This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article （e.g. in Word or Tex form） to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Fatty acid collectors for phosphate flotation and their adsorption behavior
using QCM-D
J. Kou a,b, D. Tao b,?, G. Xub
a School of Civil and Environment Engineering, University of Science and Technology Beijing, 30 Xueyuan Road, Haidian District, Beijing, 100083, PR China
b Department of Mining Engineering, University of Kentucky, Lexington, KY 40506, USA
a r t i c l e i n f o a b s t r a c t
Article history:
Received 31 July 2009
Received in revised form 5 March 2010
Accepted 11 March 2010
Available online 25 March 2010
Keywords:
Fatty acid
FTIR
Hydroxyapatite
Phosphate flotation
QCM-D
Zeta potential
In this paper the relationship between the flotation performance of phosphate collectors and their adsorption
behavior was evaluated using a variety of techniques including the Crystal Microbalance with Dissipation
technique （QCM-D）. The adsorption of the collectors on the surface of hydroxyapatite was primarily
characterized using QCM-D, which is a high sensitivity in-situ surface characterization technique. Additionally,
the collectors were evaluated via zeta potential and FTIR analyses. The flotation performance of the collectors
was evaluated using a laboratory mechanical flotation cell at different process parameters such as pH, collector
dosage, diesel dosage and flotation time. The two collectors evaluated were a commercial plant collector and a
refined tall oil fatty acid. The QCM-D data showed that the refined tall oil fatty acid adsorbed on phosphate
more readily and produced stronger hydrophobicity and better flotation performance than the plant collector.
The chemisorption and surface precipitation mechanisms of the refined tall oil fatty acid on the surface of
hydroxyapatite were demonstrated by means of zeta potential measurements and FTIR analysis.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
In the conventional phosphate flotation （Crago） process, a
significant amount of the silica present in the feed is floated twice,
first by fatty acid, and then by amine （Zhang et al., 1997）. The Crago
process is therefore inefficient in terms of collector efficiency. The
phosphate mining industry is faced with higher fatty acid prices, lower
feed grade, and stricter environmental regulations （Sis and Chander,
2003）. To meet the market demand for higher effectivity, lower cost
and better selectivity of phosphate flotation collectors, there is a need
to evaluate surface adsorption techniques that may help researchers
develop better collectors by understanding how the adsorption
behavior of materials affects their performance as flotation collectors.
In order to evaluate this relationship, a plant collector of proprietary
composition and a refined tall oil fatty acid were compared. The
refined tall oil fatty acid, referred to as GP193G75, was comprised of
47% oleic and 33% linoleic acids.1 Flotation tests were performed at
varying process parameters such as pH, collector dosage and flotation
time with phosphate ore from CF Industries' phosphate rock mine in
Hardee County, Florida. To better understand the behavior of the
collectors on an apatite surface, their adsorption on the surface of a
hydroxyapatite-coated sensor was studied using the QCM-D technique.
The adsorption and flotation characteristics of the two
collectors were then compared.
Most of the studies about the adsorption mechanism of collectors
on mineral surface were conducted based on ex-situ measurements
such as contact angle, adsorption isotherm, FTIR spectroscopy, and
zeta potential, which unfortunay cannot monitor the real-time
formation process and characteristics of adsorbed layer. QCM-D is the
second generation of QCM, which has been shown by many
investigators to be a sensitive tool for studying the behavior of protein
and surfactant adsorption in aqueous solutions, with sensitivity in the
ng/cm2 （submonolayer） region （Hook et al., 1998）. It can simultaneously
determine changes in frequency and energy dissipation of a
quartz crystal at nanoscale in real-time and derives valuable in-situ
information on adsorbed mass as well as the mechanical （viscoelastic）/
structural properties of the adsorbed layer from experimentally
obtained data of energy dissipation in relation to frequency shift （Paul
et al., 2008）. The purpose of this study was to investigate in-situ the
adsorption behavior of two collectors on the hydroxyapatite surface by
means of QCM-D technique and to determine whether the differences
observed may lead to differences in flotation performance.
2. Experimental
2.1. Materials
The phosphate from the CF Industries' phosphate rock mine was a
mixture of apatite with gangue minerals such as quartz and clay
International Journal of Mineral Processing 95 （2010） 1–9
? Corresponding author. .: +1 859 257 2953; +1 859 323 1962.
address: dtao@engr.uky.edu （D. Tao）.
1 The refined tall oil fatty acid was supplied by Georgia-Pacific Chemicals, LLC under
the name GP 193G75.
0301-7516/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.minpro.2010.03.001
Contents lists available at ScienceDirect
International Journal of Mineral Processing
journal homepage: www.elsevier.com/ locate/ijminpro
Author's personal copy
minerals. The moisture of the sample was 20.88%. Wet sieving was
conducted with the as-received ore and the particle size distribution
for this particular phosphate sample is shown in Table 1. As shown,
the sample contained only 5.47% +30 mesh materials and 0.49%
smaller than 200 mesh particles. The dominant size fractions were
−30+50 mesh and −50+70 mesh fractions which accounted for
29.06% and 51.36% of the total sample, respectively. Each size fraction
of the feed sample was assayed for its grade （P2O5%） and acid
insoluble （A.I.） content （see Table 1 results）. The overall feed grade
was 7.54% P2O5% and 65.05% A.I.
The flotation collector blend was prepared using the collectors as
supplied and Chevron #2 diesel fuel. Soda ash in the form of 15% water
solution was used for pH adjustment of ore slurry.
The sensor used in QCM-D analysis was an AT-cut quartz disc with
5 nm Cr, 100 nm Au, 50 nm Ti and 10 nm hydroxyapatite that were
sputter-coated onto the crystal surface successively. The sensors and
Q-sense E4 system were supplied by Q-sense Co.
Zeta potential and FTIR tests were conducted with hydroxyapatite
powder （50% less than 16.93 μm, 90% less than 85.90 μm） made of
pulverized hydroxyapatite crystals purchased from Ward's Natural
Science, and the XRD results did not show any other impurity in the
samples.
2.2. Flotation tests
Flotation tests were conducted using a Denver D-12 lab flotation
machine equipped with a 2-liter tank and a 2–7/8 in. diameter
impeller. The slurry was first conditioned with soda ash solution for
pH adjustment. It was then conditioned in a bucket at about 70% solids
concentration after the addition of collector and fuel oil at certain
dosages. After that, the conditioned slurry was transferred to the 2-
liter flotation cell and water was added to dilute it to 25% solids by
weight. Unless otherwise specified, flotation lasted for 1–2 min. Tap
water was used in all flotation tests.
To evaluate the collector's performance and optimize the process
parameters, flotation tests were carried out at different pHs, collector
dosages, diesel percentage and flotation times （kinetic tests）. The
flotation products were assayed for grade （P2O5%） and acid insolubles
（A.I.）. Flotation recoveries were calculated based on the analyses of
the separation products （concentrates and tailings）. Flotation efficiency
is a composite parameter used for evaluating the flotation
performance. It was calculated from recovery and A.I. rejection in the
equation
Flotation efficiency = P2O5 recovery + A:I: rejection−100 ð1Þ
A.I. rejection can be calculated from Eq. （2）
A:I: rejection =
Tt′
Cc′ + Tt′ 100 ð2Þ
in which t′ and c′ are A.I. of tailing and concentrate, respectively.
2.3. QCM-D analyses
The QCM-D experiments were conducted at pH 10.0 and 25 °C
（±0.02 °C）. The stock solution was prepared by dissolving the
appropriate amount of collector with sodium hydroxide in deionized
water. To ensure dissolution and degassing, the solutions were left in
an ultrasonic bath for 5–10 min. For each experiment, the data
generated with the solvent only （sodium hydroxide solution in this
study） were accepted as baseline when the process became stable.
The fatty acid solutions were injected into the measurement system
by a chemical feeding pump capable of precise flow rate control. The
flow rate in the experiment was kept at 0.5 mL/min.
Software QTools 3.0 was used for data modeling and analysis. For a
rigid, thin, uniform film, the change in dissipation factor ΔD is smaller
than 10−6 for a 10 Hz frequency change （Paul et al., 2008）. The
Sauerbrey equation （Eq. （3）） can be used for mass determination.
Δm = −
ρqtqΔf
f0n
= −
ρqvqΔf
2f 2
0 n
= −CΔf
n
ð3Þ
where ρq and tq are the density and thickness of quartz crystal,
respectively, and vq is the transverse wave velocity in quartz. The
constant C has a value of 17.8 ng cm−2 Hz−1, and n is the harmonic
number （when n=1, f0=5 MHz）.
If the adsorbed film is “soft” （viscoelastic）, it will not fully couple to
the oscillation of the crystal （Ekholm et al., 2002）, and this will cause
energy dissipation of system. The dissipation factor D is proportional
to the power dissipation in the oscillatory system （Eq. （4）） and can
give valuable information about the rigidity of the adsorbed film
（Ekholm et al., 2002）:
D =
Edissipated
2πEstored
ð4Þ
where Edissipated is the energy dissipated during one oscillation, and
Estored is the energy stored in the oscillating system （Ekholm et al.,
2002）.
When adsorption causes a great shift in the D value （ΔD N 1 × 10−6）
as a result of the adsorption of a viscous and soft layer, Voigt modeling
（Eqs. （5） and （6）） （Voinova et al., 1999） can be used:
Δf ≈− 1
2πρ0h0
η3
δ3
+ Σ
j=1;2
hjρjω−2hj
η3
δ3
2 ηjω2
μ2
j + η2j
ω2
（ " #）
ð5Þ
ΔD≈− 1
2πf ρ0h0
η3
δ3
+ Σ
j=1;2
2hj
η3
δ3
2 μjω2
μ2
j + η2j
ω2
（ " #）
: ð6Þ
According to the Voigt model for viscous adsorption layer, Δf and
ΔD depend on the density （ρ）, thickness （h）, elastic shear modulus （μ）
and shear viscosity （η） of the adsorption layer, and j is the number of
adsorbed layers. The Sauerbrey equation and the Voigt model are the
theoretical basis for data modeling in the QCM-D analysis.
2.4. Zeta potential measurements
The zeta potential measurements were made with Zeta-plus
analyzer of Brook Haven Instruments Corporation. All experiments
were conducted with 1 mM KCl solution at laboratory atmosphere
and temperature. 1.0 g hydroxyapatite powder was first conditioned
in 50 mL of 1 mM KCl solution with magnetic stirrer for 1 h, during
which 1.25 mg GP139G75 was added into solution to make the
concentration of 25 ppm, and pH was adjusted by NaOH or HCl
solutions. The mineral suspension was filtered using Whatman filter
Table 1
Particle size distribution of phosphate sample.
Size
（mesh）
Wt
（%）
Grade
（P2O5%）
A.I.
（%）
Cumulative
overscreen （%）
Cumulative
overscreen
grade （P2O5%）
Cumulative
overscreen
A.I. （%）
N30 5.47 4.13 91.44 5.47 4.13 91.44
30–50 29.06 11.95 32.86 34.53 10.71 42.14
50–70 51.36 6.56 75.44 85.89 8.23 62.05
70–100 6.34 5.07 91.41 92.23 8.01 64.07
100–200 7.28 1.78 75.53 99.51 7.56 64.91
b200 0.49 4.76 92.89 100.00 7.54 65.05
2 J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9
Author's personal copy
paper （pore size 25 μm） and then poured into the rectangular cell for
zeta potential measurements. The pH was measured again at the end
of test as the final pH.
2.5. FTIR analysis
The infrared transmission spectra were recorded on a Thermo
Nicolet Nexus 470 FTIR spectrometer. The hydroxyapatite powders
were conditioned with 50 mL collector solution at different pHs and
different concentrations, while being agitated with magnetic stirrer
for 0.5 h to make 1% suspension. The suspension was then filtered
with Whatman filter paper （pore size 25 μm）, and the solids were airdried
overnight at the room temperature. The samples were prepared
by dispersing 0.025 g air-dried powder in 5 g KBr followed by pressing
into a transparent tablet for scanning. The untreated （initial）
hydroxyapatite powder was used as reference. Each spectrum is an
average of 250 scans.
3. Results and discussion
3.1. Effect of collector dosage
The effects of dosage on rougher phosphate flotation performance
were studied by a series of flotation tests with the refined tall oil fatty
acid in conjunction with the plant collector. The pH, conditioning time
and impeller rotating speed were fixed constant at pH 10, 6 min and
1500 rpm, respectively and the ratio of collector to diesel was 3:2. The
dependence of flotation efficiency and A.I. on collector dosage is
shown in Fig. 1. It was quite obvious that the refined tall oil fatty acid
generated higher flotation efficiency than the plant collector at all
three dosages tested. The flotation efficiency increased significantly
with increasing dosage of the refined tall oil fatty acid from 0.15 kg/t
to 0.3 kg/t, but decreased slightly when the refined tall oil fatty acid
dosage increased further from 0.3 kg/t to 0.45 kg/t. When the plant
collector dosage increased from 0.15 kg/t to 0.3 kg/t, the increase in
flotation efficiency was less significant than for the refined tall oil fatty
acid. The highest flotation efficiency of 79.5% was achieved with
0.3 kg/t of the refined tall oil fatty acid, which was about 10% higher
than the flotation efficiency with the plant collector at the same
dosage or more than 6% higher than the maximum flotation efficiency
achieved with the plant collector at 0.45 kg/t. However, a further
increase in the dosage to 0.45 kg/t decreased the flotation efficiency to
77.8%, which was still about 5% higher than the plant collector at the
same dosage. The concentrate A.I. increased with increasing the
dosage and the A.I. with the plant collector remained higher （i.e.
worse） than with the refined tall oil fatty acid. When the refined tall
oil fatty acid dosage increased from 0.3 kg/t to 0.45 kg/t, the increase
in A.I. was less significant than with the plant collector, which
indicates that the refined tall oil fatty acid showed better selectivity
and higher acid insolubles rejection than the plant collector.
Fig. 2 demonstrates the relationship between P2O5 recovery and
grade with the refined tall oil fatty acid and the plant collector to
compare the separation sharpness of these two reagents. The results
indicate that the flotation P2O5 recovery with the refined tall oil fatty
acid was higher than with the plant collector at a given P2O5 grade
ranging approximay from 20% to 30%. When the grade decreased
from 29.0% to 25.9%, the recovery with the refined tall oil fatty acid
increased from 72.1% to 82.5% while the recovery with the plant
collector showed no significant increase. All of the above data indicates
that the refined tall oil fatty acid performed better than the plant
collector.
3.2. Effect of diesel dosage
Diesel plays an important role in phosphate flotation. It acts as
solvent or booster of tall oil fatty acid collectors and has a significant
effect on foam controlling （Snow et al., 2004）. To understand the
effect of diesel on the performance of the collectors, flotation tests
were conducted by adding the mixture of the refined tall oil fatty acid
and diesel as collector at the same dosage but in different ratios （by
weight） at pH 9.5 and the results of flotation recovery and P2O5 grade
are presented in Fig. 3.
Fig. 3 shows that the P2O5 recovery increased with increasing the
diesel percentage from 10% to 50% at both dosages and the highest
recovery was achieved when the ratio of the refined tall oil fatty acid to
diesel was 1:1. However, a further increase in the diesel percentage
from 50% to 80% decreased the recovery from 88.4% to 32.9% at the
dosage of 0.9 kg/t and from 63.9% to 15.7% at the dosage of 0.6 kg/t. The
highest grade was achieved at 30% diesel percentage with 0.6 kg/t
collector and 60% diesel percentage with 0.9 kg/t collector. It is
interesting to see that 60.1% recovery and 17.5% P2O5 grade were
achieved at 30% diesel percentage with 0.27 kg/t diesel and 0.63 kg/t
GP193G75 but 63.9% recovery and 15.8% P2O5 gradewere also obtained
at the dosage of only 0.3 kg/t GP 193G75 with 0.3 kg/t diesel. This
indicates that increasing diesel ratio may decrease the consumption of
the refined tall oil fatty acid but with insufficient amount of collector,
increasing diesel ratio cannot achieve optimum performance. When
both recovery and grade are considered at the same time, a collector
dosage of 0.9 kg/t was significantly better than that of 0.6 kg/t.
Fig. 1. Effect of collector dosage on flotation efficiency and concentrate A.I.
Fig. 2. Relationship between the P2O5 recovery and grade with GP193G75 and plant
collector at different dosage.
J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9 3
Author's personal copy
3.3. Effect of pH
Flotation tests were conducted to investigate the effect of slurry pH
on the collector performance in apatite flotation by varying the pH
from 9.5, 10, to 10.5. Fig. 4 demonstrates the flotation efficiency/A.I. as
a function of slurry pH in the presence of 0.3 kg/t of the refined tall oil
fatty acid or the plant collector conditioned for 6 min. It can be seen
that the flotation efficiency increased with increasing pH values and
the refined tall oil fatty acid showed significantly higher flotation
efficiency than the plant collector at all three pHs. The flotation
efficiency obtained with the refined tall oil fatty acid increased from
72.8% to 82.6% as the pH increased from 9.5 to 10.5. Meanwhile, the A.
I. with the refined tall oil fatty acid and the plant collector increased
with increasing slurry pH but was always lower with the refined tall
oil fatty acid at all three pHs, which indicates that the refined tall oil
fatty acid results in better selectivity than the plant collector.
The relationship between P2O5 recovery and grade with the refined
tall oil fatty acid and the plant collector is shown in Fig. 5. The curves
closer to the upper right corner represent more efficient separation.
Compared with the plant collector, the refined tall oil fatty acid
collector increased P2O5 recovery by 6% at the same concentrate grade.
Obviously, the refined tall oil fatty acid has a better flotation
performance than the plant collector. It can also be concluded from
the results that higher pulp pH had positive effects on phosphate
flotation. This observation is in agreement with the previous studies.
Feng and Aldrich （2004） investigated the effect of some operating
parameters such as pulp pH and collector dosage on the kinetics of
apatite flotation and obtained the best flotation performance at a pH
level of 12.3 with fatty acid and sulphonate as collector. They reported
that an elevated pulp pH increased the flotation recovery and kinetics,
probably by softening the process water and speeding up the
electrolysis of the fatty acid. According to the work by Robert Pugh
and Per Stenius （1984）, the minimum surface tension and the
formation of pre-micella associated species occurred at higher pH in
lower concentrations of sodium oleate solution, which attributed to
the better flotation recovery of apatite.
3.4. Kinetic flotation tests
The kinetics of flotation studies the variation in floated mineral
mass as a function of flotation time. To characterize the effect of
different collectors on flotation rate, kinetic flotation tests were
performed with both the refined tall oil fatty acid and plant collector.
During the tests concentrate samples were collected at a time interval
of 10 s in the first 30 s and also finally at 90 s. Both reagents were tested
under the same conditions. The dosage of collector was 0.45 kg/t
mixed with 0.4 kg/t fuel oil and the pH of ore pulp was 10. The results
of recovery and concentrate grade versus flotation time are shown in
Fig. 6. It can be seen that the recovery increased about 30% in the first
Fig. 4. Effect of slurry pH on flotation efficiency and concentrate A.I.
Fig. 5. Relationship between the P2O5 recovery and grade with GP193G75 and plant
Fig. 3. Effect of diesel percentage on flotation recovery and grade （pH 9.5, dosage of collector at different pH.
GP193G75 and diesel mixture at 0.6 kg/t and 0.9 kg/t）.
Fig. 6. P2O5 grade and recovery versus flotation time.
4 J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9
Author's personal copy
30 s with both the refined tall oil fatty acid and the plant collector and
additional 7% and 11.7% in the last 60 s, respectively. The grade of
concentrate generated by the plant collector decreased 5% with the
increase in recovery in the first 30 s. On the other hand no grade
decrease was observed with the refined tall oil fatty acid during the
same period of time. At 90 s the refined tall oil fatty acid showed a
higher recovery of 91.7% and a higher grade of 26.5% compared to the
plant collector （88.6% recovery and 20.6% grade）. The above data
shows that the refined tall oil fatty acid had considerably better
selectivity and flotation efficiency than the plant collector.
3.5. Adsorption behavior of fatty acid onto hydroxyapatite surface
To characterize the adsorption behavior of the refined tall oil fatty
acid and plant collector on the hydroxyapatite surface, a highly sensitive
in-situ surface characterization technique QCM-D was employed
in conjunction with zeta potential measurement and FTIR
spectra analysis.
3.5.1. QCM-D measurements
Fig. 7 shows the real-time response curves of frequency shift
（Δf） and dissipation shift （ΔD） from the third overtone （15 MHz）
associated with the refined tall oil fatty acid and the plant collector
adsorption onto a hydroxyapatite surface at a concentration of
500 ppm. Fig. 7A displays the frequency shift （Δf） for the refined tall
oil fatty acid and the plant collector adsorption on the hydroxyapatite
surface. Arrow a indicates the beginning of injection of collector
solution into the system. It can be observed that after the injection of
the refined tall oil fatty acid, Δf had an immediate sharp decrease
simultaneous with a sharp increase in ΔD （Fig. 7B, arrow a）. These
sharp changes in Δf and ΔD indicate the quick adsorption of the refined
tall oil fatty acid on the apatite surface and the high ΔD value （ΔDN1E
Fig. 7. QCM-D experimental data of frequency shift （A） and dissipation shift （B） （measured at 15 MHz） of GP193G75 and plant collector adsorption onto a hydroxyapatite surface at
concentration of 500 ppm.
J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9 5
Author's personal copy
−6） suggests the formation of a dissipated layer. The highest
adsorption thickness obtained at arrow b was 70 nm which was
calculated using the Voigt model. It should be noted that Δf continued
to increase after the adsorption thickness reached the maximum and a
steady state was not reached until after 90 min at arrow c. The
adsorption thickness at the stable frequency shift was 11 nm. This
phenomenon has rarely been reported in the previous studies of
adsorption of proteins and surfactants. According to the work by
Mielczarski and Mielczarske （1995）, who applied the FTIR reflection
spectroscopy in the analysis of the molecular orientation and thickness
of oleate adsorption on the apatite surface, the oleate monolayer on
the apatite was well-organized, but the second layer adsorbed on top
of the first well-ordered layer was randomly spread and oriented
almost parallel to the interface. Since the FTIR is an ex-situ analysis
method and QCM-D is an in-situ analysis technique, this phenomenon
of increase in Δf （Fig. 7A, from arrow b to arrow c） probably reflected
the orientation process of the second and subsequent layers of
molecules, which led to a low free energy state in the solution system.
As reported by Mielczarski and Mielczarske （1995）, water molecules
and calcium ions on the surface of apatite formed an intramolecular
layer, which was not uniform for the carboxylate group. The higher ΔD
at arrow b probably indicated the formation of soft and water-rich
multi-layers of fatty acid on the apatite surface at the very beginning.
ΔD decreased with the increase of Δf （Fig. 7B from arrow b to arrow c）,
which indicated the formation of more rigid multi-layers due to the
change in orientation of top layers of molecules and the loss of water
caused by layer compression.
Compared with the refined tall oil fatty acid, the plant collector
generated a gradual decrease of Δf and increase of ΔD. However, Δf
reached a stable value after 30 min when the ΔD value was still
increasing. This indicates a slow adsorption process as well as the
formation of much softer and more porous multi-layers. Rodahl et al.
（1997） postulated that if the porous film is deformed by shear
oscillation, liquid can be “pumped” in and out of the film and also be
subjected to oscillatory “flow” within the film as the pores in the film
change in shape and size. This may cause the continuing increase of ΔD
even when the Δf was stable. The final thickness of the plant collector
adsorption layer on the surface of hydroxyapatite calculated using the
Voigt model was 4 nm, which was not only thinner, but also more
dissipated and less rigid （plant collector ΔDNthe refined tall oil fatty
acid ΔD） than the adsorption layer for the refined tall oil fatty acid.
The frequency shift and dissipation shift caused by adsorption of
the refined tall oil fatty acid and the plant collector at a concentration
of 1000 ppm showed almost the same trend as at 500 ppm. It should
be noted that the stable Δf at 1000 ppm appeared 1.3 h earlier than at
500 ppm for the refined tall oil fatty acid. This indicates that the
refined tall oil fatty acid adsorbed on the hydroxyapatite surface more
rapidly at higher concentrations.
The measured frequency and dissipation shifts at 500 ppm and
1000 ppm were summarized as ΔD–Δf plots in Fig. 8A–D. ΔD–Δf plots
Fig. 8. ΔD–Δf plots for （A） plant collector at 500 ppm, （B） plant collector at 1000 ppm, （C） GP193G75 at 500 ppm, and （D） GP193G75 at 1000 ppm, adsorbed on the hydroxyapatite
surface.
6 J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9
Author's personal copy
can provide information on the energy dissipation per unit mass added
to the crystal. The slope of the plots was defined as K （K=ΔD/Δf,
absolute value） which is indicative of kinetic and structure alternation
during adsorption process （Paul et al., 2008）. A more rigid and compact
adsorption mass is expected to yield a small K value, and a soft and
dissipated layer is expected to yield a higher K value. Only one slope is
associated with the adsorption process without kinetic or conformational
change. More than one slope suggests the adsorption associated
with orientation change or hydrodynamically-coupled water in the
adsorbed layer （Paul et al., 2008; Rodahl et al., 1997）. Fig. 8A and B
shows the ΔD–Δf plots for the adsorption of the plant collector on the
hydroxyapatite surface at 500 ppm and 1000 ppm, respectively. Two
slopes were observed based on the gross trend of curves in Fig. 8A and
B. It can be observed that the 500 ppm and the 1000 ppm tests have
similar K1 values of 0.106 and 0.092, respectively, while K2 was higher
than K1 at both concentrations, and the trendlines were almost vertical
to the X axis. This indicates that the initially-formed layer was less
dissipated, but with the increase of adsorbed mass and trapped water
molecules, it was getting softer and more dissipated. At the second
stage with a slope of K2, the ΔD increased rapidly while the Δf was
stable at about −10 Hz. As mentioned before, this elevated ΔD was
probably due to the structural change of the adsorbed layer caused by
trapped water. Fig. 8C and D shows the ΔD–Δf plots for the adsorption
of 500 ppm and 1000 ppm of the refined tall oil fatty acid on the
hydroxyapatite surface. The trend of the curve shows three distinguishable
stages （three K values） in the adsorption process. Comparing
the K values in Fig. 8C and D reveals that K2 （0.159 at 500 ppm and
0.029 at 1000 ppm） was smaller than K1 （0.305 at 500 ppm and 0.220
at 1000 ppm）. This indicates that although the initial adsorption of the
refined tall oil fatty acid on hydroxyapatite was rapid, the adsorbed
layer was soft and dissipated. With decreasing Δf, ΔD increased slowly
at the second stage, which yielded the lower K2.
It must be noted that, in comparison with the plant collector, the
refined tall oil fatty acid has a third adsorption stage at which the Δf
increased as the ΔD decreased to smaller than 1E−6. This indicates
the loss of water caused by layer compression or the change in
orientation of molecules. The dissipation shift per unit mass at this
stage was the same as the first stage.
3.5.2. Zeta potential measurements
Fig. 9 shows the zeta potentials of pure hydroxyapatite before and
after conditioning with the refined tall oil fatty acid or the plant
collector at the constant ionic strength with respect to pH. It can be
seen that the zeta potential of pure hydroxyapatite was zero at pH 4
and the shapes of curves agreed well with the results reported by Rao
et al. （1990）. The results indicate that after conditioning with
collectors, the zeta potential of hydroxyapatite became more negatively
charged and shifted the i.e.p. towards lower than pH 4. In
addition, the zeta potential with the refined tall oil fatty acid was lower
than that with the plant collector at pHs ranging from 5 to 11, which is
in good agreement with the QCM-Dresults that showed the refined tall
oil fatty acid is associated with higher affinity and adsorption density
than the plant collector. According to Rao et al. （1990）, the more
negative charge of apatite as a result of conditioning with sodium
oleate indicates the high affinity of oleate with surface Ca-sites. Either
the chemisorption of the refined tall oil fatty acid, which forms the
monocoordinated complex （i.e. 1:1 oleate-lattice calciumcomplex） or
the precipitation of its calcium salt may take place, especially at high
concentrations of the collector. As shown inQCM-Dresults, the sharply
decreased frequency shift from the beginning of adsorption demonstrates
that the consistent mass increase happened on the hydroxyapatite
surface due to the simultaneous chemisorption and surface
precipitation on hydroxyapatite, which formed the 70 nm adsorption
layer at 500 ppm solution with the refined tall oil fatty acid.
The QCM-D results show not only the adsorption process, but also
the process of layer compression and molecular orientation （K2bK1）.
According to Mielczarski et al. （1993）, the hydrophobicity of apatite is
closely related to the structure of the adsorbed layer on the surface,
wherein the higher packing density involves hydrophobic character of
sample, a poorly organized structure does not produce a high
hydrophobicity of apatite. When the adsorption became stable, the
adsorbed layer of the refined tall oil fatty acid had a lower ΔD than the
plant collector, which indicated better organized structure of the
refined tall oil fatty acid on hydroxyapatite than the plant collector,
which caused higher hydrophobicity of hydroxyapatite conditioned
with the refined tall oil fatty acid.
3.5.3. FTIR spectra
Infrared transmission spectra of hydroxyapatite after adsorption in
the refined tall oil fatty acid solution at different pHs and concentrations
are shown in Fig. 10 and Fig. 11. According to many researchers
（Antti and Forssberg, 1989; Mielczarski et al., 1993; Rao et al., 1991）,
the absorbance bands which can provide useful information about
the nature of the adsorbed fatty acid such as oleic acid can be observed
in two regions of frequency, i.e., （1） most peaks at wavenumbers
from 3100 cm−1 to 2900 cm−1 are related to the hydrocarbon chain.
（2） Most peaks at wavenumbers from 1700 cm−1 to 1400 cm−1 are
Fig. 9. Zeta potential of hydroxyapatite and hydroxyapatite conditioned with 25 ppm
GP193G75 and plant collector at constant ionic strength as a function of pH.
Fig. 10. Infrared transmission spectra of hydroxyapatite after adsorption of GP193G75
at different concentrations and pHs （concentration ppm/pH） for the alkyl chain region
2750–3100 cm−1.
J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9 7
Author's personal copy
related to the carboxylate radical. It must be noted that the infrared
spectra in the case of calcite are found to be hard to interpret due to the
highly interfering carbonate absorption in the same frequency region
as that of the carboxylate radical （Rao and Forssberg, 1991）.
Fig. 10 shows the typical absorption bands of alkyl chain in the
region from 2750 cm−1 to 3100 cm−1, wherein the 2923 cm−1 and
2852 cm−1 bands were the asymmetric and symmetric stretching
vibrations in the CH2 radical, respectively and 2956 cm−1 was the
asymmetric stretching vibration in the CH3 radical. Even though the
pH did not show significant effect on the intensities of adsorption
peaks, the spectra in alkyl region showed absorption bands in the
range of 2750–3100 cm−1 and the intensities of these bands were
found to increase with increasing concentration of the refined tall oil
fatty acid, as shown in Fig. 10.
Fig. 11 shows the infrared transmission spectra of hydroxyapatite
after adsorption of the refined tall oil fatty acid at different concentrations
and pHs for the carboxylate group region from 1400 cm−1 to
1600 cm−1. Since the hydroxyapatite sample itself shows characteristic
absorption bands between 1400 cm−1 and 1500 cm−1 and the carboxyl
absorbance bands appear in the narrow range between 1540 cm−1 and
1580 cm−1, it is difficult to identify the typical absorption bands of
carboxyl. It was observed that the carboxylate group in the refined tall
oil fatty acid gave rise to characteristic absorption bands in the region of
1600–1400 cm−1 at the higher concentration of 1000 ppm,wherein the
1450 cm−1 bandcorresponds to the asymmetric deformation of theCH3
radical. In ionized fatty acids, the band at 1710 cm−1 is replaced by two
new bands due to vibrations in the COO– radical including 1610–
1550 cm−1 and 1420–1300 cm−1 （Antti and Forssberg, 1989）. The
adsorption bands that occur at 1576 cm−1 and 1537 cm−1 are assigned
to the asymmetric stretching vibration of COO– radical. This corresponds
to themonocoordinated complex or the surface precipitation of
calcium salt fatty acid at the concentration of 1000 ppm, which is also
demonstrated by QCM-D measurements that showed the frequency
shift decreased significantly at the beginning of adsorption as a result
of the formation of a thick and soft adsorption layer.
According to Mkhonto et al. （2006）, all surfactants containing
carbonyl and hydroxy groups interact strongly with the apatite
surfaces and bridging between two or more surface calcium ions of
apatite is the preferred mode of surfactant adsorption.
In summary, infrared studies showed that the adsorption of the
refined tall oil fatty acid on apatite involves both chemical reaction
between the carboxylate species and surface calcium ions and surface
precipitation, which agrees well with the zeta potential and QCM-D
measurements.
4. Conclusions
The performance of a commercial phosphate collector and a
refined tall oil fatty acid was investigated in this study. Flotation tests
were performed under different process parameters. It was found that
under the same process conditions, the refined tall oil fatty acid
achieved a considerably better flotation performance than the plant
collector. The recovery and concentrate grade were 91.7% and 26.5%,
respectively when the refined tall oil fatty acid was used at a dosage of
0.45 kg/t at pH 10.0 with a 9:8 （by mass） concentration ratio of fatty
acid to diesel. High pulp pH was found to have positive effects on
phosphate flotation. Use of diesel as synergist can significantly
decrease the dosage of the refined tall oil fatty acid, but an increase
in diesel dosage cannot achieve good performance without a sufficient
amount of the refined tall oil fatty acid.
The QCM-D results indicated that the refined tall oil fatty acid had
different adsorption behavior than the plant collector. The plant
collector had two adsorption stages and K2 was greater than K1, which
means the adsorbed layer become more dissipated with increasing
adsorption time. On the contrary, the adsorption layer of the refined
tall oil fatty acid became more rigid when K2 was smaller than K1.
It also showed a compression stage represented by K3. Therefore,
the adsorption layer of the refined tall oil fatty acid formed on a
hydroxyapatite surface was more organized and rigid than the plant
collector at the concentrations of 500 and 1000 ppm. Since the higher
packing density of adsorbed layer results in hydrophobic character and
a poorly organized structure does not produce a high hydrophobicity,
the hydroxyapatite surface adsorbed by the refined tall oil fatty acid
had higher hydrophobicity than the plant reagent after the adsorption
reached the steady state. The results of zeta potential measurements
and FTIR analyses indicated strong electrostatic interaction and high
affinity of the refined tall oil fatty acid on the hydroxyapatite surface.
Both chemisorption and surface precipitation mechanisms were
demonstrated by FTIR analysis and QCM-D measurement.
References
Antti, B.M., Forssberg, K.S.E., 1989. Pulp chemistry in industrial mineral flotation,
studies of surface complex on calcite and apatite surfaces using FTIR spectroscopy.
Miner. Eng. 2, 217–227.
Ekholm, P., Blomberg, E., et al., 2002. A quartz crystal microbalance study of the
adsorption of asphaltenes and resins onto a hydrophilic surface. J. Colloid Interface
Sci 247, 342–350.
Feng, D., Aldrich, C., 2004. Influence of operating parameters on the flotation of apatite.
Miner. Eng. 17, 453–455.
Hook, F., Rodahl, M., Brzezinski, P., Kasemo, B., 1998. Energy dissipation kinetics for
protein and antibody–antigen adsorption under shear oscillation on a quartz
crystal microbalance. Langmuir. 14, 729–734.
Mielczarski, J.A., Mielczarske, E., 1995. Determination of molecular orientation and
thickness of self-assembled monolayers of oleate on apatite by FTIR reflection
spectroscopy. J. Phys. Chem 99, 3206–3217.
Mielczarski, J.A., Cases, J.M., Bouquet, E., Barres, O., Delon, J.F., 1993. Nature and structure
of adsorption layer on apatite contacted with oleate solution. 1. Adsorption and
fourier transform infrared reflection studies. Langmuir. 9, 2370–2382.
Mkhonto, D., Ngoepe, P.E., Cooper, T.G., Leeuw, N.H., 2006. A computer modeling study of
the interaction of organic adsorbates with fluorapatite surfaces. Phy. Chem. Minerals
33, 314–331.
Paul, S., Paul, D., Tamara, B., Ray, A.K., 2008. Studies of adsorption and viscoelastic
properties of proteins onto liquid crystal phthalocyanine surface using quartz crystal
microbalance with dissipation technique. J. Phys. Chem. C. 112, 11822–11830.
Pugh, R., Stenius, P., 1984. Solution chemistry studies andflotation behavior of apatite, calcite
and fluorite minerals with sodium oleate collector. Int. J. Miner. Process. 15, 193–218.
Rao, K.H., Forssberg, K.S.E., 1991. Mechanism of fatty acid adsorption in salt-type
mineral flotation. Miner. Eng. 4, 879–890.
Rao, K.H., Antti, B.M., Forssberg, E., 1990. Mechanism of oleate interaction on salt-type
minerals, part II. Adsorption and electrokinetic studies of apatite in the presence of
sodium oleate and sodium metasilicate. Int. J. Miner. Process. 28, 59–79.
Fig. 11. Infrared transmission spectra of hydroxyapatite after adsorption of GP193G75
at different concentrations and pHs （concentration ppm/pH） for the carboxylate group
region 1600–1400 cm−1.
8 J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9
Author's personal copy
Rao, K.H., Cases, J.M., Forssberg, K.S.E., 1991. Mechanism of oleate interaction on salttype
minerals V. Adsorption and precipitation reactions in relation to the solid/
liquid ratio in the synthetic fluorite-sodium oleate system. J. Colloid Interface Sci
145, 330–348.
Rodahl, M., Hook, F., Fredrikson, C., et al., 1997. Simultaneous frequency and dissipation
factor QCM measurements of biomolecular adsorption and cell adhesion. Faraday
Discuss. 107, 229–246.
Sis, H., Chander, S., 2003. Reagents used in the flotation of phosphate ores: a critical
review. Miner. Eng. 16, 577–585.
Snow, R., Zhang, P., Miller, J.D., 2004. Froth modification for reduced fuel oil usage in
phosphate flotation. Miner. Eng. 74, 91–99.
Voinova, M.V., Rodahl, M., et al., 1999. Viscoelastic acoustic response of layered polymer
films at fluid–solid interfaces: continuummechanics approach. Phys. Scr. 59, 391–396.
Zhang, P., Yu, Y., Bogan, M., 1997. Challenging the “Crago” double floate process II.
Amine-type-fatty acid flotation of silicious phosphates. Miner. Eng. 10, 983–994.
J. Kou et al. / International Journal of Mineral Processing 95 （2010） 1–9 9