INVESTIGATION OF SURFACTANT SURFACE TENSION AND ITS

CORRELATION WITH TEMPERATURE AND CONCENTRATION

Li Yaw (Michael) Cheng

Su Min (Lisa) Wei

Science One Program

The University of British Columbia
Vancouver, Canada
June 2012
Cheng and Wei (2012)

1 Abstract
Surface tension is a property that defines the energy per unit area of a liquid-gas interface. The purpose of this study
is to investigate the surface properties and the process of aggregation of surfactant micelles in detergents. By using a
drop-weight method, the surface tension of detergent solutions was determined under standard pressure condition 1
bar. A 100% concentrated detergent solution was subject to heating. Data analysis showed that surface tension de-
creased linearly with increasing temperature, as represented by the equation γ = – 0.1766 mN/(m⋅K)(x) + 94.654
mN/m, where x is temperature. When subjected to variations in concentration, surface tension of the detergent de-
creased until reaching the critical micelle concentration, as modelled by the segmented linear regression γ = – 0.565
mN/(m⋅%)(x) + 61.7 mN/m when x < 50%, and γ = 33.45 mN/m when x ≥ 50%, where x is concentration.

2 Introduction
Surface tension of a solvent or solution, denoted by the Greek letter gamma γ, represents a well-defined physico-
chemical property. It is caused by cohesion of fluid molecules and allows for its resistance of an external force [1]. The
unit of measurement has dimensions of force per unit length, or equivalently, energy per unit area. Molecules at the
surface experience a net inward force which results in internal pressure. Considering the isoperimetric inequality,
liquid surfaces minimize its energy of configuration by contracting to minimal area [2]. Surface tension is ubiquitous
in the natural world. For instance, water striders are able to float on top of water due to the surface tension of water
providing the restoring force to the insect’s weight. In surface and colloidal chemistry, surface tension can be useful in
characterizing the chemical activity, adsorption, and dissolution of materials [3].

A specific branch of study in surface chemistry is the surface tension of surfactants, which are surface-active agents
that adsorb to a fluid surface. Besides the common usage of surfactants in food additives, detergents, and cosmetics,
these molecules are often used as surface active agents in the production of drugs within both medicinal and phar-
maceutical sciences [4]. When dissolved in solvent, surfactants disrupt the intermolecular interactions of the liquid,
thereby decreasing the solvent’s surface tension. Water-soluble polymers have amphipathic structures, with non-polar
hydrophobic parts and hydrophilic parts that favour interaction with polar solvents such as water [3]. The formation
of micelles can be understood through principles of thermodynamics; micelles can form spontaneously because of a
balance between entropy and enthalpy. In water, hydrophobic effects and London van der Waals forces between water
and surfactant molecules drive the micellization process.

In the present study, we have selected a particular detergent brand named “Palmolive Original.” The primary surfac-
tant agent is lauramidopropylamine oxide, a non-ionic surfactant. As its name suggests, non-ionic surfactants do not
possess electrical charges [9]. The substance’s surface properties were characterized by measuring the surface tension
of the detergent at different temperatures and concentrations. The du Nuöy ring method was adopted as a model for
the conduction of surface tension measurement [5].

3 Method
3.1 Equipment Calibration
The main component of the apparatus is the dial-o-gram (Ohaus, #1650-WO). Figure 1 depicts the setup of our ex-
periment. A stiff copper wire of negligible mass was used to construct a ring and secured to the beam. To ensure the
copper ring was circular and completely flat, the wire was tied around a cylindrical container and hammered into
circular conformation. The ring was flattened using a hammer on a smooth table. The ring’s flatness and circular
shape were crucial to achieving even distribution of the pulling force upon the liquid surface. When securing the
extended arm of the ring onto the beam, tweezers were used to minimize the wiggle room of the twisted wire. A 1 kg
calibration weight was placed on the stainless steel plate and weight was adjusted on the beam to create an initial
rough balance. A clean and lightweight Styrofoam cup was placed onto the plate. The fine-tuning knob was adjusted
so the beam hit the zero-mark, which indicated the exact point of balance. This marked the initial position.

1
Cheng and Wei (2012)

Figure 1: Experimental Setup. From left to right, the Styro- Figure 2: Visual representation of torque calculation. M is
foam cup for addition of water as mass applied, calibration bulk the mass added to maintain and eventually disrupt the equilib-
weight of 1 kg, minor control knob to obtain equilibrium posi- rium position. L is the mean perimeter of the ring. Fmass is the
tion, three beams (two are used for balancing and the third for gravitational force exerted by the applied mass. Ftension is the
the attachment of ring), and the ring for contacting the periph- force exerted by fluid periphery against Fmass. rR is the arm
ery of the fluid. length between the axis of rotation and the ring. rP is the arm
length between the axis of rotation and the centre of the stain-
less steel plate.

3.2 Measurement of Surface Tension

A 250 mL beaker was used to hold the solution subjected to measurement. The beaker was filled to precisely the 200
mL mark. This mark was the level that allowed the ring to rest on top of the liquid surface without disrupting the
equilibrium position of the beam. The cup was pre-weighed on an analytical balance to obtain the mass to 5 signifi-
cant digits. Since surface tension is considered a weak force compared to a gravitational pull by the applied force, a
dropper bottle was used to add water drops to the Styrofoam cup in small increments until the beam at the delicately
balanced position was tipped and the ring detached from the liquid surface. Clean water from the tap was used. The
cup and added water were weighed on the analytical balance, and the mass of water added was calculated. Five repli-
cates were carried out for each trial. Note the dial-o-gram was used because it has a high degree of sensitivity to min-
uscule additions of mass. An addition of one hundredths of a drop of water can be sufficient to tip the balance.

3.3 Mathematical Derivation of Surface Tension Formula

Considering the du Noüy formula [5], the surface tension is given by γ = Mg/(2L) where γ is the surface tension, M is
the mass used to applied force, g is the gravitational acceleration (9.807 m/s2 was used throughout the experiment),
and L is the mean circumference of the ring. The unit of measurement is force per unit length with SI units N/m. The
surface tension of the liquid was calculated using principles of torque and rotational dynamics [5]. Torque is the prod-
uct of arm length r with the force component F that is perpendicular to the radial line. The formula is given by τ ≡
rFsinθ, where θ is the angle between the moment arm and the force applied. In the experiment, θ = 90°, since the
surface tension and the weight of water drops were both tangent to the beam. Initially, the system was at rotational
equilibrium. As shown in Figure 2, the addition of drops of water to the Styrofoam cup resulted in a gravitational
torque greater than the torque exerted by surface tension. This caused the system to rotate counter-clockwise about
the fulcrum (the pivot point). Given τR is depicted as the torque on the ring side where tension exerts force and τP as
the torque on the plate side where mass exerts a downward force, then the formula for surface tension can be derived
as follows:

( Ftension )( rR ) = ( Fmass )( rP )
γ 2LrR = MgrP
MgrP
γ = (1)
2LrR .

2
Cheng and Wei (2012)

3.4 Statistical Analysis

The uncertainty (σm) in each measurement of surface tension was derived from standard deviation. The formula is σm
= σ/ N , where the standard deviation is calculated as follows:

1 N
σ= ∑
N − 1 i=1
( yi − µ )
2
(2)
,

where the mean (µ) is

1 N
µ= ∑ yi
N i=1 .
(3)

For each linear regression model in the form of y = mx + b, the best-fit line was generated on Excel 2010 by using the
weighted best-fit slope, its uncertainty δm and y-intercept b as displayed below:

1 ⎡ N 1 N xi yi N
xi N yi ⎤
m= ⎢∑ 2∑ 2 −∑ 2∑ 2 ⎥
(4)
Δ ⎢⎣ i=1 (δ yi ) i=1 (δ yi ) i=1 (δ yi ) i=1 (δ yi ) ⎥⎦,
1 N 1
δm = ∑
Δ i=1 (δ yi )2
(5)
,
1 ⎡ N x2 N y N
x N
xy ⎤
b = ⎢∑ i 2 ∑ i 2 − ∑ i 2 ∑ i i 2 ⎥ (6)
Δ ⎢⎣ i=1 (δ yi ) i=1 (δ yi ) i=1 (δ yi ) i=1 (δ yi ) ⎥
⎦,

where

2
N
1 N
xi2 ⎛ N x ⎞
Δ=∑ ∑ −⎜∑ i 2 ⎟ (7)
i=1 (δ yi )2 i=1 (δ yi )2 ⎝ i=1 (δ yi ) ⎠ .

Each resultant linear regression model is then evaluated using weighted least square analysis derived as follows:

2
1 N ⎡ y − f (xi ) ⎤
χ = ∑⎢ i
2
(8)
N i=1 ⎣ δ yi ⎥⎦ .
w

3.5 Varying Temperature

The temperature of a 100% concentration surfactant solution and of clean tap water was varied. Identical procedures
of temperature variations were carried out for both types of solutions. A 250mL beaker was filled with 200mL of solu-
tion and put in a standard refrigerator to be cooled to 273K. A thermometer was used to monitor the temperature
drop. Then the beaker was taken out and placed on an insulating mat to prevent heat to be transferred to or from the
solution to the surroundings. Surface tension was measured using standard protocols as delineated in the above pro-
cedural section. The solution was heated in 5K increments with the glass beaker placed in a pan. The pan is then
placed on an electric stove to prevent the glass from cracking due to large pressure changes from cooling or heating.
The temperature was monitored using a thermo-anemometer (Extech Instruments), which detects air velocity, rela-
tive humidity and temperature of the system. This instrument was used to ensure the uniformity of environmental
conditions other than experimental variable, which is temperature.

3
Cheng and Wei (2012)

3.6 Varying Concentrations

Surface tension was also correlated with surfactant concentration. “Palmolive Original” detergent was used as the
surfactant. Multiple concentrations of detergent were employed to ensure accuracy of the direct relation between
surfactant and change in magnitude of surface tension. Varying concentrations without affecting volume was
achieved by the volume replacement method, where the volume was set to 200mL. The following equation was used
for concentration and dilution, respectively:

200 ( xk+1 − xk )
ΔV (xk , xk+1 ) =
1− xk
(9)
200 ( xk − xk+1
′ )
ΔV (xk , xk+1
′ )=
xk ,

where ∆V is the volume replaced to achieve a certain concentrated or diluted concentration. This volume of solution
was either discarded and replaced with pure detergent when increasing surfactant concentration or replaced with
water when performing dilutions. The calculated amount should result in a mixture of desired concentration while
returning its volume to 200mL as the control volume. xk is the initial concentration, xk+1 is the increased concentra-
tion, and x'k+1 is the decreased concentration upon dilution. Concentrating from 0% detergent to 100% detergent and
its reverse (i.e., diluting) were carried out to optimize the precision of each concentration data point. This resulted in
the number of replicates doubling to ten.

4 Results and Analysis

4.1 Temperature Dependence of Surface Tension
80 50
y = -0.2866x + 152.86 y = -0.1766x + 94.654
Surface Tension (mN/m)

Surface Tension (mN/m)

75
45
70
65 40
60 35
55
30
50
45 25
270 285 300 315 330 345 360 375 270 285 300 315 330 345 360 375
Temperature (K) Temperature (K)

Figure 3: Surface Tension of Water at Various Temperatures Figure 4: Surface Tension of Detergent at Various Tempera-
at Standard Pressure Condition 1 bar. Bars represent stan- tures at Standard Pressure Condition 1 bar. Bars represent
dard errors. Each data point represents total number of five standard errors. Each data point represents a total number of
replicates (n = 5). χw2 = 8.361, m = – 0.288 mN/m, b = 152.86 five replicates (n = 5). χw2=2.310, m = – 0.177 mN/m, b =
mN/m. 94.654 mN/m.

Linear relation is observed and hypothesized for both Figure 3 and Figure 4 using y = mx + b, where m is the slope
and b is the y-intercept of the model function. Using the weighted least square fitting as the linear model, it was de-
termined that for water, the linear model yields a minimum weighted chi-square value of 8.361, with m = – 0.288
mN/(m⋅K) and b = 152.86 mN/m, resulting in a critical temperature of 531.024K. The critical temperature marks the
point at which surface tension is expected to drop to zero (equivalent to the x-intercept). However, the extrapolated
value of the critical temperature for water has no physically relevant meaning because water would already be in gas
phase at 373K. With the same method, it was determined that for detergent, the linear model yields a minimum
weighted chi-square value of 2.310 with m = – 0.177 mN/(m⋅K) and b = 94.654 mN/m, resulting in a critical tem-
perature of 535.896K. The linear model for water shows a steeper slope and a higher y-intercept than the parameters
of the linear model for detergent.

4
Cheng and Wei (2012)

Distinct discretization of surface tension decrement can be observed in Figure 3 and Figure 4. Consecutive differ-
ences are not uniform. Significant drops in value occur in a stepwise fashion. The cause was first hypothesized to be
the discrete mass added for each drop of water and the sensitivity of the equipment used. To test the hypothesis, the
weight of several drops of water was determined and was divided by the number of drops to obtain mass per drop.
Mean mass is applied to equation (2) to obtain the mean gravitational force per length per drop. This renders the
value 0.212 mN/m per drop, which was compared to mean significant changes in surface tension in Figure 3 and
Figure 4. Mean significant changes are quantified as the mean of the greatest ten consecutive fluctuations of data. In
Figure 3, the mean significant change in surface tension is 4.422 mN/m. In Figure 4, the mean significant change in
surface tension is 2.122 mN/m. Both results show low correlations to the mean gravitational force exerted per drop.
The higher values in fluctuation can be rationalized with the additional applied force when squirting the water into
the Styrofoam cup.

4.2 Concentration Dependence of Surface Tension

80 70.00
y = -0.1766x + 94.654

Surface Tension (mN/m)

Surface Tension (mN/m)

70
58.75
60
50 47.50
40
36.25
30
y = -0.2866x + 152.86
20 25.00
270 285 300 315 330 345 360 375 0 25 50 75 100
Temperature (K) Concentration of Detergent (%)

Figure 5: Surface Tension of Water and Detergent at Various Figure 6: Surface Tension of Water-Detergent Mixture Meas-
Temperatures at Standard Pressure Condition 1 bar. Bars ured with Different Concentrations at Room Temperature
represent standard error calculated according to the appendix. 298K and Standard Conditions with Varying Concentra-
Each data point represents a total number of five replicates (n = tions. Bars represent standard errors. Each data point repre-
5). sents a total number of ten replicates (n = 10). χw2 = 20.217, m
= – 0.565 mN/m⋅K, b = 61.700 mN/m, C = 50%.

Considering the formula for uncertainty shown in the procedural section, the surface tension of water at pressure 1
bar and room temperature (273K) was measured to be 70.000 ± 3.200 mN/m and the surface tension of detergent is
36.000 ± 1.700 mN/m. By comparing Figures 3 and Figure 4 in the composite graph Figure 5, the surface tension of
water is consistently higher than the detergent, by nearly 50%. There is no ambiguity with this statement since the
uncertainty ranges do not overlap. However, as shown by the small error bars, this absence of ambiguity may be due
to underestimation of error propagation in measurements or instrumental error (Please refer to Qualitative Error
Analysis for in-depth discussion).

In Figure 6, the surface tension after the critical point decreases steadily with increasing surfactant concentration.
However, close inspection reveals that the decreasing nature halts at about 50% concentration. Surface tension after
this concentration lacks a particular trend; it could be oscillating, increasing, or constant. Most of the data points
demonstrate significance with no overlaps. Minimum uncertainty is observed at 50% concentration and maximum
uncertainty is determined to be at 62.500%. Hence, linear regression termed the “hockey stick function” was used to
fit the data shown in Figure 6:

⎧⎪ m1 x + b2 x<C
y=⎨ (10)
m x + b2 x≥C
⎩⎪ 2 ,

where m1 and m2 are slopes, b1 and b2 are y-intercepts, and C is the critical breakpoint of the segmented linear regres-
sion where behaviour changes dramatically. The parameters are generated by determining the least chi-square value.
For this segmented linear regression model, since the best-fit slope and y-intercept cannot be determined with the

5
Cheng and Wei (2012)

aforementioned formula, manual minimization of fit parameters was employed. The minimality of the slopes, y-
intercepts and value of C were confirmed by plotting trial values to generate parabolic functions. The point at which
the derivative is zero depicts the minimum value of the parameter. The set of graphs are shown in Figure 7 below.

Optimal Slope (m) Optimal y-Intercept (b) Optimal Breakpoint (C)

20.45 22.5 22.0
y = 0.0019x2 - 0.0433x + 20.459 y = 0.0132x2 - 0.3532x + 22.583
20.40 22.0 21.5
χ2 Values

χ2 Values

χ2 Values
20.35 21.5
21.0
20.30 21.0
20.25 20.5 20.5
20.20 20.0 20.0
-0.555 -0.56 -0.565 -0.57 60.5 60.9 61.3 61.7 62.1 62.5 49 49.3 49.6 49.9 50.2 50.5 50.8
Values of m Values of b Values of C

Figure 7. Manual Minimization of Chi-Square Values. The parameters were generated by determining the minimum points on
the graphs, each corresponding to a specific parameter. As shown in Figure 6, the values are χw2 = 20.217, m = – 0.565 mN/m⋅%, b
= 61.7 mN/m, C = 50%.

The uncertainty associated with each data value in Figure 6 is one sigma bar, which suggests two-thirds of the data
values should cross the fit line. However, only two-ninths of the error bars cross the line. This may be due to the
aforementioned underestimation of error.

5 Discussion
5.1 Surface Activity of Surfactant
The cause of the general decreasing trend in the surface tension of detergent-water mixture may be due to the pres-
ence of the cleaning and foaming agent lauramidopropylamine oxide (C17H36N2O2). This molecule is a nitrogenated,
non-ionic surfactant [9]. As shown in Figure 8, non-ionic surfactants contain tertiary amine oxides that are often used
as emulsifiers since they have high solubility in water [9]. Lauramidopropylamine oxide contains a long hydrophobic
alkyl chain and a highly polar N-O bond. The nitrogen atom involved in this bond readily donates its electrons, re-
sulting in a negative oxygen atom which readily captures a proton in aqueous solution [7]. The highly polar N-O bond
drives the hydrophilic interactions with water, subsequently modifying and disrupting the hydrogen bonding be-
tween water molecules. The amphiphilic property of the surfactant molecule also allows this hydrophilic end to ad-
sorb to the fluid surface. The collective energy that generates water’s high surface tension decreases.

Figure 8: Chemical structure of lauramidopropylamine ox-

ide. The highly polar N-O bond in lauramidopropylamine ox-
ide disrupts the hydrogen bonding between water molecules,
causing surface tension of the aqueous solution to decrease
linearly until reaching the critical micelle concentration.

5.2 Critical Micelle Concentration

The position upon which the plateau is reached in Figure 6 can be attributed to the critical micelle concentration
(CMC). In colloidal and surface chemistry, CMC is defined as the concentration of surfactant above which micelles
spontaneously form in solution [1][7]. The value of CMC and surface tension itself are intensive thermodynamic pa-
rameters [4]. These quantities are therefore dependent upon temperature, pressure, and on the presence and concen-
tration of other surface-active electrolytes [8]. The value of the critical micelle concentration quantifies the degree of
packing and the orientation of the adsorbed surfactant molecules [1]. As surfactant concentration increases, surface
coverage of surfactant molecules increase, and the maximum coverage are reached at CMC [1][9]. At that point, mole-
cules begin to aggregate into micelles. Monomers and micelles thereafter exist in dynamic equilibrium, and the spac-
ing between adjacent molecules attains its most favourable distance to minimize the intermolecular potential energy.
Any subsequent formation of micelles will have minor effects on surface tension.

The surface tension of the surfactant at which the experimental CMC value was reached was compared to those at-
tained by Atta et al. [1] who worked with non-ionic surfactants under similar conditions. Shown in Figure 6, the criti-
cal micelle concentration was determined to be 50% surfactant concentration. The average constant surface tension

6
Cheng and Wei (2012)

was 33.45mN/m at 298K based on the segmented linear regression model. Atta et al. [1] showed that the non-ionic
surfactant CMC was reached at surface tension 34.8 ± 0.5mN/m at 303K.

When comparing Figure 3 and Figure 4, the detergent surface tension decreased much more slowly than did water.
This is because in the 100% concentrated surfactant solution, the area per molecule at the surface decreased at the
surface due to increased dehydration of the hydrophilic group at higher temperature [7] yet increased due to enhanced
molecular motion and thermal agitation at higher temperatures. The two opposing forces result in lower decreasing
rate for surfactant surface tension when compared to that of water.

5.3 Qualitative Error Analysis

Some transparent soap bubbles formed around the copper ring and between the ring and the surface of the surfactant
as the ring was tilted out of the solution. This made it more difficult to distinguish the exact point at which the ring
loses contact with the liquid surface. This would result in a slightly greater amount of water being added to the cup
than the volume sufficient to break surface tension. As observed in Figure 3, there are large fluctuations from 273K
to 308K, and the values do not correspond to the slope of the general decreasing trend of the measurements above
308K. Bubble-formation was frequent in the range 273K to 308K. Large uncertainties were associated with measure-
ments at lower temperature, as can be seen in Figure 3. The maximum error was detected at 288K temperature, with
surface tension 43.460 ± 2.460mN/m.

Due to the ring’s frequent contact with the detergent, the ring’s smooth wax layer was observed to dissolve. The
strength of contact between the liquid surface and the wax layer of the ring would be significantly different from the
tension between the surfactant surface and the rougher, grey metal layer underneath the wax. The rough layer of the
wire would, in principle, have greater interaction with the water as the wax layer is smoother. This would result in the
surfactant clinging more tightly to the ring and exert a larger resistant force to the upward pull. The wax layer has a
more hydrophobic surface [4][8]. This results in hydrophobic repulsion that increases the fluid’s energy per unit of sur-
face area. The fluid would have a greater tendency to minimize its contact with the waxy surface than to the rougher
ring surface [6][7]. As can be observed in Figure 3 and Figure 4, the values remain at a generally constant value with
smaller fluctuations. The significantly stronger pull exerted by the non-waxy layer of wire may have contributed to
similar measurements.

Tap water contains impurities such as electrolytes, sediments and minerals. This would cause the measured surface
tension of water to deviate from literature values (see Appendix). The experimental surface tension values are lower
than literature values, especially at temperatures above 298K. The greatest deviation from literature value can be seen
at 368K: the measured tension was 50.610 ± 1.250mN/m, whereas the value listed in the CRC Handbook was
59.87mN/m. The 9.260 ± 1.250mN/m difference marks a significant deviation. The general decreasing trend in Fig-
ure 3 and Figure 4 may be attributed to increased thermal motion and energy of the molecules with increasing tem-
perature. However, the presence of electrolytes disrupts hydrogen bonding in water molecules. Since intermolecular
forces were already weakened before heating, the molecules could attain greater degree of freedom, resulting in lower
than predicted surface tension.

5.4 Future Directions

Discrepancy in concentration of mixture may be due to measurement of insufficient accuracy. A graduated cylinder
with smaller and more accurate increments of measurement should be used in further investigations. The material of
the ring should be switched plastic to prevent degradation of the ring's surface layer. To increase confidence in the
uniform sizes of water drops added as weight, a Pasteur pipet should be used.

When detergent solution was heated to 353K, intense bubbling occurred and the detergent was observed to undergo
drastic compositional change, resulting in a more opaque and deeper green solution. The cause of this phenomenon
was not known. It may be useful to utilize spectroscopy, generating reflection imaging using lasers, to detect the de-
composition mechanism of the surfactant surface layer or layers [1][5]. Further studies of this surface tension of the
surfactant solution and of water-surfactant mixtures should be conducted via measurement of the fluid’s contact an-
gle in a capillary tube [1][3].

7
Cheng and Wei (2012)

6 Conclusion
We have concluded that under standard pressure condition 1 bar and 298K, the surface tension of “Palmolive Origi-
nal” surfactant solution decreases linearly as temperature increases, as indicated by the weighted least square fit γ = –
0.177 mN/(m⋅K)(x) + 94.654 mN/m, where x is temperature. When concentration was varied at constant tempera-
ture, surface tension decreased linearly until a particular concentration, which is C = 50% for the non-ionic surfac-
tant. Beyond that threshold concentration, the surface tension was interpreted to remain constant. The “hockey stick
function” was used to model the trend, with the initial linear γ = – 0.565 mN/(m⋅%)(x) + 61.7 mN/m when x < 50%,
and γ = 33.45 mN/m when x ≥ 50%, where x is concentration. This model adsorption isotherm is extremely useful in
characterizing the surface properties of the surfactant solution. The surface tension of water was observed to be con-
sistently higher than the surface tension of surfactants. These model parameters allow us to make reasonable predic-
tions of the mechanisms of micelle-formation and the energy at the liquid-gas and liquid-solid interface.

7 Acknowledgement
We would like to thank the UBC Chemistry labs for lending us the hotplate and the Physics department for lending
us various measuring equipment. Our mentor James Charbonneau has been a tremendous help throughout this
process. We would also like to acknowledge our peer reviewers, Samantha Tan, Liz Geum and Gurkaran Singh, for
their numerous suggestions and invaluable advice.

8 Literature Cited
[1] Atta, A. M., El-Kafrawy, A. F., Abdel-Rauf, M. E., Maysour, N. E., & Gafer, A. K. Surface and thermodynamic properties of nonionic surfactants based on rosin-
maleic anhydride and acrylic acid adducts. J. Dispersion Sci. Technol. 31, 567-576 (2010).
[2] Bhairi, S. M., & Mohan, C. (2007). Detergents - A guide to the properties and uses of detergents in biology and biochemistry. Retrieved February 28, 2012 from
The Wolfson Centre for Applied Structural Biology: http://wolfson.huji.ac.il/purification/PDF/detergents/CALBIOCHEM-DetergentsIV.pdf
[3] Billings, B. H., & Gray, D. E. (1972). American Institute of Physics handbook (3rd ed.). New York: McGraw-Hill.
[4] Carazzo, D., Wohlfeil, A., & Ziegler, F. Dynamic surface tension measurement of water surfactant solutions. J. Chem. Eng. Data 54, 3092-3095 (2009).
[5] Du Noüy, P. L. An interfacial tensiometer for universal use. J. Gen. Physiol. 7, 625-631 (1925).
[6] Lange, N. A., & Speight, J. G. (2005). Lange's handbook of chemistry (16th ed.). New York: McGraw-Hill.
[7] Liao, X., Gautam, M., & Zhu, H. J. Effect of position isomerism on the formation and physicochemical properties of pharmaceutical co-crystals. J. Pharm. Sci. 99,
264-254 (2010).
[8] Lunkenheimer, K., Lind, A., & Jost, M. Surface tension of surfactant solutions. J. Phys. Chem. B, 107, 7527-7531 (2003).
[9] Salager, J.-L. (2002, December 15). Surfactants - Types and Use. Retrieved February 28, 2012 from Nanoparticles.org:
http://nanoparticles.org/pdf/Salager-E300A.pdf

9 Appendix
9.1 Literature Values for Water Surface Tension
Temperature Experimental Literature Value Table 1: Literature Values for Water Sur-
ºC K Result CRC Handbook AIP Handbook Lange’s Handbook face Tension from CRC Handbook, AIP
0 273.15 76.17 ± 1.04 75.65 75.60 75.83 Handbook, and Lange’s Handbook. The
5 278.15 73.87 ± 0.48 74.94 74.90 75.09 values are derived from three published
10 283.15 67.48 ± 1.69 74.22 74.22 74.36 handbooks. Highlighted values demon-
15 288.15 70.04 ± 2.60 73.49 73.49 73.62 strate complete agreement with experimen-
20 293.15 68.51 ± 2.08 72.74 72.75 72.88 tal data range. Since CRC Handbook pro-
25 298.15 70.04 ± 3.22 71.97 71.97 72.14
vides temperatures with increments of 2K,
69.27 ± 2.26 odd number temperatures are estimated by
30 303.15 62.88 ± 0.48 71.19 71.18 71.40 averaging adjacent values with common
35 308.15 65.44 ± 1.53 70.40 - 70.66 difference of 1K. 273.16K is estimated to be
40 313.15 67.23 ± 1.43 69.60 69.56 69.92 273.15K and 371.05K is estimated as
45 318.15 63.65 ± 1.30 68.78 - - 371.15K. The first surface tension for
50 323.15 64.16 ± 2.23 67.94 67.91 68.45 298.15K is taken from initial measurement
55 328.15 62.88 ± 1.30 67.10 - - of water surface tension at room tempera-
60 333.15 61.12 ± 1.11 66.24 66.18 66.97 ture and standard condition. The bottom
65 338.15 65.95 ± 1.43 65.37 - - value is taken during heating the substance.
70 343.15 58.79 ± 2.29 64.48 64.40 65.49
75 348.15 53.17 ± 2.38 63.58 - -
80 353.15 51.64 ± 2.27 62.67 62.60 64.01
85 358.15 48.82 ± 1.10 61.75 - -
90 363.15 47.29 ± 0.40 60.82 - 62.54
95 368.15 50.61 ± 1.74 59.87 - -
97.9 371.05 50.36 ± 1.25 59.30 - -