Anda di halaman 1dari 100

Page 83

3
The Analysis of Surfactants in Cosmetics
Jane M. Eldridge
Analytical Services, Rhône-Poulenc, Inc., Cranbury, New Jersey

I. Introduction 84

II. Anionics 84

A. Sulfates and Sulfonates 85

B. Phosphate Esters 86
C. General Tests 86

D. Isethionate Esters 87

III. Nonionics 87

A. Ethoxylated Alcohols and Phenols 88

B. Ethoxylated Fatty Acids 90


C. Alkanolamides 91

IV. Cationics 91

A. Amine Oxides 91

B. Quaternary Ammonium Compounds 91


V. Amphoterics 92

A. Active 93

B. Water 93

C. Salt 94

D. Amines 94
E. Acids 94

VI. Preservatives 94

A. General 94

B. Formaldehyde and Formaldehyde Releasers 95


C. Parabens 95
Page 84

D. Triclosan 95

E. Miscellaneous 96
VII. Impurities 96

A. Ethylene Oxide and Acetaldehyde 96

B. Dioxane 97

C. Nitrosamines 98

D. Sultones 98
VIII. Analysis of Formulated Products 98

A. General Separation Schemes 98

B. Preliminary Screening 99

C. Separation and Quantitation 100


D. Alkoxylate Ratios 101

Abbreviations 101

References 101

I
Introduction
The purpose of this chapter is to acquaint the cosmetic chemist with the methods used to analyze surfactants. The exact
conditions for carrying out the procedures are not given here but rather a general description that will allow the chemist to
select the appropriate analysis and obtain the specific conditions from the literature.

Since surfactants are raw materials in many cosmetic products the chemist should be aware of the quality control tests that are
used to ensure that the product is consistent from lot to lot. These tests confirm the nature of the surfactant itself, the character
and quantity of the impurities from the manufacture, and the amount of any preservative that may have been added to ensure
the absence of microbial contamination. The first section of this chapter is concerned with the quality control tests that are
used. It will describe the traditional methods and the newer techniques that are being used to reduce the time for the analyses
and to better characterize the surfactant. This section is organized by surfactant type for those analyses that are specific for a
particular type of surfactant and by analyte for the impurities and preservatives that may be found in various types of
surfactants. Cosmetic chemists may also want to identify and quantitate a surfactant present in a formulated product. The
second section of this chapter will deal with the separation of surfactants from mixtures and the identification and quantitation
of these surfactants.

Several useful texts are available that describe in detail the traditional methods used in the analysis of surfactants [110]. These
references include those most useful for the separation and identification of unknown surfactants [4,9,10].

II
Anionics

Anionic surfactants are those which owe their water solubility to the negative charge on the molecule and their lipid
compatibility to a long hydrocarbon chain.
Page 85

A
Sulfates and Sulfonates

A large number of anionic surfactants are either sulfates or sulfonates of fatty alcohols or ethoxylated fatty alcohols. The
quality control tests for these two groups have traditionally been very similar.

Active. The concentration of the anionic in the water solution is determined by a two-phase titration using a cationic surfactant
(usually Hyamine 1622) as the titrant and various indicators (usually methylene blue) to determine when the endpoint is
reached. Longman [3] explains how this titration works and how to select the best titrant and indicator. He also gives the
specific details for carrying out the assay. The ASTM procedure [11] provides the specific details for carrying out this analysis.
Although this two-phase technique has been relatively successful and gives good, reproducible results when carried out by an
experienced operator, it is subject to interferences. It requires a subjective decision by the operator and it requires a significant
amount of operator time.

In order to resolve some of these problems several electrode methods have been developed for the determination of anionic
active. The American Society for Testing and Materials (ASTM) method D 4251 [11] outlines a method for determining the
active by titrating an aqueous solution of the surfactant with Hyamine 1622 using a nitrate-ion-selective electrode. Orion has
developed a surfactant electrode that determines the endpoint when an anionic surfactant is titrated with benzethonium
chloride (Hyamine 1622). Oei et al. [12] have compared the standard two-phase titration using visual judgement of the
endpoint with the use of the electrode to determine the endpoint. Their work showed that the electrode gave a lower relative
standard deviation and a slightly higher concentration of anionic than the standard two-phase titration.

Other Components
Chlorides, sulfates, and moisture are also commonly determined in the quality control of sulfates and sulfonates. Chlorides and
sulfates are usually determined by potentiometric titration as described by Schmitt [1]. Chlorides and sulfate may also be
determined by manual titration if a potentiometric titrator is not available. Moisture is usually determined either by Karl Fisher
titration (either manually or potentiometrically) or oven drying at 105°C for two hours or to constant weight. Schmitt [1]
describes the Karl Fisher technique using hydranal (pyridine free titration reagent) as the titrant.

Characterization

The HPLC technique is being used to better characterize commercially produced sulfates and sulfonates. Several research
groups have developed methods for determining the alkyl chain length distribution in alkyl sulfates and sulfonates and for
separating the sulfate peaks from the sulfonate peaks. The methods discussed here show several ways to detect the non-UV-
absorbing alkyl sulfates and sulfonates. One method [13] uses iron 1,10-phenanthroline salts as a mobile-phase
additive for the separation on a reversed-phase PRP-1 column with a photometric detector at 510 nm. Another method [14] for
separating alkyl sulfates and sulfonates based on alkyl chain length uses a reversed-phase PRP-1 column and an
acetonitrile/water mobile phase modified with LiOH. A conductivity detector is used to detect the analytes.
Larson [15] has developed an interesting technique for the characterization of alkyl sulfates by alkyl chain length. Using
indirect photometric chromatography he adds a UV-absorbing counter ion to the mobile phase and employs a UV-detector. The
peaks of interest appear as negative peaks in an elevated baseline as they are displaced from the column.
Page 86

Kunitani and Kresin [16] use an ion chromatographic system with a conductivity detector to achieve separation of alkyl chain
lengths ranging from C-8 to C-20. The separation is performed on a hydrophobic resin-based column utilizing tetrabutyl
ammonium hydroxide as an ion-pair reagent and acetonitrile as an organic modifier.

Another method [17] that depends on a conductivity detector uses a porous resin column with ammonium and tetra-alkyl
ammonium salts as the ion-pairing reagents to separate the alkyl sulfates and sulfonates by chain length and from each other.

A method [18] has also been developed that can be used to determine whether the alkyl group in an alkyl benzene sulfonate is
branched or linear. The method uses a reversed-phase column and a water/methanol mobile phase that is buffered with sodium
acetate.
Castles et al. [19] describe a method for determining the concentration of linear alkylbenzesulfonates using a reversed-phase
column and a THF/water mobile phase modified with sodium perchlorate. In this system the benzene sulfonate
chromatographs as a single peak. Although they developed this method to determine the concentration of
alkylbenzenesulfonates in environmental samples it could be modified to determine the benzene sulfonate in other mixtures
and perhaps used in place of active titrations.

One of the newest techniques for the characterization of sulfates and sulfonates is capillary electrophoresis. Chen and Pietrzyk
[20] have developed a separation scheme for sulfate and sulfonate surfactants using capillary electrophoresis. They separate
sulfates from sulfonates and then separate each on the basis of alkyl chain length. They use a UV detector to detect both
aromatic and nonaromatic surfactants. The aromatic surfactants are detected by direct UV at 220 nm and the nonaromatic are
detected by indirect UV at 230 nm, using salicylate anion as a chromophoric buffer additive. Desbene et al. [21] have
developed a system for the characterization of alkyl aromatic sulfonates using high-performance capillary electrophoresis.
Using acetonitrile as the organic modifier, they achieve baseline separation for alkyl chain lengths between C-2 and C-12.

B
Phosphate Esters

Phosphate esters are also anionic surfactants but require somewhat different quality control tests from those used for S-
containing esters.

Ester Ratio. For the phosphate esters, the ratio of monoester to diester is important as well as the concentration of residual
phosphoric acid. Schmitt [1] describes how to do this titration and how to carry out the calculations. This titration is only
useful if the material was not neutralized as a part of the manufacturing process and if there is no residual pyrophosphate in the
surfactant solution. An HPLC method has been developed [22] which separates the monoester and diester on a reversed-phase
column. The esters are first derivatized to produce UV-absorbing species and are then chromatographed using a UV detector
for detection. This method requires extracting the surfactant into diethyl ether before derivatization.
C
General Tests

Starting Alcohol

The concentration of unreacted starting alcohol must be determined in phosphates, sulfates, and sulfonates. These nonionics
(often called free fats) have traditionally been quantitated gravimetrically. The ASTM procedures [11] D1570 (alkyl sulfates),
D1568 (alkyl aryl sulfonates), and D3673 (2-olefinsulfonates) describe
Page 87

the extraction procedures. The extraction and quantitation of the unreacted alcohol in phosphate esters is described by Schmitt
[1]. When the starting alcohol is ethoxylated, the extraction procedure may not extract all of the nonionic since this alcohol is
more water soluble. It may thus be more appropriate to use an ion-exchange technique in which all of the ionic materials are
removed by an ion-exchange column while the nonionics that are eluted from the column are determined gravimetrically.
Longman describes such a procedure [3]. It is possible to develop modifications of this procedure using mixed-bed resins and
simply stirring an aliquot of the resin in a solution of the surfactant.

If a new procedure using ion exchange techniques is being developed for the determination of nonionics it is necessary to be
sure there is no anionic in the residue. This may be confirmed with the aid of an IR spectrum of the residue or by doing a
standard anionic titration of the residue. All of these techniques calculate everything that is nonionic or which extracts into the
nonpolar solvent as unreacted alcohol.

The HPLC technique is used to obtain more reliable assays of the unreacted alcohol and to reduce the time required for the
determination. When HPLC is used for the quantitation, the alcohol standard must be the exact alcohol (same carbon and
ethoxylate distribution) as that used in the synthesis of the anionic.

Yoshimura et al. [23] have developed a procedure using a back-flush technique for chromatographing an alcohol ethoxylate in
which the alcohol chromatographs as a single peak regardless of carbon chain or ethoxylate distribution. It would seem
reasonable to use this technique to determine unreacted starting alcohol.

It should be noted that the concentration of alcohol determined by HPLC will probably not equal the concentration determined
gravimetrically since the gravimetric analysis will include materials that are not starting alcohol.

Characterization
The determination of alkyl chain-length distribution is an important part of the characterization of sulfates, sulfonates, and
phosphates. Usually this is done by the hydrolysis of the product and gas or liquid chromatography of the hydrolyzed product.
It is possible to simply chromatograph the extracted unreacted alcohol but this may not give a true distribution since all the
chain lengths may not have reacted at the same rate. Schmitt [1] provides the conditions for the hydrolysis and desulfonation of
sulfonates and the chromatographic conditions for these products.

D
Isethionate Esters

Another group of anionic surfactants are the isethionate esters. In general, testing is very similar to that of other sulfonates. It is
important to determine the concentration of sodium isethionate, which is usually performed by HPLC using an anion column
and a conductivity detector [24].

III
Nonionics

Nonionic surfactants have no charge on the molecule but owe their hydrophilic character to a polar group in the molecule,
while their hydrophobic properties result from the presence of a long chain hydrocarbon. Only a few major groups of
nonionics are discussed below. Ethoxylated fatty alcohols and phenols form one of the major classes of nonionic surfactants.
Two other classes of nonionics are the ethoxylated fatty acids and the alkanol amides.
Page 88

A
Ethoxylated Alcohols and Phenols

Determination of Average EO Chain Length

Traditionally the tests for assessing the level of ethoxylation in nonionics determined only an average value for the number of
moles of added ethylene oxide. The easiest test for this determination is the cloud point. Longman [3] describes this technique
and tells how to run the test. Cloud point is a valuable test for quickly determining the approximate number of ethylene oxide
units in the molecule. Another somewhat more accurate method for determining ethylene oxide content of the molecule is
hydroxyl number. The molecular weight of the starting alcohol must be known in order to determine the degree of ethoxylation
by hydroxyl number. Cross [25] describes the procedures that are used to determine hydroxyl numbers and thus the average
number of moles of ethylene oxide that have been added.

Spectroscopy has also been used to determine the average EO chain length in ethoxylated alcohols and phenols. Meszlenyi et
al. [26] described an IR technique for determining the average ethylene oxide chain length in nonyl phenol ethoxylates by
calculation using the ration of the absorbance at 1350 and 1610 cm-1. Nuclear Magnetic Resonance has been used to
characterize the hydrophobic group and to determine the average number of moles in the ethoxylate chain. Cross and Mackay
[27] have described a method that allows determination of the average chain length of the alkyl group as well as the degree of
polymerization of the ethylene oxide chain by forming the trimethylsilyl derivative of the ethoxylate before carrying out the
proton NMR analysis.

These techniques are useful for providing information about the average molecular weight of the surfactant, but some recently
introduced techniques provide information about both the ethoxylate distribution and the alkyl chain length distribution.
Chain length distribution. The distribution of the carbon chain lengths in the alkyl ethoxylates has usually been established by
gas chromatography. If the alcohol has been ethoxylated, some cleavage is required before chromatography to determine the
carbon chain-length distribution. Cross [25] gives a summary of these techniques and the conditions for carrying them out.

Recently it has become more important to know both the carbon chain-length distribution and the ethoxylate distribution for
the quality control of ethoxylated surfactants and for matching formulated products. Gas, liquid, and supercritical fluid
chromatography have all been used to better define these distributions.

High performance liquid chromatography has been used to identify and quantitate the components of alcohol ethoxylates. In
general normal-phase chromatography has been used to determine the ethylene oxide distribution and reversed-phase
chromatography has been used to determine the hydrophobe distribution. Kudoh [28] has developed a method that can separate
alcohol ethoxylates according to their alkyl chain lengths. Using a C-18 column with an acetone/water mobile phase he
separates the ethoxylates by alkyl chain length and not by ethylene oxide distribution. He uses refractive index to detect the
peaks. A method for characterizing fatty alcohol (C16C18) ethoxylates (up to 30 moles of EO) by ethoxylate chain length has
been developed by Desmazieres et al. [29] They use a cation exchange column with an acetonitrile/water (98:2) mobile phase
to which they add sodium acetate. They find that elevated temperatures improve the separation. Since there is no mobile phase
gradient, it is possible to use a refractive index detector. Okada [30] describes an HPLC method for separating both the
hydrophobic and ethoxylated oligomers. He uses a cation exchange resin in the alkali metal form to
Page 89

separate the ethylene oxide units and a reversed-phase system for the hydrophobic separations. He emphasizes the need for
temperature control and sometimes the use of temperatures above ambient. Trathnigg and his coworkers [31] have developed a
technique using two-dimensional chromatography to determine both the carbon chain-length distribution of the alkyl group
and the ethoxylate distribution in fatty alcohol ethoxylates. Liquid Chromatography under critical conditions is used as the first
dimension and size-exclusion chromatography as the second.

The ethoxylate distribution is also important in the phenol ethoxylates. Wang and Fingas [32] have developed a method for
separating the oligomers of octyl phenol ethoxylates by ethylene oxide adduct up to 40 moles of ethylene oxide using a
reversed-phase system (Supelcosil LC-1 column and a water/methanol mobile phase). A UV detector is used, and quantitation
is carried out using molar response factors.

Supercritical Fluid Chromatography is being used to determine the ethoxylate distribution in alcohol ethoxylates. Geissler [33]
has developed a method for both carrying out the separation of the ethoxylate (by EO unit) and calculating the mole percent of
each oligomer. His method separates branched- from straight-chain alcohol ethoxylates but there is some overlap when the
straight-chain alcohol is based on a mixture of carbon chain lengths. He has developed a calculation scheme that allows the
determination of the individual components even without complete separation. Pinkston et al. [34] use SFC with an FID to
separate both the ethoxylate and carbon chain oligomers. When the chromatogram becomes too complicated to permit simple
interpretation, they use mass spectrometry to identify the peaks. Another group (Kalinoski and Hargiss [35]) also uses capillary
SFC to characterize alcohol ethoxylates and mass spectrometry to identify the peaks. The paper describes the specialized
conditions necessary to obtain usable mass spectra.

Several researchers have compared HPLC and SFC in the analysis of alcohol ethoxylates. The evaporative light scattering
detector (ELSD) has increased the possibilities for developing methods to determine the ethoxylate distribution of an alcohol
ethoxylate using both HPLC and SFC. Since the ELSD allows the detection of non-UV absorbing species and since its
performance is not affected by changes in the mobile phase, it allows much more versatility in the method development.
Lafosse et al. [36] developed schemes carrying out the separation of ethylene oxide oligomers by both HPLC and SFC using
the ELSD. Using the ELSD as the detector, Brossard et al. [37] compare HPLC and SFC in the analysis of ethoxylated fatty
alcohols. They show that they can carry out separations for higher molecular weight ethoxylates using SFC rather than HPLC.

Methods have also been developed for separating the ethylene oxide oligomers of the phenol ethoxylates using SFC and
HPLC. Wang and Fingas [38] have developed a procedure using SFC for separating the ethylene oxide oligomers of
octylphenol from 0 to 25 ethylene oxide units. They identify the individual adducts by comparison to the peak for octylphenol
and octylphenol plus one EO. The area percent method was used to calculate the oligomer distribution. The paper outlines the
detailed conditions used for the analysis and shows the effect of varying the SFC conditions. The same group has also
developed a simple, rapid, and reproducible capillary SFC method for the separation and identification of the ethoxylate
distribution in nonylphenol ethoxylates [39]. They compared the separation achieved by SFC to that achieved by HPLC and
find the SFC separation is better and the analysis time is shorter.
Page 90

Desbene and Desmazieres [40] developed an HPLC procedure that provides a satisfactory separation of the ethoxylate
oligomers for both alkyl alcohol ethoxylates and phenol ethoxylates. They have achieved a baseline resolution of 1 to 80 EO
units using a p-nitrophenyl-bonded silica with an n-propyl spacer and an n-heptane-dichloromethane-methanol mobile phase at
45°C. They used either a UV detector (derivatizing where necessary) or a lightscattering detector. They found the separation
was satisfactory when they derivatized before chromatography.

Some work has also been done to compare SFC to high-temperature gas chromatography (GC) in the analysis of alcohol
ethoxylates. Silver and Kalinoski [41] show the advantages and limitations of both techniques for the quantitative
characterization of alcohol ethoxylates. Sandra and David [42] have also compared the two techniques for the analysis of
nonionics. They conclude that the two techniques are complementary and the main advantage of SFC is the lower temperature
required for analysis, which permits the analysis of thermally labile compounds. Chromatographic conditions are shown for
both SFC and high temperature GC.

Carbon NMR has been used to characterize alcohol ethoxylates. Kalinoski and Jensen [43] developed a method for
determining the distribution of alcohol ethoxylates and compared the information so obtained to that obtained from SFC.

Impurities

When alcohols and phenols are ethoxylated there is always some polyethylene glycol (PEG) formed. The concentration of
PEG in the ethoxylate is critical to some applications; so testing for PEG is part of the quality control of the product. The most
widely used method for determining PEG is an extraction procedure in which the PEG is extracted from the surfactant and its
concentration determined gravimetrically. In an HPLC technique a reversed-phase column and a methanol/water mobile phase
are used. This HPLC method is not totally reliable since the PEG is eluted from the column almost in a void volume and thus
is subject to interferences. Both these procedures are described in detail by Schmitt [1].

Several researchers have used the evaporative light scattering detector as the detector in their work to separate the PEG from
the ethoxylate and to determine the distribution of the PEG. Brossard et al. [37] showed that they could elute higher molecular
weight PEG better by SFC than by HPLC. They were also able to resolve the distribution of the PEG, but the PEG and
ethoxylated alcohol distribution overlapped. Lafosse et al. [36] have also developed some schemes for separating PEG from
the surfactant using HPLC and SFC. Neither of these methods is being used routinely for PEG determinations at this time.

B
Ethoxylated Fatty Acids

Another group of nonionic surfactants is the ethoxylated fatty acids that are really esters of fatty acids and polyethylene glycol.
In addition to the monoester these products may also contain some diester, some free acid, and some PEG. These esters can be
analyzed by gas chromatography for low-molecular-weight ethoxylates. The ratio of monoester to diester can be calculated
from hydroxyl number and ester-number determinations performed on the product after the removal of the PEG [1].

Ethoxylated fatty acids may also be analyzed by HPLC. Aserin et al. [44] developed a method for characterizing the monoester
by separating the EO adducts up to 20 EO
Page 91

units. They used normal-phase chromatography with gradient elution (isopropanol/methanol/hexane mobile phase) and a UV
detector at 220 nm.

Kudoh et al. [45] developed an HPLC technique to separate and quantitate the polyethylene glycols, the monoester and the
diester. They used a preparative scheme to separate and then quantitate gravimetrically; their system employed a reversed-
phase column with acetone/water as the mobile phase.

C
Alkanolamides

Alkanolamides and ethoxylated alkanolamides comprise another group of nonionics that are widely used in cosmetic products.
Several researchers have developed HPLC techniques to analyze for alkanolamides. Cross [25] described several methods for
determining the concentration of alkanolamides. Ben-Bassat and Wasserman [46] developed a reversed-phase HPLC procedure
that can be used to quantitate alkanolamides. They used a reversed-phase column, a ternary mobile phase
(tetrahydrofuran/acetonitrile/water) and a refractive index detector. This method also separated the fatty acids and could be
used to quantitate unreacted fatty acid in the amide.

It is often necessary to know the concentration of ester in alkanol amides. This is most often determined by the Cosmetics,
Toiletry and Fragrance Association (CTFA) method, which uses IR spectroscopy [6]. Residual amine and fatty acid may be
determined by titration or chromatography as described by Schmitt [1].

IV
Cationics

Cationic surfactants are those that owe their water solubility to a positively charged nitrogen and any fat-like properties to a
hydrophobic long-chain group that usually consists of a mixture of homologs. The major classes of these cationics are
quaternary ammonium compounds and amine oxides. Amine oxides are sometimes classified as nonionics or amphoterics
since their character changes with pH.

A
Amine Oxides

Amine oxides are widely used in cosmetic applications. Since they are made by the oxidation of tertiary amines and since the
tertiary amine is usually insoluble in water it is important to determine the concentration of the tertiary amine and that of the
amine oxide. Wang and Metcalfe [47] developed a nonaqueous titration technique that permits the determination of both the
amine and the amine oxide. The potentiometric titration shows two breaks that allow the calculation of both. Schmitt [1] also
outlines several titration techniques for determining both the tertiary amine and the amine oxide in the same solution. Recently,
some methods have been developed for determining the tertiary amine using HPLC. Both Schmitt [1] and Metcalfe [48]
outline systems that may be used.

B
Quaternary Ammonium Compounds

Quaternary ammonium compounds (quats) made from long-chain fatty acids are the most widely used cationic surfactants. The
most common method for the determination of the concentration of the quaternary compounds is a two-phase titration in which
the cationic surfactant is titrated with an anionic surfactant. Both Metcalfe [48] and Schmitt [1]
Page 92

describe the procedure and give the exact conditions. Metcalfe [48] also describes several direct single-phase titration methods.
The surfactant electrode developed by Orion is currently being used to determine the endpoint in the titration of cationics with
anionics. As noted above Oei et al. [12] describe the use of this electrode as well as one that they developed in house. They
describe how the titration is performed and compare the results obtained by the electrode titration with those obtained from the
standard two-phase titration. They show that the electrode method has many advantages over the standard titration: it gives
better precision, it is less time consuming, there is no organic waste, and the presence of nonionic substances does not
interfere. It does not always give the same concentration as that found from the standard two-phase titration so it can not be
substituted for that titration in the quality control lab without prior validation and perhaps adjustment of the specification
range.

Liquid chromatographic methods are being developed for the analysis of the quaternary ammonium compounds as well as
amines that may be present. Metcalfe [48] references procedures that will separate the quats by class or by chain length. He
also describes a method that is being used routinely to separate the desired quat from the impurities in the reaction mix.
Schmitt [1] gives the conditions and references for a number of HPLC systems that have been developed for chromatographing
cationic surfactants. Wilkes et al. [49] have developed a normal phase HPLC method for the separation of quaternary
ammonium surfactants that uses a silica gel column with gradient elution and an evaporative light scattering detector. They use
trifluoroacetic acid as an ion-pairing reagent in the nonpolar mobile phase to improve peak shape and to avoid column
contamination with amines that are often present in the quaternary ammonium salts. The method is useful for the
characterization of the surfactant as well as the quantitation of the surfactant in a formulated product. The procedure separates
compounds on the basis of the number of alkyl groups on the nitrogen. When there is only one alkyl group it also separates on
the basis of chain length; when there is more than one alkyl group only some separation by chain length occurs and the system
needs to be refined to separate these components completely.

It may be necessary to both quantitate the amine in the surfactant and to determine the alkyl chain length of the amine and
some ion chromatographic procedures have been developed to carry out these determinations. Vialle et al. [50] have designed a
method using alkali metal ions as eluting ions and an organic solvent in the aqueous mobile phase to improve the efficiency for
long chain amines. They find the system is very efficient for separating amino compounds but the sensitivity of the detection
of the amines may not be great enough for quantitating low levels of amines. Other detection modes will be necessary to
improve the detection for low levels. Krol et al. [51] have also used ion chromatography and conductivity detection for the
analysis of alkyl and alkanol amines. Their method allows the detection of low ppm levels of the amines by using an indirect
conductivity detection system.

V
Amphoterics

Amphoteric surfactants contain at least two oppositely charged ionizable sites. These are usually provided by a ternary amine
group (cationic) and a carboxylate or sulfonate group (anionic).
For analysis purposes the amphoterics may be divided into three groups [1]. First is
Page 93

the amino acid type, which is usually made by the addition reaction of a fatty amine with an unsaturated methyl ester followed
by hydrolysis of the ester. The general structure is

where R is a long chain alkyl group.

Another group of amphoterics are the betaines, which are usually prepared by quaternization of a tertiary amine with
chloroacetic acid. The surfactant betaines are compounds in which one of the methyl groups in betaine (trimethyl glycine) has
been replaced by a fatty chain. The general structure is

where R is long chain alkyl group.

The third group of amphoterics is imidazoline derived, and they contain an amido group in their structure. These products are
mixtures of the general structure

and

A
Active

The methods used for the routine analysis of amphoterics are not well defined at this time, but new techniques are being
developed for improved assays of the product and its impurities. The required analyses performed on any amphoteric depend
on the process used for the production. There are currently no direct methods used routinely to determine the active ingredient
in the amphoteric surfactant. In the customary indirect assay, water and any known impurities are subtracted from 100%.
Schmitt [1] lists some HPLC methods that have been developed for amphoterics, but they are rarely being used for quality
control of the product. He discusses a method developed by Koenig and Strobel [52] that seems to have some potential as the
basis of a quantitative test. Oei et al. [12] describe the use of the Orion surfactant electrode to determine the concentration of
betaines. Capillary electrophoresis is another technique that would seem perfectly suited to the analysis of amphoteric surfacts
that are very similar to amino acids in structure. Chadwick et al. [53] developed a procedure that uses capillary zone
electrophoresis to separate and quantitate sodium cocoamphocarboxyacetate in less than 12 minutes.

B
Water
The determination of water is a critical part of the analysis of amphoterics. It is most commonly done by an oven technique
(convection, microwave, moisture balance). Since amphoterics decompose easily, it is critically important to remove the water
using the gentlest conditions possible. Karl Fisher analysis is difficult since the concentration of water is so high.
Page 94

C
Salt

The concentration of sodium chloride must be determined in all types of amphoterics. This is usually done by direct
potentiometric titration as described by Longman [3].

D
Amines

All amphoterics contain some level of amines or amido amines. One method of determining the amine is extraction from a
basic solution into a nonpolar solvent and titration to quantitate. The concentration of a specific amine may also be determined
by chromatographic techniques. Amines may be determined by gas chromatography using a column specifically designed for
amine analysis and preferably a nitrogen-specific detector. Some ion chromatographic procedures have recently been
developed for amine quantitation. Although they may need to be modified for the particular amine of interest, the methods
should provide a general starting place in the analysis to determine the concentration of amine in amphoterics. Vialle et al. [50]
use alkali metal ions as eluting ions and add an organic solvent to the aqueous mobile phase to improve the efficiency for long-
chain amines. Krol et al. [51] also use ion chromatography for the quantitation of amines but they use indirect conductivity
detection to provide detection limits in the ppm range.
E
Acids

Since monochloroacetic acid is used in the synthesis of amphoterics and it is undesirable in the finished product, the surfactant
is often analyzed to determine the concentration of monochloroacetic acid and its hydrolysis product, glycolic acid. This is
usually done by some form of liquid chromatography. Very recently Nair et al. [54] developed two ion chromatographic
techniques that may be used to determine the concentration of monochloroacetic acid. One method is based on an anion-
exchange separation with suppressed electrical conductivity detection, and the second is based on anion-exclusion separation
with UV detection. Another method uses an organic ion column and a conductivity detector to determine the concentration of
both chloroacetic and glycolic acids [55].

Acrylic acid or methyl acrylate are also used in the synthesis of some amphoterics, so it may be necessary to determine the
concentration of acrylic acid in the surfactant. This is easily done by HPLC using a reversed-phase column and an ion-pair
reagent in the water/tetrahydrofuran mobile phase [56].

VI
Preservatives

Preservatives are sometimes added to surfactants to ensure they will not support microbial growth. The cosmetic chemist may
need to analyze for the presence of preservatives either in the surfactant raw material or in the formulated product. In some
cases it may be necessary both to identify and quantitate the preservative that is present.

A
General

Several schemes have been published for the separation, identification, and quantitation of preservatives. Richard et al. [57]
used thin layer chromatography to determine what preservative may be present. De Kruijf and Schouten [58] developed a very
extensive program for the identification and quantitation of many of the more commonly used preservatives. Although their
methods are designed for determining preservatives in
Page 95

cosmetics they may easily be adapted to identify and quantitate the preservatives in surfactants. They use TLC and HPLC for
screening and HPLC for quantitation. Both Liem [50] and Wilson [60] have developed TLC schemes for identifying the
preservative that is present in a surfactant. De Kruijf et al. [61] also developed a method for identifying preservatives by the
use of HPLC. The chromatographic systems described in this publication could be used to quantitate as well as identify the
preservatives in the surfactant raw material.

B
Formaldehyde and Formaldehyde Releasers

Formaldehyde or formaldehyde-releasing preservatives are widely used in the surfactant industry. In many cases the analysis to
determine the concentration of formaldehyde in surfactants is quite easily carried out using standard wet chemical methods.
The best test for the determination is the Hantzsch reaction method developed by Nash, which is based on the reaction of
formaldehyde with acetyl acetone and an ammonium salt to give diacetyldihydrotoluidine, which absorbs at 412 nm [62]. This
same method can be used to determine the concentration of a formaldehyde-releasing preservative if the solution is first treated
to release the formaldehyde. Benassi et al. [63] developed a method for determining formaldehyde concentration in which the
formaldehyde is derivatized with 2,4-dinitro phenylhydrazine and HPLC is used to separate and quantitate the hydrazone. The
quantitation is carried out by the method of standard additions. Again the method could be used for determining the
concentration of formaldehyde donors if the solution is pretreated.

Summers [64] has developed a method to separate and quantitate imidazolidinyl urea (Germall 115), DMDM Hydantoin
(Glydant), trans-1-(3-chloroallyl)-3,5,7-triaza-1-azoniadamantane chloride (Quaternium 15, Dowicil 75) and formaldehyde
using HPLC with post-column derivatization. This system allows the quantitation of formaldehyde in the presence of these
formaldehyde-releasing preservatives since it separates formaldehyde from the others. The derivatization uses Nash's reagent
and the derivative produced absorbs at 410 nm and fluoresces at 510 nm. Semenzato et al. [65] developed an HPLC method for
determining the concentration of Quaternium 15 without decomposing it to formaldehyde. They used a Li-CN column, a UV
detector at 200 nm and a mobile phase of acetonitrile and phosphate buffer. Another method for determining the concentration
of formaldehyde in the presence of Quaternium 15 was developed by Benassi et al. [66]. In this method the Quaternium 15
(Dowicil 200) is retained on a cationic stationary phase while free formaldehyde is eluted and determined by HPLC after
derivatization.
C
Parabens

The esters of p-hydroxybenzoic acid (parabens) are also used as preservatives in surfactants. The methyl, ethyl, propyl and
butyl esters are most commonly used. Masse et al. [67] outlined a procedure for identifying the presence of parabens by TLC
and quantitating by HPLC. De Kruijf and Schouten [58] have also developed a method for quantitating the parabens by HPLC
using a reversed-phase system.

D
Triclosan

Irgasan DP-300 (Triclosan) is sometimes used as a preservative in surfactants. Methods have been published for determining
Irgasan concentration using both gas and liquid
Page 96

chromatography. Hoar and Sissons [68] developed a GC method that uses a mixed SE-30-QF1 stationary phase and either an
FID or electron capture detector (ECD). The ECD is more sensitive, but this sensitivity is usually not necessary for the
concentrations that would be expected. Marquardt et al. [69] also described a method for quantitating DP-300 in various
substrates using gas chromatography. They extracted the DP-300 from the substrate, formed the acetyl derivative and analyzed
by gas chromatography using an OV-17 column and an ECD. Several researchers have developed methods for quantitating
Triclosan using HPLC. Achari and Chin [70] reported on an HPLC technique. The separation and quantitation were carried out
using a micro Bondapac Alkylphenyl column with a 1:1 (v/v) acetonitrile/water mobile phase and a UV-vis detector. George et
al. [71] developed a quick, accurate, and reproducible method for determining the concentration of Triclosan in soaps. They
used a reversed-phase column with a water/THF mobile phase and a UV detector at 280 nm.

E
Miscellaneous

Another preservative used in surfactants is PCMX (p-chloro-m-xylenol). The USP [72] describes a standard gas
chromatographic method for the assay of PCMX that could be adapted for the quantitation of PCMX in a surfactant.

Kathon CG (5-chloro-2-methyl-4-isothiazoline-3-one and 2-methyl-4-isothiazdin-3-one) is being used to some extent to


preserve surfactants where the chemistry of the surfactant will permit. Since Kathon is used at low levels, its analysis in a
surfactant may sometimes be very difficult. Rohm and Haas [73] developed a reversed-phase HPLC method for determining
the concentration of Kathon in solution. They recommend a Spherisorb ODS column and an isocratic mobile phase of 50:50
methanol and water. Matissek et al. [74] developed a method in which the Kathon CG is first derivatized and then quantitated
by ion-pair HPLC.

VII
Impurities

There are some impurities in surfactants that may have harmful effects if they are present in cosmetics. Reliable methods to
quantitate these impurities in the raw material are critical to good quality control.

Ethoxylated materials may contain trace levels of ethylene oxide (EO), acetaldehyde, and dioxane.

A
Ethylene Oxide and Acetaldehyde
Most of the methods for determining residual ethylene oxide use headspace sampling and gas chromatography to determine the
EO concentration. The EO standards are somewhat difficult to prepare and various techniques have been used to prepare them.
Acetaldehyde will almost certainly also be present and may chromatograph very close to the ethylene oxide so acetaldehyde
standards should be chromatographed to determine retention time and verify that it is not cochromatographing with EO. When
using a headspace sampling technique it is very important that the standards and samples are in the same size container and the
headspace above the liquid is the same for both. It is also important that the standard be in the same matrix as the analytes.

Dahlgran and Shingleton [75] reported a headspace sampler/GC technique using a Chromasorb 102 column and a flame
ionization detector. They prepared the primary
Page 97

ethylene oxide standard by adding ethylene oxide gas to hexane but the working standards were prepared by diluting this
primary standard in a surfactant that had been vacuum stripped to remove all traces of EO. They reported a quantitation limit
of 1.0 ppm and a detection limit of 0.5 ppm. Leskovsek et al. [76] also used a headspace technique to quantitate residual EO.
They prepared their standards and samples in gas-tight system vials, which they heated for a specified time in a temperature-
controlled sonic bath. Aliquots of the headspace were injected into the gas chromatograph with gas-tight syringes. They
chromatographed the EO using a column packed with 0.8% THEED on Carbopak C and used the method of standard additions
for their quantitation. Cardeal et al. [77] described a novel way of preparing the EO standards by synthesizing the EO in situ.
They compared a headspace technique to a direct injection method for quantitating residual EO.
Reaction products of ethylene oxide (ethylene chlorohydrin and ethylene glycol) may also be present in the surfactant and may
need to be quantitated. Sasaki et al. [78] have developed a procedure for detecting and quantitating very low levels of ethylene
oxide (EO) and ethylene chlorohydrin (ECH) in surfactants (detection limits 0.0050.03 µg/g of EO and 0.010.07 µg/g for
ECH). They desorb them from the solution and convert the EO and ECH to ethylene iodohydrin, which is then quantitated by
gas chromatography using an electron capture detector. Danielson et al. [79] describe a procedure for quantitating ethylene
oxide, ethylene chlorohydrin, and ethylene glycol by a single gas chromatographic analysis using a DB-Wax Column.

B
Dioxane

1,4-Dioxane must also be quantitated in ethoxylated products, both nonionic and sulfated. Usually the concentration of dioxane
is higher in the sulfated products than in the nonionics. Gas chromatography is most commonly used in the analysis of dioxane
with various sample preparations by either headspace or direct injection. The USP [72] procedure requires that the dioxane be
removed from the surfactant by vacuum distillation and that the distillate be gas chromatographed using a column packed with
a cross-linked copolymer of acrylonitrile and divinyl benzene. The quantitation is carried out using a standard prepared in
water, and the quantitation limit is 10 ppm.
Scalia et al. [80] use a solid-phase extraction procedure before direct injection onto the Poroplot Q capillary column. They use
a selected ion-monitoring mass spectrometry detector and achieve a quantitation limit of 3 mg/kg. Although this method was
developed for determining dioxane in cosmetics it could be easily adapted to surfactants. Italia and Nunes [81] have
determined the concentration of 1,4-dioxane in shampoos by direct injection onto a column packed with OV-1 as the stationary
phase and the use of a flame ionization detector. They use an internal standard for the quantitation.

A number of methods have been reported in the literature for determining dioxane concentrations by gas chromatography using
headspace sampling. Goetz et al. [82] use a column packed with SP-1000 on Carbopak C and a flame ionization detector.
Using the method of standard addition for the quantitation they could quantitate 1 µg/g and detect 0.5 µg/g. They showed that
dioxane was not being generated during the warming process before injection. Beernaert et al. [83] use a capillary column
coated with CP sil 8 CB film and a flame ionization detector. They prepare their working standards in a product that is similar
to the one being analyzed but which contains no ethoxylated material. They report a detection limit of 2 mg/kg. Rastogi [84]
uses a Supelcowax 10
Page 98

capillary column and mass spectrometry for the identification and quantitation of the dioxane. He found a detection limit of 0.3
ppm. He prepared the standards in a material similar to those that were to be analyzed but which did not contain any
ethoxylated products.

High performance liquid chromatography has also been used to quantitate dioxane. Scalia [85] reported a procedure for
determining dioxane in sulfated, ethoxylated alcohol surfactants in which the impurities are removed by solid phase extraction
and are then quantitated by HPLC using a reversed-phase column and an acetonitrile/water mobile phase with a UV detector at
200 nm. The minimum quantifiable amount was 18 µg/g.

C
Nitrosamines

Nitrosamines may be found in any surfactant that is synthesized from secondary amines (or a tertiary that contains some
secondary). Since nitrosamines are considered hazardous, it may be necessary to quantitate any nitrosamine in the surfactant.
Quantitation may be carried out by first separating and purifying the amine and then quantitating by a colorimetric technique.
This is a time-consuming technique that must be carried out by an experienced analyst to avoid loss and contamination. It
does, however, determine all nitrosamines that are present.

Probably the most common way to analyze for nitrosamines is to separate them by gas or liquid chromatography and detect
and quantitate the nitrosamines by use of the TEA analyzer. This analyzer was developed and produced by Thermedics, Inc.
(Waburn, MA) specifically for the analysis of nitrosamines and it is very sensitive and selective for nitrosamines.

Erickson et al. [87] described a method for analyzing for N-nitroso-diethanolamine (NDELA) in cosmetic ingredients. After a
clean-up step the solutions were chromatographed by HPLC, and NDELA was detected and quantitated by use of a TEA
analyzer. A similar procedure could be used for the determination of nitrosamines in surfactants, but less sample preparation
would be necessary.

Meili et al. [88] have developed a method of analyzing for nitrosamines using gas chromatography and a photoionization
detector. They use a capillary column coated with octoxynol-3 for the separation. Gorski and Cox [89] have reported an
amperometric method for determining the concentration of nitrosamines using an electrode coated with a ruthenium-based
inorganic polymer.

D
Sultones
Sultones are also harmful compounds that may be present in certain anionic surfactants. Porter [4] describes several techniques
that may be used for the quantitation of sultones. The most commonly used technique is to extract the neutral oil from the
anionic and quantitate the sultones in the neutral oil by gas chromatography.

VIII
Analysis of Formulated Products

A
General Separation Schemes
It is often necessary for the cosmetic chemist to determine the composition of formulated products. There are many schemes
for the complete breakdown and analysis of these
Page 99

products but these schemes should not be followed routinely without thought. Decisions must be made at each step of the
procedure to decide on the best next step. It is also useful to work with a chemist who knows at least generally what
ingredients may be in the formulation in order to reduce the time required for the analysis and to make a more accurate
determination. Schmitt [1] outlines a good general approach for identifying the components of an unknown mixture and
provides a general list of the most common ingredients in a cosmetic product.

There are several authors who have developed separation and analysis schemes for these products. Both Longman [3] and
Hummel [9] outline general schemes for the breakdown and analysis of formulated products using standard wet methods.

Milwidsky and Gabriel [2] outline methods necessary to determine the composition of unknown surfactant mixtures. They
include an ion-exchange scheme for the separation of the surfactant groups and depend on standard wet techniques for the
analysis of the separated surfactants. Porter [4] also shows an ion-exchange system for separating surfactants from each other.
He presents the use of many techniques in the analysis of the products including many chromatographic and spectroscopic
techniques.
B
Preliminary Screening

It is generallly useful to conduct some preliminary screening of a formulated product before deciding how to proceed with the
analysis. Probably the most widely used tool for this purpose is infrared spectroscopy. A number of volumes of spectra of
surface-active agents have been published. In general these volumes not only show the spectra but indicate the characteristic
frequencies for each group of surfactants. Stadler's book of IR spectra [10] groups the spectra by surfactant class and notes the
characteristic group frequencies. Hummel [9] has also published a volume of spectra that is part of the two-volume set
Identification and Analysis of Surface-Active Agents. There is an ASTM monograph (D2357-74) [90] which includes a table of
the infrared absorption bands of commercial detergents. Nettles [91] has published a summary of the IR spectra of some
commonly used surfactants with a discussion of the distinguishing characteristics of each spectra.

Several schemes have also been reported for using thin layer chromatography to screen a product for the tentative
identification of the surfactants present. Desmond and Borden [92] developed a thin-layer chromatographic system that uses
both Rf value and a special color development technique to identify the surfactants present by type. The system separates
anionics from nonionics, and the color development further identifies the specific type of anionic or nonionic. They also
suggest a column chromatographic scheme that will separate anionics from nonionics and will allow enough sample to be
collected for IR identification of the surfactant. Henrich [93] has chromatographed over 150 surfactants in six different thin-
layer chromatographic systems, and an unknown surfactant may be tentatively identified by comparison of the Rf of the
unknown to the Rf of the knowns in the various systems. Spray reagents are used for visualization and characterization.
Armstrong and Stine [94] describe a technique that uses a reversed-phase thin-layer system to separate the surfactants by class
and a silica-gel system to separate individual anionic or cationic surfactants from others that are similarly charged. They
described a two-dimensional technique that allows separation of complex mixtures of surfactants. These thin-layer techniques
provide the chemist with a quick and easy
Page 100

scan of an unknown formulation for surfactant composition without the use of expensive instrumentation.
Nuclear Magnetic Resonance Spectroscopy (NMR) has also been suggested as a tool for preliminary screening of formulated
products. Carminati et al. [95] have shown that it is possible to use 13C NMR to identify the surfactants in commercial
products. Where this is possible it eliminates the necessity for the time-consuming separation steps.

C
Separation and Quantitation

The thin-layer and NMR schemes may give an idea what surfactants are present but it is impossible to quantitate using these
methods. High performance liquid chromatography (HPLC) may be used to both identify and quantitate the surfactants.
Nakamura and Morikawa [96] have developed a system for separating nonionic, anionic, and amphoteric surfactants from each
other using a reversed phase technique with sodium perchlorate in the mobile phase. The surfactants, all of which contained n-
dodecyl groups, were separated from each other and determined in formulations without any pretreatment such as ion
exchange. Ban et al. [97] have shown that simultaneous and isocratic separations of alkybenzenesulfonates, nonylphenol
ethoxylates, fatty acid ethoxylates, and fatty amine ethoxylates can be achieved using a reversed-phase HPLC technique. In
some cases a preliminary separation on an anion exchange resin was necessary.

Yoshimura et al. [23] have developed an HPLC method that allows the quantitation of alkyl ethoxylates in formulated
products. Using a backflush technique, they developed a method in which the ethoxylate chromatographs as a single peak with
no splitting due to either alkyl chain length or ethoxylate distribution. Since the ethoxylate chromatographs as a single peak,
quantitation is fairly straightforward.

Some work has been done using HPLC to identify and quantitate surfactants in environmental samples. These methods might
be adapted to quantitate the surfactant in a formulation. Matthijs and Hennes [98] describe an HPLC procedure for the
determination of anionic, nonionic, and cationic surfactants and compare the technique to the curretly used colorimetric
techniques. Evans et al. [99] present a method for the quantitative determination of linear primary-alcohol ethoxylates using
LC/MS. This method allows the quantification of the individual alcohol ethoxylates. Some HPLC techniques have also been
used to characterize the alkyl and EO distribution in ethoxylated alkyl amines. Schreuder et al. [100] describe a system in
which they use a cyano-modified silica column to determine the alkyl distribution and an amino-modified silica to determine
the ethylene oxide distribution. Henrich [93] describes an HPLC system for identifying octyl and nonylphenols by their
oligomer distribution.

Nakamura and Morikawa [101] developed a methodology that can separate surfactants from each other and can also separate
them by their individual homolog distribution. They used representatives from each class of surfactants to illustrate their
separations but did not use any ethoxylated surfactants. Konig and Strobel [102] describe an HPLC method that separates,
identifies, and quantitates the surfactants found in toothpaste. They use a reversed-phase column with a methanol/water mobile
phase (modified with sodium perchlorate) and a refractive index detector to carry out the analysis. Since they can separate the
surfactants based on the carbon chain length, they can use this system to identify as well as quantitate the surfactant.
Page 101

D
Alkoxylate Ratios

When alkoxylated surfactants have been used it is often advantageous to know the number of moles of EO or PO that are
present in the molecule. There are several approaches used to determine the ratio of alkoxylate to hydrophobe. Gas
chromatography was used by Kusz et al. [103] to determine the EO/PO ratio in block copolymers. The gas chromatography
was carried out after degradation using acetyl chloride.

Another approach uses FTIR spectroscopy. Das and Kumar [104] developed a method to determine the moles of EO on the
lauryl alcohol in lauryl alcohol ethoxylates using a ratio of the peak absorbances at 720 cm-1 and 843 cm-1. The ratio of peak
heights can be used to predict the moles of EO on an unknown ethoxylate once a calibration curve has been established using
known ethoxylates. Nuclear Magnetic Resonance has been widely used to determine the ethylene oxide content of ethoxylated
surfactants. Proton NMR was used by Hammond and Kubik [105] to determine the EO content of alcohol ethoxylates. They
also determined the precision and accuracy of their method. Carbon-13 NMR was used by Gronski et al. [106] to characterize
ethylene-oxide/propylene-oxide adducts. They were able to determine whether it was a block or random polymer, the mean
sequence lengths of the EO/PO sequences, the number of blocks per 100 monomeric units, and the starter and end groups.

Abbreviations

ECD Electron Capture Detector


ELSD Evaporative Light Scattering Detector
EO Ethylene Oxide
FID Flame Ionization Detector
FTIR Fourier Transform Infrared Spectroscopy
GC Gas Chromatography
HPLC High Performance Liquid Chromatography
IR Infrared Spectroscopy
LC/MS Liquid Chromatography/Mass Spectrometry
NMR Nuclear Magnetic Resonance Spectroscopy
PEG Polyethylene Glycol
PO Propylene Oxide
SFC Supercritical Fluid Chromatography
TEA Thermal Electron Analyzer
THF Tetrahydrofuran
TLC Thin Layer Chromatography

References

1. T. M. Schmitt, Analysis of Surfactants, Marcel Dekker, New York, 1992.

2. B. M. Milwidsky, and D. M. Gabriel, Detergent Analysis, Halsted Press, New York, 1982.

3. G. F. Longman, The Analysis of Detergents and Detergent Products, Wiley-Interscience, London, New York, Sydney,
Toronto, 1975.

4. M. R. Porter, Recent Developments in the Analysis of Surfactants, Elsevier Applied Science, London & New York, 1991.
Page 102

5. Official Methods & Recommended Practices of the American Oil Chemists Society, 4th ed., D. Firestone, (ed.), American
Oil Chemists Society, 1990.

6. CTFA Compendium of Cosmetic Ingredient Composition, Methods, J. Nikitakis, and G. N. McEwen, (eds.), Cosmetic,
Toiletry and Fragrance, Association, Washington, DC. (1990).

7. Newburger's Manual of Cosmetic Analysis, A. J. Senzel, (ed.), Association of Official Analytical Chemists, Washington,
DC, 1977.

8. Official Methods of Analysis of the Association of Official Analytical Chemists, K. Helrich, (ed.), Association of Official
Analytical Chemists, Arlington, VA, 1990.

9. D. Hummel, Identification and Analysis of Surface-Active Agents, vol. I and II, Interscience Publishers, New York, 1962.
10. The Infrared Spectra Atlas of Surface Active Agents, Stadtler Research Laboratories, Philadelphia, PA, 1982.

11. American Society of Testing and Materials, 1916 Race St., Philadelphia, PA, 19103.

12. H. H. Y. Oei, I. Mai, and D. C. Toro, J. Soc. Cosmet. Chem. 42:309316, (1991).

13. D. J. Pietrzyk, P. G. Rigas, D. Yuan, J. Chromatogr. Sci. 27:485490 (1989).


14. D. Zhou, and D. J. Pietrzyk, Anal. Chem. 64:10038 (1992).

15. J. R. Larson, J. Chromatogr. 356:37981 (1986).

16. M. G. Kunitani, and L. M. Kresin, Anal. Biochem. 182:1038 (1989).

17. J. Weiss, J. Chromatogr. 353:3037 (1986).

18. C. Kwan, S. Lin, Chieh Mien K'o Hsueh 24:1627 (1984).


19. M. A. Castles, B. L. Moore, and S. R. Ward, Anal. Chem. 61:253440 (1989).

20. S. Chen, and D. J. Pietrzyk, Anal. Chem. 65:27705 (1993).

21. P. L. Desbene, C. Rony, B. Desmazieres, and J. C. Jacquier, J. Chromatogr. 608:37583 (1992).

22. M. Kudoh, and K. Tsuji, J. Chromatogr. 294:45659 (1984).


23. H. Yoshimura, T. Sugiyama, and T. Nagai, J. Am. Oil Chem. Soc. 64:55055 (1987).

24. Rhone Poulenc Standard Method, RP-1079, Rhone Poulenc, Cranbury, NJ.

25. J. Cross, in Chemical Analysis, Marcel Dekker, New York, 1987.

26. G. Meszlenyi, G. Kortvelyessy, E. Juhasz, and M. Eros-Lelkes, Acta Chim. Hung. 128:17981 (1991).
27. C. K. Cross and A. C. Mackay, J. Am. Oil Chem. Soc. 50:24950 (1973).

28. M. Kudoh, J. Chromatogr. 291:32730 (1984).

29. B. Desmazieres, F. Portet, and P. L. Desbene, Chromatographia 36:30717 (1993).

30. T. Okada, J. Chromatogr. 609:21318 (1992).

31. B. Trathnigg, D. Thamer, X. Yan, B. Maier, H. R. Holtzbauer, and H. Much, J. Chromatogr. 657:36575 (1993).
32. Z. Wang, and M. Fingas, J. Chromatogr. 673:14556 (1993).

33. P. R. Geissier, J. Am. Oil Chem. Soc. 66:68589 (1989).

34. J. D. Pinkston, D. J. Bowling, and T. E. Delaney, J. Chromatogr. 474:97111 (1989).

35. H. T. Kalinoski and L. O. Hargiss, J. Chromatogr. 505:199213 (1990).


36. M. Lafosse, C. Elfakir, L. Morin-Allory, and M. Dreux, J. High Res. Chromatogr. 15:31218 (1992).

37. S. Brossard, M. Lafosse, and M. Dreux, J. Chromatogr. 591:14957 (1992).

38. Z. Wang and M. Fingas, J. Chromatogr. 641:12536 (1993).

39. Z. Wang, and M. Fingas, J. Chromatogr. Sci. 31:50918 (1993).


40. P. L. Desbene and B. Desmazieres, J. Chromatogr. 661:20713 (1994).
Page 103

41. A. Silver and H. T. Kalinoski, J. Am. Oil Chem. Soc. 69:599608 (1992).
42. P. Sandra and F. David, J. High Res. Chromatogr. 13:41417 (1990).

43. H. T. Kalinoski and A. Jensen, J. Am Oil Chem. Soc. 66:117175 (1989).

44. A. Aserin, N. Garti, M. Frenkel, J. Liq. Chromatogr. 7:154557 (1984).

45. M. Kudoh, M. Kotsuji, S. Fudano, and K. Tsuji. J. Chromatogr. 18791 (1984).


46. A. A. Ben-Bassat and T. Wasserman, J. Liq. Chromatogr. 10:293950 (1987).

47. C. N. Wang and L. D. Metcalfe, J. Am. Oil Chem. Soc. 62:55860 (1985).

48. L. D. Metcalfe, J. Am. Oil Chem. Soc. 61:36366 (1984).

49. A. J. Wilkes, G. Walraven, and J-M. Talbot, J. Am. Oil Chem. Soc. 69:60913 (1992).

50. J. Vialle, P. Navarro, T. T. Nguyet, P. Lanteri, and R. Longeray, J. Chromatogr. 549:15974 (1991).
51. J. Krol, P. G. Alden, and J. Morawski, J. Chromatogr. 626:16590 (1992).

52. H. Koenig and W. Strobel, Proc. 2nd World Surfact. Congress 3:10822 (1988).

53. R. R. Chadwick, J. C. Hsieh, K. S. Resham, and R. B. Nelson, J. Chromatogr. 671:40310 (1994).

54. L. M. Nair, R. Saari-Nordhaus, J. M. Anderson, J. Chromatogr. 671:30913 (1994).


55. Rhône Poulenc Standard Method, RP-0108 or RP-1109, Rhône Poulenc, Cranbury, NJ.

56. Rhône Poulenc Standard Method, RP-0016, Rhône Poulenc, Cranbury, NJ.

57. G. Richard, P. Gataud, J. Arnaud, P. Bore, in Cosmetic Analysis (P. Bore, ed.), Marcel Dekker, New York, pp. 15777
(1985).
58. N. de Kruijf and A. Schouten, Parfumerie and Kasmetik 72:38692, 394, 396, 398 (1991).

59. D. H. Liem, Cosmet. Toiletries 92:5972 (1977).

60. C. H. Wilson, J. Soc. Cosmet. Chem. 26:7581 (1975).

61. N. de Kruijf, A. Schouten, M. A. H. Rijk, and L. A. Pranoto-Soetardhi, J. Chromatogr. 469:31728 (1989).

62. J. Walker, Formaldehyde, 3rd ed., Reinhold, New York, 1964.


63. C. A. Benassi, A. Semenzato, and A. Bettero, J. Chromatogr. 494:38793 (1989).

64. W. R. Summers, Anal. Chem. 62:13971402 (1990).

65. A. Semenzato, C. A. Benassi, G. Rossi, A. Bettero, M. Lucchiaro, and R. Cerini, Int. J. Cosmet. Sci. 12:26572 (1990).

66. C. A. Benassi, A. Semenzato, F. Zaccaria, and A. Bettero, J. Chromatogr. 502:193200 (1990).


67. M. O. Masse, B. Wyhowski de Bukanski, and C. Gilquin, Cosmet. Toiletries 99:46, 48, 50, 556, 589 (1984).

68. D. R. Hoar, D. J. Sissons, Methodol. Dev. Biochem. 5:2216 (1976).

69. F. H. Marquardt, J. Schulze, and D. Smith, NBS Special Publication, 422, (1976).

70. R. G. Achari and D. Chin, J. Soc. Cosmet. Chem. 32:16373 (1981).

71. E. D. George, E. J. Hillier, and S. Krishnan, J. Am. Oil Chem. Soc. 57:13134 (1980).
72. United States Pharmocopeia, XXII, USP Inc., Rockville, MD, 20852.

73. In House Publication Rohm and Haas Company, 1988. Independence Mall W, Phila, PA. 19105.

74. R. Matissek, R. Nagorka, I. Wengatz, and J. Rohde, Fresenius Z. Anal. Chem. 332:81316 (1988).

75. J. R. Dahlgran and C. R. Shingleton, J. Assoc. Off. Anal. Chem. 70:79698 (1987).
76. H. Leskovsek, A. Grm, and J. Marsel, Fresenius Z. Anal. Chem. 341:72022 (1991).

77. Z. L. Cardeal, D. Pradeau, B. Lejeune, and M. Hamon, Analusis 22:2326 (1994).


Page 104

78. K. Sasaki, K. Kijma, M. Takeda, and S. Kojima, J. Assoc. Off. Anal. Chem. Int. 76:29296 (1993).
79. J. W. Danielson, R. P. Snell and G. S. Oxborrow, J. Chromatogr. Sci. 28:97101 (1990).

80. S. Scalia, F. Testoni, G. Frisina, and M. Guarneri, J. Soc. Cosmet. Chem. 43:20713 (1992).

81. M. Italia, and M. Nunes, J. Soc. Cosmet. Chem. 42:97104 (1991).

82. N. Goetz, G. Kaba, H. Burgaud, and N. Paoletti, in Cosmetic Analysis (P. Boré, ed.), Marcel Dekker, New York, 1985, p.
139148.

83. H. Beernaert, M. Herpol-Borremans, and F. DeCock, Belg. J. Food Chem. Biotechnol. 42:13135 (1987).

84. S. C. Rastogi, Chromatographia 29:44145 (1990).

85. S. Scalia, J. Pharm. Biomed. Anal. 8:867870 (1990).


86. D. B. Black, R. C. Lawrence, E. G. Lovering, and J. R. Watson, J. Assoc. Off. Anal. Chem. 66:18083 (1983).

87. M. Erickson, D. B. Lakings, A. D. Drinkwine, and J. L. Spigarelli, J. Soc. Cosmet. Chem. 36:22330 (1985).

88. J. Meili, P. Bronnimann, B. Brechbuhler, and H. J. Heiz, J. High Res. Chromatogr. and Chromatogr. Comm 2:47580
(1979).

89. W. Gorski, and J. A. Cox, Anal. Chem. 66:277174 (1994).


90. Standard for the Classification of Surfactants by Infrared Absorption, ASTM.D 235774, American Society for Testing and
Materials, 1916 Race St., Phila., PA, 19103.

91. J. E. Nettles, Infrared Spectroscopy for Identifying Surfactants 1:43041 (1969).

92. C. T. Desmond, and W. T. Borden, J. Am. Oil Chem. Soc. 41:55253 (1964).
93. L. H. Henrich, J. Planar Chromatogr. 5:103117 (1992).

94. D. W. Armstrong, and G. Y. Stine, J. Liq. Chromatogr. 6:2333 (1983).

95. G. Carminati, L. Cavalli, and F. Buosi, J. Am. Oil Chem. Soc. 65:66977 (1988).

96. K. Nakamura, and Y. Morikawa, J. Am. Oil Chem. Soc. 61:113035 (1984).
97. T. Ban, E. Papp, and J. Inczedy, J. Chromatogr. 593:22731 (1992).

98. E. Matthijs, and E. C. Hennes, Tenside Surf. Det. 28:2227 (1991).

99. K. A. Evans, S. T. Dubey, L. Kravetz, I. Dzidic, J. Gumulka, R. Mueller, and J. R. Stork, Anal. Chem. 66:699705 (1994).

100. R. H. Schreuder, and J. Martijn, Chromatogr. 368:33950 (1986).

101. K. Nakamura, and Y. Morikawa, J. Am. Oil Chem. Soc. 59:6468 (1982).
102. H. Konig, and W. Strobel, Fresenius Z. Anal. Chem. 331:43538 (1988).

103. P. Kusz, J. Szymanowski, K. Pyzalski, and E. Dziwinski, LC-GC 8:48, 50 (1990).

104. S. Das and V. V. Kumar, Indian Journal of Chemistry 32A:10045 (1993).

105. C. E. Hammond and D. K. Kubik, J. Am. Oil Chem. Soc. 71:11315 (1994).
106. W. Gronski, G. Hellmann, and A. Wilsch-Irrgang, Makromol. Chem. 192:591601 (1991).
Page 105

4
Principles of Emulsion Formation
Thomas Förster
Chemical Research, Henkel KGaA, Düsseldorf, Germany

I. Basic Mechanisms to Produce Emulsion Droplets 105

A. Energy Scale for Droplet Break-up 106

B. Droplet Break-up under Shear 106

C. Droplet Break-up under High Pressure 108


D. Role of Surfactant: Droplet Size Reduction and Stabilization 108

II. Phase Behavior and Structure of Emulsions 109

A. Representation of Phase Behavior by Phase Diagrams 109

B. Properties of Microemulsions and Lamellar Phases Relevant to Emulsion 114


Formation

III. Phase-Inversion Emulsification 119

A. Balanced Surfactant Systems and Optimum Formulation 119

B. PIT Emulsification and Gel-Phase Emulsification 120

References 123

I
Basic Mechanisms to Produce Emulsion Droplets

Emulsions are disperse systems in which twoor sometimes severalalmost insoluble liquid phases are intimately mixed. Except
for special cases such as microemulsions or high internal phase emulsions with a foam structure, the internal phase is
contained in the external phase in the form of spherical droplets. Thus in the simplest case there is either an oil in water (o/w)
or a water in oil (w/o) emulsion. However for cosmetic and technical applications, multiphase emulsions are also quite
common; for example a water in oil emulsion can be dispersed in water to obtain a w/o/w emulsion.
Page 106

A
Energy Scale for Droplet Break-up

Because of their dispersed character, emulsions have a large internal surface, which is energetically unfavorable. This has two
basic consequences: first, energy is consumed during emulsion preparation, and second, emulsions are thermodynamically
unstable since there is practically no entropic gain that would compensate the energy consumption (internal surface A ·
interfacial tension (go/w). An appreciable positive mixing entropy only results in the case of colloidal mixing in a
microemulsion. Therefore, apart from oil and water, the preparation of emulsions also requires surfactants and energy.

The actual energy demand for the preparation of an emulsion is several orders of magnitude higher than the amount of energy
stored in the emulsion in the form of surface energy because of the droplet break-up mechanism [1]. During the preparation of
an emulsion, large oil droplets are first deformed and then broken up. The Laplace pressure (p = 4 go/w/D, with droplet diameter
D) acts against the droplet deformation. For a w/o emulsion with a typical interfacial tension of approximately 2.5 mN/m and a
droplet diameter of 1 µm, a Laplace pressure of 10,000 Pa has to be overcome. In order to obtain such a pressure gradient over
a length in the order of magnitude of the droplet diameter, a very high energy input is required, the greatest part of which
dissipates as heat.

B
Droplet Break-up under Shear

A large number of emulsifying processes are based on droplet break-up under shear [2]. The break-up of a droplet is possible
when the deforming force exceeds the interfacial force that maintains the shape. This ratio is described by the Weber index. In
the case of laminar flowe.g., in the clearance of a colloid mill (see Fig. la)it is

Fig. 1
Types of emulsifying
machines: (a) colloid mill

(b) stirrer
(c) homogenizer.
(For detailed survey see Ref. 2.)
Page 107

t is the shear tension that results from the product of emulsion viscosity hE and shear rate v.
A droplet break-up is only possible up to a minimal diameter Dmin, which is characterized by the critical Weber index Wecr [3,4].
Through individual droplet examinations under laminar shear and emulsifying experiments in a colloid mill, Schubert and
Armbruster showed that the critical Weber index only depends upon the viscosity ratio between the dispersed phase and the
emulsion hd / hE [5,6]. The critical Weber index passes a distinct minimum in the range 0.1 < hd / hE < 2 (see Fig. 2). Only in
this range of minimal Weber indices can finely dispersed emulsions be produced under laminar shear. For Wecr = 1

With an emulsion viscosity hE of 0.1 Pa·s a shear rate v of 5000/s in the clearance of a colloid mill, and an interfacial tension
go/w of 2.5 mN/m, a simple estimation yields a minimum droplet diameter of 10 µm.
One precondition for an effective emulsification in colloid mills is therefore a sufficiently high emulsion viscosity.
Furthermore, the viscosity ratio hd / hE is often too high for o/w emulsions. In this case, the oil droplets behave as hard spheres,
i.e. they are not deformed. Thus another emulsifying process must be applied, which is either not

Fig. 2
Critical Weber number as a function of viscosity ratio for o/w emulsions
with C12E10 (laureth-10) and C12E20 (laureth-20) as emulsifier.
(Reprinted from Ref. 5 by courtesy of VCH Verlagsgesellschaft.)
Page 108

based on laminar shear or utilizes an emulsion inversion to the type w/o (see below) during preparation.
During the process of stirring (see Fig. 1b), the droplets are broken up by pressure gradients in a turbulent flow. Instead of
shear stress, now the Reynolds tension 1/2 Dv2 · r, with the mean square velocity fluctuation Dv2 and the density r, must be
used to calculate the Weber index [5]. According to Kolmogorov, for the simplifying assumption of a homogeneous, isotropic
turbulence, the mean square velocity fluctuation Dv2 can be calculated from the product of power density P/V and droplet
diameter D. The smallest droplet size attainable with stirrers is [5,7].

C
Droplet Break-up under High Pressure

In high pressure homogenizers the emulsion is forced through a homogenizing nozzle (radial gap in Fig. 1c). Here
homogenizing pressures of 100 to 1000 bar are reached with volume flows of up to 50,000 L/h [5]. Equation (3) can be used
again for the calculation of the droplet size. The power P is calculated from the product of pressure drop p and volume flow
. Because of the high pressure drop and the short residence time of the emulsion in the homogenizing nozzle, very high power
densities are achieved. In contrast to the colloid mill, the power density is independent of the emulsion viscosity and the
viscosity ratio hd / hE . Therefore, high-pressure homogenizers are particularly effective when finely dispersed low-viscosity
emulsions are emulsified (e.g., homogenization of milk). A detailed comparison of the power densities of different emulsifying
devices can be found in Reference [7].

D
Role of Surfactant:
Droplet Size Reduction and Stabilization

During the preparation of an emulsion, droplets not only break up, but at the same time they may recoalesce to larger droplets.
Both phenomena are influenced decisively by surfactants.

Surfactants reduce the interfacial tension between the oil and the water phase from approximately 25 to below 2.5 mN/m for
w/o emulsions and below 0.25 mN/m for o/w emulsions, thus facilitating droplet break-up. In the colloid mill the theoretically
attainable droplet diameter is 10 to 100 times smaller [see Eq. (2)]; in high-pressure homogenization it is 4 to 16 times smaller
with than without surfactant [see Eq. (3)].

However, these theoretically expected minimal droplet sizes are not achieved in practice. One important reason is
recoalescence into larger droplets during the preparation of an emulsion [7,8]. Surfactants stabilize the droplets against
coalescence as a result of a ''self-healing mechanism" that is known as the GibbsMarangoni effect (see Fig. 3) [9,10].

The surfaces of two emulsion droplets that have just been separated arebecause of the sudden surface increasecovered
incompletely with surfactant molecules (Fig. 3, left). When these two droplets approach each other again, surfactant molecules
from the external phase will adsorb unevenly at the droplet surface (Fig. 3, center). The adsorbed surfactant quantity is
smallest at the point where a thin film forms between the two emulsion droplets. This results in a surfactant concentration and
interfacial tension
Page 109

Fig. 3
Scheme of the Gibbs-Marangoni effect.

gradient that causes the surfactant molecules to move from areas rich in surfactant to the contact surface depleted of surfactant
molecules (Fig. 3, right). The surfactants drag along molecules of the external phase, and as a result, the droplets are separated,
i.e., recoalescence is prevented.

A precondition for the stabilizing GibbsMarangoni effect is a surfactant concentration gradient in the droplet surface that
develops due to the uneven adsorption of the surfactant from the external phase. If the surfactant is dissolved in the interior of
the emulsion droplets, the surfactant concentration gradient cannot develop, and coalescence is not prevented. In this way the
GibbsMarangoni effect also supplies the explanation for the well-known Bancroft rule: The phase with the better surfactant
solubility becomes the external phase in an emulsion [11]. Dispersed droplets that contain dissolved surfactant are not
protected against recoalescence.

II
Phase Behavior and Structure of Emulsions

Simple o/w or w/o emulsions contain two phases (water and oil) whose thermodynamically unstable intimate mixture is
stabilized with surfactants.

Due to their amphiphilic (hydrophilic and lipophilic) molecular structure, surfactants have a tendency to aggregate in aqueous
or oily environments, e.g., to form micelles or liquid crystals. These aggregates form thermodynamically stable phases that can
also change the macroscopic appearance of an emulsion.
A
Representation of Phase Behavior by Phase Diagrams

The type of surfactant and oil, the mixing ratios, and the external conditions such as temperature and salt or solvent content
determine which macroscopic phases can occur in a certain surfactant-oil-water system. These complex relationships can be
represented graphically in the form of phase diagrams, as shown for a simple three-component system of C12E7
(dodecylpolyoxyethylene(7)glycol or laureth-7), decane, and water in the diagram in Fig. 4. In order to provide a clear
overview, only single-phase areas are represented, not the different multiphase emulsion types.
In this emulsion system, there are three single-phase areas at 20°C: the cubic gel phase, the hexagonal phase, and the lamellar
liquid-crystalline phase. In each corner, one of the three components is present in its pure form. The composition of mixtures is
read
Page 110

Fig. 4
Ternary phase diagram of C12E7, decane, and
water at 20°C and liquid crystalline structures.
(Reprinted from Ref. 12 by courtesy of Steinkopff Verlag.)

with the aid of parallels to the three edges. A system of 10% decane, 60% water, and 30% C12E7 therefore forms a transparent,
high-viscosity cubic gel I, which could, for example, be used as a basis for a sunscreen gel or a topical preparation (1315).
Other common names for the cubic gel phase are transparent oil-water gel [16], microemulsion gel [17] orin the case of special
acoustic propertiesringing gel [18].

In this simple case the cubic gel phase can be imagined as a crystal of closely packed oil-swollen surfactant micelles [12,18].
The curvature radius of the surfactant film is very large.

With an increasing surfactant/water ratio, the hydration and therefore the polyoxyethylene head group area decline, and the
curvature radius and the packing conditions in the liquid crystal change. A hexagonal liquid crystal HI results (where I
indicates that water is the external phase), which has a high viscosity because of its rod-like structure. The curvature radius
decreases further during further exchange of surfactant for water, and a planar structure, the lamellar liquid crystal La, forms.
The planar lamellar layers are easily movable, and as a result, the viscosity is distinctly lower than that in the hexagonal or
cubic liquid crystal.
It is apparent from the phase diagram that the cubic phase exists only in a narrow concentration range, while the hexagonal and
the lamellar phases cover large areas. Lamellar and hexagonal liquid crystals in particular also occur in the oil-free system (for
an overview of phase behavior of nonionic surfactant water systems see Reference 20).

Two-dimensional triangular diagrams usually show the influence of the mixing ratio
Page 111

of oil, water, and surfactant on the phase behavior of ternary systems. In the case of four-component systems the phase
behavior must be represented in the form of three-dimensional phase tetrahedra [21,22]. As a rough approximation the four-
component systems can be considered as so-called pseudo three-component systems in which two componentsin most cases
the two emulsifying agentsare summed to form a pseudo component exhibiting a constant mixing ratio [23,24,27].
In order to represent the influence of process variables such as temperature or salt content on the phase behavior of a ternary
system, a so-called phase prism is often used [25].

The schematical phase prism in Fig. 5 demonstrates the decisive influence of temperature on the phase behavior of ternary oil-
water-ethoxylated nonionic surfactant mixtures; in this case the emulsion and microemulsion phases with a low surfactant
content were studied (in contrast to Fig. 4 where the single-phase liquid crystalline phases rich in surfactant were examined).
Higher ethoxylated, nonionic surfactants form oil-swollen micelles in water (wm) at low temperatures and stabilize o/w
emulsions. With increasing temperature, oil solubilization in the micelles increases and finally a third phase (D) with a high
surfactant content forms. At low surfactant concentration, the system separates into three phases: an oil phase, a water phase
poor in surfactant, and a middle phase rich in surfactant, which solubilizes large quantities of water and oil (Winsor III). With a
further rise in temperature, the ethoxylated surfactant head groups are dehydrated even more, thus increasing the surfactant
solubility in oil. Finally inverse, water-swollen

Fig. 5
Phase prism, Winsor types, and dispersion structure of an emulsion composed of nonionic surfactant, oil, and water.
(Reprinted from Ref. 27 by courtesy of VCH Verlagsgesellschaft.)
Page 112

surfactant micelles are present in the oil (om), and w/o emulsions result. An increase in temperature therefore induces a phase
inversion from an o/w into a w/o emulsion; in the range of the phase inversion temperature (PIT) the system passes through a
microemulsion phase.

Perpendicular cuts through the phase prism yield clear two-dimensional representations. The cutas indicatedthrough the
surfactant corner with fixed o/w-ratio (often 1:1) shows the dependence of the phase behavior of an emulsion system on
temperature (on the ordinate) and surfactant concentration (on the abscissa) (see Fig. 6). The phase inversion temperature
range in which a microemulsion exists appears as the well-known "Kahlweit fish" [25,27]. The body of the fish corresponds to
the three-phase-microemulsion range (Winsor III), the tail of the fishat higher surfactant concentrationscorresponds to the
single-phase microemulsion (Winsor IV) or the lamellar liquid crystal. The cusp of the single-phase microemulsion, where tail
and body meet, specifies the minimum surfactant concentration that suffices to solubilize the entire oil and water quantity in
the D-phase, i.e. it is a measure of the solubilization capacity of the surfactant used. At low temperatures around 40°Cbelow
the lamellar liquid crystal areathe single-phase microemulsion is water-continuous and consists of oil-swollen, strongly bent
micelles [28]. The curvature radius of the surfactant layer decreases with increasing temperature, and a lamellar phase results
in which the surfactants are packed in parallel, stiff layers [29]. The stiffness of the lamellar layers declines with a further rise
in temperature while the mean curvature radius remains constant and a single-phase

Fig. 6
"Kahlweit-fish" in the phase diagram of
C12E5, tetradecane, and water.
(Reprinted from Ref. 27 by courtesy of VCH
Verlagsgesellschaft.)
Page 113

bicontinuous microemulsion forms [29,30]. Freeze fracture electron microscopy pictures of different microemulsion structures
are shown in References 28 and 31.

Another common cut through the phase prism runs parallel to the oilwater axis with a constant surfactant content. This
representation (Fig. 7) demonstrates solubilization limits and cloud points [32]. On the ordinate, decisive formulation variables
such as temperature (for ethoxylated nonionic surfactants [3234]), salt content (for ionic surfactants [35]) or surfactant mixing
ratios [32,34] are specified. The structure of the microemulsion phase changes with the oil-water ratio [29,33,36]: The water-
rich side shows oil-swollen micellar droplets in water (wm), which are referred to as an o/w microemulsion. With an increasing
o to w ratio the o/w microemulsion changes into a three-phase microemulsion (w + D + o) orin the case of higher surfactant
contentinto a bicontinuous microemulsion (D). On the oil-rich side, a w/o microemulsion (om) forms with droplet structure.

The third important representation type finally is the water content map, i.e., a perpendicular cut through the water corner with
a constant oilemulsifier ratio (Fig. 8). This representation is particularly suitable for the planning of production processes,
since the usual production methodpreparation of an oilemulsifier mixture, heating, and addition of watercan be taken into
consideration [37]. At water contents in the range from 15 to 80% o/w emulsions exist, which invert nearly constantly between
78 and 98°C. At extreme water contents the phase inversion is no longer possible due to

Fig. 7
Phase diagram of tetradecanewater emulsions containing 5% C12E5.
(Reprinted from Ref. 32 by courtesy of Steinkopff Verlag.)
Page 114

Fig. 8
Temperaturewater-content map of an emulsion containing a fatty alcohol,
nonionic surfactant, and paraffin oil in a mixing ratio of 3/4/20.
(Reprinted from Ref. 37 by courtesy of Allured Publishing.)

geometric reasons, and unstable w/o emulsions (w/o*) or o/w emulsions form at very low or very high water contents,
respectively. The temperaturewater-content map shows which phases are passed through with certain production paths (see e.g.
path 1 for a one-step hot process and path 2 for a two-stage hot-cold process, where an emulsion concentrate with only 20%
water is diluted at 85°C with cold water). In addition, Fig. 8 contains information about the consistency of the systems at 25°C
as a function of the water content. At water contents below 40%, close packing of the oil droplets is approached, resulting in a
drastic increase in the viscosity and yield value.

B
Properties of Microemulsions and Lamellar Phases Relevant to Emulsion Formation

Why are microemulsions interesting for the preparation of emulsions? Microemulsions are specific systems in which the
emulsifying agents are optimally adapted to the oil and water phase and therefore show maximum interfacial activity. This is
manifested in minimum interfacial tension and maximum solubilization [25,32,3840]. In the phase diagram in Fig. 7, the
single-phase ranges wm and om specify the solubilization limits in the case of extreme o to w ratios. Along the perpendicular
line for an o to w ratio of 5:95 at temperatures below 35°C an oil in water emulsion (o + wm) forms. As the phase inversion
temperature range is approached, oil solubilization increases, and in the phase inversion temperature (PIT) range (36 to 41°C)
the micelles can solubilize more than 5%
Page 115

oil, resulting in a single-phase range of oil-swollen micelles (wm), an o/w microemulsion. The micelles aggregate above the
PIT range, and a phase with a high surfactant content (w + D) separates. Oil-swollen micelles are oil microdroplets with a
radius between 2 and 30 nm coated with a surfactant monolayer [40,41]. The surfactant/oil ratio determines the droplet radius
[40,41].
Similar conditions prevail in the case of the reverse o to w ratio (95:5). In the phase inversion range (50 to 55°C in Fig. 7)
water-swollen inverse micelles in oil (om) form. The radius of the water microdroplets lies in the range of 5 to 30 nm and is
determined by the surfactant/water ratio [41].

With a medium o to w ratio, a three-phase Winsor III microemulsion develops in the phase-inversion temperature range. Figure
9 shows the interfacial tension curve and the solubilized oil and water volume fractions with increasing temperature along the
line in Fig. 7 with an o to w ratio of 1:1. When the phase inversion temperature range is approached, the interfacial tension
between the aqueous micellar solution (wm) and the oil phase decreases by several orders of magnitude while the oil
solubilization in the micellar phase increases. In the phase inversion temperature range (45 to 50°C) a microemulsion phase D,
rich in surfactant, is present in addition to the almost surfactant-free oil and water phases [42]. The interfacial tension is
minimal [25,40,4346] and the water and oil solubilization in the surfactant phase D is maximal [25,32,38,39]. Since low
interfacial tension facilitates the mechanical droplet break-up (see Eqs. [23]), formation of a microemulsion phase during the
preparation of emulsions can improve the result. Above the phase-inversion temperature range the emulsion inverts into w +
om, the interfacial tension increases, and the water solubilization in the inverse micelles (om) declines.

During passage through the microemulsion range (Winsor I-Winsor III-Winsor II) the emulsion system inverts from o/w to
w/o, because the distribution of the surfactants between oil and water phase changes drastically (Bancroft rule).
At surfactant concentrations encountered in practice, the distribution of the surfactant between the oil and water phase is not
determined by the monomer solubility but rather by the critical micelle concentrations in oil and water [4750]: The surfactant
accumulates in the phase in which it first forms micelles, independent of whether the solubility of the surfactant monomers is
better in oil or in water. Thus, the Bancroft rule can be reformulated: The phase in which the surfactant forms micelles
becomes the external emulsion phase [50].

Generally in the phase-inversion range the major portion of surfactant is dissolved in oil and in the microemulsion phase
[22,51,52]. The surfactant concentration in the water is near the critical micelle concentration and is negligible. Surfactant
distribution between oil and microemulsion phase depends on the hydrophilic/lipophilic balance of the surfactant and the
dissolving properties of the oil. For commercial surfactants and surfactant mixtures of hydrophilic and lipophilic surfactants,
the surfactant composition in the microemulsion phase can deviate considerably from the weighed surfactant composition
since the lipophilic surfactants accumulate preferably in oil [22,53]. As a consequence, the surfactant mixture in the
microemulsion phase becomes poorer in the lipophilic components, and the phase-inversion temperature increases. This
depletion effect is more pronounced at lower ratios of surfactant to oil. Thus for mixtures of high-and low-ethoxylated fatty
alcohols the PIT declines with increasing overall surfactant concentration (Fig. 10). As can be seen, the depletion effect is
more strongly marked in
Page 116

Fig. 9
Influence of temperature on volume fractions of water,
oil;and surfactant phases (schematic) and interfacial
tension between phases in the system composed
of 5% C12E5, 47.5% water, and 47.5% tetradecane.
(Reprinted from Ref. 32 by courtesy of Steinkopff
Verlag and with permission from Ref. 40,
copyright American Chemical Society.)

the case of a higher portion of the lipophilic surfactant C12E4 in the surfactant mixture. Saturation of the lipophilic surfactant in
oil is achieved only at high surfactant levels, and the surfactant composition of the microemulsion phase then approaches the
weighed surfactant composition [22,5254].

The minimal interfacial tension attainable with microemulsions is the result of close surfactant packing in the fluctuating,
planar configuration of the bicontinuous D-phase. Another phase in which the surfactant molecules are closely packed in a
planar config-
Page 117

Fig. 10
Phase inversion temperature (PIT) as a function of mixed surfactant
concentration in an emulsion containing mixed fatty alcohol
ethoxylates, heptane, and water with a fixed-oil/water ratio of 1:1.
(Reprinted from Ref. 22 by courtesy of Academic Press.)

uration is the lamellar phase (see above, Fig. 6). Like the microemulsion phase, the lamellar liquid crystal also reduces the
interfacial tension between the oil and water phase [21,55]. Thus, after passing through a lamellar phase during the emulsion
preparation, finely dispersed emulsions are obtained. Furthermore, lamellar phasesunlike microemulsion phasescontribute
considerably to emulsion stabilization.

On account of their amphiphilic character, droplets of the dispersed phase may be covered with lamellar layers. This viscous,
lamellar film on the droplets is several layers thick and reduces the attraction potential between the droplets [21,56]. As a
result, the lamellar layer acts as a barrier against coalescence.

At usual surfactant concentrations the volume proportion of the lamellar phase can become so large in the emulsion system
that it not only covers the droplets as a stabilizing film but forms a macroscopic phase in which the droplets are dispersed [57].
Multi-phase emulsions of the type o/lc/w exist in which several oil droplets are trapped in one secondary liquid crystal (1c)
droplet of the lamellar phase (Fig. 11). Cosmetic emulsions often contain self-bodying agents, which, when mixed with
hydrophilic surfactants, build a lamellar gel phase [44,5865]. The structure of the lamellar gel phase resembles that of the
lamellar liquid crystal [65]. Because of the portion of high-melting self-bodying agents, e.g. cetostearyl alcohol or glycerol
monostearate, the lamellar layers are solid at room temperature [44,59,60,62,64]. At still higher proportion of viscosity-
contributing
Page 118

Fig. 11
Structure of an oil/liquid crystal/water emulsion.
(Reprinted from Ref. 58.)

ingredients, crystals of this ingredient separate from the lamellar gel phase (for details see Chapter 7).
The formation of the lamellar gel phase leads to important changes in the macroscopic emulsion properties. In cosmetic
applications changes of the rheological properties (viscosity and yield point) through the structural buildup in the emulsion
[58,65,66] and the increased water-bonding capacity of the lamellar gel phase [58,62,67] are most important. Figure 12 shows
the influence of the addition of cetostearyl alcohol on the phase-inversion temperature and on the yield point (a measure of the
consistency of the o/w emulsion). In a mixture with a fatty alcohol ethoxylate as hydrophilic emulsifier, cetostearyl alcohol
acts as a lipophilic coemulsifier that reduces the phase-inversion temperature from 85°C with 3% fatty alcohol to 65°C with
6%. In the range of 2 to 4%
Page 119

Fig. 12
Phase inversion temperature and yield value as a function of fatty alcohol content in an o/w
emulsion containing 4% C16/18E12 and 20% paraffin oil.
(Reprinted from Ref. 66 by courtesy of Chapman and Hall.)

cetostearyl alcohol, the yield point is zero and the emulsion viscosity is low. At higher concentrations, the yield point increases
linearly with the fatty alcohol content, resulting in o/w lotions or creams.

III
Phase-Inversion Emulsification
A
Balanced Surfactant Systems and Optimum Formulation

The formation of a lamellar phase or a microemulsion indicates that the hydrophilic/lipophilic properties of the surfactant
mixture are balanced and exactly adjusted to the water and oil phases, which can be seen from the minimal interfacial tension
and phase inversion (see Fig. 13) [25,32,3840]. In such an optimum formulation, the different formulation variables [e.g.
temperature, salt content, surfactant composition, ACN (alkane carbon number)] of the oil mixture, surfactant-to-oil ratio, and
water content follow specific rules [22,54,70,71]. Within the framework of the so-called R-theory (ratio R of the interaction
energies of the surfactant with the oil and the water) the influence of the different formulation variables as well as the
surfactant molecular structure on the phase behavior of microemulsions can be estimated semiquantitatively [72].

In practice these regularities are used in order to characterize surfactants and oils with a specificity that reaches far beyond the
long-known HLB system [73].
For ethoxylated nonionic surfactants there are characteristic HLB temperatures that can be combined according to a linear
mixing rule for the estimation of the emulsifying properties of surfactant mixtures [22,54,74]. A more general system,
originally developed for applications in tertiary oil recovery, is based on EACN (equivalent alkane carbon number) values for
the characterization of the surfactants [7577]. Especially for

Page 120

Fig. 13
Oil water interfacial tension (in mN/m) and macroemulsion type
(from electrical conductivity) against different formulation variables.
(Reprinted from Ref. 47 by courtesy of Royal Society of Chemistry, from Ref. 68
by courtesy of Elsevier Science Publishers and with permission from Ref. 69,
copyright American Chemical Society.)

cosmetic applications, the CAPICO (calculation of phase inversion in concentrates) concept was developed from these two
systems (for further details see Chapter 9) [78].

B
PIT Emulsification and Gel-Phase Emulsification

The relationship between microemulsion phase behavior and emulsions is most obvious in those emulsifying processes that
use the optimum formulation in a phase inversion. The phase-inversion temperature (PIT) emulsification is based on a
heatingcooling cycle in which the system passes through a microemulsion range (see Fig. 14). At increased temperatures o/w
emulsions that are stabilized with ethoxylated nonionic surfactants can invert into w/o emulsions. For surfactant mixtures with
long-chain nonionic surfactants and lipids commonly used in cosmetics, the phase-inversion temperature (PIT) can be adjusted
to range between 60 and 100°C (see Fig. 14). In this PIT range the hydrophilic/lipophilic properties of the surfactant mixture
are balanced, and a microemulsion or a lamellar phase with minimal interfacial tension forms [44]. In the subsequent cooling
process to room temperature, very little mechanical energy is required to break up the oil phase into minute oil droplets
[45,79]. As a result, a coarse o/w emulsion is converted into a finely dispersed bluish o/w emulsion through temperature-
induced phase inversion (see Fig. 15); it is stable against sedimentation because of the small oil-droplet size (diameter
approximately 120 nm) [66,7881]. Concerning coalescence, PIT emulsions are generally considered to be stable if the phase-
inversion temperature is far higher than the storage temperature [45,78,80]. Therefore, production temperatures (= PIT) are
usually between 75 and 90°C [78].

In practice, the PIT emulsifying process is preferably applied as a hot-cold process that allows considerable savings in energy
and processing time [37,8284]. The temper-
Page 121

Fig. 14
Principle of PIT emulsification: A course o/w emulsion is transformed into a blue
fine-disperse o/w emulsion by passing a microemulsion
during a heating-cooling cycle.
(Reprinted from Ref. 66 by courtesy of Chapman and Hall.)

Fig. 15
Microscopic appearance of o/w emulsions at 25°C as a function of
preparationcourtesy temperature.
(Reprinted from Ref. 66 by courtesy of Chapman and Hall.)
Page 122

aturewater-content map in Fig. 8 shows such a two-stage hotcold process. The single-step PIT process shown in Fig. 14 is
marked as production path 1 in Fig. 8. The whole emulsion with 7% mixed surfactant, 20% oil, and 73% water is heated to
85°C (to reach the microemulsion range) and cooled down. In the two-stage process (path 2), only an emulsion concentrate
with 20% water has to be heated to 85°C. In a second step the hot microemulsion concentrate is diluted with water of 40°C to
reach the final concentration while cooling. As a result, as with path 1, a finely dispersed and long-term stable o/w emulsion is
obtained [37].

In gel-phase emulsification, a microemulsion or a lamellar gel phase is not induced through temperature variation but through
polyol addition [8587]. The gel phase emulsification is also a two-stage process (see Fig. 16). In the first stepat room
temperatureoil is dispersed in a lamellar phase of surfactant, glycerol, and a small quantity of water so that a transparent o/lc
gel emulsion is obtained. The surfactant can be anionic [86] or an ethoxylated nonionic surfactant [85,87]. In the second step
this o/lc gel emulsion is diluted with water to obtain the final formulation. An o/lc/w emulsion forms in which the lamellar
liquid-crystalline phase protects the oil droplets against coalescence. The resulting oil droplet size depends primarily on the
surfactant/oil ratio and can be adjusted in a wide range from over 1000 nm to less than 100 nm [86]. During

Fig. 16
Principle of gel phase emulsification: (1) An o/lc gel
emulsion containing oil, monoarginine hexyldecyl phosphate,
glycerol, and a minor amount of water is formed; and (2) The
o/lc gel emulsion is diluted with cold water.
(Reprinted from Ref. 86 by courtesy of Academic Press.)
Page 123

the second dilution step, polyol concentration declines drastically. In the presence of especially suitable surfactants, the polyol
concentration in the final formulation can be lowered sufficiently to make this gel-phase emulsification suitable for the
preparation of cosmetic o/w emulsions. Both processes, the PIT emulsification as well as the gel-phase emulsification, show
that by utilizing the microemulsion phase behavior, finely dispersed o/w emulsions can be prepared by a simple emulsification
procedure without the need for highly sophisticated emulsifying equipment.

References

1. P. Walstra, in Encyclopedia of Emulsion Technology, vol. 1 (P. Becher, ed.), Marcel Dekker, New York, 1983, pp. 57127.

2. Ch. Fox, in Emulsions and Emulsion Technology, part 2 (K. J. Lissant, ed.), Marcel Dekker, New York, 1974, pp. 702933.
3. G. T. Taylor, Proc. Roy. Soc. 138:4148 (1932).

4. H. P. Grace, Chem. Eng. Commun. 14:22527 (1982).

5. H. Schubert and H. Armbruster, Chem. Ing. Tech. 61:70111 (1989).

6. H. Armbruster, H. Karbstein, and H. Schubert, Chem. Ing. Tech. 63:26667 (1991).


7. P. Walstra, Chem. Eng. Sci. 48:33349 (1993).

8. P. Becher and M. J. McCann, Langmuir 7:132531 (1991).

9. H. Lange, J. Soc. Cosm. Chem. 16:697714 (1965).

10. E. H. Lucassen-Reynders, Food Structure 12:112 (1993).

11. W. D. Bancroft, J. Phys. Chem. 17:501519 (1913).


12. J. A. Bouwstra, H. Jousma, M. M. van der Meulen, C. C. Vijverberg, G. S. Gooris, F. Spies, and H. E. Junginger, Colloid
Polymer Sci. 267:53138 (1989).

13. F. Comelles, V. Megias, J. Sanchez, J. L. Parra, J. Coll, F. Balaguer, and C. Pelejero, Int. J. Cosm. Sci. 11:519 (1989).

14. F. Comelles, J. Caelles, J. L. Parra, and J. Sanchez Leal, Int. J. Cosm. Sci. 14:183195 (1992).
15. T. H. El-Faham, S. M. El-Shanawany, and M. G. Abdel-Mohesen, Eur. J. Pharm. Biopharm. 38:180185 (1992).

16. C. Provost, Int. J. Cosm. Sci. 8:22347 (1986).

17. E. Nürnberg and W. Pohler, Progr. Colloid Polymer Sci. 69:6472 (1984).

18. M. Gradzielski, H. Hoffmann, and G. Oetter, Colloid Polymer Sci. 268:16778 (1990).

19. D. J. Mitchell and B. W. Ninham, J. Chem. Soc., Faraday Trans. 2, 77:60129 (1981).
20. D. J. Mitchell, G. J. T. Tiddy, L. Waring, T. Bostock, and M. P. McDonald, J. Chem. Soc., Faraday Trans. 1, 79:9751000
(1983).

21. S. E. Friberg and M. A. El-Nokaly, in Surfactants in Cosmetics (M. M. Rieger, ed.), Marcel Dekker, New York, 1985, pp.
5586.

22. H. Kunieda and K. Shinoda, J. Colloid Interface Sci. 107:10721 (1985).

23. S. Friberg, I. Lapczynska, and G. Gillberg, J. Colloid Interface Sci. 56:1932 (1976).

24. M. Kahlweit, R. Strey, and D. Haase, J. Phys. Chem. 89:16371 (1985).

25. M. Kahlweit, R. Strey, D. Haase, H. Kunieda, T. Schmeling, B. Faulhaber, M. Borkovec, H.-F. Eicke, G. Busse, F. Eggers,
T. Funck, H. Richmann, L. Magid, O. Söderman, P. Stilbs, J. Winkler, A. Dittrich, and W. Jahn, J. Colloid Interface Sci.
118:43653 (1987).

26. M. Kahlweit, Tenside Surf. Det. 30:8389 (1993).

27. M. Kahlweit and R. Strey, Angew. Chem. 97:66569 (1985).


Page 124

28. J. F. Bodet, J. R. Bellare, H. T. Davis, L. E. Scriven, and W. G. Miller, J. Phys. Chem. 92:18981902 (1988).
29. W. Sager and H.-F. Eicke, Colloids Surfaces 57:34353 (1991).

30. B. P. Binks, J. Meunier, O. Abillon, and D. Langevin, Langmuir 5:41521 (1989).

31. W. Jahn and R. Strey, J. Phys. Chem. 92:22942301 (1988).

32. K. Shinoda, Progr. Colloid Polymer Sci. 68:17 (1983).


33. U. Olsson, K. Shinoda, and B. Lindman, J. Phys. Chem. 90:408388 (1986).

34. K. Shinoda, H. Kunieda, T. Arai, and H. Saijo, J. Phys. Chem. 88:512629 (1984).

35. K. Shinoda and Y. Shibata, Colloids Surfaces 19:18596 (1986).

36. J. F. Billman and E. W. Kaler, Langmuir 7:160917 (1991).

37. T. Förster and H. Tesmann, Cosmetics Toiletries 106 (12):4952 (1991).


38. K. Shinoda and S. Friberg, Adv. Colloid Interface Sci. 4:281300 (1975).

39. H. Kunieda and K. Shinoda, Bull. Chem. Soc. Jpn. 55:177781 (1982).

40. R. Aveyard, B. P. Binks, and P. D. I. Fletcher, Langmuir 5:121017 (1989).

41. R. Aveyard, B. P. Binks, S. Clark, and P. D. I. Fletcher, Progr. Colloid Polymer Sci. 79:2027 (1989).
42. M. Kahlweit, R. Strey, and G. Busse, J. Phys. Chem. 94:388194 (1990).

43. R. Aveyard, B. P. Binks, and J. Mead, J. Chem. Soc., Faraday Trans. 1, 82:175570 (1986).

44. F. Schambil, F. Jost, and M. J. Schwuger, Progr. Colloid Polymer Sci. 73:3747 (1987).

45. T. Förster, F. Schambil, and W. von Rybinski, J. Disp. Sci. Techn. 13:183193 (1992).

46. R. Aveyard, B. P. Binks, and J. Mead, J. Chem. Soc., Faraday Trans. 1, 83:234757 (1987).
47. R. Aveyard, B. P. Binks, T. A. Lawless, and J. Mead, J. Chem. Soc., Faraday Trans. 1, 81:215568 (1985).

48. R. Aveyard, B. P. Binks, S. Clark, and J. Mead, J. Chem. Soc., Faraday Trans. 1, 82:12542 (1986).

49. J. S. Maryland and B. A. Mulley, J. Pharm. Pharmac. 24:72934 (1972).

50. F. Harusawa, T. Saito, H. Nakajima, and S. Fukushima, J. Colloid Interface Sci. 74:43540 (1980).
51. M. Tagawa, K. Shinozaki, Y. Tabata, and N. Ohba, J. Soc. Cosm. Chem. Japan. 17:4551 (1983).

52. H. Kunieda, K. Hanno, S. Yamaguchi, and K. Shinoda, J. Colloid Interface Sci. 107:12937 (1985).

53. H. Kunieda and M. Yamagata, Colloid Polymer Sci. 271:9971004 (1993).

54. H. Kunieda and N. Ishikawa, J. Colloid Interface Sci. 107:12228 (1985).

55. O. Ghosh and C. A. Miller, J. Colloid Interface Sci. 116:59397 (1987).


56. S. E. Friberg and C. Solans, Langmuir 2:12126 (1986).

57. S. Friberg, J. Soc. Cosmet. Chem. 30:30919 (1979).

58. T. Suzuki, H. Tsutsumi, and A. Ishida, J. Disp. Sci. Techn. 5:11941 (1984).

59. S. Fukushima and M. Yamaguchi, Cosmetic Toiletries 98 (5):89102 (1983).


60. N. Krog and K. Larsson, Chem. Phys. Lipids 2:12943 (1968).

61. N. Krog and A. P. Borup, J. Sci. Fd. Agric. 24:691701 (1973).

62. H. Junginger, C. Führer, J. Ziegenmeyer, and S. Friberg, J. Soc. Cosmet. Chem. 30:923 (1979).

63. T. de Vringer, J. G. H. Joosten, and H. Junginger, Colloid Polymer Sci. 262:5660 (1984).
Page 125

64. T. de Vringer, J. G. H. Joosten, and H. Junginger, Colloid Polymer Sci. 264:691700 (1986).
65. G. M. Eccleston, J. Soc. Cosmet. Chem. 41:122 (1990).

66. T. Förster, F. Schambil, and H. Tesmann, Int. J. Cosm. Sci. 12:21727 (1990).

67. H. Junginger, A. A. M. D. Akkermans, and W. Heering, J. Soc. Cosmet. Chem. 35:4557 (1984).

68. B. P. Binks, Colloids Surfaces 71:16772 (1993).


69. B. P. Binks, Langmuir 9:2528 (1993).

70. J. L. Salager, I. Loaiza-Maldonado, M. Minana-Perez, and F. Silva, J. Disp. Sci. Techn. 3:27992 (1982).

71. R. E. Anton, P. Castillo, and J. L. Salager, J. Disp. Sci. Techn. 7:31929 (1986).

72. M. Bourrel, J. Biais, P. Bothorel, B. Clin, and P. Lalanne, J. Disp. Sci. Techn. 12:53145 (1991).

73. W. C. Griffin, J. Soc. Cosmet. Chem. 5:24956 (1954).


74. H. Arai and K. Shinoda, J. Colloid Interface Sci. 25:396400 (1967).

75. M. Bourrel, A. Graciaa, R. S. Schechter, and W. H. Wade, J. Colloid Interface Sci. 72:16163 (1979).

76. M. Bourrel, J. L. Salager, R. S. Schechter, and W. H. Wade, J. Colloid Interface Sci. 75:45161 (1980).

77. J. L. Salager and R. E. Anton, J. Disp. Sci. Techn. 4:25373 (1983).


78. T. Förster, W. von Rybinski, H. Tesmann, and A. Wadle, Int. J. Cosmet. Sci. 16:8492 (1994).

79. T. Mitsui, Y. Machida, and F. Harusawa, Amer. Cosmet. Perfum. 87:3336 (1972).

80. K. Shinoda and H. Saito, J. Colloid Interface Sci. 30:25863 (1969).

81. S. Friberg and C. Solans, J. Colloid Interface Sci. 66:36768 (1978).

82. T. J. Lin, J. Soc. Cosmet. Chem. 29:74556 (1978).


83. T. J. Lin, T. Akabori, S. Tanaka, and K. Shimura, Cosmetics Toiletries 95 (12):3339 (1980).

84. T. J. Lin, T. Akabori, S. Tanaka, and K. Shimura, Cosmetics Toiletries 96 (6):3139 (1981).

85. H. Sagitani, J. Disp. Sci. Techn. 9:115129 (1988).

86. T. Suzuki, H. Takei, and S. Yamazaki, J. Colloid Interface Sci. 129:491500 (1989).
87. T. Suzuki, M. Nakamura, H. Sumida, and A. Shigeta, J. Soc. Cosmet. Chem. 43:2136 (1992).
Page 127

5
Emulsifier Selection/HLB
Donald L. Courtney, SR.
Emulsions REZ, Landenberg, Pennsylvania

I. Emulsifier Selection with the Aid of HLB 128

II. Required HLB of Blends of Lipids 131

III. Experimental Determination of Required HLB Number 131

IV. Published Data 132


V. Theoretical Validation of Griffin's HLB Values 132

A. J. T. Davies's Group Numbers 132

B. Alternate Determination of Required HLB 135

C. The PIT Method 135

VI. An Outstanding Pair of Emulsifiers 136


VII. Amount of Emulsifier 136

A. Required Stability 137

B. Difficulty of Ingredients to Be Emulsified 137

C. Amount of Oil Phase Ingredients 137


D. Efficiency of the Emulsifiers 137

E. Accuracy of the HLB Number Determination 137

F. Use of a Stabilizer 137

G. Preparation Method 137

References 138
Page 128

I
Emulsifier Selection with the Aid of HLB

When formulating an emulsion, the formulator has numerous choices to make in selecting the ingredients that will produce the
desired qualities the formulator has been called upon to develop. Many ingredients are quite predictable in their contribution to
the final product relative to viscosity, tactile qualities, appearance, rub-in, after-feel, irritation, odor, etc. An experienced
formulator can generally select and formulate ingredients that will approximate the desired properties in a few days time. But
selecting the system that will modify this complex mixture of ingredients into a stable, homogeneous, and elegant emulsion is
usually the most elusive task of the entire process. Of course we are referring to those extremely complex chemical mixtures
known as emulsifiers, and the search is more difficult because a pair works better than one. Besides, there are thousands from
which the formulator can choose. Selection of the proper emulsifiers for a specific blend of particular ingredients can be a
ponderous, confusing, and time-consuming process.
How does a formulator select emulsifiers for a brand new mixture of lipids? A popular method is to use those that have been
used successfully in a past product. This is a logical step since the compounder knows the raw materials; they are in inventory;
medical has approved them for use; and the supplier is a known and reliable source. The drawback is they probably won't work
satisfactorily in this new system.
A resourceful formulator will read suppliers' literature, technical magazines, and textbooks for alternate ideas. But what are the
chances of finding an existing formula identical to his? If he did he would change his formula. However, he may come up with
several new candidates that he tries with only modest success. Fortunately there is help. It is known simply as the HLB system.
In the 1940s, William C. Griffin pioneered the creation and manufacture of emulsifiers (surfactants). They were being
generated so fast that he was faced with the dilemma of determining their utility and how they related to each other. He was
primarily concerned with nonionics made from polyols (glycerin, propylene glycol, and sorbitol), fatty acids and alcohols, and
ethylene oxide adducts. At the time (1949) it was known that the greater the amount of polyol or ethylene oxide (EO) in the
molecule, the more hydrophilic or water soluble it would be. Conversely, the greater the proportion of fatty acid or fatty
alcohol, the more lipophilic or oil soluble the emulsifier would be. Griffin was looking for some way to classify or ''rate" these
products, and he ultimately reasoned that comparing the hydrophilic "strength" of each emulsifier would be a meaningful way
to relate their performance. If three emulsifiers had the same percent hydrophilic content they should orient themselves in a
micelle or on oil droplets in a similar fashion. Differences in the polyols and lipid moieties would affect emulsifying efficiency,
but performance should be similar.

Since the hydrophilic portion of these products can readily be calculated and measured, Griffin set out to evaluate his premise.
To simplify the system he divided the weight percent of the hydrophilic portion of the molecule by five and called the resulting
number the HLB number of the surfactant. He derived the following formulas for the calculation [1].

where S is saponification number of the ester and A is acid number of the recovered acid.

Page 129

When the hydrophilic portion of the molecule consists only of ethylene oxide, the following formula pertains [2].

where E is weight percent of oxyethylene content. The abbreviation HLB stands for Hydrophile Lipophile Balance, and
"Balance" refers to the relative portions of lipidic vs. hydrophilic segments.
Fortunately, this method of "rating" or classifying emulsifiers turned out to be quite useful and meaningful. A system for
organizing emulsifiers had been discovered and is summarized in Table 1.

The next step was to relate the HLB number of these emulsifiers to materials to be emulsifiedfats, oils, fatty acids, waxes,
lipids, etc. Not surprisingly (in hindsight), each material to be emulsified has a required HLB number, i.e., each requires an
emulsifier (or emulsifier blend) with a specific HLB number for optimum emulsification. This was an amazing and invaluable
discovery! Use of an emulsifier with an HLB number identical to the required HLB number of a lipid should yield a good
emulsionor at least a better one than any emulsifier with the same chemical components but a different HLB number. For
example, cetyl alcohol has a required HLB number of 15.5. An ethoxylated stearyl alcohol (emulsifier) with 21 moles of
ethylene oxide has an HLB of 15.5 and emulsifies cetyl alcohol better than stearyl alcohol containing 16, 18, 20, 22, or 25
moles of ethylene oxide.
TABLE 1 HLB of Various Surfactants
(ICI America Literature)
Surfactant HLB value
Sorbitan trioleate 1.8
Glyceryl oleate 2.8
Sorbitan oleate 4.3
Sorbitan stearate 4.7
Steareth-2 4.9
Laureth-4 9.7
PEG-8 Stearate 11.1
Nonoxynol-5 10.0
Nonoxynol-9 13.0
PEG-4 Sorbitan peroleate 9.0
PEG-25 Hydrogenated Castor Oil 10.8
Triethanolamine Oleate (TEA oleate) 12.0
Polysorbate 60 14.9
Polysorbate 80 15.0
PEG-40 Stearate 16.9
PEG-100 Stearate 18.8
Sodium Oleate 18.0
Potassium Oleate 20.0
Page 130

Once the required HLB of a material to be emulsified and the HLB number of emulsifiers are known, producing a good
emulsion can be accomplished easily and quickly. The remaining task is to find the correct chemical type of emulsifier at the
determined HLB value for the lipid to be emulsified. There is a best chemical type for each lipid.

What makes the HLB System even more useful to the formulator are its alligative relationships. It is well established that a
pair of emulsifiers makes a better emulsion than a single emulsifier. If the formulator selects a pair of emulsifiers so that one
has a higher and the other a lower HLB number than the required HLB of the lipid to be emulsified, they can be blended to the
exact required HLB number of the lipid.

The exact blend of the two emulsifiers is easily calculated since it is an arithmetic relationship. For example, if emulsifier A
has an HLB number of 5 and emulsifier B has an HLB number of 15, the 50/50 (weight) blend, has an HLB number of 10. The
following ratios of this pair of surfactants produce the resulting HLB values shown in Table 2.

If the required HLB of the lipid is 8.5, the following formula allows a simple calculation:

Proof:

TABLE 2 HLB Values from Surfactant Pair


Surfactant A: HLB 5 Surfactant B: HLB 15 Resulting HLB of blend
10% = 0.5 + 90% = 13.5 = 14.0
20% = 1.0 + 80% = 12.0 = 13.0
30% = 1.5 + 70% = 10.5 = 12.0
40% = 2.0 + 60% = 9.0 = 11.0
50% = 2.5 + 50% = 7.5 = 10.0
60% = 3.0 + 40% = 6.0 = 9.0
70% = 3.5 + 30% = 4.5 = 8.0
80% = 4.0 + 20% = 3.0 = 7.0
90% = 4.5 + 10% = 1.5 = 6.0
Page 131

The ability to blend emulsifiers to the correct required HLB is enormously important and valuable for the following three
reasons.

1. As stated earlier, two emulsifiers make a more stable emulsion than one, almost without exception. So it is desirable to use a
pair of emulsifiers.

2. By blending emulsifiers one can achieve the exact HLB number required. Even though there are thousands of emulsifiers
available, the number providing a specific HLB number can be quite small or nonexistent. Without blending, the formulator
would have to have an enormous number of emulsifier samples on hand for optimum emulsification of the numerous lipids
used in formulating.
3. The chemical composition of the emulsifier plays an important role in its efficiency for specific lipids. For example, oleic
acid derivatives are very good for emulsifying white oils (mineral oil). However, there may not be any available with the exact
required HLB of a given white oil.

For your convenience, a table of emulsifier HLBs (Table 1) is included herein.

II
Required HLB of Blends of Lipids

Cosmetic formulators are noted for using a myriad of lipid components to produce the desired elegance. So far we have dealt
with the required HLB number of a single lipid. How does the HLB system cope with a combination of lipids? Quite simplythe
combination of the lipids, no matter how complex, has a required HLB! This can be calculated quite readily, or better yet,
determined experimentally. It is calculated as follows [3]:
Calculation of Required HLB Number of a Complex Mixture
Ingredient Required HLB number % Oil phase HLB contribution
A 12.0 × 30 = 3.6
B 6.0 × 50 = 3.0
C 15.0 × 20 = 3.0
Total 9.6

Thus 9.6 is the required HLB number of the blend.

III
Experimental Determination of Required HLB Number
The following procedure is followed to determine the required HLB number of a single lipid or complex mixture of lipids.

Any pair of emulsifiers can be used, but one should have a low HLB number (i.e. < 6) and the other a high HLB number (i.e. >
14). For this illustration, the pair consisting of Surfactant A and Surfactant B in Table 2 is selected. A series of emulsions is
prepared with each blend using a total of 2% surfactant, 20% lipid and 78% water. The surfactant and the oil are blended in
one container and water is added. Simple uniform shaking is adequate (at least ten shakes). If either the surfactant or lipid is a
solid, it is melted and warm water is used so that the system is liquid.
Page 132

Since the required HLB number is not known, emulsions are made with blends ranging from 10/90 (A/B) to 90/10 (A/B). The
resulting emulsions are allowed to stand and are observed over a period of several hours. Ideally most, if not all, will separate
in varying degrees. The one with the least separation is the required HLB number; that is, if the best emulsion is the 50/50
blend, the required HLB number is 10.0 [4].
If the required HLB had been known in advance to be somewhere around 10.0, one would only have to make up five
emulsionstwo with lower required HLB values (8 and 9) two with higher (11 and 12) and one at 10.

If all the emulsions broke rapidly and equally, the process is repeated using 3% emulsifier or only 10% lipids. If the initial
emulsions are creams, and thus hindered from separating, the procedure is repeated using reduced levels of lipids. If HLB 10
and HLB 11 yield equally stable emulsions, one may assume that the required HLB number is 10.5.

Now one needs to find that optimum chemical type emulsifiers for the chosen lipid(s). (Reference 4 includes specific
recommendations.) Surfactant types can be selected as desired. Ordinarily, stearic acid derivatives and stearyl alcohol
derivatives are superior emulsifiers and are good choices.
Pairs are chosen so that one has an HLB higher than 10 and one an HLB lower than 10. Each pair is blended to HLB (the
required HLB number just determined), and emulsions are made with each. This time a mechanical stirrer is used to make each
system as consistent as possible. The emulsion showing the least separation contains the chemical type to pursue.

It is only necessary to make one emulsion with each pair of emulsifiers, as each pair blended to HLB 10 should yield a better
emulsion than any other blend. This is an important, time-saving benefit of the HLB System. When the best pair of emulsifiers
has been identified, it would be judicious to finalize the system by comparing HLB 9, 10, and 11 due to slight variations
among chemical types. The HLB system is not as accurate as titrating an acid with a baseGriffin originally claimed an
accuracy of ± 1 unit [1].

IV
Published Data

Fortunately, the required HLB numbers of many common lipids have been determined and published [4]. Following is an
excerpt of an extensive list of required HLB numbers of widely used ingredients (Table 3).

V
Theoretical Validation of Griffin's HLB Values
The concept put forth by W. C. Griffin in 1949 has persisted so long for one reasonit works! Many formulators can attest to
this fact because of years of personal use. (The author has had great success with it during thirty-five years of formulation
work.) However, inquisitive theorists have scrutinized its concept from numerous approaches to validate it or to improve it.
Validation has been successful, improvement is probably marginal. Some of these efforts will be reviewed.
A
J. T. Davies's Group Numbers

Davies attempted to explain Griffin's HLB numbers (shown in Table 4) thermodynamically and by relating it to structural
groups [6]. He produced the following equation to explain his concept:
Page 133

TABLE 3 Required HLB Numbers for O/W Emulsions


Material Required HLB
Acetylated lanolin 14
Acid, isostearic 1516
Acid, oleic 17
Alcohol, cetyl 1516
Alcohol, isohexadecyl 1112
Alcohol, stearyl 1516
Arlamol, C12C15 benzoate 13
Arlamol E 7
Beeswax 9
Caprylic/Capric Triglyceride 5
Carnauba wax 15
Castor oil 14
Cocoa butter 6
Coconut oil 5
Corn oil 6
Cottonseed oil 56
Dimethyl silicone 9
d-Limonene (varies widely) 67
Isopropyl myristate 1112
Isopropyl palmitate 1112
Jojoba oil 67
Lanolin, anhydrous 9
Mineral oil light (naphthenic) 1112
Mineral oil light (paraffinic) 1011
Mineral spirits 14
Mink oil 5
Nonyl phenol 14
Paraffin wax 10
Petrolatum 78
Pine oil 16
Silicone oil 5
Silicone oil (volatile) 78
Soybean oil 6
Vitamin A plamitate 6
Vitamin E 6
Source: Ref. 1.
Page 134

TABLE 4 Davies's HLB Group Numbers


Hydrophilic group Numbers
SO4Na 38.7
COOK 21.1
COONa 19.1
N (tertiary amine) 9.4
Ester (sorbitan ring) 6.8
Ester (free) 2.4
COOH 2.1
OH 1.9
O 1.3
OH (sorbitan ring) 0.5
Lipophilic Groups
CH 0.475s

Derived Group
(CH2CH2O) 0.33
0.15

Becher notes, "Unfortunately, HLB values for ethoxylates calculated by this method are erroneous, since this method assumes
that each additional EO group adds the same increment to the HLB." [6]
As Hans Schott commented, "Since Griffin validated his HLB scale for nonionic emulsifiers by measuring optimum
emulsification for a wide variety of lipids with water it should be the scale of choice. The Davies' scale seriously distorts the
range of the experimental HLB values required for emulsifying lipids in water and vice versa. Therefore it is not suited to
select nonionic emulsifiers."[7]
Becher identified several alternate methods for HLB determination [6]:

Spreading coefficient
Polarity index
Polarography
NMR and mass spectroscopy
Calorimetry
Dielectric concept
Page 135

Partition coefficient
Heat of (interfacial) adsorption
Solubility parameter
Partial molar volume
Monolayer properties
Critical micelle concentration
Water number, cloud point
Phase Inversion TemperaturePIT
Interfacial tension
Foaming

B
Alternate Determination of Required HLB

Vaughan and Rice [8] have recently developed a method for determining the required HLB values of lipids based on their
solubility parameters using the following equation:

The solubility parameter is derived from molecular weight, specific gravity, and boiling point, which are readily available for
numerous compounds. This allows one to predict the required HLB of lipids without experimentation.

Required HLB values calculated by this technique generally have surprisingly good correlation with the Griffin method for
linear emulsion ingredients. However, for chain-folded structures there is considerable deviation.
C
The PIT Method

The phase inversion temperature or HLB temperature has intrigued many physical chemists interested in producing ultimate
emulsions. It is based on the fact that o/w emulsions when heated reach a temperature at which they suddenly invert to w/o
emulsionsthe PIT. This concept was announced by Shinoda and Arai in 1964. It is claimed that an emulsion should be formed
(phases mixed) at 24°C below the PIT. This allows an o/w emulsion to form initially yielding smaller particle sizes than if the
emulsion is made above the PIT and inverts from w/o to o/w on cooling [9]. See Chap. 4, pages 120122, for additional
comments.

This property can also be used to assist in selecting the optimum chemical type of emulsifier by making a series of emulsions
with various emulsifiers and observing the viscosity changes as water is added to the oil and as the emulsion inverts from w/o
to o/w. If the system passes from a low viscosity w/o system through a viscoelastic gel stage while inverting to o/w and then to
a fluid emulsion, the resulting emulsion should have good stability [9]. It appears that PIT augments the HLB system and is
used in conjunction with it.

The method involves a substantial amount of test work that may yield a more accurate HLB value for the emulsion. The
method implies that the mixing temperature of the phases may play a more important role in the stability of an emulsion than
previously thought and that determination of the optimum emulsification temperature may be of value. The author has not used
this method as an adjunct to the HLB system as described
Page 136

earlier but has produced hundreds of excellent emulsions. It is his belief that PIT is not widely used.
It has been noted that the phase inversion temperature is assigned to the particular emulsion system and accounts for oil type,
electrolyte concentration in the aqueous phase and other factors which influence surfactant partitioning at the o/w interface,
whereas HLB is assigned to the surfactant molecule. For homogeneous nonionic surfactants (i.e. pure or single specie
products) there is a linear relationship between HLB and PIT:

where Koil is approximately 17°C/HLB unit NHLB is Griffin HLB number, and Noil is the HLB value of the oil at THLB = 0°C [9].

VI
An Outstanding Pair of Emulsifiers

The author has had great success with a pair of emulsifiers for a wide variety of lipids and mixtures of lipids. The formulator is
urged to evaluate this pair when trying to determine the correct chemical type of emulsifier. This pair is steareth-2 (Brij 72)
(HLB 4.9) and steareth-21 (Brij 721) (HLB 15.5) [5]. For the sake of convenience the following blends yield HLB values from
5 to 15.
% Steareth-2 % Steareth-21 HLB
100 0 5
90 10 6
80 20 7
70 30 8
60 40 9
50 50 10
43 57 11
33 67 12
23 77 13
15 85 14
5 95 15

VII
Amount of Emulsifier

The amount of emulsifier needed in a formula varies appreciably from formula to formula with no hard and fast rules. The
variable factors governing the amount to use include:

ARequired stability

BDifficulty of ingredients to be emulsified

CAmount of oil phase ingredients


DEfficiency of the emulsifiers

EAccuracy of the HLB number determination

FUse of a stabilizer

GPreparation method

Fortunately, guidelines can be followed to make a reasonable estimate of the amount,


Page 137

and actual determination is not a Herculean task. Some of the factors involved are reviewed below.
A
Required Stability

Very simply, the more severe the stability requirements of the formula the greater the amount of emulsifier needed; the less
restrictive the requirements the lower is the level of required emulsifier. Ordinarily manufacturers will require two to three
years' stability of a formula at room temperature. Ninety days' stability at 40°C is a reasonable high-temperature requirement.
One must also consider the stability at 50°C since this temperature is easily reached in the real world. A month's protection
against separation should be adequate; and some will accept less. Low-temperature stability is also important. Four or five
freeze-thaw cycles are normal, and refrigerator temperatures are also important.

B
Difficulty of Ingredients to Be Emulsified

The difficulty of the lipid to be emulsified and the amount are very important, but efficient emulsifiers (previously discussed)
should prevail.

C
Amount of Oil Phase Ingredients

The amount of lipids used is directly related to the amount of emulsifier required although the relationship is not linear.
D
Efficiency of the Emulsifiers

Emulsifiers can vary significantly in efficiency, i.e., the amount required to produce the desired stability. Normally a
formulator would want to use a highly efficient pair. Usually stearic acid and stearyl alcohol derived emulsifiers are superior.
E
Accuracy of the HLB Number Determination

The correct HLB number is very important as discussed above.

F
Use of a Stabilizer

Sometimes emulsifiers alone will not yield the desired stability. With nonionic emulsifiers, Carbomer 934 at levels of 0.1% to
0.5% usually will help achieve the desired stability.

G
Preparation Method

Good preparation would include adding the water phase to the oil phase (containing the emulsifiers) to provide inversion and,
possibly homogenization. Having said this, a minimum of 2% emulsifier is probably required. It should be rare to exceed 5%,
but this is not unheard of. For example, if white oil is considered a benchmark, 3% oil could probably be emulsified with 2%
emulsifier; 2050% white oil may require 45% emulsifier. The author has actually emulsified 70% oil with 7% emulsifier
resulting in an extremely stable emulsion. Obviously, a straight-line relationship does
Page 138

not exist. A suggested procedure is to use 3 or 4% emulsifier with and without Carbomer 934 for normal lipid levels. This
assumes, of course, that very efficient emulsifiers are used.

In closing:
If emulsion problems keep you awake
Because of the speed with which they break,
Say a prayer on bended knee,
And quickly check your HLB!

References

1. The HLB System: A Time-Saving Guide to Emulsifier Selection, ICI Americas, Wilmington, DE, 1992, p. 19.

2. The HLB System: A Time-Saving Guide to Emulsifier Selection, ICI Americas, Wilmington, DE, 1992.

3. The HLB System: A Time-Saving Guide to Emulsifier Selection, ICI Americas, Wilmington, DE, 1992.
4. The HLB System, (51-0010-304), June 1990, ICI Americas, Wilmington, DE.

5. Brij 721 Polyoxyethylene 21 Stearyl Ether, (51-0001-228), Revised 1988, ICI Americas, Wilmington, DE.

6. J. Becher, J. Dispersion Science and Technology, 5:8196 (1984).

7. H. Schott, Journal of Pharmaceutical Science/87, 79:(1990).

8. Vaughan and Rice, J. Dispersion Science and Technology 11:8391 (1990).


9. M. Johnston, in Surfactant Technology, ICI Australia.

10. K. Shinoda and S. Friberg, Emulsions and Solubilizations, John Wiley, New York, 1986, pp. 6, 13032.
Page 139

6
Multiple Emulsions in Cosmetics
Monique Seiller and Francis Puisieux
Physico-Chimie-Pharmacotechnie-Biopharmacie, Université de Paris-Sud, Châtenay-Malabry, France

J. L. Grossiord
Physique Pharmaceutique, Université de Paris-Sud, Châtenay-Malabry, France

I. Introduction 139

II. Structure and Composition 140


III. Preparative Method 141

IV. Characterization 144

A. Microscopic and Particle Size Analysis 144

B. Rheological Analyses 145


C. Titration of a Tracer or of an Active Substance 147

V. Stability 148

VI. Release of the Active Substance 149

VII. Dermatological Applications 150

VIII. Conclusions 153


References 153

I
Introduction

As a result of their structure, multiple emulsions, i.e., emulsions of emulsions, can contain active water-soluble and lipid-
soluble substances in each of the three constituent phases. Although potentially advantageous as systems for application to the
skin and mucous membranes, multiple emulsions are not currently used as therapeutic dosage forms. This is due to the fact that
such systemsinvestigated for only about ten yearsare new vesicular forms that are still not fully controllable in terms of
formulation, manufacture, characterization, and performance after application to skin or mucous membranes.
Page 140

Yet there are reasons to believe that these limitations and stability problems may be overcome shortly and that considerable use
will be made of this dosage form in the coming years, particularly if its ability to prolong the release of active ingredients is
confirmed.

The use of multiple emulsions could become nearly as widespread as the use of simple emulsions from which they are derived
and may equal that of other similar vesicular systems. One argument in their favor is that they exhibit approximately the same
properties as these better known systems, and like simple emulsions, they can be dispensed directly to the skin because they
are oily creams with the proper consistency for easy spreading.

II
Structure and Composition

Multiple emulsions are emulsions in which a dispersed phase contains another dispersed phase [1]. Thus a w/o/w emulsion is a
system in which water globules are dispersed in oil globules, the latter being themselves dispersed in an aqueous phase (Fig.
1). By analogy, multiple emulsions of the o/w/o type exist, in which an internal oily phase is dispersed in the aqueous globules,
which in turn are dispersed in an external oily phase. The emulsions described here are chiefly w/o/w emulsions with improved
production and application properties for pharmaceuticals and cosmetics.

Multiple emulsions are sometimes called three-phase emulsions or triple-phase emulsions. In fact, the term double emulsion
would be more appropriate, because two emulsions coexist in these systems: an emulsion with a continuous oily phase and an
emulsion with a continuous aqueous phase. The term triple emulsion is nonetheless acceptable and sanctioned by custom. Yet
it is not recommended, although often used, when it applies to miscellaneous preparations consisting, for example, of a gelled
simple emulsion or an emulsion containing microparticles in suspension, or even a simple emulsion in which a different active
substance is incorporated into each phase (e.g., retinyl palmitate in the oily phase and urea and aloe extract in the aqueous
phases).

Since multiple emulsions consist of at least two immiscible liquids, their preparation demands the presence of emulsifiers,
which are usually synthetic but sometimes are of

Fig. 1
Oily globule in w/o/multiple emulsions.
Page 141

natural origin. At least two emulsifiers are used. They are called the primary surfactant (S1) and the secondary surfactant
(S11). For multiple w/o/w emulsions, the molecules of S1, with a lipophilic tendency, are oriented at the internal w/o interface
and those of S11, with a hydrophilic tendency, at the w/o external interface. If the concentrations and the types of surfactants
are properly adapted to that of the oily phase, they give rise to two monomolecular films. The apolar parts of the emulsifiers
are found in the oil, and the polar parts are found in the internal (S1) or external (S11) aqueous phase. It follows that the
emulsifiers exist in oriented layers, forming an actual envelope of the vesicle around the oil.

Although some surfactants and oily phases are better adapted than others, all normal components of simple emulsions can be
used to obtain multiple emulsions. For this purpose certain conditions must be satisfied concerning the ratio between the
concentration of S1 and S11 or of the maximum HLB of their mixture.

The oily phases most often used to form multiple w/o/w emulsions arein decreasing order of usehydrocarbons such as liquid
paraffin, triglycerides (mostly vegetable oils), followed by esters, fatty alcohols or acids, and silicones. Emulsifiers are
primarily nonionic surfactants. Among the most effective for S1 are esters of sorbitan with a long hydrocarbon chain, perfluoro
derivatives, and, above all, polymeric surfactants such as cetyl dimethicone copolyol. For S11, the most effective are
polyethoxylated esters of sorbitan, copolymers of ethylene oxide and propylene oxide, highly ethoxylated fatty acids, and
polyglyceryl condensates.
As in other vesicular systems, various additives are often introduced. At very low concentrations, they play the role of markers,
making it possible to quantify the stability of the systems. They include electrolytes (NaCl), sugars (glucose), and fluorescent
substances (carboxyfluorescein). In higher concentrations, some additives tend to increase stability. They may include
hydrophilic polymers (xanthan gum, alginates, cellulose derivatives, and carboxyvinyl compounds) incorporated into one of
the aqueous phasesbut usually in the external aqueous phaseor lipophilic substances (waxes, fatty acids or alcohols, silicone
derivatives) introduced into the intermediate oily phase.

The following is an example of the formula of a typical w/o/w multiple emulsion.

Primary emulsion
almond oil 24.0%
purified soy lecithin 5.0%
magnesium sulphate 0.7%
deionized water q.s.p. 100.0%

Multiple emulsion
Primary emulsion 70.0%
polysorbate 60 1.0%
polysorbate 80 1.0%
deionized water q.s.p. 100.0%

III
Preparative Method

The procedure is crucial for obtaining a multiple emulsion and is even more important than that for a simple emulsion. Four
main procedures can be applied.
Page 142

The two-step process, as shown in Fig. 2. The most common method. Note that, strictly speaking, this name is inaccurate,
because the other processes also demand a double manufacturing step.

The phase inversion process, as shown in Fig. 3. This process is increasingly used. This name is also inappropriate because it
does not involve a true phase inversion.

Lamellar phase dispersion process, as shown in Fig. 4. This method is very rarely used. It is comparable to one of the
processes employed to obtain vesicles with nonionic surfactants.

The oily isotrope dispersion process. According to the authors, who recommend it, this process is more appropriate for o/w/o
type microemulsions.

These four methods are applied in a virtually identical manner.


Initially, a simple emulsion with an oily continuous phase (for the first two processes) or the lamellar phase or the oily isotrope
(for the latter two processes) is prepared at 70 to 80°C. Whatever the system involved, the water, oil and emulsifiers, whose
proportions vary according to the method employed, are mixed using a standard turbine agitator for about 30 min with high
speed of about 103 rpm.

In the second phase, for the two-step procedure, the emulsion containing an oily continuous phase is slowly transferred into
the aqueous phase. For the remaining three processes, the water is gradually introduced, either into the emulsion with an oily
continuous phase, into the lamellar phase, or into the oily isotrope. The second dispersion is also performed with a turbine
agitator, usually at ambient temperature for about 30 min but at a lower spread of rotation of a few hundred rpm.

Each process obviously has its own advantages and drawbacks.

Fig. 2
Two step emulsification method.
Page 143

Fig. 3
Phase inversion method.

The main advantage of the two-step process is fully controlled application. It is possible, at least in theory, to control the
quantity of internal water. Its disadvantage is that it is not reproducible, if only because of the second emulsification is a
critical step. In fact, the primary w/o emulsion, which is highly viscous, is difficult to disperse. During this step, which
requires intense shear forces, there is risk of breakage of some newly formed oil globules, resulting in the mixing of part of the
internal water with the external water.
The advantage of the phase-inversion process is that it is very easy to apply and provides a specific proportion of internal
aqueous phase exactly as in the two-step process. Unlike the latter, it is also reproducible. Yet it has the drawback of requiring
two emulsification steps and, above all, of being difficult to implement. In fact, a very slight excess of water suffices to
transform the multiple emulsion into a simple emulsion of the aqueous type (w/o). On the other hand, excessively slow
incorporation of water causes the formation of a simple emulsion of the oily type (o/w).

Dispersion of a lamellar phase in water offers the advantage of requiring a single emulsification step. The initial phase formed
by a concentrated surfactant solution is thermodynamically stable and can be obtained rapidly. One limitation of this process
stems from the fact that not all surfactants form a lamellar phase. If this phase exists,

Fig. 4
Ternary phase diagram.
Page 144

the HLB of the surfactants is often high, which is disadvantageous for the stability of a multiple emulsion. The quantity of
oil incorporated in the lamellar phase is always low, rarely higher than 10%.

The process by dispersion of an oily isotrope in water offers approximately the same advantages as the above process. Only
one emulsification is required. The initial phase is a pure phase, stable and simple to obtain. The main drawback is the low
level of solubilized water in the inverse micelles, which can be raised above 10% only with difficulty.

The latter two processes also present common drawbacks resulting from the necessarily large quantities of surfactant
employed. Water is added to the lamellar phase or to the oily isotrope, in which the initial water is not truly emulsified in the
form of globules. It then is difficult to determine whether the specific quantity of water, to which this dispersion gives rise, is
still solubilized and to determine the total amount of this water in the internal phase.

Also worth mentioning is another process recommended by Kavaliunas et al. [6]. It calls for the mixing, in preset proportions,
of an oily isotrope, an aqueous isotrope, and a lamellar phase. While this process appears attractive, it is rarely applicable
because it requires constituents that can yield these three phases simultaneously.

IV
Characterization

It is difficult to include all available physicochemical analyses for the characterization of multiple emulsions; these analyses
can be applied upstream to optimize formulation and manufacture. This section, therefore, is limited to a discussion of methods
for characterizing multiple emulsions from the time they are formed and during their aging. Most of the assays are
microscopic, particle-size, and rheological analyses as well as analytical determinations of tracers and of encapsulated active
substances.
A
Microscopic and Particle Size Analysis [7,8]

Microscopic examination is the first test performed to identify the systems obtained. It also offers an excellent means to
observe the stability of these systems during aging.
Optical microscopy is a standard analytical method for checking the multiplicity of the systems, as well as the particle size
distribution. It allows direct measurement of the size of multiple globules greater than 0.5µm in diameter as well as an
evaluation of the ratio of multiple globules to single globules. This method also provides an idea of the size of the internal
aqueous globules, often about 1µm in size (Fig. 5). Before observation and measurement of the actual size of the globules, the
multiple emulsion must be extensively diluted to yield a solution that is iso-osmotic with the internal aqueous phase. This is
required because the migration of water from the external phase to the internal phasein case of dilution with a solution of lower
osmolaritymight cause swelling, sometimes followed by bursting of the oily globules. Conversely, dilution with a solution of
higher osmolarity might cause loss of internal water and hence a contraction of the internal aqueous droplets.

Examination with polarized light under crossed Nicols sometimes helps to identify a
Page 145

Fig. 5
Optical microscopy photograph of a multiple emulsion.

texture corresponding to a lamellar phase, which indicates the orientation assumed by the two emulsifiers in the oily phase.

After freeze etching of the multiple emulsion, microscopy is sometimes used to visualize the rupture of the oily globules (Fig.
6). Rather than attempting a direct estimation after observation, some authors prefer to take measurements from photographs in
order to determine the average particle-size distribution and the size dispersion on a sufficiently large sample. To obtain a still
more accurate quantitative characterization of the sizes of the multiple oily globules, use is often made of particle-size analyses
by particle counting or by the diffusion or diffraction of light. Observations of this type require high dilution of the multiple
emulsions, whose osmolarity should be adjusted as closely as possible to that of the internal aqueous phase to avoid disturbing
the size of the globules.

B
Rheological Analyses [913]

The descriptive power and the versatility of rheological analyses are systematically employed to characterize multiple
emulsions. In fact, their wide scope makes it possible to

identify the multiple emulsion by using nondestructive viscoelastic analysis (linear microshear regime), which gives an
accurate signature of the structure at rest (Fig. 7);

simulate by oscillatory shear the conditions of skin application and to characterize its effect over time;

cause accelerated aging by the application of intense shear forces and to identify a number of changes in multiple emulsions
(flocculation, fracture, phase inversion)

Page 146

Fig. 6
Freeze etching photograph of a multiple emulsion.

the effects of skin application or artificial aging can be quantified by dynamic viscoelastic analysis (Fig. 7) or by the use of
steady-state flow tests (Fig. 8); and

characterize the kinetics of swelling and subsequent breakdown of the oil globules. Recording the changes in viscosity over
time and correlating swelling with the volume fraction of the dispersed phase (w/o primary emulsion), serves to quantify these
kinetics (Fig. 9).

Fig. 7
Oscillatory viscoelastic strain sweep.
Page 147

Fig. 8
Examples of steady state flow rheograms without (a) and with (b) hysteresis loop.

C
Titration of a Tracer or of an Active Substance

The titration of a compound (tracer or active substance), initially incorporated into the internal aqueous phase, assesses the
amount of the compound that remains in the internal phase at the time of measurement by comparison to the theoretical
quantity in the encapsulated mass. This determination thus serves to

determine not only the encapsulation yield of production but also the analysis of the stability of the multiple emulsions over
timeprovided the compound determined is sufficiently water-soluble to prevent its diffusion across the oily membrane (for
example, strong electrolytes); and
observe the kinetics of release of the compound by bursting of the oily globules and/or

Fig. 9
Characterization of a swelling-breakdown kinetics by recording of the viscosity.
Page 148

by diffusion across the oily membrane (if permitted by the lipophilic character of the compound).
The tracer is assayed either directly in the multiple emulsion (in the external aqueous phase after removal of the primary
emulsion) or in the dialysis medium (after dialysis of the multiple emulsion).

Analysis of the Multiple Emulsion

Various ionized substances can be titrated in the multiple emulsion itself. Magdassi et al. [14], for example, used a sodium salt
or an ephedrine salt. The chloride ions were determined by the mercuration method or by potentiometric titration directly in
the multiple emulsion.

Analysis of the External Aqueous Phase

Davis et al. [15] encapsulated carboxyfluorescein in the internal aqueous phase. The external aqueous phase was separated by
filtration after creaming of the primary emulsion. The marker was then detected in the clear external aqueous phase by
fluorescence spectrophotometry.

Fukushima et al. [16] and Ratz and Cueman [17] used cytabarine and dextran blue or polyporphyrin. After separating the
external aqueous phase by sedimentation of the oily globules (with or without centrifugation of the multiple emulsion) the
concentrations of each substance that passed into the external phase was measured by spectrophotometry.

Adeyeye and Price [18] used sodium salicylate. The multiple emulsion was diluted in the dispersant aqueous phase to which a
small amount of hydrophilic surfactant had been added. The mixture was allowed to rest until the supernatant oily layer could
be aspirated. The conductivity was then determined in the external phase, and its variation over time was measured at regular
intervals.

Analysis of the Dialysis Medium

Matsumoto et al. [19] employed the encapsulation of glucose. A certain amount of emulsion was dialysed against a volume of
distilled water. The migrated glucose was then assayed by measuring its reducing power. Matsumoto et al. [20] also used
sodium chloride. It was titrated by conductimetric measurements in the dialysis medium.

Obviously, spectrophotometry continues to be the most common means of titration for identifying the release of the
encapsulated active substance after dialysis. Thus, for example, Law et al. [21], using sulphane blue (C.I. No. 53430),
Omotosho et al. [2224], using fluorouracil, methotrexate and chloroquine, and Fredo-Kumbaradzi et al. [25], using sodium
sulphacetamide, resorted to these types of titrations.

V
Stability

Since multiple emulsions are thermodynamically unstable systems, they unavoidably change over time until they break. The
first sources of instability of w/o/w multiple emulsions include those that are common to all dispersed systems: creaming,
flocculation and coalescence of the oily globules, and phase inversion. In addition, multiple emulsions may also exhibit
specific instability. Besides the coalescence of the internal aqueous globules, it has already been pointed out that, in the
presence of an osmotic gradient, an aqueous flow from the internal phase to the external phase or a reverse flow may occur,
leading to the formation of a simple w/o or o/w emulsion.
At the present time, the use of high-performance polymeric surfactants, thickeners that considerably increase the viscosity of
the different phases, and cross-linking of the
Page 149

intermediate oily phase by various methods [26,27] has helped to retard the destabilization of these systems. Stable multiple
emulsions have actually been stored for two years at room temperature and for six months at 40°C without any change in their
characteristics or in the release of the encapsulated substance [1].

VI
Release of the Active Substance

The release mechanisms of an active substance encapsulated in the internal aqueous phase are complex, difficult to control,
and not fully understood. Two main mechanisms are usually considered. The first, which is the more likely after skin
application, is mixing of the internal phaseand hence of the active substancewith the external phase, caused by the breakage of
the oily membrane separating them. The second is due to diffusion of the active substance across this membrane.

It has already been stated that breakage of the intermediate membrane can occur under the effect of an osmotic gradient
between the two aqueous phases. This takes place after the globules attain a critical size due to swelling. This swelling is
caused by the aqueous flow from the external phase to the internal phase, generated by the concentration gradient of all the
soluble species existing in the two phases. Breaking can also be caused by sufficiently intense shear, particularly during skin
application.

The diffusion mechanism essentially concerns relatively lipid-soluble substances that have a significantly high oil/water
partition coefficient. For this type of substance, diffusion is passive and is governed by Fick's law. It is also possible to
consider the diffusion of highly water-soluble compounds due to the lipophilic surfactants present in the intermediate oily
phase, which could act as carriers. For example, inverse micelles could solubilize the water and thus transport a part of the
active substance from the internal phase to the external phase via facilitated diffusion. Activated diffusion does not appear to
occur; such a mechanism would require the action of carriers of different affinities at the two interfaces.

The behavior of a multiple emulsion after application to the skin is a conjectural question. Two assumptions can be made. (1)
The water evaporates before breaking or diffusion becomes apparent. In this case, the partitioning of drugs between oil and
skin predominates, a behavior similar to that observed with simple emulsions. (2) The evaporation takes place slowly enough
to allow breakage or diffusion. Two principal mechanisms may be operative.

1. The encapsulated active substance displays virtually no diffusion across the oily membrane. The active substance is then
only released from the internal phase by breaking of the multiple oil globules. This rupture takes place either by shearing of the
preparation or after swelling of the internal aqueous globules. Shear could be caused by massage or by rubbing during the
spreading of the cream.

Swelling could be caused by dilution of the multiple emulsion with water. In fact, during administration it is conceivable that
the user is required to blend a given dose of the multiple emulsion in a certain volume of water just before application. This
blending could even take place extemporaneously by means of dual compartment packaging. Due to the osmotic gradient thus
created or enhanced, and according to a mechanism already analyzed, an aqueous flow is then directed towards the internal
Page 150

phase, causing swelling of the internal water globules. When these globules reach their maximum size, they burst.
In both casesbreakage by shear or breakage after swellingthe active substance, which is now present in the external aqueous
phase, is immediately available. The multiple emulsion thus behaves as a simple emulsion with an aqueous continuous phase.
In this case, the chief benefit of the multiple emulsion is to protect an active substance by encapsulation or to permit the
introduction, in a single preparation, of active substances that will come into contact only when the cream is used.

2. The encapsulated active substance diffuses across the oily membrane, and when it is applied to the skin the multiple
emulsion preserves its multiple globules intact. The release of the active substance from the multiple emulsion is then no
longer identical to that of a simple emulsion. It is gradual and progressive and depends specifically on:
the type and thickness of the oily membrane,

the possible existence of paracrystalline phases in the interfacial film, especially lamellar phases obtained from surfactant
bilayers,

the type and concentration of the surfactants,


the partition coefficient of the active substance and its diffusion rate across the oily membrane, and

the particle size distribution of the oily globules.

The advantage of a multiple emulsion accordingly resides in its ability to delay the release of the encapsulated active substance
in comparison with simple emulsions.

VII
Dermatological Applications

To date, dermatological applications of multiple emulsions for the administration of active substances have been considered
only rarely and have been assessed infrequently. At this time, only a limited number of investigations have been published on
the subject.

The first study employing the application of multiple emulsions by a localized method was that of Attia et al. in 1986 [28]. It
concerned the release of pilocarpine hydrochloride as a myotic agent in eye infections. The active ingredient was added to
water or to castor oilpossibly containing a sorbitan ester as a surfactantin o/w or w/o emulsions and in w/o/w and o/w/o
multiple emulsions. Attia et al. conducted tests in vitro and in vivo.

For the in vitro tests, performed by dialysis, the authors showed that when the active ingredient was initially in the internal oily
phase of the o/w/o emulsion, release was approximately as fast as that from the o/w primary emulsion but slower than that of
the oily suspension containing the surfactant and that of the aqueous solution.

For the in vivo tests, based on the diameter of the pupil and the intraocular pressure, the authors made the following
observations: the shortest response times occur when the active substance is contained in the aqueous solution. Conversely, the
longest response times occur when the active substance is included in the o/w/o emulsion and in the oily suspension (1.5 times
more for the myotic response and 3.5 times more for the intraocular pressure). The active substance must first cross the o/w
interface and then the external oily phase before being released in the conjunctiva. According to the authors, these results are
caused either by a salting out effectprovoked by the presence of sodium chloride in the emulsion, causing a precipitation of the
active substanceor to prolonged
Page 151

contact of the emulsion with the corneal surfacecausing in situ sedimentation of the active substance.
The second study was a skin application conducted by Kundu Subhass in 1990 [29] with w/o/w and o/w/o types of multiple
emulsions. These consisted of liquid paraffin and sorbitan esters and contained either sodium or methyl salicylate or salicylic
acid as tracers. These systems were investigated in vitro in a Franz cell equipped with a semipermeable membrance of
cellulose acetate and in vivo after application to pig abdomens. The percentage of tracer passing into the receiving medium or
into the blood of the animals was measured by ultraviolet spectrophotometry.

In vitro, the author showed that the active substances are released more slowly from multiple emulsions than from pure
solutions containing the same substances. This release was even slower if thickenerssuch as silica, xanthan gum, or a
carboxyvinyl polymerwere added to the external phase of the w/o/w emulsions. This effect was especially pronounced if their
concentration is high. Similarly, if sodium choride was incorporated in the internal aqueous phase, the release of the active
substance was also modified. The release of the encapsulated tracer in the internal aqueous phase was characterized by the
existence of a linear relationship between the quantity released and the square root of time. For the w/o/w emulsions
investigated, the author concluded that the transport of the tracer across the external aqueous phase was the limiting factor.

In the case of o/w/o type of emulsions, the release of the active substances is also delayed if the carboxyvinyl derivative is
introduced into the intermediate aqueous phase. In this type of system, the limiting factor appears to be the water/oil interface.
In vivo, the author showed that, regardless of the type of emulsion, no systemic permeation of the active substances appeared
after local application.

In order to create a skin cream and to understand the mechanism of formation of multiple emulsions, Raynal et al. [30]
developed a multiple emulsion for the treatment of acne. It contained spironolactone in the intermediate oily phase, a
chlorhexidine salt in the external aqueous phase, and sodium lactate in the internal aqueous phase. This multiple emulsion with
three active substances was presumed to exert a threefold action: to treat the acne with spironolactone, to fight the bacteria
always present in this infection with chlorhexidine salt, and to moisturize the often dry skin by this treatment with sodium
lactate. Yet studies to evaluate the activity of this multiple emulsion after skin application still remain to be performed.
The most important studies conducted on the behavior of multiple emulsions for cosmetic applications are those of Ferreira et
al. [3133]. In all these studies w/o/w emulsions were compared with simple w/o and o/w emulsions. The originality of the
work arises from the fact that these authors attempted to obtain significant comparative results and to examine the influence of
vesicles. They developed simple emulsions with an oily continuous phase (w/o), simple emulsions with an aqueous continuous
phase (o/w), and multiple w/o/w emulsions, each containing the same components in identical amounts. Different procedures
were followed to obtain these three types of emulsion with the same formula. The oil phase is a liquid paraffin, the primary
surfactant a modified nonionic polyester of the polymer type (Hypermer A60nd)* and the secondary surfactant a copolymer of
ethylene oxide and propylene oxide (Poloxamer 407)*.
* I.C.I., 1 Avenue Newton, 92142 Clamart
Page 152

In the first study [31], Ferreira investigated the release of an antibacterial agent (metronidazole), which could be used topically.
The tests, performed with the Franz cell, compared the behavior of the three types of emulsions. Different membranes were
used: a hydrophilic cellulose membrane, a lipophilic silicone membrane, and biopsied rat skin. The three types of emulsion
were applied in an infinite dose with occlusion. The active substance was determined by High Performance Liquid
Chromatography.

The influence of the type of emulsion on the release of metronidazole was abundantly clear when the emulsions were
deposited on a membrane, like the cellulosic membrane, which offers only weak resistance to the passage of the hydrophilic
active ingredient. With this membrane, release is fast and nonlinear for the multiple emulsions (w/o/w) and for the simple o/w
emulsion, but it is slow and linear for the simple w/o emulsion. The total quantity of active substance released immediately
after application is about 15 times greater from the first two emulsions than from the w/o emulsion. After 5 h, however, the
proportion of metronidazole released is much higher from the w/o emulsion. The authors believe that this is due to depletion of
the active substance from the two emulsions with an aqueous external phase. This is not the case with the w/o emulsion; in
fact, only 3% of the dose applied is released from this form. These results can be explained on the basis of the partition
coefficient of metronidazole, which is relatively low and more favorable to water.

The influence of the type of emulsion on release is much less clear if the emulsions are deposited on membranes that control
diffusion kinetics of metronidazole: silicone and biopsied skin. Nevertheless, due to hydration, the passage of the active
substance from aqueous emulsions appears to be much easier across the skin than through the silicone membrane. Yet the
active substance always appears to be released slowly from the w/o/w multiple emulsion, slightly more slowly from the o/w
emulsion, but much faster from the w/o emulsion. Thus, with a hydrophilic active ingredient exhibiting weak affinity for oil,
the behavior of the multiple w/o/w emulsion is always intermediatè between that of the other two emulsions but closer to that
of the o/w emulsion.

In a subsequent study [32], Ferreira et al. conducted the same protocol but replaced metronidazole with glucose, an active
substance that possesses a much lower partition coefficient into oil. The glucose was determined by radioactivity. The authors
found the same release profile from the three types of emulsions with glucose as with metronidazole. The in vitro flux is high
for the o/w emulsion, low for the w/o emulsion, and intermediate for the multiple emulsion, irrespective of the type of
membrane. Thus, for example, absorption across hairless rat skin is equal to 72 µg/cm2 for the o/w emulsion, 0.8 µg/cm2 for
the w/o emulsion, and 7.8 µg/cm2 for the w/o/w emulsion. The authors show that although the release profiles are identical for
all three emulsions, the differences in flux are greater with glucose than with metronidazole. For example, release through rat
skin biopsy is practically identical for the o/w and the w/o/w emulsions immediately after application. After the first few hours
(312 h), however, the o/w emulsion exhibits a high flux and then reaches equilibrium after 12 h. These differences between the
emulsions are explained by the differences in the partition coefficient of glucose, favoring water more than metronidazole.
With such a strongly water-soluble substance, the behavior of the multiple emulsion is, therefore, close to that of the w/o
emulsion.
Thus the behavior of the different types of emulsion on the skin appears to depend equally on the physicochemical properties
of the active ingredient and on the vehicle itself. The behavior of a multiple emulsion, while always intermediate between the
other
Page 153

two types of emulsion, can vary considerably according to the type of active substance encapsulated. This observation was
confirmed by Ferreira et al. in a final study concerning moisturizing active substances [33]. The authors show that, irrespective
of the type of moisturizer, the occlusivity of the multiple w/o/w emulsion is always lower than that of a simple w/o emulsion.
By contrast, it differs from the effect of a simple o/w emulsion. Thus, when non-film-forming moisturizers are included in the
internal aqueous phase (sodium lactate, glycerol, urea, sodium pyrrolidone carboxylate), the occlusivity of the multiple
emulsion is higher than that of the simple o/w emulsion. On the other hand, with film-forming moisturizers (chitosan,
hydroxypropylcellulose), the occlusivity is lower. This can be explained as follows: in the multiple emulsion and the o/w
emulsion, these moisturizers are in the internal and external aqueous phase respectively. After application, when the water
evaporates, a network of polymers exhibiting some occlusivity is formed from the o/w emulsion, but such a network is lacking
in the multiple emulsions.

VIII
Conclusions

No therapeutic applications of multiple emulsions in new medicinal products have been proposed so far. The rarely marketed
multiple emulsions are designed for cosmetic application, with claims for refreshing, moisturizing, perfuming properties, and
the like. And yet it is likely that multiple emulsions may soon achieve wide acceptance as pharmaceutical dosage forms for
topical use.

Multiple emulsions for topical use can be compounded as creams of varying consistency and can be used as soon as they are
formed. Unlike other vesicular systems, it is not necessary to disperse them in a gel or in a cream in order to obtain a suitable
topical form. Moreover, recent research has demonstrated that multiple w/o/w emulsions display better performance than
simple emulsions with an aqueous continuous phase and are also more pleasant to use than simple emulsions with an oily
continuous phase, because they are less oily to the touch.

Multiple w/o/w emulsions provide water and oil to the skin, just like simple emulsions, and may contain various hydrophilic
and lipophilic components. They are easy to administer and also offer excellent cosmetic qualities. Multiple emulsions
represent a new and interesting form in several respects. For example, they offer protection for encapsulated active substances
and their prolonged release. When formulated with the proper components, they could be identified as targeted dosage forms
like other vesicular systems.

References

1. M. De Luca, C. Vaution, A. Rabaron, and M. Seiller, STP Pharma 4:67987 (1988).

2. D. Attwood and A. T. Florence, Multiple Emulsions, Surfactant Systems, Their Chemistry, Pharmacy and Biology, Chapman
and Hall, London, 1983, pp. 50966.

3. S. Matsumoto, J. Colloid Interf. Sci. 94:36268 (1983).

4. M. Frenkel, R. Schwartz, and N. Garti, J. Colloid Interf. Sci. 94:17478 (1983).

5. A. T. Florence, and D. Whitehill, J. Colloid Interf. Sci. 79:24356 (1981).


6. D. R. Kavaliunas, and S. G. Franck, J. Colloid Interf. Sci. 66:58688 (1978).

7. S. S. Davis, and A. S. Burbage, in Particle Size Analysis (M. S. Groves, ed.), Heyden, London,1978, pp. 395410.
Page 154

8. S. S. Davis, and A. S. Burbage, J. Colloid Interf. Sci. 62:36162 (1977).


9. S. Matsumoto, I. Takeshi, K. Masanori, and T. Ota, J. Colloid Interf. Sci. 77:56465 (1980).

10. Y. Kita, S. Matsumoto, and D. Yonezawa, J. Colloid Interf. Sci. 62:8794 (1977).

11. A. A. Elbary, S. A. Nour, and S. A. Ibrahim, Pharm. Ind. 52:35763 (1990).

12. J. L. Grossiord, M. Seiller, and F. Puisieux, Rheol. Acta 32:16880 (1993).


13. I. Terrisse, M. Seiller, A. Rabaron, A. Magnet, C. Le Hen-Ferrenbach, and J. L. Grossiord, Int. J. Cosm. Soc. 15:5362
(1993).

14. S. H. Magdassi, M. Frenkel, and N. Garti, J. Dispersion Sci. Technol. 5:4959 (1984).

15. S. S. Davis, and I. Walker, Int. J. Pharm. 17:20313 (1983).


16. S. Fukushima, M. Nishida, and M. Nakamo, Chem. Pharm. Bull. 35:337581 (1987).

17. J. I. Ratz, and G. H. Cueman, J. Soc. Cosm. Chem. 39:21122 (1988).

18. C. M. Adeyeye, and J. C. Price, Drug Dev. Ind. Pharm. 16:105378 (1990).

19. S. Matsumoto, T. Inoue, M. Kohdo, and K. Ikura, J. Colloid Interf. Sci. 77:55559 (1980).

20. S. Matsumoto, Y. Kita, and D. Yonesawa, J. Colloid Interf. Sci. 57:35361 (1976).
21. T. K. Law, T. L. Whateley, and A. T. Florence, J. Controlled Release 3:27990 (1986).

22. J. A. Omotosho, T. L. Watheley, and A. T. Florence, Biopharm. Drug Dispos. 10:25768 (1989).

23. J. A. Omotosho, T. L. Whateley, and A. T. Florence, J. Microencapsulation 6:18392 (1989).

24. J. A. Omotosho, T. L. Watheley, and A. T. Florence, J. Pharm. Pharmacol. 38:86570 (1986).


25. E. Fredo-Kumbaradzi, and A. Simov, Pharmazie 47:38889 (1992).

26. P. Oza, and S. G. Frank, J. Dispersion Sc. Technique 10:16385 (1989).

27. A. T. Florence, and D. Whitehill, Int. J. Pharm. 11:277308 (1982).

28. M. A. Attia, and F. S. Habib, STP Pharma 2:63640 (1986).

29. C. Kundu Subhass, Preparation and evaluation of multiple emulsions as controlled release topical drug, delivery systems.
Thesis, St. John's University, USA, 1990.

30. S. Raynal, J. L. Grossiord, M. Seiller, and D. Clausse, J. of Controlled Release 26:12940 (1993).

31. L. Ferreira, M. Seiller, J. L. Grossiord, J. P. Marty, and J. Wepierre, Int. J. Pharmaceutics 109:25159 (1994).

32. L. Ferreira, M. Seiller, J. L. Grossiord, J. P. Marty, and J. Wepierre, J. Controlled Release, accepted for publication,
September 1994.

33. L. Ferreira, M. Seiller, J. L. Grossiord, C. Vaution, J. P. Marty, and J. Wepierre, VI ème Congrès International de
Technologie pharmaceutique, APGI, Paris, 1992.
Page 155

7
Multiphase Emulsions
H. E. Junginger
Department of Pharmaceutical Technology, Leiden/Amsterdam Center for Drug Research, Leiden, The Netherlands

I. Introduction 156

II. Structural Elements of O/W Creams As Multiphase Emulsions 157

A. Creams with Ionic Surfactants 157

B. Colloidal Structures of Nonionic Hydrophilic Creams 163


C. Conclusions 167

III. Colloidal Structures of W/O Creams 168

IV. Colloidal Gel Structures of Amphiphilic Creams 169

V. Formation of Colloidal Crystalline Gel Phases during Manufacturing 171

A. Hydrophilic Ointment, DAB 10, and Water-Containing Hydrophilic 173


Ointment DAB 10

B. Stearate Creams 175

C. Nonionic Hydrophilic Cream DAC 176

VI. PhysicoChemical Stability and Aging of Colloidal Crystalline Gel Structures 176
VII. Surfactant Systems Used in Cosmetic Multiphase Emulsions 179

A. Consistency Increasing Agents 179

B. Nonionic O/W Systems 180

C. Nonionic O/W Systems (Alkylsulfates) 180


D. Nonionic O/W Systems (Soaps) 181

E. W/O Systems 181

References 181
Page 156

I
Introduction

For many centuries ointments and creams have been used to improve the healing of wounds, to treat skin diseases in empirical
ways, and, moreover, to retard the aging process of the skin and to preserve its natural beauty. Modern cosmetics and
dermatological preparations are identified as semisolids [1] due to their unique property of being in the solid state under
ambient conditions and being transformed to the liquid state when mechanically stressed during application on the skin. This
property allows the systems to spread easily on the surface of the skin. This semisolid state is also the main difference between
fluid liquid emulsions (in cosmetics often named milks) and creams, although in modern cosmetic preparations the transition
from a cream to an emulsion may be gradual. This semisolid state is attributed to particular structural elements, namely
crystallineand in some cases liquid crystalline gelstructures of colloidal dimensions that form a three-dimensional network
within the system. They are responsible for the consistency and stability of the creams, for their application properties such as
proper feel, spreadability, and cooling effects and their possible interactions with skin lipids. Furthermore, their formation
during the manufacturing process requires special attention and mixing speed, and shear stresses have to be adapted to not
interfere adversely in the crystallisation process of these crystalline structures. Additionally, the colloidal gel structures are
primarily responsible for the physical aging of topical preparations. As a consequence these colloidal gel structures form the
inherent networks of semisolid preparations, whereas they are absent in liquid emulsions. Because these colloidal gel
structures form additional phases in semisolid systems, they may be defined as multiphase emulsions.

In general, colloidal gel structures may be defined as special systems, consisting of at least two components, which by
themselves consist of one or more phases. By applying the classical definition of a gel, given by Wolfgang Ostwald, semisolids
may be described as two-component lyogels: the first component, being in the solid state, builds up a coherent three-
dimensional networkalso called matrix or texturein which a liquid, the second component, is immobilized as the other coherent
medium. A typical gel may be compared with a water-soaked sponge. Since both phases completely interpenetrate (bi-coherent
system), a differentiation between the inner and the outer phaseas is possible for liquid emulsions and suspensionscannot be
made. This three-dimensional structure of the gel network is built up in semisolids by secondary valence forces, particularly by
van der Waals forces and hydrogen bonds.
Creams as multiphase systems are defined as water-containing ointments and hence are dispersed systems, in which the
dispersed phase is stabilized by such a gel structure. Depending on the type of cream, the continuous liquid phase is always
part of the three-dimensional gel structure, i.e. in oil in water (o/w) creams the oily phase is the dispersed phase and water the
continuous phase, both stabilized by colloidal structures. In water in oil (w/o) creams the aqueous phase is the dispersed one,
and the oily phase is the coherent liquid phase of the colloidal gel structure. In amphiphilic creams special colloidal gel
structures exist that allow either the transition to an o/w cream if water is added or to a w/o cream if the amount of the oily
phase is increased.
If the colloidal structures of the different types of cosmetic creams are understood and known, systematic development of
creams with desirable properties becomes possible, and a basic understanding of these multiphase emulsions will be the key to
modern formulation approaches. Most research dealing with characterization of the colloidal gel
Page 157

structures of semisolids as ointments and creams has been conducted with pharmaceutical formulations that will be described
in the following paragraphs. However, the basic results obtained for special formulations (which are sometimes taken up as
official monographs in pharmacopoeias) have general applicability and are also valid for cosmetic preparation because they
contain either the same or similar ingredients. Hence the colloidal gel structures of multiphase emulsions can be viewed as
basic systems valid for both cosmetics and pharmaceutical preparations.

Generally, in pharmaceutical formulations, the oily phase of these systems consists mainly of petrolatum or paraffinum sub- or
perliquidum. The cosmetic properties, however, may be improved (resulting in a less occlusive effect) by using more skin-
compatible emollients such as isopropyl myristate, oleyl oleate, and triglycerides with fatty acids of medium chain length. In
the following structural models, these oily components are regarded as inert, simple paraffins; although due to their weak
polarity, these compounds may show some slight interference with the crystalline colloidal gel structure. On the other hand,
humectants such as glycerol, sorbitol, and natural moisturizing factors (NMF) in low concentrations are treated as belonging to
the water phases.

In this chapter the structural elements of multiphase emulsion as well as their implications on rational cream design and
manufacture will be discussed.

II
Structural Elements of O/W Creams as Multiphase Emulsions

A
Creams with Ionic Surfactants

1
Colloidal Structures of Cetostearyl* Sulfate Creams

As a model system for cetostearyl creams (o/w creams), the Water Containing Hydrophilic Ointment DAB 10 (German
Pharmacopoeia, 10th ed. Govi Verlag, Frankfurt/Main, 1992) will be discussed. Its formula is as follows:
Emulsifying Wax 9.0% wt/wt
Liquid Paraffins 10.5% wt/wt
White Petrolatum 10.5% wt/wt
Water 70.0% wt/wt

Emulsifying Wax (DAB 10) itself consists of


Cetyl sulfate sodium 5% wt/wt
Stearyl sulfate sodium 5% wt/wt
Cetyl alcohol 45% wt/wt
Stearyl alcohol 45% wt/wt

Small angle X-ray diffraction (SAXD) and wide angle X-ray diffraction (WAXD) in combination with quantitative differential
scanning calorimetry (DSC), thermogravime-
*The term cetostearyl is used here as shorthand for blends of C16 and C18 alkyl derivatives.
Page 158

Fig. 1
Gel structures of the Water Containing Hydrophilic Ointment
DAB 10. a: mixed crystal bilayer of cetostearyl alcohol and
cetostearyl alcohol sulfates, b: interlamellarly fixed water
layer, a + b: hydrophilic gel phase, c: lipophilic gel phase
(cetostearyl alcohol semihydrate), d: bulk water phase,
e: lipophilic components (dispersed phase).
(From Ref. 5.)

tric analysis (TGA), and polarization microscopical techniques led to the following structure of the Water Containing
Hydrophilic Ointment DAB 10 [25]. It was found that such o/w creams may be regarded as four-phase systems (Fig. 1). The
dominant matrices are the hydrophilic and the lipophilic gel phases. Both gel phases consist of surfactant bilayers of
colloidally sized mixed crystals. The surfactants in the bilayers are oriented in such a way that the hydrocarbon tails are
directed towards each other, as are the polar groups (Fig. 1, region a). The hydrophilic gel phase consists of cetostearyl alcohol
and all of the ionic sodium n-alkylsulfates, which are randomly distributed between the cetostearyl alcohol molecules. The
latter act as lateral spacers for the strong polar sodium n-alkylsulfate molecules. In the crystalline bilayer structure, therefore,
strong hydrophilic moieties and hydrophobic cores counteract each other (Fig. 1, region a).
One part of the total water amount of the system is interlamellarly inserted between the polar groups of the surfactant
molecules (Fig. 1, region b). This part of the water is named interlamellarly fixed water. Regions a and b together form the
hydrophilic gel phase. The water molecules interlamellarly fixed in the hydrophilic gel phases are in equilibium with water
molecules in the other aqueous part, the bulk water phase (Fig. 1, region d). The bulk water phase is the liquid component of
the gel structure, and the solid phase is the hydrophilic gel phase (although it contains part of the water interlamellarly bound).
The bulk water is fixed within the network of the hydrophilic gel phase mainly by capillary attraction forces. Furthermore, it is
assumed that the interlamellarly fixed water molecules exhibit physicochemical properties differing from those of the bulk
water phase.

The surplus of cetostearyl alcohol not incorporated in the hydrophilic gel phase builds up a separate matrix with lipophilic
properties (Fig. 1, region c) called lipophilic gel phase. The inner or dispersed phase (Fig. 1, region e) is to a large degree
immobilized mechanically by this lipophilic gel phase. The lipophilic gel phase consisting of pure cetostearyl alcohol is only
able to form a semihydrate with water [3].

Freeze fracture electron microscopy (FFEM) has added a new dimension to the studies
Page 159

Fig. 2
Freeze fracture micrograph of Emulsifying Wax DAB 10 (main structural component of
the Water Containing Hydrophilic Ointment DAB 10) with 70% wt/wt of water. a: mixed
crystal bilayer, b: interlamellarly fixed water, a + b: hydrophilic gel phase, c: fracture
edge of a lipophilic plane, d: bulk water phase. Magnification 48,000-fold.
(From Ref. 5.)

of colloidal o/w cream organization. This technique allows the visualizationat an ultra-structural levelof the previously
mentioned structural elements [58]. From Fig. 2 the hydrophilic gel phase can be recognized very clearly. In this micrograph
the alternating layers of the hydrophilic gel phase are nearly at right angles to the fracture plane. Together with areas of bulk
water (d) entrapped in the hydrophilic gel phase, the interlamellarly bound layers of water (b) and the bilayers of the surfactant
molecules (a) are visible. Together, (a) and (b) form the hydrophilic gel phase.

Investigations into the swelling ability of the Emulsifying Wax DAB 10 (main surfactant component of the Water Containing
Hydrophilic Ointment DAB 10) with water show a characteristic swelling behavior of the lamellar gel structure resulting in a
straight line (Fig. 3), when the long spacings, as obtained from small angle X-ray diffraction (SAXD), are plotted versus the
water/surfactant ratio (wt/wt) (see Fig. 3, where Ca is the weight fraction of surfactants; 1-Ca is the weight fraction of water).
At a water content of 70% wt/wt the thickness of the interlamellar fixed water layer is about 15 nm (total long spacing minus
long spacing of cetostearyl alcohol). For comparison, the molecular sizes for the cetostearyl alcohol (lipophilic gel phase) as
semihydrate and for the hydrophilic gel phase are given on the right-hand side of Fig. 3.

It must be emphasized that the degree of swelling of the hydrophilic gel phase depends on the total water content of the cream.
Hence a dynamic equilibrium exists between the bulk water and the interlamellarly fixed water. The bulk water phase forms
Page 160

Fig. 3
Swelling behavior of Emulsifying Wax DAB 10 with water.
Ca: weight fraction of surfactant, 1-Ca: weight fraction of

water, : interplanar spacings of cetostearyl alcohol

semihydrate (lipophilic) gel, : interplanar


spacings of the hydrophilic gel phase.
(From Ref. 3.)

the continuous phase of the system, but the interlamellar water fraction also contributes to this continuity. The capacity of the
hydrophilic gel phase to incorporate interlamellar water is high enough to obtain clearly defined melting and recrystallization
peaks measured by DSC, which strongly vary from the water-free systems (see Sec. V.A.). At a certain water content of the
system, maximum swelling of the interlamellar water layer is reached. Beyond this point the water molecules within the
interlamellar layer possess the same mobility as those of the bulk phase. Consequently, the colloidal structure of the
hydrophilic gel phase breaks down, and the three-dimensional gel structure is lost. This physical change represents the
transition from the cream into the (unstable) state of a suspension (emulsion). As a result of this transition, the plastic flow
behaviour properties of the cream are lost, and the system exhibits the pseudo-plastic flow behaviour of an emulsion or a
suspension.

The lipophilic gel phase (Fig. 1, region c) can only form a semihydrate with water, which is independent from the total water
present in the system. After the transition from a cream into an emulsion (suspension) state, the lipophilic gel phase still
surrounds and stabilizes the dispersed inner phase (Fig. 1, region e).

To characterize these o/w creams further, a technique was sought that allows a more
Page 161

quantitative differentiation between the different types of water in these systems. It may be speculated that the swelling
capacity of the hydrophilic gel phase may influence the water release to the skin and consequently the cooling efficiency of the
creams. By means of a dynamic thermogravimetric analysis (TGA), a method was developed that enabled us to differentiate
between interlamellarly fixed water and the bulk water fraction [3,4]. The TGA results form the Water Containing Ointment
DAB 10 with 70% water (wt/wt) and systems with lower water content are summarized in Fig. 4. The total water contents of
the different creams are plotted against interlamellarly fixed water fractions. At a total water amount of 70% wt/wt about 40%
wt/wt is present interlamellarly in the hydrophilic gel phase and only 30% wt/wt belongs to the bulk water phase. If the total
water exceeds about 80% wt/wt the hydrophilic gel phase reaches a saturation state, in which the water molecules in the
middle of the water layer between the bilayers are found to have the same free energy as those of the bulk water molecules at
the same temperature. This superhydrated state of the hydrophilic gel phase by water represents the transition from a cream
into an emulsion. If this over-hydrated state is exceeded, the fraction of interlamellarly fixed water decreases markedly in
favour of the fraction of the bulk water phase (Fig. 4 open symbols).

2
Colloidal Structures of Stearate Creams

For stearate creams a formulation proposed by Tronnier and used in cosmetic formulations was chosen as a model system [9].
The stearate creams investigated had the following compositions:

Fig. 4
Amount of interlamellarly fixed water in Water Containing
Hydrophilic Ointment DAB 10 depending on the total water
content of the system ( = unstable systems).
(From Ref. 3.)
Page 162

Stearic acid 12% wt/wt


Palmitic acid 12% wt/wt
Triethanolamine 1.2% wt/wt
Glycerin 13.5% wt/wt
Water 1061.3% wt/wt

Addition of water to the water-free system that consists of stearic acid, palmitic acid, and triethanolamine results in lamellar
mixed crystals in which the water is incorporated between the hydrophilic moieties of the lamellae. With increasing water
concentrations (1061.3% wt/wt) the swelling capacity of this hydrophilic gel phase is enhanced (Fig. 5). The thickness of the
interlamellar water layer is about 6.5 nm at a water content of 60% wt/wt (total long spacing equals 12.6 ± 0.2 nm) [5,10].

The results obtained by means of SAXD, DSC, and TGA led to a structure model for stearate creams as given in Fig. 6
[3,5,10]. The part of the lamellar mixed crystals that consists of free fatty acids and their triethanolamine salts is able to form
the hydrophilic gel phase. Between the polar moieties of the mixed crystals (Fig. 6, region a) water molecules are present (Fig.
6, region b). This water of the hydrophilic gel phase is in equilibrium with the bulk water of the continuous phase (Fig. 6,
region d).

Fig. 5
Swelling behavior of the gel forming components (triethanolamine stearate/palmitate)
of a stearate cream with water. Ca: weight fraction of surfactants, 1-Ca: weight fraction
of water, : interplanar spacing of stearic/palmitic acid mixture (1:1 moles); lipophilic
gel phase, : interplanar spacing of triethanolamine stearate/palmitate (water free), :
interplanar spacing of triethanolamine stearate/palmitate swollen with water
(hydrophilic gel phase of the stearate cream).
(From Ref. 3.)


Page 163

Fig. 6
Gel structures of stearate creams. a: mixed crystal
bilayer of triethanolamine stearate/palmitate, b:
interlamellarly fixed water, a + b: hydrophilic
gel phase, c: lipophilic gel phase (''stearate"),
d: bulk water phase, e: isolated "stearate."
(From Ref. 3.)

The second part of the gel network, consisting of pure mixed crystals of palmitic and stearic acid, which cannot retain water
interlamellarly, forms a lipophilic gelphase (Fig. 6, region c). If a dispersed (lipophilic) phase should be present in such a
system, it would be essentially immobilized by the lipophilic gel phase. Stearate creams can show a special pearl effect due to
the crystallization of very small isolated platelets (Fig. 6, region e), depending on the amount of added triethanolamine and on
the manufacturing conditions. Platelets are formed preferentially instead of a coherent lipophilic gel phase, especially in the
absence of a lipophilic phase.

The thermogravimetric results of the stearate cream are depicted in Fig. 7. It becomes apparent that at a high water content
only one-third of the water is fixed interlamellarly, but two-thirds of the water is present as a bulk water phase, which is
directly available for skin hydration. These facts may explain the results of Tronnier, who stated that the skin hydration rate by
stearate creams is much higher in comparison to other o/w creams [9].

These facts could also be confirmed by isothermal TGA, comparing systems with different ratios of interlamellarly fixed water
in their hydrophilic gel phases. At ambient temperatures, stearate creams lost a substantial part of the incorporated water much
more quickly than Water Containing Hydrophilic OIntment DAB 10 with a high amount of interlamellarly fixed water.

At a total water amount higher than 55%, these systems become unstable (Fig. 7), and a transition takes place from a cream
with a coherent three-dimensional hydrophilic gel network to an emulsion without these structural elements [4, 5].
B
Colloidal Structures of Nonionic Hydrophilic Creams

The o/w creams with crystalline gel structures in which nonionic emulsifiers are present are widely used in topical and
cosmetic preparations. This is easily explained by the fact that nonionic emulsifiers cause less irritation on sensitive skin
compared to ionic
Page 164

Fig. 7
Ratio of bulk and interlamellarly fixed water of a stearate
cream, which depends on the total water amount of the
systems ( , = unstable systems).
(From Ref. 3.)

emulsifiers. They tolerate addition of ionic compounds as active ingredients much better, and the stability of the system is not
affected as severely as in the case of ionic creams. Last but not least, a large number of nonionic emulsifiers that meet all
formulation requirements are on the market today.

To exemplify an o/w cream with nonionic emulsifier a water-containing nonionic hydrophilic ointment with the following
formula is described:
PEG-20 glyceryl stearate (PGM 20) 7.5% wt/wt
Liquid paraffin 7.5% wt/wt
Cetyl alcohol 5.0% wt/wt
Stearyl alcohol 5.0% wt/wt
Glycerin 8.5% wt/wt
White soft paraffin 17.5% wt/wt
Water 51.5% wt/wt

This particular system is known as Unguentum Hydrophilicum Nonionicum Aquosum, DAC (Deutscher Arzneimittel Codex,
1979). Similar formulations appeared in the Swiss Pharmacopoeia, 6th edition, and in the Formulary of the Dutch Pharmacists
(FNA).

Investigations using SAXD showed that mixtures of PGM20 and cetostearyl alcohol crystallize as mixed crystals (Fig. 8). In
the water and glycerin free system the diffraction peaks of pure PGM20 and cetostearyl alcohol are also found.
Page 165

Fig. 8
Swelling characteristics of cetostearyl alcohol, PEG-20 glyceryl stearate and water mixtures.
In the water-free state two diffraction peaks are found for cetostearyl alcohol (4.7 nm) and
cetostearyl alcohol/PEG-20 glyceryl stearate mixed crystals (6.3 nm). Up to 25% wt/wt
water, irreproducible diffraction peaks are found due to insufficient hydration of the
polyoxyethylene groups. Between 2560% wt/wt water, continuous swelling of the
hydrophilic gel phase takes place. At a water content higher than 60% wt/wt
the hydrophilic gel phase becomes unstable and breaks down.
(From Ref. 3.)

With SAXD no reproducible repeating distances are found up to 25% wt/wt if water is added to the water free system (Fig. 8).
The samples remain solid-like, and polarization microscopy shows anisotropic structures. At a water content of 25% wt/wt the
polyoxyethylene chains are just surrounded by the minimum of hydration water needed to build up a homogeneous lamellar
structure. As a consequence, any reduction of the water content leads to formation of mixed crystals containing partially
hydrated or incompletely hydrated polyoxyethylene chains as well as cetostearyl alcohol and PGM 20. This picture (Fig. 8) is
reinforced by the observation that samples containing less than 25% wt/wt water show several endothermal peaks with DTA,
and no electrical conductivity is observed. Thus, the coherency of the water layer has not been reached [11,12] due to
insufficient hydration of the polyoxyethylene groups of PGM20.
Between 2560% (wt/wt) water, continuous swelling of the hydrophilic gel phase
Page 166

takes place (Fig. 8). In this system the nonionic surfactant PGM20 crystallizes with cetostearyl alcohol in the form of mixed
crystals (Fig. 9, region a). The degree of swelling with water depends on total water content. The length of the
polyoxyethylene unit determines the maximum swelling capacity of the systems. Together with the water bound to the
polyoxyethylene units of the PGM20 molecules, the lamellar mixed crystals build up the hydrophilic gel phase (Fig. 9, regions
a and b) into which some bulk water may also be incorporated. The hydrophilic gel phase together with (part of) the bulk water
(Fig. 9, region d) are the components of the three-dimensional gel network. The surplus of cetostearyl alcohol again forms the
lipophilic gel phase (Fig. 9, region c), which immobilizes the lipophilic dispersed phase (Fig. 9, region e) consisting mainly of
white petrolatum and liquid paraffins. As a result of these investigations it is concluded that nonionic o/w creams may also be
regarded as four-phase systems consisting of the same structural elements as the ionic o/w creams [5,13]. Many nonionic
surfactants with various PEG-chain-lengths are available. Thus it becomes possible to develop nonionic o/w creams with any
desired ratio of interlamellarly fixed water and bulk water fraction. Knowledge of these gel structures is of fundamental
importance for developing formulations with desired properties such as controlled water release, especially in light of the
interactions between the vehicles and the skin [14,15].

Fig. 9
Schematic presentation of the gel structures of nonionic hydrophilic
cream (DAC 79). a: mixed crystal bilayer of cetostearyl alcohol and
PEG-20 glycerylmonostearate (PGM 20), b: interlamellarly fixed
water, a + b: hydrophilic gel phase, c: lipophilic gel phase
(cetostearyl alcohol-semihydrate), d: bulk water phase, e:
lipophilic components (dispersed phase).
(From Ref. 5.)
Page 167

C
Conclusions

The o/w creams with colloidally sized crystalline gel structures do not represent classical o/w emulsions in which typical oil
droplets, the inner phase, are stabilized by surfactant layer(s). On the contrary, they form coherent colloidal networks with
water filled pores. A large proportion of the water present in these creams is interlamellarly fixed.

The following conclusions may be drawn from the structure models for o/w creams with nonionic and ionic emulsifiers.

Interlamellarly Fixed Water

Interlamellar water, which is stabilized by interfacial forces, has different physicochemical properties than free water. During
storage of the creams it has been shown (see Sec. V of this chapter) that there is a shift in the ratio of interlamellarly fixed
water to bulk water in favor of the bulk water [15]. This may have some influence on the release of active substances and on
the cooling effect of the creams.

Cooling Effect
The gel structures of the various o/w creams described here always contain a portion of interlamellarly fixed water in the
hydrophilic gel phase, while another portion represents the bulk water phase. An o/w cream is smoothly applied to the skin
surface, but only the free water (bulk water) may be readily available. In the above-mentioned stearate cream two-thirds of the
total water is directly available for evaporation, while only one-third is interlamellarly fixed. Therefore, such a stearate cream
shows excellent cooling properties. In o/w creams with nonionic emulsifiers the ratio of bulk water and interlamellarly fixed
water depends largely on the polyoxyethylene chain length: with increasing chain length the interlamellarly fixed water
increases, and the extent of the cooling effect can thus be simply determined by the chain length. It must be noted, however,
that interlamellarly fixed water, which is expected to be released slowly, can have a sustained release effect on the skin only as
long as the colloidal structures are not destroyed during application to the skin. By selecting the appropriate components of the
hydrophilic gel phase, the interlamellarly fixed water amount and hence the cooling effect of the system can be varied to a
certain extent.
Transition from Cream to a Liquid Emulsion

When a large amount of water is added, the hydrophilic gel phase swells until a threshold value is reached. At this value the
water molecules in the middle of the swollen layer have the same free energy as the molecules of free, unfixed water (bulk
water) at the same temperature. In this transition state, the same high mobility of the water molecules in the hydrophilic gel
phase as that in the bulk water phase is reached, which causes the break-down of the hydrophilic gel phase.

The transition is characterized by the fact that the cream is transformed into a state which is described as a "liquid emulsion"
or, in cosmetic terms, as a "milk." The plastic flow behavior properties of the cream (due to the gel structures) and the cream's
yield value are lost; the resulting system exhibits the pseudo-plastic flow behavior of an emulsion.

Improved Viscosity of O/W Creams

Fatty alcohols play an important role in the formation of the hydrophilic gel phase. The surplus of fatty alcohols, which is not
integrated into the hydrophilic gel phase, forms the lipophilic gel phase. An increase in the proportion of fatty alcohols in such
creams always leads to increased formation of the lipophilic gel phase and is accompanied by an increase in the viscosity of
the entire system. This facilitates control of the viscosity of o/w creams. A large proportion of a lipophilic gel phase can
influence the interaction of the system with the skin. Systems
Page 168

with a very viscous lipid film have occlusive effects on the skin surface and may enhance therapy. For example, users of a sun
protection cream with a high sun protection factor do not like to have a white fat film residue on the skin, which has to be
intensively massaged into the skin. In addition, the long-term efficacy of the sun protection cream is improved.

Salt Intolerance

Addition of 2- or 3-valent ions to the hydrophilic gel phase of systems with ionic surfactants that are stabilized via interfacial
forces triggers a very sensitive reaction: when Ca2+, Mg2+, or Al3+ (as well as high concentrations of Na+ and K+ are added, the
stabilizing interfacial forces of the hydrophilic gel phase are lost, and the interlamellarly fixed water is released. An
irreversible phase separation occurs. The fact that white spots may appear on the skin surface after the application of such a
cream can be attributed to this destabilization of the cream. If the salt content on the skin is too high (e.g. after severe
sweating), phase separation of interlamellar water from the hydrophilic gel phase may occur, and the fatty residues of the
cream become visible as white spots. Creams with nonionic emulsifiers in general, however, do not react as strongly to salt
addition as systems with ionic emulsifiers.

III
Colloidal Structures of W/O Creams

By definition w/o creams are hydrophobic systems, the continuous phase of which is lipophilic. A general formula of a w/o
system (according to DAB 10) is as follows:
Anhydrous lanolin (wool fat)* 3.00% wt/wt
Cetostearyl alcohol 0.25% wt/wt
White petrolatum 46.75% wt/wt
Water 50.00% wt/wt
*with cholesterol as the most important ingredient.

A schematic presentation of w/o cream gel structures is given in Fig. 10. The w/o surfactants (cholesterol and other sterols as
well as cetostearyl alcohol) accumulate primarily at the interface between the water droplets and the oily phase of white
petrolatum, forming a monomolecular mixed layer of surfactants at the water/oil interface. Experimental work has proven that
the water capacity strongly increases when mixtures of fatty alcohols and sterols are used. It seems to be important to create a
liquid crystalline monolayer at the oil/water interface. Crystallization of the surfactant film at the interface drastically reduces
the system's capacity to take up water. As expected from their solubilities the sterols and fatty alcohols are dissolved in the
paraffin mixture (white petrolatum). A surplus of these o/w emulsifiers may cause separate crystallization in the lipophilic
phase and may, in addition, strengthen the (para)crystalline gel structures of white petrolatum [5].

The w/o creams represent classical w/o emulsions that are stabilized by a highly viscous gel of paraffins. The long-chain
paraffins are able to build up a three-dimensional solid gel network in which the short-chain liquid paraffins are mainly
immobilized by lysosorption [5].

In contrast to the o/w creams described in Sec. II, the crystalline gel structures of w/o creams do not contain any water between
the paraffin lamellae, nor does the crystalline
Page 169

Fig. 10
Schematic presentation of the gel structures of a w/o cream: long paraffin chains
are forming the solid gel in which liquid paraffin chains are immobilized by lyosorption.
Both cetostearyl alcohol ( ) and cholesterol (derivatives) ( ) accumulate
at the waterparaffin interface, and both are molecularly dispersed in the paraffin
gel according to their solubilities: a surplus of cetostearyl alcohol may
crystallize as separate lamellar crystals.
(From Ref. 5.)

gel structure show swelling properties upon addition of the liquid component (short-chain liquid paraffins).

IV
Colloidal Gel Structures of Amphiphilic Creams

Amphiphilic creams are colloidal systems that transform by addition of an oily phase to a w/o cream and by water addition to
an o/w cream. They therefore represent a special transition state between the other two cream types. A colloidal state is
obligatory for the existence of special gel structures in amphiphilic creams. Another prerequisite for the existence of a cream
with amphiphilic properties is the presence of lamellar mixed crystals that exhibit only limited swelling upon water addition.
Examples of suitable surfactants fulfilling these requirements are glyceryl stearate and esters or ethers of PEG with fatty
alcohols or fatty acids. The degree of polymerization of the PEG should be low with
Page 170

n = 23. All these compounds are surfactants of the w/o type. The colloidal structure of an amphiphilic cream is illustrated with
aid of the "basic cream" of DAC (Deutscher Arzneimittel-Codex, Govi Verlag, Frankfurt/Main, 1986) having the following
formulations:
Glyceryl stearate 4.0% wt/wt
Cetyl alcohol 6.0% wt/wt
Medium chain triglycerides 7.5% wt/wt
White petrolatum 25.5% wt/wt
PEG-20 glyceryl-stearate (PGM20) 7.0% wt/wt
Propyleneglycol 10.0% wt/wt
Water 40.0% wt/wt

The limited swelling ability of glyceryl stearate is well documented in the literature [16,17]. Melted glyceryl stearate together
with water swells continuously by interlamellarly incorporating water molecules between the hydrophilic glycerin residue until
a water content of 30% wt/wt is reached. At higher amounts of water the degree of swelling remains constant, and the excess
water is incorporated mechanically as droplets (bulk water phase) in the glyceryl stearate gel structure [5].

Conductance measurements of mixtures of glyceryl stearate, liquid paraffins, and water with a constant amount of glyceryl
stearate of 30% wt/wt show no conductivity below a water content of 20% wt/wt (Fig. 11). Systems containing less than 20%
wt/wt of water exhibit w/o characteristics [5].

On the other hand, mixtures with a water content between 60 and 70% wt/wt show o/w characteristics. In the range between
2050% wt/wt water, the systems behave as amphiphilic systems, i.e., addition of water results in o/w systems, while addition
of liquid paraffins generates w/o systems. However, none of these simple systems meet the requirements of well-developed
creams for pharmaceutical or cosmetic use. The resulting o/w systems are generally unstable, and phase separation of the
mechanically stabilized bulk water phase occurs. Therefore, well-compounded formulations of these amphiphilic systems
contain added nonionic o/w emulsifiers with high water binding capacity. In the above mentioned formulation of DAC, PGM
20 is added, which can form stable o/w compounded creams at high water content, i.e. exceeding 50% wt/wt. Amphiphilic
creams commonly contain relatively high amounts of o/w and w/o surfactants, and the oil and water phases are approximately
equal.

Amphiphilic creams are tri-coherent systems:

the dominant coherent gel phase is built up by glyceryl stearate lamellae,


the coherent continuous water phase consists mainly of interlamellarly bound water between the glyceryl stearate lamellae, and

the lipophilic phase is also coherent.

The liquid paraffins are predominantly mechanically fixed by the glyceryl stearate lamellae. A schematic representation of the
proposed gel structures of an amphiphilic cream and the transitions to a w/o system and o/w system are given in Fig. 12 A, B,
and C, respectively.

Starting with the above described amphiphilic system (Fig. 12B), the degree of water swelling of the glyceryl stearate lamellae
is strongly reduced by addition of oil. Due to
Page 171

Fig. 11
Conductivity of paraffinwater mixtures with a constant
amount of glyceryl stearate (Tegin M) of 30% wt/wt.
(From Ref. 5.)

the favorable phasevolume ratio with respect to the lipophilic phase, a w/o cream results automatically (Fig. 12A). On the
other hand, an increase of the water level of the amphiphilic basic cream allows strong swelling of the PGM 20 surfactant-cetyl
alcohol lamellae, which additionally stabilize the bulk water phase (Fig. 12C). As a result, water addition yields a stable o/w
cream.

V
Formation of Colloidal Crystalline Gel Phases during Manufacturing
In the previous parts of this chapter it has been shown that lamellar mixed crystals of colloidal size form the most important
structural elements of o/w creams: the hydrophilic and the lipophilic gel phases that are the solid parts of the gel structure of
these systems. In amphiphilic creams a hydrophilic gel structure with limited swelling potential coexists with precursors of
hydrophilic and lipophilic gel phases, which may transform to these crystalline phases after addition of water or oil to the
amphiphilic starting systems. All solid components of the gel structure in creams consist of lamellar crystals, which in the case
of hydrophilic gel phases, may be able to incorporate large amounts of water between the lamellae.

During the manufacturing process these lamellar structures must recrystallize from
Page 172

Fig. 12
Schematic presentation of gel structures existing during the transition from a w/o-cream
(A) to an amphiphilic cream (B) and to a o/w-cream (C). a: Mixed crystals consisting of
PEG-20 glyceryl stearate and cetostearyl alcohol (A in the water-free state, B in the
partly swollen state, C in the swollen state). b: Mixed crystals of glycerol stearate and
cetostearyl alcohol with limited swellability (A in the water-free state, B and C in the
state of limited swelling). c: Lipophilic phase (A as coherent continuous phase,
B as coherent phase, C as dispersed inner phase).
(From Ref. 5.)
Page 173

the melted mixture of all components of the system, and care must be taken during the cooling part of the manufacturing
process in order not to destroy these lamellae by excessively high shear forces. As a matter of fact, in all the systems
investigated, incorporation of water between the lamellae (hydrophilic gel phase) yields higher melting (recrystallization)
peaksmeasured by differential scanning calorimetry, DSCby comparison to the water free systems (lipophilic gel phase). Both
hydrophilic and lipophilic gel phases are present, and the different recrystallization temperatures of these gel phases must be
considered during the manufacturing process. In the following paragraphs the formation of the crystalline gel phases is
discussed.

A
Hydrophilic Ointment, DAB 10, and Water-Containing Hydrophilic Ointment DAB 10

Differential scanning calorimetry measurements of Hydrophilic Ointment DAB (i.e., the Water-Containing Hydrophilic
Ointment in the absence of water (the formulation presented in Sec. II.A.l. of this chapter)) have shown the following results
(18).
The heating curve shows two distinct endothermic peak maxima at 23°C and 41°C, respectively (Fig. 13a). The first peak at
23°C represents the polymorphic phase transition of the b0-modification to the a-modification of the mixed crystals consisting
of 90% wt/wt of cetostearyl alcohol and 10% wt/wt of sodium cetostearyl sulfate. The weak shoulder at 37°C represents the
phase transition of the g4-polymorph to the a-modification. The broad peak with a peak maximum at 41°C represents melting
of white petrolatum components and the phase transition from the a-modification to the liquid phase of cetostearyl alcohol
(sulfate sodium). Upon cooling of the completely molten systems, the correspondent peak for recrystallization and
polymorphic phase transitions are obtained (Fig. 13a).
After addition of 20% wt/wt of water, the Hydrophilic Ointment DAB 10 shows completely different DSC scans (Fig. 13b):
the polymorphic phase behavior of the fatty alcohol components is shifted to lower temperatures. The melting point of the a-
phase together with the white petrolatum components is slightly increased from 41°C to 43°C, and a new peak is seen at 68°C.
Similar exothermic peak patterns are obtained during cooling of the sample.

The Water-Containing Hydrophilic Ointment DAB 10 with 70% wt/wt water shows the following DSC scans (Fig. 13c). Upon
heating, a very shallow peak in the temperature range of 28°C up to 42°C indicates melting of the white petrolatum
components of the system. The peak at 54°C results from the melting of the cetostearyl alcohol (semihydrate), which is the
lipophilic gel phase of the system. The peak at 72°C results from the melting of the hydrophilic gel phase of the systems. Note
that the area under the peaks is reduced due to the large amount of water added to the system [18].

Upon cooling of the molten system the hydrophilic gel phase recrystallizes at about 70°C, whereas the lipophilic gel phase
solidifies at 54°C and the white petrolatum components finally stiffen the system in the temperature range between 42°C and
28°C (Fig. 13c).

In the manufacturing process, the lipophilic components are heated to about 80°C, to which water at the same temperature is
added under vigorous stirring in order to achieve homogeneous mixing. At this temperature a w/o emulsion exists. At 75°C
phase inversion to an o/w emulsion takes place, subsequently followed by recrystallization of the hydrophilic gel phase of the
system. Below that temperature the shear stress applied to the system must be reduced in order to ensure the interlamellar
insertion
Page 174

Fig. 13
DSC heating and cooling curves (heating and cooling rate: 2 K/min) of
a: water free formulation (Hydrophilic Ointment DAB 10), b: water Containing
Hydrophilic Ointment with 20% wt/wt water, and c: water Containing
Hydrophilic Ointment of 70% wt/wt water. DAB 10 formula.
(From Ref. 18.)

of water molecules between the mixed crystals of the hydrophilic gel phase. At about 54°C the lipophilic gel phase crystallizes
and stabilizes the dispersed lipophilic phase. During manufacturing, stirring should still be high enough to remove solidifying
material from the vessel's wall and distribute it homogeneously among the other (still liquid) material.

High speed and high shear processes, however, may lead to complete destruction of the colloidal structures of such an o/w
cream: whereas the ratio of interlamellarly fixed to bulk water in the Water Containing Hydrophilic Ointment DAB 10 is about
40 to 30% wt/wt as shown in Fig. 4 of this chapter, the ratio becomes 61 to 9% wt/wt when the cream is subjected to high
shear stress (Ultra Turrax) during the complete cooling cycles [19]. Such treatment results in a completely inhomogeneous
cream with phase separation of the water phase. If such a ''destroyed" cream is reheated to 80°C and cooled down with
appropriate shear stress, a homogeneous o/w cream is obtained.
Page 175

B
Stearate Creams

Differential scanning calorimetry investigations of the water-free formulation (see Sec. II.A.2.) show a strong endothermic
peak at 57°C [10] (results not shown). Polymorphic phase transitions are less pronounced in mixtures of fatty acids compared
to the corresponding fatty alcohols. When water is added in increasing amounts to the waterfree components, DSC scans show
two clearly distinguishable peaks at 57°C and 60°C respectively (Fig. 14a, b, and c). They can be attributed to the melting of
the lipophilic gel phase (57°C) and of the hydrophilic gel phase (60°C) of the stearate creams. The corresponding cooling
curves in Fig. 14 show the same exothermic peaks even more clearly, indicating crystallization of the hydrophilic gel phase at
about 56°C and recrystallization of the lipophilic gel phase at about 47°C. During manufacture, high mixing speed with high
shear stress should be applied during the cooling process until 60°C is reached. Then mixing with low shear stress should be
continued until crystallization of all components of the stearate cream is complete [10].

Fig. 14
DSC heating and cooling curves (heating and cooling rate: 2 K/min)
of a: water containing stearate cream with 20% wt/wt water, b: water
containing stearate cream with 30% wt/wt water, and c: water
containing stearate cream with 50% wt/wt water.
(From Ref. 10.)
Page 176

C
Nonionic Hydrophilic Cream DAC

The water-free formulation of the Nonionic Hydrophilic Cream DAC (See Sec. II.B. of this chapter) shows three distinct
endothermic peaks at 32°C, 40°C, and 51°C, as depicted in Fig. 15 [12]. The peak at 51°C denotes the melting of cetostearyl
alcohol. The other peaks belong to the b-to-a transition of cetostearyl alcohol, to melting of cetostearyl alcohol and PGM20,
and possibly to melting of separated PGM 20. The latter three processes overlap [12]. In Fig. 16 the heat flow scan of a sample
containing 42.1% (wt/wt) water is given. The endothermal heat effect at 61°C is due to the transition of the hydrocarbon chains
in the lipophilic bilayers from the a-modification to the liquid state. Notice the weak shoulder at the low-temperature side of
the peak. The shoulder is presumably connected with regions of hydrated cetostearyl alcohol in which no PGM20 is
incorporated (melting point 56°C). This peak represents the lipophilic gel phase, whereas the very sharp peak at 61°C is
connected with the melting of the hydrophilic gel phase [12,20]. Complementary DSC peaks are obtained when the molten
cream is cooled down. Also in the case of the Nonionic Hydrophilic Cream DAC, high-shear mixing could be applied during
the cooling process until a temperature of 61°C is reached. Thereafter low-shear mixing should be applied to obtain a cream
with the desired properties.

It is concluded from this discussion that the hydrophilic gel phases crystallize at relatively high temperatures (72°C59°C). For
the manufacture of large-scale batches, in-process control of the mixing speed at temperatures above the crystallizing
temperatures of the hydrophilic gel phases is required (at high-shear rates). After recrystallization of the gel structures, low-
shear stress should be maintained in order to obtain proper colloidal crystals of the most important hydrophilic gel phases of
these o/w creams.

VI
PhysicoChemical Stability and Aging of Colloidal Crystalline Gel Structures

Little information is currently available about the physico-chemical stability and the aging processes of colloidal crystalline gel
structures in o/w creams. The dominant effect

Fig. 15
DSC heating curve of a mixture containing cetyl alcohol,
stearyl alcohol, and PGM 20. Heating rate: 5 K/min.
(From Ref. 12.)
Page 177

Fig. 16
DSC heating curve of Nonionic Hydrophilic Cream
containing 42.1% wt/wt water. Heating rate: 5 K/min.
(From Ref. 12.)

described for both Water Containing Hydrophilic Ointment [21] and Nonionic Hydrophilic Cream DAC [14] is continuous
repair of crystal defects of the lamellar sheets during storage. In both systems cetostearyl alcohol is present in excess with
respect to the formation of an "ideal" hydrophilic gel phase. The excess of cetostearyl alcohol results in the formation of the
lipophilic gel phase: see Sec. II.A.1. and II.B. of this chapter. During manufacture, stirring causes shear stresses. As has been
described in the previous section of this chapter, the crystallization process of the hydrophilic gel phase is rapid, and hence
mixed crystals with a nonideal stochastic distribution of cetostearyl alcohol molecules and their sodium sulfate esters (Water
Containing Hydrophilic Ointment DAB 10 or PGM20 (Nonionic Hydrophilic Cream DAC) are formed. The stochastic
distribution of both components will be different from that of the most stable (ideal) arrangements of both components in these
primary crystals and will contain numerous crystal defects. Physicochemical aging processes in such systems mean
predominantly healing processes of these crystals. As a consequence of these aging processes the ratio of interlamellarly fixed
water to bulk water is shifted in favor of the amount of bulk water. Measurements employing SAXD clearly show a decrease
of the long spacings of hydrophilic gel phases of these systems. Upon heating and melting of the o/w creams and subsequent
cooling, the maximum insertion of water molecules between the lamellar mixed crystals is again obtained.

De Vringer et al. [15] have described in detail the two extreme situations (depicted in Fig. 17). The already mentioned excess
of cetostearyl alcohol can be distributed among the bilayers in two possible situations. Three types of water are distinguished:
F stands for free water, that is, water inside or outside the lamellae into which no polyoxyethylene chains are incorporated; H
stands for water bound to the hydroxyl groups of cetylstearyl alcohol; and O represents water in an aqueous polyoxyethylene
solution. In situation 1 (Fig. 17) the excess of cetostearyl alcohol is distributed inhomogeneously. In fact, two phases coexist.
Small patches of hydrated cetostearyl alcohol, into which no PGM 20 is incorporated, alternate with gel patches in which
cetostearyl alcohol and PGM 20 are present in the ideal ratio. It has been shown that only a one-molecule-thick water layer
Page 178

can be incorporated between the hydroxyl groups of pure cetostearyl alcohol [12]. Since in this case the patches of swollen
cetostearyl alcohol do not collapse, it is likely that they are small and are stabilized by the surrounding gel patches.

In situation 2 (Fig. 17) the hydrocarbon tails of the PGM20 molecules are distributed homogeneously among the cetostearyl
alcohol molecules in the lipophilic bilayers. Consequently, the polyoxyethylene chains are also distributed homogeneously
among the hydrophilic layers. Obviously, the two situations described here are the extremes. In fact, a mixture of all
intermediate situations could exist.

If we consider only the hydrocarbon bilayers, situations 1 and 2 are equal. For this reason we cannot differentiate situation 1
from situation 2 by means of DSC heating experiments or with WAXD experiments. Both methods yield information about the
lipophilic sheets, Still, this information is very valuable. Since the WAXD reflection and the melting enthalpies of the
hydrocarbon sheets of old samples resemble the data obtained from freshly prepared samples, we can conclude that the
structure of the hydrocarbon sheets does not change on aging [14]. In SAXD experiments, we notice, however, that the long
spacings decrease on aging. Since the structure of the lipophilic sheets does not change, the decrease must be related to
changes of the hydrophilic layer. Experiments on PGM20/water mixtures with DSC [15] showed that two water molecules are
tightly bound to the ether oxygens of the polyoxyethylene unit, and although excess water still contributes to the hydration of
the polymer chain, further addition of water does not lead to an increase in the fraction of nonfreezing water. For this reason,
and because the hydration of polyoxyethylene is stronger at ambient temperature than at the elevated preparation temperature,
it was expected that on aging, the nonfreezing water fraction, which equals the minimum amount of water required to retain
the gel structure, should remain unchanged. Indeed, DSC cooling experiments showed that in both old and freshly prepared
samples, two water molecules per oxyethylene unit did not freeze [15].

In order to gain more insight into the mechanism involved in the aging processes, spin-lattice relaxation experiments have been
performed [15] with Nonionic Hydrophilic Cream DAC containing 60.1% wt/wt D2O. In fact these measurements prove that
the inhomogeneous distribution of the polyoxyethylene chain in the mixed crystals is becoming more homogeneous. The
change of the distribution is caused by the temperature dependency of the hydration of polyoxyethylene. At the preparation
temperature, the

Fig. 17
Possible distributions of cetylstearyl alcohol among the lipophilic
sheets. 1 = inhomogeneous situation; 2 = homogeneous situation;
F, H, and O represent free water, water bound to the
lipophilic-hydrophilic interface (////), and water in an aqueous
polyoxyethylene solution, respectively.
(From Ref. 15.)
Page 179

TABLE 1 distribution of Water in Colloid Crystalline Gels


Age of the samples = 8 days Age of the samples = 185 days
Total water, Bulk water, Interlamellarly fixed Total water, Bulk water, Interlamellarly fixed
% wt/wt % wt/wt water, % wt/wt % wt/wt % wt/wt water, % wt/wt
63 21 42 63 30 33
70 33 47 70 31 39
77 25 52 77 28 47
80 27 53 80 38 42

hydration is relatively weak, and the hydrated polyoxyethylene chains form clusters. These clusters are fixed at the moment the
lipophilic chains of PGM20, which stick in the hydrocarbon sheets, solidify. At room temperature, the hydration of
polyoxyethylene is stronger, and hydration forces tend to separate the polyoxyethylene chains. However, this separation is only
possible if the hydrocarbon chains of PGM20, which are tied up in the lipophilic sheets, also part. Because of the crystalline
character of the lipophilic sheets, lateral diffusion of the hydrocarbon tails is very slow. Consequently, the aging process is also
slow [15]. Within 190 days the thickness of the interlamellarly fixed water inserted between the lamellae decreased about 50%.
With thermogravimetric analysis (TGA) similar results have been obtained with different amounts of water, as depicted in
Table 1 [19].

During physicochemical aging studies of these creams, no macroscopic change of the systems is observed. Hence the
"primary" quality of the creams is not affected. The "secondary" quality, however, may be affected in such a way that the shift
to a higher amount of bulk water may change the desired application properties of the o/w creams to some extent. However, it
should be stressed that during skin application most of their colloidal structures are either mechanically destroyed or changed
by quick water evaporation because the thickness of the creams is only about 20 µm on the skin surface [22].

VII
Surfactant Systems Used in Cosmetic Multiphase Emulsions

Surfactant systems currently used in cosmetics as ingredients for multiphase emulsions are mostly complex multicomponent
formulations. Division into the following categories is possible: consistency increasing agents, nonionic o/w systems, anionic
o/w systems (alkylsulfates), anionic o/w systems (soaps), and w/o systems. Some examples are listed below.
A
Consistency Increasing Agents
CFTA designation Composition
Cetyl Palmitate Cetyl palmitate

(table continued on next page)


Page 180

(table continued from previous page)


CFTA designation Composition
Glyceryl Stearate Glycerine monostearate
Glyceryl Stearates Mixture of mono-, di-, and triglycerides of
palmitic and stearic acid
Glyceryl Stearate (and) Cetearyl Alcohol (and) Cetyl
Esters (and) Coco-Glycerides
Sorbitan Stearate Sorbitan monostearate
Cetearyl Alcohol
Myristyl Alcohol
Cetyl Alcohol
Stearyl Alcohol

B
Nonionic O/W Systems
CFTA designation Composition
PEG-20 Glyceryl Stearate Polyoxyethylene
glycerine monostearate
with approx. 20 moles
EO
Cetearyl Alcohol (and) Ceteareth-20
Cetearyl Isononanoate (and) Glyceryl Stearate (and) PEG-20 Glyceryl
Stearate (and) Cetearyl Alcohol (and) Ceteareth-
20 (and) Cetyl Palmitate
Glyceryl Stearate (and) Ceteareth-20 (and) Ceteareth-12 (and) Cetearyl
Alcohol
(and) Cetyl Palmitate
Cetearyl alcohol (and) Cetearyl Glucoside
Ceteareth-12
Ceteareth-20
Ceteareth-30
PEG-40 Hydrogenated Castor Oil
PEG-60 Hydrogenated Castor Oil
Polysorbate 20 Polyoxyethylene sorbitan
monolaurate
Polysorbate 60 Polyoxyethylene sorbitan
monostearate
Polyglyceryl-2-PEG-4 Stearate Polyoxyethylene
polyglyceryl stearate

C
Nonionic O/W Systems (Alkylsulfates)
CFTA designation
Glyceryl Stearate (and) Sodium Cetearyl Sulfate
Cetearyl Alcohol (and) PEG-40 Castor Oil (and) Sodium Cetearyl Sulfate
Cetearyl Alcohol (and) Sodium Lauryl
Sulfate

Anda mungkin juga menyukai