The monolayer structure of the branched nonyl phenol oxyethylene glycols at the air-water interface

Article abstract of DOI:10.1021/jp9919009

The behavior of the para-substituted nonyl phenol ethoxylates with branched nonyl chains at the air-water interface was studied to examine the effect of branching in the alkyl chain and the insertion of a benzyl ring in the alkyl chain on the monolayer structure of the nonionic surfactant. A series of para-(C₄H₉)₂CHC₆H₄(OC₂H₄)_nOH (BNPE), where n=4,8, and 12 were used in the surface tension measurements. The CMC was 1.1 ± 0.3 x 10^-5 M for BNPE₄, 4.0 ± 0.3 x 10^-5 M for BNPE₈, and 8.0 ± 0.3 x 10^-5 M for BNPE₁₂; the limiting area per molecule (A_cmc) at the cmc was 46 ± 3, 61 ± 4, and 75 ± 5 sq A, respectively. For the same size of headgroups, the A_cmc values were almost the same as those obtained from dodecyl ethoxylates. This indicated that A_cmc for these nonionic surfactants are dependent on the size of headgroups and independent of the chemical structure of the hydrophobic chains. Nonyl phenol layers projected onto the surface normal direction were 18 ± 3 A (BNPE₄), 19 ± 3 A (BNPE₈), and 22 ± 3 A (BNPE₁₂). These layers, in all cases, were nearly twice as thick as the fully extended chain, suggesting that the hydrophobic chain are widely distributed across the layer an extensive resulting in a strong mixing of the chain with the ethoxylate groups.

Full text of DOI:10.1021/jp9919009

J. Phys. Chem. B 2000, 104, 1507-1515
1507
The Monolayer Structure of the Branched Nonyl Phenol Oxyethylene Glycols at the
Air-Water Interface
S. R. Green, T. J. Su, and J. R. Lu*
Department of Chemistry, UniVersity of Surrey, Guildford GU2 5XH, U.K.
J. Penfold
ISIS Facility, Rutherford Appleton Laboratory, CCLRC, Chilton, Didcot OX11 0QX, U.K.
ReceiVed: June 10, 1999
The structure of the monolayers formed by a group of monodisperse nonyl phenol ethoxylates with para-
substituted 1-butylpentyl chains (para-(C4H9)2CHC6H4(OC2H4)nOH, abbreviated to BNPEn, n ) 4, 8, and
12) at the air-water interface has been determined by surface tension measurements and neutron reflection.
The critical micellar concentration (cmc) was found to be 1.1 ( 0.3 × 10 -5 M for BNPE4, 4.0 ( 0.3 × 10-5
M for BNPE8 and 8.0 ( 0.3 × 10-5 M for BNPE12 and the limiting area per molecule (Acmc) at the cmc to
be 46 ( 3, 61 ( 4, and 75 ( 5 Å2, respectively. The values of Acmc are almost identical to those obtained
from dodecyl ethoxylates (C12En) for the same size of the headgroups, suggesting that Acmc for these nonionic
surfactants is determined by the size of the headgroups and is not affected by the chemical structure of the
hydrophobic chains. The thicknesses of the nonyl phenol layers projected onto the surface normal direction
were found to be 18 ( 3 Å for BNPE4, 19 ( 3 Å for BNPE8, and 22 ( 3 Å for BNPE12. In all cases they
were about twice as thick as the fully extended chain, suggesting a broad distribution of the hydrophobic
chain across the layer and hence a strong mixing of the chain with the ethoxylate groups.
Introduction
C8H17C6H4(OC2H4)9-10OH. Our NMR measurements suggest
that the octyl chain has at least four different alkyl chain
Alkyl phenol ethoxylates (APEs) have been widely used in
various domestic and industrial applications, ranging from heavy
duty detergents and car shampoos to stabilizers for colloidal
dispersions. APEs are also used in many biological processes
to mediate the activities of proteins and enzymes. Unlike sodium
dodecyl sulfate (SDS), APE surfactants are mild and their
binding to many globular proteins does not usually cause the
breakdown of the globular framework.1 However, APEs are
effective in deactivating a number of biological molecules such
as cytochrome oxidase, adenylate cyclase.2,3 These properties
are thought to be related to the specific binding of APE
molecules to the active sites of these biological molecules. Such
selective binding is in sharp contrast to the cooperative mode
of interaction between protein and ionic surfactants such as SDS,
which usually deteriorate the native or near-native structures.
Many investigations have been made to characterize the
adsorption and aggregation of APE surfactants in aqueous
solutions using surface tension measurements, NMR and
dynamic light scattering.4-12 These studies have demonstrated
how the critical micellar concentration (cmc), surface and
interfacial tension, the size and shape of surfactant aggregates,
and cloud points are affected by temperature, salt type and
concentration, and the size of ethoxylate headgroups. The main
drawback in these studies is that impure samples have been
used. For most commercial APE samples, the alkyl chains
vary in the number of carbons as well as in the degree of
branching and the ethoxylate groups are polydisperse. For
example, Triton X-100 has a nominal chemical formula of
structures and liquid chromatography shows that the length of
the ethoxylate headgroups is highly polydisperse. The polydis-
persity in these surfactant samples has undermined the reliability
of many results of such studies reported in the literature. In
some cases, the uncertainty in the chemical structure of the
surfactant samples also constrains the interpretation of the
data.5,6,9
In a systematic investigation of the adsorption of nonionic
surfactants, we have recently studied the structural conformation
of alkyl ethoxylates C12En at the air-water interface using
neutron reflection. We found that for this series of surfactants
a strong mixing exists between the alkyl chains and the
ethoxylate headgroups.13-17 The extent of mixing increases with
the size of the headgroups. In an effort to examine the effect of
the chemical structure of the alkyl chain on the conformational
structure of the adsorbed monolayers we have recently studied
the adsorption of alkyl phenol ethoxylates (CmH2m+1C6H4-
(OC2H4)nOH) at the air-water interface using neutron reflection
combined with surface tension measurements. Pure monodis-
perse APE samples have been used throughout this work. The
results from this study together with those of CmEn series will
allow a direct comparison to be made between the chemical
configuration of the hydrophobic groups and the structure of
the adsorbed nonionic surfactant layers.
Most existing synthetic routes for making APE samples utilize
AlCl3 or ZnCl2 as an alkylation catalyst, which causes a random
branching of the alkyl chain and substitution at both para and
ortho positions.18 We have hence devised new synthetic routes
so that only one isomeric structure is obtained in a given
synthesis. We have prepared monodisperse para and ortho nonyl
* To whom all correspondence should be addressed: Department of
Chemistry, University of Surrey, Guildford GU2 5XH, U.K. E-mail:
j.lu@surrey.ac.uk.
10.1021/jp9919009 CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/28/2000

1508 J. Phys. Chem. B, Vol. 104, No. 7, 2000
Green et al.
deuterated bromobutane. Proton NMR spectra have been taken
for all the samples and Figure 2 shows the result for BNPE4 as
an example. Within the spectrum the ratios between different
protons follow the stoichiometric relation within the experi-
mental resolution. The peak at 0.8 ppm corresponds to -CH3,
around 1.2 ppm to the middle -(CH2 CH2)-, at 1.5 ppm to the
end -CH2- on the butyl chain, at 2.4 ppm to the single H on
the tert carbon at the end of the two butyl chains, the three
peaks around 4 ppm to the whole oxyethylene units, and the
two peaks at 6.8 and 7 ppm to the phenol ring. The peak at 7.3
ppm arises from the deuterium chloroform and the small one
at 2.7 ppm may originate from the labile OH group. The two
small ones around 2 ppm are possibly associated with traces of
ethyl ether remaining in the sample or other unknown impurities.
In the case of BNPE8 and BNPE12, the positions of the peaks
are almost identical and the only changes are the integrated areas
for the headgroups around 4 ppm that increase in proportion to
Figure 1. Molecular configuration of the branched nonyl phenol
ethoxylates with n ) 4, 8, and 12 and with the alkyl chains
hydrogenated and deuterated.
phenol ethoxylates in fully hydrogenated and alkyl chain
deuterated forms. In this work we present the results on the
surface behavior of the para-substituted nonyl phenol ethoxylates
with branched nonyl chains at the air-water interface. Our main
aim is to examine how the branching in the alkyl chain and the
insertion of a benzyl ring in the alkyl chain affect the monolayer
structure of the nonionic surfactant. The measurements recorded
in this paper have been made using a series of BNPE samples
with different ethoxylate headgroups.
Experimental Section
1
n, as expected. H NMR spectra have also been taken for the
Two isotopic species of the surfactants were synthesized,
para-C9H19C6H4(OC2H4)nOH, abbreviated to h-BNPEn, n )
4, 8, and 12, and para-C8D17CH2C6H4(OC2H4)nOH, abbrevi-
ated to d-BNPEn, n ) 4, 8, and 12. The chemical structure of
the two labeled species is shown in Figure 1. The main reaction
steps are outlined as follows:
alkyl chain deuterated samples and the integrated areas below
2 ppm (corresponding to the residual protons in the butyl groups)
allow a direct determination of the degree of deuteration. The
level of deuteration in the butyl groups was found to be 92 (
2%. Bromobutane (99% pure) and 4-methoxybenzylaldehyde
(98% pure) were obtained from Aldrich. Deuterated bromobu-
tane was made by brominating deuterated butanol (BOSTI,
Moscow) following the procedures given in ref 18.
p-CHOC6H4OCH3 2C H MgBr8
4
9
The neutron reflection measurements were done on the white
beam time-of-flight reflectometer SURF at the Rutherford
Appleton Laboratory, ISIS, Didcot, UK.19 To avoid mechanical
vibration, the sample holder was mounted on an anti-vibration
table. Positive liquid meniscus was obtained by pouring sur-
factant solutions into Teflon troughs. The position of each liquid
surface was aligned using a laser beam that shares the same
beam path as neutron radiation. Neutron beam intensities were
calibrated with respect to the reflectivity from a D2O surface.
A flat background determined by extrapolation to the high values
of momentum transfer, κ (κ ) (4π sin θ)/λ), where λ is the
wavelength and θ is the glancing angle of incidence, was
subtracted. All the experiments were performed at 25 ( 2 °C.
Ultrapure water was used for all the measurements (Elga
Ultrapure) and all the glassware and Teflon troughs for the
reflection measurements were cleaned using alkaline detergent
(Decon 90) followed by repeated washing in ultrapure water.
Surface tension measurements were made using a Kruss K10
maximum pull tensiometer with a Pt du Nouy ring. For a given
concentration several readings were typically taken and the
difference between the repeated measurements was usually less
than 0.5 mN m-1 near the cmc and was about 1 mN m-1 over
the dilute concentration region.
p-(C4H9)2COHC6H4OCH3 (1)
BBr3
NaBH /CF COOH8 p-(C4H9)2CHC6H4OCH3
8
4
3
p-(C4H9)2CHC6H4OH
TsCl/C H ) N8 p-(C4H9)2CHC6H4OTs
2
5 3
+
-
K O (C H O) H8 p-(C4H9)2CHC6H4(OC2H4)4OH
2
4
4
A Grignard reaction coupled bromobutane and 4-methoxy-
benzylaldehyde to afford a tertiary alcohol. Reduction of the
hydroxyl group with sodium borohydride in trifluoro acetic acid,
followed by deprotection of the latent phenol functionality with
boron tribromide, gave branched chain nonyl phenol an overall
40% yield. After purification by flash column chromatography
the tosylate of the phenol was formed by reaction with para-
toluenesulfonyl chloride (abbreviated to Ts) in the presence of
triethylamine using dichloromethane as solvent. In a separate
flask, potassium tert-butoxide (twice in excess of phenol in
molar ratio) was reacted with dry tetraethylene glycol (six times
in excess of phenol) to form the mono potassium salt. The tert-
butyl alcohol released was removed under vacuum. The tosy-
lated phenol was then added to the monopotassium salt of
tetraethylene glycol and with vigorous stirring the reaction
mixture maintained 80 °C under vacuum for 4 h. The reaction
mixture was neutralized using concentrated hydrochloric acid,
and was then extracted several times with diethyl ether, followed
by a concentration of the combined organic extracts and flash
column chromatography to give BNPE4 in a 40% yield to the
starting phenol. BNPE8 was prepared by coupling BNPE4 with
tetraethylene glycol in a similar manner, and BNPE12 was then
prepared by a repeated coupling of BNPE8 with tetraethylene
glycol.
Results and Discussion
A. Surface Tension Measurements. The cmc for the nonyl
phenol ethoxylates was determined by surface tension measure-
ments. Figure 3 shows the variation of surface tension with log
[concentration] for the fully hydrogenated surfactants. The
reading of surface tension was found to vary with time for all
the nonionic surfactants studied in this work. It typically took
about 5-10 min for the adsorption to reach a constant value at
the concentration around the cmc, while over the dilute
concentration range, it took some 30 min for the saturated
adsorption to be reached. Time-dependent adsorption for this
type of surfactants was also observed by Crook et al.4 and Schick
et al.5 Our observed time scale for reaching constant surface
tension values is broadly consistent with that reported by these
Alkyl chain deuterated samples were made using the same
procedure, except the starting Grignard reagent was made from

Branched Nonyl Phenol Oxyethylene Glycols
J. Phys. Chem. B, Vol. 104, No. 7, 2000 1509
Figure 2. Proton NMR spectra for BNPE4.
Figure 4. Plots of log[cmc] versus n for OPEn (O), C12En (4), and
BNPEn (+). The straight lines through each set of data are the best
fits.
Figure 3. Surface tension plots for the branched nonyl phenol
ethoxylates with the number of the ethoxylate units equal to 4 (4), 8
(+), and 12 (b). The dashed lines are the limiting slopes at the cmc
converted from the surface excesses determined from neutron reflection.
concentration under these conditions. Nevertheless, the limiting
values of surface excess (Γcmc) for these surfactants can still be
estimated from the tension plots by taking the limiting slopes
just below the cmc. These values can then be directly compared
with the surface excesses obtained from neutron reflection. The
dashed lines shown in Figure 3 represent the neutron surface
excesses measured at the cmc for each of the surfactants. These
straight lines are in reasonable agreement with the limiting
slopes of the surface tension plots at the cmc, suggesting that
the surface excesses from both surface tension measurements
and neutron reflection are in a reasonably good agreement.
There are no cmc values for these surfactants in the literature
for comparison. It is, however, useful to note that Crook et al.4
have measured surface tension for a series of pure octyl phenol
ethoxylates (abbreviated to OPEn) and the commercial Triton
series at the air-water and oil-water interfaces. Comparison
of our data with the results from their pure para-substituted OPEn
will enable us to examine both the effect of the number of
carbons in the alkyl chain (m) and the effect of the length of
ethoxylate groups (n) on the cmc. Both sets of alkyl phenol
results are shown in Figure 4. It can be seen from Figure 4 that
an increase in m by one has resulted in the decrease of the cmc
by almost a factor of 10. This trend is consistent for all the
surfactants studied with different sizes of ethoxylate headgroups.
Klevens20 has shown that for many surfactants containing
straight alkyl chains as their hydrophobic groups, the variation
authors. The results shown in Figure 3 are the equilibrated
values. The absence of a minimum around the cmc in each
tension curve is consistent with the samples prepared being of
high purity. The cmc was found to be (1.1 ( 0.3) × 10-5
M
for h-BNPE4, (4.0 ( 0.3) × 10-5 M for h-BNPE8 and (8.0 (
0.3) × 10-5 M for h-BNPE12. The surface tension plots for the
corresponding chain deuterated surfactants were found to be
identical to the hydrogenated ones, within an error margin of 1
mN m-1. The agreement was better in the region of the cmc,
suggesting that both deuterated and hydrogenated surfactants
were of similar purity. There was no observable difference in
the surface activity, caused by either the traces of impurities or
the possible difference in surface activity arising from the
deuterium labeling.
We have previously shown that analysis of the surface tension
data below the cmc using the Gibbs equation leads to a reliable
estimate of surface excess (the adsorbed amount) at different
bulk surfactant concentrations,13,15 provided that measurements
are made over a sufficiently wide range of surfactant concentra-
tion to ensure that accurate slopes are obtained. For the
surfactants studied in this work, the cmc is relatively low and
when the bulk concentration is below 5 × 10-6 M the
reproducibility in the tension measurement was found to be poor.
The surface tension data are thus not accurate enough to render
a reliable derivation of the variation of surface excess with bulk

1510 J. Phys. Chem. B, Vol. 104, No. 7, 2000
Green et al.
of the cmc with the length of the alkyl chain follows a half
logarithmic relation: that is, log[cmc] decreases linearly with
m. Meguro et al.21 have examined the cmc variation of the CmEn
series with the alkyl chain length and found that the slope of
the plot of log[cmc] versus m is about -0.5, suggesting that
with an increase of one carbon in the alkyl chain, the cmc is
also decreased by a factor of 10, an observation consistent with
that from the alkyl phenol surfactants. It should, however, be
noted that for most of the straight chain ionic surfactants the
slope was found to be around -0.3, showing that an increase
of one carbon decreases the cmc by a factor of 5.22 Because
the dependence of the cmc on the number of carbons in the
hydrophobic group is the same for CmEn, OPEn, and BNPEn,
the results thus suggest that the inclusion of the phenol group
does not alter the rate of decrease of the cmc on the number of
carbons in the hydrophobic chain. It is also interesting to
mention that while the alkyl chain in CmEn is straight, the alkyl
chains in APE samples are branched, and the branching between
the OPEm and NPEm surfactants is different. Such differences
do not appear to affect the rate of variation of the cmc on m
either.
The dependence of the cmc on the size of the hydrophilic
headgroups for nonionic surfactants has also been discussed in
the literature.22,23 An empirical linear relationship between log-
[cmc] and n has also been proposed, but the slope of the plot
in this case is positive and is much smaller than the absolute
value corresponding to the plot of log[cmc] versus m, showing
that the increase of cmc with n is much slower than its decrease
with m. The fact that two sets of the data for BNPEn and OPEn
in Figure 4 can be approximated to two parallel straight lines
suggests that the proposed linear relation holds and the
difference in the branching of alkyl chains does not affect the
dependence of the cmc on n. The slope of the two lines was
found to be about 0.10. For comparison, the variation of the
cmc for C12En series is also shown in Figure 4 and the slope of
the best linear fit to the data was found to be 0.03, showing
that the cmc for APE series increases 3 times faster with n than
that for the CmEn series. However, it should be noted that both
Hsiao et al.24 and Crook et al.4 have shown that the linear
relation for APE series only holds when n is less than 10, beyond
which the rate of increase in cmc with n tends to slow. This
trend is almost visible in Figure 4 in that the data do not exactly
follow a linear relation with n.
Figure 5. Surface excess plots for the branched nonyl phenol
ethoxylates with n ) 4 (b), 8 (+), and 12 (4). The arrows indicate the
cmc and the solid lines are to guide the eye.
specular reflectivity will arise from the interface. This mixed
water is usually termed null reflecting water (NRW). When a
deuterated surfactant is adsorbed on the surface of NRW, the
surfactant is the only species that contributes to the specular
reflectivity. Under these conditions, and if the adsorbed layer
is assumed to be uniform, the area per molecule A is given by
m b
∑
i
i
A )
(3)
Fτ
where mi is the number of ith atom with scattering length bi
and τ is the thickness of the layer. The relationship between A
and surface excess Γ is
1
ANa
Γ )
(4)
where Na is the Avogadro’s constant. In fitting the reflectivity
profiles with a uniform layer model, it is usually found that τ
and F can be varied over a limited range, but the variations
cancel in their contribution to A so that A is to a good
approximation independent of the assumption that the layer is
uniform.
B. Surface Excess from Neutron Reflection. The surface
excess can be determined more reliably in neutron reflection
and the principle can be outlined as follows.
In a neutron reflection experiment the intensity of the
incoming beam (Io) and that of the exiting beam (I) are measured
and the ratio of I/Io is known as neutron reflectivity (R(κ)). In
a specular reflection R(κ) is a function of the scattering length
density distribution (F(z)) normal to the interface,25,26 which
depends on the number densities of each atom species ni and
their known scattering lengths bi:
The surface excesses for BNPEn have been obtained by
performing neutron reflectivity measurements using the alkyl
chain deuterated surfactants in NRW. Figure 5 shows the
resulting surface excesses as a function of bulk surfactant
concentration in NRW at 25 °C. The arrows in Figure 5 mark
the cmc for each of the surfactants. An interesting feature of
Figure 5 is that the change of surface excess with concentration
becomes more gradual as n increases. The same trend has been
observed for the monodisperse C12En series.17 The slower
variation of surface coverage and surface tension with bulk
concentrations of BNPEn and C12En at large values of n
resembles the behavior of poly(ethylene oxide) (PEO),27 sug-
gesting that as the size of the headgroups increases the effect
of the hydrophobic groups becomes less significant. A further
interesting feature is that the limiting surface excess above the
cmc decreases with n, showing that the limiting area per
molecule increases with the size of the headgroups. Acmc was
F ) n b
(2)
∑
i
i
Since the value of bi varies from isotope to isotope, the use of
isotopic labeling can alter neutron reflectivity for a given
chemical structure. For a hydrocarbon system, the most straight-
forward application is the substitution of hydrogen by deuterium.
Since the scattering lengths of D and H are of opposite sign,
the scattering length density of water can be varied over a wide
range. For example, if one mole of D2O is mixed with 11 mol
of H2O, the mixed water will have zero scattering length. When
neutron reflection is performed at the air-water interface no

Branched Nonyl Phenol Oxyethylene Glycols
J. Phys. Chem. B, Vol. 104, No. 7, 2000 1511
TABLE 1: Physical Constants Used in Data Analysisa
fragment
∑bi × 105/Å volume/Å3 F × 106/Å-2
lc
(C4D9)2CH-
(C4H9)2CH-
-C6H4-
161.2(92%D)
-11.2
24.9
19.4
35.9
265
265
160
260
510
760
30
6.1
-0.4
1.6
0.75
0.7
0.7
7.8
7.8
3
15.5
30
-(OC2H4)4OH
-(OC2H4)8OH
-(OC2H4)12OH 52.5
H2O
D2O
44
-1.68
19.1
-0.56
6.35
30
a The values of scattering lengths were from ref 33; the fragment
volumes and fully extended lengths were from refs 34 and 35; lc denotes
the fully extended length.
TABLE 2: Structural Parameters for BNPEm Obtained
from Uniform Layer Model Fittinga
isotope
A/Å2
τ(σ′c) ( 2/Å
n
d-BNPE4/NRW
d-BNPE4/D2O
h-BNPE4/D2O
d-BNPE8/NRW
d-BNPE8/D2O
h-BNPE8/D2O
d-BNPE12/NRW
d-BNPE12/D2O
h-BNPE12/D2O
46 ( 3
46 ( 3
46 ( 3
62 ( 4
61 ( 4
61 ( 4
73 ( 5
76 ( 5
75 ( 5
22(20)
23
12.5
24(22)
25.5
17
27(24)
28
21
-
7 ( 1
7 ( 1
-
16 ( 1
16 ( 1
-
Figure 6. Neutron reflectivity profiles for the alkyl chain deuterated
nonyl phenol ethoxylates in NRW with n ) 4 (b), 8 (+), and 12 (4).
The continuous lines are calculated using uniform layer model and the
structural parameters are given in Table 2.
24 ( 2
24 ( 2
a n denotes the number of water molecules required to give
appropriate scattering length density to fit the reflectivity profile at a
given isotopic substitution.
found to be 46 ( 3 Å2 for BNPE4 as compared with 44 ( 3 Å2
for C12E4, 61 ( 4 Å2 for BNPE8 as compared with 62 ( 3 Å2
for C12E8, 75 ( 5 Å2 for BNPE12 as compared with 72 ( 5 Å2
for C12E12. The almost identical Acmc obtained for the same size
of the ethoxylate headgroups suggests that the change in the
structure of the hydrophobic chain in the form of the alkyl chain
branching and the insertion of the benzyl chain has negligible
effect on the limiting area per molecule. Acmc must be
predominantly determined by the size of the ethoxylate head-
groups.
C. Structure Determination. Although the difference in the
chemical configuration of the hydrophobic chains does not affect
the limiting area per molecule at the cmc, it is interesting to
examine if such difference affects the distributions of different
fragments within the adsorbed layers. For each surfactant studied
in this work, reflectivity profiles were measured under three
different isotopic contrasts: the alkyl chain deuterated in NRW
and in D2O, and the fully hydrogenated in D2O. These
measurements allow the determination of the distributions of
the hydrophobic groups and their locations relative to the water
across the interface. Although these measurements do not
include the use of headgroup deuterated surfactants, some useful
information about the distributions of the ethoxylate headgroups
can still be gained, as will be discussed later.
Neutron reflectivity is often analyzed using the optical matrix
method.28 The procedure for extracting structural information
is through model fitting to the different reflectivity profiles. A
structural model is usually assumed for an interface and its
corresponding reflectivity is then calculated using the optical
matrix formula. The calculated reflectivity is then compared
with the measured one and the structural parameters are
modified in a least-squares criterion until a good fit is obtained.
The parameters directly used in the calculation are mainly the
thicknesses of the layers τ and the corresponding scattering
length densities F, from which A and Γ can be obtained through
eqs 3 and 4. Figure 6 shows the best fits to the measured
reflectivity profiles at the cmc for the three alkyl chain
deuterated surfactants in NRW assuming uniform layer distribu-
tions. The scattering lengths and volumes for different fragments
used in the fitting are given in Table 1 and the resultant structural
parameters are listed in Table 2. The calculated reflectivity
profiles fit all the three measured ones well within the
experimental error, suggesting that the uniform layer model is
a reasonable approximation for describing the distributions of
the layers in NRW. The fitting to the reflectivity profiles shown
in Figure 6 gives the thicknesses of 22 ( 3 Å for BNPE4, 24 (
3 Å for BNPE8, and 27 ( 3 Å for BNPE12. Since the scattering
length from the hydrogenated ethoxylate groups is relatively
small, the thicknesses obtained mainly represent the distributions
of the nonyl phenol groups with a small fractional contribution
from ethoxylate headgroups.
Although the hydrogenated ethoxylate groups in NRW are
hardly visible, the situation is different if the measurement is
made in D2O. The association of D2O molecules with the
ethoxylate groups serves to highlight the distributions of the
headgroups. Since the scattering length density for one ethoxy-
late unit is close to zero (see Table 1) and its volume is
equivalent to two water molecules, the association of two D2O
molecules to each ethoxylate unit will produce an average
scattering length density that is about half of the value for the
pure D2O. When a fully hydrogenated surfactant is adsorbed at
the surface of D2O, the region above water will have an average
scattering length density close to zero and is virtually invisible.
The measurement with this contrast offers reliable information
about the distribution of the ethoxylate groups. Figure 7 shows
the measured reflectivity profiles for the fully hydrogenated
surfactants in D2O. The reflectivity from the surface of pure
D2O is also shown (dashed line) for comparison. The large
difference between the water profile and those in the presence
of fully hydrogenated surfactants shows that the measurement
in D2O is quite sensitive to the distributions of the ethoxylate
heads. As the size of the ethoxylate headgroups becomes greater,
the reflectivity profiles decay faster, suggesting that the thickness
of the layer mixed with water is increased. The continuous lines
are the best fit assuming uniform layer distributions for the
headgroups. The good consistency between the calculated curves
and the measured ones indicates that to a good approximation

1512 J. Phys. Chem. B, Vol. 104, No. 7, 2000
Green et al.
Figure 8. Neutron reflectivity profiles for the nonyl chain deuterated
phenol ethoxylates in D2O with n ) 4 (b), 8 (+), and 12 (4). The
continuous lines are calculated using uniform layer model and the
structural parameters are given in Table 2. The dashed line is the
calculated reflectivity from the surface of pure D2O.
Figure 7. Neutron reflectivity profiles for the fully hydrogenated nonyl
phenol ethoxylates in D2O with n ) 4 (b), 8 (+), and 12 (4). The
continuous lines are calculated using uniform layer model and the
structural parameters are given in Table 2. The dashed line is the
calculated reflectivity from the surface of pure D2O.
chains. As both the hydrophobic and the hydrophilic groups
contribute to the reflectivity, the thickness of the layer offers a
good measure of the dimension for the whole layer. Indeed,
we have found that in the case of C12En the thickness obtained
from the chain deuterated surfactant in D2O is within 3 Å
identical to that of the fully deuterated surfactant in NRW.15,16
Figure 8 shows a similar trend of depression of reflectivity with
the size of the headgroups to that observed in Figure 7, except
the distributions of the headgroups are well represented by the
uniform layer model.
The increase in the thickness of the water soluble layer may
arise from two contributing factors: the broadening of the
ethoxylate group layer with n and the increased extent of
immersion of the whole surfactant layer into water as a result
of the increased affinity of surfactant molecules toward water.
The latter has been shown to be unlikely for C12En layers,26,29
because the distances between the centers of headgroup distribu-
tions and those of water were always found to be the same
within (2 Å. Hence, the thickness obtained under this water
contrast must represent the distributions for the ethoxylate
groups. The values were found to be 16 ( 2 Å for BNPE4, 18
( 2 Å for BNPE8, and 21 ( 2 Å for BNPE12, showing that as
the length of the headgroup is tripled, the thickness of the water-
miscible layer is only increased by some 30%. The scattering
length densities for the layers are all between 3.1 × 10-6 Å-2
to 3.6 × 10-6 Å-2, and are close to the half value of F for bulk
D2O. The immersed layers are hence well highlighted by the
underlying D2O subphase. The actual number of water molecules
associated with each headgroup can be obtained from the values
of F. Each ethoxylate group was found to have two hydrated
water molecules, an observation entirely consistent with previous
results for C12En series.17 The physical significance of this value
can be assessed by checking the volume constraint within the
immersed layer. For example, the volume of E4 in BNPE4 is
about 250 Å3 and that for eight water molecules is 240 Å3,
making a total of 490 Å3, as compared with the value of 575
Å3 (12.5 Å × 46 Å2) for the total space available. The difference
is likely to be caused by the mixing of the hydrophobic
fragments into the headgroup as will be discussed later. A similar
trend was observed for BNPE8 and BNPE12. It should be noted
that such treatment is only approximate as the model itself is
very coarse.
that the extent of depression is more intensified. The sharp
-1
interference fringes occur at a smaller κ close to 0.1 Å
,
consistent with the whole layers being thicker than the regions
immersed in water. The continuous lines shown in Figure 8
represent the uniform layer fitting and the thickness of the layers
was found to be 23 ( 3 Å for BNPE4, 25.5 ( 3 Å for BNPE8,
and 28 ( 3 Å for BNPE12. These values together with those
obtained from the other two contrasts provide a brief account
of the dimensions of the whole layer and the hydrophobic and
hydrophilic regions across the interface.
We have previously shown that neutron reflectivity profiles
can be alternatively analyzed using kinematic approximation.26
In the kinematic approach the reflectivity profiles are usually
converted into the structural factors, hii and hij, from which
structural parameters are derived more directly. For a surfactant
layer adsorbed on the surface of water, the scattering length
density profile can be written as
F(κ) ) bana(z) + bwnw(z)
(5)
where a and w refer to surfactant and water. The reflectivity
can be approximately expressed as25,26
16π2
κ2
2
R(κ) )
[ba2haa(κ) + bw hww(κ) + 2babwhaw(κ)] (6)
The single uniform layer model also fits the reflectivity
profiles of chain deuterated surfactants in D2O, as shown in
Figure 8. This is because the association of D2O with the
hydrogenated headgroups makes the scattering length density
of the headgroup region comparable to that of the deuterated
where haa(κ) and hww(κ) are the partial structure factors
describing the distribution of the surfactant and that of the water,
respectively, and haw(κ) is the partial structure factor describing
the relative location between the surfactant layer and water
distribution across the interface. For alkyl chain deuterated

Branched Nonyl Phenol Oxyethylene Glycols
J. Phys. Chem. B, Vol. 104, No. 7, 2000 1513
structure factor only contains information about the distribution
for a given component at the interface. Uncertainty caused by
coupling between different structural parameters is greatly
reduced.
The layer fitting based on the optical matrix method does
not offer any information about the relative locations between
any given pair of components across the interface. Such
information can be obtained directly from kinematic approach.
Equation 6 can be modified so that the combined reflectivity
measurements under the three different isotopic contrasts can
be used to determine the cross term in the partial structure factors
between the center of the hydrophobic chain and that of water
and the new equation is given as follows
16π2
κ2
[bc′2hcc′(κ) + bw hww(κ) + 2bcbwhcw′(κ)]
(9)
2
R(κ) )
where “′” again indicates the contribution from the hydrophilic
headgroup. The cross distance between the center of the
hydrophobic region to that of water, δcw, is obtained from the
following equation26,30
Figure 9. Plots of ln[hcc′] vs κ2 for the alkyl chain deuterated nonyl
phenol ethoxylates in NRW with n ) 4 (b), 8 (+), and 12 (4). The
continuous lines are calculated using eq 8 and the structural parameters
are given in Table 2.
h ′(κ) ) - h ′(κ)hww(κ) sin(κδcw)
(10)
x
cw
cc
The assumption relating to eq 10 is that the distribution of the
hydrophobic region is symmetrical and is close to an even
function and that the solvent resembles an odd function. The
validity of the assumption has already been previously assessed30
and for surfactant monolayers the accuracy of δcw is usually
very high. The values of δcw were found to be 11.5 ( 1 Å for
BNPE4, 12.5 ( 1 Å for BNPE8, and 13 ( 1 Å for BNPE12.
The presence of surfactant at the surface of water also affects
the water distribution at the interface. The decay of water density
can be represented by a tanh function26 and hww(κ) can be written
as
surfactant in NRW, bw ) 0 and eq 6 becomes
R(κ)κ2
hcc′ )
(7)
16π2bc′
where c denotes the hydrophobic chain and “′” indicates that
there is a contribution of scattering length from the hydrogenated
ethoxylate group to the hydrophobic chain. Although the number
density distribution can be obtained from hcc′ through Fourier
transformation it is usually more convenient to use analytical
expressions. For example, if a Gaussian distribution is used to
represent surfactant distribution the following relationship exists
between hcc′ and κ26
h
ww(κ) ) nwo2(πκδcwê/2)2csch2(πêκ/2)
(11)
where ê is the width parameter of the tanh function and nwo is
the number density of bulk water. For a pure water surface, ê
is about 2 Å. The penetration of the surfactant headgroups
disrupts the distribution of water, resulting in a broader width
of water density profile; and accordingly, ê is much greater.
The values of ê were found to be 8 ( 1 Å for BNPE4, 10 ( 1
Å for BNPE8, and 13 ( 1 Å for BNPE12, showing that the
adsorption of surfactant has significantly disturbed water
distribution.
The analysis described above offers no direct information
about the distribution of the ethoxylate head and how it mixes
with water. We have previously shown that useful information
relating to the headgroup distribution can be extracted with
measurable reliability from the simultaneous least-squares fitting
to the three reflectivity profiles using the analytical expressions
to represent the distributions of the three components and the
relations between them. In this analysis, the distributions of the
hydrophobic chains and the ethoxylate headgroups are repre-
sented by the Gaussian function and that of water by tanh
function. The relations between them can be described by eq
10 except for the relative location between the hydrophobic
chain and the hydrophilic headgroup. The term of sin(κδcw)
should be replaced by cos(κδcw) because both distributions are
now even functions.26 The combined equations lead to five
unknown parameters: the thicknesses of the nonyl phenol layer
and the ethoxylate head, the width parameter for water distribu-
tion, and the distances between the hydrophobic layer and water,
and between the head layer and water. Table 3 lists the
ln hcc′(κ) ) 2ln Γ - σc2κ2/8
(8)
where σc is the full width at the height of n°c/e (n°c is the
maximum number density). Thus the surface excess is obtained
from the intercept of the straight line plot of ln hcc′(κ) versus
κ2 and the thickness of the layer from the slope. Figure 9 shows
the plots of ln[hcc′(κ)] versus κ2 for the three surfactants at the
cmc. The straight lines through each set of the measured data
were calculated using eq 8. The intercept gives Γ, which is
within the experimental error consistent with the values obtained
from the optical matrix method, as described previously. σc′
was found to be 20 ( 3 Å for BNPE4, 22 ( 3 Å for BNPE8,
and 25 ( 3 Å for BNPE12. These values are on average about
90% of the corresponding values obtained from the uniform
layer model, and the difference is entirely due to the definition
of the two models.26 The good fit of eq 8 to the experimental
data shown in Figure 9 shows that the Gaussian distribution is
equally good for describing the distribution of the hydrophobic
region. Although the measured reflectivity data do not allow a
clear distinction to be made between the two models, Gaussian
distribution is more realistic because of the effect of roughness
and the structural disorder within the surfactant layer, as will
be further discussed later. Although model fitting is also used
in the kinematic approach, fitting to the partial structure factors
is more direct than to the reflectivity profiles because each partial

1514 J. Phys. Chem. B, Vol. 104, No. 7, 2000
Green et al.
TABLE 3: Structural Parameters for BNPEm and C12Em at
Their cmcs
parameters
BNPE4/C12E4
BNPE8/C12E8
BNPE12/C12E12
A/Å2
46 ( 3/44 ( 3
18/16.5
14/14
11/9.5
0.3/2
61 ( 3/62 ( 3
19/15
18/19
13/11
1/0.5
72 ( 5/75 ( 5
22/15
22/21.5
14/13
1/1
13/12
σc ( 2/Å
σh ( 4/Å
δcw ( 1/Å
δhw ( 1/Å
ê ( 1/Å
8/7.5
10/9
parameters obtained and the values of δcw, ê, and σc are
consistent with those obtained from the direct analysis based
on eq 9. The values of δhw are always around 1 ( 1 Å, showing
that the centers of the headgroup distributions are almost at the
same positions as those for the water. The sensitivity of the
fitting to the structural parameters was tested by varying a given
parameter while fixing the others at their optimal values. It was
found that the reflectivities were sensitive to the variation in
δcw, δhw, ê, and σc, but insensitive to σh. The errors quoted in
Table 3 reflect the range of variation over which no obvious
difference between the measured reflectivities and the fitted ones
was observed. It should be pointed out that such a treatment
can only provide some approximate estimate of the distributions
of the headgroup and its location relative to other components.
More reliable information on σh and δhw can only be obtained
by measuring the reflectivity profiles using the head deuterated
samples.
The distributions of different fragments in terms of volume
fractions along the surface normal are plotted in Figure 10 for
all three surfactants. Gaussian distributions have been used for
the hydrophobic chain and the head, and tanh function for the
distribution of water. The extent of mixing between different
fragments can be seen from the overlapped areas in Figure 10.
The distribution of the total volume fraction is also plotted for
each surfactant and the variation is within (10% when
approaching the bulk density, showing a good self-consistency
between the data. We have previously discussed that the main
errors arise from the insensitivity of the reflectivities to the shape
of the distributions, e.g., Gaussian distributions for the hydro-
phobic chain and hydrophilic head, and tanh function for water.26
Although the trend of water distribution resembles a tanh
function, the exact shape may not follow a tanh profile precisely.
The deviation is usually the main source of error contributing
to the variation in the total volume fraction distribution.
The fully extended length of the branched nonyl phenol group
is about 10 Å. The thicknesses for the nonyl phenol layers
obtained are around 20 Å, showing that the hydrophobic groups
are twice as thick as their fully extended length. As the size of
the headgroups increases the width of the hydrophobic layers
also increases, clearly showing that as n increases the distribu-
tions of the hydrophobic chains are further broadened. The
thicknesses measured on the surface of water contain contribu-
tions arising from capillary wave roughness and structural
disorder. The relationship between the intrinsic thickness of the
surfactant monolayer in the direction normal to the surface (lz)
and the roughness (ω) can be approximately expressed by29
Figure 10. Volume fraction distributions for BNPE4 (a), BNPE8 (b),
and BNPE12 (c) at their cmcs. The continuous lines denote the alkyl
chain; the long dashed lines denote the headgroups; the short dashed
lines denote the water and the long and short dashed lines denote the
total volume fraction distributions. The reference points are at the
centers of the hydrophobic chain distributions.
σ2 ) lz2 + ω2
(12)
The value of lz depends on contributions from the extended
component lengths, different chain conformations, and the
average distribution of orientations with respect to the surface
normal. There are two types of roughness: structural disorder
within the layer and thermal roughness arising from capillary
waves. The contribution of thermal roughness can be estimated
using the relationship developed by Schwartz et al.31 for the
surface of pure liquid. Schwartz et al.31 have shown that the
capillary wave amplitude of a pure liquid varies inversely as
the square root of surface tension. For example, the surface
tension of BNPE4 at its cmc is 29 mN m-1, and this will give
a capillary wave roughness of about 4.4 Å. Since unit mean
square amplitude in terms of capillary waves is equivalent to
2.3 units in terms of the distributions used in the present work,
the thermal roughness of 4.4 Å at the cmc is then equivalent to

Branched Nonyl Phenol Oxyethylene Glycols
J. Phys. Chem. B, Vol. 104, No. 7, 2000 1515
10 Å in our analysis and will contribute to the broadening of
the whole monolayer in the form as described in eq 12.31 As
the thickness for the hydrophobic chain is 18 Å, removal of 10
Å according to eq 12 leads to a length of 14.5 Å projected onto
the surface normal. This thickness is still much greater than
the value of about 10 Å for the fully extended length of the
branched hydrophobic chain. The result hence implies consider-
able structural roughness within the surfactant layer.
the distribution of the hydropbobic layer, but does not alter the
limiting area per molecule at the cmc. The much thicker
hydrophobic layer caused by the branching of the alkyl chain
and the insertion of the benzyl ring underlines a large extent of
mixing between the hydrophobic groups and the ethoxylate
headgroups, and is indicative of a high affinity of the hydro-
phobic alkyl phenols toward the hydrophilic ethoxylate head-
groups.
The effect of alkyl chain length on the thickness of the
adsorbed surfactant layers has been examined in our previous
studies on the adsorption of the cationic series Cm TAB on the
surface of water.32 It was found that when m ) 12 the thickness
of the alkyl chain layer is effectively the same as the fully
extended length. As the chain length is increased, the alkyl chain
layers become thinner than the fully extended lengths. For
example, the thickness of the alkyl chain layer formed by C18
TAB at its cmc is about 16.5 ( 1 Å and the fully extended
length is 24.3 Å, showing that the layer is dense and is
effectively tilted. However, in the case of C8 TAB whose fully
extended alkyl chain length is 10.5 Å, the alkyl chain layer was
found to be 15.5 Å, suggesting an increased level of structural
disorder within the adsorbed layer. In a separate study of the
adsorption of short chain alcohols on the surface of water we
also found that for both ethanol and butanol the adsorbed layers
are thicker than the fully extended lengths of the alcohol
molecules.32 For the nonionic surfactant series, we have only
studied the C12En series. Hence the possible effect of the length
of the alkyl chains on the distributions of the fragments in the
layers is not well established. Nevertheless, for all the C12En
species studied, the thickness of the dodecyl chain layer is
always less than its fully extended length. This is especially so
when the size of the ethoxylate group is large. In comparison
with these results it appears that the behavior of the branched
nonyl phenol ethoxylates resembles those with short alkyl
chains. It is not clear at this stage how the inclusion of the benzyl
ring affects the distribution of different fragments within the
adsorbed layer, but what is clear is that the insertion of the
benzyl ring has dramatically reduced the cmc for this series of
surfactants.
It is useful to compare BNPEn distributions with those of
C12En layers at the same n. The comparison is given in Table
3. It can be seen that although there are differences between
cross distances between the centers of distributions, the differ-
ences are mostly within 2 Å. For both series of surfactants, the
extent of mixing increases with the size of the headgroups. All
the width parameters are comparable except the thickness of
the hydrophobic layers. The thicker hydrophobic chain region
within BNPEn layers clearly indicates a stronger mixing between
the hydrophobic chains and the hydrophilic headgroups, and
between the hydrophobic chains and water. The increased
mixing may arise from the presence of phenol groups that offer
high affinity to the ethoxylate headgroups, but it should be
remembered that in comparison with the straight dodecyl chains,
the nonyl phenol groups differ in the length of alkyl chain, the
branching and the insertion of benzyl ring. Further work is
required to disentangle the effects of these structural differences
on the extent of mixing within the adsorbed layers.
Acknowledgment. We thank Dr. R. K. Thomas at the
Department of Physical and Theoretical Chemistry, University
of Oxford for many valuable discussions. We also thank the
U.K. Engineering and Physical Sciences Research Council
(EPSRC) for support.
References and Notes
(1) Makino, S.; Reynolds, J. A.; Tanford, C. J. Biol. Chem. 1973, 248,
4926.
(2) Rubin, M. S.; Tzagoloff, A. J. Biol. Chem. 1973, 248, 4269.
(3) Garland, R. C.; Cori, C. F. Biochemistry 1972, 11, 4712.
(4) Crook, E. H.; Fordyce, D. B.; Trebbi, G. F. J. Phys. Chem. 1963,
67, 1987.
(5) Schick, M. J.; Atlas, S. M.; Eirich, F. R. J. Phys. Chem. 1962, 66,
1326.
(6) Almgren, M.; Lofroth, J.; Stam, J. J. Phys. Chem. 1986, 90, 4431.
(7) Greenwald, H. L.; Kice, E. B.; Kenly, M.; Kelly, J. Anal. Chem.
1961, 33, 465.
(8) Kubeck, E.; Naylor, C. G. J. Am. Oil Chem. Soc. 1990, 67, 400.
(9) Podo, F.; Ray, A.; Nemethy, G. J. Am. Chem. Soc. 1973, 95, 6164.
(10) Schott, H. J. Colloid Interface, Sci. 1995, 173, 265.
(11) Schott, H. J. Colloid Interface, Sci. 1997, 192, 458.
(12) Schott, H. J. Colloid Interface, Sci. 1998, 205, 496.
(13) Lu, J. R.; Li, Z. X.; Su, T. J.; Thomas, R. K.; Penfold, J. Langmuir
1993, 9, 2408.
(14) Lu, J. R.; Hromadova, M.; Simister, E.; Thomas, R. K.; Penfold,
J. J. Phys. Chem. 1994, 98, 11519.
(15) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Staples, E. J.; Thompson, L.;
Tucker, I.; Penfold, J. J. Phys. Chem. 1994, 98, 6559.
(16) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Staples, E. J.; Tucker, I.;
Penfold, J. J. Phys. Chem. 1993, 97, 8012.
(17) Lu, J. R.; Su, T. J.; Li, Z. X.; Thomas, R. K.; Staples, E. J.; Tucker,
I.; Penfold, J. J. Phys. Chem. 1997, 101, 10332.
(18) Furness, B. S.; Hannaford, A. J.; Smith, P. W. G.; Tatchell, A. R.
Vogel’s Textbook of Practical Organic Chemistry, 5 ed.; Langman: London,
1989.
(19) Penfold, J.; Richardson, R. M.; Zarbakhsh, A.; Webster, J. W. P.;
Bucknall, D. G.; Rennie, A. R.; Jones, R. A. L.; Cosgrove, T.; Thomas, R.
K.; Higgins, J. S.; Fletcher, P. D. I.; Dickinson, E.; Roser, S. J.; McLure,
I. A.; Hillman, A. R.; Richards, R. W.; Staples, E. J.; Burgess, A. N.;
Simister E. A.; White, J. W. J. Chem. Soc., Faraday Trans. 1997, 93, 3899.
(20) Klevens, H. B. J. Am. Oil Chem. Soc. 1953, 30, 74.
(21) Meguro, K.; Takasawa, Y.; Kawahashi, N.; Tabata, Y.; Ueno, M.
J. Colloid Interface Sci. 1981, 83, 50.
(22) Rosen, M. J. Surfactants and Interfacial Phenomena, 2 ed.;
Wiley: New York, 1989.
(23) Myers, D. Surfaces, Interfaces and Colloids: Principles and
Applications; VCH Publishers: New York, 1991.
(24) Hsiao, L.; Dunning, H. N.; Lorenz, P. B. J. Phys. Chem. 1956, 60,
657.
(25) Crowley, T. L. Physica 1993, A195, 354.
(26) Lu, J. R.; Lee, E. M.; Thomas, R. K. Acta Crystallogr. 1996, A52,
11.
(27) Lu, J. R.; Su, T. J.; Thomas, R. K.; Penfold, J.; Richards, R. W.
Polymer 1996, 37, 109.
(28) Born, M.; Wolf, E. Principles of Optics; Pergamon: Oxford, 1970.
(29) Lu, J. R.; Simister, E. A.; Thomas, R. K.; Penfold, J. J. Phys.:
Condens. Matter 1994, 6, A403.
Conclusions
(30) Lu, J. R.; Simister, E. A.; Lee, E. M.; Thomas, R. K.; Rennie, A.
R.; Penfold, J. Langmuir 1992, 8, 1837.
Changes in the chemical configuration of the hydrophobic
chain affects the adsorption and aggregation properties of
nonionic surfactants. Although the rate of decrease of the cmc
with the increase of carbon number in the hydrophobic chains
is the same for CmEn and APE surfactants, the cmc of APEs
increases 3 times faster with the length of the ethoxylate units.
The change in the structure of the hydrophobic chain affects
(31) Schwartz, D. K.; Schlossman, M. L.; Kawamoto, E. H.; Kellogg,
G. J.; Pershan, P. S.; Ocko, B. M. Phys. ReV. 1990, A41, 5687.
(32) Lu, J. R.; Thomas, R. K. J. Chem. Soc., Faraday Trans. 1998, 94,
995.
(33) Sears, V. F. Neutron News 1992, 3, 26.
(34) Tanford, C. J. J. Phys. Chem. 1972, 76, 3020.
(35) Takahashi, Y.; Sumita, I.; Tadokoro, H. J. Polym. Sci. 1973, 11,
2113.