interactions between keratin and surfactants
TRANSCRIPT
Interactions between Keratin and Surfactants:
A Surface and Solution Study
A thesis submitted to the University of Manchester for the degree of
Doctor of Philosophy
in the Faculty of Engineering and Physical Sciences
by Zhiming Lu
2015
School of Physics and Astronomy
1
Contents
Abstract 7
Declaration 8
Copyright 9
Acknowledgements 10
1 Chapter 1: Introduction to surfactants, biosurfactants, proteins and their
complexes 11
1.1 Conventional surfactants 13
1.1.1 Sodium dodecyl sulfate, SDS 13
1.1.2 Dodecyl trimethyl ammonium bromide, DTAB 14
1.1.3 Hexaethylene glycol monododecyl ether, C12E6 14
1.2 Biosurfactants 15
1.2.1 Rhamnolipids 16
1.2.2 Mannosylerythritol lipids (Mels) 17
1.3 Proteins 18
1.3.1 Protein structure 20
1.3.2 Keratin, a fibrous structural protein 21
1.4 Surfactant and protein behaviour at interfaces and in solution 22
1.4.1 Surface adsorption behaviour 23
1.4.2 Self-assembly behaviour in solution 24
1.5 Protein/surfactant complexes 24
References 27
2
2 Chapter 2: Theoretical background and experimental procedures 31
2.1 Surface tension 31
2.1.1 Static surface tension 31
2.1.2 Dynamic surface tension 32
2.2 Spectroscopic ellipsometry, (SE) 33
2.2.1 Background theory 33
2.2.2 Data analysis 34
2.2.3 Instrument 35
2.3 Quartz crystal microbalance with dissipation, (QCM-D) 36
2.4 Dual polarization interference, (DPI) 38
2.4.1 Background 38
2.4.2 Theoretical principles 39
2.5 Neutron reflection 43
2.5.1 Theoretical background 44
2.5.1.1 Calculation of the refractive index 45
2.5.1.2 Reflectivity for uniform layer 46
2.5.1.3 Reflectivity for multi layers 47
2.5.2 Instrumentation 48
2.5.3 Experimental procedures 50
2.5.3.1 Air/water interface 50
2.5.3.2 Solid/water interface 51
2.5.4 Data analysis 52
2.6 Small-angle neutron scattering 53
2.6.1 Theoretical background 53
2.6.2 Data analysis 54
3
2.6.3 Instrumentation 57
2.6.4 Experimental procedures 58
2.7 Dynamic light scattering, (DLS) 58
References 60
3 Chapter 3: Adsorption of C12E6 at the SiO2/water interface: a combined
study by DPI, SE, QCM-D and NR 61
3.1 Literature review 62
3.2 Experimental 65
3.2.1 Sample preparation 65
3.2.2 Experimental method 65
3.3 Results and discussion 67
3.3.1 Neutron reflection measurements 67
3.3.2 Dual polarisation interferometry measurements 76
3.3.3 Quartz crystal microbalance with dissipation measurements 82
3.3.4 Spectroscopic ellipsomtry measurements 85
3.4 Comparison of the four techniques 86
3.5 Conclusion 89
References 90
4 Chapter 4: Surface adsorption and solution aggregation of wool keratin
4.1 Theoretical background 92
4.2 Materials 94
4.2.1 Keratin production 94
4.2.2 Preparation of keratin solutions 95
4.3 Results 96
4.3.1 Surface tension 96
4
4.3.2 Neutron reflectivity 97
4.3.2.1 Reflectivity profiles of keratin in the null reflecting water subphase 97
4.3.2.2 Reflectivity profiles of keratin in the D2O subphase 100
4.3.2.3 Reflectivity profiles of keratin with 0.5 M NaCl in null reflecting water
subphase 103
4.3.3 Dynamic light scattering, (DLS) 104
4.3.4 Small-angle neutron scattering, (SANS) 106
4.4 Discussion 110
4.5 Conclusion 114
References 116
5 Chapter 5: Interaction of keratin and surfactants of sodium dodecyl sulfate
and dodecyl trimethyl ammonium bromide at the air/water interface 117
5.1 Literature review 118
5.2 Experimental procedures 120
5.3 Results and analysis 121
5.3.1 Surface tension measurements 121
5.3.2 Neutron reflectivity 122
5.3.2.1 Reflectivity profiles of the keratin/h-SDS complexes in NRW 122
5.3.2.2 Reflectivity profiles of the keratin/d-SDS complexes in NRW 126
5.3.2.3 Reflectivity profiles of the keratin/h-DTAB complexes in NRW 131
5.3.2.4 Reflectivity profiles of the keratin/d-DTAB complexes in NRW 132
5.4 Conclusion 137
References and support information 138
6 Chapter 6: Binding of cationic surfactant DTAB onto coated keratin film
at the solid/water interface studied by SE, QCM-D and NR 140
5
6.1 Literature review 141
6.2 Experimental procedures 142
6.3 Results and discussion 143
6.3.1 SE measurement 143
6.3.2 QCM-D measurement 146
6.3.3 NR measurement 151
6.4 Conclusion 158
References and support information 160
Chapter 7: The Interfacial interaction of keratin and rhamnolipids at the
solid/liquid interface as studied by NR, QCM-D, SE and DPI 165
6.5 Experimental procedures 166
6.5.1 Mass spectroscopy of rhamnolipid 1 and 2 166
6.6 Results and discussion 168
6.6.1 Film stability and reproducibility check by NR 168
6.6.2 Adsorption of rhamnolipids on the coated film of keratin 173
6.6.2.1 QCM-D measurements 173
6.6.2.2 SE measurements 176
6.6.2.3 NR measurements 178
6.6.2.4 DPI measurements 184
6.7 Conclusion 185
References and support information 187
7 Chapter 8: An initial study of Mel-C at the solid/water interface and in
solution 189
7.1 Results and discussion 190
6
7.1.1 Surface tension measurements 190
7.1.1.1 Static surface tension 190
7.1.1.2 Dynamic surface tension 191
7.1.2 Adsorption of Mel-C at the SiO2/water interface 194
7.1.3 Adsorption of Mel-C on the coated film of keratin 195
7.1.3.1 SE measurements 195
7.1.3.2 QCM-D measurements 199
7.1.4 Size measurements of Mel-C in solution 200
7.1.5 Zeta potential measurements 203
7.2 Conclusion 204
References and support information 205
Chapter 9: Summary 208
References 213
7
Abstract
The University of Manchester
A thesis for the degree of Doctor of Philosophy
Interactions between Keratin and Surfactants: A Surface and Solution Study
by Zhiming Lu
Keratins are important structural components of hair and skin. There has been extensive study of
keratins from the health and medical perspectives, although little work has been done to date to
investigate their basic physicochemical properties in the form of biomaterials. The work presented
in this thesis aimed to study surface and interfacial adsorption and solution aggregation of water
soluble keratin polypeptides (made available by previous work within the research group). A
range of physical techniques were employed including spectroscopic ellipsometry (SE), neutron
reflection (NR), dual polarisation interferometry (DPI), quartz crystal microbalance with
dissipation (QCM-D), dynamic light scattering (DLS) and small-angle neutron scattering (SANS).
A major technical advantage of the neutron techniques is the use of hydrogen/deuterium
substitution to enhance structural resolution. This approach was explored to study the interaction
of keratins with both conventional surfactants and novel biosurfactants. The work presented
comprises four results chapters. The first examines and compares four widely used interfacial
techniques, SE, DPI, QCM-D and NR, by studying the adsorption of C12E6 at the silicon
oxide/water interface. Whilst the data exhibits a large degree of consistency in the interfacially
adsorbed amount, each technique helped reveal unique structural information with a high degree
of complementarity. The second results chapter reports on findings regarding the properties of
keratin polypeptides in surface adsorption and solution aggregation. It was found that the keratins
adsorbed strongly on the surface of water, and formed rugby-shaped nanoaggregates in solution,
the size and shape of which responded to salt concentration. The third results chapter reports on
the interfacial behaviour of keratin/surfactants complexes in bulk solution, with cationic DTAB
and anionic SDS as model conventional surfactants. It was found that both the electrostatic and
hydrophobic forces contributed strongly to the surface adsorption processes. The final results
chapter reports on interactions of a coated keratin film with novel biosurfactants including
rhamnolipids (R1 and R2 with 1 and 2 sugar head(s), respectively) and Mel-C. The keratin films
formed were found to be exceptionally stable and reproducible below pH 8, and these films could
be widely used as model keratin substrates for screening their binding with surfactants and
bioactive molecules. Both rhamnolipids and Mel-C exhibited strong adsorption onto the keratin
substrate and interestingly, whilst R1 exhibited a completely reversible adsorption, R2 showed
only a partially reversible adsorption. Mel-C showed some degree of irreversible adsorption
similar to R2 and exhibited the strongest adsorption at around pH 4-5. These results show mild
interactions with the keratin substrate, but indicate that the extent of adsorption and desorption
could be manipulated by surfactant structure or solution conditions.
The findings presented in this thesis are fundamental in aiding the development of the use of
keratin polypeptides as biomaterials, in applications such as personal care. The work is also
highly relevant to the understanding of the interactions between surfactants and keratin
molecules at interfaces and in solution.
8
Declaration
PhD Candidate Declaration
I declare that the work presented in this thesis is entirely my own, except where otherwise stated,
and has not been submitted in support of an application for another degree or qualification at this
University or any other institution.
Zhiming Lu
August 2015
9
Copyright
The author of this thesis (including any appendices and/or schedules to this thesis) owns any
copyright in it (the "Copyright") and s/he has given The University of Manchester the right to use
such Copyright for any administrative, promotional, educational and/or teaching purposes.
Copies of this thesis, either in full or in extracts, may be made only in accordance with the
regulations of the John Rylands University Library of Manchester. Details of these regulations
may be obtained from the Librarian. This page must form part of any such copies made.
The ownership of any patents, designs, trade, marks and any and all other intellectual property
rights except for the Copyright (the "Intellectual Property Rights") and any reproductions of
copyright works, for example graphs and Tables ("Reproductions"), which may be described in
this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual
Property Rights and Reproductions cannot and must not be made available for use without the
prior written permission of the owner(s) of the relevant Intellectual Property Rights and/or
Reproductions.
Further information on the conditions under which disclosure, publication and exploitation of this
thesis, the Copyright and any Intellectual Property Rights and/or Reproductions described in it
may take place is available from the Head of School of Physics and Astronomy (or the Vice-
President) and the Dean of the Faculty of Engineering and Physical Sciences, for Faculty of
Engineering and Physical Sciences candidates.
10
Acknowledgements
First and foremost, I would like to thank my supervisors Prof. Jian Lu and Prof. Henggui Zhang
for accepting me into the Biophysics Group, and for their valuable help and guidance over the last
four years. I feel very privileged to have had the opportunity to work with you.
Next, I would like to thank my industrial supervisors Prof. Jordan Petkov, Dr Ian Tucker and Dr
Hosking at Unilever, Port Sunlight, for their great support and the opportunity to work on such a
great project. They have given me a lot of new ideas and the opportunity to explore my own.
I am very thankful to all of the members of the Biological Physics Group for their help and our
valuable friendship: Jon, Lisa, Mario, Jing, Jiqian, Zongyi, Ruiheng, Mohammed, Jamie, Elias,
Charlie, Daniela and Robert. We had such a great time in the lab and at work.
I also thank all the scientists at ISIS, especially Dr John Webster, Dr Max Skoda, Dr Ann Terry
and Dr Rogers for their extensive assistance and support in setting up the instruments and Peixun
Li for MS measurements. It’s always a pleasant time working at ISIS.
I’d like to thank the ORS and Unilever for funding my PhD and the STFC for funding my neutron
experiments.
Thanks to my wife Rong Huang for comforting me when I was in bad moods. I could not have
made it without her support and back up.
Finally I would like to thank my parents. They have always given me the confidence and the huge
drive to finish my work.
11
Chapter 1
Introduction to surfactants, biosurfactants, proteins and their complexes
This thesis explores the interaction of surfactants with keratin polypeptides at interfaces and in
aqueous solution using a range of physical techniques, including surface tension measurements
(ST), neutron reflection (NR), spectroscopic ellipsometry (SE), quartz microbalance with
dissipation (QCM-D), dynamic light scattering (DLS) and small-angle neutron scattering (SANS).
Keratins usually exist in the form of fibres. This work has benefitted from the provision of a
sufficient quantity of water soluble and water dispersible keratin polypeptides from an innovative
approach involving the dissolving of wool into aqueous solution. To my knowledge, there has
been no such study reported in the past. Information about the adsorption of keratin polypeptides
and their interactions with surfactants at interfaces and in solution provides the first round of
insight about some of the basic features. This research is hence highly relevant in understanding
how keratins interact with different surfactants and how keratin polypeptides could be exploited
in technological applications.
In this chapter, basic background information about surfactants, proteins and their interactions is
provided and aims and plans for the research are introduced.
Surfactants are one of the most important types of chemicals that have undergone intensive study
over the last few decades. They are widely used in our daily life in applications ranging from
detergents, food and personal cosmetics, to fabric softeners, textiles, oil dispersants and
emulsions.1-3 The remarkable properties of surfactants include their low surface tension, good
solubility, amphiphilicity and foaming capacity. In many of these fundamental applications
especially in personal care, surfactants generally appear as mixtures or form complexes with other
materials to provide comprehensive performance. As conventional surfactants are compounds
12
some of which are toxic to animals, humans and the ecosystem4-6, their negative impact is of
growing concern. Biosurfactants, the surfactants produced by bacteria, yeasts and fungi, have
gained more and more attention in recent years due to their environmental impact 7. They have a
promising future in the replacement of synthetic surfactants in a wide area ranging from cosmetics
and personal care to water treatment. However, a significant disadvantage of biosurfactants is
their extremely high price due to high material costs and difficulties in the separating and
purifying procedures from living cells. These disadvantages hinder their replacement of
conventional surfactants.
Proteins are large biological molecules. They consist of many amino acids linked together to form
one or more long chains. They form the very basis of life and perform a variety of activities within
living organisms, from making up human skin to replicating DNA. Although extensive effort is
applied in the study of proteins, very little is focused on their surface adsorption and solution self-
assembly behaviour. Studying proteins at interfaces and in solutions is of great importance in
understanding their structures and functions.
The protein/surfactant complexes and their interactions have been of great interest over recent
decades and they have been utilized in a wide variety of uses in daily life from toiletries and
cosmetics to washing powder enzymes8, 9. However, there is a deficiency in the study of their
adsorption and interactions at interfaces and in solutions, which is essential in understanding their
functions and how they interact with each other.
The aim of the work presented in this thesis is to address the deficiency in the studies mentioned
above. The research firstly focused on the surface adsorption of a conventional surfactant, C12E6,
secondly on the interactions of a fibrous structural protein, keratin, with conventional surfactants,
SDS and DTAB, and the most interesting biosurfactants, rhamnolipids, at interfaces and in
solutions.
13
1.1 Conventional surfactants
Surfactants are usually amphiphilic compounds, which means that they consist of a hydrophobic
group (tail) as a water-insoluble component and a hydrophilic group (head) as a water-soluble
component. The ‘tail’ of a surfactant is an alkyl chain that can be branched, linear or aromatic,
whilst the ‘head’ defines the type of the surfactant: anionic, cationic, nonionic and zwitterionic.
Anionic surfactants, cationic surfactants and nonionic surfactants contain negatively charged head
groups, positively charged head groups and non-charged head groups respectively. Zwitterionic
surfactants have both cationic and anionic centres on head groups. They are usually neutral,
although if one or both of the charged groups is a weak electrolyte they can be net positively
charged or negatively charged depending on the solution pH. This thesis primarily focuses on
anionic, cationic and nonionic surfactants.
1.1.1 Sodium dodecyl sulfate, (SDS)
Figure 1.1 Molecular structure of charged SDS- and its counterion Na+.
The anionic surfactant sodium dodecyl sulfate (SDS) is the most commonly used anionic
surfactant model. Its homologues or derivatives are widely used in many home care and hygiene
products due to its highly effective effect on oily spots. It has a 12-carbon chain tail as a
hydrophobic part and a sulfate group as a hydrophilic part, giving the material amphiphilic
properties. In pure water, the critical micelle concentration (CMC) of SDS at 25 ˚C is quoted at
8-8.2 mM 10, 11 and the aggregation number at or above CMC is measured to be 6312. Previous
studies reported that the addition of salt in solution could considerably decrease the CMC of SDS.
It was found that the addition of 5 mM sodium chloride (NaCl) in solution can decrease the CMC
14
of SDS by 25% 13. Temperatures in the range of 20 ˚C to 40 ˚C have little effect on the CMC of
SDS.14 The solutions used in this work are all of 5 mM NaCl at 25 ˚C.
1.1.2 Dodecyl trimethyl ammonium bromide, (DTAB)
Figure 1.2 Molecular structure of charged DTAB+ and its counterion Br-.
Dodecyl trimethyl ammonium bromide (DTAB) is a common cationic surfactant. Its homologues
are widely used in wetting agents and hair conditioning products. It has a 12-carbon hydrophobic
chain attached to a hydrophilic ammonium head. The CMC of DTAB in pure water is quoted as
14.3 mM at 25 ˚C and changes from 14.3 mM to 15.3 mM with a temperature rise from 20 ˚C to
40 ˚C 15. Corrin et al. 13 found that the addition of 5 mM NaCl can decrease the CMC of DTAB
to 11 mM. In this work, DTAB is used in 5 mM NaCl solution at 25 ˚C.
1.1.3 Hexaethylene glycol monododecyl ether, C12E6
Figure 1.3 Molecular structure of C12E6。
Hexaethylene glycol monododecyl ether, C12E6, is a polyoxyethylene nonionic surfactant widely
used in many detergents. It has a 12-carbon chain and its head consists of six ethylene oxide (EO)
groups and a hydroxide. It has a very low CMC compared to SDS and DTAB, ranging from
7.0×10-5 M16, 17 to 8.7×10-5 M18.
15
The increase of the PEO head length would increase the hydrophilicity of the molecule and thus
increase the CMC of C12Em surfactants. For example, the CMC of C12E3 (5.5×10-5 M) is lower19
while the CMC of C12E8 (1×10-4 M)20 is higher than that of C12E6. On the other hand, the increase
of alkyl tail length would decrease the CMC of CnE6 surfactants21 since it increases the
hydrophobicity of the molecule.22
Since the nonionic surfactant is neutrally charged, it is expected that the solution pH would not
affect their micellar formation like SDS and DTAB. However, there is high sensitivity to pH with
respect to adsorption at interfaces18. Penfold et al.23 found that different methods of surface
treatment had a strong effect on the adsorption of nonionic surfactants greatly, which revealed
some of the difficulties in the measurements of nonionic adsorption.
1.2 Biosurfactants: rhamnolipids
Biosurfactants are surface-active substances that are produced from bacteria, yeasts and fungi.
They have the potential to replace conventional surfactants in industrial sectors such as personal
care. They also have potential for use in enhanced oil recovery, lubrication, food additives,
cosmetics and household detergents24, due to their biodegradability7, low toxicity25, low CMC26
and high emulsifying abilities27. Recent advances in the medicinal and therapeutic production of
biosurfactants has further broadened their range of potential uses.28 Conventional surfactants and
biosurfactants are categorised differently: conventional surfactants are classified by head and tail
groups, whereas biosurfactants are often classified by molecular weight.29 The biosurfactants with
low molecular weights include phospholipids, glycolipids and lipopeptides whereas those with
high molecular weights include amphipathic lipopolysaccharides, polysaccharides and
lipoproteins.30
Glycolipids are the most widely studied biosurfactants among all the classifications. They include
rhamnolipids, trehalose lipids, mannosylerythritol lipids (Mels), polyol lipids and sophorose lipids.
16
The work presented in this thesis is focused on the biosurfactants rhamnolipids and Mel-C, the
most widely used and studied glycolipids, and their interaction behaviours with a fibrillar
polypeptide, keratin.
1.2.1 Rhamnolipids
Rhamnolipids produced by Pseudomonas aeruginosa are amongst the most extensively
investigated glycolipids31 and the best characterised of the bacterial surfactants.32 Bergstrom et
al.33 initially discovered the rhamnolipids in 1946, as reported in a paper describing a glycolipid
extracted from p. aeruginosa. Modern techniques with higher resolution and sensitivity have
given scientists the opportunities to find out about sixty different rhamnolipid congeners34
extracted from various bacteria. The 60 congeners are mainly categorised as six types: R3 and R4
contain one carbon chain; R1 and R2 have two carbon chains; while Ra and Rb have three carbon
chains.
Figure 1.4 Molecular structures of rhamnolipids, R1 and R2. R1 had one sugar head while R2 had 2 sugar head.
R1 and R2 are the two main types produced commercially in large quantities. R1 and R2, also
known as α-L-rhamnopyranosyl-β-hydroxydecanoyl-β-hydroxydecanoate (Rha-C10-C10) and α
-L-rhamnopyranosyl-α-L-rhamnopyranosyl-β-hydroxydecanoyl-β-hydroxydecanoate (Rha-Rha-
C10-C10), consist of one and two rhamnose group(s), respectively.35 Since R1 has only one
hydrophilic rhamnose group compared to the two rhamnose groups of R2, it has a lower CMC
17
and a lower hydrophilicity. Therefore, R1 is more surface active and has greater rigidity than R2,
which can also be measured by QCM-D. Ozdemir et al.36 found by surface tension that the CMCs
of R1 and R2 are 1×10-4 M and 1.5×10-4 M with area per molecule (A) of 135 Å2 and 131 Å2, at
pH 6.8. The CMCs of R1 and R2 become 4×10-5 M at pH 5 for both of them, with area per
molecule of 59 Å2 and 64 Å2, respectively. Chen et al.37 found that the CMCs of R1 and R2
are1.8±0.2×10-4 M and 1.1±0.2×10-4 M at pH 7 while Peker et al.38 reported the CMCs of R1 and
R2 both to be 1.5×10-4 M at pH 6.8. The discrepancy of CMCs of R1 and R2 is rather small, but
as naturally produced biosurfactants are highly diverse, it is very difficult to have monodisperse
chains and sugar heads. Thus, the CMCs from different batches or different groups may well be
different due to different cuts and some degree of chain and head structural distributions.
1.2.2 Mannosylerythritol lipids (MELs)
Figure 1.5 Molecular structure of MELs39. The molecule is neutrally charged.
Mannosylerythritol lipid (MEL) is a yeast biosurfactant and is extracted from Candida strains. It
is widely used in industry and medical science, ranging from antimicrobial to immunological uses.
Haskins et al.40 first reported MELs in a paper concerning the cultured suspension of Ustilago in
1955. MELs are attracting much attention because they are widely used in pharmaceutical
applications41, 42 and in the treatment of diseases caused by microbial infections,43 in drug
delivery39, 44, 45 and other applications.
18
MEL is classified as a glycolipid by Bhatttacharjee et al.46, which consists of a variety of
hydrophobic fatty acids groups and acyl residues that range from 7 to 14 carbon chains.47 The
measurements of the thin layer chromatography (TLC) show that there are mainly four types of
MELs: MEL-A, MEL-B, MEL-C and MEL-D, respectively. MEL-A is described as 4-O-(4’, 6’-
di-O-acetyl-2’, 3’-di-O-alkanoyl-β-D-mannopuranosyl)-D-erithritol where R1=R2=Ac.
Conversely, MEL-B and MEL-C are described as 4-O-(6’-O-acetyl-2’, 3’-di-O-alkanoyl-β-D-
mannopuranosyl)-D-erithritol, while MEL-D has a completely diacetylated structure.47, 48 The
surface and solution properties of various MELs differ and have been investigated over recent
decades. Kitamoto et al.43 found the CMC of MEL-A to be 2.7×10-6 M while the CMC of MEL-
B was found to be 4.5×10-6 M, and both have a low surface tension of 28.2 mN/m. The CMC of
MEL-C is quoted as 4×10-6 M,49 4.5×10-6 M50 to 6×10-6 M51 with surface tensions ranging from
24.2, 30.7 to 25.1 mN/m, respectively.
MEL-C produced from Pseudozyma hubeiensis has the fatty acids on its 6-, 12- and 16-carbon
chains52. In this work, MEL-C was used to investigate the self-assembling properties and its
interaction with proteins at interfaces.
1.3 Proteins
1.3.1 Protein structure
Proteins are an extremely interesting component of, and the major structural constituent of living
beings. They take part in various actions and processes within living organisms, ranging from
replicating DNA,53 transporting molecules54 and catalysing metabolic reactions55 to playing a part
in structural support and defence against germs56. Proteins consist of amino acids, which consist
of extraordinarily complex chains of smaller molecules. Amino acids are shaped like strings
folded into specific forms and create billions of critical and essential components of living beings.
19
Polypeptides are made up of a linear chain of amino acid residues and a protein consists of at least
one polypeptide. A polypeptide with less than 30 amino acid residues is commonly called a
peptide instead of a protein. To be able to form a peptide or protein, the peptide bond is used to
bond the amino acids. Two amino acids interact through the covalent chemical bond between the
carboxyl group and the amino group, which produces a molecule of water (H2O) as a result. In
general, there are 20 standard amino acids as shown in Figure 1.6, with their sequence in a protein
encoded on organisms DNA.57 Each amino acid has a special structure and property. The
combined effect of many amino acids in a protein determines the protein’s chemical and biological
properties.58
Figure 1.6 Chemical structures of the 20 standard amino acids59. They can be divided into three types: Amino acids
with electrically charged side chain, polar uncharged side chains and hydrophobic side chains.
20
The unique three-dimensional structure of proteins allows them to carry out their unique functions.
The question of how the amino acids arranges and folds in a protein and subsequently affects its
function has been raised since the 1960’s.60 A study on the mechanisms that determine the folding
of protein chains was first carried out by Anfinsen et al.61, who explained the assembly of proteins
depend both on the amino acids and their surroundings.61, 62
Figure 1.7 Three representations of the three-dimensional structure of the triose phosphate isomerase protein.63 The
type of a protein is defined by its tertiary structure.
In general, a protein is composed of three degrees of structures. The protein’s primary structure
is defined by the positions of disulfide bonds and the amino acid sequence. The secondary
structure is defined by patterns of bonding formed by the peptide bonds. There are three types of
secondary structure called coils, α-helices and β-sheets.64 The tertiary structure is defined by the
geometric shape of the protein after the protein folds into its unique state. Some proteins may have
a quaternary structure, which is created by the aggregation of a number of polypeptides (tertiary
structures). The type of a protein is defined by its tertiary structure. For example, keratin belongs
to the type of fibrous proteins; Haemoglobin 65 to the type of globular proteins and colic colicin B
in B 66 is categorised in the type membrane proteins.
21
1.3.2 Keratin, the fibril structural protein
Keratins are widely distributed in skin, hair and nails and provide both the epithelial and
endothelial coverage of organs. They are a member of the family of fibrous structural proteins 67.
Recent studies have mostly focused on better understanding of its role in health and diseases. For
example, its interactions with other types of peptides have implicating roles in diseases such as
rheumatoid arthritis 68, and the study of DNA sequence coding for keratins and their mutations
has informed the induction of skin conditions and related diseases 69-71. Extensive research has
also been undertaken in other fields including cosmetics, animal foods and textiles 72, 73, with
many of these studies focusing on fibres and nails. As a multifunctional biomaterial, there is
increasing interest to explore its use as scaffolds in tissue engineering and regenerative medicine.
Recent studies have identified two fundamental roles of keratin: structural support and metabolic
processes, which have important repercussions in tissue reconstruction, cell seeding and diffusion,
and implantable biomaterials 74.
Most known keratins belong to the type I and type II classes of currently recognized intermediate
filament proteins, with their molecular weights spanning from 40 kDa to 70 kDa.75 The type I
keratins often are of smaller size with molecular weights of 40 – 48 kDa and acidic isoelectric
points, whereas the type II keratins are larger with molecular weights of 59 – 62 kDa and neutral
or slightly basic isoelectric points. Human hair or sheep wool contain examples of these two
classes, which can be further divided into hard keratins (type Ia and IIa) and soft keratins (type Ib
and IIb). Eight major hard keratins containing four type Ia (a1, a2, a3, a4) and four type IIa (b1,
b2, b3, b4) have been identified in different species.76
Common to many of the intermediate filament proteins such as keratin is the densely packed
central alpha-helical domain, which is comprised of four coiled-coil segments and non-helical
terminal domains of varying lengths and sequences.77
22
The acidic and basic soft keratins interact to firstly form the basic heterodimers. These can then
pair up to produce tetramers, and the process continues to finally polymerise into 10 nm filament
structures observed78. It is widely thought that wool and hair share a similar process in keratin
filament formation because of their closely related secondary structures.
Yu et al. 80 have undertaken amino acid sequence comparisons between a human type Ia keratin
a3 with another four type Ia hair keratins, including two from sheep and two from mice. The a3
hair keratin was found to have 404 amino acids, a molecular weight of 45,914 Da and an
isoelectric point of 5.6, in contrast to the 412 amino acids and the molecular weight of 48,300 Da
(including an acetyl group on its N-terminal) of a wool microfilament keratin 8c-1 79 and the
molecular weight of 47,600 Da of another wool type I microfilament keratin 80. The comparisons
show predominant sequence homologies with a small frequency of variations.
1.4 Surfactant and protein behaviour at interfaces and in solution
The surface behaviour of molecules in aqueous solution is important in a large range of
applications from biological processes to chemical reactions such as emulsion and stabilisation.
The adsorption of surfactants at interfaces is mainly driven by the hydrophobic effect due to their
amphiphilic properties. A surfactant in water displays the tendency of the movement of its
hydrophobic part away from the water solution. Thus at concentrations of the surfactant below its
CMC, the amphiphilic molecules would rearrange themselves to form a monolayer at the air/water
interface with their hydrophobic tails away from water; at concentrations of the surfactant above
its CMC, the molecules would rearrange to form micelles in solution with their hydrophobic parts
inside the micelle and their hydrophilic parts attached to water. The molecules form either
monolayers or micelles in order to minimise the free energy of the whole structure.
The surface adsorption of proteins is usually an irreversible process. Once the protein molecules
are adsorbed at an interface, their conformation may change considerably, which may be
23
considered as the denaturation of the protein. The adsorption of proteins at an interface is very
slow depending on their molar masses and it usually takes hours for proteins to reach adsorption
equilibrium.81 The aggregation of protein molecules is strongly influenced by the pH of the
solution, which determines whether the protein molecules are positively charged or negatively
charged and thus affects the electrostatic driving force 82 between the protein molecules.
1.4.1 Surface adsorption behaviour
The term ‘adsorption’ originates from the German physicist Heinrich Kayser.83 Adsorption is the
process by which molecules attach to a surface, and is a consequence of reducing the surface
energy. The adsorption of surfactants and proteins at an interface is a dynamic process and
dynamic equilibrium will be reached when the surface is saturated with molecules. It is usually
described using isotherms, the amount or mass of molecules at the interface as a function of
concentration at constant temperature. There are several isotherm models describing the
adsorption process such as the Langmuir model,84 Freundlich model,85 BET model 86 and Kisliuk
87 model. Among these models, the Langmuir isotherm model is the most widely used model for
surfactant adsorption at an interface.
In order to study the amount of molecules adsorbed at an interface, the Gibbs adsorption equation
is used to describe the relationship between the change in surfactant concentrations in solution
and the change in surface tension. At constant temperature, the equation can be written as88:
where г is the surface excess of the surfactant; dγ is the change of the surface tension; dlnC is the
change in surfactant concentration in log operator; R is the gas constant (8.31 J mol-1 K-1) and T
is the temperature in kelvins. The area per molecule (A) can be calculated as89:
1
ln T
d
RT d C
24
where N is Avogadro number (6.02×1023); г is in moles/m2 and A is in Å2. It can be noted that the
Gibbs equation is directly applicable to small nonionic surfactants when they behave in ideal
conditions. Any deviation from the ideal would usually result in the necessity for the actual
effective concentration C to be calibrated by a coefficient f leading to the term activity a. In
addition, in the case of ionic surfactants in pure water, the dissociation of surfactant molecules
into free ions results in two entities contributing to the surface activity or surface tension changes.
Thus, a prefactor of 2 must be used.
1.4.2 Self-assembly behaviour in solution
In the case of the concentration of the surfactant being above the CMC in solution, the amphiphilic
molecules will self-assemble to form micelles in order to minimise interactions between the
hydrophobic fragments of the surfactant and the solution. This results in all the micelles having a
‘core-shell’ shape with a super hydrophobic core and a hydrophilic shell.
Many parameters determine the structure of self-assembled micelles, such as pH, ionic strength,
surfactant concentrations and temperature, in addition to their own molecular structures. The
mechanism of micelle formation has been investigated since the 1950s. However, the
development of new techniques such as dynamic light scattering (DLS) and small-angle neutron
scattering (SANS) has intensified their study in the last two decades.
1.5 Protein/surfactant complexes
Protein-surfactant interactions at interfaces are a topic of great interest over recent decades. The
technology is widely used in many applications ranging from efficient washing powders to
products for personal hygiene.90 Protein/surfactant interactions are very complex. The effective
201 10A
N
25
way to characterize their interactions is through their size and shape.91 An increased volume of
investigations into these interactions took place between the 1960s and 1970s which resulted in a
greater understanding of the basic principles of how charged proteins and charged surfactants
interact.92, 93 Many of these studies focused on investigating whether the monomeric or micellar
surfactants denatured proteins and how proteins were unfolded by surfactants.
Figure 1.8 An example of T1 and T2 breaking points of surface tension against surfactant concentration at a fixed
PVP polymer concentration. CT and γ in the graph are SDS concentration and surface tensions, respectively. This
graph is referenced from Goddard et al. 94
A widely used method of studying protein/surfactant interactions in these studies was by surface
tension. Normally there are two breaking points in the surface tension graph plotted against
surfactant concentrations at a fixed protein concentration. Figure 1.8 is an example of this, and
shows the T1 and T2 breaking points of surface tensions measured with polymer/surfactant
complexes. It can be seen that in the case of a higher concentration of polymer, the T1 break point
occurs at a higher surfactant concentration, but this may result in the T2 break point occurring at
a lower surfactant concentration. It can also be seen that the surfactant has only a small effect on
the polymer with regards to surface activity. Therefore, the T1 is defined as the critical point
where the surfactant has a small influence on the polymer. In general, the concentration of the
26
surfactant at T1 break point is less than its CMC, as the surfactant above the CMC would
considerably increase the surface activity compared to the complex. Beyond the T1 break point is
a flat region where the surface tension does not change with surfactant concentration. This
indicates that the surfactant molecules in that area are more likely to interact with polymers rather
than be adsorbed at the interface. Once the complexes have become saturated with surfactant
molecules, it can be clearly seen that the surfactant molecules will once again be adsorbed with
other monomers to form micelles, which will further increase the surface activity of the system
until the T2 break point. Thus, T2 represents the saturation status of the interface when the surface
tension does not change any further. Cockbain et al. 95 investigated the T2 value and found that it
to be proportional to the concentration of the polymers.
As a result of these studies, it became apparent that electrostatics are the main driving force for
binding monomeric surfactants to proteins below the CMC.96 Another important driving force is
hydrophobic interactions, which was introduced by Tanford’s classic monograph from 198097, 98.
Tanford et al. also demonstrated that the surfactant-protein interactions are highly affected by the
type of surfactant headgroups and the pH values of the solutions99, 100.
Since the electrostatic driving force plays an important role in surfactant-protein interactions, the
charged surfactants are more likely to interact with charged proteins than non-ionic surfactants101.
Therefore, it can be stated that the surfactants with opposite charges to the protein are more bound
to the protein than surfactants with the same charge, under a given pH. Many studies have shown
that the electrostatic and hydrophobic forces occur at the very first step of interactions in most of
the cases, especially at low concentrations of surfactants below the CMC. Above the CMC, the
micellar interactions with proteins become more complex and cause shape changes of the protein-
surfactant complexes9. Anionic surfactants such as SDS can aggregate to form clusters on proteins
at relatively high concentrations 102-104, resulting in an apparent lowering of the CMC105. It can
27
therefore be concluded that the presence of surfactants can significantly affect how a protein or
polypeptide behaves in both bulk solutions and at interfaces.
References
1. K. Bouchemal, S. Briançon, E. Perrier and H. Fessi, International Journal of Pharmaceutics, 2004, 280, 241-251.
2. A. Özcan, Ç. Ömeroğlu, Y. Erdoğan and A. S. Özcan, Journal of hazardous materials, 2007, 140, 173-179.
3. M. Gallarate, M. Carlotti, M. Trotta and S. Bovo, International Journal of Pharmaceutics, 1999, 188, 233-241.
4. T. L. Metcalfe, P. J. Dillon and C. D. Metcalfe, Environmental Toxicology and Chemistry, 2008, 27, 811-818.
5. E. Emmanuel, K. Hanna, C. Bazin, G. Keck, B. Clément and Y. Perrodin, Environment international, 2005, 31, 399-406.
6. M. Murphy, M. Al-Khalidi, J. Crocker, S. Lee, P. O’Regan and P. Acott, Chemosphere, 2005, 59, 235-246.
7. R. S. Makkar and K. J. Rockne, Environmental Toxicology and Chemistry, 2003, 22, 2280-2292. 8. A. Valstar, M. Almgren, W. Brown and M. Vasilescu, Langmuir, 2000, 16, 922-927. 9. D. Otzen, Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics, 2011, 1814, 562-591. 10. P. Mukerjee and K. J. Mysels, Critical Micelle Concentrations of Aqueous Surfactant Systems: By
Pasupati Mukerjee and Karol J. Mysels, National Bureau of Standards, 1970. 11. A. Dominguez, A. Fernandez, N. Gonzalez, E. Iglesias and L. Montenegro, Journal of Chemical
Education, 1997, 74, 1227. 12. N. J. Turro and A. Yekta, Journal of the American Chemical Society, 1978, 100, 5951-5952. 13. M. Corrin and W. D. Harkins, Journal of the American Chemical Society, 1947, 69, 683-688. 14. E. Matijević and B. Pethica, Transactions of the Faraday Society, 1958, 54, 587-592. 15. S. Mehta, K. Bhasin, R. Chauhan and S. Dham, Colloids and Surfaces A: Physicochemical and
Engineering Aspects, 2005, 255, 153-157. 16. D. Anghel, S. Saito, A. Băran, A. Iovescu and M. Corniţescu, Colloid and Polymer Science, 2007,
285, 771-779. 17. V. M. Garamus, Langmuir, 2003, 19, 7214-7218. 18. D. McDermott, J. Lu, E. Lee, R. Thomas and A. Rennie, Langmuir, 1992, 8, 1204-1210. 19. J. Lu, E. Lee, R. Thomas, J. Penfold and S. Flitsch, Langmuir, 1993, 9, 1352-1360. 20. J. Lu, Z. Li, R. Thomas, E. Staples, L. Thompson, I. Tucker and J. Penfold, The Journal of Physical
Chemistry, 1994, 98, 6559-6567. 21. J. Penfold, R. Thomas, C. Dong, I. Tucker, K. Metcalfe, S. Golding and I. Grillo, Langmuir, 2007,
23, 10140-10149. 22. M. Ueno, Y. Takasawa, H. Miyashige, Y. Tabata and K. Meguro, Colloid and Polymer Science,
1981, 259, 761-766. 23. J. Penfold, E. Staples and I. Tucker, Langmuir, 2002, 18, 2967-2970. 24. R. Reis, G. Pacheco, A. Pereira and D. Freire, Biodegradation—Life of Science, 2013, 31-61. 25. C. N. Mulligan, Environmental pollution, 2005, 133, 183-198. 26. L. Fracchia, I. M. Banat, M. G. Martinotti and M. Cavallo, Biosurfactants and bioemulsifiers
biomedical and related applications-present status and future potentials, INTECH Open Access Publisher, 2012.
28
27. G. Bognolo, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 1999, 152, 41-52.
28. S. S. Cameotra and R. S. Makkar, Current Opinion in Microbiology, 2004, 7, 262-266. 29. M. Pacwa-Płociniczak, G. A. Płaza, Z. Piotrowska-Seget and S. S. Cameotra, International journal
of molecular sciences, 2011, 12, 633-654. 30. J. D. Desai and I. M. Banat, Microbiology and Molecular biology reviews, 1997, 61, 47-64. 31. T. J. Smyth, A. Perfumo, R. Marchant, I. M. Banat, M. Chen, R. K. Thomas, J. Penfold, P. S.
Stevenson and N. J. Parry, Applied microbiology and biotechnology, 2010, 87, 1347-1354. 32. G. Soberón-Chávez, M. Aguirre-Ramírez and R. Sánchez, Journal of Industrial Microbiology and
Biotechnology, 2005, 32, 675-677. 33. S. Bergstrom, H. Theorell and H. Davide, Arkiv for kemi mineralogioch geologi, 1947, 23, 1-12. 34. A. M. Abdel-Mawgoud, F. Lépine and E. Déziel, Applied microbiology and biotechnology, 2010,
86, 1323-1336.
35. R. Rahim, U. A. Ochsner, C. Olvera, M. Graninger, P. Messner, J. S. Lam and G. Soberón‐Chávez, Molecular microbiology, 2001, 40, 708-718.
36. G. Ö zdemir, S. Peker and S. Helvaci, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2004, 234, 135-143.
37. M. Chen, J. Penfold, R. Thomas, T. Smyth, A. Perfumo, R. Marchant, I. Banat, P. Stevenson, A. Parry and I. Tucker, Langmuir, 2010, 26, 18281-18292.
38. S. Peker, S. Helvaci and G. Ö zdemir, Langmuir, 2003, 19, 5838-5845. 39. D. Kitamoto, 2008, 128, 695-706. 40. R. Haskins, J. Thorn and B. Boothroyd, Canadian journal of microbiology, 1955, 1, 749-756. 41. M. Shibahara, X. Zhao, Y. Wakamatsu, N. Nomura, T. Nakahara, C. Jin, H. Nagaso, T. Murata and
K. Yokoyama, Cytotechnology, 2000, 33, 247-251. 42. X. Zhao, T. Murata, S. Ohno, N. Day, J. Song, N. Nomura, T. Nakahara and K. K. Yokoyama,
Journal of Biological Chemistry, 2001, 276, 39903-39910. 43. D. Kitamoto, H. Yanagishita, T. Shinbo, T. Nakane, C. Kamisawa and T. Nakahara, Journal of
Biotechnology, 1993, 29, 91-96. 44. Y. Ueno, Y. Inoh, T. Furuno, N. Hirashima, D. Kitamoto and M. Nakanishi, Journal of Controlled
Release, 2007, 123, 247-253. 45. Y. Ueno, N. Hirashima, Y. Inoh, T. Furuno and M. Nakanishi, Biological and Pharmaceutical
Bulletin, 2007, 30, 169-172. 46. S. S. Bhattacharjee, R. H. Haskins and P. A. J. Gorin, Carbohydrate Research, 1970, 13, 235-246. 47. J. I. Arutchelvi, S. Bhaduri, P. V. Uppara and M. Doble, Journal of industrial microbiology &
biotechnology, 2008, 35, 1559-1570. 48. U. Rau, L. A. Nguyen, S. Schulz, V. Wray, M. Nimtz, H. Roeper, H. Koch and S. Lang, Applied
microbiology and biotechnology, 2005, 66, 551-559. 49. T. Morita, M. Konishi, T. Fukuoka, T. Imura, S. Yamamoto, M. Kitagawa, A. Sogabe and D.
Kitamoto, Journal of oleo science, 2007, 57, 123-131. 50. L. Rodrigues, I. M. Banat, J. Teixeira and R. Oliveira, Journal of Antimicrobial Chemotherapy,
2006, 57, 609-618. 51. M. Konishi, T. Morita, T. Fukuoka, T. Imura, K. Kakugawa and D. Kitamoto, Applied microbiology
and biotechnology, 2008, 78, 37-46. 52. M. Konishi, T. Morita, T. Fukuoka, T. Imura, K. Kakugawa and D. Kitamoto, Applied microbiology
and biotechnology, 2007, 75, 521-531. 53. M. S. Wold, Annual review of biochemistry, 1997, 66, 61-92. 54. T. Maack, V. Johnson, S. T. Kau, J. Figueiredo and D. Sigulem, 1979, 16, 251-270. 55. H. Jeong, B. Tombor, R. Albert, Z. N. Oltvai and A.-L. Barabási, Nature, 2000, 407, 651-654. 56. L. V. Hooper, T. S. Stappenbeck, C. V. Hong and J. I. Gordon, Nature immunology, 2003, 4, 269-
273. 57. F. H. Crick, Journal of molecular biology, 1968, 38, 367-379.
29
58. A. Gutteridge and J. M. Thornton, Trends in biochemical sciences, 2005, 30, 622-629. 59. E. Matzner, Introduction Scoring Tools Phylogenetics Applications Self Test, 2002. 60. C. Levinthal, Mossbauer spectroscopy in biological systems, 1969, 22-24. 61. C. B. Anfinsen, Studies on the principles that govern the folding of protein chains, 1972. 62. C. A. Floudas, H. K. Fung, S. R. McAllister, M. Mönnigmann and R. Rajgaria, Chemical
Engineering Science, 2006, 61, 966-988. 63. Wikipedia, Protein. 64. N. V. Dovidchenko, N. S. Bogatyreva and O. V. Galzitskaya, Journal of bioinformatics and
computational biology, 2008, 6, 1035-1047. 65. M. Perutz, British medical bulletin, 1976, 32, 193-194. 66. E. Schramm, J. Mende, V. Braun and R. Kamp, Journal of bacteriology, 1987, 169, 3350-3357. 67. S.-i. Iwasaki, T. Asami and H. Aoyagi, 1989, 141-158. 68. P. Birkenfeld, N. Haratz, G. Klein and D. Sulitzeanu, Clinical Immunology and Immunopathology,
1990, 54, 14-25. 69. N.-O. Ku, J. M. Darling, S. M. Krams, C. O. Esquivel, E. B. Keeffe, R. K. Sibley, Y. M. Lee, T. L.
Wright and M. B. Omary, Proceedings of the National Academy of Sciences, 2003, 100, 6063-6068.
70. Y.-N. Wang and W.-C. Chang, Journal of Biological Chemistry, 2003, 278, 45848-45857. 71. R. C. Betz, L. Planko, S. Eigelshoven, S. Hanneken, S. M. Pasternack, H. Büssow, K. Van Den
Bogaert, J. Wenzel, M. Braun-Falco and A. Rütten, The American Journal of Human Genetics, 2006, 78, 510-519.
72. J. M. Cardamone, Journal of molecular structure, 2010, 969, 97-105. 73. G. Secchi, Clinics in Dermatology, 2008, 26, 321-325. 74. A. K. Smiley, J. M. Klingenberg, S. T. Boyce and D. M. Supp, Burns, 2006, 32, 135-138. 75. J. F. Conway and D. A. Parry, International Journal of Biological Macromolecules, 1988, 10, 79-
98. 76. H. W. Heid, E. Werner and W. W. Franke, Differentiation, 1986, 32, 101-119. 77. P. M. Steinert, A. C. Steven and D. R. Roop, Cell, 1985, 42, 411-419. 78. P. Steinert and D. Roop, Annual review of biochemistry, 1988, 57, 593-625. 79. L. M. Dowling, W. G. Crewther and A. S. Inglis, Biochem. J, 1986, 236, 695-703. 80. J. Yu, D.-W. Yu, D. M. Checkla, I. M. Freedberg and A. P. Bertolino, Journal of Investigative
Dermatology, 1993, 101, 56S-59S. 81. M. A. Bos and T. van Vliet, Advances in Colloid and Interface Science, 2001, 91, 437-471. 82. W. Wang, S. Nema and D. Teagarden, International Journal of Pharmaceutics, 2010, 390, 89-99. 83. H. Kayser, Annalen der Physik, 1881, 248, 526-537. 84. A. Halperin, A. Buhot and E. Zhulina, Journal of Physics: Condensed Matter, 2006, 18, S463. 85. C. Ng, J. N. Losso, W. E. Marshall and R. M. Rao, Bioresource technology, 2002, 85, 131-135. 86. A. Lecloux and J. Pirard, Journal of Colloid and Interface Science, 1979, 70, 265-281. 87. K. Schönhammer, Surface Science, 1979, 83, L633-L636. 88. D. Bangham, Transactions of the Faraday Society, 1937, 33, 805-811. 89. X. Zhao, F. Pan, B. Cowsill, J. R. Lu, L. Garcia-Gancedo, A. J. Flewitt, G. M. Ashley and J. Luo,
Langmuir, 2011, 27, 7654-7662. 90. R. Green, T. Su, H. Joy and J. Lu, Langmuir, 2000, 16, 5797-5805. 91. N. J. Turro, X.-G. Lei, K. Ananthapadmanabhan and M. Aronson, Langmuir, 1995, 11, 2525-2533. 92. F. W. Putnam, Advances in protein chemistry, 1948, 4, 79-122. 93. J. Steinhardt and J. A. Reynolds, Multiple equilibria in proteins, Academic Press, 1969. 94. E. D. Goddard, Colloids and Surfaces, 1986, 19, 255-300. 95. E. Cockbain, Transactions of the Faraday Society, 1953, 49, 104-111. 96. E. B. Abuin and J. Scaiano, Journal of the American Chemical Society, 1984, 106, 6274-6283. 97. S. Magdassi, Surface activity of proteins: chemical and physicochemical modifications, CRC
Press, 1996.
30
98. C. Tanford, The Hydrophobic Effect: Formation of Micelles and Biological Membranes 2d Ed, J. Wiley., 1980.
99. C. Tanford, Journal of molecular biology, 1972, 67, 59-74. 100. J. A. Reynolds and C. Tanford, Journal of Biological Chemistry, 1970, 245, 5161-5165. 101. W. L. Mattice, J. M. Riser and D. S. Clark, Biochemistry, 1976, 15, 4264-4272. 102. M. Schwuger, Journal of colloid and interface science, 1973, 43, 491-498. 103. M. L. Smith and N. Muller, Journal of Colloid and Interface Science, 1975, 52, 507-515. 104. M. Fishman and F. Eirich, The Journal of Physical Chemistry, 1971, 75, 3135-3140. 105. M. N. Jones, Biochimica et Biophysica Acta (BBA)-Protein Structure, 1977, 491, 121-128.
31
Chapter 2
Theoretical background and experimental procedures
This chapter describes the main techniques used in the research presented in this thesis, including
surface tension (ST), spectroscopic ellipsometry (SE), dual polarization interferometry (DPI),
quartz crystal microbalance with dissipation (QCM-D), neutron reflection (NR), small-angle
neutron scattering (SANS) and dynamic light scattering (DLS). The chapter covers the basic
theoretical background and introduces the specific instruments and their use.
2.1 Surface tension
Surface tension is the surface energy that tends to hold a liquid in a shape with the least surface
area. It involves both static surface tension (SST) and dynamic surface tension (DST). The SST
is the value of the surface tension when the interface reaches in thermodynamic equilibrium,
whilst the DST is a measurement of the bubble pressure with a bubble produced by a capillary
inserted in the solution. Both methods provide a way to investigate the adsorption behaviours of
materials.
2.1.1 Static surface tension
At air/liquid interfaces, the water molecules are more attracted to each other by cohesive forces.
There is a net attraction of water molecules towards bulk solution that causes water to behave like
a droplet or a membrane. Static surface tension (SST), is usually represented by the symbol γ and
is expressed in units of mN/m or J/m2. Assuming a liquid is in a rectangular frame with one side
movable, let F be the force required to stop the movable side which is pulled by the liquid film
and L be the length of the movable side. The surface tension can then be calculated as:
32
(2.1.1.1)
where E is the energy due to the force F and A is the area moved by the side. Pure water has a
high surface tension of 71.99±0.05 mM/m at 20 °C 1 mainly due to hydrogen bonds between
molecules. Surface-active materials such as surfactants can significantly reduce the surface
tension of the solution by adsorbing at interfaces since this lowers the energy of the system. In the
current work presented in this thesis, the surface tension was used to determine CMCs of studied
surfactants and to investigate the adsorption behaviour of proteins.
The tensiometer used in this work was the Krüss K100 which has an automatic pump. All the
measurements were performed using a Wilhelmy plate at 25 °C, which is thought to be preferable
to the Du Nouy ring method for use with high surface tension liquids. Prior to each measurement,
the tensiometer was calibrated by measuring UHQ water. A solution of the highest concentration
was measured first, followed by measurements in diluted solutions. The dilution process of the
solution was carried out automatically with a pump, and was followed by stirring for 60 seconds
with a magnetic stirrer.
2.1.2 Dynamic surface tension
The technique used to measure dynamic surface tension is also termed the maximum bubble
pressure measurements. This technique can measure the bubble time with an accuracy of
milliseconds. The tip of a capillary in solution continuously exposes air and generates bubbles in
solution. A sensor is connected to the tiny tube in order to measure the bubble pressure. The
surface tension (γ) can be calculated from the Laplace equation, which is expressed as:
γ=Pr/2 (2.1.1.2)
where Pr is the bubble pressure and r is the bubble radius.
2 2
F F x E
L L x A
33
Dynamic surface tension measurements are widely used for processes including printing and
coating, as these are all performed under dynamic conditions. In some cases, there may not be
enough time for the surface-active materials to reach surface equilibrium. DST is more appropriate
and provides more information than SST in such cases.
2.2 Spectroscopic ellipsometry, (SE)
2.2.1 Background theory
Spectroscopic ellipsometry (SE) is an optical technique that is widely used to investigate
interfacial properties of thin films. It is a non-destructive and efficient technique, which measures
the polarization changes of a polarized light after reflection by a sample of interest. The resolution
of SE ranges from several angstroms to several micro meters. Surface properties such as thickness
and refractive index can be obtained by appropriate data analysis.
Figure 2.1 Schematic representation of wave reflection for spectroscopic ellipsometry. The incident light with a
linearly polarized plane wave is reflected by the sample surface and is converted into an elliptically polarized wave.
Figure 2.1 shows the schematic representation of an incident linearly polarized light wave
undergoing reflection and being converted into an elliptically polarized beam. The incident light
is linearly polarized and consists of two polarizations: the p-polarization and the s-polarization,
which are taken to be perpendicular and parallel to the sample surface respectively. The reflection
34
of the two polarized waves is described by Fresnel reflection coefficients, resulting in the elliptical
polarization for the reflected wave. The Fresnel’s reflection coefficients rs and rp are given by
(2.2.1.1)
where n0, n1, θ0 and θ1 are denoted in Figure 2.4. The Fresnel’s transmission coefficients are given
by
(2.2.1.2)
SE can be used to measure Δ and Ψ, which are the phase shift between incident and reflected
beams and the amplitude ratio of reflected and incident beams, respectively. These two parameters
are related to the ratio of the amplitudes of s to p components, as given by2-4
(2.2.1.3)
where and are the phase changes of the p and s waves, respectively. The ellipsometer
measures the ratio of rp and rs values, and the results are very robust and reproducible.
2.2.2 Data analysis
Ellipsometry measures the phase (Δ) and amplitude (Ψ) changes of the polarized light to determine
the surface properties of film thickness and refractive indices. Δ and Ψ can only be converted
directly into optical constants in the case of a homogeneous material. In all other cases, a model
rs=
n0
cosq0- n
1cosq
1
n0
cosq0+ n
1cosq
1
rp
=n
0cosq
1- n
1cosq
0
n0
cosq1+ n
1cosq
0
ts=
2n0
cosq1
n0
cosq0+ n
1cosq
1
tp
=2n
0cosq
1
n0
cosq1+ n
1cosq
0
tan(y )eiD =r
p
2
rs
2e
i(dp-d
s)
d p d s
35
is required to provide estimation and comparison with the measured data. The procedures of the
data analysis is as shown in Table 2.15,
Table 2.1 Data fit procedures of the ellipsometer.
Step two and step three are usually repeated until a good fit is produced. A regression analysis is
performed during the procedure. The Mean Squared Error (MSE) method is applied on every fit
iteration to find the best fit with the minimum MSE value.
After determining the coefficients of the thickness (τ) and refractive index (n) of the layer, the
mass of material per area can be calculated using the equation proposed by de Feitjer et al.5,
(2.2.2.1)
where n0 is the refractive index of the solution; C is the concentration of the sample solution and
dn/dC is the rate of change of refractive index with respect to the solution concentration.
2.2.3 Instrumentation
In this research, the ellipsometer M-2000 (J.A Woollam Co.) was used to determine the properties
of interfacial layers at solid/liquid and solid/air interfaces.
Measurement
•Experimental data
Choose model
•Optical constants n, k
Fit
•Comparison with model and measured data
Produce results
•Film thickness
•Refractive index
G =
t(n- n0)
(dn/dC)
36
Figure 2.2 Schematic representation of the ellipsometry cell. The plastic cap on the right is used to sample wafers
in the cell.
All ellipsometry measurements were performed in a specially made cell shown in Figure 2.2. Two
windows were installed on both sides of the cell for incident and exiting light and were accurately
adjusted at an angle of 70o. The windows are made of amorphous quartz in order to minimise their
effect on the incident and reflected light. A silicon wafer was cut to the size of 12×12 mm to fit
into the illumination area inside the cell. Prior to measurements, the silicon wafers were wiped
with 5% Decon90 solution and rinsed with plenty of water. The thickness of the SiO2 layer was
then measured by ellipsometry in air. The cleaning process was repeated until the SiO2 thickness
became constant at 11-14 Å.
2.3 Quartz crystal microbalance with dissipation, (QCM-D)
The quartz crystal microbalance with dissipation, (QCM-D), is an acoustic technique and was
used in this research for obtaining interfacial acoustic measurements. The technique is an upgrade
of QCM, which has been used over previous decades. By measuring the change of the sensor’s
frequency (∆f), the dissipation (∆D) produced by the adsorbed material on the surface of the sensor
can be deduced and the information about the adsorbed material can be determined, including the
adsorbed mass, layer thickness and layer rigidity.
37
A simple Sauerbrey model is used for non-dissipative rigid films. It relates the frequency change
to the hydrated mass of the deposition layer in a simple way, with the adsorbed mass directly
proportional to the change of the frequency. The Sauerbrey equation6 is written as
(2.3.1)
where ∆m is the adsorbed layer mass, C is the mass sensitivity constant (17.7 ng/cm2/Hz at f = 5
MHz), and ∆fn is the change in frequency at harmonic n. The values of layer thickness and
adsorbed mass generally have a ±15% difference by using different harmonics. This is due to the
fact that the softness of the film does not account for the calculation, but has a different effect
with different harmonics.
A more precise estimation for soft layers was achieved using the Kevin-Voigt model with
combined measurement of frequency and dissipation.6 The dissipation (D) is defined as
(2.3.2)
where Elost is the energy lost during each sensor oscillation and Etotal is the total energy in the
oscillation. A simple way to determine whether an adsorbed layer is rigid or soft is by calculating
the ratio ∆D/∆f. If the value of this ratio is high (above 2), the Kevin-Voigt model is more
appropriate. The ∆f and ∆D data were measured at several harmonics (n= 3, 5, 7, 9, 11, 13) and
were used to calculate viscoelastic properties on the adsorbed layer. The thickness, viscosity and
shear modulus can be related to ∆f and ∆D by the Kevin-Voigt model given by
(2.3.3)
( / )nm C f n
2
lost
total
ED
E
22
3 3 11 0 1 2 2 2
0 0 3 3 1 1
2
3 3 11 2 2 2
0 0 3 3 1 1
12
2
12
f h hh
D hf h
38
where is the crystal’s density, the crystal’s thickness, is the viscosity of the bulk solution,
is the density of the solution and is the angular frequency of the oscillation.7 It can be seen
that the thickness, viscosity and shear modulus are scaled by the calculated film density. In this
thesis, the Q-sense (Biolin Scientific) was used with the software package “QTools”. The model
of QCM sensors used was ‘QSX335’, which has a SiO2 layer coated on a thick titanium layer.
These specific types of sensor can also be used in optical instruments such as Ellipsometry.
2.4 Dual polarisation interferometry, (DPI)
2.4.1 Background
Dual polarisation interferometry (DPI) is a powerful technique used to study properties of
solid/liquid interfaces. It is a more recently developed technique than reflectivity and ellipsometry
that addresses the problems of the high cost of neutron reflection and the low resolution of
ellipsometry.
Figure 2.3 Schematic representation of the illuminated slab waveguide interferometer. Five thin films of silicon
oxynitride layer are coated on the silicon substrate. Two sample channels can be used in parallel to provide
reproducible measurements.8
0 0h 3
3
39
The DPI consists of an illuminated slab waveguide interferometer that is used to produce
interference patterns in the far field (Figure 2.3). Two orthogonal polarisations of light, the
transverse electric mode (TE) and the transverse magnetic mode (TM), are produced
simultaneously. The refractive index, thickness and density of the adsorbed layer can then be
calculated by analysing the interference patterns.
Figure 2.3 shows a representation of five thin films of silicon oxynitride on a silicon substrate.
Each layer is stacked on another layer with a lower or higher refractive index. The second and
fourth layers have a respectively higher refractive index than the others since they function as
sensing and reference waveguides. The cladding layers are used to isolate the sensing and
reference waveguides from each other and from samples. As the length between the far field
region and the waveguides is much longer than the thickness of the waveguide layers, the light
emitted from the two waveguides can be regarded as two point sources. The divided polarized
light beam travels simultaneously and the two modes diffract and produce Young’s fringes in the
far field, where they are recorded by camera.9 The centre of the sensing waveguide is exposed to
adsorbed samples. As the two modes of the waveguides are very sensitive, small changes of the
sample on the sensing waveguide can produce an effect on the light phase and intensity, thus
producing a final change on the interference patterns. Therefore, changes on the sensing
waveguide surface such as adsorption and desorption can be calculated by analysing changes in
the interference patterns.
2.4.2 Theoretical principles
When light is travelling through the waveguide, Snell’s law 10 applies for light reflection and
refraction inside the waveguide. Thus the refractive index and thickness of each slab and
waveguide layer is critical for the light transmission. The values are given in the Table below:
40
Layer Refractive index
±0.001
Thickness (Å)
±1 Å
Sensing waveguide 1.520 1×104
Cladding 1.470 3×104
Reference waveguide 1.520 1×104
Cladding 1.485 2×104
The value of the refractive index of each layer is controlled by doping nitrogen molecules. The
refractive index of the waveguide layers is higher than that of the cladding layers, which confines
the light propagating through the waveguides.
In this case, the light can appropriately by regarded as a wave, as the boundary decay of the wave
is calculated to obtain information concerning the adsorbed layers. The light that propagates
through the waveguides exhibits exponential decay at the boundaries and produces an evanescent
wave field. The evanescent field decays as a function of the distance from the boundary (interface)
and thus can interact with the electrons provided by the adsorbed surface of the samples. The
propagation constant β may be introduced here and represents the rate of decay of an
electromagnetic wave propagating in a particular direction. In the case of the reference waveguide
layer, the propagation constant β remains constant because its boundaries are isolated from any
changes caused by surface adsorption. In the case of the sensing waveguide layer, β changes with
different samples since its top boundary is exposed to the sample solution. Therefore, the phase
difference between the sensing and reference waveguides, ∆φ, can be calculated as:
(2.4.2.1) L
41
where L is the interaction length of the sensing waveguide. After a Fourier transformation, the
phase shift ∆φ may be obtained by the change in interference fringes, which is continuously
monitored by a CMOS camera. Since L is a known parameter, the measured data can be
transferred to changes in the propagation constant ∆β. Thus the change in refractive index on the
surface layer, ∆N, can be obtained using
(2.4.2.2)
where k0 is the free space wavenumber.
In order to analyse the TE and TM modes, Maxwell’s equations are applied to the light
transmission to describe how the magnetic and electric fields interact in the waveguide. The
Maxwell equations in differential form can be written as:
(2.4.2.3)
where E is the electric field, B is the magnetic flux density, ρ is the electric charge density, μ is
the permeability and ε is the permittivity. Gauss’s law (the first equation) shows that an electric
field that left a 3-dimentional space is proportional to the charge inside. Integration of the second
equation shows that the total magnetic flux of a closed surface is zero. The Faraday’s law of the
third equation implies that a change in magnetic fields can produce electric fields, while the forth
0
Nk
0
0 0
0
( )
E
B
HE
t
EB J
t
42
equation (Ampere’s law) implies that changes of electric fields or electric current can produce
magnetic fields.
In order to obtain appropriate values for the TE and TM modes, the following equations may be
solved using Maxwell’s equations:
(2.4.2.4)
where D and E are constant coefficients, kx is the wave vector perpendicular to the waveguide
surface, kz is the wave vector parallel to the surface and w is the light frequency. At the interface,
the tangential parts of the magnetic field (H) and the electric field (E) must be continuous. For
each layer, kx can be determined by kz:
(2.4.2.5)
where n is the refractive index of the layer. In order to obtain values of the coefficients A and B,
the evanescent wave function at the surface of the sensing waveguide is first solved to match the
tangential components of E and H for both modes. For each TE and TM polarization, there will
be a range of layer refractive index and thickness combinations that satisfy the solutions of
Maxwell’s equations. However, there will be only one set of values for the layer thickness and
refractive index that corresponds to both TE and TM modes simultaneously in the case of the
adsorbed layers being homogenous.11
( )
( )
( , , ) ( )
( , , ) ( )
x x z
x x z
ik ik i k z wt
z
ik ik i k z wt
z
H x z t D Ae Be e
E x z t E Ae Be e
2 2 1/2
0( )x zk nk k
43
2.5 Neutron reflection, (NR)
Neutron reflection (NR) is a powerful technique that is widely used to investigate the structure of
thin films at solid/liquid and air/liquid interfaces. Over the last two decades, the technique has
been intensively used for determination of the adsorption of surfactants, lipids, proteins and their
mixtures.11, 12 The following provides a full description of the NR technique, which is the primary
technique used in the work presented in this thesis.
Due to the wave-particle duality principle, neutrons can be reflected as waves, as in the case of x-
rays and visible lights. The wavelength of neutrons ranges from 1 to 10 Å, giving a high resolution
of 2-3 Å when used for surface layer measurements. Compared to other techniques such as
ellipsometry, DPI and x-ray, NR has some unique features. Most notably, as the neutron is neutral
it interacts with the nucleus of the sample rather than the electron. It is more sensitive for the
measurement of light elements including hydrogen, carbon and oxygen. It is therefore more
convenient to use neutrons in biological measurements. Secondly, the isotopes of hydrogen (H)
and deuterium (D) have large differences in their scattering length (-3.74×10-5 Å-1 and 6.67×10-5
Å-1 respectively), thus giving H2O and D2O different scattering length densities (-0.6×10-6 Å-2 and
6.35×10-6 Å-2). Therefore, D2O with high scattering length density (SLD) can be used to label
samples of molecules to highlight the sample parts of interest. By deuterating part of the samples
or tuning the H2O/D2O ratios of solvent, it also provides a method of carrying out parallel
experiments under the same conditions in order to obtain precise conformation of samples. Finally,
the neutral character gives neutrons the ability to highly penetrate the sample without perturbing
its structure. However, the high cost of the neutron infrastructure and maintenance is the main
disadvantage of this technique. Other disadvantages include concerns over the radioactive nature
of the neutron beam and the relatively low efficiency compared to optical techniques.
44
2.5.1 Theoretical background
Neutron reflection is a technique with high sensitivity that provides precise information on the
structure of the sample film, provided that the film is homogeneous. Neutron can be regarded as
a quantum particle and its reflection described by its interaction with an energy barrier. The time-
dependent Schrӧdinger’s equation describes principles of neutron reflection 13. Schrӧdinger’s
equation in one dimension is given by
(2.5.1.1)
where ħ is the Planck constant divided by 2π, Ψ is the wave function of the quantum system, m
is the particle mass, z is the direction, V is the potential energy barrier and E is the neutron energy,
described as
(2.5.1.2)
where k is the wave vector perpendicular to the surface and ρ is the scattering length density,
defined as
(2.5.1.3)
where bj is the scattering length of the nuclei and nj is the number of nuclei per unit volume. The
scattering length varies dramatically between isotopes such as hydrogen and deuterium.
2.5.1.1 Calculation of the refractive index
Assuming light of wavelength λ is travelling from media 0 to media 1, the refractive index n is
can be expressed as
2 2
2( ) 0
2
dV E
m dz
2 2 22;
2
kV E
m m
j j
j
b n
45
(2.5.1.1.1)
where A=Nb/2π, C=Nσ/4π, N is the atomic number density, b is the coherent scattering length, σ
is the adsorption cross section and λ is the wavelength of neutron. A more detailed description of
the relation between refractive index and wavelength is given by Hayter et al. 14 For atoms in most
polymer molecules, the adsorption cross sections are negligible. Therefore the imaginary part of
the equation above may be neglected. Using Snell’s law, it can be determined that the incident
and refraction angles are related as
(2.5.1.1.2)
where n0, n1, θ0 and θ1 are shown in Figure 2.4. In the case where media 0 is air, then n0 equals
1. Since the critical angles of neutron reflection is very small, equation 2.5.1.1.2 may be
approximated as
(2.5.1.1.3)
The small-angle approximation may be applied as
(2.5.1.1.4)
Combining equations 2.5.1.1.2, 2.5.1.1.3 and 2.5.1.1.4 then
(2.5.1.1.5)
For incident angles higher than θc, the reflectivity R can be calculated using Fresnel’s law,
(2.5.1.1.6)
21n A i C
0 0 1 1cos cosn n
1cos c n
2
cos 12
1
2( )c Nb
2
0 0 1 1
0 0 1 1
sin sin
sin sin
n nR
n n
46
2.5.1.2 Reflectivity for a uniform layer
The key aim of the neutron reflection technique is measurement of the specular reflection as a
function of neutron wavelength perpendicular to the surface. There are two methods of fit the
reflectivity profiles: the direct analytic method and the optical matrix method. The specular
neutron reflection of a uniform layer is shown in Figure 2.4.
The amplitude of the reflected neutron beam is given by Fresnel’s law 15,
(2.5.1.2.1)
where rij is the Fresnel reflection coefficient at the ij interface and β is the optical path length,
(2.5.1.2.2)
where pi and pj are wave vectors parallel and perpendicular to the surface respectively; n1 and τ1
are layer refractive index and thickness as shown in Figure 2.4. Specular reflection is then described as
the square of the amplitude,
(2.5.1.2.3)
The direct analytic method can be extended for up to three layers. Beyond this the calculation
becomes unwieldy.
2
01 12
2
01 121
i
i
r r eR
r r e
1 1 1
2; sin
i j
ij
i j
p pr n
p p
22
01 12
2
01 121
i
i
r r eR
r r e
47
Figure 2.4 Schematic representation of specular reflection of neutron from a uniform layer.
2.5.1.3 Reflectivity from multi layers
The optical matrix method is widely used for multiple layers in the case where the boundaries of each layer
are continuous. By applying Maxwell’s equations at boundaries, the amplitude of the reflected beam may
be calculated for 2×2 matrices, expressed as
(2.5.1.3.1)
where pj and βj are defined in equation 2.5.1.2.1 and 2.5.1.2.2. The final amplitude of the reflected
beam is
(2.5.1.3.2)
where Mf is the resultant matrix and Mij refers to the matrix element of Mf. The reflectivity of the
whole multilayer is then given as
(2.5.1.3.3)
1cos ( )sin
sin cos
j j
jj
j j j
pM
ip
11 21
1 2
12 22
f n
M MM M M M
M M
2
11 12 21 22
11 12 21 22
( ) ( )
( ) ( )
s i s
s i s
M M p p M M pR
M M p p M M p
48
where pi and ps represent the initial and substrate phases. The optical matrix method can, in theory,
deal with infinite layers, the only limitation being the speed of computing. Although the
reflectivity can be calculated precisely, it is difficult to invert the reflection results to provide
information on each layer including the layer thickness and scattering length density. Therefore
the kinematic (Born) approximation is introduced.
2.5.2 Instrumentation
The INTER 16 and SURF 17 reflectometers at ISIS, Rutherford Appleton Laboratory (RAL),
Didcot, UK, are the main instruments used in the work presented here. The neutron beam is
produced by a spallation source with high flux in which highly energetic protons from an
acceleration hit a target material and emit neutrons from the nuclei of the target atoms. INTER is
a pioneering instrument at ISIS that uses a lower repetition rate of a neutron source in order to
produce a wider range of neutron wavelengths in each pulse. Both instruments use the neutron
time of flight (TOF) method to measure the neutron wavelength at a fixed angle of incident. Figure
2.5 shows a schematic design used in both reflectometers at ISIS, UK.
Figure 2.5 Schematic representation of the INTER reflectometer at the ISIS pulsed neutron source, UK. 18 S1-S4 are
the four slits used to define neutron beams.
49
In the work presented here, SURF was mainly used at the solid/liquid and air/liquid interfaces. In
the SURF reflectometer, initially, a collimation tube guides the neutron beam to the controller
chamber with an angle of 1.5° to the horizontal. Subsequently, a nimonic chopper defines the
beam by cutting off the fast neutrons and gamma rays of each pulse. The double disc chopper is
further applied at a rotation rate of 25 or 50 Hz to result in neutrons at the wavelength ranging
from 0.5 Å to 6.7 Å or 0.5 Å to 13.6 Å depending on the chopper frequency 19. After a coarse
collimation the neutron beam has dimensions of 60 mm × 10 mm. The sample is then mounted on
a platform (10.25 m from the neutron source and 1.75 m form the detector), a fine collimation is
applied by tuning the height of the sample, the lateral position of the sample and the widths of
four cadmium/B4C slits, S1, S2, S3 and S4, as shown in Figure 2.5. All the procedures of fine
collimation are computer controlled in order to provide high precision (the sample position is
accurate to 10 μm and the detector angle to 0.001°). There are two options for the neutron detectors:
a single 3He gas detector or a two-dimensional area detector based on a ZnS scintillator design.19
Both detectors are mounted separately, with a variable sample to detector distance of 1.0 m to
2.43 m.19 The SURF reflectometer also provides a ‘supermirror’ (2.5 times the critical angle of
nickel with a radius of curvature of 160 m) 19 that can be positioned on the neutron source in order
to change the incident angle and give a wider range of Q values (wave vector transfer). In the
work presented here, a single detector was used at neutron beam angles of 1.5° and 2.3° (with the
supermirror). The lower angle provides a Q range from 0.05 Å to 0.57 Å whilst the higher angle
provides a Q range from 0.01 Å to 0.5 Å, which is useful in some systems to determine the ‘critical
edge’ of a reflectivity profile.
50
2.5.3 Experimental procedures
2.5.3.1 Air/water interface
A tray consisting of five polytetrafluoroethylene (PTFE or Teflon) troughs was used as the sample
holder in the SURF reflectometer. The INTER reflectometer had a similar tray consisting of seven
Teflon troughs. Each Teflon trough had a dimension of 235 mm × 50 mm × 2 mm. The trough
was filled with 30-50 ml solution, dependent on the surface activity of the sample. A lid sealed
with a rubber ring was put on each trough after the changing of a sample to minimise evaporation
of the solution. The time that the sample remained in the trough was minimised, to avoid any
evaporation changing the height of the interface and thus the collimation of neutron reflection.
In the work presented in this thesis, all measurements were performed at 20 °C (room temperature).
Prior to each experiment, the Teflon trough was cleaned by a solution of 5% Decon90 for a few
minutes and rinsed with copious UHQ water. After the washing procedure, the trough was
thoroughly dried with nitrogen gas to prevent H/D exchange with the sample solution. For each
measurement, the sample solution was squeezed carefully into the trough from a 20 ml syringe to
make sure the trough was filled thoroughly. Any bubbles produced during the sample filling were
removed using a pipette.
Prior to the sample measurement, a direct neuron beam was measured for later subtraction from
all the data. The reflectivity profile of a pure D2O solution was obtained to provide the scaling
factor for all the subsequent experiments. For the fine collimation procedure, a height scan of the
sample by the neutron beam was applied to determine the position of maximum reflection. For
sample solutions with weak reflection signals, the height of the sample was determined by eye
with a reflected laser beam instead of with the neutron beam scan. After each measurement, the
data was firstly roughly fitted by the software ‘OpenGenie’ to make sure that no large scale errors
51
had been made during the preparation process. Once this had been established, a further more
precise fit procedure was obtained using the program ‘motofit’ in the Igor software.
2.5.3.2 Solid/water interface
The tray used for the solid/water interface consisted of four sample positions. The rectangular
sample cell was tightened with two aluminium plates, each with a screw hole on each of the four
corners. When installing the cell, four long screws were tightened simultaneously with the
aluminium plate to clamp a transparent Perspex trough onto a silicon block. This step required
care in order to prevent any imbalance of pressure on the silicon block. The silicon block was 60
(length) × 50 (width) × 12 (thickness) mm and provided an area of 40 × 30 mm2 for neutron
reflection after removing the sealed edge. When the cell was positioned in the tray, the polished
surface of the neutron block was put upside down, along with the sample solution. Therefore, the
beam travelled firstly through the silica block and it was then reflected from the solid/liquid
interface.
Prior to the measurements, all silicon blocks were cleaned with Piranha solution (a mixture of 98%
H2SO4 and 30% H2O2 with a ratio of six to one) for 90 seconds at 90 °C. The blocks were then
rinsed with copious water and dried with nitrogen. Each block was then measured with
ellipsometry to maintain the thickness of the SiO2 layer at between 12 and 16 Å.
Fine collimation of neutron beam for reflection at the solid/liquid interface is more complex than
in the case of air/liquid, since the illumination area is smaller and the interface had to be adjusted
to be horizontal, which was not an issue in the case of the air/water interface. The collimation
process consisted of three steps. The first step was to move the detector parallel to the interface in
order to roughly scan the height of the sample and obtain an approximate position. The next step
was to move the detector to its primary place and to gradually change the angle of the sample.
This involved scanning each angle of the sample with the neutron beam and obtaining the angle
52
of maximum reflection. The third step was to scan the height of the sample once more with a
precision of 0.1 mm to calculate the final sample position.
2.5.4 Data analysis
The neutron reflection data were analysed using the program ‘motofit’ in the Igor software.
Calculation of the fit profile was based on the optical matrix method, as mentioned in section
2.5.1. The fit parameters consisted of the number of layers, the layer thickness and the layer SLD
of each layer. Using the least-squares routine, the program determines the best fit.
Assuming a uniform monolayer consisting of water and surfactant molecules, the molecular
volume fraction of the material ( ) is given by
(2.5.4.1)
where , and are the scattering length densities for the surfactant, the buffer solution (-0.56
× 10-6 Å for H2O and 6.35 × 10-6 Å for D2O) and the layer in total, respectively. The area per
molecule (A) can then be obtained through
(2.5.4.2)
where is the volume of the surfactant molecule and is the thickness of the layer. The coated
amount of surfactant ( in the unit of mg/m2) can be expressed as
(2.5.4.3)
where the molecular weight of surfactant is expressed in g/mol and A is in Å2.
s
ws
s w
s w
;s sV A
sV
6.02
MolecularWeight
A
53
2.6 Small-angle neutron scattering, (SANS)
Small-angle neutron scattering (SANS) is widely used to measure both the shape and size of
particles in solution such as polymers, surfactants and biomaterials. The technique is able to probe
structures at length scales ranging from 5 Å to more than 1000 Å. In contrast to small-angle x-ray
scattering (SAXS), SANS is a non-destructive technique for fragile samples, which makes it
suitable for use in biological systems. However, disadvantages of SANS include the high cost of
neutron sources and model dependent analysis.
2.6.1 Theoretical background
Figure 2.6 Schematic representation of the small-angle neutron scattering measurement.20 ki and ks are vectors of
incident and attenuated transmitted beams, respectively; q is the wave vector where q= ks- ki.
Figure 2.6 is a schematic representation of SANS measurement. As the neutron wavelength (≈10-
10 m) is much larger than the scattering nucleus ((≈10-15 m), the scattered wave is spherical and
may be described by the planar monochromatic sphere wave equation,
(2.6.1.1) ( )
0( , ) i t krr t e
54
where Φ0 is a constant, r is the wave vector, t is time, ω is the angular frequency and k is the wave
number, 2π/λ. Since in SANS, only the coherent elastic interaction between the neutron and
sample particles is considered, the wave vector k parallel to the incident beam will be changed
according to
(2.6.1.2)
where ki and ks are the incident and scattered neutron wave vectors, respectively. From Figure 2.6
it follows that
(2.6.1.3)
where the angle between ki and ks are 2θ. Bragg’s Law of Diffraction gives
(2.6.1.4)
where d is the distance between scattered particles. By combining equation 2.6.1.3 with equation
2.6.1.4, the following equation is obtained,
(2.6.1.5)
The above equation relates the distance in reciprocal spaces to changes of the wave vectors, which
allows the observation area during the experiment to be defined. It also indicates that the larger
the q range the larger the range of sample particle sizes that can be measured.
2.6.2 Data analysis
SANS is a scattering technique that measures the scattering cross section, which contains
information about particle size, shape and particle interactions21. The scattering intensity from a
sample solution can be expressed as
i sq k k
4sinq
2 sind
2q
d
55
(2.6.2.1)
where q, N, V and are the momentum transfer, the number of particles in a volume that can be
scattered, the volume of a particle and the difference in scattering length density between the
scattering particles and the bulk solution, respectively. P(q) is termed the form factor which
denotes the degree of neutron interference scattered from different parts of the same object, whilst
S(q) is the structure factor that describes the scattered neutron interference from different objects.
B is the background signal.
There are two methods available to analyse SANS data profile: the model-independent and the
model-dependent methods.22 The first deals with the data directly while the latter uses
mathematical models such as cylinder, core-shell and ellipsoidal models to manipulate the data.22
The analysis in this thesis uses the latter method with the ellipsoidal model as the fit model for
56
protein and protein/surfactant solutions. The output of the 2D scattering intensity function for an
ellipsoid is given by 23
(2.6.2.2)
(2.6.2.3)
(2.6.2.4)
where scale is the scale factor, is the angle between the axis of the ellipsoid and the beam, V is
the volume of the ellipsoid, Ra and Rb are the length along and perpendicular to the rotation axis
of the ellipsoid respectively, and ∆is the difference in the scattering length density between the
scattered particle and the bulk solution.
The structure factor S(q) is determined by the interference effect between scattering particles. This
is therefore dependent on the concentrations of scattered particles in solution. The structure factor
S(q) is given by
(2.6.2.5)
Where r is the distance to the centre of a scattered particle and g(r) is the pair distribution function
of two scattering particles. In the work presented here, the concentration of sample solutions was
usually very small, so that S(q) could therefore be regarded as unity. Thus the scattered intensity
I(q) in equation 2.6.2.1 was only determined by the contribution of the variable P(q).
The fit process used the iterated method until an acceptable fit was produced. By comparing the
calculated profile from the ellipsoidal model with the measured scattering profile, the geometrical
shape and size of the protein molecule in bulk solution was obtained.
0
4( ) 1 [ ( ) 1] sin( )
pNS q g r r qr dr
q
57
The normal ellipsoidal model did not contain the volume fraction of the scattered particles in the
solvent. Thus, firstly, a Hayter-MSA structure factor was used, that was introduced to complement
the fit parameters required in this study. In general, the mean spherical approximation (MSA)24
method is able to describe well the charged colloidal particles that interact with a Coulombia
potential. It was found from MSA fits that the scattered particle in solvent was very weakly
charged but could be treated as no charge in the case of salt concentrations in the range of 5 mM
to 0.5 M. This meant that the MSA method was not appropriate in high salt systems. In the zero
charge limit, the Percus-Yevick hard sphere solution was recovered.25 This assumes the inter-
particle potential is given by
(2.6.2.6)
The Percus-Yevick method provides very accurate approximations for particle volume fractions
of up to ~0.45.26 The hard sphere structure method combined with the ellipsoidal model allows
for calculation of the inter-particle structure factor for mono-disperse spherical particles that
interact through hard sphere interactions. In this case, it is not appropriate for the form factor S(q)
to be approximated as unity and it is approximated as 27
𝑆(𝑞) = [1 +24η𝐻𝑆𝐺(2𝑅𝐻𝑆𝑞)
2𝑅𝐻𝑆𝑞]−1 (2.6.2.7)
The two variables in the S(q) equation are the hard sphere volume fraction HS and the
corresponding hard sphere radius RHS. 𝐺(2𝑅𝐻𝑆𝑞) is fully described by Stieger et al.27. In the data
analysis, the fit parameters included the background, the two radiuses of the ellipsoid, the scale
factor, the volume fraction and the scattering length densities of the scattered particles and the
solvents. The scale factor was fixed at 1. Therefore the remaining variables for the fit process
were the ellipsoidal radii a and b, the scattering length density and the volume fraction of the
scattered particles.
58
2.6.3 Instrumentation
The SANS experiments presented in this thesis were performed on the LOQ diffractometer at the
ISIS Neutron Facility, Rutherford Appleton Laboratory, Didcot, UK. The neutron beam had a
wavelength range of 2.2-10 Å with a chopper which cut the overlapped area between the two
pulses of neutrons. The detector had a dimension of 64 cm2 and was located 4.1 m from the source.
It had a qrange of 0.006-0.28 Å-1. Sample measurements were undertaken in 2.0 mm path length
quartz cells, with a 12 mm diameter neutron beam. Data were corrected by the standard
procedures,28 before subtraction of the comparative D2O backgrounds. The software ‘Sasview’
(provided by RAP) was used for data fit.
2.6.4 Experimental procedures
The 2.0 mm quartz cells were cleaned with piranha solution, rinsed with copious UHQ water and
dried with nitrogen prior to use. In the measurement of protein/surfactant complexes, all
surfactants were prepared in stock solutions for later use. A 1 ml solution of each sample was
prepared for each measurement, which then underwent ultrasonic vibration for 15 minutes. After
each use of the cell, it was cleaned by injecting a 5% decon90 solution, using the ultrasonic water
bath for 15 minutes.
2.7 Dynamic light scattering, (DLS)
Dynamic light scattering (DLS) is a technique which can be used to determine the size of particles
in solution by measuring the particles’ mobility and diffusion, and has been widely used as a
convenient method in the study of surfactants and biomaterials. Detailed principles of DLS and
its applications are decribed by Berne et al.29 In the general case, particles with Brownian motion
in the solution are measured using light with a specific wavelength, and DLS calculates the
59
intensity changes of the particles between 1ns and 1ms. In order to measure changes in solution
intensity, laser light is first directed from the instrument to the sample solution. Assuming the
intensity of the signal at time t is I(t), the intensity is I(t+τ) after a short interval τ. The changes
are automatically analysed by a correlator and quantified using the second order autocorrelation
function: 30
(2.7.1)
where q is the particle wave vector and τ is the delay time. The brackets in the above equation
denote the expected value. If the delay time is short, there is insufficient time for particles to have
moved a large distance from their initial position and thus the correlation is high. As the delay
time increases, the correlation tends to zero since it decreases. For monodisperse particles the
decay is simply exponential, as decribed by
(2.7.2)
where A and β are constants and Γ is the decay rate. The most important application of the function
is in determination of particle sizes. The relation between the decay rate (Г) and the diffusion
constant (D) is given by
(2.7.3)
where n0 is the refraction index of the buffer, λ0 is the laser light wavelength and θ is the scattering
angle. For spherical particles the relation between the particles radius (r) and the diffusion
constant D is given by the Stokes-Einstein equation,
(2.7.4)
(2)
2
( ) ( )( ; )
( )
I t I tg q
I t
(2) 2( ; )g q A e
2
0
2 2 2
016 sin ( )2
D
n
6
Bk TD
r
60
where kB is the Boltzmann’s constant, η is the dynamic viscosity and T in the temperature in
Kelvin (absolute temperature). This analysis is only accurate for spherical particles, meaning that
the measured hydrodynamic radius r is not equal to the actual radius, but an estimation.
All DLS measurements described in this thesis were carried out using the Malvern Nanosizer
instrument. The Nanosizer has a 633 nm laser with the capability of detecting within the range of
0.6 nm to 6 m. The incoming beam has a detection angle of 173°. An isotropic quartz cell with
a 1 cm path length was used to contain the samples. The software ‘Malvern Instruments
Dispersion Technology Software’ was used to analyse the data. The refractive index used in this
work is set to be 1.45. Water has a refractive index and viscosity of 1.33 and 0.8872 cPa,
respectively.
References
1. N. Pallas and Y. Harrison, Colloids and Surfaces, 1990, 43, 169-194. 2. M. Malmsten, Journal of Colloid and Interface Science, 1994, 166, 333-342. 3. Y. Tang, J. R. Lu, A. L. Lewis, T. A. Vick and P. W. Stratford, Macromolecules, 2002, 35, 3955-
3964. 4. Y. Tang, T. J. Su, J. Armstrong, J. R. Lu, A. L. Lewis, T. A. Vick, P. W. Stratford, R. K. Heenan and J.
Penfold, Macromolecules, 2003, 36, 8440-8448. 5. Ellipsometry tutorial, http://users.aber.ac.uk/tej/ellipso.html. 6. M. Westwood, T. R. Noel and R. Parker, Soft Matter, 2010, 6, 5502-5513. 7. S. X. Liu and J.-T. Kim, Journal of the Association for Laboratory Automation, 2009, 14, 213-220. 8. J. R. Lu, M. J. Swann, L. L. Peel and N. J. Freeman, Langmuir, 2004, 20, 1827-1832. 9. G. H. Cross, Y. Ren and N. J. Freeman, Journal of applied physics, 1999, 86, 6483-6488. 10. J. W. Shirley, American Journal of Physics, 1951, 19, 507-508. 11. J. Lu, T. Su, P. Thirtle, R. Thomas, A. Rennie and R. Cubitt, Journal of colloid and interface
science, 1998, 206, 212-223. 12. R. K. Thomas, Annu. Rev. Phys. Chem., 2004, 55, 391-426. 13. V. Gudkov, G. Opat and A. Klein, Journal of Physics: Condensed Matter, 1993, 5, 9013. 14. J. B. Hayter and H. Mook, Journal of Applied Crystallography, 1989, 22, 35-41. 15. M. Born and E. Wolf, Foundations of Optics, 1970. 16. http://www.isis.stfc.ac.uk/instruments/inter/inter2471.html, Inter. 17. http://www.isis.stfc.ac.uk/instruments/surf/surf2469.html, Surf-Liquids and interfaces
reflectometer. 18. A. Holt, http://www.isis.stfc.ac.uk/instruments/inter/publications/inter-science-case6738.pdf. 19. J. Penfold, R. Richardson, A. Zarbakhsh, J. Webster, D. Bucknall, A. Rennie, R. Jones, T. Cosgrove,
R. Thomas and J. Higgins, Journal of the Chemical Society, Faraday Transactions, 1997, 93, 3899-3917.
61
20. CD. Dewhurst, I. Laue-Langevin, 2008, Volume 19, Number 3. 21. R. A. Pethrick and J. Dawkins, Modern Techniques for Polymer Characterisation, 1999. 22. L. Feigin, D. I. Svergun and G. W. Taylor, Structure analysis by small-angle X-ray and neutron
scattering, Springer, 1987. 23. L. Feigin and D. Svergun, Plenum Press, New York. 24. J. B. Hayter and J. Penfold, Molecular Physics, 1981, 42, 109-118. 25. J. K. Percus and G. J. Yevick, Physical Review, 1958, 110, 1. 26. D. Frenkel, R. Vos, C. De Kruif and A. Vrij, The Journal of chemical physics, 1986, 84, 4625-4630. 27. M. Stieger, J. S. Pedersen, P. Lindner and W. Richtering, Langmuir, 2004, 20, 7283-7292. 28. R. Heenan, J. Penfold and S. King, Journal of applied crystallography, 1997, 30, 1140-1147. 29. B. J. Berne and R. Pecora, Dynamic light scattering: with applications to chemistry, biology, and
physics, Courier Dover Publications, 2000. 30. B. G. Vertesse, S. Magazu, A. Mangione, F. Migliardo and A. Brandt, Macromolecular bioscience,
2003, 3, 477-481.
62
Chapter 3
Adsorption of C12E6 at the SiO2/water interface: a combined study by
DPI, SE, QCM-D and NR
Spectroscopic ellipsometry (SE), dual Polarisation interferometry (DPI), quartz crystal
microbalance with dissipation (QCM-D) and neutron reflection (NR) have all been used to study
interfacial adsorption but their complementarity has not yet been well explored. In the work
described here, this is explored by studying the widely known non-ionic surfactant hexaethylene
glycol monododecyl ether (C12E6) at the silica oxide/aqueous solution interface, over a wide range
of concentrations of 0.5 to 160 CMC (critical micelle concentration, 1 CMC= ~8.7x10-5 M). This
work highlights the unique information that each technique is able to provide. SE, DPI and QCM-
D detected dynamic interfacial adsorption, whilst NR revealed changes in the equilibrium layer
structure and volume fraction of the interfacially adsorbed C12E6 layers as a function of
concentration. Specifically, NR, with the addition of deuterium labelling, determined the structure
of the surfactant layers and the water volume fraction across the interface. The adsorbed surfactant
layers were characterised by formation of a bilayer of overall thickness of approximately 40 ± 4
Å, with an inner head group layer of 13 ± 1 Å, middle chain layer of 16 ± 1 Å and outer head layer
of 13 ± 1 Å, which is consistent with an interdigitated or tilted bilayer chain structure.1 The volume
fraction of the head and chain segments in each layer increased with bulk concentration, but that
of the outer head layer tended to be lower than that of the inner layer, and exhibited slight
asymmetrical packing. SE and DPI provided consistent adsorbed quantities, but DPI offered more
effective real time monitoring, with the additional structural information of layer thickness and
density, more consistent with a micellular structure in the form of the tilted bilayer. QCM-D
measured the real time adsorbed hydrated mass that includes the trapped water incorporated into
the bilayer. Changes in the QCM-D’S dissipation indicated a gradual transition from a more rigid
63
layer to a less rigid and more hydrated adsorbed layer as the concentration increased, over the
range of 5 CMC to 10 CMC. Furthermore, the inverted order of dissipation as a function of
harmonic frequency compared to that typically observed, could be fitted to a low viscous inner
head layer of 13 Å with a higher viscosity secondary layer that combines the middle chain and
outer head groups. This further supports the proposition that the structure involves highly
hydrated PEG headgroups closest to the surface. DPI clearly showed the dynamic formation,
build-up and subsequent dissolution of the bilayer with various structural changes, including
thinning of the bilayer followed by breakdown of the hydrated fluid interfacial layer, which is also
corroborated by QCM-D.
3.1 Literature review
The properties of surfactants at a solid/water interface are an important area of interest in surface
science with wide ranging application in a range of industrial applications including foaming or
coating processes. When the concentration of a surfactant in the bulk water reaches a particular
value, its adsorption process is accompanied by a self-assembly behaviour in which the surfactants
form aggregates in the bulk solution with their hydrophobic tails packed in the core of the
aggregate and their hydrophilic parts packed outside, facing the aqueous solution.
A number of scientific techniques have been developed over recent years to investigate adsorbed
layers at interfaces, and to determine the amount of adsorption and investigate the dynamic
process. In the work presented in the thesis, dual polarization interference (DPI), quartz crystal
microbalance with dissipation (QCM-D), ellipsometry (SE) and neutron reflection (NR) were the
main techniques implemented, and were used in the investigation of the adsorption of C12E6 at a
solid/water interface, utilising hydrophilic silica as the solid substrate.
64
Since the end of the last century, neutron reflection has been developed to measure interfacial
adsorption both in terms of the amount of material adsorbed, as well as its structural distribution2.
The adsorption behaviour of surfactants, proteins and their mixtures at air/liquid and solid/liquid
interfaces has been extensively studied. A typical NR measurement takes around 1 hr to complete,
and therefore cannot readily offer real time information. In contrast, SE probes changes in the
polarization of the light reflected from sample layers. In order to determine the adsorbed mass,
the thickness and refractive index of the adsorbed layer must both be fitted simultaneously. The
thickness is ordinarily not reliable, and is usually achieved by an assumed refractive index for the
layer.3
QCM-D is an acoustic surface sensitive technique. It monitors changes in the resonant frequency
and dissipation of quartz crystals at a number of overtones which can then be used to calculate the
hydrated, acoustically coupled mass of an adsorbed layer and viscoelastic or structural properties
with the use of the Voigt model. The Voigt model is used to calculate the hydrated mass of formed
layers that exhibit viscoelastic behaviour, which can be determined by changes in dissipation.4, 5
DPI is an optical waveguide technique which uses the transverse electric and transverse magnetic
modes excited in a waveguide by a He-Ne laser. The waveguide modes produce a decaying
evanescent field at the waveguide surface, which is used to probe real time signals of the effective
refractive index generated by the adsorption of a film.6 Compared to SE, which can only derive a
mass per unit area, DPI allows calculation of both thickness and RI (proportional to density) from
the raw data. Acquisition of a mass is then achievable by combining the thickness and RI with the
use of the de Fejiter formula and refractive index increment value (dn/dc) of the material. A
uniform single slab layer model is ordinarily used for the analysis of adsorbed films as it is the
most simple. However, it assumes an average layer structure, which may be unrealistic if, for
example, the layer has a variable density profile perpendicular to the surface.7
65
More than 20 years ago, McDermott et al. 8 were among the first to report C12E6 self-assembled
structures on silicon oxide surfaces by neutron reflection. They revealed that the adsorbed
surfactant formed a bilayer with a thickness of 49 ± 4 Å above the CMC. In the work presented
in this thesis, the refractive indices of C12E6 were measured in different concentrations by DPI,
which were used in the Cauchy model with a range of values from 1.39 to 1.42. Tiberg et al. 9, 10
demonstrated the dynamic adsorption of C12E6 and the effect of ethoxylate chains on adsorption
for C12En. The results revealed that the adsorption of C12E6 happens in the first few minutes, which
is in agreement with the findings in the work presented here. Penfold et al. 11 demonstrated that
non-ionic surfactant adsorption at a solid/liquid interface is sensitive to different surface treatment
and variation in the solution pH. The adsorption of C12E6 on SiO2 (pre-treated with Piranha)
exhibited large differences in measurements from low pH to high pH and high pH to low pH.
The aim of the work presented here was to study C12E6 adsorption on silicon oxide by combining
four different techniques, in order to provide a more precise view of the layer information. When
comparing measurements across different techniques, care is required in order to ensure that the
conditions are matched as well as possible, in order to avoid the potential issue of different
information provided by different techniques appearing to be inconsistent. As C12E6 is a non-ionic
surfactant, its adsorption on silicon oxide with weakly negative charges is very sensitive to pH
values. Therefore all the solutions of C12E6 were maintained between pH 6.8 and pH 7.2.
Differences in experimental conditions such as surface substrates can mean that it is not feasible
or straightforward to compare data from separate experimental setups. Therefore, in this work, all
the silicon substrates used in SE, QCM-D, NR and DPI were treated and cleaned in the same
manner. In this way, the different techniques were applied under conditions as closely matched as
possible, and data was measured on the same or similar model surface where possible. The SE
measurement was obtained in combination with a QCM-D cell at an incident angle of 65o, thus
allowing the two techniques to be performed simultaneously on the same surface.
66
3.2 Experimental
3.2.1 Sample preparation
Hexaethylene glycol monododecyl ether with full hydrogenation (hC12hE6) was obtained from
Sigma Aldrich (>99.9%). The chain deuterated dC12hE6 was provided by Dr Peixun Li at ISIS,
Rutherford Laboratory, UK. In order to prevent this surfactant becoming wet, both the hC12hE6
and the dC12hE6 were freeze-dried (using a freeze dryer made by ESCO) prior to use. The Decon90
was provided by Decon Laboratories Limited, U.K. The water used was UHQ grade obtained
from a MilliQ system. The D2O was purchased from Sigma Aldrich (containing 99.9% D).
It was ensured that the samples were adequately sealed. A stock solution of the surfactant was
firstly prepared at 100 times the CMC by dissolving the surfactant into 2ml of D2O. The C12E6
solutions at different concentrations were prepared precisely, by dilution.
3.2.2 Experimental method
NR measurements were performed on the SURF instrument at the ISIS Neutron Facility,
Rutherford Appleton Laboratory (RAL), which has a q-range of 0.01 to 0.5 Å-1.12 The standard
procedures for data treatment have been described previously in Section 2.5.1. The cells and
silicon blocks used in the measurements were as described in Section 2.5.2. The temperature for
all experiments was set at 20 °C to provide for consistency. The background of D2O was
subtracted from the data after completion of the measurements.
DPI measurements were performed using the AnaLight Bio200 dual polarization interferometer
(Farfield Group Ltd, Manchester, U.K.) with a helium-neon laser at a wavelength of 632.8nm.
Operating principles and theoretical details of DPI have been introduced previously in Section 2.4
and are also described by Coffey et. al.13 and Lu et. al14 in the case of protein adsorption. In brief,
the technique allows for measurement of changes in the propagation of light in a mono-mode
67
waveguide in both Transverse Electric (TE) and Transverse Magnetic (TM) polarisation modes
upon adsorption of a molecular layer onto the exposed waveguide surface. The adsorbed layer can
then be fitted by a uniform layer including information regarding both the layer thickness and
layer refractive index. From the values of thickness and RI, the mass density and adsorbed amount
can be calculated according to the de Feijter equation15 in the same way as previously described
for SE in Section 2.2. A PHD2000 pump was used to provide a slow and steady injection of
running buffer and sample. Prior to measurements, the DPI chip waveguide structure was
calibrated using 80% ethanol (w/w) in UHQ water and the UHQ water, and followed by a buffer
rinse. The chip was cleaned by three injections of 2% diluted Decon90 solution (w/w), 80%
ethanol (w/w) and rinsed with buffer after each measurement.
The QCM-D experiment was performed using the Q-sense E1 and E4 equipment from Biolin
Scientific AB and carried out in combination with spectroscopic ellipsometry (SE). Silica coated
QCM sensors (QSX335) were used as a substrate during the adsorption measurements. A simple
Sauerbrey model is ordinarily used for non-dissipation and rigid films. In the work described here,
the Kelvin-Voigt model was applied to the viscoelastic layers of the surfactant. ∆f and ∆D data
were measured at several harmonics (n= 3, 5, 7, 9, 11, 13) in order to calculate the viscoelastic
values of the adsorbed layer. The layer density was set to 970 kg/m3 as measured by DPI. The
adsorbed C12E6 layer was fitted using the three factors of shear elasticity ( ), viscosity ( ) and
thickness ( ).16
The SE used in this work was model M-2000 made by J.A. Woollam and had a wavelength range
of 370 to 1000 nm. The measurements were performed in a QCM-D cell in order to combine them
with QCM-D measurements. Two windows were installed on both sides of the cell to allow
transmission of incident and exiting light and fixed at an angle of 65o. The sensor was specifically
designed for suitability with both SE and QCM-D measurements. It consists of a 50 nm silicon
1 1
1
68
dioxide layer with a 100 nm Ti adhesion layer below. A four layer model was used: B-Spline
model,17 for the substrate; TiO2 for the second layer; SiO2 for the third layer; and the Cauchy
model for the fourth layer. The amount of protein adsorbed was then obtained using the equation
proposed by de Feijter et al.15
3.3 Results and discussion
3.3.1 NR measurements
Figure 3.1 shows the NR profiles of a silicon oxide/D2O interface at a C12E6 concentration of 20
times CMC in UHQ water at pH 7. The dashed red line is the best fit under the assumption of
formation of a monolayer for C12E6 adsorption. The solid blue line is the best fit under the
assumption of the formation of a sandwiched bilayer. In this case, the hydrophobic parts exist in
the middle layer and hydrophilic head groups orient towards the SiO2 surface and bulk solution
phase. Both assumptions appear to fit the data well, within the Q range measured. However, the
sandwiched bilayer is the most widely accepted structural model for the non-ionic surfactant layer
structure formed.
69
Figure 3.1 NR profiles of C12E6 adsorption at 10 CMC in D2O at pH 7. The dashed red line is the best fit for a
single slab monolayer for dC12hE6. The solid blue line is the best 3 layer fit, assuming a sandwiched bilayer type
structure with the hydrophobic parts in the middle layer and hydrophilic parts of the head groups oriented towards
the SiO2 surface and the bulk solution phase.
The structures of C12E6 layers can be revealed more precisely by the analysis of several contrasting
measurements, for example, by inclusion of the chain deuterated dC12hE6 in H2O and D2O
together with hC12hE6 in D2O, which shows different contrasting variations between the head and
chain regions (Figure 3.2). Under the application of a monolayer assumption, a thickness of 45 ±
3 Å was fitted for the measured profile for a volume fraction of 0.39. This thickness corresponds
to micelles of C12E6 formed on the silica surface with gaps, or a sandwiched bilayer. A three layer
model better fits the measured data than the monolayer model with different volume fractions and
SLDs for each layer. The SLDs of the inner layer close to the silica surface and the outer layer
towards the bulk buffer are low compared to the middle layer, which indicates that a sandwiched
bilayer type structure has been formed. These results are highly consistent with previously
published results.8, 10
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.06 0.11 0.16
Ref
lect
ivit
y
Momentum Transfer /Å-1
70
Figure 3.2 Bilayer model fit to reflectivity data as a function of momentum transfer for the C12E6 layers adsorbed at
the SiO2/water interface: diamonds represent the dC12hE6 in D2O; triangles represent hC12hE6 in D2O; circles
represent dC12hE6 in H2O.
The molecular volume fraction ( ) of C12E6 in the layer is calculated using equation 2.5.4.1 in
Chapter 2. The scattering length densities can be calculated with -0.42 × 10-6 Å-2 for protonated
C12 tails, 6.5 × 10-6 Å-2 for the deuterated C12 tails, and 0.71 × 10-6 Å-2 for EO6 head groups.
Specifically, 0.19 × 10-6 Å-2 was found for the protonated hC12hE6 and 3.64 × 10-6 Å-2 for the
deuterated dC12hE6, respectively. The area per molecule (A) was obtained by the application of
equation 2.5.4.2 where the volume of C12E6 ( ) is 710 Å3, with 330 Å3 for the chain group and
380 Å3 for the head group.
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum Transfer/Å-1
s
sV
71
Table 3.1. Structural parameters of C12E6 layers adsorbed at the silicon/water interface from NR at a concentration
of 0.82mM (10 CMC) using a sandwiched bilayer model, with layer 1 closest to the oxide.
layer τ (Å) ±1 ρs (×10-6Å2) ±0.1 φs ±0.01 Γ (mg/m2) ±0.1
dC12hE6 in D2O 1 13 4.3 0.37 1.73
2 16 5.3 0.43
3 12 4.8 0.27
hC12hE6 in D2O 1 13 4.0 0.41 1.8
2 16 3.3 0.43
3 15 4.5 0.33
dC12hE6 in H2O 1 14 0.3 0.42 1.68
2 14 1.8 0.4
3 15 0.5 0.36
Table 3.1 shows the structural parameters of the C12E6 layers adsorbed at the silicon/water
interface at a concentration of 0.82mM (10 CMC) in three different contrast conditions. It shows
that the chain deuterated surfactant in D2O has lower volume fractions for the first and the third
layers than the protonated C12E6. In addition, as the SLDs of the deuterated chain and D2O are 6.5
× 10-6 Å-2 and 6.35 × 10-6 Å-2 respectively, the middle layer of the deuterated surfactant is expected
to be between 6.5 × 10-6 Å-2 and 6.35 × 10-6 Å-2. However, the fitted SLD of the middle layer is
5.3× 10-6 Å-2, which is lower than predicted. The only viable reason for this result is that there are
protonated head groups of C12E6 in the middle layer, which decrease its total SLD. In addition,
there are also chain tails, which are present in the first layer and the third layer. In this case, the
adsorbed surfactant is assumed to form micelles instead of an orderly bilayer over the whole silica
72
surface. From the SLD of the middle layer, a volume fraction of (EO)6 groups in the middle layer
can be calculated (0.12) given that the SLDs of the deuterated chains and D2O are almost
equivalent.
It is also possible to fit a single distribution of head (E6) and tail (C12) groups across the three
contrast conditions as shown in Table 3.1. In this case, the number of head groups and tail groups
was fixed as equal (the difference was controlled within 0.1 in volume fraction) and the total error
between the calculated SLDs of the modelled structure and the data was minimised. The
calculated volume fractions of head, tail and water in each of the three layers are shown in Table
3.2a, whilst the calculated SLDs for each layer and the errors are shown in Table 3.2b. In this case,
the same structure exhibited a good fit for two of the contrasts, whilst one exhibited greater errors.
This is the deuterated surfactant in D2O, which is also the condition where the contrast difference
between the layers and the solvent was lowest and the error in the layer fit likely to be greater also.
Table 3.2a Fitted volume fractions of head and tail groups from NR across the three layers of the ‘Bilayer model’
fitted to the C12E6 adsorbed at the silica/water interface.
Layers Head
Volume Fraction
±0.02
Tail
Volume Fraction
±0.02
Water
Volume Fraction
±0.02
1 0.309 0.090 0.601
2 0.134 0.341 0.524
3 0.205 0.124 0.671
73
Table 3.2b Calculated scattering length densities and measured SLDs from NR for the single surfactant structure
model shown in Table 3.2a.
layer τ (Å) ±1 ρs (×10-6Å2) ±0.1 ρs calc (×10-6Å2) ±0.1
dC12hE6 in D2O 1 13 4.3 4.620
2 16 5.3 5.645
3 12 4.8 5.215
hC12hE6 in D2O 1 13 4.04 3.999
2 16 3.3 3.282
3 15 4.5 4.356
dC12hE6 in H2O 1 14 0.3 0.466
2 14 1.8 2.021
3 15 0.5 0.576
74
Figure 3.3 NR profiles as a function of momentum transfer for the silicon oxide/water interface with surfactant
concentrations of 3.3 mM (4 CMC) (blue diamonds), 8.2 mM (10 CMC) (red triangles) and 33 mM (40 CMC)
(black circles).
Figure 3.3 shows the NR profiles at the silicon/water interface with surfactant concentrations at
3.3 mM (4 CMC), 8.2 mM (10 CMC) and 33 mM (40 CMC). Table 3.3 shows the fitted structural
parameters of C12E6 layers adsorbed at different concentrations at this interface. The results
revealed that the interfacially adsorbed amount increased slightly from 1.75 to 1.9 mg/m2 as the
concentration increased from 10 to 80 CMC. The total thickness of the three layers also increased
slightly from 41 ± 1 Å to 44 ± 1 Å. The greatest difference was seen in the external head group
layer, where at 80 CMC, the asymmetry in volume fraction between layer 3 and layer 1 was almost
eliminated.
1E-8
1E-6
1E-4
1E-2
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum Transfer/Å-1
75
Table 3.3 Structural parameters of the C12E6 layers adsorbed at the silicon/water interface at different
concentrations from NR experiments.
C12E6 mM Layer τ(Å)
±1
ρs (×10-6Å-2)
±0.1
φs
±0.01
Γ (mg/m2)
±0.1
0.8 in D2O 1 13 4.3 0.37 1.75
2 16 3.3 0.43
3 12 4.8 0.27
3.3 in D2O 1 13 4.04 0.41 1.85
2 16 3.3 0.43
3 15 4.5 0.33
6.6 in D2O 1 14 4.1 0.42 1.9
2 14 3.4 0.42
3 15 4.3 0.39
Figure 3.4 shows the NR profiles at the silicon/water interface with the dC12hE6 concentration at
3.3 mM (4 CMC) from two experimental methods. In the first method (blue line), the cell buffer
was replaced by a single volume exchange of 2 ml surfactant solution, whilst in the second
exchange, a tenfold volume of the sample was used in rinsing and replacement (red line, 20 ml).
The surface adsorbed amount in the case of the blue line is 0.4 mg/m2, whilst that in the case of
the red line (under excess volume exchange) is 1.8 mg/m2. This difference illustrates that there is
far more adsorption in the case of the 20mL solution injection than that with 2mL injection. Fit of
the two sets of data shows over 4 times differences in the surface adsorbed amounts (1.8 mg/m2
and 0.4 mg/m2 respectively). This demonstrates the importance of allowing a sufficient quantity
76
of the sample to fully exchange with the solution, and these results are likely to be predominantly
due to mixing and dilution of the sample on solution exchange. This result may also be caused by
adsorption in the plastic tubing and cell walls from the diluted solutions, resulting in insufficient
surfactant in the 2 mL solution for adsorption equilibrium to be achieved. This result is slightly
less than the value reported by Mcdermott et al. 8, with 2.04 mg/m2 at the CMC of C12E6 at the
same silicon oxide/water interface. It was however far greater than that reported by Penfold et al.
18 of the adsorbed amount of C12E6 at 4 CMC concentration of 0.5~1.5 mg/m2. Such differences
could well arise from differences in levels of hydration resulting in differences in apparent CMC
values.
Figure 3.4 NR profiles at the silicon oxide/water interface with surfactant concentrations at 0.33 mM (4 CMC) at
pH 7. Blue squares represent measurements from solution injection of 2 ml into the NR cell without overflow. Red
circles represent measurements from solution injection into the NR cell of 20 ml flow for 10 minutes. The capacity
of the NR cell was 1.3-1.5 ml.
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer/Å-1
77
3.3.2 DPI measurements
DPI is an analytical optical technique that is used for simultaneously probing the thickness,
density and mass of materials during interfacial adsorption.19 It was utilised in the work presented
here to reveal the time dependent evolution of the mass and thickness of the adsorbed C12E6 layer.
Figure 3.5 DPI transverse electric polarisation phase changes as a function of time for adsorbed C12E6 layers on
SiO2 substrate from 0.5 CMC to 160 CMC at solution pH 7. The inflections at A and B upon rinsing were
associated with dilution of the bulk surfactant concentration.
Figure 3.5 shows the transverse electric polarization changes after injection of different
concentrations of C12E6 solutions. The magnitude of the phase change is small at 0.5 CMC and
below, increases slowly at 0.8 CMC and more quickly up to 2 CMC. Above 2 CMC, there is a
point of inflection and another increase of the TE signal. The first rapid increase takes place on
the timescale of the bulk solution exchange and exhibits concentration dependence at higher
concentrations due to the accompanying bulk refractive index contribution from the C12E6 solution.
After rinsing off at the end of a sample injection, a reversal of the adsorption processes is observed.
-0.5
0.5
1.5
2.5
3.5
4.5
5.5
0 200 400 600
TE
Pola
riza
tion
Time (Sec)
160CMC
80CMC
40CMC
10CMC
2CMC
1CMC
0.8CMC
0.5CMC
A B
78
There is a fast drop in the signal due to the bulk refractive index change as the solution is
exchanged (A to B in Figure 3.5), followed by a sloping shoulder, which is of similar size for all
the higher concentration injections, prior to a more rapid drop as the bulk of the material is
removed from the surface.
Detailed analysis requires a value for the dn/dC of C12E6. This is determined from the data
presented here by applying the assumption that the response at higher concentrations corresponds
to the bulk RI contribution. Figure 3.6 shows the maximum response from a series of C12E6
injections from 0.5 CMC to 320 CMC analysed as a bulk refractive index change. The initial
sharper increase corresponds to the effect of surface physio-sorption and the increase at higher
concentrations corresponds to the bulk refractive index change once the surface has become
saturated.
Figure 3.6 Refractive index (RI) as a function of concentration of the C12E6 solution (unit of CMC) measured by
DPI. The slope over the higher concentration range provides the value of dn/dC for the surfactant.
Figure 3.6 shows the refractive index as a function of C12E6 concentration. The refractive index
increment (dn/dC) can be calculated from the slope over the high C12E6 concentration range. This
79
gave a value of 0.1215 mL/g of dn/dC, which was used to determine the adsorbed amount for both
DPI and SE techniques.
Figure 3.7 Mass density of the adsorbed layer as a function of time for a range of bulk concentrations of C12E6
measured by DPI. The different coloured lines represent different bulk concentrations corresponding to those in
Figure 3.5.
Figure 3.7 shows the mass density of the adsorbed layer as a function of time for a range of bulk
concentrations of C12E6. It can be seen that there is a relative constancy of the mass density as a
function of concentration. This illustrates that the surface is essentially uniformly covered with
surfactant over the range of concentrations. It is, however, the structure of that layer that changes
significantly.
Figure 3.8 shows the surface adsorbed amount and thickness for selected concentrations of C12E6,
calculated from the data in Figure 3.6. The bulk refractive index has been accounted for at 20
CMC and above, and applied from the start of the injection until after the initial fast drop following
rinse-off. The data during this transition is not shown. The rinse-off times have also been adjusted
80
to overlay the different traces. There are breaks in the data for very low coverage (thickness below
1 Å), where no layer is fitted with the slab model.
Figure 3.8 Surface adsorbed amount (left graph) and thickness (right graph) as a function of time for a range of
bulk concentrations of C12E6 from 0.5 CMC to 160 CMC measured by DPI. The different coloured lines represent
different bulk concentrations corresponding to those described in Figure 3.5.
It can be seen that at low CMC (0.8 CMC and below) there is little adsorption with an average
layer thickness of less than 10Å. At higher CMC the layer build up is slow, but reaches a layer
thickness of 35 Å. At 2 CMC, a layer of 34 Å thick is rapidly formed, followed by a slower further
increase in both mass and thickness before a stable plateau is reached at around 36 Å. At higher
concentrations, these values become a little higher, increasing from 38 Å to 40 Å at 10 CMC for
example, and reaching a maximum plateau of 43.6 Å at 320 CMC. These values are very close to,
although slightly lower than, the accumulated values obtained by NR in Table 3.3. As the DPI
layer is a single homogeneous layer fit and the mass density is not homogeneous throughout the
whole film, this slightly lower average value of the mass density is to be expected.
On rinse-off, after the initial fast bulk solution exchange the initial adsorbed amount and thickness
values remain unchanged, which indicates a persistent surface layer. There is then a slow decay
81
in both thickness and surface adsorbed amount. Although this onsets at the highest value for 320
CMC, all the concentrations above the CMC lie at different levels overlaid on this same rinse-off
curve. This common rinse-off is the same for regardless of the surface adsorbed amount and the
mass density. This suggests that the surface structure as it undergoes rinse-off at 320 CMC evolves
to be equivalent to that at 160 CMC (and other values) until the structure at CMC is reached.
Beyond this, the thickness reduces more rapidly to a very low level, at which point only a small
amount of residual material is being rinsed from the surface.
The detailed relationship between the three calculated parameters is illustrated in Figure 3.9 for
10 CMC, at which concentration no bulk RI effect is observed. It can be seen that during initial
adsorption the mass density is lower than the final equilibrium value, which may reflect the
adsorption of discrete micelles onto the surface, although the very low and increasing thickness
does suggest that the initially adsorbed C12E6 monomer or micelles was confined to the surface.
After the initial adsorption the slower increase to the plateau is accompanied by a small further
increase in mass density suggesting a rearrangement of the solution based micelles into more
stable surface bound structures.
On rinse-off, this process is not reversed. The density remains almost constant or increases slightly
during the first slow dissolution phase. After this phase is complete and the faster dissolution step
begins, the density increases to a peak of 0.53 g·cm-3 (at a thickness of 21 Å) prior to a reduction
to approximately 0.4 g·cm-3 (and a thickness is well below 5 Å).
82
Figure 3.9 Surface adsorbed amount, thickness and mass density as a function of time for the 10 CMC sample. The
first vertical line indicates the start of injection, the second line indicates the start of rinsing, the third and fourth
lines indicate the removal of the adsorbed layer, and the fifth line indicates the completion of the rinse.
The data allows for potential structures for the adsorbed C12E6 at equilibrium on the surface to be
considered. NR determined the adsorbed layer to be formed as a sandwiched bilayer with C12E6
tails interdigitated in the middle at a length of 15±1 Å. Given that the extended length of a C12E6
tail is of the order of 17 Å, the middle layer could be formed either as a tail to tail structure with
a variety of angles on the substrate or by pairs of parallel interdigitated molecules. It should be
emphasised that these structures are stylised representations using linear chains, whereas the tilted
chains may in fact be of zig zag formation, coiled or otherwise disordered and still have the same
average tilt angle.
83
By taking the concentration dependent NR and DPI data together, it can be suggested that the
slow decay may reflect a loss of C12E6 predominantly from the side away from the surface, with
the resultant layer becoming thinner and more tilted (33 Å). As the mass reduces further, the
density increases to a peak (at a thickness of 21 Å) which may be due to a transition to a thinner
transient monolayer structure. With a further reduction in thickness and a small reduction in
density it can be suggested that the layer becomes broken up with residual monomer covering the
surface.
3.3.3 QCM-D measurements
Figure 3.10 shows the time dependent frequency change and dissipation for C12E6 adsorption on
SiO2 at a concentration of 10 CMC. It can be seen that the adsorption of C12E6 very quickly reaches
equilibrium. The dissipation altered by about 0.5×10-6 at harmonic 5 after the injection of C12E6
solution, implying that the adsorbed layer is not a strictly rigid layer. However, if the Sauerbrey
relation is correct and applicable, the value of frequency change observed at each harmonic should
be a constant value. In the data presented here, the frequencies for different harmonics changed
by the order of 10%. Camille et al. 20 reported that different harmonics at one concentration had
the same frequencies. One explanation would be that this may be due to a lower sensitivity of the
QCM-D in detecting small changes. In addition to this 10% difference, there was exhibited a clear
reversal in the sequence of the dissipation for the harmonics, which indicates that there may be an
occurrence of a rather interesting phenomena. The results are different to those to be expected for
a typical film. This is not due to an artefact, since it was consistently reproducible with different
instruments. It is therefore appropriate to use the Voigt model rather than the Sauerbrey Model
even though the D/F ratio is low (0.035). The density of the layer used in the Voigt model is
calculated from the DPI measurements. DPI determined that the density of C12E6 adsorbed onto
the substrate was 0.43±0.02 g/cm3, excluding water inside the layer. Therefore a layer density of
84
0.917 g/cm3 was obtained given the pure C12E6 density of 0.9 g/cm3.
Figure 3.10 Dissipation (left) and frequency (right) of Harmonic 3, 5 and 7 for the C12E6 adsorption at a
concentration of 10 CMC (0.82 mM) at pH 7 measured using QCM-D.
It can be seen from Figure 3.10, that the magnitude of dissipation at the 7th harmonic was the
highest, which is not a common sequence of harmonics. The results cannot be fitted using a single
adsorbed layer with the Voigt model, which would give the magnitude of dissipation in the reverse
order with the nth harmonic. Attempts to do this were found to yield unrealistically thick layers.
This indicates a more complex shape for the adsorption layer. In order to fit the disordered
magnitudes of dissipation, the adsorbed layer was split into two layers with an inner layer at the
silica surface and an outer layer towards bulk solution.
From DPI and NR measurements, the thickness of the adsorbed layer at 10 CMC was found to be
between 42 Å and 50 Å. This provides a guidance to obtain the best fit for the Voigt model of
QCM-D. An inner layer of thickness 10 Å, a viscosity of 0.0016 kg/ms, and a shear rate of 300
Pa and an outer layer with thickness of 38 Å, a viscosity of 0.008 kg/ms and a shear rate of 3×106
Pa were obtained from the fit. This is well explained by a structure of adsorption comprising head
85
groups of C12E6 projected towards the hydrophilic silica surface, forming a thin layer of low
viscosity and low shear with a thickness around 10 Å. This may be further accentuated since the
C12 inner layer formed by back-to-back intermixing is a dense fluid monolayer. Any shear at the
bottom of it would be transferred to the top, whereby the molecules protruding from the bottom
would drag the molecules protruding from the top along on the bulk water side. This would
prevent any slipping in the middle. However, the (EO)6 layer next to the surface is likely to be
much more flexible. As measured by NR, the (EO)6 chains had a length of around 14 Å, which is
not fully extended. The (EO)6 chains are also highly hydrated and not coupled directly to the silica
surface by any strong force such as charge interaction. Therefore, the whole interfacially bound
micelle may be able to flow over the chip surface, but the C12 layer would not be able to have a
high value of shear. Another point of note is that the order of dissipation appears to reverse after
rinse-off. This is also consistent with an increase in rigidity and a change of structure without the
more fluid interfacial layer.
86
3.3.4 SE measurements
Figure 3.11 Adsorbed amount of C12E6 as a function of time for three bulk concentrations measured by
ellipsometry (CMC given by the blue line, 4 CMC by the red line, 10 CMC by the purple line and 20 CMC by the
green line).
In the work presented here, spectroscopic ellipsometry (SE) was used to study the dynamic
adsorption at several fixed concentrations of C12E6. As the ellipsometry used the same QCM-D
cell with a flowing buffer, the reflected light signals from the buffer are not as stable as the parallel
-0.5
0
0.5
1
1.5
2
0 200 400 600 800 1000 1200 1400
Ad
sorb
ed a
mou
nts
(m
g/m
2)
Time (Sec)
-0.5
0
0.5
1
1.5
2
0 50 100 150 200 250 300
Ad
sorb
ed a
mou
nts
(m
g/m
2)
Time (Sec)
Rinse
Rinse
87
QCM-D measurement. The refractive index for SE ranges from 1.371 to 1.38 with different
concentrations of C12E6. The refractive index increment (dn/dC) used in the de Feijter formula is
0.1215.
Figure 3.11 illustrates that interfacially adsorbed amounts reached their equilibrated values faster
with higher bulk concentrations (the time to reach equilibrium decreased with increased C12E6
concentration). The data shows that at the CMC adsorption reached equilibrium after a much
greater time. This observation is in good accordance with DPI and QCM-D data.
3.4 Comparison of the four techniques
-
Figure 3.12 Changes of the interfacially adsorbed amount as a function of C12E6 concentration measured by DPI (),
QCM-D (□), SE (Δ) and NR (Ο).
Figure 3.12 shows change in the interfacially adsorbed amount as a function of C12E6
concentration as measured by DPI, QCM-D, SE and NR. The data shows that adsorbed amounts
above 4 CMC concentrations measured by SE and NR are consistent. The adsorbed amounts
above 1 CMC concentrations (1.7 ± 0.15 mg/m2) from SE are slightly lower than the results
0
1
2
3
4
5
0.4 2 10 50
Mass
(m
g/m
2)
Concentration (CMC)
88
obtained by Tiberg et al. 9, 10, 21 (1.9 ± 0.1 mg/m2). This difference may be due to different pH
values, with pH 7 being used in this work whilst in Tiberg’s measurements the pH was below 7
with C12E6 solutions in water. The trend in the data is in agreement with the observation by Lee
et al. 22, who reported that the adsorbed amount of C12E6 increased at low pH, but that the thickness
remained the same. The adsorbed amount at a 4 CMC concentration (1.9 mg/m2) is also consistent
with the result obtained by Mcdermott et al. 8 at 1 CMC concentration for 2.04 mg/m2, indicating
that the adsorption reached saturation at and above the CMC. At the 4 CMC of C12E6, the surface
adsorbed amounts from DPI were found to be slightly greater than in the case of NR and SE data
above the CMC. This may be due to the slightly higher surface roughness of the DPI sensor
surface. The interfacially adsorbed amounts obtained by QCM were found to be 2.5 times greater
than those obtained by the other three techniques. This clearly demonstrates that water contained
within the layer is also accounted for in the QCM masses. The ratio between the QCM masses
and those obtained from the other techniques is consistent with that reported by Lubica et al.23
derived from QCM and SE data.
Figure 3.13 Changes in adsorbed layer thickness as a function of C12E6 concentration measured by DPI (), QCM-
D (□), SE (Δ) and NR (Ο).
0
10
20
30
40
50
60
0.4 4 40
Th
ick
nes
s (Å
)
Concentration (CMC)
89
Figure 3.13 shows the changes in adsorbed layer thickness as a function of C12E6 concentration
as measured by DPI, QCM-D, SE and NR. The refractive index of the adsorbed layer used for SE
was 1.39 ± 0.05, which was obtained from the DPI measurements. The DPI is able to provide
accurate measurement of the refractive indexes for different surfactant concentrations and
therefore contributes to obtaining more accurate thicknesses for the SE measurements. The
adsorbed layer density used in QCM measurements was assumed to be 0.95g/mL. The thicknesses
obtained from the four techniques are all consistent, indicating that the adsorbed layer thickness
is 45 ± 3 Å above CMC. These results are also consistent with previously published data.8, 9, 18, 21
Figure 3.14 Schematic representation illustrating the structure of the surfactant layer at the SiO2/water interface,
based on the DPI, QCM-D, SE and NR study.
Figure 3.14 shows the schematic diagram illustrating the structure of the surfactant layer at the
SiO2/water interface. From the previous analysis, the surfactant molecules form micelles on the
silica surface with a denser part close to the surface and a comparatively more dilute part towards
the bulk solution.
90
3.5 Conclusion
The four different surface characterisation techniques utilised in the work presented here were
shown to be capable of providing consistent data with a high degree of complementarity. Whilst
NR was found to be capable of providing precise static equilibrium structure information, DPI
was found to be a reliable method for measurement of the dynamic evolution of the thickness and
refractive index of an adsorbed layer. Given the correct dn/dC values for the solutions, DPI was
shown to be able to provide for precise calculation of the adsorbed amount, the thickness and the
mass density, in high accordance with neutron reflection results. These results were able to be
used to gain information about the net orientations of molecules in the layer. From neutron
reflection results, an asymmetric bilayer structure was derived, which proved useful for QCM-D
analysis. Subsequent to a more detailed modelling with a constrained two layer Voigt model, the
QCM-D data provided a compelling corroboration of the bilayer structure by NR and DPI. QCM-
D is a well-established method for the detection of molecule adsorption at solid/liquid interfaces.24
The adsorbed masses measured by QCM-D sensors also include part of the solution coupled to
the adsorbed layers.25 There have been many previous studies that reported large differences
between the masses measured by QCM-D and those measured by other optical techniques for
adsorption from systems such as polypeptide and lipids.26, 27 Stalgren et al. 28 showed that the
adsorbed mass of C14E6 on silica measured by QCM-D was overestimated by approximately 100%
compared to the mass obtained by Ellipsometry. This result is consistent with the models
presented in the work here. The SE adsorbed masses and thickness data agree well with the DPI
and neutron data, where the same RI is used as that determined from DPI.
91
References
1. D. McDermott, J. Lu, E. Lee, R. Thomas and A. Rennie, Langmuir, 1992, 8, 1204-1210. 2. J. R. Lu and R. K. Thomas, Journal of the Chemical Society, Faraday Transactions, 1998, 94, 995-
1018. 3. H. Arwin, Thin Solid Films, 1998, 313–314, 764-774. 4. F. Höök, B. Kasemo, T. Nylander, C. Fant, K. Sott and H. Elwing, Analytical Chemistry, 2001, 73,
5796-5804. 5. F. Höök, M. Rodahl, P. Brzezinski and B. Kasemo, Langmuir, 1998, 14, 729-734. 6. P. D. Coffey, M. J. Swann, T. A. Waigh, F. Schedin and J. R. Lu, Opt. Express, 2009, 17, 10959-
10969. 7. P. D. Coffey, M. J. Swann, T. A. Waigh, Q. Mu and J. R. Lu, RSC Advances, 2013, 3, 3316-3324. 8. D. C. McDermott, J. R. Lu, E. M. Lee, R. K. Thomas and A. R. Rennie, Langmuir, 1992, 8, 1204-
1210. 9. F. Tiberg and M. Landgren, Langmuir, 1993, 9, 927-932. 10. F. Tiberg, B. Lindman and M. Landgren, Thin Solid Films, 1993, 234, 478-481. 11. J. Penfold, E. Staples and I. Tucker, Langmuir, 2002, 18, 2967-2970. 12. J. Penfold, R. Richardson, A. Zarbakhsh, J. Webster, D. Bucknall, A. Rennie, R. Jones, T. Cosgrove,
R. Thomas and J. Higgins, Journal of the Chemical Society, Faraday Transactions, 1997, 93, 3899-3917.
13. P. D. Coffey, Interfacial Measurements of Colloidal and Bio-colloidal Systems in Real-Time, 2011.
14. J. R. Lu, M. J. Swann, L. L. Peel and N. J. Freeman, Langmuir, 2004, 20, 1827-1832. 15. J. De Feijter, d. J. Benjamins and F. Veer, Biopolymers, 1978, 17, 1759-1772. 16. S. X. Liu and J.-T. Kim, Journal of the Association for Laboratory Automation, 2009, 14, 213-220. 17. B. Johs and J. S. Hale, physica status solidi (a), 2008, 205, 715-719. 18. J. Penfold, E. Staples, I. Tucker and R. K. Thomas, Langmuir, 2004, 20, 1269-1283. 19. B. Lillis, M. Manning, H. Berney, E. Hurley, A. Mathewson and M. M. Sheehan, Biosensors and
Bioelectronics, 2006, 21, 1459-1467. 20. C. Gutig, B. P. Grady and A. Striolo, Langmuir, 2008, 24, 4806-4816. 21. F. Tiberg, B. Joesson and B. Lindman, Langmuir, 1994, 10, 3714-3722. 22. E. M. Lee, R. K. Thomas, J. Penfold and R. C. Ward, The Journal of Physical Chemistry, 1989, 93,
381-388. 23. L. Macakova, E. Blomberg and P. M. Claesson, Langmuir, 2007, 23, 12436-12444. 24. M. Rodahl, F. Hook, A. Krozer, P. Brzezinski and B. Kasemo, Review of Scientific Instruments,
1995, 66, 3924-3930. 25. E. Reimhult, C. Larsson, B. Kasemo and F. Höök, Analytical Chemistry, 2004, 76, 7211-7220. 26. T. J. Halthur and U. M. Elofsson, Langmuir, 2004, 20, 1739-1745. 27. R. P. Richter and A. R. Brisson, Biophysical Journal, 2005, 88, 3422-3433. 28. J. J. R. Stålgren, J. Eriksson and K. Boschkova, Journal of Colloid and Interface Science, 2002, 253,
190-195.
92
Chapter 4
Surface adsorption and solution aggregation of wool keratin studied by
neutron reflection and small-angle neutron scattering
The surface adsorption of keratin polypeptides extracted from wool was studied using surface
tension (ST) and neutron reflectivity (NR). Solution aggregation was studied by dynamic light
scattering (DLS) and small-angle neutron scattering (SANS). ST showed a steady surface tension
reduction with increasing concentration and NR revealed a steadily rising surface adsorption of
up to 0.1 mg/ml. Since no specular NR signal arose from the air/null reflecting water (NRW)
interface, the NR measurements at this interface offered the highest sensitivity to the adsorbed
layers. It was found that the interfacial layers were comprised of two main regions, a dense top
layer of 18-25 Å and a loose bottom layer of 25-30 Å. Parallel measurements at the air/D2O
interface revealed that approximately half of the top dense layer was exposed to air with the
remainder of the top layer, and the diffuse bottom layer immersed in the D2O substrate. Both the
volume fraction and the layer thickness were found to increase with keratin solution concentration
as did the adsorbed amount which was seen to plateau just above 2 mg m-2 at approximately 0.1
g dm-3. It was found that an increase in NaCl was found to decrease surface adsorption, in
association with thinning of the top layers. However, DLS revealed that the occurrence of
aggregates with their hydrodynamic radii peaked at approximately 100 Å with a density of
approximately 0.1 g dm-3. This is consistent with SANS modelling of ellipsoidal aggregates,
which had a major radius of 140 Å and the minor radius of 60 Å. With increasing NaCl
concentrations, it was found that the ellipsoids became thinner but longer, which is consistent with
the electrostatic effect as observed from the surface adsorption. This implies that as the
polypeptide chains become stiffer, they were more readily aligned, resulting in the formation of
thinner layers and longer aggregates.
93
4.1 Theoretical background
Keratins are widely distributed in skin, hair and nails and provide both epithelial and endothelial
coverage of organs. The keratin studied in the work presented in this thesis has been described in
Section 1.3.2. To study keratin adsorption at different interfaces, a stable keratin solution is
required. An important feature of keratin as compared to other proteins such as collagen and
elastin is the existence of a large number of disulphide bonds and hydrophobic amino acids. This
makes it difficult to dissolve keratin in most solutions or solvents including water 1. The keratin
used in this work was extracted from sheep wool by utilising reducing agents to break the
disulphide bonds. This produces keratin which is readily soluble in water. In this process, the
reducing agents cause little chemical alteration or damage to the protein2. An interesting
application of the work presented here is in the context of the use of keratin materials in personal
care. Although keratins have previously been rendered water soluble by a range of approaches,
information regarding their adsorption at interfaces and aggregation in solution is required for
their use in this field. Of current interest is also investigation of their interactions with various
formulation ingredients such as surfactants. The work reported here presents an initial study of
the surface adsorption and solution aggregation of keratin derived from wool using a combination
of neutron reflection (NR) and small-angle neutron scattering (SANS).
Protein folding and related structural changes in solution have been of interest over recent
decades3-5. Due to difficulties in the measurement of structural changes with proteins
experimentally, a combination of techniques is usually required. Various modes of nuclear
magnetic resonance techniques including nuclear overhauser effect spectroscopy (NOESY) have
provided a great wealth of information regarding intra- and inter-molecular structures with a high
degree of resolution 6. However, these methods have limitations in the detection of interfacially
adsorbed protein layers, particularly at the air/liquid interface. Circular dichroism (CD)
94
determines a protein’s secondary structures by detecting the impact on the optical polarisation
associated with chiral centres under different environments 7. Fourier transform infrared
spectroscopy (FTIR) is also able to reveal secondary structural information at both substrate
interface and solution 8, 9. Additional techniques including ellipsometry and dual polarisation
interferometry (DPI) are only able to measure the adsorbed amount of materials at an interface,
and are unable to detect structural changes in the protein layers 10, 11.
Neutron reflection has been widely used in thin film studies and is one of the most commonly
used techniques for the study of the adsorption of surfactants, polymers, proteins, and their
mixtures, at the air/water interface 12, 13. This technique, when used in conjunction with selective
deuterium labelling14 has the unique advantage of allowing for the determination of each
individual component in a mixture. A further advantage of neutron reflection is that it allows
measurement of both the amount of surface adsorption and the thickness of the adsorbed layer
with Ångstrom resolution15, 16. Because NR is highly sensitive to H/D isotopic substitution,
simultaneous measurements can be performed by labelling the surfactants, or by using H2O, D2O
or any of their mixtures as a solvent. This results in significant improvements in the sensitivity
and resolution of the interfacial structures. Whereas NR at an air/water interface provides a useful
tool in determining the structural conformations of keratin molecules at the interface, the
scattering profile from small-angle neutron scattering provides a way to estimate the dimensional
conformations of the keratin molecules. The effect of salt in producing structural changes in the
keratin molecules can also be investigated.
95
4.2 Materials
4.2.1 Keratin production
Keratin was obtained by following a recognised procedure with the application of some
modifications 17, 18. The modification involved the immersion of 1 g of degreased wool (fine cuts)
in 10 mL of the dissolving solution. The solution comprised of 8 M urea, 0.2 M SDS and 0.5 M
Na2S2O5. The mixture was then heated to 100 °C for a duration of 30 min and filtered and washed
through a stainless steel mesh. The filtered solution was then encased in cellulose tubing
(molecular weight cut-off of 12000-14000 Da (Sigma)) at an ambient temperature of 18-20 °C for
dialysis against distilled water. The water was replaced every 4~5 hours during dialysis for
approximately 4 days in order to remove excess reagents (urea, SDS and Na2S2O5), with the final
conductance close to that of the pure water. The obtained keratin stock solution was stored at 4°C
prior to subsequent use.
Investigation of the extracted keratin molecular weight distribution was carried out by SDS
polyacrylamide gel electrophoresis (SDS-PAGE), using the Mini-PROTEAN 3 Cell system from
Bio-Rad. Stacking gels (6% acrylamide of about 0.75 mm thickness) and resolving gels (12%
acrylamide of about 0.75 mm thickness) were prepared according to a standard method described
by the Mini-PROTEAN 3 Cell Instruction Manual (run at a constant voltage of 150 V). Keratins
were visualized by Coomassie Brilliant blue G 250 stain using a protein marker (Biolabs) for
calibration. The molecular weight distribution of extracted wool keratins was observed as two
main bands, at approximately 45 kDa (equivalent to human type Ia keratins) and 60 kDa
(equivalent to human type IIa keratins). In addition, several weak bands were observed,
corresponding to the low molecular weights at approximately 6-9 kDa and 10-20 kDa, which were
attributed to the high-sulphur and high-glycine/tyrosine proteins of the matrix and the low-sulphur
96
intermediate filament proteins. The findings of these keratins from wool extraction is consistent
with results reported previously by other groups17, 18.
For ease of data analysis in neutron reflection, it is necessary to use one keratin as a model
molecule and to estimate its scattering length density (SLD) and to calculate the amount of surface
adsorption. Table 4.1 shows how SLD values vary with the ratio of H2O and D2O due to the labile
H/D exchanges. An important observation can be made from the data in Table 4.1, in that the SLD
values change little for different proteins in a given solvent. This illustrates that although proteins
differ in sequence and in their physical and biological properties, all their amino acid compositions
tend to be similar 19.
Table 4.1 Comparison of the scattering length density (SLD) values as calculated from three different keratins and
BSA. This illustrates that despite different sequences their SLDs in a given solvent (H2O, D2O or their mixture) are
virtually the same. 20, 21 NRW is null reflecting water.
SLD in D2O
(×10-6 Å -2)
SLD in H2O
(×10-6 Å -2)
SLD in
NRW (×10-6 Å -2)
Molecular
Weight (g/mol)
Molecular
Volume (Å3)
Human Hair Keratin (a3) 3.46 1.94 2.07 46174 54871
Mouse Hair Keratin (a1) 3.42 1.94 2.06 47859 56842
Sheep Wool Keratin (47.6) 3.40 1.91 2.03 47332 56620
Bovine Serum Albumin 3.45 1.97 2.09 69323 83421
4.2.2 Preparation of keratin solutions
The sample solutions of keratin were prepared with ultrapure water (Elga, Vivendi Water Systems
Ltd.) containing 5 mM NaCl. Solution pH was adjusted to the required value by use of the
minimum amount of HCl or NaOH. All measurements were carried out at an ambient temperature
of approximately 25˚C unless otherwise stated.
97
4.3 Results
4.3.1 Surface tension
Surface tension measurements were carried out in order to understand the interfacial behaviour of
the keratin solutions and to assess their dynamic adsorption with time. These surface tension
studies also provide an estimate of the concentration range required in the subsequent neutron
reflectivity work. Figure 4.1 shows the time dependent surface tension changes for a series of
solutions prepared at pH 6. At the lowest concentration of 310-4 mg/ml, no surface tension
reduction was observed over the entire measurement period. As the keratin concentration was
increased to 110-3 mg/ml, there was seen an initial induction period of some 3000s, after which
the surface tension showed a clear trend of decline. With further increases in keratin concentration,
this induction period was seen to reduce, with the interfacial adsorption accelerating. As the
keratin concentration was increased up to 0.1 mg/ml, there was a further reduction of the induction
time (Figure 4.1b). Further increases in keratin above that point was seen to produce very small
changes in the surface tension profile, indicating fast saturation of interfacial adsorption. On the
basis of these and other surface tension measurements (not presented here), a concentration of 0.1
mg/ml was taken as the concentration required to produce saturated keratin adsorption under these
experimental conditions.
98
Figure 4.1 Surface tension measured over time from keratin solutions at pH 6. From top to bottom, concentrations
were fixed at 310-4 mg/ml, 110-3 mg/ml (0.21 uM), 310-2 mg/ml, 0.1 mg/ml and 0.3 mg/ml (6.3 uM). (a) Linear
plots; (b) Logarithmic plots highlight the concentration-dependent lag time for interfacial adsorption; (c)
equilibrium surface tension as a function of keratin concentrations.
4.3.2 Neutron reflectivity
4.3.2.1 Reflectivity profiles of keratin in the null reflecting water subphase
Neutron reflectivity was initially used to determine the adsorbed amount of keratin at the air/water
interface as a function of increasing protein concentration. This is usually carried out using a
mixture of approximately 8% D2O in H2O, termed null-reflecting water (NRW). NRW has the
same scattering length density as air, and hence all reflectivity originates from the adsorbed
interfacial layer. This technique is therefore able to provide invaluable information regarding both
the adsorbed amount and the interfacial thickness 22, 23. All measurements were conducted at 25˚C
with a pH of 6 and 5 mM NaCl. Reflectivity profiles for a series of keratin solutions ranging from
310-3 to 0.3 mg/ml are shown in Figure 4.2. The reflectivity can be seen to increase with
increasing keratin concentration, and to reach almost full adsorption at a concentration of 0.1
99
mg/ml. Small changes in reflectivity were observed when the concentration was further increased
to 0.3 mg/ml. It should be noted that time dependent effects were avoided by ensuring that after
the solutions had been loaded into the liquid troughs, a state of equilibrium was reached, the
timescales for which were obtained from the surface tension data shown in Figure 4.1.
Figure 4.2 NR profiles as a function of momentum transfer at the air/NRW interface with keratin concentrations at
0.003 (red), 0.01 (purple), 0.03 (blue) and 0.3 (black) mg/ml, with the NaCl concentration constant at 5mM and a
temperature of 25˚C. The solid lines indicate the optimum two layer fits.
1E-7
1E-6
1E-5
1E-4
1E-3
0.02 0.20
Ref
lect
ivit
y
Momentum transfer, Q (Å-1)
100
Figure 4.3 Neutron reflectivity measured from 0.1 mg/ml keratin solution adsorbed at the air/NRW interface. The
red dashed line indicates the best uniform layer fit with a thickness of 30 Å and a volume fraction of the protein at
0.6. The green dotted line also shows a uniform layer fit with a thickness of 25 Å and a volume fraction of protein
also at 0.6. The continuous line represents the best fit of a two layer model with a top dense layer of 25 Å on the air
side and a bottom dilute layer of 30 Å in NRW.
An initial fit to the profiles using a one layer fit proved unsatisfactory. An example of this can be
seen in the reflectivity profile for 0.01 mg/ml as shown in Figure 4.3. It can be seen that the red
(long dashed) line and green (short dashed) line, both of which represent the best one layer fits to
the measured profile under 2 layer thicknesses, fail to adequately model the data. Introduction of
a second layer in the interfacial region with a lower scattering length density produced an
improved fit to the data as shown by the continuous line in Figure 4.3. The fit parameters used are
given in Table 4.2.
The adsorbed interfacial layers were therefore modelled by a dense top layer with a thickness
increasing from 18 to 25 Å and a concentration ranging from 0.003 to 0.01 mg/ml. The
corresponding second layers were much looser, extending towards the bulk aqueous phase with a
layer thickness between 18 to 26 Å over the same concentration range. The volume fraction of
1E-7
1E-6
1E-5
1E-4
1E-3
0.02 0.20
Ref
lect
ivit
y
Momentum transfer, Q (Å-1)
101
keratin in both the top and bottom layers increased with increasing concentrations, ranging from
0.3 to 0.6 for the dense top layer and 0.05 to 0.11 for the loose bottom layer. The adsorbed amount
of keratin increased with increasing protein concentration and reached the highest value of 2.29
mg/m2 at a concentration of 0.1 mg/ml. No further increase was observed at the highest
concentration of 0.3 mg/ml studied. Figure 4.2 shows the NR profiles at the air/NRW interface
with the 4 representative keratin concentrations studied.
Table 4.2 The best two layer model fits to reflectivity profiles shown in Figure 4.3 under different keratin
concentrations in NRW, NaCl of 5 mM, at 25˚C.
[Keratin]
mg/ml
layer Thickness (Å)
±1
SLD
(×10-6 Å -2)
±0.01
Volume
fraction
±0.01
Layer mass
(mg/m2)
±0.05
Total mass
(mg/m2)
±0.1
0.003 1 18 0.6 0.30 0.75 0.88
2 20 0.11 0.05 0.13
0.01 1 22 1 0.50 1.53 1.81
2 24 0.16 0.08 0.28
0.03 1 24 1.06 0.53 1.76 2.06
2 26 0.17 0.08 0.30
0.1 1 25 1.1 0.55 1.90 2.29
2 26 0.22 0.11 0.39
0.3 1 25 1.1 0.55 1.90 2.29
2 26 0.22 0.11 0.39
4.3.2.2 Reflectivity profiles of keratin in the D2O subphase
Reflectivity profiles at the air/water interface were also measured from keratin solutions in D2O,
as shown in Figure 4.4. The use of D2O as the subphase allows more insight to be gained regarding
the interfacially adsorbed structure of keratin. D2O possesses a very different SLD from air, and
102
thus the reflectivity profiles in Figure 4.4 include contributions from the adsorbed keratin layer,
the D2O subphase and interferences between them. This allows information to be obtained
regarding the proportions of keratin in air and in D2O. The change in SLD for keratin upon moving
from NRW to D2O (2.03×10-6 Å-2 in NRW to 3.40×10-6 Å-2 in D2O) was accounted for in the
models.
Figure 4.4 NR profiles as a function of momentum transfer at the air/D2O interface with keratin concentrations of
310-3, 110-2, 310-2, 0.3 mg/ml. Solutions were prepared in 5 mM NaCl. Solid lines indicate the best fits to the
data.
In order to reduce the number of fit parameters involved, the reflectivity profiles were modelled
using the layer thickness and adsorbed keratin amount as determined from the NRW
measurements, with the only variable remaining, being the extent of layer immersion into D2O.
The data indicate that the proportion of keratin immersed in the aqueous phase is much greater
that in air. As a result, it was necessary to split the top layer as fitted from the NRW into two
separate layers, representing the fraction exposed to air and the fraction immersed in water. The
overall interfacial region was therefore described by a three layer model, in which the top layer is
5E-7
5E-6
5E-5
5E-4
5E-3
5E-2
0.02 0.20
Ref
lect
ivit
y
Momentum transfer, Q (Å-1)
K1S 0.003 g/l D2O 5 mM NaCl pH 5.6
K1S 0.01 g/l D2O 5 mM NaCl pH 5.8
K1S 0.03 g/l D2O 5 mM NaCl pH 5.8
K1S 0.3 g/l D2O 5 mM NaCl pH 5.8
103
exposed to air and the remaining 2 layers are fully immersed into the aqueous phase. The layer
exposed to air was found to increase in thickness from 10 to 13 Å as the concentration increased
from 310-3 mg/ml to 0.03 mg/ml. No further increase was found as the concentration increased
to 0.3 mg/ml. For all concentrations studied, approximately half of the top layer was exposed to
air and it was found to contain approximately half of the total adsorbed mass. The fit parameters
used are given in Table 4.3.
Table 4.3 Three layer model fits to the reflectivity profiles of keratin measured in D2O aqueous subphase.
[Keratin]
mg/ml
layer Thickness
(Å)
±1
SLD
(×10-6 Å -2)
±0.01
Volume
fraction
±0.01
Layer mass
(mg/m2)
±0.05
Total mass
(mg/m2)
±0.1
0.003 1 10 1.1 0.3 0.42 0.92
2 8 5.2 0.3 0.33
3 25 6.15 0.05 0.17
0.01 1 12 1.6 0.47 0.78 1.63
2 10 4.8 0.49 0.68
3 25 6.1 0.06 0.17
0.03 1 13 1.8 0.53 0.96 2.08
2 12 4.75 0.57 0.95
3 25 6.1 0.06 0.17
0.1 1 13 1.9 0.56 1.02 2.25
2 12 4.75 0.57 0.95
3 30 6.1 0.07 0.29
0.3 1 13 1.9 0.56 1.02 2.25
2 12 4.75 0.57 0.95
3 30 6.1 0.07 0.29
104
4.3.2.3 Reflectivity profiles of keratin with 0.5 M NaCl in the null reflecting water subphase
Table 4.4 Two Layer Model fits to the reflectivity profiles of keratin measured with 5 mM and 500 mM NaCl
under null reflecting water.
[NaCl]
(mM)
layer Thickness
(Å)
±1
SLD
(×10-6 Å -2)
±0.01
Volume
fraction
±0.01
Layer mass
(mg/m2)
±0.05
Total mass
(mg/m2)
±0.1
500 1 19 1.05 0.52 1.3
1.65 2 25 0.2 0.1 0.35
5 1 25 1.1 0.55 1.9
2.29 2 26 0.22 0.11 0.39
In order to investigate the effect of electrolytes on the adsorption of keratin at the air/water
interface, NR measurements in NRW were replicated in the presence of 500 mM NaCl. Table 4.4
shows the reflectivity profiles for 0.3 mg/ml of keratin in 5 mM NaCl (illustrated by black circles)
and 500 mM NaCl (illustrated by blue circles). The fit parameters are given in Table 4.4. The
decrease in reflected intensity indicates a lower adsorbed amount at higher salt concentrations.
The thickness of the top dense layer was also found to decrease from 25 Å to 19 Å with increasing
concentration of NaCl, whereas the keratin volume fraction changed little. The thickness of the
bottom layer and the scattering length densities of both the top and bottom layers changed slightly.
Therefore, it can be concluded that the reduced adsorption was directly linked to the loss of the
hydrophobic segments exposed to the air, which suggests that the keratin polypeptides may
become more hydrophilic as a result of increased ion binding.
105
Figure 4.5 Neutron reflectivity as a function of momentum transfer measured from 0.3 mg/ml keratin in NRW in
5 mM NaCl (black circles) and 500 mM NaCl (blue circles). The lines indicate the best fits to the data: the
thickness of the dense top layer reduces from 25 Å to 19 Å with NaCl concentrations being increased from 5 mM to
500 mM.
4.3.3 Dynamic light scattering (DLS)
Figure 4.6 shows comparison of the hydrodynamic diameters of nanoobjects measured from
keratin solutions in 1 mg/ml solution for different salt concentrations. It can be seen that the size
of molecules was increased from 140 Å to 300 Å, by an increase in the salt concentration from 5
mM to 0.5 M. These results are in strong agreement with the SANS experiments, which will be
discussed later. It is well accepted that the volume intensity of DLS is proportional to the total
number of particles and the volume of the particle.
1E-7
1E-6
1E-5
1E-4
1E-3
0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q (Å-1)
106
Figure 4.6 DLS size distribution of keratin polypeptides in 1 mg/ml solution with salt concentrations at 5 mM (red
line), 100 mM (blue line), 200 mM (green line) and 500 mM (purple line), respectively. The dashed line and dotted
line represent keratin solutions of 0.3 mg/ml and 0.1 mg/ml with 5 mM NaCl, respectively.
According to information in the DLS manual, concentrations of keratin solutions below 0.1 mg/ml
are not recommended for estimating sizes in the range of 10 to 100 Å. However, a set of DLS
measurements was performed from 0.0001 mg/ml to 0.1 with 10 points of data in total. This
showed that the particle sizes varied at values of 15±5 Å and 140±20 Å and were not consistent,
and indicated that the keratin solution had already reached its CAC lower than 0.1 mg/ml, which
is consistent with the surface tension measurement.
0
4
8
12
16
0 10 20 30 40 50 60 70
Volu
me
(per
cen
tage)
Hydrodynamic radius /nm
107
4.3.4 Small-angle neutron reflection (SANS)
Figure 4.7 The effect of keratin concentration on its aggregation studied at 0.25, 0.5 and 1 mg/ml by SANS. The
intensity I(Q) is shown as a function of momentum transfer. All solutions were prepared in buffers with 5 mM
NaCl at pH 6.1. All the scattering curves were able to be fitted with the same size and shape of aggregates,
indicating that they remained the nanostructure over this concentration range.
Figure 4.7 shows the scattering intensity profiles of keratin solutions with concentrations of 0.25
mg/ml to 1 mg/ml. The intensity profiles from three concentrations of keratin were different,
however, the profiles remained the same shape, indicating that the size and shape of the scattering
objects were the same. The optimally fitted parameters as shown in Table 4.5 revealed the
aggregates were ellipsoidal particles with a long radius of 138±10 Å and a short radius of 60±2
Å. It can also be noted that the fit was very sensitive to the short radius of the ellipsoid and less
to the long radius. The only difference in the parameters fitted for the scattering profiles measured
at the three concentrations was found to be the volume fraction. This implies that the polypeptides
start to aggregate below 0.25 mg/ml and that changes in the keratin concentration over this range
produces no change in the size and shape of the aggregates.
1E-3
1E-2
1E-1
1E+0
1E+1
0.006 0.06
I(Q
) (c
m-1
)
Q (Å-1)
k1s 1gL pH 6.1
k1s 0.5gL pH 6.1
k1s 0.25gL pH 6.1
108
Table 4.5 Ellipsoidal model fits to the small-angle neutron scattering profiles of keratin measured with 5 mM NaCl.
Ellipsoid model Keratin 1 mg/ml Keratin 0.5 mg/ml Keratin 0.25 mg/ml
SLD Ell (±0.02 /e-6 Å-2) 4.48 4.48 4.48
SLD Solv (e-6 Å-2) 6.35 6.35 6.35
Radius a (±1 /Å) 60 60 59
Radius b (±2 /Å) 138 139 138
Volume Fraction 0.001 0.0005 0.00025
Figure 4.8 Effect of NaCl on keratin solution aggregation measured by SANS with NaCl concentrations varied
from 5 mM to 0.5 M at pH 6 in D2O. The intensity is shown as a function of momentum transfer.
Figure 4.8 shows the scattering profiles of keratin solutions at 1 mg/ml of keratin concentration
with NaCl over a range of 5 mM to 0.5 M in D2O. It can be seen that an increase in the salt
concentration reduced the scattered intensity gradually, which suggests that the size and shape of
the ellipsoidal particles changed gradually with increasing NaCl concentrations. Interestingly, the
scattering profiles changed little over the low NaCl concentration range, with larger reductions in
scattering seen when NaCl concentration was above 100 mM. A clear trend can be seen which
1E-3
1E-2
1E-1
1E+0
1E+1
0.008 0.08
Inte
nsi
ty (
cm-1
)
Q (Å-1)
K1S+500mM NaClK1S+200mM NaClK1S+100mM NaClK1S+40mM NaClK1S+20mM NaCl
109
indicates that the presence of NaCl in protein solution stretched the ellipsoidal particle, making it
thinner and longer with increasing NaCl concentration. In the presence of 0.5 M NaCl, the keratin
molecules formed an ellipsoidal shape with a short radius of 36 Å and a long radius of 306 Å,
compared to the radii of 58 and 138 Å with 5 mM NaCl. As the salt concentration increased, the
short axial radius remained almost constant up to 100 mM NaCl. Above this concentration, the
short axial radius began to decrease. The long axial radius showed a steady increase over the entire
salt concentration range studied. The shape change of keratin aggregation with salt is illustrated
in Figure 4.9. Table 4.6 shows the ellipsoidal model fits of keratin measured from Figure 4.8. The
volume of the keratin aggregate can be calculated using
V =4
3𝜋𝑎2𝑏 (4.3.4.1)
where a and b are the short and long radiuses of the ellipsoid, respectively. The percentage of
keratin in each aggregate can also be calculated as
(4.3.4.2)
The volumes of the scattered particles under different NaCl concentrations were calculated using
equation 4.3.4.1 and the results are listed in Table 4.6. Whilst the aggregate shape showed a steady
transition with NaCl concentration, the volume peaked around 100 mM NaCl. The volume
fractions of keratin in the keratin aggregates occupied an almost constant value of 0.6, except at
the highest salt concentration of 500 mM, where it reduced to 0.4. Given that the molecular
volume of keratin was taken to be approximately 56,600 Å3, the molecular number of keratin in
an aggregate can be calculated. Over the entire salt concentration studied, the average number of
keratin molecules per micellar aggregate began at 21 at NaCl = 5 mM, peaked at 38 at NaCl =
100 mM and then reduced to 14 at NaCl = 500 mM. This shows that over the entire concentration
110
range studied, keratin formed aggregates and the size and shape of the aggregates changed in
response to changes in the salt concentration.
Table 4.6 Ellipsoidal model fits to the small-angle neutron scattering profiles of keratin at 1 mg/ml measured with
NaCl changed from 5 mM to 0.5 M in D2O, pD 6.3.
Ellipsoid model [NaCl] =
5mM
[NaCl] =
20mM
[NaCl] =
40mM
[NaCl] =
100mM
[NaCl] =
200mM
[NaCl] =
500mM
SLD Ellip (±0.02 /10-6 Å-2) 4.48 4.5 4.53 4.58 4.69 5.28
Radius a (±2/Å) 60 59 59 55 50 37
Radius b (±5/Å) 138 168 201 260 280 350
Volume fraction 0.001 0.001 0.001 0.001 0.001 0.001
Particle volume
(±0.02 /10-6Å 3)
1.94
2.37
2.55
3.54
2.93
2.01
Volume fraction of keratin
±0.02
0.58
0.59
0.6
0.61
0.61
0.39
No. of keratin molecules in
one aggregate ±1
21
25
27
38
31
14
A discussion of the fit sensitivity is useful at this point. The ellipsoidal model is very sensitive to
the short radius and the errors can be controlled within ±2 Å. However, the long radius has much
less sensitivity, hence the fits within ±10 Å were all acceptable. The values of radius b listed in
Table 4.6 are the minimum values able to be used to produce acceptable fits. Despite these
uncertainties, the size and shape as observed must be valid. The geometrical shape changes in the
aggregate were also well supported by DLS data as shown in Figure 4.6, where the measured
hydrodynamic diameters of the aggregates show substantial increases in peak values as well as
their distributions with NaCl concentrations.
111
4.4 Discussion
Figure 4.9 Schematic representation of a steady shape change of keratin aggregates with increasing NaCl
concentration deduced from SANS experiments.
On the basis of SDS-PAGE analysis, the K1S keratin polypeptide sample is comprised of hard
keratins (Ia and IIa) with two main MW bands around 45 and 62 kDa. Although the polypeptides
are able to can adsorb and reduce surface tension over a wide concentration range, the surface
tension reduction is low, with the lowest surface tension of 55 mN/m being reached at 0.3 mg/ml
(Figure 4.1c). The surface tension after 4 hours showed a steady decline with no occurrence of
any break point to indicate the start of solution aggregation. DLS and SANS measurements
confirmed the formation of aggregates at keratin polypeptide concentrations of around 0.1 mg/ml,
which was marked by the large hydrodynamic diameters peaked at approximately 20 nm (Figure
4.6) and the ellipsoids with long axial lengths of approximately 28 nm (Figure 4.7, Table 4.5). At
lower peptide concentrations, it proved difficult to detect aggregation by SANS, but DLS showed
clear signs of aggregates forming as the polypeptide concentration were reduced to 0.03-0.1
mg/ml, which indicates that although the polypeptides were made soluble in the aqueous phase,
they were highly prone to aggregation.
112
The NR work carried out explored both the concentration and salt effects of keratin adsorption at
the air/water interface, with NRW and D2O contrasts to highlight the layer structure and
composition and their extent of immersion in water. All of the reflectivity profiles measured under
different conditions revealed a two layer structure for the adsorbed keratin polypeptides: a top
dense layer of 18-25 Å with a keratin volume fraction of 0.3-0.55; and a bottom loose layer of 25-
30 Å with a keratin volume fraction of 0.05-0.11. The top layer was found to contain 80-85% of
the polypeptides with the bottom layer contained the remaining 15-20%. The parallel
measurements in D2O revealed that almost half of the top layer was exposed to air and that this
exposed region also contained almost half of the adsorbed polypeptides. Figure 4.10 shows the
total surface adsorbed amount as a function of keratin concentration. It can be seen that the
adsorption plateaued approaching 0.1 mg/ml, which is consistent with the detection of aggregation
by both DLS and SANS measurements above this concentration.
It is also evident from Figure 4.10, that an increase in salt concentration caused a reduction in the
surface adsorption. However, the main part that contributed to the reduction of the adsorbed
amount was the top dense layer, whereas the bottom loose layer remained broadly the same as a
function of salt concentration. The loss of the adsorbed polypeptides on the air side implied that
salt addition caused them to become more hydrophilic, possibly through charge association or
binding. However, the basic feature of a dense top layer and a loose bottom layer still remained.
This feature of adsorbed protein layers has also been observed using human lactoferrin24 and
human serum albumin (HSA) and bovine serum albumin (BSA) by NR12, although it was thought
that the adsorption of the albumins did not lead to any major structural unfolding.
113
Figure 4.10 Keratin surface adsorbed amounts obtained from NR plotted against solution keratin concentration,
with all data measured after 60 minutes of adsorption. The blue dots are measured from keratin solution with 5 mM
NaCl at keratin concentrations from 0.003 to 1 mg/ml, whilst the red dots are keratin solution with 0.5 M NaCl at
keratin concentrations of 0.03 and 0.3 mg/ml.
The aggregation of keratin polypeptides was first studied by DLS. At the low NaCl concentration
of 5 mM the hydrodynamic diameters were peaked around 200 Å. Increasess in salt concentration
above 100 mM led to the average diameters centering around 300 Å. The SANS work revealed
an ellipsoidal shape with a short radius of approximately 60 Å and a long radius of approximately
140 Å under 5 mM NaCl. An increase in the ionic strength of keratin solution was found to change
the radii of the ellipsoids formed. It was found that the keratin molecules were stretched into a
thinner, longer ellipsoid with increasing salt concentration, consistent with the thinning of the
adsorbed layer. Thus the SANS and NR work were broadly consistent in that the salt caused the
polypeptide to stiffen and become more hydrophilic.
The effect of ions on the conformation of polyelectrolytes has been extensively analysed25. With
reference to the interfacial adsorption and formation of multiple films of synthetic polyelectrolytes
0
0.5
1
1.5
2
2.5
1 10 100 1000
mass
(m
g/m
2)
Concentration (mg/l)
114
detailed by McAloney et al.26, observations described in the work here can be explained in terms
of the influence of the concentration of ions on the polyelectrolyte conformation in aqueous
solution. Specifically, the differences between the adsorbed layers and size and shape transitions
of micellar aggregates formed under different salt concentrations in this work can be accounted
for by considering a conformational transition from extended rod to globular coil in the bulk
polyelectrolyte solution, leading to a reduced effective persistence length and chain stiffening.
The assumption can be made from the results that the bulk conformation is largely retained during
adsorption and aggregation. At low salt concentrations, the extended rod configuration of the
polyelectrolytes adsorbs and aggregates. At high salt concentration, the polyelectrolytes have a
globular coil conformation, forming different layers and aggregates due to altered packing and
interaction between adjacent molecules. It can be noted that, unlike synthetic polymers, the keratin
polypeptides used in this work contained several polypeptide chains, each bearing both positive
and negative charges. The transitions with respect to ion concentration changes show close
similarity to the features observed by McAloney et al.26 in synthetic polyelectrolyte films.
115
Figure 4.11 Schematic diagram of the keratin distributions at air/water interface and in bulk solution.
4.5 Conclusion
The keratins extracted from wool, although water soluble up to relatively high concentrations,
were found to be readily adsorbed and aggregated. Surface adsorption was found to increase
steadily and plateau at approximately 0.1 mg/ml, at which point solution aggregation was seen to
occur. However, surface tension exhibited only a modest decline with no clear sign of a break
point to indicate an occurrence of aggregation in solution. Neutron reflection showed that the
adsorbed keratin formed a two-layer27 structure with a dense top layer of 18-25 Å and a loose
bottom layer of 25-30 Å. The keratin occupied approximately 55% of the dense top layer at
solution concentrations above 0.03 mg/ml and 6-11% of the area of the loose layer. Parallel D2O
116
measurements found that half of the dense top layer is exposed to air, highlighting the strong
amphiphilic nature of the keratin molecules. The SANS measurements found that the keratin
aggregates adopted an ellipsoidal shape with a major radius of approximately 140 Å and a minor
radius of approximately 60 Å in the low NaCl concentration of 5 mM, consistent with the DLS
measurements. Increase in NaCl was found to increase ion binding and in association with keratin
and chain stiffening, resulted in reduced adsorption and in thinning of the adsorbed layers and a
reduction in aggregation number and stretching of the ellipsoid. The discussions above concluded
that keratin conformation in solution is affected by the concentration of ions and that the adsorbed
layer conformation is more or less retained (adsorbed amount reduced to approximately 20% at
high NaCl concentrations). This can be accounted for by a conformational transition of keratin
particles in solution from globular to ellipsoidal with different ionic concentrations. The results
of adsorbed layers indicate that at low concentrations, the globular-shaped particles adsorb to form
a thicker layer while at high concentrations, the ellipsoidal-shaped particles produce a thinner
layer. The salt effects on polyelectrolyte multilayers have also been observed by McAloney et al.
26 and by Dubas and Schlenoff.27
117
References
1. P. M. Schrooyen, P. J. Dijkstra, R. C. Oberthür, A. Bantjes and J. Feijen, Journal of colloid and interface science, 2001, 240, 30-39.
2. A. Gupta, N. B. Kamarudin, C. Y. G. Kee and R. B. M. Yunus, Journal of Chemistry and Chemical Engineering, 2012, 6, 732-737.
3. P. Privalov and N. Khechinashvili, Journal of molecular biology, 1974, 86, 665-684. 4. G. M. Clore, A. T. Brünger, M. Karplus and A. M. Gronenborn, Journal of molecular biology,
1986, 191, 523-551. 5. A. Roy, A. Kucukural and Y. Zhang, Nature protocols, 2010, 5, 725-738. 6. H. Roder, G. A. Elöve and S. W. Englander, Nature, 1988, 335, 700. 7. N. Sreerama and R. W. Woody, Analytical biochemistry, 2000, 287, 252-260. 8. Y. Jiang, C. Li, X. Nguyen, S. Muzammil, E. Towers, J. Gabrielson and L. Narhi, Journal of
Pharmaceutical Sciences, 2011, 100, 4631-4641. 9. B. R. Singh, Infrared Analysis of Peptides and Proteins, 2000. 10. J. Armstrong, H. J. Salacinski, Q. Mu, A. M. Seifalian, L. Peel, N. Freeman, C. M. Holt and J. R. Lu,
Journal of Physics: Condensed Matter, 2004, 16, S2483. 11. B. J. Cowsill, T. A. Waigh, S. Eapen, R. Davies and J. R. Lu, Soft Matter, 2012, 8, 9847-9854. 12. J. Lu, T. Su and R. Thomas, Journal of Colloid and Interface Science, 1999, 213, 426-437. 13. D. Lyttle, J. Lu, T. Su, R. Thomas and J. Penfold, Langmuir, 1995, 11, 1001-1008. 14. D. Cooke, C. Dong, J. Lu, R. Thomas, E. Simister and J. Penfold, The Journal of Physical Chemistry
B, 1998, 102, 4912-4917. 15. I. Purcell, J. Lu, R. Thomas, A. Howe and J. Penfold, Langmuir, 1998, 14, 1637-1645. 16. J. Lu, Annu. Rep. Prog. Chem., Sect. C: Phys. Chem., 1999, 95, 3-46. 17. C. Tonin, M. Zoccola, A. Aluigi, A. Varesano, A. Montarsolo, C. Vineis and F. Zimbardi,
Biomacromolecules, 2006, 7, 3499-3504. 18. A. Vasconcelos, G. Freddi and A. Cavaco-Paulo, Biomacromolecules, 2008, 9, 1299-1305. 19. A. Cooper, M. W. Kennedy, R. I. Fleming, E. H. Wilson, H. Videler, D. L. Wokosin, T.-j. Su, R. J.
Green and J. R. Lu, Biophysical journal, 2005, 88, 2114-2125. 20. J. Yu, D.-W. Yu, D. M. Checkla, I. M. Freedberg and A. P. Bertolino, Journal of investigative
dermatology, 1993, 101, 56S-59S. 21. K. Hirayama, S. Akashi, M. Furuya and K.-i. Fukuhara, Biochemical and biophysical research
communications, 1990, 173, 639-646. 22. Z. Li, J. Lu and R. Thomas, Langmuir, 1997, 13, 3681-3685. 23. J. R. Lu, S. Perumal, E. T. Powers, J. W. Kelly, J. R. Webster and J. Penfold, Journal of the
American Chemical Society, 2003, 125, 3751-3757. 24. J. R. Lu, S. Perumal, X. Zhao, F. Miano, V. Enea, R. R. Heenan and J. Penfold, Langmuir, 2005, 21,
3354-3361. 25. W. F. Reed, S. Ghosh, G. Medjahdi and J. Francois, Macromolecules, 1991, 24, 6189-6198. 26. R. A. McAloney, M. Sinyor, V. Dudnik and M. C. Goh, Langmuir, 2001, 17, 6655-6663. 27. S. T. Dubas and J. B. Schlenoff, Langmuir, 2001, 17, 7725-7727.
118
Chapter 5
Interaction of keratin and surfactants of sodium dodecyl sulfate and
dodecyl trimethyl ammonium bromide at the air/water interface
Interactions of keratin polypeptides and the surfactants of sodium dodecyl sulfate (SDS) and
dodecyl trimethyl ammonium bromide (DTAB) were studied by surface tension (ST) and neutron
reflectivity (NR) at the air/water interface. ST showed apparently different features between
keratin/SDS and keratin/DTAB complexes due to different hydrophobic and electrostatic forces.
Keratin/SDS system has a steady surface tension reduction with increasing concentration of SDS
at a fixed concentration of keratin while keratin/DTAB system has two minimum points with
increasing concentration of DTAB, coinciding with the occurrence of precipitation at the charge
neutralization point. Since no specular NR signal arose from the air/NRW interface, the NR
measurements at this interface offered the highest sensitivity for the measurements of the adsorbed
layers.
The results revealed that the addition of a small amount of SDS (0.1 mM) significantly reduced
the adsorbed amount of keratin, whereas DTAB had less of an effect on keratin adsorption until
its concentration reached 1 mM. In the case of the keratin/SDS complexes, the molar ratio of
SDS/keratin was approximately 25 at the interface at 0.01 mg/ml keratin and 0.1 mM SDS
concentrations. This ratio was found to increase to approximately 240 with the addition of 2 mM
SDS, indicating that SDS significantly suppressed the adsorption of keratin.9 The amount of
adsorbed keratin decreased from 1.81 g dm-3 to 0.28 g dm-3 with the addition of 2 mM SDS. In
the case of the keratin/DTAB complexes, the ratio of DTAB/keratin was approximately 7 with
0.1 mM DTAB and increased to approximately 19 with the addition of 0.4 mM DTAB. This ratio
was further increased to approximately 135 with the addition of 2 mM DTAB. The adsorbed
119
keratin was found to remain constant with the addition of 0.7 mM DTAB, and to undergo a
reduction to 0.67 g dm-3 with the addition of 2 mM DTAB. The whole process indicated a complex
interaction due to the opposite charges of the two components. The results also indicated that the
combined electrostatic and hydrophobic forces have combined effects on the keratin/SDS and
keratin/DTAB complexes.
5.1 Literature review
Protein/surfactant interactions at interfaces have been a topic of great interest in recent decades.
Results of research carried out have found a wide range of applications from efficient washing
powders to products for personal hygiene.1 A large number of investigations on the interactions
were performed between late 1960s and early 1970s and determined the general principles of how
charged proteins and charged surfactants interact2, 3. Several techniques have been used to
investigate protein-surfactant complexes in bulk solutions such as calorimetry, fluorescence4, 5,
viscometry6, dynamic light scattering7 and small-angle neutron scattering8. Much of the work was
focused on the central areas of interest of whether the monomeric or micelle surfactant denatured
proteins and how proteins are unfolded by surfactants. Among these studies, a widely accepted
model is one in which the surfactants form micelles to bond to the polypeptide chains with a
smaller aggregation number compared to that in the free micelles6. The studies also revealed that
electrostatics is the main driving force for binding the monomeric surfactants to proteins below
the CMC. Hydrophobic interactions were found to be another important driving force, as
introduced and explained by Tanford’s classic monograph from 1980.9, 10 Tanford et al. also
demonstrated that the surfactant/protein bindings are strongly affected by the type of surfactant
headgroups and the pH of solutions.11, 12
Since the electrostatic driving force plays an important role in surfactant/protein bindings, the
charged surfactants are more likely to interact with charged proteins than non-ionic surfactants13.
120
Therefore, surfactants with opposite charges to a protein are more strongly bonded to the protein
than the same charged surfactants. Many studies have illustrated that electrostatic and
hydrophobic forces occur at the very first step of interactions, especially at low concentrations of
surfactants below the CMC. Above CMC, the micellar interactions with proteins becomes more
complex and produces more shape changes to the protein-surfactant complexes.14 Anionic
surfactants such as SDS can aggregate to form clusters on proteins at relatively high
concentrations,15-17 resulting in an apparent lowering of the CMC18. The structure of SDS/protein
complexes has been intensively studied since the 1960s. Reynolds and Tanford et al. demonstrated
that the shape of the complexes can generally be described as a ‘rod-like’ ellipsoidal aggregate of
protein-surfactant mixtures.10, 12, 19 Both electrostatic and hydrophobic properties of protein and
surfactant can be reflected by the dimensions of the complexes. The length of the ellipsoid
depends on the molecular weight of the protein. The width is influenced by the SDS alkyl chain
length, which is usually around 20 Å. Moren et al. 20studied the SDS-lysozyme interactions and
showed that precipitates can be formed at concentrations where charges between lysozyme and
SDS are neutralized. Green et al. 1 described the SDS/lysozyme interactions at air/liquid interfaces
and also demonstrated precipitations at the charge neutralization point of the complexes. The
studies all revealed that the electrostatic forces play an important role in SDS/lysozyme
interactions with opposite charged lysozyme/SDS complexes. Fewer studies have been carried
out on DTAB-protein interactions than on SDS-protein complexes at both bulk solutions and
interfaces. Monteux et al.21 reported on the oppositely charged polyelectrolyte/DTAB complexes
at the air/water interface and revealed that precipitations occurred at very low concentrations of
DTAB at interfaces.
It can be concluded from the studies described that the presence of a surfactant can significantly
affect how a protein behaves in both bulk solutions and at interfaces. However, much of the
interfacial work relating to protein-surfactant complexes has been performed at solid/liquid
121
interfaces in order to study the removal function of the surfactant on pre-adsorbed protein layers.22
In the work presented in this thesis, both the cationic surfactant of dodecyl trimethyl ammonium
bromide (DTAB) and the anionic surfactant of SDS were used as charged surfactants.
Investigations were carried out in order to examine how the interfacial adsorption of the
complexes behave and how the complexes interact at interfaces. Unlike the oppositely charged
interactions of lysozyme/SDS complexes, the keratin studied in the current chapter carried the
same negative charges of SDS at pH 6 solutions with an isoelectric point around pH 4.5. Both
keratin/SDS and keratin/DTAB systems were selected and used in this work in order to compare
how the adsorption is affected by electrostatic forces.
5.2 Experimental procedures
Neutron reflection experiments were performed on SURF at the ISIS Neutron facility. Typical
backgrounds in null reflecting water (NRW) and D2O were 5e-6 Å-2 and 2e-6 Å-2, respectively. For
a solution with a comparatively low concentration, the time effect of adsorption was checked first,
by repeating NR measurements at different time intervals (60, 120 and 180 minutes). It was found
that there was no time effect of adsorption after 60 minutes. All of the NR measurements were
performed after 90 minutes, in order to ensure solution equilibrium. Hydrogenated SDS (h-SDS)
and DTAB (h-DTAB) (purchased from Sigma) were recrystallized several times until the surface
tension profiles were seen to be consistent with previous published literature. The deuterated SDS
(d-SDS) and DTAB (d-DTAB) were made at Oxford University. They were also recrystallized
before use. A buffered solution with 5 mM NaCl and pH controlled at 6 was used for all samples.
Stock solutions of 0.97 mg/ml keratin and varying concentrations of SDS and DTAB were diluted
and prepared separately for each measurement. The solutions were then mixed with keratin
solutions prior to each experiment. All mixed samples appeared transparent without visible
precipitates.
122
5.3 Results and analysis
5.3.1 Surface tension measurements
The surface tensions of pure SDS and DTAB were measured in 5 mM NaCl buffered solutions,
followed by the measurement of keratin/SDS and keratin/DTAB complexes. Measurements of
surface tensions of SDS and DTAB in pure water have been previously reported, but their profiles
in 5 mM NaCl buffered solutions at pH 6 have not been specified.
It can be seen from Figure 5.1 that the CMCs of SDS and DTAB in 5 mM NaCl buffered solution
are 4±1 mM and 10±1 mM respectively. The surface tensions of keratin/SDS and keratin/DTAB
complexes are indicated by the purple and red lines. The results indicate that the keratin/SDS
complexes showed a steady surface tension reduction with increasing concentrations of SDS at a
fixed concentration of keratin, and that the keratin/DTAB complexes exhibited two minimum
points with increasing concentration of DTAB.
123
Figure 5.1 Surface tension measured as a function of surfactant concentrations. The grey line shows the isotherm of
pure SDS. The light blue line is the isotherm of pure DTAB. The purple line is the isotherm of SDS with a fixed
concentration of keratin at 0.01 mg/ml. The red line is the isotherm of DTAB with a fixed concentration of keratin
at 0.01 mg/ml. All measurements were performed in 5 mM NaCl buffer.
5.3.2 Neutron reflectivity
5.3.2.1 Reflectivity profiles of keratin and h-SDS complex in the NRW subphase
The NRW consists of 8.01% D2O by volume and the scattering length density is 0 under this
isotopic contrast, which makes the NRW transparent to neutrons. This implies that the reflectivity
profile is fully accounted for by the adsorbed protein at interfaces, enabling the adsorbed amounts
and structural conformations of the layer to be determined, and allows for a relatively simple data
analysis procedure.23, 24 The measurements were are all conducted at 25˚C and a pH of 6 with the
total ionic strength of NaCl fixed at 5 mM. All samples were equilibrated for 1.5 hours prior to
the measurements to ensure that the experimental results were obtained at the steady state.
124
Figure 5.2 shows the reflectivity profiles of the adsorbed keratin/h-SDS complex against wave
vector at a fixed keratin concentration of 0.01 mg/ml and a range of SDS concentrations from 0.1
mM to 1 mM with 5mM NaCl at pH 6. Since the air and NRW subphase are transparent to
neutrons, there is no reflection in the absence of any interfacial adsorption. Therefore, a flat
background line would be expected with an intensity of 5-9×e-6 Å
-2. Table SI 5.1 shows the fit
parameters of protonated and deuterated surfactants including the molecular volume, molecular
weight and molecular scattering length density. It can be seen from Table SI 5.1 that the
protonated surfactants have very low scattering length densities, which are almost invisible to
neutrons. Therefore the increases in reflectivity seen in Figure 5.1 can be assumed to be primarily
due to adsorbed keratins at the interface. It was found that the intensity of the signals decreased
with an increase of SDS concentrations and that the reflectivity signal became virtually zero with
the addition of 1 mM SDS.
The fit applied to the reflectivity data allows the structural conformation and composition of the
adsorbed layer to be revealed. The fitted model uses the optical matrix equation which requires
the assumption that wave functions and their gradients are continuous at each boundary for
multiple layers25. The least-squares iteration method was used for the fit processes with a
hypothetical layer structure as a starting point. In this process, determination of the number of fit
layers is based on the degree of inhomogeneity perpendicular to the interface. In the case where
the raw data cannot be fitted with the application of a one-layer model, multiple layers can be
used. The molecular weight of the keratin was calculated by addition of the weight of all the
individual amino acid inside the molecule, which gave a total weight of ~47,000 g/Mol.
125
Figure 5.2 Neutron reflectivity as a function of momentum transfer measured for keratin/h-SDS complexes with h-
SDS concentrations of 0 ( ), 0.1 (×) 0.3 ( ) and 1 mM ( )with a fixed keratin concentration of 0.01 mg/ml.
NaCl at 5mM was in the buffer throughout the experiments. Solid lines indicate the best uniform layer fits.
The first set of data in Figure 5.2 provides direct information on the surface excesses of keratin
with different SDS concentrations, since both h-SDS and NRW are invisible to neutrons. It was
found that the level of reflectivity decreased and the reflectivity curves became flat with increased
SDS concentration. The results reveal a clear decreasing trend between the adsorbed amount and
the thickness of the keratin layer. The reflectivity profiles can be seen to reduce almost to
background level at the SDS concentration of 1 mM, indicating that the adsorbed amount of
complexes had fallen to a negligible level. The adsorbed amount of the complexes can be seen to
have decreased from 1.54 mg/m2 in the absence of SDS to approximately 0.6 mg/m2 in the
presence of 1 mM SDS.
3E-7
3E-6
3E-5
3E-4
0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
126
Determination of the fit parameter of the keratin/h-SDS complexes measured at the NRW
interface, accounted for the small contributions of h-SDS reflectivity, by combining the parallel
measurements of the keratin/d-SDS complexes. There was found to be little difference (within
experimental error) produced by assuming that the h-SDS contributions of reflectivity were
negligible, which simplified the determination of the adsorbed amount of keratin at the interface.
It has previously been described in Chapter 4 that under the same conditions as the results here,
the pure keratin adsorption at 0.01 mg/ml was seen to form a two-layer structure with a dense
layer of 0.5 mg/m2 on the top and a very loose layer of 0.08 mg/m2 at the bottom. The fits with
the absence of the bottom loose layer are also acceptable, but the two-layer structure provides for
a more accurate fit. In the case of the keratin/h-SDS complexes, a uniform layer provides a good
fit since SDS interacts with the keratin and produces a much looser layer at the bottom that can
be assumed to be negligible. The adsorption of pure keratin at 0.01 mg/ml was found to lead to a
dense layer with an area per molecule of 5100 100 Å2 and a thickness of 22 Å. The area per
molecule and layer thickness was found to change to 5700 100 Å2 and 18 Å by the addition of
0.1 mM SDS. The results indicate that even a small amount of SDS can inhibit the adsorption of
keratin. Green et al.1 showed that SDS below 0.25 mM enhances the adsorption of lysozyme, and
proved that further addition of SDS (above 0.25 mM) resulted in a fast decrease of the adsorbed
amount of lysozyme. They found that the first increase in adsorption followed a stoichiometric
relation with the same molar ratios of protein and surfactant. This process is caused primarily by
electrostatic forces. Further addition of SDS broke the lysozyme structure. Once the structural
configuration of the lysozyme has been altered, interactions between SDS and the lysozyme would
gradually shift from an attraction by electrostatic forces to a combined interaction by both
electrostatic and hydrophobic forces. This causes a sharp decrease in adsorption subsequent to
adsorption maximum.
127
As in the case of the keratin/SDS complexes, SDS was shown to reduce the adsorption of keratin
at 0.1 mM SDS concentrations. Unlike in the lysozyme/SDS interactions where the surface
excesses initially increased with increasing SDS concentration, and were followed by a decrease
with additional SDS, the adsorbed amounts of keratin were found to be always decreased by SDS
at concentrations ranging from 0.1 mM to 1 mM. The lack of the first process in the keratin/SDS
systems can be well explained by the lack of electrostatic forces between keratin and SDS since
they all carried the same negative charges. This theory is further strengthened by the
keratin/DTAB systems, as discussed later in the thesis.
5.3.2.2 Reflectivity profiles of keratin and d-SDS complexes in the NRW subphase
Keratin/d-SDS complexes were also investigated under the same conditions as described in Figure
5.1. Combined measurements of keratin/h-SDS and keratin/d-SDS complexes allow for precise
determination of the surface excesses of SDS at the interface. Table 5.2 shows the combined fit
parameters for the structural conformations. As NRW contains 8.1% D2O, the scattering length
density of keratin takes into account the H/D exchanges between labile hydrogens in keratin and
deuterium in NRW. The H/D exchanges of keratin are assumed to be complete, which has
previously been described in work of Lu et al.26
The same one-layer model was used to fit the NR profiles performed at the same concentration
but with different isotopic labelling. It was found that the fits at higher concentrations of SDS
were less good than the fits at low concentrations, to an acceptable extent. To simplify data
analysis, the uniform layer model was also used in the keratin/d-SDS complexes.
128
Figure 5.3 Comparison of neutron reflectivity profiles as a function of momentum transfer. The black line was
measured with keratin at 0.01 mg/ml; the blue line was measured with 1 mM d-SDS; and the red line was measured
with a mixture of keratin and d-SDS. A NaCl concentration of 5mM was in the buffer throughout the experiments.
Figure 5.3 shows a comparison of the reflectivity profiles between SDS and keratin and their
mixture. The solid lines are fitted with a uniform layer model. The graph illustrates that the
reflectivity intensity of the keratin/SDS complexes was much greater than that of keratin, and less
than that of SDS. Table 5.1 shows the adsorbed layer conformation as described in Figure 5.3 as
fitted by a uniform layer model. The results indicate that the presence of SDS reduced the
thickness of the keratin layer from approximately 45 Å to approximately 32 Å and the adsorbed
amount of keratin from 20% to 9%. The area per molecule can be obtained using the equation
, which is described in equation 2.5.4.2. The area per molecule of keratin and SDS in
Figure 5.3 was found to be 6290 Å2 and 45 Å2 respectively while in the case of the mixture the
values were found to be 19660 Å2 and 82 Å2 respectively. This calculation shows that a single
keratin molecule interacts with approximately 240 SDS molecules.
3E-7
3E-6
3E-5
3E-4
0.05 0.1 0.2
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
129
Table 5.1 Best mono-layer fits to the reflectivity profiles in Figure 5.3. All measurements were performed with 0.5
M NaCl under NRW at pH 6.
Keratin
/mg•m2
d-SDS
/mM
Thickness /Å
±2
SLD
/10-6 Å-2
±0.02
SDS Volume
Fraction /10-6 Å-2
±0.02
Keratin Volume
Fraction /10-6 Å-2
±0.02
K1S Mass
/mg•m2
±0.05
Mass SDS
/mg•m2
±0.05
0.01 0 45 0.65 0 0.32 1.91 0
0 1 19 2.93 0.55 0 0 1.14
0.01 1 32 1.30 0.18 0.16 0.70 0.64
Figure 5.4 Neutron reflectivity as a function of momentum transfer measured for keratin/d-SDS complexes with d-
SDS concentrations of 0 ( ), 0.1(×), 0.3 ( ), 1 ( ), 2 ( ) and 5 mM (□) with a fixed keratin concentration of
0.01 mg/ml. NaCl at 5 mM was in the buffer throughout the experiments. Solid lines are the best uniform layer
model fits.
3E-7
3E-6
3E-5
3E-4
0.05 0.1 0.2
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
130
In order to further verify this behaviour, interfacial adsorption of keratin/d-SDS complexes over
a range of SDS concentrations was investigated as shown in Figure 5.4. It can be seen that the
reflectivity profiles are increased by increasing the d-SDS concentrations, which is a trend in
opposition to that seen in the case of the reflectivity profiles of keratin/h-SDS complexes.
Comparison of the data illustrated in Figures 5.2 and 5.4, clearly indicates that the addition of
SDS obstructed the adsorption of keratins at the interface. It can also be noted from the data in
Table 5.2 that SDS occupied more spaces with increasing concentrations up to 5 mM.
Table 5.2 Best mono-layer fits to neutron reflectivity profiles of keratin/d-SDS complexes and keratin/h-SDS
complexes. All measurements were performed with 0.5 M NaCl under NRW at pH 6.
Keratin Conc. of
SDS
/mM
Thickness
/Å
±2
SLD
/10-6 Å-2
±0.02
SDS
Volume
Fraction
±0.02
SDS
Mass
/mg•m2
±0.02
APM
SDS
/Å2
±5
Keratin
Volume
Fraction
±0.02
Mass
K1S
/mg•m2
±0.02
APM
K1S
/Å2
±20
d-SDS 0 45 0.65 0 0 -- 0.32 1.91 3932
0.1 37 0.86 0.04 0.16 320 0.31 1.59 4936
0.3 34 1.10 0.11 0.42 127 0.25 1.18 6661
1 32 1.30 0.18 0.64 82 0.16 0.70 11058
2 26 1.83 0.30 0.89 61 0.08 0.28 27221
5 22 3.15 0.58 1.40 37 0 0 --
h-SDS 0 45 0.65 -- -- -- 0.32 1.91
0.1 37 0.63 -- -- -- 0.31 1.59
0.3 34 0.5 -- -- -- 0.25 1.18
1 32 0.32 -- -- -- 0.16 0.70
131
Figure 5.5 Schematic representation of the keratin/SDS complexes and their structural changes at the air/liquid
interface with increasing SDS concentration.
Figure 5.5 is a schematic representation of the keratin/SDS complexes and their structural changes
at the air/liquid interface with increasing SDS concentration. The adsorbed amount of keratin is
reduced by increasing SDS concentration. At 1 mM SDS concentration, the layer thickness is
reduced from approximately 45 Å to approximately 32 Å. At the CMC concentration (5 mM), the
surface is fully occupied by SDS molecules and the keratin is squeezed out from surface.
132
5.3.2.3 Reflectivity profiles of keratin and h-DTAB complexes in the NRW subphase
Figure 5.6 Neutron reflectivity as a function of momentum transfer measured for keratin/h-DTAB complexes with
h-DTAB concentrations and a fixed keratin concentration of 0.01 mg/ml. NaCl at 5 mM was in the buffer
throughout the experiments. Solid lines are the best two-layer model fits.
Measurements of keratin/DTAB adsorption in NRW with the same buffers as with the
keratin/SDS complexes were performed at different DTAB concentrations, as shown in Figure
5.6. As described previously, reflectivity profiles are able to detail the amount of keratin
adsorption at the interface. It can be seen that the reflected NR signal increased and the reflectivity
curves became steeper with the presence of 0.1 mM DTAB. This indicates that more keratin was
adsorbed by its interactions with DTAB. However, unlike keratin/SDS adsorption, the reflectivity
profile reduced significantly with the further addition of 0.3 mM DTAB, which indicates that the
amount of surface complexes decreased by a very great amount. This observation is entirely
consistant with the lysozyme/SDS adsorption measured by Green et al1. Besides Green’s work,
there has been much published literature relating to interactions between oppositely charged
3E-7
3E-6
3E-5
3E-4
0.05 0.1 0.2
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
K1S 0.01mg/mlk1S 0.01 mg/ml+ h-DTAB 0.08 mMk1S 0.01 mg/ml+ h-DTAB 0.1 mMk1S 0.01 mg/ml+ h-DTAB 0.3 mMk1S 0.01 mg/ml+ h-DTAB 0.4 mMk1S 0.01 mg/ml+ h-DTAB 1 mM
133
complexes at air/solution interfaces. Penfold et al.27 demonstrated poly(ethyleneimine)/SDS
adsorption at air/solution interfaces with different strengths of electostatic interactions by
manipulating the pH. Staples et al.28 reported the interactions of SDS and a cationic polymer,
poly(dimethyldiallylammonium chloride) at air/water interfaces. The studies all provided
evidence for both electrostatic and hydrophobic forces and demonstrated the ‘enhancement’
followed by ‘restraint’ effects for interfacial adsorption with an increase of surfactant
concentrations. As in this case, the adsorbed keratin at the interface remained constant at
approximately 1.5 mg/m2 with the addition of 0.03 mM DTAB and increased to a maximum of
approximately 2.4 mg/m2 with the addition of 0.1 mM DTAB. The further addition of 0.3 mM
DTAB led to a sharp decline of keratin adsorption at approximately 1.2 mg/m2.
5.3.2.4 Reflectivity profiles of keratin and d-DTAB complexes in the NRW subphase
Figure 5.7 Neutron reflectivity as a function of momentum transfer measured for keratin/d-DTAB complexes with
d-DTAB concentrations from 0 to 5 mM and a fixed keratin concentration of 0.01 mg/ml. The NaCl concentration of
5 mM was used in the buffer throughout the experiments. Solid lines are the best two-layer model fits.
3E-7
3E-6
3E-5
3E-4
0.05 0.1 0.2
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
0K1s 0.01mg/ml+d-DTAB 0.08mMK1s 0.01mg/ml+d-DTAB 0.2mMK1s 0.01mg/ml+d-DTAB 0.7mMK1s 0.01mg/ml+d-DTAB 2mMK1s 0.01mg/ml+d-DTAB 8mM
134
A parallel study was performed under identical conditions as those in the keratin/d-DTAB
complexes, using the keratin/h-DTAB complexes, the results of which are shown in Figure 5.8.
By comparison of the adsorption of the keratin/h-DTAB complexes, the amounts of adsorbed
keratin and DTAB can be calculated. The calculations showed similar results in that the level of
the reflectivity profiles increased with the increase of DTAB concentrations and reached a
maximum with the addition of 0.1 mM DTAB. The reflectivity reduced by a significant amount
in the presence of 0.3 mM DTAB. Table 5.3 shows the fit parameters for a two-layer model. The
results reveal that the amount of adsorbed DTAB increased steadily from 0.2 mg/m2 to 0.5 mg/m2
with increasing DTAB concentrations.
Figure 5.8 Neutron reflectivity as a function of momentum transfer measured for keratin/h-DTAB complexes with
h-DTAB concentrations from 0 to 0.3 mM and a fixed keratin concentration of 0.01 mg/ml in D2O. A NaCl
concentration of 5 mM was used in the buffer throughout the experiments. Solid lines are the best two-layer model
fits to the data.
1E-06
1E-05
1E-04
1E-03
1E-02
0.04
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
K1S 0.01 mg/ml
K1S 0.01 mg/ml+0.2mM h-DTAB
K1S 0.01 mg/ml+0.4mM h-DTAB
135
Table 5.3 Best uniform layer fits to the neutron reflectivity profiles of the keratin/d-DTAB complexes and the
keratin/h-DTAB complexes in NRW, and three layer model fits of the keratin/h-DTAB complexes in D2O. All
measurements were performed with 0.5 M NaCl under NRW at pH 6. APM refers to area per molecule.
Keratin Conc. of
DTAB /mM
Number of
Layer
Thickness
/Å
±2
SLD
/10-6 Å-2
±0.02
Mass
DTAB
/mg•m2
±0.02
DTAB
Volume
Fraction
±0.01
APM
DTAB
/Å2
±5
Keratin
Volume
Fraction
±0.02
Mass K1S
/mg•m2
±0.02
APM
K1S
/Å2
±20
d-DTAB 0 1 45 0.65 -- -- 0.29 1.81 3932
(in NRW) 0.03 1 42 0.61 0 0 -- 0.29 1.69 4648
0.10 1 38 0.84 0.09 0.02 638 0.33 1.74 4515
0.20 1 44 0.91 0.29 0.06 184 0.28 1.71 4595
0.40 1 50 0.92 0.28 0.05 194 0.31 2.15 3652
0.70 1 50 0.95 0.31 0.05 194 0.29 2.01 3904
1.00 1 45 1.01 0.37 0.07 154 0.28 1.75 4493
2.00 1 40 1.10 0.65 0.14 87 0.12 0.67 11795
8.00 1 26 1.71 0.81 0.28 67 0.02 0.07 108880
h-DTAB 0 1 45 0.61 -- -- -- 0.29 1.81 --
(in NRW) 0.03 1 43 0.64 -- -- -- 0.31 1.85 --
0.1 1 41 0.67 -- -- -- 0.33 1.85 --
0.2 1 44 0.57 -- -- -- 0.28 1.69 --
0.4 1 50 0.63 -- -- -- 0.31 2.12 --
1 1 45 0.58 -- -- -- 0.28 1.76 --
2 1 40 0.25 -- -- -- 0.12 0.68 --
h-DTAB 0 1 12 1.4 -- -- -- 0.41 0.68 --
(in D2O) 2 12 5.6 -- -- -- 0.40 0.65 --
3 21 6.1 -- -- -- 0.13 0.37 --
total 45 -- -- -- -- -- 1.70 --
0.2 1 12 1.33 -- -- -- 0.41 0.68 --
2 12 4.55 0.25 0.20 -- 0.40 0.65 --
3 22 6.1 -- -- -- 0.13 0.37 --
0.4 1 12 1.33 -- -- -- 0.41 0.68
2 12 4.6 0.23 0.19 -- 0.40 0.65
3 22 6.1 -- -- -- 0.13 0.37
136
Figure 5.8 shows parallel experiments to those shown in Figures 5.6 and 5.7 using D2O as the
substrate. This provides additional information to explore the layer structure of the mixtures.
Figure 5.9 Schematic representation of the keratin/DTAB complexes and their structural changes at the air/liquid
interface with increasing SDS concentration.
Figure 5.9 is a schematic representation of the keratin/DTAB complexes and their structural
changes at the air/liquid interface with increasing SDS concentration. The results shown in Table
5.3 are thus represented in Figure 5.9. The keratin/DTAB complexes dominated the variation in
the DTAB concentration ranges from 0.1 mM to 2 mM. The area per molecule of DTAB varied
from 638 Å2 to 87 Å2 with DTAB concentrations from 0.1 mM to 2 mM. The area per molecule
of keratin changed little (5960-7560 Å2) with DTAB below 1 mM. At the CMC concentration (8
137
mM), the surface was fully occupied by DTAB molecules and the keratin squeezed out from
surface.
Figure 5.10 Variation of surface compositions of keratin/SDS complexes (top graph) and keratin/DTAB complexes
(bottom graph) as a function of surfactant concentrations. The red line indicates the volume fractions of keratin at the
interface; the blue line indicates the volume fractions of SDS (above) and DTAB (below) at the interface respectively.
The dashed line indicates the ratio of volume fractions of keratin/SDS (DTAB) (calculated from Table 5.2 and Table
5.3).
138
5.4 Conclusion
Figure 5.10 shows the variation in volume fraction of keratin and SDS as well as DTAB against
surfactant concentrations. The volume fraction of the keratin/SDS complexes is shown in Figure
5.10 (top) and that of keratin/DTAB complexes is shown in Figure 5.10 (bottom). The red lines
indicate volume fractions of keratin and the blue lines indicate volume fractions of SDS and
DTAB. The dashed lines indicate volume fraction ratios of keratin to surfactants. The adsorbed
amount of keratin/DTAB complexes oscillates from 0.1 mM DTAB to 0.7 mM DTAB, indicating
that the surface composition is determined by the surface activity of the complexes, and the
surface activity is determined by the addition of DTAB. The fluctuations of the surface
composition of DTAB represents structural variation of the complexes. The corresponding
measurements of surface tension shown in Figure 5.1 also reveal the regions in which the different
modes of interaction occurs.
It is shown in Figure 5.10 (above) that the surface volume fraction of keratin began to decrease
with the addition of 0.1 mM SDS. At 1 mM SDS, the volume fraction of keratin reduced from
0.31 to 0.16. At this point, the molar ratio of SDS to keratin was 135 and it indicated a strong
interaction of the complexes. The thickness of the keratin/SDS complexes also decreased from
approximately 45 Å to approximately 32 Å in the presence of 1 mM SDS, indicating a structural
deformation of the complexes. Further addition of SDS lead to further deformation of the
complexes and a denaturation of the keratin. As a result, the surface was fully occupied by SDS
with the addition of 5 mM SDS.
In the case of the keratin/DTAB complexes, it can be seen from Figure 5.10 that although the
addition of DTAB at approximately 0.1 mM produced an increase in the surface excess of keratin,
there was a decrease of DTAB adsorption, suggesting that the complexes had started to deform
139
on the surface. The complexes composition on the surface oscillated with increasing DTAB
concentration up until 0.7 mM, indicating that the activity of the adsorbed complexes at the
surface was changed by the addition of DTAB. The volume fraction ratio of keratin to DTAB is
shown in Figure 5.10 (bottom) with dashed lines to indicate the relative vibration in the surface
composition. In comparison to changes in the keratin/SDS complexes with increasing SDS
concentrations, the keratin/DTAB complexes indicated a charge neutralization point that may
have caused the oscillation of surface compositions in the region of approximately 0.2 mM DTAB.
At this point the protein has been neutralised by DTAB, suggesting a decrease in the surface
adsorption of the keratin. Further addition of DTAB lead to its full interaction with keratin
molecules and a deformation of the keratin structures at the surface. Since the thickness of the
adsorbed layers changed from approximately 45 Å to 26 Å, it can be assumed that the complexes
had a high density distribution over the surface. This phenomenon has also been described by Lu
et al29. The volume fraction ratios of keratin to DTAB in Figure 5.10 reduced quickly after the
addition of 0.7 mM DTAB, coincident with the charge neutralization point where the surface
excess of the complexes is not increasing with increasing DTAB concentration. The results
indicate that the combined electrostatic and hydrophobic interactions take place after the
neutralization point.
References
1. R. Green, T. Su, H. Joy and J. Lu, Langmuir, 2000, 16, 5797-5805. 2. F. W. Putnam, Advances in protein chemistry, 1948, 4, 79-122. 3. J. Steinhardt and J. A. Reynolds, Multiple equilibria in proteins, Academic Press, 1969. 4. P. Hansson and M. Almgren, Langmuir, 1994, 10, 2115-2124. 5. K. Thalberg, J. Van Stam, C. Lindblad, M. Almgren and B. Lindman, The Journal of Physical
Chemistry, 1991, 95, 8975-8982. 6. E. B. Abuin and J. Scaiano, Journal of the American Chemical Society, 1984, 106, 6274-6283. 7. Y. Li, J. Xia and P. L. Dubin, Macromolecules, 1994, 27, 7049-7055. 8. P. M. Claesson, M. Bergström, A. Dedinaite, M. Kjellin, J.-F. Legrand and I. Grillo, The Journal of
Physical Chemistry B, 2000, 104, 11689-11694.
140
9. S. Magdassi, Surface activity of proteins: chemical and physicochemical modifications, CRC Press, 1996.
10. C. Tanford, The Hydrophobic Effect: Formation of Micelles and Biological Membranes 2d Ed, J. Wiley., 1980.
11. C. Tanford, Journal of molecular biology, 1972, 67, 59-74. 12. J. A. Reynolds and C. Tanford, Journal of Biological Chemistry, 1970, 245, 5161-5165. 13. W. L. Mattice, J. M. Riser and D. S. Clark, Biochemistry, 1976, 15, 4264-4272. 14. D. Otzen, Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics, 2011, 1814, 562-591. 15. M. Schwuger, Journal of colloid and interface science, 1973, 43, 491-498. 16. M. L. Smith and N. Muller, Journal of Colloid and Interface Science, 1975, 52, 507-515. 17. M. Fishman and F. Eirich, The Journal of Physical Chemistry, 1971, 75, 3135-3140. 18. M. N. Jones, Biochimica et Biophysica Acta (BBA)-Protein Structure, 1977, 491, 121-128. 19. W. W. Fish, J. A. Reynolds and C. Tanford, Journal of Biological Chemistry, 1970, 245, 5166-
5168. 20. M. D. Lad, V. M. Ledger, B. Briggs, R. J. Green and R. A. Frazier, Langmuir, 2003, 19, 5098-5103. 21. C. Monteux, C. E. Williams, J. Meunier, O. Anthony and V. Bergeron, Langmuir, 2003, 20, 57-63. 22. P. M. Claesson, E. Blomberg, J. C. Fröberg, T. Nylander and T. Arnebrant, Advances in Colloid
and Interface Science, 1995, 57, 161-227. 23. Z. Li, J. Lu and R. Thomas, Langmuir, 1997, 13, 3681-3685. 24. J. R. Lu, S. Perumal, E. T. Powers, J. W. Kelly, J. R. Webster and J. Penfold, Journal of the
American Chemical Society, 2003, 125, 3751-3757. 25. J. Penfold and R. Thomas, Journal of Physics: Condensed Matter, 1990, 2, 1369. 26. J. R. Lu, S. Perumal, X. Zhao, F. Miano, V. Enea, R. R. Heenan and J. Penfold, Langmuir, 2005, 21,
3354-3361. 27. J. Penfold, I. Tucker, R. K. Thomas and J. Zhang, Langmuir, 2005, 21, 10061-10073. 28. E. Staples, I. Tucker, J. Penfold, N. Warren, R. K. Thomas and D. J. F. Taylor, Langmuir, 2002, 18,
5147-5153. 29. J. Lu, T. Su, R. Thomas, J. Penfold and R. Richards, Polymer, 1996, 37, 109-114.
Support Information
Table SI 5.1 List of protonated and deuterated surfactants with regard to neutron reflection and scattering parameters.
Formula Molecular
Volume (Å3)
Molecular
Weight
(g/Mol)
Scattering
Length Density
(10-6)
h-SDS CH3(CH2)11OSO3Na 474 288.4 0.34
d-SDS CD3(CD2)11OSO3Na 474 313.4 5.36
h-DTAB CH3(CH2)11N(CH3)3Br 485 308.3 -0.22
d-DTAB CD3(CD2)11N(CD3)3Br 485 342.3 6.09
141
Chapter 6
Binding of the cationic surfactant DTAB onto a coated keratin film
at the solid/water interface studied by SE, QCM-D and NR
The binding of the cationic surfactant dodecyltrimethylammonium bromide (DTAB) onto a
coated keratin film at the SiO2/water interface was investigated by three techniques: spectroscopic
ellipsometry (SE), quartz crystal microbalance with dissipation (QCM-D) and specular neutron
reflection (NR). The keratin film was produced by spin-coating from a keratin solution at 1 mg/ml
onto the hydrophilic SiO2 substrate. It was found that the film became compacted when dry in air,
but hydrated once immersed in water. The film under water was found to be well described by a
two-layer model, with an inner dense layer of 29 Å and an outer loose layer of 35 Å and was found
to be very stable. Subsequent binding to the keratin by DTAB at the solid/water interface was
investigated by comparisons before and after DTAB addition. SE demonstrated that the refractive
index of the entire layer increased with increasing DTAB concentration below 10 mM when the
thickness of the layer was kept constant, indicating penetration of the DTAB molecules into the
coated film. Further increase in DTAB concentration caused the disintegration of the film. QCM-
D revealed that the coated film had a high dissipation/frequency ratio, and indicated that the coated
layer was very soft and abundant in water. The dissipation/frequency ratio was decreased by the
addition of DTAB below CMC, indicating that most of the DTAB molecules penetrated the coated
film by squeezing water molecules out of the film. Above the CMC, addition of DTAB removed
most of the keratin, with the remaining materials forming a very soft and dilute layer with high
dissipation at the interface. Finally, NR revealed that three quarters of the DTAB molecules
penetrated into the coated film in the low DTAB concentration range and that the remaining
DTAB formed a very dilute layer on the outer surface of the coated film with a thickness of
approximately 30 Å. At DTAB concentrations above 10 mM, some of the keratin was removed
142
and the thickness of the coated layer started to decrease, accompanied by a significant structural
deformation of the keratin film. The work presented here shows that the formation of keratin films
can be used in the systematic study of the structural details related to DTAB binding.
6.1 Literature review
An overview of keratin and its interactions with surfactants is described in Chapters 4 and 5.
Among the techniques used in the study of interfacial properties, spectroscopic ellipsometry (SE)
can be used to study protein/surfactant interactions.1-3 However, SE can only determine an
approximate adsorption trend over a timescale of the order of 10 seconds and is unable to
distinguish the components and conformation of the adsorbed layers. Quartz crystal microbalance
with dissipation (QCM-D) has been used in recent years to identify the dynamic process of
protein/surfactant interactions on a time scale of the order of micro seconds, and also provides the
layer softness.4-7 However, in most cases QCM-D measures adsorbed layers with trapped water
inside and can only function with a flowing system. Specular neutron reflection (NR) is also
widely used for detection of the structure of interfacial layers for protein/surfactant interaction.8-
10 NR is efficient in investigating the structural conformation and composition of mixed
protein/surfactant layers by varying H/D labelling of surfactants, but lacks the ability to detect the
dynamics of adsorption. Therefore, there is no single experimental method that can provide high
resolution (generally under a few monometers), interfacial dynamics and interfacial
conformations. Furthermore, interaction between the two components often produces disordered
and inhomogeneous adsorbed layers.
The work presented here investigated the interaction of keratin and the cationic surfactant DTAB
at the solid/water interface using a combination of techniques: SE, QCM-D and NR. The
interaction of keratin and non-ionic C12E6 was also studied by SE in order to provide a comparison
with DTAB. The procedures used in the production of keratin solutions are described in Chapter
143
4. The keratin in this work was spin-coated onto the SiO2 surface from solution before its
interactions with surfactants are studied. The coated films were checked by SE and NR prior to
use. The coated film of protein is more reproducible and time-efficient than the pre-adsorption of
protein on a substrate described by previous studies, and provides an improved system for further
exploration with surfactants.
6.2 Experimental procedures
Spin-coating of keratin
The spin-coating process involved the dissolving of keratin into 5 mM NaCl buffer solution at 1
mg/ml. Silica wafers for SE were gently cleaned with 5.0% (v/v) Decon90, and subsequently
rinsed with UHQ water to ensure good hydrophilicity of the surfaces. The thickness of the oxide
layer of the silicon wafers was 14±2 Å as determined by SE. 30 μl of the solution was taken using
a pipette and dropped onto the silicon wafer, which had been fixed at the centre of the spin coater.
The spin coater was set to 3000 rmp for 20 seconds for each coating. The coated keratin films
were then dried for 10 minutes in air and annealed for 2 hours at 40˚C in vacuum in order to
remove the resident solvent. After the heating process, the coated wafers were stored in sealed
plastic bottles prior to later use.
144
6.3 Results and discussion
6.3.1 SE measurements
Figure 6.1 Phase changes of DTAB adsorption as a function of wavelength on the coated film of keratin in buffer
solutions at 20 minutes measured by SE: the coated film in air (circles), the coated film in water (triangles), DTAB
adsorption at 3 mM (squares) and DTAB adsorption at 14 mM (crosses). The blue symbols represent φ and the
orange symbols represent ∆. The solid lines are the best fits to the data.
Spectroscopic ellipsometry (SE) is a powerful interfacial technique, widely used in the
investigation of adsorption from biomolecules such as peptides and biosurfactants at interfaces.11-
19 Figure 6.1 shows the phase changes of φ and ∆ as measured by SE. Prior to each measurement,
the coated film was immersed in water for 30 minutes in order to allow the film to swell to its
equilibrium thickness. The process was followed by a SE measurement in order to set the standard
zero value, shown as triangles in Figure 6.1. Firstly, a two-layer model was used for DTAB
adsorption on the coated layer: the inner layer was assumed to be the coated film with a pre-fitted
refractive index and layer thickness and the outer layer was assumed to be the adsorbed layer of
110
120
130
140
150
160
170
0
10
20
30
40
300 400 500 600
∆ (
°)
φ(°
)
Wavelength /nm
Coated film in air
Coated film in water
3 mM DTAB
14 mM DTAB
145
DTAB. However, it was found that the fits were biased for measurements at higher concentrations
of DTAB. It was found that changing the refractive index of the inner layer improved the fits.
Therefore, a one-layer model was used, applying the assumption that most of the DTAB molecules
penetrated into the coated film and changed the refractive index of the film.
Table 6.1 The best fitted parameters of the SE data represented in Figure 6.2
DTAB conc.
/mM
Refractive
index
±0.002
Thickness /Å
±2
Total mass
/mg•m2
±0.02
DTAB mass
/mg•m2
±0.05
0 1.434 51 2.55 0
1 1.445 51 2.84 0.29
3 1.461 51 3.28 0.73
7 1.475 50 3.70 1.15
10 1.533 50 4.81 2.26
14 1.565 48 5.74 3.39
14 1.488 73 6.29 3.41
The solid lines in Figure 6.1 represent the best one-layer fits using the parameters listed in Table
6.1. It was found that a small addition of DTAB shifted ∆ downwards to a lower value, whilst the
overall shape of φ remained the same. It was found that a good fit could be obtained for both the
coated keratin films and the subsequent adsorbed layers of DTAB at low concentrations. At
concentrations of DTAB around the CMC, ∆ was further shifted downwards with its overall shape
less steep. It can be seen that the fits are biased with the addition of 10 mM DTAB, which indicates
a significantly structural change of the overall layer. The fit results indicate that both the one-layer
model and the two-layer model were unable to provide a fit to the data at higher concentrations
of DTAB, suggesting that DTAB damaged the homogeneity of the coated film. The fit parameters
listed in Table 6.1 were calculated after 20 minutes of adsorption. In the calculation of the
146
adsorbed mass of DTAB, the refractive index of water used was 1.344 at a wavelength of 400 nm.
Table 6.1 indicates that the small addition of DTAB changed only the refractive index of the
overall layer, and any change in the thickness was negligible, indicating that most of the DTAB
molecules remained inside the coated layer. As DTAB concentrations increased to around the
CMC, the fits deviated from the raw data, resulting in a large errors. However, the fitted data
suggests a high increase in the adsorbed amount of DTAB.
Figure 6.2 Dynamic adsorption of DTAB as a function of time at concentrations of 0.5 mM to 14 mM on the
coated film of keratin at the solid/water interface. All DTAB solutions were controlled at pH 5.
Figure 6.2 shows the time dependent adsorption of DTAB on the coated film of keratin over a
wide range of concentrations of DTAB at the SiO2/water interface. The results show that the
adsorbed amount of DTAB was dependent on both solution concentration and adsorption time.
Below the critical micelle concentration (CMC) of DTAB (11 ± 1 mM in 5 mM NaCl buffer as
shown in Support Information), the adsorbed amount of DTAB increased steadily with increasing
solution concentration and reached the maximum value around the CMC. It was also found that
0
1
2
3
0 5 10 15 20
Ad
sorb
ed a
mou
nt
(mg/m
2)
Time (min)
0.5 mM
1 mM
3 mM
7 mM
11 mM
14 mM
147
the fast adsorption in the first 2 minutes was followed by a slow and steady increase at around the
CMC concentration of DTAB. The results show that DTAB has a strong interaction with the
negatively charged film of keratin at high concentrations, forming surface complexes and causing
swelling of the film. At DTAB concentrations of above 20 mM (data not shown), the signals based
on changes in φ and became irregular, and the adsorbed amount of DTAB began to be poorly
reproducible, indicating that DTAB caused a deformation in and removed of the coated film.
6.3.2 QCM-D measurements
Figure 6.3 Frequency and dissipation shifts as a function of time for the dynamic adsorption of DTAB on coated
keratin films measured by QCM-D. The shifts of frequency (F) and dissipation (D) prior to 25 minutes are due to
the swelling of the coated film in water. The further shifts of F and D after 25 minutes correspond to the injection of
DTAB solution, followed by a water rinsing after 2 minute. (a): 3 mM DTAB; (b): 12 mM DTAB; (c): 20 mM
DTAB.
-0.7
0.3
-5
0
5
0 5 10 15 20
Dis
sip
ati
on
(e-
6)
Fre
qu
ency
(H
z)
Time (min)
F3
F5
F7
F9
F11
Injection of 3 mM DTAB
Rinsing
(a)
Remaining mass
after washing
148
-1
0
1
2
3
4
5
-30
-25
-20
-15
-10
-5
0
5
0 10 20 30 40 50
Dis
sip
ati
on
(e
-6)
Fre
qu
ency
(H
z)
Time (min)
F_3
F_5
F_7
F_9
F_11
D_3
D_5
D_7
D_9
D_11
-5
0
5
-15
-10
-5
0
5
10
15
0 5 10 15
Dis
sip
ati
on
(e
-6)
Fre
qu
ency
(H
z)
Time (min)
F3
F5
F7
F9
F11
(b)
Film swelling in UHQ UHQ rinsing DTAB injection
(c)
149
Quartz crystal microbalance with dissipation (QCM-D) is a complementary technique for
detecting the real time adsorption of proteins, peptides and surfactants with a time scale of micro
seconds upwards.7, 20-24 The measurement of DTAB adsorption followed the same procedure as
SE: the coated film was immersed in water for 20 minutes in order to be swelled to equilibrium
before each measurement. Unlike SE, QCM-D detects the total mass with water that is trapped
inside the coated layer.24
Figure 6.3(a) shows the adsorption of DTAB at 3 mM on the coated film of keratin. The rapid
adsorption occurred in the first 2 minutes, followed by a very small desorption during the
subsequent 15 minutes. The small desorption would be caused by the fluid flow that removed
some of the keratin molecules. The dissipation also dropped from about 0.3 to 0, indicating that
the adsorbed layer became more compacted with increasing time of adsorption. The rinsing
process was performed after the equilibrium of adsorption, indicating that some DTAB molecules
remained in the film after rinsing and that DTAB has strong interaction with keratins.
Figure 6.3(b) shows the adsorption of DTAB at around the CMC on the coated film. A 10 Hz shift
in frequency was seen at 25 minutes, due to the film swelling, with a shift in the dissipation of 2
×10-6. Theory states that the ratio of dissipation/frequency represents the rigidness of the adsorbed
layer. The higher this ratio is, the higher the softness (more diffuse) the layer.21, 25 In practice, the
layer is assumed to be soft if the ratio value is above 0.1 ×10-6 and solid if the ratio value is below
0.1. In this study, the ratio of D/F for the coated film reached 0.5×10-6, indicating that the coated
film formed a very soft layer after swelling in water. The large diversity for different harmonics
reveals that the layer was not homogeneous in the vertical direction. The results show that the
addition of DTAB shifted the frequency by approximately 15 Hz with a negligible shift of the
dissipation (about 0.5×10-6), indicating that the adsorption of DTAB did not change the softness
of the whole layer. This implies that most of the DTAB molecules penetrated into the coated film.
The rinsing process after the adsorption of DTAB is also illustrated in Figure 6.3(b). The
150
frequency returned to a higher value than its initial value after rinsing, suggesting that part of the
keratin/DTAB complexes was removed. The dissipation dropped to half of the initial value of
their shifts, indicating that the remaining part of the layer was more rigid. This result is consistent
with the NR data, which revealed a thinner but denser layer with the addition of DTAB.
Figure 6.3(c) shows the adsorption of DTAB above the CMC (20 mM) on the coated film. It
indicates significantly higher amounts of adsorption in the first 1 minute, followed by a rapid
desorption over the next 5 minutes and a slow desorption over the subsequent 15 minutes up until
equilibrium. Different harmonics were altered differently during the process of desorption but
only the 3rd harmonic lower than the initial value of adsorption. All dissipations remained at the
same level, indicating that the remaining film was highly soft, unstable and inhomogeneous.
Figure 6.4 Dynamic adsorption of DTAB as a function of time at concentrations of 3 mM to 40 mM on the keratin
coated film measured by QCM-D. All DTAB solutions were controlled at pH 5.
-0.5
1.5
3.5
0 5 10 15
Ad
sorb
ed a
mou
nt
(mg/m
2)
Time (min)
3 mM
6 mM
14 mM
20 mM
40 mM
151
Figure 6.4 shows the adsorbed amount of DTAB on the coated film at different concentrations
over time. It can be seen that at low concentrations below the CMC, there was a rapid adsorption
for the first 2 minutes, followed by a rapid equilibrium for the next 20 minutes. At the CMC of
DTAB, there was also a rapid adsorption process in the first 2 minutes, followed by a further
increase with increasing adsorbed amount from 1.5 mg/m2 to 2.1 mg/m2. This abnormal
phenomenon at the CMC is also shown for the DTAB adsorption at the SiO2/water interface,
which is shown in the support information as measured by DPI. For the adsorption of DTAB
above the CMC, the profiles for dynamic adsorption differ greatly to those at low concentrations
due to the deformation process of the coated film. The adsorbed amount of DTAB at 20 CMC
reached a value of 3.3 mg/m2 in 2 minutes, followed by a sharp decrease over the next 5 minutes
and a subsequent trend towards a plateau. Figure 6.3(c) also shows that the dissipation increased
dramatically to a value of 4.5 ×10-6 and stayed at the same level after rinsing, indicating that the
keratin/DTAB complexes formed a very dilute layer with high softness. At 40 mM of DTAB, the
profile of dynamic adsorption follows the same trend as at 20 mM but has lower adsorption in the
first 1 minute and less desorption afterward. However, the remaining adsorbed mass on the
substrate is higher than that upon the addition of 20 mM DTAB, suggesting that DTAB at the
CMC has the highest washing efficiency.
152
6.3.3 NR measurements
Figure 6.5(a) NR profiles as a function of momentum transfer for DTAB adsorption on the coated film of keratin at
DTAB concentrations of 0.32, 0.8 and 3.2 mM at the silicon/water interface.
Figure 6.5(b) NR profiles as a function of momentum transfer DTAB adsorption on the keratin coated film at
DTAB concentration of 7, 12, 20 and 40 mM at the silicon/water interface.
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
SiO2 in D2O
Coated film in D2O
DTAB 0.32 mM
DTAB 0.8 mM
DTAB 3.2 mM
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentrum transfer, Q /Å-1
SiO2 in D2O
Coated film in D2O
DTAB 7 mM
DTAB 12 mM at pH 5
DTAB 20 mM
DTAB 40 mM
153
Neutron reflection is a powerful technique for looking at the structural conformation of adsorbed
layers with high sensitivity and resolution26-29. The technique was used in the work presented here
to investigate the keratin/DTAB interaction. Figure 6.5(a) shows the reflectivity profiles for
DTAB adsorption on the keratin coated films at different DTAB concentrations. The solid lines
on the figure are the best fits for the data. The reflectivity profiles of the bare SiO2 layer and the
SiO2/coated film layer are also included on Figure 6.5 in order to allow comparison between them
and the profiles of DTAB adsorption.
Figure 6.5(b) shows the reflectivity profiles of DTAB adsorption onto the keratin coated film in
D2O at comparatively high concentrations of DTAB. It was found that the reflectivity profiles
follow the same trend at low DTAB concentrations until beyond the CMC. At 20 mM, the profile
shows a bump at a Q value of approximately 0.14 Å-1, indicating a significant structural change
in the overall adsorbed layer. The profile for the addition of 40 mM DTAB shows the same shape
of bump but at a higher reflectivity level, suggesting that the keratin/DTAB formed more elaborate
complexes.
Table 6.2 The best fitted parameters of the neutron reflectivity data for the keratin coated film as represented by the
purple circles in Figure 6.5(a).
Coated layer of
keratin
thickness /Å
±2
SLD
/e-6 Å-2
±0.02
Volume
Fraction
±0.02
Mass
mg•m2
±0.03
SiO2 15 3.41 -- --
1 29 5.19 0.401 1.63
2 35 5.86 0.166 0.81
sum 64 2.44
Table 6.2 gives data for the best fit parameters of the coated film after it had swelled to equilibrium
at the silicon/water interface. The results reveal that the coated film formed two layers on the
154
substrate with a dense layer attached to the subphase and a top loose layer exposed to water. The
inner layer had a thickness of 29±2 Å with a volume fraction of 0.37, whilst the outer layer had
a thickness of 35±2 Å with a volume fraction of 0.17. The overall mass of the coated layer was
2.31±0.2 mg/m2, which is consistent with the SE results within experimental error (2.55 ± 0.2
mg/m2).
Table 6.3 The best fit parameters of the neutron reflectivity data as shown in Figure 6.5(a) upon adsorption of
DTAB onto the keratin film over the low DTAB concentration range.
DTAB
Conc.
/mM
Layer thickness
/Å
±2
SLD
/10-6 Å-2
±0.02
keratin
Volume
Fraction
±0.003
DTAB
Volume
Fraction
±0.001
Keratin
mass
/mg•m2
±0.02
DTAB mass
/mg•m2
±0.02
0.32 1 29 4.98 0.401 0.04 1.63 0.12
2 35 5.65 0.166 0.02 0.81 0.08
3 31 6.19 0 0.02 0 0.07
Sum 95 2.44 0.27
0.80 1 29 4.78 0.401 0.07 1.63 0.21
2 35 5.67 0.166 0.02 0.81 0.08
3 35 6.14 0 0.03 0 0.09
Sum 99 2.44 0.38
3.2 1 29 4.46 0.401 0.12 1.63 0.37
2 35 5.61 0.166 0.03 0.81 0.11
3 32 5.99 0 0.05 0 0.17
Sum 96 2.44 0.65
Table 6.3 gives the best fitted parameters for the reflectivity profiles of DTAB adsorption as
shown in Figure 6.5(a). The adsorbed amount of DTAB increased from 0.27 mg/m2 to 0.65 mg/m2
with increasing concentrations of DTAB. Among the adsorbed DTAB molecules, over half of the
molecules interacted with keratins in the inner layer of the film whilst a quarter of the molecules
155
interacted with keratins in the outer layer of the film. A small number of the DTAB molecules
formed a third dilute layer on the outer surface of the coated film. The keratin film remained
unchanged with the addition of DTAB at low concentrations.
Table 6.4 The best fitted parameters of the neutron reflectivity data as shown in Figure 6.5(b) for DTAB adsorption
onto keratin films over the high DTAB concentration range.
DTAB
Conc.
/mM
Layer thickness /Å
±2
SLD
/10-6 Å-2
±0.02
keratin
Volume
Fraction
±0.003
DTAB
Volume
Fraction
±0.001
Keratin
mass
/mg•m2
±0.02
DTAB
mass
/mg•m2
±0.02
10 1 29 3.84 0.305 0.29 1.24 0.89
2 35 5.65 0.09 0.07 0.44 0.26
3 35 5.99 0 0.05 0 0.18
Sum 99 1.68 1.33
20 1 21 3.54 0.18 0.35 0.53 0.77
2 25 5.28 0.08 0.13 0.28 0.34
3 57 5.91 0 0.07 0 0.42
Sum 103 0.81 1.53
40 1 25 2.47 0.09 0.55 0.29 1.44
2 18 5.57 0.04 0.10 0.11 0.19
3 58 5.86 0 0.07 0 0.43
Sum 101 0.4 2.06
Table 6.3 gives the best fitted parameters of the measured data as shown in Figure 6.5(b). The
coated layer started to become deformed from 2.31 mg/m2 to 1.91 mg/m2 with the addition of 10
mM DTAB (~CMC). The mass of the coated film was further decreased to 0.4 mg/m2 by the
further addition of DTAB. The adsorbed mass of DTAB increased from 1.25 mg/m2 to 2.06 mg/m2
with increasing concentrations of DTAB. The molar ratio of DTAB/keratin changed from 18.4 to
788 as DTAB concentrations increased from 0.32 mM to 40 mM, suggesting that DTAB
156
molecules were gradually dominant in keratin/DTAB complexes as well as in the total surface
coverage.
Figure 6.6 NR profiles of DTAB adsorption as a function of momentum transfer onto the coated film of keratin at
10 mM DTAB concentration for four different contrasts: h-DTAB in D2O (circulars, black); d-DTAB in D2O
(diamonds, cyan); d-DTAB in H2O/D2O mixture (ρ=3.41 × 10-6 Å-2, crosses, green) and d-DTAB in H2O (squares,
orange).
Figure 6.6 shows the adsorption of 10 mM DTAB on the coated film with four different contrasts:
h-DTAB in D2O; d-DTAB in D2O; d-DTAB in H2O/D2O mixture (ρ=3.41 × 10-6 Å-2) and d-DTAB
in H2O. The solid lines are the best fits the three-layer model. Results from data analysis indicate
that the interfacial area consisted of a dense inner layer of 29±2 Å with approximately 30%
keratin and 26% DTAB attached to the SiO2 substrate, a less dense layer in the middle of 35±2
Å with approximately 9% of keratin and 7% of DTAB, and a very dilute layer of approximately
35±2 Å containing 5% of DTAB exposed to the solution side.
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q /Å-1
157
Table 6.5 The best fitted parameters of the measured data as shown in Figure 6.6 for d- and h-DTAB adsorption at
10 mM under different water contrasts onto the keratin film.
DTAB
10 mM
Layer thickness
/Å
±2
SLD
/10-6 Å-2
±0.02
keratin
Volume
Fraction
±0.003
DTAB
Volume
Fraction
±0.001
Keratin
mass
/mg•m2
±0.02
DTAB
mass
/mg•m2
±0.02
h-DTAB in
D2O
1 29 3.84 0.305 0.29 1.24 0.89
2 35 5.65 0.09 0.07 0.44 0.26
3 35 5.99 0 0.05 0 0.18
Sum 99 1.68 1.33
d-DTAB in
D2O
1 29 4.89 0.305 0.28 1.24 0.85
2 34 5.80 0.09 0.03 0.44 0.11
3 34 6.29 0 0.05 0 0.18
Sum 98 1.68 1.14
d-DTAB in
CMSi
1 29 3.70 0.305 0.27 1.24 0.82
2 36 3.39 0.09 0.03 0.44 0.11
3 34 3.36 0 0.05 0 0.18
Sum 98 1.68 1.10
d-DTAB in
H2O
1 29 1.82 0.305 0.26 1.24 0.79
2 35 0.02 0.09 0.03 0.44 0.11
3 33 -0.10 0 0.05 0 0.18
Sum 97 1.68 1.08
158
Figure 6.7 Comparison of the adsorbed amount of DTAB on the coated film measured by SE (orange diamonds),
QCM-D (green squares) and NR (blue triangles). The black squares correspond to the secondary y-axis and
represent the remaining keratin of the coated film measured by NR. The two sets of points of QCM-D at the
presence of 20 and 40 mM of DTAB represented the adsorption process at 1 minute and 20 minutes respectively.
Figure 6.7 shows the adsorbed amount of DTAB measured by SE, QCM-D and NR. In general,
the data measured by QCM-D are much higher than from other techniques, since QCM-D
accounts for water molecules incorporated in the measured layers.30-32 Macakova et al.24 studied
the trapped water in the DTAB and its adsorption on a silica substrate through comparison of
QCM-D and SE measurements, and found that 30%-60% water was trapped inside the adsorbed
layer under general conditions. However, the results from the three techniques used in the work
presented here are consistent below 9 mM DTAB. This consistency indicates that the adsorption
of DTAB below 9 mM did not cause more water to be incorporated into the coated layers. It can
therefore be stated that most of the DTAB molecules became trapped into the coated layer by
squeezing water molecules out of the film. At around the CMC of DTAB, the adsorbed mass
measured by QCM-D (1 minute after adsorption) was much higher than SE and NR data,
0
1
2
3
0
1
2
3
0.25 2.5 25
Ker
ati
n r
ema
ined
(m
g/m
2)
Ad
sorb
ed a
mo
un
t o
f D
TA
B (
mg
/m2)
Concentration of DTAB (mM)
NR
SE
QCM-D
Keratin remained
159
indicating that at and above this point, increasing concentrations of DTAB sharply swelled the
coated film of keratin and trapped more water in the measured layers. This result is consistent
with the SE measurement as shown in Figure 6.1, where the coated layer began to be dissolved
by the addition of 10-14 mM DTAB. Figure 6.7 indicates that the remaining amount of keratin in
the film began to decline above the CMC. The remaining mass of keratin in the surface complexes
with the addition of 20 mM DTAB was 1.0 ±0.1 mg/m2, reducing to 0.5 ±0.1 mg/m2 after the
rinsing process (data not shown).
Figure 6.7 also shows that whilst the amount of keratin remaining on the surface reduced rapidly,
the adsorbed DTAB increased sharply above the CMC, indicating that the keratin/DTAB
complexes consisting of more DTAB molecules were more likely to be bound on the SiO2
substrate. This observation is interesting from a practical application viewpoint. It implies that
cationic agents could be concentrated onto a given surface when and if an anionic keratin is pre-
deposited.
6.4 Conclusion
The interfacial interaction and deformation of the wool keratin with surfactant DTAB at the
SiO2/water interface was investigated. Figure 6.8 is a schematic representation of DTAB
adsorption on the keratin coated film consisting of three steps. The coated keratin firstly interacted
with DTAB at low concentrations. The film is then gradually removed with increasing
concentrations of DTAB. Both SE and QCM-D showed a fast initial adsorption of DTAB in the
first 2 minutes, followed by a plateau in the next 20 minutes at DTAB concentrations below 10
mM. At around the CMC of DTAB, SE showed an inhomogeneous film at the interface while
QCM-D revealed a rapid adsorption, followed by a steady and slow increase. The results also
showed that the rinsing process after adsorption removed part of the keratin/DTAB complexes at
the interface. At DTAB concentrations above the CMC, QCM-D revealed a sharp increase of
160
adsorption in the first 2 minutes, followed by a rapid desorption of the coated film in the next 20
minutes. The high dissipation of the coated film indicate a very soft layer, whilst the negligible
change of dissipation of DTAB molecules indicates that most DTAB molecules penetrated into
the coated layer but had limited changes to the layer structure.
Figure 6.8 Schematic representation of DTAB adsorption on the coated film of keratin with increasing DTAB
concentration.
Analysis of the layer structure from NR indicated an inhomogeneous film in the vertical direction,
consisting of a three-layer structure of DTAB adsorption with a dense layer attached to the SiO2
containing keratin/DTAB complexes, a middle layer with medium density and an outer loose layer
containing a small fraction of DTAB. QCM-D also indicated an inhomogeneous film by revealing
a vast diversity of frequency and dissipation shifts with different harmonics. NR also revealed the
desorption process of the coated film with the addition of 10 mM DTAB with negligible keratin
161
remaining with a further addition of 40 mM DTAB, which supports the desorption process
measured by QCM-D.
Finally, there was no adsorption of C12E6 on the coated film of keratin (shown in support
information in Figure SI 6.3), indicating that the main driving force of the keratin/DTAB
interaction is electrostatic.
References
1. H. Elwing, A. Askendal and I. Lundström, Journal of colloid and interface science, 1989, 128, 296-300.
2. M. Malmsten, Journal of colloid and interface science, 1998, 207, 186-199. 3. M. C. Wahlgren and T. Arnebrant, Journal of colloid and interface science, 1991, 142, 503-511. 4. C. Ayela, F. Roquet, L. Valera, C. Granier, L. Nicu and M. Pugnière, Biosensors and Bioelectronics,
2007, 22, 3113-3119. 5. R. Bordes, J. R. Tropsch and K. Holmberg, Langmuir, 2010, 26, 10935-10942. 6. R. P. Richter and A. R. Brisson, Biophysical Journal, 2005, 88, 3422-3433. 7. A. Dolatshahi-Pirouz, K. Rechendorff, M. B. Hovgaard, M. Foss, J. Chevallier and F. Besenbacher,
Colloids and Surfaces B: Biointerfaces, 2008, 66, 53-59. 8. D. Horne, P. Atkinson, E. Dickinson, V. Pinfield and R. Richardson, International dairy journal,
1998, 8, 73-77. 9. J. Lu, T. Su and R. Thomas, The Journal of Physical Chemistry B, 1998, 102, 10307-10315. 10. J. Lu, T. Su, R. Thomas and J. Penfold, Langmuir, 1998, 14, 6261-6268. 11. X. Zhao, Z. Zhang, F. Pan, Y. Ma, S. P. Armes, A. L. Lewis and J. R. Lu, Langmuir, 2005, 21, 9597-
9603. 12. X. Zhao, F. Pan, P. Coffey and J. R. Lu, Langmuir, 2008, 24, 13556-13564. 13. X. Zhao, F. Pan, S. Perumal, H. Xu, J. R. Lu and J. R. Webster, Soft Matter, 2009, 5, 1630-1638. 14. X. Zhao, F. Pan, B. Cowsill, J. R. Lu, L. Garcia-Gancedo, A. J. Flewitt, G. M. Ashley and J. Luo,
Langmuir, 2011, 27, 7654-7662. 15. S. R. Goates, D. A. Schofield and C. D. Bain, Langmuir, 1999, 15, 1400-1409. 16. J. L. Keddie, Current opinion in colloid & interface science, 2001, 6, 102-110. 17. C. B. Walsh, X. Wen and E. I. Franses, Journal of colloid and interface science, 2001, 233, 295-
305. 18. E. Bellet-Amalric, D. Blaudez, B. Desbat, F. Graner, F. Gauthier and A. Renault, Biochimica et
Biophysica Acta (BBA)-Biomembranes, 2000, 1467, 131-143. 19. X. Zhao, F. Pan, S. Perumal, H. Xu, J. R. Lu and J. R. P. Webster, Soft Matter, 2009, 5, 1630-1638. 20. J. Malmström, H. Agheli, P. Kingshott and D. S. Sutherland, Langmuir, 2007, 23, 9760-9768. 21. A. A. Feiler, A. Sahlholm, T. Sandberg and K. D. Caldwell, Journal of colloid and interface science,
2007, 315, 475-481. 22. P. Roach, D. Farrar and C. C. Perry, Journal of the American Chemical Society, 2005, 127, 8168-
8173. 23. A. Naderi and P. M. Claesson, Langmuir, 2006, 22, 7639-7645. 24. L. Macakova, E. Blomberg and P. M. Claesson, Langmuir, 2007, 23, 12436-12444.
162
25. Q. Chen, W. Tang, D. Wang, X. Wu, N. Li and F. Liu, Biosensors and Bioelectronics, 2010, 26, 575-579.
26. J. R. Lu, A. Marrocco, T. J. Su, R. K. Thomas and J. Penfold, Journal of colloid and interface science, 1993, 158, 303-316.
27. J. Lu, E. Lee, R. Thomas, J. Penfold and S. Flitsch, Langmuir, 1993, 9, 1352-1360. 28. T. J. Su, J. R. Lu, R. K. Thomas, Z. F. Cui and J. Penfold, Journal of Colloid and Interface Science,
1998, 203, 419-429. 29. J. Lu, T. Su, R. Thomas, J. Penfold and R. Richards, Polymer, 1996, 37, 109-114. 30. F. Höök, J. Vörös, M. Rodahl, R. Kurrat, P. Böni, J. Ramsden, M. Textor, N. Spencer, P. Tengvall
and J. Gold, Colloids and Surfaces B: Biointerfaces, 2002, 24, 155-170. 31. J. Stålgren, J. Eriksson and K. Boschkova, Journal of colloid and interface science, 2002, 253, 190-
195. 32. G. Cheng, Z. Liu, J. K. Murton, M. Jablin, M. Dubey, J. Majewski, C. Halbert, J. Browning, J.
Ankner and B. Akgun, Biomacromolecules, 2011, 12, 2216-2224.
Support information
Figure SI 6.1 Surface tension as a function of DTAB concentration in 5 mM NaCl buffer to give the CMC just
above 10 mM.
30
40
50
60
0.1 1 10
Avg.
SF
T [
mN
/m]
Concentration [mM]
163
Figure SI 6.2 Dynamic adsorption of DTAB on the SiO2 substrate in pH 5 buffer solution measured by DPI. The
adsorbed amount is shown as a function of time.
Figure SI 6.3 Dynamic adsorption of C12E6 on the keratin coated film in pH 5 buffer solution as measured by SE,
showing very little adsorption. The adsorbed amount is shown as a function of time.
0
0.5
1
1.5
-10 40 90 140 190 240 290
Ad
sob
ed a
mo
un
t (m
g/m
2)
Time (min)
3.5 mM
7 mM
11 mM
14 mM
28 mM
56 mM
140 mM
164
Film stability check by SE
Figure SI 6.4 Thicknesses of spin-coated keratin films as a function of the concentration of keratin solution. The
thickness was measured by ellipsometry with a film reflective index of 1.458. The red dots represent film
thicknesses before they were heated in the vacuum oven; the black dots represent film thicknesses after they were
heated in the vacuum oven at 60˚C for 1 hour.
Figure SI 6.4 shows the thicknesses of spin-coated films against the concentration of the solution
measured at solid/air interface by SE. Each point was repeated 5 times to ensure that the film was
reproducible. The data indicate that the coated films underwent slight shrinkage after they were
heated, due to removal of the solvent. The measurements also indicate a linear relation between
the film thickness and concentration, which in some respects illustrate that the film was
homogeneous.
0
50
100
150
0 0.5 1 1.5 2 2.5 3
Fil
m T
hic
kn
ess
(Å
)
Concentration (mg/ml)
165
Film stability check by QCM-D
Figure SI 6.5 Film stability test for different pH buffers (pH 3 buffer (orange), pH 4 buffer (red), pH 7 of UHQ
water (blue), pH 9 buffer (green), pH 10 buffer (purple), pH 11 buffer (cyan). The keratin solution used was at 1.2
mg/ml. The film thickness is shown as a function of time.
Figure SI 6.5 shows checks in the stability of spin-coated keratin films in different pH buffers at
the solid/water interface. Since the buffer may destroy film uniformity and flatness, SE was
restricted and did not provide reliable measurements. The coated films were measured by QCM-
D, with an initial layer condition set to zero before each measurement. It was therefore very
sensitive in detecting the changes of each layer. It was found that the coated films stayed stable
below pH 7 and started to dissolve at pH 9. The dissolving process was characterised by a rapid
initial process within the first 25 minutes, followed by a slower process. At pH 11, most of the
keratin film was dissolved within the first 30 min.
25
35
45
55
0 10 20 30 40 50 60 70 80
Th
ick
nes
s (Å
)
Time (min)
uHQ
pH 3
pH 4
pH 9
pH 10
pH 11
166
Chapter 7
The interfacial interaction of keratin and rhamnolipids at the
solid/liquid interface as studied by NR, QCM-D, SE and DPI
In the investigations presented here, the spin-coating method described in Chapter 6 was used
with different concentrations of keratin solutions to form isotropic films of keratin on SiO2. In
addition to the checks of SE and QCM-D for the coated films described in Chapter 6, NR was
also used in order to determine the film’s conformational structure both in dried air and in water.
NR exhibited the high stability and reproducibility of the coated film in both dried air and in water.
The fits from NR also revealed that the coated film formed a three-layer structure with an inner
dense layer, a middle layer with medium density and an outer dilute layer.
Biosurfactants have gained increased attention in recent years due to their multi-functional
properties and environmental advantages as compared to synthetic surfactants. Chemical
synthesis of biosurfactants in recent years has become more efficient and practical as a result of
analysis of their structural information. The most important quality of biosurfactants is their
environmental compatibility, they are readily biodegradable and less toxic than conventional
surfactants.
Adsorption of biosurfactant rhamnolipids on the coated film of keratin was investigated by four
techniques: QCM-D, SE, DPI and NR. Both QCM-D and SE revealed that the dynamic adsorption
of rhamnolipid 1 and 2 was identified with a rapid process at the start of the first 20 minutes, and
was followed by a slower adsorption over the next 50 minutes accompanied by molecular
rearrangement. Both QCM-D and NR showed that rhamnolipid 1 was able to be completely
washed off by water after adsorption whereas only 70% of rhamnolipid 2 was able to be removed
after adsorption, indicating that rhamnolipid 2 has a stronger interaction with keratin than
167
rhamnolipid 1. QCM-D also showed that the adsorption of rhamnolipid 2 produced a high
dissipation of the film whereas the adsorption of rhamnolipid 1 produced a negligible increase of
dissipation, indicating that the addition of rhamnolipid 2 changed (swelled) the structure of the
coated film. By comparison of the SE and QCM-D data, the percentage of trapped water in the
coated film was determined, and the results are consistent with NR measurements. DPI provides
another means of investigation and demonstrated that the coated film was composed of a dense
part and a dilute part, and that the adsorption of rhamnolipid 1 was reversible.
7.1 Experimental procedures
7.1.1 Mass spectroscopy of rhamnolipids 1 and 2
Figure 7.1(a) Mass spectroscopy of rhamnolipid 1 consisting of one head. The units of axes X and Y are m/z
(mass-to-charge ratio) and relative abundance (%), respectively.
168
Figure 7.1(b) Segmental analysis of rhamnolipid 1 from mass spectroscopy. The red numbers refer to molecular
masses of segments.
Figure 7.1(a) shows the results of mass spectroscopy of rhamnolipid 1 that consists of one head
group. The red numbers in brackets in Figure 7.1(a) are calculated from segments of rhamnolipid
1. Figure 7.1(b) shows the molecular weight analysis of different segments and 532g/mol for
rhamnolipid 1 was obtained.
Figure 7.2(a) Mass spectroscopy of rhamnolipid 2 consisting of two heads. The units of axes X and Y are m/z
(mass-to-charge ratio) and relative abundance (%), respectively.
169
Figure 7.2(b) Segmental analysis of rhamnolipid 2 from mass spectroscopy. The red numbers refer to molecular
masses of segments.
Figure 7.2(a) shows the results of mass spectroscopy of rhamnolipid 2 consisting of two head
groups. The red numbers in brackets in Figure 7.2(a) were calculated from segments of
rhamnolipid 2. Figure 7.2(b) shows molecular weight analysis of different segments and 678
g/mol for rhamnolipid 2 was obtained.
7.2 Results and discussion
7.2.1 Film stability and reproducibility explored with NR
The film stability was checked with SE and QCM-D as described in the support information in
Chapter 6. The 1 mg/ml keratin solution was used in Chapter 6 for the coating process. In the
work presented here, 1.2 mg/ml keratin solution was used to form a thicker film in order to enable
better examination of how rhamnolipids remove keratins. Figure 7.3 shows the NR profiles of the
coated film in different contrast buffers. The black line is a reference measurement for the bare
SiO2 layer in D2O; the blue line represents the coated film on the SiO2 layer in D2O and the red
line is the coated film in H2O. The H/D exchange in keratin molecules in different buffer solutions
(D2O and H2O) would alter its scattering length density and thus affect the reflectivity profiles of
keratin. The changing mechanisms were described in section 4.2.1.
170
Figure 7.3 Reflectivity profiles as a function of momentum transfer of the keratin coated film measured by NR at
the solid/water interface (bare SiO2 in D2O (black), keratin film in D2O (blue), and keratin film in H2O (red)).
Table 7.1 gives the fit parameters from Figure 7.3. In this Q range, the reflectivity difference
between the blue and black lines is determined by the thickness and SLD of the coated film. The
single layer model with a homogeneous composition is the most convenient way to analyse the
reflectivity data. However, both the single-layer model and the two-layer model were unable to
provide good fits in this case. Therefore a three-layer model was used comprising a dense layer
attached to the SiO2 substrate, a dilute layer close to the bulk and a middle layer of medium
density. The scattering length densities of the keratin was 1.97 × 10-6 Å-2 in H2O, 2.07 × 10-6 Å-2
in NRW and 3.45 × 10-6 Å-2 in D2O.
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q/Å-1
171
Table 7.1(a) Fit parameters from the analysis of neutron reflectivity data for the coated film of keratin in D2O.
Layer Thickness/Å
±2
SLD/×10-6 Å-2
±0.02
Volume
Fraction
±0.02
Mass /mg•m2
±0.02
1 19 3.41 -- --
2 15 4.7 0.52 1.08
3 34 5.5 0.29 1.37
4 40 6 0.12 0.67
sum 49 3.12
Table 7.1(b) Fit parameters from the analysis of neutron reflectivity data for the coated film of keratin in H2O.
Layer Thickness/Å
±2
SLD/×10-6 Å-2
±0.02
Volume Fraction
±0.02
Mass /mg•m2
±0.02
1 15 3.41 -- --
2 15 1.4 0.55 1.14
3 35 0.55 0.28 1.36
4 40 0.24 0.11 0.62
sum 50 3.11
Table 7.1 shows that the film thickness of the densely inner layer and the middle layer was 50±2
Å while the outer layer had a thickness of approximately 45±3 Å with a volume fraction of
0.11±0.02. By calculating the film thicknesses and volume fractions of the fitted layers, the total
coated mass was obtained as 3.1±0.2 mg/m2. In comparison, the film thickness measured by
ellipsometry was 48±3 Å with a coated mass of 3.2±0.3 mg/m2.
172
Figure 7.4 Reflectivity profiles as a function of momentum transfer of the keratin coated film: the dried film
(protonated) at the air/solid interface (diamonds); the wet film (deuterated) in D2O at the solid/water interface
(triangles); the dried film (deuterated) at the air/solid interface (circles).
In order to detect the structure of the coated film in dried air and to check whether the buffer
solution is able to change the film structure, three steps were performed. Firstly, the coated film
was heated in the vacuum oven at 60˚C for an hour and subsequently measured in the sealed
neutron cell. Since the volume of the cell is very small, the cell can protect the coated film from
damp. This step was measured at the solid/air interface and is presented by blue diamonds in
Figure 7.4.
Secondly, the film was immersed into D2O for an hour and subsequently measured in D2O buffer
solution. Measurement was at the solid/water interface and is represented by red triangles in
Figure 7.4. In this step, the hydrogens in keratins were partly replaced by deuterium. Thus the
SLD of keratin was changed from 1.91×10-6 Å-2 to 3.4×10-6 Å-2.
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q/Å-1
173
In the final step, the film was dried and heated by vacuum oven for a duration of 2 hours at 60 °C
after the immersion in D2O described in step 2. NR measurements were carried out in the sealed
neutron reflection cell, the results of which are represented as black circles in figure 7.4. The
difference in the two reflectivity profiles (blue diamonds and black circles) is not only affected
by structural changes in the film, but also by H/D exchanges of the keratin molecules.
Table 7.2(a) Protonated keratin film fits at air/solid interface.
Layer Thickness/Å
±2
SLD/×10-6 Å-2
±0.02
Volume
Fraction
±0.02
Mass /mg•m2
±0.02
1 15 3.41 SiO2 SiO2
2 27 1.38 0.58 1.18
3 25 0.45 0.27 1.27
4 39 0.08 0.07 0.57
sum 91 3.02
Table 7.2(b) Deuterated keratin film fits at air/solid interface.
Layer Thickness/Å
±2
SLD/×10-6 Å-2
±0.02
Volume
Fraction
±0.02
Mass /mg•m2
±0.02
1 15 3.41 SiO2 SiO2
2 29 1.69 0.49 1.34
3 29 0.61 0.24 1.20
4 35 0.11 0.08 0.52
sum 93 3.05
174
Table 7.2 gives the fit parameters of Figure 7.4 at the solid/air interface. The results indicate that
the dense inner layer of the coated film became slightly denser after reheating and that the overall
thickness of the film had a 10 % increase with a sparser outer layer after the immersion. However,
there was a negligible change of the overall mass of the film, indicating that the coated film was
stable and reproducible.
7.2.2 Adsorption of rhamnolipids on the coated film of keratin
7.2.2.1 QCM-D measurements
Figure 7.5 Frequency and dissipation changes as a function of time for rhamnolipid 1 adsorption on the coated film
measured by QCM-D: frequency (blue); dissipation (red). The four valleys represent the injections of rhamnolipid 1
solutions at 0.01, 0.03, 0.05 and 0.1 mg/ml.
QCM-D monitors the frequency and energy dissipation response of the vibrating sensor.1 The
frequency of the vibrating sensor measures the total oscillating mass adsorbed on the surface,
including the water coupled inside the layer.2 Figure 7.5 shows the frequency and dissipation
shifts for the adsorption of rhamnolipid 1 on the coated film. The four valleys represent the
injections of rhamnolipid 1 solutions at 0.01, 0.03, 0.05 and 0.1 mg/ml. It was found that the
dissipation remained at around zero whereas the frequency increased with increasing
concentrations of rhamnolipid 1 and reached a maximum at 0.1 mg/ml. For samples with high
-3
-2
-1
0
1
2
3
-40
-30
-20
-10
0
10
0 100 200 300 400 500
Dis
sip
ati
on
(10
-6)
Fre
qu
ency
(H
z)
Time (min)
175
dissipations it proves necessary to use the more complicated Voigt model for viscoelastic layers.
In general, the Sauerbrey model is used when Hz-1 where D and F represent
dissipation and frequency of the layer, respectively. In this work, there was negligible change of
dissipation but a significant change of frequency, implying that the addition of rhamnolipid 1 did
not change the softness of the coated film. Therefore, it could be the case that rhamnolipid 1 either
formed an extremely rigid layer on the dilute film of keratin or most of the rhamnolipid 1
molecules penetrated into the coated film by squeezing water molecules out. The first explanation
is not physically feasible, which implies that most rhamnolipid 1 and keratins formed complexes
in the coated layer by limited changes to the film structure.
Figure 7.6 Frequency and dissipation changes as a function of time for rhamnolipid 2 adsorption on the coated film measured by
QCM-D: frequency (blue); dissipation (red).
Figure 7.6 shows the frequency and dissipation changes of the adsorption of rhamnolipid 2 on the
coated film. It was found that the shifts of dissipations were much bigger than rhamnolipid 1 at
the same concentration, revealing that rhamnolipid 2 either formed a viscoelastic and soft layer
-6
-2
2
6
-40
-20
0
20
40
0 100 200 300 400 500
Dis
sip
ati
on
(10
-6)
Fre
qu
ency
(H
z)
Time (min)
176
on the film or swelled the film. Calculation shows that Hz-1, implying that the
Sauerbrey model is still applicable in this case as the dissipation is negligible.
Figure 7.7 Adsorbed amount for rhamnolipid 1 (top) and 2 (bottom) adsorption as a function of time on the coated
film measured by QCM-D.
-2
0
2
4
6
8
0 100 200 300 400 500
Mass
(m
g/m
2)
Time (min)
Rhamnolipid 1
0.1g/L pH 5Rhamnolipid 1
0.05g/L pH 5
Rhamnolipid 1
0.03g/L pH 5
Rhamnolipid 1
0.01g/L pH 5
0
2
4
6
8
0 100 200 300 400 500
Mass
(m
g/m
2)
Time (min)
Rhamnolipid 2
0.01g/L pH 5
Rinse
Rhamnolipid 2
0.03g/L pH 5
Rinse
Rhamnolipid 2
0.06g/L pH 5
Rinse
(a)
(b)
177
Unlike the adsorption of rhamnolipid 1, the frequency did not return to zero after the adsorption
of rhamnolipid 2, suggesting that rhamnolipid 2 had a stronger interaction with keratin and could
not be removed. The dissipation shifts were seen to return to their initial value, implying that the
total structure of the adsorbed layer remained virtually unchanged by the washing process.
Figure 7.7 shows the adsorbed mass of rhamnolipids on the coated film as measured by QCM-D.
The residual mass of rhamnolipid 2 after rinsing off was approximately 0.9±0.2 mg/m2, indicating
that rhamnolipid 2 had a stronger binding to keratins compared to rhamnolipid 1. However, the
adsorbed amounts of the two rhamnolipids at the same molar concentration were found to be
similar.
7.2.2.2 SE measurements
Figure 7.9 The adsorbed masses measured by QCM-D and SE as a function of time. The blue line represents the
data from QCM-D, while the red line represents the data from SE.
Dynamic measurements were performed involving recording of the SE signals every 10 seconds
from the beginning of the adsorption. The fit parameter for the layer thickness was also obtained
-2
0
2
4
6
8
0 100 200 300 400 500
Ad
sorb
ed m
ass
(m
g/m
2)
Time (min)
Rhamnolipid 1
0.1g/l pH 5Rhamnolipid 1
0.05g/l pH 5
Rhamnolipid 1
0.03g/l pH 5
Rhamnolipid 1
0.01g/l pH 5
178
simultaneously with the mass. Figure 7.9 shows the adsorbed mass of rhamnolipid 1 measured by
the two techniques of QCM-D and SE. The QCM-D data was analysed by the Sauerbrey model
at the 3rd frequency. The four peaks in Figure 7.9 were injections of rhamnolipid 1 solutions at
concentrations of 0.01 mg/ml, 0.03 mg/ml, 0.05 mg/ml and 0.1 mg/ml. It was found that the two
techniques resulted in a similar trend in adsorption profile, but gave different adsorbed amounts.
As QCM-D also measures water molecules that are coupled inside the adsorbed layer, the
percentage of water can be calculated by comparing the mass obtained from the two techniques.
Hook et al.3 showed that the amount of water coupled in the layer lies on different kinds of
adsorbed molecules and can be as high as 95%. It was found in this work that the difference in
the adsorbed mass measured by the two techniques increased with increasing concentrations of
rhamnolipid 1, indicating that the addition of rhamnolipid 1 swelled the coated film and allowed
more water inside the keratin film. At 0.1 mg/ml of rhamnolipid 1, the adsorbed mass measured
by ellipsometry was about 50% compared to the measurement by QCM-D, suggesting that
approximately 50% of water is coupled into the adsorbed layer. Measurements of rhamnolipid 2
by SE however, were restricted due to poor signals. It was assumed that rhamnolipid 2 caused
more damage to the coated film, thus causing the inhomogeneity in the horizontal direction.
179
7.2.2.3 NR measurements
Figure 7.10 Neutron reflection as a function of time for rhamnolipid 1 adsorption on the coated films at pH 5: SiO2
in D2O (triangles); the coated film in D2O (rectangles); 0.26 mg/ml (5 CMC) RL 1 in D2O (circles); the coated film
in D2O after the adsorption of RL 1 (crosses).
In the work described here, neutron reflection (NR) was used to investigate the conformational
structures of adsorbed rhamnolipids on the coated films. In order to reproduce the QCM-D results
that the adsorption of rhamnolipid 1 is able to be washed off thoroughly while rhamnolipid 2 is
not, solutions of rhamnolipids 1 and 2 at high concentrations were first performed onto the coated
film. The concentrations of rhamnolipid 1 and rhamnolipid 2 used were 0.26 mg/ml (0.5 mM and
5 times CMC) and 0.4 mg/ml (0.6 mM and 4 times CMC) respectively. The CMCs and other fit
parameters of rhamnolipids are described in the support information. Figure 7.10 shows the
adsorption of rhamnolipid 1 on the coated film in D2O buffer. The triangles represent the bare
SiO2 in D2O. The rectangles represent the coated film. The circles represent the adsorption of
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum tansfer, Q (Å-1)
180
rhamnolipid 1 on the coated film, and the crosses represent the adsorption profile after the rinsing
process.
Table 7.3(a) Structural parameters of the coated film at the solid/water interface from NR.
Layer Thickness/Å
±2
SLD/×10-6
Å-2
±0.02
Volume
Fraction
±0.02
Mass /mg•m2
±0.02
SiO2 16 3.41 -- --
1 15 4.7 0.52 1.08
2 34 5.5 0.29 1.37
3 40 6 0.12 0.67
sum 3.12
Table 7.3(b) Structural parameters of adsorbed rhamnolipid 1 at 0.26 mg/ml on the coated film.
Layer Thickness/Å
±2
SLD/×10-6
Å-2
±0.02
Keratin
Volume
Fraction
±0.02
RL 1
Volume
Fraction
±0.02
Keratin
mass
/mg•m2
±0.02
RL 1
mass
/mg•m2
±0.02
SiO2 16 3.41 -- -- -- --
1 15 3.12 0.52 0.24 0.53 0.37
2 34 3.76 0.29 0.25 1.19 0.87
3 40 5.34 0.12 0.10 0.57 0.41
4 59 5.59 0 0.15 0 1.07
Sum 3.12 2.72
Table 7.3(a) gives the fit parameters for the coated film and Table 7.3(b) describes the interaction
of the rhamnolipid 1 with the coated film. By using the same fit thicknesses of the coated film in
181
Table 7.3(a), a four-layer model was chosen for use for the adsorption of rhamnolipid 1 with the
first three layers of the same thickness as the coated film and a fourth dilute layer on top. The
results reveal that most of the rhamnolipid 1 was immersed into the coated film. The molar ratio
of rhamnolipid 1/keratin for the inner layer was approximately 41 whereas it was 77 for the middle
layer and 75 for the outer layer of the film. A loose layer was also formed containing rhamnolipid
1 with a thickness of 59 ± 2 Å. QCM-D showed the total thickness of the layer to be approximately
120 Å, which is consistent with the NR results. The totally adsorbed mass of the rhamnolipid 1
on the coated film was 2.72 mg/m2, the same value as the SE results (2.82 mg/m2). It is interesting
to note from Figure 7.10 that the profiles of the crosses and the rectangles were close, implying
that the adsorbed rhamnolipid 1 on the coated film can be readily washed off by pure water.
However, the results also indicate that the rhamnolipid 1 did not change the structure of the coated
film.
Figure 7.11 Reflectivity profiles as a function of momentum transfer for the adsorption of rhamnolipid 2 on the
coated film: SiO2 in D2O (triangles); coated film in D2O (rectangles); 0.4 mg/ml RL 2 in D2O (circles); the coated
film in D2O after the adsorption (crosses).
1E-6
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
0.01 0.02 0.04 0.08 0.16
Ref
lect
ivit
y
Momentum transfer, Q(Å-1)
182
Figure 7.11 shows the adsorption of rhamnolipid 2 on the keratin coated film. The triangles
represent the bare SiO2 in D2O, the rectangles represent the coated film, the circles represent the
adsorption of rhamnolipid 2 and the crosses represent the profile after rhamnolipid 2 had been
washed off. Unlike rhamnolipid 1, the results show that the coated film did not return to the initial
condition after the rinsing process of adsorption, revealing that rhamnolipid 2 had a stronger
interaction with keratin.
Table 7.4(a) Structural parameters of the adsorbed layer of rhamnolipid 2 at 0.4 mg/ml on the coated film
Layer Thickness/Å
±2
SLD/×10-6
Å-2
±0.02
Keratin
Volume
Fraction
±0.02
RL 1
Volume
Fraction
±0.02
Keratin
mass
/mg•m2
±0.02
RL 1
mass
/mg•m2
±0.02
SiO2 16 3.41 -- -- -- --
1 15 2.52 0.52 0.42 0.53 0.70
2 34 4.76 0.29 0.13 1.19 0.50
3 40 5.59 0.12 0.07 0.57 0.54
4 69 5.89 0 0.10 0 0.77
Sum 3.12 2.51
Table 7.4(a) shows a four layer structure of adsorbed rhamnolipid 2 on the coated film. The
thicknesses of the first three layers were kept the same as in Table 7.3(a) in order to provide a
comparison with the adsorbed mass of rhamnolipid 1. The results revealed that 85% of
rhamnolipid 2 was penetrated into the coated film. Of these rhamnolipid 2 molecules, 34%
remained in the inner keratin layer and 51% remained in the outer layers. The molar ratios of
rhamnolipid 2/keratin in the first three layers were found to be 57, 32 and 41, respectively. In
comparison to rhamnolipid 1/keratin ratio (41, 77 and 75), it was found that a higher percent of
rhamnolipid 2 penetrated into the dense inner layer of the film. The results also indicate that the
183
percentage of rhamnolipid 2 immersed into the coated film was higher than rhamnolipid 1,
implying that rhamnolipid 2 interacted more strongly with keratin molecules. There was found to
be 15% of the materials that remained outside the coated film. The overall mass of the adsorbed
rhamnolipid 2 was 2.51 mg/m2, which is of the same level as rhamnolipid 1 (2.72 mg/m2). The
main cause for the small difference is that the outer layer of rhamnolipid 2 was more dilute than
rhamnolipid 1.
Table 7.4(b) Structural parameters of the coated film from NR after the adsorbed rhamnolipid 2 was washed off.
Layer Thickness/Å SLD/×10-6 Volume
Fraction
(keratin)
Volume
Fraction
(RL 1)
Keratin
mass
(mg/m2)
RL 1
mass
(mg/m2)
1 16 3.41 -- -- -- --
2 15 4.25 0.52 0.10 0.53 0.15
3 34 5.15 0.29 0.06 1.19 0.21
4 40 5.96 0.12 0.01 0.57 0.04
5 69 6.20 0 0.01 0 0.07
Sum 3.12 0.47
Table 7.4(b) gives the structural parameters of the coated film after the adsorbed rhamnolipid 2
was washed off, as represented by crosses in Figure 7.11. The mass of the remaining rhamnolipid
2 after the rinsing process was 0.47 mg/m2, which is about one fifth of the total adsorbed mass.
The results also show that the remaining rhamnolipid 2 was mainly in the inner layer of the coated
film. The total mass of the residuals was lower compared to QCM-D results (0.9 mg/m2), implying
that a proportion of the keratin layers was removed at the same time.
184
Figure 7.12 The scattering length density as a function of distance from the silicon interface: keratin films spin-
coated on SiO2(red); adsorbed rhamnolipid 1 at 0.3 mg/ml on the coated film (green); adsorbed rhamnolipid 2 at 0.4
mg/ml on the coated film (orange); keratin films after adsorbed rhamnolipid 2 had been washed off (blue).
Figure 7.12 shows the scattering length densities of adsorbed rhamnolipid 1 and 2 on the coated
film as a function of distance from the silicon interface. The green and orange lines represent the
adsorption of rhamnolipid 1 and 2 on the coated film respectively. It can be seen in the graph that
rhamnolipid 2 is more likely to penetrate into the inner layers of the coated film than the
rhamnolipid 1, and that over half of the rhamnolipid 2 remained in the inner layers of the coated
film. The blue line indicates that the SLD shifted upwards after the rhamnolipid 2 was washed off,
implying that only part of rhamnolipid 2 adsorbed in the inner layer of the coated film was washed
off.
185
7.2.2.4 DPI measurements
Figure 7.13 Transverse magnetic (TM) changes as a function of time for rhamnolipid 1 adsorption on the coated
film measured by DPI.
DPI is a powerful instrument used to study the dynamic adsorption at the solid/liquid interface. It
is able to provide data of higher resolution than SE.4 However, its use is complex, especially for
adsorption on pre-coated films, which has restricted its use. In this work, manipulation of the
coating process on the DPI chip was difficult, due to the small size of the chip. In order to obtain
the correct information regarding the coated film with DPI measurement, there must be no bubble
throughout the whole process. Thus only one measurement was obtained at one concentration of
rhamnolipid 1 adsorption to demonstrate the coated film, the adsorption process of rhamnolipids,
and to get a comparison with the other three techniques. Figure 7.13 shows the shifts of the
transverse magnetic signal for rhamnolipid 1 adsorption on the coated film. The results revealed
that rhamnolipid 1 could be readily washed off by water. The information of the film structure
was partly determined by the value of film birefringence, showing that there was negligible
structural change of the film after the rinsing process. After the rinsing process, the coated film
was washed off by 5% w/w diluted Decon90. Under the assumption that the keratin layer was
186
homogeneous, the TM trend in the graph is able to represent the trend of adsorbed amount of
rhamnolipid 1 and the coated layer. In the Decon rinsing process, the TM signal dropped
dramatically at the first stage and decreased more slowly in the middle of the process. NR showed
that the coated film had a dense inner layer and a dilute outer layer, corresponding to the two-
stage washing process as shown in Figure 7.13. The analysis of the dynamic adsorption of
rhamnolipid 1 measured by DPI is shown in the support information.
7.3 Conclusion
The spin-coated film of keratin on the SiO2 substrate proved to be a useful and efficient method
to investigate keratin/surfactant interactions. NR revealed that the coated film formed a three-
layer structure with an inner dense layer, a middle layer with medium density and a dilute outer
layer. NR also revealed the high stability and reproducibility of the coated film both air dried and
in water. The adsorption of rhamnolipid 1 and 2 on the coated film was examined by the combined
techniques: QCM-D, SE, DPI and NR. It was found that rhamnolipid 1 was able to be completely
washed off by water after adsorption whereas only 70% of rhamnolipid 2 was removed after
adsorption under the same conditions, indicating that rhamnolipid 2 has a stronger interaction with
keratin than rhamnolipid 1. QCM-D also showed that the adsorption of rhamnolipid 2 produced a
high dissipation of the layer while the adsorption of rhamnolipid 1 had negligible dissipation,
implying that the addition of rhamnolipid 2 swelled the structure of the coated film. By comparing
SE and QCM-D, the percentage of trapped water in the coated film with the addition of
rhamnolipids was determined and the results show good consistency with NR measurements. DPI
provides another method of investigation, demonstrating that the coated film was composed of a
dense part and a dilute part, and the adsorption of rhamnolipid 1 was reversible.
187
Figure 7.14 Adsorbed amount of rhamnolipid 1as a function of bulk concentration for four different techniques:
NR, QCM-D, SE and DPI.
Figure 7.14 shows the adsorbed amount of rhamnolipids on the coated film. The results from
rhamnolipid 1 reveal that SE, DPI and NR were highly consistent in relation to the adsorbed
amount. The comparatively higher values measured by QCM-D indicate that there was 50% water
inside the adsorbed layers. The adsorbed amount measured by NR shows that both rhamnolipid 1
and 2 had similar adsorbed amount at the keratin film interface at the same molar concentration.
QCM-D data shows that rhamnolipid 1 had a slightly higher adsorbed amount than rhamnolipid
2, indicating that the addition of rhamnolipid 1 trapped more water inside the coated film. A large
188
pH effect was also seen on the adsorption of rhamnolipid 1 but a small pH effect on the adsorption
of rhamnolipid 2. This aspect of the investigation will constitute part of future work.
References
1. R. P. Richter and A. R. Brisson, Biophysical Journal, 2005, 88, 3422-3433. 2. L. Macakova, E. Blomberg and P. M. Claesson, Langmuir, 2007, 23, 12436-12444. 3. F. Höök, B. Kasemo, T. Nylander, C. Fant, K. Sott and H. Elwing, Analytical chemistry, 2001, 73,
5796-5804. 4. M. J. Swann, L. L. Peel, S. Carrington and N. J. Freeman, Analytical biochemistry, 2004, 329, 190-
198.
Support information
Table SI 1 Parameters of rhamnolipids used in NR fits
SL /10-4 Å
±0.02
volume CMC at pH5 CMC at pH6.8 Molar mass
/g•Mol
SLD /10-6 Å-2
±0.02
h-Rhamnolipid
1
-4.50 813 0.04mM (0.02
mg/ml)
0.1mM (0.05
mg/ml)
532 -0.554
h-Rhamnolipid
2
6.40 1052 0.04mM (0.026
mg/ml)
0.15mM (0.098
mg/ml)
678 0.608
189
Figure SI 1 Surface tension measurements of rhamnolipid 1 (blue points) and rhamnolipid 2 (red points) as a
function of bulk concentration in the pH 5 buffer.
Figure SI 2 Adsorbed amount of rhamnolipid 1 on the coated film as a function of time in the pH 5 buffer
measured by DPI. The concentration of rhamnolipid 1 solution is 0.06 mg/ml.
20
30
40
50
60
70
0.01 0.1 1 10 100 1000
Su
rface
ten
sion
(m
N/m
)
Concentration (mg/l)
0
1
2
3
4
0 200 400
Mass
(m
g/m
2)
Time (Sec)
190
Chapter 8
An initial study of Mel-C adsorption at solid/water interfaces and its
behaviour in solution
Mels are mannosylerythritol lipids that are classified as glycolipids. They are widely used in
medical science and in the chemical industries, in applications ranging from immunological to
antimicrobial1, 2. In this chapter, an initial study of Mel-C adsorption was carried out at both
SiO2/water and the keratin/water interfaces. There was found to be little adsorption of Mel-C at
the SiO2/water interface due to its hydrophobicity and slightly negatively charge. The adsorbed
amount of Mel-C onto a coated keratin film however, was found to be 0.68 mg/m2 at pH 6, 2.17
mg/m2 at pH 5, 2.68 mg/m2 at pH 4 and 1.91 mg/m2 at pH 3, revealing that adsorption is highly
dependent on pH. At pH 5, the adsorbed amount of Mel-C was found to reach saturation at 0.05
mg/ml (11 CMC). Dynamic light scattering (DLS) showed that Mel-C formed vesicular structures
in solution with vesicular sizes ranging from 30 to 3000 nm. The sizes of the vesicles increased
by increasing pH of the solution. SANS measurements showed that Mel-C formed stacked discs
with disc thicknesses or radii of approximately 900-3000 nm. Zeta potential measurements
showed that Mel-C molecules were negatively charged and that the charges were reversible by
the addition of a cationic surfactant DTAB in the solution.
191
8.1 Results and discussion
8.1.1 Surface tension measurements
8.1.1.1 Static surface tension
Figure 8.1 The surface tension isotherm of Mel-C as a function of bulk concentration in pH 5 buffer solution. The
solid lines are fitted linearly by eye.
The surface tension isotherm of Mel-C was measured in order to check its purity and to get a
broad understanding of its interfacial behaviour. The structure of Mel-C is described in section
1.2.2 where the CMC of Mel-C is quoted from 4×10-6 M 3, 4.5×10-6 M 4 to 6×10-6 M 5 with surface
tensions ranging from 24.2, 30.7 to 25.1 mN/m, respectively. Figure 8.1 shows the surface tension
isotherm of Mel-C in pH 5 buffer solutions. The CMC of Mel-C was found from the data to be
approximately 4 mg/l, which is 5.8×10-6 M, given the assumption that the molecular weight of
Mel-C is 680 Dalton.12 This result is consistent with the measurements by Konishi et al.5
20
30
40
50
60
70
0.01 0.1 1 10 100
Su
rface
ten
sion
(m
N/m
)
Concentration (mg/l)
192
8.1.1.2 Dynamic surface tension
Figure 8.2 Dynamic surface tension of Mel-C as a function of bubble life time with different Mel-C concentrations
in pH 5 buffer solution.
In processes such as foaming, emulsifying, coating and spraying, interfaces are produced
extremely quickly. Most surfactants also have a very rapid time constant to reach interfacial
equilibrium due to their small molecular weights. Such processes are not only affected by the
equilibrium value of the surface tension, but also by the kinetics of the formation of the interface.
Figure 8.2 shows the dynamic surface tension of Mel-C at 0.02, 0.1 and 0.5 mg/ml bulk
concentration in the pH 5 buffer solution. The data shows that the surface tension reduced more
rapidly at higher concentrations of Mel-C solution. At the concentration of 0.1 mg/ml (1.5×10-4
M), the surface tension began to reduce at 10 seconds of bubble life time, indicating a slow
dynamic adsorption process of Mel-C molecules.
55
60
65
70
75
0.001 0.1 10
Su
rface
ten
sion
(m
N/m
)
Life time (sec)
MelC 0.5 mg/ml
MelC 0.1 mg/ml
MelC 0.02 mg/ml
193
Figure 8.3 Dynamic surface tensions as a function of life time for Mel-C, SDS, and their mixture in the pH 5 buffer
solution. The concentrations of Mel-C and SDS were 0.1 mg/ml (1.47×10-4 M) and 0.3 mg/ml (1×10-3 M),
respectively.
The anionic surfactant SDS was also used to investigate its interaction with Mel-C molecules.
Figure 8.3 shows the dynamic surface tension of Mel-C, SDS, and their mixture. The results show
that the profile of the mixture is similar to the profile of SDS prior to 4 seconds, indicating that
the surface tension was dominated by the adsorption of SDS. The data implies that the adsorbed
layer was then replaced by Mel-C molecules, and that after 50 seconds of the bubble life time, the
surface was dominated by Mel-C molecules. The values for both the mixture and Mel-C were
similar at 50 seconds, indicating that there is little interaction between Mel-C and SDS.
60
65
70
75
0.001 0.1 10
Su
rface
ten
sion
(m
N/m
)
Life time (sec)
MelC 0.1mg/ml
MelC 0.1mg/ml+SDS 0.3mg/ml
SDS 0.3mg/ml
194
Figure 8.4 Dynamic surface tensions as a function of life time for Mel-C, DTAB, and their mixture in the pH 5
buffer solution.
Figure 8.4 shows the data from measurements of dynamic surface tensions for Mel-C, DTAB and
their mixture, and allows for a comparison with the Mel-C/SDS mixture. The data show the same
profile as the Mel-C/SDS mixture, which indicates that there was little interaction between Mel-
C and DTAB. It can therefore be concluded that Mel-C molecules are relatively stable and their
charge effect on other molecules is negligible.
60
65
70
75
0.001 0.1 10
Su
rface
ten
sion
(m
N/m
)
Life time (sec)
MelC 0.1 mg/ml uHQ
MelC 0.1gL+DTAB 0.3gL
DTAB 0.3gL pH 4
195
8.1.2 Adsorption of Mel-C at the SiO2/water interface
Figure 8.5 Mel-C adsorption (0.04 mg/ml, 10 CMC) as a function of time for at the SiO2/water interface at three
different pH values measured by SE.
Spectroscopic ellipsometry (SE) has been used to examine the adsorption of Mel-C at the
hydrophilic silicon oxide/water interface. The treatment of the silica wafers has been described in
Section 2.2.3. Figure 8.5 shows the profiles of Mel-C adsorption in pH 3, pH 4 and pH 5 solutions.
Under these conditions, the silicon oxide layer bears negative charges. The results show that there
was no adsorption of Mel-C on the silica surface, indicating that the hydrophobic and electrostatic
interactions at the interface were very weak. This is consistent with the slow dynamics of Mel-C
measured by dynamic surface tension.
SE, however, does not allow for the detection of very dilute layers. Therefore, parallel QCM-D
measurements were conducted. By measuring the change in the sensor’s frequency (∆f) and
dissipation (∆D), the adsorbed amount, thickness and softness of the adsorbed layer can be
determined. Prior to each measurement, the sensor was calibrated to provide the zero frequency
and the dissipation in an aqueous environment. The results show that there was little adsorption
-1
-0.5
0
0.5
1
0 10 20 30
Ad
sorb
ed a
mou
nt
(mg/m
2)
Time (min)
pH 5pH 4pH 3
196
of Mel-C during the first 6 minutes, whereas the dissipation was increased to 1×10-6, indicating a
negligible layer with high softness.
8.1.3 Adsorption of Mel-C on the coated film of keratin
Few studies have been conducted into the interaction of Mel-C and proteins. In the literature
Konishi et al.13 studied Mel-A and its high binding affinity to lectins. Kitamoto et al.11 showed
the binding affinity of Mel-A towards the protein concanavalin A. They also reported the
antimicrobial activity of Mel-A and Mel-B by modifying the sugar moiety.14 Inoh et al.15 reported
the function of Mels for gene delivery, in that they are able to immensely enhance the adhesion
of liposome-DNA mixture into the cell. In the field of cosmetic applications, Morita et al.16 found
that Mel-A could recover the human skin cell damaged by SDS treatment. Therefore Mel-A has
great potential for skin-care. In the work presented here, an initial study of the interaction of Mel-
C and keratin by SE and QCM-D was carried out.
8.1.3.1 SE measurements
Figure 8.7 Psi and delta profiles as a function of wavelength for the SiO2 layer in air, the coated keratin film at the
solid/air interface and the coated film at the solid/water interface measured by SE. The psi and delta are described
in Chapter 2.2.
197
Figure 8.7 shows the three profiles of a SiO2 layer in air, the coated keratin film at the solid/air
interface and the coated keratin film at the solid/water interface. The blue and red lines represent
psi and delta and the dotted line represents the best fit of the measured data as found using the
Cauchy model. The perfect fits indicate that the coated film is homogenous and flat. The best fits
of the three profiles are listed in Table SI 8.1 in Support Information.
Figure 8.8 Psi and delta profiles as a function of wavelength for the coated keratin film and the adsorption of Mel-
C at 20 minutes and 40 minutes measured by SE. The psi and delta are described in Chapter 2.2.
Figure 8.8 shows the phase changes of the psi and delta profiles measured from the adsorption of
Mel-C on the coated keratin film. The adsorption reached equilibrium after 40 minutes. The dotted
lines are the best fits to the measured data. The fits became less good with increasing time of Mel-
C adsorption, indicating that Mel-C adsorption led to a change in the structure of the film. The
best fitted parameters are listed in Table SI 8.2, showing that the thickness of the total layer
increased from 38 Å to 70-80 Å. Neutron reflectivity would be required in order to obtain further
structural information of both the coated layer and the adsorbed layer, which will be the aim of
future research.
198
Figure 8.9 Adsorbed amount of Mel-C against bulk concentration of 0.01 mg/ml, 0.02 mg/ml, 0.03 mg/ml, 0.05
mg/ml and 0.1 mg/ml taken upon 30 min of adsorption (1.47×10-4 M and 25 times CMC), respectively (left). Layer
thickness of adsorbed Mel-C as a function of time (right). The thickness was calculated by taking the refractive
index fixed at 1.4.
Figure 8.9 shows the adsorbed amount of Mel-C as a function of time in the pH 5 buffer solution.
The results show that the adsorbed amount of Mel-C reached equilibration at 0.05 mg/ml (11
CMC) of Mel-C. In the right graph the thickness was greatly affected by the refractive index of
the total layer (the best fits can be obtained with the refractive index changed from 1.39 to 1.42),
which produced the large error bars as shown in Figure 8.9.
199
Figure 8.10 Adsorbed amount of Mel-C as a function of time at 0.1 mg/ml in pH 3, 4, 5 and 6 solution, respectively
(left). Adsorbed amount of Mel-C against pH (right).
Figure 8.10 shows the pH effect of Mel-C adsorption. Since the coated film would begin to
dissolve when the pH of the solution is above 7 (as described in Section 6.3.1), only the pH effect
up to pH 6 was measured. The results show that the adsorbed amount of Mel-C reached the
maximum at a pH of approximately 4 to 5, indicating that the keratin at its isoelectric point most
strongly attracts the Mel-C molecules.
0
1
2
3
0 10 20 30
Am
ou
nt
of
Mel
-C (
mg/m
2)
Time (min)
Ph 6
Ph 5
Ph 4
Ph 3
0
1
2
3
4
3 4 5 6
Am
ou
nt
of
Mel
-C (
mg
/m2
)
pH
200
8.1.3.2 QCM-D measurements
Figure 8.11 Frequency and dissipation shifts of Mel-C as a function of time at 0.04 mg/ml (10 CMC) in the pH 5
solution measured by QCM-D. F and D represent the frequency and the dissipation, respectively.
Figure 8.11 shows the frequency and dissipation shifts of Mel-C adsorption in the pH 5 buffer
solution as a function of time. The results show that Mel-C formed a layer with very high
dissipation compared to its frequency change, indicating a very soft layer. The adsorption reached
equilibration at around 20 minutes, 5-10 minutes faster than the SE measurement. This
phenomenon has also been observed in other studies by comparing SE with DPI and QCM-D
measurements. The reasons for the difference of adsorption time are not certain. However, the
differences may be explained by considering that SE is a static measurement with limited solution
in the cell whereas the QCM-D is a dynamic measurement with the solution flowing through the
surface. It is noteworthy that the frequency and dissipation did not return to their initial values
after the rinsing process, indicating a strong binding between the keratin and Mel-C molecules.
The increased dissipation also indicates that Mel-C hugely softened the coated film compared to
rhamnolipids.
201
8.1.4 Size measurements of Mel-C in solution
There are few studies to date concerned with the behaviour of Mel-C in solution. However, some
studies have focused on Mel-A and Mel-B in solution. Tomohiro et al. 6 investigated the
distinctive self-assembling properties of Mel-A and Mel-B using dynamic light scattering (DLS),
transmission electron microscopy (TEM) and small-angle x-ray scattering (SAXS) analysis, and
showed that the vesicular sizes of both Mels in solution were around 160-180 nm above their
critical aggregation concentration (CAC). They also showed that the structure of Mel-A changed
drastically from large unilamellar vesicles to sponges with changes in the concentration of the
solution, whereas the structure of Mel-B changed gradually with the concentration of solution.
Such structures provide excellent moisture retention and can find application in ceramide creams
for human skin and hair.7, 8 Considering such useful structures of Mels, Yoshikazu et al.9 further
reported that Mel-A can dramatically increase gene transfection via membrane fusion and
Kitamoto et al.10 tested the antimicrobial activities of Mel-A and Mel-B.
Mel-C normally has a long-chain acid attached to the mannose moiety, which makes its
hydrophobic part drastically different from that of Mel-A and Mel-B. Kitamoto et al.11 reported
that both Mel-B and Mel-C formed giant vesicles in a phosphate buffer (pH 8.0) with the vesicular
sizes approximately 10 µm by using fluorescence microscopy. They also showed that the giant
vesicles had particular binding affinity towards the mannose-binding protein, concanavalin A. In
the work described here, the size of Mel-C vesicles was investigated by DLS and SANS. In
addition to the giant vesicles, it was found that Mel-C also formed some small vesicles ranging
from 30 to 3000 nm in solution.
202
Figure 8.12 The volume percentage as a function of hydrodynamic radius for Mel-C at 0.1 mg/ml (250 CMC) as
measured by DLS (red line: pH 9; blue line: pH 7; green line: pH 5).
Figure 8.12 shows the distribution of vesicular size of Mel-C in different pH values. It was found
that Mel-C molecules formed bigger vesicles with increasing pH values in solution. The average
size of the Mel-C vesicles was around 100 nm at pH 5, 110 nm at pH 7 and 400 nm at pH 9. The
high pH contributed to a large Gaussian distribution of vesicular sizes. The results showed that
Mel-C also formed small vesicles in addition to the giant vesicles reported by Kitamoto et al.11
As DLS is limited to the measurement of particle sizes up to only a few microns, the giant vesicles
(of the order of 10 μm) formed by Mel-C cannot be detected.
Figure 8.13 shows SANS measurements of Mel-C at 25, 100 and 250 CMC, respectively. Initial
attempts using the lamellar, vesicular and other widely used models gave very poor fits. The
stacked disc model provided the form factor, P(q), for stacked discs with a core/layer structure,
where the form factor was normalized by the volume of the cylinder. It is also assumed that the
next neighbour distance (d) in a stack of parallel discs obeys a Gaussian distribution. The best
fitted parameters for 250 CMC Mel-C are given in Table 8.1. Since there are 7 fit parameters for
this model, it is difficult to determine the precise structure of the vesicles. However, the radii of
the discs can be well restricted from 50 to 1500 nm. The best fits for 100 CMC and 25 CMC Mel-
C are shown in Support Information (Table SI 8.3 and SI 8.4).
203
Figure 8.13 Left: SANS profiles of Mel-C in D2O at 25, 100 and 250 CMC, respectively. Right: schematic graph of
the best fitted model. All solutions are in pH 5 solution.
Table 8.1 The best fitted parameters for SANS measurements of 250 CMC Mel-C in D2O.
204
8.1.5 Zeta potential measurements
Figure 8.14 The zeta potential of Mel-C at 0.1 mg/ml with and without DTAB. The curves from left to right
represent the added DTAB concentrations at 0, 0.01, 0.03, 0.05, 0.1 and 0.3 mg/ml concentrations, respectively.
Zeta potential is an important factor in the estimation of the stability of colloidal dispersions. For
particles in a colloid dispersion, higher absolute values of the zeta potential imply higher stability
of the colloidal system. The cell with two gold electrodes was used for the zeta potential
measurements. Figure 8.14 shows the zeta potential shifts of Mel-C solutions with the addition of
DTAB at different concentrations. It was found that the Mel-C vesicles have a zeta potential of -
50 mV, which indicates a very good stability of the solution. By the addition of cationic DTAB,
Mel-C vesicles are neutralised from negative charges to positive charges with increasing DTAB
concentrations. The mixture is neutralised to be zero zeta potential with the addition of 0.1 mg/ml
DTAB.
205
8.2 Conclusion
The interfacial adsorption of biosurfactant Mel-C at the SiO2/water interface and its interaction
with the coated keratin film at the coated film/water interface were investigated. Time-dependent
adsorption demonstrated little adsorption of Mel-C on the bare SiO2 substrate. However, both
Mel-C concentration and adsorption time affected the adsorbed amount onto the coated film of
keratin. The adsorption tended to equilibrate after 30 minutes, with higher concentrations reaching
equilibrium faster. The adsorbed amount of Mel-C on the coated film increased with increasing
Mel-C concentration. The adsorbed amount of Mel-C on the coated film is pH dependent and
reached its maximum at pH 4-5. The adsorbed Mel-C molecules onto the coated keratin film
formed a soft layer with high dissipation. Part of the adsorbed materials penetrated into the coated
film and formed a strong binding with the protein that could not be washed away by UHQ water.
DLS and NR results showed that Mel-C formed vesicles in solution with unit sizes ranging from
50 to 150 nm. Mel-C formed bigger vesicles in solution at higher pH values and the Gaussian
distribution of particle sizes ranged from 20 to 300 nm. The zeta potential results showed that
Mel-C aggregates are very stable with a value of -50 mV. The addition of DTAB was able to
minimise the zeta potential value and cause precipitation.
206
References
1. S. S. Cameotra and R. S. Makkar, Current opinion in microbiology, 2004, 7, 262-266. 2. P. Singh and S. S. Cameotra, TRENDS in Biotechnology, 2004, 22, 142-146. 3. T. Morita, M. Konishi, T. Fukuoka, T. Imura, S. Yamamoto, M. Kitagawa, A. Sogabe and D.
Kitamoto, Journal of oleo science, 2007, 57, 123-131. 4. L. Rodrigues, I. M. Banat, J. Teixeira and R. Oliveira, Journal of Antimicrobial Chemotherapy,
2006, 57, 609-618. 5. M. Konishi, T. Morita, T. Fukuoka, T. Imura, K. Kakugawa and D. Kitamoto, Applied microbiology
and biotechnology, 2008, 78, 37-46. 6. T. Imura, Y. Hikosaka, W. Worakitkanchanakul, H. Sakai, M. Abe, M. Konishi, H. Minamikawa
and D. Kitamoto, Langmuir, 2007, 23, 1659-1663. 7. T. Fukuoka, T. Morita, M. Konishi, T. Imura and D. Kitamoto, Carbohydrate Research, 2008, 343,
555-560. 8. D. Kitamoto, T. Morita, T. Fukuoka, M.-a. Konishi and T. Imura, Current Opinion in Colloid &
Interface Science, 2009, 14, 315-328. 9. Y. Inoh, D. Kitamoto, N. Hirashima and M. Nakanishi, Journal of controlled release, 2004, 94,
423-431. 10. D. Kitamoto, H. Yanagishita, T. Shinbo, T. Nakane, C. Kamisawa and T. Nakahara, Journal of
Biotechnology, 1993, 29, 91-96. 11. D. Kitamoto, S. Ghosh, G. Ourisson and Y. Nakatani, Chem. Commun., 2000, 861-862. 12. S. Hewald, U. Linne, M. Scherer, M. A. Marahiel, J. Kämper and M. Bölker, Applied and
environmental microbiology, 2006, 72, 5469-5477. 13. M. Konishi, T. Imura, T. Fukuoka, T. Morita and D. Kitamoto, Biotechnology letters, 2007, 29,
473-480. 14. D. Kitamoto, H. Yanagishita, T. Shinbo, T. Nakane, C. Kamisawa and T. Nakahara, Journal of
Biotechnology, 1993, 29, 91-96. 15. Y. Inoh, D. Kitamoto, N. Hirashima and M. Nakanishi, Biochemical and biophysical research
communications, 2001, 289, 57-61. 16. T. Morita, M. Kitagawa, M. Suzuki, S. Yamamoto, A. Sogabe, S. Yanagidani, T. Imura, T. Fukuoka
and D. Kitamoto, Journal of oleo science, 2009, 58, 639-642.
Support information
Figure SI 1 Mel-C colloids at 0.5 mg/ml at pH 5. The left bottle was at normal temperature and the right bottle was
heated to 80˚C over a duration of 1 hour. No difference between the two bottles mean that the colloid is very stable
at both temperatures.
207
Table SI 8.1 The best fits of the three profiles as shown in Figure 8.6
Cauchy Layer
Model Layer Thickness /Å
±0.2
A B C
1 SiO2 14.2 - - -
2 SiO2 14.2 - - -
Keratin Film 34 1.41 0.01 0.00001
3 SiO2 14.2 - - -
Keratin Film 37 1.41 0.011 0.00001
Table SI 8.2 The best fits parameters of the three profiles as shown in Figure 8.7
Cauchy Layer
Model Layer Thickness /Å
±0.2
A /1 B /1 C /1
1 Coated film 38 1.41 0.01 0.00001
2 Coated film 38 1.41 0.01 0.00001
Mel-C layer 27 1.40 0.01 0.00001
3 Coated film 38 1.41 0.01 0.00001
Mel-C layer 37 1.40 0.01 0.00001
208
Table SI 8.3 The best fits parameters for SANS measurement of 100 CMC Mel-C in D2O.
Table SI 8.4 The best fits parameters for SANS measurement of 25 CMC Mel-C in D2O.
209
Chapter 9
Summary
The aim of the research presented in this thesis was to study the interfacial adsorption and solution
self-assembly properties of three conventional surfactants (C12E6, DTAB and SDS), two
promising biosurfactants (rhamnolipids and Mel-C) and their interactions with the keratin protein.
The research area is of interest for commercial applications such as cosmetics and personal
hygiene products. The use of biosurfactants to replace synthetic surfactants has become an
increasingly appealing topic in recent years, due to the highly degradable and environmentally
friendly properties of biosurfactants. Many techniques have been developed in recent years for
investigating adsorbed layers at surfaces as well as for characterising the adsorption process.
By studying the surface adsorption of C12E6 at the SiO2/water interface, the research presented
here examined and compared the four most widely used techniques in surface physics, namely the
quartz crystal microbalance with dissipation (QCM-D), dual polarization interference (DPI),
spectroscopic ellipsometry (SE) and neutron reflection (NR).
SE, DPI and NR were found to provide consistent adsorbed amounts, DPI offered more effective
real time monitoring, and NR offered additional structural information on the adsorbed layer.
QCM-D measured the real time adsorbed hydrated mass that includes the trapped water
incorporated in the layers. Changes in the QCM-D’s dissipation indicated a gradual transition
from a more rigid to a less rigid layer as the concentration of C12E6 was increased. The combined
result of a micellular structure with the tilted bilayer possibility of the adsorbed layer proved the
consistency and effectiveness of the four techniques.
The research also investigated the surface adsorption of keratin polypeptides extracted from wool
by surface tension (ST) and neutron reflectivity (NR) and their solution aggregation by dynamic
210
light scattering (DLS) and small-angle neutron scattering (SANS). ST showed a steady surface
tension reduction with increasing concentration and NR revealed a steadily rising surface
adsorption up to 0.1 mg/ml1, 2. It was found that the interfacial layers were comprised of two main
regions, a dense top layer of 18-25 Å and a loose bottom layer of 25-30 Å. In the case of the top
dense layer, approximately half of it was exposed to air and the remaining part of it along with
the diffuse bottom layer were immersed into the solution. DLS revealed the occurrence of
aggregates with their hydrodynamic radii peaked around 100 Å with a density of around 0.1 g dm-
3, consistent with the SANS modelling of ellipsoidal aggregates, with the major radius of 140 Å
and the minor radius of 60 Å. The addition of NaCl caused the ellipsoids to become thinner but
longer, consistent with the electrostatic effect as observed from the surface adsorption.3, 4 The
results indicate that at low concentrations, the globular-shaped particles adsorb a thicker layer
whereas at high concentrations, the ellipsoidal-shaped particles produce a thinner layer.
Keratin/surfactant interactions at the air/water interface were investigated by ST and NR. As
described previously, the addition of surfactant can significantly affect how the protein behaves
in both bulk solutions and interfaces6-8. The results revealed that the small addition of SDS (0.1
mM) significantly reduced the adsorbed amount of keratin, whereas DTAB had less impact on
keratin adsorption until its concentration reached 1 mM. For the keratin/SDS complexes, the
molar ratio of SDS/keratin is of the order of 25 at the interface at 0.01 mg/ml keratin and 0.1 mM
SDS concentrations. The ratio raised to approximately 240 with the addition of 2 mM SDS,
indicating that SDS notably suppressed the adsorption of keratin.9 The amount of the adsorbed
keratin decreased from 1.81 g dm-3 to 0.28 g dm-3 with the addition of 2 mM SDS. For the
keratin/DTAB complexes, the ratio of DTAB/keratin is approximately 7 with 0.1 mM DTAB and
increased to approximately 19 with the addition of 0.4 mM DTAB. The ratio was further increased
to approximately 135 with the addition of 2 mM DTAB. In the meantime, the adsorbed keratin
remained constant with the addition of 0.7 mM DTAB, followed by a reduction to 0.67 g dm-3
211
with the addition of 2 mM DTAB. The whole process of the DTAB/keratin complex indicated a
complex interaction due to the opposite charges of the two components. The two results indicated
that the combined electrostatic and hydrophobic forces have a combined effect on the keratin/SDS
and keratin/DTAB complexes.
The binding of cationic surfactant DTAB onto the coated keratin film at the SiO2/water interface
was investigated using three techniques (QCM-D, SE and NR). A spin coater was used to coat the
keratin film on silicon oxide wafers. The film structure and its stability and reproducibility were
checked with precision using SE, QCM-D and NR. NR showed a three layer structure of the
coated film with a dense inner layer and a loose outer layer. Measurement revealed that a
homogeneous and reproducible film could be achieved by using the same solution of keratin and
the same settings of the spin coater. NR also revealed that the coated film could be dried after its
immersion into water and was able to be heated up to 80 °C without undergoing structural changes.
This finding is of great importance as it opens up a method to investigate the keratin/surfactant
interaction at the solid/water interface.
The subsequent binding to the keratin by DTAB at the solid/water interface was investigated by
comparing the difference before and after DTAB addition. Firstly, SE showed that the refractive
index of the entire layer increased with increasing DTAB concentration below 10 mM whereas
the thickness of the layer remained constant, indicating penetration of DTAB molecules into the
coated film. Further increase in DTAB concentration deteriorated the film. Secondly, QCM-D
revealed that the coated film had a high dissipation/frequency ratio, indicating that the coated
layer was very soft and abundant in water. The dissipation/frequency ratio was decreased by the
addition of DTAB below CMC, indicating that most of the DTAB molecules became penetrated
into the coated film by squeezing water molecules out of the film. Above the CMC, the addition
of DTAB removed most of the keratin, with the remaining materials forming a very soft and dilute
layer with high dissipation at the interface. Finally, NR revealed that three quarters of the DTAB
212
molecules penetrated into the coated film over the low DTAB concentration range and the
remaining DTAB formed a very dilute layer on the outer surface of the coated film with a
thickness of approximately 30 Å. At DTAB concentrations above 10 mM, some of the keratin had
already begun to be removed and the thickness of the coated layer started to decrease,
accompanied by significant structural deformation of the keratin film.
The two most widely studied biosurfactants, rhamnolipids and Mel-C, and their interactions with
the keratin protein were investigated. Two main obstacles in the use of biosurfactants in industrial
applications are their expensive large-scale production and difficulties in the separation and
purification processes. They are however promising in the replacement of synthetic surfactants
due to many positive attributes including outstanding biodegradability10 and significantly high
surface activity.11
The interaction of keratin and rhamnolipids was studied at the solid/liquid interface. Both QCM-
D and SE revealed that the dynamic adsorption of rhamnolipid 1 and 2 was identified with a rapid
process at the onset of the first 20 minutes, followed by a slower adsorption over the next 50
minutes with molecular rearrangement. Both QCM-D and NR found that rhamnolipid 1 was able
to be completely washed off by water after adsorption, whereas only 70% of rhamnolipid 2 was
able to be removed after adsorption, indicating that rhamnolipid 2 has a stronger interaction with
keratin than rhamnolipid 1. QCM-D also showed that the adsorption of rhamnolipid 2 produced a
high dissipation of the film while the adsorption of rhamnolipid 1 had negligible increase of
dissipation, implying that the addition of rhamnolipid 2 changed (swelled) the structure of the
coated film. By comparing SE and QCM-D, the percentage of trapped water in the coated film
was determined and the results were found to be in high consistency with NR measurements. DPI
provides another method to investigate the studied system, demonstrating that the coated film was
composed of a dense part and a dilute part, and that the adsorption of rhamnolipid 1 was reversible.
213
A preliminary study of the biosurfactant Mel-C was carried out. The mannosylerythritol lipid is
widely used in applications ranging from immunological to antimicrobial.11-13 The interfacial
adsorption of Mel-C was studied at both the SiO2/water and the keratin/water interfaces.
Compared to the low adsorption at the SiO2/water interface, the amount of adsorption at the
keratin/water interface was highly dependent on concentration and the pH of the solution buffer.
The amount of adsorption reached a maximum at pH 4-5, indicating that keratin at its electrostatic
point has the strongest interaction with Mel-C. DLS and SANS showed that Mel-C formed stacked
discs in solution with the disc radius ranging from 30 to 1500 nm. Zeta potential measurements
showed that Mel-C molecules were negatively charged and can be reversed by the addition of the
cationic surfactant DTAB in the solution.
The work reported in this thesis provides an overall and detailed picture of keratin/surfactant
interactions at interfaces studied by combining four powerful techniques (SE, DPI, QCM-D and
NR). Some SANS and DLS results of their solution properties are also included in order to provide
additional information about their interaction. The work has highlighted a number of areas of
interest for further study:
o The coated keratin film is exceptionally stable and reproducible with its layer structure
determined by multiple techniques, and as such is a good substrate to investigate other
materials besides the rhamnolipids.
o Mel-C is a promising biosurfactant with many outstanding properties. However, its study
is very limited at this stage. Its strong interactions with the keratin shows it has great
potential in the chemical industries. Mechanisms of the interactions between Mel-C and
the keratin at interfaces and in solution may be further investigated by additional
techniques such as NR and SANS.
o The NR studies of the keratin and the h-rhamnolipids only provide a general idea of how
they interact with each other at the solid/water and air/water interfaces. To further
214
investigate the mechanisms of their complexes, the deuterated d-rhamnolipids are also
required. This future work could produce a clearer graph of the functions of the
rhamnolipids.
References
1. Z. Li, J. Lu and R. Thomas, Langmuir, 1997, 13, 3681-3685. 2. J. R. Lu, S. Perumal, E. T. Powers, J. W. Kelly, J. R. Webster and J. Penfold, Journal of the
American Chemical Society, 2003, 125, 3751-3757. 3. R. A. McAloney, M. Sinyor, V. Dudnik and M. C. Goh, Langmuir, 2001, 17, 6655-6663. 4. S. T. Dubas and J. B. Schlenoff, Langmuir, 2001, 17, 7725-7727. 5. R. Green, T. Su, H. Joy and J. Lu, Langmuir, 2000, 16, 5797-5805. 6. M. D. Lad, V. M. Ledger, B. Briggs, R. J. Green and R. A. Frazier, Langmuir, 2003, 19, 5098-5103. 7. C. Monteux, C. E. Williams, J. Meunier, O. Anthony and V. Bergeron, Langmuir, 2003, 20, 57-63. 8. P. M. Claesson, E. Blomberg, J. C. Fröberg, T. Nylander and T. Arnebrant, Advances in Colloid
and Interface Science, 1995, 57, 161-227. 9. J. Lu, T. Su, R. Thomas, J. Penfold and R. Richards, Polymer, 1996, 37, 109-114. 10. S. S. Cameotra and R. S. Makkar, Current opinion in microbiology, 2004, 7, 262-266. 11. D. Kitamoto, H. Yanagishita, T. Shinbo, T. Nakane, C. Kamisawa and T. Nakahara, Journal of
Biotechnology, 1993, 29, 91-96. 12. Y. Inoh, D. Kitamoto, N. Hirashima and M. Nakanishi, Biochemical and biophysical research
communications, 2001, 289, 57-61. 13. T. Morita, M. Kitagawa, M. Suzuki, S. Yamamoto, A. Sogabe, S. Yanagidani, T. Imura, T. Fukuoka
and D. Kitamoto, Journal of oleo science, 2009, 58, 639-642.