounts of mine tailings of certain chemical characteristics that can be added for specific
construction material purposes. For example, if it found that the leaching of heavy metals
of mine tailings concrete specimens that are composed of 30% sulfide bearing minerals is
not potentially hazardous, then this material can only be used for interior use, minimizes
the contact with water from the natural environment. It is important to mention that
governments will need to work side by side with industry and academic partners.
Table 2.1: USEPA Drinking Water and Aquatic Life Criteria Standards
Drinking Water
Water Quality Criteria (Fresh Water)
Metal Standards
MCL, mg/L CMC, mg/L CCC, mg/L
Arsenic 0.010 0.340 0.150
Cadmium 0.005 0.002 0.00025
Chromium III - 0.570 0.074
Chromium VI - 0.016 0.011
Chromium Total 0.1 - -
Copper 1.3 0.013 0.090
Lead 0.015 0.065 0.0025
Mercury 0.002 0.0014 0.00077
Nickel - 0.470 0.052
Selenium 0.05 - 0.005
Silver 0.1 0.0032 -
Thallium 0.002 - -
Zinc 5 0.120 0.120
26
|
Colorado School of Mines
|
CHAPTER 3
STATEMENT OF EXPERIMENTAL OBJECTIVES
The re-use of mine tailings with the purpose of creating structural concrete is a
strategy for reducing the amount of mine waste that ends up being disposed, thus
reducing the environmental impact of one of the largest sources of waste in the world.
There could be a potential source of revenue generation, reduction of the carbon foot
print associated with the mining of aggregates for mining, build and design firms, and
concrete companies. To examine the feasibility of the re-use of mine tailings, hypotheses
are tested by accomplishing the following objectives:
1. Examining the compressive strength of concrete cylinders with and without mine
tailings substituted for the fine aggregate. The first criterion is that the compressive
strength with tailings substituted for the fine aggregate has at least the same
compressive strength as the control concrete sample. The second criterion set as a
goal is that concrete specimens with mine tailings meet the design parameters as
specified in Table 3.1.
Table 3.1: Design criteria for structural concrete
Design Parameter Goal
Uniaxial compressive strength 4000 psi
Slump 3-4”
2. The application of leaching tests to simulate the weathering due to acid rain and
alkaline conditions. When using batch leach extraction methods, the leachate
concentration of heavy metals will not be above the regulatory standards.
Table 3.2: Design criteria for leaching tests
Design Parameter Goal
Reduced particle size of concrete specimens 2-5 mm
pH = 4.2, 60/40 sulfuric/nitric acid
Extraction fluid pH = 5.0, 60/40 sulfuric/nitric acid
pH = 10, NaOH
27
|
Colorado School of Mines
|
The uniaxial compressive strength tests are completed on small batches. The curing
time depends on what parameters are of interest. For example, a study can be conducted
to look at different tailing to fine aggregates (T:FA) ratios to see if there is an optimal
amount of mine tailings that should be added to a batch of concrete. Or, a study can be
utilized to look at how the strength and leachability changes over time.
The goal of manufacturing concrete with compressive strength consistently over
4,000 psi is done by designing a control concrete mix that meets the specifications by
conducting and undergoing iterative pours and adjustment of the concrete mix design.
The batch leach shake extraction tests are modified TCLPs, which accelerate
weathering process due to the reduced particle size of the contaminated material. By
reducing the size of the particles, the surface to volume ratio increases, thus exposing
more of the contaminated material to the acid conditions of the extraction fluid. The pH
of the extraction fluid varies, simulating different regions of the United States. The East
Coast has rain water that is often times more acid that rain water found West of the
Mississippi. This is due to the higher concentration of sulfur dioxide from industrial
process, mainly the off gas from coal fired power plants. A more alkaline extraction fluid
is also used, this if for testing the leachability of amphoteric metals.
28
|
Colorado School of Mines
|
CHAPTER 4
EXPERIMENTAL PROCEDURE
The four components that make up the concrete mix design for this project are:
cement, coarse aggregate (gravel), fine aggregate (sand and mine tailings), and water.
The cement was donated by Quikrete, a manufacturing company located in Denver, CO,
where 418 lb of type I/II OPC was transported to the Colorado School of Mines (CSM).
Roughly 1250 lb of gravel and 250 lb of sand was donated by Metro Mix, LLC in
Denver, CO. The gravel and sand were shoveled into 55 gallon drums and transferred to
CSM campus. The aggregate was placed in a metal trough, where polyethylene lined the
bottom of the container. A wood separator was placed between the gravel and sand so
that cross contamination was minimized. The mine tailings were collected from a TSF in
Silverton, CO.
Mine tailings were collected from the Pride of the West TSF in Silverton, CO. There
is one active storage area that splits the two inactive areas, Figure 4.1. The samples were
collected from the Southern inactive storage facility, which is highlighted in O. During
the field sampling, the Operations Director, John Ferguson, mentions that the holding
pond East of the TSF has a pH around 4.5. The stream’s yellow/orange color was
observed during the time of sampling, indicating a high concentration of iron(III)
hydroxide precipitates.
Seven 10’ x 10’ sampling plots were constructed. Random sampling was conducted
so that the composite sample collected was representative of the 10’ x 10’ plot. The
method of random sampling was the use of a blind throw of a stone over the shoulder. If
the stone landed within the 10’ x 10’ plot, the stone was removed and a shovel full of
tailings was collected. Three large stainless steel, acid washed, shovels were was used to
excavate the mine tailings. If the stone landed outside the plot, the stone was thrown
again. If a stone landed in a previously dug hole, the stone was thrown again. This
process was repeated until a 5 gallon, acid washed, bucket was filled, between 8 to 10
stone throws were required to fill one bucket. The plots were spaced to test if the particle
sizes of mine tailings changed over the distances sampled. After the seven plots were
sampled, seven breathable sampling bags were filled with a small stainless steel, acid
washed shovel.
29
|
Colorado School of Mines
|
Spigot
Figure 4.1: Sampling plots at Pride of the West TSF - Silverton, CO
Before placing mine tailings in their respective 5 gallon buckets, each 5 gallon
bucket and lid was cleaned with a rag to remove any loose debris, rinsed once with tap
water, acid washed with one rinse of 10% hydrochloric acid, and rinsed three times with
RO water from a reverse osmosis (RO) filter. The hydrochloric acid was purchased from
JT Baker Chemical Company and is 37.7% by weight. The Pure-FlowII reverse osmosis
filter was purchased from Coralife and has a 1 µm filter cartridge and a carbon block
water filter. The RO water meets the Specification for Reagent Water (ASTM Standard
D1193-06 2011).
After the buckets and lids were cleaned they were air dried. This method was also
used to clean the four shovels that were used for excavation. When acid washing
laboratory equipment; however, substitutions were made for the hydrochloric acid and
RO water with nitric acid, sulfuric acid, and deionized water respectively.
The sieving process was conducted at the CSM campus. A 3’ x 3’ sieve was
constructed out of wood slats, screen door mesh and screws. The screen mesh had
dimensions of 1.5 mm x 2 mm. Each 5 gallon bucket was sieved and mixed by hand with
a small stainless steel shovel to homogenize the sample. The mixing occurred in a
30
|
Colorado School of Mines
|
cleaned concrete mixer drum. The tailings were then placed back into their respective
buckets.
4.1 Physical Characterization
To meet the objectives of this project the preparation, testing and analysis were
organized so that experiments were performed and data was collected in systematically.
The first set of experiments was to collect the necessary physical parameters. Each of the
seven mine tailing samples were subjected to seven physical characterization tests.
4.1.1 Particle Size Distribution
The first experiment that was conducted was estimating the particle size distribution
of the fine aggregates. The purpose is to observe the distribution, calculate the fineness
modulus, and estimate the specific surface area (SSA).
A dry sieve approach was used for the larger particles, whereas a hydrometer test
was conducted for the fines, smaller than 0.074 mm. The set of sieves that were used for
the dry sieve are: #10, 20, 30, 40, 50, 80, 100, and 200. The mass of each sieve was
recorded prior to the experiment. The balance that was used for all the physical
characterization and concrete mixing was the RADWAD, model WLC 10/A2 balance,
with a maximum capacity of 10 kg and a resolution of 0.01 g. After the sieves were
weighed and recorded a homogenized fine aggregate sample was weighted on the
RADWAD balance. The set of sieves were placed on a mechanical shaker, and the
weighed sample was slowly poured into the staked sieves. A cap was placed on top of the
staked sieves and the shaker was operated for 10 minutes. After 10 minutes, each sieve
was re-weighted then emptied and cleaned with a brush. Everything that past the #200
sieve underwent a hydrometer test. The hydrometer type used to perform each
measurement was 152H.
To perform a hydrometer test the following materials were used: two 1000 mL
sedimentation graduated cylinders, rubber stopper, malt mixer, mixing cup, deflocculated
agent , stop watch, hydrometer, wash bottle, and RO water. Each component was rinsed
and cleaned before each experiment. The deflocculated agent that was used was a reagent
grade salt, hexametaphosphate. This salt was slowly mixed with RO water at a
concentration of 40 g/L. The procedure for conducting the hydrometers experiments meet
the Test Method for Particle-Size Analysis of Soils (ASTM Standard D422-63 2007)
31
|
Colorado School of Mines
|
where the geometry of the 152H hydrometer is shown in Figure 4.2. The hydrometer is
slowly placed in the sedimentation column filled with well dispersed sediment sample,
Figure 4.2: 152H Hydrometer
which was mixed for 4 minutes, and 125 mL of deflocculated solution. The hydrometer
reading occurs at time intervals, which are recorded in minutes. At this time the
temperature is also recorded. The effective depth of the hydrometer can be calculated
from the following equation:
[ ( )] (4.1)
Where, V is the volume of the hydrometer bulb, which equals 67.0 cm3 and A is
B
the cross sectional area of the sedimentation cylinder, which equals 27.8 cm3. L equals
2
27.0 cm3 and L is the distance from the top of the bulb to the reading of the hydrometer.
1
The tabulated data for calculating the effective depth can be found in is shown in ASTM
D422 standard. It is important to note that the effective depth has to be corrected because
of the variation due to the meniscus, the zero effect, and temperature.
After each reading the hydrometer is placed in the second sedimentation cylinder,
which is filled with RO water. The experiment is completed after 1.5 hours after the
32
|
Colorado School of Mines
|
correcting the depth of the hydrometer reading during the particle size distribution
experiments. Second, the specific gravity is used to calculate the amount of mine tailings
that could substituted for sand within the concrete mix.
Every ingredient in the concrete mix design was proportioned by weight except the
sand, which was estimated by volume. To perform the specific gravity experiment, two
pycnometers, two rubber stoppers, vacuum pump, RADWAD balance, and water were
used. The vacuum pump is manufactured by Robinair Vacumaster, Model 15400, with a
capacity of 4 CFM with a 1/3 HP motor. An adapter was used to split the vacuum line in
two. The two pycnometers were connected to the pump for 1 hour. As the vacuum pump
ran, the pycnometer was shaken side to side gently to help with removing the trapped air.
These experiments were conducted in duplicate.
The specific gravity was calculated from the following equation,:
(4.5)
( ) ( )
Where, M is the mass of empty, clean pycnometer, M is the mass of empty
P P
pycnometer plus the dry tailings, M is the mass of pycnometer plus the water, and M is
A B
the mass of pycnometer plus the dry tailings MIW the water.
4.1.4 Bulk Density
The purpose of calculating the bulk density was to estimate the contribution of the
coarse aggregate. Unlike the specific gravity of the gravel, the bulk density includes the
void volumes of air between the pieces of coarse aggregate. The reason why the bulk
density is used instead of the specific gravity is because the voids with in the aggregate
are still present when the concrete is curing; these voids fill up with the cement paste and
fine aggregates.
To measure this parameter the Test Method for Bulk Density (“Unit Weight”) and
Voids in Aggregates (ASTM Standard C29/C29M-09 2009) was followed. To perform
this experiment a container of a known volume, RADWAD balance, and water were
used. The container was filled evenly three times; each portion was rodded 25 times
before adding the next portion of gravel. The rodding technique can be compared to
34
|
Colorado School of Mines
|
jigging and shoveling, where rodding is the process of tapping the side of the container so
that the aggregate pieces are compacted, thusly reducing the void space. This procedure
was done two more times so that the container was full, where the top voids in the
container roughly equaled the parts of the gravel that stuck above the container. The
weight of the gravel was calculated and then divided by the known volume. This
procedure was conducted in triplicate.
4.1.5 Moisture Content
The purpose of measuring the moisture content of the aggregates is to gain an
understanding of how much water is retained in the fine aggregates. This amount of water
needs to be taken into consideration when measuring the weight of the fine aggregates.
This experiment requires the use of an oven operating at 105 °C, moisture cans,
polyethylene drying sheet, and water. The procedure of this experiment meets the Test
Methods for Laboratory Determination of Water (Moisture) Content of Soil and Rock by
Mass (ASTM Standard D2216-10 2010). The cans were pre-weighted before the
experiment and the fine aggregate were prepared so that surface dry conditions were met.
The samples were placed in the oven for 24 hours and then the moistures cans were re-
weighted. The moisture content experiments were calculated in duplicate by the
following equation:
( )
(4.6)
Where, S is the weight of the moist sample and A is the weight of the oven dry
sample.
4.1.6 Atterberg Limits
The purpose of performing Atterberg Limit tests is to measure the plastic limit,
liquid limit, and plasticity index of the finer particles in mine tailings and sand. These
limits are based on the moisture content of the fine aggregates. The plastic limit, liquid
limit, and plasticity index can be related to the amount of clay or silt that is present in the
sample. The equipment needed to perform Atterberg Limit tests are: liquid limit device,
evaporating dish, standard groove tool, balance, moisture cans, spatula, plate glass, wash
35
|
Colorado School of Mines
|
bottle, purified water, and drying oven. The liquid limit device that was used is
manufactured by ELE International and is a part of the Soiltest Production Division. The
base rises 10 mm above the rubber base and the cup moves with the utilization of an
electric motor. The cup is dropped onto the rubber base twice per second. The detailed
procedure of conducting these experiments can be found in the Test Methods for Liquid
Limit, Plastic Limit, and Plasticity Index of Soils (ASTM Standard D4318-10 2012). The
experiment was not conducted in duplicate due to the success of the experiment. The
outcomes of this experiment are further explained in the results and discussion.
4.2 Concrete Encapsulation
After the physical characterization, the seven mine tailings samples were mixed
homogeneously so that a batch of finer mine tailings and one batch of coarser mine
tailings were created. The coarsest mine tailing samples, 1, 2, 3, and 4 were combined in
equal proportions, by weight. The finest mine tailing samples, 4, 5, 6, and 7 were
combined in equal proportions, by weight.
The objective of this second section of systematic experiments is to create
structural concrete that meets strength standards. The mine tailings are substituted for
standard fine aggregate. Two sets of experiments were conducted. The first, was intended
to understand how the tailings to fine aggregate (T:FA) ratio affects the unconfined
compressive strength and the leachability. The variation was 10%, 30%, and 50% of mine
tailings to total fine aggregate. The second was an experiment examining the unconfined
compressive strength and the leachability after the concrete specimens cured for 3 days, 7
days, 14 days, and 28 days. The mine T:FA ratio was 30% for the second experiment.
Before making and pouring the concrete can occur, the concrete mix was designed.
The reason why the experiments lasted 28 days is because that is the time it takes
for the capillaries to mature and become segmented, (Neville 1996), for a water to cement
ratio equal to 0.57. This means that the water that was introduced to the cement has
hydrated the significant portion of the calcium silicates and aluminates. There will be
some pore water left, but the reaction is considered completed for the water to cement
ratio stated above. If the water to cement ratio was above 0.70 there would never be
segmented capillaries and water would have would be able to travel through the
cemented paste.
36
|
Colorado School of Mines
|
4.2.1 Mix Design
To meet the first objective, non-air entrained, structural concrete must be constructed
to withstand 3,500 psi or pressure. Therefore a 4000 psi concrete with a 3-4” slump was
chosen as design parameters. Important constraints such as the largest size of the coarse
aggregate, fineness modulus of fine aggregate, specific gravity, and the bulk density of
the coarse aggregate, will dictate the proportions of the four ingredients. The concrete
mix design parameters were selected from the “Design and Control of Concrete
Mixtures” (Kosmatka S.H. and W.C. 2003). For example, to design a 4000 psi concrete
with a 3-4” slump, the water to cement ratio must be equal to 0.57. The amount of water
per yard was also estimated, which for a 1.5” coarse aggregate equals 300 lb/yd3 for a
3-4” slump. However, it was later found that the coarse aggregate was better represented
by a value equal to 325 lb/ yd3. The bulk volume of dry rodded coarse aggregate per unit
volume of concrete was estimated from the fineness modulus. This value changes
depending on the fineness modulus and nominal maximum size of the gravel, (Kosmatka
S.H. and W.C. 2003).
These steps enable the proportioning of water, cement, and coarse aggregate by
weight. The proportions of the four ingredients for the T:FA variation and strength
development studies can be found in the results and discussion section. To solve for the
amount of fine aggregate needed, the T:FA ratio was used in conjunction with the
specific gravity of each component. This volumetric approach is more accurate then
directly calculating the mass of sand needed because the volume of the concrete
specimen is known. The mixing of concrete specimens was conducted after an initial
design was completed and the mix design was refined in an iterative process.
4.2.2 Making and Curing
Preparing the ingredients before mixing concrete batches is necessary to insure
proper conditions. For example, the cement was kept in dry conditions to make sure that
no large clumps consolidated in the concrete batches. The mine tailings were stored in
5 gallon buckets, keeping in moisture and keeping out contaminants. The sand and gravel
were kept in dry conditions. Three days before mixing, the sand is kept at saturated
conditions. The sand is placed on top of a tarp and the other end of the tarp covers the top
of the sand. On the third day the top portion of the tarp is removed so that the sand may
37
|
Colorado School of Mines
|
become surface dry. The coarse aggregate undergo similar preparations, however, instead
of absorbing water on top of a tarp, the gravel is placed in large clean containers filled
with water. The gravel is allowed to absorb water for two days and on the third day the
gravel is set on the same tarp as the sand, allowing surface dry conditions.
The rotary concrete mixer, manufactured by ELE, has a capacity of 3.5 ft3.
Therefore, each mix can be completed in one batch. The batches are made under the
accordance of the Practice of Making and Curing Concrete Specimens in the Laboratory
(ASTM Standard C192/C192M-12 2012). Each ingredient is weighed with the
RADWAD balance. First, the fine aggregate is added to the mixer and then mixer is
turned on. Then, the water is added. After the fine aggregates and water are
homogenized, the cement is slowly added, preventing fine particles emissions. After
well-mixed conditions the coarse aggregate is slowly added. Once this mixture is
homogenized the concrete mixer is turned off.
Each batch undergoes a slump test, where the slump cone meets the Test Method for
Slump of Hydraulic-Cement Concrete (ASTM Standard C143/C143M-12 2012). The
slump cone is filled in three even portions, where each portion is tamped 25 times and
rodded 20 times. After the slump cone is filled, the top of the cone is striked. The cone is
then slowly removed in a vertical fashion. The slump is recorded.
After the slump is complete, three cylindrical molds are filled. The cylinders
dimensions are 4” in diameter and 8” tall. The cylinders undergo similar tamping and
rodding as the slump cone, however, the cylinders are poured in two even portions
instead of three. The cylinders are then striked and capped. The caps are 4” diameter
plastic and tightly cover the cylinders so that no moisture escapes.
In between each batch, the concrete mixer is cleaned and all other tools that are used
during the mixing and pouring. The cylinders are transported as little possible and left to
cure for their designated time periods.
4.2.3 Unconfined Compressive Strength (UCS)
Once the curing time has been achieved the concrete cylinders undergo UCS tests.
The reason for conducting UCS tests is to understand how the T:FA ratio affects the
strength of concrete. The uniaxial compressive strength indicates if the concrete batches
meet the structural concrete requirements. The cylinders are prepared by first removing
38
|
Colorado School of Mines
|
the cylinders from their molds. A sharp knife is used to cut away the cap. Once the cap is
removed, compressed air is used to remove the cylinders from their molds. If the
cylinders do not lie flat, they are carefully sanded up to 3 mm. The height is measured
twice, 180° separating each measurement. The diameter of the cylinder is measured with
combination square in two places, at the midpoint between the top and the bottom at 90°
sections. The cylinder is then weighed and the density is calculated from the previously
measured dimensions. If the dimensions and the density of the concrete cylinder meet the
requirements from the Test Method for Compressive Strength of Cylindrical Concrete
Specimens (ASTM Standard C39/39M-12a 2012); then cylinder undergoes a UCS test.
After the maximum load is reached and has decreased by 5% or more, the test is over.
The cylinder is observed for facture type, segregation, large air voids, fractures that pass
through aggregate. The UCS test equipment, Figure 4.3, that was used is an ELE
International AccuTek 250 with a 100 kN maximum load capacity.
Figure 4.3: UCS test equipment
39
|
Colorado School of Mines
|
4.2.4 Reduced Particle Size Fractionation
After the maximum load was measured during the UCS tests, each concrete
specimen is reduced in size so that the shake extraction test could be conducted. The size
fractionation was completed so that the target range of 2-5 mm diameter particles was
captured. At first a hammer and chisel were used to break the concrete specimens into
smaller chunks, small enough to fit into the jaw crusher, Figure 4.4, which is stored in the
Mineral Processing Lab at CSM. The initial fine particles were separated from the rest of
the 2 inch chunks with a 3mm sieve. After the initial sieving, the concrete pieces were fed
into the jaw crusher. The jaw crusher was modified with shims so that the size of the
mouth of the jaw crusher was slightly larger than 5 mm. After the specimen was crushed,
the particles were placed in a sieve set that consisted of a 3/8 in a No. 4, 6, 8, and 10
sieves. The sieve set was placed on a Ro-Tap, Figure 4.5, for 6 minutes.
Figure 4.4: Jaw crusher
40
|
Colorado School of Mines
|
Each sieve was weighed on a Sartorius Universal balance, with a resolution of 0.1g,
before and after the shaking process. The particles were separated according to size. The
reduced size particles were combined with samples that originated from the same
concrete batches. In these experiments three cylinders were poured from a single batch,
therefore, three 2-5 mm fractionations were combined for future chemical
characterization.
Figure 4.5: Ro-tap
4.3 Chemical Characterization
The third stage of systematic experiments is the chemical characterization of the
mine tailings with the use of shake extraction tests. The significance of conducting
leaching tests on the mine tailings and concrete specimens is to observe the acid
generating potential and the ability for concrete to encapsulate heavy metals.
4.3.1 Synthetic Precipitation Leaching Procedure (SPLP)
The extraction fluid used during this experiment is simulated acid rain and alkaline
water. The pH values of the extraction fluid used during these experiments were 4.2, 5.0,
and 10. The acidic extraction fluids were synthesized by diluting a mixture of 60%
sulfuric acid to 40% nitric acid, by weight. The nitric acid was purchased by Macron Fine
Chemicals and is 68.010% by weight. The sulfuric acid was purchased by EMD and is
41
|
Colorado School of Mines
|
95.0-98.0% by weight. The basic extraction fluid is synthesized from diluting NaOH
pellets. The NaOH pellets were purchased from Macron Fine Chemicals.
The materials used for this experiment are: a balance, 1000 mL beaker, 500 mL
beaker, crucible, soil scoop, magnetic Teflon stir bar, pipets, pH meter, wash bottle,
deionized water, rotary agitator, 125 mL plastic extraction vessels, 2 L filter flask, 0.7 µm
glass microfiber filter, clamp, vacuum pump, test tube holder, and 20 and 40 mL glass
vials.
The rotary agitator, Figure 4.6, is manufactured by Associated Design and MFG
Company and the 4.7 cm diameter glass microfiber filters are manufactured by Whatman.
The vacuum pump is manufactured by Gast and the pH meter is manufactured by Denver
Instruments, Model 215. The Acculab balance was used for all SPLP tests is
manufactured by Sercom, Model V1-1mg, with a maximum capacity of 120 g and a
resolution of 0.001 g. All materials that come in contact with the samples are acid washed
with a dilute concentration of the 60/40 mixture of sulfuric and nitric acid. The procedure
of conducting this extraction test is in accordance to the Test Method for Shake
Extraction of Mining Waste by the Synthetic Precipitation Leaching Procedure (ASTM
Standard D6234-98 2007) and Synthetic Precipitation Leaching Procedure (USEPA
Method 1312 1994).
First, a 5 g sample is prepared and placed in a 125 mL extraction vessel. Next,
100 g of extraction fluid is weighed and added to the extraction vessel with a pipet,
therefore, creating a 20:1 ratio of extraction fluid to mine waste sample by weight. The
contents are then sealed and placed tightly in the place assembly of the rotary agitator.
Once all the extraction vessels are safely attached to the place assembly, the shake
extraction begins. The rotary agitator operates for 18 ± 2 hours and rotates at 2 rpm.
After the shake extraction is complete, the leachate must be separated from the
mine waste slurry. Before filtering the sample, the pH and specific conductance is
recorded and the glass microfiber filter is acid washed with 0.5 L of sulfuric/nitric
solution and rinsed with 1.5 L of deionized water. The leachate is then sucked through
the filter apparatus at -20 psi, leaving behind the mine waste. Each analytical batch has
one duplicate and one blank.
42
|
Colorado School of Mines
|
Figure 4.6: SPLP rotary agitator
4.3.2Total Metals Analysis
To understand the constituents of the raw mine tailings and the finer and coarser
mixture of the tailings, a total metals analysis was completed by Kate Fenlon from the US
EPA. The method that was followed for conducting this experiment can be found in the
US EPA Method 3051A Microwave Assisted Acid Digestion of Sediments, Sludges,
Soils, and Oils, (USEPA Method 3051A 2007). The constituents in the leachate were
measured with inductively couple plasma (ICP).
4.3.3 Specific Conductivity
The use of this test indicates the amount of free ions in solution relative to other
samples. The ability to understand what compounds are in solution is can be estimated
with comparing the ICP-AES results and the concentration of hydroxide ions, however,
this modeling exercise is not as accurate as measuring the mineralogy and specific
complexes in solution. Therefore, this experiment is to compare the overall
electrochemical activity, not to measure the specific complexes of heavy metals, alkali
metals, alkaline earth metals, and non-metals.
To perform this experiment the following equipment is needed: 40 mL graduated
cylinder, specific conductivity meter, wash bottle, deionized water, and Kim wipes, and
50 mL plastic beaker. The procedure from the Test Methods for Electrical Conductivity
and Resistivity of Water was followed, (ASTM Standard D1125-95 2009). The
43
|
Colorado School of Mines
|
conductivity meter was calibrated with respect to specific conductance and temperature.
The temperature was calibrated with deionized water at temperature of 20 °C. The
specific conductance of the standard KCl solution is 2764 µS at 25 °C, where the
conductance was adjusted to 2550 µS due to the solutions temperature of 21 °C. The
standard conductivity fluid was manufactured by Oakton with a cell constant of 1.
4.3.4 Inductively Coupled Plasma – Atomic Emission Spectroscopy (ICP-AES)
The purpose of using this spectrophotometric technique is to chemically analyze
the raw leachate from the collected mine tailings. This analysis returns base line
concentrations of heavy metals, alkali metals, alkaline earth metals, and non-metals that
leach off of the mine waste and end up in the aqueous phase.
The ICP-AES is a Perking Elmer Optima 5300 DV instrument. The Cyclonic spray
chamber is non-baffled and the nebulizer is a Type A Meinhard. The argon gas flow is
16 L/min for the plasma, 65 L/min for the nebulizer, and 0.5 L/min for the auxiliary gas
flow. The leachate from the SPLP test is transferred to 40 mL vials. The leachate is
prepared with Opitma grade HNO , which is 67-70% by weight and is manufactured by
3
Fisher Scientific. The pH is measured during preparation so that the concentration of
hydrogen ions falls below 10-2 M, making sure that most metals are found in the aqueous
phase. The duplicates and the blank undergo the same procedure as all other samples.
4.3.5 Hardness and Alkalinity Analysis
Water quality parameters can be calculated from the results of the ICP-AES
analysis. The two parameters of interest are the hardness and the alkalinity of the water.
The hardness is a measure of multivalent ions, but for these studies calcium and
magnesium are assumed to be the source of hard water. The alkalinity is the buffering
capacity; therefore water with a higher alkalinity requires more acid to lower the pH. In
these studies the environmental condition of an open system is assumed. Therefore, the
partial pressure of CO is in equilibrium with the carbonate system in the leachate from
2
the SPLP test.
To find the hardness the milliequivalent, m concentration of calcium and
eq,
magnesium are calculated from the mass of analyte per volume of filtered leachate with
the application of stoichiometry. The milliequivalent concentrations of the two
multivalent ions are summed, and then this value is converted into the common units that
44
|
Colorado School of Mines
|
are used to report water hardness, mg/L of CaCO . The conversion can be completed with
3
the following relationship of 50 mg/m of CaCO .
eq 3
Since an open system is assumed to occur when the samples rest on the auto
sampler during ICP-AES analysis, the following equation is used to calculate the
alkalinity, (Brezonik and Arnold 2011):
[ ] [ ] (4.7)
[ ] [ ] [ ]
Where, K and K are the equilibrium constants for the carbonate system. These
a1 a2
are the points where the concentration of carbonic acid equals the concentration of
bicarbonate, K , and where the concentration of bicarbonate equals the concentration
a1
carbonate, K . P is the partial pressure of carbon dioxide in the atmosphere, K is the
a2 CO2 H
Henrys Law constant for carbon dioxide, K is the equilibrium constant for the dissociate
W
of water, and [H+] is the concentration of hydrogen ions in the sample. The values are
stated below in Table 4.1.
Table 4.1: Values of parameters to calculate alkalinity
K K K , mol/L∙atm P , atm K
a1 a2 H CO2 W
4.5E-07 4.7E-11 3.2E-02 3.2E-04 1.0E-14
45
|
Colorado School of Mines
|
diffusion model. This is a gross under estimation because a spherical shape has the
optimum volume to surface area. Any blemishes, cracks, fissures, or holes will increase
the surface area per unit volume.
Table 5.2: Specific surface area for various fine aggregates
Fine Aggregate Specific Surface Area, m2/kg
Sand 5.4
Tailings Sample 1 21.6
Tailings Sample 2 13.8
Tailings Sample 3 20.5
Tailings Sample 4 29.0
Tailings Sample 5 30.7
Tailings Sample 6 42.5
Tailings Sample 7 55.6
Coarser Tailing Mixture 21.2
Finer Tailing Mixture 39.5
One important parameter that pertains to the design of concrete specimens is the
water to cement ratio. This ratio can be manipulated by the amount water that one adds to
the mix, however, the moisture that sorbs to the fine aggregate needs to be accounted.
Therefore, the preparation of the fine aggregates dictates the moisture content which
influences the water to cement ratio. Also, the absorption of water can reduce the
workability of the mixture because of the relative moisture content percentages. For
example, the fine aggregates can absorption 10% of the fine aggregates weight in water
which less water is less available for the hydration process. In practice the internal
moisture is neglected in the total amount of water added to the mixture, where the surface
moisture is subtracted from the total amount of water when the aggregates are prepared at
saturated surface dry conditions.
This is how the mix design was first calculated when mixing and pouring the T:FA
variation experiment. However, the UCS test results showed that the uni-axial
compressive strength did not meet the design criteria. Creating iteration of the concrete
mix was necessary to achieve the design criteria of compressive strength of at least 4000
psi.
48
|
Colorado School of Mines
|
Table 5.5: Total metals analysis for raw mine tailings samples, concentration of dry
tailings in mg/kg
Average concentrations (dry basis)
Sample Cd Mn Pb S Zn
Name mg/Kg mg/Kg mg/Kg mg/Kg mg/Kg
Finer 33.33 7476.7 4345.2 35023.2 5095.3
Coarser 38.48 5331.6 5884.8 46238.9 5715.5
Tailing 1 24.48 1100.0 5877.3 37528.8 4249.0
Tailing 2 7.45 933.1 5916.3 31742.9 1299.0
Tailing 3 18.82 1035.2 5836.6 35292.7 3272.8
Tailing 4 65.19 8350.0 6847.6 67065.9 7363.4
Tailing 5 17.88 5296.3 4418.3 25557.3 2998.6
Tailing 6 22.39 5988.3 3071.3 22604.1 3774.5
Tailing 7 41.19 8649.1 3954.2 37298.0 5886.5
The concentration of analytes with respect to the leachate is not only important, the
concentrations with respect to the solids, Table 5.5, are used to understand how the
introduction of mine waste can change the hardening process of concrete. For example,
the amount of sulfur can be used to assume the amount of sulfide bearing minerals
present. This is an over estimation because not all sulfur is in sulfide form. The presence
of the sulfide bearing mineral is necessary to form MIW and produce sulfuric acid, which
will degrade the concrete specimens. Also, the amount of calcium present in the raw mine
tailings will indicate the amount of alkali material that may have been added to help
neutralize and inhibit the formation of MIW.
The second chemical characterization that was conducted on the leachate from the
SPLP test is the measurement of specific conductance. This represents the relative
amounts of dissolved ions in solution. This experiment cannot extrapolate which ions are
in solution, but is an indicator of the relative amounts if dissolved material found in the
sample.
The higher the specific conductance represents the more leaching that occurred or
the higher the chance of contamination after filtration through the glass microfiber filter.
For example, if it was observed that an analytical blank has high specific conductance,
relatively speaking >0.1 mS/cm, then this is an indication that contamination occurred.
Also, if the samples specific conductance varied from the mean of the analytical batch,
then some form of contamination occurred during the SPLP experiment. The specific
52
|
Colorado School of Mines
|
When developing a concrete mix design the conditions of the environment and the
ingredients must be accounted. The water to cement ratio is an important factor and in
this study was corrected during the iterative pours after this study. During the T:FA
variation study the water to cement ratio was defined as the water added to saturated
surface dry aggregates. Therefore, the aggregates were prepared so that saturated surface
dry conditions were met. This enabled a workable concrete mix with slumps between
0.5”-3.5”, depending on the ratio of tailings to fine aggregate, Table 5.7.
The more mine tailings that were encapsulated in the concrete specimens, the lower
the slump. This is due to the absorption of water that was retained in the finer particles.
Since, the mine tailings are finer than the sand; the relative surface area available is
increased per total volume of fine aggregate. Therefore, the slump is reduced when more
mine tailings are added.
5.2.1 Structural Integrity
After 28 days of curing, the aged specimens did not meet the compressive strength
design criteria for the uni-axial compressive strength, Table 5.9. Therefore, the approach
that was used to achieve this mixture needed to be refined before moving onto the
strength development study. However, inferences
about the leachability could still be made for these
specimens. It is important to note that the
cylindrical specimens were made from molds that
were past the life-time usefulness. These cylinders
showed significant chipping at the bottom of the
cylinders. This was corrected by using new molds
that had thicker plastic at the bottom section of the
cylinder. These molds were used for future
experiments.
The result of the low compressive strength
could have been due to human error, air voids, and
Figure 5.1: Air void found
segregation. Figure 5.1 and Figure 5.2 show the
in sample with T:FA = 50% with
findings of air voids and segregation, respectively.
finer mine tailings mix, S-F-0.5-2
To minimize large air voids and segregation, the
54
|
Colorado School of Mines
|
Table 5.9: UCS test results for T:FA variation study
Tailings/Fine Age, Avg. Compressive Std Deviation, Coefficient of
Aggregate days Strength, psi psi Variation, %
Control Mix
0 28 2350 76.9 3.3
Finer Mine Tailing Mix
0.1 28 2540 430 16.9
0.3 28 2820 574 20.3
0.5 28 2370 1230 52.1
Coarser Mine Tailing Mix
0.1 28 1730 271 15.6
0.3 28 2380 161 6.8
0.5 28 3060 635 20.7
5.2.2 Leachability
After the completion of the UCS tests, the concrete specimens were reduced to a
size between 2-5 mm, where on average 43% of the crushed specimen fell within this size
range, and underwent SPLP batch leach extractions.
Most of a single cylinder was crushed to particle sizes smaller than 2 mm, 47% on
average. This was due to the act of the jaw crusher pulverizing the hardened cement,
large aggregate pieces, sand, and mine tailings. The particle size distribution was similar
for the control specimens and the specimens containing mine tailings. This is can be
observed by the size fractionation in Table 5.10.
The initial characterization of the raw mine tailings showed concentrations of
cadmium, lead, manganese, and zinc were above the MCLs and recommended
concentrations for drinking water. Therefore, these constituents were analyzed by
comparing the raw mine tailing leachate to the finer and coarser encapsulated mine
tailings and the control specimens, see Figure 5.3 and Figure 5.4, respectively. Most
constituents saw a 2-log removal for the leachate. There were exceptions, zinc for the
concrete specimens containing finer mine tailings and cadmium for the concrete
specimens containing coarser mine tailings, where a 1-log removal was observed.
The cemented paste material visibly noticeable but the metals that were originally
found in the mine tailings such as lead, cadmium, and zinc were also measured in the
leachate after batch extraction tests. This is an indication that the dissolved metals were
from the tailings. However, the metals could be coming from the cement itself.
57
|
Colorado School of Mines
|
There are instances where the control specimens show higher concentrations of
metals compared to the ones containing mine tailings. These constituents could be
leaching from the cementitious material. For example, the measured concentration of
arsenic in the raw mine tailings samples was below the detection limit, where in T:FA
variation study arsenic was measured around 0.01 mg/L in most leachates. The source of
arsenic most likely came from the cement, due to the processing of various minerals
during the manufacturing process of Portland cement.
There were outliers, the most interesting being a sample from the concrete
specimens containing coarser mine tailings, where the concentration of zinc was more
than 200% more that the raw mine tailing samples. This is most likely because the
amphoteric zinc leached due to the pH 10 extraction fluid. The concentration of calcium
in the leachate was almost 5 orders of magnitude lower than other samples, where levels
were measured around .03 mg/L compared to 250 mg/L. Therefore, the high
concentration of zinc was due to the human error of not collecting a sample that properly
represented the concrete specimen.
Table 5.10: Size fractionation after crushing concrete specimens with jaw crusher
Sample T:FA, % > 4.76 mm, % 2-4.76 mm, % < 2.0 mm, %
Control Specimens
S-1 0 10 43 47
S-2 0 9 42 49
S-3 0 9 41 50
Finer Mine Tailing Specimens
S-F-0.1-1 10 10 43 47
S-F-0.1-2 10 10 43 47
S-F-0.1-3 10 9 41 50
S-F-0.3-1 30 10 44 46
S-F-0.3-2 30 11 43 46
S-F-0.3-3 30 11 43 46
S-F-0.5-1 50 11 44 45
S-F-0.5-3 50 10 43 47
Coarser Mine Tailing Specimens
S-C-0.1-1 10 9 41 51
S-C-0.1-2 10 9 40 51
S-C-0.1-3 10 9 43 48
S-C-0.3-1 30 11 42 47
S-C-0.3-2 30 10 42 48
S-C-0.3-3 30 10 42 48
S-C-0.5-1 50 11 44 45
S-C-0.5-2 50 11 44 45
S-C-0.5-3 50 10 43 47
58
|
Colorado School of Mines
|
There was no correlation between the concentrations of heavy metals when using
different extraction fluids varying in pH. The use of alkaline extraction fluid was to see if
amphoteric metals such as zinc or aluminum were observed to have higher concentrations
compared to batch extraction conducted with acidic fluids. This was not the case for zinc,
or any metal of concern, except for the possibility of the outlier mentioned above. The
important finding is the capacity level of encapsulation. In the case of not including large
aggregate, this is similar finding from a study that looked at the leaching of Cd, Cu, Zn,
and Pb, (Choi, et al. 2009), for small mortar specimens.
Water quality parameters, Table 5.11, were either measured in the laboratory or
estimated from the ICP result. The specific conductance show relatively high
concentrations of ions of the non-filtered leachate after the SPLP test was conducted. The
blanks, which only contained the extraction fluid, show relatively low concentrations of
ions. This indicated that little to no contamination occur during the SPLP experiments.
Finer Mine Tailing Mix
L Avg. = 0.93
/0.005
0.008
g
m ,n0.004 Cd Pb 0.006L / g
m
o
i t0.003 ,n
a r t n0.002 0.004o i t a
e r
c t
n
o0.001
0.002n
e
c c
d C 0 0
n
o c
0 0.1 0.3 0.5 raw b
0 0.1 0.3 0.5 raw P
Tailings : Fine aggregate
Tailings : Fine aggregate
Avg. = 6.53 Avg. = 1.21
0.05 0.2
L L
/ Mn Zn /
g 0.04 Extraction g
m 0.15m
,n
0.03
pH = 4.2 ,n
o o
i t a r 0.02 pH = 5.0 0.1 i t a r
t n
e c 0.01 pH = 10
0.05t n
e c
n n
o o
c 0 0 c
n n
M 0 0.1 0.3 0.5 raw 0 0.1 0.3 0.5 raw Z
Tailings : Fine aggregate Tailings : Fine aggregate
Figure 5.3: Concentrations of metals from fine mine tailing concrete specimens after
SPLP experiments for the T:FA variation study
59
|
Colorado School of Mines
|
The alkalinity was estimated from the pH of the unfiltered sample for an open
system. Thusly, the equilibration of CO from the atmosphere is an assumption that is
2
used during this calculation.
If the pH of the unfiltered leachate was below 7 then Equation (4.7 returns a
negative number because the concentration of hydrogen ions out compete the carbonate
system. Every blank during this study observed a pH less then 7, therefore instead of
reporting a negative alkalinity value, 0 mg/L of CaCO3 is reported.
Coarser Mine Tailing Mix
Avg. = 0.27
Avg. = 1.47
0.012
L
/ g
m
0.01 Cd Pb
00 .. 00 00 67
L /
g
m
,n o0.008 0.005 ,n
i
t a0.006
0.004o
i t
r a
t 0.003r
n e0.004 t n
c 0.002e
n c
o0.002 n
c 0.001o
d
C 0 0
c
b P
0 0.1 0.3 0.5 raw
0 0.1 0.3 0.5 raw
Tailings : Fine aggregate
Tailings : Fine aggregate
Avg. = 48.1 Avg. = 108
0.03
L L
/ Mn 40 /
g g
m Zn m
,n 0.02 30 ,n
o Extraction o
i t a r pH = 4.2 20 i t a r
t n e 0.01 t n e
c n pH = 5.0 10 c n
o o
c
n M 0
pH = 10
0
c
n Z
0 0.1 0.3 0.5 raw 0 0.1 0.3 0.5 raw
Tailings : Fine aggregate Tailings : Fine aggregate
Figure 5.4: Concentrations of metals from concrete specimens containing coarser
mine tailings after SPLP experiments for the T:FA variation study
60
|
Colorado School of Mines
|
Finally, the hardness represents to amount of potential free calcium and magnesium
that was found in each sample. These samples show high concentrations of these ions. In
the case of leaching of concrete in an urban setting, the leachate would most likely be
diluted by other surface runoff so that the concentration falls below 150 mg/L of CaCO
3
equivalent. However, the values reported show that specimens leach water that would be
considered very hard. Therefore, industrial process that may indirectly use this water
downstream may have to treat this water due scaling issues. If the problem persists,
precipitate of minerals like gypsum may occur and create clogging of pipes. Again, the
dissolved metals could be diluted due to mixing with other sources of water.
5.3 Iterations of Concrete Mixes
Conducting consecutive mixtures of concrete batches is necessary when using the
ACI method for optimizing the strength. The main goal during these iterations was to
reduce the water to cement ratio while not inhibiting the workability of the material. This
strategy is the most common method for creating a mix design for normal strength
concrete in the US, therefore, it used for this project for the purpose of consistency.
Each iteration process includes three steps to obtain a new estimated amount of water
to add to the mixture. The first step is to measure the slump of the each batch. The slump
is an indication of the workability, the goal for this mix design was to have a moderately
workable, with a slump between 3”-4” concrete with reasonable compressive strength,
4000 psi. If the workability is too high, such as slump of 0” for example, this means there
is not enough water in the mixture. Likewise, if the slump is greater than 4”, this
indicates a mixture with too much water.
The slump is not a dependable parameter to run iterations of concrete mixtures. This
fact is shown in the results from the T:FA variation experiment, Table 5.7. The slump
ranged from 0-3.5”, showing a reasonably workable concrete, but the strength did not
meet the design conditions of 4000 psi. Therefore, water was removed from the mix
while compromising the workability of the concrete. To help with the workability issue a
water reducer was added to the mix. The super plasticizer targets a lower water-cement
ratio by 5-12% and increases the slump. Therefore, the concrete can be pumped more
easily with the same amount of water. The main ingredients for water reducers, which are
Type A admixtures, are organic. Such admixtures use lignosulfonates, hydroxycarboxylic
62
|
Colorado School of Mines
|
acids, and hydroxylated polymers, (Rixom and Mailvaganam 1999). A superplasticizer is
a material derived from these compounds in a polymerized form. The benefits for using
this additive are shown in the first iteration that was conducted, where water was
removed from the T:FA variation mix and the slump ended up to be 1.25”. The
compressive strength met the design conditions, Table 5.12 & Table 5.13.
Table 5.12: Concrete mix design and parameters for iterative batches
Tailings/Fine Volume, Water, OPC, Tailings, Gravel, Sand, Super Slump,
w/c
Aggregate yd3 lb. lb. lb. lb. lb. Plasticizer, lb. in.
Iteration 1
0 1 300 526 0 1736 1461 19 0.57 1.25
Iteration 2
0 1 325 570 0 1637 1517 0 0.57 -
Fine 0.1 1 325 570 148 1686 1323 0 0.57 -
The second step in during the iteration process was to measure the compressive
strength of the cured concrete specimens. This was completed either on the 7th or 14th day
of curing. During the second iteration, the super plasticizer was removed from the mix to
see if the design criteria for the compressive strength could be achieved. Even though the
workability was sacrificed, the compressive strength for the second batch showed to have
failure loads comparable to the first batch. It is important to note that the cure time was
shorter for the batch containing the super plasticizer. When organic matter is added to a
concrete mix, the setting time is often times increased; however, in this case we saw
acceleration in the setting time. Therefore, it is possible that CaCl , Ca formate,
2
triethanolamine, and/or sodium thiocyanate were present in the super plasticizer, (Rixom
and Mailvaganam 1999).
Table 5.13: UCS test results for iterative batches
Tailings/Fine Age, Avg. Compressive Std Deviation, Coefficient of
Aggregate days Strength, psi psi Variation, %
Iteration 1
0 7 4320 236 5.5
0 7 3330 621 18.7
Iteration 2
0 14 3950 119 3.0
Fine 0.1 14 3930 149 3.8
63
|
Colorado School of Mines
|
5.4.1 Structural Integrity
The intent of these experiments was to understand the strength development of the
materials produced and to see if there is a correlation between the age of the specimen
and the potential to leach heavy metals. The time between measuring compressive failure
loads, reducing the particle size, and conducting leaching tests was minimized. Therefore,
when the curing time was complete these tests all took place on the same day.
The UCS tests show that the strength developments over 28 days are not only
similar for the specimens containing finer and coarser mining tailings, but also similar to
the control specimens, Error! Reference source not found..
One issue with the results from the UCS tests is the coefficients of variation
(COV). The limit that is stated by the ASTM standards is 10.6%, (ASTM Standard
C39/39M-12a 2012), where the sample sets are often larger than this value. This issue
could be caused from inclusion of air voids and presence of segregation. However, during
the reduction of particle size, the specimens were inspected and often times did not have
either of the strength reducing criteria. For example, sample L-28 d-C-0.3-2, which is the
second specimen out of three that was tested after 28 days of curing that contained
coarser mine tailings that substituted for 30% of the fine aggregate, did not show signs of
air voids or segregation and failed at relatively low loads compared to the other
specimens in the batch. These anomalies were observed to have typically type III
fractures, which vertical fractures are propagating from the top of the cylinder towards
the bottom.
Even though that the COV values were often over the ASTM standards, the general
trend, Figure 5.5, shows that the concrete specimens containing finer and coarser mine
tailings maintained comparable strength to structural concrete up to 28 days. That is not
to say the degradation of the cemented material would not occur over the lifetime of the
concrete. For example, the specimens containing coarse mine tailings do show
degradation in strength between 14 and 28 days of curing. However, since the COV is
large for this sample set, then it is difficult to extrapolate conclusions. Therefore, longer
term studies are needed to further investigate the potential of using this as a construction
material.
65
|
Colorado School of Mines
|
One reason why the strength development is comparable for the specimens
containing mine tailings versus the controls might be because of the chemical make-up of
the tailings. Large amounts of pyritic material, which is the source of MIW, could be
reactive and produce sulfuric acid. However, on a 28 day time frame, this problem did
not occur. This would have been visible with the observation of pop-outs and stained
from iron hydroxides. Thusly, the pyritic material present in the tailings must not have
been reactive during the time of the experiment.
Using the information
5000
in found Table 5.5 the total 4500
4000
amount of sulfide per unit
3500
weight of the tails was found i 3000
s
p
to be 3.5% for the finer mine
,S2500
C
U2000
tailings batch and 4.6% for Control
1500
the coarser mine tailing 1000 Fine, 30% substitution
500 Coarse, 30% substitution
batch. Therefore, there could
0
be the possibility to form 0 10 20 30
Ages, days
MIW products such as
Figure 5.5: UCS test results for strength
sulfuric acid and dissolved
development study
heavy metals. However, there
was the presence of calcium, which indicates if neutralizing material may have been
added to the tails, which were found to be 6.3% and 8.6% by weight of the finer and
coarser tails respectively. Due to compressive strength results, inferred from Figure 5.5,
the formation of MIW products did not cause the specimens containing tails to degrade.
This does not mean for longer curing times this trend will stay the same, however, up to
28 days this material could be feasibly used as structural concrete.
5.4.2 Leachability
Again, once the UCS tests were finished, the concrete specimens immediately were
prepared for undergoing a batch leach extraction tests. The cylinders were fed through a
jaw crusher and particles between 2-5 mm were collected and used for the acid digestion
experiment. The results show that cadmium, lead, and zinc were encapsulated below
either the MCL levels or the recommended concentrations for drinking water. One
66
|
Colorado School of Mines
|
concrete specimen that contained the fine mine tailings did show a concentration above
the recommended level. This data point occurred when an extraction fluid of pH equal to
10 was used during the digestion period. This is an interesting observation because
manganese oxides are not amphoteric and do not dissolved in a basic solution. However,
even at a concentration of 0.06 mg/L, this is orders of magnitude below the concentration
of Mn leaching from the raw mine tailings. The results for the strength development
study are shown in Figure 5.6, Figure 5.7, Figure 5.8, and Figure 5.9.
When comparing the results for the SPLP compared to the total metals analysis, Cd
showed a 3 log encapsulation capacity, Pb showed almost a 4 log encapsulation capacity
(every measurement was below the detection limit), and Mn and Zn both showed a 3-log
encapsulation capacity. The effectiveness of this treatment could be feasible for this site
specific material.
Again, water quality parameters such as the specific conductance, hardness, and
alkalinity, Table 5.11 and Table 5.16, were either measured in the laboratory or estimated
from the ICP result. The specific conductance show samples in the range of 2.03-3.26
mS/cm, where the blanks were orders of magnitude less, around 20 µS/cm. This is
indication that contamination of the samples was minimized during the SPLP
experiments. The alkalinity was estimated from the pH of the unfiltered sample for an
open system. Thusly, the equilibration of CO from the atmosphere is an assumption that
2
is used during this calculation. The pH of the leachate from the samples was found to be
consistently above 12.
72
|
Colorado School of Mines
|
CHAPTER 6
RECOMMENDED FUTURE RESEARCH
This study demonstrates the fundamental feasibility of using mine tailings as a
substitute for fine aggregates in making structural concrete when considering remediation
strategies for managing the environmental impacts of mine tailings. The first experiment
showed that a variable amount of mine tailings can be solidified and stabilized in
concrete with similar encapsulation capacities. The ratio of tailings to fine aggregates
were relatively low compared to the amount of sand, but was necessary to meet the
fineness modulus requirements for ASTM standards.
To have a deeper understanding of the economics and the technical viability of
encapsulating mine tailings in structural concrete, long-term durability studies should be
conducted. The study presented above is, in essence, a snapshot in time combined with a
limited period of curing, with a maximum of 28 days. One option would be to take
specimens and apply thermocycling to simulated seasonal temporal changes. This
increased rate of erosion will age the concrete and possibly help predict future strength
and leachability outcomes. The second option would be to have the cylinder age much
longer in the laboratory before undergoing UCS and SPLP testing.
There are parameters for stabilization and solidification of wastes that control the
model more than others. The leachability of heavy metals is dependent on list of physical
characteristics of the curing of concrete, leachant, and waste composition. The physical
nature of the solid matrix is linked to the chemical properties of the leaching process, as
noted with the importance of porosity, particle size distribution and moisture content.
Some of these factors can be controlled in the experiment such as the curing temperature,
curing time, curing moisture, mass of cement, water/cement ratio, type of leachant,
leaching time, leaching temperature, and liquid/solid ratio, (Heimann, et al. 1992).
However, the type of matrix cannot be controlled, optimizing structural integrity, because
the waste composition often varies. Modeling the solid matrix is difficult, (Hills and
Pollard 1997), some instruments such as an electron microscope can show the
microstructure of the solid matrix, but knowing the chemical makeup takes a deeper
understanding.
73
|
Colorado School of Mines
|
The structure of the solid matrix will change the permeability due the change in
pore structure. To further understand the corrosion of concrete, the permeability of the
specimens could be measured. This would be useful when developing leaching models.
For metals, the tortuosity is also an important factor, dictated by the structure of the solid
matrix. A retardation factor can be estimated from the leached porosity, ε, and the
tortuosity, τ, where the retardation factor equals ε/τ, (Baker and Bishop 1997). The
retardation factor is often used in mechanistic models for predicting the leachability of
heavy metals from concrete. Another parameter of interest is the acid neutralizing
capacity. This parameter is measured so that the buffer capacity can be estimated for the
waste solid. The buffer capacity will dictate the solid form’s susceptibility to reacting
with acids and minimizing alkali conditions, (Barth, et al. 1990). The higher the buffer
capacity the more alkali material is needed to lower the pH and thus increasing the
leachability.
A common mechanistic model used for predicting the leachability is the bulk
diffusion model. This type of model is used for the leachability of radioactive wastes
from solid materials, however, has been applied to hazardous wastes such as heavy
metals, (Cheng and Bishop 1990). The physical and chemical properties are different
between radioactive and heavy metals wastes. The pH of the system must be taken into
consideration due to acid/base reactions. The Fickian diffusion model is dependent on the
concentration gradients of the waste material between the solid matrix and the leachate.
The observed diffusion coefficient is an important parameter estimated in this model. The
following equation represents the leaching of a contaminant in a mass to mass basis,
(Cheng and Bishop 1990)is:
∑
(6.1)
Where, is the contaminant loss in mg, is the initial contaminant in mg, is
the volume in cm3, is the surface area in cm2, is the leaching period in s, and the
observed diffusion coefficient in cm2/s. The overall bulk diffusion equation in three
dimensions is as follows, (Barna, et al. 1997):
74
|
Colorado School of Mines
|
( ) (6.2)
The effective diffusion, D , is a parameter which describes the physical properties
e
of the system, but, does not describe the chemical properties of the leaching process. The
chemical and physical properties both need to be represented in a model. An observed
diffusion coefficient can be measured, where both the chemical and physical properties
are included in the term, however, the chemical and physical immobilization mechanisms
must be separated for this type of model to fully work, (Batchelor 1992). The solution to
the overall bulk diffusion equation in one dimension under unsteady-sate conditions can
be described as:
√
( ) [ ( )] (6.3)
√
Where, x is the penetration distance of the leachant into the solid matrix, d is the
density of the sample, and k is a first order reaction rate. The penetration distance must be
adjusted with the retardation factor because the porosity of the solid matrix creates a long
tortuous path for the heavy metals to find its way into the bulk leachate. There are
boundary conditions stated before this solution was solved from the overall bulk diffusion
equation. When time equals zero for all penetration distances, the concentration in the
leachant is zero. When time is greater than zero at penetration distance equals to zero, the
concentration equals the initial concentration. Finally, when penetration distance equals
infinity, concentration equals zero.
A common leachant used in experiments is acetate. Therefore, the four elements
modeled for the leachability of wastes in concrete are Ca2+, H+, Ac-, and the contaminant,
such as Pb2+, (Batchelor 1992). When using this model, it is important to consider the
activity versus the concentration because the ionic strength of the leachate solution,
(Shukla, et al. 1992), may not be negligible.
The bulk diffusion model has limitations not addressed due to the complexity of the
chemical interactions between the leachate and the solid matrix. The shrinking unreacted
75
|
Colorado School of Mines
|
core (SUC) leaching model is more comprehensive because the buffer capacity is
incorporated into the mathematics. The overall view, Figure 6.1, of the leaching process
is not different from the bulk diffusion model except leachate is considered to be
diffusing into the solid matrix, instead of the waste diffusing out to the leachant, (Malviya
and Chaudhary 2006). There are limitations to this model such the inability of measuring
the diffusion through the boundary layer, measuring the diffusion through the leached
layer and the ability to understand the chemical reactions at the leaching boundary,
(Baker and Bishop 1997). However, the use of an acidic leachant does not over
complicate the testing procedures. The exposure of the acid must be calculated, very
similar to the concentration time (CT) calculation for chlorination in a waste water
treatment plant. Therefore the leachability is a function of CT not time, when compared
to the bulk diffusion model. One problem arises with this model; a lechant of neutral pH
is used than the model breaks down since the exposure integral must be calculated by
taking periodic pH measurement throughout an experiment. Also, an understanding of the
kinetics must be known so that the acid neutralizing capacity is known with respect to
time.
Figure 6.1: A diagram of the conceptual process of the SUC leaching procedure,
(Baker and Bishop 1997)
It would also be advantageous to look at different cementitous material and see the
different encapsulation properties that exist for these materials, particularly geopolymers.
76
|
Colorado School of Mines
|
Dr. Claire White, a postdoctoral fellow at Los Alamos National Laboratory, mentions
that mine tailings can be used as a precursor to geopolymers, not only used as a substitute
for the small aggregate in concrete. Therefore, a two-fold benefit can be utilized when
incorporating mine tailings as a filler and a binder. However, there is little research
contributed to this idea, (Pacheco-Torgal, Castro-Gomes and Jalali 2008).
The amount of research that has developed has shown great promise for slag, fly-
ash, and metakaolin precursors to geopolymers. However, the use of mine tailings as a
precursor needs more attention. Characterizing the physical and chemical properties is
important to develop this idea further. Also, more focus needs to be aimed at studying
how mine tailings can be encapsulated in concrete. There is a lot of work geared towards
mortars and grout, however, if large volumes of mine tailings are going to transported
from storage facilities to construction sites, concrete is a material that has the possibility
of retaining a large percentage of mine wastes. Therefore, the interactions between large,
coarse aggregate, compressive strength and leachability need more attention.
It would also be advantageous to understand the mineralogy of the mine tailings
and see what these structures look like when combined with cemented paste. This could
be done by using x-ray diffraction (XRD) in combination of using a scanning electron
microscope. With this method one could visually see what the hardened material and
have an understanding of the amounts of hydrates present in the specimen.
Since the assumption the fine aggregates could be geometrically considered
spheres, the SSA was grossly overestimated. However, using the Brunauer, Emmett, and
Teller (BET) method for gas absorption experimentation often yields an underestimate
for the SSA. Therefore, a range of SSA values could be obtained.
77
|
Colorado School of Mines
|
CHAPTER 7
CONCLUSION
The scope of this project was to investigate the feasibility of re-using mine tailings as
an input material for structural concrete. The factors that were examined were the
unconfined compressive strengthen and the concentration of heavy metals in leachate
from batch leach extraction tests. It was hypothesized that the compressive strength of
concrete containing mine tailings would meet the design criteria for the control
specimens, which was a compressive strength of 4000 psi and a slump between 3-4”. Due
to the encapsulation capabilities, it was also hypothesized that the concentration of heavy
metals will be below the regulatory levels for drinking water standards in the United
States.
The initial characterization of the raw mine tailings showed concentrations of
cadmium, lead, manganese, and zinc were above the MCLs and recommended
concentrations for drinking water.
A trend in the size distribution was correlated to the locations where the mine tailing
samples were taken at the Pride of the West TSF. As the distance increased from the
spigot, from which the tailings were deposited, the particle sizes distribution got wider
and the relative amount of fine material increased.
It was observed that the fineness modulus is not the best indicator for the compressive
strength; it was observed that for the coarse specimens that had 50% of mine tailings
substituted for sand had average compressive strength significantly higher than the
control specimens. Where the fineness modulus limit is 2.3, the value for these specimens
was 1.9.
When varying the T:FA ratios no correlation was observed in the concentration of
heavy metals in the leachate for substitution ratio less than 50%. Since, there no
correlation found in the optimal portion of mine tailings and the leachabiltiy, the fineness
modulus was used to calculate the maximum amount according to the ASTM standards,
which was found to be 35% for this site specific material.
The heavy metals of interest constituents were analyzed when varying the T:FA
ratios by comparing the raw mine tailing leachate to the fine and coarse encapsulated
mine tailings and the control specimens. Most constituents saw a 2-log removal for the
78
|
Colorado School of Mines
|
leachate. There were exceptions, it was observed that zinc for the fine concrete specimens
and cadmium for the coarse concrete specimens achieved 1-log removal.
Since, the uni-axial compressive strength did not meet the design criteria during the
T:FA study, the concrete mix design was perfected so that the compressive strength was
4000 psi or greater. It was found that a slump of 3-4” could not be achieved unless
additives were used, such as a superplasticizer. To minimize ingredients, especially
relatively expensive ones, the superplasticizer was not used during the strength
development tests so that minimum bias was present.
It was found during the strength development tests that the compressive strength for
the specimens containing 30% substitution of mine tailings for the fine aggregate
maintained similar failure loads to the controls. This shows that the strength of this
concrete meets the ASTM standards for compressive strength up to 28 days. This opens
up the possibility for future research; however, the long-term durability is of grave
interest before ever commercially utilizing this material.
When comparing the results for the SPLP compared to the total metals analysis, Cd
showed a 3 log encapsulation capacity, Pb showed almost a 4 log encapsulation capacity
(every measurement was below the detection limit), and Mn and Zn both showed a 3-log
encapsulation capacity. The effectiveness of this treatment could be feasible for this site
specific material.
Concrete specimens that contained mine tailings can leach heavy metals above the
MCL level for drinking water standards. The one of most concern is the level of cadmium
from a specimen that contained the coarse mine tailing mixture at a T:FA of 50%. Even
though there was no correlation between the T:FA ratio and the concentration of heavy
metals, this shows that it is possible to not encapsulate heavy metals well enough when
using the most conservative standards for decision making. Therefore, more research
needs to be conducted on the different types of cementitous material and additives for the
purposed use of this remediation strategy.
79
|
Colorado School of Mines
|
14 Day
Table H.5: SPLP analysis for strength development study, 14 day setting time
A n a ly t e N a m e L - 1 - 4 .4 2d ( m g /L ) L - 1 - 5 4 (d m g /L ) L - 1 - 1 04 d ( m g /L ) L - 1 - F - 0 - 4 . .4 3 2d ( m g /L ) L - 1 - F - 0 - 4. .4 D 3 2d ( m g /L
)
L - 1 - F - 0 - 5 . 4 3 (d m g /L ) L - 1 - F - 0 - 1 . 04 3 d ( m g /L ) L - 1 - C - 0 - 4 . .4 3 2d ( m g /L ) L - 1 - C - 0 - 5 . 4 3 (d m g /L ) L - 1 - C - 0 - 1 . 04 3 d ( m g /L ) L - 1 - B - 4 .4 2d ( m g /L )
Ag BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL 0.0011
Al 1.32 1.35 1.33 1.35 1.44 1.25 1.04 1.34 1.37 1.14 0.0412
As 0.0146 0.0128 0.0136 0.0109 0.0089 0.0151 0.0185 0.0112 0.0073 0.0139 BDL
B 0.561 0.511 0.273 0.446 0.357 BDL BDL 0.0159 BDL BDL 0.0102
Ba 0.549 0.503 0.391 0.525 0.396 0.274 0.309 0.189 0.150 0.194 0.027
Be BDL BDL BDL BDL BDL 0.0001 0.0001 BDL BDL BDL BDL
Ca 258 242 215 197 179 199 222 168 159 175 0.271
Cd 0.0004 0.0004 0.0006 0.0003 0.0004 0.0004 0.0010 0.0004 0.0003 0.0003 BDL
Co BDL 0.0003 0.0002 BDL BDL BDL 0.0003 BDL BDL BDL 0.0003
Cr 0.0162 0.0156 0.0151 0.0123 0.0128 0.0109 0.0123 0.0078 0.0074 0.0082 BDL
Cu 0.0043 0.0031 0.0044 0.0039 0.0054 0.0037 0.0041 0.0069 0.0040 0.0045 0.0020
Fe 0.106 0.102 0.0625 0.0866 0.0804 0.0073 0.0067 0.0246 0.0087 0.0036 0.0109
K 35.8 30.7 30.8 29.8 24.5 28.0 35.3 27.5 23.2 33.1 1.24
Li 0.0283 0.0267 0.0228 0.0183 0.0160 0.0192 0.0234 0.0216 0.0202 0.0258 BDL
Mg 0.0221 0.0215 0.0172 0.0208 0.0186 0.0087 0.0095 0.0144 0.0116 0.0105 0.0063
Mn 0.0005 0.0002 0.0002 0.0006 0.0002 0.0028 0.0627 0.0012 0.0149 0.0002 0.0068
Mo BDL BDL BDL 0.0053 0.0057 0.0038 0.0044 0.0044 0.0041 0.0042 BDL
Na 5.87 5.53 8.09 5.23 4.14 4.36 8.38 4.05 3.46 7.90 0.0077
Ni 0.0012 0.0009 0.0013 0.0015 0.0016 0.0011 0.0011 0.0014 0.0010 0.0016 0.0007
P 0.0068 0.0089 BDL BDL BDL 0.0109 0.0099 0.0159 BDL 0.0112 BDL
Pb BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL 0.0017
S 7.57 6.91 6.74 7.03 7.20 6.94 7.03 7.13 6.67 7.41 1.10
Sb BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL
Se BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL
Si 2.34 2.39 3.04 2.87 3.24 2.59 2.35 3.37 3.68 3.34 0.336
Sn BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL
Sr 1.37 1.31 1.12 1.10 0.951 1.13 1.31 0.934 0.855 1.03 0.0011
Ti 0.101 0.0952 0.0515 0.0805 0.0743 BDL BDL 0.0153 BDL BDL 0.0043
Tl BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL 0.0044
V 0.0019 0.0021 0.0022 0.0020 0.0027 0.0011 0.0009 0.0018 0.0019 0.0016 BDL
Zn 0.0373 0.0248 0.0299 0.0279 0.0250 0.0221 0.0276 0.0480 0.0238 0.0285 0.0428
143
|
Colorado School of Mines
|
ABSTRACT
In environments where mechanical excavation systems break rock, airborne rock dust is
generated and could pose respiratory health threats to workers. These environments include
mining and civil applications with roadheaders, continuous miners, or similar machines used in
underground operations. Therefore, this study aims to compare the characteristics of rock dust
generated by conical pick cutters at various wear conditions. With supplemental experiments, the
results can aid future evaluations of proper bit management, dust suppression, and control
systems.
The work conducted under this study included full-scale laboratory cutting of concrete,
limestone, and sandstone samples. Three symmetrically worn conical picks cut each sample at a
new, moderately worn, and fully worn stage of wear. Equipment collected airborne dust samples
and fines during and after cutting for qualitative and quantitative characteristics. Analysis of
collected dust revealed the dust's characteristics, including the concentrations, silica contents,
particle size distributions, and particle shapes.
Findings indicated that the dust generated in the cutting process increases as the pick tip
wears. With a superimposed circle at the pick tip, further analyses show that as the tip radius
increased by one millimeter during excavation, the dust generation at the pick increased by an
average of 50 mg/m3 for all the samples. Furthermore, analysis of the cutting forces and specific
energy of cutting (the amount of energy used to excavate a unit volume of rock) show that as the
pick radius (mm) increases, the concentration (mg/m3) of dust per specific energy (kW-hr/m3)
increases linearly. The correlation between dust and specific energy showed an average increase
of 3.0 [(mg/m3) / (kW-hr/m3)] / mm of tip radiusin the rock samples tested. In terms of silica, the
silica content is a function of mineralogy, and all the rock types contained traces of quartz.
The airborne respirable particle size distributions insignificantly shifted between pick
wear. The fines size distributions slightly increased as the pick wear level increased. However,
the differences are also deemed insignificant and, therefore, are negligible. Finally, the pick wear
does not influence change in the particle shapes. All the picks consistently generated suspended
respirable particles with similar particle shapes that are slightly oval with mostly smooth edges.
iii
|
Colorado School of Mines
|
CHAPTER 1: INTRODUCTION
1.1 Introduction
Exposure to airborne rock dust particles containing silica, coal, or other minerals, can cause
irreversible damages to the respiratory systems and lead to diseases such as coal workers' pneumoconiosis
[1], silicosis [2], and other lung complications [3], [4]. Airborne rock dust can be generated during
excavations with longwall shearers or roadheaders, as in Figure 1.1. However, even with the regulations
implemented by the US Mine Safety and Health Administration (MSHA) to limit the concentration
exposure and exposure to silica, there has been an unfortunate increase in lung disease cases in the United
States since the 1990s with a continued rise in numbers in recent years [5]–[9] as seen in Figure 1.3 and
Figure 1.4.
Researchers do not fully understand why modern miners’ lung diseases are increasing with the
MSHA concentration and silica exposure limits in place [10], [11]. Therefore, various research
investigations are currently underway to analyze the presence of nanoparticles, particle size distributions,
mineralization, and how particle shape affects deposition into the human lung [12]–[16]. Additionally,
there is limited understanding and implementation of dust suppression management concerning the pick
cutter's life.
The primary excavation system for soft to moderate-strength rock fragmentation is mechanized
excavators such as roadheaders, continuous miners, or drum shearers. The primary rock-cutting tool used
in these types of mechanical machines is conical picks. Rock dust is created and released at the interface
between the conical pick and rock surface, where some of the fines generated can be transported and re-
introduced into the air during the muck handling process. Conical picks create dust at the carbide tip
because the concentrated loads they apply penetrate the rock surface and fragment the rock into smaller
pieces and dust particles. The geometry of the pick tip will change over time because they wear down
[17], [18], but it is uncertain if the changing of the pick tip changes dust characteristics. The limited
evidence connecting pick wear to dust characteristics is that coal dust concentration increases with the
wear of pick tips [19]–[23]. There are limited further investigations on other rock types, nor is there a
clear quantitative relationship between the dust concentration generated and the pick tip's geometry.
Therefore, this study investigates the concentration, mineral composition, particle size
distributions, and particle shapes of airborne and fines dust generated from three different conical pick
wears during laboratory full-scale cutting of a concrete, limestone, and sandstone rock block. The conical
pick is used for cutting because these picks are regularly implemented in concrete, limestone, and
sandstone excavation operations in various applications [24]–[27]. The study allows for comparing the
1
|
Colorado School of Mines
|
dust particle characteristics generated from rock cutting with three pick-tip geometries as it wears out
during operation. Results can eventually guide operators to strategize bit management concerning
production rate and machine utilization, considering dust suppression measures when cutting soft to
medium-strength rock.
1.2 Mechanical Excavation of Rock
Cutting, drilling, tunneling, and excavating rock is common practice in today’s modern world.
With advancing technology and growing society, humans are more frequently interacting with the earth
by constructing tunnels, building caverns, and extracting ore. For example, the United States produced
430 million metric tons of limestone in 2022 from underground and surface operations [28]. Rock
excavation in an integral part of various applications including civil and mining, such as building tunnels
for subway transportation or cutting rock to obtain metals and non-metals.
The standard method for removing rock without explosives includes using mechanical energy and
force to break the rock surface. Longwall shearers, roadheaders, and tunnel boring machines (TBMs) are
today's most common machines to excavate rock. Longwall shearers and roadheaders commonly use
picks with carbide tips to contact the rock surface, while tunnel boring machines commonly use disc
cutters. Picks widely excavate rocks up to 90 MPa in unconfined compressive strength, whereas disc
cutters typically excavate any harder rocks [29]–[32]. As seen in Figure 1.1, the cutting drum of longwall
shearers and roadheaders impact and drag picks along the rock's surface to break rock chips off the
cutting face.
Figure 1.1 Longwall shearer (left) cutting coal in an underground environment and a roadheader (right)
cutting a coal seam from a highwall excavation [33], [34].
2
|
Colorado School of Mines
|
The picks that strike the rock surface are made of hardened steel and have durable carbide tips.
The carbide tips minimize wear and come in various shapes and sizes depending on the application and
the geology of the rock excavated. The two main types of picks used in mechanical excavation include the
conical-shaped and the radial-shaped picks. Conical picks rotate in their respective blocks to allow for
uniform wear on the pick tip so that they last longer. On the other hand, radial bits do not rotate once
mounted on the cutterhead and, therefore, do not experience uniform wear, leading to a much shorter life.
Although carbide tips and rotating picks distribute wear, significant amounts of wear on the picks
still occur and can occur at different rates. The operators replace worn picks to maintain machine
productivity while minimizing operating costs and maximizing mechanical availability. In other words,
pick replacement occurs when the drag force and specific energy applied to the rock increase with the
wear to a point when the amount of rock excavated is insufficient to warrant increasing the machine
forces to penetrate and cause breakage. For this reason, picks have a life span from new, when the pick
tips are sharp, to worn, when the pick tips become blunt, or the carbide tip falls out. The lifespan of a pick
depends on many factors, with the main contributor being the rock type, strength and abraisivity, the
penetration and spacing implemented by the machine, the design of the cutterhead, and the amount of heat
generated during the cutting process.
Rock breakage and dust generation occur at the interface between the rock surface and the pick.
Rock cracks and chips are formed during cutting between the normal and drag forces applied to the rock
surface through the pick tip. The cracks and chips release finer rock dust particles upon breaking from the
main body of the host rock. Rock dust particles can also be crushed beneath the pick tip if small particles
remain from the previous cut lines.
1.3 Respirable Particles
Respirable dust particles that can reach the alveolar region of the lungs and are commonly
defined as particles smaller than either 4µm or 10µm in aerodynamic diameter by various sources [35].
For this study, respirable particles are defined as particles less than 10µm in aerodynamic diameter. The
aerodynamic diameter is the diameter at which a particle of a specific size and density has the same
settling velocity as a water droplet with fixed density. In other words, aerodynamic diameter considers the
density of the particle along with the physically measured diameter to provide a standardized diameter.
Respirable dust exposure to underground workers is an ongoing issue in the mining and tunneling
industry because inhaling these particles can cause irreversible lung diseases such as coal-workers
pneumoconiosis [1], [7], silicosis [2], [36], and other life-threatening lung diseases [3]. Additionally,
3
|
Colorado School of Mines
|
inhaling particles from 0.1 to 100 µm in aerodynamic diameter can cause coughing and irritation. As seen
in Figure 1.2, particles above 10 µm in aerodynamic diameter are the particles deposited in the thoracic
and the inhalable areas of the lungs or within the trachea and nose or mouth, respectively [35].
Figure 1.2 Fractions of where the respirable, thoracic, and inhalable particles will deposit in the lung
regarding their aerodynamic diameter [35].
Since the late 1970s and early 1980s, US Mine Safety and Health Administration (MSHA) has
implemented regulations to limit workers’ exposure to respirable dust [10]. The mandated exposure limits
(requiring exposer limit average concentration for a work shift to be 3 mg/m3) showed significant
progress in increasing industrial hygiene standards as the cases of lung diseases dropped drastically for
the following twenty years [8]. However, in the 2000s, cases of lung diseases in underground mine
workers started to increase again [4]–[6], [37] as seen in Figure 1.3 and Figure 1.4, especially in coal
mines. Participants in surveys presented in the figures are miners from mines all over the United States
who participate in regular chest x-ray and health screenings at the Coal Worker’s Surveillance Program.
4
|
Colorado School of Mines
|
concentrations over 1.5mg/m3; this method will continue to defend occupational workers underground.
However, it is still uncertain what is the underlying cause for the increase in lung-related health problems.
Researchers are investigating other possible causes for worker’s lung diseases beyond the effects
of dust concentration. There are studies on the shape of particles, the size distribution of particles,
nanoparticle influence, mineralogy affects, changes in excavation into roof strata or varying seams, and
location of mine [13]–[16], [39], [40]. Although these comprehensive studies have increased
understanding of the characteristics of dust particles, it is still uncertain what types of exposures
potentially harm workers. Additionally, there are limited studies on how a pick tip influences dust
generation. The investigations on pick tip shape influencing dust generation include analysis of the
concentration of dust [17]–[19], [22].
1.4 Research Objectives
The primary objective of this research is to quantify, characterize and compare the respirable rock
dust associated with rock fragmentation during mechanized excavation with conical picks at various
stages of wear. The overarching research question is: How do respirable dust characteristics change and
compare with the symmetrical increase of pick tip radii for conical picks used in mechanical excavations
from cutting medium-strength rock?
Therefore, the following are the research objectives:
Verify the changes in respirable dust concentrations as the pick wear increases;
Determine and establish a quantitative measure to track the rise in dust concentration as
the pick wear increases;
Examine the presence of silica within the respirable airborne dust generated from the
rock samples during the cutting process;
Investigate the shifts, or lack thereof, in airborne respirable dust particle size distributions
generated from the three pick wears;
Investigate the particle size distributions of the fines generated from the three pick wears;
Investigate the change, or lack thereof, in respirable airborne particle shapes generated
from the three pick wears.
With empirical experiments to achieve the research objectives, full-scale cutting tests with
individual conical picks at different wear levels in a linear cutting machine allowed for dust collection and
6
|
Colorado School of Mines
|
characterization. Multiple devices and standardized techniques subsequently analyzed the dust to
determine the concentration, silica presence, particle size distributions, and particle shapes.
1.5 Hypothesis
Research reveals that the dust generated by dull or worn pick cutters is higher than that of sharp,
new ones. However, the quantification is not well understood. Additionally, there has never been a study
or quantification of airborne particle size distributions or particle shapes concerning pick wear. The
assumption is that particle size increases with wear, and particle shapes would become less jagged and
elongated with the increase in pick wear, which requires verification. Therefore, with various
characteristics of dust analyzed in this study, a few hypotheses are provided:
Regarding dust concentrations, it’s hypothesized that as the pick tip radius increases,
respirable dust concentration will increase.
The particle sizes will increase as the pick tip radius increases.
The particle shapes will become smoother around the edges and have an aspect ratio closer to
one as the pick-tip radius increase due to the creation of the significant pressure bubble and
partial movement of material within this zone.
1.6 Methodology
Compared to field measurements or modeling, this study included laboratory full-scale cutting
tests of rock samples with collecting and analyzing respirable dust as the picks wear out and become dull.
Standard dust monitoring devices collected the samples. Then, the dust was subject to tests and analyses
that quantified their intake volumes, silica contents, particle size distributions, and particle shapes. The
NIOSH Manual of Analytical Methods 0600 and 7500, optical systems, and laser diffraction techniques
with standardizations analyzed the particle characteristics.
The foundation of the analysis resides on full-scale cutting tests performed on various rock
samples. Quantitative data is collected and used for statistical comparisons between the dust generated
from the three different pick wears throughout cutting the three different rock types. Therefore, the
research methodology includes the following tasks:
1. Conduct a preliminary literature review to determine the knowledge gaps on this topic in the
past and current literature;
7
|
Colorado School of Mines
|
2. Establish research methodology, including the measurement methods and procedures for dust
characterization;
3. Setup equipment at the Earth Mechanics Institute and develop the design and fabrication of
the automated dust collection system;
4. Secure selected conical picks at various wear levels for testing in a controlled environment;
5. Perform preliminary cutting tests with dust collection equipment and collect the fines from
the cut lines to evaluate the characteristics of the fines;
6. Perform concentration data analysis of respirable airborne dust samples with the NIOSH
Manual of Analytical Methods 0600 and 7500 techniques;
7. Perform particle shape data analysis of the respirable airborne dust samples with the Field
Emission Scanning Electron Microscope and Clemex photo analyzer;
8. Perform particle size distribution analysis of the respirable airborne and fines dust with the
standardized laser diffraction methods;
9. Run further analyses, including statistical analyses, of compiled characterization data;
10. Compare the results of the measurements made in various rock types and at different pick
wear levels to observe general trends to seek the possibility of developing overarching
models.
Under the testing methodology, a microscope measured three individual conical picks comprised
of carbide tips for their pick tip radii at the Earth Mechanics Institute on the CSM campus in Golden, CO.
A lathe and Dremel worn down two of the pick tips in a controlled environment to obtain two separate
larger pick tip radii. After measuring and characterizing these three pick wears quantitatively as “new,”
“moderately worn,” and “worn,” the picks cut concrete, limestone, and sandstone samples with the Linear
Cutting Machine at full-scale.
Concrete, limestone, and sandstone are all medium-strength samples with uniaxial compressive
strengths between 30 and 90 MPa. The experiments required a selection of similar strength samples to
obtain consistency and samples that conical picks commonly cut in the industry. The experiments selected
these samples for additional consistency because they have primarily uniform and homogenous interiors.
The automated dust collection system collected airborne dust generated during cutting with
various instruments, including a Tsai Diffusion Sampler and 10-mm Dorr-Oliver cyclones. This system
collected representative dust samples consistently, and the design details are in CHAPTER 3:
METHODOLOGY AND EXPERIMENTAL SETTINGS. The total dust collected included the airborne
samples and the fines left on the surface.
8
|
Colorado School of Mines
|
After collection, the researcher analyzed the airborne dust and fines via a field emission scanning
electron microscope and standardized methods to capture the dust characteristics in terms of
concentration, silica content, particle size distributions, and particle shapes. The analysis included the
standardized NIOSH Manual of Analytical Methods 0600 and 7500 to determine the various
concentrations and the silica contents of samples. Further research utilized standardized laser diffraction
methods to determine the particle size distributions. Then, a field-emission scanning electron microscope
paired with an image analysis program allowed for the analysis of the particle shapes. Statistical studies,
such as the two-way ANOVA and Kolmogorov-Smirnov test, were utilized to highlight any significance
between pick wears or rock types. In the end, the resulting data can provide the future foundation for
subsequent research activities in dust characterization.
1.7 Originality
A literature review, as seen in detail in CHAPTER 2: LITERATURE REVIEW, identified a
limited number of papers and understanding of how respirable rock dust characteristics change, if at all,
concerning pick wear during excavation processes. Of the two primary articles identified, experiments
analyzed data using non-standardized methods, tested minimal rock samples, provided no quantitative
data on the pick wear values, and analyzed minimal characteristic parameters. Therefore, the originality
of the research presented in this thesis offers new methods, techniques, and dust characterization
parameters. The unique aspects of this research include:
Performing full-scale rock cutting tests with three symmetrical conical pick wears and
measuring the wears to obtain quantitative wear data;
Collecting the dust generated from three different rock types;
Collecting and analyzing both airborne dust and fines material with the most recent and
accurate instruments and standardizations;
Characterizing dust beyond concentration, focusing additionally on particle size distributions
and particle shapes;
Conducting an analysis that quantitatively connects pick wear to concentration, size
distribution, particle shape, and its relationship with the specific energy input of the cutting
system.
1.8 Limitations
The research conducted for this project has the following limitations:
9
|
Colorado School of Mines
|
Given the objectives, this research did not attempt to perform a toxicology study on the rock
dust samples generated. Such an attempt would require significant additional testing than the
project allows;
The collected and analyzed samples were representative dust samples from the larger body of
dust generated. However, the sample collection process was kept consistent within the testing
program to allow for a comparative study;
During testing, potential contamination of the samples occurred because of the casting of the
rock specimens in concrete jackets, which would lead to the collection of unwanted concrete
dust during the limestone and sandstone cutting. Therefore, the experiments enabled energy
dispersive x-ray spectroscopy analysis during the field emission scanning electron
microscope photo capture, and the researcher removed particles that contained an elemental
analysis of concrete from the limestone and sandstone datasets;
Cutting three full-scale rock samples allowed three duplicate trials in similar-strength rocks.
Future experimental programs can consider additional rock types with various rock strengths;
During full-scale cutting tests, some particles would coagulate on the filter surfaces, such that
during image capture or laser diffraction analysis, particles appeared larger and were
unrepresentative during the examination. Therefore, a Tsai Diffusion Sampler, a newer
instrument that more readily distributes particles on the collection surface [41], was used to
collect particles. During the field emission scanning electron microscope photo capture, the
researcher took images manually and visually checked for coagulation. Additionally, a light
ultrasonic sample prep was added to the laser diffraction methods to “shake” particles apart as
best as possible before analysis;
The researcher conducted the experiments with the limitations in mind to keep the testing as
representative as possible. Future research with an altered methodology may eliminate some limitations,
as provided in Section 7.3 Recommendations.
1.9 Organization of the Thesis
Chapter 1 of this doctoral dissertation introduces current rock excavation practices and issues in
the underground working environment due to dust exposure. Within this introduction, the chapter outlines
the primary objectives of this research, which focuses on the characterization of respirable rock dust
generated while cutting various rock types with multiple pick wear levels. Additionally, Chapter 1 details
the research hypothesis, overarching methodology, originality, and limitations.
10
|
Colorado School of Mines
|
Chapter 2 outlines the literature search, including current dust controls, previous cutting studies,
and current dust and lung interaction studies. With these topics outlined, Chapter 2 highlights the
knowledge gaps within the literature and provides the originality of this doctoral dissertation with the
research question and objective.
Chapter 3 describes the research methodology for full-scale cutting tests performed with three
different pick wear levels on three different rock types. This chapter includes the details of the developed
automated dust collection system, the specific dust collection equipment utilized, and the specific
standardized analyses performed.
Chapter 4 provides the data and results based on the standardized tests and the field emission
scanning electron microscope image analyses. Raw data is presented in addition to graphs and figures to
represent the data visually.
Chapter 5 discusses the research results and further analyzes the raw data. The chapter reveals the
statistical analyses and results, along with discussions of the statistical results.
Chapter 6 addresses the connection between the dust concentrations and the specific energy
required during cutting tests. Although the material in this chapter is outside the scope and proposal of the
thesis, it provides a foundation for further research. Figures and a discussion in the chapter reveal the
trends found.
Chapter 7 outlines the contributions and conclusions of the research project. There are also
recommendations for future work based on research findings.
11
|
Colorado School of Mines
|
CHAPTER 2: LITERATURE REVIEW
2.1 Current Dust Controls
The National Institute of Occupational Health and Safety (NIOSH) recommends implementing a
hierarchy of dust exposure controls to protect underground workers. As seen in Figure 2.1, their
recommended techniques range from the most effective, where dust elimination occurs at the source, to
the least effective, where workers wear personal protective equipment, such as a respirator [42].
Figure 2.1 NIOSH's hierarchy of controls for human hazards [42].
The least effective method to control workers' dust exposure requires personal protective
equipment, such as respirators. A more effective mode is administrative controls, such as hazard
awareness training and job rotations. Engineering controls are the next most effective because they reduce
the hazard at the source, such as installing a dust curtain between the dust source and the worker.
Furthermore, substitution controls are among the most effective because they replace the threat with a
different hazard, such as using a less toxic material in building operations. Lastly, the most effective
control for dust hazards is to eliminate dust from being generated in the first place. A great example is
designing and implementing transfer points that do not cause dust when material transfers. Therefore,
during rock excavation, cutting, and drilling, the ideal situation is minimizing the dust generation at the
source.
12
|
Colorado School of Mines
|
2.2 Previous Cutting Studies
As noted in Section 1.3 Respirable Particles, there are health concerns for workers exposed to
high rock dust concentrations. Therefore, many research projects in the 1970s set out to see what
connections they could make between pick cutters and dust concentrations. For example, the Bureau of
Mines performed multiple studies in the 1970s and 1980s on the airborne coal dust concentrations
generated from various pick geometries and pick wears on coal samples [17]–[20]. One research team in
the 1980s collected sandstone dust parameters from cutting with various conical pick-tip angles and point-
attack wear amounts [21]. Another group, most recently in 2019, performed small-scale cutting tests with
different pick tip angles to see the changes in concentration and particle size distributions of the dust [22].
Table 2.1 summarizes the previous experiments with various pick types and wear levels and the dust
parameters analyzed. Additionally, some previous experiments studied the forces and specific energy
generated from picks with different wear levels [43], [44].
Table 2.1 Previous studies of varying pick geometries compared to dust characteristics
Paper Title Picks tested Image of picks Rock and Dust
cut type parameters
Reduction of Dust Various pick Coal Concentration
and Energy During types:
Coal Cutting Using
Point-attack Bits Small
with an Analysis of Small bullet, scale
Rotary Cutting and Pencil, cutting
Development of a Minibob,
New Cutting Plum bob
Concept [18]
Effect of symmetric Plumb bob Coal Concentration
bit wear and attack at three even
angle on airborne wears:
respirable dust and Small
energy scale
consumption [19] New sharp, cutting
Rounded,
Worn blunt
13
|
Colorado School of Mines
|
In 1976, Roepke performed small-scale cutting tests on coal samples to determine the connection
of pick tip geometry to cutting forces and dust concentrations. He used a small bullet, pencil, minibob,
and plum bob (conical) pick to cut coal in his methods. He then collected dust with 3µm sized pore filters
and analyzed the concentration with an optical particle size monitor. Roepke's main conclusion is that
dust concentration is lower when cutting small-scale coal samples with sharper pick tips [18].
In 1979, Hanson performed small-scale cutting tests on a coal sample to determine the connection
between the symmetrical conical pick wear and dust concentration. He used three pick wear levels which
were qualitatively named “New sharp,” “rounded,” and “worn blunt,” and cut samples at multiple depths.
A machine artificially wore picks to obtain the rounded and worn blunt picks. A modified Bausch and
Lomb optical particle size monitor determined the concentration of dust. In the end, he concluded that the
depth of cut was the most significant factor in dust, and wear was a secondary factor [19].
In 1983, Roekpe and Hanson performed small-scall cutting tests on coal samples to determine the
connection between asymmetrical pick wear to conical picks on dust concentration. They used five pick
wear levels ranging from a new to a thoroughly worn conical cutter. Machines artificially wore the picks
asymmetrically, and cutting tests ensured the penetrations into the samples remained consistent. A
modified Bausch and Lomb optical particle size monitor determined the concentration of dust. Ultimately,
they concluded that as the picks wear down, the concentration of coal dust increases [20].
In 1988, Plis investigated the relationship between airborne respirable dust to a “new” and
symmetrically “worn out” conical pick with full-scale coal-cutting tests. He collected dust with 3µm sized
pore filters and analyzed the concentration with an optical particle size monitor. In the end, he concluded
that the new bits generally entrained more dust for a given cutting distance than worn-out bits.
In 1984, Fowell performed full-scale cutting tests on sandstone with three conventional wedge
types and one point-attack tool to compare the dust generation to the pick type. He used new, sharp, and
blunt tools to cut sandstone. The results simply state that “a unique relationship was established between
specific dust make and specific energy input of each tool type” [21].
In 2020, Zhou investigated the effect of the pick tip geometry on dust generation and particle size
distributions when cutting coal samples. He used small-scale cutting methods to cut coal with three
different conical pick geometries. The three geometries included pick angles at 87, 100, and 110 degrees.
An anti-static tray collected dust and deposited material during cutting, and a mass analysis provided the
dust generated. A laser particle size analyzer (Winner 2000, Jinan Winner Particle Instrument Stock Co.,
Ltd., Jinan, China) determined the particle size distribution of the airborne dust. The team concluded that
the mass rate of dust generation increases as the pick tip angle increases [22].
15
|
Colorado School of Mines
|
Of these previous rock-cutting tests that studied dust concerning pick tips, most studies used
various pick tip geometries. These studies analyzed the differences in dust concentration compared to the
angle of the different pick tips. Only two studies performed in the 1970s and 1980s used the same pick
with multiple stages of wear, one focusing on symmetrical wear and the other on asymmetrical wear. For
the experiment cutting with symmetric pick wears, the team used “new sharp,” “rounded,” and “worn
blunt” plumb-bobs (also referred to as conical cutters of point attack bits) to cut coal samples on a small
scale. They did not quantify the “new sharp” pick wear; however, they measured the radii of the
“rounded” and the “worn blunt” picks by superimposing a circle at the tip to find 6.4 mm and 9.5 mm
radii, respectively. Their results concluded that “Symmetric wear … proved to be a significant factor in
airborne respirable dust per centimeter [of cutting], specific energy, and average cutting force. The results
for airborne respirable dust per centimeter show a relatively large increase between the ¼-inch and 3/ -
8
inch radius bits”. These are positive findings and reveal that airborne respirable dust and energy input will
generally increase in small-scale coal cutting. However, it is uncertain how the airborne respirable dust
will increase when translated to full-scale cutting if this trend applies to non-coal samples and what the
quantitative increase of dust is concerning pick wear and tip geometry. This is because of the impact of
scale in rock excavation and drilling where the change in scale can impact the resultant forces and the true
size of the cutting tool, as well as cutting geometry can alter the results.
Additionally, of the previous experiments conducted, most of the collection equipment was
rudimentary and is now outdated. For example, in the investigation conducted by the Bureau of Mines
with different pick types, they collected dust with 3µm sized pore filters [18]. Filter media with pores as
large as 3µm will lose a significant amount of respirable dust in the size range that is potentially harmful
to humans. In the other experiments by the Bureau of Mines with various pick wears, the team used a
modified Bausch and Lomb optical particle size monitor from the 1970s, which only recognized particles
0.4 to 6.9µm in physical diameter [19], [20]. Using unregulated equipment makes it uncertain how
reliable the equipment collected and presented results. For example, laser diffraction is widely accepted
and used in today’s research studies, where particles are more accurately measured and counted with
standardized procedures and processes [45]–[48]. Recognizing particle sizes larger than 7µm in diameter
is vital because they can be inhaled into the respiratory tract and potentially cause coughing or mild health
effects [42].
Previous studies also only characterize the airborne dust in terms of concentration, whereas only
one study [44] on a small-scale setup analyzed the size distribution. A significant gap exists in
understanding the specific particle sizes within airborne dust concentrations. Only Zhou’s recent study
investigated the dust left on the surface, or the fines material.
16
|
Colorado School of Mines
|
2.3 Current Dust and Lung Interaction Studies
Exposure to high dust concentrations of particles less than 10µm is most likely harmful to long-
term human lung health [35]. However, it is unclear how specific particle shapes deposit throughout the
lungs and how they affect lung health.
Early studies involving respirable asbestos aspect ratio metrics revealed that the particles with
large aspect ratios, meaning aspect ratios greater than 40, were more readily entrained in the lungs [49].
Recent studies revealed more specific details of particle shapes, such as smaller aspect ratios reaching
deeper regions of the lungs. For example, one team found that “For spherical shape particles, traveling
time [throughout the lung] is less and it increases by increasing particle aspect ratio 10 to 1000” [50].
Another group of researchers found that particles with an aspect ratio as low as two will travel deeper into
the respiratory system: “… the rod-like particles with a larger aspect ratio [of 2 and 4] indicated higher
reachability into the depth of the simple respiratory model. This was attributed to the high velocity motion
of the particles whose long axis was in the direction of the deep region” [51]. Therefore, particles with
aspect ratios of 2 or more have higher chances of getting deeper into the lungs and can have a longer
residence time.
In another recent study in 2019, the findings revealed that when clustered nanoparticles create
more rough particle edges, the dynamic factor of the collective particle increased and resulted in a higher
percentage of deposition in the alveolar region of the lungs [52]. Therefore, particles with rougher edges
are deposited more readily in the deepest region of the lung. With additional research determining how
specific roughness values relate to lung deposition, future experiments can generate a more accurate
deposit model of particles with known roughness values.
The available research and publications do not draw any conclusions regarding particle
characteristics as they pertain to the incidence of pulmonary disease or causation factors to health
outcomes. The topic connecting the dust particle’s characteristics to human health is outside the scope of
the research. Instead, the study provides particle characteristic information with discussions of possible
implications.
2.4 Knowledge Gaps
There is a gap in the literature on the quantitative change in dust concentration, particle size
distribution changes, and particle shapes for dust generated with different pick wear levels. There are also
17
|
Colorado School of Mines
|
gaps in understanding of how particle shapes influence the deposition in human lungs, such as the
roughness of a particle. Therefore, the experiments in this thesis characterize and then compare rock dust
particles in terms of concentration, particle size distributions, and particle shapes for particles generated
with three pick wear levels. The list in Table 2.2 highlights the knowledge gaps in previous and current
rock-cutting and dust analysis studies that research aims to fulfill:
Table 2.2 Content of previous studies and the knowledge gaps present in the literature
Previous studies Knowledge gaps
Most of the previous tests were Lack of information and verification in full-scale
conducted in a small-scale setting. testing.
Most of the previous tests were focused Lack of information and measurements in different
on coal samples to analyze coal dust. types of rock samples excavated in mining operations.
Most of the previous tests used a variety There is a limited continued dust measurement on a
of pick types and pick tip geometries. single conical pick geometry at multiple stages of
symmetrical wear.
The cutting tests were performed in the There are limitations in testing and analyzing data with
1970s and 1980s with outdated and less more accurate and reliable instruments and accepted
accurate instruments for analyzing dust. regulated standardized tests.
The previous tests only analyzed dust The understanding of the dust in terms of quantitative
concentration. concentration data, particle size distributions, and
particle shapes is incomplete.
There is no quantitative analysis of the The interaction and quantitative connection between
connection between dust concentration dust concentration, machine input-specific energy, and
and machine-specific energy input to pick tip wear is poorly defined.
pick wear level.
All previous tests only collect and The investigation of the fines’ particle size distributions
analyze the airborne dust generated is incomplete.
during cutting.
18
|
Colorado School of Mines
|
Due to the knowledge gaps in the previous studies, various aspects of this research plan to
address these issues. By analyzing the concentration of dust for three non-coal samples (concrete,
limestone, and sandstone), the previous trends observed in coal dust concentration concerning pick wear
can be confirmed and extrapolated to other rock types. Moreover, analyzing airborne and fines particle
size distributions provides an understanding of particle shape concerning pick wear. The airborne particles
are the respirable particles in the air during cutting. The fines are the particles generated directly beneath
the crushing pick tip that remain on the sample's surface after cutting.
Since none of the previous studies examined the respirable particle shapes from cutting various
rock types with multiple pick wears, the particle shape analysis will offer new prospects for further
evaluating the relationship between the particle shape and its impact on the respiratory system. Overall,
the following list of project aspects is unique to the experiments conducted in this thesis:
Performing full-scale cutting tests at the Earth Mechanics Institute (EMI) to evaluate dust
generation in a controlled environment;
Using various symmetrical conical picks with quantification of the pick tip radii;
Cutting and generating dust with three new, different rock types to complement previous
measurements of coal dust;
Collecting and analyzing both the airborne and fines material with the most recent and
accurate instruments and standardizations;
Characterizing dust beyond concentration, focusing additionally on particle size distributions
and particle shapes;
Conducting an analysis that quantitatively connects pick wear to concentration and specific
energy input.
2.5 Research Question and Objective
With knowledge gaps in respirable dust characteristics from various rock samples, the driving
question for this research is as follows: How do respirable dust characteristics change and compare with
the symmetrical increase of pick tip radii for conical picks used in mechanical rock excavations from
cutting medium strength rock? Therefore, the thesis aim is to characterize and compare concentrations,
particle size distributions, and particle shapes of respirable dust generated from three medium-strength
19
|
Colorado School of Mines
|
rocks when cutting with three conical pick wear levels. The comparison between the three pick wear
conditions for each characteristic will reveal any potential trends.
By creating a compilation of dust characteristics in terms of concentration, silica presence, size
distribution, and shape of particles between 0.1µm and 10µm in aerodynamic diameter, the comparisons
drawn between the dust generated from different picks can guide bit management and dust suppression
practices. Because a toxicology study is outside this project's scope, there are no concrete
recommendations provided regarding which pick wear causes the most potentially harmful dust to
workers. Additionally, there is insufficient research and evidence regarding which specific dust
characteristics are potentially most harmful to the human respiratory system.
Therefore, observation of the shape of the dust particles can help in future studies and analysis of
the impact of dust from various rock types on workers’ health. The shape data can lead to developing
specific guidelines for cutter management, changing picks, or increasing dust suppression levels in the
future. The health issues are not thoroughly analyzed, and this thesis will not address bit management
guidelines. Instead, the document provides an analysis of the characteristics and comparisons of the dust
particles with the values and numerical trends. With the data provided in this thesis with discussions
provided in Chapter 5 and Chapter 6, follow-up studies can focus on bit management and dust
suppression decisions.
20
|
Colorado School of Mines
|
CHAPTER 3: METHODOLOGY AND EXPERIMENTAL SETTINGS
3.1 Full-scale Cutting Tests
Utilizing full-scale cutting tests at the Earth Mechanics Institute (EMI) allowed representative
dust particles to be generated similar to their generation in a working operation with mechanical
excavations. Three main constituents drove the full-scall cutting at EMI. These include the large samples
cut, the actual conical pick cutter tool used in practice to cut the rock, and the Linear Cutting Machine
(LCM) to provide power and consistency in cutting. Conical picks with various wear levels cut three
samples with duplicate cut lines to generate dust. Instruments then analyzed the dust to determine
differences in dust characteristics for each sample and tool wear.
3.1.1. Rock Samples Tested
The experiments tested and cut concrete, limestone, and sandstone. These samples have similar
strength properties and do not contain many rock joints or fractures. Additionally, the concrete, limestone,
and sandstone selected were primarily homogenous regarding mineral content and did not illustrate strong
signs of bedding.
The samples did not contain any joints or discontinuities and were structurally intact, which
provided uniform and homogeneous samples for consistent cutting tests. The concrete sample had a
uniform small aggregate mixture poured at the Earth Mechanics Institute (EMI). The limestone sample
was a uniform Indian limestone rock, and the sandstone sample was a less consistent sample, with a few
spots of discoloring.
Figure 2.2 A) Concrete sample freshly poured and leveled top B) Concrete sample after cutting.
21
|
Colorado School of Mines
|
Table 3.1 Rock properties of the three samples tested
Rock Types Cut UCS BTS CAI Punch Pen. energy slope index
(MPa) (MPa) (kN/mm)
Concrete 34.2 3.1 1.13 6.0
Limestone 43.4 3.0 0.57 9.4
Sandstone 80.3 6.4 1.7 13.9
The LCM cut the rock samples with all fresh and conditioned surfaces before dust collection
commenced. In other words, the cut lines generated and set into the samples before experimentation
eliminated interaction from previous cuts. Section 3.1.3. Linear Cutting Machine provides conditioning
details.
3.1.2. Conical Pick Cutters
New, moderately worn (assumed mid-life), and worn (assumed end-life) conical picks cut each
sample. Conical picks were chosen for experimentation because they are typical cutting tools for breaking
hard rock. Conical picks are used on various machines and in many mining and civil applications [26],
[53].
The same brand and make of conical pick and block cut each sample. The U92 4.0 is a conical
pick manufactured by Kennametal Inc with a 105 mm (4 in) gauge and 90-degree tip angle cut the
samples. The corresponding block with a straight sleeve shank, called K35, was used to hold the pick and
was welded to the saddle, as seen in Figure . A microscope quantified the rounded pick cone nose, as seen
in Figure 3.1. For example, a circle was superimposed at each pick tip to show the rounded nose radius
changing with wear. The new pick tip had a radius of 1.8 mm (0.07 in), the moderately worn pick had a
radius of 3.7 mm (0.15 in), and the worn pick tip had a radius of 5.2 mm (0.2 in). The tip's radius
quantitively represents pick wear because the tip cone angle will stay the same throughout cutting, and
only the tip itself will become more blunt or rounded over time. The research team did not alter the new
pick and cut the samples straight out of the box from the manufacturer. On the other hand, the research
team fabricated the moderately worn and fully worn picks by wearing down a duplicate new pick by hand
with a diamond-tipped Dremel® and lathe, as seen in Figure 3.23.2. The fully worn pick is referred to as
“fully worn” or “worn” because this pick wore to the point where the cap on the tip is gone.
With this, the tip radius contacting the rock face surface determines the size of the pressure
bubble under the pick tip at the contact point and the amount of surface area in contact with the sample.
The wider the tip, the higher the forces and the specific cutting energy [14], [51]. Therefore, it is most
likely worn picks will generate a higher volume of fines in cutting the same rock at identical spacing and
penetration.
23
|
Colorado School of Mines
|
The main components of the LCM include the sled, load frame, and saddle, as shown in Figure .
The sled is the platform the sample sits on once cast into a large metal box. Hydraulic cylinders push the
sled during testing to move the sample linearly while in contact with the cutting tools. The load frame
controls the penetration depth into the sample and ensures consistent penetration with hydraulic cylinders.
Inserting steel plates between the cross frame and the machine’s main structure fixed the penetration into
the rock sample. For a 5 mm (0.2 in) penetration depth, a 5 mm plate assured the accuracy of the
measured penetration. The plates allow stiffness in the cutting unit since the vertical load spreads over a
large area the metal sheet provides. Multiple 5 mm sheets or inserts with the appropriate thickness (i.e.,
2.5, 10, 12.5 mm) installed in the cross frame allow for applying selected penetration. Additionally, the
saddle is the interface connecting the load frame to the sample on the sled. The saddle holds the full-scale
cutting tools, as seen in Figure 3.5, and the dust collection equipment is later attached.
25
|
Colorado School of Mines
|
Figure 3.3 Linear Cutting Machine. A) A potash rock block cast into the metal box. Red square) Location
of the pick within the dust curtain. Green arrows) The movement of the sled and box under the stationary
pick to generate cuts. Grey arrows) Indicate the horizontal direction of the box to index the pick for the
next cut. Yellow arrows) Show the movement of the frame to control the penetration of cuts.
Each sample block was cast into its respective LCM steel box. The limestone and sandstone
samples were cast with concrete, poured, and mixed into the steel box. Once the concrete cured after a
few weeks, a forklift placed the sample onto the LCM sled for full-scale cutting. For these experiments,
the sled moved the rock box linearly such that the pick contacted and cut the sample’s surface at a
constant speed of 250 mm/sec (10 in/s) to simulate rock cutting in a full-scale cutting process. The length
of each cut line was 1.1 m (3.5 ft); therefore, each cut line took 4.2 seconds. Multiple cuts were made
across the sample’s surfaces, as seen in Figure 3.4, at 3.81 cm (1.5-inch) spacings, a typical cut spacing
for a machine cutting a similar rock with a UCS between 30 and 80 MPa. The group of cut lines made
26
|
Colorado School of Mines
|
horizontally across the surface of the samples at the same elevation created one pass across the surface
where dust was generated and collected. One dust sample was collected while cutting all the horizontal
lines for one pass at the desired spacing. A “test set” is the dust collected during one pass at the same
spacing and penetration for each of the three pick wears.
Figure 3.4 Concrete (left), limestone (center), and sandstone (right) samples cast in concrete and cut with
multiple lines across the surface along one elevation for one pass of cutting.
The picks' penetration, or indentation depth, into each sample, was 0.76 cm (0.3 inches) for
experiments. These parameters represent typical spacings and penetrations used for the specific rock types
in similar operations. Optimization of the specific energy and normal forces when cutting various rocks
guides the spacing and penetration combinations [54]. The angle of attack, or the angle of the pick in
contact with the sample, was 45 degrees from the horizontal. To simplify the drag and normal forces, a
45-degree angle of attack cut samples in these experiments. As seen in Figure 3.5, the team welded the
Kennametal block to the saddle at 45 degrees, and the conical pick inserts into the saddle.
27
|
Colorado School of Mines
|
Figure 3.6 Front view of the calibration with force at 9 degrees at the side.
Conditioning cuts across the sample surface before cutting addressed the potential impact of
propagation of horizontal and vertical fractures from previous cuts when starting at the top of the sample.
Conditioning provides the consistent internal propagation of fractures for the specific set of penetration
and depth of cuts, essential for simulating rock-cutting tests. The relief cuts, or the edge cuts on the far
left and right of the sample, were not used when collecting dust. Instead, dust samples were collected only
from the cuts made in the middle of the sample to ensure consistent samples from a realistic simulation of
an excavator.
3.2 Dust Collection
When cutting generates dust, some particles become airborne, while others remain on the
sample's surface. Therefore, this study analyzes both dust that becomes airborne and the material
remaining on the sample’s surface, called the fines material. The airborne dust is collected with
instruments that obtained particles 10 µm in aerodynamic diameter and less. The fines dust could also
29
|
Colorado School of Mines
|
contain deposited particles that were once airborne, and the collection instruments obtained particles 75
µm in physical diameter and less. The methods for collecting the airborne dust differed from those for
collecting the fines. There was some overlap with the instruments used to analyze airborne and fine dust,
which the following sections discuss.
Airborne dust was one of the main focuses of these studies because airborne dust can be
suspended near the nose and mouth, allowing for direct inhalation by workers. Airborne dust can also
contain particles with aerodynamic diameters 10µm or less such that these particles will stay airborne for
extended periods and reach the alveolar region of the lungs if inhaled. A slow settling velocity and
potential to reach the alveolar region of the lungs are two unideal characteristics for worker exposure to
dust because there is evidence that the presence of these airborne, respirable particles can lead to
compromising health effects [35].
Although currently on the sample's surface away from the breathing zone, the fines material can
quickly become airborne during operations and then inhaled. For example, material left on the surface of
rocks can be transported throughout a mine and, at transfer points, re-entrained into the air [42] Therefore,
the fines, which are defined as the particles 75 µm in physical diameter and less, were also collected and
analyzed because it could become airborne dust in real-world applications [42].
3.2.1. Airborne Dust
Two collection methods were used to obtain representative airborne dust from cutting the
samples, as seen in Figure 3.7 and Figure 3.8. The first method used nylon 10-mm Dorr-Oliver cyclones
with a 50 percent cut point of 4µm aerodynamic diameter particles. The second method utilized a Tsai
Diffusion Sampler (TDS) with a cutoff aerodynamic diameter of 3.8 µm particles [41].
30
|
Colorado School of Mines
|
Polycarbonate (PC) filters were used to collect dust and then analyze particles in a field emission
scanning electron microscope (FE-SEM) because the PC substrate is a viable background during FE-SEM
image capture [53], [55], [56]. PVC filters collected dust during experimentation to analyze concentration
and silica content. The PVC filters’ stable weight and high collection efficiency allow for consistent and
reliable gravimetric and mineral analysis [57]–[60].
Figure 3.25 shows three cyclones and one TDS to collect airborne dust while cutting the samples.
Two cyclones contained PC filters, and one cyclone included a PVC filter. There was an extra cyclone
with a PC filter to act as a backup. Because the nature of this study is comparative between samples and
the same cyclones were used throughout tests and cleaned between collection test sets, the efficiencies of
the cyclone collections were not measured or analyzed.
The three cyclones and one TDS were equidistant for the experimental setup and evenly spaced
around the pick to collect dust. Looking straight at the LCM as in Figure 3.3, the TDS was in the front left
position, the cyclone with PVC filter was in the front right position, and the two cyclones with PC filters
were placed at the back left and right positions.
The TDS and front right cyclone were attached to the LCM saddle with unique fabricated pieces,
and the two cyclones slid into yellow welded holding cylinders. The fabricated components were
designed, 3D printed, and installed with high-strength neodymium magnets to connect to the metal body
of the saddle. As seen in Figure 3.9, a small blue bracket clip threaded the TDS tubing and held the
instrument in place during testing. As seen in Figure 3.10, a black 3D-printed piece held the cyclone's
bottom. The bottom of the cyclone was covered with electrical tape to ensure a snug fit into the black 3D-
printed piece and to limit rattling during full-scale rock-cutting tests. As seen in Figure 3.11, fabricated
metal cylinders welded onto the saddle held the two cyclones with PC filters. The bottoms of the cyclones
were also covered with electrical tape to ensure a snug fit and to limit rattling during full-scale rock-
cutting tests.
Tygon® tubing connected the collection equipment to air pumps to draw in the dusty air once
attached to the saddle. The pumps connected to the cyclones ran at 1.7 liters per minute, the
recommended flow rate [61]. The pump connected to the TDS ran at 1.0 liters per minute, the
recommended flow rate for short-term sampling with high concentrations [41]. Due to the humidity of air
altering the collection efficiency of cyclones, experiments ran within a range that does not affect
collection efficiency, which was between 20% and 30% humidity of air [62]–[65]. This humidity range
was the environmental condition of the lab space at the Earth Mechanics Institute and does not necessarily
represent average working conditions.
32
|
Colorado School of Mines
|
Figure 3.11 Installed cyclone with the PC filter into the yellow metal holding piece welded to the LCM
saddle.
Once the experimenter installed the dust collection equipment, they placed the dust curtain
around the saddle, as seen in Figure 3.12 and Figure 3.13. The purpose of the dust curtain was to confine
the dust generated from the sample's surface during cutting and not allow outside unrepresentative dust
from being collected. It will enable a fixed volumetric space for the airborne dust collection equipment to
obtain dust samples. The dust curtain materials are thick, sturdy polycarbonate plastic on top and flexible
plastic sheeting at the bottom. The top part was attached to the saddle with high-strength neodymium
magnets and contained tight-fitting holes for the Tygon® tubing to thread through. The black ribbed flush
tubing also tightly fits into the top part of the dust curtain to provide high-efficiency particulate air
(HEPA) filtered air into the space after each cut line. The bottom was fringed so that the curtain smoothly
ran along the sample's surface during cutting movement.
34
|
Colorado School of Mines
|
Figure 3.13 Side view of the dust curtain installed around the saddle with air tubes piped out the side.
The pumps used to collect air also needed to be checked and maintained throughout testing to
ensure consistency in flow. Gillian BDX II pumps paired with the PC filters, and AirChek® XR5000
pumps paired with the TDS and cyclone with PVC filters. Before testing, the airflow rate calibrator
validated the pumps’ flow rates. The BIOS international DryCal DC-1 flow calibrator, as shown in Figure
3.14, was used to verify flow rates. In series, the dust collection equipment was installed with respective
filters, connected with Tygon® tubing to the calibration bubble cell, and then connected with more
Tygon® tubing to the pump. As seen in Figure 3.15, a rotameter in series checked the calibration such that
the rotameter verified pump rates during testing. Lastly, after charging the pumps, they ran for 10 minutes
to obtain a stable battery and flow rate before installation and experimental use.
36
|
Colorado School of Mines
|
Shop-vac® with HEPA filter. Vacuum serial number 16016 R 1052 Model WD09701,
120-Volt, 3.8 Amps, 60 Hz. RIGID® HEPA filter VF6000, 5-layers from Home Depot.
o The purpose of the shop-vac® was to take air from the lab space, pass the air
through a HEPA filter, and then push fresh air into the dust curtain space after
each cut line to clear out the area for the next cut.
Four electric ball valves. Motorized ball valve ¾”, stainless steel with full port, 9-24 V
AC/DC, two-wire auto return from U.S. Solid.
o The purpose of the electric ball valves was to start and stop the pumps from
pulling in air through the airborne dust collection equipment after each cut line.
The display monitor provides a simple check to validate the consistency of the
ball valve closing. Each time the ball valves close, the pumps’ flow rates would
decrease, and the display monitor would read “closed” for each valve.
Low flow constant pressure controllers. Used for 5 to 200 ml/min flow rates, part number
224-26-CPC from SKC.
o The pressure controllers placed in line with the Tygon® tubing between the
pumps and the electric ball valves were to keep the Gillian BDX II, and AirChek
pumps from stalling and shutting off when the ball valves closed or cutting the
flow.
As seen in Figure 3.21, the assembled automated dust collection system provided four separate
dust collection routes, one for each of the dust collection equipment instruments inside the dust curtain. A
rotameter was added in series between the electric ball valve and the TDS to ensure the desired flow rate.
A real-time dust concentration monitor was also added where Tygon® tubing connects the monitor to the
inside of the dust curtain near the airborne dust collection equipment. The purpose of the real-time dust
concentration monitor was for the operator to ensure that the shop-vac® flushing system purges enough
dust out of the dust curtain before the next cut began.
Figure 3.22 shows the display and control monitor and the pumps with the constant pressure
controllers installed immediately downstream of the pump inlet. The display control monitor was
programmed to keep track of the amount of time the pumps were pulling in air through the airborne dust
collection equipment. It only logs the time when the electric ball valves are open, so the experiment can
be closely monitored with the accurate amount of time the dust collection equipment collected air. There
41
|
Colorado School of Mines
|
is a screen with a timer visible to see the dust collection time and a button to reset the timer if desired.
Other screens that the control monitor include:
Home screen for navigating to desired screens;
Ball valve indication screen to show the operator which ball valves are open or closed;
Ball valve override screen to allow the operator to override the electric ball valves to
either open or closed;
Flushing system, or shop-vac®, indication and override page to see if the shop-vac® is on
or off, and gives the operator option to override the shop-vac®;
The laser reading screen shows what values the laser is outputting and the PLC is
reading;
A total systems page allows the operator to shut down or turn off the system if desired.
Figure 3.21 (A) Data acquisition interface and power unit that controls when the dust is collected and
flushed out of the system; (B) Electric ball valves to control the on and off of air flow through the
cyclones and TDS; (C) Pumps running at 1.7 or 1.0 L per minute; (D) Real-time dust concentration
monitor, (E) Rotameter currently placed in-line before testing to ensure pumps are running at an
appropriate flow rate.
42
|
Colorado School of Mines
|
Figure 3.22 The pumps installed with the pressure controllers to the left and the display screen to the
right.
The laser measured the location of the rock box concerning the start and end of a cut line, and the
PLC data interface, programmed with ladder logic, used these readings to open and close the electric ball
valves. The PLC also utilized the laser reading to determine the timing of the flushing system. When the
LCM cut the sample, and the rock box was near the laser, the PLC commanded the ball valves to be open
and the flushing shop-vac® to turn off. After the rock box left the laser’s sight, such that the laser read a
significant distance, a timer started in the PLC. After 30 seconds on the timer, as reached, the PLC closed
the electric ball valves to stop collecting airborne dust within the dust curtain and turned on the flushing
system to deliver fresh air into the dust curtain.
Another timer started once the rock box retracted after the cut to reset for the next cut. During
retracting, the rock box was pulled in front of the laser such that the laser read a close distance. The PLC
then started a 45-second timer after receiving the “short distance” reading. After the countdown, the PLC
opened the electric ball valves and stopped the flushing system from pushing fresh air into the dust
curtain space. The system was then ready for the following cut line to commence and collect dust.
The automated dust collection system ensured consistency and more reliable comparability
between dust generated with the various picks on multiple samples. A rotameter verified pump flow rates
before and after cutting. A real-time concentration monitor ensured the flush system removed all particles
43
|
Colorado School of Mines
|
in the dust curtain before the next cut began. Ultimately, the automated dust collection system was a vital
tool for airborne dust collection in these studies and for future dust studies at EMI.
3.2.3. Fines Generated by Cutting Tests
The method to collect fines material dust, or the crushed material left on the sample’s surface,
used a vacuum or 5-gallon shop-vac®. This vacuum collected the particles left behind along the lines of
cuts, or the crushed zone under the bit, on the rock's surface. The fines remain on the surface of the rock
block after cutting because the rock sample is crushed underneath the pick and does not become airborne.
The purpose of collecting and analyzing the fines is that in an actual mine operation, these particles could
be transported into downstream processes and potentially introduced into the air at transfer points. The
particles that would most likely become airborne and stay airborne for an extended time would be the
respirable particles less than 10µm in aerodynamic diameter.
At the end of the cut lines across the sample's surface, the boundary lines marked the consistent
area to collect fines. The boundary lines were the same for each sample to provide the same collection
area with the vacuum for each pass. As seen in Figure 3.23, two measuring tapes perpendicular to the cut
lines marked the boundaries for the area of fines material collected. Additionally, the dust generated from
the relief cuts, or the first and last cuts, was not collected. This is because the edge relief cuts have
different internal fracturing characteristics compared to the interior cuts [29], [66], [67].
44
|
Colorado School of Mines
|
Figure 3.24 The vacuum used to collect the fines on the surface of the samples.
3.3 Dust Analyses
Multiple analyses obtained a broad characterization of the dust particles. The experiments
characterized the airborne respirable particles less than 10µm in aerodynamic diameter by concentration,
mineral content, particle shape, and particle size distribution. Particle size distributions characterized the
fines left on the sample's surface.
The suspended particle characteristics analysis included the concentration, percent silica, particle
size distribution, and particle shapes. The NIOSH Manual of Analytical Methods 0600 determined
concentration, and the NIOSH Manual of Analytical Methods 7500 determined the silica contents. Laser
diffraction methods determined the particle size distributions, and a field-emission scanning electron
microscope (FE-SEM) paired with an image analyses program characterized the particle shapes.
On the other hand, the fines characteristics analysis included the particle size distributions and
particle shapes. The laser diffraction methodology paired with an internal optical meter obtained the
particle size distribution and took images of the particles for particle shape analysis. As seen in Figure
3.25, each dust collection equipment references a specific dust analysis technique.
46
|
Colorado School of Mines
|
Figure 3.25 Matrix of the purposes and analyses performed on dust once collected with the specific dust
collection equipment.
3.3.1. NIOSH Manual of Analytical Methods
A professional third-party laboratory performed the NIOSH Manual of Analytic Methods
(NMAM) analyses on samples. The NMAM 0600 analysis provides a gravimetric technique to determine
the mass of the respirable dust fraction in a sample collected with a 10-mm nylon cyclone with PVC pre-
weighed filter [61]. The mass concentration is measured with this method for any non-volatile respirable
dust and then converted to mass per volume concentration with the known pump run time. For example,
one test might run 11 minutes and another 9 minutes, with the pumps running at 1.7 L per minute. The
concentration values are converted by dividing the dust mass collected by the air volume that passes
through the respective pump, knowing the run time and pump flow rate. The NMAM 0600 is a standard
test used in the United States by industry to regulate and control dust exposures. Note that the samples are
not considered “end of shift” but rather an average per sample duration.
A professional third party performed the NMAM 7500 and provided the micrograms of
cristobalite, quartz, and tridymite detected per sample. The NMAM 7500 analysis provides an x-ray
powder diffraction technique to determine the amount of crystalline silica of the respirable dust fraction in
a sample collected with a 10-mm nylon cyclone with a PVC filter [61]. The NMAM 7500 is a standard
test used in the United States by industry to regulate and control dust exposures. Note that the samples are
not considered “end of shift” but rather an average per sample duration.
47
|
Colorado School of Mines
|
For the gravimetric analysis in the NMAM 0600, the lab used a Mettler model XP6 balance with
0.001 mg accuracy. The media blanks used for the level of detection and limit of quantitation were
provided from the laboratory as pre-weighed PVC filters from the same filter series, as seen in Figure
3.26. Additionally, some samples underwent a replicated analysis, where all the replicate results were
within the 20% relative percentage difference limit.
Figure 3.26 The pre-weighted filters in closed cassettes obtained from the third-party lab.
Once the raw data was returned from the third-party lab in mass units per sample, the researcher
calculated the concentration per sample with the known pump run time and flow rate. For example, if the
pump flow rate was 1.7 liters per minute and the run time was 9 minutes for a specific sample, the total
air volume passed through the sample was 15.3 liters. The researcher then converted liters to cubed
meters and divided the mass for the particular sample by the cubed meters volume. The result is the mass
per volume, the concentration value for the specific sample, or milligrams per cubed meter. This is the
standard method for representing airborne dust in mining and construction.
One cassette was filled with dust per test set, resulting in a total of 27 cassettes analyzed. In other
words, three rock types, three pick wears, and three duplicates resulted in a 3x3x3 testing matrix with 27
cassette samples. As seen in Section 5.1 NIOSH Manual of Analytical Methods, the R2 values between
test sets range between 0.81 to 0.99, with over half the values above 0.90.
3.3.2. Field Emission Scanning Electron Microscope and Image Analysis
A field emission scanning electron microscope (FE-SEM) and image analysis program captured
and processed images from the polycarbonate (PC) filter surfaces to determine the particle shapes. A
Tescan FE-SEM at a voltage of 15 kV with back scattering electron (BSE) detection, as seen in Figure
3.27, was used to obtain images, with twenty photos taken per PC filter, as seen in Figure 3.28. Then, the
48
|
Colorado School of Mines
|
Figure 3.29 Raw SEM image (left), the image under grey thresholding (middle), and the program
identifying and highlighting individual particles in blue (right).
The program calculated the aspect ratio, roughness, and roundness values for particles less than
10µm in aerodynamic diameter. The aspect ratio refers to a particle’s elongation. This value is essential
because particles with higher aspect ratios, such as asbestos fibers, can more readily be trapped deep into
the lungs and cause adverse health effects [49], [71]–[73]. Roughness refers to the undulations along the
perimeters. This value is crucial because particles with more rough edges have a higher percentage of
deposition in the alveolar region of the lungs due to the change in the particle’s dynamic shape factor
[52]. For example, as the dynamic shape factor changes, the particle’s settling velocity changes [35] and it
is possible particles can travel differently and deeper through the respiratory system [50], [52] causing
varying health effects. Lastly, roundness refers to how close a particle is to a perfect circle and considers
both the aspect ratio and the roughness. Although this value is less used when studying particle deposition
in the human lung, it is analyzed and provided in this research.
The program removed particles less than 0.25µm in physical diameter from the data set. This is
because the image-analysis program detected and counted the unrepresentative filter media holes as
particles because the holes were 0.2µm in diameter. Additionally, the program removed particles greater
than 10µm in aerodynamic diameter from the data set because particles from 10µm to 100µm in
aerodynamic diameter are more likely to travel and deposit in the thoracic (trachea) and nose region of the
respiratory system, where irritation effects occur [35]. Particles less than 10µm in aerodynamic diameter
can penetrate deep into the alveolar region of the lungs, where most long-term and harmful lung diseases
occur [35]. The lower boundary limit of 0.25µm does not impact the overall conclusions of the results, as
the scope covers particle shapes from 0.25 to 10µm in physical diameter. When more accurate
instrumentation is readily available, researchers can analyze nanoparticles.
50
|
Colorado School of Mines
|
The program calculated particle roundness to determine how close to a perfect circle the particle
of interest was via Equation 3.1 [68] to obtain a value incorporating the aspect ratio and roughness into
one value. The program then calculated the particle aspect ratio to determine the elongation of particles
via Equation 3.2 [68]. Finally, it calculated particle roughness to assess the smoothness of the particle
perimeter via Equation 3.3 [68]. The convex perimeter was the particle roughness, where the convex
perimeter was the object's perimeter if a rubber band was placed around the particle, as shown in
Equation 3.4 [68]. With this, length is the longest measurement across an object, and width is the shortest
distance measured across an object.
4 (area)
Roundness = (3.1)
π (length)2
length
Aspect Ratio = (3.2)
width
Convex Perimeter
Roughness = (3.3)
Perimeter
π
ConvexPerimeter = ∑ferets(cid:4674)2tan(cid:4672) (cid:4673)(cid:4675) (3.4)
2(number of ferets)
3.3.3. Laser Diffraction
The laser diffraction method analyzed the suspended airborne dust and the fines. The values
obtained via laser diffraction methods include the particles' physical diameters, which are put into size
bins to create distribution curves. The researcher converted the physical diameters into aerodynamic
diameters, as detailed below.
A Microtrac SYNC instrument ensured valid laser diffraction methods, as seen in Figure 3.30.
The Microtrac SYNC follows the international standard ISO 13320:2020 for laser diffraction analysis of
particle size. The physical diameters of particles were obtained from the instruments and converted to
aerodynamic diameters. The conversion to aerodynamic diameters is vital because aerosol studies of
particles’ behavior within the human lung use the aerodynamic diameter parameter. Aerodynamic
diameter considers the particle density and shape such that industrial hygienists can predict where the
particles will travel and deposit within the lung [74], [75].
51
|
Colorado School of Mines
|
(cid:3435)6.21× 10-4(cid:3439)T
Cunningham correction factor = C = 1+ (3.8)
da da, no slip
aerodynamic diameter without slip correction= d =d ×(cid:3495)
(cid:3096)(cid:3174)
(3.9)
(cid:2911),(cid:2924)(cid:2925) (cid:2929)(cid:2922)(cid:2919)(cid:2926) (cid:2926)
(cid:3096)(cid:3116)×(cid:3102)
A = area
d = physical diameter of the particle measured
p
T = temperature = 293K = 20˚C
= density of concrete particle = 2,400 kg/m3 [76]
p
= density of water = 1,000 kg/m3
0
= dynamic shape factor = 1.36 for quartz [35]
3.4 Ambient Air and Background Examination
Before any testing, the ambient air analysis determined if the methods could accidentally collect
additional particles in the laboratory environment. Ambient air is the air inside the laboratory before rock
cutting commences. It is critical to determine the possibility of foreign particles collected during full-scale
cutting to remove these unrepresentative particles from the analysis. For this reason, ambient air samples
were collected with PVC and PC filters in-line with cyclones and analyzed with NMAM 0600, NMAM
7500, and the FE-SEM.
Collection times were between 5 and 7 minutes, a typical interval for dust collecting during full-
scale cutting tests. Pumps connected to the cyclones and filter ran at a consistent 1.7 liters per minute, and
experiments collected dust samples for various conditions present in the laboratory environment. For
example, collections were made for environments when:
The doors by the LCM were open and closed;
The far ventilation fan was on and off;
The garage door was open and closed;
The purge system was running for a specific typical number of passes;
The front panel of the purge system removed.
53
|
Colorado School of Mines
|
As seen in Table 3.2, the ambient air collected for all combinations of environments contains an
undetectable concentration of particles. Sample PZ234770295 is an outlier because it is the only sample
that detected any concentration. The particles detected on this sample could have been accidentally
deposited from pre- or post-handling and not during actual collection time.
Table 3.2 Data results for NMAM 0600 and NMAM 7500 when collecting various scenarios of ambient
air in the laboratory
In addition to the NMAM standard analyses, an FE-SEM analysis tested the configurations in
Table 3.3. The researcher compared ambient air samples to cutting concrete and potash samples. Ambient
air Sample 1 ran for five minutes with the automated dust collection system running for five hypothetical
lines and collected particles with a cyclone, PC filter, and pump running at 1.7 LPM. Then, with a new set
of filters, the concrete sample cut five lines at 0.2” penetration depth, with 1.5” spacing, using a new
conical pick. Equipment collected concrete dust for five minutes with a cyclone, PC filter, and pump
running at 1.7 LPM during cutting.
Ambient air Sample 2 ran for seven minutes with the automated dust collection system running
for five hypothetical lines and collected particles with a cyclone, PC filter, and pump running at 1.7 LPM.
Then, a new radial Sandvik pick cut five lines of a potash sample at 0.6” penetration depth and 4”
spacing. Potash dust is collected for 7.5 minutes with a cyclone, PC filter, and pump running at 1.7 LPM.
54
|
Colorado School of Mines
|
Table 3.3 Summary of the samples collected with the respective test setups and collection times with the
pump rate at 1.7 liters per minute
Sample Run time Penetration Spacing Pick type
(Minutes) (inches) (inches)
Ambient air 1 5 - - -
Concrete 1 5 0.2 1.5 New conical
Ambient air 2 7 - - -
Potash 1 7.5 0.6 4 New radial
After collecting samples, nine images were taken at evenly spaced locations on each filter once
loaded into the FE-SEM. All photos captured the filter surfaces at a 50-micron scale bar, and the Clemex
program described in Section 3.3 Dust Analyses processed the images. The program detected particles
and determined the amount of filter surface area covered by particles in each image. As seen in Figure
3.31, there is visually a significant difference in the number of particles present on the filter used to
collect ambient air compared to the filter used to collect dust generated from full-scale concrete cutting. A
200-micron scale bar revealed a large area of the filter surface with the additional availability to see
individual particles. Figure 3.32 displays the energy-dispersive x-ray spectroscopy (EDS) results of the
ambient air filter that obtained particles of various compositions versus the concrete sample containing
only particles of SiO as expected.
2
55
|
Colorado School of Mines
|
Table 3.4 reveals the numerical analysis of the surface area covered by dust on the various filters.
The comparison of surface areas collected for ambient air Sample 1 and Sample 2 to the concrete and
potash testing reveals minimum ambient air particles. In other words, comparing the appropriate ambient
air test to the respective rock dust test shows that the ambient air samples add insignificant particles.
Ambient air Sample 1 would cover 0.3% surface of the concrete sample, and ambient air Sample 2 would
cover 0.2% surface of the potash-filled filter sample.
Table 3.4 Results from Clemex image analyses of the various air collection setups
Filter Area covered (µm^2)
Ambient air Sample 1 2,752
Concrete 822,847
Ambient air Sample 2 61
Potash, new pick 28,892
In the end, ambient air tests collected negligible airborne dust particles. Therefore, the ambient air
is neglected during testing and analysis for the experiments performed.
3.5 Statistical Analyses
Various statistical analyses were performed with the data after collection to test differences
between dust data collected from cutting with the three pick wear levels. A two-way ANOVA statistical
test analyzed the effect of pick wear and rock type on concentration, silica content, and particle shapes for
all three pick cutters. Two-way ANOVA tests are reliable in testing the effect of one input variable on an
output variable. The p-value threshold value for the tests performed is 0.05, such that values below the
threshold correspond to sufficient evidence that the output variable is affected by the input variable.
Kolmogorov-Smirnov (KS) tests compared the airborne and fines particle size distributions
generated from the three pick wears for the three rock types. KS tests are reliable in providing evidence of
57
|
Colorado School of Mines
|
CHAPTER 4: REVIEW OF THE RESULTS
With methods outlined in Chapter 3, the full-scale cutting tests utilized specific parameters and
metrics. Table 4.1 reveals the particular testing parameters ran on the Linear Cutting Machine (LCM), and
the following sub-sections provide the specific samples collected and raw data. The precise parameters
stemmed from the literature review and previous studies as detailed in Section 3.1 Full-scale Cutting
Tests.
Table 4.1 Consistent parameters used throughout the testing
LCM parameter Fixed value
Cutting speed 10 inches per second
Line spacing 1.5 inches
Angle of attack 45 degrees
Penetration 0.3 inches
Number of cut lines per dust collection (“test set”) 5 lines (6 lines for sandstone samples)
Number of dust collection sets 3 repeated tests for each pick wear scenario
4.1 NIOSH Manual of Analytical Methods
The NIOSH Manual of Analytical Methods (NMAM) 0600 determined the concentration of
airborne respirable dust generated from each pick wear when cutting each rock type. The NMAM 7500
determined the micrograms of silica per filter sample of airborne dust and the percentage of silica per
sample. One cassette was filled with dust per test set, resulting in a total of 27 cassettes analyzed. In other
words, three rock types, three pick wears, and three duplicates resulted in a 3x3x3 testing matrix with 27
cassette samples. Note that the sampling times are slightly different because it is difficult to align all the
moving parts of the linear cutting machine during full-scale cutting tests in each test set.
The NMAM 0600 and NMAM 7500 analyzed each cassette, yielding a value for the
concentration and percent silica for each test set cassette. The raw test setup data, the determined
respective concentration, and the percent silica are provided in Table 4.2, Table 4.3 and Table 4.4. The
color coding of the tables follows the pick tip wear parameters. For example, the dust collected from
cutting with the new pick colored rows in shades of green, the dust collected from cutting with the
59
|
Colorado School of Mines
|
moderately worn pick colored rows in shades of yellow, and the dust collected from cutting with the worn
pick colored rows in shades of orange.
Note that the environmental data of temperature and percent humidity of air for the concrete dust
test sets were not collected. It was not fully understood then that these parameters were necessary to
ensure cyclone consistency until after the concrete cutting tests. Lastly, Table 4.5 and Table 4.6 provide
the raw final concentration and percent silica data organized for all test sets for all the rock types cut in
one organized location.
Table 4.2 Collected test data when cutting concrete to obtain dust concentration and percent silica
Average
Pump run time Pump rate Concrete Testing Volume Mass Concentration Silica per sample
min L/min Names L mg/sample mg/m3 % silica
9 1.7 New pick, Test set 1 15.3 2.3 150 25
6 1.7 New pick, Test set 2 10.2 1.5 147 23
10 1.7 New pick, Test set 3 17 1.4 82 27
5 1.7 Mod pick, Test set 1 8.5 2.2 259 23
4 1.7 Mod pick, Test set 2 6.8 1.1 162 28
8 1.7 Mod pick, Test set 3 13.6 2.0 147 22
5 1.7 Worn pick, Test set 1 8.5 2.7 318 27
4 1.7 Worn pick, Test set 2 6.8 1.5 221 25
5 1.7 Worn pick, Test set 3 8.5 2.6 306 26
Table 4.3 Collected test data when cutting limestone to obtain dust concentration and percent silica
Pump run Pump Humidity Average Silica per
time rate Limestone Testing Volume Temp of air Mass Concentration sample
min L/min Names L °F % mg/sample mg/m3 % silica
19 1.7 New pick, Test set 1 32.3 78.2 21% 0.81 25 11
26 1.7 New pick, Test set 2 44.2 78.9 30% 1.1 25 16
15 1.7 New pick, Test set 3 25.5 77.9 32% 1.0 39 21
14 1.7 Mod pick, Test set 1 23.8 73.5 41% 0.95 40 9
13 1.7 Mod pick, Test set 2 22.1 76.4 26% 1.5 68 22
14 1.7 Mod pick, Test set 3 23.8 79.8 24% 3.1 130 14
18 1.7 Worn pick, Test set 1 30.6 73.7 23% 2.6 85 24
15 1.7 Worn pick, Test set 2 25.5 75.2 23% 3.8 149 16
13 1.7 Worn pick, Test set 3 22.1 75 22% 3.1 140 7
60
|
Colorado School of Mines
|
Table 4.4 Collected test data when cutting sandstone to obtain dust concentration and percent silica
Pump run Pump Humidity Average Silica per
time rate Sandstone Testing Volume Temp of air Mass Concentration sample
min L/min Names L °F % mg/sample mg/m3 % silica
17 1.7 New pick, Test set 1 28.9 68.3 20% 1.3 45 42
15 1.7 New pick, Test set 2 25.5 69.2 20% 2.1 82 44
15 1.7 New pick, Test set 3 25.5 66.3 19% 3.0 118 54
19 1.7 Mod pick, Test set 1 32.3 64.9 21% 1.9 59 43
21 1.7 Mod pick, Test set 2 35.7 65.3 21% 8.0 224 55
23 1.7 Mod pick, Test set 3 39.1 65.8 21% 17.0 435 56
15 1.7 Worn pick, Test set 1 25.5 58.8 21% 3.4 133 48
20 1.7 Worn pick, Test set 2 34 60.8 21% 12.0 353 46
20 1.7 Worn pick, Test set 3 34 59.5 21% 17.0 500 59
Table 4.5 Combined data results for the dust concentrations obtained in various rock types, units of mg/m3
NMAM 0600 Concrete Limestone Sandstone
(mg/m3)
New Mod Worn New Mod Worn New Mod Worn
Test set 1 150 259 318 25 40 85 45 59 133
Test set 2 147 162 221 25 68 149 82 224 353
Test set 3 82 147 306 39 130 140 118 435 500
Table 4.6 Combined data results for the percent silica measured by sampling dust during the full-scale
testing in different rock types on each of the filters
NMAM 7500 Concrete Limestone Sandstone
(% silica) New Mod Worn New Mod Worn New Mod Worn
Test set 1 25 23 27 11 9 24 42 43 48
Test set 2 23 28 22 16 22 16 44 55 46
Test set 3 27 25 26 21 14 7 54 56 59
From the results obtained in Table 4.5 and Table 4.6, visual representations of the data Figure 4.1
and Figure 4.2 as histograms. The histogram data is organized by rock type and test set groups.
61
|
Colorado School of Mines
|
4.2 Particle Size Distributions
Laser diffraction determined the particle size distribution for the airborne particles and fines
material, as outlined previously. The “percent channel” used for analysis results refers to the bins during
particle counting, where each particle was individually analyzed and placed into bins dependent on size.
The physical diameters obtained were then corrected to aerodynamic diameters using the Cunningham
and slip correction outlined in Chapter 3, sub-section “Laser Diffraction.”
The laser diffraction instrument detected around 10,000 individual particles for each test set
analyzed. With the 10,000 particles per test set, the program determined the average curve for each pick
wear. In other words, each colored curve displayed in the following figures is the average of three
duplicate sample runs.
4.2.1. Airborne Particles
Figure 4.3 reveals the suspended particle size distributions for the three pick wears generated
from the three rock types. The raw physical diameter particle size bins in µm used for percent channel
data collection include the following: 26.16, 22, 18.5, 15.56, 13.08, 11, 9.25, 7.78, 6.54, 5.5, 4.63, 3.89,
3.27, 2.75, 2.31, 1.95, 1.64, 1.38, 1,16, 0.97, 0.818, and 0.688 µm. These twenty-two bins allowed an
optimal balance between processing time and a fine enough resolution to obtain a valid particle size
distribution. Once the instrument collected data, the researcher converted the physical diameters to their
aerodynamic diameters.
66
|
Colorado School of Mines
|
4.3 Particle Shapes
The field emission scanning electron microscope obtained images of the dust particles, as seen in
Figure 4.5. The Clemex program processed the images and calculated the particle shapes regarding
roundness, aspect ratio, and roughness. The number of particles detected and analyzed are in Table 4.7 for
each rock type and respective pick wear levels. Table 4.8 reveals the average roughness, aspect ratio, and
roughness values of the dust generated from cutting the various rock types with the three stages of pick
wear. The histogram-based visual representation of the roundness, aspect ratio, and roughness is provided
in Figure 4.6, Figure 4.7, and Figure 4.8, respectively.
Figure 4.5 FE-SEM sample images of suspended dust generated from the moderate pick while cutting
concrete (left), limestone (center), and sandstone (right).
Table 4.7 Number of airborne particles analyzed for particle shapes from the Polycarbonate filter surfaces
captured with the images using the FE-SEM
Rock type New pick Moderately worn pick Worn pick
Concrete 2,578 2,298 1,101
Limestone 1,043 1,377 1,416
Sandstone 779 1,112 1,123
Table 4.8 Average data results for various particle shape characteristics
Rock type New Mod Worn
Roundness
Concrete 0.57 0.57 0.56
Limestone 0.57 0.58 0.57
Sandstone 0.57 0.57 0.56
Aspect Ratio
72
|
Colorado School of Mines
|
CHAPTER 5: DATA ANALYSIS
This chapter provides further insight into the data results by including statistical analyses
performed on the data sets and discussions on the research and statistical results.
5.1 NIOSH Manual of Analytical Methods
5.1.1 NMAM 0600 Dust Concentration Analyses
The NIOSH Manual of Analytical Methods 0600 determined the concentration of airborne
respirable dust generated from each pick wear when cutting each rock type. Table 4.5 and Figure 4.1
reveal the concentration values obtained for each test set.
A two-way ANOVA statistical test conducted in MATLAB® tested if rock type and pick wear
level affect dust concentration and if there is a meaningful interaction effect between rock type and pick
wear. The p-value for rock type is 0.0128, the p-value for pick wear is 0.0064, and the p-value for their
interaction was 0.7023. The p-value threshold for this analysis is assumed to be 0.05, where values below
0.05 correspond to sufficient evidence that the output variable is affected by the input variable. Therefore,
strong statistical evidence suggests that both pick wear and rock type individually affect the concentration
of dust generated while cutting. However, there is weak statistical evidence that the rock type and the pick
wear interact to affect the dust concentration. This could be expected due to the limited number of cutters
and rock types in the experimental data.
As revealed in Table 4.5, Figure 4.1, and the ANOVA statistical tests, the concentration of
suspended respirable dust consistently increased as the wear of the pick increased for all three rock types
cut. The trend is repeated and confirmed throughout all nine test sets and agrees with previous studies of
coal samples cut with various pick wears [19], [20], [22]. Therefore, operators could consider increasing
measures to mitigate and suppress dust as picks age and wear down with more rounded tip radii.
Further analysis reveals the concentration data trending between the measured amount of pick
wear. The quantitative concentration data plotted and compared to the quantitative pick tip radii data is
displayed in Figure 5.1 A, B, and C for all rock types. This figure shows consistent positive linear trends
between all the test groups for each rock sample. The smallest slope increased by 17.197 mg/m3 of dust
generated per millimeter of radius wear (R2 value 0.882) during the first test set on the limestone sample.
The most significant slope increased by 114.83 mg/m3 of dust generated per millimeter of radius wear (R2
value 0.915) during the third test set of the sandstone sample. The average slope for all nine test groups
75
|
Colorado School of Mines
|
was 48.7 mg/m3 per millimeter of radius wear, with an average R2 value of 0.908 for all nine test groups.
Additionally, for all cases, it is noted that the dust concentration at least doubled from the new pick to the
worn pick.
As revealed from the comparison analysis between the dust concentration and pick tip radii, there
were consistent positive linear trends of dust concentration increasing with pick wear. The average
positive linear trend was 48.7 mg/m3 per millimeter of radius wear with an average R2 value of 0.908
using all nine test sets. This data suggests that as the pick tip radius symmetrically grows by one
millimeter of wear during concrete, limestone, or sandstone excavations, there will be an expected
additional 50 mg/m3 increase in dust at the face. Therefore, this anticipated increase could be utilized
when making informed decisions about controlling worker exposures in the future.
5.1.2 NMAM 7500 Percent Silica Analysis
The NIOSH Manual of Analytical Methods 7500 determined the percent silica in the suspended
respirable dust. Table 4.6 and Figure 4.2 reveal the percent silica determined in each dust sample collected
during tests for the three pick wears when cutting each rock type.
A two-way ANOVA statistical test analyzed if rock type and pick wear affect the percent silica
generated and if there is a meaningful interaction effect between rock type and pick wear. The p-value for
rock type is 6.7E-10, the p-value for pick wear is 0.86, and the p-value for their interaction is 0.91. The p-
value threshold for this analysis is also assumed to be 0.05, where values below 0.05 correspond to
sufficient evidence that the output variable is affected by the input variable. Therefore, statistical evidence
suggests that the rock type affects the percent silica generated in dust but that the pick wear does not
influence the percent silica generated in dust while cutting. Additionally, there is weak statistical evidence
that the rock type and the pick wear interact to affect the percent-silica generated within the dust.
As revealed in Table 4.6, Figure 4.2, and the two-way ANOVA test, the percent silica in
suspended respirable dust is inconsistent and appears unaffected by pick wear. On the other hand, the two-
way ANOVA test provides evidence that the percent silica generated is affected by the rock cut. This
connection is reasonable because each rock type is mineralogically different from one another and most
likely contains varying elemental compositions. In the end, the purpose of the NIOSH manual of
analytical methods 7500 analyses was to observe the presence of silica, which is valid for all rock types,
whether in the form of cristobalite, quartz, or tridymite. Dust mitigation is less of an issue in limestone
and is more significant in concrete, which can contain aggregates containing silicates, or sandstone, which
can have a notable amount of quartz and silicates.
76
|
Colorado School of Mines
|
5.2 Particle Size Distributions
5.2.1. Airborne Particles
Visually analyzing Figure 4.3, the particle size distributions generated were single modal for each
pick wear in every rock type, and all distributions reside close to one another without any major shifts or
trends. The instruments used to collect dust ranged from 0.2µm to 10µm because the cyclones eliminated
dust larger than 10µm and the filter sizes went down to 0.2µm pores. For this reason, it is hypothesized
that the peaks of particle sizes are around 2µm.
Kolmogorov-Smirnov (KS) statistical tests performed in MATLAB® between particle size
distributions tested the null hypothesis that the airborne particle sizes generated from each pick are the
same. A significance level of 0.05 is used again in the statistical analysis. Therefore, p-values below 0.05
provide strong statistical evidence to reject the null hypothesis.
The KS statistical test results are provided in Table 5.1 and suggest that the particle size
distributions are statistically different between different pick wears, with this trend confirmed for each
rock type cut. For example, all p-values determined between particle size distributions generated from the
three different pick wears are 1.5 x 10-10 or less, which provides strong evidence to reject the null
hypothesis.
Table 5.1 P-values obtained from KS statistical tests between the various pick wears for airborne particle
size distributions
Pick Comparison Concrete Limestone Sandstone
p-values p-values p-values
New to Moderate 1.5 x 10-10 1.5 x 10-28 1.0 x 10-11
New to Worn 3.8 x 10-23 2.6 x 10-112 1.8 x 10-12
Moderate to Worn 1.5 x 10-47 8.0 x 10-39 5.2 x 10-10
The statistical KS tests provide evidence that the pick influences the airborne particle size
distributions wears, most likely due to the peak value variance and high number of particles analyzed.
However, there is no significant shift in particle size distributions within the respirable range for these
particles less than 10µm in aerodynamic diameter. All distributions have an average particle size between
1.5µm and 2.3µm. Therefore, the airborne particle size distributions do not shift significantly enough to
77
|
Colorado School of Mines
|
warrant implementing dust control and bit management decisions based on particle size distribution
changes.
5.2.2. Fines
Visually analyzing Figure 4.4 reveals that the fine material particle size distributions slightly shift
according to pick wear for all rock types. The new picks generated slightly smaller particles than the worn
picks, which generated slightly larger particles, and the moderately worn picks, which generated particle
sizes between the two. In other words, the distributions generated from the new picks were all slightly
shifted to the left (indicating smaller particle sizes) compared to the distributions generated from the worn
picks, which were shifted to the right (indicating larger particle sizes) beyond the distributions generated
by the moderately worn pick.
The averages and standard deviations, as provided in Table 5.2, confirm the trend between the
particle sizes and the pick wears. For example, the average fines concrete particle aerodynamic diameter
from the new, moderately worn, and fully worn pick was 2.47, 2.74, and 2.80, respectively, revealing that
particle size increases as the pick wears.
Table 5.2 Average aerodynamics diameter values for the fines particle size distributions with standard
deviations
Concrete Limestone Sandstone
Data analyses New Mod Worn New Mod Worn New Mod Worn
Standard Deviation 1.91 2.07 2.20 1.07 1.26 1.34 2.07 2.59 2.59
Average aerodynamic 2.47 2.74 2.79 1.61 1.73 1.96 2.01 2.30 2.30
diameter (µg)
KS statistical tests in MATLAB® were also performed between the fines particle size distributions
generated by the three pick wear levels. Of the about 10,000 particles analyzed for each pick wear
condition in each rock, the sizes were input to the KS test to test the null hypothesis that the distributions
of particle sizes generated from each pick are the same.
The KS statistical test results in Table 5.3 suggest that the particle size distributions generated
from different pick wear levels are statistically different, assuming the 0.05 significance level. This trend
is confirmed for each rock type cut. For example, all p-values determined between particle size
distributions generated from the three different pick wears are 9.1 x 10-27 or less, which provides strong
evidence to reject the null hypothesis. Therefore, there is strong statistical evidence to suggest that the
fines particle size distributions are not the same. Visually observing the distributions in Figure 4.4, there
78
|
Colorado School of Mines
|
is a slight shift revealing that as the pick tip wears, the particle size will respond by slightly increasing. It
is theorized that the new pick generated smaller fines particles due to the smaller contact area (or
concentrated forces over a smaller area) during rock cutting with newer picks. The smaller point of
contact might lead to more rock crushing under the pick tip due to a higher force concentration in the
pressure bubble beneath the pick tip. Particles may become larger as the pick wears down because the
worn pick tips are rounder and distributes the load over a larger contact area during cutting [32], [44],
[77], [78].
Table 5.3 P-values obtained from KS statistical tests between the various pick wears for fines particle size
distributions
Pick Comparison Concrete Limestone Sandstone
p-values p-values p-values
New to Moderate 4.5 x 10-29 9.1 x 10-27 2.3 x 10-56
New to Worn 2.0 x 10-78 4.2 x 10-311 2.1 x 10-223
Moderate to Worn 1.1 x 10-23 5.1 x 10-167 8.6 x 10-100
Although particles shifted to larger sizes as the pick wear level increased in all cases, the shifts
are insignificant and do not warrant continued use of the worn-out bits. In other words, the increased
volume (concentration) of fines generated under the worn cutters outweighs the generation of smaller
particles under new bits since the size fraction of the smaller particles relative to the volume of fines
generated would increase the net amount of dust in the mine workings.
5.3 Particle Shapes
A two-way ANOVA statistical test in MATLAB® analyzed if rock type and pick wear affect dust
particle shapes. The p-value threshold for this analysis is 0.05, where values below 0.05 correspond to
sufficient evidence that the output variable of particle shape affects the input variable. The outputs of the
statistical tests are provided in Table 5.4.
Table 5.4 Results of the two-way ANOVA tests for comparing particle shapes to rock type and pick wear
Particle p-values
shape Rock type variance Pick wear variance
characteristic
Roundness 0.032 0.11
Aspect Ratio 0.025 0.43
Roughness 0.001 0.69
79
|
Colorado School of Mines
|
The results of the two-way ANOVA statistical tests between the datasets provide weak evidence
that the particle shapes differ between pick wear levels. In other words, with all the pick wear p-values
above the 0.05 threshold, as seen in Table 5.4, the statistical test fails to provide strong evidence that the
pick wear levels affect the roundness, aspect ratios, or roughness particle shapes. On the other hand, with
the rock type p-values all slightly below the 0.05 threshold, as seen in Table 5.4, the statistical test
provides evidence that the rock type causes a slight effect on the roundness, aspect ratio, and roughness
particle shapes.
Further analyses determined the average particle shapes. The particle shape values obtained for
the new, moderately worn, and worn pick were averaged together for each respective rock type and are
provided in Table 5.5. This average clearly explains the particle shapes generated throughout cutting
operations for each rock type with pick cutters. Therefore, the quantitative particle shape average data
reveals that the concrete, limestone, and sandstone particles generally have a slightly oval shape and
somewhat smooth edges throughout cutting operations.
Table 5.5 Average particle shape values for the three various rock types cut
Shape Characteristic Concrete Limestone Sandstone
Roundness 0.56 0.57 0.57
Aspect Ratio 1.55 1.52 1.56
Roughness 0.96 0.94 0.97
Ultimately, the two-way ANOVA test fails to provide evidence that the pick wear variable affects
the suspended dust particle shapes. In other words, as the pick wears, the shapes of the particles do not
appear to change drastically. However, some statistical evidence shows that changing the rock type
slightly changes the particle shapes in terms of roundness, aspect ratio, and roughness.
There are implications for the determined particle shapes in the analysis. The particles in the
study have aspect ratios near 1.5, and particles with aspect ratios as low as two suggest deeper penetration
into the lungs [51]. Therefore, the particles in the study are nearing an aspect ratio that suggests these
particles can penetrate deeper into the lungs. On the other hand, the particles have roughness values close
to one, pointing to particles with less-rough edges. Smoother edges are favorable because previous studies
find that rougher particles are more readily entrained in the deeper regions of the lungs [52]. Note that
80
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.