University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Virginia Tech | 1.7.3 Detailed Testing
Two series of detailed test programs were conducted using the pilot-scale test unit. The
first series of tests were performed to investigate the effects of the key design variables on
separator performance and to simultaneously define the overall grade and recovery curve. The
subsequent series of testing was used to investigate the effects of key operating parameters.
Tests were conducted primarily as a function of teeter bed pressure and fluidization water rate.
The coal/rock interface, or teeter bed, was adjusted to different levels (i.e. different bed pressure)
for each steady-state test. Fluidization water was adjusted to fine tune the separation. For each
test, samples were taken from the feed, overflow, and underflow streams after conditions were
stabilized. Five test runs were completed during the on-site test work.
1.7.4 Process Evaluation
Due to the low percent solids, the fine size distribution, and the low specific gravity of
the material, bed development in the CrossFlow separator was very difficult for this particular
application. Initial plans to feed the unit at 1 tph/ft2 could not be obtained due to the turbulence
occurring in the bed formation area. Feed rates were slowly reduced over time until a 0.10
tph/ft2 feed rate with a water addition rate of 1.5 gpm produced a stable bed in the unit. Even at
this low feed rate, an appreciable amount of plus 100 mesh material was still reporting to the
overflow. Five sets of samples were collected of the feed, overflow, and underflow streams.
However, laboratory analyses were not conducted on these samples because visual observations
of the product streams indicated poor performance at attempting to classify the feed stream.
The coal slurry evaluated in this series of experiments possessed a mean particle size of
0.075 mm. Table 1.14 is a summary of the array of operating parameters that were investigated
54 |
Virginia Tech | 1.8 Conclusions
1. A comprehensive study of the CrossFlow separator was conducted at four coal
preparation plants on the east coast. In-plant testing of a 9 x 16 inch unit resulted in
separation efficiencies at or above existing classification equipment in the size class of
0.2 to 1.0 mm.
2. The data demonstrate that for any given product ash content or sulfur content, the
CrossFlow separator can produce a higher clean coal yield and higher combustible
recoveries at higher feed rates when compared to the existing coal spirals. The
CrossFlow also demonstrated its ability to handle the entire flow of multiple spirals in a
single-stage circuit.
3. In the instance where the ultimate goal was to compare results against the existing clean
coal effluent cyclones (28 mesh by zero material at 100 mesh), it was determined that the
material was too fine to develop the necessary teeter-bed, and the project was therefore
abandoned.
4. The test work conducted in this series of tests supports the replacement of spirals with the
CrossFlow technology for several applications. As a result, several full scale installations
of the unit are being planned in the near future. Based on the successful installation of
these full scale units, further implementation of additional units can be utilized in a broad
spectrum of companies and industries.
56 |
Virginia Tech | Chapter 2
In-Plant Testing of HydroFloat Separator in Phosphate Industry
2.1 Introduction
2.1.1 General
Teeter bed technologies can only be applied for gravity concentration when the particles
in the feed stream have a relatively narrowly size distribution and moderately large difference in
component densities. These units inherently accumulate low density coarse particles at the top
of the teeter bed which are too light to penetrate the bed, but at the same time, too heavy to be
carried by the rising water into the overflow. As a result, misplacement of low-density, coarse
particles to the high-density underflow can occur. This inefficiency can be partially corrected by
increasing the elutriation water, to try to carry the low density coarse particles into the overflow.
However, this action often causes the fine, high-density particles to also report to the overflow,
thereby impacting the quality of the products. As a result, the widespread application of
traditional hydraulic separators is greatly limited by these physical constraints.
The limitations of traditional hydraulic separators were recently recognized and
overcome through the design of the HydroFloat separator. This technology effectively combines
the flexibility of a flotation process with the high capacity of a density separator to overcome
barriers that commonly limit conventional teeter bed separators. The HydroFloat can
theoretically be applied to any mineral classification system where differences in apparent
density can be created by the selective attachment of air bubbles. Figure 2.1 provides a
schematic drawing of the HydroFloat separator.
58 |
Virginia Tech | through the teeter bed and are eventually discharged through the control valve at the bottom of
the separator.
2.1.2 Advantages of a Hydraulic Separator
Compared to traditional froth flotation, the use of a fluidized bed within the HydroFloat
significantly improves the recovery of particles by (i) reducing turbulence, (ii) enhancing
buoyancy, (iii) increasing particle retention time, and (iv) improving bubble-particle contact. The
presence of the high-solids teeter bed reduces the turbulence commonly associated in traditional
flotation units and therefore enhances the buoyancy of the particles. The teetering effect of the
hindered-bed relinquishes the need for bubble-particle aggregates to have sufficient buoyancy to
rise to the top of the cell. The low density agglomerates can easily overflow into the product
launder, where as the hydrophilic particles move through the teeter bed and eventually discharge
through the control valve at the bottom of the separator.
Other benefits of the HydroFloat separator versus traditional froth flotation cells include
increases in particle retention time by producing a counter-current flow of particles settling in a
hindered state against an upward rising current of water, and the increased probability of bubble-
particle contacting in the teeter-bed due to the high-solids content. A higher production rate is
possible with the HydroFloat separator than in traditional froth flotation cells due to the high
percent solids in the compact teeter bed.
The HydroFloat separator is ideally suited to recover coarse particles that traditional froth
flotation cells cannot efficiently recover for several reasons. One reason for the improved
recovery of coarse particles is the upward flow of elutriation water in the HydroFloat separator
helps lift the larger particles into the product launder. The teeter bed also produces ideal
60 |
Virginia Tech | conditions for bubble-particle interactions by maintaining high solids content and quiescent flow
conditions. In addition, the high solids content within the teeter-bed separator makes it possible
to treat large tonnages in a very compact volume as compared to conventional flotation
separations which are conducted at very low solids contents using large volume cells.
2.1.3 Project Justification
One of the driving forces behind the HydroFloat separator is the phosphate industryβs
need to recover coarse particle phosphate (28 x 35 M size fraction) from the feed matrix. It is
estimated that 10% of feed material to a Florida phosphate plant is in the plus 35 mesh fraction,
which is virtually impossible to recover with present classification equipment. An improvement
in coarse particle recovery with the HydroFloat alone corresponds to an additional $7.5-15
million of revenues.
As in the coal industry, the energy benefits of the HydroFloat over conventional
equipment are related to the reduction in pumping requirements and water usage which is a
direct result of the higher feed ton rate. The lower operating and maintenance cost per ton of
product is significantly reduced with the HydroFloat versus conventional equipment. Overall,
the implementation of the HydroFloat separator will allow operations to become more profitable
and more competitive by utilizing reserves more effectively, reducing waste and increasing
productivity.
61 |
Virginia Tech | 2.2 Literature Review
2.2.1 General
The recovery of minerals by flotation is one of the most versatile mineral-processing
techniques used in industry today. Flotation methods are utilized throughout the mining industry
to treat sulfide ores such as copper, lead and zinc, oxide ores such as hematite and cassiterite and
non-metallic ores such as phosphate and coal (Wills, 1992). Since its inception in the early
1900βs, improvements in the flotation process have long been a goal within the industry and
numerous studies have been financed to overcome the inefficiencies inherent in the process.
Industry and government sponsored research programs have focused on all areas of the flotation
process to improve recoveries including advancements in chemical reagents, adaptations to
existing equipment and introduction of novel equipment.
2.2.2 History of Flotation
Although a subject of considerable debate, flotation was believed to be first utilized in the
mining industry by T.J. Goover, who in 1909 patented (British Patent No. 27-02-1909) the first
multi-cell impeller-type apparatus for froth flotation (Rubinstein, 1995). However, research into
the relationship between particle size and floatability did not begin until 1931, when Gaudin, et
al. (1931) showed that coarse and extremely fine particles are more difficult to recover as
compared to intermediate size particles. Twenty years after this original work, Morris (1952)
arrived at the same conclusion, that particle size is one of the most important factors in the
recovery of ores by flotation. Intermediate size particles will achieve the highest recovery,
where as very fine particles (d <20Β΅m) will have the lowest recovery. In addition, as the particle
p
diameter begins to increase, the recovery will start to decline. This reduction in recovery on the
62 |
Virginia Tech | fine and coarse size fractions is indicative of a reduction in the flotation rate of the particles
(Jameson, 1977). It can be seen that the efficiency of the froth flotation process deteriorates
rapidly when operating in the extremely fine or coarse particle size ranges, which is considered
between 10 Β΅m and 200 Β΅m. This is evidence that conventional flotation practices are optimal
for the recovery of particles between 65 to 100 mesh.
According to Soto and Barbery (1991), conventional flotation cells operate with two
contradictory goals. First, a conventional cell has to provide enough agitation to maintain
particles in suspension, shear and disperse air bubbles, and promote bubble-particle collision.
However, for optimal recovery, a quiescent system is required to reduce detachment and
minimize entrainment. As a result, coarse particle flotation is more difficult since increased
agitation is required to maintain particles in suspension. Furthermore, coarse particles are more
likely to detach under turbulent conditions. To compensate for the lack of recovery, some
installations are using relatively small flotation devices operated at low feed rates (Lawver,
1984).
As particle size is reduced, two dominating characteristics begin to emerge, i.e., the
specific surface becomes large and the mass of the particle becomes very small (Abdel-Khalek,
et al., 1990). These are the dominating factors affecting fine particle recovery in flotation
systems. Virtually all ores are associated with a clay mineral, which is ultimately transferred to
the preparation plant with the mineral of interest. The clay minerals associated with the fine
fractions will reduce mineral recovery by inhibiting bubble-particle attachment, and consuming
flotation reagents.
The variety of flotation machines available on the market today can be classified into two
distinct groups: pneumatic and mechanical machines (Wills, 1992). Pneumatic machines
63 |
Virginia Tech | commonly utilize air that is blown in or induced, where it must be dissipated through a series of
baffles or some form of permeable base within the cell. Since air is used not only to produce the
froth and create aeration but also to maintain the suspension and to circulate it, an excessive
amount is usually introduced (Wills, 1992). Complications directly related to the excessive
amount of air limited the use of pneumatic machines until the development of the flotation
column.
Mechanical flotation machines are the most common and widely used flotation machine
on the market today. The units are characterized by a mechanically driven impeller which
agitates the slurry and disperses the incoming air into small bubbles (Wills, 1992). Air addition
into the cell can either be forced through an external blower, or self-aerating. Typically most
mechanical flotation cells are set up in a series of βbanksβ, where several cells will allow free
flow from one cell to the next down the bank.
Performance is generally based on three factors including: (i) metallurgical performance,
i.e., product recovery and grade, (ii) capacity, and (iii) operating and maintenance costs (Wills,
1992). An analysis of the effectiveness of the various types of flotation machines was made by
Young (1982), who discusses performance with regard to the basic objectives of flotation, which
are the recovery of the hydrophobic species into the froth product, while still achieving a high
selectivity by retaining as much as the hydrophilic species as possible in the slurry. Recovery is
directly related to particle-bubble attachment and requires quiescent conditions, which is not
found in conventional mechanical flotation devices. The mechanical impellers found in typical
flotation cells are not ideal for particle-bubble contact, which has led the industry to utilize
column cells for a variety of mineral applications that, up until the past decade or two, was
unheard of.
64 |
Virginia Tech | Column cells are considered to be ideal displacement machines, where as mechanical
cells are ideal mixers (Wills, 1992). A column cells improves flotation performance by
minimizing turbulence within the cell and reducing entrainment using froth washing. In 1914,
G.M. Callow patented the first apparatus with air sparging through a porous false bottom,
(Rubinstein, 1995), which would become the basis for future column cell designs. By 1919, M.
Town and S. Flynn had developed the first design involving a countercurrent of slurry and air
within a column. While pneumatic Callow apparatuses were very popular in the early 1920βs
and 30βs, the lack of technological progress in the area of reliable pneumatic air spargers and
lack of process control systems forced the introduction of impeller-type apparatuses. It wasnβt
until the mid 1960βs that column cells began to be intensively developed and extensively
introduced into the industry, when practically all the work on updating other types of flotation
cells ceased (Rubinstein, 1995).
The advantages of column cell technology over conventional mechanical cells are
directly related to the direction of flow of the slurry and air. The counter-current regime
provides for more ideal bubble-particle attachment and enhanced aggregate stability. The
likelihood of bubble-particle detachment is minimized due to low turbulence of slurry flows
within the column. These benefits have prompted the phosphate industry to implement column
flotation cells into the industry for fine and coarse particle flotation.
2.2.3 Phosphate Flotation
Phosphate beneficiation plants are designed to process run-of-mine ore, typically called
the ore matrix, into a sellable product for use in either the fertilizer market or as an integral part
or the production of phosphoric acid. The ore matrix is upgraded by separating the phosphate
65 |
Virginia Tech | grains from other impurities such as clay and silica. Beneficiation plants in the southeastern
United States (Florida and North Carolina) generally use sizing and classification processes to
concentrate the phosphate rock and separate it from impurities.
Florida beneficiation plants typically wash and deslime the ore matrix at 150 mesh. The
material finer than 150 mesh is considered tailings and is pumped to settling ponds.
Approximately 30% of the phosphate contained in the original ore matrix is lost to the tailings
ponds. The remaining rock is separated into three size classes, a pebble size fraction, coarse and
fine size fractions. The pebble is a high phosphate content rock (-3 ΒΌ x 14 mesh) that requires
no further processing. The coarse size fraction (14 x 35 mesh) and fine size fraction (35 x 150
mesh) are treated separately in different flotation circuits.
Historically, fine phosphate flotation is an efficient process with recoveries from
conventional froth flotation in excess of 90% for most ores. Recoveries will vary depending on
the ore type, with recoveries dropping slightly for some high manganese or dolomitic ores. In
contrast, froth flotation recoveries for coarse phosphate are generally much lower than those of
fine phosphate ores. Typical recoveries for coarse flotation are less than 50%. Historically,
hammer mills were used for size reduction, but due to high maintenance costs and loss of fines,
this practice has been discontinued (Soto, 1992).
The industry, however, has taken other approaches to circumvent the problem of low
floatability of coarse particles. For instance, such approaches are exemplified by the use of
gravitational devices such as spirals, tables, launders, sluices and belt conveyors modified to
perform a "skin flotation" of the reagentized pulp. Although a variable degree of success is
obtained with these methods, they have to be normally supplemented by scavenger flotation. In
addition, some of them require excessive maintenance, have low capacity, or involve high
66 |
Virginia Tech | 2.3 In-Plant Testing at Phosphate Plant A
The Phase I field-testing of the HydroFloat separator involved equipment setup,
shakedown and detailed testing at the Phosphate Plant A. The goal of this effort was to compare
the unit to existing conventional cells in several different areas of the plant by analyzing the
anticipated product grade and recovery, insol content, reagent consumption, and feed capacity at,
and above, design feed rates of the unit. The three areas of the plant where the HydroFloat
separator was tested included the fine feed, amine flotation and coarse feed circuits.
The main objective of the fine and coarse phosphate testing was to demonstrate the
potential of the unit as a candidate for the process equipment in a proposed plant design with
both fine and coarse circuits. The main objective of the amine flotation testing was to
demonstrate the feasibility of using the unit for silica flotation and to develop data to determine
its potential application for use in the amine flotation circuit at Phosphate Plant A.
Approximately 6 months was allocated to this task.
Individuals from Eriez Magnetics and Virginia Tech participated in the testing at
Phosphate Plant A with cooperation from key personnel at the processing plant. Additional tests
were conducted by Phosphate Plant A representatives to expand the data base for evaluating the
potential of incorporating the HydroFloat separator into proposed circuit upgrades.
2.3.1 Equipment Setup
2.3.1.1 Fine Circuit
The installation of the pilot-scale unit in the fine feed circuit at Phosphate Plant A was the
main objective of this task. The separator was transported from the Eriez Magnetics Central
Research Lab in Erie, Pennsylvania to the processing plant. With cooperation from the operators
68 |
Virginia Tech | and mechanics at the plant, the 18-inch diameter pilot-scale HydroFloat separator was installed at
the fine circuit at Phosphate Plant A as shown in Figure 2.2. Reagentized feed was supplied to
the HydroFloat separator through a 2 inch line connected to the existing plant conditioning tanks.
Concentrate and tailings streams were discharged into floor sumps.
The unit was operated as a column flotation cell, utilizing the HydroFloat separator air
sparging system. The test unit included 3 compartments that allowed more water and air to be
added (up to 60 gpm water and 10 cfm air). There was no teeter-bed required in this system.
Plant compressed air and 115 volt electrical power were connected to the separator for the
automated control system. The separator was automatically controlled through the use of a
simple PID control loop which includes a pressure sensor mounted on the side of the separator to
measure the relative pressure (level), a single loop PID controller, and a pneumatic pinch valve
to control the underflow discharge to maintain a constant bed pressure (level).
69 |
Virginia Tech | added (up to 60 gpm water and 10 cfm air). There was no teeter-bed required in this system.
Plant compressed air and 115 volt electrical power were connected to the separator for the
automated control system. The separator was automatically controlled through the use of a
simple PID control loop which includes a pressure sensor mounted on the side of the separator to
measure the relative pressure (level), a single loop PID controller, and a pneumatic pinch valve
to control the underflow discharge to maintain a constant pressure (level).
2.3.1.3 Coarse Circuit
The same separator used in the fine and amine flotation circuits was also used in the
coarse circuit, with one modification. The center compartment was removed from the unit, so as
to allow the unit to operate with a typical teeter-bed (a total of 2 compartments). With
cooperation from the operators and mechanics at the plant, the 18-inch diameter pilot-scale
HydroFloat separator was installed in the coarse circuit at Phosphate Plant A. Reagentized feed
was supplied to the HydroFloat through a 2-inch line connected to existing plant conditioning
tanks. Concentrate and tailings streams were discharged into floor sumps.
Electrical power at 115 volt and plant compressed air were connected to the separator for
the automated control system. The separator was automatically controlled through the use of a
simple PID control loop which includes a single loop PID controller, a pressure sensor mounted
on the side of the separator to measure the relative pressure, and a pneumatic pinch valve to
control the underflow discharge to maintain a constant bed pressure.
71 |
Virginia Tech | 2.3.2 Shakedown Testing
Preliminary shakedown testing was conducted after completing the installation of the test
HydroFloat unit to resolve any unexpected operational problems that could arise. These tests are
normally necessary to resolve any problems that may have been overlooked in the initial
engineering and to confirm that feed capabilities, pipe sizes, electrical supplies, control systems,
etc., are adequate. An average of six shakedown tests per circuit was conducted with the unit.
2.3.3 Detailed Testing
Two series of detailed test programs were conducted using the pilot-scale test unit. The
first series of test were performed to investigate the effects of the key design variables on
separator performance and to simultaneously define the overall grade and recovery curve.
The HydroFloat separator is designed for feed rates of 2 tph/ft2 and 1 tph/ft2 rougher
concentrate, which allows the test unit to operate at 4 tph feed and 2 tph concentrate,
respectively. The initial testing in the fine and coarse circuit evaluated the unit at loading rates
much higher than design to establish the recovery fall-off. The design rates for the amine
flotation circuit were not precisely known going into the testing, but were thought to be similar
to those for rougher flotation. Part of the amine testing program was devoted to determining the
design rates and evaluating the HydroFloat separator performance across the board, both at the
design rate and above.
With the recovery fall-off determined for each circuit and unit configuration, the
subsequent series of testing was used to investigate the effects of key operating parameters.
Tests were conducted to establish reagent consumption (fatty acid, surfactant, amine and diesel
oil), to investigate the bed levels and sparger water required for the best unit operation and to
72 |
Virginia Tech | investigate the variability associated with the overall system. For each test, samples were taken
from the feed, concentrate and tailings streams after conditions were stabilized. The samples
were analyzed for BPL, MgO and insol contents.
2.3.4 Process Evaluation
All as-received results were analyzed and adjusted using mass balance software to ensure
the test data was reliable and self-consistent. Any experimental values that were deemed by the
mass balance routines to be unreliable were removed from the data set. The participating mining
company used the compiled data to establish the metallurgical improvement, operating savings
and economic payback that may be realized by implementing the proposed high-efficiency
technologies.
The process evaluation has been divided into three sections including (i) fine feed circuit,
(ii) amine flotation circuit, and (iii) the coarse feed circuit.
2.3.4.1 Fine Feed Circuit
Fifty-three tests were conducted during the fine circuit testing at Phosphate Plant A.
Testing in the fine circuit produced an average of 10% higher BPL recoveries with a 0.8% lower
BPL rougher tail in the HydroFloat separator than in the plant Wemco cells. Figure 2.3 displays
the HydroFloat separator and plant tails percent BPL for each test. The plant Wemco cells
averaged only about 0.7% BPL higher-grade rougher concentrates than the HydroFloat as shown
in Figure 2.4. An average HydroFloat separator rougher concentrate grade of 54.9% BPL is
satisfactory considering the test feed grade only average 8% BPL through most of the testing.
73 |
Virginia Tech | During testing, several attempts were made to obtain final grade concentrates (7% insol)
with one stage of flotation. The results show that insol concentrates between 9-10% produced
only 74-76% recoveries, and dropping the insol to 7-8% reduced the recoveries to 70% or less.
Further testing in this area needs to be conducted utilizing more selective reagents or higher feed
grades to achieve the desired 7% insol concentrates in a one step flotation process with the
HydroFloat separator.
One of the most important operating parameter to consider for fine flotation is the ability
of the process equipment to recover coarser material into an acceptable concentrate: i.e., recover
coarse phosphate without recovering fine silica. Comparison testing of the HydroFloat separator
with the Wemco Cell produced promising results. As shown in Figure 2.5, the HydroFloat
separator recovered 80%, 83%, and 88% of the plus 20 mesh, 20 x 28 mesh, and 28 x 35 mesh
phosphate, respectively. The performance values were well above those established for the plant;
the plant recovered only 24% of the plus 35 mesh and 67% of the plus 48 mesh phosphate.
Percent solids in the tailings averaged between 20-30% at optimum testing conditions.
During less than optimum conditions, the solids were as high as 53%. Optimum conditions
occurred at 70-75 bed levels, with between 50-60 gpm of sparger water, and 4 tph feed. While
higher bed levels and less sparger water could produce a slightly higher percent solids in the
tailings, this adversely affected the recovery and concentrate grades. Using the unit with 3
compartments and with bed levels of 70-75, the optimum froth depths were 15-20 inches.
Reagent dosages were affected by the poor water quality and excessive slimes in the feed
during the testing program. The fatty acid dosage in the plant ranged from 0.80 to 1.20 lb per ton
of fine feed during testing, whereas the fuel oil dosage in the plant ranged from 0.35 to 0.55 lb
75 |
Virginia Tech | optimum operating conditions. The optimum conditions for the HydroFloat separator for use in
fine flotation as defined by this testing program are: 3 compartment unit, with bed level between
70-75, a froth depth of 15-20 inches, sparger water between 55-60 gpm, air flow of 10 cfm, and a
surfactant dosage of at least 0.2 lb per ton of feed. The measured recovery values and
concentrate grade at these design rates were acceptable. Based on this data, the HydroFloat
separator can successfully be implemented into the Phosphate Plant A fine flotation circuit.
2.3.4.2 Amine Circuit
Twenty-four tests were conducted during the amine flotation circuit testing at Phosphate
Plant A. HydroFloat separator testing in the amine flotation circuit produced an average of 1.3%
higher insol concentrate and recovered about 8% less insol to the amine tailings than in the Plant
Wemco Cell. Figure 2.6 displays the concentrate grade for the HydroFloat separator and the
plant for each test. The plant Wemco cells averaged only about 0.5% higher BPL recovery than
the HydroFloat separator as shown in Figure 2.7.
The HydroFloat separator performed virtually the same as the plant Wemco cell for
amine flotation over the range 3 to 18% concentrate insol and 95 to 99% BPL concentrate
recovery. The unit demonstrated it could effectively recover coarse silica. The HydroFloat
separator insol recovery values were about 3% lower on average than those in the plant at above
design feed rates. The differences ranged from 6% to 11% in the 35 mesh and 48 mesh fractions
to 2% in the finer fractions. The HydroFloat separator insol recovery values were about 2%
higher on average than those in the plant at the lower feed rates.
77 |
Virginia Tech | 100
95
90
85
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Test No.
79
)%(
yrevoceR
LPB
Plant
HydroFloat
Figure 2.7. BPL Recovery Comparisons HydroFloat Separator vs. Existing Plant Cells.
One of the most important operating parameters to consider for amine flotation is the
ability of the process equipment to recover coarse silica without recovering phosphate.
Comparison testing of the HydroFloat separator with the Wemco Cell produced promising
results. As shown in Figure 2.8, the HydroFloat separator had just slightly less recoveries than
the plant for all of the size fractions except the 35 mesh, where it had a nearly 6% increase in
BPL recovery than the plant.
Reagent dosages were affected by the poor water quality and excessive slimes in the feed
during the testing program. The surfactant dosage for the HydroFloat separator ranged from 0.13
to 0.40 lb per ton of feed. The recommended dosage was 0.14 lb per ton at design rates. |
Virginia Tech | 100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00
35 48 65 100 150 -150
Size Class (Mesh)
80
)%(
yrevoceR
LPB
HydroFloat
Plant Cells
Figure 2.8. Comparison of Test Results for Amine Phosphate (Plant Circuit #2).
The interactions of varying diesel fuel dosage rates were studied during the amine circuit
testing. Amine flotation circuits use diesel oil or polymer occasionally to modify the froth when
slimy water is present. Froth stability was investigated, but was difficult to determine due to the
lack of air flow measurement available at the time of testing. Exact diesel fuel dosage rates are
unknown at this time.
While the operation of the HydroFloat separator for amine flotation was difficult to
optimize due to various outside variables affecting the system, a significant number of tests were
conducted at differing operating variables under varying operating conditions to achieve
optimum operating conditions. The optimum conditions for the HydroFloat separator for use in
amine flotation as defined by this testing program are: 3 compartment sections, with bed level
between 70-75, a froth depth of 15-20 inches, sparger water at 25 gpm, air flow of 10 cfm, and a |
Virginia Tech | surfactant dosage of at least 0.2 lb per ton of feed. Additional testing will be needed in the future
to validate these recommendations. The measured silica recovery values and concentrate grades
at these design rates were acceptable. Based on this data, the HydroFloat separator can
successfully be implemented into the Phosphate Plant A amine flotation circuit.
2.3.4.3 Coarse Circuit
Twenty-four tests were conducted during the coarse circuit testing at Phosphate Plant A.
Testing in the coarse circuit produced an average 12% higher BPL recovery with a 3.5% lower
BPL rougher tail in the HydroFloat separator than in the Plant Wemco Cell. Figure 2.9 displays
the HydroFloat separator and plant tails percent BPL for each test. Figure 2.10 displays the
concentrate recovery for the HydroFloat separator and Plant Wemco Cell. The plant average
about 6% BPL higher-grade rougher concentrates than the HydroFloat separator as shown in
Figure 2.11. However, the average concentrate grade of 62.6% BPL was still considered
satisfactory for the testing.
As with the fine and amine flotation circuits testing, poor water quality played an
important role in the overall performance of the reagents during testing. Fatty acid dosage in the
plant ranged from 2.04 to 3.61 lb per ton of coarse feed during testing, while fuel oil dosage
ranged from 1.06 to 1.68 lb per ton of feed. Both of these values are considered high for
Phosphate Plant A, and hindered recoveries as a result.
Surfactant dosage for the HydroFloat ranged from 0.23 to 0.77 lb per ton of feed, which
was also considered to be a high dosage, mostly attributable to the high fatty acid-fuel oil dosage
in the plant. Other contributing factors were the poor water quality and the need to set the
81 |
Virginia Tech | 2.4 In-Plant Testing at Phosphate Plant B
Equipment setup, shakedown testing, and detailed testing comprised the phase I field-
testing of the HydroFloat separator at Phosphate Plant B. The goal of this effort was to compare
the unit to existing hydroclassifiers and conventional cells by analyzing the anticipated product
grade and recovery, insol content, reagent consumption and feed capacity at, and above, design
feed rates of the unit. The main objective of testing was to determine if the HydroFloat separator
could achieve higher recoveries of the ultra-coarse particles than the existing second-stage
hydroclassifer at the plant. Further investigations of the coarse and fine matrices were
conducted, comparing results against the existing conventional cells currently in operation at the
plant. Approximately 12 months was allocated to this task. Individuals from Eriez Magnetics
and Virginia Tech participated in the testing at Phosphate Plant B with cooperation from key
personnel at the processing plant.
2.4.1 Equipment Setup
The separator was transported from the Eriez Magnetics Central Research Lab in Erie,
PA to the processing plant. With cooperation from the operators and mechanics at the plant, the
1-foot diameter pilot-scale HydroFloat separator was installed at each circuit (ultra-coarse,
coarse and fine) for a period of several weeks for each circuit at Phosphate Plant B as shown in
Figure 2.13. Reagentized feed was supplied to the HydroFloat separator through a 2-inch line
connected to the existing plant conditioning tanks. Concentrate and tailings streams were
discharged into floor sumps.
Plant compressed air and 115 volt electrical power were connected to the separator for
the automated control system. The separator was automatically controlled through the use of a
86 |
Virginia Tech | 2.4.2 Shakedown Testing
After completing the installation of the test HydroFloat unit in each circuit, preliminary
shakedown testing was conducted to resolve any unexpected operational problems that could
arise. Shakedown test are commonly utilized to resolve any problems that may have been
overlooked in the initial engineering and to confirm that feed capabilities, pipe sizes, electrical
supplies, control systems, etc., are adequate.
2.4.3 Detailed Testing
Two series of detailed test programs were conducted for each circuit using the pilot-scale
test unit. The first series of test were performed to investigate the effects of the key design
variables on separator performance and to simultaneously define the overall grade and recovery
curve.
The HydroFloat separator is designed for feed rates of 2 tph/sqft and 1 tph/sqft rougher
concentrate, which allows the test unit to operate at 4 tph feed and 2 tph concentrate,
respectively. The initial testing in the coarse circuit evaluated the unit at loading rates much
higher than design, to establish the recovery fall-off.
With the recovery fall-off determined for each circuit and unit configuration, the
subsequent series of testing was used to investigate the effects of key operating parameters.
Tests were conducted to establish reagent consumption (fatty acid, surfactant, and diesel oil), to
investigate the bed levels and sparger water required for the best HydroFloat separator operation,
and to investigate the variability associated with the overall system. For each test, samples were
taken from the feed, concentrate and tailings streams after conditions were stabilized. The
samples were analyzed for BPL, MgO, and insol contents.
88 |
Virginia Tech | 2.4.4 Process Evaluation
To ensure the test data was reliable and self-consistent, all as-received results were
analyzed and adjusted using mass balance software. Experimental values that were deemed by
the mass balance routines to be unreliable were removed from the data set. The participating
mining company used the compiled data to establish the metallurgical improvement, operating
savings and economic payback that may be realized by implementing the proposed high-
efficiency technologies.
The process evaluation has been divided into three sections including the (i) ultra-coarse
rock feed, (ii) the coarse rock feed, and (iii) the fine feed circuits.
2.4.4.1 Ultra Coarse Feed
Grade versus recovery data for the in-plant evaluation of the HydroFloat had BPL
recoveries of 87% to 99% with product grades ranging between 5% and 14% insols. The
resulting products contained, on average, 67% BPL. Figure 2.14 is a graph of the grade versus
recovery data for the in-plant testing and earlier laboratory-scale testing. Size-by-size analysis of
the HydroFloat was conducted and results are presented in Figure 2.15. The HydroFloat is
capable of high BPL recoveries for even the coarsest size fractions, where 96.7% of the available
BPL in the +16 mesh size class was recovered.
89 |
Virginia Tech | 2.4.4.2 Coarse Feed
Figure 2.16 summarizes the grade and recovery data for the coarse feed test work. BPL
recoveries ranged from 90% to 98% while product grades averaged 24.7% insols. The resulting
products contained, on average, 55% BPL by weight. Figure 2.16 also illustrates that the results
for the laboratory evaluations were superior to those produced for the in-plant trials. This
occurrence is a direct result of the mean particle size difference found between the samples used
for the laboratory and in-plant testing. It was calculated that the sample used for the coarse
matrix laboratory testing was as coarse (mean size: 0.706 mm) as the sample provided for the
ultra-coarse testing (mean size: 0.721 mm). During the in-plant trials, it was observed that the
coarse matrix was significantly finer, amplifying any occurrence of hydraulic carry-over or
activation of fine floatable insols.
100
95
90
85
80
75
70
0 10 20 30 40 50 60
Product Insols (%)
91
)%(
yrevoceR
LPB
Plant Testing
Laboratory Testing
Figure 2.16. BPL Recovery vs. Product Insol Grade for Coarse Matrix. |
Virginia Tech | 2.4.4.3 Fine Feed
The results from the in-plant testing on the fine matrix are shown in Figure 2.17. BPL
recovery ranged from 88% to 97% using the HydroFloat. When operated as an open column,
BPL recoveries ranged from 85% to 92%, though at a significantly lower product insol (37% vs.
22%, respectively). Results from samples collected around the existing plant rougher-scavenger
swing circuit are also presented in Figure 2.17 for comparison. The findings indicate that the
open column cell (w/ HydroFloat sparging system) is able to achieve incrementally higher BPL
recoveries at lower product insol grades compared to either the HydroFloat or the existing
column technology. The corresponding product grade (%BPL) averaged 55% for the open
column system as seen in Figure 2.18. As with the ultra-coarse and coarse circuits, the
HydroFloat achieved an acceptable product grade and recovery in the fine circuit.
100
95
90
85
80
75
70
0 10 20 30 40 50 60
Product Insols (%)
92
)%(
yrevoceR
LPB
HydroFloat Plant Testing
HydroFloat Laboratory Testing
Existing Swing Cells
Open Column w/HF Sparging System
Figure 2.17. BPL Recovery vs. Product Insol Grade for Fine Matrix. |
Virginia Tech | 2.5 Conclusions
1. The in-plant evaluation of the HydroFloat separator demonstrated that this novel
separation device can successfully treat the three different size fraction in a typical
phosphate processing plant. For the ultra-coarse rock, the separator produced a high
grade phosphate product (+66% BPL) at BPL recoveries exceeding 95%. For the coarse
sized feed fraction, the separator produced a 99% BPL recovery at an 8% insol grade.
Significant improvements were also achieved in the fine feed fractions where a BPL
recovery greater than 90% was achieved with product insoles ranging between 22-25%.
2. Several advantages can be realized through implementation of the HydroFloat system.
The system can provide a higher product mass recovery, superior metallurgical results,
lower reagent costs and lower power requirements, with the greatest advantage being the
higher separation efficiency. A higher product mass recovery with a better product
quality is a significant achievement for this application. The HydroFloat has a
substantially lower operating cost due to reduced reagent consumption and power
requirements compared to conventional equipment.
3. One of the goals of this project is to successfully prove the technology in a sufficient
period of time to minimize the financial risk that will be taken by industry. The previous
years test work has eliminated the uncertainties associated with the HydroFloat separator
by proving plant scale units do in fact work. This can be seen by the fact that industry
leaders have submitted purchase requests for full scale units in their preparation plants.
95 |
Virginia Tech | Chapter 3
Beneficiation of Ultra-Fine Phosphate Streams
3.1 Introduction
Phosphate beneficiation plants are engineered to recover phosphate occurring in the 0.1
mm size fraction. In many cases, phosphate ore coarser than 1mm (i.e., ultra-coarse) is
stockpiled and either ground and routed through the beneficiation plant or blended directly with
the final product. Unfortunately, the phosphate that occurs in the finer fractions (i.e., finer than
0.1 mm) is generally rejected as waste in the clay slimes. Approximately 100,000 tons of these
phosphatic clay slimes in dilute (3-5%) slurry are pumped to tailings impoundments in Florida
each day. In addition to the expense of disposing of the waste clay, mining companies are
throwing away millions of dollars in phosphate to the refuse stream. It is estimated that 27% of
the current annual production of phosphate is lost to this refuse stream. Economic recovery of
phosphate from the phosphatic clays could extend Floridaβs phosphate resource life by decades.
Slimes are a natural component of phosphate ore. The slimes (phosphatic clays) consist
of both phosphate and clay particles. According to Zhang (2001), there are three major
characteristics of refuse slimes cause extreme difficulty in recovering the phosphate
economically: the ultra-fine particle size (35-50% below 1 micron), the even distribution of
phosphate particles among the various size fractions and the high clay content (30-50%). As
particle size is reduced, two dominating characteristics begin to emerge: the specific surface
becomes large and the mass of the particle becomes very small (Abdel-Khalek, et al., 1990).
The ultra-fine particles in a slimes stream increase the chance of the fine particles getting carried
into the froth as they are either entrained in the liquid or mechanically entrapped with the
98 |
Virginia Tech | particles being floated. Material from a slimes stream will have excess amounts of gangue
minerals, which will ultimately reduce the grade of the concentrate if they are carried into the
froth. Excessive slimes can significantly reduce flotation performance and simultaneously
increase reagent consumption. Because of their high dolomite content, acidulation of such
concentrates consumes excessive sulfuric acid and causes problems during phosphoric acid
manufacture such as increased acid viscosity, precipitation of insoluble Mg-phosphates and
difficulties in filtration and clarification of the final product.
As a result, the phosphatic clays are continuously rejected from the beneficiation plant
using both vibrating screens and hydrocyclones. The slimes, including the fine phosphate, are
ultimately placed in large tailings ponds, which act as large holding cells that are used for both
water clarification and refuse storage. Misplaced coarse phosphate values are also present in the
refuse stream. The coarse values are predominately a result of inefficient sizing or desliming.
For instance, if an excess amount of material is pumped to a hydrocyclone, the cyclone will
βropeβ causing vast misplacement of coarse material. βRopingβ is a common occurrence in the
phosphate industry due to the natural settling characteristics of the coarse ore.
A typical Florida phosphate plant discards 100,000 gpm to the tailings ponds, of which a
significant portion of is fine phosphate (passing 0.1 mm). Combined with the misplaced coarse
particles, the discarded phosphate represents a significant reduction in overall plant efficiency
and a resultant increase in the required volume of the refuse tailings ponds. Recovering this
material would increase overall plant production, while ultimately reducing operating costs.
While previous research has been conducted in this area, there is still not an economical recovery
system and as a result, this phosphate is currently considered unrecoverable by industry.
99 |
Virginia Tech | An urgent need exists to develop a low-cost recovery circuit to recover a significant
amount of the phosphate values that are currently being thrown away. One circuit configuration
that may meet this requirement is shown in Figure 3.1. The circuit, which incorporates simple
low-cost unit operations such as hydraulic classifiers, hydrocyclones, and flotation cells, can be
tailored to exploit particle characteristics particular to the size range in question.
The proposed circuit uses a large free-settling tank to collect the entire flow of phosphatic
clays that are currently rejected from the beneficiation plant. Tanks similar to these are already
extensively used by industry. The tanks allow the coarse fraction (plus 400 mesh) of the
phosphatic clays to settle out in a free-settling regime. The settling tank is equipped with an
overflow launder to collect the finest fraction, which is directed to plant tailings. The plus 400
mesh fraction is pumped to a bank of hydrocyclones. The hydrocyclones reject as overflow the
remaining slimes that inherently remain with the coarse fraction.
At this point in the circuit, the cyclone underflow will be significantly deslimed and will
have relatively high percent solids. However, as flotation feed, the cyclone underflow will still
contain slimes (due to bypass) that will prove detrimental to flotation performance and reagent
consumption. To overcome this problem, the cyclone underflow is further treated using a
CrossFlow hydraulic separator. The teeter bed in the CrossFlow separator removes the majority
of slimes via countercurrent washing. The underflow from the CrossFlow, which is optimal
material for flotation feed, will be discharged through the bottom of the unit at a high percent
solids (i.e., up to 75%).
The final step of the process involves column flotation to recovery the phosphate ore in
the presence of the clay slimes. Unlike coarse phosphate ore, it is expected that the fine nature of
100 |
Virginia Tech | 3.2 Literature Review
3.2.1 General
Florida has enormous reserves and will continue to produce a third of the worldβs
phosphate supply simply by making changes in the present system of mining, beneficiation,
waste disposal and land reclamation (Lawver et al, 1984). Beneficiation plants continue to
optimize separation processes to improve recoveries and minimize gangue minerals in the final
product. From improved separation technologies to better flotation reagents, all aspects of the
beneficiation process are being analyzed to determine where improvements can be made. As
waste disposal becomes a more and more urgent issue, new techniques are being proposed to
increase recovery of P O in the clay slimes stream. Current beneficiation plants dispose of one-
2 5
third of the current annual phosphate production through the clay slimes stream.
The source of the problematic recovery of phosphate from the slimes stream is the
existence of almost 50% clay minerals in the stream. These slimes are removed, not only
because they are refractory toward presently known upgrading processes, but they also interfere
with the operation of the flotation on the remainder of the material (Hazen, 1969). As a result,
researchers have been trying for decades to develop efficient and economical processes to
recover the phosphate from the clay slimes stream.
Beneficiation has become increasingly more difficult as reserves continue to have more
dolomite content, lower ore grades and increased clay content. As production continues into the
Hawthorn Formation in Florida, the silicate contamination is compounded with additional
magnesium contamination. Beneficiation plants can no longer blend lower MgO ore with higher
dolomitic ores (containing the MgO) to alleviate this problem. Both calcite and dolomite
consume sulfuric acid and can cause excessive foaming during rock processing. Dolomite is
103 |
Virginia Tech | particularly detrimental to the process because magnesium passes into and degrades both liquid
and solid fertilizer products (McCullough et al., 1977).
Historically, the separation of dolomite and calcite from phosphate has primarily been
conducted through various flotation processes. Three reagents are historically used for rougher
flotation of phosphate from most of the siliceous gangue: a fatty acid type collector, an oil-type
extender, and a pH regulator (Lawver et al., 1984). Fatty acid collectors consist of crude tall oil,
reconstituted tall oil, blends of tall oil with vegetable fatty acids, or tall oil soap skimmings. Oil-
type extenders are commonly fuel oils, reclaimed motor oils, or mixtures of these two. The use
of fuel oil extenders with fatty acids as an auxiliary collector in flotation is well known and
practiced in the industry. Fuel oil does not adsorb onto phosphate and quartz in the absence of
fatty acids, and as a result, flotation cannot be achieved with fuel oil alone. Fuel oil will adsorb
synergistically with fatty acids on apatite. Ammonia and soda ash are the two most common pH
modifiers used in the industry. Sodium silicate is one of the most widely used depressants for
floating non-sulphide mineral. Sodium silicate is primarily a strong and selective depressant
(Shin and Choi, 1985). The use of selective flocculation has also been studied on the phosphatic
clays. While studies have been conducted with all of these reagents, today there is still no
industry-wide standard of separating phosphate from dolomite, calcite and other clay
constituents.
3.2.2 Advances in Flotation Reagents
3.2.2.1 Use of Sodium Silica
Recovery of phosphate from fine clay streams has historically been very problematic due
to the high concentration of gangue minerals. Sun and Smit (1963) were able to recover fine
104 |
Virginia Tech | phosphate from Florida washer slime material using various fatty acid collectors and tall oil.
Various conditions were tested including effects of pH, fuel oil, pulp density and addition of
sodium metasilicate. Flotation results show that phosphate flotation decreases in the order of
linolenic, linoleic, oleic and stearic acid, which have 3, 2, 1, and 0 double bonds respectively
(Sun and Smit, 1963). Sun and Smit (1963) concluded that unsaturated acids are preferred for
most flotation purposes. Duplicate test runs achieved recoveries of 29.8 and 31.0% P O with
2 5
13.0 and 12.7% insolubles, respectively with test conditions including: 8.5 pH, 2.24 lb/ton fatty
acid, 6.72 lb/ton fuel oil, 0.10 lb/ton pine oil, and 12 minutes of conditioning at 25% solids (Sun
and Smit, 1963). The fatty acid used had between 1.3 and 1.5% rosin acids, with the highest
overall fatty acid content of the materials tested (96.5-96.9%) to achieve the best recoveries.
Mishra (1982) was one of the first who studied the electrokinetic properties of sodium
metasilicate, to be used as the modifying agent for the separation of apatite from calcite by
depressing calcite when using sodium oleate as the collector. Mishra determined that the degree
of polymerization of sodium silicate is influence by its SiO to Na O ratio, where polymerization
2 2
increases with the increase in SiO to Na O ratio. The optimum rate of polymerization was
2 2
found to be at pH 8.6 (Mishra, 1982). While sodium metasilicate and sodium oleate additions
were kept constant at 5 X 10-3 M and 5 X 10-4 M, respectively, flotation test work was carried out
at various pH. The addition of sodium metasilicate produced a depression zone between pH 6.0
and 9.0, but flotation recovery increased to 100% between pH 9.0 and 10.0. Recoveries
decreased to zero by pH 11.7. Mishra related the differences in recovery to changes in zeta
potential, and predicted two causes for its occurrence: (1) the adsorption of cationic species at the
apatite surface or (2) the adsorption of colloidal molecular species of sodium silicate. Flotation
results suggested that the selective action of sodium metasilicate on calcite, with sodium oleate
105 |
Virginia Tech | as a collector, apatite and calcite could be separated in the alkaline pH environment at about pH
10.0, depressing calcite with sodium metasilicate (Mishra, 1982).
Anazia and Hanna (1987) also found that the addition of sodium silicate improved the
P O recovery in the phosphate concentrate and increased the rejection of siliceous gangue in the
2 5
tailings in Florida phosphate operations. Their research concentrated first on the removal of
dolomite from the feed, followed by silica removal. Testing procedures included 250-g batches
of 48 x 400 mesh flotation feed in a D-1 Denver laboratory flotation machine with the impeller
speed set at 100 rpm (Anazia and Hanna, 1987). Oleic acid was used as the collector and pine
oil as the frother.
Dolomite flotation was conducted by adjusting the pH to between 4.5 and 5.0 with
various dilute acids. The fatty acid collector (at 3.3 lb/ton collector addition) and frother were
injected into the pulp and air was immediately introduced into the pulp to float the carbonate
gangue minerals. After separation of dolomite, as froth, the cell product was conditioned for
three minutes with 1.1 lb/ton sodium silicate at a pH of 6.0 to 7.0. This was followed by another
three minute conditioning with 1.1 lb/ton fatty acid collector and flotation of the phosphate
minerals. The silica was depressed by the sodium silicate and the phosphate was floated to give
concentrates of 29% P O and 0.8% MgO, with a recovery of 76% (Anazia and Hanna, 1987).
2 5
This process is also referred to as the Mineral Resource Institute (MRI) process. Since there was
no phosphate depressant used, the pH had to be strictly controlled, which resulted in large acid
consumption. As a result the project was considered unsuccessful.
Anazia and Hanna concluded that carbonate minerals (dolomite and calcite) readily react
with inorganic acids, resulting in the preferential dissolution or removal of mixed surface
contaminants on the carbonate particles and exposure of fresh clean surface sites suitable for
106 |
Virginia Tech | fatty acid adsorption. Carbon dioxide microbubbles are then generated by the carbonate
minerals, allowing for enhanced oleic acid adsorption at the solid-liquid-gas interface, and
βinstantβ flotation of the mineral (Anazia and Hanna, 1987). Oleic acid will still have an affinity
for the apatite particles, but it will not be as great as for that of the microbubble-encapsulated
dolomite particles.
During the same time period, the Bureau of Mines conducted flotation tests on fine
phosphate recovered from Florida operation slimes utilizing sodium silicate. The clay was first
sized using hydrocyclones and hydroseparators, recovering up to 96% of the plus 400 mesh
material, which was then used as a basis for the flotation test work (Zhang, 2001). The study
used a fatty acid flotation at pH 9.0 with sodium silicate to depress the quartz in the rougher and
cleaner flotation stages. At pH 9.0 the phosphate recovery was above 80% and the rougher
concentrate grade was above 26% P O (Jordan et al, 1982). The rougher concentrate grade fell
2 5
quickly as the pH rose above 9.0. At a 3 lb/ton fatty acid and 2 lb/ton sodium silicate addition
rates, a concentrate grade of 30% P O and 90% recovery were obtained. Half the sodium
2 5
silicate was added before the fatty acid collector; the rest was added prior to the cleaner flotation
stage (Jordan, et al, 1982). While P O concentrates were high, the overall P O recovery from
2 5 2 5
the total clay stream was low.
Shin and Choi (1985) continued the investigation of sodium silicate to better understand
its selective flotation behavior among calcium minerals. Materials tested were wet-ground and
sized to obtain a minus 270 mesh to plus 400 mesh fraction. The maximum amount of
adsorption of sodium silicate was found to occur at pH 9.8. The amount of adsorption of sodium
silicate increased with the increase in temperature of sodium silicate solutions (Shin and Choi,
1985).
107 |
Virginia Tech | Several years later, another in-depth study of the role of sodium silicate was conducted
by Dho and Iwasaki (1990) on Florida phosphate ore. Their research led them to conclude
sodium silicate can enhance flotation efficiency through: (1) the removal of impurity minerals
and calcium-bearing precipitates from quartz surfaces by dispersion, (2) drier and more
persistent froths stabilized by oily droplets containing calcium silicate precipitates, and (3)
higher specific flotation rates of phosphate relative to quartz, leading to faster flotation rates and
increased selectivity of separation (Dho and Iwasaki, 1990). Their research involved frothability
tests and continuous and batch flotation tests to compare flotation results with and without
sodium silicate.
Testing was conducted with a (1:1) fatty acid-fuel oil mixture, a 3.22 SiO /Na O sodium
2 2
silicate ratio and ammonia as the pH modifier. The material was conditioned for 90 seconds at
pH between 9.2 and 9.4 and diluted to 25% solids prior to flotation in a 2-liter Denver laboratory
flotation cell. Several observations were noted during the flotation testing, including:
β’ The tests showed more stable froths in the presence of sodium silicates than in the
absence of sodium silicates.
β’ The use of sodium silicate improved both the specific flotation rates and the coefficients
of mineralization of phosphate, thereby leading to increased relative floatability of the
phosphate.
β’ Increasing the amount of sodium silicate prevented the entrainment of quartz particles in
the froth, making it more stable (Dho and Iwasaki, 1990).
108 |
Virginia Tech | Factors adversely affecting the efficiency of anionic flotation of phosphate in the absence
of sodium silicate can decrease the overall concentrate grade (Dho and Iwasaki, 1990). CaCO is
3
a common precipitate in plant water, caused by the presence of Ca++ ions, which adversely
affects flotation. When Ca++ is mixed with sodium silicate, calcium and silicate ions interact,
resulting in the formation of calcium silicate precipitates. The main dispersive action of sodium
silicate on quartz is produced by electrostatic repulsion due to calcium silicate precipitates
formed on the surface of quartz (Dho and Iwasaki, 1990). The calcium silicate eventually
detaches as the zeta potential of calcium silicate and quartz decreases, leaving the quartz surfaces
virtually free of precipitates and thus is depressed (Dho and Iwasaki, 1990).
The aforementioned test work was conducted using the conventional froth flotation
technique. Several researchers have successfully floated fine phosphate from clay slimes in a
column flotation cell. Fine phosphate particles (<45Β΅m) from an Egyptian phosphate mine were
recovered utilizing a laboratory flotation column with a diameter of 5.04 cm by 361 cm high.
Testing included the use of oleic acid as the phosphate collector and sodium silicate as the silica
depressant. Test results included P O concentrate of 25.3% with a 51.52% recovery (Abdel-
2 5
Khalek et al., 2000). Similar results could not be attained with conventional flotation techniques.
3.2.2.2 Dolomite Recovery
More recently, the Florida Institute of Phosphate Research (FIPR) conducted a study to
optimize the Chinese dolomite collector, βPA-31β for use in the United States phosphate pebble
industry. The Chinese Lianyungang Design and Research Institute (CDRI) had previously
demonstrated the ability to reduce dolomite content in mainland China phosphate ores and were
willing to assist the United States phosphate industry with their findings. The dolomite
109 |
Virginia Tech | collectors USPA-31 and FAS-40A, developed by two Florida local reagent producers, produced
concentrates of more than 30% P O and less than 1.0% MgO, with overall P O recoveries
2 5 2 5
averaging approximately 79% (FIPR 02-150-197, 2003).
In 2003, another FIPR project (No. 00-02-145, 2003) concluded it was possible to
separate dolomite from apatite by coating it with a surfactant and immersing it in a dilute acid
solution where the dolomite generates carbon dioxide gas that is trapped by the surfactant and
floated to the surface (FIPR 00-02-145, 2003). The preliminary study examined the effects of
over a dozen different surfactants and dosages on their amenability to dolomite flotation. El-
Shall and Stana found polyvinyl alcohol (PVA) to be the most promising surfactant for coating
the mixture of phosphate rock and dolomite to effectively separate the two minerals. The ore is
first immersed in a 3% PVA solution and mixed well prior to being added to acidic water (3%
sulfuric acid). The sink product of dolomite flotation was further upgraded by either silica or
phosphate flotation, achieving concentrates near 64% BPL and 1-2% MgO. While this research
looks promising, additional testing will be needed on finer material to determine its feasibility in
the industry.
3.2.3 Other Recovery Mechanisms
One of the first tests conducted on phosphate slimes material that didnβt incorporate
solely flotation testing was in 1992 at PCS Phosphate in North Carolina (TexasGulf). The North
Carolina Minerals Research Laboratory (MRL) conducted a preliminary study on the
effectiveness of using hydraulic classifiers (Linatex Hydrosizer) to improve the beneficiation
process of North Carolina phosphate ores. The main objective of the research was to efficiently
deslime feed ore (14 mesh x 0) at a cut size of 200 mesh. Less than 2% of the P O value was
2 5
110 |
Virginia Tech | lost to the overflow, while over 96% of the minus 200 mesh was removed at a throughput of 0.27
stph of feed per ft2 of overflow area (Schlesinger, L.; Hutwelker, J, 1992). Prior to this study, the
Linatex Hydrosizer had only been fully tested on 28 x 100 mesh material. The initial test work
was deemed a success and additional testing was proposed but never carried out.
While flotation is the most typical separation process, there have been several other
mechanisms proposed to facilitate the separation of phosphate and gangue minerals. One of the
earliest accounts of this is a patent issued by Hazen Research in 1969 (Patent No. 3,425,799) to
leach the slimes with sulfuric acid under conditions that allowed crystals of calcium sulfate to
form and function as a filter aid. The leach liquor was then treated with an amine solvent and
processed through a solvent extraction step using ammonia. The phosphate values were
recovered as diammonium phosphate.
The Florida Institute of Phosphate Research sponsored a project that utilized an autoclave
acidulation technique to recover phosphate from the clay stream (Zhang, 2001). At high
temperature and pressure, P O was quickly recovered and minimal clay residue (35-45%)
2 5
remained as a by-product. The project was never brought to full scale due to the high capital
investment cost and inconsistent acid availability.
111 |
Virginia Tech | 3.3 Testing
3.3.1 Equipment Setup and Sample Acquisition
Sample acquisition took place during January 2004 at PCS Phosphate in White Springs,
Florida. A 2-inch line from the clay launder located fifty feet above the washer floor fed a 6-
inch diameter Krebs hydrocyclone and sump as shown in Figure 3.2. The hydrocyclone was
initially set up with a 1.25-inch apex and a 2.5-inch vortex finder. The cyclone/sump
configuration was a semi-closed system as a majority of the cyclone overflow was circulated
through the system while samples were collected.
Eight 5-gallon buckets of fine clay refuse stream were collected over the four day period.
Three 55-gallon drums of cyclone underflow were collected and dewatered during the 4-day
testing period for the subsequent conditioning and flotation test work. Additional samples of the
cyclone underflow were collected, to be used as a composite of the larger bulk sample and for
subsequent βsettlingβ tests. These tests were used to determine the size of the required tanks for
a full-scale operation. The material was screened at 400 mesh and analyzed for BPL and insol
content.
After sampling was completed, all samples were shipped to Eriez Magnetics Central
Research Lab (CRL) in Erie, Pennsylvania. Here, the samples were thoroughly characterized to
determine the size distribution of the phosphate and gangue minerals. The concentrated
underflow material was run through a 2 x 6 inch CrossFlow separator to wash the slurry of the
minus 400 mesh clay material as shown in Figure 3.3. The plus 400 mesh material obtained
from the CrossFlow unit was dried, riffle-split and divided into 250 gram charges to be used as
flotation feed.
112 |
Virginia Tech | 3.3.2 Plant Hydrocyclone Testing
A 6-inch diameter Krebs Hydrocyclone was utilized during the on-site testing and sample
acquisition for future evaluations of a full-scale unit. The hydrocyclone was placed in series
with the large mixing sump, and the circuit was set up to run continuously in a closed-loop
configuration. The cyclone was situated such that the overflow and underflow could return to
the feed sump by gravity.
The cyclone was optimized to give the optimum size separation by varying feed pressure,
vortex finder and apex parameters. The test rig was run at pressures between 20 β 28 psig with
three different apexes and two different vortex finders. An analysis of these tests parameters
would determine the best suited configuration to maximize the recovery of plus 400 mesh
phosphate values to the underflow while simultaneously rejecting the maximum amount of clay
slimes to the overflow. Five tests were conducted using the parameters as defined in Table 3.1.
Table 3.1: HydroCyclone Testing Parameters at Florida Phosphate Plant.
Test Operating Apex Vortex Feed U/F O/F
No. Pressure Size Finder Flow
(psi) (inch) (inch) (gpm) (gpm) (gpm)
1 27 1.00 2.5 190.7 16.70 174.00
2 20 1.00 2.5 na 27.30 na
3 26 0.75 2.5 na 7.70 na
4 20 0.75 2.0 160.30 10.30 150.00
5 28 0.75 2.0 250.00 10.30 239.70
Underflow and overflow samples were taken for each set of test parameters and assayed
for a complete characterization and material balance at plus 150 mesh, 150x270, 270x325,
325x400 and minus 400 mesh. Table 3.2 summarizes the size distribution for each test
parameter.
114 |
Virginia Tech | Table 3.2: Size Distribution for Hydrocyclone Tests 1-5.
Cyclone Underflow Percent (%) Retained
Size
Test Number
Passing Retained Passing Retained Mean 1 2 3 4 5 Average
*** 100 250 150 193.6 1.52 0.53 1.13 0.19 0.13 0.70
100 150 150 106 126.1 3.48 1.00 1.01 0.58 0.26 1.27
150 270 106 53 75.0 82.25 19.70 9.54 7.30 6.41 25.04
270 325 53 45 48.8 2.51 15.30 18.45 4.42 5.51 9.24
325 400 45 37 40.8 1.24 6.88 2.02 1.92 1.15 2.64
400 *** 37 *** 37.0 8.99 56.58 67.85 85.59 86.54 61.11
As described later, the cyclone underflow collected in this step was further processed in a
hindered-bed classifier.
3.3.3 Detailed Testing
The concentrated bulk cyclone underflow material was next run through a hindered-bed
classifier. The objective of this task was to remove any residual slimes that were inherently
misplaced to the coarse underflow by the cyclone. Several hindered-bed evaluations were
considered to optimize the condition of the underflow material, ideally producing the minimum
amount of slimes. The operating parameters were bed level (pressure), fluidization rate, and feed
characteristics (i.e., rate and percent solids).
The underflow from the hindered-bed classifier was then upgraded through conditioning
and flotation evaluations. During conditioning evaluation, the optimum residence time, agitation
intensity, and reagent dosage was established. The flotation testing includes determination of the
optimum operating parameters, including aeration rate, feed rate and froth level. While
ultimately the circuit will include a column cell, the initial test work was conducted with
conventional flotation cells.
115 |
Virginia Tech | Table 3.3: Reagents Used in Phosphate Flotation Test Work.
Reagent Source Description
Fatty acids: Arizona Chemical Sylva FA-1. Contains a low rosin percent
PCS Phosphate Standard FA/FO. Current flotation reagent
Amine: PCS Phosphate Current reagent
Formula βDβ and βNβ. Used as a silica
Na SiO : PQ Industries depressant
2 3
Alum: Fischer Scientific Used as a dolomite depressant.
H O: Erie, PA water system Used in all test work.
2
3.3.4 Process Evaluation
A representative sample of the fine clay refuse stream was screened at plus 150, 150x270,
270x325, 325x400 and minus 400 mesh. The size distribution of the feed is summarized in Table
3.4. A screen analysis of the cyclone underflow material was performed and size fractions were
assayed for BPL, acid insol, Fe O , Al O and MgO content as shown in Table 3.5. Settling tests
2 3 2 3,
were then conducted on this material for future scale up evaluation. The results of the settling
tests are summarized in Table 3.6.
Table 3.4: Average Size Distribution of Feed Sample.
Size Individual Cumulative
Mesh Microns Mass Mass
Passing Retained Passing Retained (%) (%)
*** 100 *** 150 1.0 1.0
100 150 150 106 3.3 4.2
150 270 106 53 3.5 7.8
270 325 53 45 0.8 8.6
325 400 45 37 0.9 9.5
400 *** 37 *** 90.5 100.0
117 |
Virginia Tech | Table 3.5: Cyclone U/F Analysis by Size Fraction (Flotation Feed).
Size (Mesh) Size (Microns) BPL Insol Fe O Al O MgO
2 3 2 3
Passing Retained Passing Retained (%) (%) (%) (%)
*** 100 *** 150 26.64 51.61 1.16 5.94 2.00
100 150 150 106 15.11 74.28 0.98 2.98 0.58
150 270 106 53 30.25 49.64 1.09 5.86 1.45
270 325 53 45 27.57 48.02 0.89 5.34 2.70
325 400 45 37 25.79 50.18 0.61 4.89 2.97
400 *** 37 *** 18.31 58.86 2.07 8.21 2.14
Table 3.6: Cyclone Underflow Settling Tests.
Post
Size Pre Post Settling
Settling Settling Normalized Percent
Mesh Microns Average Average Average Passing
Passing Retained Passing Retained Mean (%) (%) (%) (%)
*** 100 265 150 199 4.1 2.8 6.0 100.0
100 150 150 106 126 0.9 0.6 1.3 94.0
150 270 106 53 75 3.5 2.4 5.1 92.7
270 325 53 45 49 2.1 1.4 3.0 87.6
325 400 45 37 41 1.0 0.7 1.4 84.6
400 *** 37 *** 37 88.4 38.7 83.1 83.1
Of the initial five tests conducted, an unacceptable 36% BPL was achieved with very
high insol and magnesium content. The second series of testing was conducted several weeks
later with alternative reagents and flotation techniques that provided more promising results. A
total of 21 tests were conducted, with the best test achieving a 58% BPL, 1.75% MgO and 5.5%
Al O . A summary of the concentrate and tailings BPL is shown in Figure 3.4, showing none of
2 3
the tests achieved the 65% BPL minimum concentrate grade. As a result, a mineralogical study
was undertaken to develop a better understanding of the potential reasons for the poor results
(Section 3.3.6).
118 |
Virginia Tech | 3.3.6 Mineralogical Investigation
3.3.6.1 SEM/EDX Introduction and Setup
The clay tailings sample acquired at PCS Phosphate in White Springs, Florida in January
2004 was analyzed by at the Department of Geological Sciences at Virginia Tech on the
SEM/EDX. The Geological Sciences Department at Virginia Tech owns and operates a
Cambridge Instruments Camscan II Scanning Electron Microscope (SEM) outfitted with an
American Nuclear Systems System 4001 EDX spectroscopy system. The system produces a
qualitative, not quantitative image by either back-scattered or secondary electrons. Over a dozen
different particles were thoroughly analyzed during the 2 day period, of which this report
describes nine of the most common. For each particle, a snapshot was taken along with the
corresponding spectrum plot. A software program called Quantum Excalibur was utilized to
view the spectrum. The spectrum plot measures the energy of each peak (in keV) versus the
intensity of each mineral.
3.3.6.2 Results
Some general comments about what was seen on the SEM. First of all, most of the
individual particles were covered with fine clay, making it difficult to get the exact components
of the particle. The coating caused a significant amount of interference in the analysis. There are
few samples that could be easily identified as exactly one mineral; virtually everything was a
composition of Al, Si, P, and Ca, with small amounts of Mg, Cl, K, Fe and Ti.
Aluminum was associated with virtually all the particles that were analyzed, as most
commonly aluminum phosphate and aluminum silicate clay. There were several calcium
particles that were obviously dead marine life that were virtually all calcium. Some particles
120 |
Virginia Tech | 3.5 Conclusions
1. Based on the SEM/EDX analysis, the clay refuse stream sample appears to be composed
of about equal amounts of calcium phosphate, aluminum phosphate and silica. The
presence of potassium feldspar and calcium carbonate (seashell material) was visible, but
not as prominent as the aforementioned. The presence of aluminum was significant in
virtually all of the particles identified, where as very little magnesium was found in
comparison. As expected, the presence of βpureβ particles is minimal as the ultra-fine
material has coated all particles of any significant size.
2. The flotation results follow what is shown in the mineralogical analysis. The absence of
pure apatite particles and the presence of an assortment of clay particles inhibit the
successful concentration of phosphate particles by flotation. A better understanding of
the interactions between the clay and apatite particles is necessary before successful
completion of this task can occur.
3. After a better understanding of the clay interactions is conceived, it is believed that an
acceptable concentrate grade can be achieved with further refinement of the test criteria
to minimize the clay/apatite interactions. Additional flotation test work is scheduled for
the near future while further research is currently being conducted on the clay
interactions.
134 |
Virginia Tech | Development of a Novel Air Sparging Device
Andrew Reid Hobert
ABSTRACT
Column flotation is commonly employed in the processing and recovery of fine mineral
particles due to an increase in flotation selectivity unattainable using conventional flotation
methods. Such an increase in selectivity is due to the employment of wash water, minimizing
hydraulic entrainment of fine gangue particles, and the presence of quiescent operating
conditions assisted by the use of various air sparging technologies. High performance air
spargers increase the probability of collision and attachment between air bubbles and particles,
thereby improving recovery of fine and coarse mineral particles otherwise misplaced to the
tailings fraction in conventional flotation cells. Although many high-pressure spargers, including
the static mixer and cavitation tube, are currently employed for the aeration of column cells, a
low pressure sparger capable of providing equivalent performance while resisting a reduction in
aeration efficiency does not exist.
In light of escalated energy requirements for operation of air compressors necessary to
provide high pressure air to existing external and internal spargers, a low-pressure and porous
sparger capable of resisting plugging and scaling was developed. Following the design,
construction, and optimization of such a prototype, air holdup and flotation performance testing
was completed to verify the viability of the sparger as a replacement to existing aerators.
Performance evaluations suggest that the sparger is capable of providing similar functionality to
currently employed sparging technologies, but further work is required with regards to
manipulation of the porous medium to prevent sparger fouling and sustain high aeration
efficiencies. |
Virginia Tech | 1.0 INTRODUCTION
1.1 Background
Froth flotation is a method of fine particle separation, physical or chemical, which
utilizes differences in surface chemistry of minerals within a mineral/water slurry. Flotation is
employed for the recovery of valuable fine grained ores, often less than 100 microns in size and
either technically or economically unrecoverable by gravity concentration or other separation
techniques such as magnetic separation. Through the introduction of air to a liquid pulp, air
bubbles selectively adhere to naturally, or chemically altered, hydrophobic minerals and carry
those solids to a surface froth phase for removal. Easily wetted, or hydrophilic, material remains
in the pulp phase for removal via a tailings or refuse stream. Froth flotation can be performed
using an array of established flotation technologies and methods, but is most commonly executed
using mechanical and column flotation cells. Conventional froth flotation, also known as
mechanical flotation, utilizes a mechanical agitator to disperse air into a mineral slurry using a
rotating impeller. Conventional flotation cells are capable of yielding high mineral recoveries
when operated in series, but suffer from limited product grades and non-selectivity due to short
circuiting of gangue laden feed water, poor recovery of fine particles less than 20 micron, and
entrainment of fine waste particles. The efficiency of fine particle flotation using conventional
flotation is also poor due to the low probability of collision between fine particles and bubbles.
To solve such issues and improve process efficiencies, column flotation is performed
using quiescent countercurrent flows of air and feed slurry in a taller cell to eliminate intense
shearing and increase flotation selectivity. The quiescent conditions provided by this flotation
method improve the selective flotation of both fine and coarse particles (Luttrell & Yoon, 1993).
Downward flowing wash water is also added to the froth phase to minimize the hydraulic
1 |
Virginia Tech | entrainment of fine gangue particles. As a result, column flotation has become widely accepted
for its ability to produce higher grade products at increased product yields.
To improve mass recoveries and minimize the misplacement of high grade fine particles
to the tailings or refuse stream, column flotation employs a more unique aeration method.
Conversely to the employment of mechanical agitation in conventional froth flotation for the
aeration of a mineral pulp or slurry, column flotation uses an array of air spargers to introduce a
fine upward rising air bubble distribution at the base of a column flotation cell. The introduction
of a finer air bubble distribution improves flotation kinetics and increases the total bubble surface
area flux, or total available bubble surface area for mass transfer (Laskowski, 2001). Efficient
and proper air sparging is vital to the success of column flotation as an increase in bubble size
promptly decreases the probability of bubble-particle collision. Existing sparging technologies
include, but are not limited to, porous spargers, one-phase and two-phase jetting spargers,
hydrodynamic cavitation tubes, and static in-line mixers, many of which are operated using high-
pressure compressed air.
Given the significant horsepower requirements necessary to supply compressed air to
currently operated high-pressure spargers, a low-pressure sparger operable by use of an air
blower offers potentially substantial economic gains. Additionally, the capital cost of an air
blower is considerably lower than that of a high horsepower air compressor. Existing low
pressure spargers, such as sintered metal porous spargers, are operable at significantly lower
pressures, but often suffer from plugging and diminishing performance when introduced to a
mineral pulp, thereby reducing sparging efficiency and increasing air pressure demands over
time. As a result, the development of a non-plugging, low-pressure sparger capable of providing
equivalent metallurgical performance to both existing jetting and dynamic external spargers
2 |
Virginia Tech | would provide both operational and capital cost savings in many global processing beneficiation
applications that employ column flotation.
1.2 Project Objective
The objective of this project was to design and develop a low pressure porous sparger
capable of resisting a diminishment in performance over extended periods of use in column
flotation applications. Such a sparger could be used as an alternative to existing in-line spargers
employed in column flotation applications, but would be capable of operation by means of a low-
pressure blower. The tasks completed in this research and development project include the
design of a low pressure sparger, construction of a sparger with alterable parameters, and
completion of an array of test work to verify the viability of the spargerβs performance.
The sparger designed in this work effort utilizes magnetism to retain a magnetic media
bed through which air is dispersed into a moving slurry. By use of a porous medium, incoming
air is distributed and broken into fine air streams before introduction to a recirculated mineral
pulp. Similarly to the Microcel design, air bubbles directly contact moving particles to increase
the probability of bubble-particle attachment. Magnetism was strictly chosen with a goal of
manipulating or rotating the internal magnetic material by movement of external magnets or an
alteration in magnetic fields. Although the overall objective of this project was to design a long
term non-plugging porous sparger, test work was devoted to proving the viability of the designed
sparger as a means of aeration in column flotation processes. This information then makes it
possible to test the sparger with several proposed cleaning concepts in both a laboratory and pilot
plant setting with knowledge that the sparger is capable of providing sufficient fine and coarse
particle flotation performance.
3 |
Virginia Tech | 2.0 LITERATURE REVIEW
2.1 Introduction to Flotation
Throughout history, numerous forms of technology or processes have been developed for
the separation of minerals by density, size, and chemical properties. To concentrate fine particles
unrecoverable by existing technology, mineral flotation was tested and established in the mid
1800βs for the separation of minerals using differences in surface chemical properties. Following
the initial patenting of a flotation concept used for the separation of sulfides in 1860 by William
Haynes, the Bessel brothers designed and constructed the first commercial flotation plant, for the
purpose of cleaning graphite minerals, in Germany in 1877 (Fuerstenau, Jameson, & Yoon,
2007). In addition to their innovative use of nonpolar oils to improve the process kinetics of
graphite by mineral agglomeration, the Bessel brothers were the first to reportedly use bubbles,
resulting from boiling, to increase flotation rates of graphite in water. Flotation continued to
develop throughout the late 1800βs as multiple methods of sulfide flotation began to expand. For
example, in 1898, Francis Elmore patented and implemented a process utilizing oil to
agglomerate pulverized ores and carry them to the surface of water for the concentration of
sulfide minerals at the Glasdir Mine in Wales (Fuerstenau, Jameson, & Yoon, 2007).
The physical separation or concentration of fine particles by true froth flotation was first
utilized in 1905 for the separation of lead and zinc ores from tailings dumps at Broken Hillβs
Block 14 mine in Australia (Hines & Vincent, 1962). Shortly after, the Butte and Superior
Copper Company built the first froth flotation plant in the United States in 1911 (Hines &
Vincent, 1962). Due to the success of sulfide flotation in the early 1900βs, the use of copper in
the United States grew by approximately 5.8 percent annually during that time period (Hines &
Vincent, 1962).
4 |
Virginia Tech | Although numerous flotation methods and flotation cells have been developed throughout
mineral processing history, the conventional mechanical cell and the column cell are most
common in present day mineral processing. The mechanical cell was invented in 1912 and is the
most widely implemented or accepted form of flotation today. To aerate a slurry, mechanical
cells, as shown in Figure 1, utilize an agitator consisting of a stator and impeller. Air is naturally
drawn down the stator or delivered using a low-pressure blower, and dispersed by an impeller
which agitates, circulates, and mixes the flotation pulp with the introduced air. As a result of
high intensity mixing between air and solids, physical contact between particles and air bubbles
occurs.
Figure 1. Conventional Flotation Cell Schematic (Luttrell G., Industrial Evaluation of the StackCell Flotation
Technology, 2011), Used with permission of Dr. Gerald Luttrell, 2014
Mechanical cells are beneficial in that they are capable of treating high material
throughputs, but struggle with lower concentrate grades due to short circuiting of feed water to
the froth phase and non-selective entrainment of fine particles. To overcome these challenges
and improve overall recovery and grade, cells can be operated in series or various flotation
circuits can be implemented. For example, in a rougher-cleaner circuit, the concentrate of a
5 |
Virginia Tech | single rougher bank is re-floated to βcleanβ or refine the product, improving concentrate grade.
Additionally, a rougher-scavenger circuit can be implemented to re-float the tailings from the
rougher bank to improve overall recovery as high solids loading in a rougher bank can lead to
inefficiencies in separation and misplacement of coarser material to the tailings due to froth
crowding. Vast circuit configurations consisting of multiple cells and recirculation of material, as
shown in the rougher-scavenger-cleaner circuit in Figure 2, provide improved recoveries and
higher product grades, but equipment and operational expenditures significantly increase and
efficiencies in fine and coarse particle flotation remain low. Due to its ability to improve
separation efficiencies, minimize hydraulic entrainment, increase fine particle flotation
selectivity, and yield higher product grades, while often requiring fewer flotation cells and
reagent volumes, column flotation has flourished in the mineral processing industry.
Figure 2. Rougher, Cleaner, Scavenger Flotation Circuit
2.2 Column Flotation
Column flotation is a form of innovative froth flotation that uses the countercurrent flow
of air bubbles and solid particles in a pneumatic cell. Pneumatic flotation, performed in a column
6 |
Virginia Tech | like structure with an air sparging device, was first developed by Callow in 1914 and the concept
of countercurrent flow of slurry and air within a flotation column was later introduced in 1919 by
Town and Flynn, as described by Rubenstein (Rubenstein, 1995). The column flotation concept,
as it is known today, was further investigated and patented in the 1960βs by Boutin and Tremblay
and is currently employed in the roughing, scavenging, and cleaning of valuable minerals such as
gold, copper, coal, and zinc (Finch & Dobby, 1990).
Although various forms of column designs were developed in the 1970βs and 1980βs,
including the Hydrochem and Flotaire column cells, the Canadian column was the first to be
developed and is the most commonly implemented form of column cell in todayβs processing
applications, as reviewed by Dobby (Finch & Dobby, 1990). The Canadian column was first
tested in the late 1960βs by Wheeler and Boutin and the first commercial column cell was
installed for the cleaning of Molybdenum ore in 1981 at Les Mines Gaspe in Quebec, Canada
(Wheeler, 1988). Following its successful application in molybdenum cleaning, the column cell
became more widely applied for the flotation of sulfide and gold ores, as well as coal, and the
cleaning of copper, lead, zinc, and tin in the late 1980βs and early 1990βs (Wheeler, 1988). The
rapid employment of the Canadian column, and development of the column flotation method, in
many mineral processing applications can be attributed to its ability to yield improved product
grades, while increasing the recovery of both fine and coarse particles.
Separation of fine particles with high specific surface areas, resulting from crushing and
grinding to liberate mineral value, while concentrating a high grade product requires control of
hydraulic entrainment of fine gangue particles. In comparison to the flotation performance
offered by conventional, mechanically agitated flotation cells, columns yield a higher quality
concentrate grade in a single flotation stage due to the removal of entrained fine gangue particles
7 |
Virginia Tech | reporting to the froth through the use of wash water. Wash water showers the froth bed of the
column vessel to eliminate entrained gangue minerals that degrade the product grade and
replaces pulp that normally reports to the concentrate in conventional flotation methods with
fresh water (Kohmuench, 2012). The flow rate of pulp to the froth concentrate must be less than
the countercurrent flow of wash water to minimize non-selective recovery of ultrafine gangue
(Luttrell & Yoon, 1993). Column cells, as shown in Figure 2, are also built with a smaller cross
sectional area to maintain a deeper and more stable froth bed necessary for froth washing. In
addition to froth washing and the use of a deep froth, column flotation promotes quiescent
operating conditions and utilizes air spargers to improve flotation selectivity and fine particle
recovery, respectively.
Unlike conventional flotation cells, taller column cells, reaching up to 16 meters in height
to permit necessary particle residence times, utilize high-pressure internal or external spargers
for very fine air bubble introduction. Air, or an air/water mixture, is injected at the base of the
cell, via an arrangement of spargers, as feed is introduced below the froth bed, developing a
countercurrent flow of feed particles and air bubbles. Due to lower traveling velocities of both
bubbles and particles during column flotation, collision and attachment between the two are
more likely. Increased contact time between air bubbles and particles under quiescent operating
conditions decreases the probability of hydrophilic particle attachment. Such conditions also
greatly improve coarse particle collection efficiency as coarse particles are less likely to detach
from air bubbles under less turbulent conditions present within column cells. Due to the use of
wash water and a deeper froth bed, a large quiescent pulp or contact zone, and various air
sparging technologies, column flotation presents the most ideal separation environment in a
single flotation stage.
8 |
Virginia Tech | Essentially, column flotation presents a multi stage flotation circuit within one cell,
illustrated by the development of the MicrocelTM by Luttrell et al. (United States of America
Patent No. 5761008, 1988). For example, the pulp zone or bubble-particle contacting region
represents a rougher stage as hydrophobic particles adhere to air bubbles and are carried to the
froth phase in this zone. Additionally, a deep froth and wash water are used to clean the
concentrate of hydraulically entrained fine gangue particles to represent the cleaning phase of a
multi-state circuit process. Lastly, external air spargers are often utilized to directly introduce air
to a circulated tailings stream, scavenging possibly misplaced fine hydrophobic particles. As a
result, column flotation is increasingly preferred for the flotation of finer particle size classes to
improve recovery and concentrate grade.
Figure 3. Flow of Water in a Column Flotation Cell (Luttrell G., Industrial Evaluation of the StackCell Flotation
Technology, 2011), Used with permission of Dr. Gerald Luttrell, 2014
Although column flotation produces superior product grades than those yielded by
conventional flotation, consideration must be given to carrying capacity for proper cell design.
9 |
Virginia Tech | Carrying capacity, in pounds or tons per hour per square foot, is the mass rate of floatable solids
that can be carried by a given superficial gas velocity. As described by Luttrell, a column cell
must be scaled according to its carrying capacity due to an inherently smaller ratio between
column cross sectional area and volume when compared to conventional flotation cells (Luttrell
& Yoon, 1993). The equation for carrying capacity, in mass rate of concentrate solids per unit of
cell area, is as shown:
C = 4 Q D Ο Ξ² / D [1]
g p b
where Ξ² is a packing efficiency factor, Ο is the particle density, D is the particle diameter in the
p
froth, D is the bubble diameter, and Q is the gas flow rate. To achieve optimal carrying capacity
b g
conditions, column cell spargers are operated at maximum allowable air velocities, while
maintaining the minimum average bubble size. The maximum air flow rate is governed by the
bubble size and V , or superficial liquid velocity in the cell. Particle residence times are also
L
higher in column flotation due to a taller pulp zone and the naturally slower rise of small
bubbles. Given increased particle residence times present using column cells given their
geometry, work has been completed to develop flotation technology which offers column-like
performance with significantly reduced particle residence times.
In addition to the column cell, further work has been done in recent years by the Eriez
Flotation Division to develop an innovative form of flotation technology labeled the StackCell
(Kohmuench, Mankosa, & Yan, 2010). The StackCell, as shown in Figure 4, offers column-like
performance with shorter particle residence times, improved bubble-particle contacting, and a
reduced unit footprint per processed ton of material (Kiser, Bratton, & Kohmuench, 2012).
Unlike typical column cells, the StackCell utilizes an aeration chamber to agitate and mix the
feed with air in a high intensity shearing zone. By the act of intense agitation, low pressure air,
10 |
Virginia Tech | introduced to the slurry before entrance to the aeration chamber, is sheared into small bubbles for
collection of fine particles. High particle concentration and gas fraction within the chamber
greatly reduces particle residence times. Additionally, the use of low pressure air and the
turbulent environment within the pre-aeration chamber decreases energy requirements as energy
is primarily consumed in bubble-particle contacting instead of particle suspension (Kiser,
Bratton, & Kohmuench, 2012). The mixture of slurry and air lastly overflows into a short,
column like tank, where froth and pulp are separated. A deep froth bed is maintained and froth
wash water is employed with the StackCell design to reduce hydraulic entrainment, similarly to
column flotation. Due to their compact size and ability to be stacked in unison, stackcells offer
much friendlier orientation in a processing plant than column cells, which are much larger and
require more structural steel for support and allowance of de-aeration of the froth before
reporting to the dewatering circuit (Kohmuench, Mankosa, & Yan, 2010).
Figure 4. Schematic Illustration of a Single Eriez StackCell (Kohmuench, Mankosa, & Yan, Evalutation of the
StackCell Technology for Coal Applications, 2010), Used with permission of Dr. Jaisen Kohmuench, 2014
11 |
Virginia Tech | 2.3 Column Flotation Performance
Flotation performance is influenced by many factors including froth depth and structure,
slurry flow characteristics, flotation cell dimensions, wash water utilization, chemical additions,
and gas holdup. Air holdup in a gaseous-liquid mixture is primarily controlled by bubble size or
frother dosage, superficial gas velocity, slurry density, and superficial or discharge liquid
velocity, as detailed in studies completed by Yianatos et al, and Finch and Dobby (Finch &
Dobby, 1990; Yianatos, Finch, & Laplante, 1985). Column cells are typically operated with a 12
to 15 percent gas holdup, or percentage of air in a gaseous-liquid volume. Superficial gas
velocity and gas holdup maintain a positive linear relationship, in what is known as the
homogenous bubbly flow regime, until gas flow rate becomes too significant (Finch & Dobby,
1990). At this point, air begins to coalesce, bubble size uniformity is lost, and water is displaced
to the froth phase.
To illustrate the reaction of a typical flotation bank as gas flow rate is increased, the
effect of superficial gas velocity on the recovery and grade curve of a Mt. Isa copper rougher
flotation bank is shown in Figure 5. As superficial gas velocity was increased, copper recovery
also increased, but copper grade diminished due to increased recovery of gangue material.
Although not evidenced in Figure 5, instigation of air coalescence quickly decreases recovery.
As stated by Finch and Dobby, development of very large and quickly rising air bubbles will
create a churn-turbulent regime within a flotation column as superficial gas velocity exceeds
approximately 3 to 4 cm/s (Finch & Dobby, 1990). Coalescence of air quickens the rise of air in
a column and increases mean bubble diameter, decreasing total bubble surface are and
diminishing bubble-particle collision efficiency.
12 |
Virginia Tech | Figure 5. Effect of Aeration rate on the Grade and Recovery Curve for a Copper Ore (Finch, Mineral and
Coal Flotation Circuits, 1981), Used under fair use, 2014
As gas holdup increases, the probability of bubble-particle collision also increases as the
availability of bubble surface area for mass transfer escalates. This is due to a decrease in bubble
size or increase in gas flow rate, both of which control the bubble surface area rate, or S . The
b
bubble surface area rate is the ratio between superficial gas velocity and bubble sauter diameter.
The equation for S is as follows:
b
S = 6V /D [2]
b g b
where D is the diameter of bubbles and V is the superficial aeration rate (Luttrell & Yoon,
b g
1993). Probability of collision between a particle and bubble is predominantly dependent upon
particle diameter, bubble diameter, and bubble Reynolds number. Yoon and Luttrell (1989)
derived an equation for collision probability that states that as bubble diameter decreases or
bubble Reynolds number increases, the probability of collision between a bubble and particle
increases. Their derived equation for probability of collision is written as follows:
π =
[3
+
4π
π0.72 ](π·π)2
[3]
π
2 15 π·
π
where D represents the particle diameter, D is the bubble size, and Re is the bubble Reynolds
p b
number. As particle size decreases, the probability of collision and attachment between a particle
13 |
Virginia Tech | and bubble also decreases (Yoon & Luttrell, 1989). Ralston et al. also derived mathematical
equations which suggest that the probability of bubble-particle attachment decreases if particles
are too coarse as air bubbles become unable to retain such heavy particle loads (Ralston, Dukhin,
& Mischuk, 1999). Although bubble size must be minimized at a maximum allowable superficial
air velocity to increase the probability of fine and coarse particle collision and attachment,
hydraulic entrainment of gangue laden water must also be managed to maximize product grade.
During operation of a column flotation cell, proper bias water rates and froth depths must
be utilized to minimize hydraulic entrainment. The effect of wash water utilization on the
hindrance of short-circuited feed water to the concentrate is shown in Figure 6. As evidenced by
the column flotation tracer study displayed in Figure 6 (Left), employment of wash water
cultivates a clean interface free of pulp water contamination.
Figure 6. Tracer Study showing the effects of Wash Water Utilization
Bias is a measurement of the percentage of wash water which reports to the pulp; the
remainder of water reporting to the froth zone. According to Dobby and Finch (1990), an
adequate bias rate and froth depth are essential to control concentrate grade as gas flow rates are
commonly maximized to improve column carrying capacity. Although wash water is required to
optimize the particle cleaning process, a minimum bias rate is recommended, up to 80 percent of
14 |
Virginia Tech | which should flow to the concentrate, to prevent short circuiting of material to the overflow,
maximize capacity, and ensure mobility of the froth (Finch & Dobby, 1990). Insufficient wash
water will lead to a reduction in product grade and a concentrate flow rate greater than that of the
wash water. In laboratory and pilot scale testing, a froth depth of greater than one half meter is
suggested for gas rates exceeding 2 cm/s to diminish the feed water concentration in the froth
zone (Finch & Dobby, 1990). In addition to the many column cell operational parameters which
effect flotation performance, chemical reagents are most intrumental in the flotation of many
naturually hydrophilic materials.
Although some minerals or rock types, such as coal, are naturally hydrophobic, bubble-
particle attachment is strongly dependent upon chemical reagents such as collectors, activators,
depressants, and pH modifiers. Collectors (anionic, cationic, or nonionic) are used to generate a
thin, nonpolar hydrophobic layer around a particle, rendering it hydrophobic. Selection of
collector is dependent upon the charge, positive or negative, or the chemical make-up of the
mineral to be floated. Activators and depressants are then used to allow or prevent the collector
from physically or chemically adsorbing to a mineral surface, respectively. Lastly, and very
importantly, pH modifiers are necessary to control the charge of minerals as a mineralsβ charge
often becomes more positive as a solution decreases in pH from alkaline to acidic conditions.
In addition the importance of collectors and other chemical reagents in promoting the
development of bubble-particle aggregates, frother type and dosage dictate both bubble size and
rise velocity within a flotation cell. Frother can be either a water soluble or insoluble polymer
that stabilizes the dispersion of air bubbles in a flotation pulp by decreasing its surface tension.
As surface tension declines, bubble population grows and average bubble diameter decreases.
Although many frothers have been developed, the two main classes of frother are alcohols and
15 |
Virginia Tech | polyglycols. Polyglycols help to quickly stabilize a froth, while alcohols are used to expedite the
increase of gas holdup (Cappuccitti & Finch, 2009). Much work has been performed to generate
relationships between gas velocity and gas holdup using numerous frother types and
concentrations (Yianatos, Finch, & Dobby, 1987; Finch & Dobby, 1990; Lee, 2002). Typically,
as frother concentration increases, in parts per million, mean bubble diameter reduces and gas
holdup rises. In addition to the calculation of gas holdup within a laboratory column using a
measurement of pressure differential and pulp density, methods of measuring and
mathematically estimating bubble sizes using photography have also been developed to better
understand the effects of numerous flotation operating parameters (Yianatos, Finch, & Dobby,
1987). Although operational set-points can be altered to impact flotation recovery and grade, the
actual method of bubble generation is integral in obtaining desired flotation performance.
2.4 Flotation Sparging
In column flotation processes, internal and external spargers are utilized to introduce and
disperse air into a liquid-mineral pulp. Proper sparger design and performance is essential to
column flotation as spargers are used control bubble size, air distribution, and air holdup within
the flotation column. External spargers are used to aerate a moving slurry which is pumped from
a flotation cell bottom and recirculated as a pulp-air mixture to the column; whereas internal
spargers inject air or an air-water mixture directly to the flotation cell. External spargers and
some internal spargers, such as the Eriez SlamJet, have expedited the development of column
flotation as they can be maintained during operation of the column and are easily operated. Since
the development of column flotation, numerous spargers have been established and industrialized
to improve bubble dispersion, minimize bubble size, decrease operational costs, and reduce
maintenance difficulties.
16 |
Virginia Tech | As expressed by Rubenstein (1995), Callow developed the first pneumatic column
flotation sparger in 1914 using a perforated metal frame that was wrapped in a woolen cloth. Air
was then introduced to a slurry through the covered frame. Many similar spargers were proposed
and tested in the early to mid-1900βs, but all suffered from plugging, improper distribution of air,
and poor reliability. As sparging technologies have rapidly developed and improved in the last
few decades, the popularity of column flotation has grown. Although the overall goal of air
sparging remains constant for each type of sparger; the design, sparging method, and features
vary substantially between sparger types. Due to differences in operating conditions present in a
laboratory in comparison to those found in industrial applications, certain low pressure sparging
devices are only feasible in a laboratory setting. Though sparging is applied to industries outside
of mineral processing, column flotation sparging for the purpose of valuable mineral recovery
will be strictly examined in this report. A detailed explanation of the design and operation of
existing spargers in mineral flotation applications is provided to illustrate the advancements and
differences in column flotation sparging technologies.
2.5 Internal Spargers
2.5.1 Low Pressure Perforated and Porous Spargers
The perforated sparger characterizes the beginning of sparging in a pneumatic flotation
cell. Sparging devices developed in the early 1900βs for use in pneumatic columns often
consisted of a perforated metal frame or structure through which low pressure air was
introduced, sometimes inclusive of a porous filter cover. Filter cloth covers were used to
generate finer bubble sizes at low pressures, but suffered from fouling or degradation over
extensive periods of use. As a result, various materials, such as glass, ceramic, and fabric, have
been used in construction of more rigid porous spargers, but the sintered metal sparger has
17 |
Virginia Tech | become the most widely accepted form of porous sparger. This is due to its rigid construction
and ability to produce the most uniform dispersion of fine air bubbles of existing low pressure
porous aerators. Sintered metal spargers are comprised of powdered metal which has been fused
together due to subjection to heat near the metalβs melting point (Mott Corportation, 2014). The
average pore size of sintered metal spargers ranges from 60 to 100 microns; therefore allowing
the production of extremely fine bubbles (Mott Corportation, 2014). Mouza and Kazakis (2007)
studied the effects of porous sparger aperture size and found that sintered metal spargers with a
smaller average pore diameter have a more uniform porosity and therefore maintain a more even
air distribution (Kazakis, Mouza, & Paras, 2007). According to a study performed by the
University of Florida, the average diameter of a bubble emitted from a sintered aluminum or
stainless steel sparger ranges from 0.7 to 0.9 millimeters (El-Shall & Svoronos, 2001). Although
sintered metal spargers are capable of producing a more fine bubble distribution when compared
to other internal and external sparging methods, sintered metal spargers likewise possess the
inability to resist plugging when exposed to a slurry or pulp in an industrial environment. Such
sparging inefficiencies have primarily been documented in wastewater treatment applications
and the separation of oil and water.
As reviewed by Rosso (2005), periodic cleaning of fine pore spargers using water and
acid is necessary to prevent a rapid performance decline in wastewater treatment applications
due to slime plugging. In a study of porous sparger aearation efficiencies in wastewater treatment
applications, diminishing sparger performance was obvious in 21 analyzed wastewater facilities
(Rosso & Stenstrom, 2005). Porous spargers require filtered air and water to promote successful
continuous flotation and minimize performance deterioration, both of which are not feasible in
most industrial beneficiation plants. Single and two-phase porous spargers are sometimes applied
18 |
Virginia Tech | in de-inking flotation, wastewater treatment, and in the separation of oil and water, but are
primarily used in a laboratory setting in mineral flotation efforts due to uneconomical
maintenance requirements. A single phase sintered metal sparger introduces air only through a
porous membrane within a column cell, whereas a two phase sintered metal sparger, as shown in
Figure 7, injects air through a porous medium surrounding the circumference of a moving stream
of water. The aerated liquid then flows into the column or aerated tank.
Figure 7. Two Phase Sintered Metal Porous Sparger (El-Shall & Svoronos, Bubble Generation, Design,
Modeling and Optimization of Novel Flotation Columns for Phosphate Beneficiation, 2001), Used under fair use,
2014
2.5.2 Single and Two Phase Jetting Spargers
In addition to low pressure porous spargers or bubblers, the US Bureau of Mines
(USBM), Cominco, and Canadian Process Technologies (CPT) have developed various forms of
high pressure jetting internal spargers. In contrast to the operation porous spargers, the jetting
action of these high pressure spargers allows for the emergence of numerous air bubbles from a
single or multiple orifices with a reduced risk of plugging. Although sparger fouling is less likely
at higher pressures, horsepower requirements for the generation of higher air pressures greatly
increase operational costs. Cominco and USBM produced the first two phase, high velocity
internal sparger that mixes both water and high pressure air before injecting the air/water mixture
into the column cell through a perforated pipe. To improve the distribution of water in air,
USBM also formulated a model that uses a bead filled mixing chamber to mix water and high
pressure air streams. Water addition is used to shear the incoming air, therefore creating a finer
19 |
Virginia Tech | bubble distribution (Finch, 1994). To improve the concept of on-line maintenance, unattainable
by USBM and Cominco spargers, CPT later industrialized a single air phase SparJet sparger. The
SparJet is a removable air lance that ejects high velocity air from a single orifice through the
column wall. Multiple air lances of varying length can be instrumented around the column
perimeter to aerate the full cross sectional area of the column. To adjust the air flow through the
sparger orifice or to close the orifice in the event of pressure loss, a t-valve is located at the
opposite end of the sparger to increase or decrease the total orifice area. CPT later replaced the
needle valve with a high tension spring that controls the orifice area depending upon the
provided air pressure. The spring is located at the sparger end opposite the orifice and is used to
control the position of an internal rod relative to the sparger tip. If air pressure is lost, the spring
closes the orifice of the SlamJet to prevent the backflow of slurry into the air system
(Kohmuench, 2012).
As detailed by Finch, a long air jet length stretching from the sparger orifice is desired to
increase the total population of bubbles. By increasing the density of air by addition of water, the
jet length increases and bubbles become finer (Finch, 1994). As a result, EFD enhanced the
SlamJet, pictured in Figure 8, with the addition of water as high pressure air and water enter the
lance together. Multiple SlamJet spargers consisting of unique orifice sizes exist for any
specified flotation duty. The largest SlamJet is operated at pressures in excess of 80 psi for
optimal performance in fine coal flotation. The SlamJet sparger is commonly used in flotation of
a somewhat coarser feed or deslime circuits which require less collision energy (Kohmuench,
2012). For the flotation of finer particles less than 325 mesh in size, external spargers using
direct rapid bubble-particle contacting have been developed.
20 |
Virginia Tech | Figure 8. Eriez SlamJet Sparger (CPT, Canadian Process Technologies, Sparging Systems - SlamJet Series), Used
with the permission of Dr. Michael Mankosa, 2014
2.6 External Spargers
2.6.1 Static Mixer/Microcel
Many internal spargers suffer from limited control of bubble size, plugging of openings,
low collision energies, and poor on-line maintenance capabilities; therefore the development of
external spargers such as the static mixer and cavitation tube have greatly expanded the use of
column flotation through the use of microbubbles and picobubbles in mineral processing
applications. In the mid 1980βs, Luttrell et al. invented and patented the MicrocelTM flotation
column using a static in-line mixing sparger for the purpose of bubble particle contacting (United
States of America Patent No. 5761008, 1988). A static mixer, as shown in Figure 9, is a tube
consisting of a series of geometric shapes used as air-slurry mixing components. Potential
tailings slurry is removed from the column bottom using a pump and is delivered to a static
mixer in addition to high pressure air supplied before the mixer inlet by an air compressor.
Significant air pressure of at least 50 to 60 psi is required for proper operation of
industrial scale static mixers to provide 40 to 50 percent air in slurry by volume and to overcome
the 20 to 25 psi pressure drop experienced by the air-slurry mixture following its movement
through the static mixer. The aerated slurry is then recirculated to the flotation column.
Microbubbles generated using a static mixer range from 0.1 to 0.4 mm in size and vastly increase
the rate of flotation as bubbles remain small at increased superficial gas velocities (Luttrell,
21 |
Virginia Tech | Yoon, Adel, & Mankosa, 2007). Using a static mixer, high separation efficiencies are realized as
a result of a decrease in bubble diameter which increases probabilities of collision and
attachment and decreases the probability of bubble-particle detachment.
Figure 9. Microcel Static Mixer Sparger (Luttrell, Yoon, Adel, & Mankosa, The Application of Microcel Column
Flotation to Fine Coal Cleaning, 2007, pp. 177-188), Used with permission of Dr. Gerald Luttrell, 2014
As reviewed by Luttrell (2007), the use of a static mixer, or other external sparging
technologies, in column flotation applications allows for the development of a three stage
flotation process within one cell (Luttrell, Yoon, Adel, & Mankosa, 2007). This is illustrated by
the successful functionality of the MicrocelTM. As air rises within a flotation column, downward
flowing hydrophobic particles collide and attach to air bubbles in the pulp zone in a roughing
stage. Risen bubble-particle aggregates are then washed or cleaned in the froth bed to minimize
hydraulic entrainment in a cleaning stage. Lastly, the implementation of direct rapid particle
contact within a static mixer recycle circuit gives particles a final opportunity for attachment to
air bubbles in a scavenger phase. This multi-stage process, as shown in Figure 10, represents a
distinct advantage of column floation that is made possible using external in-line spargers such
as the static mixer and cavitation tube.
To demonstrate the value of a static mixer sparging apparatus in the cleaning of fine coal,
Luttrell et al. (2007) completed a flotation test program on multiple minus 28 mesh coal samples
22 |
Virginia Tech | using both column and conventional flotation methods. The separation efficiency, or difference
between the combustible recovery of coal and ash recovery, was 6 to 16 percent greater using a
single twelve inch diameter column equipped with a static mixer than a bank of four, three cubic
foot conventional cells (Luttrell, Yoon, Adel, & Mankosa, 2007). In addition to its application in
fine coal flotation, the static mixer has also displayed proven performance in coarse particle
flotation. In a coarse phosphate flotation study performed at the University of Kentucky (2006),
both a static mixer and porous sparger were individually employed for the flotation of a minus
1.18 mm phosphate ore. The effect of a given sparger type on the concentrate grade and total
phosphate recovery was determined at varying gas velocities. At the optimum operating point, or
elbow of the P O recovery and grade curve, established for each sparger, the static mixer
2 5
provided twelve percent higher phosphate recovery at a slightly better concentrate grade than the
porous bubbler (Tao & Honaker, 2006).
Figure 10. MicrocelTM Process Diagram (Luttrell, Yoon, Adel, & Mankosa, The Application of Microcel Column
Flotation to Fine Coal Cleaning, 2007, pp. 177-188), Used with permission of Dr. Gerald Luttrell, 2014
23 |
Virginia Tech | 2.6.2 Cavitation Sparging
In addition to the static mixer, hydrodynamic cavitation based spargers function as in-line
aerators utilized in fine particle flotation. Cavitation is the formation of cavities or bubbles
within a liquid due to rapid changes in fluid pressure. Bubbles or cavities begin to open up within
a liquid at the location of highest fluid velocity where pressure is negative in an attempt to
relieve pressure (Zhou, Xu, & Finch, 1993). Such fluctuations in pressure are obtained by the
alteration of liquid velocities using a venturi tube configuration. To induce hydrodynamic
cavitation in sparging applications, slurry pressure is reduced below its vapor pressure through
an area constriction, increasing slurry velocity, and is then returned above its vapor pressure
following an increase in slurry flow path cross sectional area. In addition to deviations in fluid
pressure, the presence of solid particles and high gas flow rates both help to promote the
development of cavities. As discovered by Zhou (1993), high dissolved gas volumes and the
addition of frother to decrease the surface tension of a slurry prevent the collapse or implosion of
bubbles in the cavitation process, allowing for the successful flotation of fine particles. Using
hydrodynamic cavitation, the bubble-particle collision stage is further assisted as bubbles directly
form on hydrophobic surfaces immediately during cavitation (Zhou, Xu, & Finch, 1993).
Although several spargers are operable using the principles of hydrodynamic cavitation,
the cavitation tube, developed by Canadian Process Technologies, is the most cost effective, high
performing, and wear resistant cavitation based sparger. Other spargers which rely upon
hydrodynamic cavitation include the eductor and two phase ejector, both of which use a parallel
throat and diffuser to promote cavitation, but are not operated in-line with recirculated middlings
or tailings streams (El-Shall & Svoronos, 2001). The cavitation tube, as shown in Figure 11, is a
completely in-line aerator constructed of a wear resistant material in the shape of an hour glass.
24 |
Virginia Tech | Using an hour glass configuration, slurry pressure is rapidly lowered and increased to support
cavitation. Wear resistivity of the inner hour glass lining and the lack of internal mechanisms
within the cavitation tube are noteworthy as the static mixer relies upon a series of mixing
components in the direct line of slurry flow. Unlike the eductor or two phase ejector, the
cavitation tube, now produced by the Eriez Flotation Division, does not employ a feed jet nozzle
or air-slurry mixing chamber before entrance to a cavitation inducing structure.
Figure 11. Eriez Cavitaton Tube Sparger (Canadian Process Technologies, Cavitation Sparging System), Used
with the permission of Dr. Michael Mankosa, 2014
Similar to the operation of the static mixer, a pressure drop of approximately 20 to 25 psi
occurs across the length of a cavitation tube as a result of rapid modifications in liquid velocity.
An operating pressure of 50 to 60 psi is also recommended for industrial applications. As
reported by Kohmuench, cavitation tube sparging is common to fine, by-zero, coal circuits that
are operated under significant material throughputs as a result of the spargerβs ability to
formulate numerous bubbles less than 0.8 mm in size (Kohmuench, 2012). The benefits of
cavitation and picobubble sparging in fine particle flotation are also evidenced by improvements
in product recovery in the flotation of both zinc sulfides and phosphates (Zhou, Xu, & Finch,
1997; Tao & Honaker, 2006). A cavitation tube was used to pre-aerate the feed to a conventional
cell and in unison with a static mixer in each scenario, respectively.
25 |
Virginia Tech | 2.7 Sparger Comparisons
The air holdup, or percentage of air within an aerated pulp, is directly related to the
bubble size generated by a given sparger type and the associated superficial gas velocity. As
average bubble size decreases, the air or gas holdup directly increases. Air holdup also increases
as a result of an increase in superficial air velocity, until air begins to coalesce. In a study of
various sparging technologies conducted at the University of Florida (2001), the performance of
multiple sparger types was reviewed under changing operating conditions. Most importantly, the
test effort analyzed the effects of increasing superficial air velocity on both the air holdup and
bubble size produced by each analyzed sparger. During the sparger performance analysis, a one-
phase porous sparger provided the greatest air holdup as gas flow rates were increased, but
cavitation based spargers generated the finest bubble sauter diameter, as small as 0.4 mm, at low
superficial air velocities (El-Shall & Svoronos, 2001). A complete assessment of the air holdup
produced by porous, perforated tube, static mixer, hydrodynamic cavitation based, and jetting
spargers as superficial air velocity was increased is shown in Figure 12. As evidenced by this
figure, sparging method strongly dictates the relationship between a specified superficial air
velocity and resulting air holdup within a flotation column cell.
26 |
Virginia Tech | Figure 12. Comparison of Performance for Various Spargers (El-Shall & Svoronos, Bubble Generation, Design,
Modeling and Optimization of Novel Flotation Columns for Phosphate Beneficiation, 2001), Used under fair use,
2014
In an analysis of the sparging performance generated by the selection of aerators, in terms
of air holdup, it is apparent that similar sparging methods provide varying results at increasing
superficial air velocities. For example, the one and two phase porous spargers are comprised of
an identical porous sintered medium, yet the air holdup generated by the single phase sparger
was two to three times greater at increased air velocities due to an escalation in bubble
population and lower bubble entrance velocities. In comparison to the two phase porous sparger,
the static mixer produced a similar minimum bubble size of 0.7 mm, but at slightly greater gas
holdup percentages as air rates were increased. Although the air holdup induced by the static
mixer and two phase porous sparger were much less than that of a one phase porous sparger, this
study did not consider the effect of increased bubble rise velocities resulting from the use of high
velocity fluid flows. It is also possible that an insufficient recirculation liquid velocities were
delivered to the static mixer as the superficial air velocity was increased.
27 |
Virginia Tech | Although an in-line cavitation tube was not evaluated in this test effort, other
hydrodynamic cavitation based aerators were assessed, such as the eductor and two phase ejector
spargers. Each of these cavitation sparging technologies supplied an average bubble diameter
and air holdup equivalent to those yielded by a single phase porous sparger. Typically, external
spargers provide superior air holdup, fine bubble diameters, and increased bubble-particle
contacting necessary to optimize flotation performance, but operational costs are great due to the
utilization of significant compressed air volumes. To further support the advantages of both static
mixer and cavitation tube technologies, the University of Kentucky conducted a phosphate
flotation review using multiple sparger types. As reviewed by the University of Kentucky, the
use of a static mixer or cavitation tube to decrease particle detachment increases the recovery of
a 16 x 35 mesh phosphate ore considerably (Tao & Honaker, 2006). However, using a single
phase porous sparger, phosphate recovery remained significantly lower, as apparent in Figure 13.
Figure 13. Effect of Implemented Sparger Types on Phosphate Recovery (Tao & Honaker, Development of
Picobubble Flotation for Enhanced Recovery of Coarse Phosphate Particles, 2006), Used under fair use, 2014
28 |
Virginia Tech | 3.0 EXPERIMENTAL
Seeing an opportunity to reduce operational costs associated with column flotation
sparging systems, the Eriez Flotation Division designed an innovative low pressure drop sparger
capable of resisting plugging and degradation. In design planning of an in-line, non-plugging,
and porous sparger, it was determined that magnetic material be employed as a medium through
which air be dispersed and introduced to a moving slurry. Such a sparger would maintain an
inherently low pressure drop due to the use of a slurry flow path of a uniform cross sectional area
and the absence of any in-line mixing components. Alternatively, such a porous sparger would
rely upon the dispersion of air through more narrow paths, allowing for the introduction of fine
bubble streams from a magnetic media bed. Upon entrance to the flow path, bubble streams
would undergo further shearing as a result of high velocity liquid flows. Magnetic material was
chosen for the fact it can be manipulated using changes in magnetic fields or movement of host
magnets to clean the material during operation. In addition, magnetic material can be structured
without need for screens or mesh material to hold the media in place; therefore eliminating the
possibility of plugging of apertures through which air must travel. Using the theoretical concept
of directing air through a porous magnetic media for the purpose of aerating a recycled slurry
stream, a lab-scale magnetic sparger was designed, constructed, and tested at the Plantation Road
Facilities in the Mining and Minerals Engineering Department at the Virginia Polytechnic and
State University. Although the full scope of the project is to ensure the magnetic material can be
cleaned, the first stage of the project was focused on designing of the sparger itself and testing
the concept of aerating a slurry through in-line injection of air through a porous magnetic media
bed. Before flotation testing progressed, sparger aeration performance was first evaluated under
air, water, and frother only conditions.
30 |
Virginia Tech | 3.1 Sparger Development
A preliminary prototype was first constructed using a ring magnet assembly, as shown in
Figure 14. This approach was first taken to stem from previous research completed by the Eriez
Flotation Division, which consisted of aerating a moving coal slurry by the injection of air
through the crevices of multiple βin-lineβ flexible discs. Each disc was identically milled or
grooved to promote undeviating aeration of the coal slurry. Although such a novel sparging
device yielded both a fine bubble dispersion and high fine particle recovery, grooved air paths
began to plug within hours of being subjected to a 15% solids coal slurry. Aperture plugging
resulted in a decrease in recovery and air distribution. Using this aeration concept, it was
believed that ceramic magnets, protected or layered by a magnetic medium, could be used to
alter the previously tested sparger design to prevent the plugging of crevices and create a filter
through which air must travel.
Figure 14. Conceptual Diagram of In-Line Magnetic Ring Sparger
The constructed prototype, inclusive of an inner perforated pipe to introduce air from
behind a surrounding group of ceramic ring magnets, is shown in Figure 15. Both ends of the
sparger were tapered to allow moving slurry to more evenly flow past the surface of the sparger
for uniform aeration. Rubber gaskets were employed at either end of the ring magnet collection,
simulating a spring, to allow for expansion of gaps between magnets to support air flow. For the
purpose of this effort, sparger air distribution was observed within a clear water tank. Analogous
to the performance of the disc sparger, the ring magnet sparger produced a predominantly fine
31 |
Virginia Tech | bubble flux, but air distribution was uneven and large bubbles were frequently emitted from the
sparger surface at confined locations due to inconsistent magnet contruction and magnetic grit.
During limited static water testing, ceramic magnets also became worn quickly, further
hampering aeration uniformity.
Figure 15. In-Line Magnetic Ring Sparger
Due to the rapid deterioration of ceramic magnets when introduced to water or high
velocity slurries, an improved magnetic sparger was designed to ensure that primary magnetic
elements be kept external to slurry or liquid flows in a dry environment. Preservation of magnets
external to slurry flow also allows for easier manipulation of magnetic fields or the magnetic
medium itself for cleaning purposes. To protect ceramic magnets, improve air distribution, and
simplify operation, an external in-line sparging flow box was constructed, as shown in Figure 16.
To aerate a recycled slurry or liquid using the designed magnetic sparger, a magnetic medium is
held perpendicular to liquid or slurry flows by use of external ceramic magnets as incoming air is
dispersed through the medium to dynamically aerate a liquid or solids-liquid mixture. A plexi-
glass side housing was used on either side of the sparger to observe the aeration process
throughout the testing effort.
32 |
Virginia Tech | Figure 16. MagAir Sparger Prototype Design
The external magnetic sparger was designed with completely alterable parameters to
better understand their effect on sparging performance due to a lack of research or pre-existing
information regarding design of such an aerator. These parameters included cross sectional area
of the liquid or slurry flow region, magnetic medium bed depth, and magnetic bed area. The
sparger, as presented in Figure 17, was assembled with a two inch by two inch cross sectional
area and a length of 28 inches. At the spargerβs inlet, a 1 ΒΌ inch pipe nipple and union were
utilized to feed the sparger and easily remove it from the testing set-up, respectively.
Additionally, a two inch by two inch square pipe was affixed to the outlet end of the sparger in
replacement of a succeeding pipe nipple. This allowed a constant cross sectional area to be
maintained following the introduction of air to prevent the coalescence of air prior to departure
of the aerated liquid from the sparger. A one inch interchangeable spacer, as shown in Figure 17
(left), was also fabricated in order to decrease the cross sectional area of flow to increase the
liquid velocity across the magnetic media bed to overcome pumping limitations.
33 |
Virginia Tech | Figure 17. External Magnetic Flow-Box Sparger Design
Within the air inlet chamber perpendicular to the slurry flow region, an adjustable
perforated pedestal and slots of fixed height increments were engineered to allow for preparation
of a magnetic bed of desired depth. The air chamber was constructed 2 Β½ inches in length and
assembled equal in width to the flow region (2β) to ensure complete and even air distribution
across the recirculated slurry or liquid. The perimeter of the perforated pedestal was wrapped in
an aluminum tape and caulked to promote air distribution and prevent the short-circuiting of air
around the media bed along the chamber walls. Magnetic material was contained by the air inlet
compartment and held flush with the liquid flow region boundary by use of two large ceramic
magnets on either side of the air chamber, external to the sparging device. Due to the generation
of a strong magnetic field, the aluminum pedestal was magnetized and therefore created a level
foundation for the media bed to compact and rest upon. Opposite the air pedestal, a threaded
magnetic feed hole was created to allow magnetic material to be fed onto the perforated pedestal
without needing to disassemble the sparging apparatus. This permitted easier transitions between
testing of different media types during aeration performance testing.
34 |
Virginia Tech | 3.2 Equipment Setup and Test Work
3.2.1 Gas Holdup Testing
To verify the magnetic spargerβs aeration capabilities and optimal operating parameters, a
test assembly was developed to quantify the air holdup produced by the sparger under various
equipment arrangements and functional conditions. In air holdup testing, water and an MIBC
frother were strictly utilized. Mineral slurry or pulp was not employed for these exercises to
allow for complete visualization of the process within the sparger and the succeeding de-aeration
liquid holding cell. Dynamic external spargers are often operated with a combined air and liquid
velocity of 17 to 20 ft/s, therefore a variable speed centrifugal pump was instrumented to
produce an array of water flows. Optimal arrangement and installation of the sparger was first
defined using a simple recycle system consisting of a 20 gallon sump and centrifugal pump. Both
horizontal and vertical sparger orientations were implemented to understand the effect of sparger
positioning on air pressure requirements and liquid-air mixing, as shown in Figure 18. Horizontal
orientation of the sparger was quickly neglected as air naturally rose and coalesced at the ceiling
of the sparger upon introduction through the magnetic medium. A bottom fed vertical orientation
promoted the mixture of air and water and decreased the coalescence of air, but increased the air
pressure required for aerator operation. Such an increase in air pressure was attributed to an
increase in water pressure which acted in opposition to the incoming air. To solve this issue, a
top fed vertical orientation was chosen. Using this method, air was more naturally drawn into the
liquid flow zone due to the downward flow of a high velocity liquid, significantly decreasing the
necessary air pressure by a factor of three to four. Following the selection of an optimal sparger
orientation necessary for the aeration of a recycled liquid or slurry, a more complex test
35 |
Virginia Tech | A de-aeration tank was applied to the test circuit to increase fluid retention time, increase
total fluid capacity, and allow water to de-aerate before recirculation to a feed sump and
succeeding variable speed centrifugal pump. In addition, controlled liquid de-aeration in
conducted performance evaluations was necessary to properly quantify the air holdup produced
by the sparger and to prevent cavitation within the pump head. To determine air holdup, change
in liquid elevation within the tank was measured with respect to a fixed measurement port at a
known depth of the tank using several elevation rods or measurement tubes. This method of air
holdup measurement is illustrated in Figure 20. Using the difference in water elevations
determined at transparent measuring ports located just above the sparger outlet inside the tank
and below the overflow flume, and the known distance between each portβs centers, the
percentage of air holdup within the tank was calculated. To ensure accurate and precise
measurements were ascertained, level readings were taken at a collection of time intervals and
then averaged. To determine the effect of superficial air velocity and recirculated liquid velocity
on air holdup, a multi-facet air manifold and magnetic flow meter were employed, respectively.
Figure 20. Holding Cell Air Holdup Measurement Method
Compressed air was provided to the experimental system via an air manifold which
consisted of a regulator, variable area flow meter, and pressure gauge, as shown in Figure 19 and
37 |
Virginia Tech | Figure 21. Using air pressure and flow rate measurements, the actual air flow rate delivered to
the magnetic media bed in standard cubic feet per minute was calculated. Air was supplied to the
manifold at approximately 110 psi from a nearby air compressor. In addition to the employment
of an air manifold to measure supplied sparger air velocities, a magnetic flow meter was
calibrated and utilized to measure water flow rates for each test effort. The maximum water flow
produced by the centrifugal pump was approximately 90 gallons per minute. Once steady state
air and water flows to the de-aeration tank were obtained, air holdup measurements were taken
and recorded using a documented group of sparger operating conditions.
Figure 21. Magnetic Sparger Air Holdup Measurement System
Before air holdup measurements could be taken, sparging operational parameters were
chosen and employed. Operational parameters which were altered during testing included:
- Magnetic media type (magnetite, steel shot powder, spherical magnetic media),
- Magnetic medium bed depth,
- Air fraction (percentage of air in water within sparger),
- Cross sectional area of liquid flow region,
38 |
Virginia Tech | - Flow rate of water (gallons per minute), and frother dosage.
Due to the substantial quantity of operational parameters, an optimal magnetic media
type was chosen using visual observations. During preliminary evaluations it was concluded that
dense magnetite and fine steel shot mediums were incapable of providing a uniform air
distribution due to significant compaction of particles upon introduction to water, as depicted in
Figure 22. As a result, air quickly short circuited to the perimeter of the impenetrable bed when
such forms of media were implemented. Coarse magnetite, crushed in two stages to a size
fraction of 1000 x 500 micron, promoted air flow and consisted of particles diverse in shape and
size, but increased air distribution variability due to a lack of uniform porosity. Additionally,
magnetite and steel shot mediums became oxidized, or rusted, within 24 hours of water
submersion, further compacting magnetic particles and obstructing air paths.
Figure 22. 250 x 150 Micron Wet Magnetite Media Sample
To improve air dispersion and maintain a more uniform porous filter medium, a variety
of ferritic stainless steel magnetic spheres were employed. A stainless steel coating allows these
magnetic spheres to better resist oxidation following extensive subjection to liquid or solids. To
perform the testing effort, 1.0 and 1.6 mm magnetic beads, as shown in Figure 23, were utilized.
39 |
Virginia Tech | tailings sump. A variable speed peristaltic pump was used to pump feed material, flotation
concentrate, and recycled tailings slurry at a rate of two gallons per minute from the sump to the
external porous magnetic sparger. A rate of two gallons per minute permitted the necessary
mixing or residence time within the sump and was sustained to maintain a consistent pulp-froth
interface level. In addition to this mixture of slurry streams, a high velocity refuse slurry stream,
removed from the lowermost zone of the flotation cell by use of variable speed centrifugal pump,
was delivered to the recessed sparging unit from above at approximately 14 gallons per minute to
simulate a column flotation cell external sparging arrangement. Together, these slurry streams
provided the required downward constant traveling fluid velocity to the sparger air interface. The
aerated coal slurry was lastly injected at the cell bottom. Due to the scale of this test, a smaller
form of the magnetic sparger was fabricated using a hose barb inherent of an inlayed perforated
screen and two external ceramic magnets used to retain the magnetic media, as shown in Figure
26. Compressed air was delivered from behind the magnetic bed using a ΒΎβ air-line.
Figure 26. Flotation Test Magnetic Sparger Media Housing
42 |
Virginia Tech | Laboratory flotation testing was carried out at an external sparging slurry flow rate of
16.38 gallons per minute. This flow rate was chosen to solicit similar combined air and slurry
velocities implemented in previously conducted air holdup evaluations and industrial external
sparging applications. To establish a grade and recovery curve for a minus 150 micron coal
sample using the magnetic sparger prototype, the provided aeration rate, or air flow rate, was
varied from 0 to 2 cubic feet per minute (cfm) using a variable area rotameter affixed to the
assembly frame. As air rate was increased, combustible recovery also increased. Air rate was
limited to no greater than 2 cubic feet per minute during the test effort to sustain an air fraction,
or percentage of air within the aerated recycle stream, of less than 50%. Given a 0.45 in2 cross
sectional area of the recirculated tailings and feed line, an air flow rate greater than
approximately 2 cfm would have yielded an air fraction in excess of 50 percent and produced
excessive burping within the flotation cell. Such burping or air bubble coalescence increases the
misplacement of water to the froth launder and decreases bubble particle collision and
attachment. To identify the maximum air fraction allowable, gas flow rate was increased until
substantial disturbance of the froth was visualized as a result of the coalescence of air or burping.
In addition to the variance of aeration rate to control air holdup within the flotation cell,
Methylcyclohexanemethanol, or MCHM 8-carbon alcohol frothing agent, was supplied to the
feed sump at an addition rate of 10 parts per million (ppm) by volume to decrease the surface
tension of the slurry and promote the development of smaller air bubbles. Wash water was not
employed in this test program, but an insignificant flow of water was added in the froth launder
to provide mobility to the dry froth concentrate to promote drainage and circulation to the feed
sump. Minimal water flow was provided to the launder to help maintain the circuit feed percent
solids and prevent frother dilution. In an attempt to uphold a constant froth level, a manual gate
43 |
Virginia Tech | valve was located and controlled opposite the high velocity tailings outlet in substitution of a
loop controlled pressure transmitter and bladder valve combination. Slurry removed from this
valve was deposited in the feed sump below.
Following the designation of an aeration rate at a constant tailings recirculation and feed
pumping rate, the flotation cell was operated until a steady state process was achieved. Holding a
froth level within the cell was proven difficult due to a lack of visibility of the pulp-froth
interface given the short height of the cell, turbulence within the cell, and size of the launder, as
presented in Figure 27. Once steady state conditions were assumed, samples of the concentrate
and tailings were taken for equal time durations using two gallon buckets of documented
weights. This process was conducted at five aeration rates to determine the effect of gas flow rate
on combustible recovery and concentrate ash content using the magnetic sparger. Additional
operating points on the grade and recovery curve were identified without using magnetic
material in the sparger air inlet to verify the effect of the employment of a porous medium on
flotation performance.
Figure 27. Flotation of a Minus 150 Micron Coal using a Magnetic Sparging System
44 |
Virginia Tech | 4.0 RESULTS AND DISCUSSION
4.1 Air Holdup Performance
Residual gas holdup or aeration performance evaluations were conducted using two
MIBC frother concentrations, 12 and 18 ppm by volume. In most industrial flotation
applications, the employment of additional reagents, such as collectors and extenders, also alter
the surface tension of a liquid or pulp; therefore higher concentrations of MIBC frother were
utilized in this test effort in their absence. For each designated frother addition rate and media
size, 1.6 mm, 1.0 mm, or 0 mm, the gas holdup within the succeeding de-aeration holding cell
was measured as the sparger air rate was increased to determine the effect of aeration rate on air
holdup. Air velocity was varied from 3 to 15 ft/s for three recirculated liquid flow rates of 90, 81,
and 72 gallons per minute to also define the relationship between recycled liquid velocity and air
holdup. A liquid flow rate range of 72 to 90 gpm, or 11.5 to 14.5 ft/s, was chosen to produce
similar liquid velocities implemented in industrial external sparging applications. In addition, air
velocities were maintained between 3 and 15 ft/s to ensure a sparger air fraction range of 20 to
50%. Liquid and air velocities were calculated using the geometry of the main sparger flow
region, 2 in2. The maximum operating air pressure observed throughout all test programs was
approximately 6 psi at a combined air and water velocity of 28.6 ft/s. The presence of a magnetic
medium negligibly increased air pressure by up to 0.25 at the greatest air velocities. Using air
holdup measurements collected at increasing air velocities for each of three liquid velocities, air
velocity was plotted versus air holdup for each individual liquid flow rate.
Tabulated data and graphs depicting the effect of increases in air and liquid velocities on
air holdup during the employment of 1.6 mm magnetic beads, 1.0 mm magnetic beads, and
without magnetic media at a frother dosage of 18 ppm are shown in Figures 28, 29, and 30,
46 |
Virginia Tech | respectively. The effect of air velocity on air holdup at increasing liquid velocities when using 12
ppm of frother was also evaluated and is illustrated in Appendices A, B, and C. As evidenced by
the air velocity and gas holdup relationships depicted in Figures 28-30, the gas holdup produced
by the magnetic sparger increased proportionally to air velocity equally no matter the media type
or liquid flow rate. Given a minimal top to bottom ratio of liquid velocities utilized, 1 ΒΌ, the
effect of the recycled liquid rate on generated air holdup was difficult to identify.
Air Velocity (ft/s) vs. Residual Gas Holdup (%): 1.6 mm Beads -18 PPM
18
16
)%
14
(
p
u 12
d
lo
10 72 GPM
H
s a 8 81 GPM
G
la 6 90 GPM
u
d 4
is
e
R 2
0
0 2 4 6 8 10 12 14 16
Air Velocity (ft/s)
Figure 28. 1.6 mm Beads: Air Velocity (ft/s) vs. Residual Gas Holdup (%) β 18 ppm Frother
Air Velocity (ft/s) vs. Residual Gas Holdup (%): 1.0 mm Beads -18 PPM
20
18
)%
16
(
p u 14
d
lo 12
90 GPM
H
s 10
a 81 GPM
G 8
la
72 GPM
u 6
d
is 4
e
R
2
0
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00
Air Velocity (ft/s)
Figure 29. 1.0 mm Beads: Air Velocity (ft/s) vs. Residual Gas Holdup (%) β 18 ppm Frother
47 |
Virginia Tech | Air Velocity (ft/s) vs. Residual Gas Holdup (%): No Beads -18 PPM
18
16
)%
14
(
p
u 12
d
lo
H 10
s 72 GPM
a G 8
la 81 GPM
u 6
d 90 GPM
is
e 4
R
2
0
0 2 4 6 8 10 12 14 16
Air Velocity (ft/s)
Figure 30. No Beads: Air Velocity (ft/s) vs. Residual Gas Holdup (%) β 18 ppm Frother
For a given air addition flow rate, gas holdup did not significantly differ as liquid velocity
was increased due to the limited range of liquid velocities employed and an increase in bubble
rise velocity within the holding cell. As liquid velocity was increased, bubble rise velocity also
increased, therefore decreasing the measured gas holdup. If much lower liquid velocities were
assessed, it is possible that a difference in the yielded air holdup for a specified aeration rate
could be more easily distinguishable. To illustrate the effect of bubble rise velocity on gas
holdup, a population balance around the de-aeration tank was formulated to calculate bubble rise
velocity as liquid velocity was increased. The derivation for bubble rise velocity is shown in
equations 4 through 7 and is as follows:
ππππ’ππ πΈππ‘πππππ ππππ = ππππ’ππ πΈπ ππππππ ππππ [4]
(π +π )(π ) = (π +π + π’ π΄)(π ) [5]
πΊ πΏ π πΊ πΏ π π
π =
(ππΊ+ππΏ)
(π ) [6]
π π
(ππΊ+ππΏ+ π’ ππ΄)
π’ =
(ππΊ+ (ππ ππΏ ))(ππ) β(ππΊ+ππΏ)
[7]
π
π΄
48 |
Virginia Tech | where Q is the gas flow rate, Q is the liquid flow rate, π is the air fraction of the introduced or
G L π
incoming aerated liquid, π is the air holdup within the holding cell, A is the cross sectional area
π
of the tank, and u is the bubble rise velocity. As shown in equation 6, air holdup is dependent
b
upon air, liquid, and bubble rise velocities. As air flow rate is increased, air holdup increases,
whereas air holdup decreases as liquid and bubble rise velocities increase. This is because
retention time of air within a tank or de-aeration cell decreases as the introduced liquid flow rate
is increased. Using the derived equation for bubble rise velocity, air rise velocity was calculated
for each completed test, as depicted in data summary tables in the attached Appendices.
To demonstrate that smaller air bubbles were generated at greater recycled liquid
velocities due to an increase in bubble shearing, bubble rise velocities were analyzed for an
equivalent introduced air fraction at increasing liquid velocities. Firstly, air fraction was plotted
versus air holdup at increasing liquid flow rates, as shown in the Appendix for each testing
effort. Although these relationships indicate that air holdup was nearly equal for a given air
fraction at increasing liquid flows, bubble rise velocity greatly increased as liquid flow rates
increased. For example, at an air fraction of 40%, the air holdup yielded by the magnetic sparger
during the employment of a 1.6 mm bead porous medium was nearly 14% at variable liquid
velocities. As liquid flow rate was increased from 72 to 90 gpm, bubble rise velocity increased
by 1.25 ft/min. In conclusion, lower air bubble rise velocities present at decreased liquid flow
rates yielded air holdup values equivalent to those provided at increased liquid flow rates where
the average bubble diameter was smaller. At increased liquid velocities, air arose more quickly
within the holding cell, but a greater air shear rate at the liquid to air interface generated bubbles
much smaller in size.
49 |
Virginia Tech | Although it is difficult to differentiate the aeration performance of the sparger while
operated at varying liquid flow rates, the air holdup provided by the magnetic sparger at
combined air and liquid velocities employed in industrial external sparging applications is quite
distinguishable. A summary of air holdup values produced at increasing combined liquid and gas
velocities and frother addition rates for each testing effort is provided in Tables I, II, and III in
Appendices A, B, and C. For example, using a 1.6 mm spherical media diameter, the air holdup
within the holding cell increased directly proportional to the combined air and water velocity.
Using a frother addition rate of 18 ppm, and a combined air and water velocity of approximately
18 feet per second, the air holdup measured within the tank was 12.75 percent. Such an operating
point is similar to that employed in an industrial setting. As liquid flow rate was increased using
a 1.6 mm magnetic spherical media, the air holdup realized at the burping point also increased to
as great as 14.85 percent. Although gas holdup increases as liquid velocities increase, an
understanding of pumping economics is necessary to determine an economical operating point of
the sparger.
In addition to the understanding of sparger performance under varying operating
conditions while employing a single porous media type, comparison of performance using
differing media yielded a recognizable difference in the sparger generated gas holdup, as
represented in Figures 31 through 33. To distinguish the difference in performance of the sparger
during the employment of each media type or lack thereof a porous medium, introduced air
fraction was plotted versus residual gas holdup for each liquid flow rate utilized. As shown in
Figure 31, the use of a 1.6 or 1.0 mm magnetic bead porous medium increased gas holdup by
almost two percent at a water rate of 72 gallons per minute. To show the effect of liquid velocity
or flow rate on this difference, similar data was plotted for liquid flow rates of 81 and 90 gallons
50 |
Virginia Tech | per minute, as shown in Figures 32 and 33, respectively. As liquid flow rate was increased, the
difference in air holdup produced with or without magnetic media was less distinguishable, but
still evident. Using 81 gallons per minute of water and a bead size of 1.6 mm, air holdup
remained almost two percent greater than without the implementation of magnetic media. In
contrast, the difference in air holdup reduced to one percent at combined air and water velocities
greater than 25 feet per second as realized when using a liquid flow rate of 90 gallons per
minute, as shown in Figure 33. This is most likely due to an increase in the shearing effect at the
air and liquid interface.
In a comparison of the air holdup performance yielded by each magnetic medium, the 1.6
mm magnetic bead medium continuously produced slightly greater air holdup values than the 1.0
mm bead medium. Because both media types were similar in size, it is difficult to classify the
effect of media size on aeration performance. However, preliminary selection of a porous
medium did evidence that the use of a uniformly sized porous media is beneficial in the
production of a more even air distribution and minimization of short circuited air flows. From
optical inspection, the larger media size yielded an improved average air holdup due to an
increase in the coalescence of air following its departure from the magnetic bed when using a
reduced size of bead. When using a smaller spherical media type, the porosity within the bed
remains equal, approximately 26 percent according to a face centered cube lattice, but ejected air
streams are located more closely which inherently heightens the risk of coalescence. It is
possible that an increase in magnetic bead size could improve air distribution, but bubble size
would be expected to increase as well.
51 |
Virginia Tech | Introduced Air Fraction (%) vs. Residual Gas Holdup (%) -90 GPM
(18 PPM)
20
18
16
)%
14
(
p
u 12
d
lo
H 10 No Beads
s
a G 8 1 mm
la
u 6 1.6 mm
d
is
e 4
R
2
0
20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0
Introduced Air Fraction (%)
Figure 33. Introduced Air Fraction (%) vs. Residual Gas Holdup (%) β 90 GPM (18 PPM)
Air holdup evaluations demonstrate that an external porous magnetic sparger using this
design concept is a capable and economical sparging device in terms of air dispersion and
distribution, and air pressure requirements. According to gathered results, air holdup percentages
generated by the magnetic sparger using each media type are quite comparable to those provided
by currently employed sparging technologies, but at much lower air pressures. As shown in
Figure 34, the air pressure required to operate the sparger during the employment of a 1.6 mm
magnetic bead porous medium did not surpass 6 psi at a combined air and water velocity greater
than 30 ft/s. At this air pressure, a low horsepower blower can be utilized to supply necessary air
flow rates. A similar trend in the relationship between the combined velocity of air and water and
air pressure was realized in all testing efforts. Necessary air pressures at increasing combined air
and water velocities during the implementation of 1.0 mm magnetic beads and no magnetic
beads are further detailed in the Appendices B and C.
53 |
Virginia Tech | 1.6 mm Beads -Combined Velocity [Water+Air] (ft/s) vs. Air Pressure (psi)
100
)
is
p
(
e
r
u 10
s
s
e
r y = 0.0074x2.1057
P
r
iA
1
1.000 10.000 100.000
Combined Velocity: Air + Water (ft/s)
Figure 34. 1.6 mm Beads: Combined Velocity [Water + Air] (ft/s) vs. Air Pressure (psi)
Due to a reliance upon low pressure air, the dimensions of the magnetic bed and the fluid
flow region must be properly engineered to minimize required air pressure and promote air
distribution. From optical observations of air flow through each tested magnetic medium, the
depth and geometrical dimensions of the bed strongly governed the distribution of air at the air
inlet. Using a deeper porous bed, air more easily short circuited to a confined location of the bed,
thus a minimum bed depth of one inch was utilized. During industrial operation of the sparger,
perfect compaction and distribution of material retained at the air inlet will more likely be
difficult when a deep bed is maintained. Air pathways increase in size throughout the bed as
bead size increases, therefore an increase in bed depth is less detrimental to air distribution. An
increase in the width of the bed also increased the short circuiting of air. This was evidenced by
the collection of air flow on the low pressure outlet side of the bed at increased fluid pressures.
Naturally, air will flow to the path of least resistance where acting pressure is lowest. For the
construction of an industrial unit prototype, air inlets should be fabricated with a small cross
sectional area of approximately 1.0 square inch or less and a width to depth ratio of
54 |
Virginia Tech | approximately 1.5 to 1.0 to prevent the short circuiting of air. Using multiple air inlets in series
as described across the flow region would increase the complete air distribution from the bed and
ensure proper aeration of a moving slurry. To prevent the short circuiting of air to the perimeter
of the bed, the surfaces of surrounding air inlet walls should be textured roughened to establish
similar friction present between individual magnetic beads. The height and hydraulic diameter of
the fluid flow region should also be minimized to increase the mixing of air and slurry and to
limit the water pressure internal to the sparger to decrease air pressure requirements,
respectively.
4.2 Flotation Performance
In addition to the completion of air holdup and aeration performance evaluations,
flotation scoping tests were conducted to demonstrate the spargerβs ability to yield high mineral
recoveries of a finely sized coal sample. Testing was conducted with and without the use of a
magnetic porous medium to characterize the benefit of using such material. Steady-state flotation
tests were performed at a fixed feed rate as aeration rate was increased from approximately 0.5 to
2.0 scfm to define the relationship between sparger air fraction and mineral recovery. A fixed
slurry feed rate of 16.38 gpm, or approximately 12 ft/s, was sustained to reproduce recirculated
liquid rates utilized in air holdup evaluations. Once steady-state conditions were established for a
given air velocity, samples of the froth concentrate and tailings were procured. An ash analysis
was then conducted on each sample to determine their respective ash contents. Ash content
values were then employed to calculate the combustible recovery realized for each assumed air
flow rate and pressure.
The combustible recoveries and concentrate ash contents achieved at variable air addition
rates, with and without the employment of a porous magnetic medium, are displayed in Tables I
55 |
Virginia Tech | and II, respectively. Using the mass yield to the concentrate and ash contents of both the tailings
and concentrate fractions, a feed ash content was also back calculated for each test. As evidenced
by the flotation test data, combustible recovery improved by up to 20% as aeration rate was
increased to an air fraction of nearly 50% by volume. As expected, an increase in air flow rate
also increased the average concentrate ash content due to an increase in the hydraulic
entrainment of high ash slimes or fines. Given the small volume of the employed flotation cell
and limited particle residence time, combustible recoveries yielded during the employment of a
porous medium can be considered favorable.
Table I β Sparger Flotation Results Without use of Magnetic Beads
No Beads - Sparger Flotation Results
True Slurry Air
Pressure Feed Ash Tail Ash Con Ash Combustible
Air Flow Fraction
(psi) (%) (%) (%) Recovery (%)
(scfm) (gpm) (%)
0.51 0.75 16.38 18.97 15.86 21.41 5.66 39.54
1.03 1.00 16.38 32.06 20.02 24.65 9.13 33.92
1.30 1.25 16.38 37.28 18.52 25.77 8.98 48.22
1.57 1.50 16.38 41.83 16.93 25.30 8.04 53.66
2.13 2.00 16.38 49.32 16.71 27.74 9.64 66.12
Table II β Sparger Flotation Results with use of Magnetic Beads
1 mm Beads - Sparger Testing
Slurry Air
Air Pressure True Air Flow Fraction
(scfm) (psi) (scfm) (gpm) (%) Feed Ash Tail Ash Con Ash
0.50 0.75 0.51 16.38 18.97 12.32 21.44 7.62
1.00 1.00 1.03 16.38 32.06 16.38 25.24 9.32
1.50 1.50 1.57 16.38 41.83 16.11 26.29 13.60
2.00 2.00 2.13 16.38 49.32 17.21 34.98 11.86
In addition to the effect of increased aeration rate on recovery, the employment of a 1.0
mm magnetic bead porous medium also greatly improved combustible recovery. Using
equivalent air flow rates in both test programs, the utilization of a porous magnetic bed increased
combustible recovery by 15 to 25%, as illustrated by the air rate versus combustible recovery
56 |
Virginia Tech | plot illustrated in Figure 35. For example, at an air addition rate of 2.13 scfm, the use of a porous
medium increased combustible recovery by 15%. Such an increase in recovery resulted from the
production of a finer air bubble distribution, which improved the likelihood of collision and
attachment between air bubbles and fine particles. Due to an increase in combustible recovery
using a porous medium, yielded concentrates were of a much higher ash content. If a taller
column cell had been utilized to increase particle retention time, while also boasting the use of a
deep froth and wash water, a more satisfactory concentrate grade and combustible recovery
could have been achieved.
Air Flow (scfm) vs. Combustible Recovery (%) with and
without a Porous Medium
90.00
80.00
)%
70.00
(
y
r e 60.00
v
o
c 50.00
e
R
e 40.00 No Beads
lb
it s 30.00 1.0 mm Beads
u
b 20.00
m
o
C 10.00
0.00
0.00 0.50 1.00 1.50 2.00 2.50
Air (scfm)
Figure 35. Air Flow (scfm) vs. Combustible Recovery (%) with and without a Porous Medium
Although an improvement in flotation performance was observed during the employment
of a porous medium, inconsistencies in calculated combustible recoveries and measured ash
contents are apparent. Most importantly, the calculated feed ash did not remain continuous
throughout all test work. This was most likely due to inefficient steady state sampling resulting
from difficulties experienced in upholding the froth level within the flotation cell by use of a gate
57 |
Virginia Tech | valve at lower aeration rates. Although material was frequently removed from the system, a
proper test should sustain a more fixed feed ash content. Due to the size of the sparging
prototype, the maximum allowable air addition rate also hindered the development of a more
stable froth at lower sparger air fractions. In column flotation applications, approximately 4 scfm
of air per square foot is required to aerate the cell. In this study, a maximum of 2.13 scfm was
allowable before burping commenced. To appropriately aerate the flotation cell utilized, a one
inch recirculation pipe should have been exercised. Such an increase in sparger size would have
increased the permissible air addition rate at both low and high air fractions. Although
experimental faults were realized, this study demonstrates the value of a porous medium for its
ability to increase recovery as a result of an increase in air holdup. To more fully understand the
sparging capabilities of the magnetic porous sparger, a laboratory column cell, incorporating
wash water and a loop controlled pressure transmitter to maintain a froth level, must be
employed.
58 |
Virginia Tech | 5.0 CONCLUSIONS
The engineered and performance evaluated low pressure porous sparger demonstrated
itself as a viable sparging prototype in terms of bubble generation and distribution, and flotation
performance. The air holdup provided by the sparger, ranging from 12 to 16 percent, was similar
to that generated by current industrially employed external spargers, such as the static mixer, at
comparable combined air and liquid velocities. Due to the presence of a negligible pressure drop
at the air-to-liquid interface, the magnetic porous sparger was also capable of yielding such air
holdup values at much lower air pressures of 5 to 7 psi. As a result of low operating air
pressures, this sparger is operable by means of a low horsepower blower, reducing operational
costs significantly.
Following the completion of numerous sparging assessments, practical sparger operating
parameters, such as a proper porous medium, correct sizing of the chosen media, necessary
combined velocities of air and water, orientation of the sparger, and required air pressure were
determined. Because air travels to the path of least resistance, a round magnetic media with a
bulk uniform porosity is necessary to avoid the short circuiting of air streams to the perimeter of
the porous medium or confined locations of the magnetic bed. To provide complete dispersion of
air through the magnetic medium, air inlets with a diameter of 1.5 inches or less, and a bed width
to depth ratio of 1.5 to 1.0 are also recommended. Subsequently, a top-fed, vertical orientation of
the sparger was deemed optimal to more naturally draw air into the sparger at lower air
pressures.
In addition to the identification of ideal sparger operational parameters, conducted
aeration evaluations also confirmed the improvement of the aeration of a recycled liquid or slurry
by use of a porous magnetic medium. During the employment of a porous magnetic medium in
both air holdup and flotation performance test efforts, an increase in air holdup of up to 2 to 3%
59 |
Virginia Tech | DESIG(cid:6) A(cid:6)D TESTI(cid:6)G OF A HYPERBARIC HORIZO(cid:6)TAL
BELT FILTER FOR FI(cid:6)E COAL DEWATERI(cid:6)G
by
Jeffery Salomon
Committee Chairman: Gerald H. Luttrell
Department of Mining and Minerals Engineering
ABSTRACT
This objective of this project was to develop a new dewatering device that could produce
a lower moisture content and better fine particle recovery than current technology. To meet this
goal, a hyperbaric horizontal belt filter was designed and constructed over the course of 18
months. Once built, the filter was then thoroughly tested to determine operational capabilities.
The test data showed that the lowest moisture content that could be achieved with a coarse feed
(minus 1 mm screen-bowl centrifuge feed) was 8.8%. This value could be further reduced to
8.2% and capacity increased with the use of dewatering aids. When testing with a fine feed
(minus 0.15 mm column product feed), the lowest moisture content was 35% without chemicals
and 29% with chemicals. A 50/50 mixture by volume of coarse and fine feeds was artificially
created and provided a moisture of 10.8%, which was reduced using reagents to 8.4%. The
machine provided a very high recovery rate for all feed materials. Of the coal input, no less than
94% of it reported to the dry product. The pressure used to dewater the coal was the controlling
factor for the air consumption of the unit. The data from these tests suggest that a full size
production unit is feasible, although the power requirements for gas compression would be high.
ii |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.