Colloids 2013

Colloidal Systems
Frank D. Blum
Oklahoma State University, Stillwater, OK, USA
NMR experiments on colloidal systems (ca 10–1000 nm in size) are discussed with a special emphasis on the species that interact with
colloids. Many NMR experiments can be used on these systems; from relaxation times to diffusion to solid-state experiments. The results show
how different techniques can be used to size colloidal materials and emulsions, understand the interactions of colloidal particles with small
molecules, polymers, or surfactants, and characterize the interactions of polymers with surfactants. NMR experiments seem well suited for the
characterization of these interfacial systems as long as the particles have sufficient surface area for the NMR experiments to have the necessary
sensitivity.
Keywords: NMR, colloidal systems, particles, polymers, surfactants, emulsions
How to cite this article:
eMagRes, 2013, Vol 2: 427–436. DOI 10.1002/9780470034590.emrstm1328
Introduction
Colloidal systems, because of their size, are systems that are
inherently complex. The complexity comes in the form of
heterogeneity, which can arise in the form of physical (size,
shape, or interface of particles) or chemical (multicomponent
materials) dispersion. This heterogeneity complicates understanding of the behavior of particles; however, it also offers
opportunities for magnetic resonance because of the different
techniques and probes that can be used.1 The applications of
these techniques are very broad and have, especially, significant
application in colloidal systems such as foods,2 oil production
and use,3 and surfactant systems.4,5
The focus of this article is on the use of NMR spectroscopy
of chemically heterogeneous systems (i.e., multicomponent
systems). There are a few areas that may be of additional
interest, but are not discussed here, including the use of ESR to
study the behavior of heterogeneous systems, especially spinlabeled polymers.6,7 In addition, the use of NMR to study
particles has also been reviewed.8 Consequently, this article
reports the study of ‘soft’ materials in colloidal systems. These
topics include dispersions of colloidal particles in fluids, and
particles treated with polymers and surfactants. The goal is
to understand how NMR can be used to characterize these
interfacial materials.
The tools of the NMR spectroscopist sometimes lend themselves nicely to heterogeneous systems, although these do not
necessarily mesh with the capabilities of high-resolution instruments. Virtually all NMR parameters can be used in the study
of colloidal systems. Lineshapes, chemical shifts, and relaxation
effects provide evidence for structure and dynamics (usually
Update based on original article by T. Cosgrove and T. M. Obey, Encyclopedia of Magnetic Resonance, © 1996, John Wiley & Sons, Ltd.
Volume 2, 2013
rotational). External fields and probes can also extend the
studies to probe translational motions, with diffusion allowing
measurements of partitioning and size.
Special Considerations
The unfortunate consequences of Boltzmann distributions in
NMR pose some special problems for colloidal systems. The
fight against low sensitivity is especially apparent when the
species of interest are interfacial. Strategies to overcome these
difficulties are focused on increasing the concentration or sensitivity through field strength, isotopic labeling, or magnetization
transfer. In some favorable cases, however, the results can be
quite stunning in terms of the quality of the spectra from colloidal systems. A few of these special variables will be discussed.
Surface Area
For the colloid scientist, the surface area of particles is perhaps
the most critical property to be understood when plotting out a
characterization strategy, followed by morphology. NMR studies are generally not appropriate for macroscopic substrates
because the substrate consumes much of the sample volume,
leaving little for the adsorbent, although for some porous materials, interstitial volumes can be of high concentration resulting
in good sensitivity for surface studies. Unfortunately, most
macroscopic particles have very small amounts of interfacial
area making them difficult to use as a substrate for NMR studies.
Using particles that are small may facilitate the study of interfacial phenomena. From 1000 to 10 nm, particles in the colloidal
domain can have large interfacial areas, especially compared to
macroscopically sized materials. The specific surface area of a
sphere is proportional to the inverse of its radius. At the lower
size end, for example, a 10 nm sphere of silica, with a density of
© 2013 John Wiley & Sons, Ltd.
427
FD Blum
2.5 g cm−3 , a specific surface of around 120 m2 g−1 is expected.
These types of high surface area particles can be an effective way
to highlight the effects of surface behavior. Particle structure
also plays an important role. The morphology determines the
available surface area and the nature of the interstitial regions,
and depends on several factors such as how the particles are
made, their stability, and their attraction to other particles.
Stability
The stabilities of dispersions are also closely related to their
surface areas. Aggregation/flocculation of the colloidal particles
can dramatically reduce the area available to adsorbing species
such as solvent or polymer. For charged particles in aqueous
dispersions, this problem can often be neglected over the
timescale of an NMR experiment, although settling may be
problematic if the density of the particles is considerably
greater than that of the surrounding dispersion medium. In
nonaqueous dispersions, it is essential that particles be stabilized
in some manner, as a charge stabilization mechanism is usually
inoperative in the low dielectric constant media used in these
formulations. In these cases, stabilization may be achieved by
matching the Hamaker constant of the particle with that of the
dispersion medium, or by adding a suitable polymeric stabilizer
or surfactant.9
The destabilization of particles through aggregation/
flocculation/precipitation causes a reduction of their effective
surface area. In addition, the adsorbed surface layer can
become compressed, desorbed, or otherwise modified by
potentially high degrees of particle packing. This effect is
particularly noticeable in measurements of the linewidths of
the adsorbed species, as the surface species may now have a
considerably reduced mobility, dependent on the exact nature
of the aggregates formed.10
Exchange
In many colloidal dispersions, the continuous phase (e.g.,
water) will be in dynamic exchange between the bulk and
the surface. In the fast-exchange limit, the average relaxation
rate of the solvent will depend on the accessible surface area
and its residence time/interaction mechanism at the particle
surface. The residence time of the solvent will also be sensitive to
competitive solutes such as surfactants and polymers. Examples
of these measurements will be given in later sections. In cases
where the lifetime of the adsorbed species is very long or where
the exchange rate is limited by strong surface interactions
or physical barriers, complex exchange equations must be
employed.11 Direct observation of spectra from surface species
that are not in fast exchange can also be used to probe both the
surface orientation and the dynamics of the adsorbed species.
Susceptibility/Homogeneity/Anisotropy
In any multiphase system, there will be abrupt changes in magnetic susceptibility upon passing from one phase into another.
This may have severe consequences for conventional highresolution NMR studies in terms of resolving and interpreting
the chemical shifts or (in solid-state studies) chemical shift
428
tensors. In such systems, however, these exchange-broadened
resonance lines may be the only route to obtain useful information. In extreme cases, it may be impossible to obtain spectra
from the surface layers.
The anisotropic nature of the mobility of the adsorbed
species must also be considered. If such motion occurs,
detailed quadrupolar relaxation measurements can be used
rather effectively to probe the dynamics of the surface layer. In
many systems, a lack of mobility of the surface-adsorbed phase
means that it is impossible to obtain spectra from that phase
directly using conventional high-resolution spectrometers, and
thus solid-state techniques are required. However, both methods have associated problems, such as the centrifugation of
the colloidal dispersion during magic-angle spinning and the
shortening of rotating frame relaxation times on adsorption.
The consequence is that some cross-polarization experiments
become difficult.
Techniques
In colloidal dispersions, the full panoply of NMR methods can,
in principle, be used to some advantage, although generally the
more conventional methods of relaxation analysis, diffusion,
and high-resolution/solid-state chemical shift measurements
have been employed most extensively. Other approaches such
as imaging have a role to play, although by their very nature
colloidal particles are too small for the spatial resolution currently available, and thus a detailed structural analysis is not
possible. Techniques that operate on the molecular length scale
(e.g., relaxation) are often more appropriate in such cases.
Combined electrophoresis and NMR self-diffusion techniques
may have significant roles to play; although at present these
methods are not widely used, they can be very powerful.12 One
interesting facet of the study of particulate colloidal dispersions
by NMR is the spectrum that arises from the substrate itself.
The methods and techniques that can be used to study the
substrate directly rely on dipolar decoupling; however, below a
certain size (ca 1–2 nm), the rotational diffusion of the particle
itself may be sufficient to narrow the resonance line.
Interactions of Small Molecules with
Colloidal Particles
The average relaxation rate of solvent molecules in colloidal
dispersions consisting of particulate matter is a source of direct
information on the particle structure and the affinity of the
surface for the solvent molecule. Figure 1 shows a set of water
spin–spin relaxation R2sp data expressed as a specific relaxation
rate, R2sp = (R2d − R2s )/R2s , where d relates to the overall colloidal dispersion and s to the bulk solvent. Several interesting
features can be seen in these data. Firstly, data taken from different sized particles of the same substrate (e.g., silica sols) fall on
the same line, demonstrating that NMR can be used to monitor
the total available surface area in a dispersion in situ, which is
not straightforward to achieve by other means. Secondly, and
possibly of more general interest, the gradient and nature of
each line are informative: the straight-line dependencies indicate that the samples satisfy the fast-exchange limit, and the
© 2013 John Wiley & Sons, Ltd.
Volume 2, 2013
Colloidal Systems
8
7
6
R2sp
5
4
3
2
1
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
2
Surface area (m )
Figure 1. Normalized specific solvent (water) relaxation rates for hydrophobic and hydrophilic surfaces as a function of surface area: () alumina; ()
silica; (•) positive polystyrene latex; (◦) negative polystyrene latex
gradients of the lines are related to the relative affinities of the
particles for the solvent. A comparison of the silica and alumina data shows a stronger solvent–surface interaction for the
latter. Care must, however, be taken when interpreting these
data, as paramagnetic impurities may be present in the sols
and these may artificially (in this case) enhance the relaxation
rate. Comparing the two polystyrene latex samples with the
inorganic sols demonstrates the more hydrophobic nature of
the polymer particles and, in the case of the negatively charged
polystyrene latex, its lower surface charge density. In mixed
particulate dispersions, the specific relaxation rates are additive, and these simple measurements can be used to monitor
heteroflocculation and competitive solvations. For example,
competitive solvation has been investigated in the system comprising water, dimethylformamide, and silica, using relaxation
rate measurements (T. Cosgrove, unpublished data). More
detailed studies of the structure of the solvent layer (water) can
be made using 17 O and 2 H relaxation rates, which are sensitive
to changes in intramolecular motion.13 These studies make it
possible to distinguish effects such as proton exchange with
surface hydroxyl groups and the role of anisotropic surface
mobility of adsorbed solvent. The extension of these methods
to other colloidal systems (micelles, vesicles, microemulsions,
etc.) is well established,14 as is the use of paramagnetic probes
as relaxation rate enhancers.
NMR translational diffusion experiments may also probe
adsorption phenomena and may be analyzed somewhat like
relaxation times.15 For a small molecule in a system with
two environments labeled A and B in fast exchange (which
is sometimes an oversimplification), the measured diffusion
coefficient for a small molecule (for simplicity, this is referred to
as the solvent), is given by D = XA DA + XB DB . In the presence
of colloidal particles, the A environment may be the bulklike solvent, often a continuous phase. The B may represent
the solvent or small molecule associated with the colloidal
particle. This situation then leaves two unknowns, the fraction
of molecules in or associated with the particle and the diffusion
coefficient of the particle. In effect, knowledge of either of these
variables allows the calculation of the other.
Volume 2, 2013
For particles that are big and slow compared to the solvent,
the diffusion of the solvent in the B environment may be
taken as zero. In that case, the measured diffusion coefficient
is dictated by the fraction of the solvent able to diffuse, XA .
The diffusion coefficient of the solvent in a continuous phase
can then be taken as equivalent to the bulk solvent, subject
to obstruction effects.5 In that case, estimation of the fraction of the solvent associated with the particle should include
any surface solvent, plus any other solvent entrapped in the
particle. This method then becomes a good way, for example,
to measure the hydration of particles, especially when polymers are involved which may trap solvent. If the particles are
small enough and their diffusion is not negligible, it may be
possible to estimate the diffusion coefficients of the particles
through a probe molecule that resides within the particle. This
approach can then determine partitioning between the two
environments, and also works quite well for microemulsions
and swollen micelles.15
Several studies have focused on the adsorption of small
molecules (as distinct from solvent) from solution onto colloidal particles. One interesting example is the use of 13 C NMR
to follow the adsorption of carbon monoxide onto colloidal
palladium, a system of commercial interest in the automotive industry.16 Figure 2 shows a 75 MHz 13 C spectrum of
2 nm diameter particles stabilized by poly(vinyl pyrrolidone)
dispersed in methylcyclohexane. A standard high-resolution
spectrometer was used to obtain the data. The adsorbed CO
resonance occurs at 200 ppm, but shows no Knight shift. The
latter observation was explained by the particle being of insufficient size to have metallic character. In a later paper,17 the
authors presented indirect evidence that indicated that larger
particles did show such a shift. The effects of exchange with CO
in the bulk, of course, must be allowed for when interpreting
these chemical shifts. The adsorption of CO onto colloidal
platinum18 can also be seen by direct observation of the 195 Pt
spectrum, as shown in Figure 3, where a shoulder appears to
the left of the main peak. These two examples indicate the
complementary nature of multinuclear magnetic resonance in
surface studies of this type.
© 2013 John Wiley & Sons, Ltd.
429
FD Blum
(a)
(b)
(c)
42.306
250
200
42.506
Frequency (MHz)
42.706
Figure 3. 195 Pt NMR spectrum obtained from platinum colloid after
bubbling with CO gas for 30 min. The resonance at 42.460 MHz [full width
at half maximum (FWHM) of 8 kHz] is due to platinum surface atoms
bound to CO
150
δ
Figure 2. Variable temperature 75 MHz 13 C NMR spectra of CO adsorbed
on colloidal palladium (ca 1% in methylcyclohexane) at (a) 220 K (364
transients), (b) 298 K (8000 transients), and (c) 333 K (50 000 transients).
Spectra are proton decoupled
than that perpendicular to (through) the surfactant layers. The
decay curves in Figure 4 for the room temperature sample
show the composite decay of the signal, which can be used
to estimate the parallel and perpendicular components relative to the bilayers.19 The analysis considers the orientations
of the different domains, as well as the anisotropic diffusion
coefficient. The diffusion of water in these systems is about
an order and a half of magnitude slower than for water in
a comparable salt solution, whereas the activation energy for
diffusion perpendicular to the bilayer is similar to water in a
comparable salt solution. An additional review on anisotropic
diffusion, with many examples of colloidal materials, and an
analysis of the experiments are available.20 It is noted that the
presence of a nonexponential decay curve is not necessarily an
indication of restricted diffusion.
In some cases, the solvent may actually be inside the colloidal
particle. NMR self-diffusion coefficients have been used to
measure the translational diffusion of water inside liquid crystals. For annealed sodium 4-(1-heptylnonyl)benzenesulfonate
(SHBS) (a double-tailed surfactant) with water, large domains
of liquid crystals were formed with water in small (about
10 Å) planar layers. These domains were large enough that the
water retained its average orientation relative to the surfactant
bilayer throughout the experiment. The decays of the diffusion
curves were consistent with diffusion parallel to the surface of
the bilayers (within the water layer) being considerably faster
Log intensity
1.6
1.4
27 °C
1.2
51 °C
68 °C
1
Dzz
0.8
Dxx
0.6
0.4
Dyy
0.2
0
0
(a)
50
100
150
200
250
Beta × 106 (s3)
(b)
Figure 4. Pulsed gradient NMR diffusion echo decay of water in the aqueous layers of smectic liquid crystal of SHBS/water as a function of Beta, the time
variable for the diffusion experiment. The curved decay, in this case, was indicative of anisotropic diffusion with Dxx = Dyy Dzz
430
© 2013 John Wiley & Sons, Ltd.
Volume 2, 2013
Colloidal Systems
Interaction of Surfactants with Colloidal
Particles
It is important to mention two approaches to monitoring
the structures of adsorbed surfactants in colloidal dispersions.
Firstly, if the added surfactant alters the exchange rate of the
solvent, then a measurement of the exchange-averaged relaxation rate of the solvent will give information on the surfactant,
provided that any bulk solvent–surfactant interaction is taken
into account. One system where this approach has proven itself
useful is in the study of the adsorption of nonionic surfactants onto silica.21 Figure 5(a) shows the solvent relaxation
enhancement R2sp = (R2d − R2p )/R2p (where d represents the
total colloidal dispersion and p is a ‘bare’ particle dispersion
at the same solids concentration) for the nonionic, dodecylpoly(ethylene oxide) surfactant C12 EO6 as a function of the
solution equilibrium concentration. The results show a sharp
increase in the solvent relaxation rate enhancement at a solution concentration of ca 0.8 mmol dm−3 [approximately equal
to the critical micelle concentration (CMC) of the surfactant],
which then becomes independent of solution concentration.
These changes follow the adsorption isotherm rather closely. In
the ‘plateau’ region of the adsorption isotherm, C12 EO6 forms
a bilayer. In Figure 5(b), data for two other poly(ethylene
oxides) (PEOs), an EO22 homopolymer and a C12 EO25 copolymer, are shown. In the latter case, the solvent relaxation rate
enhancement is slightly lower than that for C12 EO6 , and the
sharp increase is slightly less pronounced. Again, the changes
follow the adsorption isotherm, this time forming surface
aggregates (or a ‘broken bilayer’) at the silica surface. The
single layer formed by the PEO homopolymer E22 shows only
a very weak relaxation rate enhancement, suggesting that the
bilayer formed by C12 E6 and the surface aggregates formed
by C12 EO25 immobilize much more water at the silica–water
interface.
A second approach is to study the adsorbed layer through
lineshape analysis. This has been carried out in a study of
the adsorption of SHBS (a double-tailed surfactant) at pH 4
onto γ-alumina using 2 H NMR.22,23 Figure 6 shows a series
of spectra obtained at 30.7 MHz using a quadrupolar echo
sequence and 20 000 scans. Spectra from the surface at two
R 2sp /R 02sp
2.5
coverages are shown, together with the spectra for the same
molecule in solution and in liquid crystalline domains. At
low surface coverage (Figure 6a), the spectrum consists of a
single broad line [broader than that of the molecule in a 2%
(w/w) solution (Figure 6c)]. Given that a proportion of the
signal must come from free surfactant, the observed spectrum
(Figure 6a) must be exchange broadened. The absence of any
quadrupolar splitting suggests that the surfactant motion is
somewhat isotropic. At high coverage (Figure 6b), a distinct
Pake doublet and a sharp central resonance appear. This can be
compared with Figure 6(d), which shows the same surfactant in
a liquid-crystalline phase, where the motion is very anisotropic.
These results suggest that there are two environments: an inner
ordered layer and a more mobile outer layer, that is in fast
exchange with the bulk. It is also likely that there will be
slow exchange between the two domains. In other words,
the authors suggest the formation of a surfactant bilayer at
the alumina–water interface. They were unable, however, to
indicate more precisely the exact nature of this bilayer. The
spectral detail that results from this study again indicates the
complementary nature of the two approaches.
An elegant technique developed by Macdonald et al.24 – 26
used the quadrupolar splitting observed in the 2 H spectra of
an adsorbed surfactant [deuterium-labeled hexadecylphosphocholine (HDPC)] to monitor the surface charge of the substrate
or changes in the surface charge upon adsorption of another
species, say, a polyelectrolyte. If an external force alters the
conformation of HPDC, then a change will be observed in the
quadrupolar splitting in its 2 H NMR spectrum. Clearly, though,
some assumptions have to be made regarding the structure of
the mixed layer if species other than HPDC are present at
the particle surface. Nevertheless, this technique offers great
potential for studies of surface electrostatics in colloid science.
Interaction of Polymers with Colloidal
Particles
Polymers are used in a multitude of colloidal formulations to
modify the properties of both the substrate and the dispersion
medium. The addition of polymers can increase the viscosity
of the dispersion medium and hence increase the stability with
(a)
(b)
2.0
1.5
1.0
10−3 10−2 10−1 100 101
c (mmol dm−3)
0
0.2
0.4
0.6
0.8
1.0
c (mmol dm−3)
Figure 5. Normalized specific solvent relaxation rates of surfactant/water/silica systems as a function of surfactant equilibrium concentration: (a) C12 EO6
and (b) C12 EO25 (Δ) and EO22 (◦)
Volume 2, 2013
© 2013 John Wiley & Sons, Ltd.
431
FD Blum
Brushes
Tails
Mushrooms
Loops
Trains
Substrate
(a)
Figure 7. Schematic representation of adsorbed polymers showing, for
randomly attached polymers (left), trains, loops and tails, and mushrooms
and brushes for terminally attached polymers (right)
(b)
(c)
(d)
−20
−10
0
20
10
ν (kHz)
Figure 6. 2 H spectra of SHBS-d4 at 25 ◦ C: (a) adsorbed on γ-alumina at
low surface coverage, (b) adsorbed at high coverage, (c) in a 2% (w/w)
water solution, and (d) in liquid crystals with water
respect to settling. Alternatively, they can be added to promote
or discourage stabilization via adsorption onto the colloidal
particles. The properties that the adsorbing polymers impart to
the substrate are directly related to the structure and dynamics
of the adsorbed polymer layer. These relationships have been
discussed in detail in an excellent book.27 The NMR techniques
available to monitor these adsorbed, polymeric structures are
solvent relaxation and diffusion, adsorbed polymer relaxation,
multiple pulse methods, and cross polarization; each of
these will be discussed briefly. The NMR characterization
of adsorbed polymers has been reviewed.28 – 30 Of particular
interest is the structure of the polymers at the interface. The
nature of the polymer, solvent, and surface can cause the
polymer segments to behave in either a solid- or liquid-like
manner. As long-chain molecules, there are many ways in
which these species can be adsorbed onto surfaces. Randomly
adsorbed polymers typically adsorb with segments of differing
mobility, the so-called trains, loops, and tails.27 Alternatively,
432
terminally attached polymers, depending on their surface
density and molecular mass may adsorb to form mushrooms to
brushes. These kinds of polymers can have different behavior
depending on their adsorption characteristics. A schematic of
these different polymers is shown in Figure 7.
The first observation of polymer adsorption using NMR
showed that the spectrum arising from the adsorbed layer
disappeared at low polymer coverages.31 It was proposed that
the adsorbed chains were lying flat at the interfacial plane,
sufficiently dipolar broadened that they were not visible in a
normal high-resolution spectrum. This led to the idea that,
by making quantitative measurements of signal intensity, and
knowing the adsorption isotherm, it should be possible to
determine the fraction of polymer segments immobilized at
the interface (p) as a function of the amount adsorbed.32 The
data derived from multiple-pulse experiments using both solid
and liquid echoes are shown in Figure 8. In this instance, p was
defined as the height of the solid echo divided by the height
of the solid and liquid echoes combined. The results showed
that as the surface occupancy increased, the chains became
more extended and a smaller proportion of the segments was
immobilized. For comparison, ESR data on the same system are
shown, as is a theoretical lattice model calculation based on the
Scheutjens–Fleer theory of polymer adsorption.27 The agreement between the two experimental methods is rather pleasing
and shows the consistency in using a mobility criterion to distinguish different parts of the adsorbed chain. The qualitative
agreement with a theoretical model is also rather good.
More detailed experiments focusing on the relaxation time of
the adsorbed segments are also possible, and a range of CPMG
experiments on adsorbed terminally grafted PEO chains on
polystyrene latex have been reported.33 As strongly dipolarcoupled spins are not visible in these experiments, the results
reflect the mobility of segments that are part of the anchored
chains but are not in direct contact with the particle surface. The
mobility of the polymer segments is complex and anisotropic,
although some qualitative description of the dynamics is possible, as the major contribution to the relaxation process is
from segmental dynamics. The results show that as the surfacegrafted amount increases, the average spin–spin relaxation time
decreases, corresponding to increased intermolecular interaction. On increasing the chain length, the average relaxation
time increases, because the longer chains have more mobility,
especially at the ends far from the interface. These changes
reflect the change in overall mobility of the so-called loops and
tails of the adsorbed polymer layer.
© 2013 John Wiley & Sons, Ltd.
Volume 2, 2013
Bound fraction
Colloidal Systems
1.0
1.0
0.5
0.5
0.0
0.0
0
(a)
1
2
Absorbed amount
3
(b)
0
0.5
1.0
Fraction of a monolayer
Figure 8. (a) Scheutjens–Fleer lattice model calculation for the polymer bound fraction as a function of adsorbed amount. (b) Experimental data as a
function of surface coverage: (•) NMR and () ESR
In addition to counting the segments with liquid- and solidlike mobilities, relaxation experiments can be important in
the understanding of the breadth of the motions of adsorbed
polymers. As with small molecules, this technique works well
for mobile species; however, unlike small molecules, the interpretation of the relaxation times can be more complicated. In
order to extract information on motions from the relaxation
times, motional models can be used to determine the behavior
of swollen adsorbed polymers. An example of the model of
Hall and Helfand34 has been used in the understanding of the
behavior of both terminally attached and randomly attached
polymers. The parameters from the model yield both a mean
correlation time and a distribution width.30
For randomly attached homopolymers of poly(methyl
acrylate)-d3 , deuterium relaxation measurements clarified
some of the motional differences between the solution- and
surface-bound polymers in the presence of toluene.30,35 For
the solution- and surface-bound polymers, the T1 relaxation
times were similar, with the surface T1 s being a little smaller
because of the restrictions placed by binding to the surface.
In contrast, the T2 s were much slower on the surface than
in solution. When analyzed via the Hall–Helfand model, the
data at room temperature were consistent with the mean
correlation times being about a factor 5 lower on the surface.
As expected, on the surface, the distribution of motions was
much broader. The fact that T2 s were much more affected than
T1 s meant that the longer range, slower motions were more
affected than the faster ones.
In contrast to randomly attached homopolymers, block
copolymers offer the opportunity for a different type of adsorption. Block copolymers of vinyl pyridine (VP) and styrene (S)
on silica have been amenable to study using similar deuterium
relaxation time experiments. High-resolution 13 C spectra of the
polymer on silica swollen with toluene show no resonances for
VP, and relatively narrow ones for S, consistent with the binding
of the VP groups.36 – 39 The relaxation measurements for this
system were consistent with the S segments being in brushes.
Based on the relaxation measurements, it could be argued that
the S segments in the bound, swollen polymer had enhanced
mobility compared to the corresponding polymer in solution.
This could be based on the ratios of T1 /T2 or the Hall–Helfand
Volume 2, 2013
analysis, which demonstrated that the adsorbed polymer had
faster correlation times and a narrower distribution than the
corresponding polymer in solution. This enhanced mobility
occurs because of the differences in packing in brushes (i.e.,
segment–segment interactions) compared to solution and
highlights that dynamics on surfaces can have surprising results.
The behavior across the interface can also be monitored.
Cross polarization between 29 Si atoms in the substrate and
adsorbed poly(vinyl alcohol) has also been used to view surface
and near-surface polymer segments selectively, and to estimate
the distance ratios of the hydroxyl groups and the polymer
backbone to the silica surface.40
The solvent-relaxation method can also be applied to
adsorbed polymers, and the results show that the observed
relaxation rate enhancement is due to adsorbed polymer segments in close proximity to the surface.41,42 One very useful
application of this method is to find the critical displacement
condition. This determination may be made by measuring the
relaxation enhancement as a function of displacer concentration (in this case, dimethylformamide). Figure 9 shows such a
plot, clearly indicating the point at which all the polymers have
been desorbed and the relaxation rate reverts to that associated
with the bare particles.
NMR relaxation methods have been applied extensively to
investigate the structure and dynamics of polymer melts near
particle surfaces. Several zones have now been identified where
the bulk polymer mobility has been slowed by interacting
either directly with the particle surface or indirectly through
entanglement with the surface-adsorbed chains.43 The detailed
analysis of the relaxation times originating from these different
regions is complex, and thus far only a qualitative description
has been given. A full description of these systems is beyond
the scope of the present article.
Polymer–Surfactant Interactions
NMR is an ideal technique for investigating polymer–surfactant
interactions because it has the ability to distinguish between
different species on the basis of chemical shift or mobility.
Many reports appear in the literature that describe the use of
these methods to study these important systems. We shall cite
© 2013 John Wiley & Sons, Ltd.
433
FD Blum
0.6
A (mg m−2)
Atot
0.3
Atr
0.0
−3
−2
−1
0
log(φ (d))
Figure 9. PEO (molecular weight 20 kDa) desorption from silica by increasing the volume fraction of displacer ϕ(d): (◦) adsorbed amount of train
segments for dimethylformamide as displacer, (•) total adsorbed amount for dimethylformamide as displacer, and (Δ) adsorbed amount of train segments
for dimethyl sulfoxide as displacer
two recent examples of the detail that can be elicited from
NMR measurements of these systems.
The interactions of surfactants with gelatin are of commercial
importance to the photographic industry, and hence a detailed
understanding of how the surfactant binds to the protein
under different solution conditions (pH and ionic strength)
is of continuing interest. High-resolution 13 C chemical shift
data show that for sodium dodecyl sulfate (SDS), only the
two carbon atoms nearest to the headgroup are affected by
interactions with gelatin.44
The low-frequency shifts (increased shielding) seen for both
polymers are likely to be caused by desolvation of the headgroup
water by the polymer. The high-frequency shifts (decreased
shielding) may be associated with the fact that more SDS
micelles are present in the system, as the gelatin (or, more
specifically, its associated counterions) lowers the CMC of
the SDS by nearly an order of magnitude. The spectrum of
the gelatin itself is also quite revealing, showing that specific
resonances (e.g., the aspartic and glutamic acid groups) are
unaffected by SDS, but that the linewidths associated with
the cationic residues and the hydrophobic groups become
progressively broader with increasing SDS concentration.
In another 13 C study,45 the interaction of SDS with PEO
has been investigated as a function of surfactant concentration.
When the polymer is saturated with SDS (above its CMC), the
NMR relaxation results indicate a dramatic slowing down of
the motion of the polymer segments attached to the micelles,
even though these represent only a small fraction of the total
number of segments. However, the whole system is dynamic,
and fast exchange occurs between free monomer and free and
adsorbed micelles.
Particle/Emulsion Characterization
In any colloidal dispersion, it is essential to know the surface
area and dispersed state of the system. For particle systems,
the relaxation rate enhancement provides a monitor of the
aggregation state (available surface area), but other NMR
0.3
Probability (μm−1)
0.25
0.2
0.15
0.1
0.05
0
0
3
6
Diameter (μm)
9
12
Figure 10. Size distribution curve obtained with a microscope for an emulsion containing 10% water. (– — –) The best fit of the data with a log–normal
size distribution and (—) the size distribution obtained by NMR
434
© 2013 John Wiley & Sons, Ltd.
Volume 2, 2013
Colloidal Systems
methods can be used. For example, the flow of solvent through
a suspension can be used to infer the volume fraction of particles in the colloidal dispersion. In particular, NMR diffusion
experiments have shown that the average diffusion coefficient
of the solvent depends on the particle concentration and also
on the structure of any adsorbed layer present.46 Effectively,
the diffusion path is lengthened by the obstacle.
For emulsions, where it is difficult to measure the size
and/or volume fraction, NMR diffusion offers many advantages over more conventional methods such as light scattering,
where the sample must be diluted. NMR diffusion has been
extended to obtain particle-size distributions by choosing
a specific distribution function to interpret the attenuation
data.3,47,48 Figure 10 shows data obtained from an NMR analysis assuming a log–normal distribution, and results from
optical microscopy. The two approaches are in reasonable
agreement, but, of course, the NMR approach can be used in
many systems that are inaccessible to microscopy or any other
optical technique. This work has been extended to use in rather
complicated crude oil emulsions in brine.3
10. E. Söderlind and P. Stilbs, Langmuir , 1993, 9, 2024.
11. A. Carrington and A. D. McLachlan, ‘Introduction to Magnetic Resonance
with Applications to Chemistry and Chemical Physics’, Harper & Row: New
York, 1967Chapter 12.
12. P. Stilbs and I. Furó, Curr. Opin. Colloid. Interface. Sci., 2006, 11, 3.
13. L. Piculell, J. Chem. Soc., Faraday Trans. I, 1986, 82, 387.
14. J. B. Nagy, Colloids Surf., 1989, 35, 201.
15. P. Stilbs, Prog. Nucl. Magn. Reson. Spectrosc., 1987, 19, 1.
16. J. S. Bradley, J. Millar, E. W. Hill, and M. Melchior, J. Chem. Soc., Chem.
Commun., 1990, 705.
17. J. S. Bradley, J. M. Millar, and E. W. Hill, J. Am. Chem. Soc., 1991, 113, 4016.
18. R. D. Newmark, M. Fleischmann, and B. S. Pons, J. Electroanal. Chem.,
1988, 255, 325.
19. F. D. Blum, A. S. Padmanabhan, and R. Mohebbi, Langmuir , 1985, 1, 127.
20. I. Furó and S. V. Dvinskikh, Magn. Reson. Chem., 2002, 40, S3.
21. M. R. Bohmer, L. K. Koopal, R. Janssen, E. M. Lee, R. K. Thomas, and A. R.
Rennie, Langmuir , 1992, 8, 2228.
22. E. Söderlind and F. D. Blum, J. Colloid Interface Sci., 1993, 157, 172.
23. K. Nagashima and F. D. Blum, Colloids Surf. A , 2001, 176, 17.
24. P. M. Macdonald and Y. H. Yue, Langmuir , 1993, 9, 1206.
Biographical Sketch
25. S. C. Kuebler and P. M. Macdonald, Langmuir , 1992, 8, 397.
Frank D. Blum was born in 1955. He obtained his BS, 1976, and
MS, 1977, from Eastern Illinois University, and a PhD, 1981, from
University of Minnesota. He worked as assistant professor, 1981–1986,
in Drexel University, and associate professor, professor, and curators’
professor, 1986–2009, in University of Missouri-Rolla/Missouri S&T.
He is the Harrison I. Bartlett Chair and Regents professor in Oklahoma
State University, 2010–present. He has contributed approximately
225 publications, book chapters, and symposium proceedings. He is
a fellow of the American Chemical Society. Fellow of the Division
of Polymer Chemistry, Inc. His research activities include dynamics
in interfacial materials, polymers and surfactants, composites, and
conducting polymer nanocomposites.
26. P. M. Macdonald, D. Staring, and Y. Yue, Langmuir , 1993, 9, 381.
27. G. J. Fleer, M. A. Cohen-Stuart, J. M. H. M. Scheutjens, T. Cosgrove, and B.
Vincent, ‘Polymers at Interfaces’, Chapman & Hall: London, 1993.
28. F. D. Blum, Colloids Surf., 1990, 45, 361.
29. F. D. Blum, Annu. Rep. NMR Spectrosc., 1994, 28, 277.
30. F. D. Blum, in ‘Colloid-Polymer Interactions : From Fundamentals to Practice’, eds P. Dubin and R. S. Farinato, Wiley: New York, 1999, Chap. Nuclear
Magnetic Resonance of Surface Polymers, 207.
31. T. Miyamoto and H. J. Cantow, Makromolekul. Chem., 1972, 162, 43.
32. K. G. Barnett, T. Cosgrove, B. Vincent, M. Cohen-Stuart, and D. S. Sissons,
Macromolecules , 1981, 14, 1018.
33. T. Cosgrove and K. Ryan, Langmuir , 1990, 6, 136.
Related Articles
34. C. K. Hall and E. Helfand, J. Chem. Phys., 1982, 77, 3275.
Spiess, Hans Wolfgang: Multidimensional Solid State NMR
of Polymers; Amphiphilic Liquid Crystalline Samples: Nuclear
Spin Relaxation; Deuterium NMR in Solids; Polymers: Relaxation and Dynamics of Synthetic Polymers in Solution
36. F. D. Blum, B. R. Sinha, and F. C. Schwab, Macromolecules , 1990, 23, 3592.
References
39. M. Xie and F. D. Blum, Macromolecules , 1996, 29, 3862.
35. M. Liang and F. D. Blum, Macromolecules , 1996, 29, 7374.
37. B. R. Sinha, F. D. Blum, and F. C. Schwab, Macromolecules , 1993, 26, 7053.
38. M. Xie and F. D. Blum, Langmuir , 1996, 12, 5669.
1. M. L. Johns, Curr. Opin. Colloid. Interface. Sci., 2009, 14, 178.
2. F. Mariette, Curr. Opin. Colloid. Interface. Sci., 2009, 14, 9.
3. C. P. Aichele, M. Flaum, T. M. Jiang, G. J. Hirasaki, and W. G. Chapman, J.
Colloid Interface Sci., 2007, 315, 607.
40. N. Zumbulyadis and J. M. O’Reilly, Macromolecules , 1991, 24, 5294.
41. G. P. Vanderbeek, M. A. C. Stuart, and T. Cosgrove, Langmuir , 1991, 7, 327.
42. T. Cosgrove, T. M. Obey, and M. Taylor, Colloids Surf., 1992, 64, 311.
43. J. P. Cohenaddad and A. Viallat, Polymer , 1986, 27, 1855.
4. O. Söderman and P. Stilbs, Prog. Nucl. Magn. Reson. Spectrosc., 1994, 26,
445.
44. D. D. Miller, W. Lenhart, B. J. Antalek, A. J. Williams, and J. M. Hewitt,
Langmuir , 1994, 10, 68.
5. O. Söderman, P. Stilbs, and W. S. Price, Concepts Magn. Reson. A , 2004,
23A, 121.
45. K. Chari, B. Antalek, M. Y. Lin, and S. K. Sinha, J. Chem. Phys., 1994, 100,
5294.
6. H. Hommel, Adv. Colloid Interface Sci., 2008, 141, 1.
46. T. Cosgrove and P. C. Griffiths, Colloids Surf. A , 1994, 84, 249.
7. H. Hommel, Adv. Colloid Interface Sci., 1995, 54, 209.
47. B. Balinov, O. Urdahl, O. Söderman, and J. Sjoblom, Colloids Surf. A , 1994,
82, 173.
8. C. Mayer, in ‘Annual Reports on NMR Spectroscopy’, ed. G. A. Webb,
Elsevier: Oxford, UK, 2005, Vol. 55, 205.
9. R. J. Hunter, ‘Foundations of Colloid Science’, 2nd edn, Oxford University
Press: Oxford; New York, 2001.
Volume 2, 2013
48. M. L. Johns and K. G. Hollingsworth, Prog. Nucl. Magn. Reson. Spectrosc.,
2007, 50, 51.
© 2013 John Wiley & Sons, Ltd.
435