Minerals Engineering 61 (2014) 1–8

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Frother function–structure relationship: Dependence of CCC95 on HLB

and the H-ratio
James A. Finch a,⇑, Wei Zhang b
a
Department of Mining and Materials Engineering, McGill University, Montreal, QC H3A 0C5, Canada
b
Mining & Mineral Resources Division, Department of Business Administration, Chinalco China Copper Corporation Limited, Beijing 100082, China

a r t i c l e i n f o a b s t r a c t

Article history: Although the number and diversity of commercial frothers has steadily increased to meet flotation indus-
Received 2 September 2013 try demands, frother selection is still mainly empirical. As part of a general structure–property study, the
Accepted 13 February 2014 paper presents a correlation between the critical coalescence concentration (CCC95) and H-ratio for sur-
Available online 24 March 2014
factants used as flotation frothers. The CCC95 were determined in 0.8 m3 mechanical flotation cell. The H-
ratio was a substitution of hydrophile-lipophile balance (HLB) and determined through high resolution
Keywords: proton nuclear magnetic resonance (1H NMR) spectrometry. A large data set, consisting of 45 surfactants
Flotation
from four frother families, was used to develop the correlation. It is shown that the H-ratio can substitute
Frothers
CCC
for HLB. The potential of NMR both to identify the frother family and to derive the H-ratio in predicting
CCC95 CCC95 for commercial frothers is discussed.
HLB Ó 2014 Elsevier Ltd. All rights reserved.
H-ratio

1. Introduction The bubble size reduction function in some manner must depend
on frother structure (Pugh, 2007), e.g. the relative strength of the
Flotation treats billions of tonnes of mineral ores a year. The hydrophobic and hydrophilic groups. The relative strength can be
process exploits differences in mineral wettability to capture modified by changing the number of –CH2– groups in the alkyl chain
hydrophobic mineral particles on bubbles dispersed in slurry. Fro- in alcohols and polyglycols, and the number of propylene oxide (PO)
thers are usually added to help reduce bubble size and increase or ethylene oxide (EO) groups in the case of polyglycols.
particle collection kinetics. This function is commonly related to Varying the relative strength of the hydrophobic to hydrophilic
the ability of frother to inhibit coalescence (Harris, 1976). groups in a molecule changes its hydrophile-lipophile balance, or
A measure of this bubble size reduction function is the critical HLB. The common calculation of HLB is the Davies method where
coalescence concentration (CCC) (Cho and Laskowski, 2002a). The group numbers are assigned and summed (Davies, 1957; Davies
CCC represents the concentration giving the minimum Sauter and Rideal, 1961). There have been some attempts to link frother
mean bubble diameter (D32). It is determined from the plot of Sau- functions to HLB (Laskowski, 2003; Pugh, 2007; Nesset et al.,
ter mean diameter vs. concentration either graphically (Cho and 2012; Zhang et al., 2012a).
Laskowski, 2002b) or by fitting a function and deriving the The ambition is to provide fundamental input to the selection of
CCC95, the concentration giving 95% reduction in D32 compared frother for a given duty, which otherwise remains a largely empir-
to water alone (Nesset et al., 2007). The CCC is considered the con- ical exercise (Nagaraj and Ravishankar, 2007). Zhang et al. (2012a)
centration at which coalescence is completely prevented. modeled the relationship between CCC95 and HLB for 1-alcohols
The major commercial frothers today are alcohols and polygly- and polypropylene glycols and their ethers by a series of exponen-
cols (Klimpel and Isherwood, 1991; Laskowski, 1998). Alcohol fro- tial functions dependent on the n and m parameters in the struc-
thers (CnH2n+1OH) usually contain a single hydroxyl (OH) group tural formulae. In the case of commercial frothers, Zhang et al.
and are restricted to 5–7 carbons either straight or branch-chained. (2012a) noted the need to first identify the structure (i.e., n and
Polyglycol frothers include polypropylene glycols (H(OC3H6)mOH), m) to use the model to predict CCC95. To identify they suggested
polypropylene glycol alkyl ethers (CnH2n+1(OC3H6)mOH) and poly- using NMR and further suggested 1H NMR might permit HLB to
ethylene glycol alkyl ethers (CnH2n+1(OC2H4)lOH) which form a be replaced by the H-ratio. The H-ratio can be derived from the
large class with varying molecular structure and molecular weight. spectrum to represent the balance of hydrophilic to hydrophobic
parts of a molecule. The purpose of this paper is to determine
⇑ Corresponding author. Tel.: +1 514 398 1452; fax: +1 514 398 4492. the HLB–H-ratio relationship and whether the H-ratio could substi-
E-mail address: jim.fi[email protected] (J.A. Finch). tute for HLB in the previous CCC95–HLB relationships.

http://dx.doi.org/10.1016/j.mineng.2014.02.006
0892-6875/Ó 2014 Elsevier Ltd. All rights reserved.
2 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8

2. Background 3  IH
HLB ¼ ð7Þ
15  IH þ 10  IL
2.1. Bubble size and CCC95
where IH and IL represent the integration area of hydrophilic and
hydrophobic (lipophilic) protons, respectively.
The function used to fit the Sauter mean diameter vs. concen-
Following this work, Berguerio et al. (1978) determined HLB
tration (D32–C) data is (Comley et al., 2002; Nesset et al., 2007):
using 1H NMR for polysorbates and related compounds and found:
D32 ¼ DL þ A  ExpðB  CÞ ð1Þ
60  H
HLB ¼ ð8Þ
where DL is the minimum (limiting) bubble size, A = D0–DL where D0 Hþ2
is D32 in water alone, and B the decay constant. From the fit CCC95
where
is determined which permits a normalized concentration to be
introduced: IH
  H¼ ð9Þ
C Itot
D32 ¼ DL þ A  Exp b  ð2Þ
CCC95 where H is the H-ratio where Itot = IH + IL.
There is an attraction to using NMR in the study of frothers. NMR
where
(proton and carbon) provides the necessary structural information
b ¼ LN0:05=CCC95 ð3Þ and the H-ratio might substitute for HLB. Two issues in using the
H-ratio are: deciding on the chemical shift in the spectrum that di-
Determined at a standard superficial gas velocity (Jg) of 0.5 cm/s
vides the hydrophilic from hydrophobic entities; and recognizing
the CCC95 results for 36 surfactants from the 1-alcohol and poly-
that the relationship to HLB might be ‘family’ dependent.
propylene glycol families were fitted to (Zhang et al., 2012a):
CCC95 ¼ a  Expðb  HLBÞ ð4Þ 3. Experimental
where a and b are constants that depend on the frother family.
Zhang et al. (2012a) showed the limiting bubble size DL also de- 3.1. Apparatus and operation
pended on HLB:
3.1.1. Bubble size determination
DL ¼ 0:072  HLB þ 1:43 ð5Þ The same rig as previously described was used (Zhang et al.,
By predicting CCC95 and DL from HLB (Eqs. (4) and (5), respec- 2009; Zhang et al., 2010). It consists of a mechanical flotation cell
tively) we can estimate b (Eq. (2)) and A (D0–DL) and thus bubble (0.8 m3 in volume) with a McGill Bubble Size Analyser (MBSA) (Go-
size (D32) can be predicted for any C (Eq. (2)). Zhang et al. mez and Finch, 2007) (Fig. 1). The cell was operated with 700 L
(2012a) illustrated the approach for some commercial frothers. water at room temperature (22–24 °C) with air dispersed through
The D32 refers to Jg = 0.5 cm/s but can be scaled to other gas veloc- the 21 cm diameter impeller set at 1500 rpm (tip speed 5.73 m/s)
ities (Nesset et al., 2012). at a superficial air rate (volumetric air rate divided by cell cross-
sectional area) of 0.5 cm/s. Bubbles were sampled 20 cm below
2.2. HLB and H-ratio the water surface, well above the turbulence of the impeller–stator
region. The images were processed using in-house software to give
The Davies’ method calculates HLB from the following: the D32. Further details are given in Zhang et al. (2012a).

HLB ¼ 7 þ Rðhydrophilic group numbersÞ 3.1.2. NMR spectroscopy

þ Rðlipophilic group numbersÞ ð6Þ The NMR analysis was performed using a 400 MHz NMR spec-
trometer (Bruker AVANCE-400) at a frequency of 400.27 MHz for
The group numbers relevant to the surfactants tested here are
protons and with temperature controlled at 298 K. Samples were
given in Table 1. By inspection, a decreasing HLB indicates increas-
ing hydrophobicity.
Various methods have been suggested to estimate HLB directly Variable speed
(Berguerio et al., 1978). Ben-Et and Tatarsky (1972) proposed controller Compressed air
determining HLB using high resolution 1H NMR spectrometry. 1H
NMR is widely used in the qualitative analysis of nonionic surfac-
Mass flow
tants, the characteristic functional groups being readily identified meter
from the spectrum. The method described by Ben-Et and Tatarsky
consisted of determining HLB from the integration area of the
hydrophilic and hydrophobic (lipophilic) proton signals (the total Baffle ring
area under the NMR spectrum represents the total proton signal
of the molecule). The empirical relationship determined for poly-
ethoxylated alkylphenolic compounds was:

Table 1
Group numbers in Davies method of estimating HLB.

Functional group Group contribution number

Hydrophilic
–OH 1.9 McGill Bubble Size
–O– 1.3
Analyzer (MBSA)

Lipophilic (or hydrophobic)

ACHA; CH2A; ACH3A; @CH 0.475
Fig. 1. Top view of Metso RCSTM 0.8 m3 mechanical cell and MBSA (CAD drawing).
 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8 3

Table 2a
Frother families and range of surfactants (n, m, l and HLB) used in the study.

a
HLB numbers obtained by Davies’ method.

Table 2b
Commercial frothers used in the study.

Frother family Commercial frother type Supplier N m l Molecular weight HLB

Aliphatic alcohols FX120-01 Flottec 6 – – 102 6.05
Polypropylene glycol (PPG) F150 Flottec 0 7 – 425 8.625
PolyFroth W31 Huntsman 0 9 – 540 8.375
Polypropylene glycol alkyl ether (PPGAE) DowFroth 200 Dow Chemical 1 3 – 206 7.975
DowFroth 250 Dow Chemical 1 4 – 264 7.83
DowFroth 1012 Dow Chemical 1 6.7 – 398 7.48
FX160-01 Flottec 1 3.8 – 251 7.86
FX160-05 Flottec 3 2.5 – 207 7.11
F160 Flottec 4 2.5 – 217 6.63
Polyethylene glycol alkyl ether (PEGAE) FX120-02 Flottec 6 – 2 190 6.75
NasFroth 240 Nasaco 4 – 3 206 8.05
NovelFroth 234 Sasol 6 – 3 234 7.1

Note structures (n, m, l) determined/confirmed by NMR.

prepared by dissolving 20 mg of reagent in 0.7 mL deuterated chlo- aided by the relatively simple molecules studied and their ready
roform (99.96 atom%, from Sigma Aldrich) and introducing into a dissolution in the deuterated chloroform. The range in chemical
high quality 5 mm NMR (Wilmad 528, limit 500 MHz frequency, shift was only from 0 to 4 ppm with little signal overlap.
L 7 in.) tube. Data acquisition (e.g. mathematical conversion of
NMR frequency signals into interpretable spectra) and integration 4.2. CCC95 and HLB
were performed using the commercial software MestRe-C.
The CCC95 vs. HLB data for the 45 surfactants were organized
3.2. Reagents
into 1-alcohols (Fig. 2), polypropylene glycols and their alkyl ethers
The surfactants (45 in total) from four frother families are iden-
tified in Table 2a. They cover a wide range in n, m and l (and cor- 4.5
responding HLB number). Extra pure analytical grade (98–99.9%
4.0
purity from Aldrich–Sigma) surfactants were used. Several com-
3.5
CCC95 (mmol/L)

mercial frothers (Table 2b) were included to test the relationships

3.0
derived using the pure compounds.
2.5
4. Results 2.0
1.5
4.1. Reliability 1.0
0.5
The precision of CCC95 determination was established in the 0.0
earlier publication (Zhang et al., 2012a): the 95% confidence limit 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
on CCC95 was better than ±0.003 mmol/L. HLB
The 1H NMR measurements showed a precision (relative stan-
dard error) better than 1% (of the integration area under the peaks), Fig. 2. CCC95 vs. HLB for the 1-alcohols (n from 3 to 8).
4 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8

Table 3
Polyethylene glycol alkyl ether HLB numbers obtained by Davies’ method and revised
HLB.

Frother type N L HLB (Davies’ method) HLBrev

m=1
m=2 Ethylene glycol ethyl ether 2 1 8.28 7.95
Ethylene glycol propyl ether 3 1 7.805 7.475
Ethylene glycol butyl ether 4 1 7.33 7
m=1

m=1 2 3 HLBrev ¼ HLB  0:33  l ð10Þ

2 3 4
2
3 4 Fig. 5 shows that using HLBrev the data for polyethylene glycol
3 17 7 7
12 alkyl ethers reduces to a single curve. In this manner all the frother
families can be represented by the exponential equation (Eq. (4)).

4.3. HLB and H-ratio

Fig. 3. CCC95 vs. HLB for the polypropylene glycols (n = 0) and their alkyl ethers (n
from 1 to 4) as function of m for a given n.
How the division into hydrophobic and hydrophilic entities was
established to calculate the H-ratio (Eq. (9)) is illustrated by exam-
ple 1H NMR spectra. The result for 1-hexanol is shown in Fig. 6. The
signals at 0.8, 1.2 and 1.4 ppm are assigned to the two methylenes
n=1 of the alkyl group and the terminal methyl. The signals of the ter-
n=1 minal OH group and one methylene conjoint with the OH are ob-
tained at 3.5 ppm. Since the hydrophobic protons in a straight
alkyl chain appear to the right side (small chemical shift) and all
n=2
other protons are considered as hydrophilic are to the left (large
2
2 chemical shift), the dividing line appears in the vicinity of
3 3 2.5 ppm chemical shift. The H-ratio (0.214) was then calculated
4
6 based on intensities of the signal areas (see Fig. 7).
4 4 The 1H NMR spectrum of tri(propylene glycol) methyl ether
(n = 1; m = 3) is shown in Fig. 8. The signal at 0.9 ppm is assigned
to the three methyls in the PO groups; the signals 3.0–3.7 ppm
are attributed to the methylenes and methanes of the PO groups
Fig. 4. CCC95 vs. HLB for the polyethylene glycol ethers as function of n for a given l. as well as the terminal methyl and OH groups. Again, the hydro-
phobic/hydrophilic division can be taken at 2.5 ppm chemical shift.
(Fig. 3) from Zhang et al. (2012a), and new data on polyethylene The H-ratio (0.591) is calculated accordingly.
glycols and their alkyl ethers (Fig. 4). The 1H NMR spectrum of tri(ethylene glycol) butyl ether (n = 4;
The results for 1-alcohols show a decreasing HLB (i.e., increas- l = 3) is shown in Fig. 8. The signals at 0.8, 1.2 and 1.4 ppm are from
ing hydrophobicity) correlates with decreasing CCC95. The same the two methylenes of the alkyl group and the terminal methyl,
is evident for the polypropylene glycols (n = 0) and their alkyl while the signals from 2.9 to 3.7 ppm are due to the protons in
ethers (n = 1–4), the trends depending on n and m. The data were the three EO groups, one methylene adjacent to the EO and the ter-
fitted to Eq. (4) (Zhang et al., 2012a). minal OH. Using the same dividing 2.5 ppm chemical shift the H-
The trend for polyethylene glycol alkyl ethers (Fig. 5) is superfi- ratio (0.653) is determined.
cially similar to the alcohols, but inspection shows that the value of Using the division at 2.5 ppm chemical shift the H-ratio was
l at a given n is not a factor; that is the number of ethylene oxide determined for the 45 surfactant structures (Table 4). The H-ratio
(EO) groups in polyethylene glycol alkyl ethers does not change data were fitted to HLB by the following relationship:
the CCC95. Since the HLB group number of EO is 0.33 from the Da-
vies method, in order to correlate against HLB we subtract the 0.33
to give a revised HLB (HLBrev), as demonstrated in Eq. (10) and
Table 3. Hydrophilic Hydrophobic
Fragment Fragment

n=1

3
4
5

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0

Chemical Shift (ppm)

Fig. 5. CCC95 vs. HLBrev for the polyethylene glycol ethers as function of n for a Fig. 6. 1H NMR spectrum of 1-hexanol and division into hydrophilic and hydro-
given l. phobic entities (fragments).
 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8 5

4.5. Predicting CCC95 for commercial frothers

Hydrophilic Hydrophobic
Fragment Fragment
The practical development is to validate the model on commer-
cial frothers. Unfortunately, the structure of many (if not most)
commercial frothers is seldom disclosed being identified by trade
names.
The structure can be identified by NMR spectrometry. Fig. 11
shows an example 1H NMR spectrum for one commercial frother,
DowFroth 250. The characteristic signal of four methyls (i.e.,
11.96/3  4) in PO groups can be observed at 0.9 ppm. The signals
of methylenes and methanes of PO groups, one terminal methyl
([15.013  4]/3  1) and one OH can be observed in the range
3.0–3.7 ppm. DowFroth 250 is thus identified as belonging to the
polypropylene glycol methyl ether family, specifically tetra(pro-
pylene glycol) methyl ether (as is known to be the case in this
4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 example). Accordingly the H-ratio (0.57) can be determined.
Chemical Shift (ppm) The identity and H-ratio for the commercial frothers was deter-
mined and CCC95 predicted from Eq. (12) using the parameters in
Fig. 7. 1H NMR spectrum of tri(propylene glycol) methyl ether and division into
Table 5. Fig. 12 confirms the model can be extended to commercial
hydrophilic and hydrophobic entities (fragments).
frothers.

Hydrophilic Hydrophobic 5. Discussion

Fragment Fragment
The general ambition is to predict frother functions such as
bubble size reduction and froth stabilization from structural infor-
mation. This and the previous work (Zhang et al., 2012a) have fo-
cused on the bubble size reduction function. This function is
conveniently quantified by the critical coalescence concentration
CCC or in this case the CCC95. The CCC95 can be related to the
hydrophile-lipophile balance (HLB) provided the frother family is
identified (Nesset et al., 2012; Zhang et al., 2012a). The present
work has extended the database to include polyethylene glycol al-
kyl ethers. It was evident (Fig. 4) that the CCC95 remained unaf-
fected by varying the number of EO groups (l). There is some
4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 support for this finding from Cappuccitti and Finch (2007) who
Chemical Shift (ppm) showed when varying EO from l = 1.5 to 3 in ethoxylated alcohols
that gas holdup, which reflects bubble size, remained almost
Fig. 8. 1H NMR spectrum of tri(ethylene glycol) butyl ether and its corresponding constant.
structure formula. As the measure of structure the hydrophile-lipophile balance
(HLB) is used. Using the Davies method to calculate HLB it is shown
that CCC95 can be related to HLB provided the frother family is
HLB ¼ A  ½1  lnðB  H ratioÞ ð11Þ
identified. The extension here to include polyethylene glycol alky
ethers required a revised HLB number. The major contribution of
where A and B are constants that depend on the family, as given in this work, however, is the substitution of HLB by the H-ratio deter-
Table 5. mined from 1H NMR spectrometry.
Fig. 9 plots HLB from the Davies method against HLB calculated Analysis of the 1H NMR spectra of the surfactants studied
from the H-ratio using Eq. (11) and the parameters in Table 5. Eq. showed the proton signals divide into two band groups: the hydro-
(11) represents an empirical estimate of HLB from the H-ratio phobic protons in the 0–2.5 ppm band (assigned to the CH3 and
appropriate for surfactants of the four frother families. CH2 protons in the alkyl group) being well separated from the
hydrophilic protons in the range of 2.5–4.0 ppm (assigned to the
CH2 and CH protons of the PO unit, CH2 protons of the EO unit,
4.4. CCC95 and H-ratio and CH3 or CH2 conjoint with the O atom as well as terminal
OH). The division was taken at 2.5 ppm chemical shift. Ben-ET
The correlation between HLB and CCC95 is Eq. (4); we can thus and Tatarsky (1972) and Berguerio et al. (1978) located the hydro-
substitute Eq. (11) into Eq. (4) to derive a dependence of CCC95 on phobic part in the 0.5–3 ppm region and the hydrophilic part in the
H-ratio, namely: 3.5–5 ppm region, i.e., a division similar to that employed here.
With the division established at 2.5 ppm the H-ratio is calculated
as the ratio of the integrated area of the hydrophilic fragments
CCC95 ¼ a  Expfb  A  ½1  lnðB  H ratioÞg ð12Þ
(chemical shift >2.5 ppm) relative to the total integrated area (Eq.
The constants a, b, A and B (Table 5) relate to the frother family. (9)). The H-ratio could be related to HLB and substituted for HLB
Statistical analysis (measured CCC95 vs. predicted CCC95) indi- in the model to predict CCC95 (Eq. (12)). From CCC95 the Sauter
cates that the polypropylene glycol data had slightly more variabil- mean bubble diameter (D32) can be estimated for any concentra-
ity than the other families. Overall, the model is judged a good fit tion (Eq. (2)), remembering that this refers to Jg = 0.5 cm/s.
having an R2 in the range from 0.971 to 0.997, with predictions The prediction of CCC95 from the H-ratio was tested on 12 com-
for CCC95 distributed either side of the 1:1 line (Fig. 10). mercial frothers. The structure was first determined using NMR, as
6 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8

Table 4
Summary of data: n, m, l, HLB, CCC95 and H-ratio for the surfactants and commercial frothers.

Frother family Frother type n m l HLB from Measured NMR method

Davies’ CCC95
Ihydrophilic Itot H- Predicted Predicted
methoda (mmol/L)
ratio HLB CCC95
(mmol/L)
1-Alcohol 1-Propanol 3 – – 7.48 3.92 3 8 0.38 7.59 4.41
1-Butanol 4 – – 7.00 0.85 3 10 0.30 6.95 0.79
1-Pentanol 5 – – 6.53 0.29 3 12 0.25 6.42 0.23
1-Hexanol 6 – – 6.05 0.11 3 14 0.21 5.97 0.08
1-Heptanol 7 – – 5.58 0.072 3 16 0.19 5.57 0.058
1-Octanol 8 – – 5.10 0.060 3 18 0.17 5.23 0.029
Polypropylene glycol Propylene glycol methyl ether 1 1 – 8.28 0.48 7 10 0.70 8.39 0.59
alkyl ether Propylene glycol propyl ether 3 1 – 7.33 0.25 6 14 0.43 7.33 0.26
Propylene glycol butyl ether 4 1 – 6.85 0.16 6 16 0.38 6.84 0.15
Di(propylene glycol) methyl ether 1 2 – 8.13 0.18 10 16 0.63 8.07 0.17
Di(propylene glycol) propyl ether 3 2 – 7.18 0.094 9 20 0.45 7.15 0.091
Di(propylene glycol) butyl ether 4 2 – 6.70 0.066 9 22 0.41 6.70 0.062
Tri(propylene glycol) methyl 1 3 – 7.98 0.073 13 22 0.59 7.91 0.076
ether
Tri(propylene glycol) propyl ether 3 3 – 7.03 0.045 12 26 0.46 7.05 0.051
Tri(propylene glycol) butyl ether 4 3 – 6.55 0.029 12 28 0.43 6.58 0.032
Polypropylene Di propylene glycol – 2 – 9.25 0.40 8 14 0.57 9.43 0.53
glycols Tri propylene glycol – 3 – 9.13 0.17 11 20 0.55 9.02 0.14
Tetra propylene glycol – 4 – 9.00 0.088 14 26 0.54 8.81 0.068
Polypropylene glycol 425 – 7 – 8.63 0.014 23 44 0.52 8.49 0.016
Polypropylene glycol 725 – 12 – 8.00 0.0091 38 74 0.51 8.18 0.0077
Polypropylene glycol 1000 – 17 – 7.38 0.0084 44 86 0.51 8.14 0.0066
Polyethylene glycol Ethylene glycol ethyl ether 2 – 1 7.95 0.38 7 10 0.70 7.96 0.38
alkyl ether Ethylene glycol propyl ether 3 – 1 7.48 0.22 7 12 0.58 7.45 0.20
Ethylene glycol butyl ether 4 – 1 7.00 0.12 7 14 0.50 7.01 0.11
Di(ethylene glycol) methyl ether 1 – 2 8.43 0.70 12 12 1.00 8.62 0.78
Di(ethylene glycol) ethyl ether 2 – 2 7.95 0.36 11 14 0.79 7.79 0.31
Di(ethylene glycol) propyl ether 3 – 2 7.48 0.21 11 16 0.69 7.33 0.19
Di(ethylene glycol) butyl ether 4 – 2 7.00 0.11 11 18 0.61 6.93 0.10
Di(ethylene glycol) hexyl ether 6 – 2 6.05 0.068 11 22 0.50 6.23 0.056
Tri(ethylene glycol) methyl ether 1 – 3 8.43 0.68 16 16 1.00 8.50 0.71
Tri(ethylene glycol) ethyl ether 2 – 3 7.95 0.34 15 18 0.83 7.81 0.32
Tri(ethylene glycol) butyl ether 4 – 3 7.00 0.11 15 22 0.68 7.06 0.12
Commercial frothers FX120-01 6 – – 6.05 0.10 3 14 0.21 5.97 0.09
FX160-01 1 3.8 – 7.86 0.048 15 27 0.56 7.73 0.036
DowFroth 200 1 3 – 7.98 0.084 13 22 0.59 7.91 0.076
DowFroth 250 1 4 – 7.83 0.038 16 28 0.57 7.81 0.045
DowFroth 1012 1 6.7 – 7.48 0.014 24 45 0.53 7.60 0.017
FX160-05 3 2.5 – 7.11 0.074 10 22 0.45 7.11 0.072
F160 4 2.5 – 6.63 0.037 14 33 0.42 6.65 0.043
F150 – 7 – 8.63 0.014 23 44 0.52 8.49 0.016
PolyFroth W31 – 9 – 8.38 0.0099 29 56 0.52 8.37 0.011
FX120-02 6 – 2 6.05 0.066 11 22 0.50 6.23 0.056
NasFroth 240 4 – 3 7.00 0.11 15 22 0.68 7.06 0.12
NovelFroth 234 6 – 3 6.05 0.067 15 26 0.58 6.17 0.054
a
Note, the HLB for polyethylene glycol alkyl ethers and the corresponding commercial frothers are the revised values.

Table 5
Estimated constants in Eqs. (11) and (12) and their statistical characteristics.

Family aa ba A B Precision
Data points, N R2 Std error (±mmol/L)
1-Alcohol 1.52E10 3.207 2.917 97.94 6 0.982 0.064
Polypropylene glycol 4.76E17 3.951 10.804 4.191 6 0.971 0.039
Polypropylene glycol methyl ether 1.61E18 4.855 2.846 74.05 6 0.989 0.025
Polypropylene glycol propyl ether 3.15E19 5.624 3.895 0.965 4 0.997 0.0061
Polypropylene glycol butyl ether 9.58E20 6.125 1.697 0.129 4 0.997 0.0035
Ethylene glycol alkyl ether 1.37E5 1.286 2.817 65.57 3 0.997 0.011
Di(ethylene glycol) alkyl ether 3.435 33.41 5 0.986 0.040
Tri(ethylene glycol) alkyl ether 4.355 19.45 3 0.995 0.031
a
The a and b are from Zhang et al. (2012a).

illustrated for DowFroth 250 for which the structure is known. A (Fig. 12). The model limitation noted previously that HLB is the
good correlation against measured CCC95 was demonstrated same for isomers but CCC95 varies (Zhang et al. 2012a) remains.
 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8 7

Fig. 12. Comparison of measured vs. model predicted CCC95 for 12 commercial
Fig. 9. HLB (Davies) vs. HLB calculated from 1H NMR for all 45 surfactants. frothers with varying structures.

et al., 2012b, 2013). Altogether these applications suggest in-

creased use of NMR in frother-related studies.

6. Conclusions

This paper correlates critical coalescence concentration (CCC95)

against the hydrophile lipophile balance (HLB) in a series of expo-
nential functions for four frother families representing a total of 45
surfactants. To extend to polyethylene glycol alkyl ethers a revised
HLB was introduced. It is shown that H-ratio which quantifies the
fraction of hydrophilic component in the molecule from the 1H
NMR spectrum could substitute for HLB. The resulting CCC95–H-
ratio model permits prediction of CCC95 as illustrated for 12 com-
mercial frothers.

Acknowledgements
Fig. 10. Comparison of measured vs. model predicted CCC95 for surfactants.

The authors would like to thank Metso Minerals for providing

the 0.8 m3 RCS™ flotation cell unit for testing and the chemical
companies for supplying the commercial frothers. The authors
acknowledge financial support from two sources: the Chair in Min-
eral Processing sponsored by Vale, Teck Resources, Barrick Gold,
Xstrata Process Support, Shell Canada, SGS Lakefield, COREM and
Flottec; and from the AMIRA International P9 project both under
the NSERC (Natural Sciences and Engineering Research Council of
Canada) CRD (Collaborative Research and Development) program.
Special thanks to Siying Zhu for her assistance in the NMR tests
and Zao Chen for his assistance in making the mechanical
drawings.

References

Ben-Et, G., Tatarsky, D., 1972. Application of NMR for the determination of HLB
values of non-ionic surfactants. J. Am. Oil Chem. Assoc. 49, 499–500.
Berguerio, J.R., Bao, M., Casares, J.J., 1978. Determination of HLB of non-ionic
4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 dispersants by NMR. Anal. Quim. 74, 529–530.
Cappuccitti, F., Finch, J.A. 2007. Development of new frothers through
Chemical Shift (ppm) hydrodynamic characterization of frother chemistry. In: Proceedings of 39th
Annual Meeting of Canadian Mineral Processors, Ottawa, pp. 399–412.
Fig. 11. 1H NMR spectrum of DowFroth 250. Cho, Y.S., Laskowski, J.S., 2002a. Effect of flotation frothers on bubble size and foam
stability. Int. J. Miner. Process. 64 (2–3), 69–80.
Cho, Y.S., Laskowski, J.S., 2002b. Bubble coalescence and its effect on bubble size and
foam stability. Can. J. Chem. Eng. 80, 299–305.
Comley, B.A., Harris, P.J., Bradshaw, D.J., Harris, M.C., 2002. Frother characterisation
NMR does offer the researcher a means to include commercial using dynamic surface tension measurements. Int. J. Miner. Process. 64, 81–100.
frothers, for which the structure is often unknown, in generating Davies, J.T., 1957. In: Proc. 2nd Int. Congr. Surface Activity, vol. 1, London, p. 426.
Davies, J.T., Rideal, E.K., 1961. Interfacial Phenomena. Academic Press, New York, p.
correlations. The methodology could find use to predict the
371.
CCC95 of new frother formulations reducing the need for experi- Gomez, C.O., Finch, J.A., 2007. Gas dispersion measurements in flotation cells. Int. J.
mentation (which is quite extensive to determine CCC95) and Miner. Process. 84, 51–58.
may open up new frother types. In a forthcoming paper some Harris, C.C., 1976. Flotation machines. In: Fuerstenau, M.C. (Ed.), A.M.Gaudin
Memorial Volume, vol. 2. SME of AIME, pp. 753–815.
new structures will be examined. We have also used NMR in quan- Klimpel, R., Isherwood, S., 1991. Some industrial implications of changing frother
titative analysis of frothers, specifically dual frother blends (Zhang chemical structure. Int. J. Miner. Process. 33, 369–381.
8 J.A. Finch, W. Zhang / Minerals Engineering 61 (2014) 1–8

Laskowski, J.S., 1998. Frothers and frothing. In: Laskowski, J.S., Woodburn, E.T. Pugh, R.J., 2007. The physics and chemistry of frothers. In: Fuerstenau, M.C.,
(Eds.), Frothing in Flotation II. Gordon and Breach, New York, pp. 1–50. Jameson, G., Yoon, R.H. (Eds.), Froth Flotation: A Century of Innovation. SME
Laskowski, J.S., 2003. Fundamental properties of flotation frothers. In: Proceeding of Publications, pp. 259–281.
the 22nd International Mineral Processing Congress, Cape Town, South Africa, Zhang, W., Kolahdoozan, M., Nesset, J.E., Finch, J.A., 2009. Use of frother with
pp. 788–797. sampling-for-imaging bubble sizing technique. Miner. Eng. 22 (5), 513–515.
Nesset, J.E., Finch, J.A., Gomez, C.O., 2007. Operating variables affecting the bubble Zhang, W., Nesset, J.E., Finch, J.A., 2010. Water recovery and bubble surface area flux
size in forced-air mechanical flotation machines. In: Proceedings AusIMM 9th in flotation. Can. Metall. Q. 49 (4), 353–362.
Mill Operators’ Conference, Fremantle, Australia, pp. 55–65. Zhang, W., Nesset, J.E., Rao, S.R., Finch, J.A., 2012a. Characterizing frothers through
Nesset, J.E., Zhang, W., Finch J.A. 2012. A benchmarking tool for assessing flotation critical coalescence concentration (CCC95)-hydrophilic-lipophilic balance
cell performance. In: Proceedings 2012 – 44th Annual Meeting of the Canadian (HLB). Minerals 2 (3), 208–227.
Mineral Processors (CIM), Ottawa, January 17–19, pp. 183–209. Zhang, W., Zhou, X., Finch, J.A., 2012b. Determining independent control of dual-
Nagaraj, D.R., Ravishankar, S.A., 2007. Flotation reagents – a critical overview from frother systems – gas holdup, bubble size and water overflow rate. Miner. Eng.
an industry perspective. In: Fuertenau, M.C., Jameson, G., Yoon, R.-H. (Eds.), 39, 106–116.
Froth Flotation: A Century of Innovation. Society for Mining, Metallurgy, and Zhang, W., Zhu, S.Y., Finch, J.A., 2013. Frother partitioning in dual-frother systems:
Exploration, Inc., Littleton, p. 375. development of analytical technique. Int. J. Miner. Process. 119, 75–82.