Characterising the phase behaviour of stearic acid and its triethanolamine soap and acid-soap by infrared spectroscopy

Article abstract of DOI:10.1039/b819582j

The behaviour of stearic acid neutralised by triethanolamine to form soap and its acid-soap has been examined by infrared spectroscopy. It was found that not only could the neutralisation behaviour be characterised, but the thermotropic behaviour could also be followed. The neutralisation confirmed the formation of a fixed stoichiometeric ratio, 2:1, acid-soap. When following the thermotropic behaviour the break up of the acid-soap could be followed along with various disordering and melting transitions of the alkyl chain tail. This allowed all the thermal transitions that have been observed to be characterised in terms of the type of molecular rearrangement that was occurring and also the transition temperature at which they occurred. This allowed the binary phase diagram to be plotted and understood for this system. This is the first time IR has been used to measure a whole phase diagram of this type. The nature of the acid-soap complex itself was also characterised, with very short hydrogen bonds being present as well as a free, non-hydrogen bonded, hydroxyl group. the Owner Societies 2009.

Full text of DOI:10.1039/b819582j

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Characterising the phase behaviour of stearic acid and its
triethanolamine soap and acid–soap by infrared spectroscopy
Paul D. A. Pudney,* Kevin J. Mutch and Shiping Zhu
Received 4th November 2008, Accepted 13th March 2009
First published as an Advance Article on the web 3rd April 2009
DOI: 10.1039/b819582j
The behaviour of stearic acid neutralised by triethanolamine to form soap and its acid–soap has
been examined by infrared spectroscopy. It was found that not only could the neutralisation
behaviour be characterised, but the thermotropic behaviour could also be followed. The
neutralisation confirmed the formation of a fixed stoichiometeric ratio, 2 : 1, acid–soap.
When following the thermotropic behaviour the break up of the acid–soap could be followed
along with various disordering and melting transitions of the alkyl chain tail. This allowed all the
thermal transitions that have been observed to be characterised in terms of the type of molecular
rearrangement that was occurring and also the transition temperature at which they occurred.
This allowed the binary phase diagram to be plotted and understood for this system. This is the
first time IR has been used to measure a whole phase diagram of this type. The nature of the
acid–soap complex itself was also characterised, with very short hydrogen bonds being present as
well as a free, non-hydrogen bonded, hydroxyl group.
used to investigate the chemical state and interactions of
1. Introduction
carbonyl and carboxylate groups12 and there is also an
extensive literature on how alkyl chain behaviour can be
studied using IR, including much work on fatty acids and
soaps.13,14 The polymorph of stearic acid can also be
determined by IR spectroscopy.15 It would therefore appear
to be a good method to apply to acid–soap complexes, as has
previously been observed by Lynch in his review on
acid–soaps.2 It should in principle be possible to characterise
both the carbonyl/carboxylate head group behaviour and also
that of the alkyl chains. There have been a limited number of
FTIR studies on acid–soaps, which further highlight its
usefulness; these mainly concentrate on the head group
behaviour at ambient temperatures.5 Two studies looked at
acid–soaps and their thermotropic behaviour,3,16 however
doubt has been cast on the nature of the starting complexes
due to the method of preparation.2 Despite these doubts, this
work showed that much information is available from the
infrared spectra in particular: the intermolecular interactions
between the acid and the soap as a function of the
neutralisation degree and temperature, which are impossible
or difficult for the conventional techniques such as
microscopy, DSC and XRD which are used in construction
of the binary phase diagram. Also, as the nature of the
acid–soap itself is not totally understood, some further insights
into this should be obtainable from the IR spectra.
The first allusion to acid–soaps was made by Chevreul in
1823,1 and although their existence has often been disputed,
there is now comprehensive proof.2 These compounds are
of great interest academically but also very important
commercially, being found in superfatted, transparent and
traditional soap bars, cleaning products, cosmetics, facial
cleaners, shaving creams, deodorants and topical-delivered
products.2 Additionally these types of molecules are thought
to be of biological importance.3 Generally, acid–soap
complexes are a group of molecular compounds that form as
a consequence of a strong hydrogen bond between the fatty
acid and its soap (carboxylate ion).4 They consist of crystals
containing carboxylic acid and, usually, metal carboxylate
ions. The amount of these components is stoichiometrically
discrete and are given by the general formula MxHyAz with X,
Y and Z integers (X + Y = Z), M usually being the metal
counter ion, mostly K+ or Na+5–9 and A is the alkanoate ion
of the form CH3(CH2)nCOOꢀ with n between 6 and 20.2 This
study differs in that the neutralisation is of stearic acid
(CH3(CH2)16COOH) by triethanol amine (TEA, N(C2H4OH)3).
There are very few studies published on amine soaps,10 despite
their wide use in the cosmetic industry where they are
commonly used to replace alkali metal neutralised fatty
acids.11
The conventional methods of studying these and similar
lipid/fat systems are to use XRD to give crystallographic
spacings and DSC to observe the thermal events during phase
transitions. These methods usually allow a phase diagram to
be produced; indeed recent work has used these methods for
this stearic acid/TEA system.10 FTIR has for many years been
It has recently been shown that neutralisation of stearic acid
by TEA produces a stable 2 : 1 acid–soap complex and its
thermotropic behaviour was explored.10 Here we show in
detail how a full range of neutralisations of stearic acid with
TEA and their thermotropic behaviour can be understood by
using FTIR. This is to see if insights that have been gained in
studies by infrared spectroscopy in acid and lipid systems can
be used to obtain further information for the type of systems
under study here.
Unilever Discover, Colworth Science Park, Sharnbrook, Bedford,
UK MK44 4LQ. E-mail: Paul.Pudney@unilever.com
ꢁc
This journal is the Owner Societies 2009
5010 | Phys. Chem. Chem. Phys., 2009, 11, 5010–5018

View Article Online
2. Experimental
Stearic acid and triethanolamine (TEA) were provided by
Fisher Scientific. The purity of stearic acid and TEA were
98% and 99% respectively. The neutralisation degree of
stearic acid is calculated based upon the molar ratio of stearic
acid to TEA and samples with different neutralisation degrees
were prepared. These were, 5% (TEA), 13%, 20%, 25%, 33%,
40%, 50%, 60%, 70%, 80%, 85%, 90% and 100%. To
prepare them, stearic acid was melted at 80 1C in a water bath
and TEA was then added to the melt whilst being stirred. The
temperature of the mix was then increased to 95 1C to obtain
an isotropic solution. The solution was then cooled and
solidified in an ambient environment. The resulting solid was
ground down to form a powder. The powder was stored in a
closed container for at least 7 days before it was analysed.
The infrared measurements were carried out using a Biorad
(now Varian) FTS 6000 FTIR spectrometer. The samples were
measured using a diamond ATR ‘golden gate’ with a uniform
and reproducible covering of sample; as the ATR method
measures the same volume above the surface (given the same
refractive index), this implies a constant amount of sample is
examined. The temperature was controlled using a heatable
top-plate with a quoted temperature accuracy of 1 1C
(Greasby Specac UK. Ltd). Spectra were taken at resolution
of 2 cmꢀ1 and each spectrum consisted of 256 scans co-added
and ratioed against an air background spectrum. A series of
background spectra were collected before each experiment
according to the temperatures at which measurements were
to be made. The correct background was then chosen to ratio
against the sample spectra according to the temperature.
Fig. 2 Infrared spectra of the fingerprint region at room temperature
of 100% TEA neutralised sample through to stearic acid (0% TEA
neutralised). Spectra have been offset for clarity.
shows a carbonyl peak at 1698 cmꢀ1 and shows no carboxylate
peak (see later, 1575 cmꢀ1 for the fully neutralised soap). This
is expected as most fatty acids exist in dimeric form, hydrogen
bonded to another fatty acid molecule and hence have no
dissociated COOH groups.17,18 The carbonyl band does have a
low frequency shoulder, which is consistent with the stearic
acid being in its most stable C crystal form. In this form,
there is a cis–trans isomerisation around the C1–C2 bond
(i.e. carbonyl carbon and the first carbon in the alkyl chain)
and hence the carbonyl is in two different environments thus
leading to the spilt into two bands. This is confirmed by the
CQO in-plane deformation which also appears as a doublet
at 690 cmꢀ1 and 670 cmꢀ1 due to cis and trans isomers
respectively,14,19–22 see Fig. 2.
The 100% neutralised sample shows a carboxylate peak at
1576 cmꢀ1 as expected for the soap. However when TEA is
added at 5% there is no band at 1575 cmꢀ1 but a band appears
at 1530 cmꢀ1. This is a similar frequency to that observed by
Lynch et al.6 for a 1 : 2 sodium palmitic acid–soap. As more
TEA is added the carbonyl band decreases in intensity and
shifts to a slightly lower frequency, typically 1692 cmꢀ1, whilst
the carboxylate peak increases in intensity and also shifts to a
slightly lower frequency. Above 33% however, the nature of
the reaction changes; the further addition of TEA produces a
band close to 1575 cmꢀ1 due to the presence of soap. The
lower frequency carboxylate peak due to acid–soap is still
present and can still be observed when 90% TEA has
been added.
3. Results and discussion
3.1 Stearic acid and neutralisation with TEA
As the starting molecule for this study is stearic acid, its
infrared spectrum from 600 to 3500 cmꢀ1 is shown in Fig. 1
and the main observed bands, their frequencies and
assignments with appropriate references are given in Table 1.
The spectra from the fingerprint region of the resulting
neutralisation of stearic acid to different degrees are shown
in Fig. 2. A number of changes can be seen in the spectra,
but the most diagnostic for acid–soaps are changes in the
carbonyl/carboxylate region.2,3,5 For clarity this region is
shown in close up in Fig. 3. It can be seen that stearic acid
Thus initially when stearic acid is neutralised it produces an
acid–soap, giving a mixture of acid–soap and unreacted stearic
acid. This continues until 33% after which soap is produced,
giving a mixture of soap and acid–soap. As 33% is the critical
point, this confirms other work10 that the acid–soap complex
consists of a 2 : 1 mixture of acid and soap. On plotting the
intensity of the carbonyl band, ICQO, there is a linear
relationship between ICQO and the % TEA (not shown) and
allows the direct deduction of the degree of neutralisation for
any unknown sample, if so desired.
The observation of the frequency shifts provides
information on the nature of the acid–soap complex. The
carboxylate of the acid–soap is likely to be more hydrogen
bonded than soap as the frequency is lower, similar to that
observed previously. The carbonyl shows a lowering of the
Fig. 1 Infrared spectrum of stearic acid at room temperature.
ꢁc
This journal is the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 5010–5018 | 5011

View Article Online
Table 1 Band assignments
Wavenumber/cmꢀ1
Assignment
Strength
B688
B717
B719/728
B760
B783
B809
B888
B912
B940
O–CQO in plane angle deformation; SA, 5%, 13%, 20%, 25%9,29,36
All in phase CH2 rock; 33% and above9,37
Split all in phase CH2 rock; SA, 5%, 13%, 20% and 25% (sh)2,5
CH2 rock36
m-np
vs
vs
m-w
m-vw
m-w
m-w
m-s
vs-np
vw
vw-s
vw-s
m-s
w-np
np-w
s
CH2 rock36
CH2 rock36
CH3 rocking mode; SA, 5%, 13%, 20%9,32,36,37
CH2 rock36
COH out of plane deformation (distinct up to 25%)17,29,36
CH2 rock36
B1017
B1033
B1058
B1104
B1123
B1129
B1186–1348
B1297
B1371
B1378
B1397
B1410
B1424–1430
B1439
B1462/1471
B1467
B1528
B1576
B1693
B1698
B2848
B2871
B2894
B2915
B2932
B2953
2961
C–C stretch36
Antisymmetric CH2 rock: strong for high % (signifies 8 or more trans-CH in methylene chain)34,38
Symmetric CH2 rock with CH3 sided chain38,39
C–C stretch36
Symmetric CH2 rock with COOꢀ sided chain: only high% (16–20 or more t-CH)36,38
CH2 wagging progression9,36
Mostly C–OH stretch (SA, 5%, 13%, 20%)17,29,36
CH3 symmetric deformation18,27,36
s-w
vw
CH3 umbrella deformation9,28
np-vw
np-s
m-np
m-np
np-w
s
COOꢀ symmetric stretch-appears as shoulder for low %, not SA9
Bending of CH2 adjacent to carboxyl group (a-CH2)28,29,36
C–O stretching and COH in plane bending (up to B70%)6,17,29,36
COH in plane bending (25–60% sh, 70% -)29
CH2 scissor (up to 25%)2,5
CH2 scissor (higher than 25%)2,5
m
COOꢀ antisymmetric stretch, acid–soap: all, but 100, 90, 85, 80 are sh5,6
COOꢀ antisymmetric stretch, soap: only above 33%5,6
CQO high %6
w-m-sh
np-w-s
w-np
vs-w
vs
m
sh
vs
sh
CQO low %6
CH2 symmetric stretch9,38,40,41
CH3 symmetric stretch9,38,40,41
Fermi resonance40,41
CH2 antisymmetric stretch9,38,40,41
Fermi resonance40,41
CH3 antisymmetric stretch (out of plane)9,38,40,41
CH3 antisymmetric stretch (in plane)9,38,40,41
m
w
High T
B846
B1211
B1280
B1301
B1340
B1367
CH3–CH2 rocking mode for tgttt (high %)32
w
w
w
w
w
w
CH2 wag; all except SA, 5%, 13%32
CH2 wag; SA, 5%, 13%, 20% and 70%, 80%, 85%, 90%, 100%32
CH2 wag-gtg0 sequence: 40% @ 1297, 50% @ 1299, 70–100% @ 1301–13039,28,32,34,35
CH2 wag end-tg (etg) sequence; all, mainly shoulder for high%9,28,32,35
Out of phase CH2 wag of gtg0 sequences9,28,32
Abbreviations: t = trans; g = gauche; s = strong; m = medium; w = weak; v = very; sh = shoulder; np = not present; SA = stearic acid. Where
strength assignments vary, the first is for low % TEA and the changes occur as % TEA increases.
should be available from the n OH, although this is compli-
cated by the presence of OH groups in the cation. The full
acid–soap complex spectrum is shown in Fig. 4. There are
three distinct bands present above 3000 cmꢀ1 for the
acid–soap complex, the lowest of which is likely to be from
the TEA counter ion. The band at approx. 3500 cmꢀ1 is at the
right frequency for a free OH in a carboxylic acid, suggesting
one OH is sterically restricted from interacting. As the
carbonyl peak suggests the presence of a hydrogen bonding
interaction, the other OH should be observed as a broad band
at lower frequency. Hydrogen bonded OH bands in carboxylic
acids would normally be expected to occur well below
3000 cmꢀ1, but no significant broad band is observable in
Fig. 3 The CQO (B1710 cmꢀ1) and COOꢀ stretching region
(B1580 to 1530 cmꢀ1) for all samples at room temperature, labelled
according to their origin, see text.
the spectrum displayed in Fig. 4.
In the acid–salts of carboxylic acids, the short chained
equivalents of acid–soaps have been observed to have
‘very short’ hydrogen bonds.23 These have been observed to
give rise to extraordinary lowering of the frequency of the OH
stretch.24 Two types of these ‘very short’ and hence very strong
frequency from that of the acid alone, which is in contrast to
the observation for the akali metal acid–soaps where the
frequency was shown to increase.3,6 Further information
ꢁc
This journal is the Owner Societies 2009
5012 | Phys. Chem. Chem. Phys., 2009, 11, 5010–5018

View Article Online
agree, as for low % TEA samples up until 33% there is still
free stearic acid present, but as the amount of stearic acid
decreases the degree to which splitting occurs also decreases.
The similar splitting of the CH2 scissor at 1462 cmꢀ1 and
1472 cmꢀ1 echoes the observation of the splitting of the CH2
rock. Again, this is only observed clearly up to 20% and for
25% it is a shoulder. Thereafter this region shows only a single
peak at 1467 cmꢀ1. This confirms the conclusion from the CH2
rock behaviour.
3.2 Thermotropic behaviour
It is known that alkali n-alkanoates generally undergo more
than one phase transition between a crystalline solid and
isotropic liquid. The number of transitions depends on n, the
number of carbon atoms; for members with n 4 6, an
anisotropic liquid phase (liquid crystal) is present before
reaching the isotropic melt.9 It is apparent therefore that for
complex acid–soap structures, it can be expected that there will
be more than one structural transition between solid and
liquid. These type of phase diagrams have been described
before30,31 and are called peritectic. The transitions which
are expected have been ascribed by DSC measurements for
this system and are published in detail elsewhere.10 The
temperatures of the transitions observed in this previous work
will be compared to the FTIR results described here. These
different transitions in other fatty acid/soap/acid–soap systems
have been found to be due to a number of different types of
transition. These transitions have different effects on the head
groups and the fatty acid chains and so different changes could
be expected for vibrations originating from those parts of the
molecules. Initially we will describe changes of vibrational
modes from the head group and the fatty acid chains
separately.
Fig. 4 FTIR spectrum of acid–soap. The black line shows location of
possible underlying strong hydrogen bonded OH band.
hydrogen bonds have been classified by Speakman23 as either
type A or B, where the bonds are either symmetrical (i.e. the
hydrogen is equidistant for the two oxygens) or unsymmetrical
(where they are not). These show quite different IR spectral
effects; a good summary of how the OH stretch changes with
bonding distance and symmetry is given by Bratos.25 Upon
examining Fig. 4 it can be seen that there appears a very broad
band underlying the spectrum from approximately 1600 cmꢀ1
to below 700 cmꢀ1 in the acid–soap, shown by an added black
curve. A band of this type would also expect to have ‘Evans
holes’ or windows,26 but these would be extremely difficult to
observe with so many other bands being present, although
there are possible candidates which appear to be below the
general level of background. A band structure of this type
would be produced by nearly symmetrical hydrogen bonds
with a length of 2.5 A or below, i.e. very short and strong in
nature. This corresponds well to the distance determined in
NaHP2 (P = palmitate) by X-ray diffraction of 2.44 A.6 The
IR spectrum of this acid–soap was also shown, and it
was suggested that the OH stretch is evident between
3000–2200 cmꢀ1, despite its presence not being obvious.
However, an underlying broad band appearing largely in the
fingerprint region could also be seen. This is consistent with a
short hydrogen bond of the length suggested and hence we
would assign this to the OH stretch. This was also observed
and pointed out by Hadzi,16 despite some concerns over the
nature of these complexes.2
The vibrational spectroscopic work of Zerbi et al.19 showed
that the C form of stearic acid exists as a trans planar molecule
with the carboxyl group co-planar with the chain, as has also
been suggested by other authors.14,27,28 This work demon-
strates that, as is known for alkanes and other lipid systems,
studying the carbon hydrogen vibrations of the methylene
chains can tell you much about its conformation. Stearic acid,
neutralised by 5%, 13% and 20% TEA, shows definite
splitting of the in-phase CH2 rock at 718 and 728 cmꢀ1
(see Table 1 and Fig. 2). As the amount of TEA is increased,
the intensity of the 728 cmꢀ1 peak decreases until it is only a
shoulder for 25% and is not observed for 33%, leaving only a
single peak around 717 cmꢀ1. This splitting of the in
phase rock indicates that the sample is crystalline with an
orthorhombic sublattice.29 The C form of stearic acid exists in
a monoclinic form with the polymethylene chain comprising
the sublattice of the orthorhombic type. These observations
3.2.1 Fatty acid alkyl chains/C–H vibrations. Following on
from the description above of the behaviour of the vibrations
associated with the methylene chain with neutralisation,
further changes would be expected with temperature, such as
various melting and disordering/reordering phenomena. Thus
this section will detail the observations that have been made on
the various CH vibrations and how they can be used to clearly
characterise the changes that occur in the fatty acid methylene
chains with temperature.
On heating, the fatty acid chains generally change in
conformation, often losing their ordered structure with the
introduction of gauche bonds and rotation of the end methyl
group. The effect of these changes should be primarily on the
spectral regions dealing most clearly with the C–H bonds.
There is much work in the literature describing vibrational
spectroscopy on this type of change, mostly developed from
the seminal work of Snyder.32
3.2.1.1 CH stretching region. It is known that the main C–H
stretching region between 2800 and 3000 cmꢀ1 yields informa-
tion about the degree of disorder in the structure.33 Some
changes were observed in this region with neutralisation.
Observing the CH2 symmetric stretch for all samples at
room temperature, it was noted that there is no systematic
correlation between the amount of TEA present and the
ꢁc
This journal is the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 5010–5018 | 5013

View Article Online
three. A similar form has been observed for some alkali metal
acid–soaps, but the origin of the transitions was not fully
identified.3 The origin of these transitions is due to disordering
and/or melting, however to determine the precise nature of the
different transitions, these changes need to be considered
in conjunction with the observations of the head group
vibrations.
Looking at the other C–H vibrations, it is clear that the CH2
bend and CH2 scissors both reflect the transitions for low%
TEA samples. As was mentioned, these bands both show
factor group splitting for stearic acid and the 5%, 13% and
20% TEA samples due to the orthorhombic packing of the
polymethylene chains. On heating, the peak at higher
frequency begins to decrease in intensity around the
temperature of the first transition and is not observed at all
when it is an isotropic liquid. This also reflects the disordering
of the polymethylene chain which results in the break up of the
orthorhombic packing.
Fig. 5 The main C–H spectral region as measured for 85% TEA on
increasing temperature. Showing how the CH2 bands shift and broaden
as the alkyl chain goes through a disordering/melting transition.
intensity of this peak. Although a gradual decrease in intensity
as the amount of TEA increases was observed, it was not a
linear relationship. The exact frequency of this peak was
observed to increase until 40% TEA, whereby it reaches a
maximum and then begins to decrease again. This band was
observed for all samples with increasing temperature, with
changes being taken relative to those at room temperature. An
example of the typical behaviour observed is shown in Fig. 5,
displaying the C–H stretching region for 85% TEA. It can be
seen that along with a gradual decrease in intensity, there is
also a shift to higher frequency as the temperature increases.
This occurs smoothly until it reaches a critical point, i.e. a
3.2.1.2 C–H wagging region: 1150–1350 cmꢀ1. One of the
regions which is of most interest in understanding structural
changes is between about 1150 and 1350 cmꢀ1. This series of
bands arises due to the methylene wag. Chapman concluded
that the number of bands in the series is half the number of
CH2 groups in the acid chain.13 The wag arises when all of the
H atoms of the CH2 groups move in the same direction along
the whole chain and this in phase bend gives rise to the series
which is observed. The wagging progression is much stronger
in fatty acids than in the corresponding hydrocarbons for two
reasons. Firstly, CH2 waggings are strongly coupled with the
motions of the CQO and C–O bonds which contribute
significantly to the changes in the transition moments, and
secondly, the carboxyl group induces a strong polarity of
the C–C bonds along the chain; this polarity decreases as the
distance from the carboxyl group increases. This makes the
wagging much stronger for acids than the virtually non-polar
hydrocarbons. 13 Not all the bands in this region arise from the
wags; the origins of some of the others are documented in
Table 1.
temperature at which
a transition occurs, whereby it
undergoes a large change. The shift of the frequency of the
symmetric CH stretch peak at B2850 cmꢀ1 was plotted and is
shown in Fig. 6. Similar plots were obtained for all of the
samples.
It can be clearly seen from Fig. 6 that there is more than one
gradient change between the lower and higher frequency
regions (from B2848.5 cmꢀ1 to B2852 cmꢀ1), showing that
at least two transitions are involved. The temperature at which
these changes occur was obtained by taking the midpoint of
the two lines involved. A third change can be observed on the
higher temperature part and is discussed below, but it is not
clear in this plot. For stearic acid, 25%, 33% and 40%, one
transition was clearly observed. For 5%, 13%, 20%, 50% and
100% there were two clear transitions, and for the remaining
compositions (60%, 70%, 80%, 85% and 90%) there were
As the substances are heated, the methylene groups of the
long fatty acid chains begin to change conformation such that
they are no longer all trans. As this all trans structure is broken
up, it causes the discrete methylene wagging structure in the
spectra to lose its form, as it relies on the coupling between the
vibrations of adjacent methylene groups. As the temperature is
increased, the degree of disorder in these chains increases.
Hence, the methylene wagging structure is no longer apparent
in the spectrum as the isotropic liquid is reached. Instead, this
is replaced by a number of very small bands which arise from
other conformations of the acid chain (as assigned in Table 1).
New bands are observed at B1211 cmꢀ1, B1280 cmꢀ1
,
B1301 cmꢀ1, B1340 cmꢀ1 and B1367 cmꢀ1 as the samples
are heated. These bands are not uniformly present for all of
the samples, sometimes being of too small intensity or being
masked by other larger peaks. The band found around
1301 cmꢀ1 has been suggested as arising from the in phase
wagging motion of the two CH2 groups involved in the
gauche–trans–gauche0 (gtg0) conformation. Similarly, the band
around 1367 cmꢀ1 is the out of phase form of this vibration.24
Fig. 6 Plot of the CH2 symmetric band frequency for 85% TEA
sample, as shown in Fig. 5, with increasing temperature.
ꢁc
This journal is the Owner Societies 2009
5014 | Phys. Chem. Chem. Phys., 2009, 11, 5010–5018

View Article Online
Fig. 9 Plot of intensity of the 1367 cmꢀ1 band relative to the
1378 cmꢀ1 band, I1367/I1378, for 80% TEA with diamonds being initial
measurements and triangles repeated low temperature measurements.
Fig. 7 Shows the changes in the bands at 1301 cmꢀ1 and 1276 cmꢀ1
for 85% TEA between 74 1C and 104 1C.
The same authors suggest that the mode near 1340 cmꢀ1 is
caused when the CH2 group adjacent to the terminal CH3
group is placed between a trans and a gauche conformation
(end tg, etg group). The bands at around 1211 cmꢀ1 and
1280 cmꢀ1 have been attributed to C–C–C bending and
methylene wagging, respectively34 although the exact
conformations giving rise to these is unknown.
and a methylene wag arising from an end trans–gauche
sequence (1341 cmꢀ1). In this case however, it is not possible
to observe the band at 1341 cmꢀ1 for low temperatures because
this is obscured by the main CH2 wagging progression. There is
a band at 1367 cmꢀ1 however which, according to Snyder,32
indicates a gtg0 sequence. This band is out of the region of the
CH2 wag and hence is not obscured. The relative intensities
of these changes with temperature can be determined by
examining the spectra (not shown). The CH2 etg band will
be affected considerably by temperature, whereas the CH3
umbrella will be affected differently as it is less strongly
coupled to the adjacent groups/alkyl chain. Fig. 9 shows a
On observing the spectra at higher temperatures, it is clear
that the frequency behaviour of the band at approximately
1301 cmꢀ1, correlates to the weak transition that was observed
at the higher temperature in the CH stretch region. This is the
change upon going from liquid crystal to isotropic liquid
phases. Plotting this change allows an observation to be made
about the temperature of the transition. It is also observed that
in the liquid crystal phase, the peak at 1301 cmꢀ1 is generally
more intense than the adjacent peak at 1276 cmꢀ1, as shown in
Fig. 7 for 85% TEA. Upon melting however, the reverse is
true. A plot of I1301/I1276 is shown in Fig. 8, making the phase
transition much clearer to observe, and thus allowing the
transition temperature to be determined more accurately.
plot of the relative intensities—I1367/I1378
.
This graph shows a decrease in the intensity ratio between
about 60 1C and 68 1C. The midpoint of this decrease is
approximately 64 1C which agrees well with the previous
observation (see Table 3). The next major change is between
69 1C and 72 1C and the midpoint of this is 70 1C, also in good
agreement. For the higher temperature transition, the change
is less clear. However, given the knowledge that there is a
transition taking place around 86–87 1C, it can be seen that at
this point in the above graph there is a slight increase in the
relative intensity, which indicates a transition. Thus the plot of
the relative intensities of the 1378 cmꢀ1 and 1367 cmꢀ1 bands
does allow the prediction of all transitions taking place; it
confirms the behaviour observed and gives further information
about the type of disorder processes taking place.
3.2.1.3 Other bands revealing heating phenomena. Although
the main bands that were used in determining the temperatures
of transition have been described, it is useful to look at bands
which may uncover some further details about the structure of
these compounds.
Work by Snyder and co-workers detailed the degree of
conformational disorder of waxes by comparing the intensities
of two bands at 1341 cmꢀ1 and 1375 cmꢀ1 35 These bands were
.
assigned as being the CH3 umbrella deformation (1375 cmꢀ1
)
3.2.2 Head group vibrations. Interesting deductions can be
made about the shift of the carbonyl and carboxylate peaks on
heating. As mentioned previously (Table 1) the peak occurring
around 1700 cmꢀ1 is due to the carbonyl group, the peak at
1530 cmꢀ1 arises from the carboxylate groups of the acid–soap
complex and the peak around 1575 cmꢀ1 is due to the
carboxylate groups of free soap.
On heating stearic acid to 68.5 1C, whereupon it melts, its
carbonyl peak shifts by 12 cmꢀ1 to 1710 cmꢀ1. This is due to
the fact that in its crystalline form the acid exists dimerised to
another stearic acid molecule, but on melting this association
is less ordered, leaving more electron density in the CQO bond
and hence making it stronger.6 The effect is also observed for
all the samples below 33%: the carbonyl peak always shifts to
approximately 1710 cmꢀ1 upon melting. The overall shift is
Fig. 8 Plot of 1301 cmꢀ1 band intensity relative to 1276 cmꢀ1 band
for 85% TEA, i.e. I1301/I1276
.
ꢁc
This journal is the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 5010–5018 | 5015

View Article Online
Table 2 Transitions occurring in acid–soap complexes
Sample
Transition
Peak used
Stearic acid
Acid/liquid
Acid–soap/acid
Acid/liquid
Acid–soap/liquid
Acid–soap/soap
Soap/liquid crystal
Liquid crystal/liquid
Soap/liquid crystal
Liquid crystal/liquid
2850 cmꢀ1 (CH2 symmetric)
2850 cmꢀ1 and carboxylate
2850 cmꢀ1
5%–25%: 1st transition
5%–25%: 2nd transition
33%
40%–90%: 1st transition
40%–90%: 2nd transition
40%–90%: 3rd transition
100%: 1st transition
100%: 2nd transition
2850 cmꢀ1
1710 cmꢀ1 (CQO) and 2850 cmꢀ1 and carboxylate
2850 cmꢀ1
1301/1276 cmꢀ1 (CH2 wags)
2850 cmꢀ1
1301/1276 cmꢀ1
larger however, given the lower frequency of the band initially
(approx. 1693 cmꢀ1).
The carbonyl group of the acid occurs at 1697 cmꢀ1 and the
same group for acid–soap is typically 1692–1693 cmꢀ1. It is
expected that if the acid–soap goes to acid, the carbonyl peak
will shift from 1693 cmꢀ1 to 1697 cmꢀ1 to reflect the changing
environment. It would then shift to 1710 cmꢀ1 on going to the
isotropic liquid. It was observed that the carbonyl peak does
not shift to the final frequency 1710 cmꢀ1 until total melting
has occurred and it changes from the acid–soap to acid, with
the acid melting occurring separately.
It is observed that the carboxylate peak of the acid–soap,
acid mixtures also change on heating. The carboxylate group
shifts considerably at the same temperature of the carbonyl
shift. For 5% TEA the shift is from 1530.2 cmꢀ1 to
1554.8 cmꢀ1. On complete melting, it then decreases consider-
ably in intensity although it does not change its position. This
shift in frequency is explained by the changing environment of
the carboxylate group. Thus when acid and acid–soap are
present, it can be seen on melting that the acid–soap complex
breaks up. This is shown by the frequency shift of the
carboxylate band to a frequency similar to that observed for
the soap, with only bands from the acid and soap observed.
Similar changes are observed for all the samples up to 33%.
For the samples of above 33% TEA, the typical behaviour
of the carboxylate region is shown in Fig. 10 (for 50% TEA).
The acid–soap carboxylate peak at 1530 cmꢀ1 and the soap
Fig. 10 Acid–soap (B1530 cmꢀ1) and soap (B1580 cmꢀ1) COOꢀ
peaks on heating for 50% TEA sample. Room temperature spectrum
at the top and highest temperature spectrum at the bottom.
Thus to summarise the observations above for both the
head group vibrations and the tail/CH vibrations, it has been
shown that all of the changes that occur for the acid/acid–
soap/soap phase diagram as identified previously10 can be
observed in the infrared spectra. These are acid melting to
liquid, acid–soap break up to acid and soap, acid–soap to
liquid, soap to liquid crystal and liquid crystal to isotropic
liquid. Also, as a lot of these changes show sharp temperature
transitions, they can be used to measure the temperature at
which they occur. This has been done in this study using the
most appropriate bands often in combination with another
band to confirm the type of transition that is occurring. The
bands that have been used are summarised in Table 2. The
temperatures obtained can then be used to plot the phase
band at 1580 cmꢀ1 merge into one at just under 1560 cmꢀ1
.
Thus the acid–soap appears to be breaking up at this
temperature. Simultaneously the carbonyl peak shifts up to
around 1713 cmꢀ1, as was observed below 33% when both the
acid–soap breaks up and the acid melts. However, unlike
previously, this is not observed at the temperature at which
the isotropic liquid is reached, as judged from the CH
vibrations described earlier, although a disordering transition
is observed. The transition is thus due to acid–soap breaking
up, producing acid in the liquid state and soap. A further
disordering (i.e. melting) of the soap is observed at a higher
temperature, producing the isotropic liquid consisting of both
acid and soap.
A plot of the shift of the carbonyl band can be used to
measure the temperature of the transitions that are occurring,
as shifts of the carbonyl band are observed for both the
acid–soap break up to soap and acid and the acid melting
transitions. An example of such a plot of the shift of this band
is given for 85% TEA in Fig. 11.
Fig. 11 Shift of the CQO band frequency for 85% TEA with
increasing temperature, change due to the acid–soap to liquid acid
transition, see text.
ꢁc
This journal is the Owner Societies 2009
5016 | Phys. Chem. Chem. Phys., 2009, 11, 5010–5018

View Article Online
Table 3 Values obtained for transition temperatures
% TEA
Acid/liquid
Acid–soap/acid
Acid–soap/liquid
Acid–soap/soap
Soap/liquid crystal
Liquid crystal/liquid
0
5
68.5 1C
67.6 1C
67.2 1C
67.0 1C
65.5 1C
65.0 1C
65.7 1C
65.5 1C
65.5 1C
13
20
25
33
40
50
60
70
80
85
90
100
66.5 1C
66.5 1C
66.5 1C
65.5 1C
64.5 1C
64.3 1C
64.0 1C
61.5 1C
66.5 1C
67.0 1C
69.0 1C
69.0 1C
69.5 1C
70.5 1C
71.5 1C
71.8 1C
73.5 1C
75.0 1C
83.5 1C
89.0 1C
89.5 1C
91.0 1C
91.5 1C
Fig. 12 Comparison of the phase diagram changes obtained from FTIR (open symbols) and DSC (filled symbols). Lines demarcating the
approximate phase regions are also drawn, and the appropriate regions labelled.
diagram. The plot of this phase diagram is shown in Fig. 12,
with the temperatures for the transitions given in Table 3.
These values compare very well to those obtained by DSC
measurements,10 which are also plotted in the same Figure.
However the DSC measurements of course do not unambigu-
ously give the type of transition that is occurring.
This shows, for instance, when the acid–soap complex breaks
up and when various disordering/melting transitions of the
alkyl chain take place, ultimately resulting in the formation of
an isotropic liquid. The temperatures at which all these
transitions occur could be measured accurately, matching well
those obtained previously by DSC. Further insights into the
nature of the acid–soap complex itself have been garnered. It
has been shown to possess very short, strong, hydrogen bonds
and a ‘free’, non-hydrogen bonded, OH group.
4. Conclusions
This investigation has shown that infrared spectroscopy is a
very good method for investigating phase behaviour and
structural changes of stearic acid and its neutralisation
products. This is the first time that such a detailed phase
diagram has been constructed of such a system using FTIR
spectroscopy. It is a quick and robust method that is sensitive
to all the states encountered in the phase diagram. It was
shown that it could not only distinguish between them, but it
can give a detailed understanding of the local molecular
environment. This was done by monitoring the changes of
multiple bands from vibrational modes originating in both
the ‘head’ group and the ‘tail’, the alkyl chain. Thus the
neutralisation of stearic acid could be followed, showing
the formation of both the acid–soap complex and the soap.
The thermotropic behaviour was also characterised in detail.
Acknowledgements
Mary Hepenstall-Butler for useful discussions and Professor
Norman Sheppard FRS for discussions especially about the
nature of the acid–soap.
References
1 M. E. Chervreul, Recherches Chimiques sur les Corps
Gras d’Origine Animales Paris, 1923, 34(53), 408.
2 M. L. Lynch, Curr. Opin. Colloid Interface Sci., 1997, 2, 495–500.
3 H. H. Mantsch, S. F. Weng, P. W. Yang and H. H. Eysel, J. Mol.
Struct. (THEOCHEM), 1994, 324, 133–141.
4 P. Tandon, S. Raudenkolb, R. H. H. Neubert, W. Rettig and
S. Wartewig, Chem. Phys. Lipids, 2001, 109, 37–45.
ꢁc
This journal is the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 5010–5018 | 5017

View Article Online
5 M. L. Lynch, Y. Pan and R. G. Laughlin, J. Phys. Chem., 1996,
100, 357–361.
6 M. L. Lynch, F. Wireko, M. Tarek and M. Klein, J. Phys. Chem.
B, 2001, 105, 552–561.
7 J. Bian, S. F. Weng, J. G. Wu, L. Q. Yang, X. Zhang and
G. X. Xu, Mikrochimica Acta, 1997, 367–368.
8 T. Ishioka, Bull. Chem. Soc. Jpn., 1991, 64, 2174–2182.
9 M. V. Garcia, M. I. Redondo and J. A. R. Cheda, Vib. Spectrosc.,
1994, 6, 301–308.
10 S. Zhu, M. Heppenstall-Butler, M. F. Butler, P. D. A. Pudney,
D. Ferdinando and K. J. Mutch, J. Phys. Chem. B, 2005, 109,
11753–11761.
11 J. Falbe, Surfacants in Consumer Products, Springer-Verlag, Berlin,
1988.
12 L. J. Bellamy, Infrared Spectra of Complex Molecules, John Wiley,
1975.
13 D. Chapman, The Structure of Lipids, Methuen and Co. Ltd.,
London, 1965, ch. 4, p. 52.
14 M. Kobayshi, in Crystallization and Polymorphism of Fats and
Fatty Acids, ed. N. Garti and K. Sato, Marcel Dekker, 1988, ch. 4,
p. 139.
23 J. C. Speakman, Struct. Bonding, 1972, 12, 141–199.
24 D. Hadzi and B. Orel, J. Mol. Struct. (THEOCHEM), 1973, 18,
227–239.
25 S. Bratos, H. Ratajczak and P. Viot, Properties of H-Bonding in the
Infra red Spectral Range in Hydrogen Bonded Liquids,
ed. J. C. Dore, T. Teixiera, 2000, Nato Series C, vol. 329, pp. 221–35.
26 M. F. Claydon and N. Sheppard, Chem. Commun., 1969,
1431–1433.
27 R. G. Sinclair, A. F. McKay and R. Norman Jones, J. Am. Chem.
Soc., 1952, 74, 2570–2575.
28 M. Delzoppo and G. Zerbi, Polymer, 1990, 31, 658–662.
29 G. Zerbi, G. Conti, G. Minoni, S. Pison and A. Bigotto, J. Phys.
Chem., 1987, 91, 2386–2393.
30 D. M. Small, The Physical Chemistry of Lipids, Phenum Press,
1986, p. 332.
31 R. G. Laughlin, The Aqueous Phase Behaviour of Surfacants,
Academic Press, 1994, p. 472.
32 R. G. Snyder, J. Chem. Phys., 1967, 47, 1316–1360.
33 A. Watkinson, R. S. Lee, A. E. Moore, P. D. A. Pudney,
S. E. Paterson and A. V. Rawlings, Int. J. Cosmet. Sci., 2002, 24,
151–161, and references therein.
15 S. Zhu, P. D. A. Pudney, M. Heppenstall-Butler, M. F. Butler,
D. Ferdinando and M. Kirkland, J. Phys. Chem. B, 2007, 111,
1016–1024.
34 T. Ishioka, Bull. Chem. Soc. Jpn., 1991, 64, 2174–2182.
35 D. Clavell Grunbaum, H. L. Strauss and R. G. Snyder, J. Phys.
Chem. B, 1997, 101, 335–343.
16 D. Hadzi, A. Grdadolnik and A. Meden, J. Mol. Struct.
(THEOCHEM), 1996, 381, 9–14.
36 R. F. Holland and J. Rud Nielsen, J. Mol. Spectrosc., 1962, 9,
436–460.
17 D. Hadzi and N. Sheppard, Proc. R. Soc. London, 1952, 216,
247–266.
37 G. S. Yu, H. W. Li, F. Hollander, R. G. Snyder and H. L. Strauss,
J. Phys. Chem. B, 1999, 103, 1046–10468.
18 P. J. Corish and D. Chapman, J. Chem. Soc., 1957, 1746–1751.
19 G. Zerbi, G. Minoni and A. P. Tulloch, J. Chem. Phys., 1983, 78,
5853–5862.
20 S. Hayashi and J. Umemura, J. Chem. Phys., 1975, 63, 1732–1740.
21 J. Umemura, J. Chem. Phys., 1978, 68, 42–48.
22 S. Hayashi, J. Umemura and R. Nakamura, J. Mol. Struct.
(THEOCHEM), 1978, 69, 123–136.
38 P. Tandon, S. Raudenkolb, R. H. H. Neubert, W. Rettig and
S. Wartewig, Chem. Phys. Lipids, 2001, 109, 37–45.
39 P. Tandon, R. H. H. Neubert and S. Wartewig, J. Mol. Struct.
(THEOCHEM), 2001, 526, 49–57.
40 S. Abbate, G. Zerbi and S. L. Wunder, J. Phys Chem., 1982, 86,
3140–3149.
41 H. Binder and H. Schmiedel, Vib. Spectrosc., 1999, 21, 51–73.
ꢁc
This journal is the Owner Societies 2009
5018 | Phys. Chem. Chem. Phys., 2009, 11, 5010–5018