Soft Matter: Paper

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Soft Matter

View Article Online


PAPER
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

View Journal | View Issue

Effects of protonation on foaming properties of


dodecyldimethylamine oxide solutions: a pH-
Cite this: Soft Matter, 2015, 11, 561
study†
Kathrin Schellmann,a Natalie Preisig,a Per Claessonb and Cosima Stubenrauch*a

The critical micelle concentration (cmc), the surface excess (G), as well as the micelle aggregation number
(m) of the surfactant dodecyldimethylamine oxide (C12DMAO) have been reported to strongly depend on
the pH-value of the aqueous surfactant solution. At high ionic strength, the cmc displays a minimum,
while both G and m have a maximum at a pH-value close to the pKa of the surfactant. These
experimental observations have been explained as being due to specific hydrogen bonds between the
head groups, which are formed once the surfactant is partly or fully protonated. This investigation
addresses the question of whether the pH also affects the foaming properties of C12DMAO solutions. To
answer this question we measured the foamability and the foam stability of C12DMAO solutions at a fixed
C12DMAO concentration of 5 cmc for five different pH-values, namely pH ¼ 2, 3, 5, 8, and 10. We found
that the foamability is hardly affected by the pH-value, while the foam stability strongly depends on the
pH. As is the case for the above mentioned properties, the foam stability also displays an extremum in
Received 7th November 2014
Accepted 17th November 2014
the studied pH-range, namely a maximum at pH ¼ 5. We discuss our results in terms of the hydrogen
bond hypothesis and show that this hypothesis indeed is in line with the observed trend for the foam
DOI: 10.1039/c4sm02476a
stability. Moreover, we discuss that hydrogen bond formation may rationalize how the molecular
www.rsc.org/softmatter structure of a surfactant affects foam stability.

surfactant mixture consisting of cationic and non-ionic species.


1. Introduction Thus, we can consider the pH-dependent properties of
The surfactant dodecyldimethylamine oxide (C12DMAO) is C12DMAO in the same way as properties of surfactant mixtures
predominantly non-ionic in aqueous alkaline solutions (ref. 1 which are of great importance in all surfactant-related elds.2 It
and references therein). A decrease of the pH-value leads to has been found that typical surfactant properties such as the
protonation of the oxygen and thus renders the surfactant critical micelle concentration (cmc), the maximum surface
cationic (C12DMAOH+). Since the protonated and the uncharged concentration (Gmax) and the plateau value of the surface
molecules are in equilibrium, both species are present over a tension (scmc) change as a function of the mixing ratio. On the
large pH-range. This equilibrium is shied towards the assumption of ideal mixing, which is almost always a reason-
protonated form with decreasing pH-value. Since the chemical able approximation for mixtures of classical low molecular
structure of the head group is almost identical for the cationic weight surfactants, the cmc-values and the scmc-values change
and the non-ionic species, the effect of the charge (here: of the continuously from the values of surfactant 1 to those of
protonation) on typical surfactant properties can be studied surfactant 2 (see e.g. data in ref. 3–5 by way of example). For the
systematically without being inuenced by other parameters Gmax-values a clear trend is normally not observed. In contrast,
such as different head group sizes and compositions. the pH-dependence of these characteristic properties for
Changing the pH-value of a dodecyldimethylamine oxide C12DMAO is different and the above mentioned properties have
solution has the same effect as changing the ratio of a an extremum close to where 50% of the surfactant is proton-
ated. This extremum is a minimum in the case of cmc and scmc,
while it is a maximum in the case of Gmax.1 Note that this
a
Universität Stuttgart, Institut für Physikalische Chemie, Pfaffenwaldring 55, 70569 behaviour is only found when the ionic strength is 0.1 M (or
Stuttgart, Germany. E-mail: cosima.stubenrauch@ipc.uni-stuttgart.de; Tel: higher) and constant. At a low ionic strength electrostatic forces
+49-711-685-64470 predominate and in this case the cmc decreases with increasing
b
KTH Royal Institute of Technology, School of Chemical Science and Engineering, pH, i.e. with increasing abundance of the cationic species.6
Department of Chemistry, Division of Surface and Corrosion Science, Drottning
To explain these extrema Maeda et al. suggested attractive
Kristinas väg 51, Stockholm, SE-100 44 Sweden
† Electronic supplementary information (ESI) available. See DOI:
hydrogen bonds between non-ionic and cationic species as well
10.1039/c4sm02476a as between two cationic species.1,7,8 The presence of such

This journal is © The Royal Society of Chemistry 2015 Soft Matter, 2015, 11, 561–571 | 561
View Article Online

Soft Matter Paper

hydrogen bonds appear to explain the pH-dependence of the processes, which we discuss in the light of the hydrogen bond
surface properties observed for C12DMAO in literature1,7–9 and hypothesis suggested by Maeda et al.
in the study at hand. The main idea is that with decreasing pH-
value repulsive and attractive interactions act simultaneously.
The repulsion is due to the charges and is thus of electrostatic 2. Experimental section
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

nature, while the attraction is a result of hydrogen bonds 2.1. Materials and cleaning procedure
between the molecules, for which, in turn, protonation of a The dodecyldimethylamine oxide (C12DMAO) was purchased
fraction of the head groups is the prerequisite. These two from Sigma Aldrich and puried by recrystallization three times
counteracting forces lead to a maximum in surface excess at from pure acetone (p.a. 99.9%, from VWR). Sodium chloride
around 50% protonation (pH  4.6): from pH ¼ 10 down to pH (NaCl) was obtained from Merck and roasted at 500  C over
¼ 5 the attraction created by the hydrogen bonds increases night to drive off organic contaminants. 1 M HCl and 0.1 M
until, at pH < 5, the electrostatic repulsion becomes increasingly NaOH aqueous solutions from Scharr and Merck, respectively,
important. were used as received to adjust the pH-values of the C12DMAO
Motivated by the strong pH-effect on the surface properties solutions. The pH adjustments of the C12DMAO solutions were
of C12DMAO in aqueous solutions we endeavoured to investi- monitored with a pH-Meter pH 197 s from WTW with 0.05
gate whether the foaming properties are also pH-dependent. If accuracy. The pH-values were controlled before and aer
so, would the foamability and the foam stability pass through experiments. Please note that under atmospheric conditions pH
an extremum at 50% ionization? The answer to this question is values higher than 6 are no equilibrium values. The dissolution
not straightforward since there are no direct correlations of CO2 in the aqueous solution leads to a continuous decrease
between the surface properties mentioned above and foaming of the pH value down to 6 as a function of time. For example,
properties. A prominent example of this lack of correlation is the value of pH ¼ 10 decreased down to pH ¼ 8.95 in 3 h. This
the foaming behaviour of the two non-ionic surfactants n- difficulty is not addressed in Maeda's work which is why the
dodecyl-b-D-maltoside (b-C12G2) and hexaethyleneglycol mono- data reported for pH > 6 should be looked at carefully. How we
dodecyl ether (C12E6).10–13 Although C12E6 has a lower cmc and a dealt with this point is explained in Section 2.2 and 2.3,
lower scmc-value compared to b-C12G2, the stability of foams respectively. All solutions were prepared with double distilled
stabilized by C12E6 is several orders of magnitude lower water and contained 0.1 M NaCl to minimize the change of the
compared to that of foams stabilized by b-C12G2. Furthermore, ionic strength of the solution with varying pH. All glassware was
we know from foam studies on this non-ionic mixture as well as cleaned before use with deconex UNIVERSAL 11 from Borer
on mixtures of non-ionic and cationic surfactants that some Chemie and rinsed several times with distilled water to remove
foam properties are unaffected by the mixing ratio while others impurities.
change continuously as a function of the composition.10,13–15‡
Again, this continuous change is not necessarily connected with
the surface properties. The only direct correlation between 2.2. Surface tension measurements
surface and foam properties that has been observed up to now is In order to discuss our results in the light of already published
between surface elasticity (or surface rigidity) and foam data we remeasured the surface tension isotherms at 22  C at
stability, where a few systematic studies have shown that the ve different pH-values utilizing the Du Noüy ring method and a
higher the surface elasticity (or surface rigidity) the better the STA1 tensiometer from Sinterface Technologies. For each pH-
foam stability.11–13 value a 100 mL ask of a stock solution with 0.1 M NaCl and a
The question this paper addresses is whether or not the pH- C12DMAO concentration above the cmc was prepared. Each
value has an effect on the foaming properties of aqueous time at least 5 mL solution for washing the glassware and 10 mL
C12DMAO solutions. In case it has an effect, how does it affect solution for a measurement was used. Aer each measurement
foamability and foam stability? To answer this question we rst the ask was lled up with 0.1 M NaCl solution with the same
measured the foamability and the initial liquid fractions 30 of pH-value and the diluted solution was used for the next
foams which were generated with aqueous solutions of measurement.
C12DMAO at a concentration of c ¼ 5 cmc for ve different pH- In Fig. 1 the surface tensions s of all C12DMAO solutions in
values, namely pH ¼ 2, 3, 5, 8, and 10. Aer the foams were the presence of 0.1 M NaCl are presented as a function of the
generated, we investigated the time evolution of the foam total surfactant concentration c for pH ¼ 2, 3, 5, 8, and 10.
volume, of the liquid fraction, and of the bubble size distribu- Measuring the pH values before and aer the experiments
tion. From these measurements we extracted pieces of infor- revealed that the drop of the pH due to dissolution of CO2 can
mation on the foam stability and on the crucial destabilizing be neglected in case of pH ¼ 8. However, this was not the case
for pH ¼ 10 which is why we measured the surface tension only
down to a surfactant concentration of c ¼ 3.5  104 mol L1.
The lower the surfactant concentration the longer it takes to
‡ To the best of our knowledge there is only one study where one foam property,
reach the equilibrium value. For pH ¼ 10 we observed a
namely the foam stability, runs through a maximum if one changes the mixing
ratio.14 In this particular case charge neutralisation leads to the formation of a
decrease, a plateau and a further decrease of the s-value which
Newton black lm, which is a sign of high lm stability and thus of high foam told us that the surface tension decreases because of surfactant
stability. adsorption and a drop of the pH. We thus stopped measuring at

562 | Soft Matter, 2015, 11, 561–571 This journal is © The Royal Society of Chemistry 2015
View Article Online

Paper Soft Matter


Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

Fig. 1Surface tensions s of C12DMAO solutions in presence of 0.1 M NaCl as function of the total surfactant concentration c for five different
pH-values, namely pH ¼ 2, 3, 5, 8, and 10. Solid lines represent second order polynomial fits of the experimental data.

the lowest concentration at which we still observed a plateau 2.3. Foam measurements
value before the decrease of the pH comes into play and further
The foams generated from the C12DMAO solutions were studied
decreases s. The solid lines in Fig. 1 represent second order
using the commercially available FoamScan device from TECLIS
polynomial ts of the experimental s(c)-data. From these ts we
(France, http://www.teclis.eu). Since the foams are generated
calculated the surface concentration according to with N2 we assumed that the pH does not change signicantly
 
1 ds during the foam measurements (see Section 2.1 for the further
G¼ : (1)
RT d ln c p;T details). The FoamScan uses conductivity measurements and
image analysis to monitor the volume and the liquid fraction of
the foam. Additionally, the bubble sizes and bubble size
The surfactant was treated as being non-ionic for the calcu- distributions can be analyzed with the cell size analysis (CSA)
lation of G, which is appropriate since, due to the presence of function. For that purpose, the number and size of bubbles
0.1 M NaCl, the concentration of the surfactant counterion [Cl] found in an area of 20 mm2 were determined and the poly-
remains essentially constant as the surfactant concentration is dispersity index PI, i.e. the normalized standard deviation, was
increased.1 determined according to
In Fig. 2 we summarize the main outcome of the surface qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
tension measurements and of the respective calculations. A hr2 i  hri
comparison with the data published by Maeda et al.1 shows the PI ¼ ; (2)
hri
same overall trend but slight deviations as regards the absolute
values of the cmc and of Gmax, while the scmc-values are in where hri is the arithmetic mean bubble radius.
perfect agreement. As regards the cmc, the differences can be A detailed description of the FoamScan method can be
explained by the total number of experimental data points, found in ref. 5, 13, 16 and 17. Briey, the FoamScan device
which are much less in Maeda's work (see e.g. Fig. 3 in ref. 7) consists of a rectangular glass column with (a) ve pairs of
compared to ours. Moreover, we extracted the cmc from the electrodes to measure the foam conductance and (b) four glass
crossover point between the scmc-value and the polynomial, prisms where pictures of the foam can be taken using one of the
while Maeda et al. most likely extracted it visually from the s(c)- two CCD cameras. To generate the same foam volume Vfoam,
curve. The discrepancies for the Gmax-values can be explained by nitrogen is sparged at a constant gas ow rate Q through a
the way they are calculated: we calculated them from the poly- porous disk (pore sizes ¼ 41–100 mm) which is located at the
nomial (gray squares ( ) in Fig. 2) since this is more accurate, bottom of the glass column. The column contains an initial
while Maeda et al. most likely used the maximum slope of the solution volume Vinitial from which the foam is generated. Aer
s(c)-curve drawn between two subsequent measuring points.§ stopping the gas input, which will be referred to as t ¼ 0 s in the
That this assumption is very likely can been seen if one looks at following, the foam conductance at xed positions along the
the data we obtained by using the maximum slope (open column is measured. From these values the liquid fraction 3 in
squares ( ) in Fig. 2) which are in perfect agreement with the foam is determined according to Feitosa's equation.18 The
Maeda's values. Finally, Maeda et al. did overlook the pH drop changes of the foam volume are detected with the second CCD
for pH values larger than 6. camera.
In all our foam measurements a foam volume Vfoam of 80 mL
was generated from an initial solution volume Vinitial of 40 mL.
To study the foamability of the C12DMAO solutions the gas ow
§ Note that neither the way they determined the cmc-value nor the way they rate Q was varied from 20 to 200 mL min1, while a constant gas
calculated the Gmax-value is explicitly mentioned in any of Maeda's papers.

This journal is © The Royal Society of Chemistry 2015 Soft Matter, 2015, 11, 561–571 | 563
View Article Online

Soft Matter Paper


Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

Fig. 3 Foamability (top) and initial liquid fraction 30 (bottom) as a


function of the N2 flow rate Q at five different pH-values, namely pH ¼
2, pH ¼ 3, pH ¼ 5, pH ¼ 8, and pH ¼ 10. The solid line (top) is calculated
according to t ¼ V/Q.

85 mm (25 mm above the liquid level) and Hprism ¼ 105 mm


(45 mm above the liquid level), respectively.
All presented foam measurements were reproduced two
times. For the sake of clarity, only one experiment is shown in
the Result section, while both measurements are plotted in
Fig. S1–S5 of the Section ESI† to illustrate the reproducibility of
Fig. 2 Critical micelle concentration cmc (top), plateau values of the the FoamScan method. All measurements were performed at
surface tension scmc (middle) and the maximum surface excess Gmax room temperature (22  C). The freeware program ImageJ was
(bottom) as a function of the pH in presence of 0.1 M NaCl. Data are used to treat the foam pictures before using the cell size analysis
taken from ref. 1 ( ) and from this work ( , ). The grey squares (CSA) function as described in ref. 13 and 17.
represent values calculated from the derivative of the polynomial fit
and the open squares are values obtained by a straight line fit between
the two last points prior to cmc.
3. Results
We systematically studied the inuence of the pH-value on the
ow rate of 50 mL min1 was used to produce foam for the foam properties of foams stabilized by C12DMAO at a surfactant
stability measurements. The conductivity measurements and concentration of c ¼ 5 cmc. All measurements were carried out
cell size analysis determination of the bubble size and bubble at ve different pH-values, namely 2, 3, 5, 8, and 10. Firstly, the
size distribution were carried out simultaneously at xed posi- foam generation was investigated by varying the N2 gas ow rate
tions along the glass column, namely, at the second electrode Q over a broad range. For that purpose, we measured the
and the second prism of the CCD camera at heights of Helec2 ¼ foamability and the initial liquid fraction 30 as a function of Q
for the ve pH-values (Section 3.1). Secondly, we monitored the

564 | Soft Matter, 2015, 11, 561–571 This journal is © The Royal Society of Chemistry 2015
View Article Online

Paper Soft Matter

time evolution of both the foam volume and the liquid fraction In conclusion one can say that at moderate ow rates up to
to gain information on the foam destabilizing processes in 100 mL min1 neither the foamability nor the initial liquid
general and about drainage in particular (Section 3.2). Finally, fraction are strongly affected by the pH-value in a range from pH
the time evolution of the bubble sizes and the bubble size ¼ 2–8. For pH ¼ 10, however, we found a slightly lower foam-
distributions were investigated, which allowed us to identify ability and – connected with this – lower 30-values for all ow
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

coalescence events (Section 3.3). rates.

3.2. Time evolution of foam volume and liquid fraction


3.1. Foamability and initial liquid fraction Aer having studied the foam generation process, we investi-
gated the time evolution of the foam volume and of the liquid
In order to examine the inuence of the gas ow rate on
fraction aer turning off the gas ow. For these studies, the
foamability,{ the time required to produce a foam volume of V
foams were always generated with the same constant N2 gas
¼ 80 mL from a given surfactant solution was measured at
ow rate of Q ¼ 50 mL min1. In Fig. 4 (le) the foam volume V
different N2 ow rates Q (20–200 mL min1) at a surfactant
is shown as a function of time t. The time t ¼ 0 s is the time at
concentration of c ¼ 5 cmc. The results are shown in Fig. 3 (top).
which the preset foam volume of 80 mL was reached and where
On the assumption that all the injected N2 gas remains trapped
the N2 ow was stopped. Thus, for t < 0 s foam generation takes
in the foam and that the liquid content of the foam is negligibly
place. As can be seen, the V(t)-curves are very different except for
small, one can approximate the time t it takes to generate a
the two curves measured at pH ¼ 8 and 10 which are hardly
foam volume V as t ¼ V/Q. This relation is a mass balance for the
distinguishable. The volume of the foams generated at pH ¼ 2,
gas, and it is shown as a solid line in Fig. 3 (top). The agreement
3 and 5 decreases only slightly over a certain period of time
between the theoretical curve and the measured points allows
before it starts decaying quickly at t  1200 s for pH ¼ 2, at t 
us to conclude that close to all the N2 indeed is located in the
1500 s for pH ¼ 3 and at t  1800 s for pH ¼ 5. On the contrary,
foam bubbles.k Looking at the data more closely one sees that
the volume of the foams generated at pH ¼ 8 and pH ¼ 10,
the foamability at pH ¼ 2, 3, 5, and 8 is the same except at the
respectively, already decays very quickly aer some tens of
lowest ow rate of Q ¼ 20 mL min1 where it can be clearly seen
seconds and the foams are completely destroyed aer t  600 s.
that the foamability increases in the following order: pH ¼ 8 <
The corresponding time evolution of the foam's liquid frac-
pH ¼ 2 < pH ¼ 3 < pH ¼ 5 (the time it took to generate 80 mL
tion 3 is shown in Fig. 4 (right). As discussed in relation to Fig. 3,
decreases in this order).** However, the foamability of the
the foams generated at pH ¼ 2, 3, 5, and 8 have similar initial
solution with pH ¼ 10 turned out to be lower than those of the
liquid fractions of 30  0.06–0.09, while for pH ¼ 10 the initial
other solutions at all ow rates. For the following discussion we
liquid fraction is considerable lower (30  0.04). The resulting
will only refer to the data obtained for Q ¼ 50 mL min1 and will
3(t)-curves reect the differences observed for the V(t)-curves.
regard the foamability to be independent on the pH in a range
For pH ¼ 2, 3, and 5 the liquid drains out continuously during
from pH ¼ 2 to pH ¼ 8 and to be slightly lower at pH ¼ 10.
the rst 1800 s (30 min), which leads to a continuous decrease
To gain further insight into the foam generation the initial
of the liquid fraction from 30  0.09 to 3  0.01. For pH ¼ 8 and
liquid fraction 30 was determined at the second electrode
10, however, a continuous drainage is observed only up to 250 s
(middle of the 80 mL foam) and plotted as a function of the ow
aer which the foam collapses so that one no longer can speak
rate Q in Fig. 3 (bottom). As expected the initial liquid fraction 30
of drainage.
increases with increasing ow rate from around 3% (for Q ¼ 20
In conclusion one can say that the combination of the V(t)-
mL min1) to 24% (for Q ¼ 200 mL min1). This is due to foam
curves and the 3(t)-curves allows us to conclude that the foam
drainage during its generation. The slower the foam is gener-
stability decreases in the following order: pH ¼ 5 > pH ¼ 3 > pH
ated, the more time it has to drain before the predetermined
¼ 2 [ pH ¼ 8  pH ¼ 10. These differences are mainly due to
nal foam volume is reached and the liquid fraction is
coalescence which increases in this order and which is much
measured. The 30(Q)-curves measured at pH ¼ 2, 3, 5, and 8 are
more pronounced at pH ¼ 8 and 10 compared to the other pH
essentially the same if one takes into account the huge scatter of
values as will be shown in the following Section.
the data. However, for pH ¼ 10 a lower initial liquid fraction was
found at all ow rates which is in line with the lower foam-
ability. In other words, the foam is more difficult to generate 3.3. Time evolution of bubble size and bubble size
since it is very unstable, which, in turn, explains the lower distribution
capacity for water uptake. Although this point is oen overlooked, all foam properties
depend on the bubble size and the bubble size distribu-
{ Foamability describes the ability of a surfactant solution to produce a certain tion.13,17,19 However, reliable and straightforward determination
foam volume in a given period of time and is therefore an important foam of bubble sizes represents an on-going challenge. Our way to
property.
measure the bubble size and the respective distribution is
k Note that the relation t ¼ V/Q also describes the results at the highest ow rate
explained in ref. 13 and 17. The results for foams stabilized by
although the initial liquid fraction at this Q is as high as 24% and thus no longer
negligible. C12DMAO at the ve different pH-values are shown in Fig. 5 and
** At low gas ow rates the error bars are larger due to the partial foam collapse Table 1 as a function of time. The pictures in Fig. 5 clearly reect
that occurs during foam generation. the decrease of the amount of liquid (black parts) with time

This journal is © The Royal Society of Chemistry 2015 Soft Matter, 2015, 11, 561–571 | 565
View Article Online

Soft Matter Paper


Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

Fig. 4 Time evolution of the foam volume V (left) and of the liquid fraction 3 (right) at five different pH-values, namely pH ¼ 2, pH ¼ 3, pH ¼ 5, pH
¼ 8, and pH ¼ 10. The gas flow rate used for foam generation was Q ¼ 50 mL min1.

caused by drainage (see Section 3.2) and they are consistent (d) Show the same drainage behaviour for pH ¼ 2, 3, 5 and a
with the V(t)- and the 3(t)-curves discussed in Section 3.2. The different one for pH ¼ 8 and 10.
foams generated at pH ¼ 2, 3, and 5 evolve in the same way up to (e) Have very different stabilities at pH ¼ 2, 3, 5 and 8, while
400 s, while the foams generated at pH ¼ 8 and 10 start the stability at pH ¼ 10 is the same as that observed for pH ¼ 8.
collapsing via coalescence already aer 200 s. The images
shown in Fig. 5 directly visualize the results seen in Fig. 4,
namely the drainage behavior as well as the sudden foam 4. Discussion
collapse.
4.1. Effect of pH-value on foaming properties of aqueous
In Table 1 we report the arithmetic mean bubble radius hri
C12DMAO solutions
and the polydispersity index PI for the ve pH-values at three
different times t. It can be seen that at t ¼ 0 s the mean radii hri Let us rst recall the main question this paper addresses before
for all foams are approximately the same. Recalling that the we discuss whether our experimental data help to answer this
initial liquid fraction 30 is also comparable (except for pH ¼ 10) question. Does the pH-value have an effect on the foaming
one can conclude that at the outset of the experiments the properties of aqueous C12DMAO solutions? The answer to this
structure of the foams generated at pH ¼ 2, 3, 5 and 8 is very question depends on the range of pH-values and on the foam
similar. The mean radii hri and the polydispersity index PI property investigated. We thus analysed the effect of the pH on
increase for pH ¼ 2, 3, and 5 within the rst 400 s. This (a) foamability, (b) drainage, and (c) foam stability.
observation indicates that drainage and coalescence events (for (a) We found that the foamabilities at moderate ow rates
coarsening the time scale is too short) take place simulta- (up to 100 mL min1) are comparable at pH ¼ 2, 3, 5 and 8,
neously with drainage being the dominant mechanism up to t  while the foamability at pH ¼ 10 is slightly lower. However, the
1200–1800 s. A comparison with the foams generated at pH ¼ 8 differences are small as reected in the fact that the measured
and 10 shows that the latter are very unstable. According to the data points all fall close to the same theoretical curve (see
data in Table 1 the PI increases signicantly in the rst 200 s Fig. 3). In other words, all foams are stable during foam
while the average bubble size remains the same. The data generation. An example of signicant differences in foamability
shown in Fig. 4 and the pictures seen in Fig. 5 reveal that foam is presented in one of our previous work (Fig. 3 in ref. 13). The
collapse is also due to both drainage and coalescence but here reason for the observation that the pH-value has no signicant
coalescence becomes the dominant mechanism as early as 200 s inuence on the foamability is the fact that foamability is
aer foam generation. mainly affected by the diffusion rate of surfactant molecules to
The main results of our foam study can be summarized as the surface, which, in turn, depends on the molecular weight
follows: foams generated with an aqueous solution of C12DMAO and on the concentration of the surfactant. These two quanti-
at a concentration of 5 cmc and at ve different pH-values, ties do not change signicantly with pH so very similar foam-
namely pH ¼ 2, 3, 5, 8, and 10 abilities are expected. In general, foamabilities of low molecular
(a) Have comparable foamabilities at moderate ow rates (up weight surfactants do not differ very much if the generated foam
to 100 mL min1) at pH ¼ 2, 3, 5 and 8, while the foamability at remains stable during the generation process, which is the case
pH ¼ 10 is slightly lower. for the foams generated in this study.††
(b) Have comparable initial liquid fractions 30 at pH ¼ 2, 3, 5
and 8, while the initial liquid fraction at pH ¼ 10 is lower. †† If the generated foam is very unstable, i.e. if it collapses partly or completely
(c) Evolve differently as a function of time as regards the during generation, the foamability becomes much lower compared to a more
stable counterpart (see Fig. 3 in ref. 13 by way of example) and can no longer be
foam volume.
described with the theoretical model.

566 | Soft Matter, 2015, 11, 561–571 This journal is © The Royal Society of Chemistry 2015
View Article Online

Paper Soft Matter


Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

Fig. 5 Time evolution of the bubble sizes and the bubble size distribution at five different pH-values, namely pH ¼ 2, 3, 5, 8, and 10. The gas flow
rate used for foam generation was Q ¼ 50 mL min1.

(b) As regards drainage, we observe that the drainage rates of pure water for all pH-values,‡‡ we speculate that the surfaces
(i.e. loss of liquid per time) are similar for pH ¼ 2, 3, and 5, of the foams generated at pH ¼ 8 and 10, respectively, have a
while the foams generated at pH ¼ 8 and 10 drain faster. For the lower rigidity and thus a lower elasticity compared to the
drainage rate h ¼ D(ln 3)/D(ln t) we obtained values of around h surfaces of the foams generated at pH ¼ 2, 3, and 5. Unfortu-
¼ 0.26  0.02 for pH ¼ 2, 3, and 5, while we found h ¼ 0.32  nately, surface rheological parameters cannot be measured. At
0.02 for pH ¼ 8 and 10. Drainage is predominantly affected by the high surfactant concentrations (5 cmc ¼ 5  103–9  103
the viscosity of the bulk solution and the surface rigidity: the
lower the viscosity and the lower the surface rigidity the faster ‡‡ Water (20  C) ¼ 1.00 mPa s; water (25  C) ¼ 0.891 mPa s; C12DMAO solution at
the drainage rate.20,21 Since the viscosities correspond to those 22  C and pH of 2 ¼ 1.09  0.05 mPa s; C12DMAO solution at 22  C and pH of 5
¼ 0.97  0.02 mPa s.

This journal is © The Royal Society of Chemistry 2015 Soft Matter, 2015, 11, 561–571 | 567
View Article Online

Soft Matter Paper

Table 1 Arithmetic mean radius (hri), total bubble number in an area of has been investigated theoretically, and it has been shown that
20 mm2 (ntotal) and polydispersity index (PI) of the foam bubbles at the hydrogen bond formed between water and the amine oxide
three different times t for the foams generated at five different pH-
group is stronger than between two water molecules. Moreover,
values
even the water–water hydrogen bond strength is increased in
pH t¼0s t ¼ 200 s t ¼ 400 s close proximity to the amine oxide head group.24 No similar
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

theoretical study has been performed to elucidate the interac-


2 hri/mm 0.152 0.158 0.176 tion between the protonated form of the head group and water,
ntotal 164 142 106
PI 0.224 0.321 0.423
and likewise no theoretical study has investigated the strength
3 hri/mm 0.154 0.145 0.163 of hydrogen bonds formed directly between a charged and an
ntotal 159 158 108 uncharged amine oxide head group.
PI 0.206 0.402 0.527 From the discussion above it seems clear that hydrogen
5 hri/mm 0.139 0.131 0.177 bonding is of great importance for the properties of an
ntotal 191 173 93
PI 0.237 0.537 0.528
interface carrying an adsorbed layer of amine oxide surfac-
8 hri/mm 0.177 0.166 — tants. One expects strong hydrogen bonds between the head
ntotal 119 103 — group and water as well as between two head groups. Spec-
PI 0.229 0.463 — troscopic studies of the hydration water, as reported for other
10 hri/mm 0.167 0.158 — non-ionic surfactants,25,26 could provide additional informa-
ntotal 135 137 —
tion, particularly if combined with theoretical calculations.
PI 0.238 0.333 —
However, as yet such data is lacking and thus we refer to the
hydrogen bond model suggested by Maeda et al.1,7,8 The key
assumption in this model is that repulsive electrostatic
Table 2 Degree of ionization in bulk solution (a)1, maximum surface
concentration (Gmax), minimum surface areas (Amin), average N–N interactions compete with attractive hydrogen bonds
distance between two molecules (Daverage), and fraction of hydrogen between the surfactant head groups, for which, in turn,
bonds (p) of C12DMAO in 0.1 M NaCl solution at different pH-values protonation of some head groups is the prerequisite. The
hydrogen bond model rationalizes why the surface properties
pH a Gmax/106 mol m2 Amin/Å2 Daverage/Å p
pass an extremum when the pH is changed (see Introduc-
2 1 4.27  0.10 38.92  0.99 6.24  0.07 0–0.07 tion). In addition, it explains why the scmc-values/Gmax-value
3 0.9 4.40  0.10 37.71  0.88 6.14  0.07 0.01–0.18 of the cationic species C12DMAOH+ are unusually low/high
5 0.44 4.89  0.10 33.97  0.71 5.83  0.06 0.41–0.55 compared to those of the non-ionic species C12DMAO.
8 0.07 4.37  0.10 37.96  0.89 6.16  0.07 0–0.15 Generally speaking, the scmc-values/Gmax-value of the
10 0 4.29  0.10 38.73  0.92 6.22  0.07 0
cationic species C12DMAOH+ are expected to be much higher/
lower than those of C12DMAO since the cationic species repel
each other at the surface. However, inspection of the data
M depending on the pH) required for the foam study the provided in Fig. 2 and Table 2 of the present work as well as
diffusion of surfactant to the surface is very fast. Since the those provided in Maeda's studies1,7,8 shows that this is not
available techniques allow measuring only in a limited range of the case – the scmc- and Gmax-values of the two species are very
frequencies the measurements would all lead to the same value, similar. These results strongly support the presence of a non-
namely to an elastic modulus of 0. electrostatic attraction between the cationic species or
(c) The foam stability, just as the drainage rate, depends on between a cationic and a non-ionic species at low and
the pH-value. Here, however, the differences are much more medium pH-values. As already mentioned in the Introduc-
pronounced. In addition, the foam stability passes a tion, Maeda et al. suggest that this attraction is the result of
pronounced extremum at around pH ¼ 5 and thus follows the hydrogen bonds between two neighbouring surfactant
trends observed for the cmc, the scmc-values, and the Gmax- molecules.1,7,8 More specically, Maeda speaks of “hydrogen
values. We will discuss and explain the pH-dependence of the bonded dimers” which he expects “to behave as if they are
foam stability using the hydrogen bond model suggested by double-chain amphiphiles”.7
Maeda et al. Since this model is not very well known and never
has been used to explain foam properties, we will rst recapit-
ulate the basic ideas (Section 4.2) before we use the model to
explain the different foam stabilities (Section 4.3).

4.2. Hydration and the hydrogen bond model


The N–O bond in the uncharged amine oxide head group is very
polar due to charge transfer from the nitrogen atom to the
Fig. 6 Schematic representation of the proposed hydrogen bonds
oxygen atom. This results in strong interactions between water
between a non-ionic and a cationic head group that leads to the
and the amine oxide head group which results in signicant formation of a “hydrogen bonded dimer”.7 The length of the N–O
repulsive forces between bilayers of C12DMAO.22,23 The interac- bond is taken from ref. 24, while the one of the hydrogen bond is taken
tion between water and the uncharged amine oxide head group from ref. 27.

568 | Soft Matter, 2015, 11, 561–571 This journal is © The Royal Society of Chemistry 2015
View Article Online

Paper Soft Matter

We rened the drawing provided in ref. 7 by adding the bond 4.3. Hydrogen bonding and foam stability
lengths and showing the tetrahedral arrangement around the
Having recapitulated the hydrogen bond model and having
central N-atom. In Fig. 6 hydrogen bonding between a non-ionic
estimated the fraction of molecules that form dimers at the
and a cationic species is shown by way of example (the distances
respective pH, we now discuss whether this model can explain
for two cationic species are the same). The gure illustrates that
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

the pH-dependence of the foam stability observed for


direct hydrogen bonding between the head groups is sterically
C12DMAO. Our discussion will be based on three thoughts: (a)
possible and may result in the formation of a dimer, while this
the degree of ionization decreases continuously from 100% at
dimer cannot form hydrogen bonds to additional surfactants. pH ¼ 2 to 0 at pH ¼ 10 (see Fig. S6, ESI†), (b) the number of
Note that the drawing provided by Maeda et al. for the two hydrogen bonded dimers displays a maximum at pH  5 (see
cationic species in ref. 7 (Fig. 4b) is misleading since it suggests
Table 2) and thus does not reect the continuous change of the
that one molecule can form hydrogen bonds to two neighbours,
degree of ionization, (c) the presence of hydrogen bonds
which is impossible for steric reasons and not in line with his
between head groups enhances the stability of a foam.
idea of “hydrogen bonded dimers”. The authors corrected this
The two key experimental observations are as follows: (1) the
drawing in ref. 8 (Fig. 1).
foam stabilities at pH ¼ 8 and 10 are signicantly lower
Let us now assume that hydrogen bonded dimers indeed
compared to those at lower pH; (2) the foam stabilities pass a
are formed and that they consist either of a non-ionic and a maximum at pH ¼ 5. Looking at Table 2 one sees that there are
cationic species or of two cationic species depending on the no (or only very few) dimers at pH ¼ 2, 8, and 10. The higher
pH. Next, we calculated the average distance between two
foam stability at pH ¼ 2 compared to pH ¼ 8 and 10 must thus
molecules in a dimer. According to the drawing in Fig. 6 the
be due to the presence of charges and thus of electrostatic
N–N distance in the hydrogen bonded dimer is Ddimer ¼ 2(1.4
repulsion between two approaching surfaces (note that we now
Å sin 109 ) + 2.8 Å ¼ 5.4 Å. The average distance between non-
refer to electrostatic repulsion between two surfactant layers
ionic amine oxide molecules, Dmonomer, not participating in
and not between neighbouring molecules in the same layer).
hydrogen bonding can be calculated from the surface excess
This repulsion slows down drainage and coalescence and thus
at high pH to be 6.22 Å. Based on this idea and on the surface increases the lifetime of the foam. At pH ¼ 3 the number of
excesses Gmax calculated from the surface tension curves, one dimers is larger compared to pH ¼ 2 which is why the foam
can quantify the number of hydrogen bonds and thus the
stability is higher. Finally, at pH ¼ 5 the attractive hydrogen
number of dimers at a given pH. From the Gmax-value we
bonds and the repulsive electrostatic interactions are counter-
calculated the respective Amin-value the square route of which
balanced in an optimum way resulting in the highest foam
is the average N–N distance Daverage between two molecules
stability.
(see Table 2). Since the N–N distance calculated from the
With these results in mind we want to recall that charges are
Gmax-values is an average over all distances, it holds in a
not necessarily needed to stabilize foams. For example, foams
simple two-state model that generated in presence of C12TAB (see Fig. 3 in ref. 14) or sodium
dodecyl sulfate (SDS)28 are much less stable compared to foams
Daverage ¼ pDdimer + (1  p)Dmonomer (3)
stabilized by a mixture of C12DMAOH+ and C12DMAO at pH ¼ 5.
where p is the fraction of molecules that form dimers which can On the other hand, the stability of foams stabilized by the non-
be calculated according to ionic surfactant b-C12G2 ref. 13 can be compared to that of
foams generated at pH ¼ 5 in the study at hand (see Fig. 7
Daverage  Dmonomer
p¼ : (4) below). Thus the lower foam stabilities observed at pH ¼ 8 and
Ddimer  Dmonomer
10 cannot be explained as a consequence of reduced electro-
static interactions. This is a very important conclusion since it is
In Table 2 we provide the calculated values for p at the just the opposite of what was argued about 35 years ago.29 Tsujii
various pH-values. The table reads as follows: at pH ¼ 2 a and Arai measured the stability of foams generated with
maximum of 7% of the molecules at the surface form dimers, at C12DMAO solutions containing 0.1 M NaCl at different
pH ¼ 5 it is up to 55% and at pH ¼ 10 no dimers are formed. In concentrations as a function of the pH which varied from 5.5 <
other words, the number of hydrogen bonded dimers displays a pH # 7. Note that they only studied the average lifetime of
maximum at pH  5. The lower tendency to form dimers at low foams generated via the Bikerman method and that they did not
pH-values is due to the electrostatic repulsion, which counter- provide any additional information about the foamability of
acts close proximity of the head groups and thus hydrogen bond their solutions, the liquid content of the foams, and the bubble
formation, while the lower tendency to form dimers at high pH- sizes. Moreover, they did not determine any other properties of
values is due to the low fraction of protonated head groups. their solutions so that one can hardly draw any conclusions
What is of importance for the following discussion is the fact from this work. Nevertheless, looking at Fig. 1 in ref. 29, one
that the values calculated for pH ¼ 2, 3, 8 and 10 are very sees that the average lifetime s of foams generated with a
similar. In view of the large uncertainties of these numbers a C12DMAO solution of c ¼ 4 mM (4 cmc) increases from s  400
quantitative analysis, unfortunately, is not possible which is s at pH ¼ 7 to s  1700 s at pH ¼ 6.3, which is in perfect
why we will discuss the data qualitatively in the following agreement with our results (see Fig. 4). Tsujii and Arai argued
Section. that it is the formation of ionic species that stabilizes the foam

This journal is © The Royal Society of Chemistry 2015 Soft Matter, 2015, 11, 561–571 | 569
View Article Online

Soft Matter Paper

5. Conclusions & outlook


We conclude by repeating our two key ndings. (1) Foams
stabilized by aqueous solutions of C12DMAO show a maximum
in the foam stability at pH ¼ 5. (2) The stabilities of foams
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

generated at low pH-values are signicantly higher compared to


those generated at high pH. The higher stability at low
compared to high pH-values is indeed due to the presence of
electrostatic repulsion between two opposing surfactant layers.
The maximum stability, however, cannot be explained by
changes in the electrostatic repulsion since this repulsion
decreases with increasing pH and does not pass a maximum.
Thus the maximum stability is linked to the possibility of
forming hydrogen bonds between the surfactant head groups at
moderate and low pH. Indeed, looking at foam stability data for
other surfactants with C12 alkyl chains, one nds that stable
foams are only observed when hydrogen bonds can be formed
Fig. 7 Time evolution of the foam volume V for foams stabilized by between the head groups. Thus stable foams are observed for
four different surfactants, namely C12DMAO at 5 cmc, 0.1 M NaCl, pH surfactants with sugar head groups,13 but not with head groups
¼ 5 ( ), C12E3I1 at 5 cmc, 0.1 M NaCl ( ), C12E6 at 10 cmc, 0.01 M NaCl consisting of ethylene oxide,13 phosphine oxide,14 trimethyl
( ), and b-C12G2 at 10 cmc, 0.01 M NaCl ( ). The data for C12E6 and b- ammonium14 or sulphate.28 These observations suggest that
C12G2 are taken from ref. 13. The gas flow rate used for foam gener-
hydrogen bonds between neighbouring molecules at the surface
ation was in all cases Q ¼ 50 mL min1.
enhance the foam stability of C12 surfactants. Formation of
hydrogen bonds between surfactant head groups gives rise to a
short-range attractive interaction1,7,8 that may render the
due to increasing repulsive interactions between the mono-
surfactant layer more rigid (in the sense that the mobility of the
layers of the foam lms. These repulsive interactions indeed
molecules is restricted) and more elastic (in the sense that
explain the higher stability at low compared to high pH but not
deformations can be counteracted), which explains the pH-
the maximum stability at pH ¼ 5, which brings us back to our
dependence of the drainage rate and of the foam stability
hypothesis, namely that foam stability is directly correlated with
observed in this study. Our hypothesis of the importance of
the number of hydrogen bonds.
hydrogen bonds for foam stability will be challenged by col-
With the hypothesis in mind that hydrogen bonds between
lecting spectroscopic data on hydrogen bonding at the
head groups enhance the stability of a foam we searched for
aqueous–surfactant–air interface e.g. with sum frequency
further foam data of surfactants with a C12-chain but different
generation spectroscopy. Theoretical investigations of interac-
head groups. Foams stabilized by aqueous solutions of C12E6
tions between surfactant head groups in the presence of water
ref. 13 and C12DMPO14 are also much less stable than the foams
molecules also have the potential to further increase our
generated at pH ¼ 5 in this study. The fact that C12TAB, SDS,
understanding.
C12E6, and C12DMPO are not capable of forming hydrogen
bonds between the head groups supports our hypothesis. On
the other hand, foams stabilized by b-C12G2 are much more Acknowledgements
stable compared to those stabilized by the above mentioned
We would like to thank the German Research Foundation (DFG)
surfactants, which, in turn, can be explained with hydrogen
for nancial support.
bonds between the sugar head groups. To further test our
hypothesis we carried out a preliminary study with the non-
ionic surfactant 1-O-(3,6,9-trioxa-henicosanyl)-myo-inositol References
(C12E3I1). The head group of this surfactant consists of an
1 H. Maeda, S. Muroi, M. Ishii, R. Kakehashi, H. Kaimoto,
ethoxylated inositol the synthesis and properties of which are
T. Nakahara and K. Motomura, J. Colloid Interface Sci.,
explained in ref. 30 and 31. What is signicant for the paper at
1995, 175, 497–505.
hand is the fact that the inositol head group has ve OH-groups
2 Mixed Surfactant Systems, Surfactant Science Series, ed. M. Abe
and is thus also capable of forming hydrogen bonds between
and J. F. Scamehorn, Marcel Dekker, New York, 2nd edn,
two neighbouring molecules. Based on our hypothesis we thus
2005, vol. 124.
expected foams generated in the presence of C12E3I1 to be more
3 N. Buchavzov and C. Stubenrauch, Langmuir, 2007, 23, 5315–
stable compared to foams stabilized by C12E6 or by C12DMAO at
5323.
pH s 5 and to have a similar stability as foams stabilized by b-
4 S. R. Patil, N. Buchavzov, E. Carey and C. Stubenrauch, So
C12G2 or by C12DMAO at pH ¼ 5. Comparing Fig. 7 with Fig. 4
Mater., 2008, 4, 840–848.
(le), one clearly sees that this is indeed the case.
5 E. Carey and C. Stubenrauch, J. Colloid Interface Sci., 2010,
343, 314–323.

570 | Soft Matter, 2015, 11, 561–571 This journal is © The Royal Society of Chemistry 2015
View Article Online

Paper Soft Matter

6 Y. Imaishi, R. Kakehashi, T. Nezu and H. Maeda, J. Colloid 19 D. J. Durian, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat.
Interface Sci., 1998, 197, 309–316. Interdiscip. Top., 1997, 55, 1739–1751.
7 H. Maeda, Colloids Surf., A, 1996, 109, 263–271. 20 S. A. Koehler, S. Hilgenfeldt and H. A. Stone, Phys. Rev. Lett.,
8 H. Maeda and R. Kakehashi, Adv. Colloid Interface Sci., 2000, 1999, 82, 4232–4235.
88, 275–293. 21 S. A. Koehler, S. Hilgenfeldt and H. A. Stone, Langmuir, 2000,
Published on 26 November 2014. Downloaded by Indian Institute of Technology New Delhi on 3/28/2022 12:11:26 PM.

9 V. Lair, S. Bouguerra, M. Turmine and P. Letellier, Langmuir, 16, 6327–6341.


2004, 20, 8490–8495. 22 C. E. Herder, P. M. Claesson and P. C. Herder, J. Chem. Soc.,
10 C. Stubenrauch, L. K. Shrestha, D. Varade, I. Johansson, Faraday Trans. 1, 1989, 85, 1933–1943.
G. Olanya, K. Aramaki and P. Claesson, So Mater., 2009, 23 L. Mol, B. Bergenståhl and P. M. Claesson, Langmuir, 1993, 9,
5, 3070–3080. 2926–2932.
11 D. Georgieva, A. Cagna and D. Langevin, So Mater., 2009, 5, 24 V. Kocherbitov, V. Veryazov and O. Söderman, J. Mol. Struct.:
2063–2071. THEOCHEM, 2007, 808, 111–118.
12 C. Stubenrauch, P. M. Claesson, M. Rutland, E. Manev, 25 E. Tyrode, C. M. Johnson, M. W. Rutland and P. M. Claesson,
I. Johansson, J. S. Pedersen, D. Langevin, D. Blunk and J. Phys. Chem. C, 2007, 111, 11642–11652.
C. D. Bain, Adv. Colloid Interface Sci., 2010, 155, 5–18. 26 E. Tyrode, M. C. Johnson, A. Kumpulainen, M. W. Rutland
13 J. Boos, W. Drenckhan and C. Stubenrauch, J. Surfactants and P. M. Claesson, J. Am. Chem. Soc., 2005, 127, 16848–
Deterg., 2013, 16, 1–12. 16859.
14 E. Carey and C. Stubenrauch, J. Colloid Interface Sci., 2010, 27 J. N. Israelachvili, Intermolecular and surface forces, Academic
346, 414–423. Press, San Diego, 2nd edn, 1991.
15 E. Carey and C. Stubenrauch, Colloids Surf., A, 2013, 419, 7– 28 http://www.itconceptfr.com/Pageshtml/Foamscan_Exemples4_
14. VA.html.
16 E. Carey and C. Stubenrauch, J. Colloid Interface Sci., 2009, 29 K. Tsujii and H. Arai, J. Am. Oil Chem. Soc., 1978, 55, 558–560.
333, 619–627. 30 G. Catanoiu, V. Gärtner, C. Stubenrauch and D. Blunk,
17 J. Boos, W. Drenckhan and C. Stubenrauch, Langmuir, 2012, Langmuir, 2007, 23, 12802–12805.
28, 9303–9310. 31 G. Catanoiu, D. Blunk and C. Stubenrauch, J. Colloid
18 K. Feitosa, S. Marze, A. Saint-Jalmes and D. J. Durian, J. Phys.: Interface Sci., 2012, 371, 82–88.
Condens. Matter, 2005, 17, 6301–6305.

This journal is © The Royal Society of Chemistry 2015 Soft Matter, 2015, 11, 561–571 | 571

You might also like