Available online at www.sciencedirect.

com

Minerals Engineering 20 (2007) 1303–1308

This article is also available online at:
www.elsevier.com/locate/mineng

Frother analysis: Some plant experiences

Stéphanie Gélinas a, James A. Finch b,*

a
Échem Technologies Inc., 558 De La Métairie, Verdun, Québec, Canada H3E 1S5
b
Department of Mining and Materials Engineering, McGill University, 3610 University Street, Montréal, Québec, Canada H3A 2B2

Received 16 March 2007; accepted 26 July 2007

Available online 11 September 2007

Abstract

A colorimetric technique, originally developed for methyl iso butyl carbinol (MIBC) based on the Komarowsky reaction, has been
extended to a wide range of common mineral flotation frothers. The procedure, which is briefly described, is suited to on-site determi-
nation as shown in a series of industrial case studies. Some observations described include: distribution of frother down a bank and
between pulp and froth; evidence of alcohol contaminants; and mass balancing. The ability to monitor frother will help interpret cell
hydrodynamics and should help optimize the type, amount and location of frother addition.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Froth flotation; Flotation frothers; Mass balancing

1. Introduction processing sulphide ores. The availability of portable visi-

ble spectrophotometers and the adaptability of the sensi-
Owing to lack of analytical methods, the task of deter- tive procedure, make on-site determination feasible.
mining frother concentration in flotation circuits is not The ability to measure frother should help plants iden-
routinely performed. Industrial sites seldom outsource tify the amount and location of frother addition to opti-
analyses that are costly and inconvenient. A colorimetric mize the process and to assess frother in reclaim water or
method has recently been demonstrated for on-site charac- in effluent streams. The information could be extended to
terization of common industrial frothers (Gélinas and help identify the type of frother to be used and to integrate
Finch, 2005; Gélinas et al., 2005). The approach, based frother use in control of cell hydrodynamics (the initial
on the Komarowsky reaction (Penniman et al., 1937; Coles impetus for the work).
and Tournay, 1942; Duke, 1947), was previously described To evaluate the colorimetric technique, the method was
by Parkhomovski et al. (1976) for methyl iso butyl carbinol first tested on-site as part of a cell hydrodynamics evalua-
(MIBC). The reaction involves the interaction of the OH tion campaign at Rio Tinto’s Copperton concentrator
group of a higher alcohol, concentrated sulphuric acid, (Utah, USA) (Gélinas et al., 2005). The procedure has sub-
and salicylaldehyde to yield coloured solutions that can sequently been exploited at other sites, which forms the
be analyzed by UV–visible spectrophotometry. basis for this communication.
Gélinas and Finch (2005) reviewed the general colori-
metric reactivity with salicylaldehyde of a range of frothers,
2. Experimental
including typical aliphatic alcohols, cyclic alcohols, phe-
nols, and alkoxyparaffins. The method proved largely
2.1. Materials
insensitive to other reagent additions typically encountered
All chemicals in the analytical procedure were used
*
Corresponding author. Fax: +1 514 398 4492. as received from Sigma–Aldrich (USA). Note that only
E-mail address: jim.fi[email protected] (J.A. Finch). chloroform stabilized with ethanol can be used to avoid

0892-6875/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2007.07.008
1304 S. Gélinas, J.A. Finch / Minerals Engineering 20 (2007) 1303–1308

cross-reaction of the stabilizer. The alcohols methyl iso below the froth surface. Froth samples were taken from
butyl carbinol (MIBC) and 3-methyl-1-butanol were also concentrate overflows using a pan.
from Sigma–Aldrich (USA). (The butanol was included
not as a frother per se but as it is a contaminant alcohol in 3. Results and discussion
the xanthate supply at one plant.) The polyglycol frothers
F-150 and Unifroth 250C (U250C) were supplied by Flottec 3.1. Frother profile in froth and pulp
(USA) and Huntsman & Reliable (Canada), respectively.
This was the first survey, conducted at Rio Tinto’s
2.2. Sample preparation for colorimetric analysis Copperton concentrator. The plant uses MIBC, an ali-
phatic alcohol, as frother and the profile (concentration
Standards are made-up of 0.05–0.2 mg of frother in down a bank) and distribution between the pulp and froth
100 mL of tap water at room temperature. Solutions from phases in a rougher/scavenger row (recovering bulk cop-
decanted supernatant pulp (or froth) samples are likewise per/moly) is shown in Fig. 1. The profiles were constructed
prepared, using appropriate dilution factors (based on esti- by analysis of samples from selected cells in the row. The
mated concentration from dosage rate and pulp density). results correspond to the calculated solution strength of
The samples are neutralized by dropwise addition of dilute around 10 ppm, obtained from an average initial MIBC
sulphuric acid (1:10), and the pH is adjusted to 6.5–7. To addition of 0.02 kg/mt (0.04 lb/st) of dry feed in cell 1
reduce foaming, 20 g sodium chloride per 100 mL sample assuming 33 wt.% solids, and a minor subsequent addition
is introduced. The samples are added to a 250 mL separa- of 0.003 kg/mt (0.006 lb/st) in cell 6.
tory funnel and two successive 10 mL extractions in chloro- The concentration in the pulp phase remains steady
form are performed. The mixed phases are vigorously throughout the row. The profile in the froth indicates an
shaken for two minutes, the aqueous phase is removed increasing concentration down the bank. Some transfer
and the bottom layer is transferred to another 50 mL separ- of frother to the froth (via adsorption on the bubble) is
atory funnel. Extraction to the reaction medium is achieved anticipated, but the increase down the bank suggests
by adding 10 mL of 3:1 sulphuric acid in water, shaking for unknown system interactions at play modulated by the
two minutes, and finally transferring the bottom layer to a presence of solids and variations in gas rate. While a con-
25 mL sealed glass vial. comitant decrease in pulp phase concentration should
result the relatively small flow of froth product in this
2.3. Reaction rougher stage compared to the pulp flow obscures any
changes. Analysis of the process water showed the presence
The reaction is initiated by the addition of 0.1 mL of 5% of 0.3 ppm MIBC, indicating the sensitivity of the
salicylaldehyde in a 1:1 solution of acetic acid in water. technique.
After mixing, the samples are placed simultaneously in a
bath of vigorously boiling water. The reaction is stopped 3.2. Frother profile following sequential addition
after exactly 20 min, by immediately immersing the sam-
ples in ice water, from which each vial is removed only This campaign was conducted at Xstrata Zinc’s Bruns-
when its turn comes for colorimetric analysis. Since the wick Mine operation (New Brunswick, Canada). The lay-
reaction is not allowed to go to completion, but is rather out of the rougher row is presented in Fig. 2. It
arbitrarily stopped, it is necessary to run standards and comprises four banks of four cells each, with MIBC (and
sample side by side under identical conditions. xanthate) addition at the head of each bank. The corre-
sponding frother distribution is roughly 55%, 20%, 15%
2.4. UV–visible analysis and 10%. Samples were taken at the top of the pulp phase

After the sample has reached room temperature, it is

transferred to a 1 cm2 visible cell for colorimetric analysis. 25
MIBC Concentration, ppm

The absorption spectra are scanned with a UV–visible

20
spectrophotometer (Thermoelectron Genesys 10 Scan-
ning). The reference sample is tap water subjected to the 15 Froth
same reaction procedure. For reference, the two alcohols
have a maximum absorption between 520 and 560 nm 10

and the polyglycols a lower range, 490 nm (U250C) to

5
500 nm (F-150). Pulp

0
1 2 3 4 5 6 7 8 9 110 111
2.5. Sampling
Cell number in Rougher/Scavenger Row

Specific pulp volumes were extracted from selected flota- Fig. 1. Pulp and froth phase frother profiles in a selected rougher/
tion cells using a manual perch sampler, at a depth well scavenger row (Rio Tinto’s Copperton concentrator).
 S. Gélinas, J.A. Finch / Minerals Engineering 20 (2007) 1303–1308 1305

Fig. 2. Row schematic showing frother addition points and (above) frother concentration for selected cells (Zinc Rougher row at Xstrata Zinc’s
Brunswick Mine).

(T) and at the bottom (B), close to the impeller, to check at Inmet’s Troilus Cu/Au concentrator (Québec, Canada).
distribution within the cell, along with samples of two Fig. 3a shows the pulp phase profile of F-150 (a polyglycol)
water sources, process water (PW) and quarry water (QW). in the rougher-scavenger row. A small variation around ca.
The results corresponding to the cell locations indicate 0.55 ppm is observed down the bank of seven cells. Traces
comparable concentrations at the top and bottom, suggest- of F-150 (0.09 ppm) were detected in the process water.
ing frother is well mixed throughout a cell. The initial con- The results for cleaner cells 3, 7, and 12 show a sharp
centration is ca. 3.5 ppm that steps up with each addition difference (Fig. 3b). Substantial amounts of frother were
to reach ca. 5 ppm. The additions, therefore, do not pri- detected in these pulp samples. The origin of the large dif-
marily replace frother (e.g., lost to froth) but progressively ference between roughers and cleaners is not fully under-
increase the pulp concentration. Process and quarry water stood. Factors being evaluated range from interference
carry approximately 1 ppm of MIBC. with the analysis (see later) to high recovery of F-150 to
the rougher froth (concentrate) to the impact of recycle
flows internal and external to the cleaner bank. The high
3.3. Frother distribution around a rougher/scavenger/cleaner values in the cleaner cells did correlate with the ‘creamy’
circuit appearance of the froth, and with hydrodynamic surveys
that showed only low gas rates could be tolerated in the
As part of a developing program to integrate frother use cleaner cells without flooding.
in control of cell hydrodynamics, a survey was conducted

1.0 100
F-150 Concentration, ppm
F-150 Concentration, ppm

0.8
75
0.6
50
0.4

0.2 25

0 0
Feed 1 2 3 4 5 6 7 Tails cleaner 3 cleaner 7 cleaner 12
Sample ID Sample ID

Fig. 3a. Frother distribution in pulp samples of a rougher-scavenger row Fig. 3b. Frother concentration in selected cleaner cells (Inmet’s Troilus
(Inmet’s Troilus Concentrator). Concentrator).
1306 S. Gélinas, J.A. Finch / Minerals Engineering 20 (2007) 1303–1308

3.4. Frother distribution around a copper/zinc circuit peak absorption), the results reported probably reflect the
total contribution of alcohol present in the pulp water. A
As part of the same program at Troilus a survey was systematic evaluation of the alcoholic components in the
conducted at Agnico-Eagle’s Laronde operation (Québec, range of reagents added at the plant is required.
Canada). The results for the distribution of MIBC at The next case study where a polygylcol frother was
selected points around the Cu-circuit are presented on the employed gave a chance to test this interference from con-
flowsheet in Fig. 4. The results indicate an apparently sig- taminant alcohol.
nificant amount of MIBC (45 ppm) is already present after
the single frother addition to the primary conditioning
tank. The concentration remains near this level through 3.5. Mass balance exercise
the roughing and scavenging sections. The cleaning circuit
shows about half this concentration level, suggesting dilu- CVRD INCO’s Clarabelle concentrator (Ontario, Can-
tion possibly by launder water. The most dilute is the final ada) uses Unifroth 250C frother (a polyglycol) and reports
cleaner, supporting that there is progressive dilution. The up to ca. 5% alcohol (3-methyl-1-butanol) in their xanthate
results (not shown) for the Zn-circuit match those for the supply. The objectives were first to demonstrate the two
Cu-circuit: about 40 ppm in roughing/scavenging decreas- reagents could be de-coupled for independent analysis
ing progressively through cleaning to ca. 10 ppm. and, second, to collaborate with a mass balance exercise
The results do not correspond to the known amount of to try to account for both reagents around the circuit. Since
MIBC added which should give ca. 6 ppm in the primary polyglycol and the alcohol have different maximum peak
conditioner. A concentration of 3.6 ppm was measured in absorption, this allowed conventional simultaneous
the recycle water thus there is no build-up from that source. analysis.
The current suspicion is that there is a contribution of alco- Fig. 5 presents the flowsheet indicating the reagent addi-
holic components originating from other reagents added to tion locations (arrows) and our sampling points (dots). The
the circuit. Xanthates, for example, often carry residual concentrations are expressed in ppm of U250C (left hand
alcohol from formulation, which is considered inert and number) and 3-methyl-1-butanol (right). The pertinent
therefore not necessary to remove (Cappuccitti, 2005). observations are: the concentration of frother and alcohol
Since the colorimetric technique cannot distinguish is similar in magnitude; after each addition frother concen-
between different aliphatic alcohols (similar maximum tration increases; frother (and to some degree the alcohol)

MIBC: 0 – 6 g/t Roughing:

Primary Cleaning:
Column cells
Column & Contact cells
COPPER
Primary Conditioning CONCENTRATE

48.3ppm
19.3ppm
45.1ppm 43.4ppm

28.3ppm

Secondary Cleaning:
Contact, Column,
Mechanical & Tank Cells

Scavenging: 44.9ppm
Contact & Mechanical Cells
24.0ppm
12.6ppm
42.4ppm

TO ZINC CIRCUIT

Fig. 4. Apparent frother distribution in the copper circuit (Agnico-Eagle’s Laronde concentrator).
 S. Gélinas, J.A. Finch / Minerals Engineering 20 (2007) 1303–1308 1307

Feed
U250C – Methyl Butanol
0 – 1.0
Mag Cleaners 0–0 Roughers/Scavengers
MAG- CLNR
1.7 -4.0
1.7-4.0 6.4 – 4.5 7.7 – 4.1 10.9 – 6.6
4.8 – 6.6 AA BB C D
SCAVENGER

MAG- RCLNR
27.9 – 10.2
4.5 – 6.9 14 5.6
14 – 5.6
11.9 – 10.4
Cleaner
“B” Cleaner

22 – 7.3
23.8 – 15.7
SCAV-- CLNR

Scavenger Cleaners
4.8 – 6.6 SCAV-- RCLNR

10.2 – 5.3
22 – 7.3

CONCENTRATE B CONCENTRATE A ROCK TAILS

Fig. 5. Flowsheet with frother addition points and sampling locations with corresponding U250C and 3-methyl-1-butanol concentrations (CVRD INCO’s
Clarabelle mill).

concentrates through the cleaning circuit; and significant to the mag circuit) and the total frother ‘‘out’’ was esti-
frother remains in the tails. mated at 15.99 kg/h, (9.66 kg/h in the rock tails, 3.5 kg/h
The de-coupling of the frother and contaminant alcohol in concentrate B, and 2.83 kg/h in concentrate A); i.e.,
is a positive for the technique but the practical concern the balance closed to within 6%. A similar exercise for
whether the alcohol has any effect on the process remains. the 3-methyl-1-butanol showed an estimated total of
The finding led to including 5-carbon alcohols in frother 11.02 kg/h ‘‘in’’ and total ‘‘out’’ of 11.94 kg/h, a discrep-
characterization work which suggested an effect on bubble ancy of 8%.
size cannot be ruled out at the concentrations detected here
(Azgomi et al., 2006). The progressive increase in concen- 4. Conclusions
tration as frother is added, as in the Brunswick case, does
suggest there is little foundation for concern sometimes The colorimetric technique for frother analysis was dem-
expressed that frother is depleted through a circuit. Indeed, onstrated in a series of on-site investigations, including
it has led to a significant concentration of frother in the simultaneous analysis in the case of an alcohol and a poly-
cleaners, which may not be desirable. glycol. Findings of potential practical importance are: for
At the same time as the frother analysis work was being MIBC and U250C (but perhaps not F-150) there is little
conducted plant personnel carried out a mass balancing evidence of frother depletion in the pulp liquor through a
exercise. Water flow rates were made available which per- circuit which is an argument against sequential addition
mitted a first attempt at mass balancing the frother (and (that may cause problems of excess frother build-up); a cor-
alcohol). Fig. 6 shows mass flows ‘‘in’’ and ‘‘out’’ of the cir- ollary is that most frother remains in the pulp, exiting with
cuit for U250C. A total of 15.11 kg/h was added ‘‘in’’ the tails; and, contaminant alcohols can be present from other
circuit (13.07 kg/h to the non-mag circuit and 2.04 kg/h reagent additions (the case of alcohol in xanthate is
detailed) that may interfere with the analysis (and may
have effects on the process). Together with the demon-
16 strated ability to mass balance, the way is opening for more
Calculated amount of U250, kg/h

Mag control over frother selection and addition.

14

12 Acknowledgement
10 Non

8
Mag
The funding for this work was initiated under a Collab-
orative Research and Development grant from the Natural
6
Sciences and Engineering Research Council of Canada
4 (NSERC) with industrial sponsorship from (giving the ori-
2 ginal names, 2001) Inco, Teck-Cominco, Falconbridge,
0
Noranda, COREM, and SGS Lakefield Research. The
"in" "out " "out" "out" authors wish to thank the many personnel who enthusias-
rock tails conc B conc A tically collaborated in the site work and the companies
Fig. 6. Overall balance obtained for U250C frother (CVRD INCO’s mentioned for supporting the project and giving permis-
Clarabelle mill). sion to present the findings.
1308 S. Gélinas, J.A. Finch / Minerals Engineering 20 (2007) 1303–1308

References Gélinas, S., Finch, J.A., Cappuccitti, F., 2005. Frother Analysis: Proce-
dure and Plant Experience. In: Proceedings 37th Annual Meeting of
Azgomi, F., Gomez, C.O., Finch, J.A., 2006. Characterizing Frothers the Canadian Mineral Processors of CIM, Jan. 18–20, 2005, pp. 569–
Using Gas Holdup. In: Xu, Z., Liu, Q. (Eds.), Interfacial Phenomena 576.
in Fine Particle Technology, 6th UBC-McGill-UA International Gélinas, S., Finch, J.A., 2005. Colorimetric determination of common
Symposium on Fundamentals of Mineral Processing Conference of industrial frothers. Min. Eng. 18 (2), 263–266.
Metallurgists (CIM). 2006, Montréal, Oct 1–4, 2006, pp. 429–442. Parkhomovski, V.L., Petrunyak, D.G., Paas, L., 1976. Determination of
Cappuccitti, F., 2005. Private communication. methylisobutylcarbinol in waste waters of concentration plants.
Coles, H.W., Tournay, W.E., 1942. The Komarowsky color reaction for Obogashchenie Rud 21 (2), 44–45.
aliphatic alcohols. Ind. Eng. Chem. Anal. Ed. 14 (1), 20–22. Penniman, W.B.D., Smith, D.C., Lawshe, E.I., 1937. Determination of
Duke, Frederick R., 1947. The Komarowski reaction. Anal. Chem. 19 (9), higher alcohols in distilled liquors. Ind. Eng. Chem. Anal. Ed. 9 (2),
661–662. 91–95.