S URFACE M ODIFICATION OF C ELLULOSE BY
C OVALENT G RAFTING AND P HYSICAL
A DSORPTION
Linn Carlsson Doctoral Thesis
AKADEMISK AVHANDLING
som med tillstånd av Kungliga Tekniska högskolan i Stockholm framläggs till offentlig granskning för avläggande av teknisk doktorsexamen fredagen den 21:a februari 2014, kl. 10:00 i Kollegiesalen, Brinellvägen 8, KTH, Stockholm. Avhandlingen försvaras på engelska. Fakultetsopponent:
Professor Mohamed Naceur Belgacem, Ecole Internationale du Papier, de
la communication imprimée et des biomatériaux, Grenoble INP – Pagora,
France.
Copyright © 2014 Linn Carlsson All rights reserved
Paper I © 2012 American Chemical Society Paper II © 2012 The Royal Society of Chemistry Paper III © 2013 American Chemical Society Paper V © 2012 The Royal Society of Chemistry TRITA‐CHE Report 2014:2
ISSN 1654‐1081
ISBN 978‐91‐7501‐987‐1
has increased tremendously over the last years. At the same time the Swedish forest industry faces new challenges in its strive to increase the utilization of cellulose fibers in high‐value end‐products. The aim of this study was to expand the toolbox for surface modification of cellulose by employing covalent surface‐initiated (SI) polymerizations or by physical adsorption of polymers.
SI‐ring‐opening polymerization (ROP) of ‐caprolactone (‐CL) was performed from filter paper (FP) and high surface area nanopaper (NP).
Larger amounts of polycaprolactone (PCL) were grafted from NP, compared to FP, owing to the higher amount of available initiating hydroxyl groups. Furthermore, the mechanical properties of PCL were improved by the grafting of FP and NP, as compared to pure PCL.
It is challenging to characterize a polymer grafted from a surface. Hence, quartz crystal microbalance with dissipation (QCM‐D) was employed to investigate SI‐ROP in real time from a cellulose model surface.
Furthermore, it was shown by colloidal probe AFM that increased length of grafted PCL, from cellulose microspheres, improved the interfacial adhesion to a pure PCL surface, suggesting that chain entanglements have a significant impact on the interfacial properties. Increased temperature and time in contact also improved the adhesion.
In order to investigate the degree of substitution (DS) and the degree of polymerization (DP), PCL‐grafted hydrolyzed cellulose cotton linters (HCCL) were studied by solid state NMR. It was found that despite a DS of only a few percent, the surface character changed considerably;
furthermore, the DS was virtually independent of the DP. To increase the amount of grafted polymer, ring‐opening metathesis polymerization (ROMP) of norbornene was performed from FP. Short polymerization times and low temperatures resulted in highly grafted surfaces.
Alternatively, physical adsorption by electrostatic interactions was employed to modify a cellulose model surface in the QCM‐D. Cationic latex nanoparticles of poly(dimetylaminoethyl methacrylate‐co‐methacrylic acid)‐block‐poly(methyl methacrylate) were produced by reversible addition‐fragmentation chain‐transfer (RAFT)‐mediated surfactant‐free emulsion polymerization by polymerization‐induced self‐assembly (PISA).
This strategy does not require any organic solvents and could potentially
be introduced in industrial processes.
markant de senaste åren. Detta samtidigt som den svenska skogsindustrin står inför nya utmaningar i sin strävan att öka användningen av cellulosafibrer i förädlade produkter. Syftet med denna studie var att utöka verktygslådan för ytmodifiering av cellulosa via kovalent ytinitierade polymerisationer eller genom fysikalisk adsorption av polymerer.
Ytinitierad ringöppningspolymerisation av ‐kaprolakton utfördes från filterpapper (FP) och nanopapper (NP) med hög ytarea. NP gav större ympad mängd på grund av fler initierande hydroxylgrupper. De mekaniska egenskaperna av polykaprolakton (PCL) förbättrades vid ympningen från FP och NP jämfört med en ren PCL matris och fuktadsorptionen reducerades avsevärt i jämförelse med omodifierade FP eller NP.
Det är en stor utmaning att karakterisera ympade polymerer, därför utvecklades en metod där ytinitierad‐ROP från en cellulosamodellyta studerades med en kvartskristallmikrovåg med dissipation (QCM‐D).
Atomkraftsmikroskopi med kolloidal sond visade att ökad ympningslängd av PCL förbättrade vidhäftningsförmågan, vilket indikerar att kedjeintrasslingar har en signifikant påverkan på gränsskiktsegenskaperna.
Ökad temperatur och kontakttid resulterade också i ökad vidhäftningsförmåga.
För att utreda inverkan av den ympade polymerisationsgraden (DP) och substitutionsgraden (DS) studerades PCL‐ympade hydrolyserade cellullosabomullslinters med fastfas‐NMR. Studierna visade att trots låg substitutionsgrad (några få procent) så ändrades ytkaraktären signifikant och substitutionsgraden var relativt konstant oberoende av DP.
För att öka mängden ympad polymer utfördes ringöppnings‐
metatespolymerisation (ROMP) från FP. Korta reaktionstider och låga polymerisationstemperaturer resulterade i kraftigt ympade ytor.
Fysikalisk adsorption genom elektrostatiska interaktioner användes för att modifiera en cellulosamodellyta i en QCM‐D. Katjoniska latexnanopartiklar av sampolymeren poly(dimetylaminoetylmetakrylat‐
metakrylsyra‐metylmetakrylat) syntetiserades med reversibel additions‐
fragmenteringskedjeöverförings‐medlad tensidfri emulsionspolymerisation
genom polymerisations‐inducerad själv‐organisation. Denna strategi
kräver inga organiska lösningsmedel och har potential att kunna tillämpas
i industriella processer.
I. “Facile Preparation Route for Nanostructured Composites: Surface‐
Initiated Ring‐Opening Polymerization of ε‐Caprolactone from High‐
Surface‐Area Nanopaper”, A. Boujemaoui, L. Carlsson, E. Malmström, M. Lahcini, L. Berglund, H. Sehaqui, and A. Carlmark, ACS Applied Materials and Interfaces 2012, 4, 3191‐3198
II. “Surface‐initiated ring‐opening polymerization from cellulose model surfaces monitored by a Quartz Crystal Microbalance”, L. Carlsson, S.
Utsel, L. Wågberg, E. Malmström,
and A. Carlmark, Soft Matter 2012, 8, 512‐517
III. “Nanobiocomposites Adhesion: Role of Graft Length and Temperature in a Hybrid Biomimetic Approach”, N. Nordgren, L. Carlsson, H.
Blomberg, A. Carlmark, E. Malmström, and M. W. Rutland, Biomacromolecules 2013, 14, 1003‐1009
IV. “Solid State NMR investigation of hydrolyzed cotton linters grafted by surface‐initiated ring‐opening polymerization of ‐caprolactone”, L.
Carlsson, P. T. Larsson, T. Ingverud, H. Blomberg, A. Carlmark, and E.
Malmström, Manuscript
V. “Surface‐initiated ring‐opening metathesis polymerisation from cellulose fibres”, L. Carlsson, E. Malmström,
and A. Carlmark, Polymer Chemistry 2012, 3, 727‐733
VI. “Modification of cellulose surfaces by cationic latex prepared by RAFT‐mediated surfactant‐free emulsion polymerization” L. Carlsson, A. Fall, I. Chaduc, L. Wågberg, B. Charleux, E. Malmström, F.
D’Agosto, M. Lansalot, and A. Carlmark, Manuscript
I. Part of the experimental work, analyses, and minor part of the preparation of the manuscript.
II. A majority of the experimental work, analyses, and most of the preparation of the manuscript.
III. Part of the experimental work, analyses, and minor part of the preparation of the manuscript.
IV. Part of the experimental work, part of the analyses, and most of the preparation of the manuscript.
V. All the experimental work and analyses, and most of the preparation of the manuscript.
VI. All the experimental work, a majority of the analyses, and most of the preparation of the manuscript.
Scientific contributions not included in this thesis:
VII. ʺAligned Cellulose Nanocrystals and Directed Nanoscale Deposition of Colloidal Spheresʺ, G. Nyström, A. Fall, L. Carlsson and L. Wågberg, submitted
VIII. ”Nano‐hybrid self‐crosslinked PDMA/silica hydrogels”, L.
Carlsson, S. Rose, D. Hourdet, A. Marcellan, Soft Matter 2010, 6,
3619‐3631
AGU anhydroglucose unit AFM atomic force microscopy
AIBA 2,2′‐azobis(2‐methylpropionamidine) dihydrochloride
ATR attenuated total reflectance BCN bacterial cellulose nanofibers BET Brunauer‐Emmett‐Teller
CA contact angle
CCL cellulose cotton linters
‐CL ‐caprolactone
CMC critical micelle concentration CMS cellulose microsphere CNC cellulose nanocrystals
CP/MAS cross‐polarized magic angle spinning CNF cellulose nanofibers
CTPPA 4‐cyano‐4‐thio‐thiopropyl‐sulfanylthiocarbonyl d
Llateral dimension
Ð
Mmolar‐mass dispersity
DCM dichloromethane
DLS dynamic light scattering
DMAEMA N,N‐dimethyl aminoethylmethacrylate DMAP 4‐(dimethylamino)pyridine
DMF N,N‐dimethylformamide
DMSO dimethyl sulfoxide DP degree of polymerization DS degree of substitution
DS
BMbulk monomer degree of substitution DS
SPparticle‐surface degree of substitution DSC differential scanning calorimetry DVS dynamic vapor sorption
FE‐SEM field‐emission scanning electron microscopy FRP free radical polymerization
FT‐IR Fourier transform‐infrared
HCCL hydrolyzed cellulose cotton linters
MAA methacrylic acid
NMMO N‐methylmorpholine N‐oxide NMR nuclear magnetic resonance PdI polydispersity (DLS) PEG poly(ethylene glycol) PET polyelectrolyte titration
PISA polymerization‐induced self‐assembly
PS polystyrene
PVAm poly(vinyl amine)
QCM quartz crystal microbalance
RAFT reversible addition‐fragmentation chain transfer RDRP reversible‐deactivation radical polymerization RH relative humidity
ROMP ring‐opening metathesis polymerization ROP ring‐opening polymerization
R
qsurface roughness
RT room temperature
SEC size exclusion chromatography SI surface‐initiated
Sn(Oct)
2tin 2‐ethylhexanoate SSA specific surface area
TBD 1,5,7‐triazabicyclo[4.4.0]dec‐5‐ene T
ccrystallization temperature TEM transmission electron microscopy TGA thermogravimetric analysis T
gglass transition temperature Ti(OiPr)
4titanium i‐propoxide Ti(OnBu)
4titanium n‐butoxide
UV ultraviolet
X
cdegree of crystallinity
TABLE OF CONTENTS
1. PURPOSE OF THE STUDY 1
2. INTRODUCTION 2
2.1 C
ELLULOSE... 2
2.1.1 Structure of cellulose ... 2
2.1.2 Cellulosic substrates ... 4
2.1.3 Surface modification of cellulose ... 6
2.2
R
ING-
OPENING POLYMERIZATION TECHNIQUES... 9
2.2.1 Ring-opening polymerization (ROP) ... 9
2.2.2 Ring-opening metathesis polymerization (ROMP) ... 12
2.3 R
EVERSIBLE ADDITION-
FRAGMENTATION CHAIN-
TRANSFER(RAFT)
POLYMERIZATION... 14
2.3.1 RAFT-mediated surfactant-free emulsion polymerization ... 17
2.3.1.1 Polymer-induced self-assembly ... 19
2.4 S
URFACE MODIFICATION OF CELLULOSE BY COVALENT GRAFTING... 21
2.5 S
URFACE MODIFICATION OF CELLULOSE BY PHYSICAL ADSORPTION... 23
3. EXPERIMENTAL 26 3.1 M
ATERIALS... 26
3.2 E
XPERIMENTAL PROCEDURES... 27
3.2.1 Cellulose surface modification by ring-opening polymerization ... 27
3.2.2 Cellulose surface modification by ring-opening metathesis polymerization ... 28
3.2.3 Synthesis of cationic latex nanoparticles ... 29
3.2.3.1 Aqueous RAFT polymerization of DMAEMA ... 29
3.2.3.2 RAFT-mediated surfactant-free emulsion polymerization of MMA . 30 3.2.3.3 Adsorption of cationic latex on cellulose model surfaces ... 31
3.3 C
HARACTERIZATION METHODS... 31
3.3.1 QCM-D ... 31
3.3.2 Colloidal probe AFM ... 32
3.3.3 Solid state CP/MAS
13C-NMR ... 32
4. RESULTS AND DISCUSSION 33 4.1 C
ELLULOSE SURFACE MODIFICATION BY COVALENT GRAFTING... 33
4.1.1 Cellulose surface modification by ring-opening polymerization ... 33
4.1.1.1 SI-ROP of -CL from FP and NP. Comparison between Sn(Oct)
2and
Ti(OnBu)
4... 33
4.1.1.3 PCL-grafted cellulose microspheres – Impact of graft length and temperature on interfacial adhesion – a study by colloidal probe by AFM ... 41 4.1.1.4 Solid state CP/MAS
13C-NMR investigation of HCCL grafted by SI-
ROP of -CL ... 45 4.1.2 Cellulose surface modification by ring-opening metathesis
polymerization (ROMP) ... 50 4.2
MODIFICATION OFC
ELLULOSE SURFACES BY PHYSICAL ADSORPTION... 53 4.2.1 Synthesis of cationic latex nanoparticles ... 53 4.2.1.1 RAFT-mediated surfactant-free emulsion polymerization of MMA . 54 4.2.2 Adsorption of cationic latex on cellulose model surfaces ... 57
5. CONCLUSIONS 61
6. FUTURE WORK 63
7. ACKNOWLEDGEMENTS 65
8. REFERENCES 67
1. PURPOSE OF THE STUDY
The interest in designing and developing new, more environmentally friendly materials from cellulose has increased immensely in the last years. Cellulose has several interesting properties, e.g., high stiffness and low density compared to commonly employed glass fibers, but cellulose is also highly hygroscopic, as a result of the large number of hydroxyl groups. These can in turn be readily utilized for surface modification; hence, the cellulose properties can be tailored and new functionalities introduced, improving the compatibility with non‐polar polymer matrices for example.
The purpose of this study was to expand the toolbox for surface modification of cellulose and thereby possibly increase the utilization of cellulose‐based materials. Different techniques were employed to achieve a better fundamental knowledge of surface modification by polymers and its effect on the surface properties. Surface modification has been performed by covalent grafting and physisorption of polymers.
The polymers have been synthesized by controlled polymerization
techniques.
2. INTRODUCTION
2.1 CELLULOSE
Cellulose is one of our most abundant polymers on earth. The polymer is renewable, biocompatible, biodegradable, and inexpensive, and has both interesting physical and chemical properties.
1, 2The cellulose fibers have both low density and high strength. The modulus of the cellulose crystal is 138 GPa, which can be compared to steel with a modulus of 200 GPa.
3The production of cellulose is mainly performed in plants, e.g., flax, hemp and jute.
4However, there are also several other sources of cellulose, e.g., wood, which has a dry content of 40 % cellulose.
5In addition, wood contains significant amounts of lignin and different types of hemicelluloses.
5Therefore, the cellulose needs to be isolated by separation from the other components prior to utilization. Bacteria, algae, and fungi can synthesize cellulose with high purity, crystallinity, and very specific morphologies can be obtained.
6The total annual production of cellulose is estimated to be over 7.5×10
10tons.
12.1.1 Structure of cellulose
Cellulose is a linear, polydisperse biopolymer; exhibiting an advanced hierarchical structure that can be attributed to the hydrogen bonds between the hydroxyl groups. The degree of polymerization (DP) varies between 300–1,700 for wood fibers and 800–10,000 for cotton and other plants, depending on origin and treatment of the cellulose raw material.
4The cellulose fibers are located in the cell wall of plants and are built‐up
of aggregates of microfibrils, see Figure 1. The bundle of microfibrils
contains 30–40 cellulose chains with different orientations and is formed
by extended cellulose macromolecules that are organized in sheets –
with both crystalline and amorphous regions – which are stabilized by
intra‐ and intermolecular hydrogen bonding.
4, 7Figure 1. Hierarchical structure of cellulose; from the tree on a macroscopic scale down to the cellulose macromolecule on nanoscale with length (L) and lateral dimension (d
L).
4, 7The schematic picture is adopted from Isogai et al.
7 andHansson
8.
The biopolymer consists of β‐
D‐glucosyl units that are linked together by
‐1,4‐glycosidic bonds. Each repeating unit in cellulose contains two anhydroglucose units (AGU). The AGU, see Figure 2, has three hydroxyl groups with a primary hydroxyl group at the carbon 6 (C6) position and two secondary hydroxyl groups at the C2 and C3 positions, respectively. Every second repeating unit is rotated 180° in the plane.
Cellulose is highly hydrophilic owing to the numerous hydroxyl groups; however, the strong inter‐ and intramolecular bonding renders cellulose insoluble in water. The exceptional mechanical properties and the lack of a melting point can also be ascribed to these hydrogen bonds.
9Figure 2. Numbering of the carbon atoms in a repeating unit of cellulose (two AGU) of the cellulose chain.
O 1
4 5
O HO
OH 6
2 O 3
OH
O 1 4 5
HO
OH
6 2 O 3
OH
n
2.1.2 Cellulosic substrates
Cellulose can be divided into groups depending on its crystal structure.
The naturally occurring cellulose is denoted cellulose I where the polymer chains are organized in a parallel arrangement. However, by treatment of cellulose I, the most thermally stable structure, cellulose II, is obtained by change of the organization of the macromolecular chains to an anti‐parallel arrangement. The transformation can be performed by conc. alkali aqueous solution or by regeneration from solutions or semi‐stable derivatives. Regenerated cellulose has a DP of 250–500. The viscous process was developed over 100 years ago to regenerate cellulosic fibers.
4This technique has been employed to produce fibers, films, membranes, sponges, and spheres. However, the method suffers from major disadvantages since it requires hazardous chemicals and only highly pure pulps can be utilized. An important breakthrough for the regenerated fibers was the Lyocell process where the cellulose fibers successfully were dissolved in N‐methylmorpholine N‐oxide (NMMO) and thereafter spun, forming regenerated fibers.
10In this process, almost all employed solvents can be recycled and the environmental impact is low. By employing the Lyocell process, Wågberg et al.
11, 12have developed cellulose model surfaces. These surfaces are smooth, have good stability in different solvents and the thickness of the films can be tuned, rendering the surfaces suitable for high‐resolution techniques, such as atomic force microscopy (AFM) and quartz crystal microbalance with dissipation (QCM‐D).
13The nanocelluloses are produced by disintegration of the wood fibers, smaller fibrils can be liberated which increases the available surface area.
14For a summary of the characteristics for the most common nanocelluloses, see Table 1. Micrographs can be observed in Figure 3.
Figure 3. TEM images of a) CNF,
15b) CNC,
16and c) SEM image of BNC.
14Reprinted with permission from (Klemm, D. et al., Angew. Chem., Int. Ed. 2011 50
p. 5438‐5466). Copyright (2011) German Chemical Society.
Table 1. CC
CNF and the other nanocelluloses are already commercialized in several products, e.g., absorbent material in hygiene products, food additives, cosmetics, and pharmaceuticals.
17CNF have recently been employed for producing nanofoams,
18, 19hydrogels,
20nanopapers,
21, 22and aerogels
23. Cellulose cotton linters (CCL) are short (2–6 mm), slightly curled, cylindrically shaped fibers with widths of 17–27 μm.
24Furthermore, CCL possess higher crystallinity and smaller lateral dimensions of the microfibrils compared to cellulose from wood fibers.
4High‐quality CCL are utilized to produce Whatman #1 filter paper which has a high cellulose content (> 98 %) and a crystallinity of 68 %.
252.1.3 Surface modification of cellulose
The first patent that reported surface modification of cellulose is dated to the 1870’s. The fibers were reacted with nitrosulfuric acid (a mixture of nitric and sulfuric acid) to form cellulose nitrate, and by adding a plasticizer, the first cellulose‐based material, celluloid, was produced.
4In recent years, the interest and utilization of cellulose fibers have increased tremendously with the aim to create new and sustainable materials developed from bio‐based resources.
4, 6, 14, 26, 27It is possible to obtain renewable composites with lower density and cost comparing to other fillers or if reinforcements, such as glass or carbon fibers, are employed. Moreover, a cellulose‐based composite can be recyclable.
Nanocelluloses with high crystallinity are of particular interest owing to their high stiffness that can reinforce polymer matrices. The extent of amorphous domains is depending on the pretreatment of the fibers. The amorphous parts enhance the flexibility.
9Cellulose can be applicable in many products; recent publications
discuss a wide range of exciting areas, e.g., inexpensive electronics,
28‐30paper‐based medical diagnostics,
31‐33functional clothes,
34, 35and
membranes
36, 37. However, it is a challenge to incorporate and disperse
cellulose due to the incompatibility of the polar cellulose with non‐polar
polymer matrices. Good compatibility is essential to draw benefits from
the fibers as reinforcement. Furthermore, the hydrophilic character will
most probably lead to moisture adsorption and swelling of the
composite. However, the available hydroxyl groups on cellulose can be
employed for modifications as they can act as chemical handles. The
number of available hydroxyl groups can be tuned by different
pretreatment methods. For example, with mercerization, i.e., treatment with a strong base followed by neutralization, the amount of hydroxyl groups is increased by breaking the hydrogen bonds, resulting in swelling of the cellulose structure and, thus, increased surface area and strength is achieved. Still, it should be taken into consideration that the mechanical properties of cellulose may be affected upon breakage of intermolecular hydrogen bonds.
4Surface modification of cellulose can be performed by physical treatments such as solvent exchange, physico‐chemical modifications by corona or plasma discharges, laser, UV‐ or γ‐irradiation, or physical adsorption. Chemical, i.e., covalent, modification is performed by attachment of small molecules or polymers.
26, 27, 38‐44The different types of modifications can also be combined.
Traditionally, surface modification of cellulose by small molecules has been the most utilized technique to introduce various functionalities on the surface. Anhydrides, isocyanates, organometallics, sulfates, and acid chlorides are examples of molecules that successfully have been employed to obtain functional groups on the fiber surface.
6, 26, 27, 38, 45‐47Cellulose esters and ethers are the most common examples of cellulose derivatives. The water‐ or organo‐soluble derivatives are obtained by substitution of the hydroxyl groups. Cellulose acetate, the most common ester, is produced by reaction with acetic anhydride in the presence of sulfuric acid as catalyst. Cellulose esters have thermoplastic properties which are influenced by the carbon number of the acyl residues. Water‐
soluble cellulose ethers are formed by alkaline treatment and subsequent substitution of an alkyl halide or addition to an oxirane.
4Free‐radical polymerization (FRP) of vinyl monomers have successfully been performed from the cellulose backbone. The initiating radicals can be formed by hydrogen abstraction (chain transfer), by employing redox systems or by utilizing substituents that can form radicals or polymerize.
48, 49However, due to poor control it is impossible to tailor the molecular structure or the molar mass by FRP. Other drawbacks can be the possible occurrence of chain scission of the cellulose backbone, which may influence the strength of the final material
50, as well as unattached polymer formed in the bulk
51.
Modification of cellulose by grafting polymer chains from its surface
enables alteration of both the physical and chemical properties and
could potentially introduce other functionalities.
41‐44, 52‐54The two most commonly employed grafting techniques are ‘grafting‐from’ and
‘grafting‐to’, see Figure 4. In the ‘grafting‐from’ method the polymer is formed by propagation of monomer from initiating species (reactive centers) on the surface, i.e., surface‐initiated (SI) polymerization. In order to fully characterize the properties of the grafted polymer chains, cleavage of the chains and subsequent isolation is necessary. This can be performed by acid hydrolysis, which decomposes cellulose, leaving the polymer grafts intact. Nevertheless, this method cannot be applied for polymers containing hydrolytically sensitive groups.
44, 55To characterize the grafted polymer without cleaving it from the surface is a great challenge; therefore, the ‘grafting‐to’ technique can be preferable.
Figure 4. Schematic illustration of the ‘grafting‐from’ and the ‘grafting‐to’
technique.
In the ‘grafting‐to’ technique, pre‐formed polymers, with an active chain‐end that can attach covalently with reactive centers present on the cellulose surface. Recently, polymers with two active chain‐ends have been reported, which enables further post modification of the grafted preformed polymer.
1, 6An advantage by employing the ‘grafting‐to’
technique is the possibility of characterizing the pre‐formed polymers,
molar mass and the molar‐mass dispersity (Ð
M) prior to the surface
attachment. However, previous studies have shown that covalent
surface modification by ‘grafting‐from’ enables higher grafting densities due to less sterical hindrance of monomers compared to when large and bulky pre‐polymers are coupled to the surface.
1, 42, 52, 53Nevertheless, there are some reports in the literature where similar grafting densities have been obtained by the ‘grafting‐to’ approach.
56‐582.2 RING‐OPENING POLYMERIZATION TECHNIQUES
Ring‐opening polymerization (ROP) and ring‐opening metathesis polymerization (ROMP) are two techniques that are suitable for SI polymerizations.
59These techniques are widely utilized for polymerization of cyclic monomers. The polymerization mechanism involves three reaction steps: initiation, propagation, and termination.
The employed initiator varies depending on the technique applied and the ring‐opening of the cyclic monomer occurs during the propagation step in presence of a catalyst. Undesired termination may occur by inter‐
or intramolecular chain transfers. Intermolecular chain transfers occur between different chains whereas intramolecular chain transfers (back‐
biting) occurs within the polymer chain, forming cyclic oligomers. The chain transfer reactions will give rise to deviation from the targeted degree of polymerization (DP) and increase the Ð
M.
2.2.1 Ring‐opening polymerization (ROP)
ROP, developed in the 1930’s by Carothers et al.
60, providing macromolecules with well‐defined end‐groups and high molar mass.
Since then, numerous cyclic monomers such as lactones, lactides, cyclic carbonates, siloxanes, and ethers have been polymerized employing this technique.
61The ease of polymerization of the cyclic monomers can be attributed to both kinetic and thermodynamic factors. The presence of a heteroatom (oxygen, nitrogen, sulfur, etc.) in the ring facilitates the ring‐
opening via a nucleophilic or electrophilic attack of the initiator.
Moreover, the reactivity of the monomers can be attributed to the ring size, i.e., thermodynamic stability decreases reactivity due to unwillingness to change state.
62Poly(‐caprolactone) (PCL) is an aliphatic polyester, and its physical,
mechanical and thermal properties are dependent on the molar mass.
63PCL possesses advantages such as good miscibility with other polymers,
biodegradability and biocompatibility. The utilized catalyst system
determines whether an anionic, a cationic, monomer‐activated or coordination‐insertion mechanism occurs.
63Several different catalytic systems have been explored; metal‐based, organic or enzymatic.
63‐66Three examples of catalysts are: the metal‐based stannous 2‐
ethylhexanoate (Sn(Oct)
2), titanium n‐butoxide (Ti(OnBu)
4), and the organic 1,5,7‐triazabicyclo[4.4.0]dec‐5‐ene (TBD).
Scheme 1. ROP of ‐CL by coordination‐insertion mechanism employing Sn(Oct)
2as catalyst and an alcohol initiator (R‐OH).
The most commonly utilized catalyst is Sn(Oct)
2, owing to its excellent performance, good thermostability, low cost, reasonably low toxicity, and approval in food and drug applications (FDA approved).
67The proposed mechanism for Sn(Oct)
2is coordination‐insertion reaction, as presented in Scheme 1. The initiation is divided into two steps ‐ formation of a metal oxide alkoxide (Oct‐Sn‐O‐R) from the catalyst (Sn(Oct)
2) and the initiator (R‐OH) which subsequently initiates the polymerization of monomer by the coordination‐insertion mechanism through coordination of the monomer to the catalyst and insertion of the monomer into a metal‐oxygen bond of the catalyst. The metal alkoxide remains as the active center throughout all the polymerization.
However, there are some drawbacks of this catalyst system ‐ intra‐ and intermolecular transesterification reactions may occur which will lead to broader molar‐mass dispersities and deviations from the targeted molar mass. The extent of the undesired termination reactions is dependent on the temperature and the monomer conversion.
68Another drawback is the difficulty in removing the catalyst, and the presence of tin in biomedical or food applications is highly undesirable; therefore, other catalyst systems are of prime importance. Hence, in recent years, a lot of
HO Sn O O R-O
O O
H R-O
O- O
Sn O O
R-O
O O
O
O Sn OR
O O
O
O Sn OR
O O
O
HO
O
O Sn OR OH
R O
O Sn
O
O tin 2-ethylhexanoate
(Sn(Oct)2)
+ +
Coordination Insertion
1. n e-CL 2. H
n +
+
poly(caprolactone) PCL ε-CL
Preinitiation
Initiation
Propagation
Oct-Sn-O-R Oct-H
Oct-Sn-OR Oct-Sn-OR
Sn O O
research has been conducted to find new, more environmentally friendly and less toxic catalysts for ROP.
63, 65, 66, 69, 70This has resulted in development of both new metal‐based
71and organic catalysts
65for ROP.
In 2010, Pärssinen et al.
72utilized titanium i‐propoxide (Ti(OiPr)
4) and titanium n‐butoxide (Ti(OnBu)
4), for ROP in air atmosphere at elevated temperatures, 70–140 °C. The chemical structure of Ti(OnBu)
4is presented in Figure 5. An advantage with those catalysts is the non‐toxic degradation products, titanium oxide and alcohol, which are approved for internal use in humans.
72They report open‐air systems requiring higher temperatures (140 °C) for efficient initiation attributed to decreasing water content. However, after initiation the reaction temperature could be decreased slightly over the melting temperature for PCL (~60 °C).
73Noteworthy, Ð
Mwas quite high and the values varied from 1.5–2, however, by
1H‐NMR it was confirmed that the PCL was only formed from one of the arms in the catalyst.
ROP has also been performed by employment of enzymatic catalysts which enables mild reaction conditions.
66Several different catalyst systems, employing amino acids and small organic molecules for ROP of cyclic lactones, have also been reported as successful.
63, 70, 74, 75Furthermore, Hedrick et al.
76have developed efficient organo‐catalysts that enable polymerization at room temperature with good accordance between theoretical and experimental molar mass, and narrow molar‐
mass dispersities. An example is 1,5,7‐triazabicyclo[4.4.0]dec‐5‐ene (TBD), which polymerizes by the dual activation mechanism, including both monomer and initiator in a transition state with the catalyst.
76Figure 5. Different catalysts employed for ROP.
SI‐ROP has been widely utilized for surface modification of different substrates, e.g., silica and gold surfaces by self‐assembled monolayers,
77clay minerals,
78nanoparticles of silica and cadmium sulfide,
79magnetite,
80and a wide range of different cellulosic substrates,
43and more recently, in ionic liquids
75.
N N
NH O
O Sn
O
O
Ti O
O O n-Bu
O n-Bu
n-Bu n-Bu
tin 2-ethylhexanoate (Sn(Oct)2)
1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD)
titanium n-butoxide (Ti(On-Bu)4)
2.2.2 Ring‐opening metathesis polymerization (ROMP)
ROMP has had a large impact in both organic and polymer chemistry for olefin metathesis over the last decade.
81‐83The polymerization technique is a chain growth process where mono‐, bi‐ or polycyclic rings are opened and converted into a polymeric material.
81In 1957, Eleuterio et al.
84at DuPont were the first to patent this technique. In 1960, Truett et al.
85reported the first polymerization of norbornene (bicyclo[2.2.1]heptane) (NBE) by ROMP. The ring‐opening mechanism of norbornene was confirmed by Dall’Asta et al.
88, 89in the beginning of the 1970s. The monomer and polymer structure can be seen in Figure 6.
ROMP is a highly interesting polymerization technique since high molar mass polymers with narrow molar‐mass dispersities can be obtained in short reaction times and under mild reaction conditions. Moreover, preserved double bonds in the polymer backbone enable facile post modification.
Figure 6. Molecular structure of norbornene (NBE), and the repeating unit of the corresponding polymer poly(norbornene) (PNBE). To polymerize NBE, Grubbs 1
stgeneration catalyst is often utilized.
Initially, ROMP required stringent reaction conditions including the use of highly pure reagents and the total exclusion of water, oxygen and other functional groups. However, owing to the Noble Prize laureates in Chemistry in 2005 – Schrock
86and Grubbs
87– and their development of less sensitive catalysts for ROMP, the interest for the polymerization technique has increased remarkably due to the increased stability of the catalysts and, thus, a wider range of polymerizable monomers with functional groups etc.
81, 88Grubbs developed ruthenium alkyl carbene
n
Ru P
Cl P
Cl Ph
norbornene
(NBE) polynorbornene
(PNBE) Grubbs' 1st generation
catalyst
that are efficient initiators with high activity for polymerization of NBE and substituted NBEs.
89In addition to organic systems, these catalysts can also be employed in aqueous systems for metathesis.
90The ROMP mechanism of NBE is based on olefin metathesis where the ring‐opening occurs at the most stable site of the monomer, i.e., the double bond, see Scheme 2.
Scheme 2. Schematic overview of the ROMP process.
The critical step upon propagation is the formation of the intermediate metallocyclobutane by coordination of the NBE by the catalytic metal center of the carbene complex, followed by [2+2] cyclo‐addition where the propagating species is formed. The metal atom separates from the olefin bond by a cyclo‐reversion.
82Commonly, trans‐vinylene units are formed during ROMP of NBE when Grubbs catalysts are employed.
Undesired secondary metathesis reactions lead to termination and occur by intra‐ or intermolecular chain transfer, yielding polymers with broader molar‐mass dispersities and ill‐defined end groups. The extent of the termination reactions is dependent on many parameters, e.g., bulkiness of the monomer, temperature, monomer concentration, and solvent.
83The pioneer work of SI‐ROMP from a gold surface by the ruthenium‐
based Grubbs’ 1
stgeneration catalyst was performed by Grubbs et al.
91in 1999. Thereafter, numerous substrates have been surface‐modified by SI‐ROMP, e.g., silicon wafers,
92quartz,
93Wang resins,
94silica nanoparticles,
95carbon nanotubes,
96gold,
97clay,
98and cellulose
99.
LnM R
LnM R
LnM
LnM R
LnM R
LnM R
m
LnM R
m+2
LnM
R m+2
X=Y MLn=X Y R
m+2 coordination
metal alkyldiene
metallacyclobutane
Termination:
Propagation:
Initiation:
[2+2]
2.3 REVERSIBLE ADDITION‐FRAGMENTATION CHAIN‐
TRANSFER (RAFT) POLYMERIZATION
During the last years, much work has been conducted in the field of polymer synthesis with the aim to design well‐defined macromolecular structures with narrow Ð
Mand end‐group functionalities.
100Tailored polymers can be synthesized by reversible‐deactivation radical polymerization (RDRP) techniques. The most commonly employed RDRP techniques includes nitroxide mediated polymerization (NMP),
101atom transfer radical polymerization (ATRP),
100, 102and reversible addition‐fragmentation chain transfer (RAFT)
103, 104. Among those techniques, RAFT is the most recently developed and probably the most versatile.
105, 106The RAFT technique has versatility for numerous different monomers and polymerization can be conducted in a wide range of reaction media, organic solution, aqueous solution, or in dispersed phase.
107, 108Therefore, this technique has been extensively employed for controlled polymerization of both, hydrophilic and hydrophobic monomers.
104, 106, 108, 109Moreover, a good control of the macromolecular structure was achieved and, as the RAFT chain‐end was preserved to a large extent, facile post‐functionalization could be performed of the polymers.
100, 106, 107In a review on RAFT, Moad et al.
104, more than 700 publications from the middle of 2009 until the beginning of 2012 was cited, which demonstrate the intense ongoing research in this field.
RAFT polymerization includes the same steps as classical free radical polymerization (FRP): initiation, propagation and termination.
However, in a RAFT system the termination reactions are suppressed by fast chain‐transfer. The polymerization starts from radical initiators that are activated in a classical manner, either by heating, redox, photochemistry, UV‐ or gamma irradiation. The addition of an external source of radicals is required to initiate polymerization and maintain the rate of polymerization.
105However, RAFT polymerization is a more sophisticated technique compared to FRP and the control is provided by degenerative (i.e., thermodynamically neutral) chain transfer, involving transfer of an atom or a group from a covalently bond and dormant species (i.e., the chain transfer agent, RAFT‐agent) to the active chain end. The RAFT agent, see Figure 7, is composed of an activating group
“Z” and a good free radical leaving group, “R”. The RAFT agent
provides control of the polymerization and should be present in large excess to the initiator
103, 104, 110to suppress irreversible termination of the active, propagating chains.
102, 103, 109Figure 7. An example of a RAFT agent, the thiocarbonylthio chain transfer
agent.
105The mechanism of RAFT polymerization employing the thiocarbonylthio RAFT‐agent is presented in Scheme 3. In the initial step, a free radical, I
·, is formed which propagates with monomer units, M. This is followed by the pre‐equilibrium step which occurs as soon as the RAFT‐agent reacts with the radical in the propagating oligomer, (P
n·). The radical position is then transferred to the RAFT‐agent, forming a relatively stable (dormant) specie. Further, the radical is either trans‐
ferred back to the propagating oligomer, which enables propagation, or to the R‐group. In this case, the R‐group will act as an initiator, and start adding monomer units (P
m), until it goes back to the dormant side by reacting with the RAFT‐agent. This process is fast which is vital for the control and, hence, the main equilibrium should be reached early in the process. The rate constants are strongly dependent on the Z‐group. The thiocarbonylthio groups are highly efficient RAFT agents, which is of high importance to obtain a good control of the polymerization.
Z
S S R
Scheme 3. Mechanism of RAFT polymerization by employing the thiocarbonylthio RAFT agent.
The versatility of the RAFT technique has permitted a wide range of monomers for polymerization, for instance, acrylates, methacrylates, and vinyl acetate.
104, 111, 112However, the environmental consciousness has increased and in this context, water is the preferred reaction medium not only due to economic reasons but also for environmental and health aspects.
106Hence, polymerization of water‐soluble monomers, such as N,N‐dimethylaminoethyl methacrylate (DMAEMA), has gained significant attention. The structure of the monomer, DMAEMA, and corresponding polymer PDMAEMA are shown in Figure 8.
Figure 8. Molecular structure of N,N‐dimethyl aminoethylmethacrylate (DMAEMA) and the repeating unit of the corresponding polymer, P(DMAEMA).
Initiation
Initiator I M M
Pn Initialization/pre-equilibrium
Pn M kp
Z S S
R kadd k-add
S S
Z R Pn
kb k-b
S S
Z R
Pn +
Reninitiation
R M
Pm Main equilibrium
Pm M kp
Z S S
Pn kaddP k-addP
S S
Z Pn Pm
kaddP k-addP
S S
Z Pn Pm
M kp Termination
Pn+Pm kt
Dead polymer
+ +
+
kd ki
M kp kp P1
kiR
M M
kp kp P1
O
N O
DMAEMA
O
N O
PDMAEMA
n
The polymer is both thermo‐ and pH‐responsive. Stimuli‐responsive polymers can change conformation, properties and interactions in response to external stimuli, e.g., switching from hydrophilic to hydrophobic.
113The molar mass of the polymer influence the responsive properties.
114Hence, they are interesting for a wide range of applications.
115PDMAEMA has been employed in biological applications, for antibacterial activity,
116, 117gene delivery,
118as well as other fields, e.g., waste‐water treatment,
119and paints
120. It is the polar tertiary amine side‐chain group that offers water solubility and, if desired, can be charged either when below its pK
avalue (close to 7.5, slightly chain length dependent)
114or irreversibly charged by quaternization. However, there are few earlier investigations available where DMAEMA has been polymerized in aqueous media by ATRP
121,122
or by RAFT
123.
Surface‐initiated RAFT (SI‐RAFT) has been performed from numerous surfaces, for instance, gold nanoparticles,
124silica nanoparticles,
125silicon wafers,
126proteins,
127and various cellulose surfaces
52, 55, 128, 129.
2.3.1 RAFT‐mediated surfactant‐free emulsion polymerization
RAFT polymerization, enables development and design of new, advanced and tailored macromolecular architectures, both for homo‐
and block copolymers.
100Amphiphilic block copolymers, composed of both a hydrophilic and a hydrophobic block,
130‐132are highly interesting in applications such as compatibilizers. However, the synthesis of amphiphilic block copolymers is demanding due to the difference in polarity between the blocks which typically causes at least two (or more) steps, including thorough purification. The synthesis of block copolymers can be performed by various methods; two examples can be seen in Figure 9.
Figure 9. Synthesis of block copolymers in a two‐step process by a) by coupling two pre‐formed macromolecular chains together or b) first forming a block with living chain‐ends that can be employed to initiate the growth of the second block.
Previously, the RDRP techniques have primarily been performed in organic solvents or bulk. However, the interest in the field of aqueous dispersions has increased tremendously.
133, 134All the RDRP techniques have successfully been employed in dispersion.
135RAFT has thus far been the most successful technique for water‐borne systems.
134, 136By employing dispersed systems, nanoparticles (d = 20 nm–10 μm) with different morphologies (core‐shell particles, hollow particles or complex multilayer structures) are formed which may be interesting for various applications, see Figure 10.
133Figure 10. P(MAA‐co‐PEOMA)‐b‐polystyrene amphiphilic block copolymer self‐
assemblies.
137Reprinted with permission from (Zhang, W. et al., Macromolecules 2012, 45, p. 4075‐4084). Copyright (2012) American Chemical Society.
Classical emulsion polymerization has been employed to a large extent
in industry since volatile organic solvents can be avoided, a wide range
of monomers can be utilized and this technique enables processing at
high solid content without any influence on the viscosity.
134The
technique requires employment of an organo‐soluble monomer, a
water‐soluble initiator and a surfactant. The three regimes of emulsion
polymerization are demonstrated in Scheme 4.
Interval I: Initially, an emulsion polymerization system consists of a dispersion of monomers in an aqueous phase. The initiation always takes place in the aqueous phase. Thereafter, two different mechanisms may occur, micellar nucleation or homogeneous nucleation, depending of the concentration of the surfactant. The monomer is entrapped in the micelles if the concentration of surfactant is above the critical micelle concentration (CMC). During the nucleation, particles are formed wherein the polymerization will proceed.
Interval II: During the second interval, the polymerization is proceeding in the particles. The growing particles are constantly fed with monomer by diffusion from the droplets through the aqueous phase into the particles. The particles are increasing in size during the polymerization.
However, the number of particles and the polymerization rate are constant throughout the process. When the monomer droplets are consumed, the third interval starts.
Interval III: The residual monomers in the particle and in the aqueous phase are consumed. At the end of interval III, the polymerization is terminated.
Scheme 4. Schematic illustration of intervals in an emulsion polymerization, adopted from Zhang.
1382.3.1.1 Polymer‐induced self‐assembly
More recently, the development of RAFT‐mediated emulsion
polymerization has led to the aqueous synthesis of amphiphilic block
copolymers, which self‐assemble to form polymer particles according to
the polymerization‐induced self‐assembly (PISA) strategy as originally
described by Hawkett et al.
139‐141The PISA process has successfully been
employed for all three RDRP techniques.
134, 140In the RAFT‐mediated
PISA process, it is crucial to choose a suitable RAFT agent that provides good control
over formation of both blocks when synthesizing block copolymers.
107Recently, new RAFT agents have been developed that can be employed for polymerization of both hydrophilic and hydrophobic monomers, where the first formed hydrophilic polymer is utilized as a macroRAFT for formation of the second hydrophobic block.
133The macroRAFT is acting as an electrosteric surfactant as it both functionalizes the surface of the particles and controls the particle growth, see Scheme 5. No additional low molar‐mass surfactant is required, which is highly advantageous since the low molar mass surfactants have an undesired impact on the latex stability under freezing conditions and diffusion upon film formation.
142Scheme 5. An overview of the PISA process. Adapted from Chaduc.
143In Figure 11 , the structure of a thiocarbonylthio RAFT agent, 4‐cyano‐4‐
thiothiopropylsulfanylthiocarbonyl (CTPPA)
142is shown, which is obtained by reacting the radical initiator, 4,4′‐Azobis(4‐cyanovaleric acid) with bis(propylsulfanylthiocarbonyl) disulfide according to the literature.
144, 145CTPPA permits polymerization in water despites its insolubility in water, as it is soluble in many hydrophilic monomers.
142,146‐151